Psychrophilic and psychrotrophic clostridia: sporulation and ...

International Journal of Food Science and Technology 2010, 45, 1539–1544 1539 

Review article 

Psychrophilic and psychrotrophic clostridia: sporulation and 

germination processes and their role in the spoilage of chilled, 

vacuum-packaged beef, lamb and venison 

Katharine H. Adam, 1 * Steve H. Flint 2 & Gale Brightwell 1 

1 Food Metabolism and Microbiology, AgResearch, Ruakura, Private Bag 3123, Hamilton, New Zealand 

2 Food, Nutrition & Human Health, Massey University, Private Bag 11222, Palmerston North, New Zealand 

(Received 27 January 2010; Accepted in revised form 20 May 2010) 

Summary Spoilage of beef, lamb and venison by psychrophilic and psychrotrophic clostridial species renders meat 

unacceptable resulting in financial losses and reduced consumer confidence. A number of clostridial strains, 

including Clostridium algidicarnis, Clostridium algidixylanolyticum, Clostridium estertheticum, Clostridium 

frigidicarnis and Clostridium gasigenes, have been implicated in red meat spoilage. Unlike other spoilers, 

these clostridia are able to grow in anaerobic conditions and at chilled temperatures (some at )1.5 °C the 

optimal storage temperature for chilled red meat). The spoilage they cause is characterised by softening of 

the meat, production of large amounts of drip (exudates), offensive odours and in the case of C. estertheticum 

and C. gasigenes production of gas. Spoilage occurs following the introduction of clostridial spores into 

vacuum packages during processing. Germination of spores is necessary for the growth of vegetative cells, 

which cause spoilage. Current mitigation strategies focus on good management practice within meat 

processing plants. However, this is not always sufficient to prevent spoilage. This review summarises the 

issues associated with meat spoilage because of psychrotolerant clostridia and discusses areas that require 

further study. 

Keywords Chilling, food quality, meat products, microbiology. 

Introduction 

Three categories of meat spoilage (as opposed to food 

safety issues such as those caused by toxin producing 

strains of Clostridium botulinum) are caused by psychrotolerant 

clostridia: ‘blown pack’, ‘surface spoilage’ 

and ‘bone taint’. A vacuum package suffering ‘blown 

pack’ spoilage is typically grossly distended to the 

point where meat would be considered off prior to 

opening. On opening the pack, strong and offensive, 

sulphurous, off odours and copious quantities of drip 

are present, and the meat may have a soft texture. 

Products affected include primal cuts of beef, lamb 

and venison, cooked dog rolls, pre-cooked turkey and 

roast beef and Sous-vide (food heat processed under 

vacuum in O2-impermeable barrier bags and stored at 

refrigeration temperature) (Broda et al., 1996; Kalinowski 

& Tompkin, 1999). Initial contamination with 

‘blown pack’ spoilers occurs on the surface of the 

*Correspondent: Fax: +64 7 838 5625; 

e-mail: katharine.adam@agresearch.co.nz 

doi:10.1111/j.1365-2621.2010.02320.x 

Ó 2010 AgResearch Ltd 

meat, and the culprit can be isolated from the drip or 

a surface swab. Post-packaging heat shrink treatments 

of vacuum packs have the potential to accelerate the 

onset of clostridial induced ‘blown pack’ spoilage (Bell 

et al., 2001). ‘Surface’ spoilage differs from ‘blown 

pack’ spoilage in that little or no gas accumulates in 

the pack. On opening, sickly spoilage odours are 

present, and the bacterium can be isolated from drip 

or a surface swab. Clostridia have also been associated 

with deep tissue or ‘bone taint’ spoilage, which is not 

discussed here as the source of the causative organisms 

differs from that of ‘blown pack’ and ‘surface’ 

spoilage. 

Clostridia associated with meat spoilage 

The first report of a Clostridial species being associated 

with the spoilage of fresh, chilled, vacuum-packaged red 

meat was published by Dainty et al. (1989). The 

bacterium involved was later described and named 

Clostridium estertheticum (Collins et al., 1992). A second 

strain of C. estertheticum, C. estertheticum sub species

1540 

Spoilage of red meat by psychrotolerant clostridia K. H. Adam et al. 

laramiense (originally Clostridium Laramie) was isolated 

from spoiled beef shortly after (Kalchayanand et al., 

1993; Spring et al., 2003). Clostridium frigidicarnis, 

Clostridium gasigenes and Clostridium algidixylanolyticum 

were all isolated from spoiled vacuum-packed meat 

originating from New Zealand. Also associated with 

chilled meat spoilage, Clostridium algidicarnis was originally 

isolated from cooked vacuum-packaged refrigerated 

pork. All strains discussed here exhibit the 

characteristics of members of the genus Clostridium as 

well as being psychrotolerant. Phylogeny of individual 

strains, within the genus, is discussed in the original 

species descriptions. Based on analysis, 16S rRNA 

sequences C. algidixylanolyticum belongs to cluster 

XIVa, which also contains C. xylanolyticum (Broda 

et al., 2000a). The other strains listed earlier belong to 

cluster I, which also contains strains of C. botulinum 

types A to F (Broda et al., 1999, 2000b; Lawson et al., 

1994; Spring et al., 2003). 

Psychrotolerance 

Historically, the definitions of the terms psychrophilic, 

psychrotrophic and psychrotolerant have varied between 

papers. Here, the following definitions have been 

adopted: psychrophilic: (cold loving) having optimum 

growth at 12–15 °C, a maximum growth temperature 

of 15–20 °C and minimum growth temperature of )5 

to 5 °C, psychrotrophic: (cold growing) having optimum 

growth at 25–30 °C, a maximum growth temperature 

of 30–35 °C and a minimum growth temperature 

at )5 to 5 °C and psychrotolerant: (cold tolerant) 

capable of growing at

Figure 1 The life cycle of spore forming 

bacteria in the environment and as spoilers of 

vacuum-packaged chilled meats. Cycling 

generally occurs in the environment. Under 

appropriate conditions transfer of spores of 

spoilage causing organisms results in spoiled 

meat. Stages of spore formation based on 

Paredes et al. (2005). 

vacuum packs of red meat, two surveys were conducted 

in New Zealand (Boerema et al., 2003; Broda et al., 

2002). Potential sources of contamination within the 

abattoir can be split into two main categories: 1. soil and 

faeces on animals and 2. equipment. Broda et al. (2002) 

relied on classical microbiological methods to isolate 

potential spoilers, which were then differentiated using 

molecular methods [restriction fragment length polymorphism 

analysis of 16S rRNA PCR (PCR-RFLP) 

product and 16S–23S rRNA internal transcribed spacer 

analysis]. Based on PCR-RFLP banding patterns, isolates 

were separated into two groups. The majority of 

gas producers belonged to the first group and were from 

either hide samples or faecal samples. This suggests soil 

particles attached to hide or present in faeces were the 

primary reservoir from which ‘blown pack’ clostridia 

were introduced on to carcasses. Boerema et al. (2003) 

increased the sensitivity of the detection methods, by 

combining sample enrichment with PCR amplification, 

in a second survey of a New Zealand meat processing 

plant from which spoilt chilled vacuum-packed meat 

had originated. Clostridium estertheticum andC. gasigenes 

were detected in soil, faecal and pelt samples but 

not in boning room or chillers and only to a limited 

extent on the slaughter floor. This survey, as well as one 

carried out by Moschonas et al. (2009) of four commercial 

beef abattoirs in Ireland, reported high levels of 

‘blown pack’ spoilage causing clostridia in samples 

taken at hide removal and in faeces, adds weight to the 

theory that contamination originated from soil or faecal 

material attached to pelts. 

Current mitigation strategies 

Chilled lamb, transported from New Zealand to international 

markets, by sea freight must reliably attain a 

storage life of between 8 and 9 weeks (Gill, 1987). Beef 

Spoilage of red meat by psychrotolerant clostridia K. H. Adam et al. 1541 

and venison transported by sea spends at least as long as 

lamb in chilled storage. Good management practice 

(GMP) including minimising initial contamination, 

reaching and maintaining a uniform, optimum temperature, 

of )1.5 °C, and the initial pH of the meat are all 

important in achieving a long shelf-life, particularly 

where product is to be sold in Europe where the use of 

chemical intervention is not permitted (Bell et al., 2001; 

Gill, 1987). Detection of high numbers of C. estertheticum 

and C. gasigenes in faeces highlights the importance 

of limiting faecal contamination of carcases 

(Moschonas et al., 2009). Prior to slaughter, the muscle 

tissue of cattle is essentially sterile becoming contaminated 

with bacteria as a result of processing. The area of 

the carcass around the opening cut has a high risk of 

becoming contaminated, and care must be taken to 

prevent rollback (the pelt around the opening cut rolling 

in causing the outside of the pelt to come in contact with 

the essentially sterile carcass) (Boerema et al., 2003; 

Moschonas et al., 2009). Maintenance of low temperature 

is particularly important when dealing with coldtolerant 

organisms. An increase in temperature from 

)1.5 to 1 or 4 °C significantly shortened the time to 

initial gas production in packs of vacuum-packed beef 

inoculated with gas producing strains of clostridia (Bell 

et al., 2001). Selection of appropriate meat for chilled 

storage also has an impact as an increase in initial pH of 

Beef from 5.5 to 6.0 was shown to result in shorter times 

to initial gas production (Bell et al., 2001). 

Although maintaining GMP can reduce spoilage by 

psychrotolerant clostridia the risk remains. Contamination 

of animals with spores is unavoidable as are low 

levels of transfer of bacteria from hide to carcass. A 

range of intervention strategies are available for reducing 

the presence of bacteria on red meat including 

trimming, the use of steam vacuum, hot water washes, 

and, where regulations permit, chemical washes (for 

Ó 2010 AgResearch Ltd International Journal of Food Science and Technology 2010, 45, 1539–1544

1542 


example lactic acid) (Jay, 1996). All intervention treatments 

have their limitations. Trimming only removes 

visible contamination and reduces the final weight of the 

carcass. Heat treating spores of C. estertheticum for 

240 s at 100 °C resulted in their inactivation in vitro 

(Broda, 2007). Further work would be required to 

determine whether this effect could be replicated on 

freshly slaughtered carcases without causing permanent 

discolouration. Steam vacuuming and hot water washing, 

while effective against E. coli (Dorsa et al., 1996), 

risks activating spores if the temperature or time of 

application is insufficient. Peroxyacetic acid-(POAA) 

based sanitizer is capable of reducing spores of C. estertheticum 

by at least 4 log CFU mL )1 in vitro (Broda, 

2007). Exposure to POAA resulted in inactivation of 

spores. It is unknown if POAA would be effective at 

reducing spores from the environment on freshly 

slaughtered carcasses. 

Sporulation 

The survival of psychrotolerant clostridial meat spoilers 

in soil and on animal hides prior to entering meat packs 

at processing plants, is dependent upon the ability to 

sporulate, because of their obligate anaerobic nature. 

The first visual sign of sporulation is the formation of an 

asymmetric forespore septum (Fitz-James & Young, 

1969). The smaller of the two halves, called the 

forespore (sometimes called the pre-spore), becomes 

the spore while the larger, called the mother cell, 

supports the spore’s formation. The mother cell engulfs 

the smaller forespore protecting it and producing 

components necessary to spore development. A thick 

peptidoglycan cortex is synthesised between the outer 

and inner forespore membranes accompanied by a large 

decrease in the water content of the forespore protoplast 

and a decrease in the forespore pH (Setlow, 2007). Later 

the forespore takes up large quantities of dipicolinic acid 

that has been synthesised by the mother cell. Finally, the 

mother cell lyses releasing the spore into the environment 

(Piggot & Hilbert, 2004). 

In Bacillus subtilis, sporulation is induced by starvation 

for carbon and ⁄ or nitrogen and is initiated by a 

multi-component phosphorelay. The phosphorelay is 

absent in Clostridium acetobutylicum; however, the 

master regulator, Spo0A is conserved. Spo0A regulates 

expression of sigma factors that drive differential gene 

expression in the forespore and mother cell (Jones et al., 

2008). During sporulation, the mother cell and forespore 

participate in extensive ‘cross-talk’ to ensure gene 

expression in both compartments remains coordinated 

(Setlow, 2007). Unlike B. subtilis, clostridial species 

commonly require complex sporulation media and the 

presence of a slowly fermentable carbon source at the 

end of the log phase of growth to allow continuation of 

energy supply during the early stages of the sporulation 

process. In the development of sporulation media, it has 

been found that even within various strains of Clostridium 

perfringens, variations in the requirements for 

sporulation occur (De Jong et al., 2002). Initiation of 

sporulation may also be affected by cell signalling, when 

cells become crowded, resulting in a rise in the concentration 

of specific signal compounds (Peck et al., 2004). 

Germination 

Spores in vacuum packs of red meat must germinate to 

grow and cause spoilage (Fig. 1). At present, there are 

no published data on germination in psychrotolerant 

clostridia. Strain-specific germinants and cellular and 

molecular processes of germination, in psychrotolerant 

clostridia, are unknown. Germination has been studied 

in a number of mesophilic strains including B. subtilis, 

Bacillus cereus, Clostridium sporogenes and C. botulinum 

typeB (Broussolle et al., 2002; Paredes-Sabja et al., 

2008). Where studied bacterial spore germination followed 

the general pattern: activation, germination and 

outgrowth. Activation is not always necessary for spore 

germination, is reversible, does not result in loss of 

resistance and is often the result of heat treatment 

(Keynan & Evenchik, 1969). The heat treatment process 

used to shrink vacuum packs may be sufficient to 

activate clostridial spores on the surface of meat as it has 

been shown to decrease the time to ‘pack blowing’ in 

packs inoculated with spores of C. estertheticum (Bell 

et al., 2001). Following activation spores will, with 

appropriate stimulation, commit themselves to germination. 

Once initiated, the germination process proceeds 

without the continued presence of the germinant or 

synthesis of new macromolecules. This indicates that 

germination is a process controlled by the sequential 

activation of a set of pre-existing germination-related 

enzymes already present in the mature spore (Johnstone, 

1994; Moir et al., 2002; Okamura et al., 2000). Germination 

results in the loss of refractility and resistance to 

a wide range of environmental assaults including heat, 

UV and solvents (Atrih & Foster, 2002; Foster & 

Johnstone, 1990). After germination, the spore is physiologically 

different from a vegetative cell despite having 

lost many spore-specific properties. Before a cell can be 

considered vegetative, it must undergo outgrowth, 

following which division can proceed (Keynan & 

Evenchik, 1969). 

Germination occurs in response to environmental 

cues called germinants. With the exception of Clostridium 

difficile, all strains of clostridia have been found to 

possess germinant receptor type proteins as part of their 

germination systems (Sebaihia et al., 2007). Germinant 

receptor proteins interact with a specific nutrient or 

nutrients to trigger germination. Studies of germination 

in psychrotolerant clostridia would highlight similarities 

to germination systems in mesophilic spore formers. 

International Journal of Food Science and Technology 2010, 45, 1539–1544 Ó 2010 AgResearch Ltd

Knowledge of the germination systems utilised by 

psychrotolerant clostridia would provide insight into 

the properties of meat packs that make them ideal 

environments for germination. Knowledge of the germination 

systems in psychrotolerant clostridia would 

also potentially lead to improved use of current intervention 

strategies or the formulation of new methods for 

the reduction in spoilage. 

Conclusion 

A number of strains of clostridia, including C. estertheticum, 

are able to spoil chilled ‘vacuum-packed’ beef, 

lamb and venison. As part of the spoilage process, they 

must germinate. The process of germination and the 

genes involved have been well studied in a number of 

strains of bacillus and clostridium. The genes involved in 

germination of psychrotolerant clostridium and the 

germinants responsible for triggering germination are 

not known. Determining the germination system(s) 

involved in the germination of cold-tolerant clostridia 

would potentially lead to methods for the reduction in 

spoilage. 

Acknowledgments 

The financial support of the New Zealand Foundation 

for Research, Science and Technology is gratefully 

acknowledged as is Pauline Hunt for help in preparing 

Figure 1. 

References 

Atrih, A. & Foster, S.J. (2002). Bacterial endospores the ultimate 

survivors. International Dairy Journal, 12, 217–223. 

Bell, R.G., Moorhead, S.M. & Broda, D.M. (2001). Influence of heat 

shrink treatments on the onset of clostridial ‘‘blown pack’’ spoilage 

of vacuum packed chilled meat. Food Research International, 34, 

271–275. 

Boerema, J.A., Broda, D.M. & Bell, R.G. (2002). PCR detection of 

psychrotolerant clostridia associated with deep tissue spoilage of 

vacuum-packed chilled meats. Letters in Applied Microbiology, 35, 

446–450. 

Boerema, J.A., Broda, D.M. & Bell, R.G. (2003). Abattoir sources 

of psychrophilic clostridia causing blown pack spoilage of 

vacuum-packed chilled meats determined by culture-based and 

molecular detection procedures. Letters in Applied Microbiology, 

36, 406–411. 

Broda, D.M. (2007). The effect of peroxyacetic acid-based sanitizer, 

heat and ultrasonic waves on the survival of Clostridium estertheticum 

spores in vitro. Letters in Applied Microbiology, 45, 336– 

341. 

Broda, D.M., DeLacy, K.M., Bell, R.G., Braggins, T.J. & Cook, R.L. 

(1996). Psychrotrophic Clostridium spp. associated with ‘blown 

pack’ spoilage of chilled vacuum-packed red meats and dog rolls in 

gas-impermeable plastic casings. International Journal of Food 

Microbiology, 29, 335–352. 

Broda, D.M., DeLacy, K.M., Cook, R.L. & Bell, G.R. (1997). 

Prevalence of cold-tolerant clostridia associated with vacuumpacked 

beef and lamb stored at abusive and chill temperatures. 

New Zealand Journal of Agricultural Research, 40, 93–98. 

Spoilage of red meat by psychrotolerant clostridia K. H. Adam et al. 1543 

Broda, D.M., Lawson, P.A., Bell, G.R. & Musgrave, D.R. (1999). 

Clostridium frigidicarnis sp. a psychrotolerant bacterium associated 

with ‘blown pack’ spoilage of vacuum-packed meats. International 

Journal of Systematic and Evolutionary Microbiology, 49, 1539–1550. 

Broda, D.M., Saul, D.J., Bell, G.R. & Musgrave, D.R. (2000a). 

Clostridium algidixylanolyticum sp. nov., a psychrotolerant, xylandegrading, 

spore-forming bacterium. International Journal of Systematic 

and Evolutionary Microbiology, 50, 623–631. 

Broda, D.M., Saul, D.J., Lawson, P.A., Bell, G.R. & Musgrave, D.R. 

(2000b). Clostridium gasigenes sp. nov., a psychrophile causing 

spoilage of vacuum-packed meat. International Journal of Systematic 

and Evolutionary Microbiology, 50, 107–118. 

Broda, D.M., Bell, R.G., Boerema, J.A. & Musgrave, D.R. (2002). The 

abattoir source of culturable psychrophilic Clostridium spp. causing 

‘blown pack’ spoilage of vacuum-packed chilled venison. Journal of 

Applied Microbiology, 93, 817–824. 

Broussolle, V., Alberto, F., Shearman, C.A. et al. (2002). Molecular 

and physiological characterisation of spore germination in Clostridium 

botulinum and C-sporogenes. Anaerobe, 8, 89–100. 

Collins, M.D., Rodrigues, U.M., Dainty, R.H., Edwards, R.A. & 

Roberts, T.A. (1992). Taxonomic studies on a phsychrophilic 

Clostridium from vacuum-packed beef: description of Clostridium 

estertheticum sp. nov. FEMS Microbiology Letters, 96, 235–240. 

Dainty, R.H., Edwards, R.A. & Hibbard, C.M. (1989). Spoilage of 

vacuum-packed beef by a clostridium sp. Journal of the science of 

food and agriculture, 49, 153–157. 

De Jong, A.E.I., Beumer, R.R. & Rombouts, F.M. (2002). Optimizing 

sporulation of Clostridium perfringens. Journal of Food Protection, 

65, 1457–1462. 

Dorsa, W.J., Cutter, C.N. & Siragusa, G.R. (1996). Effectiveness of 

steam-vacuum sanitizer for reducing Escherichia coli 0157:H7 

inoculated to beef carcass surface tissue. Letters in Applied Microbiology, 

23, 61–63. 

Fitz-James, P. & Young, E. (1969). Morphology of sporulation. In: 

The Bacterial Spore (edited by G.W. Gould & A. Hurst). Pp. 39–72. 

London & New York: Academic Press. 

Foster, S.J. & Johnstone, K. (1990). Pulling the trigger: the mechanism 

of bacterial spore germination. Molecular Microbiology, 4, 137–141. 

Gill, C.O. (1987). Prevention of microbial contamination in the lamb 

processing plant. In: Elimination of Pathogenic Organisms from Meat 

and Poultry (edited by F.J.M. Smulders). Pp. 203–219. Amsterdam: 

Elsevier. 

Helps, C.R., Harbour, D.A. & Corry, J.E.L. (1999). PCR-based 16S 

ribosomal DNA detection technique for Clostridium estertheticum 

causing spoilage in vacuum-packed chill-stored beef. International 

Journal of Food Microbiology, 52, 57–65. 

Jay, J.M. (1996). Fresh Meats and Poultry. Modern food microbiology. 

Pp. 69–96. New York: Chapman & Hall. 

Johnstone, K. (1994). The trigger mechanism of spore germination: 

current concepts. Journal of Applied Bacteriology, 76, S17–S24. 

Jones, S.W., Paredes, C.J., Tracy, B. et al. (2008). The transcriptional 

program underlying the physiology of clostridial sporulation. 

Genome biology, 9, 7. 

Kalchayanand, N., Ray, B., Field, R.A. & Johnson, M.C. (1989). 

Spoilage of vacuum-packaged refrigerated beef by Clostridium. 

Journal of Food Protection, 52, 424–426. 

Kalchayanand, N., Ray, B. & Field, R.A. (1993). Characteristics of 

psychrotrophic Clostridium laramie causing spoilage of vacuumpackaged 

refrigerated fresh and roasted beef. Journal of Food 

Protection, 56, 13–17. 

Kalinowski, R.M. & Tompkin, R.B. (1999). Psychrotrophic clostridia 

causing spoilage in cooked meat and poultry products. Journal of 

Food Protection, 62, 766–772. 

Keynan, A. & Evenchik, Z. (1969). Activation. In: The Bacterial Spore 

(edited by G.W. Gould & A. Hurst). Pp. 359–396. London & New 

York: Academic Press. 

Lawson, P., Dainty, R.H., Kristiansen, N., Berg, J. & Collins, M.D. 

(1994). Characterization of a psychrotrophic Clostridium causing 

Ó 2010 AgResearch Ltd International Journal of Food Science and Technology 2010, 45, 1539–1544

1544 


spoilage in vacuum-packed cooked pork: description of Clostridium 

algidicarnis sp. nov. Letters in Applied Microbiology, 19, 153–157. 

Moir, A., Corfe, B.M. & Behraven, J. (2002). Spore Germination. 

Cellular and Molecular Life Sciences, 59, 403–409. 

Moschonas, G., Bolton, D.J., Sheridan, J.J. & McDowell, D.A. (2009). 

Isolation and sources of ‘blown pack’ spoilage clostridia in beef 

abattoirs. Journal of Applied Microbiology, 107, 616–624. 

Okamura, S., Urakami, K., Kimata, M. et al. (2000). The N-terminal 

prepeptide is required for the production of spore cortex-lytic 

enzyme from its inactive precursor during germination of 

Clostridium perfringens S40 spores. Molecular Microbiology, 37, 

821–827. 

Paredes, C.J., Alsaker, K.V. & Papoutsakis, E.T. (2005). A comparative 

genomic view of clostridial sporulation and physiology. Nature 

Reviews Microbiology, 3, 969–978. 

Paredes-Sabja, D., Torres, J.A., Setlow, P. & Sarker, M.R. (2008). 

Clostridium perfringens spore germination: characterization of 

germinants and their receptors. Journal of Bacteriology, 190, 1190– 

1201. 

Peck, M.W., Granum, P.E., Gould, G.W. & Mainil, J.G. (2004). Food 

borne clostridia and sporulation, Pathology and Ecology of the 

genus Clostridium in Humans, Animals, and other Foodstuffs: 

Identification, Epidemiology and Prophylaxis. Concerted Action 

QLK2-CT2001-01267 Presse de la faculté de Me´decine Vétéraire de 

l’Université de Liege, 4000 Liège, Belgique. 

Piggot, P.J. & Hilbert, D.W. (2004). Sporulation of Bacillus subtilis. 

Current Opinion in Microbiology, 7, 579–586. 

Russell, N.J. (1997). Psychrophilic bacteria-molecular adaptations of 

membrane lipids. Comparative Biochemistry and Physiology, 118A, 

489–493. 

Sebaihia, M., Peck, M.W., Minton, N.P. et al. (2007). Genome 

sequence of a proteolytic (Group I) Clostridium botulinum strain 

Hall A and comparative analysis of the clostridial genomes. Genome 

Research, 17, 1082–1092. 

Setlow, P. (2007). I will survive: DNA protection in bacterial spores. 

Trends in Microbiology, 15, 172–180. 

Spring, S., Merkhoffer, B., Weiss, N., Kroppenstedt, R.M., Hippe, H. 

& Stackebrandt, E. (2003). Characterization of novel psychrophilic 

clostridia from an Antarctic microbial mat: description of Clostridium 

frigoris sp. nov., Clostridium lacusfryxellense sp. nov., Clostidium 

bowmanii sp. nov. and reclassification of Clostidium laramiense as 

Clostridium estertheticum subsp. laramiense subsp. nov. International 

Journal of Systematic and Evolutionary Microbiology, 53, 1019–1029. 

Yang, X., Balamurugan, S. & Gill, C.O. (2009a). Substrate utilization 

by Clostridium estertheticum cultivated in meat juice medium. 

International Journal of Food Microbiology, 128, 501–505. 

Yang, X., Gill, C. & Balamurugan, S. (2009b). Effects of temperature 

and pH on the growth of bacteria isolated from blown packs 

of vacuum-packaged beef. Journal of Food Protection, 72, 2380– 

2385. 

International Journal of Food Science and Technology 2010, 45, 1539–1544 Ó 2010 AgResearch Ltd


Original article 

Shelf life extension of durum semolina-based fresh pasta 

Cristina Costa, 1 Annalisa Lucera, 1 Marcella Mastromatteo, 2 Amalia Conte 1,2 & Matteo Alessandro Del Nobile 1,2 * 

1 Department of Food Science, University of Foggia, Via Napoli, 25 – 71100 Foggia, Italy 

2 Istituto per la Ricerca e le Applicazioni Biotecnologiche per la Sicurezza e la Valorizzazione dei Prodotti Tipici e di Qualità, BIOAGROMED, 

Università degli Studi di Foggia, Via Napoli, 25 – 71100 Foggia, Italy 

(Received 15 January 2010; Accepted in revised form 12 March 2010) 

Summary In this work, the combined effects of chitosan, modified atmosphere packaging (MAP) and packaging barrier 

properties on shelf life of fresh pasta is presented. In particular, all pasta samples were packaged under active 

and passive MAP in two different polymeric films with high and low barrier properties. In order to assess the 

influence of the variables described beforehand on the shelf life of pasta, the sensorial and microbiological 

quality has been monitored during storage. Results confirmed the antimicrobial properties of chitosan. 

Moreover, the findings recorded in this study suggest that the shelf life of fresh pasta is limited by the 

sensorial characteristics. Statistically significant differences between the shelf life of pasta packaged in low 

barrier and high barrier films were found. The best result was obtained for samples packaged in high barrier 

film, due to the ability of the packaging to maintain the gas headspace conditions during the storage. 

Keywords Antimicrobial agents, pasta, shelf life. 

Introduction 

Pasta is a generic word for a wide range of products with 

very different characteristics in terms of shape, colour, 

composition, storage, requirements and use. According 

to the Italian legislation ‘Pasta’ is defined as the product 

obtained by extrusion or lamination and successive 

drying (to 12.5% maximum water content) of a dough 

made of durum wheat semolina and water. Pasta with 

moisture content more than 24% is defined ‘fresh pasta’ 

and requires storage temperatures lower than 4 °C. 

For its high water content fresh pasta is a product 

easily perishable; spoilage is due to the metabolic 

activity of microorganisms (bacteria, yeasts, moulds) 

that can easily grow in a product with these characteristics 

(Del Nobile et al., 2009a; Zardetto, 2005). The 

microbial quality in fresh pasta at the end of the 

production process is strictly related to the characteristics 

of raw materials, such as durum semolina or 

alternative flours and water and to the methods used 

to make pasta (homemade pasta, pilot plant or industrial 

plant). Moreover, the cell load is also influenced by 

the methods used to sanitize the plant and to prevent 

new outer contaminations. Large differences can be 

found between pasta made with an industrial production 

line and homemade pasta (Del Nobile et al., 2009a); 

*Correspondent: Fax: (+39) 881 589 242; 

e-mail: ma.delnobile@unifg.it 

doi:10.1111/j.1365-2621.2010.02277.x 

Ó 2010 The Authors. Journal compilation Ó 2010 Institute of Food Science and Technology 

the lower microbial loads of homemade pasta can be 

attributed to the better management of the facilities. 

To prolong the shelf life of fresh pasta different 

methods can be proposed. Actually, in fresh pasta is 

allowed the use of chemical preservatives and bacteriostatic 

compounds to avoid microorganisms proliferation 

(FDA, 2006). However, due to the increasing consumer 

demand for high quality food without chemical agents, 

the use of modified atmosphere packaging (MAP) or the 

natural compounds with antimicrobial properties to 

control the microbial proliferation has been receiving 

considerable attention from the scientific researchers 

(Del Nobile et al., 2009a,b). 

Among the natural antimicrobials, chitosan has 

received considerable interest for commercial applications 

(Dutta et al., 2009; Mohy Eldin et al., 2008; Zheng 

& Zhu 2003). 

Different studies reported the implication of MAP to 

preserve the quality of fresh pasta (Del Nobile et al., 

2009b; Zardetto, 2005). In particular, low O 2 and high 

CO2 concentrations limit the development of microorganisms 

and reduce the growth and toxin production 

of different moulds, strictly aerobic and sensitive to high 

concentrations of CO2 (Zardetto, 2005). Moreover, 

the success of the MAP is closely connected to the 

permeability of the film used for packaging. A number 

of factors influences the barrier properties of packaging 

materials. In particular, gas diffusion across a film is 

determined by film structure, thickness, area, gradient

1546 

Shelf life extension of durum semolina-based fresh pasta C. Costa et al. 

concentrations, concentrations across the film, temperature, 

and differences in pressure across the film 

(Gholizadeh et al., 2007). 

In several studies, the quality of fresh pasta is 

determined by the assessment of either the microbiological 

(Alamprese et al., 2004; Zardetto, 2005) or other 

quality indices (i.e. protein contents, viscoelasticities, 

furosine, texture analysis) (Kovacs et al., 1997; Alamprese 

et al., 2005; Zardetto and Dalla Rosa, 2006). In our 

previous works, the influence of different natural antimicrobial 

compounds and MAP on microbial stability 

of amaranth-based homemade fresh pasta was evaluated. 

The recorded results highlighted that the combined 

effect of chitosan and MAP improved the microbiological 

quality of fresh pasta (Del Nobile et al., 2009a,b). 

However, the shelf life of a given food is related to both 

its sensorial and microbial quality. 

Thus, the aim of this work is to evaluate the influence 

of chitosan, MAP and film packaging on shelf life 

extension of durum semolina-based fresh pasta, made by 

a pilot plant. In particular, three different concentrations 

of chitosan and two packaging films with different 

barrier properties were tested to assess their influence on 

the microbiological and sensorial quality of fresh pasta. 

Materials and methods 

Raw materials and pasta production 

Fresh pasta samples were produced by using durum 

semolina (provided by Mulini Tandoi, Corato, Bari, 

Italy). Chitosan (Danisco, Braband, Denmark), as 

antimicrobial compound, was added to the dough at 

three different concentrations: 1000, 2000 and 

3000 mg kg )1 . Semolina and tap water (30% v ⁄ w) were 

mixed to prepare pasta dough. The samples were 

prepared with a pilot plant made of an extruder 

(60VR; Namad, Rome, Italy) and equipped with a 

bronze head to give the pasta dough the shape of 

macaroni. The kneading time applied to produce fresh 

pasta was 20 min. To obtain pasta samples loaded at 

different antimicrobial concentrations, three active solutions 

were prepared by dissolving chitosan in lactic 

acid (0.3% in the final pasta dough). The solutions 

were added to the dough, separately, to obtain final 

concentrations of 1000 mg kg )1 , 2000 mg kg )1 and 

3000 mg kg )1 

of chitosan in the fresh pasta 

(CHT1000, CHT2000 and CHT3000). As control, pasta 

samples without antimicrobial (CNT) were also prepared. 

No lactic acid was added to the control sample. 

About 200 g of pasta samples were arranged in a 

plastic tray, that in turn was packaged in a plastic bag. 

Two polymeric films with different characteristics were 

used as packaging bags. An anti-fog low-barrier film 

(Low-B), made up of polypropylene (PP) with 30 lm 

thickness (Carton Pack, Bari, Italy) and an anti-fog 

high-barrier multilayer film (High-B) made up of polyethylene 

terephthalate (PET), ethylene-vinyl alcohol 

(EVOH) and polyethylene (PE), with 90 lm thickness 

(Di Mauro Officine Grafiche s.p.a., Napoli, Italy). All 

packaged samples were sealed by means of a thermal 

sealer (Gandus sealers, Milan, Italy) under ordinary 

atmosphere (passive MAP, named as P-MAP) and 

modified atmospheric conditions (active MAP, named 

as A-MAP). To realise the modified headspace conditions 

the following gas concentrations were used: 70% 

CO2 and 30% N2. All the samples were stored at 4 °C. 

Headspace gas composition 

The changes in headspace O 2 and CO 2 concentration of 

packaged samples were measured using a PBI Dansensor 

O2 ⁄ CO2 analyzer (Checkmate 9900, Denmark). The 

volume taken from the package headspace for gas 

analysis was about 10 cm 3 . To avoid modifications in 

the headspace gas composition due to gas sampling, 

each package was used only for a single measurement of 

the headspace gas composition. Two bags were used for 

each measurement. 

Permeation tests 

The water vapour transmission rate (WVTR) was 

determined by means of a Lyssy permeabilimeter 

(Model L80-5000; PBI Dansensor, Milan, Italy). Samples 

of each film with a surface area of 50 cm 2 were 

tested at 23 °C and 85% of relative humidity (RH). 

The oxygen transmission rate (OTR) was determined 

by means of an Ox-Tran permeabilimeter (Model 2 ⁄ 21; 

Mocon, Neuwied, Germany). Samples of each film with 

a surface area of 5 cm 2 were tested at 23 °C and 0% RH 

at the upstream and the downstream sides of the sample. 

The carbon dioxide transmission rate (CDTR) was 

determined by means of a Permatran permeabilimeter 

(Mocon, Model C 4 ⁄ 41). Samples of each film with a 

surface area of 5 cm 2 were tested at 23 °C and 0% RH 

at the upstream and the downstream side of the 

sample. WVTR, OTR and CDTR tests were performed 

twice. 

Microbiological analyses 

For microbiological analyses, about 25 g of sample was 

aseptically removed from each package, placed in a 

stomacher bag, diluted with 0.9% NaCl solution and 

homogenised with a stomacher LAB Blender 400 (Pbi 

International, Milan, Italy). Serial dilutions in sterile 

saline solution were plated onto appropriate media. The 

media and the conditions were the following: plate count 

agar (PCA) incubated at 30 °C for 48 h for aerobic 

mesophilic bacteria and at 7 °C for 10 days for psychrotrophic 

bacteria; Violet Red Bile Agar (VRBA) 

International Journal of Food Science and Technology 2010 Ó 2010 The Authors. Journal compilation Ó 2010 Institute of Food Science and Technology

incubated at 37 °C for 24 h for total coliforms; Baird- 

Parker Agar, supplemented with egg yolk tellurite 

emulsion, incubated at 37 °C for 48 h for Staphylococcus 

spp.; deMan Rogosa Sharpe agar (MRS), added 

with 0.17 g L )1 cycloheximide (Sigma-Aldrich, Milan, 

Italy) incubated at 30 °C for 48 h for lactic acid 

bacteria; Sabouraud Dextrose Agar, added with 

0.1 g L )1 chloramphenicol (C. Erba, Milan, Italy), 

incubated at 25 °C for 48 h for yeasts and 25 °C for 

5 days for moulds. All media and supplements were 

from Oxoid (Milan, Italy). All microbiological analyses 

were performed twice on two different batches. 

pH evaluation 

The measurement of pH on the homogenised product 

was performed twice on two different batches by using a 

pH-meter (Crison, Barcelona, Spain). 

Sensory analysis 

Both uncooked and cooked fresh pasta were subjected 

to sensory evaluation. All the uncooked samples were 

submitted in a single session to a panel of eight trained 

tasters for estimation of colour, odour and overall 

quality. In addition, adhesiveness, bulkiness, firmness, 

elasticity, colour, odour, taste and overall quality were 

evaluated on cooked pasta. Each unpackaged sample 

(about 200 g) was cooked in a cooker containing about 

4000 mL of tap water at 100 °C. A nine-point hedonic 

rating scale, where 1 corresponded to ‘extremely 

unpleasant’ and 9 to ‘extremely pleasant’, was used to 

perform the panel test. A score equal to 5 was used as 

the threshold for product acceptability (Del Nobile 

et al., 2009a). The panelists were selected on the basis of 

their sensory skills (ability to accurately determine and 

communicate the sensory attributes, appearance, odour, 

flavour and texture of a food product) (Meilgaard et al., 

1999). Prior to testing pasta, the panelists were trained 

in sensory vocabulary and identification of particular 

attributes, by using samples of commercial pasta. The 

analyses were performed in isolated booths in a 

standard taste panel kitchen. 

Moreover, panellists were also asked to search for 

visual moulds, thus allowing determining the day 

between the latest storage time at which moulds were 

not visible and the earliest storage time at which moulds 

were visible, hereinafter referred to as VMT (Visual 

Moulds Time). 

Modelling 

To quantitatively determine the effectiveness of the 

combined effects of the two packaging films, chitosan 

and MAP in preventing microbial growth, the storage 

time at which the viable cell concentration reached its 

Shelf life extension of durum semolina-based fresh pasta C. Costa et al. 1547 

threshold value was calculated according to the Gompertz 

equation, as re-parameterised by Corbo et al. (2006): 

logðNðtÞÞ ¼ logðNmaxÞ 

A exp exp l max 2:71 

þ A exp exp l max 2:71 

k MAL 

A 

þ 1 

k t 

A 

þ 1 ð1Þ 

where N(t) is the viable cell concentration (CFU g )1 )at 

storage time t, A is related to the difference between the 

decimal logarithm of maximum bacterial growth 

attained at the stationary phase and the decimal 

logarithm of the initial cell load concentration 

(CFU g )1 ), lmax is the maximal specific growth rate 

(Dlog[CFU g )1 ] day )1 ), k is the lag time (day), t is the 

time (day), Nmax is the cell load concentration threshold 

value (CFU g )1 ), MAL is the microbial acceptability 

limit (day) (i.e. the storage time at which the N(t) equals 

Nmax). In the case of total mesophilic and psychrotrophic 

bacteria the value of N max was set to 

10 6 CFU g )1 , whereas in the case of total coliforms 

and Staphylococcus spp. it was set to 10 4 CFU g )1 

(Ministerial Health Decree 32, 1985). 

A similar approach was used to quantitatively determine 

the efficacy of the tested variables on sensorial 

quality. To this aim, the Gompertz equation, as reparameterised 

by Corbo et al. (2006), was also fitted to 

the sensorial data: 

SQðtÞ ¼SQ min 

A Q exp exp l Q max 

2:71 kQ 

SAL 

A Q 

þ 1 

þ A Q exp exp l Q kQ t 

max 2:71 þ 1 ð2Þ 

AQ where SQ(t) is the pasta sensorial quality at time t, A Q is 

related to the difference between the sensorial quality 

attained at the stationary phase and the initial value of 

pasta sensorial quality, l Q max is the maximal rate at 

which SQ(t) decreases, k Q is the lag time, SQmin is the 

threshold value, SAL is the sensorial acceptability limit 

(i.e. the storage time at which the SQ(t) equals SQmin), 

and t is the storage time. The value of SQmin was set 

equal to 5. 

Statistical analysis 

The values of MAL, SAL and shelf life of all the 

investigated samples were compared by one-way anova 

analysis. A Duncan’s multiple range test, with the 

option of homogeneous groups (P < 0.05), was used to 

determine significance among differences. To this aim, 

statistica 7.1 for Windows (StatSoft Inc., Tulsa, OK, 

USA) was used. 

Ó 2010 The Authors. Journal compilation Ó 2010 Institute of Food Science and Technology International Journal of Food Science and Technology 2010

1548 


Results and discussion 

Film barrier properties 

Table 1 shows the values of WVTR, OTR and CDTR of 

the two films. As it can be seen, WVTR values are very 

similar; on the contrary, the OTR and CDTR values 

differ by three orders of magnitude, being the Low-B 

film the most permeable to oxygen and carbon dioxide. 

In fact, these two films were chosen to test the ability of 

different plastic materials to maintain the active MAP 

during storage and consequently to preserve fresh pasta. 

Finally, it must be highlighted that the WVTR, OTR 

and CDTR values should be considered for sole 

comparative purposes as the permeation tests were 

conducted at 23 °C and not at 4 °C, the temperature at 

which shelf life tests were conducted. 

As an example, the Fig. 1 shows the headspace 

oxygen and carbon dioxide concentration plotted as a 

function of storage time for CNT samples packaged 

under active MAP. Similar trends were also obtained 

for samples with chitosan packaged in the same 

conditions (data not shown). Regardless of packaging 

film used, pasta under P-MAP conditions (data not 

shown), showed a reduction of oxygen concentration 

and an increase of carbon dioxide, principally due to 

the metabolic activity of the aerobic microorganisms 

(Cruz et al., 2006). Differently from what expected, in 

the High-B bag under P-MAP no anaerobic conditions 

were created. Concerning the product packaged under 

A-MAP, relevant differences between samples in Low- 

B and High-B bags were recorded. In particular, while 

the Low-B bag had a fast change in both carbon 

dioxide and oxygen headspace concentrations, the gas 

mixture in the High-B bag was maintained constant 

throughout the entire monitoring period. This experimental 

evidence could be related to the limited 

development of microbial groups producing carbon 

dioxide. It is also worth noting that samples packaged 

in Low-B under A-MAP were monitored for a shorter 

period, due to the proliferation of visible moulds on 

pasta. The recorded results demonstrated that film gas 

barrier properties can play a key role in preserving 

product quality (Conte et al., 2009a; Del Nobile et al., 

2009c). 

Table 1 Values of water vapour transmission rate (WVTR), oxygen 

transmission rate (OTR) and carbon dioxide transmission rate 

(CDTR) of the two selected packaging films 

Film 

WVTR 

[g ⁄ (m 2 Æday)] 

OTR 

[cc ⁄ (m 2 Æ day)] 

CDTR 

[cc ⁄ (m 2 Æ day)] 

Low Barrier (30 lm) 0.71 ± 0.04 1971.57 ± 33.15 6311.56 ± 28.92 

High Barrier (90 lm) 0.69 ± 0.01 2.64 ± 0.12 2.5 ± 0 

(a) 

(b) 

Figure 1 Headspace O2 (•) and CO2 (¤) concentration plotted as a 

function of storage time for CNT samples packaged under active MAP 

in (a) low-barrier and (b) high-barrier film. 

Microbiological stability 

The influence of chitosan concentration, MAP and 

packaging film barrier properties on microbial quality 

loss during storage of fresh pasta was assessed by 

monitoring the viable cell concentration of the main 

spoilage microbial groups (i.e., mesophilic and psychrotrophic 

bacteria, total coliforms, Staphylococcus spp., 

lactic acid bacteria, yeasts and moulds). Figures S1 and S2 

show the evolution of total mesophilic bacteria for pasta 

with and without chitosan, packaged in the two selected 

films, under P-MAP and A-MAP conditions. Similar 

microbial trends were found for psychrotrophic bacteria, 

lactic acid bacteria and yeasts (data no shown). The 

curves reported in the figures were obtained by fitting 

eqn (1) to the experimental data; the solid horizontal line 


is the threshold value as imposed by the law. MAL has to 

be intended as the time at which the fresh pasta is no 

more marketable, due to the imposed threshold (Ministerial 

Health Decree 32, 1985). MAL values are listed in 

Table 2 for each microbial group. These values were 

calculated when the microbial load in pasta samples was 

found higher than the threshold. As can be seen in the 

above figure, the total mesophilic cell load steadily 

overlaps the limit, reaching a maximum population of 

about 10 8 CFU g )1 or higher in all CNT samples. On the 

other hand, the samples loaded at different chitosan 

concentrations showed lower cell loads and consequently, 

higher MAL Mesophilic values. The microbial 

growth in samples packaged in Low-B under P-MAP 

and A-MAP, and in High-B under P-MAP was favoured 

by the presence of oxygen, which generally promotes the 

growth of aerobic microorganisms (Cruz et al., 2006). 

The oxygen was naturally present in the P-MAP and 

gradually increased in the headspace of Low-B bag 

sealed under A-MAP, due to the scarce film barrier 

properties. Conversely, for pasta packed in High-B film 

under A-MAP, the mesophilic population can be related 

to the development of anaerobic microorganisms. As can 

be inferred from data, chitosan efficiently delay the 

growth of mesophilic bacteria in pasta packaged in both 

films under P-MAP and A-MAP, if compared to the 

CNT sample. These results confirmed the well-known 

antimicrobial properties of chitosan on Gram positive 

and negative bacteria (No et al., 2007). MAL Mesophilic 

value for CHT3000 packaged in High-B under A-MAP 

was 21.51 days, whereas CHT1000 and CHT2000 sam- 


Table 2 Microbial acceptability limit (MAL) and sensorial acceptability limit (SAL) (day; mean ± SD) of pasta samples packaged in the high 

barrier film; visual moulds time (VMT) defined as the day between the latest storage time at which moulds were not visible and the earliest 

storage time at which moulds were visible; Shelf life assumed as the lowest value between MAL, SAL and VMT 

MAL (day) SAL Overall quality (day) 

MAL Mesophilic MAL Psychrotrophic MAL Coliforms MAL Staphylococcus VMT (day) SAL O.Q.Uncooked SAL O.Q.Cooked 

ples showed lower MAL Mesophilic values (11.70 and 

10.51 days, respectively). It is worth noting that the 

different time scale used in Figs S1 and S2 is related to 

the detection of visible moulds, most probably proliferated 

for the headspace oxygen concentration. 

In addition, the combination of film barrier properties 

and MAP influenced the microbial quality loss. In fact, 

higher MAL values were obtained for samples packaged 

in High-B film under A-MAP. In particular, for these 

samples visible moulds did not occur, whereas for the 

other samples visible moulds were detected at the tenth 

day storage. Moreover, results suggested that MAP and 

chitosan can act in synergic way to control the microbial 

stability, as also previously reported by Del Nobile et al. 

(2009b). 

The total coliforms and Staphylococcus spp. were 

always found below the threshold imposed by law, in all 

fresh pasta samples, except the CNT samples packaged 

in High-B under P-MAP that overlapped the limit 

imposed for Staphylococcus spp. (10 4 CFU g )1 ) already 

after 24 h of storage. Although during the entire storage 

period no moulds were recorded by plate count, visible 

moulds appeared on pasta surface, as consequence of 

numerous factors affecting moulds proliferation 

(Sautour et al., 2002; Samapundo et al., 2007). 

pH evaluation 

Shelf life (day) 

Samples in Low-B 

CNT P-MAP 0.48 ± 0.48a 1.93 ± 1.07a >12 >12 12 6.64 ± 0.35a >12 0.48 ± 0.48a 

CHT1000 P-MAP 7.28 ± 0.80b 7.83 ± 1.87b >12 >12 12 8.80 ± 0.51b >12 7.28 ± 0.80d 

CHT2000 P-MAP 9.16 ± 0.60c >12 >12 >12 12 8.55 ± 0.57b >12 8.55 ± 0.57b 

CHT3000 P-MAP >16 >16 >16 >16 16 12.16 ± 1.37d >16 12.16 ± 1.37e 

CNT A-MAP 0.61 ± 0.54a 1.02 ± 0.23a >12 >12 12 7.26 ± 0.31a 9.46 ± 0.17a 0.61 ± 0.54a 

CHT1000 A-MAP >12 >12 >12 >12 12 8.84 ± 0.46b >12 8.84 ± 0.46b,c 

CHT2000 A-MAP >12 >12 >12 >12 12 8.99 ± 0.43b,c 9.58 ± 0.11a 8.99 ± 0.43b,c 

CHT3000 A-MAP >16 >16 >16 >16 16 10.00 ± 0.09c 12.06 ± 1.62b 10.00 ± 0.09c 

Samples in High-B 

CNT P-MAP 0.69 ± 0.20a 0.86 ± 0.14a >10 0.41 ± 0.14 10 4.95 ± 0.15a >10 0.41 ± 0.14a 

CHT1000 P-MAP >13 >13 >13 >13 13 10.92 ± 3.10b >13 10.92 ± 3.10b 

CHT2000 P-MAP >13 >13 >13 >13 13 >13 >13 13 b 

CHT3000 P-MAP >17 >17 >17 >17 17 >17 >17 17 c 

CNT A-MAP 1.76 ± 1.18a 1.69 ± 0.30b >8 >8 >8 5.15 ± 0.35a >8 1.69 ± 0.30a 

CHT1000 A-MAP 11.70 ± 1.57b >37 >37 >37 >37 >37 20.93 ± 3.29a 11.70 ± 1.57b 

CHT2000 A-MAP 10.51 ± 1.40b >43 >43 >43 >43 19.54 ± 2.54c 24.31 ± 2.12a 10.51 ± 1.40b 

CHT3000 A-MAP 21.51 ± 5.10c >43 >43 >43 >43 17.75 ± 2.44c 24.03 ± 2.59a 17.75 ± 2.44c 

Some differences in pH values between samples were also 

detected. In particular, CNT samples showed pH values 

higher than samples loaded with chitosan, 6.19 ± 0.01 


1550 


Figure 2 Uncooked pasta sensorial quality during storage time for 

samples packaged in (a) low-barrier and (b) high-barrier film. The 

curves are the best fit of eqn 2 to the experimental data. 

and 4.50 ± 0.01, respectively. The lower pH values of 

samples loaded with chitosan can be attributed to the 

lactic acid used to dissolve chitosan in the dough. It is 

worth noting that Del Nobile et al. (2009a) proved that 

lactic acid did not affect, to a great extent, microbial 

growth. The pH values of all pasta samples packaged 

under P-MAP and A-MAP decreased during storage. In 

particular, CNT packaged in both MAP achieved pH 

values of 5.35 ± 0.02 and 4.25 ± 0.01, respectively. In 

the former case this was probably due to the rapid 

microbial increase, whereas in the latter case it could be 

ascribed to the presence of high carbon dioxide headspace 

concentration (Ke et al., 1991). 

Sensory evaluation 

In Fig. 2 the overall quality of uncooked pasta packaged 

in High-B film under P-MAP and A-MAP conditions is 

reported; the solid horizontal line is the overall quality 

threshold. The curves shown in this figure were obtained 

by fitting equation (2) to the sensorial data. SAL values 

of the overall quality (SALO.Q) listed in Table 2 for 

uncooked and cooked pasta were calculated only when 

the sensory attribute judgement was below 3. It is worth 

noting that SAL O.Q. is the time at which the investigated 

fresh pasta is no more marketable from a sensory point 

of view. As can be seen, the control samples under 

both ordinary atmosphere and MAP rapidly fall down 

below the threshold value. In fact, SAL O.Q. values of 

CHT1000, CHT2000 and CHT3000 were found to be 

higher than that of the control samples. It is also worth 

highlighting that tests were stopped because of the 

presence of visible moulds and ⁄ or when microbial load 

or sensorial overall quality reached the relative threshold 

value. In particular, for the uncooked pasta packaged 

in both films under P-MAP and A-MAP 

conditions, the SAL O.Q. was mainly influenced by visible 

moulds appearing on the product surface and by the 

odour, respectively. On the contrary, the odour was the 

sensorial attribute that limited the overall quality of all 

cooked pasta samples (data not shown). 

Shelf life evaluation 

In Table 2 the shelf life values of all packaged samples 

are also reported, these data representing the lowest 

values between MAL, SAL O.Q. and VMT (Conte et al., 

2009b). As can be inferred from the data, the shelf life 

values of CNT sample packaged in Low-B under P- 

MAP and A-MAP are similar, probably due to the 

scarce barrier properties of the film, that deleted the 

effects deriving from MAP. Data also show that shelf 

life values of chitosan-loaded samples were higher than 

the CNT sample. In particular, the shelf life of 

CHT3000 sample packaged under P-MAP and 

A-MAP was 12.16 and 10.00. respectively, compared 

to a shelf life less than 1 day recorded in the CNT pasta. 

For most samples the shelf life was limited by the 

sensorial quality of the uncooked pasta, except for the 

CNT under both packaging atmospheres and CHT1000 

packaged under P-MAP, that became unacceptable for a 

high microbial proliferation (MAL Mesophilic ). 

Results in Table 2 also highlight that for samples 

packaged in high barrier film under P-MAP conditions, 

microbial and sensorial quality were both responsible 

for the unacceptability of CNT sample and CHT1000 

respectively, whereas the shelf life of CHT2000 and 

CHT3000 sample was limited by the development of 

visible moulds. Conversely, for CNT, CHT1000, 

CHT2000 samples packaged under A-MAP the shelf 

life was limited by the microbial quality; instead the 

sensory characteristics limited the shelf life of CHT3000 

sample. From a sensorial point of view, no statistically 

significant differences were found between the samples 

packaged under P-MAP and A-MAP, even though the 

use of High-B film with A-MAP delayed visible moulds 

proliferation. 

Conclusion 

In this work the influence of chitosan, gas headspace in 

the package and film barrier properties on shelf life of 


durum semolina-based fresh pasta was studied. In 

particular, three different chitosan concentrations and 

two packaging films with high and low barrier properties 

were tested. To assess the influence of all the selected 

variables on the shelf life of fresh pasta the microbiological 

and sensorial quality were monitored. Results 

suggested that the sensorial quality, in particular the 

odour of the packaged product, played a significant role 

in determining the product acceptability. Moreover, 

there are statistically significant differences between the 

samples packaged in Low-B and High-B film. The high 

shelf life value of pasta packaged in High-B under 

A-MAP (17.75 days) can be ascribed to the ability of 

the packaging film in maintaining the initial modified 

headspace conditions during the entire storage period. 

Results recorded in this work highlighted that MAP, 

chitosan and high barrier packaging system can act in 

synergic way to control the quality loss of fresh pasta 

during refrigerated storage from both microbial and 

sensorial points of view. 

References 

Alamprese, C., Rossi, M., Casiraghi, E., Hidalgo, A. & Rauzzino, F. 

(2004). Hygienic quality evaluation of the egg product used as 

ingredient in fresh egg pasta. Food Chemistry, 87, 313–319. 

Alamprese, C., Iametti, S., Rossi, M. & Bergonzi, D. (2005). Role of 

pasteurisation heat treatments on rheological and protein structural 

characteristics of fresh egg pasta. European Food Research and 

Technology, 221, 759–767. 

Conte, A., Gammariello, D., Di Giulio, S., Attanasio, M. & Del 

Nobile, M.A. (2009a). Active coating and modified-atmosphere 

packaging to extend the shelf life of Fior di Latte cheese. Journal of 

Dairy Science, 92, 887–894. 

Conte, A., Scrocco, C., Brescia, I. & Del Nobile, M.A. (2009b). 

Packaging strategies to prolong the shelf life of minimally processed 

lampascioni (Muscari comosum). Journal of Food Engineering, 

90, 199–206. 

Corbo, M.R., Del Nobile, M.A. & Sinigaglia, M. (2006). A novel 

approach for calculating shelf life of minimally processed vegetables. 


Cruz, R.S., de Fa´tima Ferreira Soares, N. & de Andrade, N.J. (2006). 

Evaluation of oxygen absorber on antimicrobial preservation of 

lasagna – type fresh pasta under vacuum packed. Cieˆncia e agrotecnologia 

Lavras, 30, 1135–1138. 

Del Nobile, M.A., Di Benedetto, N., Suriano, N., Conte, A., 

Lamacchia, C., Corbo, M.R. & Sinigaglia, M. (2009a). Use of 

natural compounds to improve the microbial stability of Amaranthbased 

homemade fresh pasta. Food Microbiology, 26, 151–156. 

Del Nobile, M.A., Di Benedetto, N., Suriano, N., Conte, A., Corbo, 

M.R. & Sinigaglia, M. (2009b). Combined effects of chitosan and 

MAP to improve the microbial quality of amaranth homemade fresh 

pasta. Food Microbiology, 26, 587–591. 

Del Nobile, M.A., Conte, A., Scrocco, C., Brescia, I., Speranza, B., 

Sinigaglia, M., Perniola, R. & Antonacci, D. (2009c). A study on 

quality loss of minimally processed grapes as affected by film 

packaging. Postharvest Biology and Technology, 51, 21–26. 

Dutta, P.K., Tripathi, S., Mehrotra, G.K. & Dutta, J. (2009). 

Perspectives for chitosan based antimicrobial films in food applications. 

Food Chemistry, 114, 1173–1182. 


FDA (2006). Food and drug administration department of health and 

human services. 1CFR172.860. 

Gholizadeh, M., Razavi, J. & Mousavi, S.A. (2007). Gas permeability 

measurement in polyethylene and its copolymer films. Materials and 

Design, 28, 2528–2532. 

Ke, D., Goldstein, L., O’Mahony, M. & Kader, A.A. (1991). Effects 

of short term exposure to low O2 and high CO2 atmospheres on 

quality attributes of strawberries. Journal of Food Science, 56, 

50–54. 

Kovacs, M.I.P., Poste, L.M., Butler, G., Woods, S.M., Leisle, D., 

Noll, J.S. & Dahlke, G. (1997). Durum Wheat Quality: Comparison 

of Chemical and Rheological Screening Tests with Sensory Analysis. 

Journal of Cereal Science, 25, 65–75. 

Meilgaard, M., Civille, G.V. & Carr, B.T. (1999). Sensory Evaluation 

Techniques. 3rd edn. Boca Raton: CRC Press. 

Mohy Eldin, M.S., Soliman, E.A., Hashem, A.I. & Tamer, T.M. (2008). 

Antibacterial activity of chitosan chemically modified with new 

technique. Trends in Biomaterial and Artificial Organs, 22, 121–133. 

No, H.K., Meyers, S.P., Prinyawiwatkul, W. & Xu, Z. (2007). 

Application of chitosan for improvement of quality and shelf life 

of foods: a review. Journal of Food Science, 72, 100–187. 

Samapundo, S., Devlieghere, F., Geeraerd, A.H., De Meulenaer, B., 

Van Impe, J.F. & Debevere, J. (2007). Modelling of the individual 

and combined effects of water activity and temperature on the radial 

growth of Aspergillus flavus and A. parasiticus on corn. Food 


Sautour, M., Soares Mansur, C., Divies, C., Bensoussan, M. & 

Dantigny, P. (2002). Comparison of the effects of temperature 

and water activity on growth rate of food spoilage moulds. Journal 

of Industrial Microbiology and Biotechnology, 28, 311–316. 

Zardetto, S. (2005). Potential application of near infrared spectroscopy 

for evaluating thermal treatments of fresh egg pasta. Food Control, 

16, 249–256. 

Zardetto, S. & Dalla Rosa, M. (2006). Study of the effect of lamination 

process on pasta by physical chemical determination and near 

infrared spectroscopy analysis. Journal of Food Engineering, 74, 402– 

409. 

Zheng, L.Y. & Zhu, J.F. (2003). Study of antimicrobial activity of 

chitosan with different molecular weight. Carbohydrate Polymers, 

54, 527–530. 

Supporting Information 

Additional supporting Information may be found in the 

online version of this article: 

Figure S1. Evolution of mesophilic bacteria plotted as 

a function of storage time for fresh pasta packaged in 

low-barrier bag under (a) passive and (b) active MAP 

and in high-barrier bag under (c) passive and (d) active 

MAP. The curves are the best fit of Eq. (1) to the 

experimental data. 

Figure S2. Evolution of mesophilic bacteria plotted as 

a function of storage time for fresh pasta packaged in 

high-barrier under (a) passive and (b) active MAP. The 

curves are the best fit of Eq. (1) to the experimental data. 

Please note: Wiley-Blackwell are not responsible for 

the content or functionality of any supporting information 

supplied by the authors. Any queries (other than 

missing material) should be directed to the corresponding 

author for the article. 


1552 


Comparison of hurdle treatments for buffalo meat 

Altaf H. Malik 1 * & Brahma Deo Sharma 

Division of Livestock Products Technology, Indian Veterinary Research Institute Izatnagar Bareilly UP 243122, India 

(Received 8 December 2009; Accepted in revised form 7 April 2010) 

Summary Three types of shelf stable buffalo meat chunks (SBM) were prepared using three types of infusion solution 

formulations. These contained varying percentages of glycerol, sodium chloride, propylene glycol and honey, 

in addition to sodium nitrite @ 100 ppm and sorbic acid @ 0.2%. Meat chunks were pasteurised at 80 °C 

for 20 min and then desorbed under refrigeration, drained and dried at 80 °C. The SBM processed with 

ISF-3 (glycerol 6.0, propylene glycol 1.00 and sodium chloride 6.00%) had significantly lower salt and 

moisture content and significantly higher protein and haem pigments than the other two products. The SBM 

processed with ISF-1 (glycerol 9.5, propylene glycol 1.5, sodium chloride 7.5 and honey 2.0%) had 

significantly lower water activity but significantly higher yield and protein solubility than other two products. 

The SBM processed with ISF-2 (glycerol 2.5, propylene glycol 2.5, sodium chloride 8.5 and honey 1.0%) had 

significantly higher salt content and Thiobarbituric acid values than the other two products. All the three 

types of products had very good sensory acceptability and satisfactory microbiological qualities. 

Keywords Chemical composition, chemical preservatives, cooking, drying, electrophoresis, foods of animal origin, honey, lipid oxidation, 

microbiology, organic acids. 

Introduction 

India ranks first in buffalo population (98 million) in the 

world accounting for more than 56% of the world 

buffalo population (Food and Agriculture Organisation 

2005). During the year 2004–05 out of India’s total meat 

export of 0.316 million metric tonnes buffalo meat 

accounted for 0.307 million metric tonnes which earned 

360.06 million US$ (DGCI&S 2006). Currently India 

produces 1.9 million metric tonnes of buffalo meat out 

of which 21% (0.399 million tonnes) is exported (Anon 

2008). Meat is an important national resource and has 

good domestic and export market. Meat being perishable 

commodity needs preservation till it is transported 

and distributed to retailers and then to actual consumers. 

Refrigeration is a common method of preservation 

of meat. However, refrigeration is an energy intensive 

and costly process which is not always practicable in a 

developing country like India having enormous geographical 

area with tropical climate. 

It, therefore, becomes imperative to find some cost 

effective alternative for preservation of meat. Hurdle 

*Correspondent: E-mail: altafhussain.dr@gmail.com 

1 Associate Professor ⁄ Senior Scientist, Division of Livestock Products 

Technology, Faculty of Veterinary Sciences & AH, Sher e Kashmir 

University of Agricultural Sciences and Technology of Kashmir 

Shuhama Alusteng, Srinagar, Kashmir, India 190006 

International Journal of Food Science and Technology 2010, 45, 1552–1563 

technology can be utilised to store and transport meat at 

ambient temperatures with good microbiological safety, 

sensory and nutritional properties and convenience at 

cheaper and affordable price. Hurdle technology, also 

referred to as combination preservation technique, uses 

the parameters which hinder microbial growth and 

enzymatic deterioration. The hurdles are selected based 

on the type of food, its natural micro flora, chemical 

composition and climatic conditions of handling and 

storage. The purpose is to disturb the homeostasis of 

microbes in order to render them inactive. The hurdles 

generally used are water activity, pH, redox potential, 

mild heat treatment, refrigeration, preservatives, competitive 

flora and high pressure etc. Hurdle technology 

has made it possible to devise some semi-moist, readyto-eat, 

stable, sound and convenient meat and meat 

products to meet the requirements of a special class of 

people like space scientists, mountaineers, and defence 

personnel especially as combat ration with light weight. 

It has been possible to produce various types of shelf 

stable and intermediate moisture meat products with 

hurdle technology. Some of the popular meats produced 

using this processing technology all over the world as 

per Leistner (1985) are raw ham, fermented sausage 

(Europe), Mortadella (Italy), Bruhdauerwurst 

(Germany), Gelderse Rook worst (Netherlands), Charque 

(Brazil), Beef Jerky (North America), Pemmican 

(North America), Biltong (South Africa), Kundi (West 

doi:10.1111/j.1365-2621.2010.02291.x 

Ó 2010 The Authors. Journal compilation Ó 2010 Institute of Food Science and Technology

Africa), Dendeng Giling (Indonesia) and Tsusousan 

(China). A review of literature on the subject reveals 

that technologists have prepared intermediate moisture 

meat products in various countries using sugar and salt 

desorption process at high concentrations. But the 

humectants at higher levels affect the flavour in finished 

product (Ledward, 1981). Recent studies by Prabhakar 

& Ramamurthi (1990) have also suggested that palatability 

of intermediate moisture meat products can be 

improved by variety of ways including using of low 

levels of preservatives and cooking with traditional 

Indian method with spices. 

In view of the above as well as growing need for shelf 

stable meat and meat products, it was proposed to 

prepare shelf stable intermediate moisture type buffalo 

meat chunks which could be stored and transported to 

long distances by rail, road, sea and air without 

refrigeration and could be cooked by consumers as per 

their convenience and taste. The buffalo meat being 

comparable to beef except for its leanness was chosen 

for its easy availability and economic importance in 

India. 


Meat source 

Buffalo meat required for the experiments was procured 

from a selected retail meat shop located in Bareilly meat 

market The meat samples strictly belonged to round 

(consisting mostly of semi-membranosus, semitendinosus, 

biceps, femoris and quadriceps muscles) of carcasses 

of almost similar conformation of spent adult female 

Murrah buffaloes slaughtered according to traditional 

Halal method. Boneless meat cuts of required weight 

were procured within 4–5 h of slaughter packed in low 

density polyethylene (LDPE) bags and brought to 

laboratory within 30 min. The meat chunks were kept 

for conditioning in a refrigerator at 4 ± 1 °C for about 

24 h till further use. 

Chemicals and ingredients 

Chemicals and ingredients used in the experiment were 

all food grade from standard concerns. 

Spice mix formulation 

The composition of spice mix formulation is given in 

Table 1. 

Preparation of meat chunks 

After chilling for 24 h separable fat and connective 

tissue were removed as far as possible and the chunks of 

approximately 2.5 · 1.25 · 1.25 cm size were cut. 

Table 1 Composition of spices mix 

Ingredients % by weight 

1. Cinnamon (Dalchini) 5.15 

2. Black pepper (Kalimirch) 12.40 

3. Nutmeg (Jaifal) 2.00 

4. Mace (Jawitri) 2.00 

5. Coriander powder (Dhania) 20.60 

6. Cloves (Laung) 2.00 

7. Cumin seeds (Zeera) 15.50 

8. Caraway seeds (Ajwain) 10.40 

9. Cardamon dry (Badi Elaichi) 5.15 

10. Bay leaves (Tejpat) 2.00 

11. Capsicum (Mirch powder) 10.40 

12. Aniseed (Saonf) 12.40 

100.00 

Standardisation of pasteurisation process 

Three hundred grams of meat chunks were dispersed in 

450 mL distilled water and heated in a water bath 

maintained at 75 °Cand80°C separately for a period of 

15 min and 20 min at each temperature. Samples were 

taken for total aerobic plate count to see the effect of 

pasteurisation. Three trials were conducted. The observations 

guided us to adopt pasteurisation at 80 °C for a 

period of 20 min. 

Preparation of infusion solution 

In preliminary trials different levels of humectants were 

tried e.g. glycerol from 2.5 to 20%, sodium chloride 6.0– 

12.0%, sucrose 2.0–6.0%, propylene glycol 1.0–2.5% 

with nitrite and sorbate as preservative. The pH of 

infusion solution was adjusted with lactic acid. Since the 

higher levels of glycerol and sodium chloride were not 

liked by the sensory panellists and hence the humectants 

were used at lower levels in three different combinations 

(Table 2). Formulation 1 was exclusively used for 

standardisation of drying and rehydration process. 

The ingredients were taken on per cent weight basis of 

infusion solution and dissolved in sterile potable water. 

Table 2 Formulation of infusion solutions 

Ingredients 

Hurdle treatments of Buffalo meat A. H. Malik and B. D. Sharma 1553 

Per cent by weight 

1 2 3 

Glycerol 9.50 2.50 6.00 

Propylene glycol 1.50 2.50 1.00 

Sodium chloride 7.50 8.50 6.00 

Honey 2.00 1.00 - 

Sodium nitrite 0.01 (100 ppm) 0.01 (100 ppm) 0.01 (100 ppm) 

Sorbic acid 0.20 0.20 0.20 

Water 79.29 85.29 86.79 

Total 100.00 100.00 100.00 


1554 

Hurdle treatments of Buffalo meat A. H. Malik and B. D. Sharma 

Desorption of meat chunks 

Meat chunks were desorbed in infusion solutions in 

sterile beakers. Meat to infusion solution ratio of 1:1.5 

by weight was followed. The pH of infusion solution 

containing meat chunks was recorded after 20 min of 

immersion and stirring. Meat chunks with infusion 

solution were pasteurised in a water bath maintained at 

80 °C for 20 min. 

Adjustment of pH 

After pasteurisation and subsequent cooling to room 

temperature the pH of the solution was adjusted to 5.0 

with the help of lactic acid before keeping it under 

refrigeration at 4 ± 1 °C to allow desorption process. 

After 24 h desorbed meat was drained for 10–15 min 

and taken in a clean tray. The surface moisture was 

wiped with blotting paper and samples were taken for 

moisture determination before proceeding for drying. 

Drying process 

To achieve moisture levels of approximately 50% and 

water activity (aw) below 0.94, Infusion Solution formulation 

1 (ISF1) was used to standardise the drying 

process at 80 °C. Desorbed meat 1500 g was loaded in 

trays in the oven preheated to 100 °C. Three time periods 

90, 120 and 150 min were tried for optimum drying. The 

meat chunks were taken out at above mentioned 

intervals, their surface wiped with blotting paper and 

allowed to cool and sampled for moisture and aw. 

Rehydration time 

The dried meat chunks were immersed in potable water 

equivalent to the weight of meat and subjected to 

occasional stirring, sampled at intervals of 90, 120 and 

150 min for moisture determination. 

Preparation of hurdle treated (shelf stable) meat chunks 

Three types of infusion solution formulations (Table 2) 

were used for the preparation of hurdle treated meat 

chunks as per the processing steps standardised earlier. 

Three trials were conducted. The products were analysed 

at appropriate stages for various physico-chemicals, 

sensory and microbiological parameters and results were 

statistically analysed. 

Cooking for sensory evaluation 

The dried hurdle treated meat chunks were rehydrated for 

2 h in potable water. The rehydrated meat chunks were 

pressure cooked for 35 min in the rehydration fluid itself 

and then fried in oil in traditional way with spices and 

condiments. Frozen meat was taken as control. The product 

was served to taste panellists for sensory evaluation. 

Physico-chemical analyses 

pH 

The pH was determined by combined glass electrode of a 

digital pH metre (Model CP901; Century India, Chandigarah, 

India) using the method of (Keller et al. (1974). 

Proximate composition and nitrite 

Methods of AOAC (1995) were followed. The nitrogen 

was estimated by using the Micro-kjeldahl distillation 

assembly and fat was estimated using the Soxhlet apparatus 

(Borosil Glass Works Ltd. Worli Mumbai, India) as 

ether extract. For nitrite the OD was determined at 

540 nm using the Beckman (Model DU 640) Spectrophotometer 

(Beckman Coulter Inc., Brea, CA, USA). 

Water activity (a w) 

The water activity (a w) of meat samples was determined 

by the procedure recommended by Lerici et al. (1983), 

with slight modification. 

Approximately 50 g minced sample was tightly 

packed inside a 60 mL glass tube and mouth was corked 

air-tight. Then it was immersed in a cooling chamber. 

The cooling chamber had precooled ethanol at )30 °C 

to )35 °C in one litre glass beaker and kept in deep 

freezer (Vertical type, Vestfrost, Denmark). Resistant 

thermometer probe (Century, CT809, S.No. 101) was 

introduced inside the cork to monitor the temperature of 

the product. The rate of decrease in temperature was 

monitored by looking at electronic digital display 

(Century) which was constant up to a specific point. 

After that the rate markedly decreased. The point at 

which the rate markedly altered was taken as freezing 

point of the sample. The freezing point was converted to 

a w value using )ln a w = 27.622–528.373 (1 ⁄ T)–4.579 

lnT where, T = freezing point temperature in Kelvin. 

Warner bratzler shear force value 

The objective texture measurement of shelf stable meat 

products was done using Warner Bratzler shear press 

(Model No. 81031307, G-R Manufacturing Co., Manhattan, 

KS, USA). Muscle cores with cross-sections of 

1.27 cm · 1.27 cm were prepared by cutting the meat 

pieces through their longitudinal axis. Maximum force 

(kg) required to shear meat cores ⁄ strips along the 

transverse axis was recorded and expressed as the force 

(kg) required to shear meat. 

Sodium chloride (salt) 

Methods described by (Koniecko (1979) were followed 

for sodium chloride. 


Thiobarbituric acid (TBA) value 

The method of Tarladgis et al. (1960) with slight 

modifications was followed. Exactly 10 g sample was 

blended with 49 mL distilled water and 1 mL of 

sulphanilamide reagent in a homogeniser. The mixture 

was quantitatively transferred into a Kjeldahl flask 

(Borosil Glass Works Ltd.). Another 48 mL of distilled 

water was used for rinsing the blender and poured into 

the flask followed by addition of 2 mL of hydrochloric 

acid solution (one volume with two volume of water). 

Few chips of paraffin wax were added and the flask 

heated at high heat and 50 mL of distillate collected into 

a graduated cylinder. The distillate was mixed well, 

5 mL was pipetted into test tubes to which 5 mL of 

Thiobarbituric acid (TBA) reagent was added and 

mixed. The tubes on a stand were immersed in boiling 

water bath for 35 min followed by 10 min cooling in tap 

water. The OD was read at 538 nm against reagent 

blank in a Beckman (Model DU 640) Spectrophotometer. 

The OD was multiplied by the factor 7.8 and results 

expressed as mg malonaldehyde ⁄ kg meat. 

Total haem pigments 

The method used by Hornsey (1956) was adopted for 

measurement of total pigments. The OD was recorded 

at 640 nm using the Beckman (Model DU 640) Spectrophotometer. 

Protein solubility 

The method used by Okonkwo et al. (1992) was 

adopted. Remi T8 Centrifuge was used for centrifugation 

at 5000 · g as per Madovi (1980). 

Sodium dodecyl sulphate polyacrylamide gel 

electrophoresis (SDS-PAGE) 

SDS-PAGE of meat extractable proteins was carried out 

as per the method of Laemmli (1970) with slight 

modification using 1 mm slab gel in Genei electrophoresis 

equipment (Genei Pvt. Ltd, Banglore, India) and 

power supply apparatus (Pharmacia American Instrument 

Exchange Inc., Skokie, IL, USA). 

Preparation of sample 

Buffalo meat samples obtained from semitendinosus 

muscle of spent adult female Murrah buffaloes, used in 

experiment 1 were taken for SDS-PAGE at different 

processing stages i.e. fresh, desorbed and dried muscle. 

Meat proteins were extracted from sample following the 

procedure of Syed et al. (1995). Minced meat (5 g) was 

mixed with 50 mL of 0.01 N sodium phosphate buffer 

(pH 7.0) containing 1% sodium dodecyl sulphate (SDS) 

plus 1% 2-mercaptoethanol and incubated at 37 °C for 

2 h. The mixture was then centrifuged at 4000 r.p.m. for 

30 min. An aliquot of supernatant was dialysed overnight 

at room temperature (26 ± 2 °C) against 0.1 N 

sodium phosphate buffer containing 0.1% SDS and 

0.1% 2-mercaptoethanol. About 50 lL–100 lL of 

dialysed. 

Microbiological quality 

All the microbiological parameters viz: Total Plate 

Count, Staphylococcus aureus, and Yeast and mould 

count were determined following the APHA (1984). For 

total plate count and Staphylococcus aureus plate count 

agar and Baird–Parker agar with potassium Tellurite of 

Himedia Laboratories private limited Mumbai at incubation 

temperature of 35 °C for 24 h and 48 h were used 

respectively. For Yeast and moulds Potato dextrose 

agar of Himedia and incubation temperature of 25 °C 

was used. 


A semi-trained experienced taste panel (Seman et al., 

1987) members consisting of scientists and postgraduate 

students of the Division and Institute evaluated the 

sensory attributes of buffalo meat products on the 

basis of sensory evaluation scoring guide and scoring 

card. The 8-point descriptive scale was used (wherein 

eight is extremely desirable and one is extremely 

undesirable). Cooked chunks were served warm to the 

panellists. 


The data obtained from the various trials was pooled 

and processed at the Institute’s Computer Centre Using 

SPSS software (SPSS Inc., Chicago, IL, USA) for 

obtaining mean ± SE. The data was subjected to 

analysis of variance, least square difference and critical 

difference (Snedecor & Cochran, 1967) for comparing 

the means to find the effects between treatments. 

Results 


Standardisation of processing steps 

The results are presented in Tables 3–5. 

Pasteurisation 

Pasteurisation at 75 °C reduced mean total aerobic plate 

count (TPC) from initial log 4.50 to log 2.41 and log 

1.90 in 15 and 20 min respectively, whereas repetition of 

same process at 80 °C resulted in counts being reduced 

to log 1.66 and log 0.88, respectively (Table 3). How- 


1556 


Table 3 Effect of time-temperature combination on total plate count 

(log10 cfu g )1 ) during standardisation of pasteurisation process for 

buffalo meat chunks at 75 °C and 80 °C 

Trial 

Initial 

numbers 

ever, no colonies were detected in two out of three trials 

on pasteurisation at 80 °C for 20 min. 

Drying of meat chunks 


75 °C 80 °C 

15 min 20 min 15 min 20 min 

1 5.47 3.77 3.42 3.54 2.65 

2 4.60 3.47 2.30 1.44 ND 

3 3.42 ND ND ND ND 

Mean 4.50 2.41 1.90 1.66 0.88 

ND = Not detected. 

Table 4 Changes in moisture (%) of buffalo meat chunks during 

standardisation of drying at 80 °C and rehydration at room temperature 

Initial 

Time 

1½ h 2 h 2½ h 3 h 

Drying 

65.81 ± 0.18 d 55.83 ± 0.20 c 53.81 ± 0.04 b 52.30 ± 0.20 a 

Rehydration 

- 

52.19 ± 0.20 a 64.48 ± 0.38 b 67.48 ± 0.38 c 

- 68.24 ± 0.18 c 

n =4. 

Means (±SE) bearing different superscripts in each row differ significantly 

(P < 0.05). 

anova indicated highly significant (P < 0.01) effect on 

moisture content due to drying time. 

Moisture 

Moisture content of desorbed meat chunks decreased 

significantly (P < 0.01) (Table 4) as drying time increased. 

Moisture content of 52.30% could be achieved 

after 2½ h drying at 80 °C. 

Water activity 

During standardisation of drying, the water activity (a w) 

values (not given in tables) of desorbed meat chunks 

(0.946) after 1½, 2 and 2½ h drying at 80 °C were 

recorded as 0.923, 0.919 and 0.904, respectively. On the 

basis of observations, time for drying was adopted as 

2½ h at 80 °C. 

Rehydration of buffalo meat chunks 

The anova indicated highly significant (P < 0.01) effect 

of time of rehydration on moisture content of rehydrated 

meat chunks. Moisture increased significantly from 

the initial value of dried meat chunks (Table 4) after 

1½, 2 and 3 h rehydration in potable water. However, 

there was no significant (P > 0.05) improvement in 

moisture content during 3 h rehydration over that of 2 h 

rehydration. 

Effect of different infusion solution formulations on the 

quality of Shelf stable buffalo meat chunks (SBM) during 

and after the production processes 

The results of pH, per cent moisture and salt at various 

stages of processing are presented in Tables 5 and 6 and 

the results of quality of SBM treated with three different 

ISFs are presented in Tables 7 and 8. 

Physico-chemical characteristics during processing 

pH 

The initial pH of infusion solution formulations was 

comparable. It is evident from Table 5 that pH of all 

infusion solution formulations increased after desorption 

of meat and subsequent pasteurisation till it was 

adjusted to 5.00 with lactic acid. Same trend was 

observed in pH of meat chunks treated with three 

infusion solution formulations after 24-h desorption and 

after drying for 2½ at 80 °C. The anova indicated that 

the infusion solution formulations (ISF) had no significant 

(P > 0.05) effect on pH at any stage of processing 

of buffalo meat chunks. 

Table 5 Effect of different infusion solution formulations on pH at various stages of processing of shelf-stable buffalo meat chunks (Mean ± SE) 

Processing stage Sample 

Infusion solution formulations 

1 2 3 

Before desorption of meat Infusion solution 3.89 ± 0.03 3.84 ± 0.03 3.87 ± 0.02 

After 20 min of desorbing meat Infusion solution with meat 4.76 ± 0.04 4.77 ± 0.06 4.83 ± 0.07 

After pasteurisation and cooling for 2 h Infusion solution with meat 5.26 ± 0.04 5.17 ± 0.03 5.24 ± 0.05 

After adjustment with lactic acid Infusion solution with meat 5.00 ± 0.00 5.00 ± 0.00 5.00 ± 0.00 

After desorption for 24 h Meat chunks 5.66 ± 0.04 5.59 ± 0.07 5.57 ± 0.07 

After drying at 80 °C for 2 h -do- 5.67 ± 0.06 5.64 ± 0.09 5.62 ± 0.05 

n =3. 

Mean initial pH of fresh meat before conditioning was 6.29 ± 0.09 and after conditioning 5.77 ± 0.07. 


Table 6 Effect of different infusion solution formulations on moisture 

and salt per cent at various stages of processing of shelf-stable buffalo 

meat chunks 

Processing 

stage 

Fresh 

meat 


1 2 3 

Moisture (%) 

Desorption 75.40 c 

65.61 a ± 0.21 67.65 b ± 0.32 68.39 b ± 0.34 

Drying 52.96 b ± 0.20 53.63 c ± 0.31 52.01 a ± 0.19 

Rehydration 68.15 b ± 0.17 65.58 a ± 0.39 65.32 a Salt (%) 

± 0.32 

Desorption 4.04 b ± 0.18 4.40 c ± 0.12 3.14 a ± 0.05 

Drying 5.10 b ± 0.10 5.55 c ± 0.07 4.42 a ± 0.06 

Rehydration 2.58 ab ± 0.18 2.74 b ± 0.11 2.24 a ± 0.08 

n =6. 


(P < 0.05). 

Moisture 

The anova indicated highly significant (P < 0.01) effect 

of infusion solution formulations on moisture content of 

meat chunks after desorption, drying and rehydration 

stages of processing (Table 6). Significant decrease of 

moisture was observed after desorption. The moisture 

was significantly lower (P < 0.01) in meat chunks 

processed with ISF-1. However, no significant difference 

was recorded in moisture per cent between the chunks 

processed with ISF-2 and ISF-3. After drying the 

moisture per cent in the entire three hurdle treated meat 

chunks was significantly different. 

After rehydration of meat chunks in water, the 

moisture content was significantly higher in meat chunks 

processed with ISF-1 than the other two ISFs while 

there was no significant difference in moisture content 

between the meat chunks processed with ISF2 and ISF3. 

Sodium chloride (common salt) 

The infusion solution formulations had highly significant 

(P < 0.01) effect on per cent salt content of meat 

chunks after desorption and drying processes and a 

significant (P < 0.05) effect after rehydration. 

Physico-chemical characteristics after processing 

The anova indicated a highly significant (P < 0.01) 

effect on almost all parameters and only a significant 

(P < 0.05) effect on nitrite content due to infusion 

solution formulations. 

Yield 

The buffalo meat chunks treated with ISF-1 had 

significantly higher yield (Table 7) than SBMs treated 

with ISF-2 and ISF-3 whereas the SBM treated with 

ISF-3 had significantly lower per cent yield than the 

other two SBMs. 

Water activity (a w) 

Water activity of SBM treated with ISF-1 was significantly 

lower than the aw of SBM treated with ISF-2 and 

ISF-3 while the aw of SBM treated with ISF-2 was 

significantly higher than others (Table 7). 

Proximate composition 

The per cent protein, fat and ash were significantly 

higher in all types of SBM (Table 7) than the fresh meat. 

Among the SBM the protein per cent of SBM treated 

with ISF-3 was significantly higher than the other two 

SBMs. While the protein per cent of SBM treated with 

ISF-1 was lowest among all the SBM. There was no 

significant difference in fat and ash contents among the 

SBM treated with three types of infusion solution 

formulations. 

Table 7 Effect of different infusion solution formulations on physico-chemical characteristics of shelf-stable buffalo meat chunks 

Parameters** Fresh meat ⁄ Control 



1 2 3 

Yield (%) - 60.13 c ± 0.18 57.65 b ± 0.47 55.55 a ± 0.58 

Water activity (aw) - 0.903 a ± 0.001 0.908 b ± 0.0006 0.913 c ± 0.0009 

Moisture (%) 75.402* ± 0.584 52.96 b ± 0.12 53.63 c ± 0.31 52.00 a ± 0.19 

Protein (%) 19.21 a ± 0.21 29.21 c ± 0.07 28.60 b ± 0.13 31.81 d ± 0.26 

Fat (%) 4.88 a ± 0.38 8.66 b ± 0.08 8.65 b ± 0.10 8.43 b ± 0.04 

Ash (%) 1.00 a ± 0.04 5.43 b ± 0.00 5.80 b ± 0.01 5.20 b ± 0.17 

Shear force value (kg ⁄ 1.27 cm) 3.16 a ± 0.16 3.76 b ± 0.13 5.69 c ± 0.14 5.49 c ± 0.11 

Salt (%) - 5.10 b ± 0.10 5.55 c ± 0.07 4.42 a ± 0.06 

Nitrite (ppm) - 20.68 b ± 0.42 19.48 a ± 0.28 19.67 a ± 0.25 

Protein solubility (%) - 82.91 b ± 0.29 80.07 a ± 0.38 80.58 a ± 0.53 

Total haem pigments (ppm) 198.57 a ± 5.76 345.60 b ± 5.06 360.88 b ± 2.17 389.88 c ± 13.95 

TBA (mg malonaldehyde ⁄ kg meat) 0.24 a ± 0.01 0.36 b ± 0.02 0.40 b ± 0.02 0.39 b ± 0.02 

*Not compared. 

**n = Shear force value 96, water activity and yield 3, others 6. Means (±SE) bearing different superscripts in each row differ significantly (P < 0.05). 


1558 


Table 8 Effect of different infusion solution formulations on sensory 

attributes of shelf-stable buffalo meat chunks 

Sensory 

attributes Control 


1 2 3 

Appearance 6.52 a ± 0.13 7.00 b ± 0.14 7.00 b ± 0.14 7.14 b ± 0.13 

Flavour 6.43 ± 0.16 6.33 ± 0.19 6.10 ± 0.33 6.67 ± 0.13 

Juiciness 6.48 ± 0.16 6.53 ± 0.18 6.24 ± 0.33 6.71 ± 0.14 

Texture 6.38 ± 0.18 6.48 ± 0.19 6.67 ± 0.16 6.85 ± 0.13 

Saltiness 6.57 ± 0.20 6.48 ± 0.20 6.52 ± 0.21 6.85 ± 0.18 

Overall 

palatability 

6.43 ± 0.15 6.29 ± 0.16 6.33 ± 0.13 6.76 ± 0.12 

n = 21. 


(P < 0.05). 

Warner bratzler 

Shear force value 

The values for shear force were significantly higher 

(P < 0.01) in all types of SBM than the fresh (control) 

meat chunks. Among the SBM the values were significantly 

lower for SBM treated with ISF-1 than the other 

two types of BM which in turn did not differ significantly 

between them. 

Nitrite 

Nitrite content was significantly higher in SBM treated 

with ISF-1 than either ISF1 or ISF2. The latter, however 

did not significantly differ from each other. 


Protein solubility values ranged between 80.07 and 

82.91. Solubility was significantly higher (P 0.05) difference between the control 

and three types of buffalo meat chunks or hurdle treated 

1 2 3 4 5 6 7 8 9 10 11 12 13 

Figure 1 SDS-PAGE of proteins extracted from buffalo semitendinosus 

muscle during different stages of processing of shelf stable meat 

chunks Lane 1: Fresh meat, Lanes 2 & 3: Desorbed meat chunks 

processed with ISF-1, Lanes 4 & 5: Dried meat chunks processed with 

ISF-1, Lanes 6 & 7: Desorbed meat chunks processed with ISF-2, 

Lanes 8 & 9: Dried meat chunks processed with ISF-2, Lanes 10 & 11: 

Desorbed meat chunks processed with ISF-3, Lanes 12 & 13: Dried 

meat chunks processed with ISF-3. 


meat chunks in flavour, juiciness, texture, saltiness and 

overall palatability. However, the scores for appearance 

(Table 8) were significantly lower (P < 0.01) in control 

than three types of SBMs which did not differ significantly 

among themselves. 


In all the three types of hurdle treated meat chunks, the 

growth of microbes viz. total plate counts, yeast and 

moulds and Staphylococcus aureus was neither detected 

in the meat chunks after desorption which included 

pasteurisation of meat chunks in infusion solution nor 

after completion of drying at 80 °C for 2½ h. 

Discussion 

The procedure for the development of shelf stable meat 

chunks was standardised and various control parameters 

were established in the preliminary studies, the 

details of which have been presented in materials and 

methods. The information on this type of shelf stable 

buffalo meat chunks is very scanty. However, an 

attempt has been made to compare, wherever possible, 

the present findings with the available literature on some 

other shelf stable meats. 

Standardisation of processing steps 


The pasteurisation at 80 °C for 20 min gave the desirable 

and superior results (Table 3) than other timetemperature 

combinations. After pasteurisation, total 

plate count of only log 0.88 was detected out of initial 

mean log of 4.5. In a large number of microbial 

population cells differ in heat resistance (Jay, 1986) 

besides some cells may become heat resistant due to 

excretion of protective substances. Further, some of the 

surviving colonies could be sub lethally damaged spores 

(Fernandez et al., 1994). This type of pasteurisation or 

mild heating for killing of vegetative cells during 

desorption of meat chunks has been also carried by 

Obanu et al. (1975), Webster et al. (1986) and Okonkwo 

et al. (1992). 

Formulation of infusion solutions 

Three infusion solutions were formulated which would 

give desired water activity while maintaining the 

texture and flavour of meat chunks. Most of the 

workers have used very high levels of glycerol to 

achieve desired water activity (Brockman, 1973; Obanu 

et al., 1975, 1976; Webster et al., 1986; Okonkwo et al., 

1992). Propylene glycol has been used at 0.3–5.0% 

level (Kaplow & Hallik, 1972; Muguruma et al., 1987) 

while sugars-sucrose or sorbitol have also been used at 

high levels to achieve the desired water activity. 


However, very high levels of glycerol may give sweet 

taste (Ledward, 1981) while propylene glycol may give 

bitter after taste (Bone, 1973). Sodium chloride at 

higher levels has palatability problem. It has physiological 

implications as well in consumers (Ledward, 

1981). Sugars will naturally give a sweet taste at higher 

levels. Conventional sucrose was replaced by 0–2% 

honey because honey contains fructose as sugar and 

has been reported to possess antibiotic property 

(Sheikh et al., 1995). The concentrations of sorbic acid 

and sodium nitrite were kept constant in all formulations 

so that effects due to combination of humectants 

could be evaluated. 

Drying of meat chunks 

The moisture (65.810) content (Table 4) and water 

activity (0.946) of meat chunks after desorption in 

infusion solution were still very high which could not 

ensure the stability of meat chunks. Hence, drying was 

taken up at 80 °C to reduce moisture to nearly 50% and 

water activity to

1560 


nen et al., 1970). Similar effect was observed in meat 

chunks, treated with three infusion solution formulations 

after drying for 2½ h at 80 °C. The shelf stable 

meat chunks did not differ significantly in the pH values 

because pH of the infusion solution containing meat 

chunks was adjusted with lactic acid uniformly to 5.0 in 

all the three formulations. The final pH of finished 

product ranged from 5.62 to 5.67 which is close to the 

range of 5.40–5.90 reported for such products by Chirife 

et al. (1979). 

Moisture 

The significant decrease in moisture content of meat 

chunks after desorption (Table 6) was a direct effect of 

solutes or humectants to equilibrate the water activity of 

the solution and the meat components. Higher concentration 

of solutes in ISF-1 caused significantly higher 

moisture loss in meat chunks than other two infusion 

solutions. 

The significantly higher moisture content in SBM 

processed with ISF-2 may be due to higher concentration 

of volatile propylene glycol (Sloan & Labuza, 

1975). After the SBM chunks were rehydrated in potable 

water at room temperature for 2 h there was increase in 

moisture content of all the three samples in direct 

proportion to the initial glycerol content (Table 6). 

Glycerol has been reported to maintain the muscle 

structure near to fresh meat and prevent protein 

denaturation (Muguruma et al., 1987) and thereby 

showing better rehydration. 

Sodium chloride (common salt) 

Desorbed meat chunks had absorbed salt in proportion 

to its concentration in three infusion solutions accounting 

for the significant differences among the meat 

chunks. Due to concentration of solutes in drying, the 

salt concentration further increased in the shelf stable 

buffalo meat chunks. However, after rehydration of 

SBM in potable water for 2 h before cooking, some of 

the salt leached out into the rehydrating fluid thereby 

decreasing the salt content (Table 6) of meat chunks to 

near normal seasoning level of 2–2.5%. Of course slight 

variation in salt level acceptability is common as a 

matter of personal preference (Pearson & Tauber, 1985). 

Salt levels in the products processed by the present 

technique were far less than reported by Prabhakar & 

Ramamurthi (1990). 

Effect of different infusion solution formulations on the 

quality of shelf stable buffalo meat chunks (SBM) 

Yield 

The significantly higher yield (Table 7) in the shelf stable 

buffalo meat chunks processed with ISF-1 may be 

possibly due to the diffusion of more solutes in this 

product as compared to others. 

Water activity 

The significantly lower water activity of SBM treated 

with ISF-1 than the other two may be due to the high 

solute concentration in the infusion solution 1 especially 

that of glycerol which has better water binding and 

water activity lowering properties. The significantly 

higher water activity in the SBM treated with ISF-3 

may be due to relatively less concentration of sodium 

chloride and propylene glycol. 

Proximate composition 

The proximate composition of fresh buffalo meat 

(Table 7) was similar to that reported for buffalo meat 

by Padda et al. (1986). Since moisture content decreased 

due to desorption and drying, concentration of constituents 

such as fat, protein and ash had a proportionate 

increase in the product. 

Shear force value 

Fresh meat had significantly lower shear force value 

than the shelf stable buffalo meat chunks (Table 7) 

which may perhaps be due to the effect of drying on 

SBM. Similar results were reported by Okonkwo et al. 

(1992). The shear force value of SBM processed with 

ISF-1 were significantly lower than the other two 

possibly because the former had more glycerol concentration 

which is reported to give soft texture (Ledward, 

1982; Muguruma et al., 1987; Okonkwo et al., 1992). 

Nitrite 

The residual nitrite content in all the samples was very 

low (19.48–20.68 ppm). This has been explained by 

Cassens et al. (1979) as 5–15% complexes with myoglobin, 

1–10% becomes free nitrate, 5–20% remains as free 

nitrite, 1–5% is evolved as gas and 26–50% forms 

sulphahydryl, lipid and protein complexes. Jay (1986) 

also reported that heating causes decrease in nitrite level. 

Though nitrite was added in equal quantity in all the 

infusion solution formulations, SBM processed with 

ISF-1 had significantly higher nitrite content. It is quite 

possible that this particular humectant combination 

might have helped in retention of more nitrite. 


It was also significantly higher in SBM processed with 

ISF-1 than the other two (Table 7). This may be due to 

its higher glycerol content than the other two as glycerol 

has been reported to give increased protein solubility in 

such meats (Okonkwo et al., 1992). The protein solubility 

values of 80.07–82.91 in this case are comparable 

with 79.0–82% reported for IMM beef by Webster et al. 

(1986) and 80% by Okonkwo et al. (1992). 

Total haem pigments 

The total meat pigments had significantly increased in 

all the SBM than the fresh meat (Table 7) due to loss of 


moisture during drying. These findings are in agreement 

with Okonkwo et al. (1992) in IMM beef. 

Thiobarbituric acid value 

The significant increase in TBA value in all the three 

SBM products from the fresh meat level may be due to 

several factors such as desorption heating (Yamauchi, 

1972; Dunlavy & Lamkey, 1995), prooxidant effect of 

salt (Torres et al., 1994) and low water activity (Torres 

et al., 1994). However, the values were still below the 

threshold limit of 1–2 mg kg )1 (Watts, 1962). 

Sodium dodecyl sulphate polyacrylamide gel 

electrophoresis (SDS-PAGE) 

Protein pattern (Fig. 1) of fresh meat (control) Lane 1 

indicates that fresh meat contains number of proteins 

varying in molecular weight ranges from smaller molecular 

weight to higher molecular weight. The small 

molecular weight range proteins might have been 

produced due to proteolytic cleavage of large proteins 

as no antiproteolytic agent was included during extraction 

of proteins from fresh meat. So, many bands 

appeared in the gel with Rf values from 0.080 to 0.863. 

The high molecular weight proteins seem to be myosin 

heavy chain, light chain etc. After desorption of meat 

chunks with infusion solution-1 (ISF-1) (Lane 2–3) no 

difference was observed between this and protein pattern 

of fresh meat (Lane 1). 

When the meat chunks processed with ISF-1 were 

dried at 80 °C for 2½ h (Lane 4–5) some of the medium 

and low molecular weight proteins disappeared. The loss 

of these proteins might be due to coagulation or other 

chemical changes produced by heating. 

In proteins extracted from meat chunks desorbed in 

ISF-2 (Lane 6–7) few bands were not observed as 

compared to bands in Lane 2–3 (for meat desorbed in 

ISF-1) and Lane 1 representing fresh meat. This might 

be perhaps due to the comparatively higher concentrations 

of sodium chloride and very low concentrations of 

glycerol in this infusion solution as glycerol has been 

reported to give protection to the muscle protein against 

denaturation and proteolysis and thus maintaining 

fresh meat like characteristics (Muguruma et al., 

1987). When meat chunks processed with ISF-2 were 

subjected to drying at 80 °C for 2½ h, there was not 

much difference (Lane 8–9) in protein pattern from its 

desorbed stage (Lane 6–7) except for the disappearance 

of few low molecular weight bands. In comparison to 

protein pattern of dried meat processed with ISF-1 

(Lane 4–5), there was not much loss of protein bands in 

this case. 

Protein pattern of meat desorbed in ISF-3 (Lane 10– 

11) shows some less bands especially in the upper half of 

gel as compared to protein pattern obtained at similar 

stage in meat processed with ISF-1 (Lane 2–3) and ISF- 


2 (Lane 6–7). This difference may be due to difference in 

composition of infusion solutions. After drying of meat 

chunks processed with ISF-3, there was some loss of 

protein bands (Lane 12–13) comparable to similar stage 

in meat processed with ISF-2 (Lane 8–9) except that few 

bands of low molecular weight which disappeared in 

case of meat processed with ISF-2 were present in the 

former suggesting that processing with ISF-3 had better 

stability at drying than ISF-1 and ISF-2. 

Sensory attributes 

Sensory evaluation of shelf stable buffalo meat chunks 

did not differ significantly (Table 8) either among 

themselves or with the control cooked in the same 

ways. This shows all the SBM had almost equal 

acceptability. The significantly higher sensory scores of 

appearance (very good to excellent) of all the SBM can 

be attributed to their cured meat colour. In general, 

sensory scores of SBM processed with ISF-3 were 

slightly higher (Table 8) than others. Some panellists 

also recorded sweet taste in SBM processed with other 

two infusion solution formulations. In fact, sensory 

scores for flavour, juiciness, texture, saltiness and overall 

palatability of control as well as all the SBM were 

between moderate to very desirable. 


The absence of total plate count, yeast and moulds and 

Staphylococcus aureus after desorption (including pasteurisation) 

and drying indicates that combined effect of 

low water activity, preservatives, low pH and mild heat 

treatment (pasteurisation) had killed the vegetative cells 

and inactivated or caused sub lethal damage to the 

spores. Drying at 80 °C for 2½ h might have further 

damaged the spores resulting in absence of microbial 

growth below detection level. This sub lethal damage of 

spores (Leistner, 1985) and Perigo factor might have 

produced a combined effect (Perigo & Roberts, 1968). 

Conclusions 

With the application of hurdle technology, it was 

possible to prepare shelf stable buffalo meat chunks 

with good to very good acceptability The pasteurisation 

⁄ mild heating of meat chunks in infusion solution for 

20 min at 80 °C was enough to give a hurdle effect to 

achieve a product with good microbiological quality. A 

drying process of 2½ h at 80 °C was enough to give a 

product with water activity 0.90–0.91 and moisture 

52.0%. Before pressure cooking, a rehydration process 

of 2 h in potable water was found enough to equate 

these chunks to near fresh meat status with respect 

to sensory attributes. The products shall have the 

practical application as no refrigeration is required for 


1562 


transportation and storage and therefore can be used in 

areas of no electricity facilities. The storage studies shall 

be dealt in a separate paper. 

Acknowledgment 

The authors duly acknowledge the financial and technical 

support of the Director Indaian Veterinary 

Research Institute Izatnagar Bareilly India 

References 

Anon (2008). Ministry of Food Processing Annual report 2007–2008 

Government of India Source. http:www.indian business.nic.in/ 

industry-infrastructure/industrial – sectors/food-process.htm (last 

accessed 3 July 2009) 

AOAC (1995). Official Methods of Analysis. 17th edn. Washington, 

DC: Association of Analytical Chemists. 

APHA (1984). Microbiological tests for sanitation of equipment. In: 

Compendium of Methods for the Microbiological Examination of 

Foods 2nd edn. (edited by M.L. Speck). Pp. 310–330. Washington, 

DC: Am. Pub. Health Assoc. 

Bone, D. (1973). Water activity in intermediate moisture foods. Food 

Technology, 27, 71. 

Brockman, M.C. (1973). Intermediate Moisture Foods. In: Food 

dehydration (edited by W.B. Van Arsdel, M.J. Copley & A.I. 

Morgan Jr). Pp. 489–505. Westport. Connecticut: The AVI publishing 

Co. Vol. 2. 

Cassens, R.G., Greaser, M.L., Ito, T. & Lee, M. (1979). Reactions of 

nitrite in meat. Food Technology, 33, 46–57. 

Chirife, J., Scorza, O.C., Maria, S., Vigo, M.H., Bertoni, M.H. & 

Caltaneo, P.C. (1979). Preliminary studies on the storage stability of 

intermediate moisture beef formulated with various water binding 

agents. Journal of Food Technology, 14, 421–428. 

Cook, D.J. (1975). Intermediate moisture foods. Nutrition and Food 

Science, 40, 20–21. 

DGCI&S (Directorate General of Commercial Intelligence and 

Statistics) Ministry of Commerce and Industries Government of 

India (2006). Export of meat and meat products. In: Compendium of 

National Workshop cum Seminar on Commercial Goat and Sheep 

(edited by M.K. Agnihotri). Pp. 127–130. DGCI&S. Cited by 

Nayyar, S. S (2006). Farmer- Industry –Researcher Interface. Held 

on 4–5 March 2006 at CIRG Makhdoom UP India. Published by 

CIRG Makhdoom (ICAR), U.P, India. 

Dunlavy, K.A. & Lamkey, J.W. (1995). Dextrose level and oven 

temperature effects on warmed-over flavour development in beef top 

round roasts. Journal of Muscle Foods, 6, 63 [Food Science and 

Technology Abstracts, 27,5S 57]. 

Dymsza, H.S. & Silverman, G. (1979). Improving the acceptability of 

intermediate moisture fish. Food Technology, 33, 52–53. 

Fernandez, S.J., Gomez, R. & Carmona, M.A. (1994). Water activity 

of Spanish intermediate moisture meat products. Meat Science, 38, 

341–346. 

Food and Agriculture Organisation (2005). Food and Agriculture 

Organisation Production Data Year Book for the Year 2006. Pp. 1– 

39. Published by FAO United Nations, NHBS Environment book 

studio, FAO Rome, Italy. Food and Agriculture Organisation. 

Hornsey, H.C. (1956). The colour of cured pork. 1. Estimation of nitric 

oxide-haem pigments. Journal of the Science of Food and Agriculture, 

7, 54–58. 

Jay, J.M. (1986). Modern Food Microbiology. 3rd edn. Pp. 41, 69, 437– 

453. Delhi: CBS Publishers & Distributors. 

Kaplow, M. & Hallik, J. (1972). Determination of moisture content in 

Glycerol-containing Intermediate Moisture Foods. U.S.Patent: 

3,694,233. Sept. 26. 

Keller, J.E., Skelly, G.C. & Acton, J.C. (1974). Effect of meat particle 

size and casing diameter changes on summer sausage properties 

during drying. Journal of Milk and Food Technology, 37, 297–300. 

Koniecko, E.S. (1979). Handbook of meat chemists. Pp. 20–45. Wayne, 

New Jersey, U.S.A.: Avery Publishing Group Inc. 

Laakhonen, E., Wellington, G.H. & Sherbon, J.W. (1970). Low 

temperature, long time beating of bovine muscle. 1. Changes in 

tenderness, water binding capacity, pH and amount of water soluble 

components. Journal of Food Science, 35, 175–177. 

Laemmli, U.K. (1970). Cleavage of structural proteins during the 

assembly of the head of bacteriophage T4. Nature, 227, 680–685. 

Ledward, D.A. (1981). Intermediate moisture meats. In: Developments 

in Meat Science 2nd edn. (edited by R. Lawrie). Pp. 159–194. 

London: Applied Science Publishers. 

Leistner, L. (1985). Hurdle technology applied to meat products of the 

shelf stable product and intermediate moisture food types. In: 

Properties of Water in Foods (edited by D. Simatos & J.L. Multon). 

Pp. 309–330. Boston: Martinus Nijhoff Publishers. 

Lerici, C.R., Piva, M. & Rosa, M.D. (1983). Water activity and 

freezing point depression of aqueous solutions and liquid foods. 

Journal of Food Science, 48, 1667–1669. 

Madovi, P.B. (1980). Changes in the free NH 2-free CO 2H and 

titratable acidity of meat proteins. Journal of Food Technology, 15, 

311–318. 

Muguruma, M., Nishimura, T., Umetsu, Raizaburo., Goto, I. & 

Yamaguchi, M. (1987). Humectants improve myosin extractability 

and water activity of raw cured intermediate moisture meats. Meat 

Science, 20, 179–194. 

Obanu, Z.A., Ledward, D.A. & Lawrie, R.A. (1975). The protein of 

intermediate moisture meat stored at tropical temperature: changes 

in solubility and electrophoretic pattern. Journal of Food Technology, 

10, 657–666. 

Obanu, Z.A., Biggin, R.J., Neale, P.J., Ledward, D.A. & Lawrie, R.A. 

(1976). The protein of intermediate moisture meat stored at tropical 

temperature: nutritional quality. Journal of Food Technology, 11, 

575–580. 

Okonkwo, T.M., Obanu, Z.A. & Ledward, D.A. (1992). Characteristics 

of some intermediate moisture smoked meats. Meat Science, 

31, 135–145. 

Padda, G.S., Sharma, B.D. & Sharma, N. (1986). Quality characteristics 

of buffalo meat. Buffalo Bulletin, 5, 27–29. 

Pearson, A.M. & Tauber, F.W. (1985). Ingredients utilised in meat 

curing. In: Processed Meats (edited by A.M. Pearson & F.W. 

Tauber). Pp. 46–54. Westport, Connecticut: AVI Publishing Co. Inc. 

Perigo, J.A. & Roberts, T.A. (1968). Inhibition of Clostridia by nitrite. 

Journal of Food Technology, 3, 91–94. 

Prabhakar, K. & Ramamurthi, R. (1990). A study on the preparation 

of intermediate moisture meat. Journal of Food Science and 

Technology, 27, 162–164. 

Seman, D.L., Moody, W.G., Fox, J.D. & Gay, N. (1987). Influence of 

hot and cold boning on the palatability, textural and economic traits 

of restructured meat. Journal of Food Science, 52, 879–880. 

Shaikh, D., Shams-Uz-Zaman, S., Naqvi, B.S., Sheikh, R.M. & 

Ghulam, Ali. (1995). Studies on microbial activity of honey. 

Pakistan Journal of Pharmaceutical Sciences, 8, 51–62. 

Sloan, A.E. & Labuza, T.P. (1975). Investigating alternative humectants 

for use in foods. Food Product Development, 9, 75–88. 

Snedecor, G.W. & Cochran, W.G. (1967). Statistical Methods. 6th edn. 

Ames, Iowa: The Iowa State University Press. 

Syed, Z.K., Rao, D.N. & Ambla, B.L. (1995). Effect of lactic acid, 

ginger extract and sodium chloride on electrophoretic pattern of 

buffalo muscle proteins. Journal of Food Science and Technology, 32, 

224–227. 

Tarladgis, B.G., Watts, B.M., Younathan, M.T. & Duga, L. Jr (1960). 

A distillation method for the qualitative determination of malonaldehyde 

in rancid foods. Journal of the American Oil Chemists’ 

Society, 37, 44–48. 


Torres, E.A.F.S., Shimokomaki, M., Franco, B.D.G.M. & Landgraf, 

M. (1994). Parameters determining the quality of charqui, an 

intermediate moisture meat product. Meat Science, 38, 229–234. 

Watts, B.M. (1962). Meat products. In: Symposium on Food Lipids and 

Their Oxidation (edited by H.W. Shultz, E.A. Day & R.O. 

Sinnbubber). Pp. 202. Westport, CT: AVI Publishing Co. Inc. 

Webster, C.E.M., Allison, S.E., Adelakun, I.O., Obanu, Z.A. & 

Ledward, D.A. (1986). Reactivity of sorbate and glycerol in 


intermediate moisture meat products. Food Chemistry, 21, 133– 

143. 

Yamauchi, K. (1972). Effect of heat treatment on the development of 

oxidative rancidity in meat and its isolated tissue fraction. Bulletin of 

the Faculty of Agriculture of Miyazaki University, 19, 147 [Food 

Science and Technology Abstracts] 


1564 


Performance of different drying methods and their effects on the 

chemical quality attributes of raw cocoa material 

Tagro Simplice Guehi, 1 * Irié Bi Zahouli, 2 Louis Ban-Koffi, 2 Monké Adrien Fae 1 & Jean Gnopo Nemlin 2 

1 Unité de Formation et de Recherche des Sciences et Technologies des Aliments (UFR-STA), Université d’Abobo-Adjamé. 02 Bp 801 Abidjan 02 

Coˆ te d’Ivoire 

2 Station de Recherche et de Technologie, Centre National de Recherche Agronomique (SRT-CNRA). 08 BP 33 Abidjan 08 – Coˆ te d’Ivoire 

(Received 13 January 2010; Accepted in revised form 27 April 2010) 

Summary Studies were carried out to investigate the impact of different drying processes on the chemical quality traits 

of raw cocoa beans. The pH of less fermented cocoa is higher than the well-fermented cocoa’s. The sun-dried 

beans pH ranged from 4.5 to 5.5, while the pH of both oven- and mixed-dried beans was between 3.8 and 5.2. 

The sun-dried beans contained lower volatile acidity than oven-dried beans. Artificially dried beans resulted 

in higher free acidity content when compared to both sun- and mixed-dried beans. Ammonium nitrogen 

content in raw cocoa beans is not influenced by the drying methods. Free fatty acid content increases slowly 

but remains below the critical value of 1.75% whatever the drying processes. While oven-dried beans show 

the FFA content above 0.70% both of sun- and mixed-dried beans are associated with FFA content below 

0.70%. 

Keywords Chemical quality, cocoa, drying methods. 

Introduction 

Cocoa beans are the seeds from fruit pods of a tropical 

tree botanically known as Theobroma cacao L. (Family 

Sterculiaceae). Each pod contains 30–40 beans embedded 

in a mass of mucilaginous pulp within the pod. The 

bean pulp is rich in fermentable sugars, such as glucose, 

fructose and sucrose, and has a low pH of 3.0–3.5, 

mainly because of the presence of citric acid. Cocoa 

bean is the principal raw material of chocolate manufacture 

(Ardhana & Fleet, 2003). Theobroma cacao is 

grown mostly in the wet tropical forest climate which is 

within 20° of latitude of the equator at countries such as 

Coˆ te d’Ivoire, Ghana, Nigeria, Cameroon, Brazil, 

Equador, Papua New Guinea, Indonesia and Malaysia 

(Beckett, 1994). World production of cocoa beans was 

about 3 888 000 t in 2006 ⁄ 07 crop year, and nearly 70% 

of this quantity was produced in West Africa (World 

Cocoa Foundation, 2008). Cocoa is a crop that needs to 

be fermented and dried before export. The processing of 

cocoa beans consists of two major steps namely 

fermentation and drying (Wood & Lass, 1985). Fermentation 

and drying are both essential steps for the 

quality of final product. Upon harvesting of ripe cocoa 

*Correspondent: Fax: +22520374300; 

e-mail: g_tagro@hotmail.com 


pods, the beans and associated pulp are removed from 

the pod and subject to microbial fermentation as the 

first stage in preparation for chocolate production. 

Fresh cocoa beans are fermented for 5–7 days and dried 

immediately after fermentation to safe moisture level of 

7.5%. The importance of bean fermentation in contributing 

to chocolate quality has been recognised for over 

100 years, and numerous studies have been conducted in 

different countries to determine the microbial species 

associated with this process (Schwan et al., 1995). These 

fermentations are generally conducted as traditional, 

indigenous processes, the details of which have been well 

reviewed (Roelofsen, 1958; Lehrian & Patterson, 1983; 

Lopez & Dimick, 1995; Thompson et al., 2001). Microbial 

action during fermentation solubilises the pulp 

material surrounding the beans and produces a range of 

metabolic end-products (e.g. alcohols, organic acids) 

which diffuse into the beans to cause their death. These 

changes induce an array of biochemical reactions within 

the beans and generate the chemical precursors of 

chocolate flavour, aroma and colour (Lehrian & Patterson, 

1983; Jones & Jones, 1984; Hansen et al., 1998; 

Hashim et al., 1998; Thompson et al., 2001). During 

fermentation the temperature of the beans will rise from 

ambient to about 50–55 °C because of the exothermic 

oxidation reaction (Wood & Lass, 1985). After the end 

of fermentation, the moisture content of the whole 

doi:10.1111/j.1365-2621.2010.02302.x 

Ó 2010 CNRA-Université d’Abobo-Adjamé

eans is approximately 60% and this must be reduced to 

7–7.5% before the cocoa can be stored, sold, transported 

and exported to European countries. So the 

beans are dried immediately to avoid over fermentation, 

which could lead to product deterioration. Drying is 

usually carried out using natural sun and artificial hot 

air techniques (Mc Donald et al., 1981). Natural or 

artificial drying methods may be chosen, depending on 

characteristics of each species, the amount of harvested 

seeds, and on weather conditions prevailing after seeds 

were harvested. Natural cocoa bean drying is directly 

dependent on weather conditions. The disadvantage of 

the method lies in the need for intensive human labour 

and in turn leads to poor operational performance. 

Cocoa smallholders produce in small quantity would 

prefer sun drying, while for the bigger plantation the hot 

air (artificial) method is preferred (Wood & Lass, 1985). 

During drying, the cocoa beans undergo various chemical 

and biochemical changes that form the necessary 

flavour and aroma precursors needed during processing. 

Cocoa beans constitute an inexpensive fat source and 

are the principal raw material of chocolate from Africa 

and both Central and South America (Tafuri et al., 

2004). Nowadays, one of the most widespread concerns 

in advanced technological countries is food quality and 

safety. The economy of most developing countries, 

based primarily on their agricultural resources, is 

strongly dependent on the often rigorous and rigid 

quality standards set by developed countries. Ivorian 

raw cocoa bean quality did not escape being degraded as 

the liberalisation of the cocoa-producing chain in 1999 

and the reasons were unspecified (DGTCP (Direction 

Geńe´rale du Tre´sor et de la Comptabilité Publique de 

Coˆ te d’Ivoire), 2004). Indeed, as this liberalisation and 

dislocation of the Ivorian traditional control quality and 

the cocoa producers framing system, agricultural practices 

varied on the region and sometimes on the 

producers in the same area. Thus, in Coˆ te d’Ivoire, it 

is difficult to talk about cocoa post harvest treatments. 

However, Guehi et al. (2008) have identified the 

common practices about fermentation and drying 

methods in Coˆ te d’Ivoire. The predominant processing 

of fermentation for Ivorian raw cocoa production 

consists of ferment beans in heaps in small farms or in 

wooden boxes in big farms without turning. In Coˆ te 

d’Ivoire, cocoa fermentation usually lasts between 4 and 

5 days on weather conditions and time during the cocoa 

season. Fermentation generally takes shorter at the start 

and peak of the cocoa crop but longer towards the end 

of the crop when there is less mucilage available for 

fermentation. At the end of fermentation, Ivorian cocoa 

producers spread freshly fermented beans on a meshed 

wooden tray with area about 30–90 cm and raised 1 m 

above ground level, mats, polypropylene sheets or the 

concrete floor of a cocoa house each day to a depth of 

not less than 5 cm and mixed constantly to promote 

Performance of different drying methods T. S. Guehi et al. 1565 

uniform drying and to break agglomerates. Although 

sun drying is the preferred method for Ivorian producers, 

since some years artificial drying method is more 

and more employed in the big cocoa farms. Probably 

such variations in agricultural practices in Ivorian cocoa 

chain could be explained why the quality of Ivorian raw 

cocoa is more and more degraded. Therefore, it is 

important to identify the factors that reduce commercial 

value by studying the chemical quality of Ivorian beans 

resulted from different durations of fermentation and 

dried by different drying processes. The objective of this 

study did not focus on postharvest handling and 

technology processing of cocoa beans at the farmer’s 

level in Coˆ te d’Ivoire. 

This study aimed to evaluate the performance of some 

cocoa drying methods in terms of chemical quality of 

raw cocoa beans such as acidity (pH, free and volatile 

acidities, ammonium nitrogen contents and free fatty 

acids content) during fermentation. 


Cocoa 

The ripe cocoa pods (Theobroma cacao L.) of mixedhybrids 

were harvested by hand during the big 2005 

cocoa season (from December to February) in the 

experimental cocoa station of Centre National de 

Recherche Agronomique (CNRA) located at Bingerville 

region of Coˆ te d’Ivoire, a moderate hot rainy region 

with an average of 28–29 °C during the harvest season, a 

low altitude, below 500 m, 70–80 mm ⁄ month rainfall. 

2.2 Cocoa pod storage and breaking 

The cocoa pod storage time is 4 days at ambient 

temperature at the field. The pod storage time is the 

time that the pods were stored after harvesting but 

before breaking using a piece of wood billet as a 

bludgeon as reported earlier (Meyer et al., 1989). Pod 

breaking using wooden billet involves one or two sharp 

blows with the edge of the billet. The distal portion of 

the pod falls away and the beans remain attached to the 

placenta from which they can be easily extracted. The 

beans were removed from placenta being careful to 

exclude any germinated, black or diseased beans or 

pieces of shell or placenta fragments. 

Cocoa fermentation 

Cocoa beans were fractioned into seven same subsamples 

(about 25 kg). One type of fermentation was 

studied as reported earlier by Mounjouenpou et al. 

(2008): six boxes fermentation, where the beans of each 

fraction were put in banana leaves placed in wooden 

boxes measuring 35 · 35 · 35 cm 3 . Fermentation was 

Ó 2010 CNRA-Université d’Abobo-Adjamé International Journal of Food Science and Technology 2010, 45, 1564–1571

1566 

Performance of different drying methods T. S. Guehi et al. 

carried out in each box using 25 kg of fresh beans during 

1, 2, 3, 4, 5 and 6 days respectively for the boxes no. 1, 2, 

3, 4, 5 and 6. The heap of wet cocoa beans was then 

covered in the box with other fresh banana leaves and 

other banana leaves were used to insulate the top of 

box. Initial turning during fermentation was done after 

48 h and additional turning 48 h thereafter to facilitate 

adequate aeration of fermenting mass and to ensure that 

beans from the top and bottom are thoroughly mixed 

together. A fresh layer of banana leaves were added to 

the original leaves after each turning to ensure adequate 

insulation. Fermentation experimentations were 

conducted triplicate. 

Drying methods 

After fermentation, cocoa beans were fractioned into 

three same parts. Fermented beans of each fraction were 

dried according to a specific method. All drying methods 

processing is stopped when the moisture content of 

cocoa beans reached 7–8%. 

Sun drying methods 

Natural or sun drying process consistent to expose 

cocoa beans from 9 am until to 6 pm (C9) is considered 

as standard drying method. So characteristics of dried 

beans resulted from both other experiments were compared 

against this sun-dried beans. Two batches, one of 

whom is respectively unfermented and fermented beans 

were separately spread thinly on a meshed wooden tray 

with area about 30–90 cm and raised 1 m above ground 

level and sun-dried until they reached moisture content 

of about 7%. The beans were mixed each 1 h to ensure 

uniformity. 

Artificial hot air drying methods 

Raw cocoa beans were artificially dried using an airventilated 

oven at temperature of 60 °C (EV34) until 

moisture content of 7% as reported (Hii et al., 2009). 

The beans were spread thinly in single layer (about 

2 cm thick) on a meshed sample tray with square 

openings. Heat was generated by the heater integrated 

into the side walls of the oven and the hot air flowed 

through the samples. The exhaust air escaped through a 

ventilation hole (diameter 4 cm) at the back of the 

oven. The beans were mixed every 2 h to ensure 

uniformity. Drying was conducted for 8 h daily, and 

the beans were left to temper at room temperature 

overnight. The tempering step is a common routine in 

cocoa drying, and the purpose is to redistribute the 

internal moisture to the outer beans layer after each 

drying cycle. 

Combination of sun and artificial drying process 

The unfermented and fermented cocoa beans were dried 

by a mixed drying process consisting primarily sun 

drying from 9 am to 6 pm daily until moisture of 25% 

and consecutively by artificial drying process using an 

air-ventilated oven (C9 EV18) until moisture content of 

7%. 

Analytical methods 

Moisture content 

The beans used in each experiment were weighed prior 

to mixing during drying by using an analytical balance. 

The moisture content (%) of the beans was determined 

with reference to the dry weight of the beans as early 

reported (Guehi et al., 2007). The measurement was 

performed in triplicates. 

Chemical acidities and pH 

The methods used to quantify, volatile and free 

acidity, and ammonium nitrogen contents, and to 

determine pH were described by Pontillon & Cros 

(1998). Five grams of ground nibs was homogenised in 

45 ml boiled distilled water. The mixture was filtered 

with Whatman No. 4 filter paper and cooled to 

20–25 °C. The resulting filtrate was measured for pH 

using a pH meter (Consort P 107) which had been 

calibrated with buffers at pH 4 and 7 as described by 

Hii et al., 2009). A further 25-ml aliquot was titrated 

to an end point pH of 8.1 with 0.01 N solution of 

NaOH. Titratable acidity was calculated using the 

formula proposed by Hamid and Lopez (2000). The 

values reported as meq of sodium hydroxide per 1 g of 

dry nibs. Acidities were measured in triplicates to 

check the good fermentation and conservation of the 

samples. 

Free fatty acids content. 

About 15 g of dried cocoa beans were carefully shelled 

manually. Cocoa nibs were frozen in liquid nitrogen 

before finely grinding in a kitchen-scale coffee grinding 

(Moulinex, France) to the smallest particle (size 

< 500 lm). Ten grams of cocoa powder was put in 

Whatman cartridge and soaked in 350 ml of petroleum 

ether (Prolabo Normapur, type 40–60 °C) for one night. 

Cocoa butter was extracted on a Soxhlet apparatus for 

8 h. After eliminating the solvent in a rotary evaporator, 

FFA contents were quantified in triplicates by the 

official method 42-1993 (IOCCC, 1996) as reported 

earlier (Guehi et al., 2008). 

Statistical methods 

All analytical experiments treatments were conducted in 

three replicates. The experimental data were analysed by 

using one-way anova using the SAS software GLM 

procedure (SAS Institute, 2002) and mean comparison 

using the Newman–Keuls test at 95% confidence level 

(P < 0.05). 

International Journal of Food Science and Technology 2010, 45, 1564–1571 Ó 2010 CNRA-Université d’Abobo-Adjamé

pH 

6 

5 

4 

3 

2 

1 

0 

Day 0 Day 2 Day 4 Day 6 


Duration of fermentation (days) 

C9 EV34 C9EV18 

Figure 1 Effect of drying methods on pH of raw cocoa beans. C9: sun 

drying method EV34: artificial hot air drying method, C9EV18: 

combination of sun and artificial drying methods. 

Effect of drying methods on pH of raw cocoa material 

Figure 1 illustrates the influence of drying processes on 

the pH of raw cocoa beans. In term of pH, whatever the 

duration of fermentation process, comparison was also 

made against the sun-dried beans obtained from other 

experiments. For fresh beans (0 day fermentation) and 

cocoa beans fermented during 2 and 4 days, the pH of 

mixed-dried beans is not significantly different from that 

of sun-dried beans except the day 6 fermented beans. 

The sun-dried beans pH ranged from 4.5 to 5.5, while 

for the oven- and mixed-dried beans registered pH was 

comprised between 3.8 and 5.2. The differences may be 

explained by several factors such as exposure times to 

drying air, drying air temperature, relative humidity in 

the drying site, nature of drying air flow and the speed 

with which moisture migrated from the inner cocoa 

beans structures to their surface as previously indicated 

by Franke et al. (2008). Indeed artificial and mixed 

drying processes are faster with little time needed for the 

cocoa beans to have their moisture reduced from about 

60% to around 8% than the natural drying process. So 

the pH value for sun-dried beans is usually higher (less 

acidic) than artificially and mixed-dried beans because 

of the slow and gentle drying process that enable the 

evaporation of more acetic acid (Hii et al., 2009). The 

lower pH of unfermented cocoa beans could be 

explained by the bean pulp which has a low pH of 

3.0–3.5, mainly because of the presence of citric acid 

(Ardhana & Fleet, 2003). Generally, the pH of fermented 

beans falls within the values reported for most 

dried cocoa beans. Indeed these results were not 

significantly different (P < 0.05) from the pH values 

of both best fermented beans sourced from West Africa 

which is around 5.5 (Franke et al., 2008) and the 

standard Malaysian estate beans, which is about 4.4–4.7 

(Nazaruddin et al., 2006). In conclusion, the pH of 

Ivorian cocoa beans resulted from each treatment was 

found to be acid as well as all current raw cocoa beans 

whatever the duration of fermentation. 

Effect of drying methods on volatile acidity of raw cocoa 

material 

The impact of drying processes on the volatile acidity of 

raw cocoa beans is shown in Fig. 2. The results show 

that the changes in volatile acidity content of cocoa 

beans depend on drying methods. Indeed sun-dried 

beans contain lower volatile acidity than oven-dried 

beans for both fresh beans and fermented beans. Solardried 

beans present a volatile content below 1.0 meq of 

NaOH ⁄ g while that of oven-dried beans ranged from 

1.0 to 2.4 meq of NaOH ⁄ g whatever the duration of 

fermentation. For the beans fermented during 6 days, 

the volatile acidity content of sun-dried beans is not 

significantly different from the content both of oven and 

mixed-dried beans. 

Highest volatile acidity in beans fermented during 

2 days could be explained by the accumulation of acetic 

acid which is mainly produced through oxidation of 

ethanol in the presence of oxygen by acetic acid bacteria 

from the second day of fermentation processing (Hii 

et al., 2009). The lowest volatile acidity content in both 

sun- and mixed-dried beans is because of the fact that 

generally volatile acidity is removed as acetic acid from 

Volatile aicdity content 

(meq of NaOH per g) 

3 

2 

1 

0 





Figure 2 Effect of drying methods on volatile acidity of raw cocoa 

beans. C9: sun drying method EV34: artificial hot air drying method, 

C9EV18: combination of sun and artificial drying methods. 


1568 


fermented beans during a slow and gentle drying process 

as natural process contrary to artificial drying process 

which would dry faster and break the diffusion path of 

the acetic acid during moisture removal because of its 

high drying rate (Jinap et al., 1994a,b). Indeed during 

drying, acetic acid is evaporated off along with the 

moisture removal process because of its volatile nature. 

However, the lactic acids contained inside cannot be 

evaporated off as it is not a volatile compound (Hii 

et al., 2009). 

In conclusion, beans solar-dried from 9 am to 6 pm 

daily for 1 week and subsequently dried in an oven at 

60 °C were of comparable quality to sun-dried beans. 

Both of these were of better quality than beans oven-dried 

at 60 °C as previously concluded by Jinap et al. (1994a,b). 

A large number of reports have referred to the high acidity 

and poor flavour development of artificially dried beans 

compared to sun-dried beans with similar levels of 

fermentation (Shelton, 1967; Jinap et al., 1994a). 

Effect of drying methods on free acidity of raw cocoa 

material 

Figure 3 shows the effect of drying methods on free 

acidity of raw cocoa beans. The changes in free acidity 

content show that artificially dried beans have resulted 

in high acidity when compared to both sun- and mixeddried 

beans whatever the duration of fermentation. 

Indeed, artificially dried cocoa contains 2.5, 4.5, 4.2 and 

3.4 meq of NaOH ⁄ g respectively for fresh beans and 

beans fermented during 2, 4 and 6 days. If the fermentation 

is stopped after 4 days, there is a danger of a 

winning of acidity and flavour acidity (Chong et al., 

1978; Duncan et al., 1989) because of the important 

Free acidity content (meq of NaOH per g) 

6 

5 

4 

3 

2 

1 

0 




Figure 3 Effect of drying methods on free acidity of raw cocoa beans. 

C9: sun drying method EV34: artificial hot air drying method, 


production of acetic acid in the presence of oxygen by 

acetic acid bacteria from the oxidation of ethanol 

initially produced by yeasts (Schwan & Wheals, 2004) 

and other organic acids such as lactic acid. The slight 

free acidity content of beans fermented during 2 days 

could be explained by the fact that biochemical changes 

in the sugar of the pulp around the beans lead to the 

formation of only the ethylic alcohol and not of acidic 

compound. Acidity of beans is influenced by several 

factors and fermentation method is known to be crucial. 

Meyer et al. (1989) found that fermentation in boxes has 

been shown to produce more acid beans than either 

heap or tray fermentation. The lower free acidity 

content in beans fermented during 6 days might be 

because of biochemical changes in the heap of fermented 

beans leading to the formation of slight concentration of 

lactic acid because of the turnings made at 48 and 96 h 

of fermentation causing the slowing of lactic bacteria 

growth, the probable loss of these acidities during 

fermentation by exudation of acidic liquid and not by 

chemical degradation (Jinap et al., 1994a). After all, 

most studies have shown that removal of moisture from 

beans increases their acidities (Hii et al., 2009). Free 

acidity content of mixed-dried beans is similar to the 

quality of sun-dried beans as previously found by Jinap 

et al. (1994a) which clearly demonstrated that beans air 

blown for 72 h and subsequently dried in an oven at 

60 °C were of comparable quality to sun-dried beans. 

Both of these cocoa were of better quality than beans 

oven-dried at 60 °C (Jinap et al., 1994a). As previously 

concluded, sun-dried beans and mixed-dried beans 

have resulted in low free acidity when compared to 

artificially dried beans because of the sun drying process 

sufficient time for the volatilisation of acids (acetic acid) 

from the beans and thereby reducing the acidity. Indeed, 

during drying, acetic acid is evaporated off along with 

the moisture removal process because of its volatile 

nature. However, the lactic acids contained inside 

cannot be evaporated off as it is not a volatile compound 

(Hii et al., 2009). So sun drying, if done properly, 

produces the best quality beans (Crespo, 1985). According 

to Bonaparte et al. (1998), this method, however, is 

inefficient and produces beans of inconsistent quality 

when drying conditions are unfavourable. Furthermore, 

it appears that drying of fermented beans at higher 

temperature such as artificially or oven drying process 

resulted in inferior quality cocoa with respect to beans 

acidity. High acidic beans are always associated with 

oven drying process. 

Effect of drying methods on ammonium nitrogen of raw 

cocoa material 

The effect of drying processes on ammonium nitrogen of 

raw cocoa material is presented in Fig. 4. Ammonium 

nitrogen content increases according to the duration of 


Ammonium nitrogen content 

(ppm) 

800 

700 

600 

500 

400 

300 

200 

100 

0 



C9 EV34 C9EV 

Figure 4 Effect of drying methods on Ammonium Nitrogen of raw 

cocoa beans. C9: sun drying method EV34: artificial hot air drying 

method, C9EV18: combination of sun and artificial drying methods. 

fermentation whatever the drying method. Low ammonium 

nitrogen content is expected to the short duration 

of fermentation especially in the early stages, and high 

ammonium nitrogen content is found in cocoa beans 

fermented during long duration. Indeed ammonium 

nitrogen content varies from about 200 to above 

600 ppm from the beginning to the end of fermentation. 

These observations demonstrate clearly that drying 

methods show no significant differences in ammonium 

nitrogen content because of the fact that ammonium 

nitrogen is not a volatile compound. However, ammonium 

nitrogen might be produced during fermentation 

from the degradation of cocoa beans nitrogen compounds, 

such as total polyphenols, flavonoids, epicatechin, 

catechin; the formation of precursors of pyrazines 

(Jinap et al., 1994b; Hashim et al., 1997); and the 

proteolysis of proteins and the oxidation of amino acids 

and peptides by the suppression of nitrogen under the 

control of amino acid oxidase and glutamine synthesis 

enzyme. The presence of relative high concentration of 

ammonium nitrogen in all samples could be because of 

the storage of cocoa pods during 4 days before breaking 

and to the boxes fermentation method. In addition 

probably, long duration of fermentation could lead to 

the production of much ammonium nitrogen through 

amino acid oxidase enzymes activities as previously 

described by Jinap et al. (1994a) and glutamine synthesis 

enzyme and then cocoa beans becomes dark in 

appearance. These enzymes activities originated from 

both endogenous source (beans) and microbial source 

such us Bacillus sp. could generate ammonium nitrogen 

during fermentation (Jinap et al., 1994b). 

Effect of drying methods on free fatty acids of raw cocoa 

material 

Free fatty acids (FFA) content increases slowly but 

remains below the critical value of 1.75% in cocoa beans 

Free fatty acids content ( ) 

2.00 

1.80 

1.60 

1.40 

1.20 

1.00 

0.80 

0.60 

0.40 

0.20 

0.00 





Figure 5 Effect of drying methods on free fatty acids of raw cocoa 

beans. C9: sun drying method EV34: artificial hot air drying method, 


whatever the duration of fermentation (Fig. 5). FFA 

content varies from 0.4% to 1% for both fresh and 

fermented cocoa beans. Among three dried samples 

studied, oven-dried cocoa beans show higher FFA 

content than both solar- and mixed-dried beans. Indeed 

artificial drying process leads to the FFA content above 

0.70%, while both sun and mixed drying processes 

produce raw cocoa beans with FFA content below 

0.70%. Slighter increasing in changes FFA depending 

on the fermentation duration indicates no significant 

difference (P < 0.05) within fermented cocoa beans. So 

limitation of cocoa beans fermentation duration to 

6 days does not seem to be critical to increase the 

chances for FFA formation. FFA content of mixeddried 

beans is similar to the quality of sun-dried beans 

demonstrating that beans air blown for 7 days and 

subsequently dried in an oven at 60 °C were of comparable 

quality to sun-dried beans. Both of these cocoas 

were of better quality than beans oven-dried at 60 °C. 

This result is probably because of the break of triglycerides 

obtained from the liquefaction and the diffusion 

of cocoa butter during the faster process of oven drying 

method. Higher variations of FFA content corresponding 

high standard deviations observed in the same cocoa 

beans sample are probably because of a high heterogeneity 

of beans qualities and certainly to the growth of 

some lipolytic fungi such as Mucor sp. and Rhizopus sp. 

Indeed circumstantial evidence presented the FFA 

formation in cocoa beans is because of microbial 

enzymatic activities in the combination with certain 

factors such as the quality and the physical integrity of 

beans and not to endogenous plant lipases as previously 

demonstrated by Gueńot et al. (1976) and Guehi et al. 


1570 


(2008). Microbial lipase activities hydrolysed the ester 

bonds between fatty acids and hydroxyl functions of 

glycerol in triglycerides of cocoa butter and liberated 

free fatty acids which content increased. As the quality 

of our beans and their integrity were high, a little 

amount of free fatty acid was formed. Indeed, whatever 

the drying process, cocoa beans resulted showed FFA 

content below critical content of 1.75% according to 

European directive (EEC, 1973). According to Gueńot 

et al. (1976), high FFA contents above 1.75% were 

mostly observed in poor-quality raw cocoa beans. This 

observation confirms that all samples obtained from our 

study are of best quality in term of FFA content and 

indicates clearly that any appropriate cocoa drying 

processing do not have an appreciable impact on FFA 

formation. 

Conclusion 

Although drying processes has shown no great effect on 

ammonium nitrogen and free fatty acids formation in 

cocoa beans, the changes in acidic characteristics such as 

pH, free and volatile are largely dependent on the drying 

methods. Among cocoa samples resulted from our 

study, sun-dried beans and beans air blown for 7 days 

and subsequently dried in an oven at 60 °C were of 

better quality than beans oven-dried at 60 °C. So sun 

drying method of cocoa beans is best for optimal 

quality. Fermentation duration has a significant impact 

on the formation of ammonium nitrogen which originated 

from the decomposition of polyphenol compounds. 

The results obtained from this study are 

essential in understanding and solving the problems 

associated with the quality of raw cocoa beans material. 

Further research is needed to carry out the effect of the 

storage time before breaking pods and their sanitary 

quality on physico-chemical and microbial quality 

attributes of industrial raw cocoa material aimed to 

improve globally the quality of raw cocoa beans sourced 

from Coˆ te d’Ivoire. 


This research was supported by Centre National de 

Recherches Agronomiques (CNRA) cocoa program of 

Coˆ te d’Ivoire. The authors are grateful particularly to 

Dr. Firmin ABOUA, Senior Researcher and to the 

support given by the Unity of Research and Formation 

in Food Sciences and Technologies (UFR-STA) of 

Universite´ d’Abobo-Adjame´. 

References 

Ardhana, M.M. & Fleet, G.H. (2003). The microbial ecology of cocoa 

bean fermentations in Indonesia. International Journal of Food 


Beckett, S.T. (1994). Industrial Chocolate Manufacture and Use, 2nd 

edn. Blacki Academic & Professional, Glasgow. 

Bonaparte, A., Alikhani, Z., Madramootoo, C.A. & Raghavan, V. 

(1998). Some Quality Characteristics of Solar-Dried Cocoa Beans in 

St Lucia. Journal of the Science of Food and Agriculture, 76, 553–558. 

Chong, C.F.Shepherd, R. & Poon, Y.C. (1978). Mitigation of cocoa 

bean acidity – Fermentary investigations. In: Cocoa Biotechnology 

(edited by P.S. Dimick). Pp. 387–414. Department of Food Science. 

Pennsylvania State University, USA. 

Crespo, S. (1985). Judging the quality of cocoa beans. The Manufact 

Confectioner, May, 59–63. 

DGTCP (Direction Générale du Trésor et de la Comptabilité Publique 

de Coˆ te d’Ivoire) (2004) Commercialisation du cacao: La Coˆ te 

d’Ivoire s’apprête à relever le de´fi de la certification. Article 140 

online http://www.tresor.gov.ci/actualités. Accessed 08 ⁄ 02 ⁄ 2006. 

Duncan, R.J.E., Godfrey, G., Yap, T.N., Pettipher, G.L. & Tharumarajah, 

T. (1989). Improvement of Malaysian cocoa bean flavour 

by modification of harvesting, fermentation and drying methods – 

The Sime – Cadbury Process. Cocoa Growers’ Bulletin, 42, 43–57. 

EEC (1973). Directive 73 ⁄ 241 ⁄ EEC by European Parliament and the 

European Council relating to cocoa and chocolate products 

intended for human consumption. Official Journal of the European 

Communities L 228 of 16 ⁄ 08 ⁄ 1973, pp. 0023–0035. 

Franke, L.B., Torres, M.Â.P. & Lopes, R.R. (2008). Performance of 

different drying methods and their effects on the physiological 

quality of grain sorghum seeds (S. bicolor (L.) Moench). Revista 

Brasileira de Sementes, 30, 177–184. 

Guehi, T.S., Konan, Y.M., Koffi-Nevry, R.N’Dri, D.Y. & Manizan, 

N.P. (2007). Enumeration and Identification of Main Fungal 

Isolates and Evaluation of Fermentation’s Degree of Ivorian Raw 

Cocoa Beans. Australian Journal of Basic and Applied Sciences, 1, 

479–486. 

Guehi, T.S., Dingkuhn, M., Cros, E. et al. (2008). Impact of cocoa 

processing technologies in free fatty acids formation in stored raw 

cocoa beans. African Journal of Agricultural Research, 3, 174–179. 

Guénot, M.C., Perriot, J.J. & Vincent, J.C. (1976). Evolution de la 

microflore et des acides gras des fèves de cacao au cours du stockage: 

e´tude préliminaire. Cafe´ Cacao The´, 20, 53–58. 

Hansen, G.E., Olmo, M.D. & Burns, C. (1998). Enzyme activities in 

cocoa beans during fermentation. Journal of the Science of Food and 

Agriculture, 77, 273–281. 

Hamid, A. & Lopez, A.S. (2000). Quality and weight changes in cocoa 

beans stored under two warehouses’ conditions in East Malaysia. 

The planter, Kuala Lumpur, 76, 619–637. 

Hashim, P., Selamat, J., Ali, A. & Kharidah, S. (1997). Pyrazines 

formation in cocoa beans: changes during fermentation. Journal of 

Food Science and Technology, 34, 483–487. 

Hashim, P., Selamat, J., Muhammad, S.K.S. & Ali, A. (1998). Changes 

in free amino acid peptide-N, sugar and pyrazine concentration 

during cocoa fermentation. Journal of the Science of Food and 


Hii, C.L., Law, C.L., Cloke, M. & Suzannah, S. (2009). ‘‘Thin layer 

drying kinetics of cocoa and dried product quality’’. Biosystems 

Engineering, 102, 153–161. 

IOCCC: International Office of Cocoa, Chocolate and Sugar Confection 

(1996). Determination of free fatty acids (FFA) content of 

cocoa fat as a measure of cocoa nib acidity. Analytical Method no 

42–1993. 

Jinap, S., Thien, J. & Yap, T.N. (1994a). Effect of drying on acidity 

and volatile fatty acids content of cocoa beans. Journal of the 

Science of Food and Agriculture, 65, 67–75. 

Jinap, S., Siti, M.H. & Norsiati, M.G. (1994b). Formation of Methyl 

Pyrazine during Cocoa Bean Fermentation. Pertanika Journal of 

Tropical Agriculture Science, 17, 27–32. 

Jones, K.L. & Jones, S.E. (1984). Fermentations involved in the 

production of cocoa, coffee and tea. Progress in Industrial Microbiology, 

19, 411–433. 


Lehrian, D.W. & Patterson, G.R. (1983). Cocoa fermentation. In: 

Food and Feed Production with Microorganisms. Biotechnology, vol. 

5. (edited by G. Reed). Pp. 529–575. Verlag Chemie, Weinheim, 

Germany. 

Lopez, A.S.F. & Dimick, P.S. (1995). Cocoa fermentation. In: 

Enzymes, Biomass, Food and Feed, 2nd edn. Biotechnology, Vol. 9. 

(edited by G. Reed & T.W. Nagodawithana). Pp. 561–577. VCH, 

Weinheim, Germany. 

Mc Donald, C.R., Lass, R.A. & Lopez, A.S.F. (1981). Cocoa drying – 

a review. Cocoa Growers’ Bulletin, 31, 5–41. 

Meyer, B., Biehl, B., Said, M.B. & Samarakoddy, R.J. (1989). Post 

harvest pod storage: a method of pulp preconditioning to impair 

strong nib acidification during cocoa fermentation in Malaysia. 

Journal of the Science of Food and Agriculture, 48, 285–304. 

Mounjouenpou, P., Gueule, D., Fontana-Tachon, A., Guyot, B., 

Tondje, P.R. & Guiraud, J.-P. (2008). Filamentous fungi producing 

ochratoxin a during cocoa processing in Cameroon. International 

Journal of Food Microbiology, 121, 234–241. 

Nazaruddin, R., Seng, L.K., Hassan, O. & Said, M. (2006). Effect of 

pulp preconditioning on the content of polyphenols in cocoa beans 

(Theobroma Cacao) during fermentation. Industrial Crops and 

Products, 24, 87–94. 

Pontillon, J. & Cros, E. (1998). Me´thodes analytiques pour le cacao 

et produits dérivés. In: Cacao et chocolat (edited by J. Pontillon). 

Pp. 447–545. Paris, France: Lavoisier. 


Roelofsen, P.A. (1958). Fermentation, drying and storage of cacao 

beans. Advances in Food Research, 8, 225–296. 

SAS Institute, Inc. (2002). JMP Software Version 5, Cary, NC. 

Schwan, R.F. & Wheals, A.E. (2004). The microbiology of cocoa 

fermentation and its role in chocolate quality. Critical Reviews in 

Food Science and Nutrition, 44, 205–221. 

Schwan, R.F., Rose, A.H. & Board, R.G. (1995). Microbial fermentation 

of cocoa beans, with emphasis on enzymatic degradation of 

the pulp. Journal of Applied Bacteriology, 79, 96S–107S (symposium 

supplement). 

Shelton, B. (1967). Artificial drying of cocoa beans. Tropical Agriculture 

(Trinidad), 44, 125–131. 

Tafuri, A., Ferracane, R. & Ritieni, A. (2004). Ochratoxin A in Italian 

marketed cocoa products. Food chemistry, 88, 487–494. 

Thompson, S.S., Miller, K.B. & Lopez, A.S.F. (2001). Cocoa and 

coffee. In: Food Microbiology Fundamentals and Frontiers (edited by 

M.P. Doyle, L.R. Beuchat & T.J. Montville). Pp. 721–736. ASM 

Press, Washington, DC, USA. 

Wood, G.A.R., Lass, R.A. (1985). Cocoa, 4th edn. Longman Inc., 

New York. 

World Cocoa Foundation (2008). Cocoa market. Available from: 

http://www.worldcocoafoundation.org. Accessed 17 ⁄ 04 ⁄ 2009. 


1572 


Optimisation of Spirulina platensis convective drying: evaluation of 

phycocyanin loss and lipid oxidation 

Elizangela G. Oliveira, Jessica H. Duarte, Kelly Moraes, Valeria T. Crexi & Luiz A. A. Pinto* 

Unit Operations Laboratory, School of Chemistry and Food, Federal University of Rio Grande (FURG), P.O. Box 474, Zip 96201-900, Rio 

Grande, RS, Brazil 

(Received 26 November 2009; Accepted in revised form 30 April 2010) 

Summary The aim of the study was the optimisation of Spirulina platensis drying on convective hot air through the 

response surface methodology. The responses were thiobarbituric acid (TBA) and phycocyanin loss 

percentage values in final product. Experiments were carried out in perforated tray drier with parallel air 

flow, and the wet samples thickness and drying air temperatures were in range of 3–7 mm and 50–70 °C, 

respectively. The statistical analysis showed significant effect (P 0.05) in relation to 

fresh biomass. The lipid profile of dried product presented high percentage of polyunsaturated fatty acids 

(34.4%), especially the gamma-linolenic acid (20.6%). 

Keywords Drying, fatty acids, microalgae, phycocyanin, spirulina platensis, thiobarbituric acid. 

Introduction 

The microalgae Spirulina platensis is produced commercially 

all over the world, and the dried product is 

valuable food supplement. It is rich in proteins (60–70% 

by dry weight), vitamins (especially B12 and b-carotene) 

and minerals. It contains many essential amino acids 

and fatty acids (Jiménez et al., 2003); also it is an 

inexpensive source of pigment (Richmond, 1988). The 

Spirulina components, with antioxidants properties, are 

the polyunsaturated fatty acids and pigments (Estrada 

et al., 2001). Phycocyanin and gamma-linolenic acid 

(C18:3, x6, GLA) are the components of the Spirulina, 

which has been receiving more attention from researchers. 

Phycocyanin is the main pigment produced by the 

microalgae Spirulina platensis and reaches 20% in dry 

weight of the cell protein (Vonshak, 1997). Phycocyanin 

has a significant antioxidant, anti-inflammatory, 

hepatoprotective and free radical properties; it is also 

used in food colouring and in cosmetics as they are 

nontoxic and noncarcinogenic (Henrikson, 1994; Morist 

et al., 2001; Minkova et al., 2003; ). Phycocyanin is used 

in chewing gums, dairy products, ice creams, jellies 


e-mail: dqmpinto@furg.br; luiz.pinto@pq.cnpq.br 


(Cohen, 1986; Yoshida et al., 1996) and biomedical 

research (Glazer, 1994). It is used as potential therapeutic 

agent reducing the oxidative stress disease 

(Romay et al., 1998; Bhat & Madyastha, 2001). The 

extraction method of phycocyanin from spirulina biomass 

is suggested by several methods such as spray 

drying and oven dried which result in approximately 

50% loss of phycocyanin (Sarada et al., 1999). The 

improved drying method is important to store maximum 

amount of phycocyanin in the biomass (Doke, 2005). 

The microalgae Spirulina platensis is a potential 

source of gamma-linolenic acid (GLA), an essential 

polyunsaturated fatty acid with economic interest. GLA 

is metabolite of linolenic acid (LA) and the first 

intermediate in the conversion of LA to arachidonic 

acid (AA) (Gustone, 1992). The microbial production 

for extraction of polyunsaturated fatty acid (PUFA) is 

considered an economical alternative to produce large 

quantities of fatty acids (Kennedy et al., 1993). 

Dehydration operation is an important step in the 

chemical and food processing industries. The basic 

objective in foodstuffs drying is the removal of water in 

the solids for to minimise the microbial growth and 

deterioration by chemical reactions (Krokida et al., 

2003). Also, it leads the reduction in weight and volume 

of material, storage and transportation costs (Okos 

et al., 1992). Convective hot air drying is one of the most 

doi:10.1111/j.1365-2621.2010.02299.x 

Ó 2010 Institute of Food Science and Technology

common industrial methods used for drying organic 

material and is a simultaneous process that involves heat 

and mass transfer followed by a phase or state change 

(Babalis & Belessiotis, 2004). 

The drying of Spirulina platensis constitutes approximately 

30% of the total production cost, and the 

traditional methods used to process fresh biomass into 

dry Spirulina are spray drying, freeze drying, solar 

drying, convective hot air drying and spouted bed 

(Morist et al., 2001; Jimeńez et al., 2003; Desmorieux & 

Decaen, 2006; Oliveira et al., 2009). Sarada et al. (1999) 

found a loss of approximately 50% of the phycocyanin 

presented in the biomass when different techniques have 

been used to dry microalgae Spirulina, including the 

spray and convective dryers. Oliveira et al. (2008) 

reported that high drying temperature (>60 °C) in 

two different dryers (spouted bed and convective air 

dryers) decreased the amount of phycocyanin extractable 

from Spirulina platensis. When low-cost drying 

methods are optimised, e.g. convective dryer, have the 

potential to approach the advantages derived from 

spray drying in terms of quality and bioavailability at 

lower cost. 

Several factors can influence hot air drying of foods, 

for example: velocity and temperature of air, water 

diffusion through material, load density, thickness and 

shape of the product. The factors temperature and 

thickness are more important in the microalgae drying 

operation according to literature (Desmorieux & Decaen, 

2006; Oliveira et al., 2009). However, the water 

removal decreases the nutritional and sensorial values of 

food and leads to phenomena such as hardening and 

shrinkage (Vega et al., 2007). The microalgae Spirulina 

platensis is a high source of biocompounds, as phycobiliproteins 

(phycocyanin) and essential fatty acids; it is 

important reason to optimise the drying conditions. 

The aim of this study was the optimisation of the of 

microalgae Spirulina platensis drying through the 

response surface methodology (RSM), considering as 

independent variables the drying air temperature and 

sample thickness. The responses were phycocyanin loss 

percentage and thiobarbituric acid (TBA) values. The 

fatty acids profile was determined in the best drying 

condition and compared with the fresh biomass. 


Raw material 

Spirulina strain LEB-18 (Costa et al., 2004) was cultivated 

in a 450 L open outdoor photo-bioreactors, under 

uncontrolled conditions, in the south of Brazil. During 

these cultivations, water was supplemented with 20% 

Zarrouk synthetic medium (Zarrouk, 1966), containing 

(g L )1 ): NaHCO3, 16.8; NaNO3, 2.5; K2HPO4, 0.5; 

K2SO4, 1.0; NaCl, 1.0; MgSO4.7H2O, 0.2; CaCl2, 0.04; 

Optimisation of Spirulina convective drying E. G. Oliveira et al. 1573 

FeSO4.7H2O, 0.01; EDTA, 0.08 and micronutrients. 

The initial biomass concentration was 0.15 g L )1 .Samples 

were taken every 24 h to determine the biomass 

concentration by optical density measurements at 

670 nm using a spectrophotometer (Quimis model 

Q108-DRM, Sa˜ o Paulo, Brazil). In the end of cultivation, 

the biomass was recovered by filtration and 

pressing. 

Biomass drying 

Drying experiments were carried out at 50, 60 and 70 °C 

in the discontinuous tray dryer. The samples thicknesses 

were 3, 5 and 7 mm with load density of 4 kg m )2 , and the 

hot air velocity was 2.5 m s )1 . The drying procedure was 

continued till the moisture content of the sample was 

about 0.10 kg kg )1 (wet basis). After drying experiments, 

all products were ground in a knife mill (Willey model, 

Philadelphia, USA) and sieved (150 mesh), packed in 

plastics bags and stored at ambient temperature. Each 

drying experiment was carried out in duplicate. 

Chemical analyses 

The following parameters were determined in duplicate 

for the Spirulina platensis fresh: moisture (method 

925.10), protein (method 960.52) and ash (method 

923.03) contents. Determinations were carried out 

according to Association of Official Analytical Chemists, 

AOAC, (1995). Lipids content was through the 

methodology proposed by Folch & Lees (1957), and 

carbohydrate content was determined by difference. 

Phycocyanin content 

Quantitative analysis of phycocyanin was carried out by 

Spectrometric method, according to Boussiba & Richmond 

(1979), on fresh and dried biomass. Initially, it was 

added 2 g of the wet sample in the pan, and after, it was 

dried in the oven for 6 h. To determine the percentage of 

phycocyanin, 40 mg of dried Spirulina was weighed, 

mixed in 10 mL of phosphate buffer 0.1 m (pH = 7) and 

stirred until complete dissolution. The samples were 

stored in refrigerator at 4 °C overnight. The samples were 

subsequently mixed and centrifuged (Fanem model Baby 

I 206BL, Sa˜ o Paulo, Brazil) at 10 °C, 4000 g for 5 min. 

The absorbance was read in spectrophotometer (Quimis 

model Q-108DRM, Sa˜ o Paulo, Brazil) at 620 nm using 

phosphate buffer as blank. Phycocyanin content was 

calculated according to Eqn 1: 

% Phyco ¼ 

A620.nd 

3:39:ðmsampleÞ:ðXdry matterÞ 

:100 ð1Þ 

where % Phyco is the phycocyanin percentage in 

sample, A 620 is the absorbency in a wave length of 

Ó 2010 Institute of Food Science and Technology International Journal of Food Science and Technology 2010, 45, 1572–1578

1574 

Optimisation of Spirulina convective drying E. G. Oliveira et al. 

620 nm, nd is the dilution number (mL), 3.39 is the 

coefficient of extinction for phycocyanin at 620 nm, 

msample is the wet mass of Spirulina (g), Xdry matter is 

the Spirulina dry composition (dimensionless) and 100 

is the representative of 100%. Each analysis was 

performed in duplicate. 

TBA value 

Dried samples were carried out to the TBA determination 

of lipid oxidation. The TBA value was done 

according to Tiburcio et al. (2007) with some modifications: 

Spirulina powder (10 g) was mixed with 40 mL 

of chloroform and filtrated. The filtrate (10 mL) was 

placed in tubes of centrifuge (Fanem model Baby I 

206BL, Sa˜ o Paulo, Brazil) with 10 mL of trichloroacetic 

acid (TCA) 10% w ⁄ v and centrifuged at 2000 g for 

15 min. The supernatant (4 mL) and 1 mL of thiobarbituric 

acid (TBA) 0.02 m were stirred for 5 min and 

then incubated in boiling water bath for 40 min to 

develop the colour. Absorbance of the resulting supernatant 

was determined at 530 nm by a spectrophotometer 

(Quimis model Q-108DRM, Sa˜ o Paulo, 

Brazil). The reagents used were of analytical grade. 

The TBA value was calculated through standard curve 

obtained by reacting of tetramethoxypropane 0.01 m 

with TBA, the value was expressed as milligrams of 

malonyldialdehyde (MDA) per kg of sample in dry 

basis. 

Fatty acids profiles 

To evaluate the fatty acids profiles of the microalga 

Spirulina platensis, lipids extraction was carried out by a 

methodology proposed by Folch & Lees (1957). The 

extraction was carried out in the fresh and dehydrated 

biomass. Fatty acid identification and quantification 

were carried out by chromatographic analysis for 

Spirulina oil. Fatty acids profiles were determined by 

preparation of methyl esters as described by Metcalfe & 

Schimitz (1966). 

Fatty acid methyl esters (FAME) were identified by 

gas chromatography (chromatographer model Varian- 

3400 CX, Palo Alto, USA) equipped with a DB-17 J&W 

Scientific (50% phenyl methylpolysiloxane) capillary 

column. Fatty acid esters analyses were carried out in 

duplicate by injecting 1.0 lL, SPLIT ratio 1:50, into the 

capillary column (30 · 0.25 mm, 0.25 m film in thickness). 

GC setting conditions were as follows: injection 

temperature 250 °C and flame ionisation detector temperature 

300 °C, helium gas carrier flow rate 

1.0 mL min )1 , linear speed 24 cm s )1 and oven temperature 

held at 100 °C for 1 min, then increased to 160 °C 

at 6 °C min )1 and held at 230 °C at6°C min )1 . FAME 

were identified by direct comparison of the retention 

times with standards (SUPELCO TM 37, Bellefonte, PA, 

USA) and were quantified as the percentage area of each 

FAME mixture. 


In the drying experiments, the effects of the drying air 

temperature (X1) and the samples thickness (X2) were 

studied through a full factorial design (3 2 ) on the 

responses for phycocyanin loss percentage and TBA 

values. The optimisation procedure was by the response 

surface methodology (Myer, 1976). 

The independent variables (factors) investigated in 

this study were drying air temperature and wet samples 

thickness. The values of these factors were 50, 60 and 

70 °C for air temperature; 3, 5 and 7 mm for samples 

thickness. The coded levels of these values, for both 

factors, were represented by ()1), (0) and (+1), respectively. 

For the regression analysis of the experimental data, 

the software Statistica 6.0 for Windows (StatSoft Inc., 

USA) was used. The statistical significance of the 

second-order statistical model (Eqn 2) and the variance 

explained by this model were determined by Fisher’s test 

and coefficient of determination (R 2 ), respectively. The 

response surfaces were found to define the optimal 

drying conditions for phycocyanin loss percentage and 

TBA values. 

The second-order statistical model is presented in 

Eqn 2: 

Yn ¼ b0 þ b1X1 þ b11X 2 1 þ b2X2 þ b22X 2 2 þ b12X1X2 

ð2Þ 

where Yn is the predicted response (in real value), X1 is 

the codified temperature, X2 is the codified thickness 

and b0, b1, b11, b2, b22 and b12 are the regression 

coefficients. 

The fatty acids profile was found in the best drying 

condition, and the results were compared using the 

Tukey test at the significance level of 95% (P < 5%), 

with the aid of the software Statistica 6.0 for Windows 

(Statsoft Inc., USA). 


Characterisation of the Spirulina platensis fresh was (in 

wet basis): moisture content 75.7% ± 0.2; ash content 

1.7% ± 0.2; protein content 11.9% ± 0.3; lipids content 

3.4% ± 0.5 and carbohydrate content 7.3% ± 0.1 

(by difference). The phycocyanin content value of the 

fresh biomass was 0.11 kg kg )1 . The phycocyanin loss 

percentage values on dried products were calculated in 

relation to the fresh biomass. Table 1 shows the total 

drying time, phycocyanin loss percentage and TBA 

values of the factorial design matrix. 

The TBA value is extensively used like an indicator to 

the lipid oxidation degree, quantified the secondary 

products by oxidation. According to Boran et al. (2006), 

the TBA value required for acceptability of oils for 

International Journal of Food Science and Technology 2010, 45, 1572–1578 Ó 2010 Institute of Food Science and Technology

Table 1 Results of the 3 2 factorial design matrix for drying experiments 

Experiment 

(n°) 

X1 (Coded 

temperature) 

X2 (Coded 

thickness) 

human consumption is in the range 7–8 mg MDA kg )1 . 

Therefore, the oils obtained by dried Spirulina in all 

drying conditions (Table 1) presented oxidative quality. 

Drying process is a very important step to store 

maximum phycocyanin content in biomass and is an 

efficient extraction method for maximum recovery with 

relatively high purity ratio (Doke, 2005). According to 

Sarada et al. (1999), considerable loss of phycocyanin 

concentration was observed when wet biomass was dried 

at elevated temperature. The same occurred with the 

biomass dried at 70 °C and in the thickness of 7 mm 

(Table 1). 

Statistical analysis of the results in Table 2 showed 

that in the drying of microalgae Spirulina platensis in 

convective air drying, both the study factors (air 

temperature and sample thickness) have a significance 

effect at level of 95% (P < 0.05) on the loss of 

phycocyanin and TBA values. The results in Table 2 

show that the temperature has an important role on 

phycocyanin content during the drying process of 

Spirulina biomass. Also, it can be observed that the 

thickness presented significance in the loss of pigment, it 

can be explained due the higher thickness material had 

taken a larger time to dry (Table 1) until the commercial 

moisture content. 

Total drying 

time* (min) 

TBA value* 

(mgMDA kg )1 ) 

Phycocyanin loss 

percentage* (%) 

1 )1 )1 192 ± 9 1.43 ± 0.10 43.76 ± 0.11 

2 0 )1 150 ± 7 2.12 ± 0.03 43.84 ± 0.17 

3 1 )1 133 ± 12 2.27 ± 0.15 75.30 ± 0.12 

4 )1 0 369 ± 5 1.20 ± 0.01 56.24 ± 0.20 

5 0 0 210 ± 5 0.90 ± 0.02 40.17 ± 0.15 

6 1 0 145 ± 15 0.96 ± 0.05 76.15 ± 0.21 

7 )1 1 608 ± 10 0.88 ± 0.04 82.22 ± 0.13 

8 0 1 300 ± 6 0.76 ± 0.01 85.13 ± 0.14 

9 1 1 214 ± 5 0.65 ± 0.05 93.07 ± 0.12 

MDA, malonyldialdehyde; TBA, thiobarbituric acid. 

*Mean value ± standard error (in replicate). 

Table 2 Effects and significance estimated by nonlinear regression 

Phycocyanin loss (%) 


In Table 2, the results of TBA values show that the 

effects of the thickness were higher than temperature one 

in this response. The lipid oxidations do not occur in 

saturated fatty acids, only in drastic conditions of 

temperature. However, the TBA value in the higher 

temperatures and thickness showed a reduction, it can 

be explained due the drying experiments occurred faster 

in highest temperature, and also in the higher thickness, 

the oil was more protected of the action of the 

temperature. The results of the second-order statistical 

model of analysis of variance are shown in Table 3. 

In Table 3, the F-test leads to that the model was 

adequate because the calculated (Fc) values were 3.48 

and 3.25 times higher than the table (Ft) values for 

phycocyanin loss percentage and TBA values, respectively, 

at 95% of confidence. As a practical rule, a model 

has statistical significance when the calculated F value is 

at least 3–5 times higher than the table value (Khury & 

Cornell, 1996). These values guarantee a satisfactory fit 

of the quadratic models to the experimental data, 

Eqns 3 and 4, and indicated that 95% of the variability 

in the responses could be explained by the models. The 

coefficients of correlation (R 2 ) for phycocyanin loss 

percentage and TBA responses were 0.95 and 0.94, 

respectively. 

TBA value (mg MDA kg )1 dry sample) 

Effect Pure Error t(2) P 

Effect Pure Error t(2) 

Average 43.85 0.085 514.25

1576 


Table 3 Analysis of variance (ANOVA) for responses of the statistical model 

Y1 ¼ 43:85 þ 10:38X1 þ 17:04X 2 1 þ 16:25X2 

þ 15:33X 2 2 

SS df MS Fc Ft* Fc ⁄ Ft 

Phycocyanin loss percentage (%) 

Regression 4153.01 5 830.60 17.59 5.05 3.48 

Residual 236.07 5 47.21 – – 

Total 4389.08 10 – – 

TBA value (mg MDA kg )1 drysample) 

Regression 2.78 5 0.55 14.75 4.53 3.25 

Residual 0.18 5 0.04 – – 

Total 2.97 10 – – 

SS, Sum square; df, degree free; MS, mean square; MDA, malonyldialdehyde; TBA, thiobarbituric acid; F c, calculated value; F t, table value. 

*The Fisher F distribution at 5% probability (P < 0.05). 

5:17X1X2 

ð3Þ 

Y2 ¼ 0:96 þ 0:06X1 0:59X2 þ 0:37X 2 2 0:26X1X2 

ð4Þ 

where Y1 is phycocyanin loss percentage, in %; Y2 is the 

TBA value, in mgMDA kg )1 sample; X1 is the codified 

temperature; X 2 is the codified thickness. 

The response surface curves and the contour plot for 

drying of microalgae Spirulina platensis were obtained 

by statistical model (Eqns 3 and 4) and are show in 

Figs 1 and 2. The drying condition with the lowest 

phycocyanin loss percentage value was considered the 

best condition. Figures 1 and 2 show that the lower 

phycocyanin loss percentage and TBA values were 

obtained in air temperature 55 °C and sample thickness 

of 3.7 mm. In the best drying condition, the responses 

for the phycocyanin loss percentage and TBA values 

were 37.4 ± 0.3% and 1.5 ± 0.1 mgMDA kg )1 sample, 

respectively. 

Figure 1 Response surface curve for loss of phycocyanin. 

Tiburcio et al. (2007) studied the lipid oxidation of 

Spirulina platensis in three drying techniques, sun, solar 

and draft oven and compared with control sample 

dried in spray dryer. The lipid oxidation was evaluated 

through of TBA value. The authors found that 

the combined effect of tertiary-butyl hydroquinone 

(TBHQ) and microwave blanching was the most 

effective pre-dehydration treatment for minimising lipid 

peroxidation in drying Spirulina, and the lowest TBA 

value in product dried in sun-drying (0.47 mg MDA kg )1 ) 

also was closest to the spray-dried sample (control), 

0.43 mgMDA kg )1 . They suggest that sun-drying when 

optimised produced a dried product more stable than 

spray-dried. As the pre-dehydration treatment was not 

used in this study, the higher TBA values were observed 

(Table 1). 

The drying method used in this study was considered 

satisfactory by statistical analysis in the drying of 

microalgae Spirulina, and in the best drying condition, 

the loss of phycocyanin was about 40%. This value was 

lower than those reported by the literature using 

methods more expensive and traditional of drying of 

microalgae (Sarada et al., 1999). 

Table 4 shows the composition of main fatty acids 

present in microalgae Spirulina platensis fresh and after 

the drying process in the best condition. There were not 

significance differences (P < 0.05), through the Tukey 

test, between the oils obtained by Spirulina fresh and by 

dried product in the best drying condition. The oil 

obtained from the Spirulina is an important source of 

monosaturated and polyunsaturated fatty acids (MUFA 

and PUFA), about 51% of total fatty acids. In both, oils 

was found linoleic acid and linolenic acid about 13.9% 

and 20.6%, respectively. Thermal processes, as well as 

light exposure, may affect MUFA and PUFA contents 

because of the oxidation process (Morist et al., 2001). In 

Table 4, the drying process does not change the chemical 

structure of linolenic acid and also did not occur 

unfavourable transformations to the nutrients of this 

microalgae. 

The fatty acids are susceptible molecules to treatments 

that use high temperature, leading to chemical 


Table 4 Fatty acids compositions of microalgae Spirulina fresh and 

dried product 

Fatty Acid (%) Spirulina fresh *† 

C11:0 3.50 ± 0.01 a 

C16:0 32.50 ± 0.02 a 

C16:1x7 13.09 ± 0.02 a 

C18:0 2.89 ± 0.01 a 

C18:1x9 3.10 ± 0,02 a 

C18:2 x6 trans 13.88 ± 0.01 a 

C18:3x6 cis 20.60 ± 0.02 a 

Saturated 38.89 ± 0.01 a 

Monounsaturated 16.19 ± 0.02 a 

Polyunsaturated 34.48 ± 0.01 a 

PUFA ⁄ SFA 0.88 ± 0.02 a 

Spirulina dried *† 

3.48 ± 0.02 a 

32.53 ± 0.01 a 

13.10 ± 0.02 a 

2.90 ± 0.01 a 

3.20 ± 0.02 a 

13.86 ± 0.02 a 

20.56 ± 0.01 a 

38.91 ± 0.01 a 

16.30 ± 0.02 a 

34.42 ± 0.01 a 

0.88 ± 0.02 a 

* Tukey test: same letters indicate no significance differences (P > 0.05). 

† Values means ± standard deviation (in replicate). 

Figure 2 Response surface curve for TBA values. 

transformations, polymerisation, geometrical isomerisation 

and intramolecular cyclisation (Berdeaux et al., 

2007; Ceriane & Meirelless, 2007); however, in this best 

condition, the trans-linolenic fat acid was not changed 

after the drying process. A similar result was reported by 

Zepka et al. (2007), on drying the microalgae Aphanothece 

microscopica Na¨geli, with different conditions in 

convective air dryer. 

A higher content of PUFA increases the nutritional 

value of foods. The recommendation for the human diet 

is a polyunsaturated ⁄ saturated (PUFA ⁄ SFA) ratio 

higher than 0.45 (Cuthbertson, 1989). An analysis of 

Table 4 shows that the lipids profile of the microalgae 

Spirulina presents the (PUFA ⁄ SFA) ratio of 0.88. This 

result agrees with other authors (Campanella et al., 

1999; Tokusoglu & U¨ nal, 2003) that considered the 

microalgae Spirulina as potential sources of the essential 

fatty acid gamma-linolenic acid. 

Conclusion 

The Spirulina platensis convective hot air drying, in 

thin layer with parallel air flow, showed a significant 

effect of air temperature and samples thickness 

(P 0.05). The lipid profile of Spirulina 

platensis samples in nature and dried showed a 

polyunsaturated and saturated fatty acids (PUFA ⁄ SFA) 

ratio of 0.88, being the recommendation for the human 

diet a ratio higher than 0.45. The palmitic and gammalinolenic 

acids presented the highest values (32.5% and 

20.6%, respectively) of the fatty acids compositions of 

Spirulina fresh and dried product. 


The authors are grateful to CAPES (Brazilian Agency 

for Improvement of Graduate Personnel) and FA- 

PERGS (Rio Grande do Sul Research Support Foundation), 

Brazil, for the financial support that made this 

work possible. 

References 


AOAC (1995). Official Methods of Analysis, 16th edn. Arlington, 

Virginia, USA: Association of Official Analytical Chemists. 

Babalis, S.J. & Belessiotis, V.G. (2004). Influence of the drying 

conditions on the drying constants and moisture diffusivity during 

the thin-layer drying of figs. Journal of Food Engineering, 65, 449– 

458. 

Berdeaux, O., Fournier, V., Lambelet, P., Dionisi, F., Sebedio, J.L. & 

Destaillats, F. (2007). Isolation and structural analysis of the cyclic 

fatty acid monomers formed from eicosapentaenoic and docosahexaenoic 

acids during fish oil deodorization. Journal of Chromatograph 

A, 1138, 216–224. 

Bhat, V.D. & Madyastha, K.M. (2001). Scavenging of peroxynitrite by 

phycocyanin and phycocyanobilin from Spirulina platensis: protection 

against oxidative damage to DNA. Biochemical and Biophysical 

Research Communications, 285, 262–266. 

Boran, G., Karaçam, H. & Boran, M. (2006). Changes in the quality of 

fish oils due to storage temperature and time. Food Chemistry, 98, 

693–698. 

Boussiba, S. & Richmond, A. (1979). Isolation and purification of 

phycocyanins from the blue-green alga Spirulina platensis. Archives 

of Microbiology, 120, 155–159. 

Campanella, L., Crescentini, G. & Avino, P. (1999). Chemical 

composition and nutritional evaluation of some natural and 

commercial food products based on Spirulina. Analusis, 27, 533– 

540. 

Ceriane, R. & Meirelless, J.A. (2007). Formation of trans PUFA 

during deodorization of canola oil: a study through computational 


1578 


simulation. Chemical Engineering and Processing: Process Intensification, 

46, 375–385. 

Cohen, Z. (1986). Products from microalgae. Handbook of Microalgal 

Mass Culture. Pp. 69–99. Boca Raton: CRC Press Inc. 

Costa, J.A.V., Colla, L.M. & Duarte Filho, P.F. (2004). Improving 

Spirulina platensis biomass yield using a fed-batch process. Bioresource 


Cuthbertson, W.F.J. (1989). What is a healthy food? Food Chemistry, 

33, 53–80. 

Desmorieux, H. & Decaen, N. (2006). Convective drying of Spirulina 

in thin layer. Journal of Food Engineering, 77, 64–70. 

Doke, J.M. Jr (2005). An improved and efficient method for the 

extraction of phycocyanin from Spirulina sp. Intern. International 

Journal of Food Engineering, 1, 1–11. 

Estrada, J.E.P., Besco´s, P.B. & Fresno, A.M.V. (2001). Antioxidant 

activity of different fractions of Spirulina platensis protean extract. 

IL Farmaco, 56, 497–500. 

Folch, J. & Lees, M. (1957). A simple method for isolation and 

purification of total lipids from animal tissues. Journal of Biological 

Chemistry, 226, 497–509. 

Glazer, A.N. (1994). Phycobiliproteins: a family of valuable widely 

used fluorophores. Journal of Applied Phycology, 6, 105–112. 

Gustone, F.D. (1992). c-linolenic acid. Occurrence and physical and 

chemical properties. Progress in Lipid Research, 31, 145–161. 

Henrikson, R. (1994). Microalga Spirulina superalimento del futuro. Pp. 

112. Barcelona: Urano. 

Jiménez, C., Cossio´, B.R., Labella, D.F. & Niell, X. (2003). The 

feasibility of industrial production of Spirulina (Arthrospira) in 

Southern Spain. Aquaculture, 217, 179–190. 

Kennedy, M.J., Reader, S.L. & Davies, R.J. (1993). Fatty acid 

production characteristics of fungi with particular emphasis on 

gamma linolenic acid production. Biotechnology and Bioengineering, 

42, 625–634. 

Khury, A.I. & Cornell, J.A. (1996). Response Surfaces. 2nd edn. New 

York: Marcel Dekker. 

Krokida, M.K., Karathanos, V.T., Maroulis, Z.B. & Marinos-Kouris, 

D. (2003). Drying kinetics of some vegetables. Journal Food 


Metcalfe, L.D.A.A. & Schimitz, J.R. (1966). Rapid preparation of 

fatty acid esters from lipids for gas liquid chromatography. 

Analytical Chemistry, 38, 510 p. 

Minkova, K.M., Tchernov, A.A., Tchorbadjieva, M.I., Fournadjieva, 

S.T., Antova, R.E. & Busheva, M.C.H. (2003). Purification of 

C-phycocyanin from Spirulina (Arthrospira) fusiforms. Journal of 

Biotechnology, 102, 55–59. 

Morist, A., Montesinos, J.L., Cusidó, J.A. & Gòdia, F. (2001). 

Recovery and treatment of Spirulina platensis cells cultured in a 

continuous photobioreactor to be used as food. Process Biochemistry, 

37, 535–547. 

Myer, R.H. (1976). Response Surface Methodology. New York: 

Library of Congress. 

Okos, M.R., Narasimhan, G., Singh, R.K. & Witnauer, A.C. (1992). 

Food dehydration. In: Handbook of Food Engineering (Edited by 

D.R. Hedman & D.B. Lund). Pp. 437–562. New York: Marcel 

Dekker. 

Oliveira, E.G., Rosa, G.S., Moraes, M.A. & Pinto, L.A.A. (2008). 

Phycocyanin content of Spirulina platensis dried in spouted bed and 

thin layer. Journal of Food Process Engineering, 31, 34–50. 

Oliveira, E.G., Rosa, G.S., Moraes, M.A. & Pinto, L.A.A. (2009). 

Characterization of thin layer drying of Spirulina platensis utilizing 

perpendicular air flow. Bioresource Technology, 100, 1297–1303. 

Richmond, A. (1988). Spirulina. In: Micro-Algal Biotechnology (edited 

by M.A. Borowitzka & L.J. Borowitzka). Pp. 85–121. Cambridge: 

Cambridge University Press. 

Romay, C., Armesto, J., Remirez, D., González, R., Ledon, N. & García, 

I. (1998). Antioxidant and anti-inflammatory properties of C-phycocyanin 

from blue-green algae. Inflammation Research, 47, 36–41. 

Sarada, R., Pillai, M.G. & Ravishankar, G.A. (1999). Phycocyanin 

from Spirulina sp: influence of processing of biomass on phycocyanin 

yield, analysis of efficacy of extraction methods and stability 

studies on phycocyanin. Process Biochemistry, 34, 795–801. 

Tiburcio, P.C., Galvez, F.C.F., Cruz, L.J. & Gavino, V.C. (2007). 

Determination of shelf life of Spirulina platensis (MI2) grown in the 

Philippines. Journal of Applied Phycology, 19, 727–731. 

Tokusoglu, Ö. &U¨ nal, M.K. (2003). Biomass nutrients profiles of 

three microalgae: Spirulina platensis, Chlorella vulgaris, and Isochrisis 

galbana. Journal of Food Science, 68, 1144–1148. 

Vega, A., Fito, P., Andrés, A. & Lemus, R. (2007). Mathematical 

modeling of hot-air drying kinetics of red bell pepper (var. Lamuyo). 


Vonshak, A. (1997). Spirulina Platensis (Arthrospira) Physiology, Cell- 

Biology and Biotechnology. London: Taylor & Francis. 

Yoshida, A., Takagaki, Y. & Nishimune, T. (1996). Enzyme immunoassay 

for phycocyanin as the main component of Spirulina color 

in foods. Bioscience, Biotechnology and Biochemistry, 60, 57–60. 

Zarrouk, C. (1966). Contribution A` L’e´tude D’une Cyanophyceé. 

Influence De Divers Facteurs Physiques Et Chimiques Sur La 

Croissance Et La Photosynthe`se De Spirulina Maxima. Ph.D. Thesis. 

Paris, France: University of Paris. 

Zepka, L.Q., Lopes, E.J., Goldbeck, R. & Queiroz, M.I. (2007). 

Production and biochemical profile of the microalgae Aphanothece 

microscopica Na¨geli submitted to different drying conditions. 

Chemical Engineering and Process: Process Intensification, 47, 

1305–1310. 




Taste characterisation of green tea catechins 

Masataka Narukawa, 1,2,3 Hironobu Kimata, 1 Chiaki Noga 2 & Tatsuo Watanabe 1,2,3 * 

1 Graduate School of Nutritional and Environmental Sciences, University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka 422-8526, Japan 

2 School of Food and Nutritional Sciences, University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka 422-8526, Japan 

3 Global COE Program, University of Shizuoka, 52-1 Yada, Suruga-ku, Shizuoka 422-8526, Japan 

(Received 25 February 2010; Accepted in revised form 30 April 2010) 

Summary There has been interest in biological activities of green tea catechins. However, little is known about the taste 

characteristics of catechins. To assess the taste characteristics of catechins (())–epicatechin (EC), ())– 

epicatechin gallate (ECg), ())–epigallocatechin (EGC) and ())–epigallocatechin gallate (EGCg)), sensory 

evaluations were performed. The taste intensity increased with increased catechins concentration. Among 

them, ECg showed the strongest taste. Catechins had tastes that consisted primarily of astringency and 

bitterness. Therefore, taste palatability decreased with increasing catechin concentrations. In line with taste 

intensity, taste palatability of ECg was the lowest. Taste sensor analysis and mouse behavioural assays 

showed same results. EC and ECg were more stable in Ringer’s solution than EGC and EGCg. Furthermore, 

taste cell responses to ECg that had the strong taste and better stability among catechins used were recorded. 

Some taste cells responded to ECg. This result suggests that ECg might be recognised by taste cells. 

Keywords Catechin, EC, ECg, EGC, EGCg, taste. 

Introduction 

Green tea has long been a popular drink in East Asia. 

Many active ingredients, such as polyphenols, vitamins 

and amino acids, are present in green tea (Graham, 1992). 

())–Epicatechin (EC), ())–epicatechin gallate (ECg), 

())–epigallocatechin (EGC) and ())–epigallocatechin 

gallate (EGCg) are major components of the polyphenols 

in green tea and are referred to as tea catechins. 

Recently, the pharmaceutical effects of catechins have 

attracted attention. According to recent studies, the 

catechins have anti-hypertensive (Henry & Stephens- 

Larson, 1984), anti-oxidative (Chen & Chan, 1996) 

(Zhang et al., 1997), anti-arteriosclerotic (Hertog et al., 

1993), anti-carcinogenic (Shi et al., 1994) (Wang et al., 

1994) and hypocholesterolemic properties (Imai & 

Nakachi, 1995). There are some studies that examined 

sensory characteristics of tea catechins (Scharbert & 

Hofmann, 2005) (Rossetti et al., 2009). Although their 

reports investigated taste threshold of catechins, the 

detailed taste analysis including the relationship of 

concentration and taste intensity and ⁄ or preference is 

not performed. Therefore, we performed more detailed 

taste analysis of tea catechins in this study. First, we 

investigated the taste of catechins using human sensory 


e-mail: watanbt@u-shizuoka-ken.ac.jp 

doi:10.1111/j.1365-2621.2010.02304.x 


evaluations. We also tried an objective assessment of 

taste by a mouse behavioural assay and an instrumental 

analysis. Furthermore, we recorded the responses of 

mouse taste cells that detect the taste of food, using the 

tea catechins as tastants. 


Materials 

EC, ECg, EGC and EGCg were provided by Mitsui 

Norin Co. Ltd. (Tokyo, Japan). Collagenase type IV 

and elastase were purchased from Worthington Biochemical 

(Lakewood, NJ, USA). Trypsin inhibitor, 

DNase, amiloride and DMSO were obtained from 

Sigma (St. Louis, MO, USA), and Fluo-4AM was 

purchased from Molecular Probes (Eugene, OR, USA). 

All other reagents were of analytical grade and were 

obtained from standard suppliers. 


We recruited 11 volunteers with a mean age of 

23.7 ± 3.7 years (mean ± standard deviation). All subjects 

had normal body mass, were non-smokers and 

were in good physical health. All psychophysical tests 

were performed according to a protocol approved by the 

University of Shizuoka Ethics Committee.

1580 

Taste characterisation of catechins M. Narukawa et al. 

Taste intensity was assessed using a visual analogue 

scale (VAS). The subjects were asked to spit out each 

solution after tasting and to rate the taste intensity by 

marking the appropriate position on a 100-mm VAS. 

The right and left ends of the scale represented ‘strong’ 

and ‘weak’ tastes, respectively. The palatability was 

assessed using a scoring method in which catechin 

solutions were graded on a scale of )3 to +3: )3, 

strongly unpleasant; )2, unpleasant; )1, slightly 

unpleasant; 0, neither pleasant nor unpleasant; +1, 

slightly pleasant; +2, pleasant; and +3, strongly 

pleasant. The subjects described the taste qualities in 

terms of no taste, bitterness, sourness, saltiness, sweetness, 

umami, astringency, pungency and other tastes. 

Because tea catechins is included at the range of 

10–300 lm in green tea drink (Chen et al., 2001), we 

prepared EC, ECg, EGC and EGCg solutions at four 

different concentrations, 10, 30, 100 and 300 lm. For the 

tasting test, the subjects were presented with cups 

containing 30 mL of a sample solution. All solutions 

were poured into plastic cups and presented at ambient 

temperature. For each tasting test, the subjects were 

asked to thoroughly rinse their mouth with distilled 

water, sip the test solution, swirl it around in their 

mouth for several seconds to taste it and then spit it out. 

The subjects rinsed their mouths with distilled water 

between tasting different samples. The sensory evaluation 

was performed in a laboratory for sensory evaluation; 

the temperature and humidity were 24 ± 1 °C 

and 54% ± 1%, respectively. 

Behavioural assay 

For animal experiments, C57BL ⁄ 6J strain mice (age, 7– 

20 weeks, supplier) were used. The animal experiments 

were performed according to a protocol approved by the 

University of Shizuoka Animal Care Committee. A 

48-h, two-bottle preference test was performed as a 

behavioural assay. The mice (n = 10) were caged 

individually and given access for 48 h to two bottles; 

one contained distilled water and the other, a tastant 

solution. After 24 h, the bottle positions were switched 

to control for any positional effect. The ratio of tastant 

volume to total liquid consumed was recorded for each 

tastant. The intake of each fluid in the two-bottle 

preference test was expressed as mean ± standard error 

of the mean (SEM), and their differences were evaluated 

statistically using the t-test. Tastant solutions were 10, 

30, 100, 300-lm EC, ECg, EGC and EGCg. 

Measurement of astringency intensity of catechins by a 

taste sensor 

A taste sensor assay referred to the procedure of 

Hayashi et al. (2008) and was performed using a TS- 

5000Z taste recognition system (Intelligent Sensor 

Technology, Inc., Kanagawa, Japan), fitted with an 

AE1 probe and reference probes. The taste sensor is a 

device that analyses the taste of food on the basis of the 

affinity to a lipid layer (Toko, 2000). As a standard 

solution, 30 mm KCl + 0.3 mm tartaric acid was used. 

Sample solutions were 300-lm EC, ECg, EGC and 

EGCg dissolved in the standard solution. 

HPLC analysis 

Degradation of green tea catechins of 48 h after preparation 

was quantified using high-performance liquid 

chromatography (HPLC). HPLC analysis referred to a 

slight modification of the procedure of Maeda-Yamamoto 

et al. (2005). Solutions of 300 lm of EC, ECg, EGC 

and EGCg were prepared in distilled water and Ringer’s 

solution followed by incubation at room temperature 

(25 °C). HPLC analysis was performed using an LC-6A 

or LC-9A system equipped with a SPD-6AV spectrophotometric 

detector (Shimadzu Corporation, Kyoto, 

Japan), and a JH-392 ODS column was used (internal 

diameter, 4.6 mm and length, 150 mm; YMC Co. Ltd., 

Kyoto, Japan). The mobile phase consisted of H2Omethanol-H3PO4 

(65:35:0.17) at a flow rate of 

1 mL min )1 . The detection wavelength was 272 nm for 

EC, ECg and EGCg and 242 nm for ECg. 

Measurement of intracellular calcium concentration 

Ca 2+ measurements were measured using Fluo-4 from 

isolated single taste cells of mouse circumvallate papillae 

as described previously (Narukawa et al., 2006, 2008, 

2009). Briefly, taste cells were isolated and plated on 

concanavalin A-coated glass-bottom dishes. Then, cells 

were loaded with the cell-permeable, Ca 2+ -sensitive 

fluorescent dye Fluo-4 AM (3 lm). Images of Fluo-4 

fluorescence were acquired at 6-s intervals by an 

Olympus Fluoview FV300 laser scanning confocal 

microscope (Olympus, Tokyo, Japan). The intracellular 

Ca 2+ changes were expressed as the relative fluorescence 

change, DF ⁄ F=(F–F0) ⁄ F0, where the resting fluorescent 

F0 was determined by averaging 10 images at the 

beginning of each experiment. A response was defined 

when peak DF ⁄ F increased by 10% or more from the 

previous fluorescence image while perfused with the 

stimulating solution. Cells were perfused with Ringer’s 

solution containing 100-lm ECg. The EGC was dissolved 

in Ringer’s solution just before use. All experiments 

were performed at room temperature. 

Our Ringer’s solution contained (in mm): 136.89 

NaCl, 5.37 KCl, 20 HEPES, 5.55 d-glucose, 0.44 

KH2PO4, 0.34 Na2HPO4, 1 CaCl2 and 2.5 probenecid 

at pH 7.4. The divalent cation-free Ringer’s solution 

contained (in mm): 65 NaCl, 26 NaHCO3, 2.5 NaH2- PO4, 1 EDTA.4Na, and 20 glucose and adjusted to pH 

7.4 with NaOH. 



Data were expressed as the mean ± standard errors of 

the mean (SEM). The statistical significance of differences 

was analysed by student t-test. The differences 

were considered to be significant at P < 0.05. 

Results 

Taste analysis of catechins by human sensory evaluation 

The relationship of catechin concentration–taste intensity, 

palatability and qualities are shown in Fig. 1. The 

taste intensity increased with increased catechin concentration 

(Fig. 1a). On the other hand, the taste palatability 

decreased with increased catechin concentration 

(Fig. 1b). At 300-lm catechin, the taste intensity of ECg 

was the strongest, followed by EGCg, EC and EGC. 

Their taste intensities were 76.1 ± 7.4, 43.8 ± 8.4, 

26.2 ± 8.2 and 21.3 ± 4.9, respectively. Similarly, the 

taste palatability of ECg was the lowest, followed by 

EGCg, EC and EGC. Catechins had a complex taste 

that consisted primarily of astringency and bitterness 

(Fig. 1c). There were a few subjects who felt saltiness 

and umami to EC. When taste quality was compared 

between ECg and EGCg, ECg had relatively stronger 

astringency and EGCg had stronger bitterness. Fiftyfive 

per cent of subjects felt astringency to 300-lm ECg. 

Seventy-three per cent of subjects felt bitterness to 

300-lm EGCg. Thus, ECg had the strongest and most 

unpleasant taste among the four tea catechins tested. 

Taste sensor analysis 

Membrane potential changes with astringency taste 

(AE1 sensor) were observed. The membrane potential 

change with ECg was the biggest, followed by EGCg, 

EC and EGC (Fig. 2). It was thought that ECg had the 

strongest astringency. 

Two-bottle preference test in mice 

Next, a 48-h, two-bottle preference test was performed 

as a behavioural assay (Fig. 3). The palatability of ECg 

decreased, concentration-dependently. The mice significantly 

avoided more than 30-lm ECg. For EGCg, 

although the palatability from 10 to 100-lm EGCg did 

not change but that of 300-lm EGCg decreased significantly. 

The palatability of EGC and EC did not change. 

Thus, mice may detect the taste of ECg most strongly in 

the catechins tested. 

Stability measurements of catechins by HPLC 

It is known that catechins are rapidly degraded around 

neutral pH (Zhu et al., 1997). Thus, we sought to 

(a) 100 

Taste intensity (VAS mm) 

(b) 

Taste preference 

(c) 

80 

60 

40 

20 

0 

0 

–1 

–2 

–3 

EC 

ECg 

EGC 

EGCg 

Taste characterisation of catechins M. Narukawa et al. 1581 

: EC 

: ECg 

: EGC 

: EGCg 

10 30 100 300 

Conc. (µM) 

Conc. (µM) 

10 30 100 300 

: EC 

: ECg 

: EGC 

: EGCg 

Taste quality 

0% 20% 40% 60% 80% 100% 

: Bitter, : Astrigency, : Umami, : Salty, : No-taste 

Figure 1 Taste intensity, palatability and quality of catechins. 

(a) Change in taste intensity at 10, 30, 100 and 300-lm catechins. 

(b) Change in taste palatability at 10, 30, 100 and 300-lm catechins. 

(c) Taste quality of 300-lm catechins. The values represent the 

mean ± SE (n = 11). 

confirm their effectiveness in the foregoing experiments. 

Also, to examine the state of catechins when used in a 

physiological experiment, we measured the stability of 


1582 


Output of AE1 sensor (mV) 

0 

–5 

–10 

–15 

–20 

EC ECg EGC EGCg 

Figure 2 Membrane potential change of astringency sensor to catechins 

(300 lm). Astringency is the membrane potential values from the 

AE1 sensor probe. The y-axis indicates that the lower the value is, the 

stronger the taste is. 

Preference ratio 

0.6 

0.4 

0.2 

* 

: EC 

: ECg 

:EGC 

: 

: EGCg 

*** 

catechins in distilled water and Ringer’s solution for 

48 h. Catechins were hardly degraded in distilled water 

for 48 h (Fig. 4a). On the other hand, in Ringer’s 

** 

**** 

0 

10 100 1000 

Conc. (uM) 

Figure 3 Change of palatability to catechins in mouse. The values 

represent the mean ± SEM (n = 10). *, **, *** and **** indicate 

P

ΔF/F 

10% 

60 s 

100 µM ECg 

Figure 5 Intracellular Ca 2+ Response to ECg (100 lm). Time course 

of the representative ECg-induced Ca 2+ response (DF ⁄ F). 

This result suggests that the taste of ECg is recognised by 

taste cells. 

Discussion 

The functionality of green tea has gained attention, and 

worldwide consumption has increased. Catechins are the 

main polyphenol in green tea. It is reported that 

polyphenols influence the quality in some foods. For 

example, it affects the flavour quality of cocoa beans 

(Noor-Soffalina et al., 2009). And its concentration has 

the negative influence on perceived quality in Longjing 

tea (Wang & Ruan, 2009). Therefore, it is possible that 

catechins are the functional components in green tea 

with an unpleasant flavour, hindering further consumption. 

To further increase consumption of green tea, it 

will be necessary to understand the taste characterisation 

of the catechins to explore methods for reducing the 

unpleasant flavour. In this study, we investigated the 

taste characteristics of representative tea catechins, EC, 

ECg, EGC and EGCg. 

Taste intensity of catechins increased, concentrationdependently. 

Among them, ECg showed the strongest 

taste. Although there was no difference in taste intensities 

of EC, EGC and EGCg at 10–100 lm, the taste of EGCg 

increased at 300 lm. Taste palatability was inversely 

related to taste intensity and decreased with increased 

catechin concentrations. Thus, it is thought that catechins 

have an aversive taste. In fact, when the taste 

quality was investigated, these compounds possessed 

aversive tastes, such as bitterness and astringency. To 

evaluate the taste of catechins objectively, we performed 

an instrumental analysis and a mouse behavioural assay. 

We performed an instrumental analysis using a taste 

sensor. The experiment using the taste sensor supported 

these results. The order of astringency sensor output 

for 300-lm catechins was as follows: ECg > EGCg 

> EC > EGC. In the catechin included in green tea, 

the order of taste intensity was as follows: ECg > EGCg 

> EC > EGC. Among the catechins in green tea, the 


amount of EGCg is highest, followed by EGC, ECg and 

EC (Goto & Nagashima, 1996) (Chen et al., 2001). 

Because the taste intensity of 300-lm ECg was about 

1.8-fold higher than that of EGCg, the influence of ECg 

in contributing to the taste of tea may be stronger than 

expected. A few subjects felt saltiness and umami to 

300-lm EC solution. As for this reason, it is thought that 

the subjects did not detect clear taste because taste 

intensity of 300-lm EC is weak rather than EC has 

saltiness and umami. In fact, the subjects who did not 

detect taste to EC existed 20%. And bitterness threshold 

to EC was reported about 1.5 mm (Rossetti et al., 2009). 

Next, we examined a taste preference test with mice. 

As we have described previously, palatability partially 

reflects taste intensity. The order of avoidance ratio of 

the 300-lm catechins was as follows: ECg > EGCg 

> EC > EGC. This corresponded with the result of 

the sensory evaluation. Thus, this experiment using mice 

may be able to predict experimental results in humans. 

Generally, it is reported that the higher the hydrophobicity 

of a bitter substance is, the higher its 

interaction with bitter receptor is (Kumazawa et al., 

1986). In the catechins used, ECg and EGCg, which 

have galloyl groups, had higher taste intensities than EC 

and EGC. The hydrophobicity is increased by the 

galloyl group. As a result, the interaction with the 

receptor is expected to be stronger. Thus, the taste 

intensity of ECg and EGCg would be expected to be 

strong. Also, there was a difference in taste intensity 

between ECg and EGCg and ⁄ or EC and EGC. The 

existence of the catechol group and the galloyl group 

may influence this. In comparison with the catechol 

group, the galloyl group has more than one hydroxy 

group. Thus, it was thought that the taste intensity 

decreased in ECg and EC because the hydrophobicity of 

EGCg and EGC, with the galloyl group, may be weaker. 

Because humans detected bitterness from catechins 

and mice recognised the taste of catechins, we tried to 

record the taste response to catechins. Because it is 

known that catechins readily degrade around neutral 

pH (Zhu et al., 1997), the stability in Ringer’s solution 

used for cell assays was examined. In Ringer’s solution, 

EC and ECg were more stable than EGCg and EGC. 

Thus, taste cell responses to ECg that had the strong 

taste and better stability, of the catechins used, were 

recorded. In taste cells isolated from mouse circumvallate 

papillae, there were cells that responded to ECg. 

Thus, a part of the taste of this compound may be 

transduced via taste cells. We performed Ca 2+ imaging 

in dispersed taste cells. Not all of the cells observed were 

taste cells. Taste cells are divided into four types. It is 

reported that some taste cells respond to bitter substances 

(Kataoka et al., 2008). Thus, the level of ECgresponsive 

cells in the measured cells may have been 

low. Also, this level of responsive cells might increase 

when ECg is at a concentration higher than 100 lm. 


1584 


Bitter compounds are detected by the T2R family, 

which consists of 25 members in humans and 34 

members in the mouse (Matsunami et al. 2000) (Adler 

et al., 2000) (Chandrashekar et al., 2000) (Mueller et al., 

2005). However, a taste receptor that can detect catechins 

has not yet been identified. In a study that 

investigated the interaction of the catechins with lipid 

bilayers using liposome systems, the order of affinity of 

the catechins was reported as follows: ECg > EGCg 

> EC > EGC (Hashimoto et al., 1999). This is the 

same order as the taste intensity of the catechins. In 

intracellular bitter taste transduction, pathways that do 

not go through receptors have also been reported: these 

pathways activate some signal components directly 

(Naim et al., 1994) (Cummings & Kinnamon, 1992) 

and ⁄ or generate phase boundary potentials by absorbing 

to lipid bilayer (Kumazawa et al., 1988). This 

indicates that the taste of catechins might not be 

transduced via taste receptors. Although astringency is 

interpreted as a taste separate from the sensation of 

bitterness (Schiffman et al., 1992), the details remain 

unclear. By further investigating the taste response to 

catechins, part of the astringency mechanism may 

become clearer. 


This work was supported by Shizuoka Prefecture 

Collaboration of Regional Entities for the Advancement 

of Technological Excellence, JST. 

References 

Adler, E., Hoon, M.A., Mueller, K.L., Chandrashekar, J., Ryba, N.J. 

& Zuker, C.S. (2000). A novel family of mammalian taste receptors. 

Cell, 100, 693–702. 

Chandrashekar, J., Mueller, K.L., Hoon, M.A. et al. (2000). T2Rs 

function as bitter taste receptors. Cell, 100, 703–711. 

Chen, Z.Y. & Chan, P.T. (1996). Antioxidative activity of green tea 

catechins in canola oil. Chemistry and Physics of Lipids, 82, 163–172. 

Chen, Z.Y., Zhu, Q.Y., Tsang, D. & Huang, Y. (2001). Degradation of 

green tea catechins in tea drinks. Journal of Agricultural and Food 

Chemistry, 49, 477–482. 

Cummings, T.A. & Kinnamon, S.C. (1992). Apical K + channels in 

Necturus taste cells. Modulation by intracellular factors and taste 

stimuli. Journal of General Physiology, 99, 591–613. 

Goto, T. & Nagashima, H. (1996). Contents of individual tea 

catechins and caffeine in Japanese green tea. Tea Research Journal, 

83, 21–28. 

Graham, H.N. (1992). Green tea composition, consumption, and 

polyphenol chemistry. Preventive Medicine, 21, 334–350. 

Hashimoto, T., Kumazawa, S., Nanjo, F., Hara, Y. & Nakayama, T. 

(1999). Interaction of tea catechins with lipid bilayers investigated 

with liposome systems. Bioscience, Biotechnology, and Biochemistry, 

63, 2252–2255. 

Hayashi, N., Chen, R., Ikezaki, H. & Ujihara, U. (2008). Evaluation of 

the umami taste intensity of green tea by a taste sensor. Journal of 

Agricultural and Food Chemistry, 56, 7384–7387. 

Henry, J.P. & Stephens-Larson, P. (1984). Reduction of chronic 

psychosocial hypertension in mice by decaffeinated tea. Hypertension, 

6, 437–444. 

Hertog, M.G., Feskens, E.J., Hollman, P.C., Katan, M.B. & Kromhout, 

D. (1993). Dietary antioxidant flavonoids and risk of coronary 

heart disease: the Zutphen Elderly Study. Lancet, 342, 1007–1011. 

Imai, K. & Nakachi, K. (1995). Cross sectional study of effects of 

drinking green tea on cardiovascular and liver diseases. BMJ, 310, 

693–696. 

Kataoka, S., Yang, R., Ishimaru, Y. et al. (2008). The candidate sour 

taste receptor, PKD2L1, is expressed by type III taste cells in the 

mouse. Chemical Senses, 33, 243–254. 

Kumazawa, T., Kashiwayanagi, M. & Kurihara, K. (1986). Contribution 

of electrostatic and hydrophobic interactions of bitter 

substances with taste receptor membranes to generation of receptor 

potentials. Biochimica et Biophysica Acta, 888, 62–69. 

Kumazawa, T., Nomura, T. & Kurihara, K. (1988). Liposomes as 

model for taste cells: receptor sites for bitter substances including N- 

C=S substances and mechanism of membrane potential changes. 

Biochemistry, 27, 1239–1244. 

Maeda-Yamamoto, M., Nagai, H., Suzuki, Y., Ema, K., Kanda, E. & 

Mitsuda, H. (2005). Changes in O-methylated catechin and chemical 

component contents of ‘Benifuuki’ green tea (Camellia sinensis L.) 

beverage under various extraction conditions. Food Science and 

Technology Research, 11, 248–253. 

Matsunami, H., Montmayeur, J.P. & Buck, L.B. (2000). A family of 

candidate taste receptors. Nature, 404, 601–604. 

Mueller, K.L., Hoon, M.A., Erlenbach, I., Chandrashekar, J., Zuker, 

C.S. & Ryba, N.J. (2005). The receptors and coding logic for bitter 

taste. Nature, 434, 225–229. 

Naim, M., Seifert, R., Nurnberg, B., Grunbaum, L. & Schultz, G. 

(1994). Some taste substances are direct activators of G-proteins. 

The Biochemical Journal, 297, 451–454. 

Narukawa, M., Mori, T. & Hayashi, Y. (2006). Umami changes 

intracellular Ca 2+ levels using intracellular and extracellular sources 

in mouse taste receptor cells. Bioscience, Biotechnology, and 


Narukawa, M., Kitagawa-Iseki, K., Oike, H., Abe, K., Mori, T. & 

Hayashi, Y. (2008). Characterization of umami receptor and 

coupling G protein in mouse taste cells. Neuroreport, 19, 1169–1173. 

Narukawa, M., Minamisawa, E. & Hayashi, Y. (2009). Signalling 

mechanisms in mouse bitter responsive taste cells. Neuroreport, 20, 

936–940. 

Noor-Soffalina, S.S., Jinap, S., Nazamid, S. & Nazimah, S.A.H. 

(2009). Effect of polyphenol and pH on cocoa maillard-related 

flavour precursors in a lipidic model system. International Journal of 


Rossetti, D., Bongaerts, J.H.H., Wantling, E., Stokes, J.R. & 

Williamson, A.-M. (2009). Astringency of tea catechins: more than 

an oral lubrication tactile percept. Food Hydrocolloids, 23, 1984– 

1992. 

Scharbert, S. & Hofmann, T. (2005). Molecular definition of black tea 

taste by means of quantitative studies, taste reconstitution, and 

omission. Journal of Agricultural and Food Chemistry, 53, 5377– 

5384. 

Schiffman, S.S., Suggs, M.S., Sostman, A.L. & Simon, S.A. (1992). 

Chorda tympani and lingual nerve responses to astringent compounds 

in rodents. Physiology & Behavior, 51, 51–63. 

Shi, S.T., Wang, Z.Y., Smith, T.J. et al. (1994). Effects of green tea and 

black tea on 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone bioactivation, 

DNA methylation, and lung tumorigenesis in A ⁄ J mice. 

Cancer Research, 54, 4641–4647. 

Toko, K. (2000). Taste Sensor. Pp. 113–180. U.K., Cambridge 

University Press. 

Wang, K. & Ruan, J. (2009). Analysis of chemical components in 

green tea in relation with perceived quality, A Case Study with 

Longjing Teas. International Journal of Food Science and Technology, 

44, 2476–2484. 

Wang, Z.Y., Huang, M.T., Lou, Y.R. et al. (1994). Inhibitory effects 

of black tea, green tea, decaffeinated black tea, and decaffeinated 

green tea on ultraviolet B light-induced skin carcinogenesis in 7,12- 


dimethylbenz[a]anthracene-initiated SKH-1 mice. Cancer Research, 

54, 3428–3435. 

Zhang, A., Zhu, Q.Y., Luk, Y.S., Ho, K.Y., Fung, K.P. & Chen, Z.Y. 

(1997). Inhibitory effects of jasmine green tea epicatechin isomers on 

free radical-induced lysis of red blood cells. Life Sciences, 61, 383– 

394. 

Zhu, Q.Y., Zhang, A., Tsang, D., Huang, Y. & Chen, Z.Y. (1997). 

Stability of green tea catechins. Journal of Agricultural and Food 

Chemistry, 45, 4624–4628. 


Additional Supporting Information may be found in the 



Figure S1. Half-period of catechins in Ringer’s solution 

(semi-logarithmic scale). 

Figure S2. Fluorescence image of the ECG-responsive 

cell that loaded with the Ca 2+ -sensitive dye, Fluo-4 AM. 

Scale bar = 10 lm. 


the content or functionality of any supporting materials 

supplied by the authors. Any queries (other than missing 

material) should be directed to the corresponding author 

for the article. 


1586 


Volatile composition and descriptive sensory analysis of Italian 

vanilla torrone 

Marisa Speziale, 1 Laura Vázquez-Araújo, 2 Antonio Mincione 1 &Ángel A. Carbonell-Barrachina 2 * 

1 Università degli Studi Mediterranea di Reggio Calabria, Dipartimento di Biotecnologie per il Monitoraggio Agro-alimentare ed Ambientale, 

C.da Melissari – III Lotto – 89124 Reggio Calabria, Italy 

2 Universidad Miguel Hernández, Departamento Tecnología Agroalimentaria, Universidad Miguel Hernández, Carretera de Beniel, km 3.2, 

03312-Orihuela, Alicante, Spain 


Summary Vanilla torrone is a honey- and almond-based confectionery product very famous in Italy, especially during 

Christmas. In this study, the effects of two factors, the manufacturing company and the manufacturing 

season, on the volatile composition of vanilla torrone were studied. Volatile compounds in torrone samples 

were extracted using simultaneous distillation-extraction and finally isolated, semi-quantified and identified 

using GC-MS. A trained panel was also used to make a descriptive analysis of torrone. Benzene derivatives 

and terpenes were the predominant chemical groups in torrone samples. The addition of artificial flavourings 

(mainly vanillin and benzaldehyde) played an important role in determining the final composition of volatiles 

and their concentrations. However, the amount and quality of almonds and honey and the intensity of 

almond toasting could also play important roles. If manufacturers want to provide a more homogeneous 

product to the consumers, more attention should be paid to the composition and concentrations of the 

flavourings added. 

Keywords Almond, aroma profile, descriptive sensory analysis, honey, Italy, turrón. 

Introduction 

Torrone is a typical Italian product, mainly consumed in 

Christmas time, made from almonds (and ⁄ or any other 

nuts such as pistachio or hazelnut), sugar, inverted 

sugar, honey and egg albumin, and manufactured in a 

traditional way in several Italian regions such as 

Calabria, Abruzzo, Campania, Lombardy, etc. This 

dessert it is included in a list of ‘Traditional Italian Food 

Products’ prepared by the Italian Ministry of Agriculture, 

Food and Forestry in cooperation with all the 

Italian regions and following the regulation 2081 ⁄ 92 

(CEE) from 14th of July 1982 related to the protection 

of geographical indications and origin designations. 

In the manufacturing of torrone in Reggio Calabria, 

honey and sugars (glucose and sucrose) are warmed in a 

copper container at low temperature and mixed at 

uniform speed for 2 h in the ‘torroniera’, where fresh egg 

whites have been added. After this time, the caramel 

(previously prepared in another copper container using 

honey and sugars), the natural vanilla extract, a handful 

*Correspondent: Fax: +34 966749677; 

e-mail: angel.carbonell@umh.es 


of pistachios and toasted almonds (48 h at 60 °C) are 

added. When all the ingredients are properly mixed, the 

mass is transferred to a marble table and manually 

mixed to confer the desired shape using wood frames or 

appropriate moulds. To obtain the so-called torrone 

vaniglia (vanilla) the product is covered on both surfaces 

with wafer, cut, and packed. On the other hand, the 

mixture could be covered by chocolate and then the 

product is called torrone gianduja (chocolate). However, 

there are composition and operational differences 

between these two types of torrone. For instance, in 

general the vanilla torrone has longer times of heating 

and a higher almond content, which leads to a harder 

texture because this type of torrone has less moisture 

than chocolate torrone. 

Several authors have studied in recent years the aroma 

composition from different types of nuts, including the 

toasted almonds (Dimick, 1994; Alasalvar et al., 2003; 

Krist et al., 2004; Va´zquez-Arau´jo et al., 2008a, 2009), 

and from different types of Spanish turroń, a similar 

product to the Italian torrone (Va´zquez et al., 2007; 

Vázquez-Arau´jo et al., 2008b,c; Narbona et al., 2010). 

However, there are no studies related to the aroma of 

torrone. 

doi:10.1111/j.1365-2621.2010.02307.x 


In the Italian torrone, almonds are raw in many cases 

or only slightly toasted, and pyrazines should not have 

such an important role. Besides, Italian torrone has 

artificial flavourings added and the main aromatic note 

expected is vanilla. 

In this work, our main objective was to determine the 

general volatile composition of vanilla torrone from the 

Reggio Calabria region as affected by the manufacturing 

company and the manufacturing season. The main 

reasons for studying the effects of these factors (company 

and season) are as follows: 

(1) The quality and price of the torrone ingredients change 

from year to year; therefore, it is expected that the 

quality of the torrone may also change. For instance, 

manufacturers can reduce the amount of almond if its 

price is too high. 

(2) Different producers will have different customers ⁄ consumers 

and they can adjust the quality and properties of 

their torrone to the specific demand of their consumers. 


Torrone 

Vanilla torrone samples were kindly provided by six 

different artisanal manufacturers of Taurianova (Reggio 

Calabria, Southern Italy). Samples were letter coded (A, 

B, C, D, E and F) to avoid using company names. The 

format of vanilla torrone chosen for this study was the 

smallest one, with individual bars weighing about 15 g 

(approximately 52 · 34 · 33 mm); 20–25 units per company 

were collected. Products were sampled during two 

consecutive manufacturing seasons: 2007 and 2008. 

Products were kept until analysis at 10 °C in a chamber 

with controlled temperature and moisture content. 

Simultaneous steam-distillation extraction 

Suspension of 30 g of each type of ground torrone were 

extracted for aroma volatiles using a Likens–Nickerson 

distillator (Afora, Barcelona, Spain), as described in 

Va´zquez-Arau´jo et al. (2008a,b,c). Briefly, 30 g of confection 

was extracted with 50 mL of dichloromethane, 

Cl2CH2. The distillation-extraction process was maintained 

for 120 min. The steam-distillation extraction 

(SDE) technique was selected because of high recoveries 

allied with low variability. Precaution was taken to 

ensure that no artefact compounds were included; only 

compounds also identified by hydro-distillation using a 

Deryng apparatus and ⁄ or solid phase micro-extraction, 

using a 50 ⁄ 30 lm DVB⁄ CAR ⁄ PDMS fibre, were 

included in the list of compounds under study (Alonso 

et al., 2009). 

As internal standard, 100 lL of 1-penten-3-one 

(1 g L )1 ) was added in the sample flask; this chemical 

Aroma of vanilla torrone M. Speziale et al. 1587 

was used as internal standard after checking that it was 

absent in Italian torrone samples and, under the 

proposed conditions, it separates well from other 

volatile compounds. Experimental recovery percentage 

for 1-penten-3-one in SDE was 97.4%. 

After the 120 min of extraction, the organic solvent 

was dried over anhydrous sodium sulphate, Na2SO4, 

and concentrated to 0.4 mL. SDE experiments were run 

in triplicate. 

GC-MS analytical conditions 

The isolation, identification and quantification of the 

volatile compounds were performed on a gas chromatograph, 

Shimadzu GC-17A (Shimadzu Corporation, 

Kyoto, Japan), coupled with a Shimadzu mass spectrometer 

detector GCMS QP-5050A. The GC-MS 

system was equipped with a TRACSIL Meta.X5 column 

(Teknokroma S. Coop. C. Ltd, Barcelona, Spain; 

30 m · 0.25 mm · 0.25 lm film thickness). Analyses 

were carried out using conditions previously described 

by Vázquez-Arau´jo et al. (2008a,b,c, 2009). 

Most of the compounds were identified using three 

different analytical methods: (i) Kovats indices, 

(ii) GCMS retention indices (authentic chemicals) and 

(iii) mass spectra (authentic chemicals and Wiley spectral 

library collection) (McLafferty, 2000). Identification 

was considered tentative when it was based only on 

mass-spectral data (Table 1). 

For the semi-quantification of the volatile compounds, 

1-penten-3-one was used as internal standard. 

Data included in this study should be considered as 

semi-quantitative, because no standard curves were 

carried out for each one the quantified volatile compounds. 

Sensory evaluation with trained panel 

Sensory evaluation, with trained panel, was used to 

discriminate and quantify main sensory attributes of 

torrone samples manufactured in 2008. A panel of ten 

panellists, aged 24–53 (five women and five men, all 

members of the panel of the Regulating Council of the 

Specific Denominations of Jijona and Turroń de Alicante, 

RCSDJTA), with wide experience in sensory 

evaluation of turroń, chocolate and derivates, evaluated 

the six trademarks of vanilla torrone. 

The panel was selected and trained following the ISO 

standard 8586-1 (1993) standard. Details of panel 

selection, training and validation could be found in 

Vázquez-Arau´jo et al. (2005). Descriptive sensory analysis 

(DSA) has been successfully used for comparing 

odour and taste attributes in foods and their products 

(Alasalvar et al., 2003; Vázquez-Arau´jo et al., 2008b; 

Narbona et al., 2010). Torrone samples were assessed 

using a flavour profile method (Meilgaard et al., 1999). 


1588 

Aroma of vanilla torrone M. Speziale et al. 

Table 1 Volatile compounds studied with their identification method, Kovats indices for this study and literature values, odour thresholds in water 

and aroma descriptors 

Compound RT KI (experimental) KI (literature) † OT (mg kg )1 ) à 

Prior to DSA, panellists discussed the flavour properties 

of torrone during two preliminary orientation 

sessions, each lasting 90 min, until they had agreed on 

their use of flavour attributes. During these orientation 

experiments, panellists evaluated three different coded 

samples of torrone from different manufacturing companies. 

Twelve appearance, flavour and texture attributes 

were identified, and standards were made available 

for panellists (Table 2). 

Measurements were performed in individual booths 

with controlled illumination (70–90 fc) and temperature 

(23 ± 2 °C) (AENOR, 1997). 

The individual products were scored for the intensity 

of the sensory parameters on a scale of 0–10, where: 

Descriptor 

3-Hydroxy-2-butanone 3.75 743 743 0.8 

Hexanal § 

4.79 764 800 0.0045 Fatty, fruity, green 

Dihydro-2-methyl-3(2H)-furanone § 

6.6 801 804 

2-Hexanol 7.0 809 810 0.006 

Furfural § 

7.8 826 831 3 Sweet, woody, almond, fragrant, baked bread 

Furfuryl alcohol 9.3 856 860 1.9 Low odour, cooked sugar taste 

1-Hexanol 9.9 868 872 0.5 

2-Heptanone 10.7 885 891 0.14 Fruity, spicy, cinnamon 

Heptanal § 

11.3 896 900 0.003 Oily, woody, fatty, fruity, nutty 

2-Acetylfuran 11.8 906 910 Almond, beef, caramel, musty, coffee, 

potato, tobacco 

a-Pinene § 

13.0 926 930 0.006 Woody 

Benzaldehyde § 

14.8 960 960 0.35 Bitter almond, fragrant, aromatic, sweet 

b-Pinene § 

15.4 970 971 0.14 Woody 

1-Heptanol 15.6 973 970 0.425 

2-Amylfuran 16.4 988 988 

Octanal § 

17.1 1000 1000 0.0007 Fatty, citrus, honey on dilution 

Limonene § 

18.4 1025 1025 1 Mild, citrus, orange, lemon 

Benzyl alcohol § 

19.3 1040 1032 20 Berry, cherry, grapefruit, citrus, walnut 

b-Ocimene 20.1 1056 1055 

Linalool oxide 21.0 1072 1086 0.32 

Nonanal § 

22.8 1104 1102 0.001 Floral, citrus, orange, rose, fatty, waxy 

Phenetylalcohol 23.4 1116 1113 Honey, floral, rose 

E-2-Nonenal 25.7 1161 1158 0.00008 Penetrating, fatty, waxy, vegetable, earthy 

1-Nonanol 26.6 1178 1174 Citrus, rose 

Neral 29.9 1243 1249 Lemon 

2-Decenal 31.0 1266 1260 0.0004 Orange, slightly fatty, floral, earthy, chocolate 

(E, Z)2,4-decadienal – 

32.3 1292 1284 0.00007 Powerful, fatty, citrus 

(E, E)2,4-decadienal § 

33.7 1323 1330 0.00007 Powerful, fatty, citrus 

Methyl anthranilate 34.82 1346 1354 0.003 Chocolate, grape, jasmine, lemon, fruity, 

herbaceous, floral 

E-2-Undecenal 35.87 1369 1371 Orange, fruity, herbaceous 

Vanillin § 

37.87 1412 1406 Vanilla, caramel 

Farnesene 41.77 1499 1505 Apple, lavender, lime, green, woody, herbaceous 

Ethyl citrate – 

48.84 1669 Wine-like, plum, sweet 

† NIST (2009). 

à Buttery et al. (1999), Serra Bonvehi (2005), and Pino & Mesa (2006). 

§ Authentic standards were used. 

– Tentatively identified (only Wiley library). 

• 0 = no appreciation of this parameter; 

• 5 = moderate intensity (regular intensity in this parameter); 

• 10 = extremely intense. 

A complete block design was made and torrone 

samples were presented (coded with a three-digit 

random number), one by one, following a Williams 

latin squared design balanced for order and first-order 

carry-over effects (Meilgaard et al., 1999). Panellists 

relied on their training experience to score products. The 

entire experiment was repeated three times (all judges 

scored all six samples on each session for a total of 

three sessions), and the sensory scores were presented as 

the overall means. 


Table 2 Attributes selected for descriptive sensory analysis 

Attribute Characteristics References 


All data were subjected to analysis of variance (anova) 

and the Tukey’s least significant difference multicomparison 

test to determine significant differences among 

torrone samples as affected by the manufacturing 

company and season. Significant differences were 

reported for P £ 0.05. The statistical analyses were 

carried out using SPSS 14.0 (SPSS Science, Chicago, IL, 

USA) and figures using Sigma Plot 9.0 (SPSS Science). 


Instrumental analysis of volatile compounds 

The suitability of SDE for analysing the volatile 

composition of almond-honey confection was previously 

discussed (Va´zquez-Arau´jo et al., 2008a,b). Preliminary 

tests were carried out using deodorised almond 

matrices (extracted with 1:1 pentane-diethyl ether for 

24 h at 40 °C) to study reproducibility and recovery of 

volatile compounds. The compounds assayed included 

2-methylpyrazine, 2,5-dimethylpyrazine, 2-ethyl-3-methylpyrazine, 

furfuryl alcohol and 2-acetylpyrrole. Their 

recoveries after 120 min of distillation-extraction ranged 

from 76.3% for 2-acetylpyrrole up to 99.9% 2-methylpyrazine, 

with a mean value of 86.7%; the standard 

error of measurements was always D > E > B > F > C. 

Vázquez et al. (2007) studied the changes in volatile 

compounds during concentration of honey and sugars in 

Alicante and Jijona turroń and the concentration of total 

volatile chemicals found was 174 mg kg )1 (only turroń 

mass was studied and no toasted almonds were under 

consideration at this stage). Later, Vázquez-Arau´jo 

et al. (2008c) studied the aromatic compounds of 

Alicante and Jijona turroń and reported values 

207 ± 36 and 194 ± 26 mg kg )1 , respectively. In general, 

turroń seemed to present higher total concentrations 

of volatiles, although some of the torrone samples 

A-2007 (company-season), D-2007 and E-2007 were at 

the same level. 

Thirty-three compounds were isolated, identified and 

quantified in the volatile profile of vanilla torrone 

(Table 1). These compounds can be grouped in seven 

chemical families: aldehydes (ten compounds), benzene 

derivatives (four compounds), furans (five compounds), 

terpenes (six compounds), alcohols (four compounds), 

ketones (two compounds) and esters (two compounds) 

(Table 3); these chemical groups contributed to the total 

concentration of volatiles in the following manner: 

8.5%, 78.7%, 5.8%, 4.0%, 2.1%, 0.5% and 0.4%, 

respectively. 

The most abundant group was benzene derivatives 

(Fig. 1), because of the high concentrations of three 

compounds: benzaldehyde, benzyl alcohol and vanillin. 

It seems from the volatile profile that some compounds 

could be artificially present in this type of confection, 

especially benzaldehyde, which is a natural compound 

present in almonds, but not as such high concentrations 

(Va´zquez-Arau´jo et al., 2008a, 2009). 

The facts that no pyrazines and pyrroles were detected 

and that only low concentrations of furans were present 

in this type of torrone indicated that almonds are used 

without any toasting step. Vázquez-Arau´jo et al. (2008c) 


1590 


Table 3 Volatile compounds (mg ⁄ kg) found in six manufacturing companies of vanilla torrone 

A B C D E F 

Compound 

2007 2008 2007 2008 2007 2008 2007 2008 2007 2008 2007 2008 

Aldehydes 

Hexanal 8.56 ± 0.06 † 

0.34 ± 0.08 1.03 ± 0.19 0.49 ± 0.20 0.16 ± 0.01 0.30 ± 0.03 0.43 ± 0.02 0.14 ± 0.05 2.46 ± 0.34 0.19 ± 0.06 3.94 ± 0.34 0.28 ± 0.07 

Heptanal 0.43 ± 0.01 0.01 ± 0.01 0.10 ± 0.01 0.03 ± 0.01 0.01 ± 0.01 0.02 ± 0.01 0.02 ± 0.01 0.01 ± 0.01 0.50 ± 0.03 0.04 ± 0.01 0.34 ± 0.02 0.02 ± 0.01 

Octanal 0.54 ± 0.01 0.02 ± 0.01 0.30 ± 0.06 0.08 ± 0.05 0.01 ± 0.01 0.03 ± 0.01 0.07 ± 0.01 0.01 ± 0.01 1.16 ± 0.01 0.12 ± 0.03 0.35 ± 0.09 0.02 ± 0.01 

Nonanal 5.54 ± 0.61 0.81 ± 0.35 1.14 ± 0.01 0.74 ± 0.33 0.12 ± 0.01 0.49 ± 0.13 0.57 ± 0.02 0.51 ± 0.08 3.03 ± 0.08 0.42 ± 0.08 2.81 ± 0.02 0.61 ± 0.07 

trans-2-Nonenal 1.27 ± 0.14 0.12 ± 0.04 0.30 ± 0.01 0.58 ± 0.31 0.02 ± 0.01 0.10 ± 0.01 0.36 ± 0.03 0.18 ± 0.05 0.86 ± 0.02 0.09 ± 0.08 3.55 ± 0.28 0.15 ± 0.01 

Citral 0.10 ± 0.01 n.d. 0.02 ± 0.01 n.d. n.d. n.d. 0.05 ± 0.01 0.01 ± 0.01 0.22 ± 0.03 n.d. 0.07 ± 0.03 n.d. 

2-Decenal 0.70 ± 0.04 0.07 ± 0.02 0.33 ± 0.09 0.14 ± 0.08 0.03 ± 0.01 0.08 ± 0.01 0.15 ± 0.03 0.03 ± 0.01 0.71 ± 0.01 0.09 ± 0.04 0.70 ± 0.07 0.06 ± 0.01 

(E, Z) 2,4-Decadienal 7.60 ± 1.28 0.18 ± 0.05 0.67 ± 0.03 0.41 ± 0.10 0.15 ± 0.01 0.18 ± 0.07 0.32 ± 0.10 0.05 ± 0.01 1.62 ± 0.28 0.20 ± 0.08 1.06 ± 0.02 0.19 ± 0.01 

(E, E) 2,4-Decadienal n.d. 0.26 ± 0.26 n.d. 1.29 ± 0.42 0.50 ± 0.02 0.71 ± 0.23 0.76 ± 0.20 0.26 ± 0.02 1.02 ± 0.12 n.d. 4.96 ± 0.52 n.d. 

E-2-Undecenal 0.67 ± 0.07 n.d. 0.29 ± 0.08 0.16 ± 0.09 0.03 ± 0.01 n.d. 0.08 ± 0.01 0.03 ± 0.01 0.68 ± 0.10 0.11 ± 0.01 0.36 ± 0.33 0.07 ± 0.01 

25.4 ± 2.2 1.82 ± 0.82 4.18 ± 0.49 3.92 ± 1.59 1.01 ± 0.10 1.91 ± 0.50 2.80 ± 0.44 1.22 ± 0.26 12.3 ± 1.0 1.26 ± 0.39 18.1 ± 1.72 1.40 ± 0.20 

Benzene derivates 

Benzaldehyde 45.0 ± 2.3 2.86 ± 0.80 26.6 ± 2.8 25.2 ± 11.9 12.4 ± 0.6 4.95 ± 0.11 105 ± 13 23.8 ± 0.7 67.4 ± 10.0 5.00 ± 2.69 16.0 ± 2.7 18.9 ± 2.2 

Benzyl alcohol 12.0 ± 1.9 0.84 ± 0.44 1.53 ± 0.15 0.44 ± 0.04 0.37 ± 0.02 2.29 ± 0.42 2.55 ± 0.25 1.21 ± 0.22 5.10 ± 0.19 0.91 ± 0.22 10.2 ± 0.01 1.28 ± 0.26 

Phenetylalcohol 1.66 ± 0.12 n.d. 0.37 ± 0.21 0.01 ± 0.01 0.09 ± 0.01 0.16 ± 0.05 0.49 ± 0.04 0.08 ± 0.02 0.45 ± 0.01 0.08 ± 0.05 1.57 ± 0.43 0.20 ± 0.02 

Vanillin 145 ± 33 58.9 ± 3.0 13.1 ± 6.5 29.6 ± 13.4 1.17 ± 0.06 13.2 ± 0.1 9.61 ± 0.62 11.2 ± 2.1 2.43 ± 0.35 3.25 ± 1.00 11.7 ± 2.6 1.28 ± 0.29 

203 ± 37 62.6 ± 4.2 41.6 ± 9.8 55.2 ± 25.4 14.0 ± 0.7 20.6 ± 0.7 118 ± 14 36.3 ± 3.1 75.4 ± 10.6 9.24 ± 3.97 39.4 ± 5.7 21.7 ± 2.8 

Furans 

Dihydro-2-methyl- 1.87 ± 0.99 0.07 ± 0.03 0.20 ± 0.09 0.03 ± 0.01 0.06 ± 0.01 0.02 ± 0.01 0.08 ± 0.02 0.03 ± 0.01 0.44 ± 0.09 0.02 ± 0.01 0.90 ± 0.13 0.05 ± 0.02 

3(2H)-furanone 

Furfural 16.3 ± 2.8 0.56 ± 0.42 1.81 ± 1.29 0.28 ± 0.14 0.32 ± 0.02 0.04 ± 0.04 0.58 ± 0.22 0.23 ± 0.13 2.77 ± 0.71 0.26 ± 0.13 3.19 ± 0.49 0.41 ± 0.22 

Furfuryl alcohol 9.19 ± 3.41 0.43 ± 0.30 1.50 ± 1.11 0.18 ± 0.11 0.28 ± 0.01 0.04 ± 0.04 0.65 ± 0.47 0.19 ± 0.11 0.27 ± 0.17 0.06 ± 0.05 1.44 ± 0.18 0.34 ± 0.12 

2-Acetylfuran 1.69 ± 0.79 n.d. 0.23 ± 0.18 n.d. 0.04 ± 0.01 n.d. 0.25 ± 0.12 n.d. 0.23 ± 0.13 0.01 ± 0.01 0.39 ± 0.16 0.03 ± 0.01 

2-Amylfuran 0.99 ± 0.03 0.08 ± 0.03 0.21 ± 0.07 0.15 ± 0.06 0.03 ± 0.01 0.11 ± 0.01 0.14 ± 0.02 0.04 ± 0.01 0.73 ± 0.08 0.04 ± 0.03 0.78 ± 0.29 0.11 ± 0.01 

30.1 ± 8.0 1.14 ± 0.78 3.95 ± 2.74 0.63 ± 0.32 0.74 ± 0.06 0.21 ± 0.10 1.69 ± 0.85 0.50 ± 0.26 4.44 ± 1.18 0.39 ± 0.23 6.70 ± 1.25 0.95 ± 0.38 

Terpenes 

a-Pinene 0.41 ± 0.01 0.41 ± 0.12 0.03 ± 0.01 0.05 ± 0.02 n.d. 0.03 ± 0.01 0.06 ± 0.01 0.02 ± 0.01 0.26 ± 0.01 0.04 ± 0.01 0.22 ± 0.06 0.02 ± 0.01 

b-Pinene 0.14 ± 0.05 0.10 ± 0.01 0.04 ± 0.01 0.02 ± 0.01 0.01 ± 0.01 0.06 ± 0.01 n.d. 0.01 ± 0.01 0.17 ± 0.02 0.01 ± 0.01 0.18 ± 0.09 0.04 ± 0.01 

Limonene 6.77 ± 0.98 0.67 ± 0.26 1.56 ± 0.09 0.48 ± 0.19 0.23 ± 0.01 1.14 ± 0.18 2.25 ± 0.27 0.30 ± 0.01 6.71 ± 0.49 0.36 ± 0.14 5.53 ± 1.26 0.39 ± 0.04 

b-Ocimene 0.89 ± 0.14 0.06 ± 0.01 0.18 ± 0.01 n.d. 0.03 ± 0.01 0.10 ± 0.01 0.19 ± 0.02 0.01 ± 0.01 0.78 ± 0.01 0.02 ± 0.01 0.62 ± 0.19 0.05 ± 0.01 

Linalool oxide 2.49 ± 0.24 0.10 ± 0.01 0.23 ± 0.14 0.01 ± 0.01 0.01 ± 0.01 0.04 ± 0.01 0.14 ± 0.08 0.01 ± 0.01 0.32 ± 0.03 0.03 ± 0.01 0.31 ± 0.07 0.01 ± 0.01 

Farnesene 0.02 ± 0.02 n.d. n.d. n.d. n.d. n.d. 0.05 ± 0.01 n.d. 0.06 ± 0.06 n.d. n.d. 0.01 ± 0.01 

10.7 ± 1.4 1.34 ± 0.41 2.04 ± 0.26 0.56 ± 0.23 0.29 ± 0.04 1.37 ± 0.22 2.69 ± 0.39 0.36 ± 0.05 8.29 ± 0.62 0.46 ± 0.18 6.86 ± 1.67 0.53 ± 0.09 

Alcohols 

2-Hexanol 2.79 ± 0.37 n.d. 0.34 ± 0.01 n.d. 0.09 ± 0.01 n.d. 0.10 ± 0.03 n.d. 0.09 ± 0.02 n.d. 0.57 ± 0.43 n.d. 

1-Hexanol 7.33 ± 0.25 0.01 ± 0.01 0.22 ± 0.03 0.09 ± 0.01 0.02 ± 0.01 0.21 ± 0.09 0.18 ± 0.01 n.d. 0.47 ± 0.07 0.02 ± 0.01 0.84 ± 0.37 0.06 ± 0.03 

1-Heptanol 1.03 ± 0.24 n.d. 0.10 ± 0.03 n.d. n.d. 0.01 ± 0.01 n.d. n.d. 0.49 ± 0.01 n.d. 0.18 ± 0.01 n.d. 

1-Nonanol 1.21 ± 0.02 0.02 ± 0.01 0.20 ± 0.07 0.03 ± 0.01 0.02 ± 0.01 0.06 ± 0.01 0.10 ± 0.01 0.01 ± 0.01 0.61 ± 0.02 0.02 ± 0.02 0.78 ± 0.03 0.09 ± 0.01 

12.4 ± 0.9 0.03 ± 0.02 0.86 ± 0.14 0.12 ± 0.02 0.13 ± 0.03 0.28 ± 0.11 0.37 ± 0.05 0.01 ± 0.01 1.67 ± 0.12 0.04 ± 0.03 2.38 ± 0.84 0.15 ± 0.04 


Table 3 (Continued) 

A B C D E F 

Compound 2007 2008 2007 2008 2007 2008 2007 2008 2007 2008 2007 2008 

Ketones 

3-Hydroxy- 1.22 ± 0.40 0.01 ± 0.01 0.14 ± 0.01 0.01 ± 0.01 0.04 ± 0.01 0.01 ± 0.01 0.14 ± 0.02 0.02 ± 0.01 0.18 ± 0.06 0.01 ± 0.01 0.52 ± 0.14 n.d. 

2-butanone 

2-Heptanone 1.25 ± 0.92 0.05 ± 0.02 0.08 ± 0.02 0.01 ± 0.01 n.d. n.d. 0.09 ± 0.03 n.d. 0.30 ± 0.04 0.03 ± 0.01 0.18 ± 0.05 0.03 ± 0.01 

2.48 ± 1.32 0.06 ± 0.03 0.21 ± 0.03 0.02 ± 0.02 0.04 ± 0.01 0.01 ± 0.01 0.22 ± 0.05 0.02 ± 0.01 0.47 ± 0.10 0.03 ± 0.02 0.70 ± 0.19 0.03 ± 0.01 

Esters 

Methyl n.d. n.d. 0.29 ± 0.14 0.23 ± 0.01 0.03 ± 0.01 0.27 ± 0.01 n.d. n.d. 1.36 ± 0.38 0.54 ± 0.08 n.d. n.d. 

anthranilate 

Ethyl citrate 0.03 ± 0.01 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 0.58 ± 0.33 n.d. n.d. 

0.03 ± 0.01 n.d. 0.29 ± 0.14 0.23 ± 0.01 0.03 ± 0.01 0.27 ± 0.01 n.d. n.d. 1.36 ± 0.38 1.13 ± 0.41 n.d. n.d. 

Total 284 ± 51 67.0 ± 6.2 52.8 ± 13.6 60.5 ± 27.6 16.2 ± 1.0 24.3 ± 1.7 125 ± 16 38.4 ± 3.7 104 ± 25 12.5 ± 5.2 74.2 ± 10.3 24.7 ± 3.5 

† Values reported are the mean of three replicates (±standard error). 

Concentration (mg kg –1 ) 

100 

80 

60 

20 

10 

0 

Aldehydes 

Benzene derivatives 

Furans 

Terpenes 

Alcohols 

Ketones 

Esters 

Total 

Figure 1 Grouping of volatile compounds according to their 

chemical nature in vanilla torrone. Values represented here are the 

mean of six manufacturing companies and two seasons. 

concluded that during manufacturing of Jijona turroń, 

the most complex (more thermal steps) Spanish confection, 

the concentrations of benzene derivatives and 

terpenes decreased while concentrations of pyrazines, 

furans and pyrroles increased. 

Descriptive sensory analysis 

Because this is the first study on vanilla torrone, the most 

important thing is to characterise this typical Italian 

Christmas confection. Figure 2 summarises its sensory 

profile, which has higher values of vanilla flavour, 

Solubility 

Cohesiveness 

Fracturability 

Adhesiveness 

Hardness 


Colour 

10 

8 

6 

4 

2 

0 

Sweet 

Alm. amount 

Vanilla flav. 

Alm. size 

Almond flav. 

Honey flav. 

Figure 2 Descriptive sensory analysis of appearance, flavour and 

texture attributes of vanilla torrone. Values represented here are the 

mean of six manufacturing companies and ten panellists. 


1592 


Table 4 Descriptive sensory analysis of vanilla torrone samples from the 2008 season 

Variation source Colour 

ANOVA test 

Assessor NS † 

Almond Flavour 

Amount Size Almond Honey Vanilla Sweetness Hardness Fracturability Cohesiveness Solubility Adhesiveness 

sweetness, cohesiveness and adhesiveness, but lower 

values of almond amount, almond flavour (especially 

toasted notes) than Spanish turroń (Va´zquez-Arau´jo 

et al., 2008b; : Va´zquez-Arau´jo et al., 2008c, 2006). 

These differences are attributed to the following factors: 

(1) formulation: in Spanish turroń, no vanilla flavouring can 

be added; the amount of almond is higher because of the 

legal requirements, and therefore the amount of sugars 

and sweetness are lower, and 

(2) manufacturing conditions: the toasting of almonds is 

softer in Italian torrone and this product is intended to 

be more cohesive and chewable than Spanish turroń 

(Speziale, 2010). 

Intensities for a number of organoleptic attributes 

(‘amount of almond’, ‘sweetness’, ‘hardness’) were not 

significantly (P > 0.05) affected by the manufacturing 

company (only samples from the 2008 season were used 

for sensory studies) (Table 4). 

The colour of samples from companies F and C was 

white and changed towards a more yellowish-white in B 

and E samples (Table 3). Samples from the company D 

had a higher amount of almonds but of a smaller size 

compared to all other samples. Regarding texture, 

samples of companies E and F were different from the 

other ones because they presented higher values of 

cohesiveness and adhesiveness, and lower of fracturability 

and solubility in saliva. 

As expected, the most relevant flavour attributes 

selected by the trained panel to describe vanilla torrone 

were almond, honey and vanilla flavours (Table 4). This 

selection could be expected after seeing the sensory 

descriptors and odour thresholds included in Table 1 

and the concentrations from Table 3. The almond 

flavour obtained a mean value for all companies of 

about 3.2 ± 0.3 (5 = moderate intensity) (Table 4), 

implying once again that almost were not toasted at all 

Texture 

NS NS NS NS NS NS NS NS NS NS NS 

Company * NS *** NS * NS NS NS ** *** *** *** 

Tukey’s multiple range test 

A 3.9 à ab § 2.0 2.0 ab 3.8 2.9 a 6.4 5.9 4.8 ab 5.5 ab 5.1 b 5.4 a 5.1 bc 

B 4.9 a 2.6 2.0 ab 3.1 2.3 ab 5.9 5.8 4.4 ab 5.6 a 4.4 b 5.9 a 3.4 c 

C 1.9 b 1.9 2.3 ab 2.5 1.3 b 4.6 4.9 2.7 b 4.3 ab 5.3 b 4.4 ab 5.4 bc 

D 3.9 ab 4.7 1.1 b 3.3 1.8 ab 5.4 5.6 4.4 ab 5.2 ab 5.2 b 4.5 ab 4.7 bc 

E 4.7 ab 2.6 2.6 a 3.5 2.4 ab 4.1 5.2 5.5 a 2.9 b 7.8 a 2.6 b 8.3 a 

F 1.9 b 1.9 1.6 ab 2.9 1.7 ab 4.7 4.9 4.0 ab 2.9 b 5.8 ab 3.9 ab 5.9 b 

*, **, and ***, significant at P < 0.05, 0.01, and 0.001, respectively. 

† NS = not significant F ratio (P < 0.05). 

à Values are the mean of ten panellists (three sessions). 

§ Values followed by the same letter, within the same variation source, were not significant different (P < 0.05), Tukey’s multiple-range test. 

or just slightly toasted. On the other hand, and as 

expected, the intensity of honey flavour was low, 

2.0 ± 0.3 (Table 4), because the almond and especially 

vanilla notes will mask the perception of the honey 

ones; besides, honey is usually present in torrone at 

concentrations lower than 10%. Finally, the highest 

flavour notes were those of vanilla (mean of 5.2 ± 0.4) 

(Table 4), which was expected after seeing the high 

contents of vanillin in some of the torrone samples, for 

instance A-2007. The intensity of the vanilla flavour 

was quantified at 5.2 but comparing to other vanilla 

torrone, without any doubt the vanilla intensity in this 

type of torrone is much higher than in any other type of 

turroń, torrone or nougat. The higher values of the 

intensities of ‘vanilla flavour’ in torrone samples 

from companies A and B (season 2008) were without 

any doubt related to the high concentration of the 

group benzene derivatives (especially vanillin and 

Total volatile compounds (mg kg –1 ) 

80 

60 

40 

20 

Total volatile compounds 

Total benzene derivatives 

R 2 = 0.9778 

R 2 = 0.9877 

0 

0 

3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 

Sensory vanilla 

Figure 3 Relationship among total volatile compounds (mg ⁄ kg) and 

total benzene derivatives (mg ⁄ kg) with the intensity of the sensory 

attribute ‘vanilla flavour’. 

International Journal of Food Science and Technology 2010 Ó 2010 The Authors. Journal compilation Ó 2010 Institute of Food Science and Technology 

80 

60 

40 

20 

Total benzene derivatives (mg kg –1 )

enzaldehyde) and thus the total volatile concentrations. 

Significant positive correlations (P > 0.001) were 

found for both relationships with coefficients of determination 

higher than 0.97 (Fig. 3). 

Conclusions 

The artisan nature of the Italian vanilla torrone was 

evident in the fact that the total concentrations of 

volatiles were very heterogeneous and were significantly 

affected by both the manufacturing companies and the 

manufacturing season. For example, the total concentration 

could take a value >100 mg kg )1 on the season 

2007 and decreased as low as 12.5 mg kg )1 in the 

following season. These important differences in the 

volatile composition of torrone could be because of 

different reasons: (i) the addition of different artificial 

aromas (mainly vanillin and benzaldehyde) and at 

different concentrations by each company and depending 

on the quality of the raw ingredients, and (ii) the 

toasting of almonds, and (iii) the quality and amount of 

both almonds and honey. 

References 

AENOR (Asociación Española de Normalizacioń y Certificación). 

(1997). Sensory Analysis. UNE Standards [In Spanish]. Madrid 

(Spain): AENOR. 

Alasalvar, C., Shahidi, F. & Cadwallader, K.R. (2003). Comparison of 

natural and roasted Turkish tombul hazelnut (Corylus avellana L.) 

volatiles and flavor by DHA ⁄ GC ⁄ MS and descriptive sensory 

analysis. Journal of Agricultural and Food Chemistry, 51, 5067–5072. 

Alonso, A., Vázquez-Arau´jo, L., García-Martínez, S., Ruiz, J.J. & 

Carbonell-Barrachina, A.A. (2009). Volatile compounds of traditional 

and virus-resistant breeding lines of Muchamiel tomatoes. 

European Food Research and Technology, 230, 315–323. 

Buttery, R.G., Orts, W.J., Takeoka, G.R. & Nam, Y. (1999). Volatile 

flavor components of rice cakes. Journal of Agricultural and Food 

Chemistry, 47, 4353–4356. 

Dimick, P.S. (1994). Peanut flavor-fade research report. Manufacturing 

Confectioner, January (1), 45–48. 

Krist, S., Unterweger, H., Bandion, F. & Buchbauer, G. (2004). 

Volatile compound analysis of SPME headspace and extract 

samples from roasted Italian chestnuts (Castanea sativa Mill.). 



McLafferty, F. (2000). Wiley Registry of Mass Spectral Data, 7th edn. 

New York: John Wiley & Sons. 

Meilgaard, M., Civille, G.V. & Carr, B.T. (1999). Sensory Evaluation 

Techniques, 3rd edn. Boca Raton, FL: CRC Press. 

Narbona, E., García-García, E., Vázquez-Arau´jo, L. & Carbonell- 

Barrachina, A.A. (2010). Volatile composition of functional ‘‘a la 

piedra’’ turroń with propolis. International Journal of Food Science 

and Technology, 45, 569–577. 

NIST (National Institute of Standards and Technology). (2009). 

http: ⁄⁄webbook.nist.gov ⁄ chemistry ⁄ name-ser.html (accessed on 

July 2009). 

Pino, J.A. & Mesa, J. (2006). Contribution of volatile compounds to 

mango (Mangifera indica L.) aroma. Flavour and Fragrance Journal, 

21, 207–213. 

Serra Bonvehi, J. (2005). Investigation of aromatic compounds in 

roasted cocoa powder. European Food Research and Technology, 

221, 19–29. 

Speziale, M. (2010). Definizione della Tipicita` Sensoriale e di Struttura 

di Produzioni Artigianali Dociarie Mediterranee. Ph.D. dissertation. 

Reggio Calabria (Italy): Università degli Studi Mediterranea di 

Reggio Calabria. 

Vázquez, L., Verdú, A., Miquel, A., Burló, F. & Carbonell-Barrachina, 

A.A. (2007). Changes in physico-chemical properties, hydroxymethylfurfural 

and volatile compounds during concentration of 

honey and sugars in Alicante and Jijona turroń. European Food 

Research and Technology, 225, 757–767. 

Vázquez-Arau´jo, L., Pe´rez-Castejoń, V., Verdu´, A. & Carbonell- 

Barrachina, A.A. (2005). Reclutamiento, selección, entrenamiento y 

validacioń de un panel de catadores especializado en turroń y sus 

materias primas. Alimentacioń Equipos Tecnologıá, 10, 92–98. 

Vázquez-Arau´jo, L., Verdu´, A., Murcia, R., Burló, F. & Carbonell- 

Barrachina, A.A. (2006). Instrumental texture of a typical Spanish 

confectionery product Xixona turroń as affected by commercial 

category and manufacturing company. Journal of Texture Studies, 

37, 63–79. 

Vázquez-Arau´jo, L., Enguix, L., Verdu´, A., García-García, E. & 

Carbonell-Barrachina, A.A. (2008a). Investigation of aromatic 

compounds in toasted almonds used for the manufacturing of 

turroń. European Food Research and Technology, 227, 243–254. 

Vázquez-Arau´jo, L., Verdú, A.E. & Carbonell-Barrachina, A.A. 

(2008b). Aroma volatiles of ‘‘a la piedra’’ turroń. Flavour and 

Fragrance Journal, 23, 84–92. 

Vázquez-Arau´jo, L., Verdu´, A., Enguix, L. & Carbonell-Barrachina, 

A.A. (2008c). Investigation of aromatic compounds in Alicante and 

Jijona turroń. European Food Research and Technology, 227, 1139– 

1147. 

Vázquez-Arau´jo, L., Verdu´, A., Navarro, P., Martínez-Sánchez, F. & 

Carbonell-Barrachina, A.A. (2009). Changes in volatile compounds 

and sensory quality during toasting of Spanish almonds. International 

Journal of Food Science & Technology, 44, 2225–2233. 


1594 


Fatty acid profile and proximate composition of Pacific mullet 

(Mugil so-iuy) caught in the Black Sea 

Sevim Köse, 1 * Serkan Koral, 2 Yes¸im Özog˘ul 3 & Bekir Tufan 1 

1 Faculty of Marine Sciences, Karadeniz Technical University, Çamburnu, Trabzon, Turkey 

2 Faculty of Fisheries, Rize University, Rize, Turkey 

3 Faculty of Fisheries, Çukurova University, Adana, Turkey 


Summary This study demonstrates proximate composition and fatty acid profile of Pacific mullet caught in Turkey. 

The highest moisture and protein contents were observed with muscle tissues as 83.74 and 10.52%, while the 

highest fat and ash contents were attributed to female gonads as 11.80 and 0.94%, respectively, with a 

significant variation amongst months (P < 0.05). Significant variation (P < 0.05) usually occurred amongst 

months within the same sex for total saturated fatty acids ( P SFA), monosaturated fatty acids ( P MUFA), 

polyunsaturated fatty acids ( P PUFA), eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) in 

muscle, although overall mean values between sexes were not found significant. Except for EPA, no variation 

observed for gonads while significant changes occurred (P < 0.05) with liver samples amongst months. 

Overall total values of P SFA, P MUFA, P PUFA, DHA and EPA in muscle samples were 29.59, 29.26, 

18.06 and 4.48%, respectively, while in gonads ranged as 30.26–33.23%, 35.17–37.47%, 11.87–14.88%, 4.38– 

5.34% and 3.02–5.02%. These FAs were 21.57–33.11%, 32.89–50.96%, 14.78–20.08%, 0.89–9.94% and 

5.85–9.54% for liver, respectively. The results of this study showed that muscle and gonads of Pacific mullet 

were rich in n-3 PUFA, especially, EPA, DHA, increasing the value of this species for human consumption. 

Keywords Docosahexaenoic acid, eicosapentaenoic acid, fatty acids, omega 3, Pacific mullet, polyunsaturated fatty acids, proximate 

composition. 

Introduction 

Pacific mullet (Mugil so-iuy, Basilewski, 1855) is a 

euryhaline fish originally found in Amu Darya River 

Basin, Far East Asia (Unsal, 1992; Kaya et al., 1998; 

Bas¸çınar & Okumus¸, 2005). It was first introduced to the 

Sea of Azov for fish farming, then migrated to the Black 

Sea, and from there to the Sea of Marmara and Aegean 

Sea (Kaya et al., 1998). In the Black Sea region, 10 kg of 

weighing members was reported, and its medium size 

was known as 20–30 cm (max. 61 cm) (Starushenko & 

Kazansky, 1996; Kaya et al., 1998). Although mullet 

species are widely caught and consumed around the 

world, the actual statistical information about Pacific 

mullet in Turkey or around the world is not clear yet. 

There have been few studies related to this species, most 

of which deals with aquaculture and revised by Bas¸çınar 

& Okumus¸ (2005). They also investigated some bioecological 

properties. Omelchenko et al. (2004) studied 

the genetic structure of far eastern mullet M. so-iuy and 

*Correspondent: Fax: +90 462 7522805; 

e-mail: kosesevim@gmail.com 


its change under acclimatisation. However, there is 

scarce information relating to its nutritional contents, 

especially fatty acids. 

There has been high evidence that fish lipids are 

beneficial for human health. Polyunsaturated fatty acids 

(PUFA), especially omega 3 fatty acids (n-3 FAs), have 

been recognised to be essential components of humans’ 

diet, prevent several diseases and have a vital role in 

health promotion. These acids, particularly eicosapentaenoic, 

20:5n3 (EPA) and docosahexaenoic, 22:6n3 

(DHA), have been reported to prevent and treat of 

cardiovascular, depressions and some other diseases like 

cancer, coronary heart diseases, inflammatory and 

autoimmune disorders, inflammation and arrhythmias 

rheumatoid arthritis (Kinsella, 1987; Lees & Karel, 

1990; Simopoulos et al., 1991; Schmidt et al., 2006). 

There is a nutritional deficit of n-3 PUFA in the human 

diet; therefore, a higher consumption of food containing 

these acids is recommended (Simopoulos, 1991; Mnari 

et al., 2007). However, lipid content and FA profile of 

fish are known to vary within species, and a number of 

factors, such as the temperature, the salinity, the season, 

the type and the availability of food, the habitat, the 

doi:10.1111/j.1365-2621.2010.02309.x 


stage of maturity and the individual variability, are 

believed to be important factors contributing to these 

variations (Tanakol et al., 1999; Halilog˘ lu, 2001; Luzia 

et al., 2003; Halilog˘ lu et al., 2004; Mnari et al., 2007; 

O¨ zogul & O¨ zogul, 2007). 

Compositions of lipids and FAs present in many fish 

species in the world have been extensively studied 

(Tanakol et al., 1999; Osman et al., 2001; Luzia et al., 

2003; Mnari et al., 2007; O¨ zog˘ ul et al., 2007; Çelik, 

2008; Yildiz et al., 2008; Karl et al., 2010). However, 

there is scarce information available on this aspect 

concerning P. mullet. Therefore, investigation of FA 

profile of such species is important. 

Pacific mullet, which has appeared in Black Sea from 

beginning of 1990, had adapted to Black Sea environment 

within 10 years of time and reached to a dynamic 

stock in this area. Okumus¸ & Bas¸çınar (1997) estimated 

its spawning season starting towards the end of May, 

followed by an egg release from middle of June until the 

middle of July. Such species is known to be settled in the 

North Black Sea area and has been caught by artisanal 

fishermen using trammel gill nets during its spawning 

period. This is the only species allowed for fishing during 

prohibition fishing period of other fish species (1st May– 

1st August) in this region (Go¨ zler et al., 2005). Therefore, 

the aim of this study is to investigate the changes of 

proximate composition and fatty acid profile in the 

edible flesh (muscle), gonads and liver of P. mullet 

during its fishing season. 


Pacific mullet (Mugil so-iuy, Basilewski, 1855) samples 

were purchased from local markets in Black Sea coast of 

Turkey just after freshly caught early in the morning. 

Sampling was carried out at three different months as 

May, June and July in 2006. They were brought to the 

laboratory immediately after buying in cold storage 

conditions. The samples were directly subjected to the 

experiments, if delayed, they were kept in the freezer at 

)20 °C. 

Biometric analyses 

All fish were individually weighed and measured. 

Weight was measured using an electronic digital balance 

(±0.1 g). Total length measurement was carried out 

using a measurement scale (±1 mm). For age determinations, 

10–15 scales of fish were taken for each 

measurement. The age was measured under microscope 

using the method described by Chugunova (1959). The 

sexes were determined after dissection by simple visual 

observation of the gonad’s physical appearance 

(Okumus¸ & Bas¸çınar, 1997). Weights of female gonads 

were determined (±0.01 g) after blotting blood and 

water according to Okumus¸ & Bas¸çınar (1997). 

Fatty acid and proximate composition of Pacific mullet S. Köse et al. 1595 

Proximate compositions analyses 

Moisture content was determined by oven drying of 5 g 

of fish muscle and gonad at 105 °C until a constant 

weight was obtained (AOAC, 1995, Method 985.14). 

Ash was determined by the AOAC (1980) Method 

7.009. Fat content was determined using a solvent 

extractor Velp SER 148 ⁄ 6 (Velp Scientifica, Milano, 

Italy) with petroleum ether (130 °C). Protein content 

was determined by AOAC (1980) Method 2.507 using 

the Kjeldahl nitrogen (6.25) in an automated distillation 

unit (Buchi 339, Flawil, Switzerland). All analyses were 

conducted in triplicate. 

Analysis of fatty acid profiles 

Fatty acid profiles for muscle, gonads and liver samples 

were analysed according to O¨ zog˘ ul et al. (2007). Analysis 

of liver samples was performed only for two 

different sampling periods (May and July) from the 

mixture obtained from all samples because the size of 

liver for individual fish was very small. 

FAME analyses 

Methyl esters were prepared by transmethylation using 

2 m potassium hydroxide (KOH) (Merck, Darmstadt, 

Germany) in methanol and n-hexane (Sigma-Aldrich, 

Steinhein, Germany) according to the method described 

by Ichihara et al. (1996) with minor modification; 10 mg 

of extracted oil was dissolved in 2 mL hexane, followed 

by 4 mL of 2 m methanolic KOH. The tube was then 

vortexed for 2 min at room temperature. After centrifugation 

at 1780 · g for 10 min, the hexane layer was 

taken for GC analyses. 

Gas chromatographic conditions 

The fatty acid composition was analysed by a GC 

Clarous 500 with auto sampler (Perkin–Elmer, Shelton, 

CT, USA) equipped with a flame ionisation detector and 

a fused silica capillary SGE column (30 m 0.32 mm ID 

0.25lm BP20 0.25UM, USA). The oven temperature was 

140 °C, held 5 min, rose to 200 °C at the rate 4 °C per 

minutes and held at 220 °C at1°C per minutes, while 

the injector and the detector temperatures were set at 

220 and 280 °C, respectively. The sample size was 1 lL, 

and the carrier gas was controlled at 16 ps. The split 

used was 1:100. Fatty acids were identified by comparing 

the retention times of FAME with the standard 37 

component FAME mixture (Supelco 37 Component 

FAME Mix. Pro Num. 47885-U). Two replicate GC 

analyses were performed, and the results were expressed 

in GC area % as mean values ± standard deviation 

(SD). Three samplings were carried out for FA profile of 

each individual fish for muscle as well as gonads. 


1596 

Fatty acid and proximate composition of Pacific mullet S. Köse et al. 


Analysis of variance (anova) was used to compare 

means of the proximate composition and fatty acid data, 

and when significant differences were found, comparisons 

among means were carried out by using Tukey test. 

A significance level of 95% (P < 0.05) was used 

throughout analysis. All statistical analyses were performed 

in jmp 5.0.1 package version (SAS Institute, Inc., 

Cary, NC, USA) (Sokal & Rohlf, 1987). The results 

were presented as means ± SD. 


Table 1 demonstrates proximate composition (% wet 

basis) values of the samples representing muscle tissue 

for both sex group and female gonads of Pacific mullet 

sampled at three different months (fishing season). 

Overall mean value for moisture of muscle samples 

increased from May to June significantly from 76.14% 

to 83.73%, while a significant decrease occurred from 

May to July in the values of protein, fat and ash 

contents, from 10.52% to 6.76%, 2.56% to 1.26% and 

0.81% to 0.58%, respectively. Although some significant 

variations (P < 0.05) were observed within the individual 

sex groups representing the same lot (Table S1), the 

variation was not found significant between sexes within 

the same months (Table 1). However, significant differences 

occurred (P < 0.05) mainly for moisture and 

protein contents amongst some months for both sexes. 

Little information exists regarding proximate composition 

of P. mullet in literature. Erdem et al. (2006) 

investigated different parts of P. mullet for proximate 

Analysis 

type 

Sampling 

time Mean females Mean males 

Moisture May 76.88 ± 1.68a A 

June 79.44 ± 1.44a B 

July 

C 

83.74 ± 1.20a Protein May 

A 

10.52 ± 0.68a June 8.37 ± 0.38a B 

July 7.03 ± 0.07a C 

Fat May 

A 

2.59 ± 0.35a June 2.73 ± 0.21a A 

July 1.34 ± 0.08a B 

Ash May 

A 

0.78 ± 0.05a June 

A 

0.70 ± 004a July 0.55 ± 0.03a B 

75.41 ± 2.58a A 

77.80 ± 1.84a A 

B 

83.72 ± 2.75a A 

10.52 ± 1.22a 8.96 ± 0.59a B 

6.49 ± 0.01a C 

A 

2.54 ± 0.22a 2.83 ± 0.17a A 

1.18 ± 0.22a B 

A 

0.83 ± 0.08a A 

0.76 ± 0.06a 0.62 ± 0.08a B 

Females & males 

total mean 

76.14 ± 2.21 A 

78.62 ± 1.79 B 

83.73 ± 2.24 C 

10.52 ± 0.88 A 

8.66 ± 0.55 B 

6.76 ± 0.26 C 

2.56 ± 0.28 A 

2.78 ± 0.19 A 

1.26 ± 0.17 B 

0.81 ± 0.07 A 

0.72 ± 0.05 A 

0.58 ± 0.08 B 

composition. They observed slightly higher fat and 

protein contents. Their ash and moisture contents 

supported our findings. The differences in protein and 

fat contents may be attributed to regional fish diet. Vlieg 

(1988) reported lower moisture content for grey mullet 

from New Zealand while protein, fat and ash contents 

were close or lower than our results with a variation 

amongst months. Muscle fat values of May and June 

samples were close to fat content observed for Mugil 

cephalus by O¨ zogul & O¨ zogul (2007). 

Moisture contents of gonad samples ranged from 

76.02% to 80.25%, with significant differences between 

July and other months (Table 1). Protein, fat and ash 

contents significantly dropped from May to July, 9.23% 

to 7.77%, 11.80% to 9.72% and 0.94% to 0.78%, 

respectively. Significant variation usually occurred 

(P < 0.05) within the individual gonads for protein, 

fat and ash contents (Table S2). The proximate composition 

relating to gonads or caviar of mullet is mainly 

carried out for grey mullet. Therefore, our results 

represent new data to the literature in this respect. 

Varying contents of moisture, protein, fat and ash were 

reported by different authors for gonads or caviar of 

grey mullet. Karakas¸ (2008) obtained moisture, protein, 

fat and ash contents as 61.8%, 23.9%, 9.9% and 5.1% 

for grey mullet, respectively. Duyar et al. (2008) 

observed higher moisture and fat contents, and lower 

protein and ash values for the same species. The values 

of S¸engo¨ r et al. (2002) for the caviar of M. cephalus were 

much lower than those obtained by Karakas¸ (2008) and 

Duyar et al. (2008) while higher protein content was 

found. Varying chemical composition was also reported 

at different types of fresh and processed fish caviar 

Female gonads 

total mean 

77.19 ± 0.55 A 

76.02 ± 1.33 A 

80.25 ± 0.30 B 

9.23 ± 0.36 A 

8.43 ± 0.43 A 

7.77 ± 0.25 B 

11.80 ± 0.61 A 

10.50 ± 0.50 B 

9.72 ± 0.09 B 

0.94 ± 0.06 A 

0.85 ± 0.02 B 

0.78 ± 0.03 C 

n: 3 (mean ± SD). 

Different small subscript letters (a, b) at the same row represent significant differences (P < 0.05) 

for mean values of males and females within the same lot (same month). 

Different superscript capital letters (A,B,C) at the same column represent significant differences 

(P < 0.05) amongst the months for the same type of chemical composition relating to mean values 

within the same sex groups or total mean values. 

Table 1 Mean proximate composition (% wet 

basis) values of the samples representing 

muscle tissue and female gonads of Pacific 

mullet sampled at three different months 


Table 2 Mean values of fatty acid profile for muscle tissues of Pacific mullet representing three different months for both male and female groups 

Mean Values for each sex group for each month Overall mean values 

May June July For sexes All Samples 

Fatty acid type Females Males Females Males Females Males Males Females F&M 

C14:0 3.08 ± 1.96 aA 

C15:0 0.92 ± 0.44 aA 

C16:0 13.38 ± 3.96 aA 

C17:0 0.74 ± 0.25 aA 

C18:0 4.27 ± 1.13 aA 

C20:0 0.43 ± 0.46 aA 

C23:0 0.15 ± 0.09 aA 

C24:0 3.18 ± 0.44 aA 

P 

SFA 

aA 

25.78 ± 4.50 

C14:1 0.06 ± 0.00 aA 

C15:1 0.15 ± 0.01 aA 

C16:1 14.03 ± 2.62 aA 

C17:1 0.82 ± 0.35 aA 

C18:1 7.83 ± 2.25 aA 

C20:1 2.38 ± 0.76 aA 

C22:1 n9 0.09 ± 0.00 a 

P 

MUFA 

aA 

24.37 ± 2.23 

C18:2cis 0.98 ± 0.78 aA 

C18:3 n3 0.36 ± 0.13 aA 

C18:3 n6 0.68 ± 0.35 aA 

C18:4 n3 0.22 ± 0.00 aA 

C20:2 cis 1.47 ± 0.50 aA 

C20:3 n6 0.94 ± 0.00 a 

C20:4 n6 4.08 ± 1.61 aA 

C20:5 n3 4.63 ± 0.20 aA 

C22:2 cis 0.29 ± 0.00 aA 

C22:6 n3 12.34 ± 1.11 aA 

P 

PUFA 

aA 

23.70 ± 2.71 

4.36 ± 0.43 a A 5.67 ± 1.96 aB 

4.44 ± 0.34 a A 3.99 ± 0.71 aA 

2.82 ± 1.76 a B 4.25 ± 1.31 a 

3.87 ± 0.91 a 

4.06 ± 0.26 

1.05 ± 0.36 a A 0.44 ± 0.14 aB 

2.45 ± 0.46 b B 0.96 ± 0.33 aA 

1.46 ± 0.99 a A 0.77 ± 0.29 a 

1.65 ± 0.72 b 

1.21 ± 0.62 

12.63 ± 0.85 a A 17.58 ± 3.11 aB 20.10 ± 0.90 a B 23.84 ± 1.66 aC 21.09 ± 0.36 a B 18.27 ± 5.26 a 17.94 ± 4.63 a 18.10 ± 0.23 

0.61 ± 0.29 a A 0.91 ± 0.60 aA 

1.19 ± 0.56 a B 0.45 ± 0.08 aB 

1.13 ± 0.97 a B 0.70 ± 0.23 a 

0.98 ± 0.32 a 

0.84 ± 0.20 

4.51 ± 0.28 a A 2.66 ± 0.34 aB 

1.90 ± 1.04 a B 1.60 ± 1.96 aC 

1.89 ± 2.60 a B 2.84 ± 1.34 a 

2.77 ± 1.51 a 

2.81 ± 0.05 

0.12 ± 0.07 a A ND ND 0.11 ± 0.03 aB 

1.95 ± 0.05 b B 0.27 ± 0.23 a 

1.04 ± 1.29 a 

0.65 ± 0.54 

0.54 ± 0.43 a A 0.22 ± 0.03 aA 

0.08 ± 0.00 b B ND 0.14 ± 0.01 a B 0.19 ± 0.05 a 

0.31 ± 0.33 a 

0.25 ± 0.09 

4.34 ± 1.40 a A 2.70 ± 1.19 aB 

1.44 ± 0.63 b B 1.43 ± 0.06 aC 

2.03 ± 0.16 b C 2.44 ± 0.90 a 

2.60 ± 1.53 a 

2.52 ± 0.12 

28.08 ± 2.66 a A 30.26 ± 4.34 aB 31.58 ± 0.67 a B 32.31 ± 1.63 aB 30.86 ± 1.46 a B 29.10 ± 4.61 a 30.09 ± 2.37 a 29.59 ± 0.70 

0.06 ± 0.00 a A 0.17 ± 0.06 aB 

0.10 ± 0.01 a B 0.06 ± 0.02 aA 

0.06 ± 0.02 a A 0.10 ± 0.06 a 

0.07 ± 0.02 a 

0.09 ± 0.02 

0.21 ± 0.14 a A 0.05 ± 0.01 aB 

0.31 ± 0.06 b B 0.06 ± 0.02 aB 

0.24 ± 0.06 b A 0.09 ± 0.06 a 

0.25 ± 0.05 b 

0.17 ± 0.12 

13.54 ± 1.43 a A 20.92 ± 1.22 aC 28.58 ± 4.03 b B 18.57 ± 3.71 aB 20.24 ± 3.30 a C 17.84 ± 3.50 a 20.79 ± 7.53 a 19.31 ± 2.08 

0.82 ± 0.52 a A 0.98 ± 0.73 aA 

2.09 ± 0.44 b B 0.09 ± 0.02 aB 

0.89 ± 0.81 a A 0.63 ± 0.47 a 

1.27 ± 0.71 b 

0.95 ± 0.45 

5.36 ± 0.54 b A 6.04 ± 1.94 aA 

7.31 ± 0.12 a B 10.29 ± 1.90 aB 

8.76 ± 2.38 a B 8.05 ± 2.13 a 

7.14 ± 1.71 a 

7.60 ± 0.64 

2.69 ± 0.54 a A 1.37 ± 0.86 aB 

0.24 ± 0.17 b B 0.44 ± 0.29 aC 

1.68 ± 1.90 a C 1.40 ± 0.97 a 

1.54 ± 1.23 a 

1.47 ± 0.10 

0.08 ± 0.01 a A ND 0.05 ± 0.01A ND 0.08 ± 0.00A 0.09 ± 0.00 a 

0.08 ± 0.01 a 

0.09 ± 0.01 

22.60 ± 1.62 a A 29.48 ± 1.44 aB 38.67 ± 3.87 b B 29.47 ± 1.36 aB 31.95 ± 0.23 b C 27.56 ± 3.04 a 30.97 ± 7.61 a 29.26 ± 1.41 

1.96 ± 1.18 a A 1.15 ± 0.11 aA 

1.10 ± 0.37 a B 2.77 ± 1.80 aB 

0.58 ± 0.42 b C 1.63 ± 0.99 a 

0.21 ± 0.02 a A 0.25 ± 0.12 aB 

0.48 ± 0.25 b B 0.12 ± 0.04 aC 

0.73 ± 0.54 b C 0.35 ± 0.29 a 

0.44 ± 0.11 a A 0.36 ± 0.04 aB 

0.64 ± 0.15 b B 0.54 ± 0.08 aC 

0.51 ± 0.16 a B 0.42 ± 0.10 a 

0.22 ± 0.03 a A 0.18 ± 0.09 aA 

0.11 ± 0.03 a B 0.07 ± 0.05 aB 

0.13 ± 0.04 a B 0.16 ± 0.08 a 

1.11 ± 0.57 a A 0.45 ± 0.34 aB 

0.13 ± 0.08 b B 0.28 ± 0.16 aC 

0.11 ± 0.01 b B 0.73 ± 0.64 a 

0.57 ± 0.16 b A ND ND ND 0.15 ± 0.00B 0.94 ± 0.00 a 

2.60 ± 0.45 b A 1.59 ± 0.03 aB 

1.88 ± 0.50 a B 1.87 ± 0.91 aB 

1.44 ± 0.04 a B 2.51 ± 1.36 a 

3.52 ± 0.19 b A 6.03 ± 3.16 aB 

7.11 ± 1.15 a B 1.58 ± 0.66 aC 

4.02 ± 0.47 b C 4.08 ± 2.28 a 

1.18 ± 0.15 a 

0.90 ± 0.02 B 

ND ND ND 0.29 ± 0.00 a 

9.42 ± 1.18 b A 9.20 ± 1.13 aB 

3.83 ± 4.00 b B 5.90 ± 0.91 aC 

5.37 ± 0.51 a C 9.15±3.22 a 

(S¸engo¨ r et al., 2002; Shirai et al., 2006; Mol & Turan, 

2008). Our moisture and ash contents supported the 

values obtained by Duyar et al. (2008), and protein 

content was similar to values of S¸engo¨ r et al. (2002) and 

Karakas¸ (2008). 


1.21 ± 0.70 a 

0.55 ± 0.16 a 

0.45 ± 0.22 a 

0.15 ± 0.06 a 

0.45 ± 0.57 a 

0.57 ± 0.16 b 

1.97 ± 0.59 a 

4.88 ± 1.94 a 

1.18 ± 0.15 b 

6.21 ± 2.89 a 

1.42 ± 0.30 

0.45 ± 0.14 

0.44 ± 0.02 

0.16 ± 0.00 

0.59 ± 0.20 

0.76 ± 0.26 

2.24 ± 0.38 

4.48 ± 0.57 

0.74 ± 0.63 

7.68 ± 2.08 

20.23±2.86 a A 19.75 ± 1.73 aB 15.30 ± 2.69 b B 13.08 ± 3.52 aC 12.95 ± 0.52 a C 19.57 ± 4.90 a 16.56 ± 3.84 a 18.06 ± 1.28 

PUFA\SFA 0.91 ± 0.04 aA 

0.72 ± 0.11 b A 0.66 ± 0.01 aB 

0.48 ± 0.01 b B 0.40 ± 0.02 aC 

0.42 ± 0.01 a B 0.67 ± 0.08 a 

0.55 ± 0.06 b 

P 

n6 

aA 

4.64 ± 1.67 3.14 ± 0.95 

0.61 ± 0.05 

a A 1.84 ± 0.05 aB 

2.37 ± 0.05 b B 1.89 ± 0.75 aB 

2.24 ± 0.54 a B 2.95 ± 1.73 a 

2.63 ± 0.74 a 

P 

n3 

aA 

17.27 ± 1.14 13.24 ± 1.15 

2.79 ± 1.22 

b A 15.71 ± 0.28 aA 11.71 ± 0.27 b B 8.04 ± 0.56 aB 10.02 ± 0.68 b B 14.38 ± 4.14 a 11.86 ± 2.11 b 13.12 ± 1.78 

n6\n3 0.27 ± 0.11b aA 

0.24 ± 0.08 b A 0.12 ± 0.00 aB 

0.21 ± 0.00 b A 0.24 ± 0.04 aA 

0.23 ± 0.00 a A 0.20 ± 0.02 a 

0.22 ± 0.04 a 

0.21 ± 0.04 

n3 ⁄ n6 4.23 ± 2.06 aA 

4.51 ± 1.32 a A 8.54 ± 0.09 aB 

4.91 ± 0.01 b A 4.36 ± 0.53 aA 

4.73 ± 0.02 a A 4.87 ± 0.03 a 

4.50 ± 0.05 a 

4.70 ± 0.07 

DHA\EPA 2.68 ± 0.24 aA 

2.68 ± 0.34 a A 1.89 ± 0.01 aB 

0.62 ± 0.01 b B 3.98 ± 0.51 aC 

1.34 ± 0.02 b C 2.24 ± 0.03 a 

1.27 ± 0.02 b 

1.71±0.01 

Unidentified 26.15 ± 1.45 29.09 ± 2.02 20.51 ± 2.21 14.45 ± 1.78 25.14 ± 1.72 24.24 ± 1.22 23.77 ± 1.36 22.38±1.45 23.12 ± 1.12 

n:3 (mean ± SD.) 

Superscript ‘a, b’ in the same line followed by different letter is significantly different (P < 0.05) between mean values between sexes within the same 

month. 

Superscript ‘A, B, C’ in the same line followed by the different letter represents significant differences (P < 0.05) amongst months for female samples. 

Subscript ‘A, B, C’ in the same line followed by the different letter represents significant differences (P < 0.05) amongst months for male samples. 

ND, Not detected; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; SFA, saturated fatty acids; MUFA, monosaturated fatty acids; PUFA, 

polyunsaturated fatty acids. 

Table 2 represents the fatty acid profile (% of total 

FAs) of P. mullet for muscle tissues representing three 

different months for both sex groups. The most abundant 

FA in muscle was palmitic acid (C16:0) with the 

mean values ranging from 13.38% to 23.84% for 


1598 


Table 3. Mean values representing fatty acid profile of female gonads and livers for different months 

Fatty Acid Type 

C14:0 1.05 ± 0.56 a 

C15:0 0.04 ± 0.01 a 

C16:0 3.53 ± 1.44 a 

C17:0 0.99±0.29 a 

C18:0 20.94 ± 0.27 a 

C20:0 3.24 ± 0.82 a 

C23:0 0.16 ± 0.07 a 

C24:0 1.74 ± 0.61 a 

P 

SFA 

a 

30.26 ± 1.80 

C14:1 0.04 ± 0.02 a 

C15:1 0.03 ± 0.01 a 

C16:1 20.41 ± 2.90 a 

C17:1 0.28 ± 0.26 a 

C18:1 10.20 ± 1.82 a 

C20:1 6.38 ± 1.45 a 

C22:1 n9 0.14 ± 0.04 a 

P 

MUFA 

a 

37.47 ± 3.49 

C18:2cis 1.85 ± 0.22 b 

C18:3 n3 0.54 ± 0.07 a 

C18:3 n6 1.00 ± 0.26 a 

C18:4 n3 0.18 ± 0.04 a 

C20:2 cis 0.07 ± 0.06 a 

C20:3 n6 0.23 ± 0.07 a 

C20:4 n6 0.76 ± 0.40 a 

C20:5 n3 5.02 ± 1.92 b 

C22:2 cis 0.09 ± 0.00 a 

C22:6 n3 5.34 ± 1.65 a 

P 

PUFA 

a 

14.88 ± 2.69 

PUFA\SFA 0.49 ± 0.08 a 

P a 

n6 1.83 ± 0.78 

P a 

n3 11.09 ± 1.84 

n6\n3 0.16 ± 0.07 a 

n3 ⁄ n6 6.06 ± 0.12 a 

DHA\EPA 1.06 ± 0.06 a 

Unidentified 15.74 ± 0.73 a 

Mean values of gonads Mean values of livers 

May June July Mixed May Mixed June 

0.86 ± 0.09 a 

0.14 ± 0.08 b 

2.84 ± 0.31 a 

2.18 ± 0.57 b 

23.26 ± 0.33 a 

3.12 ± 0.21 a 

0.15 ± 0.01 a 

1.52 ± 0.48 a 

33.23 ± 0.87 a 

0.06 ± 0.02 a 

0.17 ± 0.09 b 

18.27 ± 1.07 a 

0.32 ± 0.22 a 

10.30 ± 1.13 a 

5.86 ± 1.52 a 

0.18 ± 0.06 a 

35.17 ± 3.07 a 

1.97 ± 0.21 b 

0.57 ± 0.04 a 

1.17 ± 0.15 a 

0.17 ± 0.04 a 

0.05 ± 0.05 a 

0.23 ± 0.08 a 

0.91 ± 0.27 a 

3.02 ± 0.63 a 

0.05 ± 0.01 a 

4.74 ± 0.79 a 

12.86 ± 1.19 a 

0.38 ± 0.06 b 

2.32 ± 0.21 a 

8.48 ± 0.78 b 

0.27 ± 0.01 b 

3.65 ± 0.08 b 

1.56 ± 0.08 b 

17.85 ± 1.43 a 

females and 12.63% to 21.09% for males amongst 

months. The major FA in female gonads was stearic 

acid (C18:0) ranging between 20.72% and 23.26% 

(Table 3). For liver samples, only two different months 

were analysed. While 16:0 was found as the main FA in 

May as 22.21%, palmitoleic (16:1) was the most 

abundant FA in June as 33.23% (Table 3). C16:0 was 

also primary saturated fatty acids (SFA) for both muscle 

0.93 ± 0.06 a 

0.17 ± 0.00 b 

1.90±0.32 a 

2.05 ± 0.16 b 

20.72 ± 0.04 a 

3.45 ± 0.11 a 

0.11 ± 0.01 a 

1.42 ± 0.09 a 

30.56 ± 0.70 a 

0.04 ± 0.00 a 

0.18 ± 0.01 b 

18.54 ± 0.09 a 

0.38 ± 0.01 a 

11.50 ± 0.15 a 

5.68 ± 2.04 a 

0.40±0.16 b 

36.71 ± 3.02 a 

3.47 ± 0.07 a 

2.64 ± 0.01 b 

0.50 ± 0.03 a 

1.56 ± 0.01 b 

22.21 ± 0.43 a 

14.54 ± 0.04 b 

0.22 ± 0.01 a 

0.88 ± 0.01 b 

5.55 ± 0.07 a 

1.06 ± 0.04 b 

ND ND 

ND ND 

1.17 ± 0.06 a 

0.91 ± 0.01 a 

33.11 ± 0.36 a 

21.57 ± 0.01 b 

0.07 ± 0.01 a 

0.06 ± 0.01 a 

ND 0.25 ± 0.01 

14.80 ± 0.57 a 

33.23 ± 0.04 b 

0.31 ± 0.00 a 

1.11 ± 0.01 b 

16.71 ± 0.11 a 

16.20 ± 0.00 b 

0.21 ± 0.00 a 

0.11 ± 0.02 b 

ND ND 

32.89 ± 0.67 a 

50.96 ± 0.07 b 

0.56 ± 0.01 a 

0.88 ± 0.01 a 

1.41 ± 0.01 b 

0.66 ± 0.01 a 

0.56 ± 0.01 a 

0.86 ± 0.08 b 

1.64 ± 0.02 a 

0.23 ± 0.00 a 

0.36 ± 0.01 b 

0.13 ± 0.00 a 

ND 0.08 ± 0.00 

0.04 ± 0.01 a 

ND 0.22 ± 0.03 

0.15 ± 0.01 a 

ND ND 

1.29 ± 0.25 a 

2.63 ± 0.11 a 

1.44 ± 0.02 b 

3.35 ± 0.16 a 

5.85 ± 0.01 a 

9.54 ± 0.06 b 

ND ND ND 

4.38 ± 0.28 a 

9.94 ± 0.04 a 

0.89 ± 0.03 b 

11.87 ± 0.99 a 

0.38 ± 0.02 b 

2.76 ± 0.60 a 

8.52 ± 0.42 b 

0.32 ± 0.01 c 

3.08 ± 0.06 b 

1.30 ± 0.14 a 

21.36 ± 2.83 b 

20.08 ± 0.15 a 

0.625 ± 0.01 b 

2.86 ± 0.11 a 

16.35 ± 0.06 b 

0.17 ± 0.01 a 

5.71 ± 0.01 b 

0.59 ± 0.001 a 

14.72 ± 0.46 a 

14.78 ± 0.18 b 

0.685 ± 0.01 b 

1.70 ± 0.01 b 

11.37 ± 0.16 a 

0.15 ± 0.003 b 

6.68 ± 0.14 a 

0.09 ± 0.002 b 

12.70 ± 0.27 b 

The weight of liver samples for individuals in May: Females: 53, 20 and 55 g. Males: 27, 35 and 32 g, for 1st, 2nd and 3rd samples, respectively. 

In June for Females: 32, 41 and 34 g. Males: 30, 25 and 23 g, respectively. 

n:3 (mean ± SD); ND, Not Detected. The a,b,c letters in front of data mean that letters in the same line followed by different letter are significantly 

different (P < 0.05) amongst months. 

DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; SFA, saturated fatty acids; MUFA, monosaturated fatty acids; PUFA, polyunsaturated fatty 

acids. 

and liver samples, while C18:0 represented gonad 

samples. Various authors also reported that C16:0 was 

prevalent amongst the SFA in marine species (Luzia 

et al., 2003; S¸engo¨ r et al., 2003; O¨ zyurt et al., 2005; 

O¨ zogul & O¨ zogul, 2007; O¨ zogul et al., 2007). Other 

researchers demonstrated that 16:0 as the major FAs in 

liver samples (Jankowska et al., 2009). No data exists 

for FA profile relating to M. so-iuy in the literature, 


% of fatty acids 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

Variation in fatty acid groups amongst 3 months 

May 

June 

July 

∑ SFA ∑MUFA ∑PUFA 

∑ n6 ∑ n3 

Type of fatty acids group 

Figure 1 The changes of P SFA, P MUFA, P PUFA, P n6 and P n3 

amongst the months for the muscle samples of Pacific mullet. SFA, 

saturated fatty acids; MUFA, monosaturated fatty acids; PUFA, 

polyunsaturated fatty acids. 

therefore, our results contribute new data relating to this 

species on this respect. 

Statistical differences in FA values for muscle samples 

were usually observed between different individuals 

(P < 0.05) despite their sizes and weight (Tables S3– 

S5). Statistical differences of individual gonad FAs 

occurred (P < 0.05) mainly for SFA (Table S6). 

Table 2 and Fig. 1 also demonstrate the significance 

of variation for the major total FA groups as well as 

P n6 and P n3 of muscle samples amongst the months. 

Total mean value of P SFA for muscle samples was 

29.59%, ranged between 25.78% and 32.31% for 

females and between 28.08% and 31.58% for males 

within the months (Table 2). Although no significant 

changes occurred between sexes within the months, the 

variation was found significant (P < 0.05) between 

May and other two months within the same sex of both 

males and females (Table 2, Fig. 1). Overall, mean 

values of different sex groups also confirmed our results 

with no significant variation between males and females 

in terms of P SFA (Table 2). Palmitoleic acid was 

observed as the main monosaturated fatty acids 

(MUFA) in muscle samples. Mean values of P MUFA 

significantly ranged (P < 0.05) from 24.37% to 29.48% 

for females and from 22.60% to 38.67% for males. 

Despite the significant changes (P < 0.05) also occurred 

between the sex groups within the same month, the 

changes were accounted as insignificant in the overall 

total mean values (Table 2). The mean values of 

P PUFA for each sex group significantly (P < 0.05) 

dropped from 23.70% to 13.08% for females and from 

20.23% to 12.95% for males from May to July. 

Significant variation between sex groups only occurred 

in June samples (P < 0.05). Moreover, overall mean 

values of both female and male samples showed no 

significant differences. Luzia et al. (2003) observed no 

seasonal changes for total saturated and unsaturated 


acid contents of fish samples such as sardine, croaker and 

tilapia despite the changes in EPA and DHA levels. 

Yanes-Roca et al. (2009) reported a seasonal variation in 

total SFA, PUFA and DHA levels and indicated that 

PUFA are functionally essential for normal growth, 

development and reproduction in fish. O¨ zyurt et al. 

(2005) observed a decrease in the level of P SFA in winter 

of gilthead sea bream, which was thought to be because of 

the catabolisation of saturated fatty acids to compensate 

for the extra metabolic energy required in that period. Our 

results showed that P SFA (males + females) increased 

from May (26.93%) to July (31. 59%) which might 

either caused because of significant variation within the 

individuals for specific FAs (Tables S3–S5) or other 

characteristics that might be use of SFA in such fish 

species. 

P MUFA has also increased from May to June. On 

the other hand, P PUFA was decreased sharply from 

May (21.97%) to July (13.02%) that represent the 

beginning and the end of reproduction period, respectively. 

Therefore, PUFA may be used more on reproduction 

purposes for this species that supports the 

reports of Yanes-Roca et al. (2009) in terms of PUFA 

functionality in the fish. 

Table 3 represents the FA profile of both gonad and 

liver samples. The primary SFA for gonads was found 

as 18:0 in this species while Halilog˘ lu et al. (2004) 

observed as 16:0 in trout gonads aquacultured at 

seawater, and the values were similar to their feed used 

for aquaculturing. Mean values of P SFA, P MUFA 

and P PUFA ranged from 30.26% to 33.23%, 35.17% 

to 37.47% and 11.87% to 14.88%, respectively. Despite 

significant differences (P < 0.05) on certain FAs for 

individual gonads (Table S6), the variation was not 

found significant for any total mean values of those 

types of major FA groups. Variation in such parameters 

was reported by different researchers for either processed 

(as caviars) or fresh gonads (S¸engo¨ r et al., 2003; 

Shirai et al., 2006; Mol & Turan, 2008; Yanes-Roca 

et al., 2009). S¸engo¨ r et al. (2003) observed 11.9% 

P SFA, 42.9% P MUFA and 39.3% P PUFA for grey 

mullet caviar. Our results were higher for SFA and 

lower for MUFA and PUFA than those obtained for 

grey mullet. For liver samples, high variation occurred 

(P < 0.05) between May and June samples for all main 

FA groups. The mean values for P SFA and P PUFA 

have significantly decreased from May to July, from 

33.11% to 21.57% and from 20.08% to 14.78%, 

respectively, whereas the values of P MUFA increased 

from 32.89% to 50.96%. Various values were observed 

by other researchers for such parameters (Mnari et al., 

2007; Jankowska et al., 2009). 

The n3 ⁄ n6 ratio has been suggested to be a useful 

indicator for comparing relative nutritional values of 

fish oils and a ratio of 1 ⁄ 1–1:5 recommended (Osman 

et al., 2001). The n3 ⁄ n6 ratio of for all types of samples 


1600 


was higher than the recommended ratio although female 

fish samples showed slightly higher value than males. 

The n3 ⁄ n6 ratios for female gonads were 3.08–6.06, the 

highest occurred with matured gonads (in May). These 

levels are found to be close to the levels obtained with 

fish flesh, suggesting that gonads of this species are also 

as valuable and it is advisable to be processed for human 

consumption. DHA and EPA had been shown to have 

preventive effects on human coronary artery disease and 

been suggested as a key component for a healthy diet in 

humans (Kinsella, 1987; Lees & Karel, 1990; Simopoulos 

et al., 1991). In our study, DHA was the basic 

PUFA followed by EPA for muscle of both sexes. The 

levels for gonad samples averaged from 4.38% to 5.34% 

for DHA and 3.02% to 5.02% for EPA. In May, EPA 

levels were significantly higher (P < 0.05) compare to 

other months. For liver samples, although samples in 

May showing high DHA and low EPA values, accounting 

for 9.94% and 5.85%, opposite situation was 

observed for June as 0.89% for DHA and 9.54% for 

EPA. Numerous researchers have reported that DHA 

constitutes a majority of the PUFA in most marine fish 

(Tanakol et al., 1999; O¨ zyurt et al., 2005). Although the 

overall mean values showed no significant differences 

between sexes for both DHA and EPA in muscle tissues, 

significant differences occurred (P < 0.05) in some 

months either between sexes or within the same month. 

For gonads, no differences were found except for EPA 

between May and other months. O¨ zogul & O¨ zogul 

(2007) found higher EPA levels for M. cephalus muscle 

as 10.5%. Their DHA levels supported our overall mean 

values. However, S¸engo¨ r et al. (2003) found lower levels 

of both DHA and EPA for M. cephalus for muscle and 

gonad samples. 

The addition of n3 PUFA to diet could improve the 

nutritional value and protect against diseases. A minimum 

value of PUFA ⁄ SFA ratio recommended is 0.45 

(HMSO, 1994). The PUFA ⁄ SFA levels in muscle were 

higher for females as 0.67 than males as 0.55. Significant 

variations (P < 0.05) were observed amongst individuals, 

months and sexes (Tables 2 and S3–S5). O¨ zogul & 

O¨ zogul (2007) found 0.75 for grey mullet, while 1.17 

obtained by S¸engo¨ r et al. (2003) (obtained from Aegean 

sea) indicating the regional effect on the same species. 

Our study also indicates that such levels can change 

depending on the time of the sampling, particularly 

reproduction period. The relating ratio for liver samples 

was found as 0.63 in May and 0.69 in June. S¸engo¨ r et al. 

(2003) observed higher level for the caviar of grey 

mullet. The UK Department of Health recommends an 

ideal ratio of n6 ⁄ n3 of 4.0 at maximum (HMSO, 1994), 

and higher values are suggested to be harmful to health 

and may promote cardiovascular diseases (O¨ zogul & 

O¨ zogul, 2007). We observed 0.21 and 0.22 for females 

and males, respectively, for muscle samples, 0.15–0.17 

for liver and 0.16–0.32 for gonads. 

Osman et al. (2001) reported that arachidonic acid 

(AA, C20:4n6) is a precursor for prostaglandin and 

thromboxan which will influence the blood clot and its 

attachment to the endothelial tissue during wound 

healing. Apart from that, the acid also plays a role in 

growth. They also indicated that the contents of AA in 

marine fishes were lower than freshwater fishes. This 

study represents higher AA levels than those obtained 

for marine fish by other researches (Kiessling et al., 

2001; Osman et al., 2001; O¨ zogul & O¨ zogul, 2007), 

indicating the genetic effect of this species as originated 

from freshwater environment. The values were higher 

for edible flesh of females as 4.08% than males 

(2.60%) in May samples compared to other samples 

obtained in June and July. It was suggested that such 

fatty acid is more abundant in fish species at the 

gonadial maturated period, especially in female samples. 

The overall mean value of AA was found as 

2.24% which was higher than the values obtained by 

O¨ zogul & O¨ zogul (2007) as between 0.08% and 0.61% 

for several fish species including M. cephalus. The 

values of AA in liver samples were similar to those 

obtained from muscle of male samples. However, 

lower AA observed for gonad samples averaged from 

0.76% to 1.29%. 

Processing and marketing of various types of caviars 

from different fish species including waxed grey mullet 

caviar have been reported (S¸engo¨ r et al., 2002, 2003; 

Shirai et al., 2006; Duyar et al., 2008; Karakas¸, 2008; 

Mol & Turan, 2008); however, the processing of caviar 

of P. mullet is rarely known and may exist only amongst 

the locals. 

Conclusion 

This study demonstrates significant variation in both 

proximate composition and FA of flesh, gonad and liver 

of P. mullet amongst the months, especially at the 

beginning and end of reproduction period. Although 

grey mullet is extensively studied for its nutritional 

value, there has been little information on P. mullet in 

regarding this aspect. As this species has been reported 

to have a rapid increase and expanding its population in 

different regions, biochemical composition, and FA 

profile will be valuable information for seafood industry 

and human consumption. Our study showed that both 

edible muscle and gonads of this species are useful 

source of fatty acid profile in terms of health diet for 

human consumption despite the low protein and lipid 

content in muscle tissue. Although P. mullet is commercially 

marketed for human consumption, caviar is 

not known commercially processed. This study indicates 

their commercial importance. The results may also 

contribute useful information on the euryhaline fish 

species in terms of proximate composition and FA 

profile. 



This study was supported by Karadeniz Technical 

University under the project number 2006.117.01.4. 

References 

AOAC. (1980). Official Methods of Analysis, 13th edn. Washington, 

DC: Association of Official Analytical Chemists, Official methods.7.009 

and 2.507 

AOAC. (1995). Official Methods of Analysis. Gaithersburg, MD: 

Association of Official Analytical Chemists, Official methods.985.14. 

Bas¸çınar, N. & Okumus¸, _I. (2005). I-VI yas¸larındaki Pasifik kefali 

ag˘ ırlık-karkas ve ag˘ ırlık – bas¸ oranı ilis¸kileri. Tu¨rk Sucul Yas¸am 

Dergisi. Ulusal Su Gu¨nleri Sempozyumu. Yıl. 3, 4, 539–544. In Turkish. 

Çelik, M. (2008). Seasonal changes in the proximate chemical 

compositions and fatty acids of chub mackerel (Scomber japonicus) 

and horse mackerel (Trachurus trachurus) from the north eastern 

Mediterranean Sea. International Journal of Food Science and 


Chugunova, N.I. (1959). Age and Growth Studies in Fishes. Pp. 67–110. 

Translated by Yasski, D. (1963). Jarusalem: Israel Program for 

Scientific Translations Ltd. 

Duyar, H.A., Ög˘ retmen, Y.Ö. & Ekici, K. (2008). Composition of 

Waxed caviar and the determination of its shelf-life. Journal of 

Animal and Veterinary Advances, 7, 1029–1033. 

Erdem, M.E., Kalaycı, F. & Samsun, N. (2006). Sinop Kıyılarında 

Avlanan Pasifik Kefali (Mugil so-iuy, Basilevsky, 1855) Filetolarında 

Besin _Içeriklerinin Dag˘ ılımı. EU Journal of Fisheries and Aquatic 

Sciences, 23(Suppl. 1 ⁄ 3), 421–424. In Turkish with English Abstract. 

Go¨ zler, A.M., Engin, S., Koral, S., S¸ahin, C. & Ag˘ ırbas¸, E. (2005). 

Dog˘ u Karadenizde 2004 Yılı Avlanma Sezonunda Avlanan Pasifik 

Kefali (Mugil so-iuy, Basilewski, 1885)’nin Bazı Populasyon Parametreleri, 

In Proceedings of ‘XIII. Ulusal Su U¨ru¨nleri Sempozyumu’, 

Çanakkale, Turkey. In Turkish. 

Halilog˘ lu, H.I. (2001). A research on fatty acid composition of some 

tissues of Rainbow trout’s (Onchorhynchus mykiss) raised in 

different fish farms in Ezurum. Ph.D. Thesis. Turkey: Atatu¨ rk 

Uni. Fisheries Dept., Erzurum. In Turkish with English Abstract. 

Halilog˘ lu, H._I., Bayır, A., Sirkecig˘ lu, A.N., Aras, N.M. & Atamanalp, 

M. (2004). Comparison of fatty acid composition in some tissues of 

rainbow trout (Oncorhynchus mykiss) living in seawater and 

freshwater. Food Chemistry, 86, 55–59. 

HMSO, (1994). Nutritional Aspects of Cardiovascular Disease. Report 

on health and social subjects N. 46. Pp. 37–46. London: HMSO. 

Department of Health: UK. 

Ichihara, K., Shibahara, A., Yamamoto, K. & Nakayama, T. (1996). 

An improved method for rapid analysis of the fatty acids of 

glycerolipids. Lipids, 31, 535–539. 

Jankowska, B., Zakes´, Z., _Zmijewski, T. & Szczepkowski, M. 

(2010). Fatty acid profile of muscles, liver and mesenteric fat in 

wild and reared perch (Perca fluviatilis L.). Food Chemistry, 118, 

764–768. 

Karakas¸, Y. (2008). Kefal balıg˘ ında havyar u¨ retimi ve kalitesinin 

belirlenmesi. Master Thesis. Isparta, Turkey: Süleyman Demirel 

U¨ niversitesity. In Turkish with English Abstract. 

Karl, H., Lehmann, I., Rehbein, H. & Schubring, R. (2010). 

Composition and quality attributes of conventionally and organically 

farmed Pangasius fillets (Pangasius hypophthalmus) on the 

German market. International Journal of Food Science and Technology, 

45, 56–66. 

Kaya, M., Mater, S. & Korkut, A.Y. (1998). A New Grey Mullet 

Species ‘‘Mugil so–iuy Basilewsky’’ (Teleostei: Mugilidae) From the 

Aegean Coast of Turkey. Turkish Journal of Zoology, 22, 303–306. 

Kiessling, A., Pickova, J., Johansson, L., A˚ sga˚ rd, T., Storebakken, T. 

& Kiessling, K-H. (2001). Changes in fatty acid composition in 


muscle and adipose tissue of farmed rainbow trout (Oncorhynchus 

mykiss) in relation to ration and age. Food Chemistry, 73, 271–284. 

Kinsella, J.E. (1987). Seafoods and Fish Oils in Human Health and 

Disease. Pp. 1–23. New York: Marcel Dekker. 

Lees, R.S. & Karel, M. (1990). Omega-3 Fatty Acids in Health and 

Disease. Pp. 39–69. New York: Marcel Dekker. 

Luzia, L.A., Sampaio, G.R., Castellucci, C.M.N. & Torres, E.A.F.S. 

(2003). The influence of season on the lipid profiles of five commercially 

important species of Brazilian fish. Food Chemistry, 83, 93–97. 

Mnari, A., Bouhlel, I., Chraief, I. et al. (2007). Fatty acids in muscles 

and liver of Tunisian wild and farmed gilthead sea bream, (Sparus 

aurata). Food Chemistry, 100, 1393–1397. 

Mol, S. & Turan, S. (2008). Comparison of proximate, fatty acid and 

amino acid compositions of various types of fish roes. International 

Journal of Food Properties, 11, 669–677. 

Okumus¸, _I. & Bas¸çınar, N. (1997). Population structure, growth and 

reproduction of introduced Pacific mullet, (Mugil so-iuy) in the 

Black Sea. Fisheries Research, 33, 131–137. 

Omelchenko, V.T., Salmenkova, E.A., Makhotkin, M.A. et al. (2004). 

Far Eastern Mullet Mugil soiuy Basilewsky (Mugilidae, Mugiliformes): 

the Genetic Structure of Populations and Its Change under 

Acclimatization. Russian Journal of Genetics, 40, 910–918. 

Osman, H., Suriah, A.R. & Law, E.C. (2001). Fatty acid composition 

and cholesterol content of selected marine fish in Malaysian waters. 


Özogul, Y. & Özogul, F. (2007). Fatty acid profiles of commercially 

important fish species from the Mediterranean, Aegean and Black 

Seas. Food Chemistry, 100, 1634–1638. 

Özog˘ ul, Y., Özog˘ ul, F. & Alago¨ z, S. (2007). Fatty acid profiles and fat 

contents of commercially important marine water and freshwater 

fish species of Turkey: a comparative study. Food Chemistry, 103, 

217–223. 

Özyurt, G., Polat, A. & Özku¨ tu¨ k, S. (2005). Seasonal changes in the 

fatty acids of gilthead sea bream (Sparus aurata) and white sea 

bream (Diplodus sargus) captured in Iskenderun Bay, eastern 

Mediterranean coast of Turkey. European Food Research and 

Technology, 220, 120–124. 

Schmidt, E.B., Rasmussena, L.H., Rasmussena, J.G., Joensena, A.M., 

Madsena, M.B. & Christensen, J.H. (2006). Fish, marine n-3 

polyunsaturated fatty acids and coronary heart disease: a minireview 

with focus on clinical trial data. Prostaglandins, Leukotrienes and 

Essential Fatty Acids, 75, 191–195. 

S¸engör, G.F., Cihaner, A., Erkan, N., Özden, O¨ . & Varlık, C. (2002). 

Topan kefali (Mugil cephalus, L. 1758) yumurtasından havyar eldesi, 

randımanı ve kimyasal kompozisyonunun belirlenmesi. Turkish 

Journal of Veterinary and Animal Sciences, 26, 183–187. In Turkish 

with English Abstract. 

S¸engör, G.F., Özden, Ö., Erkan, N., Tu¨ ter, M. & Aksoy, H.A. (2003). 

Fatty acid compositions of flathead grey mullet (Mugil cephalus L., 

1758) fillet, raw and beeswaxed caviar oils. Turkish Journal of 

Fisheries and Aquatic Sciences, 3, 93–96. 

Shirai, N., Higuchi, T. & Suzuki, H. (2006). Analysis of lipid classes 

and the fatty acid composition of the salted fish roe food 

products, Ikura, Tarako, Tobiko and Kazunoko. Food Chemistry, 

94, 61–67. 

Simopoulos, A.P. (1991). Omega-3 fatty acids in health and disease 

and in growth and development. American Journal of Clinical 

Nutrition, 54, 438–463. 

Simopoulos, A.P., Kifer, R.R., Martin, R.E. & Barlow, S.M. (1991). 

Health Effects of Omega 3 Polyunsaturated Fatty Acids in Seafoods, 

Vol. 66. Basel, Switzerland: S. Karger Publishing; World Review of 

Nutrition and Dietetics. 

Sokal, R.R. & Rohlf, F.J. (1987). Introduction to Biostatistics, 2nd edn. 

New York: W.H. Freeman and Company. 

Starushenko, L.I. & Kazansky, A.B. (1996). Introduction of Mullet 

Haader (Mugil soiuy Basilevsky) into the Black Sea and the Sea of 

Azov. GFCM Studies and Reviews No. 67. Rome: FAO. 


1602 


Tanakol, R., Yazıcı, Z., Sener, E. & Sencer, E. (1999). Fatty acid 

composition of 19 species of fish from the Black Sea and the 

Marmara Sea. Lipids, 34, 291–297. 

Unsal, S. (1992). A new mullet species for Turkish seas: Mugil soiuy, 

Basilevsky. Dog˘a, Turkish Journal of Veterinary and Animal 

Sciences, 16, 427–432. 

Vlieg, P. (1988). Proximate composition of New Zealand Marine 

Finfish and Shellfish, FAO Corporate Document Repository, Home 

page. http://www.fao.org/docrep/007/ae581e/ae581e00.htm#Contents, 

Accessed 28th July 2009. 

Yanes-Roca, C., Rhody, N., Nystroma, M. & Main, K.L. (2009). 

Effects of fatty acid composition and spawning season patterns on 

egg quality and larval survival in common snook (Centropomus 

undecimalis). Aquaculture, 287, 335–340. 

Yildiz, M., S¸ener, E. & Timur, M. (2008). Effects of differences in diet 

and seasonal changes on the fatty acid composition in fillets from 

farmed and wild sea bream (Sparus aurata L.) and sea bass 

(Dicentrarchus labrax L.). International Journal of Food Science and 




online version of this article 

Table S1. Proximate composition (% wet basis) values 

for muscle tissues of individual Pacific mullet samples 

obtained from three different months. 

Table S2. Proximate composition of gonads of individual 

Pacific mullet samples representing three different 

months. 

Table S3. Fatty acid profile for muscle tissues of 

individual Pacific mullet samples obtained in May. 


individual Pacific mullet samples obtained in June. 


individual Pacific mullet samples obtained in July. 

Table S6. Fatty acid profile of individual female 

gonads obtained in May and July. 









Volatile composition and nutritional quality of the edible 

mushroom Polyporus tenuiculus grown on different agro-industrial 

waste 

Alejandra Omarini, 1 Cynthia Henning, 2 Jorge Ringuelet, 2 Julio A. Zygadlo 3 & Edgardo Albertó 1 * 

1 Laboratorio de Micología y Cultivo de Hongos Comestibles, IIB-INTECH (UNSAM-CONICET) Camino Circunv. Laguna km 6, C.C. 164; 

C.P. B7130IWA, Chascomús, Argentina 

2 Laboratorio de Fitoquímica, Facultad de Ciencias Agrarias y Forestales, Universidad Nacional de La Plata, Calle 60 Nº 119, La Plata (1900), 

Argentina 

3 Instituto de Ciencia y Tecnología de Alimentos (ICTA), IMBIV-CONICET, FCEFyN-UNC, Av. Velez Sarsfield 1611, (5016), Co´ rdoba, 

Argentina 

(Received 23 December 2009; Accepted in revised form 4 May 2010) 

Summary The volatile composition and the nutritional value of Polyporus tenuiculus grown on supplemented and 

nonsupplemented wheat straw and willow sawdust were determined. Thirty-nine volatile compounds were 

detected, including acids, esters, alcohols, hydrocarbons, aldehydes and ketones. The main volatile 

compound in all samples was 1-octen-3-ol, with increasing levels in mushrooms cropped on supplemented 

substrate. In addition, several precursors of this alcohol were identified in lower percentages. Mushrooms 

grown on supplemented substrates showed lower fat (5.2–5.7%) and carbohydrate contents (48.2%) and 

higher protein content (22–22.5%). Fibre and ash contents showed some variations between types of 

substrates. Compared to other edible fungi, P. tenuiculus high fibre and protein contents point to this species 

as a healthy nutritional alternative of interest for the food industry. Moreover, the wide spectrum of volatile 

compounds of P. tenuiculus reveals great potential for biotechnological applications such as the production 

of ‘‘non artificial’’ mushroom flavour. 

Keywords Edible mushroom, flavour, nutrients, Polyporus tenuiculus, volatile compounds. 

Introduction 

Mushrooms, the fruiting bodies of fungi, are appreciated 

not only for their texture and flavour but also for their 

chemical and nutritional characteristics (Manzi et al., 

1999). Nutritionally, mushrooms provide key nutrients 

and bioactive components such as high-quality proteins, 

some vitamins (including riboflavin, niacin, thiamine, 

folic acid and ascorbic acid), minerals (potassium, phosphorus, 

magnesium, zinc, copper, and selenium), unsaturated 

fatty acids and fibre (Buswell & Chang, 1993). 

Research on the quality and diversity of the cultivated 

fungal species is as important as the conducted research 

on agronomic techniques to obtain mushrooms with 

higher nutritional and medicinal value as well as suitable 

biological material for biotechnological applications for 

industrial purposes (Liu et al., 2004; Ç ag˘ larırmak, 2007). 

Relevant targets are the pharmaceutical, diet supplements 

and food industries, which can attain products of 

*Correspondent: E-mail: ealberto@intech.gov.ar 

doi:10.1111/j.1365-2621.2010.02306.x 


higher added value by enriching their products with 

edible fungi (GRAS organisms) and therefore obtain 

new aromas, flavours and nutritional values. 

Fungal aroma is a typical and special trait of each 

edible mushroom species. The potential of basidiomycetes 

to produce natural flavours de novo or by 

biotransformation is of considerable interest for biotechnological 

purposes because, at present, flavour 

supply is quite restricted to the natural biosynthetic 

capability of plants (Abraham & Berger, 1994). As 

fungal metabolites represent a wide diversity of chemical 

species, research on fungal secondary metabolism is of 

great interest. Furthermore, along with certain Ascomycota, 

several species of the Polyporaceae family have 

been studied in recent years as a good source of 

biologically active compounds (Wu et al., 2005). Thus, 

fungal biotransformation of low-cost substrates into 

high-value flavour and aroma compounds reveals to be 

of great commercial potential. For this reason, it is of 

great importance to learn about the volatile composition 

of each fungal species under natural conditions.

1604 

Volatile composition and nutritional quality of Polyporus tenuiculus A. Omarini et al. 

Therefore, a detailed research of this kind is a basic 

starting point for revealing the underlying mechanisms 

in flavour and aroma formation in fungal fruiting bodies 

(Wu et al., 2005). 

Both volatiles and nutrients can be affected by 

substrate or growth media, which is of great importance 

given that one of the most important factor for 

consumer acceptance of food is flavour (Shashirekha 

et al., 2005; Wu et al., 2005; Valencia del Toro et al., 

2006; Ç ag˘ larırmak, 2007; Cho et al., 2007, 2008; Silva 

et al., 2007; ). Substrates used for mushroom cultivation, 

mixtures of different agro-industrial wastes (wheat, 

rice, paddy straw, sawdust, corn cobs) and nitrogen 

supplements (wheat bran, soybean hull, millet) influence 

chemical composition and, consequently, the nutritional 

value of the cultivated mushrooms (Bonatti et al., 2004; 

Shashirekha et al., 2005). 

Many studies focus on the search of new wild species 

of edible mushrooms to enrich the number of available 

species for human consumption. Mushrooms of the 

genus Polyporus are known to have a great capacity for 

producing and transforming high-value compounds 

(Wu et al., 2005), which makes them an interesting 

subject for further study. We have focused on Polyporus 

tenuiculus, a species widely distributed in America which 

is harvested by different ethnic groups for human 

consumption and to be sold in the markets in Me´ xico 

(Omarini et al., 2009). Recently, Omarini et al. (2009) 

managed to obtain fruiting bodies of P. tenuiculus on 

agriculture waste. They also studied the sensory attributes 

of dried specimens and proved that they can vary 

with the substrate where it is cultivated (Omarini et al., 

2010). This species, being massively consumed in different 

American countries, indicates a clear preference over 

other native edible mushrooms but its flavour profile 

and the nutritional traits are still unknown. Therefore, 

we examined the volatile flavour compounds and 

nutritional value of P. tenuiculus and their variations 

according to growth substrate. 


Strain 

Cultures used in this study are conserved in the IIB- 

INTECH Collection of Fungal Cultures (ICFC) reference 

in the WDCM data base (WDCM 826). P. tenuiculus 

ICFC 383 ⁄ 00, Brazil, Rı´ o Grande do Soul, Porto Alegre, 

Parque Faroupilla (fallen trunks of Musa sp. trees). 

Culture condition for mushroom cultivation 

Culture media and spawn preparation 

Potato dextrose agar (39 g L )1 PDA; Britania, Buenos 

Aires, Argentina) culture medium was used for routine 

culture and storage purposes. Spawn was prepared from 

Table 1 Substrate formulation (%) 

Substrates Main component (%) 

Supplements (%) 

Wheat 

brand 

Soybean 

flour CaCO 3 

NS-Wt Wheat straw (98) 0 0 2 

S-Wt Wheat straw (78) 15 5 2 

NS-Sd Willow sawdust (98) 0 0 2 

S-Sd Willow sawdust (78) 15 5 2 

boiled wheat seeds (Triticum sp.) supplemented with 2% 

w ⁄ w calcium carbonate (CaCO 3), placed in polypropylene 

bags and then sterilised at 121 °C for 2 h (Stamets, 

1993). As the bags were cooled, they were inoculated 

with mycelia grown on PDA and incubated in darkness 

at 25 °C until mycelia had completely covered the wheat 

grains. 

Substrate formulation 

Dried willow tree (Salix sp.) sawdust and wheat 

(Triticum sp.) straw were chopped at 1.5–3 mm and of 

30–50 mm long, respectively. These substrates were used 

with and without supplements (Table 1). Distilled water 

was added to all formulas and left overnight to 

moisturise completely to obtain 70% w ⁄ w of water 

content. Polypropylene bags (25 · 45 cm, 30 lm thick) 

were filled with 1 kg of wet substrate and stoppered with 

cotton plugs held by polyvinyl chloride cylinders, before 

they were sterilised twice at 121 °C for 2.5 h. After 

cooling, the bags were inoculated with 5% w ⁄ w spawn 

and incubated in the dark at 25 °C for 60 days until the 

mycelium completely colonised the substrate. 

Culture condition 

Cropping conditions to induce fruiting bodies formation 

for all substrates were 22 ± 2 °C, 9-h light ⁄ 15-h dark 

photoperiod (20 W fluorescent light), 75–85% relative 

humidity levels and watering by spray (fog type) and 

automatically watering with a fog system for 5 min 

every 8 h. Fruiting bodies were harvested when mature 

(see Omarini et al., 2009) and they were frozen at 

)20 °C for assays. 

Fruiting body analysis 

Analysis of volatile compounds 

Samples of fresh fruiting bodies (50 g) harvested from 

wheat straw (NS-Wt and S-Wt) and sawdust (NS-Sd 

and S-Sd) were placed in a glass recipient seal. Extraction 

of headspace volatile compounds was carried out 

using a solid-phase microextraction (SPME) device 

(Supelco, Bellefonte, PA, USA) with a 100-lm polydimethylsiloxane 

fibre. Capped vials were placed in a 

40 °C water bath, and volatiles were sampled for 20 min 


(Ouzouni et al., 2009). After extraction, volatiles from 

the SPME fibre were desorbed at the injection port of a 

gas chromatograph coupled to a mass spectrometer (HP 

5890II-HP 5971MSD, Hewlett-Packard), with a potential 

of 70 eV for ionisation by electron impact. The 

injector and detector temperatures were 200 and 270 °C, 

respectively. The column was temperature programmed 

as follows: 50 °C (2 min) to 200 °C (3 °C ⁄ min). The 

carrier gas was helium with a constant flow rate set close 

to 0.9 mL per minutes. 

Compounds were resolved on a DB-5 column (25 m 

length · 0.25 mm inside diameter · 0.25 lm film thickness) 

and identified on the basis of comparison of their 

mass spectra with authentic standards when this was 

possible; or with commercial mass spectral databases of 

the Wiley library (Wiley 275; J. Wiley & Sons Ltd, Wet 

Sussex, UK) and NIST 98 MS Library (Rev. D. 01.00; 

HP, Ringoes, NJ, USA)). Volatile analyses of the 

samples were done in triplicate. 

Proximate analysis 

Samples of fresh fruiting bodies of P. tenuiculus (250 g 

of a combination of young and old sporophores) were 

collected by culture bags of supplemented and nonsupplemented 

wheat straw (S-Wt and NS-Wt) and 

sawdust (S-Sd and NS-Sd) and were dried until constant 

weight at 50 °C. 

The proximate composition of mushroom samples 

(based on the official AOAC method, 1990) was carried 

out by triplicate to obtain the following fractions 

(expressed in percentage): (i) moisture content (drying 

in oven at 100 °C), (ii) ash content (by incineration at 

550 °C), (iii) crude fat content (determined by Soxhlet 

using hexane as solvent), (iv) crude fibre content (by acid 

and alkali digestion), and (v) crude protein content (by 

Kjeldahl method (the total nitrogen content was 

determined by the Kjeldahl (AOAC, 1990) method). 

The total protein was determined from the total 

nitrogen content, using the correction factor 4.38 

(Breene, 1990). Non-nitrogen extractive fraction (% 

ENN), which includes soluble carbohydrates, was calculated 

as dry matter not accounted in the sum of 

humidity, ash, crude fibre, crude fat and crude protein. 


Statistical analyses were performed using SigmaStat Ò 

program for windows, version 3.1 (Systat software Inc., 

Point Richmond, CA, USA). Volatile compounds analysis 

of mushrooms samples was calculated as the 

average of two replicates performed in each substrate 

sample whereas proximate analysis of mushrooms 

composition samples was calculated as the average of 

three replicates. Means and standard deviations were 

calculated, and significant differences were estimated 

using Tukey multiple-range test at P < 0.05. 

Volatile composition and nutritional quality of Polyporus tenuiculus A. Omarini et al. 1605 


Volatile compounds of fruiting bodies 

Volatile compounds identified in P. tenuiculus fruiting 

bodies grown on lignocellulose substrates (Wt and Sd) 

consisted of acids, esters, alcohols, carbohydrates, 

aldehydes and ketones. Table 2 includes the 39 identified 

volatile compounds, their retention indexes (IRs) and 

relative percentages based on GC ⁄ MS peak area. 

Independently of the culture substrate used in this 

study, the major volatile compounds identified in all 

samples was 1-octen-3-ol, widely referred in the literature 

for its characteristic ‘‘mushroom’’ aroma (Chiron & 

Michelot, 2005; Combet et al., 2006) contributing 

significantly to aroma intensity. It is frequently found 

in many mushroom species such as Agaricus bisporus, 

Fomitopsis pinicola, Piptoporus betulinus, Cantharellus 

cibarius, Polyporus sulfureus, Pleurotus ostreatus, P. florida, 

Hericium erinaceum, Boletus edulis, Tricholoma 

matsutake (Ro¨ secke et al., 2000; Liu et al., 2004; Chiron 

& Michelot, 2005; Wu et al., 2005; Combet et al., 2006; 

Guedes de Pinho et al., 2008). Also, several compounds 

with intermediate to low percentages were identified, 

such as (Z,Z)-9,12-octadecadienoic acid methyl ester, 

(Z)-9-octadecenoic acid methyl ester, (Z,Z)-9,12-octadecadienoic 

acid methyl ester, and hexadecanoic acid, and 

all of them are precursors in the synthesis of 1-octen- 

3-ol. As informed for A. bisporus, this alcohol is the 

product of a series of enzymatic reactions starting with 

the oxidation of (Z,Z)-9,12-octadecadienoic acid (linoleic 

acid), by a lipoxygenase, and the further cleavage of 

the intermediate fatty acid hydroperoxide by a hydroperoxide 

lyase (Chiron & Michelot, 2005; Wu et al., 

2005; Combet et al., 2006). The fruiting bodies of 

P. tenuiculus harvested on supplemented substrates 

(S-Wt and S-Sd) showed a small increase in 1-octen-3-ol 

percentages. As it was reported by Wu et al. (2005) and 

Ç ag˘ larırmak (2007), the flavours of mushrooms were 

affected by the compositions of growth medium, growth 

conditions, age and genetics variation. Therefore, our 

results showed that either the addition of supplements in 

the substrates of P. tenuiculus, such as wheat bran and 

soybean flour, could provide precursors for C8 acids 

synthesis leading to increasing relative percentages or 

the increase in the nutrient availability for mushrooms 

nutrition could allow the increase in certain metabolites 

level (Shashirekha et al.,2005). 

The fungal aroma of P. tenuiculus is mainly associated 

with the occurrence of C8 aliphatic chains despite the 

contribution of other kind of compounds to it. Linoleic 

acid is the precursor to eight-carbon volatiles formation 

in fungi (1-octen-3-ol, 1-octen-3-one and 3-octanol), 

acting as substrate to a fatty acid oxygenase (lipoxygenase), 

and then that of a hydroperoxide lyase (Combet 

et al., 2006). This process occurs generally in the late 


1606 


Table 2 Volatiles composition of fresh Polyporus tenuiculus fruiting bodies harvested from the different lignocellulose wastes 

Compound 

IR %* 

phase of cultivation, indicating an increasing contact 

of enzyme and substrate as a result of increased cell 

lyses (Abraham & Berger, 1994). It has also been 

demonstrated for A. bisporus that 1-octen-3-ol can be 

produced by enzymatic reduction or self-oxidation of 

1-octen-3-one (Chen & Wu, 1984). 

Among the volatiles identified in P. tenuiculus fruiting 

bodies were two benzoid compounds of pleasant weak 

floral fragrance, benzaldehyde and phenylacetic acid, 

both in a low percentage in all samples analysed but 

specially in those grown on sawdust. A reason for the 

DB-5 NS-Sd S-Sd NS-Wt WS-t 

2-methylbutanoic acid methyl ester 770 Tr Tr Tr 0.15 ± 0.02 

Hexanal 785 2.75 ± 0.35 b Tr Tr 0.35 ± 0.07 a 

2-methyl propanoic acid 793 Tr 0.25 ± 0.07 a 0.30 ± 0 a 0.4 ± 0.14 a 

4-hidroxy-4-methyl-2-pentanone 809 Tr Tr Tr Tr 

4-hexen-3-one 811 Tr 0.45 ± 0.07 a 0.55 ± 0.07 a Tr 

3-methylbutanoic acid 850 0.90 ± 0.14 b 0.05 ± 0.07 a 0.05 ± 0.07 a Tr 

5-(1-methylethylidene)-1,3-cyclopentadiene 858 Tr Tr Tr Tr 

c-butyrolactone 870 0.30 ± 0.14 c 0.15 ± 0.07 b 0.05 ± 0.07 Tr 

Hexanoic acid methyl ester 911 Tr Tr Tr Tr 

c-valerolactone 918 0.15± 0.07 ab 0.10 ± 0 a 0.25 ± 0.07 b Tr 

2-heptenal 930 Tr Tr Tr Tr 

Benzaldehyde 935 0.20 ± 0.2 b 0.05 ± 0.07 a 0.35 ± 0.07 c Tr 

2-methylhexanoic acid 941 Tr 0.51 ± 0.07 a 0.51 ± 0.07 a Tr 

1-octen-3-one 952 0.35 ± 0.07 a 1.20 ± 0 b 0.45 ± 0.07 a Tr 

3-octanone 960 12.55 ± 0.78 b 8.25 ± 0.35 a 11.55 ± 1.20 b 10.95 ± 0.21b 

1-octen-3-ol 970 38.45 ± 1.06 a 40.80 ± 0.42b 37.85 ± 0.21 a 44.05 ± 0.21c 

Octanal 980 Tr Tr Tr 0.75 ± 0.07 

3-octanol 984 5.90 ± 0.14 a 8.90 ± 0.14 c 7.90 ± 0.28 b 11.05 ± 0.21d 

Hexanoic acid methyl ester 990 Tr 0.20 ± 0 a 0.20 ± 0 a Tr 

1-octanol 1063 0.90 ± 0.14 a 2.45 ± 0.49 b 3.80 ± 0.28 c 2.50 ± 0.42 b 

Nonanal 1085 0.80 ± 0.14 a 0.25 ± 0.21 a 0.15 ± 0.07 a 2.50 ± 0.42 b 

2-phenylethanol 1088 0.70 ± 0.28 b 0.25 ± 0.07ab 0.50 ± 0.14 b Tr 

2-octenol phenylacetic acid 1248 4.95 ± 0.07 b 2.20 ± 0.28 a 2.00 ± 0.28 a Tr 

Nonanoic acid 1273 0.70 ± 0.28 a 0.70 ± 0.14 a 0.50 ± 0.14 a 1.85 ± 0.21 b 

Hexadecane 1602 Tr Tr Tr Tr 

Dodecanoic acid 1-methylethyl ester 1618 Tr Tr Tr Tr 

Heptadecane 1702 Tr Tr Tr Tr 

Octadecane 1802 Tr Tr Tr Tr 

Pentadecanoic acid methyl ester 1814 Tr 0.05 ± 0.07 a 0.15 ± 0.21 a 3.05 ± 0.21 b 

Pentadecanoic acid 1857 Tr Tr Tr Tr 

(Z)-9-hexadecenoic acid methyl ester 1890 Tr Tr 0.20 ± 0.14 Tr 

Hexadecanoic acid methyl ester 1916 Tr 0.95 ± 0.07 c 0.45 ± 0.07b 0.05 ± 0.07 a 

Hexadecanoic acid 1958 7.60 ± 0.57 c 4.90 ± 0.14 a 6.20 ± 0.28 b Tr 

Eicosane 2002 Tr Tr Tr Tr 

(Z,Z)-9,12-octadecadienoic acid methyl ester 2079 2.65 ± 0.92 a 8.60 ± 0.71bc 7.85 ± 0.21 b 9.85 ± 0.21 c 

(Z)-9-octadecenoic acid methyl ester 2087 3.15 ± 0.21 b 0.65 ± 0.35 a 0.85 ± 0.21 a 0.90 ± 0.14 a 

(Z,Z)-9,12-octadecadienoic acid 2126 10.50 ± 0.70 a 10.60 ± 0.70 10.60 ± 1.13 a 9.50 ± 0.70 a 

Octadecanoic acid 2157 1.25 ± 0.35 b Tr Tr 0.20 ± 0.14 a 

*Relative percentages of the volatiles are based on the peak areas obtained, without MS detector response factor correction. Values are average of two 

replicates. In each file, different letters indicate significant differences (P < 0.05). Substrates: NS-Wt: wheat straw; S-Wt: wheat straw supplemented with 

soybean flour (5%) and wheat brand (15%); NS-Sd: willow sawdust; S-Sd: willow sawdust supplemented with soybean flour (5%) and wheat brand 

(15%); Tr, trace. 

small number and amounts of aldehydes could be the 

rapid oxidation by different oxidising fungal enzyme 

systems, as they have particularly been reported for 

white rot fungi (Abraham & Berger, 1994). 

An odd-long-chain fatty acid, identified as the pentadecanoic 

acid, was detected in trace amounts as well as 

its methyl ester. Both were previously reported for 

P. sulfurous by Wu et al. (2005) and the relevance of this 

finding resides in the rarity of their occurrence in nature. 

The role of fatty acids in the aroma of microorganisms, 

plants and animal products is well known. Not only do 


they contribute to the odour by themselves but also act 

as flavour precursors under enzymatic oxidation, decarboxylation 

or esterification reactions. Therefore, fatty 

acid profiles and their variations in fungi are important 

factors in the synthesis and transformation of volatile 

compounds (Wu et al., 2005). Several authors have 

stated that the profile of compounds compromising 

fungal aroma varies according to species, varieties, 

grade and mushroom tissues (pileus ⁄ stipe) (Wu et al., 

2005; Cho et al., 2007, 2008). The profile of volatile 

compounds can also be influenced by substrate composition. 

Shashirekha et al. (2005) reported a 35% increase 

in total lipids of samples grown on supplemented 

substrates, with the linoleic acid (C8 volatile precursor) 

as the predominant unsaturated fatty acid. 

Data on volatile composition represents an important 

contribution to the chemosystematic approach. Callac & 

Guinberteau (2005) measured the relative amounts of 

volatile benzoic and C8 volatile compounds of several 

species of Agaricus and were able to identify a relationship 

between ratio and phylogenetic closeness, indicating 

a potential importance of volatile chemicals as 

taxonomic marker. Also, a profile of the major volatile 

enables the development of biotechnological strategies 

to obtain compounds of interest at industrial scale by 

adding precursors to the culture media in optimised 

conditions. Such an application was carried out by Liu 

et al. (2004) who incremented the content of 1-octen- 

3-ol in fresh grounded A. bisporus samples cultured on a 

hydrolysed soybean media and, after several extractionconcentration 

steps, microencapsulated the achieved 

aroma by means of a aspersion drying technique, 

obtaining a suitable product for the food industry. 

Our results reveal a wide volatile compound spectrum 

for P. tenuiculus, which could be of technological 

interest for ‘‘natural’’ fungal aroma production. Among 

these, the most abundant compounds were 1-octen-3-ol, 

3-octanol, 1-octanol, and 3-octanone, allowing us to 

conclude that their contribution to the aroma is the 

most important. 

Nutritional components of Polyporus tenuiculus 

Table 3 shows moisture content, protein, fibre, ash, 

crude fat and soluble carbohydrates (compounds without 

nitrogen) contents of P. tenuiculus fruiting bodies 

harvested from the different lignocellulose substrates: 

nonsupplemented and supplemented wheat straw 

(NS-Wt and S-Wt) and sawdust culture media (NS-Sd 

and S-Sd). 

Moisture content was not significantly different 

among fruiting bodies harvested from all substrates and 

their average, around 90%, was consistent with the typical 

high water content of these products (Breene, 1990). 

Similar moisture content (85.2–93.5%) was reported for 

Pleurotus ostreatus, P. eryngii, P. pulmonarius, P. florida, 


Table 3 Proximal composition of the fruiting bodies of Polyporus 

tenuiculus harvested from the different lignocellulose substrate (dry 

weight basis) 

Components (%) 

Substrates* 

Ash 5.7 ± 0.2 a 

Crude fat 8.6 ± 0.1 c 

Crude Protein 15.5 ± 0.3 a 

Carbohydrates 47.2 ± 0.9 a 

Crude fibre 12.1 ± 0.2 c 

Moisture 89.1 ± 0.7 a 

NS-Wt S-Wt NS-Sd S-Sd 

6.6 ± 1.1 b 

5.2 ± 0.1 a 

22.5 ± 0.6 b 

48.2 ± 1.0 a 

7.5 ± 0.4 a 

90.0 ± 0.2 a 

6.5 ± 0.2 b 

9.2 ± 0.2 d 

15.1 ± 0.2 a 

51.6 ± 0.7 c 

7.6 ± 0.2 a 

89.9 ± 0.3 a 

6.6 ± 0.1 b 

5.7 ± 0.4 b 

22.0 ± 0.3 b 

48.2 ± 0.2 a 

8.3 ± 0.4 b 

90.8 ± 0.1 a 

*Values (average of three replicates) express as percentage. In each file 

different letters indicate significant differences (P < 0.05). Protein: 

N · 4.38. NS-Wt, wheat straw; S-Wt, wheat straw supplemented with 

soybean flour (5%) and wheat brand (15%); NS-Sd, willow sawdust; S- 

Sd, willow sawdust supplemented with soybean flour (5%) and wheat 

brand (15%). 

P. sajor-caju and Lentinula edodes, none of them having 

found differences among different growth substrates 

(Manzi et al., 1999; Bonatti et al., 2004; Ponmurugan 

et al., 2007). This variable is exclusively dependent on the 

fungal species and on the temperature and relative 

humidity conditions during growth (Bano & Rajarathnam, 

1988). 

Differences in ash content among samples from 

different substrates were found lower levels of whose 

harvested from nonsupplemented wheat straw media 

(NS-Wt) (5.7%; P < 0.05). Similar results were 

reported by Bonatti et al. (2004) for P. ostreatus and 

P. sajor-caju´. It is important to point out that N2-rich 

substrates (S-Wt and S-Sd) positively condition ash 

content, inducing its increase and thus enhancing 

mineral bioavailability in human diet associated with 

edible fungi consumption (Silva et al., 2007). Still, 

higher values (by a factor of 0.3) have been reported 

in literature for P. ostreatus grown on wheat straw 

supplemented with beetroot (Manzi et al., 1999). 

Protein content was estimated as total nitrogen corrected 

by a conversion factor of 4.38% (Breene, 1990). 

Results varied within a range of 15.1–22.5% (Table 3), 

with higher percentages on fruiting bodies harvested from 

supplemented substrates, S-Wt and S-Sd (P < 0.05). 

These values are within the ranges informed for P. ostreatus, 

P. sajor caju´ and L. edodes, although a great 

variability of protein content is observed among species 

and strains (Manzi et al., 1999; Bonatti et al., 2004). 

Higher protein content was obtained for fruiting bodies 

grown on N 2-rich substrates (S-Wt and S-Sd), indicating 

that the nutritional value of mushrooms is compromised 

by substrate composition. Probably the fungus used the 

N 2 available in the substrate to produce larger amounts 

of protein. The same effect was observed in species of 

the genus Pleurotus showing varied nutritional values 


1608 


according to strain, growth substrate and fruiting body 

stage of development (Shashirekha et al., 2005; Valencia 

del Toro et al., 2006; Silva et al., 2007;). 

Crude fat content increased when nonsupplemented 

substrates were used, the highest values were obtained for 

mushrooms harvested from sawdust (P < 0.05). Apparently, 

the supplements provided nutrients that stimulated 

the fungus to produce less crude compounds but more 

protein content. In the case of supplemented substrates, 

sawdust also afforded fat-richer mushrooms than those 

grown on wheat straw (P < 0.05). The levels of P. tenuiculus 

fat components were higher than those obtained 

with Agrocybe cylindracea (Uhart et al., 2008) but similar 

to those reported for P. ostreatus and P. sajor caju, 

which have also shown substrate-dependent fat content 

variation (Ortega et al.,1993; Bonatti et al., 2004). 

The highest and lowest values of soluble carbohydrates 

were obtained for mushrooms grown on nonsupplemented 

sawdust (NS-Sd) and wheat straw (NS-Ws) 

substrates, respectively (P < 0.05). Similar data have 

been reported by other authors for several species of the 

genus Pleurotus grown on different agro-industrial 

wastes, in the range of 42.8–47.7% (Patrabansh & 

Madan, 1997; Ragunathan & Swaminathan, 2003). 

As to fibre content, the highest values were obtained 

for mushrooms harvested from nonsupplemented wheat 

straw substrate (P < 0.05). Addition of supplements to 

substrate led to opposite effects on fibre content 

depending on the substrate used, resulting in higher 

values for sawdust and lower values in the case of wheat 

straw (P < 0.05).This behaviour can be ascribed to 

differences of wheat and sawdust composition. Fibre 

percentages of P. tenuiculus were lower to those 

described for P. ostreatus grown on wheat straw, which 

is in the range of 20.0–34.8% (Justo et al., 1999). On the 

other hand, similar fibre percentages have been reported 

for other species of the genus Pleurotus grown on 

different agro-industrial residues (7.6–18.0%) (Patrabansh 

& Madan, 1997; Bonatti et al., 2004; Valencia del 

Toro et al., 2006). 

When compared to other highly consumed species like 

A. bisporus, L. edodes and P. ostreatus, the nutritional 

feature of P. tenuiculus which outstands is fibre content, 

by a factor of 2.5, approximately. Protein content, 

although lower than that of A. bisporus, has shown to be 

similar to that of L. edodes and P. ostreatus (Buswell & 

Chang, 1993), pointing to P. tenuiculus as a valuable 

specie for human consumption. 

Conclusions 

Our results reveal a wide volatile compound spectrum 

for P. tenuiculus where thirty-nine volatile compounds 

were detected. Among these, the most abundant 

compounds were 1-octen-3-ol, 3-octanol, 1-octanol, 

and 3-octanone, allowing us to conclude that their 

contribution to the aroma is the most important. The 

main volatile compound in all samples was 1-octen-3-ol, 

with increasing levels in mushrooms cropped on supplemented 

substrate. In addition, several precursors of 

this alcohol were identified in lower percentages. 

An overall view shows that, in the case of P. tenuiculus, 

supplemented substrates (S-Wt and S-Sd) promote 

protein-rich, fat-poor and carbohydrate-poor 

fruiting bodies, compared to those grown on nonsupplemented 

substrates. Also, a comparative examination 

of fat and protein levels obtained for supplemented 

substrates shows that a decrease in the former is 

associated with an increase in the latter. This correlation 

had been previously reported by Silva et al. (2007) for 

P. sajor caju´ grown on different N 2-rich substrates. On 

the other hand, fibre and ash content varied according 

to substrate type rather than according to supplements 

(Wt and Sd). According to Sturion & Oetterer (1995), 

fungi nutritional value can be greatly affected by its 

growth substrate, as supported by our results. 

Compared to other edible fungi, P. tenuiculus high 

fibre and protein contents point to this species as a 

healthy nutritional alternative of interest for the food 

industry or for biotechnological applications such as the 

production of ‘‘non artificial’’ mushroom flavour. 


This work was partially supported by the research 

project PIP 5516 from National Research Council 

(CONICET, Argentina). 

References 

Abraham, B.G. & Berger, R. (1994). Higher fungi for generating 

aroma components through novel biotechnologies. Journal of 


AOAC (1990). Official methods of the Association of Official Analytical 

Chemists, 15th edn. Arlington, VA, USA: Association of Official 

Analytical Chemists. 

Bano, Z. & Rajarathnam, S. (1988). Pleurotus mushrooms. Part ii, 

nutritional value, post-harvest physiology, preservation and role as 

human food. CRC Critical Reviews in Food Science and Nutrition, 

27, 87–158. 

Bonatti, M., Karnopp, P., Soares, H.M. & Furlan, S.A. (2004). 

Evaluation of Pleurotus ostreatus and Pleurotus sajor-caju nutritional 

characteristics when cultivated in different lignocellulosic 

wastes. Food Chemistry, 88, 425–428. 

Breene, W.M. (1990). Nutritional and medicinal value of specialty 

mushrooms. Journal of Food Protection, 53, 883–894. 

Buswell, J.A. & Chang, S.T. (1993). Edible Mushrooms: attributes and 

applications. In: Genetics and Breeding of Edible Mushroom (edited 

by S.T. Chang, J.A. Buswell & P.G. Miles). Pp. 297–306. Amsterdam: 

Gordon & Breach Science Publishers. 

Ç ag˘ larırmak, N. (2007). The nutrients of exotic mushrooms (Lentinula 

edodes and Pleurotus ostreatus species) and an estimated approach 

to the volatile compounds. Food Chemistry, 105, 1188–1194. 

Callac, P. & Guinberteau, J. (2005). Morphological and molecular 

characterization of two novel species of Agaricus section xanthodermatei. 

Mycologia, 97, 416–424. 


Chen, C.C. & Wu, C.M. (1984). Studies on the Enzymic Reduction of 

1-Octen-3-one in Mushroom (Agaricus bisporus). Journal of Agricultural 

and Food Chemistry, 32, 1342–1344. 

Chiron, N. & Michelot, D. (2005). Odeurs des champignons: chimic et 

roles dans les interactions biotiques- une revue. Cryptogamie 

Mycologie, 26, 299–364. 

Cho, I.N., Lee, S.M., Kim, S.Y. & Chol, H.K. (2007). Differentiation 

of aroma characteristics of pine-mushrooms (Tricholoma matsutake 

Sing.) of different grade using gas chromatography-olfactometry 

and sensory analysis. Journal of Agricultural and Food Chemistry, 55, 

2323–2328. 

Cho, I.H., Namgung, H.J., Choi, H.K. & Kim, Y.S. (2008). Volatiles 

and key odorants in the pileus and stipe of pine-mushroom 

(Tricholoma matsutake Sing.). Food Chemistry, 106, 71–76. 

Combet, E., Henderson, J., Eastwood, D.C. & Burton, K. (2006). 

Review: eight-carbon volatiles in mushrooms and fungi: properties, 

analysis and biosynthesis. Mycoscience, 47, 317–326. 

Guedes de Pinho, P., Ribeiro, B., Gonçalves, R.F. et al. (2008). 

Correlation between the pattern volatiles and the overall aroma of 

wild edible mushrooms. Journal of Agricultural and Food Chemistry, 

56, 1704–1712. 

Justo, M.B., Guzmán, G.A., Mejía, E.G., Díaz, C.L.G., Martinez, G. 

& Corona, E.B. (1999). Calidad proteínica de tres cepas mexicanas 

de setas (Pleurotus ostreatus). Archivos Latinoamericanos de Nutricioń, 

49, 81–85. 

Liu, Z., Zhou, J.H., Zeng, Y.L. & Ouyang, X.L. (2004). The 

enhancement and encapsulation of Agaricus bisporus flavour. 


Manzi, P., Gambelli, L., Marconi, S., Vivanti, V. & Pizzoferrato, 

L. (1999). Nutrients in edible mushrooms an inter-species comparative 

study. Food Chemistry, 65, 477–482. 

Omarini, A., Lechner, B.E. & Alberto´ , E. (2009). Polyporus tenuiculus: 

a new naturally occurring mushroom that can be industrially 

cultivated on agricultural waste. Journal of Industrial Microbiology 

& Biotechnology, 36, 635–642. 

Omarini, A., Nepote, V., Grosso, N., Zygadlo, J. & Albertó, E. (2010). 

Analysis and fruiting bodies characterization of the edible mushrooms 

Pleurotus ostreatus and Polyporus tenuiculus obtained on leaf 

waste from the essential oil production industry. International 

Journal of Food Science and Technology, 45, 466–474. 

Ortega, G.M., Martinez, E.O., Betancourt, D., Gonzalez, A.E. & 

Otero, M.A. (1993). Enzyme activities and substrate degradation 


during white rot fungi growth on sugar-cane straw in a solid state 

fermentation. World Journal of Microbiology and Biotechnology, 9, 

210–212. 

Ouzouni, P.K., Koller, W.-D., Badeka, A.V. & Riganakos, K.A. 

(2009). Volatile compounds from the fruiting bodies of three 

Hygrophorus mushroom species from Northern Greece. International 


Patrabansh, S. & Madan, M. (1997). Studies on cultivation, biological 

efficiency and chemical analysis of Pleurotus sajor-caju (FR.) Singer 

on different bio-wastes. Acta Biotechnology, 17, 107–122. 

Ponmurugan, P., SeKhar, Y.N. & Sreesakthi, T.R. (2007). Effect of 

various substrates on the growth and quality of mushrooms. 

Pakistan Journal of Biological Science, 10, 171–173. 

Ragunathan, R. & Swaminathan, K. (2003). Nutritional status of 

Pleurotus spp. grown on various agro-wastes. Food Chemistry, 80, 

371–375. 

Rösecke, J., Pietsch, M. & Ko¨ nig, W.A. (2000). Volatile constituents of 

wood-rotting basidiomycetes. Phytochemistry, 54, 747–750. 

Shashirekha, M.N, Rajarathnam, S. & Bano, Z. (2005). Effects of 

supplementing rice straw growth substrate with cotton seeds on the 

analytical characteristics of the mushroom, Pleurotus florida (Block 

& Tsao). Food Chemistry, 92, 255–259. 

Silva, G.E., Souza Díaz, E., Siqueira, F.G. & Schwan, R.F. (2007). 

Chemical analysis of fructification bodies of Pleurotus sajor-caju 

cultivated in several nitrogen concentrations. Cieˆncia e Tecnolgia de 

Alimentos, 27, 72–75. 

Stamets, P.S. (1993). Growing Gourmet and Medicinal Mushrooms. Pp. 

554. Berkeley, USA: Ten Speed Press. 

Sturion, G.L. & Oetterer, M. (1995). Composição química de 

cogumelos comestíveis (Pleurotus spp.) originados de cultivos em 

diferentes substratos. Cieˆncia e Tecnolgia de Alimentos, 15, 189–193. 

Uhart, M., Piscera, J.M. & Alberto´ , E. (2008). Utilization of new 

naturally occurring strains and supplementation to improve the 

biological efficiency of the edible mushroom Agrocybe cylindracea. 

Journal of Industrial Microbiology and Biotechnology, 35, 595–602. 

Valencia del Toro, G., Castelán Vega, R., Garín-Aguilar, M.E. & Leal 

Lara, H. (2006). Biological quality of proteins from three strains of 

Pleurotus spp. Food Chemistry, 94, 494–497. 

Wu, S., Zorn, H., Krings, U. & Berger, R. (2005). Characteristic 

volatile from young and aged fruiting bodies of wild Polyporus 

sulfureus (Bull.: FR.) Fr. Journal of Agricultural and Food Chemistry, 

53, 4524–4528. 


1610 


Use of HR-NMR to classify propolis obtained using different 

harvesting methods 

Giulia Papotti, 1 Davide Bertelli, 1 * Maria Plessi 1 & Maria Cecilia Rossi 2 

1 Dipartimento di Scienze Farmaceutiche, Università di Modena e Reggio Emilia, via Campi 183, 41100 Modena, Italy 

2 Centro Interdipartimentale Grandi Strumenti, Università degli Studi di Modena e Reggio Emilia, via Campi 213 ⁄ A, 41100 Modena, Italy 

(Received 2 February 2010; Accepted in revised form 4 May 2010) 

Summary Propolis has various biological activities closely related to the composition which varies according to 

environmental factors and also to the method of production. This study was aimed at determining whether 

or not HR-NMR and multivariate statistical analysis were able to classify propolis according to the method 

used to harvest it. Sixty propolis samples were analysed in all. The ethanolic propolis extracts were initially 

analysed for quantification of the main bioactive substances, balsams and waxes. The 1 H NMR and 

heteronuclear multiple bond correlation spectra were then acquired. Spectral data were analysed by the 

application of multivariate statistical techniques (Factor Analysis and General Discriminant Analysis). 

The best results were obtained using the 1 H NMR which furnishes a sufficiently effective model by analysing 

the spectral region between 4.50 and 13.00 ppm (predictive capacity: 96.7%). 

Keywords Chemical fingerprint, chemometrics, harvesting methods, NMR, propolis. 

Introduction 

Propolis or bee glue is a product based on resins 

collected by bees from plant exudates and contains more 

than 160 constituents. The propolis is used to defend the 

beehive from the intruders, to prevent decomposition of 

animals killed after invading the colony and for the 

thermal isolation (Greenaway et al., 1990). It is characterised 

by a mean content of 50% balsams and resins, 

30% waxes, 10% essential and aromatic oils, 5% pollen, 

5% various other substances and usually contains a 

variety of chemical compounds, such as polyphenols 

(flavonoids, phenolic acids and their esters), terpenoids, 

steroids and amino acids. Flavonoids are thought to be 

responsible for many of its biological and pharmacological 

activities including anticancer (Matsuno, 1995), 

anti-inflammatory (Wang et al., 1993), antimicrobial 

(Kujumgiev et al., 1999) and antioxidant effects (Nieva 

Moreno et al., 2000). It is generally accepted and 

demonstrated that in temperate zones, including Europe, 

Asia and North America, the bud exudates of 

Populus species are the main source of propolis (Nagy 

et al., 1986; Greenaway et al., 1987); indeed this is the 

main propolis type also in Italy and in particular in the 

limited geographic area from which the samples of this 


e-mail: davide.bertelli@unimore.it 


study derive, where the Populus species are the dominant 

ones. The poplar propolis can be characterised by its 

content of bioactive components. In 2007, Popova et al. 

suggested the following minimum contents for the 

poplar propolis, which could be used for standardisation 

and quality control: balsam 45%, total phenolics (TP) 

21%, total flavones and flavonols (TFF) 4%, total 

flavanones and dihydroflavonols (TFD) 4%. Smell, 

colour, constitution, and composition of propolis vary 

according to not only the different botanical sources, 

but also the geographical origin and the climatic 

conditions (Bankova et al., 2002). Besides, also the 

method of harvest can effect composition (Rossi, 2006; 

Ramanauskiene et al., 2008) and properties of propolis. 

Sales et al. (2006) and Bedascarrasbure et al. (2004) 

reported the effect of the method of harvest on the lead 

propolis content caused by the metal chelating capacity 

of flavonoids, and it can be concluded that the harvesting 

methods employed for gathering propolis cause 

chemical composition to vary. The influence of the 

harvesting method on composition and properties of 

propolis seems to be related to the recognised behaviour 

of the bees that cover the smooth and narrow surfaces, 

as the thin space created by the wedges, with propolis 

that is purer and richer in balsams and therefore in 

bioactive compounds such as flavonoids, whereas they 

use the propolis mixed with major amount of waxes to 

cover the largest and irregular surfaces of the beehive or 

doi:10.1111/j.1365-2621.2010.02310.x 


the propolis mat used for the production (Alı` et al., 

1997). The amount of balsams is an important characteristic 

because high percentage of balsams means the 

propolis contained a low percentage of wax and 

insoluble matter and a higher content of biologically 

active components (Kujumgiev et al., 1999). In Italy, the 

propolis is commonly harvested using one of the 

following three methods: (i) the scraping, which consists 

in scraping the inner surface of the beehive including the 

frames holding the wax combs; (ii) the wooden wedges, 

which is obtained through scraping propolis deposed by 

bees to close the space created by thin wooden wedges 

(3–5 mm thick) interposed between the super and the 

cover of the hive to maintain the microenvironment; 

(iii) the propolis mat, in which the beekeeper scrapes the 

plastic mesh placed between the super and the cover of 

the hive. The propolis has been used since the ancient 

time; however, it has recently achieved more popularity 

as it is widely used in drinks and foods to improve the 

health and prevent diseases (Banskota et al., 2001). 

Considering the widespread use of propolis and the 

increased interest on the propolis characterisation, this 

study was conducted to evaluate whether the highresolution 

nuclear magnetic resonance (HR-NMR) used 

as chemical fingerprint analysis coupled with multivariate 

statistical methods is able to classify poplar propolis 

samples according to the method of harvest. In this 

regard, sixty raw propolis, collected with different 

methods, were extracted with ethanol and, before 

performing the NMR experiments, were analysed for 

quantification of the three main groups of bioactive 

substances (TP, TFF, TFD), balsams and waxes to 

confirm the poplar source. 

In recent years, the use of much higher magnetic fields 

and the greater sensitivity that they bring have stimulated 

interest in one-dimensional (1D) and bidimensional 

(2D) NMR spectroscopy as a routine method for the 

analysis of complex mixtures (Fan, 1996; Charlton 

et al., 2002). In general, there have been several applications 

of NMR spectroscopy to propolis analysis 

(Watson et al., 2006; Cuesta-Rubio et al., 2007; 

Kumazawa et al., 2007); however the 1D and 2D 

NMR, to our knowledge, have never been previously 

applied to propolis in effort to classify samples according 

to the method of harvest. 

Material and methods 

Chemicals and apparatus 

All reagents, solvents (analytical grade) and standard 

compounds (caffeic and ferulic acids, apigenin, chrysin, 

galangin, kaempferol, pinocembrin, pinostrobin and 

quercetin) were purchased from Fluka (Buchs, Switzerland). 

Dinitrophenylhydrazine reagent was prepared by 

dissolving 1 g of 2,4-dinitrophenylhydrazine (2,4-D) in 

Classification of different propolis by HR-NMR G. Papotti et al. 1611 

2 mL of 96% sulphuric acid and then diluting to 

100 mL with methanol. The absorbances were measured 

with a Varian Cary 50 Bio UV-Visible spectrometer 

(Torino, Italy). Wilmad NMR tube, 5 mm, Ultra- 

Imperial grade, 7 in. L, 526-PP were purchased from 

Sigma-Aldrich (Milan, Italy). One-dimensional and 

bidimensional NMR spectra were acquired with a 

Bruker FT-NMR Avance 400 spectrometer (Ettlingen, 

Germany). 

Propolis samples 

To reduce the influence of the geographical origin, the 

botanical source and the climatic variability on the 

propolis composition, all the samples were provided by 

several local beekeepers and by CRA-API (Consiglio 

per la Ricerca e la Sperimentazione in Agricoltura- 

Istituto Nazionale di Apicoltura e Bachicoltura) (Bologna, 

Italy) in the early summer of 2007 and were 

obtained from colonies located in a restricted area of 

the Emilia Romagna in which it is known that bees 

produce propolis mainly from poplar exudates. Sixty 

samples were supplied in all: seventeen obtained by 

scraping, twenty-six obtained by wooden wedges and 

seventeen obtained by propolis mat. For each sample, 

about 50 g was furnished. Because the sampling 

conditions ensured minimal influence from environmental 

factors on the propolis composition, all the 

samples were considered individually. All the analyses 

were carried out in triplicate, and the results are 

reported as mean ± SD. 

Propolis extraction 

Ethanol was used because it is the extraction solvent for 

the most common commercial propolis products. Each 

propolis sample was frozen, chopped into small pieces 

and homogenised, and from this bulk material, 1 g 

propolis, exactly weighed, was extracted with 10 mL of 

solvent with continuous stirring at room temperature 

(twice after 24 h). Each extract was filtered into a 

volumetric flask, and the volume made up to 25 mL 

with ethanol. 

Total phenolics content 

Total phenolics were estimated by the Folin-Ciocalteau 

method properly modified (Singleton et al., 1999). 

A volume of 50 lL of extract diluted 1:50 (v ⁄ v) in 

ethanol was mixed with 2.5 mL of the Folin-Ciocalteau 

reagent 1:10 (v ⁄ v) and 2.0 mL of Na 2CO 3 hot saturated 

solution. Absorbance was measured at 760 nm after 

5-min incubation at 50 °C. Gallic acid was used for the 

calibration curve (20–800 lg mL )1 ). TP were expressed 

as milligram gallic acid equivalents per gram of propolis 

(GAEs g )1 ). 


1612 

Classification of different propolis by HR-NMR G. Papotti et al. 

Total flavones and flavonols content 

The procedure described by Woisky & Salatino (1998) 

was employed. For the calibration, quercetin standard 

solutions in 80% ethanol (25, 50, 100 and 200 lg mL )1 ) 

were used. The standard solutions (0.5 mL) were mixed 

with 1.5 mL of 95% ethanol, 0.1 mL of 10% AlCl3 in 

water (w ⁄ v), 0.1 mL of 1 m potassium acetate and 

2.8 mL of 80% ethanol. After incubation at 20 °C for 

30 min, the absorbance was measured at 425 nm. The 

amount of 10% AlCl3 was substituted by distilled water 

in blank. Similarly, 0.5 mL of each extract diluted 1:50 

(v ⁄ v) in 80% ethanol was reacted with AlCl3 as 

described earlier. The results were expressed as mg g )1 

propolis. 

Total flavanones and dihydroflavonols content 

The method described by Nagy & Grancai (1996) was 

used, with minor modifications. Four standard solutions 

of pinocembrin in methanol (200, 500, 1000 and 

1500 lg mL )1 ) were used for the calibration. One 

millilitre of these solutions was separately reacted with 

2 mL of 2,4-D reagent and 2 mL of methanol at 50 °C 

for 50 min. After cooling, the reaction mixture was 

mixed with 5 mL of 10% KOH in 70% methanol (w ⁄ v) 

and incubated at 20 °C for 2 min. Then, the mixture was 

mixed with 5 mL of methanol and centrifuged at 

671 · g for 5 min. The supernatant was collected and 

adjusted to 25 mL, and the absorbance was measured at 

495 nm. A blank solution with 1 mL of methanol 

instead of the pinocembrin was used in an analogous 

procedure. The extracts were reacted with 2,4-D as 

described earlier. A blank sample solution with 2 mL of 

methanol instead of 2,4-D reagent was used. The results 

were expressed as mg g )1 propolis. 

Balsams and waxes contents 

The method described by Popova et al. was used for 

balsams (2007). Two millilitres of each extract was 

evaporated under vacuum until constant weight, and the 

percentages of balsam in the extracts were therefore 

calculated as the ethanol soluble fraction. 

Waxes were estimated using the method described by 

Woisky & Salatino (1998). Each row propolis sample 

was extracted with petroleum ether at 40–60 °C, and 

after evaporation, the residue was heated under reflux 

with ethanol until clear solution was obtained. After 

cooling, the solid waxes were filtered and weighed. The 

results were expressed as mg g )1 propolis. 

Sample preparation for spectroscopic analysis 

To prepare NMR samples, 1 mL of each extract was 

transferred to a NMR tube and evaporated to dryness at 

room temperature under nitrogen flow. The residue was 

dissolved in 0.5 mL of methyl sulphoxide-d 6 (DMSOd6), 

selected for its high solvent capacity, and 20 lL of 

tetramethylsilane (TMS) was added as reference compound. 

Each sample was immediately analysed and used 

for both 1D and 2D NMR experiments. Also standard 

samples, prepared by dissolving 15 mg in 0.5 mL of 

DMSO-d6 and adding 20 lL of TMS as reference 

compound, were analysed. 

NMR experiments 

To obtain comparable spectra, the propolis and standard 

samples were analysed using the same acquisition 

parameters. All experiments were performed at 300 K. 

The 1 H NMR spectra were measured at 400.13 MHz. 

Time domain (number of data points) was 16K, the 

acquisition time 1.3665 s, the delay time 1.0 s, and the 

numbers of scans 256. Spectral width was 5995.204 Hz. 

Total acquisition time was 10 min 19 s. The 1 H- 13 C 

heteronuclear multiple bond correlation (HMBC) spectra 

are a 2D 1 H ⁄ 13 C correlation via heteronuclear zero 

and double-quantum coherence optimised on longrange 

couplings with low-pass j-filter to suppress one-bond 

correlations, no decoupling during acquisition, and 

gradient pulses for selection. The acquisition parameters 

were as follows: number of scans, 32; dummy scans, 

16; time domain (number of data points), 3K in F2 ( 1 H) 

and 100 in F1 ( 13 C); spectral width, 5592.841 Hz in F2 

( 1 H) and 20124.465 Hz in F1 ( 13 C); digital resolution, 

1.8206 Hz in F2 ( 1 H) and 201.207 Hz in F1 ( 13 C); 

acquisition time, 0.2747 s; delay time, 0.5 s; HMBC 

delay time, 62.5 ms. Total acquisition time was 82 min 

and 11 s. The choice of 1 H- 13 C HMBC instead of other 

2D NMR techniques depends on the fact that it 

provides a relatively high number of well-defined signals 

and does not need a phasing process, avoiding the 

introduction of a source of variability during the 

spectral calculation. Therefore, it appears to be suitable 

as a fingerprint technique. Besides, sensitivity appears to 

be sufficient for the aim of this study because signals are 

acquired on the proton channel, preventing partially the 

low sensitivities typical of 13 C spectroscopy. On the 

contrary, the HMBC is a time-consuming approach with 

respect to the simpler 1 H NMR. 

Spectral calculation 

The application of HR-NMR technique to propolis 

samples generates very complicated spectra that need to 

be previously processed and subsequently analysed by 

chemometric methods. First of all, 1 H-NMR spectra 

were calibrated using the TMS signal, whereas the 

1 13 

H- C HMBC spectra were calibrated using the residual 

ethanol signal (-CH3, 57.5 ppm in F1 ( 13 C) and 

1.1 ppm in F2 ( 1 H)) which is present in all the spectra. 


1 H-NMR 

The 1 H NMR spectra were used as intensity. Each 

spectrum generated a file containing 16K data points 

corresponding to time domain that is the number of 

points acquired and digitalised by the instrument along 

the spectral width of 5995.204 Hz to obtain the Free 

Induction Decay and then converted into a frequency 

domain spectrum by Fourier transform; these files were 

collected in a data set consisting of 16K spectral 

variables and sixty samples. Then, to reduce the number 

of data points, all the spectral regions devoid of signals 

and the solvents signals were not considered; besides, 

also the spectral resolution was reduced, obtaining a 

data set with 2159 variables and sixty samples. Three 

other data sets were prepared: the first one referred to 

the spectral region between 4.50 and 13.00 ppm, the 

second one that contains principally the aromatic 

compounds signals, between 4.50 and 8.25 ppm and 

the third one that contains the carbonylic and carboxylic 

protons signals between 8.25 and 13.00 ppm. These 

three data sets contain 1930, 1087 and 1350 variables, 

respectively. The 1 H NMR spectra were also integrated 

using the software Amix 3.7.10 (Bruker Biospin GMBH, 

Rheinstetten, Germany). The integration was performed 

by considering the signals present in the spectra 

obtained from all the samples analysed. Thirty-seven 

signals were integrated in all and used for chemometric 

analyses. 

1 H- 13 C HMBC 

To apply chemometric analysis to HMBC spectra 

previously calibrated, the spectral correlations were 

integrated, and the volume was calculated by the 

spectrometer software. This program adds together the 

intensity of the points located in previously manually 

defined areas surrounding the correlations; all spectra 

were processed using the same map of regions of 

interest. The integrated signals selection was performed 

by verifying the signals present in the spectra of all the 

samples analysed and the absence of overlapping 

adjacent signals, consequently 161 regions of interest 

were defined and used for chemometric analyses. 

Besides, even in this case, a data set of 112 variables 

and sixty samples was obtained, considering only the 

spectral region between 4.50 and 13.00 ppm in F2 ( 1 H). 


For the propolis composition, the analysis of variance 

(anova) was used to evaluate the statistical significance 

of measured differences between the propolis. To 

evaluate the most important variables which discriminate 

between the propolis, post hoc test was performed 

using the Tukey ‘honest significant difference test’. For 

all these monovariate tests, the P level was set at 0.05. 

Before the analyses on NMR spectra were performed, 


all data were normalised. To achieve a reliable classification 

of the different propolis samples, unsupervised 

and supervised pattern recognition procedures were 

applied to the data sets. Factor analysis (FA) (Burt, 

1950) and general discriminant analysis (GDA) 

(McLachlan, 1992) were used in this work to classify 

propolis according to their NMR fingerprint. FA is a 

multivariate procedure that allows a large portion of the 

total variance of data to be expressed with a smaller 

number of variables and can be used to identify the most 

significant of original factors. GDA is a supervised 

technique used to determine whether a given classification 

of cases into a number of groups is an appropriate 

one. In this work, to perform GDA, a reduction in 

variables with respect to complete data sets was necessary; 

therefore the number of variables was reduced, 

considering only the signals which presented a factorial 

weight during FA > |0.8|. The data set of thirty-seven 

variables and sixty samples obtained by applying the 

integration pattern was statistically analysed by anova 

test; therefore only the statistically significant variables 

(P < 0.05) were considered and used for the following 

GDA analysis. After the construction of the models, to 

evaluate the classification performance, the leave-one 

out method (Henrion & Henrion, 1994) was used as a 

validation procedure. All statistical calculations were 

performed using Statistica 6.1 (StatSoft Ò Italia, Vigonza, 

Italy) and spss 13.0 (SPSS Inc., Chicago, IL, USA). 


Propolis composition 

In Table 1 the phenolics, balsams and waxes mean 

contents are reported. As shown, the results are in large 

agreement with the literature data (Popova et al., 2007) 

for poplar propolis and this confirmed poplar as the 

source of the samples in the current study. The anova 

and the post hoc tests show statistically significant 

differences between scraping and wedges and propolis 

mat for TP, TFF and TFD. Considering these results, 

the NMR experiments were performed. 

1 

H NMR spectra (0–13.00 ppm) 

In Table 2, the NMR assignments of the most important 

propolis components, obtained by analysing 

standards, which are readily identifiable in our monodimensional 

spectra, are reported. 

A typical 1 H-NMR spectrum of a propolis extract is 

shown in Fig. 1. All the 1 H NMR spectra of the propolis 

obtained with different harvesting methods were compared 

using the spectra positional matching technique 

available with the software Amix 3.7.10 (BRUKER 

BIOSPIN GMBH). The goal of this strategy is to 

calculate similarities using the peaks distributions of 


1614 


Table 1 Phenolic compounds, balsams and waxes mean content of propolis samples a 

Samples n 

Total phenolics 

mg GAEs g )1 

Total flavones 

and flavonols 

mg g )1 

automatically calibrated spectral regions. For this purpose, 

the entire spectral region between 0 and 13.00 ppm 

was used. A match factor of 95.22% was obtained. This 

fact confirmed that the samples showed limited and not 

significant qualitative differences, and they derived from 

the same botanical species. After applying the FA, a 

data set of thirty variables and sixty samples was 

obtained. The GDA did not provide good graphic 

classification results. The best model was obtained using 

the forward stepwise method which introduced twentyfive 

variables, and it was only marginally able to group 

samples as a function of the method of harvest. The 

first canonical function (CF) explains 68.2% of the 

total variance, and the results of the leave-one out 

cross-validation show a predictive capacity of 83.3% 

(Table S1). The first CF appears to be particularly 

correlated with signals of the phenolic compounds. 

Considering these results, the next step was to analyse 

only the spectral region between 4.50 and 13.00 ppm, 

which includes the aromatic signals of flavonoids except 

the low frequency region between 0 and 4.50 ppm, 

which contains the signals of aliphatic moiety of 

different compounds and in particular of waxes. 

1 

H NMR spectra (4.50–13.00 ppm) 

After applying the FA, a data set of fifty-three variables 

and sixty samples was obtained. Figure 2 shows the 

results of the GDA. Also in this case, the best model was 

obtained using the forward stepwise method which 

introduced twenty-three variables, and it is able to 

group the propolis samples in three evident clusters. The 

first CF explains 79.4% of the total variance, and the 

results of the leave-one out cross-validation show a 

predictive capacity of 96.7% (Table S1). The first and 

the second CF are particularly correlated with the 

signals of the 5OH protons of chrysin, quercetin, 

kemferol, galangin, pinostrobin and pinocembrin. Also 

the broad and unresolved signal attributed to 7OH of 

chrysin, quercetin and pinostrobin and the evident 

signal of unknown compound at 11.40 ppm are significantly 

correlated with CFs. Besides, the phenolic 

Total flavanones 

and dihydroflavonols 

mg g )1 

Balsams 

mg g )1 

Waxes 

mg g )1 

Scraping 17 301 ± 44(A) 44 ± 10(A) 54 ± 18(A) 564 ± 87 235 ± 26 (A) 

Wedges 26 321 ± 60(A) 50 ± 11(A) 52 ± 18(A) 633 ± 82 150 ± 43 (B) 

Propolis mat 17 256 ± 54(B) 29 ± 10(B) 37 ± 14(B) 601 ± 58 235 ± 46 (A) 

P (ANOVA) b 

0.002

Table 2 Assignment of 1 H signals of phenolic compounds identified in sample spectra using DMSO-d6 as solvents 

Compound 

Chemical 

shift (d) 

1 H multiplicity Assignment Compound 

bioactive compounds of propolis, such as flavonoids, 

phenolic acids and their esters, provide signals in both 

the spectral regions considered. 

1 

H NMR spectra analysed by the integration pattern 

After applying the anova test, a data set of twenty-nine 

variables and sixty samples was obtained. Also in this 

case, the GDA did not provide good graphic results; 

Chemical 

shift (d) 

1 H multiplicity Assignment 

Apigenin 6.19 d 6H Chrysin 6.22 d 6H 

6.46 d 8H 6.51 d 8H 

6.75 s 3H 6.94 s 3H 

6.92 d 3¢H; 5¢H 7.58 m 3¢H; 4¢H; 5¢H 

7.90 d 2¢H; 6¢H 8.04 d 2¢H; 6¢H 

10.40 s 4¢OH 10.90 s (bb) 7OH 

10.75 s 7OH 12.82 s 5OH 

12.96 s 5OH Galangin 6.16 d 6H 

Kaempferol 6.20 d 6H 6.40 d 8H 

6.44 d 8H 7.44 m 3¢H; 4¢H; 5¢H 

6.93 d 3¢H; 5¢H 8.08 d 2¢H; 6¢H 

8.05 d 2¢H; 6¢H 9.59 s 3OH 

9.35 s 3OH 10.59 s 7OH 

10.10 s 4¢OH 12.31 s 5OH 

10.78 s 7OH Caffeic Acid 6.17 d C2H 

12.48 s 5OH 6.76 d 5¢H 

Quercetin 6.19 d 6H 6.96 dd 2¢H 

6.41 d 8H 7.03 d 6¢H 

6.89 d 5¢H 7.42 d 3H 

7.54 d 6¢H 9.13 s 3¢OH 

7.73 d 2¢H 9.52 s 4¢OH 

9.18 s 3OH 12.10 s COO(H) 

9.45 s 3¢OH Ferulic Acid 3.82 s 3¢O-CH 3 

9.59 s 4¢OH 6.36 d 2H 

10.75 s(bb) 7OH 6.79 d 5¢H 

12.48 s 5OH 7.08 dd 2¢H 

Pinostrobin 2.83 dd 3bH 7.28 d 6¢H 

3.29 dd 3aH 7.49 d 3H 

3.80 s O-CH3 9.54 s 4¢OH 

5.62 dd 2H 12.07 s COO(H) 

6.11 d 6H 

6.15 d 8H 

7.42 m 3¢H; 4¢H; 5¢H 

7.54 d 2¢H; 6¢H 

12.12 s 5OH 

Pinocembrin 2.79 dd 3bH 

3.23 dd 3aH 

5.58 dd 2H 

5.93 dd 6H; 8H 

7.41 m 3¢H; 4¢H; 5¢H 

7.52 d 2¢H; 6¢H 

10.79 s(bb) 7OH 

12.13 s 5OH 

s, singlet; d, doublet; dd, doublet of doublets; m, multiplet; bb, broad band. 


however the best model was achieved using the forcedentry 

method which introduces all independent variables 

that satisfy tolerance criteria simultaneously. The first 

CF explains 76.3% of the total variance, and the results 

of the leave-one out cross-validation show a predictive 

capacity of 61.7% (Table S1). The poor results could be 

explained by the fact that the propolis extracts provided 

spectra which were too complicated to be correctly 

integrated, besides also the manual phasing process 


1616 


Figure 2 Score plot of the first two canonical functions for the data set 

obtained by 1 H NMR spectra related to spectral region between 4.50 

and 13.00 ppm showing the separation of samples in three evident 

clusters corresponding to the harvesting methods: scraping (4), 

wedges (+), propolis mat (h), group centroids (d). 

introduced a source of variability during the spectral 

calculation. All this explains the importance of directly 

applying the chemical fingerprint analysis to the spectral 

signals avoiding other processing procedures. 

1 13 

H- C HMBC spectra 

A typical 1 H- 13 C HMBC spectrum of a propolis extract 

is shown in Fig. 3. After applying the FA, data sets of 

ninety-six variables and sixty samples for the complete 

spectral region and sixty-one variables and sixty samples 

for the spectral region between 4.50 and 13.00 ppm in 

Figure 1 1 H-NMR spectrum of a propolis extract. 

Figure 3 1 H- 13 C heteronuclear multiple bond correlation spectrum of a 

propolis extract. 

F2 ( 1 H) were obtained. Figures 4 and 5 show the GDA 

results. With respect to the model obtained by the first 

data set (complete spectral region), the best result was 

achieved using the forward stepwise method which 

introduced thirty-eight variables. The first CF explains 

82.1% of the total variance, and the results of the leaveone 

out cross-validation show a predictive capacity of 

91.7% (Table S1). The most correlated signals are not 

only those of the phenolic compounds, but also few 

other ones in the low frequency region, in which 

aliphatic moiety signals of other substances (i.e. terpenoids, 

alcohols, esters and acids) are present. Considering 

the model obtained by the second data set [spectral 

region between 4.50 and 13.00 ppm in F2 ( 1 H)], also in 

this case the best result was achieved using the forward 



obtained by 1 H- 13 C heteronuclear multiple bond correlation spectra 

related to complete spectral region showing the separation of samples 

in three evident clusters corresponding to the harvesting methods: 

scraping (4), wedges (+), propolis mat (h), group centroids (d). 


obtained by 1 H- 13 C heteronuclear multiple bond correlation spectra 

related to the spectral region between 4.50 and 13.00 ppm in F2 ( 1 H) 

showing the separation of samples in three different clusters corresponding 

to the harvesting methods: scraping (4), wedges (+), 

propolis mat (h), group centroids (d). 

stepwise method which introduced thirty-eight variables. 

The first CF explains 63.6% of the total variance, 

and the results of the leave-one out cross-validation 

show a predictive capacity of 81.7% (Table S1). In this 

model, the most correlated signals are those of phenolic 

compounds, in particular, aromatic protons of flavonoids. 

In general, comparing these models with the 

corresponding 1 H NMR results, it is possible to affirm 

that 1 H- 13 C HMBC technique provided useful information 

from the complete spectra because it was able to 

furnish well-defined signals also in the low-frequency 

region (0–4.50 ppm), which appeared not to be well 

resolved in the 1 H NMR spectra owing to the presence 

of confused and overlapped signals. 

Conclusion 

Considering the obtained results, it can be concluded 

that the use of HR-NMR coupled with an appropriate 

data processing procedure and multivariate statistical 

methods enabled development of sufficiently effective 

and appropriate models for classifying propolis according 

to the harvesting methods. It is interesting to note 

that the best results were obtained using the 1 HNMR 

which is the simplest and fastest technique. Furthermore, 

the differences in chemical constituents’ profiles of 

the samples analysed (Table 1) give useful information 

to partly explain the cluster classification reported. As 

shown in Figs 2, 4 and 5, the first CF is always able to 

distinguish the wedges propolis samples from the 

scraping and propolis mat ones. The wedges propolis 

samples, in particular, showed higher TP, TFF and 

balsams contents and the lowest amount of waxes 

compared with the other samples. As reported in 

literature and already mentioned in this work 

(Kujumgiev et al., 1999), the amount of balsams is an 

important characteristic because, in general, a high 

percentage of balsam corresponds to a low waxes and 

insoluble matter percentages and to a higher content of 

biologically active compounds. Moreover, the developed 

models seem to be particularly related to flavonoids as 

suggested by the fact that the most influencing and 

significant signals were identified in the spectral region 

between 4.50 and 13.00 ppm. 


The authors thank the staff of CRA-API (Consiglio per 

la Ricerca e la Sperimentazione in Agricoltura-Istituto 

Nazionale di Apicoltura e Bachicoltura) (Bologna, Italy), 

in particular Dr Anna Gloria Sabatini for assistance 

during the experimental work and for providing useful 

suggestions and discussions, and the Fondazione Cassa 

di Risparmio di Modena, for financial support given 

towards the purchase of the Bruker Avance 400 spectrometer. 

References 


Alì, A., Mearelli, F., Sgrignani, M. & Camporese, A. (1997). La propoli, 

chimica, farmacologia e terapia. Firenze, Italy: Planta Medica. 

Bankova, V., Popova, M., Bogdanov, S. & Sabatini, A. (2002). 

Chemical composition of European propolis: expected and unexpected 

results. Zeitschrift fu¨r Naturforschung, 57c, 530–533. 

Banskota, A.H., Tezuka, Y. & Kadota, S. (2001). Recent progress in 

pharmacological research of propolis. Phytotherapy Research, 15, 

561–571. 


1618 


Bedascarrasbure, E., Maldonado, L. & Alvarez, A. (2004). Preliminary 

results about method of harvest’s effect on the propolis’ content of 

lead. Honeybee Science, 25, 129–131. 

Burt, C. (1950). The factorial analysis of qualitative data. British 

Journal of Psychology, 3, 166–185. 

Charlton, A.J., Farrington, W.H.H. & Brereton, P. (2002). Application 

of 1 H NMR and multivariate statistics for screening complex 

mixtures: quality control and authenticity of instant coffee. Journal 

of Agricultural and Food Chemistry, 50, 3098–3103. 

Cuesta-Rubio, O., Piccinelli, A.L., Campo Fernandez, M., Marquez 

Hernandez, I., Rosado, A. & Rastrelli, L. (2007). Chemical 

characterization of cuban propolis by HPLC-PDA, HPLC-MS, 

and NMR: the brown, red, and yellow cuban varieties of propolis. 

Journal of Agricultural and Food Chemistry, 55, 7502–7509. 

Fan, T.W.M. (1996). Metabolite profiling by one- and two-dimensional 

NMR analysis of complex mixtures. Progress in Nuclear 

Magnetic Resonance Spectroscopy, 28, 161–219. 

Greenaway, W., Scaysbrook, T. & Whatley, F.R. (1987). The analysis 

of bud exudate of Populus x euramericana, and of propolis, by gas 

chromatography-mass spectrometry. Proceedings of the Royal Society 

of London, 232, 249–272. 

Greenaway, W., Scaysbrook, T. & Whatley, F.R. (1990). The 

composition of plant origins of propolis. Bee World, 71, 107–118. 

Henrion, R. & Henrion, G. (1994). Uberwachte klassifikation. In: 

MultiVariate Datenanaly. Pp. 71–73. Berlin, Germany: Springer- 

Verlag. 

Kujumgiev, A., Tsvetkova, I., Serkedjieva, Y., Bankova, V.S., Christov, 

R. & Popov, S. (1999). Antibacterial, antifungal and antiviral 

activity of propolis of different geographic origin. Journal of 

Ethnopharmacology, 64, 235–240. 

Kumazawa, S., Ueda, R., Hamasaka, T., Fukumoto, S., Fujimoto, 

T. & Nakayama, T. (2007). Antioxidant prenylated flavonoids from 

propolis collected in Okinawa, Japan. Journal of Agricultural and 


Matsuno, T. (1995). A new clerodane diterpenoid isolated from 

propolis. Zeitschrift fu¨r Naturforschung, 50C, 93–97. 

McLachlan, G.J. (1992). Discriminant Analysis and Statistical Pattern 

Recognition. New York: Wiley. 

Nagy, M. & Grancai, D. (1996). Colorimetric determination of 

flavanones in propolis. Pharmazie, 51, 100–101. 

Nagy, E., Papay, V., Litkei, G. & Dinya, Z. (1986). Investigation of the 

chemical constituents, particularly the flavonoid components, of 

propolis and Populi gemma by the GC ⁄ MS method. Studies in 

Organic Chemistry, 23, 223–232. 

Nieva Moreno, M.I., Isla, M.I., Sampietro, A.R. & Vattuone, M.A. 

(2000). Comparison of the free radical – scavenging activity of 

propolis from several regions of Argentine. Journal of Ethnopharmacology, 

71, 109–114. 

Popova, M.P., Bankova, V.S., Bogdanov, S. et al. (2007). Chemical 

characteristics of poplar type propolis of different geographic origin. 

Apidologie, 38, 306–311. 

Ramanauskiene, K., Savickas, A., Ivanauskas, L. et al. (2008). 

Analysis of phenolic acids in propolis using the high-performance 

liquid chromatography technique. Current Nutrition and Food 

Science, 4, 209–212. 

Rossi, P. (2006). Comparazione di metodi di produzione della propoli. 

APOidea, 3, 27–32. 

Sales, A., Alvarez, A., Rodriguez Areal, M. et al. (2006). The effect of 

different propolis harvest methods on its lead contents determined 

by ET AAS and UV-visS. Journal of Hazardous Materials, 137, 

1352–1356. 

Singleton, V.L., Orthofer, R. & Lamuela-Raventos, R.M. (1999). 

Analysis of total phenols and other oxidation substrates and 

antioxidants by means of Folin-Ciocalteau reagent. Methods in 

Enzymology, 299, 152–178. 

Wang, L., Mineshita, S. & Ga, I. (1993). Anti inflammatory effects of 

propolis. Japanese Journal of Pharmacology, 24, 223–226. 

Watson, D.G., Peyfoon, E., Zheng, L. et al. (2006). Application of 

principal components analysis to 1 H-NMR data obtained from 

propolis samples of different geographical origin. Phytochemical 

Analysis, 17, 323–331. 

Woisky, R. & Salatino, A. (1998). Analysis of propolis: some 

parameters and procedures for chemical quality control. Journal of 

Food and Drug Analysis, 37, 99–105. 




Table S1. Leave-one out cross-validation results. 









Supercritical CO 2 and N 2O pasteurisation of peach and kiwi juice 

Sara Spilimbergo 1 * & Lara Ciola 2 

1 Department of Materials Engineering and Industrial Technologies, University of Trento, via Mesiano 77, 38123 Trento, Italy 

2 Trentofrutta S.p.a., via A. Degasperi 130, 38123 Trento, Italy 


Summary The microbial inactivation and qualitative parameters (pH, sugar content, titratable acidity, absorbance at 

420 nm and turbidity) of peach and kiwi juices treated at 35 °C with supercritical carbon dioxide (SC-CO 2) 

and nitrous oxide (SC-N2O) were determined as a function of pressure and treatment time. Total inactivation 

of both naturally occurring microorganisms and Saccharomyces cerevisiae strain (10 5 cfu mL )1 ) was 

obtained after 15 min of SC-CO 2 ⁄ N 2O treatment, 10 MPa and 35 °C, for both juices. No significant changes 

in chemical-physical or in sensorial characteristics between untreated and treated juice were detected. The 

results obtained demonstrate the feasibility and the potential of SC-CO 2 ⁄ N 2O treatment as an alternative 

low temperature pasteurisation process for peach and kiwi juices. 

Keywords CO 2, juice, kiwi , N 2O, pasteurisation, peach, supercritical fluids. 

Introduction 

Research efforts for the development of nonthermal 

food preservation techniques has been growing in the 

last decade because of the consumer demand for 

natural and minimally treated food products (Raso & 

Barbosa-Canovas, 2003). Thermal pasteurisation is the 

traditional and most common method in the food 

industry to prevent microbial spoilage of juices; however, 

it may cause degradation of flavour, nutrients, 

colour and texture (Gunes et al., 2005). Nonthermal 

processes, like supercritical carbon dioxide (SC-CO2), may offer better retention of natural flavour and 

nutrients in treated foods compared to the traditional 

thermal processes (Kincal et al., 2006). Microbial and 

enzymatic inactivation with SC-CO2 has been mostly 

applied to liquid food products, such as fruit juices, 

beer, wine and milk, etc; in particular, SC-CO2 has 

been shown to be a promising alternative for pasteurisation 

of fruit juice (Arreola et al., 1991; Del Pozo- 

Insfran et al., 2006). 

Three recent reviews compile the relevant current 

knowledge about the potential of SC-CO2 as a nonthermal 

inactivation technology and summarise the most 

significant state of the art, including relevant applications 

and data, in both simple suspensions and complex 

media, for the treatment of a wide range of microorganisms 

in both liquid and solid substrates (Spilimbergo 

& Bertucco, 2003a; Damar & Balaban, 2006; Garcia- 

Gonzalez et al., 2007). 

*Correspondent: E-mail: sara.spilimbergo@ing.unitn.it 

doi:10.1111/j.1365-2621.2010.02305.x 


Little information is available on the effects of the 

perceivable and chemical-physical properties of the 

products immediately after CO2 treatment and during 

storage; mainly orange and apple juice have been tested. 

Physical and chemical properties of orange juice, pH, 

Brix values and titratable acidity (TA) do not appear to 

be influenced by CO2 treatment. Yellowness and lightness 

seem to increase whereas redness seems to decrease 

(Arreola et al., 1991; Wei et al., 1991; Park et al., 2002). 

Three recent papers report about physical changes of 

apple products: no visual changes or modifications in 

total soluble solid content of apple cider have been 

found (Van Ginneken et al., 2002), whereas Gui et al. 

(2006) reported a significant reduction of the browning 

degree in CO 2-treated cloudy apple juice processed at 

55 °C, 30 MPa for 60 min, after storage at 4 °C. 

Gasperi and co-workers investigated quantitatively the 

sensorial modification induced by CO2 treatment in 

apple juice. Sensory analyses were performed on fresh 

control (i.e., untreated) and CO2-processed juices using 

both difference-from-control and ranking tests. No 

significant differences were detected between the CO 2treated 

sample and control (Gasperi et al., 2009). 

In the last year, a few interesting papers have been 

published in the field: they confirm the feasibility and the 

effectiveness of this innovative technique on different 

foodstuff. We quote the study of Garcia-Gonzalez et al. 

(2009) who investigated the effect of CO 2 at high pressure 

on the inactivation of naturally occurring microorganisms 

in liquid whole eggs; they concluded that CO2 

processing extends the shelf-life of the sample up to 

5 weeks at 4 °C, which is the current shelf-life of heat

1620 

Supercritical CO2 and N2O pasteurisation S. Spilimbergo and L. Ciola 

pasteurised liquid whole eggs. Chen et al. (2009) studied 

the influence of thermal and dense-phase carbon dioxide 

pasteurisation on physicochemical properties and flavour 

compounds in Hami melon juice. They found CO2 

efficiently induced microbial inactivation as well as 

preserved more quality attributes, including aroma compounds, 

than thermal treatment; Zhong et al. (2008) 

demonstrated that CO 2 has potential as a pasteurisation 

technology for application to leafy green vegetables. 

Interestingly, few studies reported in literature investigate 

the bactericidal action of another gas under 

pressure, N2O: Enomoto et al. (1997) showed a 5 log 

reduction of Saccharomyces cerevisiae using N2O at 

4 MPa and 40 °C for 240 min, whereas Fraser (1951) 

observed 58% reduction of Escherichia coli after N 2O 

treatment at 38 °C and 3.45 MPa for 3 min. A recent 

study reports total inactivation of microbes in fresh 

apple juice using SC-N 2O at 10 MPa and 36 °C with no 

significant chemical-physical, nutritional or sensory 

differences between N2O-treated and untreated sample 

(Spilimbergo et al., 2007a,b; Gasperi et al., 2009). 

To our knowledge, a few observations quoted in 

literature concern other fruits juices than apple and 

orange ones, in particular carrot (Park et al., 2002), 

grape (Gunes et al., 2005), grapefruit (Ferrentino et al., 

2009), coconut (Damar & Balaban, 2005), mandarin 

(Lim et al., 2006), watermelon juice (Lecky, 2005), and 

no papers concern other kind of juices such as peach or 

kiwi juices some of the most critical products to be 

pasteurised without inducing chemical-physical modifications 

(personal communication by Trentofrutta 

S.p.A., Trento, Italy). Usually the absorbance at 

420 nm, tightly related to the oxidation of sugars and 

proteins contained in juice, is the most important 

parameter for product acceptability after pasteurisation. 

After thermal treatments an increase in absorbance in 

the range within 0.100 ‚ 0.200 and 0.200 ‚ 0.300 is 

usually detected in peach and kiwi juice respectively 

compared to 0.050 ‚ 0.100 in apple juice (confidential 

data kindly provided by Trentofrutta S.p.A.). Peach 

chemical composition includes thermostable compounds 

like vitamins (B1, C), monoterpens (linalol and geraniol) 

and carotenoids, precursors of vitamin A (Riu-Aumatell 

et al., 2005). Moreover, peach juice has a high content 

of potassium, fibres, proteins and sugars, which can be 

oxidised in severe conditions such as high hydrostatic 

pressure treatment (400 MPa, 20 °C for 10 min) (Dogan 

& Erkmen, 2003); increased levels of benzaldehyde has 

been attributed to b-glucosidase activity (Sumitani 

et al., 1994). 

Kiwi juice has significant amounts of biologically 

active compounds including ascorbic acid and antioxidant 

capacity because of the content of phytonutrients, 

such as carotenoids, lutein, phenolics and flavonoids: 

most of these compounds offer benefits for specific 

health conditions (Tasselli et al., 2007). The dark green 

colour of kiwi fruit, which is attributed to the chlorophylls, 

in plastids in the pericarp cellsonce solubilised 

can be converted into the olive brown pheophorbide by 

low pH and high temperature. Pheophorbide is oxidised 

into a colourless compound (Cano, 1992; Cano et al., 

1993). 

In this context, the aim of this paper consists on 

extending the potentiality of SC pasteurisation to other 

high-sensitive products such as peach and kiwi juices. In 

detail, 

1 The efficiency of both SC-CO 2 ⁄ N 2O on microbial 

inactivation of peach and kiwi juice will be checked; 

2 The effect of SC-CO 2 ⁄ N 2O on chemical-physical 

parameters such as Brix value, pH, TA, absorbance 

and turbidity of fresh juice will be measured and 

compared to the untreated one. 


Peach and kiwi juices 

Freshly squeezed fruit juices were produced by Trentofrutta 

S.p.A.; the juice, some of which was inoculated 

with S. cerevisiae (ATCC 9763) to obtain a bioburden, 

N0, of 10 5 CFU mL )1 , was frozen in plastic bags 

(1000 mL). The evening before the treatment, the juice 

bags were thawed at 4 °C (overnight). After the treatment, 

reference and treated samples were stored at 4 °C 

before microbiological and chemical-physical analysis. 

High-pressure equipment and procedure 

The experiments were performed with the multi-batch 

pilot plant described by Spilimbergo & Mantoan (2006). 

CO2 (4.5; Messer, Torino, Italy) or N2O (2.0; Messer) 

was cooled down to 3 °C and then pumped into highpressure 

vessels by a volumetric pump (LEWA, mod. 

LCD1 ⁄ M910 s, Leonberg, Germany) with a maximum 

flow rate of 13 L h )1 . The vessels consist of two 310-ml 

cylinders and of ten 15-ml cylinders equipped with a 

magnetic system for stirring (micro-stirrer; Velp, New 

York, NY, USA). The vessels, immersed in a water 

bath, were thermally controlled inside one small and one 

big reactor, while pressure gauges measured the operative 

pressure inside the vessels (Gefran, Brescia, Italy 

and Wika, Lawrenceville, NJ, USA). Gefran probe has a 

pressure range within 0 ‚ 50 MPa, an accuracy £0.5% 

FSO (Full Scale Outputs) while Wika one has a pressure 

range of 0 ‚ 40 MPa, an accuracy £0.25% FSO. The 

operating temperature and pressure were continuously 

recorded by a real-time acquisition data system (NA- 

TIONAL INSTRUMENTS, field point FP-1000 RS 

232 ⁄ RS 485, Austin, TX, USA) and monitored by the 

software LabVIEWÔ 5.0, (NATIONAL INSTRU- 

MENTS). 


Microbial and chemical-physical trial 

Seven millilitres of juice was introduced in each vessel 

(Vmax = 20 mL) and exposed to SC gas at 35 °C for a 

treatment time within 0 ‚ 15 min, with a stirring rate of 

300 r.p.m. 

Turbidity ⁄ pH and organoleptic trial 

Seventy-five millilitres of juice was introduced in each 

vessel (Vmax = 310 mL) and exposed to SC gas at 35 °C 

for 10 min with a stirring rate of 300 r.p.m. Two 

consecutive experimental runs were performed to produce 

the total volume of 300 mL needed for sensorial, 

turbidity and pH analyses. 

Microbiological analysis 

Before and after each treatment, the viable cell counts 

were determined by standard plating technique (Speck 

et al., 1975). 

Every sample was diluted in ringer solution, (Oxoid, 

Milan, Italy) (dilution 1:10), plated onto Plate Count 

Agar (PCA), incubated at their optimal condition 

(28 °C for 48 h) before counting. PCA is a not selective 

microbiological medium, commonly used to detect total 

microbial count of a sample. The same medium was 

used to analyse samples with inoculum to count both 

S. cerevisiae and all natural occurring microorganism 

colonies in the juice. In the result section the inactivation 

data are expressed as survival % (N ⁄ N0%) vs. time, 

where N0 is the initial number of cells in the control 

sample and N is the number of cells in the sample after 

treatment (CFU ⁄ ml). Each result is the mean value of 

four different runs. 

Chemical-physical analysis 

Typical chemical and physical parameters (Benitez 

et al., 2007) such as pH, sugar content (°BX), TA, 

absorbance (A) at 420 nm and turbidity [Nephelometric 

Turbidity Units (NTU)] were measured before and after 

each SC treatment. 

NTU is a widely used parameter in food industry to 

measure the limpidity of the fruit products (Morais 

et al., 2006). After an acceptable pasteurisation treatment 

no modification of NTU should be detected. In 

some cases, a slight decrease in NTU can be measured 

after a thermal process because of the presence of 

thermolabile proteins in the juice. 

Absorbance is a crucial parameter tightly related to 

the juice browning which induces changes in visual 

characteristics of the final product. 

The initial pH, °BX, TA, A and NTU values, reported 

in Tables 1–4 for both juices, are indicated as ‘untreated’. 

No standard deviations are indicated because the resulted 

values were particularly stable, thus the standard deviation 

was insignificant. 

Supercritical CO2 and N2O pasteurisation S. Spilimbergo and L. Ciola 1621 

Table 1 °BX, TA and A at 420 nm values in peach juice before and 

after treatments at different pressures, 35 °C and 15 min, both with 

CO 2 and N 2O 

Gas 

Pressure 

(MPa) 

Untreated Treated 

°BX TA A 

D°BX 

(SD) DTA (SD) DA (SD) 

CO 2 6.3 8.8 4.4 0.225 0 (0.1) 0.4 (0.1) 0 (0.02) 

8 9.2 4.7 0.121 0 (0.1) 0.1 (0.0) 0 (0.01) 

10 10 5.3 0.123 0 (0.0) 0.2 (0.0) 0.061 (0.02) 

N2O 5.3 8.8 4.6 0.117 0 (0.0) 0 (0.1) 0 (0.01) 

8 9.1 4.7 0.128 0 (0.1) 0 (0.0) 0 (0.01) 

10 10 5.3 0.123 0 (0.1) 0 (0.1) 0.012 (0.01) 

D indicates variation compared to the untreated sample. 

SD, standard deviation; TA, titratable acidity. 

Table 2 A at 420 nm of peach and kiwi juice before and after thermal 

treatments at 90 ± 5 °C performed for 46 s in a tube in tube heat 

exchanger 

Fruit juice 


A (SD) DA (SD) 

Peach 0.170 0.505 

Kiwi 0.147 (0.007) 0.395 (0.035) 

D indicates variation compared to the untreated. 

SD, standard deviation. 

pH was measured using a digital pH meter (WTW 

Inolab, FlightTech, Belluno, Italy), °BX determined 

with a digital refractometer (model RFM 91; Bellingham 

+ Stanley, Kent, UK) at 20 °C, TA calculated with 

an automatic titolator (DL50; Mettler, Milano, Italy) 

using a NaOH solution (0.1 N), A at 420 nm measured 

with a spectrophotometer (model lambda 25; Perkin 

Elmer, Waltham, MA, USA) and NTU calculated with 

a calibrated turbidimeter (Turbiquant 1500 T, HACH 

18900-00, Merck, Whitehouse Station, NJ, USA). 

Tables 1–4 show the variation (D) of pH, °BX, TA, A 

and NTU with respect to the ‘untreated’ value together 

with the standard deviation (SD). Each value of DBX°, 

DA, DTA, DpH and DNTU is the mean of three different 

replicates. 

Organoleptic analysis 

After microbiological and chemical-physical analyses 

(2 h after SC treatment), each treated sample (75 mL) 

were poured from the 310-mL reactor into a colourless 

glass at ambient temperature and tasted by three trained 

judges to evaluate colour, flavour and taste. The used 

method follows competent indications (Poretta, 2000), 

developed from Trentofrutta according to an established 

method (Anon., 1987). Untreated juice was taken as 


1622 


Table 3 °BX, TA, A at 420 nm, pH and NTU values in peach juice before and after treatments at 35 °C, 10 MPa and 15 min, both with CO2 and 

N2O 

Gas 


°BX TA A pH NTU D°BX (SD) DTA (SD) DA (SD) DpH (SD) DNTU (SD) 

CO 2 9.2 4.2 0.103 3.7 0.20 0 (0.0) 0.2 (0.1) 0.007 (0.001) 0 (0.0) 0 (0.0) 

N2O 9.2 4.2 0.098 3.7 0.20 0 (0.0) 0 (0.0) 0 (0.001) 0 (0.0) 0 (0.0) 


SD, standard deviation; NTU, nephelometric turbidity units; TA, titratable acidity. 

Table 4 °BX, TA, A at 420 nm, pH and NTU values in kiwi juice before and after treatments at 35 °C, 10 MPa and 15 min, both with CO2 and 

N2O 

Gas 


°BX TA A pH NTU D°BX (SD) DTA (SD) DA (SD) DpH (SD) DNTU (SD) 

CO 2 13.5 10.2 0.345 3.6 1.4 0 (0.1) 0.4 (0.0)) 0.021 (0.02) 0 (0.0) 0 (0.0) 

N2O 13.5 10.2 0.098 3.5 1.2 0 (0.1) 0 (0.1) 0 (0.01) 0 (0.0) 0 (0.0) 


SD, standard deviation; NTU, nephelometric turbidity units; TA, titratable acidity. 

reference and any changes in colour, taste and flavour 

recorded by the judges. The selection and the training of 

all of judges were considered compulsory steps to 

guarantee reproducible assessments and good discriminatory 

ability. 


Microbial inactivation 

In all experimental runs the sample volume was kept 

constant at 7 mL with a stirring rate of 300 r.p.m. and 

temperature at 35 °C. This value was considered low 

enough to maintain the sensorial properties of the 

product and sufficiently high to reach a satisfactory rate 

of inhibition in a short treatment time (Spilimbergo & 

Mantoan, 2006). Pressure conditions were varied from 

cylinder pressure (5.3 and 6.3 MPa for N 2O, CO 2, 

respectively) to 8 and 10 MPa bar. Higher pressure has 

been demonstrated not to be beneficial (Spilimbergo & 

Mantoan, 2006). 

To check SC treatments efficiency in a freshly 

squeezed juice, in the first set of experiments, SC 

treatments were performed on natural juice without 

inoculum: the initial microbial count consists of naturally 

occurring microorganisms in the untreated sample 

(roughly 10 2 cfu mL )1 ). 

The inactivation kinetics of natural microflora performed 

at different pressures are shown in Fig. 1a and b. 

CO2 and N2O inactivation kinetics appear quite similar, 

and complete inactivation was achieved at 10 MPa after 

15 min, while at lower pressures the inactivation kinetic 

results quite slow with 90% inactivation after 15 min. 

The effect of the pressure on inactivation ratio could 

be explained by considering the diffusion process of the 

SC gases inside the liquid phase. It is well known that an 

increase in pressure induces an increase in diffusivity. 

In the second set of experiments, the inactivation rate 

of S. cerevisiae was evaluated to check whether the 

kinetic ratio was dependent on N0. In other words, the 

same matrix was tested with a higher initial N 0 

compared to the previous experiments (roughly 

10 2 cfu mL )1 ). Figure 2a shows the inactivation kinetics 

in peach juice obtained with both SC-CO 2 and SC-N 2O 

at 10 MPa and 35 °C with N0 of 10 5 CFU mL )1 while 

Fig. 2b shows the inactivation kinetic in kiwi juice 

obtained with both gases at the same operative conditions 

and N0. 

It is worth noticing that the time needed to obtain 

total inactivation seems to be independent of N0, at least 

at the temperature, pressure and in the range of N 0 

considered, with total inactivation resulting after 

15 min. 

Also, in this case, peach and kiwi juice inactivation 

kinetics appear quite similar. This observation suggests 

that both the gases are able to be solubilised in different 

fruit juices with similar kinetics. Solubilisation of the 

gases into the juice has been demonstrated to be the 

limiting step of the entire process (Spilimbergo et al., 

2003b). N2O and CO2 do not have very different 

solubility in water (1.1485 and 0.739 mol kg )1 at 50 °C 

and 100 bar, 1.133 mol kg )1 and 0.936 mol kg )1 at 

60 bar and 25 °C, for CO2 and N2O respectively) 

(Sun, 2003; Arao et al., 2005), similar critical temperature 

and pressure (Green & Perry, 2008), or capability to 

suppress fungal growth (El-Goorani & Sommer, 1981). 


(a) 100 

N/N 0 % 

10 

1 

0.1 

10 MPa 

8 MPa 

1E–4 

1E–5 

6.3 MPa 

0 2 4 6 8 10 12 14 16 

100 (b) 

N/N 0 % 

10 

1 

0.1 

1E–4 

1E–5 

10 MPa 

8 MPa 

6.3 MPa 

Time (min) 

0 2 4 6 8 10 12 14 16 

Time (min) 

Figure 1 Microbial inactivation kinetics in peach juice (a) after SC-CO 2 

treatment at 35 °C and different pressures; (b) after SC-N2O treatment 

at 35 °C and different pressures. N0 (10 2 CFU mL )1 ), represents the 

number of CFU mL )1 initially present in the untreated sample. 

In conclusion, we would like to remark that in the last 

decade many studies have been addressed on microbial 

aspects of SC treatment as an alternative pasteurisation 

treatment with attractive results. 

Chemical-physical analysis 

The second aim of this study was the investigation of the 

effect of SC treatments on chemical-physical qualities of 

peach and kiwi juice. In particular, sugar content (°BX), 

acidity (TA) and absorbance (A) at 420 nm values have 

been measured. Table 1 shows °BX, TA, A variations in 

peach juice after both CO2 and N2O treatment at 

different pressure conditions. 

Treatments with SC-CO2 and SC-N2O do not cause 

any significant changes in °BX, TA and A at 420 nm in 

peach juice; however, SC-CO2 seems to cause a slight 

variation in acidity, whereas SC-N2O does not seem to 


(a) 

N/N 0 % 

(b) 

N/N 0 % 

100 

10 

1 

0.1 

1E–4 

1E–5 

0 2 4 6 8 10 12 14 16 18 20 

100 

10 

1 

0.1 

1E–4 

Time (min) 

N 2 O 

CO 2 

1E–5 

0 2 4 6 8 10 12 14 16 18 20 

Time (min) 

N 2 O 

CO 2 

Figure 2 Microbial inactivation kinetics after SC-CO2 and SC-N2O 

at 10 MPa and 35 °C, (a) in peach juice, inoculated sample, 

N 0 =10 5 CFU mL )1 ; (b) in kiwi juice, uninoculated sample, 

N0 =10 2 CFU mL )1 . 

induce any. A probable reason may be the acidic 

property of CO2, which, after dissolution in aqueous 

phase, probably induces a decrease in solution pH 

whereas N2O does not. 

In whole, both SC treatments do not cause significant 

chemical-physical alterations in peach juice, especially 

with regard to A, where a variation within 0.061 and 

0.012 after CO 2 and N 2O treatment, respectively, at any 

pressure considered, was measured. The values correspond 

to a one order of magnitude purportedly less 

compared to thermal treatment. Table 2 shows typical 

absorbance variations after a thermal treatment of both 

peach and kiwi juice at 90 ± 5 °C for 46 s in a tube and 

tube heat exchanger (confidential data kindly given by 

Trentofrutta S.p.A.). The other parameters (°BX, TA, 

NTU, pH) were not significantly influenced by the 

thermal process (data not shown). 

To check all chemical-physical parameters (°BX, TA, 

A at 420 nm, pH and NTU) of peach and kiwi juice 


1624 


after SC treatment, additionally, sets of experiment were 

performed with the 310-mL reactor to collect enough 

sample volume. The treatments were performed at 

10 MPa, 35 °C and 15 min to assure total microbial 

inactivation (see Fig. 2a and b). Tables 3 and 4 summarise 

the chemical-physical data measured in peach 

and kiwi juices; they confirm the result of Table 1 – no 

significant modification in °BX, TA or A – and 

demonstrate no NTU or pH variations in peach and 

kiwi juice after SC treatments, compared to the control. 

Our results, in accordance with previous works 

(Ferrentino et al., 2009; Gasperi et al., 2009), demonstrate 

that SC treatments have less effect than thermal 

treatments on peach and kiwi juices properties. 

Organoleptic analysis 

Organoleptic analyses have been performed on untreated 

and treated samples. Peach juice treated samples 

with both gases were compared to the control; they 

appeared slightly lighter and clouded, SC-CO2–treated samples had the taste of hazel whereas SC-N2O treated 

ones did not show any modifications in taste or flavour. 

Concerning kiwi juice, the judges could not detect any 

appreciable differences between treated and untreated 

samples either in colour, taste or flavour after any SC 

treatments. 

The results obtained show the high potential of 

SC-N2O process to be exploited as a low-temperature 

pasteurisation technique. However, a quantitative analysis 

by means of a panel test should be performed to deep 

further the sensorial effects on peach and kiwi juices. 

Conclusion 

The effect of SC-CO2 and SC-N2O on the microbial and 

chemical-physical quality of fresh peach and kiwi juices 

has been investigated. 

The feasibility of CO2 and N2O treatment as innovative 

pasteurisation technique has been confirmed: a 15-min 

treatment at 10 MPa and 35 °C is sufficient to assure a 

total inactivation of natural or inoculated (S. cerevisiae, 

N0 10 5 CFU mL )1 ) microorganisms in the samples. 

CO 2 and N 2O treatment do not modify the chemical, 

physical and organoleptic qualities of the fresh product. 

In particular, SC-N2O treated juice shows no differences 

in °BX, TA, pH and NTU when compared to fresh 

juice, whereas SC-CO 2 causes a slight impact in TA. 

Nevertheless, knowledge about the inactivation mechanism 

of SC-CO2 is still under debate. It probably 

involves the peculiar ability of CO 2, a lipophilic compound, 

to solubilise into the phospholipid bilayer of the 

cell membrane and to highly increase its permeability. So, 

CO 2 can penetrate through the membrane and hence 

accumulate in the cytoplasmic interior of the cells. If too 

much dissolved CO2 enters the cytoplasm, the cells may be 

unable to expel all the resulting protons and intracellular 

pH may start decreasing (Garcia-Gonzalez et al., 2007). 

As concerns SC-N2O, no mechanism has ever been 

postulated. What it is known so far is just that N2O isa 

lipophilic and anaesthetic gas which can easily penetrate 

through the phospholipidic cell membrane. The anaesthetic 

potency is closely correlated with the lipid 

solubility (Overton, 1901). Anaesthetics exert their 

primary effects by dissolving in the lipid bilayer of cell 

membranes inducing changes in membrane fluidity, 

volume and thickness (Miller, 1985). 

In medical practice, N2O is used as an analgesic and 

anaesthetic gas (Leshem & Wills, 1998), binds lipids and 

protein and inhibits some enzymes, like cytochrome 

c activity in mitochondrial particle isolated from seed, 

leaf or cellular suspension (Qadir & Hashinaga, 2001). 

N2O inhibits methionine synthase activity (Frasca et al., 

1986) producing a methionine deficit, that alters protein 

synthesis in the cell and giving rise to an accumulation 

of homocysteine (Young et al., 1997). It is known that 

homocysteine has pro-oxidative activity on phospholipids 

in presence of some cellular ions; this reaction 

increases cells membrane permeability. 

This could be an explanation of N2O microbial 

inactivation effect. However, experiments are needed 

to confirm this hypothesis. 


This work was supported by Provincia Autonoma di 

Trento project. The authors are grateful for the skilful 

technical assistance of Laura Genetti and Giorgia 

Bortolameotti and of all the technicians in the Microbiological 

and Chemical-Physical Laboratory (Trentofrutta 

S.p.A.). We also thank Giovanni Gallerani 

(Mace` s.r.l.) for his helpful suggestions and Claudia 

Contrini and Stefano Castioni for their help in carrying 

out the experimental runs. 

References 

Anon. (1987). RSK-Werte Die Gesamtdarstellung - Richtwerte und 

Schwankungsbreiten bestimmter Kennzahlen mit ueberarbeiteten 

Analysenmethoden - V d F Verband der deutschen Fruchtsaftindustrie 

e. V. - Bonn, Verlag Flussiges Obst, Shonborn. 

Arao, T., Hara, Y. & Tamura, K. (2005). Effect of high pressure gas on 

yeast growth. Bioscience Biotechnology and Biochemistry, 69, 1365– 

1371. 

Arreola, A.G., Balaban, M.O., Wei, C.I., Peplow, A. & Marshall, M. 

(1991). Effect of supercritical carbon dioxide on microbial populations 

in single strength orange juice. Journal of Food Quality, 14, 275–284. 

Benitez, E.I., Genovese, D.B. & Lozano, J.E. (2007). Scattering 

efficiency of a cloudy apple juice: effects of particles characteristics 

and serum composition. Food Research International, 40, 915–922. 

Cano, M.P. (1992). Freezing preservation of four Spanish kiwi fruit 

cultivars (Actinidia chinensis, Planch): chemical aspects. Zeitschrift 

fuer Lebensmittel-Untersuchung und-Forschung, 196, 142–146. 

Cano, M.P., Marin, M.A. & De Ancos, B. (1993). Pigment and colour 

stability of frozen kiwi-fruit slices during prolonged storage. 


Zeitschrift fuer Lebensmitteluntersuchung und Forschung, 197, 346– 

352. 

Chen, J.L., Zhang, J., Feng, Z.S., Song, L.J., Wu, J.H. & Hu, X.S. (2009). 

Influence of thermal and dense-phase carbon dioxide pasteurization 

on physicochemical properties and flavor compounds on hami melon 

juice. Journal of Agricultural and Food Chemistry, 57, 5805–5808. 

Damar, S. & Balaban, M.O. (2005). Cold pasteurization of coconut 

water with a dense phase CO2 system. In: IFT Annual Meeting Book 

of Abstracts. New Orleans, La. Chicago. Abstract nr 54F-2. 

Damar, S. & Balaban, M.O. (2006). Review of dense phase CO 2 

technology: microbial and enzyme inactivation, and effects on food 

quality. Journal of Food Science, 71, R1–R11. 

Del Pozo-Insfran, D., Balaban, M.O. & Talcott, S.T. (2006). Microbial 

stability phytochemical retention, and organoleptic attributes of 

dense phase CO2 processed muscadine grape juice. Journal of 


Dogan, C. & Erkmen, O. (2003). Ultra high hydrostatic pressure 

inactivation of Escherichia coli in milk, and orange and peach juices. 

Food Science and Technology International, 9, 403–409. 

El-Goorani, M.A. & Sommer, N.F. (1981). Effects of modified 

atmospheres on postharvest pathogens of fruits and vegetables. 

Horticultural Reviews, 3, 412–461. 

Enomoto, A., Nakamura, K., Nagai, K., Hashimoto, T. & Hakoda, 

M. (1997). Inactivation of food microorganisms by high-pressure 

carbon dioxide treatment with or without explosive decompression. 

Bioscience Biotechnology Biochemistry, 61, 1133–1137. 

Ferrentino, G., Plaza, M.L., Ramirez-Rodrigues, M., Ferrari, G. & 

Balaban, M.O. (2009). Effects of dense phase carbon dioxide 

pasteurization on the physical and quality attributes of a red 

grapefruit juice. Journal of Food Science, 74, E333–E341. 

Frasca, V., Stephenson Riazzi, B. & Matthews, R.G. (1986). In vitro 

inactivation of methionine synthase by nitrous oxide. Journal of 

Biological Chemistry, 261, 15823–15826. 

Fraser, D. (1951). Bursting bacteria by release of gas pressure. Nature, 

167, 33–34. 

Garcia-Gonzalez, L., Geeraerd, A.H., Spilimbergo, S., Elst, K., Van 

Ginneken, L. & Debevere, L. (2007). High pressure carbon dioxide 

inactivation of microorganisms in foods: the past, the present and 

the future. International Journal of Food Microbiology, 117, 1–28. 

Garcia-Gonzalez, L., Geeraerd, A.H., Elst, K., Van Ginneken, L., Van 

Impe, J.F. & Debevere, L. (2009). Inactivation of naturally 

occurring microorganisms in liquid whole egg using high pressure 

carbon dioxide processing as an alternative to heat pasteurization. 

Journal of Supercritical Fluids, 51, 74–82. 

Gasperi, F., Aprea, E., Biasioli, F. et al. (2009). Effects of supercritical 

CO2 and N2O pasteurisation on the quality of fresh apple juice. Food 

Chemistry, 115, 129–136. 

Green, D.W. & Perry, R.H. (2008). Perry’s Chemical Engineers’ 

Handbook, 8th edn. New York, USA: McGraw-Hill. 

Gui, F.Q., Wu, J.H., Chen, F. et al. (2006). Change of polyphenol 

oxidase activity, color, and browning degree during storage of 

cloudy apple juice treated by supercritical carbon dioxide. European 

Food Research and Technology, 223, 427–432. 

Gunes, G., Blum, L.K. & Hotchkiss, J.H. (2005). Inactivation of yeasts 

in grape juice using a continuous dense phase carbon dioxide 

processing system. Journal of Science of Food and Agriculture, 85, 

2362–2368. 

Kincal, D., Hill, W.S., Balabam, M. et al. (2006). A continuous highpressure 

carbon dioxide system for cloud and quality retention in 

orange juice. Journal of Food Science, 71, C338–C344. 

Lecky, M. (2005). Shelf life evaluation of watermelon juice after 

processing with a continuous high pressure CO 2 system. in: IFT 

Annual Meeting book of Abstract, New Orleans, La. Chicago. 

Abstract nr 54F-5. 

Leshem, Y.Y. & Wills, R.H.B. (1998). Harnessing senescence and 

delaying gases nitric oxide and nitrous oxide: a novel approach to 

postharvest control of fresh horticultural produce. Biologia Plantarum, 

41, 1–10. 


Lim, S., Yagiz, Y. & Balaban, M.O. (2006). Continuous high pressure 

carbon dioxide processing of mandarin juice. Food Science and 

Biotechnology, 15, 13–18. 

Miller, K.W. (1985). The nature of the site of general anesthesia. In: 

International Review of Neurobiology (edited by J. Smythies & R. 

Bradley). Pp. 1–61. Vol. 27. Orlando, FL: Academic Press. 

Morais, I.P.A., Toth, I.V. & Rangel, A.O.S.S. (2006). Turbidimetric 

and nephelometric flow analysis: concepts and applications. Spectroscopy 

Letters, 39, 547–579. 

Overton, E. (1901). Studien uber die Narkose Zugleich ein Beitrag zur 

Allgemeinen Pharmakologie. Jena, Germany: Gustav Fischer Verlag 

Jena GmbH. 

Park, S.-J., Lee, J.-I. & Park, J. (2002). Effects of a combined process 

of high-pressure carbon dioxide and high hydrostatic pressure on the 

quality of carrot juice. Food Engineering and Physical Properties, 67, 

1827–1834. 

Poretta, S. (2000). Analisi sensoriale & consumer science. Edited by 

Chirotti Editor, Pinerolo (TO), Italy. 

Qadir, A. & Hashinaga, F. (2001). Inhibition of postharvest decay of fruits 

by nitrous oxide. Postharvest Biology and Technology, 22, 279–283. 

Raso, J. & Barbosa-Canovas, G.V. (2003). Nonthermal preservation 

of foods using combined processing techniques. Critical Reviews in 

Food Science and Nutrition, 43, 265–285. 

Riu-Aumatell, M., Lopez-Tamames, E. & Buxaderas, S. (2005). 

Assessment of the volatile composition of juices of apricot, peach 

and pear according to two pectolytic treatments. Journal of 


Speck, M.L., Ray, B. & Read, R.B. (1975). Repair and enumeration of 

injured coliforms by a plating procedure. Applied Microbiology, 29, 

549–550. 

Spilimbergo, S. & Bertucco, A. (2003a). Non-thermal microbial 

inactivation with dense CO2. Biotechnologies and Bioengineering, 

84, 627–638. 

Spilimbergo, S. & Mantoan, D. (2006). Kinetic analysis of microorganisms 

inactivation in apple juice by high pressure carbon dioxide. 

International Journal of Food Engineering, 2, 6. 

Spilimbergo, S., Elvassore, N. & Bertucco, A. (2003b). Inactivation of 

microorganisms by supercritical CO2 in a semi-continuous process. 

Italian Journal of Food Science, 15, 115–124. 

Spilimbergo, S., Mantoan, D. & Dalser, A. (2007a). Supercritical gases 

pasteurization of apple juice. Journal of Supercritical Fluids, 40, 485– 

489. 

Spilimbergo, S., Mantoan, D. & Cavazza, A. (2007b). Yeast inactivation 

in fresh apple juice by high pressure nitrous oxide. International 

Journal of Food Engineering, 3, n2. 

Sumitani, H., Suekane, S., Nakatani, A. & Tatsuka, K. (1994). Changes 

in composition of volatile compounds in high pressure treated peach. 


Sun, D. (2003). An improved model calculating CO 2 solubility in pure 

water and aqueous NaCl solution from 273 to 533 K and from 0 to 

2000 bar. Chemical Geology, 193, 257–271. 

Tasselli, F., Cassano, A. & Drioli, E. (2007). Ultrafiltration of kiwifruit 

juice using modified poly(ether ether ketone) hollow fibre membranes. 

Separation and Purification Technology, 57, 94–102. 

Van Ginneken, L., Weyten, H., Willems, L. & Lodewijckx, B. (2002). 

In: Method and Apparatus for the Inactivation of Microorganisms in 

Liquid Food and Feed Products. Vol. (Ed. WO Patent 2004 ⁄ 039180 

A1) (May 13). 

Wei, C.I., Balaban, M.O., Fernando, S.Y. & Peplow, A.J. (1991). 

Bacterial effect of high-pressure CO2 treatment on foods spiked with 

Listeria or Salmonella. Journal of Food Protection, 4, 189–193. 

Young, P.B., Kennedy, S., Molloy, A.M., Scott, J.M., Weir, D.G. & 

Kennedy, D.G. (1997). Lipid peroxidation induced in vivo by 

hyperhomocysteinaemia in pigs. Atherosclerosis, 129, 67–71. 

Zhong, Q.X., Blanck, D.G., Davidson, P.M. & Golden, D.A. (2008). 

Nonthermal inactivation of Escherichia coli K-12 on spinach leaves, 

using dense phase carbon dioxide. Journal of Food Protection, 7, 

1015–1017. 


1626 


Determination of biochemical properties of foam-mat dried mango 

powder 

Dattatreya M. Kadam,* Robin A. Wilson & Sumandeep Kaur 

Central Institute of Post-Harvest Engineering and Technology (CIPHET), PO: PAU, Ludhiana-141004, Punjab, India 


Summary Investigations were carried out to see the impact of drying air temperature (65, 75 and 85 °C) and milk as 

foaming agent in different concentration levels (0%, 10%, 15%, 20% and 25%) on the chemical properties of 

foam-mat dried mango juice powder. Chemical properties such as total sugars, ascorbic acid, total carotenes, 

minerals, total acid, pH, total soluble solids (TSS) and microbial load (fungal and bacterial) of foam-mat 

dried mango powder were determined. Data were analysed as per two-way anova, Duncan’s multiple range 

test and l.s.d. of AgRes Software statistical package. Almost all chemical properties show decreasing trend 

with increase in drying air temperature. Microbial load was not detected in foam-mat dried mango powder. 

It was found that addition of 10% milk as foaming agent and drying at 65 °C temperature gave better 

results. 

Keywords Chemical properties, drying, foam-mat, foaming, microbial load, milk. 

Introduction 

Mango (Mangifera indica L.) is an important tropical 

fruit crop normally accredited with king of fruits. India 

is the largest producer of the mango, and it contributes 

37% of total 30.5 million tons of global production. 

Annually, India exports about 50 000 tons mangoes to 

different parts of the world including Middle East, 

Europe and United States (Pandit et al., 2009), and its 

demand is increasing all the time. Compared with 

several temperate fruits, the tropical and subtropical 

fruit such as mango presents greater problems in storage 

and transportation because of its perishable nature 

(Mitra & Baldwin, 1997). The production, marketing 

and consumption of mango fruits are restricted because 

of improper handling, inadequate transport and storage 

facility, disease problems and sensitivity to low storage 

temperature (Mitra & Baldwin, 1997). Products derived 

from mango fruit are increasingly used in beverage, 

dairy and confectionery industries, where fruit purees 

and concentrate are the major intermediates. Dehydrated 

powders can be reconstituted into juice and 

used as a starter for the preparation of products like 

beverages and baby foods. Thus, dehydrated mango 

juices have promising and potential scope in domestic 


e-mails: kadam1k@yahoo.com, kadam1k@gmail.com 


and international market (Kadam & Balasubramanian, 

2010). 

Drying is an important post-harvest technology 

among the agricultural processing technologies. The 

major objective in drying agricultural products is 

the reduction in moisture content, weight and volume, 

minimising packaging, storage and transportation costs 

(Okos et al., 1992; Singh, 1994 and Kadam et al., 2008). 

The traditional drying methods viz. sun drying require 

long drying time and final product may be contaminated 

from dusts, insects and microbes. Thus, an alternative 

method of drying such as foam mat drying is a beneficial 

process of mango juice preservation in the form of 

powder. 

Foaming of liquid and semi-liquid materials has long 

been recognised as one of the efficient methods to 

shorten drying time. Over the past decade, this 

relatively old technology, known as foam-mat drying, 

received renewed attention because of its added ability 

to process hard-to-dry materials to produce products of 

desired properties, retaining its volatiles that otherwise 

would be lost during the drying of non-foamed 

materials (Ratti & Kudra, 2006; Kadam & Balasubramanian, 

2010). Thus, current research is directed not 

only to convective drying of purposely foamed materials 

in spray dryers, plate dryers and band dryers but 

also to conventional freeze-drying, as well as microwave 

drying of frozen foams with and without dielectric 

inserts as complementary heat sources (Ratti & 

doi:10.1111/j.1365-2621.2010.02308.x 

Ó 2010 Institute of Food Science and Technology

Kudra, 2006). Foam-mat drying involves the incorporation 

of foaming agent into liquid foods with 

subsequent whipping to form stiff foam (Morgan et al., 

1961; Hart et al., 1963). In general, drying rate of 

foamed materials is faster than non-foamed materials 

and is greatly accelerated at the end. Many researchers 

have reported that the increased interfacial areas of 

foamed materials are the responsible factors for its 

induced drying rate. Certain foams such as the ones 

from soymilk (Akintoye & Oguntunde, 1991) or starfruit 

(Karim & Wai, 1999) exhibit higher drying rates in 

the beginning of foam-mat drying, whereas for other 

materials such as tomato paste (Lewicki, 1975), tomato 

juice (Kadam & Balasubramanian, 2010), bananas 

(Sankat & Castaigne, 2004) and mango (Cooke et al., 

1976), drying rates are greatly accelerated at the end of 

drying. 

Mango is prone to post-harvest losses. Thus, technology 

is needed to be developed that can reduce such 

losses during transportation and storage. Foam-mat 

drying is an alternative way to preserve the quality 

and also to add value to the mango. In spite of its 

importance, the information on the foam drying of 

mango has received less attention. The objectives of this 

study were to study the effect of foam-mat drying air 

temperature and foaming concentration on some chemical 

properties and quality of foam-mat dried mango 

juice powder. 


Foam-mat drying of mango juice experiment was 

carried out at Central Institute of Post-Harvest Engi- 

neering and Technology (CIPHET), Ludhiana, Punjab 

(India) using tray dryer to prepare a mango powder. 

Mango juice powder preparation 

Ripe mangoes (M. indica L.) were purchased from the 

local market of Ludhiana. Fruits of similar ripeness, 

having uniform visual quality and size were selected. 

Mangoes were subjected to washing with running water, 

peeling of skin and stone removal before juice ⁄ pulp 

preparation using mixer. Foaming was achieved adding 

liquid milk (0%, 10%, 15%, 20% and 25% concentrations) 

having solid non fat (SNF) – 8.5%, fat – 4.5%, 

minerals – 0.6%, carbohydrates – 4.6% and proteins – 

¼ 

Quality of foam-mat dried mango powder D. M. Kadam et al. 1627 

3.3%, pasteurised standardised milk, processed by Verka, 

Milkfed, Verka milk Plant, Ludhiana, in homogenised 

mango juice using hand blender for 3 min whipping to 

obtain consistent foam. Foamed juice was poured in 

food-grade stainless steel trays and spread to obtain 

3-mm-thick foam-mat and dried at 65, 75 and 85 °C air 

temperature in tray dryer. The dried product was scraped 

and pulverised before packing for further studies. 

Rehydration for sample preparation 

Foam-mat dried mango powder was reconstituted in 

distilled water in the ratio of 1.2: 10 (Mango Powder: 

Water) and was used for estimations of chemical 

properties such as total sugar, ascorbic acid, total 

carotene content, minerals, titratable acidity, pH and 

microbial load (fungi and bacteria). 

Chemical properties determination 

Total sugar 

Estimation of total carbohydrates was done using 

phenol–sulphuric acid method (Sadasivam & Manikam, 

1996). Reconstituted sample was appropriately diluted 

from which 0.1 mL was further diluted to 1 mL with 

distilled water. To this, 1 mL of 5% phenol was added 

followed by addition of 5 mL of H2SO4 after 10 min. 

Test tubes were incubated at 37 °C for 30 min, and 

absorbance was noted at 490 nm against reagent blank 

on UV visible spectrophotometer (LABINDIA UV- 

3000 + ,UV⁄ VIS Spectrophotometer, Thane, India). Percentage 

of total carbohydrates was worked out from the 

standard curve of glucose and calculated as: 

Sugar value from the graph ðlgÞ Total volume of extract ðmLÞ 

Total sugarð%Þ ¼ 

Aliquot of the sample used ðmLÞ Volume of sample 1000 

Ascorbic acid 

Ascorbic acid of foam-mat dried powder was estimated 

using 2, 6 dichlorophenol-indophenol visual titration 

method (Sadasivam & Manikam, 1996). Aliquot of the 

reconstituted sample was diluted to a fixed volume with 

3% HPO 3 and then titrated with 2, 6 dichlorophenolindophenol. 

Standard ascorbic acid solution of 5 mL 

was added to 5 mL of HPO3 and titrated with dye 

solution to a pink colour, which persisted for 15 s. The 

dye factor was determined, i.e. mg of ascorbic acid 

per mL of the dye, using the formula: Dye factor 

= 0.5 per titre. 

Ascorbic acid (mg per 100 ml) of reconstituted juice 

was calculated using the formula: 

Titre Dye factor Volume made up 100 

Aliquot of extract taken for estimation Weight or volume of sample taken for estimation 


1628 

Quality of foam-mat dried mango powder D. M. Kadam et al. 

Total carotene content 

Carotene was estimated following and modifying the 

method of Rangana (1995). Dry mango powder 

(500 mg) was homogenised and agitated with 5 mL of 

acetone ⁄ hexane (4:6), allowed to settle down and then 

supernatant was decanted. The procedure was repeated 

until no colour was obtained, and supernatants were 

pooled and final volume was made up to 10 mL. The 

absorption of supernatant was recorded at 490 nm on 

UV visible spectrophotometer. Carotene content was 

calculated as follows: 

Carotene ðmg per 100 gÞ ¼ 

pH and titratable acidity 

Digital pH meter was used to measure the pH of the 

samples, and titratable acidity of reconstituted sample 

was estimated diluting the aliquot of the sample with 

water to a fixed volume and then titrated with 0.1 N 

NaOH using phenolphthalein as an indicator. Percentage 

acidity was calculated as the percentage of anhydrous 

citric acid using the formula: 

Minerals 

Ash solution 

Food samples (5 g) in triplicate were taken in crucibles. 

These were burnt on hot plate and were placed in an 

electric muffle furnace at 600 °C for 6 h. After cooling 

the crucibles to room temperature in desiccator, the ash 

in the crucibles is dissolved in 25 mL of dilute hydrochloric 

acid (1:4) on a water bath. The contents are 

transferred to 250-mL volumetric flask and the volume 

is made up. The above solution was filtered and used as 

ash solution for the estimation of calcium, iron and 

phosphorous. 

Calcium (AOAC, 1980) 

To 50 mL of ash solution, 10 mL of saturated ammonium 

oxalate solution was added. Two drops of methyl 

red indicator was added after boiling, and dilute ammonium 

hydroxide (1:1) was added to neutralise. Crystalline 

precipitates were obtained after boiling followed by 

the addition of few drops of dilute HCl (1:4) until the 

colour was faint pink. After overnight standing, precipitates 

were filtered through Whatman filter paper no. 42 

and washed thoroughly with hot distilled water till the 

precipitates are free of oxalates (tested with silver nitrate 

solution). The precipitates along with the filter paper are 

added in the original beaker dissolved in 20 mL of 10% 

H 2SO 4. The contents are heated to about 70 °C and 

titrated against 0.1 N KMnO4 to faint pink colour. One 

millilitre of KMnO4 used is equivalent to the product of 

0.002 g of calcium and dilution factor. 

Iron (AOAC, 1980) 

Ten millilitres of ash solution was pipetted out into the 

25-mL volumetric flask, and 1 mL of 10% hydroxylamine 

hydrochloride solution was added. After a few 

minutes, 5 mL of acetate buffer solution and 1 mL of 

concentration of carotene in solution as read from standard curve (mg/mL) Final volume Optical density 100 

Weight of sample taken ðgÞ 1000 

0.1% ortho-phenanthroline solution were added. The 

contents are mixed, and the volume is made up with 

distilled water up to the mark. The intensity of the 

colour developed is measured at 452 nm. In the place of 

sample, 10 mL of HCl (2 mL per 100 mL) was taken 

and treated as blank. The concentration of iron in the 

known sample is calculated from the standard curve and 

is multiplied by the dilution factor. 

Titre Normality of alkali Volume made up Equivalent weight of acid 100 

Total acid ð%Þ ¼ 

Volume of sample taken for estimation Weight or volume of sample taken 1000 

Phosphorus (Fisher, 1971) 

Ash solution of 0.2 mL was diluted to 1 mL with 

distilled water. To this, 2 mL of ammonium molybdate 

solution (25 g ammonium molybdate per 300 mL + 

75 mL concentrated H 2SO 4 per 200 mL), 1 mL of 

(0.5%) hydroquinone and 1 mL of 20% sodium sulphate 

solution are added and mixed, followed by the 

addition of 5 mL of distilled water. Absorbance of the 

solution was observed at 660 nm against a reagent 

blank. The concentration of phosphorous in the known 

sample is calculated from the standard curve and is 

multiplied by the dilution factor. 

Total soluble solids 

TSS of the rehydrated mango juice sample (1.2:10 i.e. 

mango powder ⁄ water) was measured with the help of 

hand refractrometer. 

Microbial load (fungal and bacterial) 

Microbial load analyses of reconstituted samples were 

done for fungal and bacterial load. For fungal and 

bacterial load, Mortin Rose Bengal agar [peptone (5 g), 

glucose (10.0 g), KH2PO4 (1 g), MgSO4 7H2O (0.5 g), 

Rose Bengal (0.035 g) and agar (18.0 g) were dissolved 


in 1000 mL of distilled water] and standard plate count 

agar [peptone ⁄ trypton (5 g), yeast extract (2.5 g), beef 

extract (2.0 g), glucose (10.0 g) and agar (18.0 g) were 

dissolved in 1000 mL of water] media were used, 

respectively. Water blanks were prepared by adding 

1.0 g of sample to 10 mL of autoclaved water. For dry 

powder and fresh juice samples, the dilution was made 

up to 10 )1 and 10 )2 , respectively, for both enumerations 

of fungi and bacteria. From different dilutions made 

from different samples, 1 mL was poured into each petri 

dish followed by the addition of 20–25 mL of media. 

The petri dish was circumvolved for proper mixing. The 

plates were allowed to solidify and then kept in 

incubator at 37 and 30 °C for bacteria and fungi, 

respectively. Colonies were counted after 72 and 24 h 

for fungi and bacteria, respectively. 


Chemical properties and microbial load evaluation were 

carried out to check the effect of treatments and safety 

of food quality in terms of fungal and bacterial loads. 

Data were analysed as per two-way anova, Duncan’s 

multiple range test (DMRT) and l.s.d. of AgRes 

Software statistical package for microbial load and 

chemical properties, respectively, to get the best treatment. 


Results and discussions 

Total sugar 

Fresh mangoes have sugar concentration of about 

40–41%. In this study, the grand mean sugar content 

was 31.35%. Higher temperature treatments resulted in 

decrease in sugar concentrations, which could be 

attributed to Maillard reactions. Temperature treatments 

do have some negative effect on the sugar 

concentrations (Table 1). Sugar content of foam-mat 

dried mango powder at 65 °C air temperature was the 

best, followed by 85 and 75 °C air temperature. Sugar 

content was higher at 85 °C when compared to 75 °C, 

because although the temperature was high, the duration 

of drying was short at 85 °C. In case of foam-mat 

dried mango at 65 °C air temperature treatment, the 

time duration was longer than those of the two highertemperature 

treatments but was compensated by low 

temperature. Generally, there was a decreasing trend of 

sugars with increasing concentrations of milk. Among 

concentrations, 5% milk was the best. Decrease in 

sugar content with the increasing concentration of 

milk could be owed to the increase in protein content, 

which in turn was because of added milk that leads 

to higher chance of Maillard reaction. Pooled l.s.d. 

results showed that 15% milk at 65 °C was the best 

Table 1 Effect of drying air temperatures and milk as foaming agent on some chemical properties of foam-mat dried mango powder 

Temperature (°C) Milk (%) Sugar (%) Ascorbic acid (mg per 100 g) Total carotenes (mg per 100 g)* Total acid (%) pH 

65 0 41.354 a 

5.600 a 

12.055 0.079 d 

6.687 a 

10 40.080 b 

4.800 b 

10.547 0.091 c 

6.327 b 

15 39.285 c 

4.800 b 

9.719 0.091 c 

6.190 d 

20 37.344 e 

4.800 b 

9.166 0.091 c 

6.077 f 

25 35.493 g 

4.800 b 

7.236 0.091 c 

5.853 j 

75 0 41.328 h 

4.053 c 

3.700 0.117 b 

5.403 k 

10 31.504 j 

3.200 d 

3.539 0.125 a 

5.337 l 

15 28.346 m 

3.200 d 

3.294 0.125 a 

5.230 m 

20 26.133 n 

3.200 d 

3.047 0.125 a 

5.220 n 

25 22.912 o 

3.200 d 

2.880 0.125 a 

5.203 o 

85 0 41.397 d 

3.200 d 

6.489 0.065 e 

6.227 c 

10 35.562 f 

3.200 d 

3.225 0.065 e 

6.170 e 

15 35.146 i 

3.200 d 

2.936 0.090 c 

6.050 g 

20 29.168 k 

3.200 d 

2.641 0.090 c 

6.007 h 

25 27.882 l 

3.200 d 

2.452 0.090 c 

5.943 i 

SED (T) 0.011 0.004 1.595 0.001 0.002 

SED (C) 0.014 0.006 2.060 0.002 0.003 

SED (T · C) 0.025 0.009 3.568 0.003 0.004 

CD (0.05) (T) 0.023 0.009 3.259 0.003 0.004 

CD (0.05) (C) 0.029 0.011 4.207 0.004 0.005 

CD (0.05) (T · C) 0.051 0.019 7.287 0.007 0.009 

Superscripts a, b, c, d ….. indicate the best performing treatment followed by poor performing treatments. Mean values with the same superscript 

letters are not significantly different. 

*Indicates non-significant effect of milk concentrations on carotenes. 


1630 


treatment. Marwaha & Pandey (2006) reported similar 

decline in sugar content during the preparation of 

dehydrated chips, and Leite et al. (2007) in banana 

drying, and Bondaruk et al. (2007) in dried potato 

cubes. 

Ascorbic acid 

Ascorbic acid content of fresh mangoes ranged 

between 23 and 25 mg per 100 g fresh weight of 

sample. In this study, the average ascorbic acid content 

was about 3.844 mg per 100 mL of reconstituted sample 

that corresponds to about 32.03 mg per 100 g of 

dried powder. The temperature treatments resulted in 

decrease in ascorbic acid content of the mango powder 

(Table 1). Degradation was highest in case of 85 °C 

treatment, which becomes less pronounced at 75 °C 

and least at 65 °C. Mean comparison by l.s.d. showed 

that the 65 °C temperature was the best among the 

three temperatures. Additions of milk at different concentrations 

as foaming agents did not have any marked 

effect on the ascorbic acid content of the dried mango 

powder. Mean comparison by l.s.d. showed that milk 

at the rate of 10–25% has equal effect on the ascorbic 

acid content of the powder. Even the control sample 

showed higher ascorbic acid content when compared to 

the milk-added samples. Pooled l.s.d. results showed 

that 10% milk at 65 °C was the best treatment among 

milk concentrations as well as at different temperatures. 

The decrease in ascorbic acid content of foammat 

dried mango powder could be because of its 

heat-sensitive nature. Similar decline in ascorbic acid 

content was noticed in other studies with potato 

(Marwaha & Pandey, 2006), pulses (Mehta et al., 

2007), muskmelon (Fernandez et al., 2007), cauliflower 

(Kadam et al., 2005) and onion (Kadam et al., 2009) 

following heat treatment. 

Total carotenes 

The overall mean of the total carotene content was 

6.177 mg per 100 g of the dry mango powder. With the 

increase in temperature, the carotene content decreased 

with highest content at 65 °C followed by 75 °C and 

then 85 °C (Table 1). This was supported by the l.s.d. 

analysis, which showed that 65 °C treatment is the best 

among the three temperature treatments. With the 

addition of milk, there was a decreasing trend virtually, 

but statistically, addition of milk at all concentrations 

did not have any significant change in the carotene 

content at P < 0.05 level of significance. Such decline in 

carotene content was observed by Chen et al. (2007) 

in Taiwanese mango, Hymavathi & Khader (2005) in 

dehydrated ripe mango powder, Wen-pingI et al. (2008) 

in fruits of Lycjum barbarum, and Lavelli et al. (2007) in 

dehydrated carrots. 

Titratable acidity 

Mean total acid (%) in the mango powder of this study 

was about 0.097%. There was no apparent trend with 

the temperature treatments affecting the titratable acidity 

of the powder (Table 1). However, statistically, 

75 °C was the one with the highest total acid (%) and 

was the best, followed by 65 and 85 °C drying treatment. 

Addition of milk at different concentrations 

resulted in a minor increase in the total acid (%) only 

in the initial one or two concentrations. Further rise in 

concentrations did not add to any total acid (%) rise. 

Various researchers with peanuts (Tsai et al., 2007), 

apples (Pereira et al., 2008) and tomatoes (Radwan & 

Lobna, 2002) reported similar results of total acid (%) 

decline. 

pH 

The overall mean pH content of the foam-mat dried 

mango powder was 5.862. There was a decreasing trend 

in the pH of the dried mango powder with the addition 

of foaming agents (Table 1). Effect of different foaming 

agents as well as different concentration levels was 

highly significant (P < 0.01) at all temperatures. Pooled 

l.s.d. analysis showed that among different milk concentrations, 

10% milk at 65 °C showed the highest pH 

content, whereas 25% milk at 75 °C showed the lowest 

pH content. 

Minerals 

Concentration of calcium in fresh mango is about 

10 mg per 100 g fresh weight. The mean calcium content 

of the mango powder was 296.89 mg per 100 g dry 

powder. The calcium content at 65 °C was the highest 

among the three temperature treatments (Table 2). 

Mean comparison by l.s.d. showed that the temperature 

treatments differ significantly at P < 0.01 level of 

significance. Calcium concentration enhanced linearly 

with the addition of milk at all concentrations of milk. 

This could be attributed to the calcium contribution by 

milk, which has calcium concentration of about 

113 mg per 100 g. This is further supported by the 

statistical analysis showing that treatment with 25% 

milk had the highest concentration of calcium. Pooled 

analysis by l.s.d. showed that 25% milk at 65 °C was the 

best treatment regarding calcium concentration. 

Iron concentration in fresh mangoes is about 

0.13 mg per 100 g fresh weight. The overall average 

iron content in foam-mat dried mango powder was 

97.05 mg per 100 g dry powder. All treatments with 

different temperatures and different concentrations of 

milk have significant (P < 0.01) effect on the iron 

concentration of the dry powder (Table 2). The iron 

concentration usually declined with the increasing 


Table 2 Effect of drying air temperatures and milk as foaming agent on mineral properties and TSS of foam-mat dried mango powder 

Temperature (°C) Milk (%) Calcium (mg per 100 g) Iron (mg per 100 g) Phosphorous (mg per 100 g) TSS (°Brix) 

65 0 226.667 j 

128.204 a 

88.912 j 

9.100 a 

10 286.667 g 

122.653 b 

102.482 i 

9.000 b 

15 340.000 e 

115.536 c 

121.480 g 

8.800 d 

20 406.667 c 

111.123 d 

138.443 e 

8.600 f 

25 486.667 a 

106.853 f 

154.389 d 

8.400 h 

75 0 140.000 m 

109.842 e 

115.035 h 

9.000 b 

10 180.000 l 

83.509 k 

137.426 e 

9.000 b 

15 210.000 k 

80.378 l 

153.371 d 

8.867 c 

20 246.667 i 

79.523 l 

177.797 b 

8.733 e 

25 283.333 g 

78.954 m 

188.654 a 

8.533 g 

85 0 243.333 i 

98.455 g 

81.787 k 

9.000 b 

10 270.000 h 

91.053 h 

104.178 i 

9.000 b 

15 300.000 f 

86.925 i 

126.909 f 

8.800 d 

20 386.667 d 

85.359 j 

137.086 e 

8.600 f 

25 446.667 b 

77.388 n 

163.548 c 

8.400 h 

SED (T) 2.309 0.195 0.731 0.009 

SED (C) 2.981 0.253 0.943 0.012 

SED (T · C) 5.165 0.438 1.634 0.021 

CD (0.05) (T) 4.716 0.400 1.429 0.019 

CD (0.05) (C) 6.089 0.516 1.927 0.025 

CD (0.05) (T · C) 10.546 0.894 3.337 0.043 

concentrations of milk. Pooled analysis showed that 

foam-mat dried mango at 65 °C air temperature with 

10% milk as foaming agent was the best. 

Fresh mangoes have about 11 mg per 100 g (fresh 

weight) of phosphorous. In this study, the overall 

phosphorous content of the foam-mat dried mango 

powder was about 132.767 mg per 100 g dry powder. 

Phosphorous concentration augmented with the addition 

of milk as foaming agent that went on increasing as 

we move from lower milk concentration to the highest 

(Table 2). Temperatures do have a significant (P < 

0.01) effect on the concentration of phosphorous with 

65 °C being the best among the three temperatures. 

Pooled mean comparison by l.s.d. showed that foamed 

mango juice dried at 65 °C with 25% milk as foaming 

agent was the best. 

Total soluble solids 

The mean TSS of the foam-mat dried mango powder 

was 8.79°Brix ranging from 8.4 to 9.1°Brix (Table 2). 

The TSS has a decreasing trend with the increasing 

concentrations of the milk. All treatments with different 

temperatures and different concentrations of milk have 

significant (P < 0.01) effect on the TSS of the dry 

powder. Mean comparison by l.s.d. showed that foamed 

mango juice dried at 65 °C with 10% milk as foaming 

agent was the best with highest TSS among different 

temperatures and milk concentrations. 

Microbial load 

Microbial load for fungi and bacteria was determined for 

reconstituted juice of mango powder. There was no 

fungal and bacterial growth detected in the freshly 

prepared mango powder, whereas the fresh mango pulp 

showed a huge microbial load for both fungi and 

bacteria with about 3.5 · 10 4 CFU and 2.7 · 10 4 CFU, 

respectively. This clearly indicates that the drying process 

of mango juice to obtain powder greatly reduced the 

microbial load. Similar results were reported in solardried 

cauliflower (Kadam et al., 2005) and onion 

(Kadam et al., 2009). 

Conclusion 


TSS, total soluble solids. 

Superscripts a, b, c, d ….. indicate the best performing treatment followed by poor performing treatments. Mean values with the same superscript 

letters are not significantly different. 

The most serious constraint for shelf life enhancement 

is the activity of microorganisms. Drying and dehydration 

of fresh fruit juice is one of the most 

energy-intensive processes in the food industry and a 

promising method in reducing post-harvest losses. 

Foam mat drying is a profitable process for drying of 

mango in powdered form. From this study, it can be 

concluded that the powder obtained as a result of the 

addition of 10% milk as foaming agent and drying at 

65 °C air temperature was the best. Mango powder 

developed by foam-mat drying has a better nutritional 

quality, increased food safety and enhanced shelf life. 

The sugar content of the mango powder is high enough 


1632 


so that there is no need to add extra sugar to the 

powder preparations. High contents of ascorbic acid, 

total carotenes and mineral are a boon to human 

health. It can be used for the preparation of a variety 

of products viz. beverages as a juice concentrate, and 

in ice creams, jellies, cakes and candies as flavouring 

and colouring agents. 


The authors express sincere thanks to Department of 

Science and Technology (DST), New Delhi, for funding 

the subproject under cluster project on food processing. 

References 

AOAC (1980). Official Methods of Analysis, 11th edn. Washington, 

DC: Association of official analytical chemists. 

Akintoye, O.A. & Oguntunde, O.A. (1991). Preliminary investigation 

on the effect of foam stabilizers on the physical characteristics and 

reconstitution properties of foam-mat dried soymilk. Drying Technology, 

9, 245–262. 

Bondaruk, J., Markowski, M. & Baszczak, W. (2007). Effect of drying 

conditions on the quality of vacuum-microwave dried potato cubes. 


Chen, J.P., Tai, C.Y. & Chen, B.H. (2007). Effects of different drying 

treatments on the stability of carotenoids in Taiwanese mango 

(Mangifera indica L.). Food Chemistry, 100, 1005–1010. 

Cooke, R.D., Breag, G.R., Ferber, C.E.M., Best, P.R. & Jones, J. 

(1976). Studies of mango processing. 1. The foam-mat drying of 

mango (Alphonso cultivar) puree. Journal of Food Technology, 11, 

463–473. 

Fernandez, R., Norena, C.P.Z., Silveira, S.T. & Brandelli, A. (2007). 

Osmotic dehydration of muskmelon (Cucumis melo): influence of 

blanching and syrup concentration. Journal of Food Processing and 

Preservation, 31, 392–405. 

Fisher, I. (1971). Modern Food Analysis. p. 22. Washington, DC: 

Freeman & Co. 

Hart, M.R., Graham, R.P., Ginnette, L.F. & Morgan, A.I. (1963). 

Foams for foam mat drying. Food Technology, 17, 1302–1304. 

Hymavathi, T.V. & Khader, V. (2005). Carotene, ascorbic acid and 

sugar content of vacuum dehydrated ripe mango powders stored in 

flexible packaging material. Journal of Food Composition and 

Analysis, 18, 181–192. 

Kadam, D.M. & Balasubramanian, S. (2010). Foam mat drying of 

tomato juice. Journal of Food Processing and Preservations, 

(accepted). 

Kadam, D.M., Lata, D.V.K.S. & Pandey, A.K. (2005). Influence of 

different treatments on dehydrated cauliflowwer quality. International 


Kadam, D.M., Samuel, D.V.K., Chandra, P. & Sikarwar, H.S. (2008). 

Impact of processing treatments and packaging material on some 

properties of stored dehydrated cauliflower. International Journal of 


Kadam, D.M., Nangare, D.M. & Oberoi, H.S. (2009). Influence of 

pre- treatment on microbial load of stored dehydrated onion slices. 

International Journal of Food Science and Technology, 44, 1902– 

1908. 

Karim, A.A. & Wai, C.C. (1999). Foam-mat drying of starfruit 

(Averrhoa carambola L.) puree. Stability and air drying characteristics. 


Lavelli, V., Zanoni, B. & Zaniboni, A. (2007). Effect of water activity 

on carotenoid degradation in dehydrated carrots. Food Chemistry, 

104, 1705–1711. 

Leite, J.B., Mancini, M.C. & Borges, S.V. (2007). Effect of drying 

temperature on the quality of dried bananas cv. prata and d’a¢gua. 

Lebensmittel-Wissenschaft und- Technologie, 11, 38–41. 

Lewicki, P.P. (1975). Mechanisms concerned in foam-mat drying of 

tomato paste. Transactions of Agricultural Academy in Warsaw, 55, 

1–67 (in Polish). 

Marwaha, R.S. & Pandey, S.K. (2006). Suitability of cultivars and 

methods for the production of dehydrated chips. Potato Journal, 33, 

110–117. 

Mehta, M.B., Mehta, B., Bapodra, A.H. & Joshi, H.D. (2007). Effect 

of germination and heat processing on protein, riboflavin, vit-C and 

niacin content in peas, cowpea, red gram and wheat. Asian Journal 

of Home Science, 2, 34–38. 

Mitra, S.K. & Baldwin, E.Z. (1997). Mango. In: Postharvest Physiology 

and Storage of Tropical and Subtropical Fruits (edited by S.K. 

Mitra). Pp. 85–122. West Bengal, India: CAB International. 

Morgan, A.I., Graham, R.P., Ginnette, L.F. & Williams, G.S. (1961). 

Recent developments in foam mat drying. Food Technology, 15, 

37–39. 

Okos, M.R., Narsimhan, G., Singh, R.K. & Witnauer, A.C. (1992). 

Food dehydration. In: Handbook of food engineering (edited by 

D.R. Heldman & D.B. Lund). Pp. 75–79. New York: Marcel 

Dekker. 

Pandit, S.S., Chidley, H.G., Kulkarni, R.S., Pujari, K.H., Giri, A.P. & 

Gupta, V.S. (2009). Cultivar relationships in mango based on fruit 

volatile profiles. Food Chemistry, 114, 363–372. 

Pereira, E.A., Machado, M.T., Borges, S.V., Maia, M.C.A. & 

Augusta, I.M. (2008). Influence of osmotic agent on the physiochemical 

composition and drying kinetics of Gala apples. Revista 

Ciencia Agronomica, 39, 173–176. 

Radwan, H.M. & Lobna, A.M.H. (2002). Production of tomato 

powder and dried slices for use in some foodstuffs. Egyptian Journal 

of Agricultural Research, 80, 1751–1762. 

Rangana, S. (1995). Hand Book of Analysis and Quality Control for 

Fruits and Vegetable Products, 2nd edn. P. 1112. New Delhi: Tata 

McGraw-Hill publication Co. Ltd. 

Ratti, C. & Kudra, T. (2006). Foam-mat drying: Energy and cost 

analyses. Canadian Biosystem Engineering, 48, 27–32. 

Sadasivam, S. & Manikam, A. (1996). Biochemical Methods, 2nd edn. 

New Delhi and TNAU, Coimbatore: New Age International (P) 

Ltd. Publisher. 

Sankat, C.K. & Castaigne, F.F. (2004). Foaming and drying behaviour 

of ripe bananas. Lebensmittel Wissenshaft und Technolgie, 37, 517– 

525. 

Singh, K.K. (1994). Development of a small capacity dryer for 

vegetable. Journal of Food Engineering, 21, 19–30. 

Tsai, S.J., Wu, T.Y., Yang, K., Liu, H., Liaw, H. & Liang, C. (2007). 

Effects of cooking methods and rehydration on the chemical 

composition and texture of peanut. Journal of Taiwan Agricultural 

Research, 6, 189–205. 

Wen-ping, M., Zhi-jing, N., He, L. & Min, C. (2008). Changes of the 

Main Carotenoid Pigment Contents During the Drying Processes of 

the Different Harvest Stage Fruits of Lycium barbarum L. Agricultural 

Sciences in China, 7, 363–369. 




On the use of ultrafiltration for the concentration and desalting of 

phosvitin from egg yolk protein concentrate 

Bertrand P. Chay Pak Ting, 1 Yves Pouliot, 1 * Lekh R. Juneja, 2 Tutomu Okubo, 2 Sylvie F. Gauthier 1 & Yoshinori 

Mine 3 

1 Institute of Nutraceuticals and Functional Foods (INAF), Pavillon Paul-Comtois, Université Laval, Québec, QC, Canada G1V 0A6 

2 Research Laboratories, Taiyo Kagaku, Co Ltd., 1-3 Takaramachi, Yokkaichi, Mie 510 0844, Japan 

3 Department of Food Science, University of Guelph, Guelph, ON, Canada N1G 2W1 

(Received 3 March 2010; Accepted in revised form 6 May 2010) 

Summary An ultrafiltration-based approach was integrated in the preparation of phosvitin (PVs) from delipidated egg 

yolk proteins. An attempt was made to concentrate PVs as well as to desalt by means of the diafiltration 

technique. Primary experiments were devoted to optimise the ultrafiltration performance as function of 

parameters such as the effects of pH, feed concentration and transmembrane pressure on permeate flux with 

the 10-kDa molecular weight cut-off (MWCO) polyethersulfone membrane at laboratory scale. Higher 

permeate flux values were observed at low concentration and at alkaline pH, whatever transmembrane 

pressure studied. Then, desalting of PVs was carried out at 50 °C with 10- and 30-kDa MWCO membranes. 

The results showed that desalting of PVs was obtained with both the 10- and 30-kDa MWCO membranes 

and with a few loss of protein in the permeate side. 

Keywords Concentration, delipidated egg yolk proteins, desalting, phosvitin, ultrafiltration. 

Introduction 

Hen egg yolk phosvitin (PVs) is a phosphoprotein found 

in egg yolk, with a molecular mass of 35 kDa, containing 

10% of phosphorus and has a specific amino acid 

composition which comprises 123 serine residues. The 

phosphorus in PVs is found as phosphoserines (Mecham 

& Olcott, 1949) arranged in clusters that can carry up 

fifteen consecutive residues (Byrne et al., 1984). Abundance 

of phosphate groups provides PVs with a strong 

binding ability for metal ions (Fe, Ca and Mg), and they 

are acting as antioxidants (Lu & Baker, 1986). Moreover, 

the antioxidant activity of PVs can be improved by 

conjugating with galactomannan (Nakamura et al., 

1998). Egg yolk PVs and PVs–galactomannan can 

inhibit the growth of pathogenic Escherichia coli under 

thermal stress, which results in the synergic effect of the 

strong metal-chelating ability and the surface properties 

(Sattar Khan et al., 2000; Choi et al., 2004). Low 

molecular weight (1–3 kDa) phosphopeptides obtained 

by PVs tryptic hydrolysis show calcium-binding capacity 

and inhibit the formation of insoluble calcium phosphate 

(Jiang & Mine, 2000, 2001). PVs peptides also 

*Correspondent: Fax: 418 656 3353; 

e-mail: Yves.Pouliot@inaf.ulaval.ca 

doi:10.1111/j.1365-2621.2010.02311.x 


showed antioxidant activity against oxidative stress in 

human intestinal epithelial cells (Katayama et al., 2006) 

as well as against oxidation of linoleic acid and radicalscavenging 

activity (Xu et al., 2007). 

Egg yolk is a complex mixture of particles, mainly 

composed of granules, suspended in a protein solution 

called plasma. The granules can be separated from egg 

yolk after dilution into a saline solution followed by 

centrifugation and consist mainly of a and b-lipovitellins 

[high-density lipoproteins (HDL)], PVs and LDLg. The 

apoproteins of HDL consist of five major polypeptides 

with molecular weight ranging from 32 to 105 kDa, the 

latter being the main one. Egg yolk plasma also contains 

low-density lipoproteins (LDL) but also livetins (Burley 

& Cook, 1961). Different studies reported that apoproteins 

of LDL consisted in up to eighteen polypeptides of 

molecular weight ranging mostly between 15 and 

180 kDa and sometimes higher molecular weights up 

to 225–240 kDa (Mine, 1998; Anton et al., 2003). 

Water-soluble livetins consist of a, b, c-livetins with 

molecular weight of 80, 45 and 150 kDa, respectively 

(Martin et al., 1957). 

At laboratory scale, PVs has been isolated by diluting 

egg yolk with a magnesium sulphate solution to form an 

insoluble complex (Mecham & Olcott, 1949), and 

several chromatographic methods have been carried

1634 

Ultrafiltration of crude phosvitin B. P. Chay Pak Ting et al. 

out for the separation of egg yolk PVs. After defatting, 

PVs has been separated into several components with 

molecular weight from 27 000 to 38 000 by using gelpermeation 

chromatography (Tsutsui & Obara, 1984) 

and into two fractions by ion-exchange chromatography 

(Connelly & Taborsky, 1961). Abe et al. (1982) fractionated 

PVs into a and b-PVs, having different phosphorus 

content and molecular weights, by gel filtration. 

Castellani et al. (2003) have developed a new purification 

method based on the insolubility of Mg 2+ ⁄ PVs salt 

following by an ion-exchange chromatographic fractionation. 

Although chromatographic methods are suitable 

for protein purification, these methods involve the 

use of chemicals and organic solvents and strong 

chemical conditions of pH that can potentially alter 

PVs structure (Taborsky, 1968). 

Defatting egg yolk granules is considered as a 

necessary step in the purification of PVs. Losso & 

Nakai (1994) developed a simple procedure for the 

isolation of PVs from chicken egg yolk. After removal of 

water-soluble proteins, the precipitate (lipids, granules) 

was extracted with organic solvents, such as hexane ⁄ ethanol 

(3:1), to liberate proteins bound to lipids and then 

filtered. PVs was isolated from the delipidated granules 

by means of salting-out, the precipitated egg yolk 

material is discarded and the PVs solution is then 

dialysed. 

Jiang & Mine (2000) have developed laboratory-scale 

experimental conditions to separate PVs from delipidated 

egg yolk proteins by salting-out using 10% NaCl 

followed by a 5000 g · 30 min centrifugation. The 

supernatant so-called crude PVs was desalted by means 

of extensive dialysis for 3 days at 4 °C. 

Despite these advances on PVs purification, there is 

still a need for a fractionation approach that would be 

potentially cost-effective at industrial scale. 

Pressure-driven membrane processes such as ultrafiltration 

(UF) can be carried out simultaneously for 

fractionation, concentration, purification and desalting 

food proteins and are already used at industrial scale. 

However, to our knowledge, only a few studies have 

reported the using of UF membrane processes for egg 

yolk proteins. Akita & Nakai (1992) have used a 

100-kDa molecular weight cut-off (MWCO) to extract 

immunoglobulins from a yolk water-soluble fraction 

upon the separation of crude immunoglobulin by 

ammonium sulphate precipitation. A maximal recovery 

of immunoglobulin (>98%) was obtained, and this 

value was comparable to the degree of purity obtained 

by gel filtration. In another study, Kim & Nakai (1996) 

developed a serial filtration approach involving dilutions, 

paper filtration and delipidation using hydrophobic 

filters. The delipidated water-soluble fraction was 

thereafter purified using 100-kDa UF membranes, and 

both high recoveries (72–89%) and purity (74–99%) 

were obtained. 

In this study, the procedure of Jiang & Mine (2000) 

was brought to a pilot-scale by using a centrifugal 

clarifier to remove the precipitated material from 10% 

NaCl-treated egg yolk proteins, and a UF–diafiltration 

(DF) procedure was designed to concentrate and replace 

the laboratory-scale dialysis procedure. In addition, a 

commercial delipidated egg yolk proteins’ concentrate 

was used as starting material to avoid the use of solvents 

and to focus on the potential of using membrane 

separation processes to concentrate PVs. 

The objective of our study was thus to evaluate the 

performance of UF membranes process followed by DF 

as an alternative approach for the production of 

desalted PVs. Considering that PVs molecular weight 

is 35 kDa, two different UF membranes, namely 10- and 

30-kDa MWCO polyethersulfone elements, were 

selected. The effect of process parameters such as pH, 

transmembrane pressure and feed concentration on 

permeate flux was investigated. Permeation flux values, 

selectivity and yield of PVs from the desalting procedure 

were determined. 


Materials 

Defatted egg yolk proteins were a gift from Taiyo 

Kagaku Co., Ltd. (Yokkaichi, Japan). The powdered 

product contained 79.20% protein, 9.86% fat and 

5.60% water according to manufacturer’s certificate of 

analysis. PVs standards for gel electrophoresis analyses 

were purchased from Sigma Chemical Co. (# catalogue: 

P1253-50MG, St Louis, MO, USA). 

Preparation of crude PVs from egg yolk proteins 

Extraction of crude PVs was performed according to 

Jiang & Mine (2000). Egg yolk proteins were suspended 

at 10% (w ⁄ v) in 10% NaCl (20 L) and stirred overnight 

at 4 °C. The precipitate was separated by a pilot plant 

centrifuge CEPA, type TZ 5 (Carl Padberg Zentrifugenbau 

GMBH, Lahr, Germany) at 25 °C and a feed 

flow rate of 735–1450 L min )1 . The sludge was discarded, 

and the supernatant was freeze-dried and stored 

at )35 °C. 

Ultrafiltration experiments 

Membranes 

UF was carried out using a polyethersulfone UF 

membrane Prep ⁄ scaleÔ-TFF 1 ft 2 cartridge (Millipore 

Corp., Bedford, MA, USA) having a MWCO of 10 and 

30 kDa with the effective area of 0.09 m 2 . The pure 

water permeation flux (Jw) was determined volumetrically 

at 25 °C at transmembrane pressure (PT) ranging 

from 0.35 to 1.75 bar. 


Jw were determined using the following equation: 

Jw ¼ 1 DV 

Dt 

ð1Þ 

Am 

where Am is the effective membrane area (m 2 ) and 

(DV ⁄Dt) is the permeate volume DV collected over time 

Dt (L h )1 ). 

The membrane retention coefficient was calculated 

according to the following relation: 

Rð%Þ ¼1 Cp 

Cr 

ð2Þ 

in which Cp (g L )1 ) and Cr (g L )1 ) are the measured 

component concentrations of phosphorus in the permeate 

and in the retentate, respectively. 

Permeation fluxes of crude phosvitin solutions 

The study of permeation flux changes as a function of 

pH was performed with the 10-kDa MWCO UF 

membrane by preparing a feed solution to a specific 

concentration (1.0%, w ⁄ v) and operating pressure of 1.4 

bar in recirculation mode. Then, the desired pH was 

decreased or increased stepwise using HCl 0.1 N or 

NaOH 0.1 N, respectively. 

Feed phases were prepared by rehydrating crude PVs 

at 0.5%, 1.0%, 5.0% and 10.0% (w ⁄ v) and adjusted to 

pH 8.0 using NaOH 0.1 N. The solutions were ultrafiltered 

with the 10-kDa MWCO UF membrane, and 

permeate fluxes of crude PVs solution (Jp) was determined 

at PT between 0.35 and 1.75 bar. Permeate flux 

values were calculated using eqn 1. 

Concentration mode 

A 900-mL solution of crude PVs was prepared (5.0%, 

w ⁄ v and pH 8.0) and equilibrated at 50 °C. The PT was 

set at 1.4 bar. The retentate was sent back to the feed 

tank, but the permeate was not recycled. UF concentration 

was performed with the 10- and 30-kDa UF 

membrane until a volume concentration factor (VCF) of 

6X was reached following a DF step where 10 diavolumes 

(DV) of distilled water was added to retentate. 

UF re-concentration was stopped when the VCF 

reached 11X. Retentate and permeate samples were 

collected during UF for further analysis. 

Analytical methods 

Phosphorus content and protein phosphorylation 

Phosphorus (P) content was determined by the colorimetric 

molybdenum method of Allen (1940). 

The Pierce Ò Phosphoprotein Phosphate Estimation 

Assay Kit was used to determine the extent of phosphorylation 

of purified protein (# catalog: 23270, Pierce, 

Rockford, IL, USA). Briefly, protein samples were 

dissolved in Tris-buffered saline and were mixed with 

Ultrafiltration of crude phosvitin B. P. Chay Pak Ting et al. 1635 

phosphate reagent. Protein samples were incubated for 

30 min at room temperature, and absorbance was 

measured at 650 nm. 

Sodium content 

Sodium content was determined by inductively coupled 

plasma (ICP-OES, Optima 4300, Dual View, Perkin– 

Elmer, Shelton, CT, USA). The wavelength used to 

determine Na element was 589.592 nm. Sodium analyses 

were carried out in radial view. 

Total nitrogen content 

The protein content was determined by the nitrogen 

combustion method with a LECO FP-528 apparatus 

(Model 601-500, LECO Corporation, St Joseph, MI, 

USA). The instrument was previously calibrated with 

ethylenediaminetetraacetic acid (EDTA). Total nitrogen 

was converted to protein using N factor 6.25 (% 

protein = % N · 6.25). 

SDS–polyacrylamide gel electrophoresis 

SDS–polyacrylamide gel electrophoresis (SDS–PAGE) 

was carried out using a Mini Protean 3 cell (Bio-rad 

Laboratories Ltd, Mississauga, ON, Canada) under 

denaturing conditions. The stacking and running gels 

were 4.0% and 12.5% polyacrylamide, respectively. The 

electrophoresis was run at a constant voltage of 30 mA 

for 1 h 30 min. About 10 lL of protein sample was placed 

on the gel. Protein bands were first stained with the 

modified Coomassie blue method by Hegenauer et al. 

(1977) to allow a more sensitive detection of phosphorylated 

proteins and followed by the normal Coomassie 

blue procedure. Destaining procedure containing acetic 

acid ⁄ methanol ⁄ water (1:8:12, v ⁄ v ⁄ v). Molecular 

weights were estimated with molecular weight markers 

from Bio-rad Laboratories Ltd (Hercules, CA, USA) 

composed of myosin (200 kDa), b-galactosidase 

(116.25 kDa), bovine serum albumin (84.79 kDa), 

carbonic anhydrase (37.41 kDa), trypsin inhibitor from 

soybean (29.05 kDa), lysozyme (19.81 kDa) and aprotinin 

(6.84 kDa). 


Preliminary observations at laboratory scale 

The effect of PT, pH and feed concentration on the flux 

values of crude PVs from egg yolk proteins using 

10-kDa UF membrane was investigated as preliminary 

laboratory-scale experiments. 

Effect of pH 

Flux values (Jp) of crude PVs solution were found to be 

influenced by the pH values between 4 and 10 (Fig. 1). 

The curve shows that Jp increased by 60% as the pH 

was raised from 4 to 10. Flux showed a minimum value 


1636 


Flux (L h –1 m –2 ) 

40 

35 

30 

25 

20 

15 

10 

5 

0 

0 2 4 6 8 10 12 

pH 

Figure 1 Effect of pH on flux values (L h )1 m )2 ) of crude phosvitin at 

1.0% (w ⁄ v) in a 0.09 m 2 PES spiral wound ultrafiltration (10-kDa 

molecular weight cut-off) membrane at 1.4 bar. Each value is the mean 

of triplicates ± standard error. 

(J p = 15.5 L h )1 m )2 ) at pH values close to the PVs 

isoelectric point (pHi = 4.0) where PVs is at its lowest 

solubility. According to Causeret et al. (1991), yolk 

granules are insoluble at pH from 4.2 to 6.5 and low 

ionic strength because they form non-soluble HDL–PVs 

complexes linked by phosphocalcic bridges as HDL and 

PVs contain a high proportion of phosphoserine residues 

that are able to bind calcium. A study on the 

influence of pH adjustments shows a minimum flux at a 

pH equal to the isoelectric point of protein solutions. It 

was suggested that the above may be attributed to 

fouling membrane (Nystro¨ m et al., 1994). Several explanations 

could be explaining this phenomenon. 

At isoelectric point, because of reduced electrostatic 

repulsion between protein molecules, a higher protein 

concentration is found at the membrane. This may also 

explain the lower flux at pH 4.0. As PVs global net charge 

is decreased, it is likely to adsorb onto the membrane 

surface, and the constituents of granules (HDL and PVs), 

which are found as microparticles (0.3–8 lm), can form a 

densely packed layer that may explain the low fluxes 

observed at the isoelectric point (Fane et al., 1983). An 

acidic precipitation phenomenon was also mentioned by 

Taborsky (1980), and aggregates would explain the 

lower flux at acidic pH. As pH values were raised 

above the isoelectric point, the permeate flux gradually 

increased (J p = 35.10 L h )1 m )2 ), probably as a result 

of increased solubility. Castellani et al. (2003) studied 

the solubility of PVs at different pH values and showed 

the highest solubility (97%) at neutral pH. Because of the 

fact that both PVs and polyethersulfone UF membranes 

are negatively charged at neutral pH, protein interactions 

with membrane’s surface are weakened because of 

increased electrostatic repulsions (Nystro¨ m et al., 

1998). As a result, high flux levels can be achieved. 

Effect of transmembrane pressure and feed concentration 

Figure S1 illustrates the effect of operating pressure 

(between 0.35 and 1.75 bar) on Jp at pH 8.0 of 0.5%, 

1.0%, 5.0% and 10.0% (w ⁄ v) of crude PVs solution 

when 10-kDa MWCO membrane is used. The Jw value 

corresponds to the water flux of the UF membrane 

measured before the UF experiment. In accordance with 

Darcy’s law, the increase in Jw value was linear over the 

entire P T range from 6.71 to 122.15 L h )1 m )2 . At 0.5% 

and 1.0% PVs, Jp values still increased linearly with PT 

from 4.80 to 48.44 L h )1 m )2 and from 7.73 to 

46.67 L h )1 m )2 , respectively. When increasing protein 

concentration from 1.0% to 5.0% or 10.0%, the flux 

values levelled off at pressures greater than 1.4 bar. As 

feed concentration was increased, the membrane surface 

concentration increased, leading to an increase in the 

pressure osmotic of the solution to the membrane. This 

reduced the net driving force for the solvent flux. This 

effect was more prominent at feed concentration at 

10.0% over the PT studied. In the pressure-independent 

UF, a limiting concentration and operating pressure 

were considered at 5.0% (w ⁄ v) and 1.4 bar, respectively. 

From this first series of observations, a feed concentration 

of 5.0% was chosen to concentrate PVs to have a 

reasonable permeate flux for concentration and desalting 

experiments at pH 8.0, under an operating pressure 

of 1.4 bar. Also, the operating temperature was set up 

at 50 °C, which is above the optimal temperature 

(30–35 °C) for microbial growth and below denaturation 

temperature of PVs (Td = 79.7 ± 1.4 °C) (Chung 

& Ferrier, 1995). Moreover, a higher temperature is 

favourable to the membrane filtration because of the 

reduction in feed viscosity. 

Comparative performance of 10- and 30-kDa ultrafiltration 

membranes 

The objective of using UF was to increase purity of 

crude PVs by removing as much of the salt and other 

solutes as possible in a UF–DF step, whereas minimising 

protein losses. 

The effect of operating time on the permeate flux on 10- and 

30-kDa molecular weight cut-off ultrafiltration membrane 

As shown in Fig. S2, the initial permeate fluxes of 

10 and 30 kDa were similar (52 and 56 L h )1 m )2 , 

respectively), and as expected, Jp was generally higher 

with the 30-kDa MWCO membrane than with the 

10-kDa MWCO membrane as a result of its higher 

MWCO. However, an important permeate flux decline 

was observed for both membranes within the first 

12 min of filtration. The initial permeate flux of 52 

and 56 L h )1 m )2 decreased about 48% and 42% when 

a final VCF value of 6 was reached for the 10- and 

30-kDa UF membranes, respectively. This phenomenon 

could be attributed to the fact that PVs is retained by the 

membrane and accumulation on membrane surface 

form a layer at the membrane surface with increasing 

operating time, which cause the resistance of the 

permeate flux, resulting in the decline of the flux. 


Adding DF deionised water (10 DV) restored the flux 

close to initial value for the 30-kDa MWCO membrane 

but not for the 10-kDa MWCO membrane. Because of 

these discrepancies in membrane permeability, the final 

VCF (11X) was reached after 34 min with the 30-kDa 

membrane, whereas only after 56 min with the 10-kDa 

membrane. It can be thought that the greater porosity of 

the 30-kDa membrane compared to that of the 10-kDa 

membrane may have resulted in less pore blocking of the 

membrane upon UF concentration (Cheryan, 1998). 

Concentration and desalting of crude phosvitin 

The efficiency of UF membrane to desalt crude PVs has 

been evaluated by determining the changes in the 

sodium content in retentates during the UF–DF process, 

while the losses in PVs was estimated by the changes in 

phosphorus content. 

Table S1 reports the sodium content of the PVs 10and 

30-kDa UF retentates of samples collected during 

concentration and DF steps. It can be seen that, except 

for the dilution effect before DF, the Na values 

remained unchanged during concentration and DF 

mode. For a DF ratio of 1.67 (1500 mL of deionised 

water was added to the initial 900 mL of crude PVs 

solution), the Na content of crude PVs solution was 

reduced by 86% for both UF membranes. Some 

apparent limitations in the passage of Na during UF– 

DF may result from the occurrence of salt binding by 

proteins such as PVs and are known to possess mineralbinding 

activity for cations such as Fe, Ca, Mg and Na 

(Lu & Baker, 1986). 

Figure 2a–d summarises the effect of VCF on retentate 

and permeate of phosphorus with 10- and 30-kDa 

MWCO membranes before and with DF. As expected, 

with increasing VCF, the phosphorus content in the 

retentate increased from an initial value of 0.153 up to 

0.390 g L )1 when the VCF was equal to 6 for both 

membranes. A 2.6-fold increase in concentration was 

achieved. At the end of the last DF step, phosphorus 

content is slightly higher (0.441 and 0.405 g L )1 for 10 

and 30 kDa, respectively) than initial phosphorus content 

(before DF) for both UF membranes, meaning that 

few losses of phosphoproteins occurred during DF. 

Phosphorus in the permeate averaged 0.019–0.017 g L )1 

for 10- and 30-kDa MWCO membranes, respectively. 

This indicates that limited amounts of phosphorylated 

proteins permeated through the membrane. It can be 

seen that protein retention was not affected by the 

concentration of PVs in the feed and was 80% or higher 

for both UF membranes. 

Compositional characteristics of crude delipidated egg yolk 

proteins, phosvitin and ultrafiltration retentates 

The compositional data reported in Table 1 show that, 

in comparison with delipidated egg yolk proteins, the 

crude PVs had lower protein content (7.25% vs. 

(a) 

Phosphorus content (g L –1 ) 

(b) 


(c) 


(d) 


0.50 

0.40 

0.30 

0.20 

0.10 

0.00 

0.50 

0.40 

0.30 

0.20 

0.10 

0.00 

0.50 

0.40 

0.30 

0.20 

0.10 

0.00 

0.50 

0.40 

0.30 

0.20 

0.10 

0.00 


UF Retentate 10 kDa UF-Permeate 10 kDa Retention 

1.5 1.8 2.25 3 6 

VCF 

UF-Retentate 10 kDa (with DF) UF-Permeate 10 kDa (with DF) Retention 

1.32 1.94 3 6.6 11 

VCF 

UF-Retentate 30 kDa UF-Permeate 30 kDa Retention 

1.5 1.8 2.25 3 6 

VCF 

UF-Retentate 30 kDa (with DF) UF-Permeate 30 kDa (with DF) Retention 

1.32 1.94 3 6.6 11 

VCF 

Figure 2 Effect of the volume concentration factor on phosphorus 

during ultrafiltration (UF) concentration ⁄ diafiltration (DF) of crude 

phosvitin in a 0.09 m 2 PES spiral wound UF membranes: (a) UF 

concentration and (b) UF–DF 10-kDa molecular weight cut-off 

(MWCO) membrane, (c) UF concentration and (d) UF–DF 30-kDa 

MWCO membrane. 

77.31%). However, the atomic ratio of nitrogen to 

phosphorus (N ⁄ P), which reflects PVs purity (Castellani 

et al., 2003), decreased from 31.13 to 7.92 from delipidated 

egg yolk proteins to crude PVs, approaching the 

value of 2.85 for PVs standard (data not shown). The 

dissociation of granules under high saline conditions 

enabled the precipitation of non-PVs protein material 

Ó 2010 The Authors. Journal compilation Ó 2010 Institute of Food Science and Technology International Journal of Food Science and Technology 2010 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

Retention ( ) Retention ( ) 

Retention ( ) 

Retention ( )

1638 


Table 1 Comparative composition of egg yolk proteins, crude phosvitin 

and UF retentates and proteins yield (%) upon desalting of crude 

phosvitin using 10- and 30-kDa UF membranes 

Delipidated 

egg yolk 

proteins Crude PVs 

UF retentates 

10 kDa 30 kDa 

N [Protein] † (%) 12.37 [77.31] 1.18 [7.25] 7.22 [45.12] 6.86 [42.89] 

Phosphorus (%) 0.88 [11.8] à 

0.33 [50.23] 2.11 [56.47] 2.29 [60.80] 

Atomic ratio N ⁄ P 31.13 7.92 7.58 6.64 

Protein yield (%) §,– 

– 89 

100 84.7 84.4 

PVs, phosvitin; UF, ultrafiltration; DF, diafiltration. 

† 

Values in brackets are N values converted into proteins (N · 6.25). 

à 

Values in brackets are calculated phosphorus molecules per molecule of 

protein. 

§ 

Protein yield (%) = protein (g) in final retentate ⁄ protein (g) in feed 

solution · 100. 

– 

First row reports yield upon preparation of crude PVs and second row 

reports yield upon UF–DF desalting of PVs. 

from egg yolk and thus, centrifugal separation increased 

its purity. 

The apparent yield (%) of crude PVs calculated by the 

ratio (lyophilised crude PVs weight ⁄ delipidated egg yolk 

proteins weight) · 100 was estimated to 89%, but this 

value does not adequately reflect the recovery of PVs 

because protein losses may have been compensated by 

NaCl brought during the precipitation step. 

UF–DF generated an increase in the protein content 

from crude PVs by a 6.25-fold and 5.92-fold factor upon 

UF using the 10- and 30-kDa membranes, respectively. 

Desalting PVs using UF–DF permitted to further 

decrease the atomic N ⁄ P ratio to 7.58 and 6.64 by UF 

retentate of 10 and 30 kDa, respectively. Overall, the 

protein yield obtained upon UF–DF desalting of PVs 

using 10- and 30-kDa MWCO membranes averaged 84%. 

Protein profile of egg yolk proteins, crude phosvitin and 

ultrafiltration retentates 

The results of the SDS–PAGE patterns of standard PVs, 

delipidated egg yolk proteins, crude PVs and retentate 

of 10- and 30-kDa UF membranes are shown in Fig. 3. 

The commercial PVs standard (lane 2) comprises all of 

the PVs polypeptides (a1, a2 and b-PVs), whereas the 

delipidated egg yolk proteins revealed at least fifteen 

bands (lane 3) with molecular weights ranging from 

around 5 to 200 kDa. Five bands whose theoretical 

molecular weight value was estimated to be between 

35 and 45 kDa may be attributed to PVs. The remaining 

bands in the high molecular weight region could 

correspond to some protein material in the HDL region 

consisting of polypeptides between 32 and 105 kDa or 

apoproteins LDL with molecular weight between 19 and 

M.W. (kDa) 1 2 3 4 5 6 

198.5 

116.2 

84.8 

53.9 

37.4 

29 

19.8 

6.8 

Figure 3 SDS–PAGE profiles of egg yolk protein, crude phosvitin and 

ultrafiltration (UF) retentates. Lane 1: molecular weight markers; 

Lane 2: phosvitin standard; Lane 3: delipidated egg yolk proteins; 

Lane 4: crude phosvitin; Lane 5: 10-kDa UF retentate; Lane 6: 30-kDa 

UF retentate. 

225 kDa in either delipidated or non-delipidated egg 

yolk proteins (Guilmineau et al., 2005). 

The salting-out of egg yolk proteins using NaCl 10% 

induced the disruption of the phosphocalcic bridges 

binding PVs and HDL and presumably yielded to a 

pellet rich in HDL upon centrifugation. The highly 

charged crude PVs remained soluble and appeared 

separated into several sub-units with molecular weight 

ranging between 10 and 80 kDa (lane 4). Also, different 

molecular weights of PVs are reported in the literature 

because of their phosphoric acid and carbohydrate 

content varying between 13 and 40 kDa (Wallace & 

Morgan, 1986). A double band with an apparent 

molecular weight of 35–36 kDa for both commercial 

and our laboratory-prepared PVs was observed. The 

b-PVs was composed mainly of polypeptide of 45 kDa, 

and the a-PVs contained bands between 37.5 and 

45 kDa. However, a-PVs migrated faster than b-PVs, 

because its molar mass was lower (Abe et al., 

1982). However, additional polypeptides of HDL that 

migrated in the same region than PVs could be present 

given their close molecular weight. 

The SDS–PAGE patterns of UF retentates from 

10 and 30 kDa (lane 5 and 6), respectively, show that 

intensity of PVs bands increased after UF–DF, compared 

to crude PVs (lane 4) but no noticeable difference 

between the bands profile of the two retentates can be 

observed. This observation suggests that using either a 

10-kDa or a 30-kDa MWCO membrane for the UF–DF 

procedure did not alter significantly the protein profile 

of PVs and that the procedure mainly resulted in a 

desalting of the extract. 


Conclusion 

Our observations suggest that an UF-based process, 

under the conditions studied, can be used to partially 

concentrate and desalt PVs from delipidated egg yolk 

proteins. The performances, in terms of production, 

protein yield and desalting capacity, are similar for both 

UF membranes. However, UF–DF using membrane 

with MWCO of 30 kDa has been shown to result in 

increase in term of permeation flux. Further investigations 

into the influence of hydrodynamic conditions on 

fouling would be necessary to explain the differences in 

permeation flux. 


This work was supported by a grant from Advanced 

Food and Materials Network (AFMnet) and the Natural 

Sciences and Engineering Research Council of 

Canada (NSERC). 

References 

Abe, Y., Itoh, T. & Adachi, S. (1982). Fractionation and characterization 

of hen’s egg yolk phosvitin. Journal of Food Science, 47, 

1903–1907. 

Akita, E.M. & Nakai, S. (1992). Immunoglobulins from egg yolk: 

isolation and purification. Journal of Food Science, 57, 629–634. 

Allen, R.J.L. (1940). The estimation of phosphorus. Biochemical 

Journal, 34, 858–865. 

Anton, M., Martinet, V., Dalralarrondo, V., Beaumal, V., David 

Brian, E. & Rabesona, H. (2003). Chemical and structural characterisation 

of low-density lipoproteins purified from hen egg yolk. 


Burley, R.W. & Cook, W.H. (1961). Isolation and composition of 

avian egg yolk granules and their constituents a and b-lipovitellins. 

Canadian Journal of Biochemistry and Physiology, 39, 1295–1307. 

Byrne, B.M., Van het Schip, A.D., Van de Klundert, J.A.M., Arnberg, 

A.C., Gruber, M. & Geert, A.B. (1984). Amino acid sequence of 

phosvitin derived from the nucleotide sequence of part of the 

chicken vitellogenin gene. Biochemistry, 23, 4275–4279. 

Castellani, O., Martinet, V., David-Briand, E., Guerin-Dubiard, C. & 

Anton, M. (2003). Egg yolk phosvitin: preparation of metal-free 

purified protein by fast protein liquid chromatography using 

aqueous solvents. Journal of Chromatography. B, Analytical Technologies 

in the Biomedical and Life Sciences, 791, 273–284. 

Causeret, D., Matringe, E. & Lorient, D. (1991). Ionic strength and pH 

effects on composition and microstructure of yolk granules. Journal 

of Food Science, 56, 1532–1536. 

Cheryan, M. (1998). Ultrafiltration and Microfiltration Handbook. Pp. 

237–291. Lancaster, PA: Technomic. 

Choi, I., Jung, C., Seog, H. & Choi, H. (2004). Purification of 

phosvitin from egg yolk and determination of its physicochemical 

properties. Food Science and Biotechnology, 13, 434–437. 

Chung, S.L. & Ferrier, L.S. (1995). Heat denaturation and emulsifying 

properties of egg yolk phosvitin. Journal of Food Science, 60, 906– 

908. 

Connelly, C. & Taborsky, G. (1961). Chromatographic fractionation 

of phosvitin. The Journal of Biological Chemistry, 236, 

1364–1368. 

Fane, A.G., Fell, C.J.D. & Suki, A. (1983). The effect of pH and ionic 

environment on the ultrafiltration of protein solutions with retentive 

membranes. Journal of Membrane Science, 16, 195–210. 


Guilmineau, F., Krause, I. & Kulozik, U. (2005). Efficient analysis of 

egg yolk proteins and their thermal sensitivity using sodium dodecyl 

sulfate polyacrylamide gel electrophoresis under reducing and 

nonreducing conditions. Journal of Agricultural and Food Chemistry, 

53, 9329–9336. 

Hegenauer, J., Ripley, L. & Nace, G. (1977). Staining acidic 

phosphoproteins (phosvitin) in electrophoretic gels. Analytical 


Jiang, B. & Mine, Y. (2000). Preparation of novel functional 

oligophosphopeptides from hen egg yolk phosvitin. Journal of 


Jiang, B. & Mine, Y. (2001). Phosphopeptides derived from hen egg yolk 

phosvitin: effect of molecular size on the calcium-binding properties. 

Bioscience, Biotechnology, and Biochemistry, 65, 1187–1190. 

Katayama, S., Xu, X., Fan, M.Z. & Mine, Y. (2006). Antioxidative stress 

activity of oligophosphopeptides derived from hen egg yolk phosvitin 

in Caco-2 cells. Journal of Agricultural and Food Chemistry, 54, 773– 

778. 

Kim, H. & Nakai, S. (1996). Immunoglobulin separation from 

egg yolk: a serial filtration system. Journal of Food Science, 61, 

510–513. 

Losso, J.N. & Nakai, S. (1994). A simple procedure for the isolation of 

phosvitin from chicken egg yolk. In: Egg uses and Processing 

Technologies: New Developments (edited by J.S. Sim & S. Nakai). 

Pp. 150–157. Oxon, U.K: Cab International. 

Lu, C.H.-L. & Baker, R.C. (1986). Characteristics of egg-yolk 

phosvitin as an antioxidant for inhibiting metal-catalyzed phospholipid 

oxidations. Poultry Science, 65, 2065–2070. 

Martin, W.G., Vandegaer, J.E. & Cook, W.H. (1957). Fractionation of 

livetin and the molecular weights of the a- and b-components. 

Canadian Journal of Biochemistry and Physiology, 35, 241–250. 

Mecham, D.K. & Olcott, H.S. (1949). Phosvitin, the principal phosphoproteins 

of egg yolk. Journal of the American Chemical Society, 71, 

3670–3679. 

Mine, Y. (1998). Adsorption behaviour of egg yolk low-density 

lipoproteins in oil-in-water emulsions. Journal of Agricultural and 


Nakamura, S., Ogawa, M., Nakai, S., Kato, A. & Kitts, D.D. (1998). 

Antioxidant activity of a maillard-type phosvitin-galactomannan 

conjugate with emulsifying properties and heat stability. Journal of 


Nyström, M., Pihlajamäki, A. & Ehsani, N. (1994). Characterization 

of ultrafiltration membranes by simultaneous streaming potential 

and flux measurements. Journal of Membrane Science, 87, 245–256. 

Nyström, M., Aimar, P., Luque, S., Kulovaara, M. & Metsämuuronen, 

S. (1998). Fractionation of model proteins using their physiochemical 

properties. Colloids and Surfaces A: Physicochemical and 

Engineering Aspects, 138, 185–205. 

Sattar Khan, M.A., Nakamura, S., Ogawa, M., Akita, E., Azakami, 

H. & Kato, A. (2000). Bactericidal action of egg yolk phosvitin 

against Escherichia coli under thermal stress. Journal of Agricultural 


Taborsky, G. (1968). Optical rotatory dispersion and circular dichroism 

of phosvitin at low pH. The Journal of Biological Chemistry, 243, 

6014–6020. 

Taborsky, G. (1980). Iron binding by phosvitin and its conformational 

consequences. The Journal of Biological Chemistry, 255, 2976–2984. 

Tsutsui, T. & Obara, T. (1984). Preparation and characterization of 

phosvitin from hen’s egg-yolk granule. Agricultural and Biological 

Chemistry, 48, 1153–1160. 

Wallace, R.A. & Morgan, J.P. (1986). Chromatographic resolution of 

chicken phosvitin: multiple macromolecular species in a classic 

vitellogenin-derived phosphoprotein. Biochemical Journal, 240, 871– 

878. 

Xu, X., Katayama, S. & Mine, Y. (2007). Antioxidant activity of 

tryptic digests of hen egg yolk phosvitin. Journal of Agricultural and 



1640 





Figure S1. Effect of transmembrane pressure and 

phosvitin concentration (0.5%, 1.0%, 5.0% and 

10.0%, w ⁄ v) on flux values (L h )1 m )2 ) at pH 8.0 in a 

0.09 m 2 PES spiral wound ultrafiltration (10-kDa 

molecular weight cut-off) membrane. Each value is the 

mean of triplicates ± standard error. 

Figure S2. Permeation flux (L h )1 m )2 ) of crude phosvitin 

(5.0%, w ⁄ v and pH 8.0) at 1.4 bar during ultrafiltration 

(UF) concentration in a 0.09 m 2 PES spiral 

wound UF (10- and 30-kDa molecular weight cut-off) 

membranes. 

Table S1. Sodium content (g L )1 ) of the UF retentate 

during the concentration and DF steps of crude phosvitin 

using 10- and 30-kDa molecular weight cut-off 

membranes. 









Effects of cellulosic fibre on physical and rheological properties of 

starch, gluten and wheat flour 

Avi Goldstein, Lida Ashrafi & Koushik Seetharaman* 

Laboratory of Cereal Science, Department of Food Science, University of Guelph, Guelph, ON N1G 2W1, Canada 


Summary In this study, we report on the effects of cellulose fibres of different particle size on changes to dough water 

absorption and rheology; and on effects of fibre on starch and gluten, separately, at different levels of fibre 

incorporation (0.1–10%). Water absorption and dough-mixing properties were affected with fibre 

incorporation, with 40-lm fibre incorporation resulting in greater absorption values. Dough stickiness 

and extensibility were affected by cellulose fibre particle size, and decreased with increasing fibre addition. 

Flour or starch and fibre mixtures were evaluated using a Micro ViscoAmlyoGraph (MVAG), and the 

resulting gel firmness was measured using a texture analyzer. MVAG peak and final viscosities of flour 

samples decreased with increasing fibre content. Starch–fibre interactions followed a similar trend as flour– 

fibre treatments. Gluten–fibre interactions were also measured using a Gluten Peak Tester on flour–fibre and 

gluten–fibre mixtures. Cellulose fibre enhanced the kinetics of gluten aggregation. 

Keywords Cellulose, dough, extensibility, gel texture, gluten, insoluble fibre, pasting profile, starch, stickiness. 

Introduction 

The demand for healthier food products with excellent 

sensory qualities is increasing (Ang, 2001). Dietary fibre 

consumption leads to reduction in serum cholesterol 

levels and the risk of colon cancer (Seguchi et al., 2007). 

Cellulose, an insoluble, high molecular weight, linear, 

homopolymer of glucose joined by (1,4) glycosidic 

linkages increases the insoluble dietary fibre content 

when incorporated in food products. 

Several studies have demonstrated the use of dietary 

fibre in baked products (Pomeranz, 1977; Chen et al., 

1988; Gomez et al., 2003; Seguchi et al., 2007; Kohajdova 

& Karovicova, 2008). A significant problem with dietary 

fibre addition in bread-type products is the reduction in 

loaf volume and poor textural quality (Gomez et al., 

2003). Tortilla dough containing 8% soluble fibre exhibited 

poor gluten development and extensive gelatinisation 

during baking, which resulted in tortillas with a dense, 

pasty crumb (Seetharaman et al., 1997). Furthermore, 

insoluble fibres physically disrupted the gluten network in 

tortilla dough resulting in collapsed air bubbles, and 

decreased shelf stability of tortillas (Seetharaman et al., 

1997). Soluble fibres affected bread dough stability, 

provided a higher water absorption capacity, increased 

mixing tolerance, and decreased extensibility of fortified 

*Correspondent: E-mail: kseethar@uoguelph.ca 

doi:10.1111/j.1365-2621.2010.02323.x 

Ó 2010 Institute of Food Science and Technology 

samples when compared to control samples (Gomez 

et al., 2003). Overall, insoluble fibres can be used as 

bulking or as texturing agents, but they can also impart a 

mouth feel, and degrade certain functional properties of 

food products (Guillon & Champ, 2000). 

Dietary fibres from a variety of sources may be 

incorporated in tortillas acceptably at £12% fibre incorporation 

(Seetharaman et al., 1994). Seguchi et al. (2007) 

examined the effect of size of cellulose fibre (6–650 lm) 

on bread-making properties. They observed that normal 

bread-making properties were obtained with the use of 

cellulose fibres >154 lm. Cellulose fibres

1642 

Effects of cellulosic fibre on physical and rheological properties A. Goldstein et al. 

Table 1 Water and Oil retention capacities of SC40 and SC200 

cellulose fibre 

Sample 

(40 lm) and SC200 (200 lm) cellulose fibre was supplied 

by Creafill Fibre Inc. (Chestertown, MD, USA). 

The fibres are 98.4% insoluble fibre, and only size of the 

fibres was different in this study. Water and oil retention 

capacities determined using AACC method 56-11.02 

(AACC International, 2000b) are reported in Table 1. 

Hard wheat flour (HWF) (12% moisture, 12% protein) 

and soft wheat flour (SWF) (12% moisture, 8% protein) 

was provided by Griffith Laboratories (Toronto, ON, 

Canada). Gluten was provided by ADM (ADM, Montreal, 

QC, Canada). 

Starch–fibre interactions 

Water retention 

capacity (g of 

water ⁄ gof 

sample) 

Oil retention 

capacity (g 

of oil ⁄ gof 

sample) 

SC40 2.88 ± 0.09 2.30 ± 0.13 

SC200 4.30 ± 0.26 5.07 ± 0.21 

The pasting profiles of starch–fibre blends were determined 

using a Micro ViscoAmlyoGraph (MVAG) 

(Brabender Inc., Duisburg, Germany). Starch–fibre 

blends were prepared on an 8% db. Starch was replaced 

with cellulose fibre at 0.1%, 1%, 5% or 10%. Samples 

were initially held at 25 °C for 3 min; then, heated from 

25 to 95 °C at10°C min )1 ; held at 95 °C for 3 min, 

cooled from 95 to 25 °C at a rate of 10 °C min )1 ; and 

finally held at 25 °C for 3 min. The different pasting 

attributes such as peak and final viscosity were recorded 

using the software. 

The resulting pastes were immediately poured into 

35 · 10 mm petri dishes, covered with parafilm and 

stored for 1 day at room temperature or for 7 days at 

4 °C. Gel texture was determined using a TAXT2-plus 

texture analyzer (Stable Micro Systems, Surrey, UK) 

fitted with a 35-mm cylindrical compression probe. 

Immediately prior to texture analysis, the gel surface 

was sliced so that a fresh surface was available for 

texture analysis. The gel was compressed at a rate of 

1.0 mm s )1 to a distance of 3.0 mm. All measurements 

were conducted in duplicate. 

Gluten–fibre interactions 

Gluten–fibre blends were analysed using a Brabender 

Gluten Peak Tester (GPT). Gluten blends were 

prepared by replacement with cellulose fibre at 0.1%, 

1%, 5% or 10%. The GPT was operated at 30 °C for 

20 min at a speed of 3000 rpm. The sample to water 

ratio was 5.5 g gluten to 12 mL water. Data for peak 

torque, lift off time and peak maximum time were 

obtained from the software. All measurements were 

conducted in duplicate. 

Flour–fibre interactions 

The pasting profile and gel texture of flour–fibre blends 

(12% db) was determined using a Brabender MVAG 

and TAXT2-plus texture analyzer. The percentage of 

cellulose fibre replacement and methods did not differ 

from those listed in starch–fibre interactions methods. 

A Brabender Farinograph-E was used to measure the 

water absorption properties using the standard AACC 

Method 54-21A (AACC International, 2000a), but using 

a 50-g mixing bowl. Cellulose fibre was incorporated at 

0.1%, 1%, 5% or 10% replacement levels. Farinograph 

attributes that were measured included water absorption, 

dough development time, stability and mixing 

time. The stickiness of the dough was determined using 

a Chen and Hoseney Stickiness Cell (Stable Micro 

Systems) according to the method described by Chen & 

Hoseney (1995). A Brabender Extensograph-E was used 

[ICC standard method no. 114 ⁄ 1 (International Association 

of Cereal Science and Technology, 1992)] to 

further analyse the visco-elastic characteristics of flour– 

fibre samples. Flour–fibre blends were also analysed 

using a GPT. Flour blends were prepared by replacement 

with cellulose fibre at 0.1%, 1%, 5% or 10%. The 

GPT was operated at 30 °C for 20 min at a speed of 

3000 rpm. The sample to water ratio was 9.9 g SWF or 

8.5 g HWF to 10 mL water. Data for peak torque, lift 

off time and peak maximum time were obtained from 

the software. All measurements were conducted in 

duplicate. 


Statistical determinations including analysis of variance 

(anova) followed by Tukey’s test to distinguish the 

responses of different fibre treatments was performed 

using spss 16.0 (SPSS Inc., Chicago, IL, USA). The 

level at which significant differences are reported is 

0.05. 

Results and Discussion 

Starch–fibre interactions 

The replacement of starch with SC40 cellulosic fibre 

significantly (P < 0.05) affected the pasting properties 

of starch systems (Fig. 1a). The peak and final viscosities 

of starch–fibre blends decreased with increasing 

amount of cellulose fibre. The longer fibre (SC200) 

exhibited a similar trend to that reported for SC40 (data 

not shown). Decrease in viscosity is likely indicative of a 

reduced degree of starch granule swelling (Collar et al., 


(a) 

Peak viscosity/final viscosity (cp) 

(b) 

Peak viscosity/final viscosity (cp) 

1200 

1000 

800 

600 

400 

200 

1600 

1400 

1200 

1000 

Starch peak viscosity Starch final viscosity 

0 

0 2 4 6 8 10 

Cellulose fibre replacement (%) 

800 

600 

400 

200 

Soft wheat peak viscosity Soft wheat final viscosity 

Hard wheat peak viscosity Hard wheat final viscosity 

0 

0 2 4 6 8 10 


Figure 1 Peak and final viscosity values of (a) wheat starch and 

(b) hard and soft wheat flour when measured with a Micro 

ViscoAmlyoGraph at different replacement levels of SC40 

cellulosic fibre. 

2006) and ⁄ or reduced leaching of polymers into the 

extragranular spaces. 

The firmness of starch–fibre gels decreased with 

increasing amount of cellulose fibre replacement 

(Fig. 2a). The lower firmness values may be indicative 

of the ability of cellulose fibre to interfere with the 

retrogradation of starch polymers. Santos et al. (2008) 

also reported a decrease in starch retrogradation in the 

presence of insoluble fibre when measured using a 

calorimetric technique. 

Gluten–fibre interactions 

The GPT is a new instrument that measures gluten 

aggregation properties. Figure 3a,b represents a sample 

GPT curve depicting locations of lift off time, peak time, 

and maximum torque for gluten and flour samples, 

respectively. Research in our laboratory has demonstrated 

that native starch or other flour components do 

not affect attributes measured by the GPT (data not 

shown). The addition of SC40 cellulose fibre decreased 

the peak development time and lift off time of gluten– 

fibre systems (Fig. 4a). With the exception of 0.1% 

Effects of cellulosic fibre on physical and rheological properties A. Goldstein et al. 1643 

(a) 

(b) 

Figure 2 Firmness and gradient of (a) wheat starch and (b) hard wheat 

and soft wheat flour gels with SC40 replacement as determined by 

compression with texture analyzer. 

replacement level, a strong negative relationship existed 

between lift off time of gluten and replacement level of 

cellulose fibre (r 2 = 0.84). The longer fibre (SC200) 

exhibited a similar trend (data not shown) to data 

reported for SC40. The decrease in peak time is likely 

related to the ability of cellulose fibre to enhance the 

kinetics of gluten network aggregation. 

Flour–fibre interactions 

The addition of SC40 fibre to SWF resulted in significant 

differences (P < 0.05) to the flour-pasting properties. 

Significant differences (P < 0.05) in peak and 

final viscosities of SWF were observed between flour 

samples with 0.1–1% cellulose fibre replacement and 


1644 


(a) 

Torque (BE) 

(b) 

Torque (BE) 

70 

60 

50 

40 

30 

20 

Max torque (BE) 

Lift off time 

(min) 

10 

0 

Peak max time 

(min) 

0 5 10 

Time (min) 

15 20 

30 

25 

20 

15 

10 

5 

Lift off time 

(min) 

Max torque (BE) 

Peak max time 

(min) 

Soft wheat flour 

Gluten 

0 

0 2 4 6 

Time (min) 

8 10 

Figure 3 Sample Gluten Peak Tester curve of (a) gluten (5.5 g gluten 

⁄ 12 mL H2O) and (b) soft wheat flour (9.9 g flour ⁄ 10 mL H2O), 

with location of lift off time, peak time, and max torque. 

flour systems with 5–10% fibre replacement (Fig. 1b). 

Trends for HWF were similar to trends for SWF. 

Trends in the firmness and gradient of flour–fibre gels 

(Fig. 2b) were supported by results of starch–fibre gels. 

The longer fibre (SC200) exhibited similar trends to that 

reported for SC40 (data not shown). 

The data from using GPT for flour–fibre interactions 

exhibited significant differences. For SWF, a consistent 

decrease in lift off time and peak time was observed with 

an increasing percentage of SC40 fibre replacement 

(Fig. 4b), although lift off and peak times were not 

significantly different (P < 0.05) at 0.1% and 1.0% 

replacement levels. Trends for HWF were similar to 

those observed for SWF. A weak flour will have a rapid 

build up in dough consistency to a sharply defined peak 

followed by a rapid break down, while a strong flour will 

have a much slower build up in dough consistency and a 

relatively more time to achieve peak consistency. At the 

0.1% and 1.0% replacement levels, there was no 

significant (P < 0.05) effect of cellulose on the strength 

of the HWF, while the replacement levels of 5% and 

10% significantly (P < 0.05) weakened the flour, in 

particular the gluten network. The longer fibre (SC200) 

(a) 

12 

Lift off time/Peak time (min) 

10 

8 

6 

4 

2 

0 

(b) 

2.5 

Lift off time/Peak time (min) 

2 

1.5 

1 

0.5 

0 

Gluten lift off time 

Gluten peak time 

0 0.1 1 5 10 


Soft wheat lift off time Hard wheat lift off time 

Soft wheat peak time Hard wheat peak time 

0 0.1 1 5 10 


Figure 4 Effect of cellulose fibre replacement on (a) gluten and (b) 

hard wheat and soft wheat flour using Gluten Peak Tester. 

exhibited a similar trend to that reported for SC40 (data 

not shown). 

Significant differences (P < 0.05) in water absorption 

were observed between dough systems containing hard 

and soft wheat flours (Fig. 5). Systems containing HWF 

maintained greater water absorption values than those 

containing SWF. This result agrees with values reported 

in the literature (Gomez et al., 2003; Seguchi et al., 

2007). The differences in water absorption between hard 

and SWF are believed to be attributed to the protein 

content of the flours. Proteins which are naturally 

present in flour, including gluten-forming proteins, are 

able to absorb 1–2· their weight in water (Pareyt & 

Delcour, 2008). Therefore, slight changes in the protein 

content of flours can contribute to large differences in 

the water absorption of samples. 

As the amount of cellulose fibre increased, the 

increase in water absorption for the SWF to achieve a 

dough consistency of 500 BU was higher than that for 


Water absorption (%) 

75 

70 

65 

60 

Hard wheat Sc40 

55 

Soft wheat Sc40 

Hard wheat Sc200 

Soft wheat Sc200 

50 

0 2 4 6 8 10 


Figure 5 The effect of cellulosic fibre replacement on the water 

absorption of hard and soft wheat flour systems. 

HWF. The increase in water absorption is believed to be 

related to the presence of cellulose fibre (Gomez et al., 

2003). Cellulose fibres are able to hold many times their 

initial weight in water, and the hydroxyl groups present 

in cellulose fibre allows for more interactions with water 

through hydrogen bonding (Gomez et al., 2003). The 

starch content of the flour is not believed to influence the 

water absorption of flour samples as the starch granules 

will not absorb water at ambient temperature. The 

addition of cellulosic fibre to dough systems did not 

appear to have a significant (P < 0.05) influence on the 

mixing time, stability, and development time of the 

dough samples. 

The addition of SC40 and SC200 cellulose fibre to 

SWF dough resulted in dough with decreased stickiness 

values (Fig. 6). At cellulose fibre replacement levels of 

1–10%, SWF dough with SC40 replacement were 

significantly (P < 0.05) more sticky than dough with 

SC200 fibre replacement. With HWF, significant 

(P < 0.05) decreases in stickiness were observed at 

5–10% fibre replacement levels, although no significant 

(P < 0.05) differences in stickiness were observed at 

0–1% fibre replacement levels. The decrease in stickiness 

in HWF from 5% to 10% cellulose replacement was 

proportionally greater in HWF dough treated with 

SC200 when compared to SC40. In general, HWF 

dough was stickier than SWF dough. This may be 

attributed to the greater protein and gluten content 

present in HWF. Glutenin, a component of gluten, is 

believed to be contributing to stickiness of dough 

systems (Eliasson & Larrson, 1993). 

The effect of cellulose fibre replacement on the viscoelastic 

properties of HWF and SWF are listed in 

Effects of cellulosic fibre on physical and rheological properties A. Goldstein et al. 1645 

Figure 6 The effect of cellulosic fibre replacement on the stickiness 

of hard and soft wheat flour systems. 

Table 2. In general, significant (P < 0.05) reductions 

in energy required to tear both HWF and SWF dough 

was observed at 10% replacement levels. The reduction 

in the energy required was less in the case of the HWF 

dough (approximately 10% and 20% reduction was 

observed when both fibre sizes were replaced at 5% and 

10%, respectively) compared to the SWF dough. The 

energy required to tear the SWF dough was reduced 

22.5% and 47.5% when SC40 was used at 5% and 10% 

replacement level; and 15.0% & 30.0% when SC200 was 

used at 5% and 10% replacement levels, respectively. 

For SWF, the R ⁄ E ratio values were seen to increase 

with an increasing percentage of fibre replacement. At 

5% and 10% replacement levels, SC200 fibre main- 

Table 2 Comparison of the visco-elastic nature of hard and soft wheat 

flour dough with SC40 and SC200 cellulose fibre replacement 

Flour type Hard wheat flour Soft wheat flour 

Fibre replacement (%) SC40 SC200 SC40 SC200 

Energy (cm 3 ) 

0 100ab * 

100ab 40a 40a 

0.1 108a 103a 37ab 37ab 

1 114a 107a 38ab 39a 

5 90ab 92ab 31c 34b 

10 

R ⁄ E Ratio 

80b 76b 21d 28c 

0 1.1a 1.1a 0.97a 0.97a 

0.1 1.0a 1.0a 0.9a 0.89a 

1 1.2a 1.2a 1.1a 1.18a 

5 1.5b 2.4b 1.43b 1.84b 

10 2.3c 3.2c 1.49b 2.55c 

*Values in columns sharing the same measurement parameter followed 

by the same letter are not significantly different at the 95% confidence 

level. 


1646 


tained greater R ⁄ E ratio values than SC40 fibre 

replacement in SWF and HWF dough. 

The visco-elastic changes observed may be related to 

interactions between insoluble fibres and protein networks. 

The increase in the resistance ⁄ extensibility (R ⁄ E) 

ratio observed with increasing fibre replacement is 

indicative that the dough became stiffer with increasing 

fibre replacement (Sudha et al., 2007). Increase in R ⁄ E 

ratio was proportionally greater for HWF samples 

compared to SWF. Insoluble fibres such as cellulose 

are believed to disrupt and weaken the gluten matrix 

(Seetharaman et al., 1997). A symptom of the weakened 

matrix may be the reduced extensibility values and 

reduced energy required to tear dough observed with 

increasing cellulose fibre replacement. 

Conclusion 

It can clearly be seen that the addition of cellulose fibre 

affected many rheological and visco-elastic parameters 

of starch, gluten and flour systems. The size of cellulose 

fibre addition was seen to cause significant differences in 

water absorption, stickiness and visco-elastic properties 

of flour–fibre systems. Cellulose fibre affected the 

dough-mixing characteristics of wheat flour as greater 

water absorption was required to attain a consistency of 

500 BU with increasing fibre replacement. The pasting 

properties of flour and starch systems were similarly 

affected by cellulosic fibre replacement, as the competition 

for water in the presence of cellulose resulted in 

decreased starch granular swelling corresponding to 

decreased peak and final viscosities. The interaction of 

fibres with hard or soft wheat flours had similar trends, 

but the extent of change was different in the two flour 

systems. The peak and lift off time of flour–fibre blends 

decreased with increasing fibre replacement, indicating 

that cellulose fibre has the ability to enhance the kinetics 

of gluten aggregation. Further research is required to 

better understand the interactions between cellulose 

fibre and the major components of dough, including a 

better understanding regarding why in certain situations 

hard and SWF have disproportionate responses 

to cellulose fibre replacement, and a better understanding 

of the interactions between gluten and cellulose 

fibre. 

References 

AACC International (2000a). Approved Methods of the American 

Association of Cereal Chemists, 10th edn. Method 54-21A. St Paul, 

MN: The Association. 

AACC International (2000b). Approved Methods of the American 

Association of Cereal Chemists, 10th edn. Method 56-11.02. St Paul, 

MN: The Association. 

Ang, J.F. (2001). Powdered cellulose and the development of new 

generation healthier foods. Cereal Foods World, 46, 107–111. 

Chen, W.Z. & Hoseney, R.C. (1995). Development of an objective 

method for dough stickiness. Lebensmittel-Wissenschaft and Technologie, 

28, 467–473. 

Chen, H., Rubenthaler, G.L. & Schanus, E.G. (1988). Effect of Apple 

Fiber and Cellulose on the Physical-Properties of Wheat-Flour. 


Collar, C., Santos, E. & Rosell, C.M. (2006). Significance of dietary 

fibre on the viscometric pattern of pasted and gelled flour-fibre 

blends. Cereal Chemistry, 83, 370–376. 

Eliasson, A. & Larrson, K. (1993). Cereals in Breadmaking: A 

Molecular Colloidal Approach. Pp. 68–69. reissue. USA: CRC Press. 

Gomez, M., Ronda, F., Blanco, C.A., Caballero, P.A. & Apesteguia, 

A. (2003). Effect of dietary fibre on dough rheology and bread 

quality. European Food Research and Technology, 216, 51–56. 

Guillon, F. & Champ, M. (2000). Structural and physical properties of 

dietary fibres, and consequences of processing on human physiology. 

Food Research International, 33, 233–245. 

International Association of Cereal Science and Technology (1992). 

ICC Standard Methods-Principles. ICC Standard Method No. 

114 ⁄ 1. Vienna, Austria: ICC Secretariat. 

Kohajdova, Z. & Karovicova, J. (2008). Influence of hydrocolloids on 

quality of baked goods. Acta Scientiarum Polonorum – Technologia 

Alimentaria, 7, 43–49. 

Pareyt, B. & Delcour, J.A. (2008). The role of wheat flour constituents, 

sugar, and fat in low moisture cereal based products: a review on 

sugar-snap cookies. Critical reviews in food science and nutrition, 48, 

824–839. 

Pomeranz, Y. (1977). Fiber in breadmaking – a review of recent 

studies. Bakers’ Digest, 51, 94–96, 142. 

Santos, E., Rosell, C.M. & Collar, C. (2008). Gelatinization and 

retrogradation kinetics of high-fibre wheat flour blends: a calorimetric 

approach. Cereal Chemistry, 85, 457–465. 

Seetharaman, K., Waniska, R.D. & Dexter, L. (1994). An Approach to 

Increasing Fiber Content of Wheat Tortillas. Cereal Foods World, 

39, 444–447. 

Seetharaman, K., McDonough, C.M., Waniska, R.D. & Rooney, 

L.W. (1997). Microstructure of wheat flour tortillas: effects of 

soluble and insoluble fibres. Food Science and Technology International, 

3, 181–188. 

Seguchi, M., Tabara, A., Fukawa, I. et al. (2007). Effects of size of 

cellulose granules on dough rheology, microscopy, and breadmaking 

properties. Journal of Food Science, 72, E79–E84. 

Sudha, M.L., Baskaran, V. & Leelavathi, K. (2007). Apple pomace as a 

source of dietary fibre and polyphenols and its effect on the rheological 

characteristics and cake making. Food Chemistry, 104, 686–692. 




Resistance of industrial mango peel waste to pectin degradation 

prior to by-product drying 

Suparat Sirisakulwat, 1,2 Pittaya Sruamsiri, 2 Reinhold Carle 1 & Sybille Neidhart 1 * 

1 Institute of Food Science and Biotechnology, Chair of Plant Foodstuff Technology, Hohenheim University, Garbenstrasse 25, 70599 Stuttgart, 

Germany 

2 Department of Crop Science and Natural Resources, Faculty of Agriculture, Chiang Mai University, Chiang Mai 50200, Thailand 


Summary Susceptibility of industrial mango peel waste to pectin degradation during storage at ambient conditions 

(25 °C, 63% relative humidity) for up to 5 h before by-product stabilisation by drying was explored. 

Depending on the interim storage period in the wet state, pectins were recovered from the dried peels by hotacid 

extraction. Most important, pectin degradation during the temporary storage of the wet peels was 

insignificant, as revealed by yields, composition, average molecular properties, and techno-functional 

quality. Hardly acetylated (DAc 2.5–4.5%), rapid-set high-methoxyl pectins were obtained at starchcorrected 

net yields of 14.1–15.6 g hg )1 . Irrelevant de-esterification during peel storage in the wet state was 

confirmed by overall uniform setting temperatures. Arabinogalactans, uniformly indicated by high molar 

galactose ⁄ rhamnose ratios of 13.8–16.9 mol ⁄ mol and an arabinose percentage of 9.5–14.4 mol hmol )1 of 

galactose residues, affected the galacturonide contents, intrinsic viscosities, and gel strengths throughout. 

The wet peels, derived from widespread manual peeling in mango canning, tolerated intermediate storage for 

5 h, thus facilitating by-product stabilisation on smaller scales. 

Keywords By-product drying, enzymatic degradation, fruit processing, intermediate storage, mango (Mangifera indica L.), pectin, waste 

utilisation. 

Introduction 

Owing to large quantities and pectin yields up to 21% 

on a dry weight basis (Rehman et al., 2004), mango 

(Mangifera indica L.) by-products, chiefly peel waste 

from juice production, drying, and canning, have been 

deemed a promising source not yet commercially 

exploited (Panouille´ et al., 2007). Reported pectin yields 

from mango peels greatly varied, despite insignificant 

depolymerisation of pectins and galactans in the peel 

during postharvest ripening (Sirisakulwat et al., 2008). 

Moreover, the gelling capacities of the rapid-set to ultrarapid-set 

pectins from mango peels may be more or less 

limited by a characteristic almost monodisperse fraction 

that may considerably reduce the average molecular 

weight depending on its mass percentage (Neidhart 

et al., 2009). 

Processing by-products like peels are highly perishable 

in the wet state. At present, pectins are commercially 

recovered from the dried press cake of apple (Malus 

*Correspondent: Tel: +49(0)711 459 22317; Fax: +49(0)711 459 24110; 

e-mail: Sybille.Neidhart@uni-hohenheim.de 

doi:10.1111/j.1365-2621.2010.02314.x 


domestica Borkh.) juice manufacture and the peels 

and rags of citrus juice processing, chiefly lemon [Citrus 

limon (L.) Burm. F.] and lime [Citrus aurantiifolia 

(Christm.) Swingle], because their pectin properties are 

suitable for food applications (Endress et al., 2006). 

Because grinding of apples causes severe loss of cell 

compartmentalisation, apple pomace requires drying 

within a few hours (Voragen et al., 1995), preferably 

straight after de-juicing of the mash (Endress et al., 

2006), to inactivate pectin-degrading enzymes and to 

prevent microbial decay. Due to techniques restricting 

de-juicing to the citrus endocarp, tissue damage of citrus 

peels is less serious and pectins may be produced from the 

wet residues in the vicinity of a citrus processing plant 

(Endress et al., 2006), especially when the essential 

washing of citrus peels is combined with their immediate 

blanching after juice extraction (Voragen et al., 1995). 

However, shipping, storage, and year-round pectin production 

independent of raw material availability require 

instantaneous by-product drying in each case. 

Like numerous microbial enzymes, a so far limited 

number of plant enzymes are involved in pectin degradation 

(Benen et al., 2003b). Besides depolymerisation

1648 

Stability of mango peel by-products S. Sirisakulwat et al. 

through endo- (EC 3.2.1.15) and exo-acting (EC 3.2. 

1.67) polygalacturonases (PG), pectin methyl esterases 

(PME, EC 3.1.1.11) affecting the degree of esterification 

are most crucial. Without immediate drying or at least 

blanching of citrus peels, even slight de-esterification 

because of high endogenous PME activities may cause 

calcium-sensitive pectins unsuitable as gelling agents 

(Voragen et al., 1995). Apart from some soluble forms, 

PME is mainly associated with cell wall proteins by 

ionic interactions (Benen et al., 2003a). Additionally, 

pectin homogalacturonan acetylesterases (PAE, EC 3.1. 

1.6) occur in orange and a few other plant species 

(Benen et al., 2003a). Aside from PG, pectate lyase 

(PAL, EC 4.2.2.2) and endo-rhamnogalacturonan 

hydrolase (endoRGH) have been found in plants as 

further pectin backbone depolymerases (Benen et al., 

2003b). Degradation of pectin side chains involves 

arabinanases and galactanases, such as a carrot 

a-l-arabinofuranosidase (EC 3.2.1.55) and a tomato 

exo-b(1 fi 4)-d-galactanase (De Vries & Visser, 2003). 

During fruit ripening, plant b-d-galactosidases (EC 

3.2.1.23), increasing in parallel with tissue softening, 

cause pectin depolymerisation by side-chain loss and 

enhanced pectin solubilisation (Brummell, 2006). In 

ripening mango mesocarp, b-d-galactosidase activity 

even rose to the eightfold level from the mature-green 

to the ripe stage, but PG and b(1 fi 4)-d-glucanase 

(EC 3.2.1.4) only to the 1.6- and 2.4-fold activities, 

while PME declined to 41% (Ali et al., 2004). 

Although the mango cultivars may differ in the 

properties of three PG isoforms (Prasanna et al., 

2006; Singh & Dwivedi, 2008), mango PG turned out 

to be mainly exo-acting (Ali et al., 2004). Homogalacturonan 

degradation in mango mesocarp has thus 

recently been ascribed to a softening-related PAL 

rather than to PG (Chourasia et al., 2006). Consistent 

with much lower, rather constant PME and PG 

activities in the peel relative to the mesocarp (Yanru 

et al., 1995), carbohydrate changes in mango peels 

because of postharvest ripening were mainly restricted 

to the degradation of starch and matrix glycan fragments 

(Sirisakulwat et al., 2008). 

On the other hand, the mango fruit may predominantly 

be processed at higher ambient temperature and 

humidity than the citrus fruit and apples in view of the 

climate in the respective growing areas. Therefore, 

rather adverse conditions during a temporary storage 

of wet mango peels between peeling and by-product 

drying may occur in practice. Their impact on pectin 

quality must be known with respect to minimum 

requirements for sustainable by-product stabilisation 

processes when evaluating this potential raw material 

source, primarily because mango processing represents a 

very multifaceted field with a wide range of production 

scales (Va´squez-Caicedo et al., 2004, 2007; Pott et al., 

2005). 

As regards a secure and economic supply of highquality 

mango by-products for further upgrading, this 

study therefore aimed at evaluating the proneness of 

industrial peel waste to pectin degradation during a 

temporary dwell time at ambient conditions until 

drying. The tolerable period between waste production 

(peeling) and drying was to be identified, mimicking 

different operational situations in mango processing. 

The contribution of by-product handling to the variations 

of mango pectins in yield and quality was to be 

assessed. 


Sample material and experimental design 

Industrial peel waste, originating from ‘Kaew Khiew’ 

mango fruit, was obtained from a canning factory in 

Chiang Mai, Thailand, during the mango processing 

campaign in April 2005. Nine samples were consecutively 

drawn within 4 h. For each of them, 1800 g of 

fresh peel waste was quickly collected from three manual 

peeling lines by randomly picking the material from 

various positions of the lines directly after its removal 

from the fruit. To explore the tolerable period between 

peeling and waste drying, the peels of each sample were 

exposed to the ambient conditions of the peeling room 

(25 °C; 63% relative humidity, RH) in an open bowl 

for holding times (HT25 °C) of 0 (variant MPW1), 

15 (MPW2), 30 (MPW3), 60 (MPW4), 90 (MPW5), 

120 (MPW6), 180 (MPW7), 240 (MPW8), and 300 min 

(MPW9), respectively. Subsequently, the peels of each 

sample were packed as a thin layer into a polyethylene 

(PE) pouch and frozen in a freezer, which was temporarily 

installed near the peeling lines. So far, the 

experiment was performed close to the peeling lines 

when they were in full operation. Because of limited 

drying capacity, the samples were temporarily stored at 

)20 °C until drying. Without previous external thawing, 

the frozen peel samples were dried at 80 °C for approximately 

4 h in a fluidized-bed laboratory dryer (Rapid 

dryer TG 1, Retsch, Haan, Germany) until a moisture 

content of 4.5–6.6 g hg )1 . After weighing to record the 

yield of the dried mango peel waste (MPW) 

(21.6–29.8 g hg )1 of fresh peel waste), the dried peels 

were vacuum-packed into PE pouches and transported 

by airfreight to Hohenheim University, Stuttgart, 

Germany, where they were manually crushed, vacuumpacked 

into PE pouches, and stored at ambient temperature 

(approximately 22–25 °C) in a desiccator until 

pectin extraction. By using the methodology detailed 

earlier (Sirisakulwat et al., 2008), the recovered pectins 

were characterised in terms of composition as well as 

average molecular and techno-functional properties. 

Analytical and empirical methods commonly used by 

pectin producers for quality assessment were included. 


Ripeness of the fruit processed was assessed at 

Chiang Mai University, Chiang Mai, Thailand, by 

analogy to a previous study (Sirisakulwat et al., 2008) 

on the basis of five fruits that were randomly selected 

among the fruits provided for peeling. The maturityspecific 

properties of the edible part included the pH 

value (pH 3.85 ± 0.01) and the contents of total 

soluble solids (TSS; 18.7 ± 0.07 °Brix) and titratable 

acids (TA as citric acid, pH 8.1; 0.58 ± 0.01 g hg )1 ), 

corresponding to a sugar ⁄ acid ratio (TSS ⁄ TA) of 

32.2 ± 0.7. Mesocarp firmness, measured with an 

Instron Universal Texture Analyzer type 3365 (Instron, 

Canton, MA, USA) as specific maximum load (SFKS) 

by using a Kramer shear cell (Va´squez-Caicedo et al., 

2006), ranged at 472 ± 9 N hg )1 . Compared to the 

mango fruit of a related study (Sirisakulwat et al., 

2008), the mesocarp properties came overall most close 

to those of the ‘Kaew Khiew’ (KAk) batch, which had 

previously been examined after major fruit softening 

(ripeness stage II). For this batch, a TSS ⁄ TA ratio of 

55.2 ± 0.9 was reported. Its mesocarp firmness of 

4.4 ± 0.1 N, recorded with a Warner–Bratzler shear 

cell, corresponded to a SFKS of 519 ± 19 N hg )1 

measured concomitantly. For comparison, full-ripe 

fruit (stage III) had then been characterised by a 

TSS ⁄ TA ratio of 94.1 ± 2.6 and low firmness records 

of 1.8 ± 0.04 N and 373 ± 12 N hg )1 , respectively 

(Sirisakulwat et al., 2008). Compared to these two 

batches, the fruit of this study thus had a lower sugar 

content and TSS ⁄ TA ratio, but mesocarp firmness was 

intermediate. 

Pectin extraction from dried mango peel waste 

Industrial hot-acid extraction was mimicked on a 

laboratory scale. Pectin was extracted from the dried 

raw material (DRM) at pH 1.5, applying a standard 

protocol (Sirisakulwat et al., 2008) to the dried MPW 

on the double scale throughout. Per run, 40 g of MPW 

was filled up with distilled water to a total of 800 g. The 

suspension was boiled for 15–20 min and recooled. The 

pH was adjusted to 1.5 with H2SO4 (25% w ⁄ w) prior to 

extraction (90 °C, 2.5 h) in an oil bath. Separation of the 

extract as well as precipitation in propan-2-ol (‡ 99.9% 

v ⁄ v) and drying (60 °C, 1–2 h) of the pectins were 

performed as described previously (Sirisakulwat et al., 

2008). Nine runs were carried out for each MPW 

sample. For each run, the yield of dried crude pectin was 

calculated as alcohol-insoluble substance (AIS, g hg )1 of 

DRM). Per sample, the AIS of all runs was pooled and 

ground in a centrifugal mill type ZM 1 (Retsch, Haan, 

Germany) to particle sizes

1650 


USA) with a Waters 2410 refractive index detector set to 

35 °C. An Aminex HPX-87H column (300 · 7.8 mm; 

Bio-Rad Laboratories, Hercules, CA, USA) was eluted 

at 40 °C with 0.01 n H2SO4 (0.6 mL min )1 ). Samples 

were prepared in duplicate with two injections per vial. 

Lithium lactate, dissolved in alkaline propan-2-ol, 

served as internal standard. 

The moisture content (g hg )1 of AIS) was gravimetrically 

determined in duplicate as mass loss after drying 

of 1 g of AIS in a porcelain crucible at 105 ± 2 °C ina 

hot-air cabinet for 2 h (Joint FAO ⁄ WHO Expert 

Committee on Food Additives (JECFA), 2007) and 

recooling in a desiccator until constant weight. After 

slow preincineration of the moisture-free AIS on an 

electric heating plate, the ash content (g hg )1 of AIS) 

was quantified by complete incineration (550 °C, 3 and 

16 h) in a muffle furnace with intermediate wetting of 

the recooled sample [2–4 drops of ultra-pure water and 

4–5 drops of H2O2 (30% w ⁄ w)] and drying (105 ± 2 °C, 

1 h) (Sirisakulwat et al., 2008). The mass of the white 

ash was verified after further incineration (550 °C, 1 h) 

and recooling (60 min). 

Quantification of average macromolecular pectin properties 

The intrinsic viscosity ([g], mL g )1 ) was assessed as an 

average parameter influenced by all polymers of an AIS 

sample. Based on the single-point relationship (Solomon-Ciuta 

equation), it was computed from the specific 

viscosities [gsp ¼ ðgsgoÞ=go] of different AIS solutions 

as a mean of ten records and verified by the Huggins 

plot and the Kraemer plot as detailed previously 

(Sirisakulwat et al., 2008). The AIS contents (c) of five 

solutions per sample were adjusted to the gsp range of 

0.2–1 (Morris, 1995) by appropriate dilution of a sample 

stock solution (c = 0.3–0.4 g dL )1 ) with the solvent. 

The dynamic viscosities (g = Kqt, mPa s) of AIS 

solutions (g s) and solvent (g o) were calculated from the 

flow time (t, s), measured at 20 °C for a test volume of 

2mL(2· 3 runs for to and ts, respectively) by using a 

Micro-Ostwald capillary viscosimeter (capillary constant 

K = 0.01181 mm 2 s )2 ). The densities (q, gcm )3 ) 

of AIS solutions (qs) and solvent (qo) were measured at 

20 °C with a density meter DMA 48 (Anton Paar, Graz, 

Austria). The AIS was dissolved under automatic 

shaking (16 h) before adjusting the volume of the stock 

solution to 100 mL. A 0.28 m buffer (pH 3.0 ± 0.05) of 

potassium acetate (KOAc: 4.57 g L )1 ) and lactic acid 

[HOLc (90%): 23.59 g L )1 ] served as solvent, mimicking 

the gel systems used for the gelation studies in terms of 

pH and concentration. 

Characterisation of techno-functional pectin properties 

For pectin-sucrose gels (TSS = 65 °Brix, pH 3) with 

AIS doses (cp) of 0.4 and 0.55 g hg )1 in a KOAc ⁄ HOLc 

buffer (0.25 mol kg )1 of gel), the setting temperatures 

were assayed by dynamic rheometric analyses in duplicate 

or triplicate, using a Bohlin CVO 120 HR rheometer 

(Malvern Intruments, Herrenberg, Germany) with a 

Searle cylinder device (C 25) (Sirisakulwat et al., 2008). 

The AIS, premixed with a small portion of the sucrose, 

was carefully dispersed at ambient temperature in the 

diluted buffer that resulted from 68 g of distilled water 

and 17 g of a KOAc ⁄ HOLc buffer stock solution 

[KOAc: 40.48 g L )1 ; HOLc (90%): 188.05 g L )1 ]. When 

the AIS was completely dissolved under heating in 

the presence of 1–2 drops of defoaming emulsion, the 

remaining part of the sugar was stepwise added to the 

boiling mixture. TSS was adjusted by cooking to a 

precalculated net weight of 170 g under stirring. Subsequently, 

the sample was poured into the preheated C 25 

device (90 °C). To avoid evaporation and superficial 

hardening, the free sample surface was covered with hot 

low-viscous liquid paraffin. Gel formation was monitored 

by measuring the storage (G¢) and the loss (G¢¢) 

moduli as well as the phase angle [d = arctan (G¢¢ ⁄ G¢)] 

at a fixed frequency (f = 1 Hz) and strain amplitude 

(c = 0.015) as a function of time (t, s) during cooling 

from 90 to 20 °C at a constant rate (DT = )0.999 ± 

0.002 K min )1 ). In the sol–gel transition range, the d-t 

curves were iteratively approximated by a modified 

logistic four-parameter model (Neidhart et al., 2003), 

using the NLIN procedure (Marquardt method) of the 

Statistical Analysis System (SAS) 9.1 (SAS Institute, 

Cary, NC, USA). The setting time (td¼45 , s) and 

temperature (Td¼45 , °C) were deduced from the crossover-point 

of the moduli [d(1 Hz) = 45°]. 

The gel properties were characterised by the breaking 

strengths (BS) of pectin-sucrose gels (TSS = 65 °Brix, 

pH 3) with AIS doses of 0.3 and 0.35 g hg )1 in a 

KOAc ⁄ HOLc buffer (0.28 mol kg )1 of gel) on the basis 

of an empirical tension test established for the standardisation 

of pectins and the comparative evaluation of 

pectin sources (Endress & Dilger, 1990). By using the 

Herbstreith-Pectinometer type Mark III (Herbstreith & 

Fox, Neuenbu¨ rg, Germany), the BS was recorded in 

device-specific units (Herbstreith-Pectinometer units, 

HPE). The AIS dose required for a standardised gel 

(BS = 530 HPE) was calculated by interpolation as the 

breaking capacity of the AIS (BC530HPE, ghg )1 of gel). 

By dividing the sugar content of this gel (TSS) by 

BC530HPE, the gelling capacity of the AIS (SBC530HPE, 

gg )1 of AIS) was obtained as the sugar-binding 

capacity, i.e., the amount of sugar bound by this AIS 

dose in a gel of 530 HPE. Multiplication of SBC530HPE 

with the AIS content of the DRM (dried MPW) yielded 

the corresponding gelling capacity of the DRM 

extracted (SBCDRM, ghg )1 of DRM). With each AIS 

dose, a gel was prepared by cooking the AIS-containing 

mixture on the basis of 216.0 g of sugar and 147.0 g of a 

0.65 m KOAc ⁄ HOLc buffer [KOAc: 10.5 g L )1 ; HOLc 


(90%): 54.25 g L )1 ] to a final net weight of 338 g and 

subsequent hot-filling of 100 ± 1 g into each of three 

test cups (Sirisakulwat et al., 2008). After 2 h of gel 

curing in a water-bath (20 °C), BS was measured for 

each test cup. TSS and pH of all gel samples were 

experimentally verified. For their normalisation to 

constant AUA levels, BC530HPE and SBC530HPE were 

related to the AUA titr content of the AIS and expressed 

as BC530HPE (AUAtitr) [in g of AUA hg )1 of gel] and 

SBC530HPE (AUAtitr) [in g of sugar g )1 of AUA], 

respectively. 


Impact of peel storage on pectin yields 

Irrespective of the exposure of the wet peels to the 

ambient conditions of the peeling room (25 °C; 63% 

RH) for up to 5 h, the crude pectin yields (AIS) from 

hot-acid extraction of the dried by-products amounted 

to 14.7–16.2 g hg )1 of DRM, with the median ranging 

at 15.7 g hg )1 (Table 1). Because of comparatively low 

starch contents of 2.2–5.6 g hg )1 of AIS, the net pectin 

yields (AISS-corr) were 14.1–15.6 g hg )1 of DRM. The 

chronological order of peel sampling is exemplarily 

shown in Fig. 1. As revealed by respective sorting of the 

samples, the highest crude and net pectin yields were 

uniformly recorded for the first (MPW9) and the last 

sample (MPW3), whereas the largest differences were 

observed for the three samples collected within the first 

64 min (MPW9, MPW1, and MPW8). Moreover, the 

peel sample dried without previous dwell period at 

ambient temperature (MPW1) was among those showing 

the lowest net pectin yield, whereas MPW9 peaked 

despite the interim storage of the wet peels for 5 h. Thus, 

the temporary exposure of the residues to the ambient 

conditions of the peeling room did not affect the 

achievable pectin yields. In fact, it was the natural 

heterogeneity of the peel waste that was reflected by the 

observed yield variations. 

Industrial hot-acid extraction in pectin recovery from 

established sources (MacDougall & Ring, 2004) was 

mimicked by the procedure applied on the laboratory 

scale. With a median of 15.2 g hg )1 of DRM, the net 

pectin yields of the samples roughly matched the 

average reported for hot-acid extraction of dried mango 

peels (Neidhart et al., 2009), slightly exceeding the 

pectin amounts extracted with the same method from 

peels of the same cultivar (‘Kaew Khiew’, KAk; 12.9– 

13.1 g hg )1 ) when the impact of cultivar and fruit 

ripeness was studied (Sirisakulwat et al., 2008). Findings 

detailed elsewhere for peels of related and other mango 

cultivars (Berardini et al., 2005a,b; Koubala et al., 2008) 

were concordant with the median net pectin yield 

Table 1 Yields of alcohol-insoluble substance (AIS) extracted from dried peel waste (MPW) that accrued from the industrial canning of 

‘Kaew Khiew’ mango fruit after exposure of the wet peels to the ambient conditions of the peeling room (25 °C, 63% relative humidity) for 

different holding times (HT 25 °C) up to 5 h prior to waste stabilisation 

Variant MPW1 MPW2 MPW3 MPW4 MPW5 MPW6 MPW7 MPW8 MPW9 

HT25 °C (min) 0 15 30 60 90 120 180 240 300 

AIS a (g hg )1 DRM) 14.9 ± 0.1 15.6 ± 0.2 16.2 ± 0.2 15.7 ± 0.2 15.8 ± 0.2 15.5 ± 0.3 15.7 ± 0.1 14.7 ± 0.3 16.2 ± 0.1 

Starch b (g hg )1 AIS) 5.6 ± 0.1 4.0 ± 0.2 3.7 ± 0.02 3.0 ± 0.1 4.2 ± 0.2 2.2 ± 0.03 3.6 ± 0.2 3.9 ± 0.2 3.8 ± 0.002 

AIS S-corr c (g hg )1 DRM) 14.1 ± 0.1 14.9 ± 0.2 15.6 ± 0.2 15.3 ± 0.2 15.1 ± 0.2 15.2 ± 0.3 15.2 ± 0.1 14.1 ± 0.3 15.6 ± 0.1 

MPW, mango peel waste. 

a Total yield of extractable polymers as AIS in g hg )1 of dried raw material (DRM, here: MPW). 

b Starch content of the AIS according to enzymatic quantification. 

c Net pectin yield (AISS-corr) as starch-corrected AIS in g hg )1 of DRM (MPW). 

[η] (mL g –1 ) 

700 

650 

600 

550 

500 

450 

400 

350 

300 

250 

MPW 9 

Stability of mango peel by-products S. Sirisakulwat et al. 1651 

MPW 1 

MPW 8 

MPW 7 

MPW 6 

MPW 2 

MPW 5 

0 60 120 180 240 

5 

300 

MPW Sampling relative to the first sampling point (min) 

Intrinsic viscosity Gal/Rha AUA c/Rha AUA titr/Rha 

Figure 1 Variability of peel quality throughout by-product sampling 

for variants MPW1-MPW9: Intrinsic viscosities ([g]) as well as molar 

ratios of galactose to rhamnose (Gal ⁄ Rha) and colorimetric or 

titrimetric anhydrogalacturonic acid to rhamnose (AUAc ⁄ Rha, 

AUA titr ⁄ Rha) of the alcohol-insoluble substance extracted from 

dried peel waste (MPW) that accrued from the industrial canning of 

‘Kaew Khiew’ mango fruit after exposure of the wet peels to the 

ambient conditions of the peeling room (25 °C, 63% relative humidity) 

for different holding times (variants MPW1-MPW9; cf. Table 1) 

prior to waste stabilisation. For each variant, the time of sampling is 

expressed on the abscissa relative to that of variant MPW9, which 

was collected first. 


MPW 4 

MPW 3 

40 

35 

30 

25 

20 

15 

10 

Sugar residue/Rha (mol mol –1 )

1652 


mentioned above. At last, the mango peels were quite 

similar to apple pomace in terms of AIS yields (Endress 

et al., 2006). 

However, in wet mango peels, the pectic substances 

obviously degraded less readily compared to wet apple 

pomace, although the mango fruit may predominantly 

be processed at higher ambient temperature and humidity. 

Unlike apple pomace, MPW usually accrues with far 

less tissue disintegration. In manual peeling, which is 

still usual practice in many canning and drying factories, 

thick peel strips accumulate. Therefore, cell de-compartmentalisation 

is restricted to the thin cutting edges 

and the mesocarp-facing cut surface. Consistently, cold 

storage (2 ± 1 °C) of fresh and blanched MPW (moisture 

content: 85–89%) for 3 and 6 days, respectively, 

was considered tolerable without loss of pectin yield 

(Sudhakar & Maini, 2000), but pectin quality was not 

evaluated in this context. 

Impact of peel storage on pectin structure 

Like the yield, also the overall composition of the 

pectins was apparently unaffected by the exposure of 

the wet mango peels to the ambient conditions of the 

peeling room for up to 5 h (Table 2). With 4.8 ± 

0.2 g hg )1 of AIS, a low starch content within the limits 

observed for the nine AIS samples (Table 1) was 

previously reported for the pectin, which had been 

extracted in the same way (pH 1.5) from the peels of an 

experimental full-ripe fruit batch of the same cultivar 

(specified as cv. ‘Kaew Khiew’ at ripeness stage III; 

Sirisakulwat et al., 2008). Similarities between this 

sample and the nine pectins from industrial peel waste 

were furthermore observed in terms of the total neutral 

sugars (ANS, 35.9 ± 0.1 vs. 31–36 g hg )1 ) as well as the 

titrimetrically quantified anhydrogalacturonic acid 

(AUA titr, 34.7 ± 0.1 vs. 38–42 g hg )1 ) and methylester 

contents (MeOHtitr, 4.38 ± 0.02 vs. 4.7–5.2 g hg )1 ). 

The slightly higher values of the latter nine samples 

(Table 2) might largely be ascribed to improved purity, 

which became evident in lower ash contents (16.2 ± 0.2 

vs. 8–10 g hg )1 ) at comparable moisture levels (4.3 ± 

0.1 vs. 4.9–5.7 g hg )1 ). However, it should be noted 

that, unlike the previous study (Sirisakulwat et al., 

2008), mass balances of 90–93% [total (2), Table 2], 

resulting from the total neutral sugars, both titrimetrically 

analysed parameters, ash, and moisture, indicated 

slight underestimation of the composition for the nine 

AIS samples. Proteins reported for acid-extracted 

mango peel pectins might contribute further 1.1–3.3% 

(Kratchanova et al., 1991). Whereas the AUAtitr and 

MeOHtitr contents were expected to be confirmed by 

further methods applied (Sirisakulwat et al., 2008), the 

colorimetric anhydrogalacturonic acid contents of 

the nine samples (AUAc, 49–58 g hg )1 ) systematically 

exceeded the titrimetric values (AUAtitr) by approximately 

12 g hg )1 on average, the methoxyl contents 

resulting from methanol analysis by HPLC (MeOH, 

Table 2 Composition of the alcohol-insoluble substance (AIS) extracted from dried peel waste (MPW) that accrued from the industrial canning of 

‘Kaew Khiew’ mango fruit after exposure of the wet peels to the ambient conditions of the peeling room (25 °C, 63 % relative humidity) for 

different holding times (HT 25 °C) up to 5 h prior to waste stabilisation: Mass balance a 


HT 25 °C (min) 0 15 30 60 90 120 180 240 300 

Moisture 5.7 ± 0.1 4.9 ± 0.1 5.1 ± 0.1 5.3 ± 0.03 5.5 ± 0.01 5.5 ± 0.04 5.3 ± 0.1 5.5 ± 0.2 5.2 ± 0.1 

Ash 9.0 ± 0.1 8.0 ± 0.1 8.4 ± 0.004 8.2 ± 0.002 8.4 ± 0.001 8.6 ± 0.03 8.7 ± 0.1 8.9 ± 0.02 10.1 ± 0.1 

ANS b 

35.7 ± 0.5 33.4 ± 0.6 32.9 ± 0.2 32.2 ± 0.2 33.8 ± 0.2 31.1 ± 0.5 33.3 ± 0.3 32.3 ± 0.1 31.7 ± 1.1 

AUAc 48.5 ± 1.2 53.1 ± 0.7 50.7 ± 0.7 55.2 ± 1.2 49.9 ± 0.8 54.9 ± 0.8 50.3 ± 1.0 58.0 ± 0.5 51.0 ± 1.2 

MeOH c 

6.09 ± 0.23 7.27 ± 0.10 7.42 ± 0.11 7.30 ± 0.10 6.80 ± 0.10 7.16 ± 0.08 6.40 ± 0.24 6.74 ± 0.03 6.76 ± 0.08 

Ac c 

0.55 0.71 0.78 0.72 0.76 0.67 0.69 0.50 0.63 

Total (1) 105.4 107.3 105.2 108.9 105.1 107.9 104.7 112.0 105.4 

AUAtitr 38.2 ± 0.9 41.2 ± 0.2 40.2 ± 0.3 42.2 ± 0.6 37.5 ± 0.2 40.8 ± 0.1 40.5 ± 0.4 40.4 ± 0.9 38.9 ± 0.5 

MeOHtitr 4.69 ± 0.14 5.05 ± 0.03 4.93 ± 0.05 5.23 ± 0.10 4.68 ± 0.04 5.00 ± 0.02 5.06 ± 0.07 4.89 ± 0.16 4.82 ± 0.04 

Total (2) 93.1 92.5 91.5 93.1 89.9 91.0 92.9 92.0 90.7 

ANS, total content of neutral sugars, calculated as their anhydro forms; AUA c, colorimetric anhydrogalacturonic acid content; MeOH, methoxyl content 

(as methanol); Ac, acetyl content (as acetic acid); AUAtitr, titrimetric anhydrogalacturonic acid content; MeOHtitr, titrimetric methylester content; Total 

(1), mass balance based on the contents of moisture, ash, ANS, AUAc, MeOH, and Ac; Total (2), mass balance based on the contents of moisture, ash, 

ANS, AUA titr, and MeOH titr. 

a Mean content ± standard error (SE) in g hg )1 of AIS (Ac: SE £ 0.01 g hg )1 ). 

b Including the glucose content of starch (cf. Table 1) and ribose (£ 0.24 g hg )1 ). 

c Quantified by HPLC. 


Table 3 Esterification indices a of the alcohol-insoluble substance (AIS) extracted from dried peel waste (MPW) that accrued from the 

industrial canning of ‘Kaew Khiew’ mango fruit after exposure of the wet peels to the ambient conditions of the peeling room (25 °C, 63% 

relative humidity) for different holding times (HT 25 °C) up to 5 h prior to waste stabilisation 


HT25 °C (min) 0 15 30 60 90 120 180 240 300 

DE (%) 69.7 ± 0.5 69.7 ± 0.1 69.6 ± 0.2 70.4 ± 0.4 70.8 ± 0.1 69.6 ± 0.1 71.0 ± 0.3 68.8 ± 0.7 70.4 ± 0.4 

DMe (%) 69.0 ± 3.0 75.2 ± 1.4 80.5 ± 1.6 72.7 ± 1.9 74.9 ± 1.6 71.7 ± 1.3 69.9 ± 3.0 63.9 ± 0.7 72.8 ± 1.9 

DAc (%) 3.3 ± 0.1 3.9 ± 0.1 4.5 ± 0.1 3.8 ± 0.1 4.5 ± 0.1 3.6 ± 0.1 4.1 ± 0.1 2.5 ± 0.02 3.6 ± 0.1 

DE, degree of esterification (as molar MeOHtitr ⁄ AUAtitr ratio, cf. Table 2); DMe, degree of methylation (as molar MeOH ⁄ AUAc ratio, cf. Table 2); DAc, 

degree of acetylation (as molar Ac ⁄ AUAc ratio, cf. Table 2). 

a Mean ± standard error. 

6.1–7.4 g hg )1 ) surpassed the titrimetric methylester 

contents (MeOHtitr) by approximately 2 g hg )1 

(Table 2). As the respective mass balance averaged out 

at 106.8% [total (1), Table 2], especially the colorimetric 

anhydrogalacturonic acid contents were obviously overestimated 

throughout this study, most notably in case of 

sample MPW8. Despite the deficiencies of the associated 

mass balance [total (2)] in this study, the precision of the 

titrimetric results (AUAtitr and MeOHtitr) of the nine 

samples overall appeared to be better than their AUAc 

and MeOH contents. 

Because the titrimetrically quantified DE of the 

galacturonans closely varied around a median of 

69.7% (Table 3), de-esterification by PME during peel 

storage at 25 °C was irrelevant. A DE of 70% was 

uniformly observed when the wet peels had been dried 

without a dwell time at 25 °C (MPW1) and after storage 

for 5 h at this temperature (MPW9). Despite the 

problems faced in this study with the colorimetric 

AUAc and the chromatographic MeOH analysis, the 

findings in terms of DE were confirmed by the DMe 

displaying a median of 72.7% (Table 3). Evidently, the 

two extreme DMe levels of 64% (MPW8) and 81% 

(MPW3) represented outliers ascribed to overestimated 

maximum contents of AUAc and MeOH, respectively, 

but were not causally related to the storage of the wet 

peels at 25 °C. Likewise, the DAc was unaffected by the 

storage of the wet peels, as shown by its slight variation 

around a median of 3.8% (Table 3). In conformity with 

the nine AIS samples from industrial peel waste, the 

‘Kaew Khiew’ peel pectin mentioned earlier for comparison 

was characterised by a DE of 71.8 ± 0.1%, a 

DMe of 77.8 ± 2.0%, and a DAc of 3.6 ± 0.1% 

(Sirisakulwat et al., 2008). 

As detailed above, all nine AIS samples revealed 

almost equal contents of anhydrogalacturonic acid and 

total neutral sugars (Table 2). Among the latter, galactose 

(Gal) residues were by far prevailing (Table 4) with 

a median of 145 mmol hg )1 of AIS S-corr, corresponding 

to 23–25 g hg )1 of AISS-corr and 59–66 mol hmol )1 of 

AUAtitr. Although the galactose contents seemed to 

be maximum and minimum within that range after 

Table 4 Molar sugar composition a of the alcohol-insoluble substance on a starch-free basis (AISS-corr) after recovery from dried peel waste 

(MPW) that accrued from the industrial canning of ‘Kaew Khiew’ mango fruit after exposure of the wet peels to the ambient conditions of the 

peeling room (25 °C, 63% relative humidity) for different holding times (HT 25 °C) up to 5 h prior to waste stabilisation 


HT 25 °C (min) 0 15 30 60 90 120 180 240 300 

AUAc 292 ± 7 314 ± 7 299 ± 7 323 ± 9 296 ± 8 319 ± 9 296 ± 7 343 ± 11 301 ± 8 

AUAtitr 230 ± 6 244 ± 4 237 ± 5 247 ± 5 223 ± 5 237 ± 6 238 ± 3 239 ± 9 230 ± 4 

Gal 151 ± 3 146 ± 5 145 ± 3 145 ± 3 142 ± 3 144 ± 5 145 ± 2 144 ± 5 139 ± 7 

Ara 16 ± 0 15 ± 0 17 ± 1 16 ± 0 21 ± 1 14 ± 0 19 ± 1 14 ± 0 17 ± 1 

Glc b 

19 ± 2 15 ± 1 14 ± 1 14 ± 1 17 ± 1 12 ± 1 15 ± 1 15 ± 1 13 ± 2 

Rha 8.9 ± 0.1 10.6 ± 0.6 10.2 ± 0.3 9.9 ± 0.2 9.9 ± 0.2 9.8 ± 0.3 10.0 ± 0.3 9.0 ± 0.3 8.8 ± 0.7 

Man 1.8 ± 0.1 2.0 ± 0.04 2.1 ± 0.1 2.2 ± 0.1 2.4 ± 0.1 2.3 ± 0.1 2.2 ± 0.03 1.8 ± 0.1 1.9 ± 0.04 

Xyl 1.7 ± 0.03 2.0 ± 0.03 2.2 ± 0.04 1.8 ± 0.1 2.2 ± 0.1 2.0 ± 0.1 2.0 ± 0.1 1.5 ± 0.1 1.6 ± 0.03 

Fuc 0.4 ± 0.03 0.4 ± 0.01 0.4 ± 0.02 0.4 ± 0.01 0.5 ± 0.01 0.4 ± 0.01 0.5 ± 0.03 0.4 ± 0.01 0.4 ± 0.03 

AUA c, galacturonic acid (colorimetric); AUA titr, galacturonic acid (titrimetric); Gal, galactose; Ara, arabinose; Glc, glucose; Rha, rhamnose; Man, 

mannose; Xyl, xylose; Fuc, fucose. 

a Mean ± standard error in mmol hg )1 of AISS-corr. 

b Without glucose included in starch. 



1654 


exposure of the wet peels to 25 °C for 0 (MPW1) and 

5 h (MPW9), respectively, constant levels were observed 

for the other seven AIS samples. On the whole, the nine 

pectins revealed, regardless of the storage of the wet peel 

waste, an almost identical neutral sugar composition 

(Table 4) as the peel pectin of the experimental ‘Kaew 

Khiew’ fruit batch cited earlier for comparison did 

(Sirisakulwat et al., 2008). Their arabinose (Ara) contents 

slightly varied around a median of 16 mmol hg )1 

of AISS-corr (i.e. 2.1 g hg )1 ), followed by median contents 

of 15, 9.9, 2.1, and 2.0 mmol hg )1 of AISS-corr 

for glucose (Glc) not contained in starch, rhamnose 

(Rha), mannose (Man), and xylose (Xyl), respectively 

(Table 4). Strongest variation among the nine samples 

was observed for arabinose and the nonstarch glucose 

contents. However, no impact of wet peel storage was 

noticed. 

As shown previously (Neidhart et al., 2009), galactose 

contents of mango peel pectins may be considerably high, 

but they seemed to be specific to the cultivar and were 

neither affected by fruit ripeness nor by the extraction 

mode. Thus, even in the presence of minor natural starch 

contents, the galacturonide content of mango peel AIS 

may markedly be limited. This feature also applied to the 

nine AIS samples of this study. Because their galacturonic 

acid contents (GalUAtitr, Mr = 194.1 g mol )1 ) 

on an ash- and moisture-free base (AISp) after pectin 

purification in acidic alcohol amounted to 56–59 g hg )1 

of AIS p (Table 5), the nine samples did not meet the 

JECFA minimum requirement as to the galacturonide 

content (65 g hg )1 of AISp; Joint FAO ⁄ WHO Expert 

Committee on Food Additives (JECFA), 2007). However, 

the portions of homogalacturonans (HG) and rhamnogalacturonans 

(RG-I) were obviously not affected by 

the wet peel storage, because largely constant molar 

AUAtitr ⁄ Rha ratios of 23–26 occurred throughout, as 

indicated by the respective sugar amounts displayed 

in Table 4. Also the backbone indices (Table 5), which 

were estimated from the molar AUAc and rhamnose 

contents on the basis of a simplified backbone model 

consisting of HG and RG-I (Sirisakulwat et al., 2008), 

did not point to any significant backbone degradation 

due to the wet peel storage, because largely constant 

maximum percentages of the pectin backbone resulted 

for homogalactacuronans (BBP HG, 93.4–94.9%) and 

rhamnogalacturonans type I (BBPRG-I, 5.1–6.6%). If all 

galactose residues were ascribed to galactans with one 

chain branching off from each rhamnose unit of rhamnogalacturonans 

type I, long and ⁄ or highly ramified 

galactan side chains of 14–17 residues may be assumed 

according to the molar Gal ⁄ Rha ratios (Table 5). This 

striking peculiarity and the AUA titr ⁄ Rha ratio, observed 

here for AIS from industrial peel waste, confirmed 

the former report (Sirisakulwat et al., 2008) for the 

Table 5 Structural indices a of the alcohol-insoluble substance (AIS) extracted from dried peel waste (MPW) that accrued from the industrial 

canning of ‘Kaew Khiew’ mango fruit after exposure of the wet peels to the ambient conditions of the peeling room (25 °C, 63 % relative humidity) 

for different holding times (HT 25 °C) up to 5 h prior to waste stabilisation 


HT25 °C (min) 0 15 30 60 90 120 180 240 300 

GalUAtitr (g hg )1 of AISp) 

Molecular parameters 

55.5 ± 1.3 58.2 ± 0.3 56.7 ± 0.4 59.0 ± 0.8 58.5 ± 0.3 58.5 ± 0.2 59.2 ± 0.6 58.8 ± 1.4 58.1 ± 0.8 

[g] (mL g )1 ) 

Side chain indices 

282 ± 2 317 ± 2 341 ± 2 329 ± 2 334 ± 2 308 ± 2 325 ± 3 276 ± 2 271 ± 2 

Ara ⁄ Rha (mol mol )1 ) 1.8 ± 0.02 1.4 ± 0.1 1.7 ± 0.1 1.6 ± 0.04 2.1 ± 0.04 1.5 ± 0.03 1.9 ± 0.1 1.5 ± 0.03 1.9 ± 0.2 

Gal ⁄ Rha (mol mol )1 ) 16.9 ± 0.3 13.8 ± 0.8 14.2 ± 0.4 14.6 ± 0.1 14.2 ± 0.1 14.7 ± 0.4 14.5 ± 0.4 16.0 ± 0.3 15.7 ± 1.4 

Ara ⁄ Gal (mol hmol )1 ) 10.7 ± 0.2 10.4 ± 0.3 11.7 ± 0.5 11.1 ± 0.3 14.4 ± 0.3 10.0 ± 0.3 12.9 ± 0.3 9.5 ± 0.1 12.2 ± 0.7 

Backbone indices (of AISS-corr) 

HGmax b (mmol hg )1 ) 283 ± 7 304 ± 7 288 ± 7 313 ± 9 286 ± 8 309 ± 9 286 ± 7 334 ± 11 293 ± 8 

c )1 

RG-Imax (mmol hg ) 17.9 ± 0.2 21.1 ± 1.1 20.4 ± 0.7 19.8 ± 0.3 19.9 ± 0.4 19.6 ± 0.5 20.1 ± 0.5 17.9 ± 0.7 17.6 ± 1.3 

d 

BBPHG (%) 94.1 93.5 93.4 94.1 93.5 94.0 93.4 94.9 94.3 

BBPRG-I e (%) 5.9 6.5 6.6 5.9 6.5 6.0 6.6 5.1 5.7 

GalUAtitr, titrimetric galacturonic acid content (g hg )1 of AISp, i.e., of nondestarched AIS after acidic purification before titration); [g], intrinsic viscosity 

(AIS dissolved in 0.28 M KOAc ⁄ HOLc buffer, pH 3.0); Ara, arabinose (mmol hg )1 of AIS); Rha, rhamnose (mmol hg )1 of AIS); Gal, galactose (mmol hg )1 

of AIS). 

a Mean ± standard error. 

b 

Estimated maximum units of the total homogalacturonan backbone (HG) calculated from the AUAc and Rha contents in mmol hg )1 of AISS-corr (HGmax = AUAc-Rha). c )1 

Estimated maximum units of the total rhamnogalacturonan-I backbone (RG-I) calculated from the Rha content in mmol hg of AISS-corr 

(RG-Imax = 2 Rha). 

d 

Maximum homogalacturonan (HG) percentage of the pectin backbone [BBPHG = 100(AUAc ) Rha)/(AUAc + Rha)]. 

e 

Maximum rhamnogalacturonan-I (RG-I) percentage of the pectin backbone [BBPRG)I = 100(2ÆRha)/(AUAc + Rha)]. 


analogously extracted peel AIS from the experimental 

‘Kaew Khiew’ fruit batch. But galactan degradation 

during storage of the wet peels by b-d-galactosidase 

activities was obviously negligible under the conditions 

applied. Because of the acid-lability of arabinofuranosyl 

linkages, arabinose residues are more or less split off 

during hot-acid extraction, depending on the pH conditions 

applied (Sirisakulwat et al., 2008; Neidhart et al., 

2009). However, like the Gal ⁄ Rha ratios, also molar 

Ara ⁄ Rha ratios of 1.4–2.1 around a median of 1.7 mol ⁄ - 

mol and Ara ⁄ Gal ratios of 10–14 mol hmol )1 (Table 5) 

pointed to insignificant side-chain degradation during the 

storage of the wet peels. 

The clearest evidence for the preservation of pectin 

backbones and side chains during the storage of the wet 

peels at 25 °C was finally provided by the intrinsic 

viscosities ([g]) of the samples (Table 5). Within a range 

of 271–341 mL g )1 ,[g] closely varied around a median 

of 317 mL g )1 , indicating good consistency among the 

nine pectins in terms of their average molecular dimensions 

(Morris, 1995). The lowest intrinsic viscosities 

(271–282 mL g )1 ) were observed for the AIS group 

from the three peel waste samples collected within the 

first 64 min of sampling (MPW1, MPW8, MPW9, 

Fig. 1), including the extreme situations, i.e. immediate 

stabilisation and exposure to 25 °C for 5 h prior to 

drying, respectively. Thus, the slight differences among 

the nine pectins were definitely ascribed to the inherent 

heterogeneity of the fruits and the reproducibility of 

pectin recovery (Sirisakulwat et al., 2008) rather than to 

the storage of the wet peels at ambient temperature. The 

rhamnose contents were slightly lower in the AIS from 

the peel samples MPW1, MPW8, and MPW9 (approximately 

9 mmol hg )1 of AISS-corr, Table 4) compared to 

the remaining pectins. Owing to the minor variation of 

galactose contents, these three pectins were further 

characterised by maximum Gal ⁄ Rha ratios (Fig. 1). 

Thus, their arabinogalactan side chains were possibly 

more extended than in the other cases, as equally 

indicated by their molar (Ara+Gal) ⁄ Rha ratios of 

18–19 relative to 15–16 for the other pectins (Table 5). 

The intrinsic viscosity would consistently decline along 

with a relative increase in branched neutral sugar side 

chains due to overall more condensed coil dimensions 

(Morris, 1995). In the present case, this seemed to be 

partly compensated by more expanded homogalacturonan 

parts, because especially the AIS from MPW8 and 

MPW9 besides MPW1 also displayed elevated molar 

AUA ⁄ Rha ratios (both on the basis of the AUAtitr and 

the AUAc contents; Fig. 1). Nevertheless, their lower 

intrinsic viscosities were most probably owing to larger 

arabinogalactan parts, because the MPW1 AIS came 

quite close to the MPW4 and MPW6 samples in terms 

of the AUA ⁄ Rha ratios despite the significantly higher 

intrinsic viscosities of the latter group ([g] ‡ 308 mL g )1 ; 

Fig. 1). Both the maximum and minimum [g] levels in 


Table 5 clearly exceeded the intrinsic viscosity of the 

reported ‘Kaew Khiew’ peel pectin (234 mL g )1 ), which 

had displayed a large arabinogalactan fraction [(Ara + 

Gal) ⁄ Rha = 19.2] similar to the MPW1 pectin, but only 

molar AUA ⁄ Rha ratios of 23 and 24 based on AUA c 

and AUAtitr, respectively (Sirisakulwat et al., 2008). As 

previously shown for this pectin by high-performance 

size exclusion chromatography, its comparatively low 

intrinsic viscosity resulted from the co-existence of two 

distinct major fractions. Apart from a high-molecular 

weight fraction also found in commercial apple pectin, 

an additional, almost monodisperse fraction with a peak 

molecular weight of 17 000 relative to dextran considerably 

reduced the average molecular weight of this 

pectin (Sirisakulwat et al., 2008). Depending on the 

amounts of this characteristic fraction, the gelling and 

thickening properties of mango peel pectins may be 

more or less limited, as demonstrated by numerous 

samples (Neidhart et al., 2009). 

Impact of peel storage on techno-functional pectin 

properties 

From the similarities in the average composition 

(Tables 2-4) and the intrinsic viscosity reflecting the mean 

molecular weight (Table 5), no differences in the technofunctional 

properties were expected among the nine 

pectins. On the other hand, weak activities of the 

respective esterases and depolymerising hydrolyses and 

lyases may nevertheless cause enhanced variation among 

the molecules in terms of molecular size, side-chain 

lengths, and the degrees of methylation and acetylation. 

Likewise, the intramolecular distribution of methylesterified 

units might be affected, especially because of 

the blockwise mode of action of plant PME (Benen et al., 

2003a). Owing to such differences in their fine structure 

(Guillotin et al., 2007), pectins with similar average 

characteristics may display different gelling behaviour 

(Neidhart et al., 2003). 

Consistent with their low galacturonide contents 

(Table 2) and intrinsic viscosities (Table 5), the nine 

samples produced only soft gels. Low breaking strengths 

of 369–497 HPE resulted even at an AIS dose of 

0.3 g hg )1 of gel (BS 0.3%; Table 6) and 0.35 g hg )1 were 

necessary to achieve BSs of 452–594 HPE (BS0.35%). As 

displayed by the AIS amounts required for 100 g of a 

standard gel with a BS of 530 HPE, the nine samples 

were quite uniform in terms of their breaking capacities 

(BC530HPE; Table 6). With a BC530HPE range of 0.33– 

0.4 g hg )1 of gel, the breaking capacity of the ‘Kaew 

Khiew’ peel AIS reported as a reference was confirmed; 

but compared to a commercial, unstandardised apple 

pectin that was also included in the previous study 

(Sirisakulwat et al., 2008), 1.9–2.4 times more AIS was 

needed to produce this standard gel. Consequently, the 

sugar amount bound in a sugar-acid standard gel by 1 g 


1656 


Table 6 Gelling capacity of the alcohol-insoluble substance (AIS) extracted from dried peel waste (MPW) that accrued from the industrial 

canning of ‘Kaew Khiew’ mango fruit after exposure of the wet peels to the ambient conditions of the peeling room (25 °C, 63 % relative 

humidity) for different holding times (HT 25 °C) up to 5 h prior to waste stabilisation 


HT25 °C (min) 0 15 30 60 90 120 180 240 300 

Gelling capacity of the AIS and the dried raw material (DRM), respectively 

BS0.3% (HPE) 458 ± 3 a 

478 ± 2 369 ± 3 399 ± 5 463 ± 9 497 ± 8 463 ± 6 416 ± 4 394 ± 1 

BS0.35% (HPE) 586 ± 5 a 

564 ± 13 452 ± 8 594 ± 5 570 ± 6 563 ± 4 587 ± 14 528 ± 6 509 ± 2 

BC530HPE (g hg )1 ) 0.33 0.33 0.40 0.33 0.33 0.33 0.33 0.35 0.36 

SBC530HPE (g g )1 of AIS) 198 197 164 195 196 200 199 185 181 

SBCDRM (g hg )1 ) 2951 3062 2644 3066 3092 3108 3124 2717 2935 

Normalised gelling capacity of the AIS relative to the AUAtitr content of the standardised gel with 530 HPE 

BC530HPE(AUAtitr) (ghg )1 ) 0.13 0.14 0.16 0.14 0.12 0.13 0.13 0.14 0.14 

SBC530HPE(AUAtitr) (gg )1 of AUA) 519 478 407 462 523 490 491 459 465 

a Mean ± standard error. BS0.3% and BS 0.35%, breaking strength of gels with AIS doses of 0.3 and 0.35% (w ⁄ w), respectively, in Herbstreith-Pectinometer 

units (HPE); BC 530HPE, breaking capacity of the AIS as gram of AIS required for 100 g of a gel with 530 HPE; SBC 530HPE, sugar-binding capacity of 

the AIS as gram of sugar bound per gram of AIS in a gel of 530 HPE; SBCDRM, gelling units (GU) of the dried raw material (DRM; here: MPW) as its 

sugar-binding capacity in gram of sugar bound by 100 g of DRM in a gel of 530 HPE; BC530HPE(AUAtitr), normalised breaking capacity of the AIS 

(g of AUA hg )1 of a gel with 530 HPE); SBC 530HPE(AUA titr), normalised sugar-binding capacity of the AIS (g of sugar g )1 of AUA for a gel of 530 HPE). 

of this apple pectin exceeded the sugar-binding capacities 

(SBC530HPE) of the nine mango AIS samples (164– 

200 g g )1 ; Table 6) by the same factors. Because of their 

uniform AIS yields (AISS-corr; Table 1) that slightly 

exceeded that of this former ‘Kaew Khiew’ peel sample 

(13.1 ± 0.2 g hg )1 ; Sirisakulwat et al., 2008), the dried 

peels MPW1-MPW9 had the capacity to bind more sugar 

in a standard gel by the recovered AIS than the latter 

(SBCDRM; 2644–3124 vs. 2250 g hg )1 of dried peel; 

Table 6). Thus, the industrial peel waste yielded slightly 

more AIS than the cited peel sample, but of comparable 

quality. The three samples displaying minimum intrinsic 

viscosities of the extractable AIS (MPW1, MPW8, 

MPW9; Fig. 1) were among the four samples with a 

SBCDRM below the median. However, together with the 

other AIS samples from industrial peel waste, their AIS 

was almost identical to that of the cited ‘Kaew Khiew’ 

peel sample (0.14 g hg )1 ; Sirisakulwat et al., 2008) and 

even came close to this apple pectin (0.11 g hg )1 ; 

Sirisakulwat et al., 2008) in terms of the normalised 

gelling capacity [BC 530HPE(AUA titr); Table 6], when the 

breaking capacities (BC530HPE) were expressed as the 

respective AUAtitr doses required for a standard gel of 

530 HPE. Thus, the functional weakness of mango peel 

AIS shown by the non-normalised parameters in Table 6 

was neither caused by the use of industrial peel waste 

as an AIS source nor by the extended exposure of the 

wet peels to ambient temperature. Also under the given 

conditions, deficits seemed to be rather because of 

galacturonide contents, which were markedly lowered 

in favour of arabinogalactan portions, and the presence 

of a characteristic fraction that considerably reduced the 

average molecular weight, as demonstrated earlier for the 

cited AIS sample (Sirisakulwat et al., 2008) in compar- 

ison to pectins with improved techno-functionality from 

other mango peels (Neidhart et al., 2009). 

Like the gel properties, the setting behaviour of the 

AIS was also not verifiably affected by the intermediate 

exposure of the wet peels to ambient temperature. 

Despite some variation of the setting temperature that 

was more pronounced at the upper AIS dose, equally 

high setting temperatures were recorded for AIS that 

was obtained from peels after their wet storage at 25 °C 

for 0, 15, and 300 min (Fig. 2). 

T (δ 1.0 Hz = 45°) (°C) 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

c P (g hg –1 of gel): 

0.4 

0.55 

0 50 100 150 200 250 300 350 

HT 25 °C (min) 

Figure 2 Gelation of alcohol-insoluble substance (AIS) samples 

extracted from dried peel waste (MPW) that accrued from the 

industrial canning of ‘Kaew Khiew’ mango fruit after exposure of 

the wet peels to the ambient conditions of the peeling room (25 °C, 

63% relative humidity) for different holding times (HT 25 °C) upto 

5 h prior to waste stabilisation: Setting temperatures (Td=45°) of 

standard gels (64.7 ± 0.3 °Brix and pH 3.25 ± 0.02 according to 

concomitant refractometer readings and pH analyses of the cured gels; 

c p, AIS dose). Error bars represent the single standard deviations. 


Conclusion 

The findings reported for mango peel AIS from experimental 

fruit batches (Sirisakulwat et al., 2008; Neidhart 

et al., 2009) were overall confirmed by industrial 

by-products from the fruit of a further crop year. Most 

important, temporary exposure of wet mango peels to 

ambient temperature was innocuous in terms of yield, 

composition, and techno-functionality of the extractable 

AIS. The interim storage of the wet peel waste in 

the peeling room prior to drying can thus be tolerated 

even up to 5 h without notable pectin degradation. 

Limited tissue damage during manual peeling and low 

activities of degrading endogenous enzymes in the peels 

(Yanru et al., 1995) might support this storage stability. 

On a relatively large part of the peel waste surface, wax 

layers and the thickness of the cuticle initially still act as 

physical barriers. Additionally, minor antimicrobial 

effects might still emanate from polyphenolics like 

gallotannins and alk(en)ylresorcinols present in mango 

peels even at consumption ripeness of the fruit (Berardini 

et al., 2004; Engels et al., 2009; Kno¨ dler et al., 

2009). Compared to twelve other mango cultivars, peels 

of ‘Kaew’ fruit were reported to be comparatively rich 

in alk(en)ylresorcinols (1.4 g kg )1 of peel dry matter; 

Kno¨ dler et al., 2009), which play a role in plant disease 

resistance (Hassan et al., 2007). Because of the enhanced 

flexibility, the observed tolerance of intermediate 

storage is important for a future exploitation of peel 

waste from mango processing on smaller industrial 

scales, where immediate peel drying may impose considerably 

higher logistic and economic investment than 

on large scale with completely continuous processing 

and waste stabilisation. As mango residues result from 

multifaceted processing on different scales, the preconditions 

for economic by-product drying are rather 

individual. Knowledge about the susceptibility of industrial 

mango peels to pectin degradation prior to drying 

is of utmost importance for the selection of adjusted 

drying technologies and economic concepts for the 

overall processes. 

The interim storage of the wet mango peels neither 

caused perceptible variations in pectin quality nor the 

comparably inferior galacturonide contents and gelling 

capacities of the pectins, which were observed in spite of 

hot-acid extraction that mimicked industrial procedures 

for pectin recovery from established sources. Despite 

individual promising findings based on laboratory 

extraction (Rehman et al., 2004; Koubala et al., 2009; 

Neidhart et al., 2009), quality deficiencies of mango peel 

pectin, which have repeatedly been reported (Pedroza- 

Islas et al., 1994) as in the present case, require further 

studies. Research into sustainable recovery of purified 

pectins from mango peels like those used in the present 

and previous trials (Sirisakulwat et al., 2008) are 

currently under way. 


This research was funded by Deutsche Forschungsgemeinschaft, 

Bonn, Germany: Project SFB 564-E2.2. It 

is part of the Special Research Program ‘Research for 

Sustainable Land Use and Rural Development in 

Mountainous Regions of Southeast Asia’ (The Uplands 

Program). 

The authors thank Herbstreith & Fox KG, Neuenbu¨ 

rg, Germany, in particular Hans-Ulrich Endress 

and Christine Rentschler, for providing laboratory 

facilities for pectin extraction as well as Northern Food 

Co., Ltd., Chiang Mai, Thailand for supplying MPW. 

Klaus Mix and Martin Leitenberger are acknowledged 

for their technical assistance with the experiments at 

Hohenheim University. 

References 


Ali, Z.M., Chin, L.-H. & Lazan, H. (2004). A comparative study on 

wall degrading enzymes, pectin modifications and softening during 

ripening of selected tropical fruits. Plant Science, 167, 317–327. 

Benen, J.A.E., van Alebeek, G.-J.W.M., Voragen, A.G.J. & Visser, J. 

(2003a). Pectic esterases. In: Handbook of Food Enzymology (edited 

by J.R. Whitaker, A.G.J. Voragen & D.W.S. Wong). Pp. 849–856. 

New York, NY, USA: Marcel Dekker. 

Benen, J.A.E., Voragen, A.G.J. & Visser, J.. (2003b). Pectic enzymes. 

In: Handbook of Food Enzymology (edited by J.R. Whitaker, A.G.J. 

Voragen & D.W.S. Wong). Pp. 845–848. New York, NY, USA: 

Marcel Dekker. 

Berardini, N., Carle, R. & Schieber, A. (2004). Characterization of 

gallotannins and benzophenone derivatives from mango (Mangifera 

indica L. cv. ‘Tommy Atkins’) peels, pulp and kernels by highperformance 

liquid chromatography ⁄ electrospray ionization mass 

spectrometry. Rapid Communications in Mass Spectrometry, 18, 

2208–2216. 

Berardini, N., Fezer, R., Conrad, J., Beifuss, U., Carle, R. & Schieber, 

A. (2005a). Screening of mango (Mangifera indica L.) cultivars for 

their contents of flavonol O- and xanthone C-glycosides, anthocyanins, 

and pectin. Journal of Agricultural and Food Chemistry, 53, 

1563–1570. 

Berardini, N., Kno¨ dler, M., Schieber, A. & Carle, R. (2005b). 

Utilization of mango peels as a source of pectin and polyphenolics. 

Innovative Food Science and Emerging Technologies, 6, 442–452. 

Brummell, D.A. (2006). Cell wall disassembly in ripening fruit. 

Functional Plant Biology, 33, 103–119. 

Chourasia, A., Sane, V.A. & Nath, P. (2006). Differential expression of 

pectate lyase during ethylene-induced postharvest softening of 

mango (Mangifera indica var. Dashehari). Physiologia Plantarum, 

128, 546–555. 

De Vries, R.P. & Visser, J. (2003). Enzymes releasing L-arabinose and 

D-galactose from the side chains of pectin. In: Handbook of Food 

Enzymology (edited by J.R. Whitaker, A.G.J. Voragen & D.W.S. 

Wong). Pp. 867–877. New York, NY, USA: Marcel Dekker. 

Endress, H.-U. & Dilger, S. (1990). Checking pectin jelly strength with 

the Pektinometer. Food Technology International, 1, 279–282. 

Endress, H.-U., Mattes, F. & Norz, K. (2006). Pectins. In: Handbook 

of Food Science, Technology, and Engineering (edited by Y.H. Hui), 

Vol. 3. Pp. 140 ⁄ 1–140 ⁄ 35. Boca Raton, FL, USA: CRC Press, 

Taylor & Francis Group. 

Engels, C., Kno¨ dler, M., Zhao, Y.-Y., Carle, R., Ga¨ nzle, M.G. & 

Schieber, A. (2009). Antimicrobial activity of gallotannins isolated 

from mango (Mangifera indica L.) kernels. Journal of Agricultural 



1658 


Guillotin, S.E., Bakx, E.J., Boulenguer, P., Schols, H.A. & Voragen, 

A.G.J. (2007). Determination of the degree of substitution, degree of 

amidation and degree of blockiness of commercial pectins by using 

capillary electrophoresis. Food Hydrocolloids, 21(3), 444–451. 

Hassan, M.K., Dann, E.K., Irving, D.E. & Coates, L.M. (2007). 

Concentrations of constitutive alk(en)ylresorcinols in peel of commercial 

mango varieties and resistance to postharvest anthracnose. 

Physiological and Molecular Plant Pathology, 71, 158–165. 

Joint FAO ⁄ WHO Expert Committee on Food Additives (JECFA) 

(2007). Compendium of Food Additive Specifications. FAO JECFA 

Monographs 4. Rome, Italy: Food and Agriculture Organization 

of the United Nations. http://www.fao.org/docrep/010/a1447e/ 

a1447e00.htm. 

Kno¨ dler, M., Reisenhauer, K., Schieber, A. & Carle, R. (2009). 

Quantitative determination of allergenic 5-alk(en)yl-resorcinols in 

mango (Mangifera indica L.) peel, pulp, and fruit products by highperformance 

liquid chromatography. Journal of Agricultural and 


Koubala, B.B., Kansci, G., Mbome, L.I., Crépeau, M.-J., Thibault, 

J.-F. & Ralet, M.-C. (2008). Effect of extraction conditions on some 

physicochemical characteristics of pectins from ‘‘Améliorée’’ and 

‘‘Mango’’ mango peels. Food Hydrocolloids, 22(7), 1345–1351. 

Koubala, B.B., Kansci, G., Garnier, C., Mbome, I.L., Durand, S., 

Thibault, J.-F. & Ralet, M.-C. (2009). Rheological and high gelling 

properties of mango (Mangifera indica) and ambarella (Spondias 

cythera) peel pectins. International Journal of Food Science and 

Technology, 44, 1809–1817. 

Kratchanova, M., Bénémou, C. & Kratchanov, C. (1991). On the 

pectic substances of mango fruits. Carbohydrate Polymers, 15, 

271–282. 

MacDougall, A.J. & Ring, S.G. (2004). Pectic polysaccharides. In: 

Chemical and Functional Properties of Food Saccharides (edited by 

P. Tomasik). Pp. 181–195. Boca Raton, FL, USA: CRC Press. 

Morris, E.R. (1995). Polysaccharide rheology and in-mouth perception. 

In: Food Polysaccharides and Their Applications (edited by 

A.M. Stephen). Pp. 517–546. New York, NY, USA: Marcel Dekker. 

Neidhart, S., Hannak, C. & Gierschner, K. (2003). Sol-gel transition 

of high-esterified pectins and their molecular structure. In: Advances 

in Pectin and Pectinase Research (edited by A.G.J. Voragen, H.A. 

Schols & R. Visser). Pp. 431–448. Dordrecht, The Netherlands: 

Kluwer Academic Publishers. 

Neidhart, S., Sirisakulwat, S., Nagel, A., Sruamsiri, S. & Carle, R. 

(2009). Which mango processing residues are suitable for pectin 

recovery in terms of yield, molecular and techno-functional 

properties of extractable pectins? In: Pectins and Pectinases 

(edited by H.A. Schols, R.G.F. Visser & A.G.J. Voragen). Pp. 

177–195. Wageningen, The Netherlands: Wageningen Academic 

Publishers. 

Panouille´, M., Ralet, M.-C., Bonnin, E. & Thibault, J.-F. (2007). 

Recovery and reuse of trimmings and pulps from fruit and vegetable 

processing. In: Handbook of Waste Management and Co-product 

Recovery in Food Processing (edited by K. Waldron), Vol. 1. Pp. 

417–447. Cambridge, UK: Woodhead. 

Pedroza-Islas, R., Aquilar-Esperanza, E. & Vernon-Carter, E.J. 

(1994). Obtaining pectins from solids wastes derived from mango 

(Mangifera indica) processing. AIChE Symposium Series, 300, 36–41. 

Pott, I., Neidhart, S., Mu¨ hlbauer, W. & Carle, R. (2005). Quality 

improvement of non-sulphited mango slices by drying at high 

temperatures. Innovative Food Science & Emerging Technologies, 

6(4), 412–419. 

Prasanna, V., Prabha, T.N. & Tharanathan, R.N. (2006). Multiple 

forms of polygalacturonase from mango (Mangifera indica L. cv. 

Alphonso) fruit. Food Chemistry, 95, 30–36. 

Rehman, Z.U., Salariya, A.M., Habib, F. & Shah, W.H. (2004). 

Utilization of mango peels as a source of pectin. Journal of the 

Chemical Society of Pakistan, 26(1), 73–76. 

Singh, P. & Dwivedi, U.N. (2008). Purification and characterization of 

multiple forms of polygalacturonase from mango (Mangifera indica 

cv. Dashehari) fruit. Food Chemistry, 111, 345–349. 

Sirisakulwat, S., Nagel, A., Sruamsiri, P., Carle, R. & Neidhart, S. 

(2008). Yield and quality of pectins extractable from the peels of 

Thai mango cultivars depending on fruit ripeness. Journal of 

Agricultural and Food Chemistry, 56(22), 10727–10738. 

Sudhakar, D.V. & Maini, S.B. (2000). Isolation and characterization 

of mango peel pectins. Journal of Food Processing and Preservation, 

24(3), 209–227. 

Vásquez-Caicedo, A.L., Neidhart, S. & Carle, R. (2004). Postharvest 

ripening behavior of nine Thai mango cultivars and their 

suitability for industrial applications. Acta Horticulturae, 645, 

617–625. 

Vásquez-Caicedo, A.L., Heller, A., Neidhart, S. & Carle, R. (2006). 

Chromoplast morphology and b-carotene accumulation during 

postharvest ripening of mango cv. ‘Tommy Atkins’. Journal of 


Vásquez-Caicedo, A.L., Schilling, S., Carle, R. & Neidhart, S. (2007). 

Effects of thermal processing and fruit matrix on b-carotene stability 

and enzyme inactivation during transformation of mangoes into 

purée and nectar. Food Chemistry, 102(4), 1172–1186. 

Voragen, A.G.J., Pilnik, W., Thibault, J.-F., Axelos, M.A.V. & 

Renard, C.M.G.C.. (1995). Pectins. In: Food Polysaccharides and 

Their Applications (edited by A.M. Stephen). Pp. 287–339. New 

York, NY, USA: Marcel Dekker. 

Yanru, Z., Pandey, M., Prasad, N.K. & Srivastava, G.C. (1995). 

Ripening associated changes in enzymes and respiratory activities in 

three varieties of mango (Mangifera indica L.). Indian Journal of 

Plant Physiology, 38(1), 73–76. 




Comparison of the texture of fresh and preserved Agaricus bisporus 

and Boletus edulis mushrooms 

Gra_zyna Jaworska, 1 * Emilia Bernas´, 1 Adriana Biernacka 1 & Ireneusz Maciejaszek 2 

1 Department of Raw Materials and Fruit and Vegetable Processing, University of Agriculture in Krakow, 122 Balicka Street, 30-149 Krakow, 

Poland 

2 Department of Refrigeration and Food Concentrate, University of Agriculture in Krakow, 122 Balicka Street, 30-149 Krakow, Poland 


Summary This work compares changes during the production process and storage period in the texture of canned 

Agaricus bisporus and Boletus edulis mushrooms previously blanched in water, blanched or soaked and 

blanched in solutions containing citric, l-ascorbic and lactic acids. The texture was examined using 

instruments [textural profile analysis (TPA), Kramer shear cell (KSC)] and sensory analysis [five-point, 

profiling (P)]. Canning B. edulis mushrooms reduced their hardness, chewiness and gumminess (TPA), the 

values for force and work (KC), and brittleness and crispiness (P), although increasing their cohesiveness 

(TPA). Canning A. bisporus mushrooms reduced their hardness (TPA) and the expenditure of work, but 

increased their cohesiveness, hardness, crispiness and firmness (P). Twelve-month storage of both species of 

canned mushrooms led to a reduction in brittleness and crispiness (P). The type of pre-treatment applied 

affected the texture only when determined using profile analysis, and significant differences in hardness, 

crispiness and firmness between blanched-only and soaked and blanched products were mainly found in 

B. edulis. 

Keywords Canning, mushrooms, preliminary treatment, storage, texture. 

Introduction 

Edible mushrooms are popular all over the world 

because of their unique taste and texture (Chang, 

2006). The most popular species globally is Agaricus 

bisporus, whose fruiting bodies are white or creamcoloured 

with crumbly, fragile, dry and hard flesh. In 

Central and Eastern Europe, including Poland, the finest 

edible mushrooms are considered to be those of the 

Boletus family. Characterised by a particularly high 

sensory value is Boletus edulis, with caps ranging from 

light to dark brown in colour, greyish-white to greyishbrown 

stipes and delicate reticulated pattern, white to 

light brown in colour. Its flesh is firm, hard and does not 

stain when cut (Mauseth, 2003). 

Edible mushrooms are highly perishable. One of the 

most frequently used methods of preserving mushrooms 

is canning, either in brine or marinade. As mushrooms 

darken quickly during processing and subsequent storage, 

they are subject to an appropriate form of pretreatment 

prior to canning. The most common of these 

are washing, soaking, blanching or vacuum impregna- 

*Correspondent: E-mail: rrgjawor@cyf-kr.edu.pl 

doi:10.1111/j.1365-2621.2010.02319.x 


tion in solutions containing substances that inhibit 

browning of the tissue, such as table salt, citric acid, 

l-ascorbic acid, hydrogen peroxide, EDTA, sodium 

isoascorbate, cysteine hydrochloride and metabisulphites. 

According to Lin et al. (2001), soaking time of 

Agaricus bisporus in water can vary from 15 min to 2 h, 

when applied prior to sterilisation. For the blanching 

process, the cited authors recommend a water solution 

of citric acid (1%). Also, during the pre-treatment prior 

to preservation, 15-min soaking of mushrooms in a 

water solution of potassium metabisulfite (1%) had been 

applied to improve the colour of Agaricus bisporus 

fruiting bodies, which were then blanched in hot water 

for 8 or 10 min (Vivar-Quintana et al., 1999). Parrish 

et al. (1974) reported the shorter 5-min blanching time 

of Agaricus bisporus. Kotwaliwale et al. (2007) applied 

blanching in hot water for 2 min followed by soaking in 

potassium metabisulfite (0.75%, 1% or 1.5%) solutions 

for 15 min prior to drying of Pleurotus ostreatus 

mushroom. Furthermore, Vivar-Quintana et al. (1999) 

propose water solutions of NaCl (2%), citric (0.5%), 

and ⁄ or ascorbic acid (1000 ppm) as marinades’ components 

of the sterilised canned product, while Zivanovic 

& Buescher (2004) suggest a solution of NaCl (1%)

1660 

Texture of fresh and preserved mushrooms G. Jaworska et al. 

containing CaCl2 at 5 or 25 mm concentration. In 

agreement with Parrish et al. (1974), Vivar-Quintana 

et al. (1999), as well as Zivanovic & Buescher (2004) the 

process of sterilisation should be carried out at 121 °C 

for 18–25 min. 

Particular technological treatments used as part of the 

preservation process, notably pre-treatment, have an 

effect on textural changes in mushrooms, concerning 

which there are seemingly conflicting opinions in the 

available literature. It should also be pointed out that the 

source literature on this question is not extensive. 

According to Czapski (1994), blanching has a deleterious 

effect on the texture of A. bisporus mushrooms; it is, 

however, widely used because of its considerable effectiveness 

in inhibiting browning of the tissue. In Steinbuch 

(1986) opinion, blanching of A. bisporus in water prior to 

freezing results in increased hardness and gumminess of 

the mushrooms. Zivanovic & Buescher (2004), on the 

other hand, showed that sterilising A. bisporus blanched 

in water increases the firmness of mushrooms. 

The aim of the work was to investigate the effect on 

the texture of canned Agaricus bisporus and Boletus 

edulis mushrooms prepared using different methods of 

preliminary processing (blanching or soaking and 

blanching), sterilisation and long-term frozen storage. 


Material 

The material for the study was fresh and canned 

Agaricus bisporus (Lange) Sing. and Boletus edulis (Bull: 

Fr.) mushrooms evaluated directly after processing and 

after 12-month storage. The canned product was 

prepared from mushrooms blanched or soaked and 

blanched in solutions containing additional substances. 

Fresh A. bisporus mushrooms were obtained from a 

specialist farm and canned approximately 4 h after 

being harvested. B. edulis mushrooms with a cap diameter 

of 4–8 cm were obtained from pine forests in 

western Poland and canned approximately 12 h after 

being picked. The pre-treatment prior to canning 

involved sorting (rejecting any unsound or wormy 

specimens); cleaning the mushrooms of any remaining 

litter; in the case of B. edulis, separating the caps from 

the stipes; and washing in cold, running water. Each 

batch of mushroom was then divided into five parts, 

each of which is of 10 kg mass. Three parts were 

blanched: in water (product code BW); in an aqueous 

solution of 0.5% citric acid (INS 330) and 0.1% 

l-ascorbic acid (INS 300) (product code BCA); and in 

an aqueous solution of 1% lactic acid (INS 270) and 

0.1% l-ascorbic acid (INS 300) (product code BLA). 

The remaining two parts were soaked and blanched in 

the same solutions as products BCA and BLA and given 

the codes SBCA and SBLA, respectively. Soaking lasted 

1 h, the proportion of weight of the mushrooms to 

solution being 1:2. Blanching was carried out at a 

temperature of 96–98 °C, the proportion of weight of 

the mushrooms to water or solution being 1:5. The 

blanching time, which was determined experimentally, 

was 3 min for whole A. bisporus mushrooms and the 

caps of B. edulis, and 1.5 min for the stipes of B. edulis. 

For the purposes of textural profile analysis (TPA) 

analysis, whole A. bisporus mushrooms and the separated 

caps and stipes of B. edulis were preserved. For the 

textural analysis in the Kramer shear cell and using 

profile analysis, the mushrooms were cut into strips 

5 mm thick before being preserved; the cut strips of 

B. edulis were packed in the proportion of 1:1.4 caps to 

stipes. The 180 g of mushrooms was placed in 300-cm 3 

‘twist-off’ glass jars. Then, jars with mushrooms were 

poured with 100 cm 3 of hot solution containing 2% of 

salt (in a ratio by mass of pilei to the solution of 3:2), 

deaerated by immersion, sealed and sterilised. The 

process of sterilisation was carried out as follows: 

elevation of temperature up to 100 °C – 5 min; a rise 

in temperature from 100 to 118°C – 10 min, a proper 

sterilisation in temperatures from 118 to 121 °C – 

12 min; and finally, cooling down to 30 °C – 10 min. 

Sterilisation took place in a laboratory pressure steriliser 

of USA manufacture. The canned mushrooms were kept 

in a storage chamber for 12 months at temperature of 

8–10 °C. 

Methods 

Chemical evaluation 

The dry matter, ash and pH level were determined by 

AOAC (1995), the results being calculated from four 

replications. 

Texture analysis 

The textural evaluation of Agaricus bisporus and Boletus 

edulis was carried out by sensory and instrumental 

analysis. 

The sensory evaluation of texture was conducted using 

five-point and profile analysis methods. Using procedures 

in accordance with Polish standard PN-ISO 6658 

(1998), texture of mushrooms was scored on a 5–1 scale 

(5 = excellent, 4 = very good, 3 = good, 2 = bad, 

1 = very bad) by a panel of five judges, fulfilling the 

requirements for sensory sensitivity in Polish standard 

(PN-ISO 3972, 1998). Each analysis was carried out in 

five replications. 

The TPA was carried out in accordance with PN-ISO 

11036 (1999) and ISO 13299 (2003). The canned products 

were assessed by a panel of eight experts qualified to 

do so according to PN-ISO 8586-2 (1996) standards. The 

profile analysis considered hardness, brittleness, crispiness, 

firmness, wateriness, gumminess and sliminess. 

Each analysis was carried out in eight replications. 


The instrumental textural examination was carried 

out using a TA XT2 texture analyser (Stable Micro 

Systems, Haslemere, Surrey, UK). Two types of probes 

were used: a P ⁄ 45 cylinder probe (45 mm diameter) 

made from aluminium for the TPA and a modified 

Kramer shear cell with three blades, enabling the 

material to be cut. The TPA analysis was carried out 

separately on the caps and stipes, with cylinders cut out 

of the mushrooms 20 mm in length and diameter. The 

samples were then pressed twice at a rate of test speed 

1mms )1 to obtain 50% deformation, with a 20-s 

interval between pressings. Each group was analysed in 

six replications. From the curves obtained from the 

TPA, the textural parameters were determined as 

hardness [N], springiness, cohesiveness, chewiness [N] 

and gumminess. 

Cut pieces of mushroom were used for the Kramer 

shear cell analysis (KSC). Measurement consisted of 

cutting through 50 g of material placed in the cell at a 

rate of 1 mm s )1 . From the curves obtained, two 

parameters were determined: the maximum force [N] 

and work [mJ] required to cut through the sample 

material. Each analysis was carried out in six replications. 

For both TPA and KSC, 25-kg load cell was used. 


The results were statistically evaluated using the 

F-Snedecor and Student’s t-tests (Statistica 6.1 Pl 

program, Stat Soft Inc., Tulsa, OK, USA), the least 

Table 1 Content of dry matter, ash and pH level in fresh and canned mushrooms 

Texture parameter Mushroom species 

Time of 

storage S 

significant difference being calculated at a = 0.01. The 

linear correlation coefficient was determined between the 

textural parameters indicated by instrumental methods 

and those indicated by the sensory-profiling method. 

Results and Discussion 

Chemical evaluations 

Fresh A. bisporus mushrooms contained 20% and 4%, 

respectively less dry matter than the caps and stipes of 

B. edulis; however, they contained 22% and 85% more 

ash, respectively. Fresh caps of B. edulis contained 19% 

more dry matter and 51% more ash than stipes. pH 

values of fresh mushrooms were similar and fluctuated 

between 6.35 and 6.40 (Table 1). 

The production of canned mushrooms increased the 

level of dry matter in A. bisporus by 1–4% but reduced it 

in B. edulis caps by 3–19% and stipes by 5–27%. 

Adding brine during the production process increased 

ash content compared with fresh mushrooms by 

46–65% in canned A. bisporus, 64–80% in canned 

B. edulis caps and 108–126% in canned B. edulis stipes. 

An insignificant, not exceeding 3%, rise was observed in 

pH values of the canned caps, those blanched in water. 

In the remaining products, pH decreased by 3–16% in 

Agaricus bisporus and by 12–32% in caps and stipes of 

Boletus edulis, whereas no pH differentiation was found 

among products obtained from the edible parts of 

Kind of pre-treatment 

Texture of fresh and preserved mushrooms G. Jaworska et al. 1661 

BW BCA SBCA BLA SBLA 

Dry matter g ⁄ 100 g f.m. Agaricus bisporus 0 8.71 ± 0.08 9.09 ± 0.08 9.10 ± 0.72 8.76 ± 0.08 9.05 ± 0.08 8.84 ± 0.43 0.153 

12 9.08 ± 0.16 9.05 ± 0.66 8.72 ± 0.16 9.07 ± 0.74 8.86 ± 0.65 

Boletus edulis caps 0 10.80 ± 0.12 10.36 ± 0.12 10.41 ± 0.12 8.77 ± 0.09 10.50 ± 0.84 8.74 ± 0.55 0.059 

12 10.37 ± 0.11 10.43 ± 0.08 8.76 ± 0.32 10.56 ± 0.66 8.81 ± 0.78 

B. edulis stipes 0 9.08 ± 0.05 8.42 ± 0.09 8.49 ± 0.63 6.72 ± 0.61 8.63 ± 0.51 6.67 ± 0.08 0.069 

12 8.40 ± 0.45 8.50 ± 0.35 6.71 ± 0.44 8.66 ± 0.44 6.74 ± 0.66 

Ash g ⁄ 100 g f.m. A. bisporus 0 0.72 ± 0.06 1.19 ± 0.08 1.18 ± 0.04 1.05 ± 0.05 1.14 ± 0.09 1.08 ± 0.07 0.061 

12 1.17 ± 0.04 1.13 ± 0.08 1.08 ± 0.09 1.12 ± 0.11 1.07 ± 0.08 

B. edulis caps 0 0.59 ± 0.03 1.06 ± 0.07 1.04 ± 0.04 0.99 ± 0.04 1.04 ± 0.08 0.97 ± 0.09 0.035 

12 1.04 ± 0.06 1.06 ± 0.07 0.98 ± 0.04 1.02 ± 0.07 0.98 ± 0.07 


12 0.86 ± 0.04 0.87 ± 0.08 0.83 ± 0.03 0.86 ± 0.05 0.82 ± 0.04 

pH A. bisporus 0 6.40 ± 0.02 6.51 ± 0.02 5.90 ± 0.02 5.81 ± 0.02 6.10 ± 0.01 5.39 ± 0.02 0.141 

12 6.47 ± 0.03 5.84 ± 0.02 5.78 ± 0.02 6.01 ± 0.03 5.31 ± 0.03 

B. edulis caps 0 6.38 ± 0.04 6.57 ± 0.03 5.59 ± 0.09 5.22 ± 0.03 5.09 ± 0.07 4.36 ± 0.05 0.120 

12 6.49 ± 0.07 5.51 ± 0.083 5.13 ± 0.09 5.04 ± 0.04 4.28 ± 0.06 


12 6.45 ± 0.07 5.49 ± 0.08 5.13 ± 0.09 4.99 ± 0.11 4.23 ± 0.07 

f.m., fresh matter; S, fresh mushrooms; BW, blanching in water; BCA, blanching in citric acid and L-ascorbic acid water solution; SBCA, soaking 

and then blanching in citric acid and L-ascorbic acid water solution; BLA, blanching in lactic acid and L-ascorbic acid water solution; SBLA, soaking 

and then blanching in lactic acid and L-ascorbic acid water solution; ± standard deviation. 

LSD, 

a = 0.01 


1662 


Boletus edulis. No significant change was observed in 

dry matter, ash and pH level during the 12-month 

storage period. 

Texture analysis 

The canned, sterilised products obtained form Agaricus 

bisporus showed a very good and balanced consistency, 

however, slightly inferior in comparison with the products 

of Boletus edulis. During 1-year storage of the 

products, a visible worsening of mushrooms’ consistency 

was observed by 0.8–0.9 points for Agaricus 

bisporus and by 0.5 points in the case of Boletus edulis. 

Pre-treatment applied prior to sterilisation did not affect 

the mushrooms’ consistency evaluated on a five-point 

scale (Table S1). 

In the TPA, it was found that fresh A. bisporus and 

B. edulis mushrooms were moderately hard and very 

brittle, crispy and firm, but were not watery or slimy 

(Table S2). 

Compared with fresh mushrooms, the technological 

process of canning A. bisporus resulted in a significant 

increase in hardness (1.0–1.4 points) and crispiness (0.5 

points) and firmness, with the exception of mushrooms 

blanched in water. In the case of B. edulis, on the other 

hand, a significant decrease in brittleness (2.3–3.0 points) 

and crispiness (1.0–2.0 points) was observed, regardless of 

the pre-treatment applied; in canned products obtained 

after soaking and blanching (SBCA and SBLA), there was 

also a decrease in hardness (0.7 points) and firmness (1.5 

points). Moreover, canned B. edulis became slightly 

watery and slimy (Table S2). 

In some samples, the storage of canned mushrooms of 

both species resulted in significant reductions in brittleness 

and crispiness of 0.1–1.1 points and 0.2–1.3 points, 

respectively, and additionally in A. bisporus, reductions 

in hardness and firmness of 0.4–1.4 points and 0.6–1.2 

points, respectively. In the case of canned B. edulis, an 

increase in hardness of 0.1–1.2 points was observed, 

significant in three products: BCA, BLA and SBLA, and 

an increase in firmness of 0.2–0.4 points with the 

exception of mushrooms blanched in water. 

The type of pre-treatment applied affected textural 

parameters evaluated through profile analysis, but most 

of the significant differences were found at the end of the 

storage period. Canned B. edulis products previously 

soaked and blanched (SBCA and SBLA) compared with 

blanched-only products (BCA and BLA) were less hard 

(0.6–1.2 points), crispy (0.8–0.9 points) and firm (0.8 

points), but more watery (0.7–0.9 points). Canned 

A. bisporus showed decreased hardness (0.9–1.0 points) 

and brittleness (0.5–0.7 points). 

The TPA analysis showed that fresh A. bisporus 

mushrooms were 50–78% harder, 17–73% gummier 

and 6–41% chewier than B. edulis caps and stipes. They 

were also 10–19% less springy and 34% less cohesive. 

There were also significant differences between the parts 

of B. edulis: the caps were 16% less hard and 11% less 

springy than the stipes, but 73% more cohesive, 49% 

gummier and 33% chewier. Ko et al. (2007) claim that 

Flammulina velutipes caps show better hardness and 

lower springiness and cohesiveness, when compared 

with stipes. 

It should also be noted that there were relatively large 

differences in the levels of individual textural parameters 

within the A. bisporus samples and also within the cap 

and stipe groups of B. edulis. In the case of A. bisporus, 

the standard deviation in the values for these parameters 

amounted to 12–22%, while in the case of B. edulis, it 

was 6–18% for caps and 8–23% for stipes (Table 2). The 

high level of differences in standard deviation resulted 

primarily from the heterogeneous structure of individual 

mushrooms, which for the purposes of analysis were 

classified according to cap size rather than structure. 

The production process did not bring any significant 

changes in springiness, chewiness, or gumminess in the 

case of canned A. bisporus, nor in springiness in the 

case of canned B. edulis. Small changes in cohesiveness 

may have resulted from the cellular moisture and 

solutes movement into intercellular spaces after the 

collapse of cell membrane. Loch & Breene (1982) 

suggested that the moisture and solutes, which retained 

in such intercellular spaces, established hydrostatic 

pressure maintaining the springiness of mushrooms. 

Canning caused a significant increase in tissue cohesion, 

with canned A. bisporus almost three times as cohesive 

as fresh mushrooms, canned B. edulis caps 2.5 times 

and canned B. edulis stipes four times as cohesive. 

Canning resulted in a levelling of cohesiveness between 

B. edulis caps and stipes. An increase in cohesiveness of 

Flammulina velutipe pilei as a consequence of high 

temperature treatment has also been noted (Ko et al., 

2007). 

Canning also caused significant decrease in hardness: 

64–69% in A. bisporus; 82–84% in B. edulis caps and 

90–92% in B. edulis stipes. Loss of hardness probably 

resulted as an effect of protein denaturation and 

permeability changes, causing loss of cellular water 

and solutes because of high temperature treatment 

(Loch & Breene, 1982). In addition, in B. edulis caps 

and stipes, it caused an approximately 2.5-fold decrease 

in chewiness and gumminess. Ko et al. (2007) observed a 

decrease in chewiness and gumminess of caps and stipes 

of Flammulina velutipes mushrooms when treated with a 

temperature above 90 °C for 1–10 min. 

Prior to storage, canned A. bisporus mushrooms 

showed similar springiness to B. edulis caps and stipes 

but were approximately 1 ⁄ 3 less cohesive, 3–4 times as 

chewy, 3–5 times as gummy and 3.5–6.0 times as hard. 

Pre-treatment and sterilisation altered the differences in 

individual textural characteristics of B. edulis caps and 

stipes. There was less of a difference in springiness and 


Table 2 The results of the TPA measurement of fresh and canned mushrooms 

Texture 

parameter Mushroom species 

Time of 

storage S 

cohesiveness between the caps and stipes of canned 

B. edulis. Similarly to fresh mushrooms, canned caps 

were approximately 1 ⁄ 3 chewier than canned stipes. 

Canning increased the differences between the parts of 

B. edulis in gumminess and hardness, which were 2 ⁄ 3 and 

1.5 times greater, respectively in caps than in stipes. 

Ko et al. (2007) revealed that in Flammulina velutipes 

mushrooms, treated with a temperature exceeding 

90 °C within 1–10 min, the reduction in hardness and 

cohesiveness is observed accompanied by an increase in 

springiness of both caps and stipes. With the exception of 

springiness, the changes reported were higher in stipes 

than in caps. The greater effect of a high temperature 

could probably be caused by the fact that the internal 

thermal conduction in the caps was less effective and less 

destructive to caps tissue because of the thicker fruiting 

bodies and greater caps’ volume in comparison with the 

stipes. 

In the TPA analysis, 12-month storage of canned 

mushrooms had no significant effect on individual 

textural parameters. It should be noted that the storage 



LSD, 

a = 0.01 

Springiness Agaricus bisporus 0 0.446 ± 0.064 0.474 ± 0.062 0.468 ± 0.058 0.446 ± 0.093 0.467 ± 0.155 0.442 ± 0.060 NS 

12 0.493 ± 0.062 0.516 ± 0.065 0.484 ± 0.094 0.536 ± 0.070 0.486 ± 0.060 

Boletus edulis caps 0 0.494 ± 0.040 0.518 ± 0.039 0.544 ± 0.063 0.522 ± 0.076 0.530 ± 0.080 0.500 ± 0.083 NS 

12 0.503 ± 0.056 0.534 ± 0.049 0.518 ± 0.057 0.512 ± 0.081 0.482 ± 0.085 

B. edulis stipes 0 0.552 ± 0.048 0.550 ± 0.091 0.552 ± 0.059 0.621 ± 0.037 0.584 ± 0.095 0.630 ± 0.051 NS 

12 0.539 ± 0.085 0.482 ± 0.079 0.538 ± 0.060 0.518 ± 0.066 0.577 ± 0.064 

Cohesiveness A. bisporus 0 0.106 ± 0.020 0.294 ± 0.062 0.290 ± 0.056 0.291 ± 0.060 0.309 ± 0.047 0.308 ± 0.024 0.1080 

12 0.289 ± 0.059 0.267 ± 0.057 0.246 ± 0.051 0.264 ± 0.034 0.249 ± 0.036 

B. edulis caps 0 0.161 ± 0.011 0.412 ± 0.056 0.419 ± 0.031 0.377 ± 0.099 0.467 ± 0.017 0.409 ± 0.058 0.1132 

12 0.392 ± 0.073 0.386 ± 0.027 0.359 ± 0.044 0.360 ± 0.134 0.355 ± 0.070 


12 0.401 ± 0.067 0.334 ± 0.048 0.341 ± 0.044 0.351 ± 0.037 0.415 ± 0.071 

Chewiness [N] A. bisporus 0 2.62 ± 0.57 2.83 ± 0.39 2.82 ± 0.51 2.21 ± 0.46 2.73 ± 0.48 2.60 ± 0.41 NS 

12 2.73 ± 0.51 2.65 ± 0.53 2.23 ± 0.47 2.93 ± 0.55 2.37 ± 0.30 

B. edulis caps 0 2.47 ± 0.40 1.09 ± 0.08 1.07 ± 0.18 1.10 ± 0.18 1.16 ± 0.18 1.18 ± 0.15 0.659 

12 1.12 ± 0.20 1.24 ± 0.15 1.22 ± 0.19 1.21 ± 0.14 1.13 ± 0.15 


12 0.59 ± 0.09 0.57 ± 0.11 0.63 ± 0.11 0.54 ± 0.10 0.71 ± 0.08 

Gumminess A. bisporus 0 5.88 ± 1.06 5.98 ± 0.87 6.33 ± 1.09 5.04 ± 0.60 5.99 ± 1.15 5.33 ± 1.05 NS 

12 5.51 ± 1.03 5.34 ± 0.71 4.86 ± 1.06 5.45 ± 1.13 4.90 ± 0.69 

B. edulis caps 0 5.04 ± 0.90 2.18 ± 0.36 1.94 ± 0.29 2.08 ± 0.36 2.38 ± 0.41 2.30 ± 0.33 1.021 

12 2.31 ± 0.34 2.29 ± 0.32 2.29 ± 0.31 2.41 ± 0.41 2.34 ± 0.20 


12 1.28 ± 0.20 1.14 ± 0.17 1.43 ± 0.24 1.05 ± 0.18 1.57 ± 0.11 

Hardness [N] A. bisporus 0 55.68 ± 6.47 18.72 ± 2.33 18.34 ± 2.28 17.48 ± 2.78 20.15 ± 4.07 18.79 ± 3.26 7.305 

12 19.18 ± 3.30 18.52 ± 3.95 20.51 ± 3.78 20.91 ± 4.28 19.89 ± 3.15 

B. edulis caps 0 31.34 ± 5.58 5.30 ± 0.71 5.01 ± 0.71 5.59 ± 1.00 5.16 ± 0.83 5.72 ± 1.04 3.198 

12 6.00 ± 0.70 5.92 ± 0.89 6.31 ± 0.74 6.31 ± 1.08 6.58 ± 1.00 


12 3.18 ± 0.52 3.89 ± 0.62 4.24 ± 0.75 3.35 ± 0.61 3.95 ± 0.48 

S, fresh mushrooms; BW, BCA, SBCA, BLA, SBLA – see Table 1; NS, not significant; ± standard deviation. 


period increased the hardness of mushrooms, particularly 

in canned B. edulis, whose caps and stipes were 

respectively, 13–22% and 5–19% harder after 12-month 

storage compared with caps and stipes directly after 

canning. A 17–33% decrease in the chewiness of canned 

B. edulis stipes was also observed after storage. 

The type of pre-treatment applied did not have a 

significant effect on texture in the TPA analysis of 

mushrooms. Kotwaliwale et al. (2007) noted that dried 

Pleurotus sp. mushrooms blanched in water prior to 

drying were significantly harder, but less cohesive and 

springy than those soaked in a potassium metabisulfite 

solution for 15 min. 

Cutting through strips of fresh A. bisporus in the 

Kramer shear cell required 19% more force and 26% 

more work than cutting through B. edulis (Table 3). 

Czapski (1994) stated that the force required to cut 

through 50-g fresh A. bisporus was 15–20 N g )1 , 2.5 

times lower than that in this experiment. 

The canning process had a significant effect on the 

values of both the above textural parameters measured 


1664 


Table 3 The results of the Kramer shear cell measurement of canned mushrooms 

Texture 

parameter Mushroom species 

in the Kramer shear cell in the case of strips of 

B. edulis and only of the work required in the case of 

strips of A. bisporus. Compared with the fresh mushrooms, 

29–46% less force was required to cut through 

the canned B. edulis. With respect to shear work, 

44–48% less work was required to cut through the 

canned A. bisporus, while for the canned B. edulis, it 

was 65–88%. 

Twelve-month storage did not result in significant 

changes in the textural parameters measured in the 

Kramer shear cell. Zivanovic & Buescher (2004) noted 

that 3-month storage of canned A. bisporus at room 

temperature had no effect on firmness determined as 

force for puncturing; however, it caused a significant 

increase in toughness determined as shear-press force, 

from 91 to 103 kg ⁄ 100 g. According to the above 

authors, firmness and toughness are affected by the 

concentration of CaCl 2 in the canning liquid because an 

increase in the concentration of this compound from 0 

to 25 mm causes a significant increase in the level of 

these two parameters. 

The type of pre-treatment applied prior to canning 

resulted in significant differences in the values of textural 

parameters only in the case of B. edulis. Canned 

mushrooms pre-treated with lactic acid and l-ascorbic 

acid (products BLA and SBLA) required an average of 

29% more force and 24% more work than those 

blanched in water (BW) or those pre-treated with citric 

acid and l-ascorbic acid (BCA and SBCA). 

Correlation ratio 

Time of 

storage S 

According to the TPA analysis, changes in the hardness 

of B. edulis caps and stipes, resulting from the canning 

process and subsequent storage, were strongly and 

positively correlated with each other (r = 1.00, R 2 = 

1.00). The hardness of A. bisporus and B. edulis determined 

by TPA analysis was strongly enough correlated 

with the force required to cut through the material in the 

Kramer shear cell; the relationship was negative in the 



case of A. bisporus (r = )0.85, R 2 = 0.73) and positive 

in the case of B. edulis caps and stipes (r = 0.79, 

R 2 = 0.63 and r = 0.78, R 2 = 0.61, respectively). 

Hardness determined by the TPA measurement or the 

value of force defined using the Kramer shear cell did 

not correlate with the hardness assessed by a profile 

analysis. 

For the remaining texture parameters (springiness, 

cohesiveness, chewiness and gumminess) determined by 

the TPA measurement and the texture parameters 

defined in the profile analysis, only weak correlation 

was observed with the linear correlation coefficients not 

exceeding )0.77 and R 2 < 0.60. 

Conclusions 

LSD, 

a = 0.01 

Force [N] Agaricus bisporus 0 140.0 ± 13.3 162.4 ± 15.9 172.9 ± 15.6 172.3 ± 26.0 175.7 ± 17.8 169.7 ± 21.5 NS 

12 160.3 ± 17.5 171.4 ± 19.7 175.8 ± 23.8 177.2 ± 20.1 167.6 ± 20.7 

Boletus edulis 0 113.3 ± 5.5 66.5 ± 11.9 65.7 ± 14.6 61.1 ± 10.4 85.3 ± 16.0 80.6 ± 14.8 18.94 

12 64.7 ± 10.6 61.9 ± 10.1 52.1 ± 13.0 81.8 ± 13.4 71.1 ± 11.5 

Work [mJ] Agaricus bisporus 0 2689 ± 215 1409 ± 124 1423 ± 237 1510 ± 240 1488 ± 222 1424 ± 156 307.1 

12 1371 ± 200 1414 ± 115 1462 ± 219 1480 ± 228 1409 ± 186 

Boletus edulis 0 1978 ± 23 549 ± 126 583 ± 125 442 ± 67 686 ± 110 614 ± 103 140.1 

12 522 ± 92 523 ± 78 395 ± 65 600 ± 70 579 ± 91 

S, BW, BCA, SBCA, BLA, SBLA, – see Table 1; NS, not significant; ± standard deviation. 

The technological process of canning had a significant 

effect on the content of dry matter and ash, unlike the 

12-month period of storage, during which no changes 

in these features of the chemical composition were 

observed. 

The canning process led to changes in the textural 

parameters depending on the species of mushroom. The 

hardness, chewiness and gumminess of B. edulis evaluated 

through TPA analysis were reduced as a result of 

canning, as was the degree of force and work required 

for cutting through mushrooms measured using the 

Kramer shear cell and the evaluations of brittleness and 

crispiness and, in some cases, hardness and firmness 

evaluated using sensory profiling. An increase in cohesiveness 

was also observed (TPA). In the case of 

A. bisporus, there were decreases in hardness determined 

instrumentally (TPA) and in work, and an increase in 

cohesiveness (TPA) as well as in hardness, crispiness and 

firmness determined through profile analysis. 

Twelve-month storage of the canned mushrooms of 

both species significantly affected the texture of mushrooms 

assessed by sensory evaluation but had no effect 

on the texture parameters evaluated using the instrumental 

methods. After 1-year storage, the texture 


evaluated by the use of five-point scale method was 

significantly lower, and changes were noted in hardness, 

brittleness and crispiness evaluated using sensory profiling. 

The type of pre-treatment applied affected only the 

degree of force and work required to cut through canned 

B. edulis mushrooms in the Kramer shear cell as well as 

certain parameters in both species evaluated through 

profile analysis. 

Texture parameters obtained by instrumental methods 

were strongly correlated with each other and did not 

correlate with those obtained from sensory profiling. 


The study was financed by the Ministry of Education 

and Science under a research Project No. 2 PO6T 041 29. 

References 

AOAC (1995). Official Methods of Analysis, 16th edn. Arlington, USA: 

Association of Official Analytical Chemistry. 

Chang, S.-T. (2006). The world mushroom industry: trends and 

technological development. International Journal of Medicinal 

Mushrooms, 8, 297–314. 

Czapski, J. (1994). Wpyw niektórych operacji technologicznych na 

wydajnos´ć i jakosć´ pieczarek blanszowanych i skadowanych w 

zalewie [The effect of some technological treatments on the yield and 

quality of blanched mushroom stored in brine]. Vegetable Crops 

Research Bulletin, 42, 101–119 [in Polish]. 

ISO 13299 (2003). Sensory Analysis. Methodology. General Guidance 

for Establishing a Sensory Profile. Geneva, Switzerland: International 

Organization for Standardization. 

Jaworska, G., Ge˛bczyński, P. & Goyszny, A. (2003). Wykorzystanie 

pieczarek do produkcji mro_zonych i sterylizowanych farszów [The 

use of mushrooms (Agaricus bisporus) in the production of frozen 

and canned stuffings]. _Zywnosć´. Nauka. Technologia. Jakosć´. [Food. 

Science. Technology. Quality], Suplement, 36, 63–71 [in Polish]. 

Ko, W.C., Liu, W.C., Tsang, Y.T. & Hsieh, C.H. (2007). Kinetics of 

winter mushrooms (Flammulina velutipes) microstructure and quality 

changes during thermal processing. Journal of Food Engineering, 

81, 587–598. 

Kotwaliwale, N., Bakane, P. & Verma, A. (2007). Changes in textural 

and optical properties of oyster mushroom during hot air drying. 


Lin, Z., Chen, H. & Lin, F. (2001). Influence of treatments and 

blanching treatments on the yield and color of canned mushrooms. 

Journal of Food Processing and Preservation, 25, 381–388. 


Loch, J. & Breene, W.M. (1982). Between-species differences in 

fracturability loss: microscopic and chemical comparison of potato 

and Chinese water chestnut. Journal of Texture Studies, 13, 325–327. 

Mauseth, J.D. (2003). Botany: An Introduction to Plant, 3rd edn. 

Austin, TX: University of Texas. 

Parrish, G.K., Beelman, R.B., McAdle, F.J. & Kuhn, G.D. (1974). 

Influence of post-harvest storage and canning on the solids and 

mannitol content of the cultivated mushroom and their relationship 

to canned product yield. Journal of Food Science, 39, 1029– 

1031. 

PN-ISO 11036 (1999). Analiza sensoryczna. Profilowanie tekstury 

[Sensory Analysis. Methodology. Texture Profile]. Warsaw, Poland: 

Polish Committee for Standardization [in Polish]. 

PN-ISO 3972 (1998). Analiza sensoryczna. Metodologia. Metoda 

sprawdzania wra _zliwosći smakowej [Sensory Analysis. Methodology. 

Method of Investigating Sensitivity of Taste]. Warsaw, Poland: 

Polish Committee for Standardization [in Polish]. 

PN-ISO 6658 (1998). Analiza sensoryczna. Metodologia. Wytyczne 

ogo´lne. [Sensory Analysis. Methodology. General Guidelines]. Warsaw, 

Poland: Polish Committee for Standardization [in Polish]. 

PN-ISO 8586-2 (1996). Analiza sensoryczna. Ogo´lne wytyczne wyboru, 

szkolenia i monitorowania oceniaja˛cych. Eksperci [Sensory Analysis. 

General Guidance for the Selection, Training and Monitoring of 

Assessors. Experts]. Warsaw, Poland: Polish Committee for 

Standardization [in Polish]. 

Steinbuch, E. (1986). Some developments of mushroom processing. 

Proceedings of the International Symposium on Scientific and 

Technical Aspects of Cultivating Edible Fungi. Pp. 311–318. The 

Penna State University Pennsylvania, USA: Elsevier. 

Vivar-Quintana, A.M., González-San José, M.L. & Collado-Fernández, 

M. (1999). Influence of canning process on colour, weight and 

grade of mushrooms. Food Chemistry, 66, 87–92. 

Zivanovic, S. & Buescher, R. (2004). Changes in mushroom texture 

and cell wall composition affected by thermal processing. Journal of 

Food Science, 69, SNQ44–SNQ49. 




Table S1. Results of the five-point texture analysis of 

canned mushrooms 

Table S2. Results of the texture profile analysis of 

canned mushrooms. 







1666 


Chemical stability of yellow pigment extracted from the flower bud 

of Sophora japonica L. (Huaimi) 

Wanping Chen, 1 Ping Li 1 & Xiaohong Wang 1,2 * 

1 College of Food Science and Technology, Huazhong Agricultural University, Wuhan 430070, China 

2 Key Laboratory of Food Safety Evaluation of the Ministry of Agriculture, Wuhan 430070, China 


Summary The dry flower bud of Sophora japonica L., generally called Huaimi in China, has been used as traditional 

Chinese medicinal materials and folk edible yellow pigment material for a long time. In this paper, the yellow 

pigment was extracted from Huaimi, and its chemical stability related with pH value, temperature and 

metallic ions was studied systemically. When pH values changed from 2.0 to 9.0, the yellow pigment revealed 

structural transformation and colour variations. According to its mathematic model of thermal stability, the 

yellow pigment had good thermal stability, with a half-life of 918.9 h at room temperature. Effect of metallic 

ions on the yellow pigment displayed a co-pigmentation influence. HPLC analysis indicated that the content 

of rutin in it was about 42.56%. These results may be helpful for the further application investigations into 

its commercial prospects in food pigment industry. 

Keywords Chemical stability, Huaimi, kinetics, mathematical model, Sophora japonica L., yellow pigment. 

Introduction 

Pigments, as food additives to improve food colour and 

stimulate people’s appetite, are widely used in food 

industry (Gouveia et al., 2007). Pigments are classified 

as natural pigments and synthetic pigments according to 

their resources. Synthetic pigments are widely used in 

food industry because of their advantages. For example, 

synthetic pigments have brighter colour, lower cost and 

stronger stability compared with natural pigments. 

However, present studies have shown that some of 

them may have latent toxic effects on humans, such as 

carcinogenicity, mutagenicity inflammation (Gouveia 

et al., 2007; Gris et al., 2007; Castaneda-ovando et al., 

2009). By contraries, natural pigments are correspondingly 

safe, some of them even reveal nutritional and 

pharmacological effects, such as anti-oxidation, anticarcinogenicity, 

anti-mutagenicity, anti-inflammation 

(Galvano et al., 2004). Nowadays, natural pigments 

are being applied more and more in modern food 

industry because of the increase in consumer’s awareness 

of health maintenance (Gouveia et al., 2007; Gris 

et al., 2007). 

Recently, seeking for available resources of natural 

pigments is an extensive and active research field (El 


e-mail: wxh@mail.hzau.edu.cn 


gharras et al., 2008). Some natural pigments, such as 

carotenoids, anthocyanins and betalains, attract more 

consumer’s attention for their pharmacological and 

bioactive functions (Cai et al., 2003; Stintzing & Carle, 

2004). However, there are some disadvantages in the 

application and storage of natural pigments. The most 

demerits of natural pigments consist in their instability, 

which narrow their appliance in food industry (Castaneda-ovando 

et al., 2009). There are a few factors related 

with their stabilities. For instance, structure and concentration 

of natural pigments themselves are basic 

factors. pH value and temperature of environment, 

along with presence of co-compounds such as phenols 

and metallic ions, are also closely associated with their 

stabilities (Hurtado et al., 2009). Up to now, the list of 

natural pigments applying in foodstuff is mainly capsanthin, 

yellow corn, sorghum red, turmeric pigment, 

gardenia yellow pigment, b-carotene, red radish, Monascus 

pigments, and so on (Dufosse & Pintea, 2005). 

The dry flower bud of Sophora japonica L., generally 

called Huaimi in China, has been used as traditional 

Chinese medicinal materials and folk edible yellow 

pigment material by Chinese people for a long time 

(Wang et al., 2006; Qi et al., 2007; Lo et al., 2009). The 

latest pharmacological studies and clinical practices on 

Huaimi have found that it possesses anti-tumour, antifertility 

and anti-cancer activities, which can be attributed 

to its chemical components, such as flavonoids 

doi:10.1111/j.1365-2621.2010.02322.x 


components, the principal active ingredients of Huaimi 

(Kite et al., 2007; Qi et al., 2007; Xu et al., 2007; Lo 

et al., 2009). 

On record in Bencao Gangmu, also known as 

Compendium of Materia Medica, a famous Chinese 

materia medica work written by Li Shizhen in Ming 

Dynasty, local residents in southern China have used 

Huaimi as colouring agent in traditional food for almost 

400 years. Long-term application practices have proven 

that yellow pigment of Huaimi is safe and stable. 

However, studies of yellow pigment from Huaimi on its 

composition and chemical characteristics are less 

reported (Paniwnyk et al., 2001; Xu et al., 2007; Lo 

et al., 2009). Applications of Huaimi as yellow pigment 

material are mainly relied on people’s practical experiences. 

UV–VIS spectrophotometry is a feasible method for 

analysing plant extracts. In the visible wavelength range, 

the content of pigments is highly correlated with the 

absorption at the specific wavelengths (Pflanz & Zude, 

2008). In the present studies, yellow pigment extracts of 

Huaimi were prepared, and then spectrophotometric 

methods were applied to analyse its chemical stabilities 

related with pH value, temperature and metallic ions. 

The purpose was to explore the potential usage of this 

species as a source of yellow pigment and understand 

the conditions affecting its stabilities. 


Materials 

Huaimi, the dry flower bud of S. japonica L., was 

gathered from Jiangxi Province in southern China and 

stored at 4 °C for analysis. Rutin was purchased from 

Shanghai Chemical Corporation, Shanghai, China. 

HPLC-grade methanol and phosphoric acid were purchased 

from Sinopharm Chemical Reagent Co., Ltd, 

Shanghai, China. 

Preparation of Huaimi yellow pigment extracts (HYPE) 

To extract pigments, Huaimi powder of fifty meshes 

sieve was immersed into a binary solvent (ethanol– 

water, 3:2, v ⁄ v) with a twenty-fold powder weight for 

2 h at 85 °C. Extracted fluid was filtrated through a 

Whatman No.1 filter paper, and then the filtrate was 

precipitated by centrifugation at 10000 · g for 10 min. 

The supernatant was transferred into an evaporating 

flask and evaporated under vacuum at 40 °C in a rotary 

evaporator (Shanghai Ailang Instruments Co., Ltd., 

Shanghai, China). The condensate was taken out for 

freeze drying (Beijing DETIANYOU Technology 

Development Co., Ltd., Beijing, China) at

1668 

Chemical stability of Huami yellow pigment W. Chen et al. 

evaporation loss on absorbance. The thermal stability of 

HYPE solution was evaluated by calculating S, which 

was the apparent value weighing the thermal stability of 

HYPE solutions, by following eqn 1. 

S ¼ A 

A0 

ð1Þ 

Where A was the absorbance of HYPE solution at 

detecting wavelength after heat treatment, given as 

average value; A0 was the absorbance of HYPE solution 

at detecting wavelength without heat treatment at room 

temperature. 

Linear regression analysis was used to obtain the 

degradation rate constant (k) for all samples. The 

dependence of reaction rate constants on temperature 

was modelled using the exponential equation. 

k ¼ k0 expðaTÞ ð2Þ 

Where k 0, a is the coefficient of exponential equation; T 

is the treatment temperature, °C. 

Metallic ions effect on stability of HYPE 

Metallic salts, including aluminium nitrate, potassium 

chloride, magnesium sulphate, manganese sulphate, 

copper sulphate, ferric trichloride, calcium chloride 

and sodium chloride, were added to HYPE solutions, 

respectively. Concentration of these metallic ions solutions 

ranged from 1.0 · 10 )5 to 1.0 · 10 )2 m. HYPE 

solutions were placed in dark area at room temperature 

for 24 h. Then, the UV–VIS spectra of HYPE solutions 

were scanned to analyse the effect of metallic ions on the 

absorbance and colour variations. 


UV–VIS spectrum of HYPE 

UV–VIS spectrum of HYPE exhibited two major 

absorption bands in the UV region (Fig. 1), the same 

as that of most flavonoids components (Botelho et al., 

2007). One absorption peak was at about 265 nm, called 

Peak I. The other absorption peak was about at 360 nm, 

called Peak II. The maximum absorbent wavelength at 

360 nm could be chosen for HYPE stability studies. 

UV–VIS spectrophotometry was based on Lambert– 

Beer Law. In the visible wavelength range, the content 

of pigments is highly correlated with the absorption at 

the specific wavelengths (Pflanz & Zude, 2008). Linear 

regression equation of HYPE solution between A360 nm 

and c (mg mL )1 ) was A360 nm = 15.416 c + 0.008, and 

correlation coefficient R 2 was 0.9997. It indicated that 

the relationship between solution absorbance (A360 nm) 

and concentration (c, mgmL )1 ) was linear from 0.01 to 

Figure 1 Room temperature absorbance UV–VIS spectra of HYPE 

solutions at different concentrations (mg mL )1 ). 

0.1 mg mL )1 . Thus, HYPE solution at concentration of 

0.05 mg mL )1 was chosen for further stability studies. 

Rutin content of HYPE 

According to the previous reports (Paniwnyk et al., 

2001; Qi et al., 2007; Xu et al., 2007), Huaimi had a high 

content of rutin and was used as plant material for rutin 

source. To measure rutin content, HPLC analysis was 

applied for quantitative determination. The analysis of 

rutin standard showed that its retention time was 

4.9 min. The relationship between peak area (y, lmÆs) 

and rutin content (x, lg) was linear from 0.4 to 1.5 lg, 

with regression equation y =2· 10 6 x ) 27733 and 

correlation coefficient R 2 = 0.9991 (n = 5). According 

to this equation, the content of rutin in HYPE was 

measured to be 42.56%. The result indicated that rutin 

was major component in HYPE. 

Rutin, sometimes known as Vitamin P, is a kind of 

flavonoid glycoside compound. It has many bioactive 

functions, such as antiplatelet, antiviral and antihypertensive, 

especially with high radical scavenging activity 

and antioxidant capacity (Guo et al., 2007; Yang et al., 

2008). Now, it has been widely used in medical fields 

(De oliveira et al., 2006; Yang et al., 2008). 

On the other hand, consumer’s awareness of health 

maintenance increases progressively. Latest processed 

foods, containing natural bioactive ingredients, gain 

more and more favour than those without such ingredients 

(Gouveia et al., 2007). So there has reason to 

believe that HYPE will have extensive applied prospects 

in food pigment industry. 

Influence of pH on stability of HYPE 

Effect of pH on HYPE stability was studied at different 

pH values, ranged from 2.0 to 9.0 (Fig. 2). When pH 


Figure 2 Effect of pH value on UV–VIS absorbance spectra of HYPE 

at concentration of 0.05 mg mL )1 . 

values ranged from 6.0 to 2.0, the intensity of absorbance 

at both Peak I and Peak II increased slowly. 

During this process, Peak I occurred hypsochromic shift 

from 265 to 254 nm, while Peak II was unchanged. The 

colour of HYPE solution remained yellow at pH from 

2.0 to 6.0. 

When pH values ranged from 7.0 to 9.0, the 

intensity of absorbance at both Peak I and Peak II 

increased greatly, and Peak II was bathochromic shift 

from 360 to 404 nm. Meanwhile, the colour of HYPE 

solution became brighter yellow when pH value was 

over 7.0. 

This phenomenon may be attributed to the structural 

transformation of HYPE. Flavonoids compounds were 

main components of HYPE obtained from HPLC and 

UV–VIS spectrophotometry analysis. To explain this 

phenomenon, rutin could be taken as a representative 

component for illustration. 

It was known that rutin changed into chalcone in 

alkaline condition, and chalcone contributed to yellow 

colour and bathochromic shift (Fig. 3). With the 

Figure 3 Structural transformation of rutin 

at different environmental conditions. 

(a) Rutin, (b) chalcone, (c) metal 

complexation, where R was rutinoside, 

M n+ was polyvalent metal ion. 

increasing of pH value, rutin mostly changed into 

chalcone, and the content of chalcone increased, so 

yellow colour of HYPE solution became brighter 

(Castaneda-ovando et al., 2009). 

Thermal stability of HYPE 

Relationship between thermal stability and temperature 

showed that the thermal degradation of HYPE at 

concentration of 0.05 mg mL )1 was in lined with firstorder 

reaction kinetics (Lopez-malo & Palou, 2008) 

(Fig. 4). Mathematical models and correlation coefficients 

R 2 were S =e )0.0033t (n =7,R 2 = 0.9975) for 

50 °C, S =e )0.0062t (n =7, R 2 = 0.9933) for 60 °C, 

S =e )0.011t 

(n =7, R 2 = 0.9879) for 70 °C, S = 

e )0.0183t (n =7,R 2 = 0.9849) for 80 °C, S =e )0.0273t 

(n =7,R 2 = 0.9708) for 90 °C (t: time stood for heat 

treatment, hours), respectively. 

The value of degradation rate k was an exponential 

increase with the temperature, which can be fitted 

into the exponential equation (eqn 2). The formula 

and correlation coefficient was k = 0.0002e 0.0531T 

(n =5, R 2 = 0.9929). The tendency of the formula 

shows that the rate of this reaction was not strongly 

temperature dependent within treatment temperature. It 

indicated that HYPE had good thermal stability within 

treatment temperature. 

As a result, mathematical model between S, time 

(t, hours), temperature (T, °C) was deduced as S ¼ 

e ð0:0002e0:0531T Þt . According to this mathematical model, 

thermal evolution of HYPE was simulated (Fig. 5). The 

half-life of HYPE solution was calculated by following 

formula. 

t ¼ 3465:74 

e0:0531T Here, t represented half-life of HYPE solution at 

concentration of 0.05 mg mL )1 , hours; T was the temperature 

of HYPE solution, used for heat treatment, °C. 

Half-life of HYPE solution was forecast to be 918.9 h 

at room temperature. 

(a) (b) 

Chemical stability of Huami yellow pigment W. Chen et al. 1669 


(c)

1670 


Figure 4 Thermal stability of Huaimi yellow pigment extract solutions 

at concentration of 0.05 mg mL )1 (S = A ⁄ A0, absorbance retention; 

LnS, natural logarithm of S). 

Influence of metallic ions on stability of HYPE 

The influence of metallic ions on stability of HYPE 

varied with the types of metallic ions. Metallic ions, 

including Na + , K + , Ca 2+ , Mg 2+ and Mn 2+ , all at 

concentration of 1.0 · 10 )2 m, almost had no effects on 

UV–VIS spectra and colour variations of HYPE solutions 

(Fig. 6). 

However, metallic ions, such as Al 3+ , Cu 2+ and 

Fe 3+ , at different concentrations had obvious effects on 

the UV–VIS spectra and colour variations of HYPE 

solutions. Al 3+ and Cu 2+ had similar effect on stability 

of HYPE (Figs 7 and 8). When the concentration of 

Al 3+ and Cu 2+ increased from 10 )5 to 10 )2 m, Peak II 

in UV–VIS spectra of HYPE solutions (at concentration 

of 0.05 mg mL )1 ) occurred bathochromic shift from 360 

to 406 nm and 409 nm, respectively, and the colour of 

HYPE solution became brighter. UV–VIS spectrum 

of HYPE no longer changed when ion concentration of 

Al 3+ and Cu 2+ reached to 10 )5 m. 

Effect of Fe 3+ on HYPE solution was not similar to 

that of Al 3+ and Cu 2+ (Fig. 9). With the increase in the 

concentration of Fe 3+ , absorbance at 360 nm of HYPE 

increased correspondingly. When concentration of Fe 3+ 

reached to 10 )3 m, absorption curve of HYPE was not 

smooth again, and the maximum absorption wavelength 

occurred hypsochromic shift. The colour of HYPE 

solution became to brown obviously. 

Effects of metallic ions on stability of HYPE could be 

contributed to co-pigmentation effect. Co-pigmentation 

effect is a common phenomenon between pigments and 

other colourless organic compounds, or metallic ions. 

This interaction formed new molecular complex by 

chemical bond transforming and then enhanced the 

colour of pigment (Gris et al., 2007; Castaneda-ovando 

et al., 2009). In food science, this phenomenon is 

considered as a very important interaction because the 

colour of a product is one of the main crucial quality 

factors for product’s acceptance (Eiro & Heinonen, 

2002). Some chemical groups, such as O-dihydroxy, 

which had metal chelating affinity, could form metal 

complex. Co-pigmentation between HYPE and metallic 

ions could be ascribed to this interaction (Fig. 3). At this 

condition, there was a change of the spectral characteristics 

of pigment molecules because of formation of the 

Figure 5 Simulation of thermal evolution of 

Huaimi yellow pigment extract (HYPE) 

solutions at concentration of 0.05 mg mL )1 

(S: absorbance retention, weighing the 

thermal stability of HYPE solutions). 


Figure 6 UV–VIS absorbance spectra of Huaimi yellow pigment 

extract solutions (0.05 mg mL )1 ) with all metallic ion (Na + ,K + , 

Ca 2+ ,Mg 2+ and Mn 2+ included) concentration at 10 )2 m. 


extract solutions (0.05 mg mL )1 ) with Al 3+ ion at different 

concentrations. Figure 9 UV–VIS absorbance spectra of Huaimi yellow pigment 

extract solutions (0.05 mg mL )1 ) with Fe 3+ ion at different concen- 

p–p* complex. The change influenced absorption intensity 

(hyperchromic effect) and wavelength (bathochromic 

shift) of HYPE. At the same time, the yellow colour 

of HYPE became brighter (Gris et al., 2007). 

Co-pigmentation caused by metals is sometimes 

beneficial for the maintenance and increment of colour 

of pigments in food pigment industry, particularly when 

these metals are not harmful to people’s health, and 

some of them are even essential minerals in human diet. 

Conclusions 

In this paper, chemical stability of HYPE from a kind of 

traditional Chinese herb and folk pigment material was 

studied systematically for the first time. HYPE displays 

perfect chemical stability. Results are consistent with its 

trations. 

Chemical stability of Huami yellow pigment W. Chen et al. 1671 


extract solutions (0.05 mg mL )1 ) with Cu 2+ ion at different 

concentrations. 

widespread usage in Chinese folk. The reason for 

widespread usage of Huaimi as a stable yellow pigment 

material in China could be explained in theory. In 

southern China, local residents usually added Huaimi 

into some traditional food to achieve brighter yellow. 

Some even added plant ash to obtain the same effect. This 

is a co-pigmentation phenomenon because plant ash 

contains potash, a kind of alkaline agent. The main 

components of HYPE belong to a large group of 

flavonoids compounds, especially the content of rutin 

in it is nearly 42.56%, which was in agreement with 

previous reports (Paniwnyk et al., 2001; Qi et al., 2007). 

Its chemical stability primarily depends on these compounds. 

On the other hand, flavonoids compounds are 


1672 


currently used as antioxidants in food industry, and it will 

undoubtedly promote the application of HYPE in future. 


We thank Professor Chen Fusheng from Huazhong 

Agricultural University for his constructive advice on 

experimental design of this study and Dr Zhou Youxiang 

from Hubei Academies of Agricultural Sciences in 

China for his aid to HPLC analysis. This work was 

supported by Undergraduate Technology Innovation 

Fund of Huazhong Agricultural University (No 

A07071). 

References 

Botelho, F.V., Alvarez-leite, J.I., Lemos, V.S. et al. (2007). Physicochemical 

study of floranol, its copper(II) and iron(III) complexes, 

and their inhibitory effect on LDL oxidation. Journal of Inorganic 


Cai, Y.Z., Sun, M. & Corke, H. (2003). Antioxidant activity of 

betalains from plants of the Amaranthaceae. Journal of Agricultural 


Castaneda-ovando, A., Pacheco-hernandez, M.D., Paez-hernandez, 

M.E., Rodriguez, J.A. & Galan-vidal, C.A. (2009). Chemical studies 

of anthocyanins: a review. Food Chemistry, 113, 859–871. 

De oliveira, I.R.W.Z., Fernandes, S.C. & Vieira, I.C. (2006). Development 

of a biosensor based on gilo peroxidase immobilized on 

chitosan chemically crosslinked with epichlorohydrin for determination 

of rutin. Journal of Pharmaceutical and Biomedical Analysis, 

41, 366–372. 

Dufosse, L. & Pintea, A. (2005). Pigments in food, more than colours. 

Trends in Food Science & Technology, 16, 368–369. 

Eiro, M.J. & Heinonen, M. (2002). Anthocyanin color behavior and 

stability during storage: effect of intermolecular copigmentation. 


El gharras, H., Hasib, A., Jaouad, A., El-bouadili, A. & Schoefs, B. 

(2008). Stability of vacuolar betaxanthin pigments in juices from 

Moroccan yellow Opuntia ficus indica fruits. International Journal of 


Galvano, F., La fauci, L., Lazzarino, G. et al. (2004). Cyanidins: 

metabolism and biological properties. Journal of Nutritional Biochemistry, 

15, 2–11. 

Gouveia, L., Nobre, B.P., Marcelo, F.M. et al. (2007). Functional 

food oil coloured by pigments extracted from microalgae with 

supercritical CO2. Food Chemistry, 101, 717–723. 

Gris, E.F., Ferreira, E.A., Falcao, L.D. & Bordignon-luiz, M.T. 

(2007). Caffeic acid copigmentation of anthocyanins from Cabernet 

Sauvignon grape extracts in model systems. Food Chemistry, 100, 

1289–1296. 

Guo, R., Wei, P. & Liu, W.Y. (2007). Combined antioxidant effects of 

rutin and vitamin C in Triton X-100 micelles. Journal of Pharmaceutical 

and Biomedical Analysis, 43, 1580–1586. 

Hurtado, N.H., Morales, A.L., Gonzalez-miret, M.L., Escudero-gilete, 

M.L. & Heredia, F.J. (2009). Colour, pH stability and antioxidant 

activity of anthocyanin rutinosides isolated from tamarillo fruit 

(Solanum betaceum Cav.). Food Chemistry, 117, 88–93. 

Kite, G.C., Stoneham, C.A. & Veitch, N.C. (2007). Flavonol 

tetraglycosides and other constituents from leaves of Styphnolobium 

japonicum (Leguminosae) and related taxa. Phytochemistry, 68, 

1407–1416. 

Lo, Y.H., Lin, R.D., Lin, Y.P., Liu, Y.L. & Lee, M.H. (2009). Active 

constituents from Sophora japonica exhibiting cellular tyrosinase 

inhibition in human epidermal melanocytes. Journal of Ethnopharmacology, 

124, 625–629. 

Lopez-malo, A. & Palou, E. (2008). Storage stability of pineapple slices 

preserved by combined methods. International Journal of Food 

Science and Technology, 43, 289–295. 

Lu, Y.H., Liu, Z.Y., Wang, Z.T. & Wei, D.Z. (2006). Quality 

evaluation of Platycladus orientalis (L.) Franco through simultaneous 

determination of four bioactive flavonoids by high-performance 

liquid chromatography. Journal of Pharmaceutical and 

Biomedical Analysis, 41, 1186–1190. 

Paniwnyk, L., Beaufoy, E., Lorimer, J.P. & Mason, T.J. (2001). The 

extraction of rutin from flower buds of Sophora japonica. Ultasonics 

Sonochemistry, 8, 299–301. 

Pflanz, M. & Zude, M. (2008). Spectrophotometric analyses of 

chlorophyll and single carotenoids during fruit development of 

tomato (Solanum lycopersicum L.) by means of iterative multiple 

linear regression analysis. Applied Optics, 47, 5961–5970. 

Qi, Y.Y., Sun, A.L., Liu, R.M., Meng, Z.L. & Xie, H.Y. (2007). 

Isolation and purification of flavonoid and isoflavonoid compounds 

from the pericarp of Sophora japonica L. by adsorption chromatography 

on 12% cross-linked agarose gel media. Journal of 

Chromatography A, 1140, 219–224. 

Stintzing, F.C. & Carle, R. (2004). Functional properties of anthocyanins 

and betalains in plants, food, and in human nutrition. Trends 

in Food Science & Technology, 15, 19–38. 

Wang, Z.L., Sun, J.Y., Wang, D.N., Xie, Y.H., Wang, S.W. & Zhao, 

W.M. (2006). Pharmacological studies of the large-scaled purified 

genistein from Huaijiao (Sophora japonica-Leguminosae) on antiosteoporosis. 

Phytomedicine, 13, 718–723. 

Xu, J.J., Zhang, H.Y. & Chen, G. (2007). Carbon nanotube ⁄ polystyrene 

composite electrode for microchip electrophoretic determination 

of rutin and quercetin in Flos Sophorae Immaturus. Talanta, 

73, 932–937. 

Yang, J.X., Guo, J. & Yuan, J.F. (2008). In vitro antioxidant 

properties of rutin. Lwt-Food Science and Technology, 41, 1060– 

1066. 

Zude, M., Spinelli, L. & Torricelli, A. (2008). Approach for nondestructive 

pigment analysis in model liquids and carrots by means 

of time-of-flight and multi-wavelength remittance readings. Analytica 

Chimica Acta, 623, 204–212. 




ATR-FTIR spectroscopy and chemometric analysis applied to 

discrimination of landrace maize flours produced in southern Brazil 

Shirley Kuhnen, 1 Juliana Bernardi Ogliari, 2 Paulo Fernando Dias, 3 Elisangela Fabiana Boffo, 4 Isabel Correia, 5 

Antônio Gilberto Ferreira, 4 Ivonne Delgadillo 5 & Marcelo Maraschin 3 * 

1 Laboratory of Biochemistry, Department of Chemistry – Federal University of Mato Grosso, 78060-900, Cuiabá-MT, Brazil 

2 Laboratory of Researches in Agrobiodiversity, Department of Plant Science – Federal University of Santa Catarina, PO Box 476, 88049-900, 

Floriano´polis-SC, Brazil 

3 Department of Zootechny and Rural Development Federal University of Santa Catarina, PO Box 476, 88049-900, Florianœpolis-SC, Brazil 

4 Laboratory of Nuclear Magnetic Resonance, Department of Chemistry, Federal University of Sa˜ o Carlos, PO Box 676, 13565-905, São Carlos- 

SP, Brazil 

5 Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal 

(Received 21 October 2009; Accepted in revised form 20 May 2010) 

Summary This work aims at discriminating flours of 26 maize landraces from southern Brazil, by using the Attenuated 

Total Reflection Fourier Transform Infrared (ATR-FTIR) spectroscopy and chemometrics (principal 

components analysis – PCA). PCA applied to the FTIR spectra in the 3-600 (whole spectrum) and 1650– 

1500 cm )1 (fingerprint region of proteins) spectral windows clearly discriminated the Amarelão landrace. 

Quantitative and semi-qualitative analysis of proteins showed a wide range among the fractions, mainly of 

prolamine (13.47–28.43 g Kg )1 ) and glutelin (5.57–30.98 g Kg )1 ) contents. Pixurum 6, Pixurum 5, and 

MPA1 landraces are of superior nutritional value for their albumin, globulin, and glutelin contents. PCA of 

the spectral dataset in the fingerprint region to carbohydrates (1200–950 and 1065–950 cm )1 ) also including 

commercial standards of amylose and amylopectin was able in separating the Moroti genotype, which 

grouped with the amylopectin standard. Thus, ATR-FTIR and PCA showed to be useful tools for the quick 

screening and discrimination of maize with distinct chemical composition. 

Keywords Amylopectin, amylose, ATR-FTIR spectroscopy, maize landraces, protein. 

Introduction 

Maize is one of the most important crop plants in the 

world. Most of this cereal produced worldwide is used as 

animal feed, although an important amount is also used 

in human diet, seeds, and for industrial purposes, 

providing raw material for the production of various 

kinds of food, pharmaceuticals, and cosmetics. The 

diversity of applications of that cereal depends on the 

differences in relative chemical composition, e.g. protein, 

oil, and starch contents in maize grains, traits that 

show prominent genetic components (White, 2001; Baye 

et al., 2006). 

Maize was domesticated in the highlands of Mexico 

about 10 000 years ago and subsequently spread to 

North and South America (Freitas et al., 2003). Over 

the last centuries, farmers have created thousands of 

*Correspondent: Tel: +55 48 3721 5328; Fax: +55 48 3721 5335; 

e-mail: m2@cca.ufsc.br 

doi:10.1111/j.1365-2621.2010.02313.x 


maize varieties suitable for cultivation in numerous 

environments. Accordingly, it seems consensual that the 

maize landraces’ phenotypes, e.g., morphological and 

agronomic traits and special chemical characteristics of 

grains are resultant of the domestication process. 

However, nowadays very few of the world’s maize 

germplasm consist of local varieties (landraces), showing 

the genetic vulnerability of that species. 

In some regions of the world, small farmers still 

cultivate maize varieties named locally creoles or landraces, 

also generating new varieties. This way, they are 

contributing to maintain the maize’s genetic diversity, 

preserving genetical ⁄ phenotypical traits that could be 

useful in yet unforeseen circumstances (Zeven, 1998). 

Therefore, it is important to provide efforts to assess the 

landrace germplasm aiming at identifying traits such as 

agronomic yield, disease and insect resistance, and 

value-added characteristics that could be incorporated 

into breeding programs, for instance. In the western 

region of the Santa Catarina State, southern Brazil

1674 

ATR-FTIR analysis of landrace maize flours S. Kuhnen et al. 

(Anchieta, Guaraciaba, and Novo Horizonte counties, 

e.g.), maize varieties with special agronomic and nutritional 

traits have been developed and cultivated by small 

farmers over the last decades. In those counties, 

Approximately 41% of the small rural establishments 

cultivate landrace genotypes, with a prominent use as 

human food (Canci et al., 2004). 

FTIR is a physicochemical method that measures the 

vibrations of bonds within functional groups and 

generates a spectrum that can be regarded as a metabolic 

fingerprint. It is a flexible method that can quickly 

provide qualitative and quantitative information with 

minimal or no sample preparation of complex biological 

matrices (Ferreira et al., 2001). Infrared (IR) was used 

in the past years to study amylose, amylopectin, and 

starch, providing information on e.g., chain folding and 

the ratio of ordered to amorphous fractions (van Soest 

et al., 1995; Sevenou et al., 2002), maize seed composition 

(Baye et al., 2006), and carotenoid content in maize 

genotypes (Berardo et al., 2004). 

A FTIR spectrum is complex, containing many 

variables per sample and making visual analysis very 

difficult. Hence, to extract useful information from the 

whole spectra, multivariate data analysis is needed, 

particularly through the determination of the principal 

components (PCA) (Fukusaki & Kobayashi, 2005). 

Such a multivariate analysis technique could allow the 

characterisation of the sample relationships (scores 

plans or axis) and the recovery of their subspectral 

profiles (loadings). This approach has been applied to 

classify maize starches (modified or not) and for 

quantification of sugars in mango juices as a function 

of ripening (Dupuy et al., 1997; Duarte et al., 2002) as 

well as for discriminating olive pulp-cell wall polysaccharide 

extracts of pectic or hemicellulosic origin 

(Coimbra et al., 1999). 

The goal of this study was to apply ATR-FTIR 

spectroscopy and chemometric analysis to discriminate 

flour samples produced from whole maize grains of 

twenty-six landraces developed and cultivated in southern 

Brazil. To perform that, we first identified the main 

wavenumbers of FTIR spectra that characterise those 

raw materials, followed by a more detailed analysis to 

distinguish them, focusing on the fingerprint region of 

specific classes of compounds. Taking into consideration 

the wide genetic variability of the maize landraces in 

study, it was hypothesised that a meaningful variation in 

chemical traits in flour samples derived from their 

(multi)coloured grains is thought to be found. 


Maize varieties 

Samples of grains of maize landraces (n = 26, 2004) 

cultured under agro-ecological management by small 

farmers in the far-west region of Santa Catarina State 

(Anchieta county – 26°31¢11¢’S, 53°20¢26¢’W), southern 

Brazil, were donated by a local small farmer association 

(SINTRAF) to the Agro-Biodiversity Study Nucleus 

(Federal University of Santa Catarina). Table S1 shows 

the main agronomic traits of the genotypes in study 

(Canci et al., 2004). 

This study was carried out in accordance with the 

current Brazilian legislation on biodiversity usage 

(Genetic Heritage Management Council – Provisional 

Act 2.186-16, August 23, 2001) and is part of an 

agreement involving small farmer communities of Santa 

Catarina State and the Federal University of Santa 

Catarina. 

Sample preparation 

Maize grain samples (50 g, dry weight, 13% humidity) 

were milled using Cyclone Sample Mill, Udy Corporation, 

Fort Collins, CO, USA (3010 ⁄ 019 model) with a 

1-mm sieve and stored in polypropylene bottles 

at )20 °C. Prior to collecting the spectra, the maize 

flours (5 g) and commercial standards of amylose (Sigma 

Chemical Co, St. Louis, MO, USA) and amylopectin 

(Fluka Chemical Co., St. Louis, MO, USA) were 

lyophilised and kept over phosphorous pentoxide until 

further spectroscopic analysis. 

ATR-FTIR spectroscopy and chemometrics 

The theory of ATR was previously discussed by 

Harrick (1967). In ATR mode, an IR beam traverses 

a prism so that it is internally reflected from the back 

of the prism, which is in contact with the sample. The 

conditions for total internal reflection occur only when 

sin hi ⁄ n2 ⁄ n1, withhithe angle of incidence on the back 

of the prism, n1 the index of refraction of the prism 

material (diamond in this case), and n2 the index of 

refraction of the sample. In this situation, an evanescent 

wave penetrates the sample and is able to interact 

with the sample, just beyond the interface prismsample, 

leading to absorption of radiation from the 

beam. The intensity of the absorption depends first, on 

the good contact between the sample and the prism 

and secondly, on the depth of penetration of the 

evanescent wave. 

According to Harrick (1967), the penetration depth 

is directly related to the wavelength so that the larger 

the wavelength, the greater the penetration of the 

wave. For example, polysaccharides like starch absorb 

in the region 1200–800 cm )1 (k = approximately 8 and 

12 lm), in an average penetration depth of approximately 

2 lm. Thus, attenuated total reflectance (ATR) 

is often considered as a surface-sensitive method 

because of its inherent low depth of penetration 

(Sevenou et al., 2002). 


FTIR spectra of maize flours and of standards of 

amylose and amylopectin were acquired in a Bruker 

IFS-55 spectrometer with a DTGS detector equipped 

with a golden gate single reflection diamond ATR 

accessory (45° incidence-angle). A background spectrum 

of the clean crystal was acquired, and samples (100 mg) 

were spread and measured directly after pressing them 

on the crystal. The spectra were recorded at the 

absorbance mode from 3000 to 600 cm )1 at the resolution 

of 4 cm )1 . Five replicate spectra (128 co-added 

scans before Fourier transform) were collected for each 

sample, in a total of 130 spectra. 

Spectra were normalised, baseline-corrected in the 

region of interest by drawing a straight line before 

resolution enhancement (k factor of 1.7) was applied 

using Fourier self deconvolution (Opus v. 5.0, Bruker 

Biospin – van Soest et al., 1995; Rubens et al., 1999). 

The assumed line shape was Lorentzian with a half 

width of 19 cm )1 . IR absorbance values at 1045, 1018, 

and 999 cm )1 were extracted from the spectra after 

baseline correction and deconvolution. Intensity measurements 

were performed on the deconvulated spectra 

by recording the height of the absorbance bands from 

the baseline. 

Chemometric analysis 

For purpose of chemometric analysis, normalised, 

baseline-corrected (3000–600 cm )1 , 1650–1550 cm )1 – 

proteins, 1200–950 cm )1 – carbohydrates, and 1065– 

950 cm )1 – amylose and amylopectin) and deconvoluted 

spectra were transferred via a JCAMP.DX format 

(OPUS v. 5.0, Bruker Biospin GmbH, Rheinstetten, 

Baden-Wu¨rttemberg, Germany) into the data analysis 

software for PCA (The Unscramble v. 9.1, CAMO 

Software Inc., Woodbridge, NJ, USA). Previously to 

PCA analysis each spectrum within the 3000–600 cm )1 , 

1650–1550 cm )1 (proteins), 1200–950 cm –1 (carbohydrates), 

and 1065–950 cm )1 (amylose and amylopectin) 

regions was standard normal deviates corrected 

(Cˇ opı´kova´ et al., 2006). 

Protein analysis – chemical extraction 

The extraction of proteins from flours was performed 

using a sequential scheme based on the Osborne & 

Mendel (1914) method, modified by Landry & 

Moureaux (1970). Samples (100 mg) were suspended in 

1 mL of extractant solution; the incubation time (min), 

the number of extractions performed, and the identification 

of protein fractions obtained are reported on 

Table 1. All the experiments were carried out in triplicate. 

The protein content of each fraction was determined 

according to Bradford (1976) using a standard 

curve of bovine serum albumin (BSA, globulins: 0.7– 

0.01 mg mL )1 – r 2 0.96, albumins: 0.1–0.01 mg mL )1 – 

ATR-FTIR analysis of landrace maize flours S. Kuhnen et al. 1675 

Table 1 Protein fraction, extractant, and incubation conditions used 

for the fractionated extraction of proteins of maize flours 

Protein fraction Extractant Incubation conditions 

Albumin Distilled water 4 °C, 2 times, 15 min 

Globulin 0.5 M NaCl 4 °C, 2 times, 30 min 

Zein I 55% (w ⁄ w) 2-PrOH 

+ 0.6% 2-ME 

Zein II 0.5 M NaCl + 0.6% 

2-ME (pH 10) 

Glutelin 0.5% SDS (w ⁄ v, pH 10) 

+ 0.6 2-ME 

r 2 0.99; prolamines: 1–0.01 mg mL )1 – r 2 0.94; glutelins: 

1–0.1 mg mL )1 – r 2 0.97). 


RT, 2 times, 30 min 

and 1 time, 15 min 



and 1 time, 15 min 

The samples were centrifuged at 12000 g for 5 min and the precipitated 

obtained used as biomass for the following extraction. 

2-ME: 2-mercaptoethanol; SDS, sodium dodecylsulfate; RT, room 

temperature. 

Figure S1 shows the FTIR spectra of flours for 26 

varieties with the majoritarian absorption peaks 

detected at 3000–600 cm )1 spectral window. The FTIR 

spectra of maize flours look very similar and this fact 

turns any discrimination analysis based on the chemical 

traits of the genotypes a hard task. Visual inspection of 

the FTIR spectra promptly indicated the presence of 

starch (1200–800 cm )1 ), lipids (2924, 2854, and 

1740 cm )1 ), and proteins (1650–1500 cm )1 ) in all the 

samples, while minor differences, quantitative or structural, 

were difficult to detect. Accordingly, to extract 

maximum information of the FTIR spectra, principal 

component analysis (PCA) was applied to the spectroscopic 

dataset. 

Figure S2 shows a PCA scores scatter plot for flour 

samples using the whole FTIR spectral window 

dataset (3000–600 cm )1 ). The classification of flours 

in scores scatter plot (PC1 vs. PC2) that contains 78% 

of the dataset variability revealed differences in chemical 

composition, mainly for the Amarela˜ o 3 genotype, 

which was separated in PC1()) axis. Figure 1 shows 

the loadings plot of PC1, revealing the most important 

wavenumbers which explain the distinction of the 

samples previously found (scores scatter plot). The 

loadings indicated that the carbohydrates (1022, 1076, 

and 1149 cm )1 ), proteins (1647, 1535, and 1519 cm )1 ), 

and lipids (2920, 2850, and 1743 cm )1 ) can be associated 

to the discrimination of the maize flours 

observed. Both genetic and environmental effects 

create significant variation in the amount and quality 

of each of these constituents (Baye et al., 2006), but 

the variations on the metabolic profile of the landrace 

maize grains herein found reflect mainly genotypic 


1676 


differences, becausethe plants were cultivated using 

similar agronomic traits in the same geographic area, 

climate, and type of soil. In a previous study, 

differences on the chemical composition were detected 

in flours of the same maize landraces regarding the 

contents of proteins (8–10.3%), lipids (4.8–5.4%), 

starch (63–73%), and linoleic acid (43.2–50.8%) 

(Steinmacher et al., 2005). Besides, Kuhnen (2007) 

reports a clear discrimination among flour samples of 

whole ⁄ degermed-grains for the twenty-six maize landraces 

herein studied by using PCA analysis (PC1 81% 

vs. PC12%) of the whole FTIR dataset (3000– 

600 cm )1 ), with a prominent effect of the lipid 

components (2924, 2850, and 1743 cm )1 ) for the 

segregation observed. 

Although the whole spectra provide a good starting 

point for an exploratory data analysis as in a metabolic 

fingerprint approach, it is often necessary to analyse 

specific regions of the spectra to understand the 

discrimination obtained by PCA. Different classes of 

compounds produce a complex and unique pattern in 

the lower region of the spectra (

Principal component analysis – Fingerprint region of 

carbohydrates 

All the spectra show absorption bands in the 1300– 

800-cm )1 region which are sensitive and related to the 

structure of starchy polysaccharides, i.e., amylose and 

amylopectin (van Soest et al., 1995; Rubens et al., 

1999). The bands in this region of the IR spectrum 

result mainly from C-O and C-C vibrational modes that 

are highly coupled. This coupling makes the assignment 

of individual bands difficult as well as a discriminative 

analysis by direct inspection of the spectra. As the 

absorptions in this region arise largely from C–O 

stretchings of the ring, linkages (C–O–C) and COH 

groups, the positions of these bands are similar in all 

carbohydrates as the associated vibrational modes are 

unlikely to be radically affected by polymer formation 

(Belton et al., 1986). 

The PCA analysis of the spectral data for the carbohydrate 

fingerprint region, i.e., 1200–850 cm )1 , was 

performed including the FTIR dataset of amylose and 

amylopectin standards. The chemical structure and the 

purity of the standards of amylose and amylopectin were 

confirmed by 13 C NMR spectroscopy. The 13 C NMR 

spectrum of amylose typically revealed resonances at 

100.6 (C1), 78.8 (C4), 73.2 (C3), 72.0 (C2), 71.6 (C5), and 


Figure 2 Principal component analysis scores scatter plot of the FTIR dataset in the spectral window of 1650–1500 cm )1 wavenumber (fingerprint 

region of proteins) of landrace maize flours cultivated in the southern Brazil. PC1 · PC2 axes cross each other at the origin. 

60.5 ppm (C6) assigned to carbons of a-d-glucose 

residues. The same characteristic resonances of a-dglucose 

residue were detected in the 13 C NMR spectrum 

of amylopectin, as well as at 100.9 (C1 and C6 of 

branching point), 79.5 (C4 of non-reducing end group 

residue), and 61.0 ppm (C1 and C6 of branching point) 

(Qit et al., 2003). The PC1 (76%) vs. PC2 (14%) scores 

scatter plot (data not showed) allowed a clear segregation 

of the commercial standards of starchy polysaccharides, 

but a poor discrimination for the maize flours samples, 

except for the Moroti variety located near to the 

amylopectin standard, suggesting the predominance of 

that polysaccharide in relation to amylose in its starchy 

fraction. Besides, no variety grouped with the amylose 

standard. Our findings suggest that regarding the starch 

composition the maize genotypes cultivated in Anchieta 

county seem not to be high-amylose type (60–70% 

amylose). The average amylose content in normal maize 

is about 25%, but this value can vary (20–36%) among 

varieties (more details in White, 2001). Studies aiming at 

to determine the amylose and amylopectin contents of 

those maize genotypes are being performed and will be 

published elsewhere. The variable loadings correlated 

with PC1 were the bands at 1018 and 999 cm )1 , being 

these wavenumbers related to the distinction of amylose 

and amylopectin, respectively, in maize flour samples. 


1678 


2.0 

1.5 

1.0 

0.5 

0.0 

1100 

Amylose 

1018 cm –1 

1050 

Branco 

Previous studies have shown that some IR bands are 

highly sensitive to polymer conformation (Wilson & 

Belton, 1988). The bands at 1047 and 1022 cm )1 are 

sensitive to the amount of ordered or crystalline and 

amorphous starch, respectively (van Soest et al., 1995; 

Sevenou et al., 2002; Bernazzani et al., 2008). On the 

present work, the amylose spectrum shows the most 

intense band wavenumber at 1018 cm )1 and not at 

1022 cm )1 . However, according to van Soest et al. 

(1995), the position of this band (1022 cm )1 ) can shift 

towards 1015 cm )1 according to the water amouts. This 

observation can explain the absence of band at 

1018 cm )1 in the amylopectin standard because the 

short chains of this polymer are responsible for the 

crystallinity of the native starch granule. 

The differences between the IR spectra of amylose and 

amylopectin standards with respect to the bands at 1018 

and 999 cm )1 , respectively, are clearly observed upon 

visual inspection as shown in Fig. 3. These two bands 

were detected in all flour samples but with different 

intensities. Moroti variety spectrum showed the band at 

999 cm )1 with maximum intensity, similarly to the 

observed in the amylopectin spectrum, reinforcing the 

results found in PCA analysis (1200–850 cm )1 region). 

The band at 1018 cm )1 was lesser intense than the 

band at 999 cm )1 for all the flour samples but for 

the Branco, Pixurum 6, Pixurum 5, and Pixurum 4 

varieties, where the two bands showed similar intensities 

(Fig. 3). 

In the past, IR was used to study amylose, amylopectin, 

and starch providing insights on the ratio of 

ordered to amorphous fractions (van Soest et al., 1995). 

Thus, as a complementary approach to better characterise 

the polysaccharide fraction of the maize flours in 

study, the variation among starchy constituents was 

interpreted in terms of the level of ordered structure 

present on the edge of the flour grains. For that purpose, 

Pix6 

Amylopectin 

999 cm –1 

Moroti 

1000 

Wavenumber (cm –1 ) 

950 

900 

Figure 3 Partial FTIR spectra (1120–880 

cm )1 ) of maize flours Moroti, Branco, and 

Pixurum 6 (Pix6) varieties, commercial amylose 

(Sigma), and commercial amylopectin 

(Fluka, Sigma). 

the estimation of the starch crystallinity in the maize 

flour samples was calculated by the IR ratio 1018 ⁄ 

999 cm )1 and 1045 ⁄ 1018 cm )1 as previously suggested 

Table 2 Infrared (IR) ratio of the absorbances 1018 ⁄ 999 cm )1 

for the flours of different maize landraces 

Samples 

IR ratio 

1018 ⁄ 999 

(cm )1 ) 

Amylopectin 0.93 

Moroti 0.94 

Palha Roxa 2 0.97 

Amarelão 3 0.97 

Roxo 29 0.97 

Rosado 0.97 

MPA 2 0.97 

MPA 13 0.98 

Composto São Luiz 0.98 

Pires 0.98 

Rajado 8 Carreiras. 0.98 

Mato Grosso Palha Roxa 0.98 

Cateto Vermelho 0.98 

Cateto 0.98 

Roxo 41 0.98 

Mato Grosso 0.98 

Língua Papagaio 0.98 

MPA 1 0.98 

Pixurum 1 0.98 

Cunha 0.98 



Asteca 0.99 



Palha Roxa 18 0.99 

Branco 0.99 

Amylose 1.08 


(Rubens et al., 1999; Sevenou et al., 2002). Intensity 

measurements were performed on the deconvoluted 

spectra by recording the heights of the absorbance 

bands from the baseline. As shown on Table 2 from 

the IR ratio R(1018 ⁄ 999 cm )1 ), the amylose and 

amylopectin standard samples were clearly distinguished 

together with some maize flour samples. The 

values of the absorbance ratios for Moroti variety 

(0.94) and amylopectin (0.93) were similar, whereas 

for Asteca, Pixurum 4, Pixurum 5, Pixurum 6, Palha 

Roxa 18, and mainly Branco genotypes (0.99) 

were related to amylose (1.08). On average, IR ratios 

of the absorbances 1018 ⁄ 999 cm )1 (0.98) and 1045 ⁄ 

1018 cm )1 (0.73 – data not shown) found for the maize 

flours in study were identical to previously observed by 

Sevenou et al. (2002) for maize starch, i.e., 0.98 and 

0.75, respectively. 

The ATR mode is a technique in which the IR beam 

penetrates into the first few micrometres (approximately 

2lm) of the sample (Sevenou et al., 2002). This penetration 

depth is smaller than the average size of the 

maize flour granules used, implying that the IR spectra 

acquired are representative of the external part of the 

flour granules. Thus, taking into consideration that 

ATR-FTIR is a surface analytical method that acquires 

information on the outer region of a sample and that 


Figure 4 Principal component analysis scores scatter plot of the FTIR dataset in the 1065–950 cm )1 spectral window (fingerprint region of 

carbohydrates) of landrace maize flours and commercial standards of amylose and amylopectin. PC1 · PC2 axes cross each other at the origin. 

the study was not performed with isolated starch from 

the maize landraces, we assume that the differences 

in the IR spectra among the flour samples are not 

related to the organisation of the starch’s growth rings. 

Besides, the differences found by ATR-FTIR between 

the starchy components of the maize landraces flour was 

not related to long-range order and only related to 

variations in the ratio of the amounts of ordered 

to unordered fraction within the starchy polysaccharides 

in the flour granules. 

Finally, the principal components were calculated 

using FTIR data from the 1065–950 cm )1 spectral 

window, a lower region of the spectrum which contains 

only the characteristic bands of amylopectin and amylose 

as shown herein. Again, a clear distinction among 

the flour samples and amylose standard emerged 

(Fig. 4), indicating that none of the studied samples is 

high-amylose type by ATR-FTIR ⁄ PCA. On the other 

hand, PC1 (82%) vs. PC2 (11%) showed similarity 

between amylopectin standard and the Moroti and 

Roxo 41 varieties, suggesting the predominance of 

amylopectin in the constitution of their starches. In 

fact, a previous study showed that grains of Moroti 

landrace are classified into the category of farinaceous 

(Canci et al., 2004), while Roxo 41 was still not 

evaluated. The PC2 loadings of the FTIR spectral data 


1680 


matrix (data not shown) confirmed the signals at 1018 

and 999 cm )1 as diagnostic wavenumbers for the characterisation 

of amylose and amylopectin, respectively, in 

the studied samples. Our findings indicate that the 1065– 

950 cm )1 region showed itself to be more suitable to 

segregate starchy polysaccharides because more samples 

have been discriminated in this analysis when compared 

to 1200–850 cm )1 spectral window. 

In conclusion, the findings on this work reveal that the 

chemical diversity is present in the flour samples of 

maize landraces cultivated in southern region of Brazil. 

By using ATR ⁄ FT-IR ⁄ PCA and chemical analysis of 

proteins, it was detected that Amarela˜ o 3 landrace is the 

most diverse among genotypes. Besides, the protein 

analysis of the flours revealed to be of higher nutritional 

value the Pixurum 6, Pixurum 5, and MPA 1 landraces 

for containing superior amounts of globulin and 

albumin fractions with higher contents of lysine and 

tryptophan. 

The PCA analysis of the IR dataset of the fingerprint 

region of carbohydrates suggested that the flours of 

maize landraces cannot be diagnosed as high-amylose or 

waxy types, except for the Moroti and Roxo 41 varieties 

that seem to contain starch granules with superior 

amylopectin amount in respect to their amylose content. 

This finding was further confirmed regarding to the 

similarity of values found for the IR absorbance 

ratio 1018 ⁄ 999 cm )1 of the Moroti variety (0.94) and 

amylopectin (0.93). 

Conclusion 

ATR-FTIR and PCA showed to be useful tools for the 

study of the local maize diversity in southern Brasil, 

allowing a fast screening of landrace maize flours with 

distinct chemical composition, i.e., amylose ⁄ amylopectin 

ratio and protein composition. The described 

method is simple, fast, and low cost, being also a 

powerful tool to be used in quantitative assays after its 

validation. Further, the identification of interesting 

chemical traits in flours of maize landraces as herein 

shown has been claimed as a strategy to address better 

usages to those raw materials as, for instance, components 

of special foods with superior nutritional ⁄ dietary 

quality, allowing to add value to them. 


The authors are grateful to the small farmer association 

(SINTRAF) for providing the landrace maize seeds. 

References 

Baye, T.M., Pearson, T.C. & Settles, A.M. (2006). Development of a 

calibration to predict maize seed composition using single kernel 

near infrared spectroscopy. Journal of Cereal Science, 43, 236–243. 

Beekes, M., Lasch, P. & Naumann, D. (2007). Analytical applications 

of Fourier transform-infrared (FTIR) spectroscopy in microbiology 

and prion research. Veterinary Microbiology, 123, 305–319. 

Belton, P., Wilson, R.H. & Chenery, D.H. (1986). Interaction of group 

1 cations with iota and kappa carrageenans studied by Fourier 

transform infrared spectroscopy. International Journal of Biological 

Macromolecules, 8, 247–251. 

Berardo, N., Brenna, O.V., Amato, A., Valoti, P., Pisacane, V. & 

Motto, M. (2004). Carotenoids concentration among maize genotypes 

measured by near infrared reflectance spectroscopy (NIRS). 

Innovative Food Science and Emerging Technologies, 5, 393–398. 

Bernazzani, P., Peyyavula, V.K., Agarwal, S. & Tatikonda, R.K. 

(2008). Evaluation of the phase composition of amylose by FTIR 

and isothermal immersion heats. Polymer, 49, 4150–4158. 

Bradford, M.M. (1976). A rapid and sensitive method for the 

quantitation of microgram quantities of protein utilizing the 

principle of protein-dye binding. Analytical Biochemistry, 72, 248– 

254. 

Canci, A., Vogt, G.A. & Canci, I. (2004). A Diversidade das Espećies 

Crioulas em Anchieta-SC: Diagno´stico, Resultado de Pesquisa e 

Outros Apontamentos para a Conservaçaõ da Agrodiversidade. Sa˜ o 

Miguel do Oeste, Brasil: McLee. 

Coimbra, M.A., Barros, A.S., Rutledge, D.N. & Delgadillo, I. (1999). 

FTIR spectroscopy as a tool for the analysis of olive pulp-cell wall 

polysaccharide extracts. Carbohydrate Research, 317, 145–154. 

Cˇ opíková, J., Barros, A.S., Sˇ midová, I. et al. (2006). Influence of 

hydration of food additive polysaccharides on FT-IR spectra 

distinction. Carbohydrate Polymers, 63, 355–359. 

Duarte, I.F., Barros, A., Delgadillo, I., Almeida, C. & Gil, A.M. 

(2002). Application of FTIR spectroscopy for the quantification of 

sugars in mango juice as a function of ripening. Journal of 


Dupuy, N., Wojciechowski, C., Ta, C.D., Huvenne, J.P. & Legrand, P. 

(1997). Mid-infrared spectroscopy and chemometrics in corn starch 

classification. Journal of Molecular Structure, 410–411, 551–554. 

Ferreira, D., Barros, A., Coimbra, M.A. & Delgadillo, I. (2001). Use 

of FTIR spectroscopy to follow the effect of processing in cell wall 

polysaccharide extracts of a sun-dried pear. Carbohydrate Polymers, 

45, 175–182. 

Freitas, F.O., Bendel, G., Allaby, R.G. & Brown, T.A. (2003). DNA 

from primitive maize landraces and archaeological remains: implications 

for the domestication of maize and its expansion into South 

America. Journal of Archaeological Science, 30, 901–908. 

Fukusaki, E. & Kobayashi, A. (2005). Plant metabolomics: potential 

for practical operation. Journal of Bioscience and Bioengineering, 

100, 347–354. 

Harrick, N.J. (1967). Internal Reflection Spectroscopy. New York, 

USA: John Wiley & Sons Inc. 

Kuhnen, S. (2007). Metaboloˆmica e Bioprospecçaõ de Variedades 

Crioulas e Locais de Milho (Zea mays). Ph.D. Thesis, Florianópolis, 

Brasil: Universidade Federal de Santa Catarina. 

Landry, J. & Moureaux, T. (1970). Heterogeneite des glutelines du 

grain de mais: extraction selective et composition em acides amines 

dês trois fraction isolees. Bulletin de la Societe´ Chimique Biologique, 

52, 1021–1037. 

Osborne, T.B. & Mendel, L.B. (1914). Nutritive properties of proteins 

of the maize kernel. The Journal of Biological Chemistry, 18, 1–6. 

Qit, X., Tester, R.F., Snape, C.E. & Ansell, R. (2003). Molecular basis 

of the gelatinisation and swelling characteristics of waxy rice 

starches grown in the same location during the same season. Journal 

of Cereal Science, 37, 363–376. 

Rubens, P., Snauwaert, J., Heremans, K. & Stute, R. (1999). In situ 

observation of pressure-induced gelation of starches studied with 

FTIR in the diamond anvil cell. Carbohydrate Polymers, 39, 231– 

235. 

Sevenou, O., Hill, S.E., Farhat, I.A. & Mitchell, J.R. (2002). 

Organisation of the external region of the starch granule as 


determined by infrared spectroscopy. International Journal of 

Biological Macromolecules, 31, 79–85. 

van Soest, J.J.G., Tournois, H., Wit de, D. & Vliegenthart, J.F.G. 

(1995). Short-range structure in (partially) crystalline potato starch 

determined with attenuated total reflectance Fourier-transform IR 

spectroscopy. Carbohydrate Research, 279, 201–214. 

Steinmacher, N.C., Rayasduarte, P., Alves, A.C. & Francisco, A. 

(2005). Rheological properties of Brazilian maize landrace flours. 

Proceedings of the AACC Annual Meeting. Pp. 1–13. Orlando, USA: 

AACC Press. 

Vasal, S.K. (2001). High quality protein corn, In: Specialty Corns 

(edited by A.R. Hallauer), Chapter 4. Pp. 85–129. Boca Raton, FL, 

USA: CRC Press LLC. 

White, P.J. (2001). Properties of corn starch, In: Specialty Corns 

(edited by A.R. Hallauer), Chapter 2. Pp. 29–54. Boca Raton, FL, 

USA: CRC Press LLC. 

Wilson, R.H. & Belton, P.S. (1988). A Fourier-transform infrared 

study of wheat starch gels. Carbohydrate Research, 180, 339–344. 

Zeven, A.C. (1998). Landraces: a review of definitions and classifications. 

Euphytica, 104, 127–139. 




Figure S1. ATR-FTIR spectra (3000–600 cm )1 ) of 

flours produced from whole grains of maize creole 

varieties cultivated in southern Brazil. The bands in 


between 1200 and 800 cm )1 , 1650 and 1500 cm )1 , 2924, 

2854, and 1740 cm )1 indicate the presence of carbohydrates, 

proteins, and lipids, respectively. 

Figure S2. Principal component analysis scores 

scatter plot of the FTIR spectra of maize flours in 

the 3000–600 cm )1 wavenumber region. The Amarela˜ o 

3 variety (Am3) is clearly shown to be discriminated by 

PC1 ()) axis. PC1 · PC2 axes cross each other at the 

origin. 

Figure S3. Percentages of the protein fractions globulin 

and albumin (soluble in water and saline solution), 

prolamines, and glutelins extracted from flours samples 

of Zea mays landraces originated from southern Brazil. 

Numbers in top of bars are referred to the total protein 

content (g Kg )1 ). 

Table S1. Grain colour and type, cultivation cycle, 

plant height, number of cobs per plant, and productivity 

of the studied maize varieties cultivated in southern 

Brazil. 







1682 


Effect of processing on the amino acid content of New Zealand 

spinach (Tetragonia tetragonioides Pall. Kuntze) 

Jacek Słupski, Jacek Achrem-Achremowicz, Zofia Lisiewska* & Anna Korus 

Department of Raw Materials and Processing of Fruit and Vegetables, Faculty of Food Technology, Agricultural University of Kraków, Balicka 

122, 30-149 Kraków, Poland 


Summary The aim of this work was to evaluate the effect of processing on the amino acid content and protein quality of 

New Zealand spinach (Tetragonia tetragonioides Pall. Kuntze). In this research, fresh and cooked New 

Zealand spinach as well as two frozen products prepared for consumption, one obtained using the traditional 

method (blanching–freezing–storage–cooking) and the other a convenience food product obtained using a 

modified process (cooking–freezing–storage–defrosting and heating in a microwave oven), were analysed. 

Glutamic acid was the dominant amino acid in fresh New Zealand spinach, and the limiting amino acids 

were cystine and methionine. Technological and culinary processing caused a significant increase in amino 

acid content in 100 g of edible portion, with the exception of methionine and cystine in frozen products 

prepared for eating. Changes in amino acid content expressed in g ⁄ 16 g of N (which corresponded to 100 g 

tissue protein) were not significant, with the exception of the lower glutamic acid content in the frozen 

product obtained using traditional processing method. 

Keywords Amino acids, New Zealand spinach (Tetragonia tetragonioides Pall. Kuntze), technological and culinary processing. 

Introduction 

In contrast to spinach (Spinacia oleracea L.), New 

Zealand spinach (Tetragonia tetragonioides Pall. Kuntze, 

Aizoaceae) is a warm climate plant (Kawashima & 

Valente Soares, 2003). However, field experiments carried 

out in Southern Poland (Krako´w region) proved 

that this valuable vegetable could successfully be cultivated 

in the temperate climate zone (Jaworska & 

Kmiecik, 1999, 2000; Kmiecik & Jaworska, 1999). 

According to Ge˛bczyn´ski (2008), New Zealand spinach 

is suitable for processing both as a traditional frozen 

product and as a convenience food, the use of which 

becomes more and more popular because of the reduced 

preparation time. Moreover, products prepared from 

this vegetable, including frozen product prepared using 

the modified technology (convenience food), are characterised 

by high antioxidant content and sensory value 

(Ge˛bczyn´ski, 2008). It is recommended to use vegetables 

because they are valuable sources of vitamins, mineral 

salts and organic acids. In addition, they are rich in 

antioxidants. Although most research on protein content 

has focused on vegetables rich in proteins, such as 


e-mail: rrlisiew@cyf-kr.edu.pl 


legumes, other vegetables containing low protein could 

also serve as an additional source of amino acids in 

meat-free diets. Jaworska indicated that protein content 

in New Zealand spinach was 27.5–30.6 g in 100 g of dry 

matter, depending on the season, planting and harvesting 

times and the part used (Jaworska & Kmiecik, 1999). 

Amino acids are vulnerable to the effects of culinary or 

technological processing. 

There are few reports available on the effects of 

culinary and processing technology on amino acid 

content in foods, and as far as the authors know, there 

are no data concerning this question for New Zealand 

spinach. 

Amino acids play an important role for human 

nutrition not only as precursors for the synthesis of 

proteins and other nitrogen containing compounds, but 

also as key molecules that regulate some of major 

metabolic pathways (Meijer, 2003).The aim of this work 

was the assessment of changes in amino acid content in 

New Zealand spinach products caused by cooking, 

12 months’ frozen storage and preparing them for 

eating. Included for analysis were fresh and cooked 

spinach and two frozen products prepared for 

consumption using two different processes. The first 

product was obtained using the traditional method of 

blanching, freezing and, after 12 months’ frozen storage, 

doi:10.1111/j.1365-2621.2010.02315.x 


cooking in brine to the required completion; the other 

by a modified method in which the spinach was cooked 

in brine until done, then frozen, kept in frozen storage 

for 12 months and finally defrosted and heated in a 

microwave oven prior to analysis. 


The investigated material consisted of fresh leafy shoots 

of New Zealand spinach (Tetragonia tetragonioides Pall. 

Kuntze) 15–17 cm in length as the raw material, cooked 

until done, and frozen products prepared for consumption 

after 12 months of frozen storage at )20 °C. New 

Zealand spinach was grown in the experimental field of 

the Department on the western outskirts of Krako´w 

(50°08¢N, 19°86¢E). The soil was in good horticultural 

condition with neutral pH and high content of potassium, 

phosphorus and calcium. The application of 

mineral fertilisers was determined by the fertility of soil 

and the nutritional requirements of the investigated 

species. The fertilisation was N, 90 kg ha )1 ;P2O5,80 kg ha )1 ;K2O, 100 kg ha )1 . Cultivation included mechanical 

weed control, sprinkler watering and protection 

against diseases and pests. The spinach was harvested in 

mid-July. A mean sample representing the whole batch 

was taken for analysis. The leaves of New Zealand 

spinach were cut into sections 7–8 cm in length and then 

processed using two different methods. 

Production of frozen products 

Two variants were used in preparing the raw material for 

freezing. In variant I, New Zealand spinach was 

blanched in stainless steel vessel containing water (1:5, 

solid to liquid ratio) at 95–98 °C for 2 min 15 s. These 

conditions permitted a decrease in the activity of catalase 

and peroxidase to the levels below 5% of the initial 

values. After blanching, the material was immediately 

cooled in cold water and drained on sieves for 30 min. 

In variant II, New Zealand spinach was cooked in a 

stainless steel vessel in 2% brine (1:5, solid to liquid 

ratio) until done (4 min). After cooking, the cut leaves 

were drained, placed in sieves and cooled in a stream of 

cold air. 

In both variants, the blanched or cooked spinach was 

packed in 500 g portions in polyethylene bags and 

frozen at )40 °C in blast chamber (Feutron 3626-51; 

Feutron Klimasimulation GmbH, Langenwetzendorf, 

Germany) for 90 min. The obtained products were 

stored at )20 °C for 12 months. 

Preparation of frozen product for evaluation 

Frozen product blanched before freezing (variant I) was 

cooked in 2% brine; the proportion of brine to plant 

material was 1:1 (w ⁄ w). The time of cooking, measured 

Amino acids in New Zealand spinach J. Słupski et al. 1683 

from the moment when the brine was boiling again, was 

2 min. After cooking, the spinach was immediately 

drained and cooled to 20 °C for analyses of chemical 

composition. 

Frozen product cooked before freezing (variant II) 

was prepared by placing 500 g of plant material in a 

heat-resisting vessel covered with a lid, defrosted and 

heated in Panasonic Ò microwave kitchen (type NN- 

F621; Matsushita Electric, Cardiff, UK). The time of 

defrosting and heating to 75 °C was 7 min 45 s accordingly 

to Codex Alimentarius (1993). 

Analytical procedures 

The moisture content and total N were determined 

according to procedures described by the Association of 

Official Agricultural Chemists (AOAC 1984). The content 

of amino acids was determined using an AAA-400 

automatic amino acid analyser (INGOS, Prague, Czech 

Republic). The freeze-dried material was hydrolysed in 

6 m HCl for 24 h at 110 °C. After cooling, filtering and 

washing, the hydrolyte was evaporated in a vacuum 

evaporator (below 50 °C for sulphur-containing amino 

acids and below 60 °C for the other amino acids), and 

the dry residue was dissolved in a buffer of pH 2.2. The 

prepared sample was analysed using the ninhydrin 

method. The ninhydrin solution was buffered at pH 

5.5. A column 370 mm in length was filled with Ostion 

ANB INGOS Ltd., Ionex (The Czech Republic). The 

temperature of the column was 55–74 °C and that of the 

reactor was 120 °C. The determination of the sulphurcontaining 

amino acids, methionine and cystine, was 

carried out by means of oxygenating hydrolysis, using a 

mixture of formic acid and hydrogen peroxide (9:1) at 

110 °C for 24 h. After cooling, the sample was hydrolysed 

with acid. Buffers of pH 2.6 and 3.0 were used; the 

temperature of the column was 60 °C and that of the 

reactor was 120 °C. The calculations were made according 

to the external standard. 

Expression of results 

The level of amino acids was given in 100 g of edible 

parts of the products. The composition of amino acids 

was also expressed as grams per 16 g of N to estimate 

the quality of the protein in New Zealand spinach by 

comparing it with the FAO ⁄ WHO pattern (1991). On 

the basis of the amino acid composition, the chemical 

score (CS) index was calculated using the Mitchell & 

Block (1946) and the integrated essential amino acid 

index using Oser (1951). 


Statistical calculations were performed using singlefactor 

analysis of variance (anova) and Duncan’s test 


1684 

Amino acids in New Zealand spinach J. Słupski et al. 

with probability level P = 0.05 with the use of Statistica 

Ò 6.1; Stat Soft Inc., Tulsa, OK, USA software. Each 

experiment was carried out in three replications, each 

determination in duplicate. 


The chemical composition of vegetables changes during 

culinary or technological treatment, and the most susceptible 

are vegetables with a large surface area or loose 

tissue structure (Lisiewska et al., 2004; Kmiecik et al., 

2007). In our former research, we found that cooking and 

operations connected with the preparation of frozen 

products and their subsequent preparing for consumption 

brought about considerable changes of many constituents 

in New Zealand spinach (Lisiewska et al., 

2009a). In the current work, changes in amino acids 

content in processed New Zealand spinach were studied. 

Despite the fact that the processing was carried out in 

water, a considerable increase in dry matter content of 

Amino acid Raw material 

Product prepared for consumption 

Cooked 

directly from 

raw material 

processed New Zealand spinach was observed after both 

blanching and cooking (Table 1). The dry matter after 

cooking was 47% higher than in the raw material, while 

in the products prepared for consumption from frozen 

spinach that was blanched before freezing (processing 

variant I) it was 36% higher and as much as 67% higher 

from the frozen-cooked spinach (variant II). Usually 

blanching or cooking in water causes leaching of some 

soluble constituents but at the same time high temperature 

of these processing methods brings about tissue 

shrinkage and this in turn expels water from vegetable. 

The overall effect of these can be an increase in dry 

matter content or in some constituents’ concentrations 

(Oboh, 2005; Ge˛bczyn´ski & Lisiewska, 2006). 

Statistically significant changes were observed in 

amino acid content in 100 g of processed New Zealand 

spinach products prepared for consumption (Table 1). 

Products prepared for consumption contained more 

amino acids than the raw vegetable, except methionine 

(in this case, the increase was nonsignificant). The 

Prepared from frozen and 

stored material 

Blanched 

before 

freezing 

(variant I) 

Cooked 

before 

freezing 

(variant II) 

Isoleucine 71 a ± 13 104 b ± 20 103 b ± 16 116 b ±20 

Leucine 135 a ± 16 192 b ± 24 189 b ± 22 219 b ±22 

Lysine 98 a ± 14 152 b ± 21 146 b ± 19 158 b ±24 

Cystine 11 a ±4 23 b ±8 21 ab ±7 22 ab ±8 

Methionine 16 a ±6 27 a ±10 29 a ±10 26 a ±9 

Total sulphur amino acids 27 a ±10 50 a ±17 50 a ±16 48 a ±16 

Tyrosine 40 a ±11 65 b ±18 65 b ±15 74 b ±15 

Phenylalanine 83 a ± 13 118 b ± 18 115 b ± 17 132 b ±16 

Total aromatic amino acids 123 a ± 23 184 b ± 33 180 b ± 32 206 b ±30 

Threonine 75 a ± 11 117 b ± 19 104 b ± 16 116 b ±19 

Valine 91 a ± 13 137 b ± 20 134 b ± 19 155 b ±17 

Histidine 56 a ±12 81 ab ±19 74 ab ±17 91 b ±16 

Total essential amino acids 676 a ± 99 1016 b ± 147 980 b ± 147 1109 b ± 146 

Arginine 108 a ± 13 150 b ± 22 147 b ± 21 171 b ±21 

Aspartic acid 165 a ± 14 218 b ± 19 225 b ± 20 255 c ±22 

Glutamic acid 208 a ± 18 275 b ± 26 253 b ± 22 335 c ±29 

Serine 76 a ±12 91 ab ±16 96 ab ± 18 113 b ±19 

Proline 94 a ± 15 138 b ± 19 137 b ± 17 136 b ±19 

Glycine 84 a ± 12 121 b ± 18 118 b ± 17 138 b ±19 

Alanine 90 a ± 14 130 b ± 22 126 b ± 17 146 b ±20 

Total nonessential amino acids 825 a ± 98 1123 b ± 142 1102 b ± 123 1294 b ± 135 

Total amino acids 1501 a ± 192 2139 b ± 286 2082 b ± 269 2403 b ± 281 

Dry matter g ⁄ 100 g of edible portion 6.19 9.12 8.44 10.31 

Means of double measurements of three different experiments. Means in the same raw with 

different superscript letters are significally different at P < 0.05. Means in the same raw without 

superscript letters are not significally different at P < 0.05. 

Table 1 Amino acid composition of raw 

and processed New Zealand spinach, 

mg ⁄ 100 g of edible portion 


percentage increase in amino acids ranged from 20% to 

100% for the cooked fresh spinach and 22% to 91% 

(variant I) and 49% to 100% (variant II) in cooked 

products, respectively. 

The percentage changes in amino acids were different 

when the results were compared in dry matter, what was 

possible to calculate from the Table 1. The content of a 

particular amino acid in dry matter ranged from )18% 

to +45% in cooked spinach, from )11% to +40% in 

cooked spinach prepared from the frozen product 

(variant I), and from )13% to +20% in New Zealand 

spinach prepared for consumption from the frozen 

product obtained using the modified technology (variant 

II). The greatest decrease was observed in serine content 

and the highest increase in cystine content. 

The nitrogen content in dry matter of raw material 

and processed New Zealand spinach did not change 

significantly (Table 2). Neither statistically significant 

differences in amino acid content expressed as 16 g of N 

Table 2 Amino acid composition of raw and 

processed New Zealand spinach, g ⁄ 16 g N 

Amino acid Raw material 


were found between all the examined samples, that is, 

the raw material and products prepared for consumption 

(Table 2). The exception was glutamic acid, content 

of which was significantly lower in cooked spinach (but 

not in the blanched) than in the raw material; however, 

among samples prepared for consumption, the changes 

of glutamic acid were nonsignificant. 

In comparison with the raw material and depending 

on the amino acid, the cooked samples contained from 

)19% to +43% of amino acids, and in the case of 

frozen samples prepared for consumption, the levels 

were )13% to +32% (variant I) and )17% to +16% 

(variant II). As to the fresh–cooked spinach, amino acid 

levels in the frozen samples were )7% to +13% 

(variant I) and )19% to +5% (variant II). Relation 

of the two frozen samples showed that the product 

obtained using the modified technology (variant II) 

contained )25% to 12% of the amino acids found in the 

traditional product (variant I). Similar, nonsignificant 


Cooked 

directly from 

raw material 



Blanched 

before 

freezing 

(variant I) 

Cooked 

before 

freezing 

(variant II) 

Isoleucine 3.68 ± 0.66 3.60 ± 0.69 3.65 ± 0.57 3.48 ± 0.60 

Leucine 6.98 ± 0.85 6.67 ± 0.83 6.74 ± 0.80 6.58 ± 0.67 

Lysine 5.07 ± 0.75 5.26 ± 0.72 5.18 ± 0.68 4.75 ± 0.71 

Cystine 0.56 ± 0.22 0.80 ± 0.27 0.74 ± 0.24 0.65 ± 0.25 

Methionine 0.86 ± 0.30 0.93 ± 0.34 1.05 ± 0.35 0.79 ± 0.26 

Total sulphur amino acids 1.42 ± 0.52 1.73 ± 0.61 1.78 ± 0.59 1.44 ± 0.49 

Tyrosine 2.09 ± 0.55 2.27 ± 0.61 2.32 ± 0.54 2.21 ± 0.45 

Phenylalanine 4.29 ± 0.65 4.11 ± 0.62 4.09 ± 0.59 3.97 ± 0.47 

Total aromatic amino acids 6.39 ± 1.20 6.39 ± 1.16 6.42 ± 1.13 6.17 ± 0.91 

Threonine 3.88 ± 0.58 4.08 ± 0.66 3.70 ± 0.56 3.48 ± 0.56 

Valine 4.74 ± 0.69 4.75 ± 0.68 4.77 ± 0.67 4.67 ± 0.51 

Histidine 2.93 ± 0.64 2.81 ± 0.65 2.65 ± 0.62 2.72 ± 0.49 

Total essential amino acids 35.07 ± 5.14 35.28 ± 5.11 34.89 ± 5.22 33.29 ± 4.39 

Arginine 5.59 ± 0.69 5.20 ± 0.77 5.22 ± 0.74 5.14 ± 0.63 

Aspartic acid 8.55 ± 0.74 7.57 ± 0.68 7.99 ± 0.71 7.66 ± 0.67 

Glutamic acid 10.81 b ± 0.95 9.54 ab ± 0.89 9.00 a ± 0.77 10.05 ab ± 0.87 

Serine 3.93 ± 0.62 3.17 ± 0.56 3.41 ± 0.63 3.40 ± 0.56 

Proline 4.90 ± 0.76 4.79 ± 0.67 4.88 ± 0.62 4.09 ± 0.58 

Glycine 4.37 ± 0.64 4.20 ± 0.64 4.20 ± 0.60 4.14 ± 0.56 

Alanine 4.66 ± 0.73 4.53 ± 0.77 4.50 ± 0.60 4.38 ± 0.59 

Total nonessential 

amino acids 

42.81 ± 5.07 38.99 ± 4.95 39.20 ± 4.38 38.87 ± 4.07 

Total amino acids 77.88 ± 9.95 74.27 ± 9.94 74.09 ± 9.58 72.16 ± 8.44 

N g ⁄ 100 g 

dry matter 

4.98 5.05 5.33 5.17 

Means of double measurements of three different experiments. Means in the same raw with 

different superscript letters are significally different at P < 0.05. Means in the same raw without 

superscript letters are not significally different at P < 0.05. 


1686 


Table 3 Amino acid indexes of raw and processed New Zealand spinach according to FAO ⁄ WHO (1991) 

Index Amino acid 

differences in amino acid content were reported in the 

protein of raw and cooked samples of cassava leaves 

(Diasolua-Ngudi et al., 2003). 

The limiting amino acids of the first order in New 

Zealand spinach, both in fresh and in products prepared 

for consumption, were cystine with methionine and of 

the second order lysine (Table 3). Nevertheless, CS 

indexes for the sulphur amino acids methionine and 

cystine were higher in products prepared for consumption 

than in the raw material (Table 3). 

Amino acids are usually susceptible to culinary or 

technological processing, especially when heating is 

involved. According to Klein & Mondy (1981), heat 

treatment might lead to compositional changes in 

nitrogenous compounds, depending on the mechanism 

of heat transfer and the particular tissue under treatment. 

During the preparation of food, amino acids of 

protein side chains can react with each other. These 

reactions could result in changes in the composition of 

amino acids (Sherr et al., 1989). Baxter (1995) used 

amino acid standards to study their behaviour in the 

heat process and observed their partial decomposition 

ranging from 8% to 14%. The Maillard reaction would 

also contribute to the changes of amino acids. The 

initial reaction in the process involving condensation 

between amino groups of amino acids and sugars is 

reversible; however, subsequent steps are not, and some 

amino acids are lost in this way (Baxter, 1995; Candela 

et al., 1997). Eventually, decomposition of some protein 

fractions rich in amino acids can result in higher values 

for other amino acids when they are calculated on the 

basis of 16 g N (Mutia & Uchida, 1993; Espe & Lied, 

1999). 

Technological processing also plays a part in the 

changes of amino acids content. Cruz-Garcia et al. 

Raw 

material 


Cooked 

directly 

from 

raw 

material 



Blanched 

before 

freezing 

(variant I) 

Chemical score index Isoleucine 131 129 130 124 

Leucine 106 101 102 100 

Lysine 87 91 89 82 

Cystine with methionine 57 69 71 58 

Tyrosine with phenylalanine 101 101 102 98 

Threonine 114 120 109 102 

Valine 135 136 136 133 

Histidine 154 148 139 143 

Essential amino acid index 107 109 108 101 

Cooked 

before 

freezing 

(variant II) 

indicated that extending the cooking time and increasing 

the volume of water used for cooking resulted in 

greater amino acid loss by diffusion and degradation 

(de la Cruz-Garcia et al., 1999). Diasolua-Ngudi et al. 

(2003) reported that a greater volume of water led to 

increased loss of constituents because of diffusion. 

However, Chau et al. (1997) proved that doubling the 

cooking time did not decrease amino acid levels and 

actually increased them in some vegetable species. 

Candela et al. (1997) showed that after heat treatment 

in water, the amino acid content in 100 g of edible 

portion increased or decreased depending on the vegetable 

species. A similar view that heat treatment could 

either increase or decrease amino acid content was 

presented by Parihar et al. (1996) and Chau et al. 

(1997). It was also shown that with the same parameters 

of heat treatment, changes in amino acid content varied 

and depended on the variety (Kmiecik et al., 1994, 

1999), ripeness (Korus et al., 2003) and the usable part 

of the analysed vegetable (Murcia et al., 2001; Lisiewska 

et al., 2004). 

In contrast to other leafy vegetables such as kale and 

spinach, the edible parts of New Zealand spinach were 

not a rich source of amino acids (Table 1). Hundred 

grams of edible parts of New Zealand spinach contained 

64% of the amino acids present in spinach 

(Z. Lisiewska, W. Kmiecik & P. Ge˛bczyn´ski, unpublished 

data) and only 41% of those in kale (Lisiewska 

et al., 2008). The dominant amino acid, as in spinach 

(Z. Lisiewska, W. Kmiecik & P. Ge˛bczyn´ski, unpublished 

data) and kale (Lisiewska et al., 2008), was 

glutamic acid, and the limiting amino acids were cystine 

with methionine (Table 3). In other leafy vegetables, the 

limiting amino acids were methionine with cystine 

(Wallace et al., 1998; Z. Lisiewska, W. Kmiecik & 


P. Ge˛bczyn´ski, unpublished data) and also leucine 

(Lisiewska et al., 2009b) and lysine (Lisiewska et al., 

2008). Essential amino acids accounted for 45% of total 

amino acids, which did not differ greatly from the 

proportion in other species of leafy vegetables: in 

spinach, it made up 49% (Z. Lisiewska, W. Kmiecik & 

P. Ge˛bczyn´ski, unpublished data) and in kale 44% 

(Lisiewska et al., 2008). 

Conclusion 

Culinary and technological processing caused a significant 

increase in total amino acid content in 100 g of 

edible portion, with the exception of methionine and 

cystine in products prepared for consumption. Changes 

in amino acid content expressed as 16 g of N were 

nonsignificant, except for the lower content of glutamic 

acid in the traditional frozen product. Cystine with 

methionine were the limiting amino acids of the first 

order both in fresh vegetable and in products prepared 

for consumption, and lysine was the limiting amino acid 

of the second order. 

Despite the fact that the total amino acid content in 

New Zealand spinach was not high and deficient in 

sulphur amino acids, its nutritional value was good. 

Interestingly, the CS indexes of sulphur amino acids, 

methionine and cystine, were higher in products 

prepared for consumption than in the raw material. 


The authors would like to thank Prof. Robert D. Tanner 

(Vanderbilt University) for grammatical corrections of 

the manuscript. 

References 

Association of Official Agricultural Chemists (AOAC). 1984. Official 

Methods of Analysis, 14th edn. Arlington, Virginia, USA: Association 

of Official Agricultural Chemists. 

Baxter, J. (1995). Free amino acid stability in reducing sugar systems. 


Candela, M., Astiasaran, I. & Bello, J. (1997). Cooking and warmholding: 

effect on general composition and amino acids of kidney 

beans (Phaseolus vulgaris), chickpeas (Cicer arietinum), and lentils 

(Lens culinaris). Journal of Agricultural and Food Chemistry, 45, 

4763–4767. 

Chau, C.F., Cheung, P.C.K. & Wong, Y.S. (1997). Effects of cooking 

on content of amino acids and antinutrients in three Chinese 

indigenous legume seeds. Journal of the Science of Food and 


Codex Alimentarius, CAC ⁄ RCP 39. (1993). Code of Hygienic Practice 

for Precooked and Cooked Foods in Mass Catering 3:39, Pp. 18. 

de la Cruz-Garcia, C., Lopez-Hernandez, J., Gonzalez-Castro, M., 

Rodriguez-Bernaldodequiros, A. & Simal-Lozano, J. (1999). Effects 

of various culinary treatments on the amino acid content of green 

beans. Deutsche Lebensmittel-Rundschau, 95, 507–510. 

Diasolua-Ngudi, D., Kuo, Y.-H. & Lambein, F. (2003). Cassava 

cyanogens and free amino acids in raw and cooked leaves. Food and 

Chemical Toxicology, 41, 1193–1197. 


Espe, M. & Lied, E. (1999). Fish silage prepared from different cooked 

and uncooked raw materials: chemical changes during storage at 

different temperatures. Journal of the Science of Food and Agriculture, 

79, 327–332. 

FAO ⁄ WHO (1991). Protein quality evaluation in human diets. Food 

Nutr Pap 51, FAO: Rome, Italy. 

Ge˛bczyn´ski, P. (2008). Changes in the contents of selected antioxidants 

in frozen vegetables as dependent on preliminary processing, 

conditions of storage and method of preparing for consumption 

(monograph in Polish). Zeszyty Naukowe Akademii Rolniczej im. 

H. Kołła˛taja W Krakowie, No. 445, vol. 332. 

Ge˛bczyn´ski, P. & Lisiewska, Z. (2006). Comparison of the level of 

selected antioxidative compounds in frozen broccoli produced using 

traditional and modified methods. Innovative Food Science and 

Emerging Technologies, 7, 239–245. 

Jaworska, G. & Kmiecik, W. (1999). Effect of the date of harvest on 

the selected traits of the chemical composition of spinach (Spinacia 

oleracea L.) and New Zealand spinach (Tetragonia expansa Murr.). 

Acta Agraria et Silvestria Series Agraria, 37, 15–26. 

Jaworska, G. & Kmiecik, W. (2000). Comparison of the nutritive value 

of frozen spinach and New Zealand spinach. Polish Journal of Food 

and Nutrition Sciences, 9, 79–84. 

Kawashima, L. & Valente Soares, L. (2003). Mineral profile of raw 

and cooked leafy vegetables consumed in Southern Brazil. Journal of 

Food Composition and Analysis, 16, 605–611. 

Klein, L.B. & Mondy, N.I. (1981). Comparison of microwave and 

conventional baking of potatoes in relation to nitrogenous constituents 

and mineral composition. Journal of Food Science, 46, 1874– 

1877. 

Kmiecik, W. & Jaworska, G. (1999). Effect of growing methods of 

New Zealand spinach on its yield and pattern of harvests. Folia 

Horticulturae, 11, 75–85. 

Kmiecik, W., Lisiewska, Z. & Jaworska, G. (1994). Amino acids 

content in fresh and preserved broad beans (Vicia faba major). 

Polish Journal of Food and Nutrition Sciences, 3 ⁄ 44, 35–43. 

Kmiecik, W., Lisiewska, Z. & Ge˛bczyn´ski, P. (1999). Content of amino 

acids in fresh and frozen and cooked broad bean seeds (Vicia faba 

var major) depending on cultivar and degree of maturity. Journal of 

The Science of Food and Agriculture, 79, 555–560. 

Kmiecik, W., Lisiewska, Z. & Korus, A. (2007). Retention of mineral 

constituents in frozen brassicas depending on the method of 

preliminary processing of the raw material and preparation of 

frozen products for consumption. European Food Research and 

Technology, 224, 573–579. 

Korus, A., Lisiewska, Z. & Kmiecik, W. (2003). Content of amino 

acids in fresh and preserved physiologically immature grass pea 

(Lathyrus sativus L.) Seeds. European Food Research and Technology, 

217, 148–153. 

Lisiewska, Z., Supski, J., Kmiecik, W. & Gebczyński, P. (2004). Amino 

acid profiles and protein quality of fresh and frozen dill depending 

on usable part of raw material, pre-treatment before freezing, and 

storage temperature of frozen products. Food Science and Technology, 

7, 3–11. 

Lisiewska, Z., Kmiecik, W. & Korus, A. (2008). The amino acid 

composition of kale (Brassica oleracea L. var. acephala), fresh and 

after culinary and technological processing. Food Chemistry, 108, 

642–648. 

Lisiewska, Z., Gebczyn´ski, P., Bernas´, E. & Kmiecik, W. (2009a). 

Retention of mineral constituents in frozen leafy vegetables prepared 

for consumption. Journal of Food Composition and Analysis, 22, 

218–223. 

Lisiewska, Z., Supski, J., Skoczen´-Supska, R. & Kmiecik, W. (2009b). 

Content of amino acids and the quality of protein in Brussels 

sprouts, both raw and prepared for consumption. International 

Journal of Refrigeration, 32, 272–278. 

Meijer, A.J. (2003). Amino acids as regulators and components of 

nonproteinogenic pathways. The Journal of Nutrition, 133, 2057S– 

2062S. 


1688 


Mitchell, H.H. & Block, R.J. (1946). Some relationships between the 

amino acid contents of proteins and their nutritive values for the rat. 

The Journal of Biological Chemistry, 163, 599–620. 

Murcia, M.A., Lopez-Ayerra, B., Martinez-Tome, M. & Garcia- 

Carmona, F. (2001). Effect of industrial processing on amino acid 

content of broccoli. Journal of the Science of Food and Agriculture, 

81, 1299–1305. 

Mutia, R. & Uchida, S. (1993). Effect of heat treatment on nutritional 

value of winged bean (Psophocarpus tetragonolobus) as compared to 

soybean I. Chemical characteristics of treated winged bean. Asian- 

Australian Journal of Animal Sciences, 6, 19–26. 

Oboh, G. (2005). Effect of blanching on the antioxidant properties of 

some tropical green leafy vegetables. LWT-Food Science and 


Oser, B. (1951). Method for integrating essential amino acid content in 

the nutritional evaluation of protein. Journal of the American 

Dietetic Association, 27, 396. 

Parihar, P., Mishra, A., Gupta, O. & Singh, A. (1996). Effect of 

cooking on limiting essential amino acid content of common pulses. 

Advances in Plant Sciences, 9, 165–170. 

Sherr, B., Lee, C.M. & Jelesciewicz, C. (1989). Absorption and 

metabolism of lysine Maillard products in relation to utilization 

of L-lysine. Journal of Agricultural and Food Chemistry, 37, 119– 

122. 

Wallace, P., Marfo, E. & Plahar, W. (1998). Nutritional quality and 

antinutritional composition of four non-conventional leafy vegetables. 





Mango film coated for fresh-cut mango in modified atmosphere 

packaging 

Rungsinee Sothornvit* & Patratip Rodsamran 

Department of Food Engineering ⁄ PHTIC, Faculty of Engineering at Kamphaengsaen ⁄ Center of Excellent for Agricultural and Food Machinery, 

Kasetsart University, Kamphaengsaen Campus, Nakhonpathom 73140, Thailand 


Summary Mango puree is known to have good oxygen barrier properties. Therefore, mango film on its own might be 

useful to extend the shelf life of fresh-cut mango. In this study, fresh-cut mango was wrapped with mango 

film and packed in modified atmosphere packaging (MAP). Each package was stored at 30 °C or5°C and 

sensory evaluation was performed to determine its quality and shelf life. The shelf life of uncoated and coated 

fresh-cut mango pieces was 6 days for each at 5 °C and was 3 and 4 days, respectively, at 30 °C. The sensory 

evaluation indicated that coated fresh-cut mango was slower to produce an off-flavour and maintained better 

visual quality than uncoated mango at 30 °C. Nonetheless, coated fresh-cut mango was softer than uncoated 

mango because of the hydrophilic nature of the mango film. No significant difference in the oxygen and 

carbon dioxide concentrations was observed between coated and uncoated fresh-cut mango. Coating freshcut 

mango with mango film showed a similar effect to MAP in prolonging the shelf life of the fresh produce. 

The use of the coating will enhance fruit quality and lead to better acceptance by consumers. 

Keywords Edible film, firmness, fresh-cut mango, modified atmosphere, sensory evaluation. 

Introduction 

Mango (Mangifera indica L.) is a major tropical fruit 

in the domestic and export market of Thailand. 

Mango cv. Namdok Mai is the most popular variety, 

and its production ranks first among commercial 

mango varieties. Recently, consumers demand healthy 

and fresh foods including minimal processing or freshcut 

fruit. Minimal processing is defined to include all 

operations such as washing, sorting, trimming, peeling, 

slicing, coring, etc. that would not extensively affect 

the fresh-like quality of the produce (Shewfelt, 1987). 

Therefore, minimal processing is prepared without 

applying any preservation treatment which may 

change the characteristics of the product. Minimal 

processing is often used for fruits that are difficult to 

peel, have a bulky, inedible peel or have objectionable 

seeds (Fardiaz et al., 2000). However, one of the most 

critical problems in minimally processed products is 

the undesirable physiological changes, such as colour, 

texture, aroma and overall appearance that reduce 

their shelf life (Bolin & Huxsoll, 1989; Wong et al., 

1994). Especially, tropical fresh-cut fruits have shorter 

*Correspondent: Fax: 66 34351404; 

e-mail: fengrns@ku.ac.th 

doi:10.1111/j.1365-2621.2010.02316.x 


shelf life than temperate fresh-cut fruits (Gonzalez- 

Aguilar et al., 2008). Thus, it is of interest to study 

alternative treatments to maintain tropical fresh-cut 

fruit quality and extend shelf life. 

Edible films are used primarily to extend the shelf life 

and quality of foods by preventing changes in aroma, 

taste, texture and appearance of foods (Arvanitoyannis, 

1999; Tharanathan, 2003). Edible films composed of 

polysaccharides and proteins have good gas barrier 

properties, but poor water barrier properties under 

certain relative humidity (RH) and temperature conditions 

(Guilbert, 1986; Kester & Fennema, 1986; Arvanitoyannis 

& Biliaderis, 1998; Petersson & Stading, 

2005). This behaviour of edible films presents an 

effective semi-permeable barrier to the respiratory gases 

(carbon dioxide and oxygen) and can create the 

modified atmosphere (MA) when applied to fruit and 

vegetable (Baldwin, 1994). This MA slows down 

respiration, metabolism and retard ethylene production; 

therefore, edible films can improve the quality and 

extend the shelf life of fruit and vegetable (Kader, 

1986). Moreover, fruit and vegetable purees such as 

peach, strawberry, apricot, apple, pear, carrot and 

broccoli have a potential to use as alternative natural 

edible film materials (McHugh et al., 1996; McHugh & 

Olsen, 2004). They showed significant reduction in

1690 

Mango film coated for fresh-cut mango R. Sothornvit and P. Rodsamran 

moisture loss and browning in fresh-cut apples wrapped 

in apple puree-based films (McHugh & Senesi, 2000). 

However, the study related to the effect of film coating for 

fresh-cut mangoes is limited. One study found that when 

edible coatings such as carboxymethylcellulose incorporated 

with maltodextrin was applied on fresh-cut mangoes, 

they showed the highest scores on visual quality and 

favour of mangoes (Plotto et al., 2004). The other study 

reported that two commercial film coatings (Gustec and 

SemperFresh) reduced colour changes, firmness loss and 

decay with prevention of undesirable attributes for 

Ataulfo fresh-cut mangoes (Gonzalez-Aguilar et al., 

2008). From our previous study, we observed that mango 

film exhibited a potential to apply for whole and fresh-cut 

mangoes (Sothornvit & Rodsamran, 2008). Therefore, 

application of films formed by fruit puree on the same 

fresh-cut product might benefit both quality and shelf life, 

without affecting flavour. 

Modified atmosphere packaging (MAP) can prolong 

shelf life of fresh produce by decreasing oxygen (O2) and 

increasing carbon dioxide (CO 2) concentrations. Each 

commodity has its own tolerance limit to the amount of 

CO2 and O2 concentration. Mostly, minimum O2 

concentration and maximum CO2 concentration range 

in between 0.5–5.0% and 2–15%, respectively (Kader, 

1992). The O2 and CO2 concentrations within MAP 

depend on the interaction between commodity respiration 

and the permeability properties of packaging 

material (Beaudry et al., 1992; Kader, 1997). In freshcut 

products, MAP is a way to alter metabolism (e.g. 

ripening, senescence, cut surface browning, etc.) favourably. 

Therefore, the objective of this study was to extend 

the application of mango film on fresh-cut mango in 

combination with MAP to maintain its quality and shelf 

life. 


Materials 

Ripe mangoes (M. indica L.), cv Namdok Mai, obtained 

from the local market in Nakhonprathom province 

(Thailand) were selected and cleaned to form mango 

puree and to prepare fresh-cut mangoes. Ripe mangoes 

were 90% mature with 19% total soluble solid 

(determined by a portable hand-held refractometer, 

Model N32, Bellevae, WA, USA) and 0.33 N firmness 

(determined by a Universal testing machine). 

Film formation 

To obtain mango puree (19% soluble solid content), 

ripe mangoes were washed, peeled, cut into pieces and 

pulped to a mango puree by hand followed by sieving. 

Mango puree was degassed by applying vacuum to 

remove dissolved air. To maintain a constant mass of 

solids, 20 g of puree was poured and spread evenly by 

lifting a plate around in a 15 cm diameter high-density 

polyethylene plate. Films were dried in a hot air oven at 

a controlled temperature 50 ± 3 °C for 12 h until they 

could be released from the plates. The final solid content 

of the film was 3.8 g. 

Film thickness 

Film thickness was measured with a micrometre (No. 

7326; Mitutoyo Manufacturing Co., Ltd, Tokyo, Japan) 

to the nearest 0.0001 (0.00254 mm) around the filmtesting 

area at five random positions. An average of film 

thickness was obtained for each film replicate. Twenty 

replicates were used to determine film thickness. 

Fresh-cut mango wrapped with mango film 

Ripe mangoes were cleaned, peeled and cut into 

rectangular pieces (approximately 15 · 25 · 15 mm). 

Each piece of fresh-cut mango was wrapped with mango 

film. The average film thickness was 0.17 ± 0.02 mm. 

Both coated and uncoated fresh-cut mangoes were kept 

in cellophane bags (20 lm of film thickness and 

108 cm 3 lm(m 2 day kPa) )1 of oxygen permeability at 

20 °C, 85% RH) with MAP and sealed with a heatsealing 

machine. The gas mixture was composed of 

9.52 ± 0.10% O 2, 0.23 ± 0.02% CO 2 and the rest was 

N2. Finally, samples were stored in the incubator at 

30±2°C (room temperature simulating commercial 

conditions in Thailand) and in a cold room at 5 ± 2 °C 

(cold temperature as a normal practice) and quality was 

evaluated everyday. Three replications were used for 

each treatment. 


The shelf life of fresh-cut mangoes was determined by 

sensory evaluation using a descriptive test (three-point 

scale). The panel was made of six semi-trained panellists, 

i.e. laboratory staff and students whom well acquainted 

with sensory evaluation of fresh mango but not specifically 

trained on mango testing. All panellists were 

familiar with mango attributes. Assessments were carried 

out under nature light at a room temperature 

28±3°C. Six semi-trained panellists were chosen to 

determine the following mango quality attributes: 

aroma ⁄ flavour, colour, internal appearance and overall 

visual quality at the given scores from 1 (=fresh-like), 2 

(=acceptable) and 3 (=completely deteriorated). At the 

beginning of each session, fresh mango was presented to 

the judges as a calibration for the ‘fresh-like mango’ 

attribute before they rated all the attributes of the freshcut 

samples during storage. The sample was coded, 

presented in random order, and the judges scored each 

sample as it was received in each plastic bag. Fresh-cut 


mangoes were kept in the cellophane bags, which were 

opaque; thus, browning (colour change) and internal 

appearance could not be observed instantly. The first 

detectable attribute was flavour, and the judges evaluated 

the general visual appearance without taste. 

Therefore, our protocol for determining the shelf life 

of fresh-cut mangoes was based on off-flavour attribute, 

followed by colour and visual appearance, respectively. 

Firmness measurement 

Firmness was determined by measuring the force 

required for a 2 mm diameter probe to penetrate the 

fresh-cut surface, held perpendicular to the probe at 

10 mm min )1 compression speed and 5 mm depth using 

Universal testing machine with 500-N load cell 

(LLOYD Instruments, Sussex, UK, model LR 5K). 

Five pieces per treatment were randomly selected. Three 

replications were used to determine firmness. 

Ethanol concentration measurement 

Ethanol concentration in the package was determined by 

a gas chromatograph (GC; Hewlett Packard-Agilent, 

CA, USA, model 6890) with a capillary column, model 

HP-FFAP polyethylene glycol TPA, 25 m length, 

0.32 mm diameter, 0.50 lm film thickness and 

28 cm s )1 velocity using helium as a carrier gas and flame 

ionization detector (FID). The injector, oven and detector 

temperatures were 210, 60 and 210 °C, respectively. A 

sample of 1 lL was injected into the GC using GC 

autosampler syringe, which is a tapered fixed needle with 

23–26 s needle gauge, 42 mm length and HP tip. A series 

of standard ethanol concentrations (10, 50 and 

100 mg kg )1 ) were used to determine the ethanol concentration. 

The calibration curve was constructed from 

the linear regression of the standard ethanol concentrations 

and the peak area with the correlation of 0.999. 

The ChemStation version A05.0X software was used in 

the data analysis to calculate the ethanol concentration 

from the area of chromatograph. Five pieces per 

treatment were randomly selected. Three replications 

were used to determine ethanol concentration. 

Oxygen and carbon dioxide concentration measurement 

Oxygen and carbon dioxide concentrations in the 

package were determined by an oxygen ⁄ carbon dioxide 

analyzer (Quantek Instruments, MA, USA, model 

902D). The needle was plunged into the package. The 

pump was electronically timed to draw in the amount of 

sample required for the analysis, and then turn itself off 

after the preset sampling time (5–10 s). Five pieces per 

treatment were randomly selected. Three replications 

were used to determine oxygen and carbon dioxide 

concentrations inside the package. 

Mango film coated for fresh-cut mango R. Sothornvit and P. Rodsamran 1691 


A completely randomised experimental design was used 

to study coating method and storage temperature 

factors. spss for Windows software program, Release 

9.0.0 (SPSS Inc., 1999) was utilised to calculate analysis 

of variance (anova) using the general linear models 

procedure. Duncan’s multiple range test was used to 

determine significant differences between treatments at 

95% confidence interval. 



Quality attributes of fresh-cut mango changed with 

storage time (Fig. 1). All sensory attributes changed to 

unacceptable as storage time increased. Temperature 

had greatly significant effects on all sensory attributes 

(P £ 0.05). 

Off-flavour 

High storage temperature increased off-flavour of 

coated and uncoated mangoes at a faster rate than low 

temperature (Fig. 1). Coated and uncoated fresh-cut 

mangoes had no significant difference on off-flavour at 

low-temperature storage. However, coated fresh-cut 

mango in MAP was slower off-flavour than uncoated 

at room temperature storage. Off-flavour was from 

anaerobic respiration which produced and released 

alcohol and acetaldehyde. Low-temperature storage 

normally lowers respiration rate and reduces off-flavour. 

This was similar to the finding in gluten-based and 

composite coating on strawberries (Tanada-Palmu & 

Grosso, 2005), commercial edible coating on fresh-cut 

mangoes (Gonzalez-Aguilar et al., 2008) and mango 

film coating on fresh-cut mango without MAP 

(Sothornvit & Rodsamran, 2008). This was consistent 

with a decrease in O2 and an increase in CO2 concentrations 

during storage including higher ethanol 

concentration at high storage temperature. 

Colour 

Low-temperature storage lowered browning reaction 

which normally occurred with fresh-cut mangoes 

(Fig. 1). Cutting allows polyphenoloxidase to come in 

contact with phenolic compounds and oxygen and lead 

to tissue browning (Schwimmer, 1978; Brecht, 1995; 

Reyes, 1996). Generally, the coating with MAP helped to 

prevent fresh-cut produce to exposure to atmospheric, 

O 2 and low-temperature storage also reduced enzymatic 

browning (Gorny et al., 1999). This effect was found in 

the reduction of browning in whey protein–coated freshcut 

persimmon and apple at low-temperature storage 

(Perez-Gago et al., 2003a,b). Low-temperature storage 

also reduced colour change and browning for fresh-cut 


1692 


(a) (b) 

(c) 

Figure 1 Sensory evaluation [off-flavour (a), colour (b), internal appearance (c) and visual quality (d)] of fresh-cut mango in modified atmosphere 

packaging. All sensory aspects were rated in a three-point scale descriptive test; 1 = fresh-like, 2 = acceptable and 3 = completely deteriorated. 

Error bar shows standard deviation. UR, uncoated fresh-cut mango at room temperature storage (30 °C); CR, coated fresh-cut mango at room 

temperature storage (30 °C); UC, uncoated fresh-cut mango at cold temperature storage (5 °C); CC, coated fresh-cut mango at cold temperature 

storage (5 °C). 

mango (Tovar et al., 2000) and mango puree (Alaka 

et al., 2003). However, using mango film had no significant 

effect on colour change in this study. Likewise, the 

commercial edible coating (SemperFresh) did not show 

any significant effect on browning index in fresh-cut 

mangoes (Gonzalez-Aguilar et al., 2008). 

Internal appearance 

Unacceptable internal appearance was defined as the 

internal appearance (translucency) similar to what 

happens in mangosteen flesh. Temperature significantly 

affected the translucency (P £ 0.05). Fresh-cut mango 

stored at room temperature was less transparent than 

samples stored at low temperature after 2 days because 

fresh-cut mango at low temperature is susceptible of 

chilling injury which translates into undesirable appearance. 

In contrast with other results, coated fresh-cut 

mango caused transparent flesh or undesirable appearance 

more than uncoated because of the solubility of 

mango film. This drawback is being solved by incorporating 

lipids to increase the water resistance of mango 

film. 

(d) 

Visual quality 

We expected that the excellent oxygen barrier of 

mango film would extend the shelf life of fresh-cut 

mango. However, it showed that MAP with or without 

mango film coating at cold storage extended the 

shelf life of fresh-cut mango to 6 days (Fig. 1). Therefore, 

coating did not show any difference in the shelf life. 

It might be because of the greater effect of MAP over 

coating on the quality of fresh-cut mangoes at lowtemperature 

storage. Coated fresh-cut mango with 

mango film was more pronounced in maintaining the 

quality and prolonging the shelf life at low-temperature 

storage without MAP from 2–3 days of normal marketable 

shelf life to 5–6 days when wrapped with mango 

film (Sothornvit & Rodsamran, 2008). Moreover, wrapping 

extended the shelf life of fresh-cut mango at room 

temperature for 2 days comparing to unwrapped mango 

without MAP application (Sothornvit & Rodsamran, 

2008). This was implied that coating fresh-cut mango 

with mango film or edible film shows the similar effect as 

MAP in prolonging the shelf life of the fresh produce. In 

brief, coating extended fresh-cut mango shelf life from 


Figure 2 Firmness of fresh-cut mango in modified atmosphere 

packaging. Error bar shows standard deviation. UR, uncoated fresh-cut 

mango at room temperature storage (30 °C); CR, coated 

fresh-cut mango at room temperature storage (30 °C); UC, 

uncoated fresh-cut mango at cold temperature storage (5 °C); CC, 

coated fresh-cut mango at cold temperature storage (5 °C). 

2–3 days to 4 days at room temperature. Both coated 

and uncoated fresh-cut mango shelf life were 6 days at 

5 °C similar to the shelf life of low methoxypectinglycerol-coated 

fresh-cut mango (Fardiaz et al., 2000). 

However, shelf life fresh-cut mangoes can change with 

cultivars (Gonzalez-Aguilar et al., 2008). 

Firmness 

Generally, fresh-cut produce lose firmness during storage 

time, and this was also observed with fresh-cut 

mangoes (Fig. 2). Firmness loss in mango has been 

attributed to the action of pectolytic enzymes, namely 

polygalacturonase and pectinesterase on the solubilisation 

of pectic substrates (Lakshminarayana, 1980). 

Firmness loss was higher degree at low-temperature 

storage because of the chilling injury. This result was 

consistent with the occurrence of translucency in freshcut 

mango. Moreover, coated fresh-cut mango was 

softer than uncoated. We previously hypothesised that 

solubility of the hydrophilic mango film might cause 

diffusion of water to the fruit flesh, resulting in lower 

firmness (Sothornvit & Rodsamran, 2008). Furthermore, 

the effect of mango film without MAP on firmness 

of fresh-cut mango (Sothornvit & Rodsamran, 2008) 

was similar to the effect of MAP alone without coating 

in this study. This indicates that coating fresh-cut 

mango with mango film exhibited the similar effect on 

firmness as MAP regardless of temperature storage. The 

loss of mango firmness was also found in carboxymethyl 

cellulose coatings on fresh-cut mango (Plotto et al., 

2004). Additionally, SemperFresh edible coating showed 

no significant protection of firmness loss in fresh-cut 

mangoes (Gonzalez-Aguilar et al., 2008). Nonetheless, 


Figure 3 Ethanol concentration of fresh-cut mango in modified 

atmosphere packaging at the end of shelf life. Error bar shows 

standard deviation. UR, uncoated fresh-cut mango at room temperature 

storage (30 °C); CR: coated fresh-cut mango at room 

temperature storage (30 °C); UC, uncoated fresh-cut mango at 

cold temperature storage (5 °C); CC, coated fresh-cut mango at cold 

temperature storage (5 °C). 

coating fresh-cut mango with mango film might enhance 

the acceptance by consumers and reduce the loss of ripe 

mango during the harvest season. 

Ethanol, oxygen and carbon dioxide concentrations 

Ethanol content in the MAP of fresh-cut mango was 

similar for all conditions (3.32 ± 0.96 mg kg )1 ) at the 

starting day. Higher-temperature storage induced higher 

ethanol content in the package (Fig. 3), resulting in offflavour 

at the end of shelf life. There is no threshold level 

of ethanol content in fresh-cut mangoes by panellists 

(Gonzalez-Aguilar et al., 2008). Thus, the ethanol content 

of fresh-cut mango was measured at the end of shelf 

life according to the sensory evaluation by panellists in 

this study. Coating did not show any significant effect at 

5 °C. However, coating showed greater ethanol content 

than uncoated at 30 °C. This might be because of the 

good oxygen barrier of mango film (Sothornvit & 

Rodsamran, 2008) affecting the greater off-flavour as 

detected in sensory evaluation. In contrast, the commercial 

edible coatings prevented ethanol production in 

all three cultivars (Ataulfo, Keitt and Kent) at 5 °C 

storage (Gonzalez-Aguilar et al., 2008). The different 

result might be because of the behaviour of mango 

cultivars on the effect of different edible coatings. 

Nevertheless, the combination of MAP and coating is 

still advantageous at cold storage in maintaining the 

quality of fresh-cut mango. Moreover, there was a 

decrease in oxygen concentration and an increase in 

carbon dioxide concentration during the storage time in 

the MAP (Fig. 4). As known, low oxygen and high 

carbon dioxide concentrations slow down respiration 

and retard ethylene production and therefore ripening 


1694 


(Kader, 1986). During storage of mangoes, low levels of 

O 2 resulted in abnormal colour development and 

increased ethanol production. Nonetheless, the O2 level 

of fresh-cut mangoes in MAP maintained in higher 

degree than those without MAP (Sothornvit & Rodsamran, 

2008). In this study, no significant difference of 

gas concentration was observed between coated and 

uncoated fresh-cut mango samples in MAP. 

The correlation between the sensory attributes and the 

ethanol content was determined. It was found that each 

sensory attribute showed a significant correlation with 

the ethanol content. The correlations of internal appearance, 

off-flavour, visual quality and colour with the 

ethanol content were 0.675, 0.645, 0.573 and 0.543, 

respectively. It was confirmed that the unacceptability 

from the panellists was because of the ethanol content. 

Conclusion 

Low-temperature (5 °C) storage of fresh-cut mango 

prolonged the retention of desirable quality characteristics. 

Interestingly, coated fresh-cut mango extended 

shelf life and maintained sensory appearance at room 

temperature storage better than uncoated regardless of 

ethanol content detected. MAP at low-temperature 

storage for fresh-cut mango was also appropriate to 

keep good quality. Therefore, temperature is a key 

factor for fresh-cut mango quality. The timing of 

firmness loss and the extent of ethanol accumulation 

were consistent with the sensory determination at the 

end of shelf life. Mango film can function as food 

wrappers similar to a seaweed wrapper for sushi or as a 

part of packaging materials for a variety of foods such 

as confectionery products, nuts, fresh and processed 

fruits substitute the synthetic packaging materials. It is 

suitable for the protection of dried food or refrigerated 

food such as frozen carrots from the weight loss and the 

undesirable appearance such as white spots on carrot 

surfaces. 


The author thanks the Kasetsart University Research 

Development Institute (KURDI) and the Postharvest 

Technology Innovation Center (Thailand) for the financial 

support throughout this research. 

References 

Figure 4 Gas concentration of fresh-cut 

mango in modified atmosphere packaging for 

all storage conditions. UR, uncoated freshcut 

mango at room temperature storage 

(30 °C); CR, coated fresh-cut mango at room 

temperature storage (30 °C); UC, uncoated 

fresh-cut mango at cold temperature storage 

(5 °C); CC, coated fresh-cut mango at cold 

temperature storage (5 °C). 

Alaka, O.O., Anina, J.O. & Falade, K.O. (2003). Effect of storage 

conditions on the chemical attributes of ‘‘Ogbomoso’’ mango juice. 


Arvanitoyannis, I. (1999). Totally and partially biodegradable polymer 

blends based on natural and synthetic macromolecules: preparation, 

physical properties and potential as food packaging material. 

Journal of Macromolecules Science, 39, 205–271. 

Arvanitoyannis, I. & Biliaderis, C.G. (1998). Physical properties of 

polyol-plasticized edible films made from sodium caseinate and 

soluble starch blends. Food Chemistry, 62, 333–342. 

Baldwin, E.A. (1994). Edible coatings for fresh fruits and vegetables: 

past, present, and future. In: Edible Coatings and Films to Improve 

Food Quality (edited by J.M. Krochta, E.A. Baldwin & M.O. 

Nisperos-Carriedo). Pp. 25–64. Pennsylvania: Technomic Publishing 

Company. 

Beaudry, R.M., Cameron, A.C., Shirazi, A. & Dostal-Lange, D.L. 

(1992). Modified-atmosphere packaging of blueberry fruit: effect of 

temperature on package O2 and CO2. Journal of the American 

Society for Horticultural Science, 117, 436–441. 

Bolin, H.R. & Huxsoll, C.C. (1989). Storage stability of minimally 

processed fruit. Journal of Food Processing and Preservation, 13, 

281–292. 

Brecht, J.K. (1995). Physiology of lightly processed fruits and 

vegetables. HortScience, 30, 18–22. 

Fardiaz, D., Setiasih, I.S., Purwadaria, H.K. & Apriyatono, A. (2000). 

Quantitative descriptive analysis and volatile component analysis of 

minimally processed ‘Arumanis’ mango coated with edible film. In: 

ACIAR Proceeding 100. Pp. 196–200. Indonesia: Institute Pertanian 

Bogor. 

Gonzalez-Aguilar, G.A., Celis, J., Sotelo-Mundo, R.R., de la Rosa, 

L.A., Rodrigo-Garcia, J. & Alvarez-Parrilla, E. (2008). Physiological 

and biochemical changes of different fresh-cut mango cultivars 

stored at 5 °C. International Journal of Food Science and Technology, 

43, 91–101. 

Gorny, J.R., Hess-Pierce, B. & Kader, A.A. (1999). Quality changes in 

fresh-cut and nectarine slices as affected by cultivar, storage 


atmosphere and chemical treatment. Journal of Food Science, 64, 

429–432. 

Guilbert, S. (1986). Technology and application of edible protective 

films. In: Food Packaging and Preservation: Theory and Practice 

(edited by M. Matholouthi). Pp. 371–394. London: Elsevier Applied 

Science Publishing. 

Kader, A.A. (1986). Biochemical and physical basis of effects of 

controlled and modified atmospheres on fruits and vegetables. Food 


Kader, A.A. (1992). Modified atmospheres during transport and 

storage. In: Postharvest Technology of Horticultural Crops (edited by 

A.A. Kader). Pp. 85–92. CA: Division of Agriculture and Natural 

Resources, University of California. 

Kader, A.A. (1997). A summary of CA requirements and recommendations 

for fruits other than apples and pears. In: Fruits Other than 

Apples and Pears. (edited by A. Kader). Pp. 1–36. Postharvest Hort. 

Series No. 17, CA: University of California, CA’97 Proc. 2. 

Kester, J.J. & Fennema, O.R. (1986). Edible films and coatings. Food 


Lakshminarayana, S. (1980). Mango. In: Tropical and Subtropical 

Fruits (edited by S. Nagy & P.E. Shaw). Pp. 184–257. Connecticut: 

AVI Publishing. 

McHugh, T.H. & Olsen, C.W. (2004). Tensile properties of fruit and 

vegetable edible films. United States-Japan Cooperative Program in 

Natural Resources, 104–108. 

McHugh, T.H. & Senesi, E. (2000). Apple wraps: a novel method to 

improve the quality and extend the shelf-life of fresh-cut apples. 


McHugh, T.H., Huxsoll, C.C. & Krochta, J.M. (1996). Permeability 

properties of fruit puree edible films. Journal of Food Science, 61, 88– 

91. 

Perez-Gago, M.B., Serra, M. & Del Rio, M.A. (2003a). Effect of whey 

protein-beeswax edible composite coatings on color change of freshcut 

persimmon cv. ‘‘Rojo Brillante’’. Journal of Food Science, 65, 

432–436. 

Perez-Gago, M.B., Serra, M., Del Rio, M.A., Alonso, M. & Mateos, 

M. (2003b). Effect of solid content and lipid content of whey 


protein-beeswax edible composite coatings on color change of freshcut 

apple. Journal of Food Science, 68, 1–6. 

Petersson, M. & Stading, M. (2005). Water vapour permeability and 

mechanical properties of mixed starch-monoglyceride films and 

effect of film forming conditions. Food Hydrocolloids, 19, 123–132. 

Plotto, A., Goodner, K.L., Baldwin, E.A., Bai, J. & Rattanapanone, 

N. (2004). Effect of polysaccharide coatings on quality of fresh cut 

mangoes (Mangifera indica). Proceedings of the Florida State 

Horticultural Society, 117, 382–388. 

Reyes, V.G. (1996). Improved preservation systems for minimally 

processed vegetables. Food Australia, 48, 87–90. 

Schwimmer, S. (1978). Enzyme action and modifications of cellular 

integrity in fruits and vegetables: consequences for food quality 

during ripening, senescence and processing. In: Postharvest Biology 

and Biotechnology (edited by H.O. Hultin & M. Milner). Pp. 317– 

347. Connecticut: Food and Nutrition Press. 

Shewfelt, R.L. (1987). Quality of minimally processed fruits and 

vegetables. Journal of Food Quality, 10, 143–156. 

Sothornvit, R. & Rodsamran, P. (2008). Effect of a mango film on 

quality of whole and minimally processed mangoes. Postharvest 

Biology and Technology, 47, 407–415. 

SPSS (1999). SPSS for Windows Release 9.0.0. Chicago, IL: SPSS Inc. 

Tanada-Palmu, P.S. & Grosso, C.R.F. (2005). Effect of edible 

wheat gluten-based films and coatings on refrigerated strawberry 

(Fragaria ananassa) quality. Postharvest Biology and Technology, 36, 

199–208. 

Tharanathan, R.N. (2003). Biodegradable films and composite coatings: 

past, present and future. Food Science and Technology, 14, 71– 

78. 

Tovar, B., Ibarra, L.I., Garcia, A.S. & Mata, M. (2000). Some 

compositional changes in Kent mango (Mangifera indica) slices 

during storage. Journal of Applied Horticulture, 2, 10–14. 

Wong, D.W.S., Camirand, W.M. & Pavlath, A.E. (1994). Development 

of edible coatings for minimally processed fruits and vegetables. 

In: Edible Coatings and Films to Improve Food Quality (edited 

by J.M. Krochta, E.A. Baldwin & M.O. Nisperos-Carriedo). Pp. 65– 

82. Pennsylvania: Technomic Publishing Company. 


1696 


Influence of pH, NaCl and pre-incubation on utilisation of surimi 

wash water in generation of antioxidative material by using the 

Maillard reaction 

Chakree Thongraung* & Sureeporn Kangsanan 

Department of Food Technology, Faculty of Agro-Industry, Prince of Songkla University, HatYai, Songkhla, 90112, Thailand 


Summary Maillard reaction products (MRPs) were generated from reaction mixtures of surimi wash water (SWW) 

with glucose or fructose (5%w ⁄ v) heated at 95 °C for 2–12 h. The effects of pH, NaCl and pre-incubation of 

SWW on the Maillard reaction and antioxidant capacity of MRPs were investigated. The antioxidative 

capacity of MRPs was determined by measuring free DPPH° radical scavenging activity and reducing power. 

The highest colour intensity (OD 420) as well as antioxidative capacity was noted in the reaction mixture 

containing fructose at pH 9.0. The addition of NaCl (0.5–2.5%w ⁄ v) caused reduction in browning intensity 

but enhanced antioxidative capacity of the MRPs. Pre-incubation of SWW at 45 °C for 4 h decreased soluble 

protein but increased the Maillard reaction and antioxidative capacity of MRPs. A positive effect of salt or 

pre-incubation of SWW on the antioxidative capacity of MRPs was not associated with the soluble protein 

content in the reaction mixture. 

Keywords Maillard reaction, NaCl, pH, pre-incubation, surimi wash water. 

Introduction 

The annual production of surimi in Thailand is about 

140 000 tons and threadfin bream (Nemipterus sp., TB) 

accounts for approximately half of the raw material 

used for surimi (Piyadhammaviboon & Yongsawatdigul, 

2009). In commercial surimi processing, fish mince 

is washed in three times the volume of cold water, 

generating 15 L of wash water per kg kilogram of surimi 

produced (Durmaz & Alpaslan, 2007). Recovery of 

suspended fish muscle particles has been established in 

the industry. However, a significant portion of very fine 

fish muscle particles remains in the discarded wash water 

and is disposed along with soluble muscle proteins. The 

surimi wash water (SWW) contains about 0.5–2.3% 

total protein, which consist mainly of sarcoplasmic 

proteins. Consequently, several attempts were performed 

to recover these discarded materials and to 

transform them to higher value products (Huang & 

Morrissey, 1998; Mireles DeWitt & Morrissey, 2002). 

The Maillard reaction, or non-enzymatic browning, 

consists of a complex set of reactions, initially between 

carbonyl groups of reducing sugars and amino groups 


e-mail: chakree.t@psu.ac.th 


from amino acids, peptides or proteins, into subsequent 

complex reaction products, collectively known as Maillard 

reaction products (MRPs). The reaction contributes 

to qualitative attributes such as the colour, aroma and 

taste of thermally treated foods. These MRPs have been 

shown to retard lipid oxidation and play a role in 

extending the shelf-life of several food products (Augustin 

et al., 2006; Sun et al., 2007). Their potent inhibitory 

activity has been attributed to their free radical 

scavenging activity, reducing power and metal ion 

chelating activities. Thus, some food processing has 

been optimised to increase the antioxidative capacity of 

the food products via the development of the Maillard 

reaction (Durmaz & Alpaslan, 2007; Peralta et al., 

2008). There are numerous factors that influence the 

reaction and the characteristics and properties of the 

product. These include types of reducing sugar and 

amino acids ⁄ proteins, pH and temperature (Yaylayan 

et al., 1993; Villamiel et al., 2006; Lamberts et al., 2008). 

However, the Maillard reaction of SWW was not 

investigated in detail. 

The aim of the present study was thus to investigate 

how initial pH, NaCl and pre-incubation of SWW 

would affect the Maillard reaction of systems containing 

SWW. In addition, the effect of these parameters on 

antioxidative capacity of MRPs was investigated. 

doi:10.1111/j.1365-2621.2010.02318.x 



Materials 

SWW from the first washing step of the surimi processing 

was obtained from a local surimi plant. The raw 

material was threadfin bream (Nemipterus sp., TB). The 

wash water was collected after its suspended fish muscle 

particles were partially recovered. It was transported to 

the laboratory of the Faculty of Agro-Industry, Prince 

of Songkla University on ice within 2 h and was used 

for the experiment on the day of arrival. d-Glucose, 

d-fructose and 1,1-diphenyl-2-picryl-hydrazyl (DPPH) 

were purchased from Sigma ⁄ Aldrich (St Louis, MO, 

USA). All other reagents were of analytical grade. 

MRPs preparation 

Glucose or fructose (5%w ⁄ v) was dissolved in the 

SWW. The mixtures were adjusted to pH 5.0, 7.0 or 9.0 

with 1 n NaOH or 1 n HCl. All mixtures were heated in 

a water bath controlled at 95 °C for 12 h without pH 

control. The samples were removed at 1-h intervals and 

placed in an ice bath to cool before centrifugation at 

5478 · g for 20 min (Hermel-Z323, Hermle Laboratory, 

Ltd., Germany). The resulting supernatants were stored 

at 4 °C until they were used for analysis within two 

days, otherwise they were stored at )20 °C until they 

were used within a month. In addition, sodium chloride 

(0.5, 1.0, 1.5, 2.0 or 2.5%w ⁄ v) was added to the SWW 

before the Maillard reaction with fructose (5%w ⁄ v) at 

pH 9.0 and 95 °C. The effect of the added salt on soluble 

protein content of sample was evaluated by the Lowry 

method (Lowry et al., 1951). 

In one experiment, the SWW was incubated at 45 °C 

for 12 h. This incubated SWW was removed at 1-h 

intervals and used for the Maillard reaction with 

fructose (5%w ⁄ v) at pH 9.0 as previously described. 

The soluble protein in the incubated SWW at a 

predetermined heating time was quantified by the Lowry 

method (Lowry et al., 1951). 

Measurement of browning 

The progressive Maillard reaction of the samples was 

followed by reading the absorbance at 294 and 420 nm 

on an UV-visible microplate reader (PowerWaveX, 

Biotek, Winooska, VT, USA) after appropriate dilution. 

DPPH radical scavenging activity 

The DPPH radical scavenging activity of the MRPs was 

estimated according to the modified method reported by 

Jing & Kitts (2002). A diluted solution of MRPs was 

mixed with DPPH in ethanol in the well of a 96-well 

microplate to contain 0.3 m DPPH. The absorbance of 

the resulting solution was measured at 517 nm using an 

UV)visible microplate reader (PowerWaveX, Biotek, 

Winooska, VT, USA). Deionised water was used as 

control. The capability to scavenge the DPPH radical of 

the sample was expressed as the effective concentration 

needed to reduce 50% of the initial amount of DPPH° 

or EC50. 

Ferric reducing power 

The reducing power of the samples was determined 

according to the method described by Sun et al. (2007). 

Diluted solution of MRPs was mixed with deionised 

water, sodium phosphate buffer (0.2 m, pH 6.6) and 

potassium ferricyanide (1.0%). The solution obtained 

was incubated at 50 °C for 20 min. After trichloroacetic 

acid (10%) was added, the solution was centrifuged at 

7000 rpm for 5 min (Hermel-Z323, Hermle Laboratory, 

Ltd., Germany). The resulting supernatant was mixed 

with deionised water and ferric chloride (0.1%), and the 

absorbance was measured at 700 nm. 


The data in the tables were expressed as mean± SD 

(n = 3). Analysis of variance was performed by anova 

procedures. Duncan’s new multiple-range test was used 

to determine the differences between means. 

Results 

Antioxidative activity of MRPs C. Thongraung and S. Kangsanan 1697 

Effect of pH, NaCl and pre-incubation of SWW on the 

Maillard reaction 

The total protein and soluble protein content of SWW 

used for the Maillard reaction were 0.46 and 

0.17 mg mL )1 , respectively. Adjusting pH of this SWW 

from an initial value of 6.83 to 5.0 or 9.0 caused a 

reduction and an increase in soluble protein by 40% and 

5%, respectively. The absorbance at 294 nm used to 

detect intermediate stages of the reaction (Fig. 1a) 

increased progressively upon extension of the incubation. 

The highest value was found in the sample 

containing fructose at pH 9.0. There was a tendency 

that the development of intermediate products of the 

Maillard reaction levelled off after heating for 8 h. The 

significant development of melanoidin, the final stage 

Maillard reaction products, as measured by absorbance 

at 420 nm (Fig. 1b), was noticed only in the system 

containing fructose at pH 9.0. The highest absorbance 

value was also recorded after 8 h of heating. While 

pigment formation was favoured in the system with 

fructose at pH 9.0, browning of the systems at other pH 

values and systems of glucose remained low. 

As shown in Fig. 2, the contribution of NaCl to 

the Maillard reaction was verified by using the 


1698 

Antioxidative activity of MRPs C. Thongraung and S. Kangsanan 

OD 294 nm 

OD 420 nm 

OD 420 nm 

16.00 

14.00 

12.00 

10.00 

8.00 

6.00 

4.00 

2.00 

Fructose pH 7.0 Fructose pH 5 Fructose pH 9 

Glucose pH 7.0 Glucose pH 5 Glucose pH 9 

0.00 

2 4 6 8 10 12 

1.00 

0.90 

0.80 

0.70 

0.60 

0.50 

0.40 

0.30 

0.20 

0.10 

(a) 

(b) 

Heating time (h) 

0.00 

2 4 6 8 10 12 

4.00 

3.50 

3.00 

2.50 

2.00 

1.50 

1.00 

0.50 

0.00 


w/o Added NaCl 0.5% NaCl 1.0% NaCl 

1.5% NaCl 2.0% NaCl 2.5% NaCl 

2 4 6 8 10 12 


SWW-fructose mixture at pH 9.0. The addition of 

between 0.5% and 1.5% NaCl to the SWW had no 

significant effect on the total soluble protein of the 

samples. Salt at 2.0%w ⁄ v produced a modest increase 

( 2%) in the soluble protein of the sample, whereas the 

addition of salt at 2.5% w ⁄ v increased the soluble 

protein in SWW by about 8%. The presence of salt in 

the reaction mixtures had no significant effect on the 

development of OD294 throughout the heating treatment 

(data not shown). However, the presence of salt led to 

reduction in browning of about 18–29%, depending on 

Figure 1 Absorbance at 294 nm (a) and 

420 nm (b) in glucose or fructose-surimi 

wash water systems during heating at 95 °C 

at pH 5.0, 7.0 and 9.0. 

Figure 2 Effect of NaCl on absorbance at 

420 nm in fructose-surimi wash water 

during heating at 95 °C at pH 9.0. 

the concentration of added salt and the duration of 

heating (Fig. 2). There were no significant differences 

(P > 0.05) in effect among samples with different salt 

concentrations although the mixture with 0.5% w ⁄ v salt 

exhibited the lowest average OD420. 

Pre-incubation at 45 °C caused a sharp reduction in 

soluble protein in SWW during the first 2 h of the preincubation; 

thereafter, there were no significant changes 

in soluble protein. By the 12th h of pre-incubation, the 

soluble protein content of SWW was about 74% of 

initial concentration. The Maillard reaction of SWW 


OD 420 nm 

1.00 

0.90 

0.80 

0.70 

0.60 

0.50 

0.40 

0.30 

0.20 

0.10 

0.00 

Heating time 6 h 



0 2 4 6 8 10 12 

Pre-incubation time (h) 

Figure 3 Effect of pre-incubation of surimi wash water at 45 °C on 

browning intensity of MRPs by heating at 95 °C and during a reaction 

time of 6, 8 or 10 h. 

with fructose at pH 9.0 and 95 °C was enhanced by 

using the pre-incubated SWW with respect to that of the 

system using non-incubated wash water (Fig. 3). 

Effect of pH, NaCl and pre-incubation of SWW on 

antioxidative capacity of MRPs 

The effects of the initial pH values of the reaction 

mixtures on the antioxidative capacities of MRPs are 

shown in Fig. 4. A drastic reduction in EC50 of DPPH 

radical scavenging activity occurred during the first 4 h 

of heating, after which it slightly decreased with heating 

time up to 12 h (Fig. 4a). The reduction in EC50 indicated increasing in effectiveness of the MRPs to 

scavenge the DPPH radicals. The development patterns 

of DPPH radical scavenging activity of the systems 

containing fructose and glucose were similar, but the 

activity of the former system was about 2–4 times 

higher. The development of the reducing power of 

MRPs (Fig. 4b) showed similar trends with those of 

intensity of brownness (Fig. 1) and free radical scavenging 

activity (Fig. 4a). The systems containing fructose 

exhibited greater reducing power than the glucose 

systems throughout the incubation period. The results 

revealed that strong antioxidative activities of MRPs 

could be prepared by using fructose at pH 9.0. 

The DPPH scavenging activity and reducing power of 

MRPs prepared from the SWWs with added salt and 

5% w ⁄ v fructose at pH 9.0 are shown in Fig. 5. The 

addition of salt at 0.5 and 2.0% w ⁄ v lowered the radical 

scavenging activity of MPRs. However, after 4 h of 

heating, their activity was not significantly different 

from that of the control. Whereas salt at concentrations 

of 1.0%, 1.5% and 2.5%w ⁄ v promoted the development 

of stronger radical scavenging activity than those of the 

control and the other samples. In contrast, the reducing 

power of the system with added salt increased significantly 

(P < 0.05) upon extending the heating (Fig. 5b). 

(a) 

100 

DPPH Scavenging activity (EC50%) 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

(b) 

4.0 

OD 700 nm 


3.5 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

Fructose pH 5 Glucose pH 5 



2 4 6 8 10 12 


0.0 

2 4 6 8 10 12 


Figure 4 DPPH scavenging activity (a) and reducing power (b) of 

MRPs prepared by heating the mixtures of glucose or fructose-surimi 

wash water at pH 5.0, 7.0 or 9.0 and at 95 °C. 

Salt at 2.0% and 2.5%w ⁄ v increased noticeably the 

reducing power of MRPs relative to that of the control 

and other samples. 

Figure 6a depicts the EC50 of DPPH scavenging 

activity of the MRPs prepared by heating the preincubated 

SWW with fructose at pH 9.0 and 95 °C for 6, 

8 or 10 h. The pre-incubation showed either strengthened 

or deteriorated effects on the scavenging activity of 

MRPs. The strongest free radical scavenging activity of 

MRPs was obtained by using the 3rd-h-incubated SWW 

and heating for 10 h. The pre-incubation exhibited 

similar effect on the development of reducing power 

(Fig. 6b). The greatest enhancing effect (about a 10% 

increase) was also found in MRPs obtained by using the 

3rd-h-incubated SWW and a reaction time of 10 h. 

Discussion 

The results of this study revealed that SWW could be 

used as a starting material for the Maillard reaction. A 

positive correlation between antioxidative capacity and 

the browning intensity of the Maillard systems is well 

documented (Fuster et al., 2000; Yanagimoto et al., 


1700 


DPPH Radical scavenging activity 

(EC50%) OD 700 nm 

7 

6 

5 

4 

3 

2 

1 

w/o Added NaCl 0.5% NaCl 

1.0% NaCl 1.5% NaCl 

2.0% NaCl 2.5% NaCl 

0 

2 4 6 8 10 12 


8 

7 

6 

5 

4 

3 

(a) 

(b) 

2 

2 4 6 8 10 12 


Figure 5 Effect of NaCl on DPPH radical scavenging activity (a) and 

reducing power (b) of MRPs prepared by heating the fructose-surimi 

wash water at pH 9.0 and 95 °C. 

2002; Hidalgo et al., 2003). Although there have been 

some reports on loss of colour intensity and antioxidative 

capacity of MRPs because of the extension of the 

heating time (Ehling & Shibamoto, 2005). The proposed 

mechanisms are the precipitation of high molecular 

weight melanoidins caused by the lengthening of the 

heating time and the transformation of small polymers 

with antioxidative activity into higher polymers with 

less ⁄ no antioxidative activity (Gue´rard & Sumaya- 

Martinez, 2003; Hidalgo et al., 2003). It is, however, 

unclear if these mechanisms account for the stationary 

development of the browning intensity (Fig. 1) and 

antioxidative activity of MRPs (Fig. 4) after heating for 

8 h in this study. It is worth mentioning that significant 

amounts of pellets were decanted from all reaction 

mixtures. 

That pH affects the rate and extent of the Maillard 

reaction and that alkaline condition favours the reaction 

progress are well established (Ashoor & Zent, 1984; 

Brands et al., 2000; Davidek et al., 2002). The general 

mechanism proposed relay on an increase in availability 

of acyclic form, a high reactivity form, of reducing sugar 

under alkaline pH. The explanation also underlines a 

strong reactivity of fructose with respect to that of 

DPPH Radical scavenging activity 

(EC50 ,%) 

OD 700 nm 

8 

7 

6 

5 

4 

3 

2 

1 

0 

5.00 

4.50 

4.00 

3.50 

3.00 

2.50 

2.00 

(a) 

(b) 

Heating time 6 h Heating time 8 h Heating time 10 h 

1 2 3 4 5 6 7 


1 2 3 4 5 6 7 


Figure 6 Effect of pre-incubation of surimi wash water at 45 °C on 

DPPH radical scavenging activity (a) and reducing power (b) of MRPs 

prepared by heating the fructose-surimi wash water at pH 9.0 and 

95 °C during a reaction time of 6, 8 or 10 h. 

glucose of this study. Moreover, alkaline pH was also 

favour the degradation of Amadori compound in both 

buffered and non-buffered aqueous systems with an 

enhancing effect on the concomitant development of 

browning (Davidek et al., 2002). Furthermore, the 

difference in brown colour intensity observed in the 

systems of fructose and glucose of this study could be 

partly explained by the contribution of some caramelisation 

reaction products to the absorbance values. 

Ajandouz et al. (2001) found that the caramelisation 

rate of these sugars during heating at 100 °C was 

increased by the increasing of medium pH particularly 

in the pH range 8.0–12.0, and fructose showed a greater 

reactivity with respect to that of glucose. Significant 

difference in browning development at pH 9.0 among 

the systems having identical soluble protein but with 

different sugar thus highlighted the difference in reactivity 

of each reducing sugar on the Maillard browning. 

Our observation on the effect of salt on browning was 

in line with findings by Kwak & Lim (2004) and Rizzi 

(2008). In the amino acid model systems, the effect of salt 

on inhibiting the browning of MRPs was increased with 

the increasing of its concentration, and at a concentration 

of 1%w ⁄ v its effect was dependent on the type of 

amino acid. A reduction in the colour intensity of the 

samples in the presence of salt is indicative of the 


decreasing in the polymerisation extent of the preformed 

intermediate MRPs. In contrast, the addition of salt 

enhanced the antioxidative capacity of the MRPs suggesting 

that some mechanistic differences most likely 

exist among the systems with and without NaCl. It is 

likely that salt favours the reaction route that produces 

products having strong antioxidative capacity. However, 

more research is necessary to validate this hypothesis. 

Although an extensive development of browning 

colour was noticed by using pre-incubated SWW 

(Fig. 3), this positive effect could not be explained by 

the availability of soluble protein (Fig. S1). Since SWW 

contains 0.25% w ⁄ v lipid, the oxidation of lipid could be 

stimulated as development of a strong fishy odour was 

noticed during prolonged pre-incubation. These would 

cause an accumulation of preformed carbonyl compounds. 

Apart from the carbonyl group of reducing 

sugar, Amadori compounds can react with other 

carbonyls especially the oxidised lipid, and the same 

products are frequently produced by identical or very 

similar reaction pathways (Zamora & Hidalgo, 2005). 

Even if the observation implies the existence of oxidised 

lipids, further study is necessary to obtain confirmation 

of an explicit contribution of lipid oxidation to the 

Maillard reaction. 

Maillard reaction yield products based on specific time 

temperature and pressure of the heating process. These 

are commonly used in the industry and its process may 

be engineered by controlling heating condition. In this 

work, prolonged heating at 1 atm may be replaced with 

shorter heating time at higher static pressure. Therefore, 

heating time can be greatly reduced and the more 

process economical. Practical use may include developing 

antioxidative supplements that are meat base and 

can be used together with a meal e.g. soup, main course. 


This study was supported in part by the Prince of 

Songkla University and the Agricultural Research 

Development Center (ARDA). Authors are also grateful 

to the ARDA for manuscript preparation. 

References 

Ajandouz, E.H., Tchiakpe, L.S., Dalle Ore, F., Benajiba, A. & 

Puigserver, A. (2001). Effects of pH on caramelization and maillard 

reaction kinetics in fructose-lysine model systems. Journal of Food 

Science, 66, 926–931. 

Ashoor, S.H. & Zent, J.B. (1984). Maillard browning of common 

amino acids and sugars. Journal of Food Science, 49, 1206–1207. 

Augustin, M.A., Sanguansri, L. & Bode, O. (2006). Maillard reaction 

products as encapsulants for fish oil powders. Journal of Food 

Science, 71, E25–E32. 

Brands, C.M.J., Alink, G.M., Van Boekel, M.A.J.S. & Jongen, 

W.M.F. (2000). Mutagenicity of heated sugar-casein systems; effect 

of the Maillard reaction. Journal of Agricultural and Food Chemistry, 

48, 2271–2275. 


Davidek, T., Clety, N., Aubin, S. & Blank, I. (2002). Degradation of 

the amadori compound N-(1-Deoxy-d-fructos-1-yl)glycine in aqueous 

model systems. Journal of Agricultural and Food Chemistry, 50, 

5472–5479. 

Durmaz, G. & Alpaslan, M. (2007). Antioxidant properties of roasted 

apricot (Prunus armeniaca L.) kernel. Food Chemistry, 100, 1177– 

1181. 

Ehling, S. & Shibamoto, T. (2005). Correlation of acrylamide 

generation in thermally processed model systems of asparagine 

and glucose with color formation, amounts of pyrazines formed, and 

antioxidative properties of extracts. Journal of Agricultural and Food 

Chemistry, 53, 4813–4819. 

Fuster, M.D., Mitchell, A.E., Ochi, H. & Shibamoto, T. (2000). 

Antioxidative activities of heterocyclic compounds formed in 

brewed coffee. Journal of Agricultural and Food Chemistry, 48, 

5600–5603. 

Guérard, F. & Sumaya-Martinez, M.-T. (2003). Antioxidant effects of 

protein hydrolysates in the reaction with glucose. Journal of the 

American Oil Chemists’ Society, 80, 467–470. 

Hidalgo, F.J., Nogales, F. & Zamora, R. (2003). Effect of the pyrrole 

polymerization mechanism on the antioxidative activity of nonenzymatic 

browning reactions. Journal of Agricultural and Food 

Chemistry, 51, 5703–5708. 

Huang, L. & Morrissey, M.T. (1998). Fouling of membranes during 

microfiltration of surimi wash water: roles of pore blocking and 

surface cake formation. Journal of Membrane Science, 144, 113– 

123. 

Jing, H. & Kitts, D.D. (2002). Chemical and biochemical properties of 

casein-sugar Maillard reaction products. Food and Chemical Toxicology, 

40, 1007–1015. 

Kwak, E.J. & Lim, S.I. (2004). The effect of sugar, amino acid, metal 

ion, and NaCl on model Maillard reaction under pH control. Amino 

Acids, 27, 85–90. 

Lamberts, L., Rombouts, I. & Delcour, J.A. (2008). Study of 

nonenzymic browning in [alpha]-amino acid and [gamma]aminobutyric 

acid ⁄ sugar model systems. Food Chemistry, 111, 

738–744. 

Lowry, O.H., Rosebrough, N.J., Farr, A.L. & Randall, R.J. (1951). 

Protein measurement with the Folin reagent. Journal of Biological 

Chemistry, 193, 265–275. 

Mireles Dewitt, C.A. & Morrissey, M.T. (2002). Pilot plant recovery of 

catheptic proteases from surimi wash water. Bioresource Technology, 

82, 295–301. 

Peralta, E.M., Hatate, H., Kawabe, D. et al. (2008). Improving 

antioxidant activity and nutritional components of Philippine saltfermented 

shrimp paste through prolonged fermentation. Food 

Chemistry, 111, 72–77. 

Piyadhammaviboon, P. & Yongsawatdigul, J. (2009). Protein crosslinking 

ability of sarcoplasmic proteins extracted from threadfin 

bream. LWT – Food Science and Technology, 42, 37–43. 

Rizzi, G.P. (2008). Effects of cationic species on visual color formation 

in model maillard reactions of pentose sugars and amino Acids. 


Sun, Y., Hayakawa, S., Ogawa, M. & Izumori, K. (2007). Antioxidant 

properties of custard pudding dessert containing rare hexose, 

d-psicose. Food Control, 18, 220–227. 

Villamiel, M., Del Castillo, D. & Corzo, N. (2006). Browning 

reactions. In: Food Biochemistry and Food Processing (edited by 

Y.H. Hui, W.-K. Nip, L.M.L. Nollet, G. Paliyath & B.K. Simpson). 

Pp. 71–100. Oxford, UK: Blackwell. 

Yanagimoto, K., Lee, K.-G., Ochi, H. & Shibamoto, T. (2002). 

Antioxidative activity of heterocyclic compounds found in coffee 

volatiles produced by maillard reaction. Journal of Agricultural and 


Yaylayan, V.A., Ismail, A.A. & Mandeville, S. (1993). Quantitative 

determination of the effect of pH and temperature on the keto form 

of -fructose by FT IR spectroscopy. Carbohydrate Research, 248, 

355–360. 


1702 


Zamora, R. & Hidalgo, F.J. (2005). Coordinate contribution of lipid 

oxidation and maillard reaction to the nonenzymatic food browning. 

Critical Reviews in Food Science and Nutrition, 45, 49–59. 




Figure S1. Effect of pre-incubation of surimi wash 

water at 45 °C on soluble protein in surimi wash water. 









Equilibrium moisture content and heat of desorption of saffron 

Hamid Reza Gazor 1 * & Hossein Chaji 2 

1 Agricultural Engineering Research Institute, Ministry of Agriculture, PO Box 31585-845, Karaj, Iran 

2 Center of Agriculture and Natural Resources of Khorasan Razavi Province, Ministry of Agriculture, Mashhad, Iran 


Summary The equilibrium moisture contents of saffron (Crocus sativus L.) stigmas were determined experimentally 

using the standard gravimetric method at temperatures 30, 45 and 60 °C and water activity ranging from 

11% to 83%. The sorption isotherm curves of saffron were sigmoidal in shape and decreased with increased 

temperature at constant relative humidity. Five selected isotherm models GAB, modified Henderson, 

modified Chung-Pfost, modified Halsaey and modified Oswin were tested to fit the experimental isotherm 

data. Modified Oswin and modified Henderson models were found acceptable for predicting desorption 

moisture isotherms and fitting to the experimental data, respectively. The isosteric heats of desorption, 

determined from equilibrium data using the Clausius-Clapeyron equation, were found to be a function of 

moisture content. The net isosteric heat of desorption of saffron varied between 1.38 and 5.38 kJ mol )1 at 

moisture content varying between 2% and 20% (d.b). 

Keywords Desorption isotherms, equilibrium moisture content, isosteric heat of sorption, saffron. 

Introduction 

Saffron, the dried stigmas of a flower scientifically 

identified as Crocus sativus L., is one of the most 

expensive agricultural products. It has been named as 

‘red gold’ because of its high economic value. Although 

the source of saffron is historically unknown, it apparently 

originated in the area of Iran, Turkey and Greece 

(Bolandi & Ghoddusi, 2006; Caballero-Ortega et al., 

2007). While the world’s total annual saffron production 

is estimated to be 190 tons, Iran produces about 90% of 

the total, accounting for 150–170 tons per year, and 

50 000 ha is already under cultivation with this crop. 

This spice is appreciated for its unique colour, bitter 

taste and aroma, which are the main characteristics 

showing its quality (Bolandi & Ghoddusi, 2006; Caballero-Ortega 

et al., 2007). Drying using microwave oven 

appears to be a suitable way to achieve a high quality for 

saffron but it is costly. The market value potential and 

farmer interest should be further evaluated (Bolandi & 

Ghoddusi, 2006). To arrive at a conclusion, the drying 

and storage conditions are of great importance because 

they determine the quality add and hence economic 

value of the final product (Winterhalter & Straubinger, 

2000; Bolandi & Ghoddusi, 2006). The stability of 

saffron pigments is dependent on the storage conditions 

employed. Both increasing temperature and water 

activity exert a strong influence on the degradation 

*Correspondent: E-mail: hgazor@yahoo.com 

doi:10.1111/j.1365-2621.2010.02321.x 


kinetics, accelerating the rate of pigment decomposition 

(Tsimidou & Biliaderis, 1997; Winterhalter & Straubinger, 

2000). 

Water activity is a term used to indicate the relation 

between a food and the equilibrium relative humidity of 

the surrounding atmosphere (Barbosa-Cańovas & Vega- 

Mercado, 1996). The equilibrium moisture content is 

dependent upon the temperature and related humidity 

of the environment as well as species, variety and 

maturity of the plant (Iglesias & Chirife, 1982; Rahman, 

1995). Accurate information on equilibrium moisture 

contents of saffron stigma at various relative humidity 

and drying temperatures is not available in the literature. 

There is also a need for comprehensive study of the 

equilibrium moisture contents of saffron stigma to 

understand its drying behaviour. The net isosteric heat 

of sorption is also an important information for drying. 

It can be used to determine the energy requirements and 

provide information on the state of water within the 

dried product (Mujumdar, 2000). The moisture content 

level of a product at which the net isosteric heat of 

sorption reaches the value of latent heat of sorption is 

often considered as the indication of the amount of 

bound water existing in the product (Wang & Brennan, 

1991). The five empirically models Guggenheim- 

Anderson-de Boer (GAB), modified Henderson, modified 

Chung-Pfost, modified Halsaey and modified Oswin 

are the most commonly used models for sorption 

isotherm determination of food materials (Chen, 2000; 

Janjai et al., 2006; Ghodake et al., 2007).

1704 

Equilibrium moisture contents of saffron H. R. Gazor and H. Chaji 

The objectives of this research were the following: 

• Determination of the effect of temperature on the 

moisture desorption isotherms of saffron stigmas that 

here in by abbreviated to saffron in the paper. 

• Analysis of the data with the help of five sorption 

isotherm models available in the literature and find the 

most suitable model describing the desorption isotherms 

of saffron. 

• Calculation of the net isosteric heat of desorption of 

moisture from the experimental data. 


Experimental procedure 

The dried saffron was supplied from a farm in Qaen 

(Southern Khorasan province, Iran). Fresh saffron 

stigmas were dried at room temperature in the shadow 

previously. Naturally dried saffron was conditioned with 

distilled water in refrigerator at temperature 4 °C for 

2 weeks. Toluene was used in sample preparation to 

avoid sample spoilage. The initial moisture content of 

the conditioned samples was determined by using the 

standard method number 259-2 (103±2 °C 16 h). It was 

found about 54 (% d.b) at experiments start time (ISIR, 

2004). 

The sorption method used was the static gravimetric 

technique, which is based on the use of saturated salt 

solutions to maintain a fixed relative humidity when the 

equilibrium is reached. The water activity of the food is 

identical to the relative humidity of the atmosphere at 

equilibrium conditions, and the mass transfer between 

the product and the ambient atmosphere is assured by 

natural diffusion of the water vapour. Equilibrium 

relative humidity can be present as per cent of water 

activity (Rahman, 1995). Seven saturated salt solutions 

(Table 1) were prepared corresponding to a wide range 

of water activities ranging from 0.11 to 0.83 (Greenspan, 

1977; Labuza et al., 1985; Adam et al., 2000). 

Table 1 Equilibrium relative humidity of saturated salt solutions at 30, 

40, 50 and 60 °C* 

Salt solutions 

Temperature (°C) 

30 40 50 60 

LiCl 0.1128 0.1121 0.111 0.1095 

CHC 3COOK 0.2161 0.204 0.192 0.180 

MgCl 2Æ6H 2O 0.324 0.316 0.305 0.293 

K2CO3 0.431 0.4330 0.4269 0.4212 

Mg(NO3)2 0.514 0.4842 0.4544 0.473 

NaCl 0.751 0.747 0.7434 0.745 

KaCl 0.836 0.823 0.812 0.8025 

*Greenspan (1977); Labuza et al. (1985). 

Figure 1 Precision oven and sealed glass bottles that used in the 

experiments. 

Seven sealed glass bottles were used in this research. 

Each bottle was provided with a sample holder that was 

hanged up in there. Sample holder was kept above 

saturated salt solutions to avoid contact of the salt 

solution with it. For each of these experiments, about 

1.5 g of saffron stigma was taken into the respective 

weighing cups. All these seven bottles were put in an 

oven with precision 1 °C and circulation fan (Fig. 1.) 

The loss in weights of all these samples in each bottle 

was monitored daily after 10 days. The Equilibrium 

Moisture Content (EMC) was acknowledged when three 

consecutive weight measurements showed a difference of 

Table 2 Sorption isotherm models 

Equations Models 

hr ¼ exp½ expðC1 þ C2tÞM C3 

e Š Modified-Halsey 

1 hr ¼ exp½ C1ðt þ C2ÞM C3 

e Š Modified Henderson 

hr ¼ 1 

Modified-Oswin 

1þ½ðC1þC2tÞ=MeŠ C3 hr ¼ exp½ 

C1 

tþC2 expð C3MeÞŠ Modified Chung–Pfost 

M0C1C2hr 

Me ¼ ð1 C2hrÞð1 C2hrþC1C2hrÞ GAB 

parameters of the models were determined by least 

square estimation and STATISTICA software (StatSoft, 

Inc, Tulsa, Oklahoma, USA) (version 6.0, 2001). 

In this table hr is the relative humidity (decimal), Me is 

the equilibrium moisture content or EMC (% d.b) M 0 

monolayer moisture content (% d.b), t is the temperature 

(°C) and C1, C2, C3 are coefficients of models. 

The parameters C 1 and C 2 in GAB model could be 

correlated with temperature using the following Arrhenius-type 

equations (Jamali et al., 2006): 

C1 ¼ C0 expð DHc 

Þ ð1Þ 

RT 

C2 ¼ K0 expð DHk 

Þ ð2Þ 

RT 

In these equations DHc, the enthalpy difference between 

monolayer and multilayer sorption (J mol )1 ), DH k the 

difference between heat of condensation and heat of 

sorption of the multilayer sorption (J mol )1 ), T the 

temperature (K), C0 the constant, K0 the constant and R 

the universal gas constant (8.314 J mol )1 K). 

Validation of EMC models 

The coefficients of the individual equations were determined 

by means of a standard regression estimation. 

The various EMC models were evaluated for their 

suitability in predicting the EMC of the sample on the 

basis of the coefficient of determination (R 2 ), standard 

error of estimate (SEE), mean relative of error (MRE) 

and randomness of residuals (e) These statistical 

parameters were calculated employing the following 

equations (Chen, 2000; Basunia & Abe, 2001; Lahsasni 

et al., 2003; Mohamed et al., 2004; Ghodake et al., 

2007): 

Table 3 Experimental data of desorption 

isotherms of saffron 30 °C 45 °C 60 °C 

Figure 2 Influence of temperature on saffron 

equilibrium moisture content (EMC) vs. 

equilibrium relative humidity (ERH). 

ERH (decimal) EMC (% d.b) ERH (decimal) EMC (% d.b) ERH (decimal) EMC (% d.b) 

0.1128 2.563 0.1116 1.434 0.1095 0.835 

0.2161 3.567 0.197 2.575 0.1748 1.837 

0.3244 4.920 0.311 4.323 0.2926 3.517 

0.4317 6.845 0.4299 5.899 0.4212 4.795 

0.514 11.176 0.4693 8.407 0.4726 7.810 

0.7509 19.188 0.7452 18.268 0.745 17.678 

0.8362 28.441 0.8174 24.211 0.8025 20.025 

Saffron EMC (% d.b) 

30 

25 

20 

15 

10 

Equilibrium moisture contents of saffron H. R. Gazor and H. Chaji 1705 

5 

0 

30 °C 

45 °C 

60 °C 

0 10 20 30 40 50 60 70 80 90 

Ó 2010 Institute of Food Science and Technology International Journal of Food Science and Technology 2010, 45, 1703–1709 

ERH (%)

1706 


Parameters 

Modified 

Henderson 

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 

P 

SEE ¼ 

N 

ðMi;exp Mi;preÞ 

i¼1 

2 

v 

u 

t 

df 

MRE ¼ 100 X 

N 

N 

i¼1 

Modified 

Chung-Pfost 

Mi;exp Mi;pre 

Mi;exp 

Modified 

Halsey 

ð3Þ 

ð4Þ 

e ¼ Mi;exp Mi;pre ð5Þ 

Where, Mi,exp is the experimental value of EMC; Mi,pre 

is the value predicted by the model; N is the number of 

data points; n is the number of coefficients in each 

model; df the degrees of freedom of regression model 

(N-n). It is generally considered that MRE values below 

10% indicate an adequate fit for practical purposes (Liu- 

Ping et al., 2005). 

Modified 

Oswin GAB 

C1 0.000834 96.98401 2.063523 10.18627 2.231425 

C2 85.32607 287.1498 )0.005016 )0.043299 0.927414 

C3 0.869323 10.36660 1.075747 1.390257 

M0 6.734712 

C0 

675579.9 

K0 4.17306 

DHc )0.33587 

DHk 

)0.04076 

R 2 

0.9885 0.9526 0.9826 0.9910 0.9816 

SEE 0.9707 1.9104 1.1564 0.6959 1.1914 

MRE (%) 11.5269 43.7620 26.5511 11.3176 15.3654 

SEE, standard error of estimate; MRE, mean relative of error. 

Saffron EMC (% d.b) 

35 

30 

25 

20 

15 

10 

Observed (30 °C) 

Pridicted (30 °C) 


5 

0 




0 10 20 30 40 50 60 70 80 90 

RH (%) 

Table 4 Model coefficients and statistical 

results of equations 

Figure 3 Observed values and predicted 

desorption isotherms by modified Oswin 

model for saffron at 30, 45 and 60 °C. 

Determination of net isosteric heat of desorption 

The net isosteric heat of desorption can be determined 

from moisture desorption data using the following 

equation, which is derived from the Clausius–Clapeyron 

equation (Adam et al., 2000; Mohamed et al., 2004; 

Janjai et al., 2006; Ghodake et al., 2007): 

@ lnðhrÞ 

@T 

¼ Qst 

RT 2 

ð6Þ 

Assuming that the net isosteric heat of desorption 

(Q st) is temperature independent, integrating eqn 3, 

gives the following equation: 

lnðhrÞ ¼ ð Qst 1 

Þ þ K ð7Þ 

R T 

Where K is a constant. The value of Qst is calculated 

from the slope of eqn 7. Using the computer software 


Figure 4 Plot of residuals fit of the 

modified Oswin and the Hendeson models 

to desorption data of saffron. 

Statistica 6.0 and Microsoft Excel, the nonlinear 

optimisation method was used to find the best equation 

for the saffron desorption isotherms and the net isosteric 

heat of desorption. 


–0.5 

–1.5 

Desorption curves 

Experimental data on desorption isotherm of saffron are 

shown in Table 3. Desorption isotherms (points) of 

saffron are shown in Fig. 2 at three temperature levels of 

30, 45 and 60 °C, in the range of 11–83% relative 

humidity. The isotherm curves have sigmoid shape and 

similar patterns. EMC values decreases with increase in 

temperature at all levels of relative humidity. This means 

Residuals 

Residuals 

1.5 

1 

0.5 

–1 

–1.5 

–2.5 

Residuals for modified Oswin model 

0 

0 5 10 15 20 25 30 35 

2.5 

2 

1.5 

1 

0.5 

–1 

–2 


Predicted EMC (% d.b) 

Residuals for Modified Henderson model 

0 

0 

–0.5 

5 10 15 20 25 30 35 

Predicted EMC (% d.b) 

that saffron becomes less hygroscopic with increase in 

temperature. The decrease in EMC value with the 

increase in temperature at constant equilibrium relative 

humidity can be explained by the fact that the kinetic 

energy associated with water molecules present in 

saffron increases with increase in temperature. 

This, in turn, resulted in decreasing attractive forces 

and escaping of water molecules. Consequently that 

leads to a decrease in EMC values with increase in 

temperature at a given relative humidity. Furthermore, 

at constant temperature, EMC value increases with 

increase in relative humidity. Several researchers have 

reported similar trends for plants and food materials 

(Basunia & Abe, 2001; Lahsasni et al., 2003; Mohamed 

et al., 2004; Janjai et al., 2006; Ghodake et al., 

2007). 


1708 


Heat of desorption (kJ mol –1 ) 

6 

5 

4 

3 

2 

1 

0 

0 5 10 15 20 25 

EMC (% d.b) 

Fitting of sorption models to experimental sorption data 

The desorption curves for saffron are drawn as equilibrium 

moisture content against the equilibrium relative 

humidity. These curves are used to estimate the coefficients 

of different sorption models considered (Table 2). 

The coefficients of models, their MRE and SEE are 

presented in Table 4. It is clear from this table that the 

modified Oswin model gives better fit to the experimental 

data with highest R 2 and lowest values of MRE and 

SEE than other models for desorption isotherms of 

saffron for a wide range of water activity. Also, modified 

Henderson model has good fitting with experimental 

data. Comparison between observed values and predicted 

desorption isotherms by modified Oswin model 

for saffron at 30, 45 and 60 °C is shown in Fig. 3. 

The analysis of the residual values vs. the predicted 

EMC values of desorption is in agreement with the 

previous results. Figure 4 shows the residuals of the 

predicted equilibrium moisture contents for the modified 

Oswin model and the Hendeson model. The modified 

Oswin model fitted the best to the isotherm data of 

saffron, and the residuals for the Oswin model are 

random in pattern. Hence, the Oswin model is acceptable. 

Heat of desorption 

Isosteric heat of desorption of saffron for different 

moisture contents was determined using the modified 

Oswin model that gave the best fitting of equilibrium 

moisture data in the range of temperatures from 30 to 

60 °C, in combination with eqn 7. The variations of 

isosteric heat of desorption for the different moisture 

contents are shown in Fig. 5. It is observed that for the 

range of temperatures from 30 to 60 °C and moisture 

contents from 2% to 20%. (d.b.), isosteric heat of 

desorption ranged from 1.38 to 5.38 kJ mol )1 . This 

indicates that at low moisture contents, the heat of 

desorption is higher than at high moisture contents. The 

net isosteric heat of desorption in saffron can be 

expressed by an exponential function of moisture 

content (Ghodake et al., 2007): 

Qst ¼ 6:19331 expðMeÞ 0:8104 R 2 ¼ 0:997 ð8Þ 

Conclusions 

Figure 5 Net isosteric heat of desorption 

isotherms for different moisture contents. 

The moisture desorption curves of saffron stigmas 

obtained at three temperatures (30, 45, and 60 °C) 

showed a sigmoid shape, as similar to previous studies. 

Desorption equilibrium moisture data have been 

collected for an over range of temperatures and relative 

humidities commonly used in drying of saffron. These 

results represent the preliminary stage of study of the 

convective drying kinetics of saffron in a conventional 

drier working with an auxiliary heating system. 

The experimental data of the desorption of saffron 

fitted to five isotherm models, so as to determine the best 

model for predicting the desorption equilibrium moisture 

contents of saffron. Modified Oswin and modified 

Henderson models were found acceptable for predicting 

desorption moisture isotherms and fitting to the experimental 

data, respectively. 

The heat of desorption of saffron decreases with 

an increase in moisture content and was found to be 

an exponential function of moisture content for desorption. 


References 

Adam, E., Mu¨ hlbaver, W., Esper, A., Wolf, W. & Spieb, W. (2000). 

Effect of temperature on water sorption equilibrium of onion 

(Allium cepa L). Drying Technology, 18, 2117–2129. 

Barbosa-Cánovas, G.V. & Vega-Mercado, H. (1996). Dehydration of 

Foods. Pp. 32–33. New York: Chapman & Hall. 

Basunia, M.A. & Abe, T. (2001). Thin layer solar drying characteristics 

of rough rice under natural convection. Journal of Food Engineering, 

47, 295–301. 

Bolandi, M. & Ghoddusi, H.B. (2006). Flavour and colour changes 

during processing and storage of saffron (Crocus sativus L.). 

Developments in Food Science, 43, 323–326. 

Caballero-Ortega, H., Pereda-Miran, R. & Abdullaev, F. (2007). 

HPLC quantification of major active components from 11 different 

saffron (Crocus sativus L.) sources. Food Chemistry, 100, 1126–1131. 

Chen, Ch. (2000). A rapid method to determine the sorption isotherm 

of peanuts. Journal of Agricultural Engineering Research, 75, 401– 

408. 

Ghodake, H.M., Goswami, T.K. & Chakraverty, A. (2007). Moisture 

sorption isotherms, heat of sorption and vaporization of withered 

leaves, black and green tea. Journal of Food Engineering, 78, 827– 

835. 

Greenspan, L. (1977). Humidity fixed points of binary saturated 

aqueous solutions. Journal of Research of the National Bureau of 

Standards – Section A. Physics and Chemistry, 81, 89–96. 

Iglesias, H.A. & Chirife, J. (1982). Hand Book of Food Isotherms. Pp. 

170–175. New York: Academic Press. 

ISIR (2004). Saffron – Test methods. Standard No 259-2 (2nd 

revision). Institute of Standards and Industrial Research of Iran. 

Jamali, A., Kouhila, M., Ait Mohamed, L., Jaouhari, J.T., Idlimam, 

A. & Abdenouri, N. (2006). Sorption isotherms of Chenopodium 


ambrosioides leaves at three temperatures. Journal of Food Engineering, 

72, 77–84. 

Janjai, S., Bala, B.K., Tohsing, K. et al. (2006). Equilibrium moisture 

content and heat of sorption of longan (Dimocarpus longan Lour.). 

Drying Technology, 24, 1691–1696. 

Labuza, T.P., Kaanane, A. & Chen, Y. (1985). Effect of temperature 

on the moisture sorption isotherms and water activity shift of two 

dehydrated foods. Journal of Food Science, 50, 385–392. 

Lahsasni, S., Kouhila, M., Mahrouz, M. & Fliyou, M. (2003). 

Moisture adsorption desorption isotherms of prickly pear cladode 

(Opuntia ficus indica) at different temperatures. Energy Conversion 

and Management, 44, 923–936. 

Liu-Ping, F., Min, Z., Qian, T. & Gong-Nian, X. (2005). Sorption 

isotherms of vacuum-fried carrot chips. Drying Technology, 23, 

1569–1579. 

Mohamed, L.A., Kouhila, M., Jamali, A., Lahsasni, S. & Mahrouz, 

M. (2004). Moisture sorption isotherms and heat of sorption of 

bitter orange leaves (Citrus aurantium). Journal of Food Engineering, 

67, 491–498. 

Mujumdar, A.S. (2000). Drying Technology in Agriculture and Food 

Sciences. Pp. 26–27. Enfield (NH), USA: Science publisher, Inc. 

Rahman, M.D.S. (1995). Hand Book of Food Properties. Pp. 2–11. 

New York: CRC Press. 

Tsimidou, M. & Biliaderis, C.G. (1997). Kinetic studies of saffron 

(Crocus sativus L.) quality deterioration. Journal of Agricultural and 


Wang, N. & Brennan, J.G. (1991). Moisture sorption isotherm 

characteristics of potatoes at four temperatures. Journal of Food 


Winterhalter, P. & Straubinger, M. (2000). Saffron: renewed interest in 

an ancient spice. Food Review International, 16, 39–59. 


1710 


Use of visible spectroscopy to assess colour development during 

ageing of fresh pork from different quality classes 

Marta Castro-Giráldez, 1 Pedro J. Fito, 1 Fidel Toldrá 2 & Pedro Fito 1 * 

1 Instituto Universitario de Ingeniería de Alimentos para el Desarrollo, Universidad Politécnica de Valencia, Camino de Vera s ⁄ n, 46022 Valencia, 

Spain 

2 Instituto de Agroquímica y Tecnología de Alimentos (CSIC), PO Box 73, 46100 Burjassot, Valencia, Spain 

(Received 9 November 2009; Accepted in revised form 24 May 2010) 

Summary The objective of this research was to study the entire visible spectra evolution during meat ageing in different 

pork meat quality classes (PSE, DFD and RFN). The potential use of the visible spectra for discriminating 

low meat quality classes during the 24 hours postmortem (hpm) was also analysed. For these purposes, 26 

pork loins were used: 3 PSE (pale, soft and exudative), 3 DFD (dark, firm and dry) and 20 RFN (red, firm 

and non-exudative). At 12, 24, 48 h and 7 days postmortem, reflectance spectra (400–700 nm) were obtained 

by a spectrocolorimeter Minolta CM-3600D after one and a half hour of blooming time. It was 

demonstrated that the ageing time has an influence in colour parameters and in blooming ability of RFN 

loins. The evolution of visible spectra was influenced by ageing time in PSE and RFN loins, while the visible 

spectra of DFD loins showed no variation with postmortem time. The results showed the possibility of 

separating PSE meats from the other classes by using visible reflectance spectra at 24 hpm. 

Keywords Colour, colour development, dark, firm and dry meats, pale, soft and exudative meats, quality, reflectance spectra, visible 

spectroscopy. 

Introduction 

Colour is one of the most important meat characteristics 

that consumers consider before buying, and is also 

related to the sensory characteristics of meat (Norman 

et al., 2003). The colour of pork is influenced by the 

pigment content, the different redox states of hemoprotein 

myoglobin: deoxymyoglobin (DMb) (purple colour), 

oxymyoglobin (MbO2) (bright red colour) and 

metmyoglobin (MetMb) (brown colour) and by meat 

structure (Lindahl et al., 2001). During meat ageing, any 

factor that acts on the myoglobin or indirectly changes 

the pH and its rate of decline can affect meat colour 

evolution (Honikel, 1997). 

Abnormal postmortem glycolysis during the conversion 

of muscle to meat produces extreme progress in pH, 

which affects negatively to water-holding capacity and 

colour (Bendall & Swatland, 1988), defining the main 

low quality classes in pork meat. Dark, firm and dry 

(DFD) meats are characterised by an ultimate pH 

higher than the isoelectric point of proteins, which 

produces a high water-holding capacity, which is asso- 


e-mail: pfito@tal.upv.es 


ciated with greater translucence and less scatter of 

incident light, making the meat appear darker (Mac- 

Dougall & Jones, 1981; MacDougall, 1982). Moreover, 

at high pH (>6) mitochondrial oxygen consumption is 

high and remains so during some time post-rigour 

(Bendall & Taylor, 1972); consequently, in meat exposed 

to air the purple colour of deoxymyoglobin predominates, 

being the surface oxymyoglobin layer thin 

(Ashmore et al., 1972). On the other hand, pale, soft 

and exudative (PSE) meats are characterised by the fast 

decline of pH after slaughtering with a simultaneous 

high muscle temperature, producing severe protein 

denaturation (Offer et al., 1989) and, consequently, high 

exudation, which produces a high light scattering and 

reflection over moist surface, giving the pale colour to 

this kind of meats. With regard to the myoglobin forms 

in this quality class, the pH decline promotes the 

inactivation of oxygen-consuming enzymes, which 

causes the oxygenation of myoglobin yielding to more 

oxymyoglobin formation (Govindarajan, 1973). These 

low-quality classes contrast to RFN (red, firm and nonexudative) 

meats, which are the desirable class to 

consumers. These meats have a slower pH decline than 

PSE meats, which affects positively to their texture and 

colour. The macromolecular structure is less dense than 

doi:10.1111/j.1365-2621.2010.02325.x 


DFD meats, increasing the light scattering in a desirable 

way and the oxymyoglobin is the predominant pigment 

(Offer et al., 1989). 

Recently, lots of studies have been made with regard 

to colour and colour stability of pork. Some parameters 

have been studied such as the pre-slaughter stress 

(Rosenvold & Andersen, 2003), diet (Hoving-Bolink 

et al., 1998; Rosenvold & Andersen, 2003; Tikk et al., 

2008), blooming time (Brewer et al., 2001; Lindahl 

et al., 2006a; Skrlep & Candek-Potokar, 2007). However, 

the influence of ageing on colour development of 

different quality classes has not been widely evaluated. 

The aim of this paper has been to study the entire 

visible spectra of different meat classes and the evolution 

of these spectra during meat ageing to evaluate meat 

colour development. The relative amounts of the three 

myoglobin species have been also studied to compare 

the ability of these quality classes to bloom during 

ageing. Moreover, the potential use of visible spectroscopy 

for discriminating low meat quality classes during 

the 24 h after slaughter has been analysed. 


Raw material 

Twenty-six pigs (Landrace · Duroc, 16 males and 10 

females) were slaughtered in a commercial slaughterhouse 

located near the Institute of Agrochemistry and 

Food technology (Valencia, Spain). These carcasses 

were selected by trained employee from a total of 200 

carcasses, trying to obtain loins from the different 

quality classes. Carcasses were rapidly chilled at )10 °C 

for 45 min and the longissimus Dorsi muscle removed. 

The muscle was kept at 4 °C until sampled. 

Technological meat quality classification 

The pork loins were classified based on L24h* coordinate, 

pH2h, pH24h and drip loss, according to the 

classification of Table 1. Samples pH was measured 

through a portable pH-meter CRISON PH25 Ò 

(CRISON Instruments, SA, Barcelona, Spain) at 2, 4, 

6, 8, 24, 48 h and 7 days postmortem. 

A thick slice (about 100 g) was excised 24 h after 

slaughter, suspended within a net in a plastic bag and 

stored at 4 °C. The per cent change in weight over the 

Table 1 Meat classification based upon pH2h, pH24h, drip loss and 

L24h* coordinate 

Meat quality classes pH2h pH24h Drip loss L24h* 

Pale soft exudative 56 

Red firm non exudative >5.8

1712 

Visible Spectroscopy to assess colour development during ageing of fresh pork M. Castro-Giráldez et al. 

Table 2 Meat quality measurements by quality classes for assayed 

pork loins 

Technologic 

parameters 

Pale, soft and 

exudative 

(n =3) 


Dark, firm 

and dry 

(n =3) 

Red firm 

non exudative 

(n = 20) 

Drip loss (%) 6.0 a ± 0.3 1.5 b ± 0.3 3.5 c ± 0.4 

L 24h* 56.7 a ± 0.8 49.4 b ± 0.2 51.8 b ± 1.8 

pH 2 h pm 5.58 a ± 0.13 6.22 b ± 0.09 5.99 c ± 0.05 

pH 24 h pm 5.37 a ± 0.10 5.90 b ± 0.06 5.63 c ± 0.04 

n, number of pork loins; a,b,c , means in a row with different superscripts 

are significantly different (P < 0.05); pm, postmortem. 

The pork loins were classified according to the classification 

of Table 1. Based on this classification, three 

pork loins were classified as PSE, three as DFD and 20 

as RFN (Table 2). 

The pH evolution during postmortem period is shown 

in Fig. 1. The rapid drop of pH suffered by PSE samples 

immediately after slaughter, from a pH near the 

neutrality (in the living animal) to 5.58 at 2 hours 

postmortem (hpm), can be appreciated in Fig. 1. RFN 

samples showed a pH at 2 hpm of 5.99, indicating a 

progressive pH decline from slaughter. DFD samples 

showed the highest pH at 2 hpm, the value of which was 

6.22. These animals are characterised because the 

glycogen reserves are depleted before slaughter, and 

thus, these muscles produce little lactic acid during the 

conversion of muscle to meat. Furthermore, in the figure 

it can be appreciated that the pH decrease continues 

until approximately 8 hpm in the three quality classes, 

reaching a final pH close to 5.4 for PSE samples, 5.63 for 

RFN samples and close to 5.90 for DFD samples. The 

decrease of pH from 2 to 24 hpm is more pronounced in 

RFN and DFD samples than in PSE samples, whose 

highest decrease is experimented during the two-first 

hpm. 

Colour evolution during ageing 

L* co-ordinate with regard to postmortem time is 

represented in Fig. 2a. The least significance difference 

(LSD) intervals (95% confidence) are also shown. L* 

co-ordinate represents lightness, where L* = 0 is completely 

black, and L* = 100 is completely white. L* 

co-ordinate at 24 hpm is one of the parameters usually 

used for meat quality classification. PSE samples 

showed a significantly higher (P < 0.05) L24h* 

co-ordinate with respect to the other two quality classes. 

In the figure, it can also be observed that L* co-ordinate 

of PSE meats remains higher than the L* co-ordinate of 

DFD and RFN during the 48 hpm. In PSE meats, the 

pH decreases rapidly during the first hour after slaughter, 

while temperature is still high, resulting in excessive 

protein denaturation (Offer & Knight, 1988), so the 

water-holding capacity is decreased (Ledward, 1992), 

light scattering is increased (Bendall & Swatland, 1988) 

and L* co-ordinate is increased (Joo et al., 1999). On the 

other hand, DFD meats present an L* co-ordinate 

similar to RFN meats during the 24 h after slaughter as 

it was reported by Zhu & Brewer (1998). Above 24 hpm, 

RFN meats present higher L* co-ordinate than DFD 

meats, which remains almost constant and lower. L* 

co-ordinate increases gradually during ageing for RFN 

meats. 

In Fig. 2, the Hue mean angle, a* and b* co-ordinates 

are represented for each meat quality class during 

ageing. In the CIE L*a*b* space, a* represents the 

red-green colour and b* the yellow-blue colour of the 

sample. DFD and PSE quality classes do not show 

important variations in a* and b* colour co-ordinates 

during ageing. On the contrary, RFN quality class 

shows variations in both a* and b* co-ordinates. The 

value of a* co-ordinate increases significantly 

(P < 0.05) from 48 h to 7 days postmortem; this fact 

means that RFN samples showed more red colour after 

7 days of ageing than during the 2 days after slaughter. 

The b* co-ordinate shows an increase from 24 h to 

Figure 1 Evolution of pH means during 

ageing, where Dark, firm and dry samples 

(h), red, firm and non-exudative samples 

(¤), pale, soft and exudative samples ( ). 

The standard deviation of each pH is shown. 


Figure 2 Evolution of (a) L* co-ordinate, 

(b) Hue mean angles, (c) a* co-ordinate, 

(c) b*co-ordinate, during ageing where red, 

firm and non-exudative ( ), Dark, firm and 

dry ( ), and pale, soft and exudative ( ). In 

all figures, the Least Significance Difference 

intervals (with 95% confidence) are 

represented. 

7 days postmortem. Both co-ordinates, a* and b*, were 

used to calculate the hue mean angle during ageing to 

obtain more information about the effect of ageing in 

meat colour. In the Fig. 2b, it can be observed that RFN 

and DFD meats have similar hue mean angles during 

ageing (70.39° and 68.29°, respectively). On the other 

hand, PSE samples showed a hue mean angle (80.38°) 

significantly higher (P < 0.05) than RFN and DFD 

meats. Hue angle measures the degree of departure from 

the red axis of the CIE colour space (Brewer et al., 

2006). Lower hue angle is related with higher visually 

perceived redness. So, RFN and DFD samples, which 

have a significantly lower (P < 0.05) hue mean angle, 

Figure 3 Mean reflectance spectra 

(400–700 nm) and Least Significance 

Difference intervals (with 95% confidence) at 

different postmortem times, where: red, 

firm and non-exudative ( ), Dark, firm and 

dry ( ), and pale, soft and exudative ( ). 

Visible Spectroscopy to assess colour development during ageing of fresh pork M. Castro-Giráldez et al. 1713 

(a) 

62 

59 

56 

53 

50 

47 

(c) 8 

6 

4 

2 

0 

12 24 48 168 

(b) 88 

10 

12 24 48 168 12 24 48 168 

84 

80 

76 

72 

68 

64 

(d) 16 

15 

14 

13 

12 

11 

12 24 48 168 

are characterised by redder colour than PSE meats. 

Moreover, from 48 h to 7 days postmortem, the RFN 

samples showed a significantly (P < 0.05) decrease in 

hue angle, which is related with higher perceived red 

colour in those samples after 7 days of ageing. 

In Fig. 3, the average reflectance spectra between 400 

and 700 nm were represented for the studied meat 

quality classes at 12, 24, 48 h and 7 days postmortem. It 

can be observed that at 12 hpm the PSE meats reflectance 

spectrum is significantly higher (P < 0.05) than 

that of DFD and RFN meats at wavelengths above 

430 nm. It might be because of the accelerated metabolism 

that characterises the exudative meats, which 


1714 


produces a fast pH-decline while temperature is still 

high; this situation produces a decrease in the waterholding 

capacity and an excessive protein denaturation, 

increasing light scattering. The same behaviour can be 

observed at 24 and 48 hpm. It is important to remark 

that RFN meats spectrum is being progressively 

increasing with ageing, approaching to PSE spectrum. 

This is because of the fact that these kind of samples 

have a higher pH at early postmortem when compared 

to the PSE samples; moreover, its pH decline is gradual, 

approaching progressively to isoelectric point of myosin 

with meat ageing, leading to the loss of water-binding 

capacity of the protein and affecting muscle structure; 

owing tothese factors, more incident light is scattered 

and RFN meat reflectance spectrum is approaching to 

PSE spectrum. This phenomenon is reflected in the 

48 hpm spectra (Fig. 3), which show differences statistically 

significant among the three quality classes. 

Otherwise, DFD samples are characterised by a little 

production of lactic acid and, consequently, by a very 

low pH decrease (pH24 near 6.0), higher water-holding 

capacity and dark colour (Lawrie, 1998). The Fig. 3 

shows that at 24 hpm, the reflectance spectra can be 

used to discriminate only the PSE meats, not being 

useful to the discrimination of DFD meats. For this 

discrimination, 48 hpm are needed. 

Evolution of myoglobin species during ageing 

The evolution of myoglobin species of different pork 

classes was investigated (Fig. 4). The relative amounts 

of DMb, MbO2 and MetMb in meat depend on the 

oxygen availability, the autoxidation rate of myoglobin 

and the MetMb reducing capacity (Ledward, 1992). In 

Fig. 4, it can be observed that oxymyoglobin is significantly 

higher (P < 0.05) at 7 days postmortem than 

during the 48 h after slaughter in RFN meats. After 

Figure 4 Effect of ageing on the relative 

concentration of the myoglobin species 

deoxymyoglobin (DMb), oxymyoglobin 

(MbO2) and metmyoglobin (MetMb), 

calculated from Kubelka–Munk K ⁄ S-ratios 

for red, firm and non-exudative ( ), Dark, 

firm and dry ( ), and pale, soft and exudative 

( ) samples. The relative content of DMb was 

estimated by the ratio K ⁄ S 474 ‚ K ⁄ S 525, the 

relative content of MbO2 by the ratio 

K ⁄ S610 ‚ K ⁄ S525 and the relative content of 

MetMb by the ratio K ⁄ S 572 ‚ K ⁄ S 525. The 

mean values and the Least Significance 

Difference intervals (with 95% confidence) 

are represented. 

slaughtering, the muscle continues consuming oxygen in 

cell respiration (Lanari & Cassens, 1991; Madavi & 

Carpenter, 1993). Meat ageing in normal meats promotes 

a decrease in respiration owing to the loss of 

structural integrity and consequently, the oxygen consumption 

decrease, allowing it to penetrate into the 

muscle (Millar et al., 1994) and thus a more pronounced 

blooming occurs (Zhu et al., 2001; Lindahl et al., 

2006a,b), leading to a more red colour (Rosenvold & 

Andersen, 2003; Lindahl et al., 2006a,b). On the other 

hand, deoxymyoglobin decrease significantly (P < 0.05) 

from 48 to 7 days postmortem for RFN meats; this 

phenomenon can be related to the increase in oxymyoglobin 

at 7 days postmortem because of the increased 

oxygenation (blooming ability) at this time. Moreover, 

metmyoglobin suffers a slight increase at 7 days postmortem 

in RFN meats; this slight accumulation of 

MetMb can be because ofthe decrease in metmyoglobin 

reductase activity as time postmortem progresses (Mancini 

& Hunt, 2005). These significant changes in myoglobin 

species of RFN meats were reflected in a 

significant increase of L*, a* and b* co-ordinates, and 

a decrease in the hue angle at 7 days postmortem. Thus, 

more blooming can be associated with lighter and redder 

colour as was previously reported by Lindahl et al. 

(2001). 

On the other hand, low-quality meats do not have a 

significant increase in oxymyoglobin during meat ageing. 

PSE meats have a rapid pH decline, while samples 

are still at high temperatures; this situation favours the 

auto-oxidation of myoglobin (Gotoh & Shikama, 1974) 

to metmyoglobin, which increase significantly 

(P < 0.05) at 7 days postmortem (Fig. 4). DFD meats 

are characterised by high pH (pH near 6), at which 

oxygen-consuming enzymes remain active, so less oxygen 

is available to the oxygenation of myoglobin 

(Fig. 4). On the other hand, this meat class has a very 


Figure 5 Mean reflectance spectra 

(400–700 nm) for each meat quality class 

during ageing, where: 12 hours postmortem 

(hpm) ( ), 24 hpm ( ), 48 hpm ( ), 7 days 

postmortem ( ). The least significance 

difference intervals (with 95% confidence) 

are represented. 

stable colour, and no significant variations on myoglobin 

states are produced. 

When compared the three quality classes in Fig. 4 it 

can be observed that DMb is similar during the 48 hpm 

in the three classes, while there are significant differences 

(P < 0.05) at 7 days postmortem, being higher in 

DFD meats than in PSE meats, which are higher than 

RFN meats. MbO2 is significantly higher (P < 0.05) in 

RFN meats than in the low-quality classes at 7 days 

postmortem, which is related with the significantly 

(P < 0.05) higher values of a* and b* co-ordinates at 

7 days postmortem in RFN meats than in the lowquality 

classes, indicating a brighter red colour of this 

samples at 7 days postmortem. On the other hand, PSE 

meats have less MbO 2 than the other two classes from 

24 h to 7 days postmortem. With regard to the MetMb, 

PSE samples have higher MetMb at 7 days postmortem 

than the other two quality classes. 

In Fig. 5, the evolution of the mean reflectance 

spectra (400–700 nm) for each meat quality during 

ageing are shown. It can be observed that the reflectance 

of RFN samples is increasing with postmortem 

time. The 7 days postmortem spectrum has a different 

shape from the other spectra. It can be observed that 

the bump at around 560 nm is more pronounced at 

7 days postmortem which can be associated with an 

increase in MbO2, which is typically observed at such 

wavelength (Lindahl, 2005). Moreover, a little decrease 

in reflectance is observed at around 630 nm, which can 

be associated to the slight increase in MetMb (Lindahl, 

2005). On the contrary, DFD samples spectra remain 

almost constant during the postmortem, which can be 

interpreted as very stable colour characteristics of this 

Visible Spectroscopy to assess colour development during ageing of fresh pork M. Castro-Giráldez et al. 1715 

λ λ 

λ 

pork meat quality class. There exist only differences at 

7 days postmortem spectrum in which the bump at 

around 560 nm disappears; it can be associated with the 

decrease in MbO2 observed at this time, although this 

decrease was not significant (see Fig. 4). PSE samples 

presented an increase in reflectance during the 48 h after 

slaughter, but the reflectance at 7 days postmortem 

decreases significantly (P < 0.05). The 7 days postmortem 

spectrum presents a different behaviour 

between 600 and 700 nm, which can be associated to 

the significant (P < 0.05) increase of MetMb (Lindahl, 

2005). 

Conclusions 

The evolution of visible spectra was influenced by ageing 

time in PSE and RFN loins, while the visible spectra of 

DFD loins showed no variation with postmortem time. 

A more pronounced blooming was observed in RFN 

meats at 7 days postmortem than during the 48 h after 

slaughter. This phenomenon was reflected in a significantly 

(P < 0.05) higher oxymyoglobin content, L*, a* 

and b* co-ordinates and a significantly (P < 0.05) lower 

deoxymyoglobin content, and hue angle at 7 days 

postmortem with regard to the values of these parameters 

in this quality class during the 48 h after slaughter. 

Thus, it is possible to conclude that ageing time affects 

positively to the colour of RFN meats. PSE samples 

showed a significant (P < 0.05) increase in metmyoglobin 

at 7 days postmortem, and a significant (P < 0.05) 

decrease in L* co-ordinate, while DFD meats did not 

show significant variation on colour parameters or on 

myoglobin states. 


1716 


Additionally, the results showed the possibility of 

separating PSE meats from the other classes by using 

visible reflectance spectra at 24 hpm. However, at this 

time, it was also demonstrated that discriminating DFD 

meats by this method is not possible. 


Grant from Agroalimed (Conselleria de Agricultura, 

Pesca y Alimentacioń, Valencia, Spain) and FPU grant 

to MC from Ministry of Science and Innovation 

(Madrid, Spain) are fully acknowledged. Work prepared 

within the Unidad Asociada IIAD (UPV)-IATA (CSIC) 

framework. 

References 

Ashmore, C.R., Parker, W. & Doerr, L. (1972). Respiration of 

mitochondria isolated from dark-cutting beef: postmortem changes. 

Journal of Animal Science, 34, 46–53. 

Bendall, J.R. & Swatland, H.J. (1988). A review of the relationships of 

pH with physical aspects of pork quality. Meat Science, 24, 85–126. 

Bendall, J.R. & Taylor, A.A. (1972). Consumption of oxygen by the 

muscles of beef animals and related species. II. Consumption of 

oxygen by post-rigor muscle. Journal of the Science of Food and 

Agricultural, 23, 707–711. 

Brewer, M.S., Zhu, L.G., Bidner, B., Meisinger, D.J. & McKeith, F.K. 

(2001). Measuring pork color: effects of bloom time, muscle, pH and 

relationship to instrumental parameters. Meat Science, 57, 169–176. 

Brewer, M.S., Novakofski, J. & Freise, K. (2006). Instrumental 

evaluation of pH effects on ability of pork chops to bloom. Meat 

Science, 72, 596–602. 

CIE. (1978). International Commission on Illumination, Recommendations 

on Uniform Colour Spaces, Colour, Difference Equations, 

Psychometric Colour Terms. CIE publication No. 15 (Suppl.2), (E- 

1.31) 1971 ⁄ (TC-1.3). Paris, France: Bureau Central de la CIE. 

Gotoh, T. & Shikama, K. (1974). Autoxidation of oxymyoglobin from 

bovine heart muscle. Archives Biochemistry and Biophysics, 163, 

476–478. 

Govindarajan, S. (1973). Fresh meat colour. CRC Critical Reviews in 

Food Technology, 4, 117–140. 

Honikel, K.O. (1997). Reference methods supported by OECD and 

their use in Mediterranean meat products. Food Chemistry, 59, 573– 

582. 

Honikel, K.O. (1998). Reference methods for the assessment of 

physical characteristics of meat. Meat Science, 49, 447–457. 

Hoving-Bolink, A.H., Eikelenboom, G., van Diepen, J.Th.M., Jongbloed, 

A.W. & Houben, J.H. (1998). Effect of dietary vitamin E 

supplementation on pork quality. Meat Science, 49, 205–212. 

Hunt, M.C., Acton, J.C., Benedict, R.C., Calkins, C.R., Cornforth, 

D.P. & Jeremiah, L.E., et al. (1991). Guidelines for meat colour 

evaluation. In: Proceedings 44 th Annual Reciprocal Meat Conference, 

June 9–12, Kansas City University, Manhattan, KS. pp. 1–17. 

Chicago, IL, U.S.A.: Publ. National Live Stock and Meat Board. 

Joo, S.T., Kauffman, R.G., Kim, B.C. & Park, G.B. (1999). The 

relationship of sarcoplasmic and myofibrillar protein solubility to 

colour and water-holding capacity in porcine longissimus muscle. 

Meat Science, 52, 291–297. 

Lanari, M.C. & Cassens, R.G. (1991). Mitochondrial activity and beef 

muscle colour stability. Journal of Food Science, 56, 1476–1479. 

Lawrie, R.A. (1998). Lawrie’s Meat Science, 6th edn. Cambridge, UK: 

Woodhead Publishing Limited. 

Ledward, D.A. (1992). Colour of raw and cooked meat. In: The 

Chemistry of Muscle-Based Foods (edited by D. E Johnston, M.K. 

Knight & D.A. Ledward). Pp. 128–144. Cambridge, UK: Royal 

Society of Chemistry. 

Lindahl, G. (2005). Colour characteristics of fresh pork. Doctoral 

Thesis. Uppsala (Sweden): Swedish University of Agricultural 

Sciences, . 

Lindahl, G., Lundström, K. & Tornberg, E. (2001). Contribution of 

pigment content, myoglobin forms and internal reflectance to the 

colour of pork loin and ham from pure breed pigs. Meat Science, 59, 

141–151. 

Lindahl, G., Karlsson, A.H., Lundström, K. & Andersen, 

H.J. (2006a). Significance of storage time on degree of blooming 

and colour stability of pork loin from different crossbreeds. Meat 

Science, 72, 603–612. 

Lindahl, G., Enfält, A.C., Andersen, H.J. & Lundström, K. (2006b). 

Impact of RN genotype and ageing time on colour characteristics of 

the pork muscles longissimus dorsi and semimembranosus. Meat 

Science, 74, 746–755. 

MacDougall, D.B. (1982). Changes in the colour and opacity of meat. 


MacDougall, D.B. & Jones, S.J. (1981). Translucency and colour 

defects of dark-cutting meat and their detection. In: The Problem of 

the Dark-Cutting in Beef (edited by D.E. Hood & P.V. Tarrant). Pp. 

328–339. Hague, Netherlands: Martinus Nijhoff Publishers. 

Madavi, D.L. & Carpenter, C.E. (1993). Ageing and processing affect 

colour, metmyoglobin reductase and oxygen consumption of beef 

muscles. Journal of Food Science, 58, 939–947. 

Mancini, R.A. & Hunt, M.C. (2005). Current research in meat colour. 

Meat Science, 71, 100–121. 

Mancini, R.A., Hunt, M.C. & Kropf, D.H. (2003). Reflectance at 

610nm estimates oxymyoglobin content on the surface of ground 

beef. Meat Science, 64, 157–162. 

Millar, S., Wilson, R., Moss, B.W. & Ledward, D.A. (1994). 

Oxymyoglobin formation in meat and poultry. Meat Science, 36, 

397–406. 

Norman, J.L., Berg, E.P., Heymann, H. & Lorenzen, C.L. (2003). Pork 

loin colour relative to sensory and instrumental tenderness and 

consumer acceptance. Meat Science, 65, 927–933. 

Offer, G. & Knight, P. (1988). The structural basis of water-holding in 

meat. Part 2: drip losses. In: Development in Meat Science (edited by 

R. Lawrie). Vol. 4. Pp. 172–243. London: ElSevier Applied Science. 

Offer, G., Knight, P., Jeacoke, R. et al. (1989). The structural basis of 

waterholding, appearance and toughness of meat and meat products. 

Food Microstructure, 8, 151–170. 

Rosenvold, K. & Andersen, H.J. (2003). The significance of preslaughter 

stress and diet on colour and colour stability. Meat 

Science, 63, 199–209. 

Skrlep, M. & Candek-Potokar, M. (2007). Pork color measurement as 

affected by bloom time and measurement location. Journal of 

Muscle Foods, 18, 78–87. 

Tikk, K., Lindahl, G., Karlsson, A.H. & Andersen, H.J. (2008). The 

significance of diet, slaughter weight and ageing time on pork colour 

and colour stability. Meat Science, 79, 806–816. 

Zhu, L.G. & Brewer, M.S. (1998). Discoloration of fresh pork as 

related to muscle and display conditions. Journal of Food Science, 

63, 763–767. 

Zhu, L.G., Bidner, B. & Brewer, M.S. (2001). Postmortem pH, muscle, 

and refrigerated storage effects on ability of vacuum-packaged pork 

to bloom. Journal of Food Science, 66, 1230–1235. 




The capability of rosemary extract in preventing oxidation of fish 

lipid 

Yesim Ozogul, 1 * Deniz Ayas, 2 Hatice Yazgan, 1 Fatih Ozogul, 1 Esmeray K. Boga 1 & Gulsun Ozyurt 1 

1 Department of Seafood Processing Technology, Faculty of Fisheries, University of Cukurova, Adana, Turkey 

2 Department of Seafood Processing Technology, Faculty of Fisheries, University of Mersin, Mersin, Turkey 


Summary The effects of rosemary extract at different levels (%1, R1, and %2, R2) on the quality of vacuum-packed 

sardine in terms of sensory, biochemical (thiobarbituric acid, total volatile basic nitrogen, peroxide value and 

free fatty acids) and microbiological analyses (total viable counts) were investigated. Fish were filleted and 

divided into three groups. First group was used as the control (C) without rosemary extract, second group 

was treated with 1% rosemary extracts (10 g L )1 ) for 2 min (R1), and the third was treated with 2% 

rosemary extracts (20 g L )1 ) for 2 min (R2). Thirty fillets per litre were used. After that, all groups were 

vacuum-packed in polyethylene bags. The samples were stored in the refrigerator condition (4 ± 1 °C) over 

the storage period of 20 days. The results showed that the use of rosemary extract improved the sensory 

quality of both raw and cooked sardine, most preferably sardine treated with 1% of rosemary. Biochemical 

analysis showed that the use of 2% of rosemary extract were found to be most effective (P < 0.05) in 

controlling the rate of lipid oxidation. 

Keywords Lipid oxidation, quality, rosemary extract, sardine, vacuum package. 

Introduction 

Fish is one of the most highly perishable food products, 

and the shelf life of such products is limited in the 

presence of atmospheric oxygen and the growth of 

aerobic spoilage microorganisms. Fish oil contains longchain 

polyunsaturated fatty acids (PUFA) such as EPA 

(eicosapentaenoic acid, C20:5n3) and DHA (docosahexaenoic 

acid, C22:6n3) that are considered to have a 

number of health benefits. However, the desirable 

PUFA content in fish oil is highly vulnerable to 

oxidative destruction. Lipid oxidation is a series of 

chain reaction with molecular oxygen reacting with 

unsaturated lipids to form lipid peroxides resulting in 

organoleptic changes of flavour, texture and aroma of 

food (Sarkardei & Howell, 2008). Synthetic antioxidants 

such as ethylenediaminetetraacetic acid, butylated 

hydroxyanisole and butylated hydroxytoluene can be 

added to retard fish oil oxidation, but antioxidants from 

natural sources may be used to replace synthetic 

antioxidants. Recently, there is an increasing demand 

for natural antioxidants because of the concern about 

safety of synthetic antioxidants. This would not only 

*Correspondent: Fax: (90) 322 3386439; 

e-mail:yozogul@cu.edu.tr 

doi:10.1111/j.1365-2621.2010.02326.x 


prevent omega-3 fatty acid oxidation, but also enhance 

the health benefits of the foods by having the additional 

health-promoting bioactivity from the herbs or spices. It 

was also reported that rosemary extracts contain a large 

amount of phenolic compounds such as carnosic acid, 

carnosol and rosmarinic acid, which have antioxidant 

potential (Frankel, 1999). Moreover, Tironi et al. (2009) 

found that application of rosemary extract at doses of 

200 and 500 ppm prevents lipid oxidation of chilled sea 

salmon. 

Corbo et al. (2008, 2009b) have tested thymol, lemon 

extract and grape fruit seed extract, at 20, 40 and 

80 ppm, against the main spoilage microorganisms 

inoculated in fish burgers stored at 5 °C. Results 

showed that all compounds have effect in slowing down 

the growth of microorganisms, suggesting that they can 

be advantageously used to prolong the shelf life of fresh 

fish burger. The combined effect of modified atmosphere 

packaging (MAP: 40% CO 2 ⁄ 30% O 2 ⁄ 30% N 2) 

and oregano essential oil on the shelf life of lightly 

salted cultured sea bream fillets stored under refrigeration 

was studied by Goulas & Kontominas (2007). 

They found that oregano essential oil in combination 

with MAP and light salting was the most effective 

treatment for the preservation of sea bream fillets 

followed by MAP.

1718 

Rosemary extract in preventing oxidation Y. Ozogul et al. 

Mahmoud et al. (2006) established a new technology, 

using pretreatment with electrolysed NaCl solutions and 

essential oil compounds, to extend the shelf life of carp 

fillets. Samples of skinless carp fillets were treated with 

100-fold (by weight) of electrolysed NaCl solutions 

[cathodic solution, EW()), and ⁄ or anodic solution, 

EW(+)] and 1% oil (0.5% carvacrol + 0.5% thymol) 

[1%(C + T)]. Results indicated that treatment with 

EW()) ⁄ EW(+) ⁄ 1%(C + T) extended the shelf life of 

carp fillets to 16 and 1.3 days compared with 4 and 

0.3 days for the control samples during storage at 5 and 

25 °C, respectively. However, it was also reported that 

fish preservation, using electrolysed NaCl solutions and 

1% (carvacrol + thymol), did not affect the quality 

(nutritional components) of carp fillets and could be a 

good alternative to synthetic preservatives used in the 

food industry (Mahmoud et al., 2007). 

Sardine is commercially important fish species in 

Turkey as they are caught in large amounts (17.531 tons 

in 2008) (Anon., 2008). Sardine are generally consumed 

as fresh, canned or used as fish meal and oil in Turkey. 

There are many researches on the quality of sardine 

stored in ice (Campos et al., 2005) and under MAP and 

vacuum packed (VP) conditions (O¨ zogul et al., 2004; 

Mendes et al., 2008). The use of rosemary as antioxidant 

in different fish and fish products has also been reported 

(Akhtar et al., 1998; Gimenez et al., 2004, 2005; O¨ zogul 

et al., 2009). However, no information is available on 

the effects of different level of rosemary extract on the 

quality of vacuum-packed sardine stored at 4 °C. The 

principal objectives of this investigation were (i) to 

determine the shelf life of the sardine treated with 

different levels of rosemary extract; (ii) to evaluate some 

of the existing objective tests as indices of quality and 

degree of spoilage of vacuum-packed sardine; and (iii) to 

determine antioxidant effects of rosemary extract on 

vacuum-packed sardine (Sardinella pilchardus) fillets 

during refrigerated storage. 


Rosemary extract and preparation of fish 

Rosemary (Rosmarinus officinalis) used in this project 

was a powder and presented by Frey-Lau Company 

(Henstedt-Ulzburg, Germany). Rosemary extract was 

applied to the fish as described by Akhtar et al. (1998). 

Fish (26.97 ± 1.49 g and 14.13 ± 0.31 cm) were 

caught by seine net in Mersin Bay, Turkey. The duration 

of time between harvesting and arrival of the fish at the 

laboratory was

Analytical techniques 

The total volatile basic nitrogen (TVB-N) content of 

sardine was determined according to the method of 

Antonocopoulus (1973) and expressed as mg TVB-N per 

100 g muscle. The value of TBA was determined 

according to Tarladgis et al. (1960) in fish fillets to 

evaluate the oxidation stability during storage period, 

and the results are expressed as TBA value, milligrams 

of malondialdehyde per kg flesh. Free fatty acid (FFA) 

analysis, expressed as percentage of oleic acid, was 

determined by AOAS (1994). Peroxide value (PV) 

expressed in milliequivalents of peroxide oxygen per 

kg of fat was determined according to AOAS (1994). 

Microbiological analysis 

Triplicate samples were taken to estimate total viable 

counts (TVC) from each of three different groups. Fish 

muscle (10 g) was mixed with 90 mL of Ringer solution 

and then stomached for 3 min. Further decimal dilutions 

were made, and then 0.1 mL of each dilution was 

pipetted onto the surface of plate count agar (Fluka 

70152, Steinheim, Switzerland) plates in triplicate. They 

were then incubated for 2 days at 30 °C. 


A completely randomised designed was used. The data 

were subjected to analysis of variance and Duncan’s 

multiple range tests. A SPSS statistical package (version 

8.0; SPSS Inc., Chicago, IL, USA) was adapted to a 

personal computer. 


Sensory analyses 

Sensory evaluation of vacuum-packed sardine fillets (C, 

R1 and R2) reached the limits of acceptance 13, 17 and 

Table 1 Quality index method for sardine in VP 

Storage 

days 

Control (C) 

X Sx 

0 0.00 ± 0.00 a 

3 0.75 ± 0.2 a 

6 1.17 ± 0.29 b 

10 4.33 ± 0.58 b 

13 12.50 ± 2.52 b 

17 14.00 ± 1.41 b 

20 16.67 ± 0.58 a 

R1 

X Sx 

0.00 ± 0.00 a 

0.63 ± 0.21 a 

0.87 ± 0.16 a 

2.83 ± 0.47 a 

6.00 ± 0.65 a 

11.83 ± 2.32 ab 

16.50 ± 0.55 a 

R2 

X Sx 

0.00 ± 0.00 a 

0.63 ± 0.19 a 

0.83 ± 0.20 a 

2.67 ± 0.51 a 

5.50 ± 0.36 a 

10.50 ± 1.64 a 

16.33 ± 0.58 a 

Different letters in the same row indicate significant differences 

(P < 0.05). 

Maximum demerit point: 17; X Sx, average ± standard deviation; C, 

control; R1, %1 rosemary; R2, 2% rosemary. 

Rosemary extract in preventing oxidation Y. Ozogul et al. 1719 

20 days of storage, respectively (Table 1). End of shelf life 

is usually determined when spoilage-related sensory 

attributes such as trimethylamine (TMA), off-odour and 

favour become strong, caused mainly by microbial origin 

(Huss, 1995). Sardine contains high level of PUFA, which 

is susceptible to autooxidation causing off-odours and 

browning of flesh colour. The off-flavour intensity of the 

treatment groups remained at low levels compared to the 

control group until the end of the storage period (day 20). 

The application of rosemary extract to the vacuumpacked 

sardine fillets stored at 4 °C led to an improvement 

in the appearance and odour of the samples, which 

received higher scores than those of the control 

(P < 0.05) from day 6 onwards (Table 1). Off-odour 

and off-flavour were detected towards the end of storage 

period as a result of the strong rosemary flavour. 

Although the effect of 1% of rosemary extract was lower 

in the samples than those treated with 2% rosemary 

extract, no significant differences (P >0.05) were found 

between samples treated with rosemary extract. The use 

of rosemary extract improved the sensory quality of 

sardine. Similar results were obtained from the other 

studies (Vareltzis et al., 1997; Akhtar et al., 1998; 

Gimenez et al., 2004, 2005) . 

It was reported that the progress of decomposition 

showed off-odour after 9 days for sardine in VP (Özogul 

et al., 2004) and 8 days for herring in VP (O¨ zogul et al., 

2000). In this research, shelf life of sardine (the control) 

without rosemary extract was 13 days. The reason for 

this longer shelf life is that fish were iced immediately 

after harvesting and also the short time between catch 

and storage ( 0.05) until 6 day. After that, significant 

differences (P < 0.05) were found between the control 

and treatment groups. Significant differences were also 

observed (P < 0.05) between R1 and R2 on day 17 and 

20. Off-flavour and off-odour of the control group, R1 

and R2 were detected on 13, 17 and 20 day of storage, 

respectively, as found for raw sardine by QIM. Because 

the flavour and taste of rosemary extract was much 

stronger in R2 than in R1, R1 group was mostly 

preferred by the panellists. 

Chemical assessment 

The proximate composition of the sardine was the 

following: 20.60 ± 0.55% protein, 9.15 ± 0.84% lipid, 


1720 


Table 2 Sensory analyses of cooked sardine in VP 

Storage days Colour X Sx Odour X Sx Taste X Sx Firmness X Sx General acceptance X Sx Groups 

0 9.00 ± 0.00 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

3 8.57 ± 0.54 a 

8.43 ± 0.54 a 

8.71 ± 0.49 a 

6 7.57 ± 0.54 a 

7.57 ± 0.54 a 

7.57 ± 0.54 a 

10 7.00 ± 0.00 a 

7.57 ± 0.54 b 

7.57 ± 0.54 b 

13 5.43 ± 1.13 a 

6.57 ± 0.54 b 

6.71 ± 0.49 b 

17 1.86 ± 0.90 a 

3.29 ± 0.49 b 

7.00 ± 0.00 c 

20 1.57 ± 0.54 a 

2.71 ± 0.49 b 

4.00 ± 0.00 c 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

8.86 ± 0.38 a 

9.00 ± 0.00 a 

7.57 ± 0.54 a 

8.00 ± 0.00 ab 

8.29 ± 0.49 b 

6.57 ± 0.54 a 

7.57 ± 0.54 b 

8.00 ± 0.00 b 

4.14 ± 0.90 a 

6.86 ± 0.38 b 

7.00 ± 0.00 b 

2.29 ± 0.49 a 

3.57 ± 0.54 b 

6.43 ± 0.54 c 

1.29 ± 0.49 a 

2.29 ± 0.49 b 

3.71 ± 0.49 c 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

8.86 ± 0.38 a 

8.86 ± 0.38 a 

9.00 ± 0.00 a 

7.57 ± 0.54 a 

8.00 ± 0.00 a 

8.57 ± 0.54 b 

6.57 ± 0.54 a 

7.86 ± 0.38 b 

8.00 ± 0.00 b 

4.00 ± 0.82 a 

6.57 ± 0.54 b 

6.71 ± 0.49 b 

2.29 ± 0.49 a 

3.57 ± 0.54 b 

6.43 ± 0.54 c 

1.00 ± 0.00 a 

2.71 ± 0.49 b 

3.71 ± 0.49 c 

68.66 ± 0.35% moisture and 1.33 ± 0.23% ash. Variations 

in chemical composition of sardine, mainly in 

lipid and moisture, were reported (O¨ zogul et al., 2004; 

Erkan & Ozden, 2008; Mendes et al., 2008). The 

variation in the chemical composition of fish is related 

to nutrition, living area, fish size, catching season, 

seasonal and sexual variations as well as other environmental 

conditions. Because sardine contains high level 

of lipid, care must be taken to preserve the quality as 

fresh as possible after harvesting. 

TVB-N is a product of bacterial spoilage and endogenous 

enzymes action, and its content is often used as an 

index to assess the keeping quality and shelf life of 

products (EEC, 1995). TVB-N is a general term that 

includes the measurement of TMA, dimethylamine, 

ammonia and other volatile basic nitrogenous compounds 

associated with seafood spoilage (Huss, 1995). 

In this study, TVB-N concentrations of all groups 

are shown in Table 3. TVB-N content of all groups 

increased with storage time. The maximum permissible 

level of TVB-N in fish and fishery products is 

35 mg 100 g )1 (EEC, 1995). At the beginning of storage, 

the initial TVB-N value was 20.59 mg 100 g )1 flesh and 

increased to 34.29 mg TVB-N 100 g )1 at day 13 for the 

control, 33.64 mg TVB-N 100 g )1 at day 17 for R1 and 

35.82 mg TVB-N 100 g )1 , in which all samples in VP 

were rejected by the sensory panellists. The lowest TVB- 

N value (P < 0.05) was obtained from R2 followed by 

R1 and the control during storage period. In this study, 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

8.86 ± 0.38 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

7.57 ± 0.54 a 

7.57 ± 0.54 a 

8.29 ± 0.49 b 

6.57 ± 0.54 a 

7.29 ± 0.49 b 

7.71 ± 0.49 b 

4.14 ± 0.90 a 

6.86 ± 0.38 b 

6.71 ± 0.49 b 

1.57 ± 0.54 a 

3.57 ± 0.54 b 

6.43 ± 0.54 c 

1.00 ± 0.00 a 

2.14 ± 0.90 b 

3.43 ± 0.98 c 

Different letters in the same column for each storage days indicate significant differences (P < 0.05). 

X Sx, average ± standard deviation; C, control; R1, 1% rosemary; R2, 2% rosemary. 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

8.86 ± 0.38 a 

9.00 ± 0.00 a 

9.00 ± 0.00 a 

7.57 ± 0.54 a 

8.00 ± 0.00 b 

8.00 ± 0.00 b 

6.57 ± 0.54 a 

7.86 ± 0.38 b 

8.00 ± 0.00 b 

3.71 ± 0.49 a 

6.86 ± 0.38 b 

7.00 ± 0.00 b 

1.57 ± 0.54 a 

3.57 ± 0.54 b 

6.43 ± 0.54 c 

1.00 ± 0.00 a 

2.00 ± 0.00 b 

3.71 ± 0.49 c 

C 

R1 

R2 

C 

R1 

R2 

C 

R1 

R2 

C 

R1 

R2 

C 

R1 

R2 

C 

R1 

R2 

C 

R1 

R2 

Table 3 Changes in the value of total volatile basic nitrogen in sardine 

during storage period 

Storage days 

0 20.59 ± 1.20 a 

3 22.44 ± 0.44 b 

6 25.10 ± 0.12 a 

10 30.67 ± 0.40 c 

13 34.29 ± 0.69 c 

17 45.79 ± 1.13 b 

20 45.96 ± 1.42 c 

Control R1 R2 

X Sx X Sx X Sx 

20.59 ± 1.20 a 

23.10 ± 0.32 c 

24.74 ± 1.06 a 

28.13 ± 0.41 b 

32.80 ± 0.31 b 

33.64 ± 0.50 a 

40.65 ± 0.43 b 

20.59 ± 1.20 a 

21.43 ± 0.06 a 

24.79 ± 0.26 a 

25.78 ± 0.80 a 

27.15 ± 0.77 a 

33.25 ± 0.37 a 

35.82 ± 0.29 a 


(P < 0.05). 

X Sx, average ± standard deviation; C, control; R1, 1% rosemary; R2, 

2% rosemary. 

when the TVB-N level exceeded the maximum value, 

samples were already refused by the panellists. Therefore, 

TVB-N values correlated well with the results of 

sensory analyses, providing a good index for the 

assessment of sardine in VP. 

Shelf life of oily fish species is limited because of the 

oxidation of lipid. The primary product of lipid 

oxidation is fatty acid hydroperoxide, measured as 

PV. Peroxides are unstable compounds, and they break 

down to aldehydes, ketones and alcohols that are 

volatile products causing off-flavour in products 


Table 4 Changes in the value of peroxide value in sardine during 

storage period 

Storage 

days 

0 4.32 ± 1.10 a 

3 19.56 ± 1.62 b 

6 10.46 ± 0.71 ab 

10 10.24 ± 1.30 b 

13 8.72 ± 1.52 a 

17 13.48 ± 0.92 b 

20 14.28 ± 2.70 b 

Control R1 R2 


4.32 ± 1.10 a 

11.24 ± 0.21 a 

8.06 ± 1.97 a 

9.04 ± 0.80 ab 

8.93 ± 1.22 a 

14.74 ± 1.37 b 

10.73 ± 1.21 a 

4.32 ± 1.10 a 

9.12 ± 1.27 a 

11.31 ± 0.47 b 

7.12 ± 0.90 a 

8.59 ± 0.95 a 

10.77 ± 0.35 a 

9.51 ± 0.72 a 


(P < 0.05). 


2% rosemary. 

(Hamilton et al., 1997). Peroxide and thiobarbituric 

acid (TBA) values are the major chemical indices to 

measure the degree of oxidative rancidity. In this study, 

the PV of oil extracted from the sardine fillets treated 

with and without antioxidants increased up to 3 days 

for the control and R1 and 6 days for R2, after 

which the values fluctuated during storage period 

(Table 4). Although there were no significant differences 

(P > 0.05) among the groups, significant differences 

were observed (P < 0.05) on days 3 and 20 

between samples with antioxidant (R1 and R2) and the 

control without antioxidant. The initial value of 

sardine was found, 4.32 meq kg )1 , which was lower 

than 27.6 meq kg )1 reported for fresh sardine by Cho 

et al. (1989). During storage period, the sardine with 

rosemary extract showed generally low lipid oxidation 

compared to the control without rosemary extract as 

reported for other fish species (Vareltzis et al., 1997; 

Gimenez et al., 2004, 2005; Da Silva Afonso & 

Santana, 2008; Sarkardei & Howell, 2008; Tironi et al., 

2009). 

TBA is second breakdown product of lipid oxidation 

and widely used as an indicator of degree of lipid 

oxidation. The concentration of TBA in freshly caught 

fish is typically between 3 and 5 mg of malondialdehyde 

(MDA) equivalents per kilogram flesh, but levels of 

5–8 mg of MDA equivalents per kilogram of flesh are 

generally regarded as the limit of acceptability for fish 

stored in ice (Nunes et al., 1992). Table 5 also shows 

TBA contents in the different treatments during storage. 

TBA values indicating rancidity development in the all 

fish flesh remained low ( 0.05) in FFA concentrations 

were observed between the treatment groups 

(R1 and R2) throughout storage period. Although the 

release of FFA increased from the initial value of 2.88 

(expressed as percentage of oleic acid) to the final value 

of 7.23 for the control, 5.98 for R1 and 6.13 for R2 at 

the end of storage period, lipid hydrolysis developed at a 

Table 6 Changes in the value of free fatty acids during storage period 

Storage days 


0 2.88 ± 0.96 a 

3 2.79 ± 0.22 a 

6 3.11 ± 0.37 a 

10 3.40 ± 1.10 a 

13 3.39 ± 0.61 a 

17 6.67 ± 0.19 b 

20 7.23 ± 0.43 a 

Control R1 R2 


2.88 ± 0.96 a 

2.65 ± 0.26 a 

2.83 ± 0.51 a 

3.37 ± 0.27 a 

3.97 ± 1.82 a 

5.34 ± 0.29 a 

5.98 ± 1.00 a 

2.88 ± 0.96 a 

3.03 ± 0.32 a 

3.37 ± 0.07 a 

3.60 ± 0.26 a 

3.10 ± 0.22 a 

5.09 ± 0.39 a 

6.13 ± 0.90 a 


(P < 0.05). 


2% rosemary. 


1722 


Log CFU g –1 

12 

10 

8 

6 

4 

2 

0 

0 3 6 10 

Storage days 

13 17 20 

slower rate in the samples treated with rosemary extract, 

regardless of the level of antioxidant. 

Microbiological assessment 

TVC content of sardine in VP 

Control 

R1 

R2 

Figure 1 Total viable count (TVC) of sardine in VP at 4 °C. 

TVC are used as an acceptability index for fish products 

because of the effect of bacteria in spoilage. TVC 

determined in sardine fillets were initially 4.22 log 

CFU g )1 , which was higher than those reported for 

sardine (El Marrakchi et al., 1990; O¨ zogul et al., 2004). 

Comparison with the proposed limits (5–7 log CFU g )1 ) 

for fresh fish (ICMSF, 1986) shows that sardine fillets 

were of good quality. TVC increased with storage time 

for all groups (Fig. 1), and the growth of microorganisms 

exceeded the limit on day 10 for the control, on 

day 13 for R1 and on day 17 for R2. It can be 

concluded that the shelf life of sardine in VP was 

approximately 8–9 days for the control, 11–12 days 

for R1 and 13–14 for R2, indicating that sensory 

analysis did not correlated well with microbiological 

analysis. 

Lemon extract and thymol in combination with MAP 

significantly (P < 0.05) reduced the growth rate of 

bacterial population compared with the control (Del 

Nobile et al., 2009). In this study, changes in TVC 

during storage also showed the existence of a reduced 

growth in the samples with rosemary extract. The result 

obtained from sensory evaluation, after rosemary treatment, 

showed a longer shelf life when compared with 

microbiological results. The spoilage rate of fish depends 

on the species and type, the initial microbial flora, the 

location of the catch, the content of the catch, processing 

methods and method of storage. After being caught 

and handled, fish are exposed to additional contamination 

on board fishing vessels. Unhygienic practices, 

especially unwashed hands, clothes, equipment, decks 

and storage facilities, can contaminate fish. To avoid 

contamination of fish, all equipment must be cleaned 

and sanitized. 

Conclusion 

Based primarily on sensory assessment, vacuum-packed 

sardine fillets (C, R1 and R2) reached the limits of 

acceptance 13, 17 and 20 days of storage, respectively. 

The use of rosemary extract improved the sensory 

quality of both raw and cooked sardine, most preferably 

sardine treated with 1% of rosemary extract. However, 

biochemical analysis showed that the use of 2% of 

rosemary extract in combination with VP was found to 

be most effective (P < 0.05) in controlling the rate of 

lipid oxidation. Rosemary extract could be used in fish 

preservation as a natural preservative agent. 

References 

Akhtar, P., Gray, J.I., Gomaa, E.A. & Boore, A.M. (1998). Effect of 

dietary components and surface application of oleoresin rosemary 

on lipid stability of rainbow trout (Oncorhynchus mykiss) muscle 

during refrigerated and frozen storage. Journal of Food Lipids, 5, 

43–58. 

Anon. (2008). Fisheries Statistics. Ankara: State Institute of Statistics, 

Prime Ministry Republic of Turkey. 

Antonocopoulus, N. (1973). Bestmmung des Flüchhtigen Basensticktoofs. 

In: Fische und Fischerzeugnisse (edited by W. Ludorf 

& V. Meyer). Pp. 224–225. Berlin und Hamburg: Aulage Verlag Paul 

Parey. 

AOAC. (1984). Official Methods of Analysis of the Association of the 

Official Analysis Chemists. Association of Official Analytical Chemists, 

14th edn. Washington, DC: AOAC. 

AOAC. (1990). Official Methods of Analysis of the Association of the 

Official Analysis Chemists. Association of Official Analytical Chemists, 

15th edn. Washington, DC: AOAC. 

AOAS. (1994). Official Methods and Recommended Practices of the 

American Oil Chemists’ Society, the American Oil Chemists’ Society. 

Champaign, IL: AOAS. 

Bligh, E.C. & Dyer, W.J. (1959). A rapid method of total lipid 

extraction and purification. Canadian Journal of Biochemistry and 

Physiology, 37, 913–917. 

Bonilla, A.C., Sveinsdottir, K. & Martinsdottir, E. (2007). Development 

of Quality Index Method (QIM) scheme for fresh cod (Gadus 

morhua) fillets and application in shelf life study. Food Control, 18, 

352–358. 

Campos, C.A., Rodríguez, O., Losada, V., Aubourg, S.P. & Barros- 

Velázquez, J. (2005). Effects of storage in ozonised slurry ice on the 

sensory and microbial quality of sardine (Sardina pilchardus). 


Cho, S., Endo, Y., Fujimoto, K. & Kaneda, T. (1989). Oxidative 

deterioration of lipids in salted and dried sardines during storage at 

5 °C. Nippon Suisan Gakkaishi, 55, 541–544. 

Corbo, M.R., Speranza, B., Filippone, A. et al. (2008). Study on the 

synergic effect of natural compounds on the microbial quality decay 

of packed fish hamburger. International Journal of Food Microbiology, 

127, 261–267. 

Corbo, M.R., Di Giulio, S., Conte, A., Speranza, B., Sinigaglia, M. & 

Del Nobile, M.A. (2009a). Thymol and modified atmosphere 

packaging to control microbiological spoilage in packed fresh cod 

hamburgers. International Journal of Food Science & Technology, 44, 

1553–1560. 

Corbo, M.R., Speranza, B., Filippone, A., Conte, A., Sinigaglia, M. & 

Del Nobile, M.A. (2009b). Natural compounds to preserve fresh fish 

burgers. International Journal of Food Science and Technology, 44, 

2021–2027. 

Da Silva Afonso, M. & Santana, L.S. (2008). Effects of pretreatment 

with rosemary (Rosmarinus officinalis) in the prevention of lipid 


oxidation in salted tilapia fillets. Journal of Food Quality, 31, 586– 

595. 

Del Nobile, M.A., Corbo, M.R., Speranza, B., Sinigaglia, M., Conte, 

A. & Caroprese, M. (2009). Combined effect of MAP and active 

compounds on fresh blue fish burger. International Journal of Food 


EEC. (1995). Decision 95 ⁄ 149 ⁄ EC. Total volatile basic nitrogen 

TVBN limit values for certain categories of fishery products and 

specifying the analysis methods to be used. Official Journal, L 097, 

84–87. 

El Marrakchi, A., Bouchrit, N., Bennour, N., Hamama, A. & Tagafit, 

H. (1990). Sensory, chemical and microbiological assessments of 

Moroccan sardines (Sardina pilchardus) stored in ice. Journal of 

Food Protection, 53, 600–605. 

Erkan, N. & Ozden, O. (2008). Quality assessment of whole and gutted 

sardines (Sardina pilchardus) stored in ice. International Journal of 


Frankel, E.N. (1999). Food antioxidants and phytochemicals: present 

and future perspectives. European Journal of Lipid Science and 

Technology, 101, 450–455. 

Gimenez, B., Roncales, P. & Beltran, J.A. (2004). The effects of natural 

antioxidants and lighting conditions on the quality characteristics of 

gilt-head sea bream fillets (Sparus aurata) packaged in a modified 

atmosphere. Journal of the Science of Food and Agriculture, 84, 

1053–1060. 

Gimenez, B., Roncales, P. & Beltran, J.A. (2005). The effects of natural 

antioxidants and lighting conditions on the quality of salmon (Salmo 

salar) fillets packaged in modified atmosphere. Journal of the Science 

of Food and Agriculture, 85, 1033–1040. 

Goulas, A.E. & Kontominas, M.G. (2007). Combined effect of light 

salting, modified atmosphere packaging and oregano essential oil on 

the shelf-life of sea bream (Sparus aurata): biochemical and sensory 

attributes. Food Chemistry, 100, 287–296. 

Hamilton, R.J., Kalu, C., Prisk, E., Padley, F.B. & Pierce, H. 

(1997). Chemistry of free radicals in lipids. Food Chemistry, 60, 

193–199. 

Huss, H.H. (1995). Quality and Quality Changes in Fresh Fish. Pp. 195. 

Rome: FAO. 

ICMSF. (1986). Microorganisms in Foods. The International Commission 

on Microbiological Specifications for Foods of the International 

Union of Biological Societies. Pp. 181–196. Oxford: Blackwell 

Scientific Publications. 

Mahmoud, B.S.M., Yamazaki, K., Miyashita, K., Shin, I.S., Chong, 

D. & Suzuki, T. (2006). A new technology for fish preservation by 

combined treatment with 1 electrolyzed NaCl solutions and essential 

oil compounds. Food Chemistry, 99, 656–662. 


Mahmoud, B.S.M., Kawai, Y., Yamazaki, K., Miyashita, K. & 

Suzuki, T. (2007). Effect of treatment with electrolyzed NaCl 

solutions and essential oil compounds on the proximate composition, 

amino acid and fatty acid composition of carp fillets. Food 

Chemistry, 101, 1492–1498. 

Mendes, R., Pestana, C. & Gonçalves, A. (2008). The effects of soluble 

gas stabilisation on the quality of packed sardine fillets (Sardina 

pilchardus) stored in air, VP and MAP. International Journal of Food 


Nunes, M.L., Batista, I. & Morao de Campos, R. (1992). Physical, 

chemical and sensory analysis of Sardine (Sardine pilchardus) stored 

in ice. Journal of Science of Food and Agriculture, 59, 37–43. 

Özogul, F., Taylor, K.D.A., Quantick, P. & Özogul, Y. (2000). A rapid 

HPLC determination of ATP related compounds and its application 

to herring stored under modified atmosphere. International Journal 

of Food Science and Technology, 35, 549–554. 

Özogul, F., Polat, A. & Özogul, Y. (2004). The effects of modified 

atmosphere packaging and vacuum packaging on chemical, sensory 

and microbiological changes of sardines (Sardina pilchardus). Food 

Chemistry, 85, 49–57. 

Özogul, Y., Ozyurt, G. & Boga, E.K. (2009). Effects of cooking and 

reheating methods on the fatty acid profile of sea bream treated with 

rosemary extract. Journal of the Science of Food and Agriculture, 89, 

1481–1489. 

Pastoriza, L., Sampedro, G., Herrera, J.J. & Herrera, L.C.M. (1996). 

Effect of modified atmosphere packaging on shelf-life of iced hake 

slices. Journal of the Science of Food and Agriculture, 71, 541–547. 

Paulus, K., Zacharıas, R., Robınson, L. & Geıdel, H. (1979). Kritische 

Betrachtungen Zur ‘‘Bewetenden Pru¨ fung Mit Skale’’ Als Einem 

Wesentlichen Verfahren Der Sensorichen Analyse. LWT, 12, 52–61. 

Sarkardei, S. & Howell, N.K. (2008). Effect of natural antioxidants on 

stored freeze-dried food product formulated using horse mackerel 

(Trachurus trachurus). International Journal of Food Science and 


Tarladgis, B., Watts, B.M. & Yonathan, M. (1960). Distillation 

method for determination of malonaldehyde in rancid food. Journal 

of American Oil Chemistry Society, 37, 44–48. 

Tironi, V., Toma´s, M. & An˜ oń, M. (2009). Lipid and protein changes 

in chilled sea salmon (Pseudopercis semifasciata): effect of previous 

rosemary extract (Rossmarinus officinalis L.) application. International 

Journal of Food Science & Technology, 44, 1254–1262. 

Vareltzis, K., Koufidis, D., Gavriilidou, E., Papavergou, E. & 

Vasiliadour, S. (1997). Effectiveness of a natural Rosemary (Rosmarinus 

officinalis) extract on the stability of filleted and minced fish 

during frozen storage. Zeitschrift fur Lebensmittel –Untersuchung 

und -Forschung A, 205, 93–96. 


1724 


Kinetics of ascorbic acid degradation and colour change in ground 

cashew apples treated at high temperatures (100–180 °C) 

Janice R. Lima, 1 * Nadiarid J. Elizondo 2 & Philippe Bohuon 3 

1 Embrapa Agroindústria Tropical, R. Dra. Sara Mesquita, 2270, Pici, 60511-110 Fortaleza – CE, Brazil 

2 Escuela de Tecnología de Alimentos, Universidad de Costa Rica, Apartado Postal 11501–2060, San Jose´, Costa Rica 

3 Montpellier SupAgro, UMR QualiSud, Cirad, 1101 avenue Agropolis, CS 24501, 34093 Montpellier cedex 5, France 


Summary Ascorbic acid (AA) degradation and colour changes, measured by the lightness index (L*), were determined 

in cashew apples (at low dissolved O 2 concentrations) heated at high temperature (100–180 °C) in a 

hermetically sealed cell. A nonisothermal method was developed to estimate thermal degradation kinetics. 

The results showed that reaction kinetics during heat treatments were well represented by first-order 

reactions. The temperature dependence of the kinetic constants was described by an Arrhenius type equation. 

The activation energy (E a) for AA degradation and lightness index were 94 ± 3 and 98 ± 3 kJ mol )1 , 

respectively. The reaction rate constant at 140 °C for AA degradation (64 · 10 )5 ±3· 10 )5 s )1 ) was twice 

that for the lightness index change (33 · 10 )5 ±2· 10 )5 s )1 ). Results allow generating temperature profiles 

of heat processes that would help preserve the AA of cashew apples as well as control the colour formation 

during high-temperature processes. 

Keywords Anacardium occidentale L., browning, heat treatment, modelling, vitamin C. 

Introduction 

Cashew production represents an important economic 

activity for many tropical countries. Although cashew 

apples contribute to human nutrition by supplying 

ascorbic acid (AA) (averaging 200 mg (100 g) )1 ), they 

are often considered as a by-product, wasted and left to 

rot under trees after the nut harvest (Azoubel & Murr, 

2003; Azoubel et al., 2009; Oliveira et al., 2009). 

The vitamin C content, including ascorbic and dehydroascorbic 

acids, besides being an indicator of nutrient 

value, can be used as a reliable and representative index 

for estimating quality deterioration during processing. 

Through oxidation, AA is transformed into dehydroascorbic 

acid, which is irreversibly hydrolyzed to 2,3diketogulonic 

acid, which has no vitamin C activity. 

Anaerobic destruction of AA yields carbon dioxide and 

furfural and has a slower rate than the aerobic path 

(Pa´tkai et al., 2002). 

Discolouration and browning because of thermal 

treatments of fruits result from various reactions. The 

Maillard condensation between reducing sugars and 

amino acids, caramelisation, AA browning and pigment 

*Correspondent: E-mail: janice@cnpat.embrapa.br 


destruction are some examples (Damasceno et al., 

2008). 

Research has been carried out on the AA degradation 

and colour formation based on isothermal methods at 

temperatures up to 100 °C, mainly using first-order 

kinetics and modelling the temperature dependence by 

the Arrhenius equation (Ibarz et al., 1999, 2000; Manso 

et al., 2001; Chutintrasri & Noomhorm, 2007; Dhuique- 

Mayer et al., 2007). However, certain industrial processes, 

such as frying, use temperatures of up to 180 °C. 

Above 100 °C, isothermal methods cannot be used to 

estimate degradation kinetics as the duration of the 

heating up and cooling down phases are too large to be 

ignored. Garrote et al. (2009) reported the use of 

nonlinear regression to obtain kinetics parameters to 

successfully predict AA degradation in canned peas 

during sterilisation. 

The purpose of this study is, first, to experimentally 

quantify the AA content [C (t) ] and lightness index [L (t) ] 

in cashew apples during heat treatments; second, to 

develop a kinetic model to describe C (t) and L (t) in heat 

situations, using a nonisothermal method; and third, to 

discuss the advantages of using high-temperature shorttime 

(HTST) treatments for cashew apple to reduce the 

heat impact on product quality during different food 

processes. 

doi:10.1111/j.1365-2621.2010.02327.x 



Cashew apple 

The cashew apples were produced from the commercial 

early dwarf clones CCP 76, grown at the Pacajus 

Experimental Station of Embrapa Tropical Agroindustry, 

in Pacajus, Ceara´, Brazil. The cashew apples were 

harvested at the commercial maturity stage, frozen and 

sent to France. Then, the cashew apples were ground in 

a mixer (Thermomix; Vorwerk & Co. KG, Wuppertal, 

Germany) until a homogeneous pulp was obtained, next 

they were packed in plastic vessels of 200 g capacity and 

stored at )20 °C. For the thermal tests, the vessels were 

let to reach ambient temperature (20 °C), and the 

ground cashew apple was degasified by bubbling 

argon for 15 min; the dissolved oxygen concentration 

was

1726 

Cashew apples treated at high temperatures J. R. Lima et al. 

Selection of the models 

The AA content and lightness index, expressed as X (t) , 

decreased during heat treatment and were described in 

terms of irreversible first-order kinetics: 

dX ðtÞ 

dt ¼ kX X ðtÞ 

ð1Þ 

The rate constant, kX (s )1 ), varied with the system’s 

absolute temperature, T (K), according to the Arrhenius 

law, as follows: 

kX ¼ kX ref exp EaX 

R 

1 

T ðtÞ 

1 

Tref 

ð2Þ 

Where kXref, EaX and R were, respectively, the rate 

constant (s )1 ) at the reference temperature (T ref), the 

apparent activation energy (J mol )1 ) for the rate constant 

and the gas constant (8.314 J mol )1 K )1 ). The 

reference temperature was chosen from the middle of the 

temperature range (Tref = 140 °C). 

During heat treatment, nonisothermal stages (heating, 

maintaining, and cooling) were recorded. Each timetemperature 

profile (T (t) , Fig. 1) was fitted with a cubic 

smoothing spline (MATLAB Ò , version 6.5; The Math- 

Works Inc., Natick, MA, USA). 

Figure 1 Time-temperature profiles for model development and 

validation trials. 

Parameter estimation and statistical methods 

The nonisothermal degradation of compound X (t) during 

heat treatment was taken into account in eqn (1) 

with (2). Hence, X (t) value, predicted at time t, noted 

^X ðtÞ , was calculated by time integration of eqn (1) with 

(2), as follows: 

^X ðtÞ =X 0 ¼ expð kX ref bXÞ ð3Þ 

Where ^X ðtÞ ¼ ^C ðtÞ , or ^X ðtÞ ¼ ^L ðtÞ , for AA content 

kinetic or lightness index kinetic, respectively. bX is the 

time–temperature history for X (t) : 

b X ¼ 

Z t 

0 

exp EaX 

R 

1 

T ðtÞ 

1 

Tref 

dt ð4Þ 

The integral of eqn (4) was calculated as the direct 

analytical integral of the cubic smoothing spline function 

of exp EaXð1=T ðtiÞ 1=TrefÞ=R with a regularisation 

parameter of 0.99 (MATLAB Ò , version 6.5). There 

were two parameters (kXref and EaX) to be estimated 

from the collected data. 

As the model was nonlinear for parameters, the 

parameters could not be solved directly, but must have 

been solved by nonlinear regression. The parameters were 

iteratively adjusted to the goodness-of-fit merit function, 

using the minimisation procedure of the Nelder–Mead 

simplex method (Lagarias et al., 1998) using the MAT- 

LAB Ò software (MATLAB Ò , version 6.5). This merit 

function was the mean square error (MSE) between 

experimental (X (t) ) and predicted (^X ðtÞ ) data: 

MSE ¼ 

1 X 

ðn pÞ 

n 

i¼1 

^X ðtiÞ X ðtiÞ 2 

ð5Þ 

where, n is the number of data, and p is the number of 

parameters (here p = 2). 

The minimum of the merit function was searched with 

different initial values to prevent the parameters from 

obtaining a local minimum (van Boekel, 1996). As the 

model was nonlinear for parameters, no explicit analytical 

solutions could be obtained for the confidence 

intervals, resulting only in approximate values (van 

Boekel, 1996). Consequently, the confidence interval for 

each parameter was determined via Monte Carlo simulations 

(Hessler, 1997). There were four steps involved 

in our method. 

Step 1: The generation of a large number of simulated 

data sets from the experimental values. Simulated data 

(~X ðtÞ ) was generated by superposition of a pseudo 

random noise on the experimental data (X (t) ). The noise 

reflected the experimental uncertainty (u X ðtÞ). 

~X ðtÞ ¼ X ðtÞ þ u X ðtÞd ð6Þ 


where, ~X ðtÞ is the simulated data; u X ðtÞ is the uncertainty 

of X (t) , determined experimentally (equivalent to the 

standard deviation for a Gaussian distribution); and d is 

a random number whose elements are normally distributed, 

with a mean of 0 and a variance of 1. All 

uncertainties were amplified as follows: 

UXðtÞ ¼ uuXðtÞ ð7Þ 

where u is the coverage factor (u = 1.96 for 95% 

confidence interval) (ISO, 1999). 

Step 2: The estimation of kinetic parameters (kXi ref 

and Ea Xi) from each simulated data set. For a given 

operating condition, m sets of AA content (C (t) ) and 

lightness index (L (t) ) data were randomly drawn, using 

eqn (6) (m = 2,000). Then, each simulated data set was 

fitted to determine the best-fit values of the kinetic 

parameters (kXref and EaX), and the m values of 

parameters were identified from m separate data sets. 

The kXref and EaX were checked for normal distribution, 

using the Kolmogorov–Smirnov test. 

Step 3: The representation of histograms from the 

tabulated kinetic parameters, to obtain discrete 

approximations of the model parameters confidence 

probability distributions. The results were, therefore, 

expressed as mean values of parameters (kXref and 

EaX) associated with their standard deviation. All 

these uncertainties were amplified by the factor 

k = 1.96, corresponding to a 95% confidence interval 

for the parameters. Relative amplified uncertainty, 

expressed as percentage, was calculated with reference 

to the mean value obtained after averaging the m 

values. 

Step 4: Model validation. The C (t) and L (t) models 

were validated by assessing the prediction accuracy for a 

set of validation trials. The validation trials were 

conducted under conditions that differed from those 

applied to the kinetic parameter determination assays. 

The temperatures tested were 110, 130, 150 and 170 °C 

(Fig. 1). 

The equivalent isothermal time 

For nonlinear inactivation, the traditional ‘F0 value’ in 

heat processes can be replaced by the ‘equivalent 

isothermal time’ at a chosen reference temperature. 

When the isothermal curves follow the first-order kinetic 

model, the ‘equivalent isothermal time’ (tequiv) is calculated 

using (Peleg et al., 2008): 

tequiv ¼ 1 

loge kX ref 

^X ðtÞ 

X 0 

! 

ð8Þ 

Where kXref is the rate constant at the chosen 

reference temperature and ( ^X ðtÞ =X 0 ) is the solution of 

Eqns (3) and (4) for time t. 

Cashew apples treated at high temperatures J. R. Lima et al. 1727 

Table 1 Initial characteristics of cashew apple used for kinetic experiments 

(mean values ±95% confidence interval with n =3) 


Characteristics of cashew apples 

The initial characteristics of cashew apples are presented 

in Table 1. The values of AA obtained in this work were 

slightly smaller than those obtained by Maia et al. 

(2004), who reported 158 mg (100 g) )1 

Ground cashew apple 

Water content [g (100 g) )1 ] 84.5 ± 0.1 

Total soluble solids (°Brix) 13.5 ± 0.2 

Total titratable acidity 

(mg citric acid per 100 g) 

0.78 ± 0.01 

pH 4.9 ± 0.01 

Water activity 0.986 ± 0.001 

Ascorbic acid (mg (100 g) )1 ) 141.9 ± 1.8 

Colour L* (lightness index) 69.1 ± 0.1 

Colour a* (redness index) 1.5 ± 0.1 

Colour b* (yellowness index) 35.8 ± 0.1 

for cashew 

apples from CCP 76. It is important to consider that 

the samples in this study were frozen to ensure transportation 

and storage before analysis, which may 

diminish overall AA content. Nevertheless, the AA 

content was still high, averaging three or four times the 

values for lemon [74 mg (100 g) )1 ], grape fruit 

[23 mg (100 g) )1 ], kiwi [65 mg (100 g) )1 ] and orange 

[83 mg (100 g) )1 ] (Lee & Kader, 2000), which are fruits 

well known for providing good vitamin C supplies. 

Kinetic parameter estimation 

The AA content and lightness index (L*) were measured 

experimentally in cashew apples for different timetemperature 

treatments (Fig. 1). The results for a* and 

b* were not considered in this study, as neither one 

presented a defined pattern. 

The nonisothermal method used to estimate kinetic 

parameters for AA and lightness showed that the reaction 

kinetics during 100–180 °C heat treatments were well 

represented by first-order reactions as low coefficients of 

variation were obtained between experimental and predicted 

values. The results are shown in Table 2. The 

Table 2 Estimated kinetic parameters (Ea and kref) for ascorbic acid 

degradation and lightness index change in cashew apple (mean 

values ± 95% confidence interval). Reference temperature of 140 °C 

Reaction Ea (kJ mol )1 ) kref (·10 )5 s )1 ) RMSE R 2 

CV (%) 

Ascorbic acid 94 ± 3 64 ± 3 0.08 0.94 8.6 

Lightness (L*) 98 ± 3 33 ± 2 0.06 0.92 4.2 

RMSE, root mean square error between experimental and predicted 

data; CV, coefficient of variation. 


1728 


confidence intervals of these kinetic parameters were 

determined via Monte Carlo simulations, using 2000 data 

sets that generate the same number of best-fit values for 

the kinetic parameters (kref and Ea). Figure 2 depicts a 

normal distribution, with an average value and standard 

deviation. 

Ascorbic acid content 

Figure 3 illustrates the good fit obtained between the 

experimental data and predicted curves, using the 

optimised kinetic parameters shown in Table 2. The 

temperature impact on AA degradation was not as high 

as expected. Ascorbic acid content after heat treatment 

at 100 °C and 7200 s (120 min) was only 30% lower 

than the initial value. At 180 °C, after 365 s (6 min) of 

treatment, 23% of the initial AA still remained in the 

product. Low oxygen concentration ( nominal 

initial AA concentration > temperature. Oey et al. 

(2006) studied the pressure and temperature stability of 

AA in model systems, under different AA and oxygen 

concentrations. The study showed that the AA degradation 

was primarily caused by oxidation, as long as 

oxygen was present. When oxygen was consumed, 

anaerobic AA degradation occurred and this anaerobic 

reaction was slower than the aerobic reaction. Lima 

et al. (2007) reported that during the thermal processing 

Figure 2 Frequency distribution for the values of activation energy 

(Ea). Solid lines are normal distributions. 

of clarified cashew apple juice, at 100 °C for 50 min, 

only 55% of AA was lost, confirming that AA in 

cashews did not deplete easily. 

The relatively high AA stability may also be explained 

by a protecting effect of the cashew apple matrix. The 

degradation kinetics of food constituents may be related 

to the molecular mobility and the glass transition 

temperature has been used as the main indicator of this 

mobility. The glass transition involves transition from a 

solid ‘glassy’ to a liquid-like ‘rubbery’ state and is 

related to the presence of polysaccharides (Frias & 

Oliveira, 2001). Hung et al. (2007) studying the decomposition 

of l-ascorbic acid freeze-dried with saccharides 

showed that l-ascorbic acid in powder prepared with 

soluble soybean polysaccharide or gum arabic decomposed 

much faster than in powder prepared with 

maltodextrin alone or associated to a-cyclodextrin, 

trehalose and lactose. 

The result obtained in this study for activation energy 

(Ea) was higher than that reported in literature for some 

fruits (36–71 kJ mol )1 ) (Manso et al., 2001; Dhuique- 

Mayer et al., 2007; Garrote et al., 2009), indicating 

stronger temperature dependence. However, the reaction 

constant at 100 °C (k100 °C = 3.4 · 10 )5 s )1 ) was 

lower than values reported for the same authors 

(0.12 · 10 )5 to 6.3 · 10 )5 s )1 ), possibly because of the 

Figure 3 Degradation kinetics of ascorbic acid in cashew apple during 

thermal treatment. Experimental (o) and estimated (curves) data. 


test’s low O2 content condition. It is not possible to 

compare the behaviour of the AA degradation at higher 

temperatures, as no data have been reported under these 

conditions in the literature. 

Lightness (L*) 

Figure 4 illustrates the good fit obtained between the 

experimental data and predicted curves, using the 

optimised kinetic parameters shown in Table 2. As 

observed for AA, the temperature impact on lightness 

was not as high as expected. The value of L* after 

thermal treatment at 100 °C and 7200 s (120 min) was 

only 15% lower than the initial value, and at 180 °C, 

after 365 s (6 min), this value was 56% lower. As L* isa 

colour measure on the light-dark axis, this decreasing 

value indicates that the samples were turning darker. 

The variation in lightness of the heat-treated samples 

can be taken as a browning measure. 

Chutintrasri & Noomhorm (2007) reported that for 

pineapple puree, lightness, based on activation energy, 

was the most sensitive measure of colour changes at 

temperatures ranging from 95 to 110 °C and recommended 

that parameter as an on-line quality control 

parameter to monitor the processing effects on colour 

change during thermal processing. Damasceno et al. 

Figure 4 Degradation kinetics of lightness in cashew apple during 

thermal treatment. Experimental (o) and estimated (curves) data. 


(2008), working with thermal treatment (88–121 °C) on 

clarified cashew apple juice, reported that at temperatures 

above 100 °C the colour shifted from light yellow to 

a dark brown hue, which became more pronounced with 

increasing treatment temperature. These authors concluded 

that browning in clarified cashew apple juice was 

caused by degradation of AA, as there was an inverse 

correlation between browning and AA concentration. 

The activation energy (Ea) and reaction constant at 

100 °C (k100 °C = 1.5 · 10 )5 s )1 ) obtained were in the 

same range than those reported in the literature for 

some fruits (Ibarz et al., 1999, 2000; Chutintrasri & 

Noomhorm, 2007). 

Model validation 

To assess the accuracy of the predictive models, six trials 

were conducted under different conditions from those 

previously applied for developing the models. Different 

time-temperature profiles (Fig. 1) were used for the 

validation. Figures 5 and 6 show the observed and fitted 

values for AA and lightness index obtained from the 

assays used to develop the models and validation 

experiments. Almost all the validation experiments 

varied by

1730 


Figure 6 Fitted vs. experimental values of lightness (L ⁄ L 0 ) after heat 

treatment. Experimental points for model generation (s) and for 

validation (h). Dashed lines: 10% variation. Dotted lines: 15% 

variation. 

Figure 7 Simulated ascorbic acid degradation curves for different 

applications: deep-fat frying, conventional and high-temperature 

short-time (HTST) sterilisation. 

ascorbic acid-rich foodstuffs, assuming similar kinetic 

parameters (kref and Ea) as those obtained in the 

present study. 

For example, during deep-fat frying, the temperature 

range is highly heterogeneous: a dried peripheral region 

is submitted to high temperatures (close to oil temperature), 

whereas the water-rich core is maintained at 

temperatures close to 100 °C. As a result, the degradation 

rate is higher in the food’s peripheral regions than 

at the centre. Nevertheless, for frying more than 60% of 

the initial AA content still remained on the surface of 

the product, and more than 95% in the core. Other 

examples are HTST sterilisation treatments, which cause 

almost no AA destruction, and conventional sterilisation, 

with 90% of the initial ascorbic content at the end 

of the process. 

Conclusions 

The method described in this study will be useful in 

determining the change in colour and AA content of 

cashew apples during nonisothermal process, which can 

be a valuable tool to the various industries engaged in 

fruit processing. Preservation of cashew apples by heat 

treatments, such as frying for producing fruit chips or 

HTST for producing concentrated or clarified juice, is a 

good alternative for enhancing the general public’s food 

quality intake. 


We are grateful to the EU funding under the PAVUC 

Project (contract N° FP6-0015279) and to Coordenaça˜ o 

de Aperfeiçoamento de Pessoal de Nı´vel Superior 

(CAPES), for their financial support. 

References 

AOAC (1997). Official Methods of Analysis, 16th edn, 3rd revision. 

Gaithersburg, MD: Association of Official Analytical Chemists. 

Azoubel, P.M. & Murr, F.E.X. (2003). Optimisation of osmotic 

dehydration of cashew apple (Anacardium Occidentale L.) in 

sugar solutions. Food Science and Technology International, 9, 

427–433. 

Azoubel, P.M., El-Aouar, A.A., Tonon, R.V. et al. (2009). Effect of 

osmotic dehydration on the drying kinetics and quality of cashew 

apple. International Journal of Food Science and Technology, 44, 

980–986. 

van Boekel, M.A.J.S. (1996). Statistical aspects of kinetic modelling 

for food science problems. Journal of Food Science, 61, 477– 

485. 

Chutintrasri, B. & Noomhorm, A. (2007). Colour degradation kinetics 

of pineapple puree during thermal processing. LWT-Food Science 


Damasceno, L.R., Fernandes, A.N., Magalhães, M.M.A. & Brito, E.S. 

(2008). Non-enzymatic browning in clarified cashew apple juice 

during thermal treatment: kinetics and process control. Food 

Chemistry, 106, 172–179. 

Dhuique-Mayer, C., Tbatou, M., Carail, M., Caris-Veyrat, C., 

Dornier, M. & Amiot, M.J. (2007). Thermal degradation of 

antioxidant micronutrients in citrus juice: kinetics and newly formed 

compounds. Journal of Agricultural and Food Chemistry, 55, 4209– 

4216. 


Frias, J.M. & Oliveira, J.C. (2001). Kinetic models of ascorbic acid 

thermal degradation during hot air drying of maltodextrin solutions. 


Garrote, R.L., Silva, E.R. & Roa, R.D. (2009). Diffusion and thermal 

degradation of ascorbic acid during canned fresh green peas 

sterilisation. International Journal of Food Science and Technology, 

44, 1990–1996. 

Hessler, J.P. (1997). The use of Monte Carlo simulations to evaluate 

kinetic data and analytic approximations. International Journal of 

Chemical Kinetics, 29, 803–817. 

Hung, L., Horagai, Y., Kimura, Y. & Adachi, S. (2007). Decomposition 

and discoloration of l-ascorbic acid freeze-dried with saccharides. 

Innovative Food Science & Emerging Technologies, 8, 500–506. 

Ibarz, A., Pagán, J. & Garza, S. (1999). Kinetic models for colour 

changes in pear puree during heating at relatively high temperatures. 


Ibarz, A., Pagán, J. & Garza, S. (2000). Kinetic models of nonenzymatic 

browning in apple puree. Journal of the Science of Food 

and Agriculture, 80, 1162–1168. 

ISO, NF ENC 13005 (1999). Détermination de l’incertitude e´largie. In: 

Guide pour l’expression de l’incertitude de mesure. Pp. 39–134. Paris: 

AFNOR. 

Jiménez, N., Bohuon, P., Lima, J., Dornier, M., Vaillant, F. & Pérez, 

A.M. (2010). Kinetics of anthocyanin degradation and browning in 

reconstituted blackberry juice treated at high temperatures (100– 

180 °C). Journal of Agricultural and Food Chemistry, 58, 2314–2322. 

Lagarias, J.C., Reeds, J., Wright, A.M.H. & Wright, P.E. (1998). 

Convergence properties of the Nelder–Mead simplex method in low 

dimensions. SIAM Journal on Optimization, 9, 112–147. 


Lee, S.K. & Kader, A.A. (2000). Preharvest and postharvest factors 

influencing vitamin C content of horticultural crops. Postharvest 

Biology and Technology, 20, 207–220. 

Lima, E.S., Silva, E.G., Moita Neto, J.M. & Moita, G.C. (2007). 

Redução de vitamina C em suco de caju industrializado e cajuína. 

Quı´mica Nova, 30, 1143–1146. 

Maia, G.A., Sousa-Filho, M.S.M., Figueiredo, R.W. & Brasil, I.M. 

(2004). Caracterização química de peduńculos de diferentes clones 

de cajueiro anão precoce (Anacardium Occidentale L.). Revista 

Cieˆncia Agronoˆmica, 35, 272–278. 

Manso, M.C., Oliveira, F.A.R., Oliveira, J.C. & Frias, J.M. (2001). 

Modelling ascorbic acid thermal degradation and browning in 

orange juice under aerobic conditions. International Journal of Food 


Oey, I., Verlinde, P., Hendrickx, M. & Van Loey, A. (2006). 

Temperature and pressure stability of l-ascorbic acid and ⁄ or [6s] 

5-methyltetrahydrofolic acid: a kinetic study. European Food 

Research Technology, 223, 71–77. 

Oliveira, M.A., Maia, G.A., Figueiredo, R.W., Souza, A.C.R., Brito, 

E.S. & Azeredo, H.M.C. (2009). Addition of cashew tree gum to 

maltodextrin-based carriers for spray drying of cashew apple juice. 

International Journal of Food Science and Technology, 44, 641–645. 

Pátkai, G., Kormendy, I. & Kormendy-Domján, A. (2002). Vitamin C 

decomposition kinetics in solutions, modelling citrus juices. Acta 

Alimentaria, 31, 125–147. 

Peleg, M., Normand, M.D. & Corradini, M.G. (2008). Interactive 

software for estimating the efficacy of non-isothermal heat preservation 

processes. International Journal of Food Microbiology, 126, 

250–257. 


1732 


Optimisation of processing variables affecting the osmotic 

dehydration of pomegranate arils 

Manoj Mundada, 1 Bahadur Singh 1 * & Swati Maske 2 

1 Department of Food Engineering and Technology, Sant Longowal Institute of Engineering and Technology, Longowal 148 106 (Sangrur), India 

2 College of Food Technology, Marathwada Agricultural University, Parbhani 431 402, India 


Summary For the optimisation of osmotic dehydration by response surface methodology, the experiments were 

conducted according to Central Composite Rotatable Design (CCRD) with three variables at five levels. The 

low and high levels of the variables were 40 and 50 °C for osmotic solution temperature, 45 and 55°Bx for 

sucrose solution concentration, 60 and 100 min for duration of dipping in osmotic solution, respectively. The 

fruit to solution ratio was kept 1:4 (w ⁄ w) during all the experiments. Before dipping the arils in sucrose 

solution, the freezing of the whole pomegranate at )18 °C was carried out to increase the permeability of the 

outer cellular layer of the arils. The arils were further convectively dehydrated at 60 °C air temperature up to 

final moisture content of 10% (wb) to make it a shelf stable product. The optimum conditions for osmotic 

solution concentration, temperature and process duration were 55°Bx, 40 °C and 100 min, respectively. 

Keywords Anthocyanins, optimisation, osmotic dehydration, pomegranate, response surface methodology, sucrose. 

Introduction 

The pomegranate (Punica granatum L.) has been 

recently acclaimed for its health benefits, in particular 

for its disease-fighting antioxidant potential viz. ellagic 

acid, gallic acid, and punicalagin which promote health 

by destroying cell damaging free radicals. These free 

radicals can lead to oxidative stress and the accelerated 

aging of cells, thus leading to premature cell death 

(Rosenblat & Aviram, 2006; Negi et al., 2003). Pomegranate 

arils also contain other potent antioxidants like 

anthocyanins and tannins. The edible part (aril) is the 

pulp surrounding the seeds. The arils (juice sac) are filled 

with juicy red, pink, or whitish (depending on variety) 

pulp. The arils contains considerable amount of acids, 

sugars, vitamins, polysaccharides, polyphenols, and 

important minerals (Vardin & Fenerciog˘ lu, 2003). 

In India, this valuable fruit is available only during 

the months – September to January. The fruit is 

consumed directly as fresh arils as well as juice. The 

arils are used as a garnish for desserts and salads (Al- 

Maiman & Ahmad, 2002) and can also be utilised as an 

anardana. An anardana is a product prepared by drying 

arils (Kingsly et al., 2006). Among different methods of 

drying, the convective dehydration is the most popular 

*Correspondent: Fax: 01672 280057; 

e-mail:bshathan@yahoo.com 


and efficient way to reduce the moisture content and 

preserve foods (Madamba & Lopez, 2002). The product 

quality depending notably on its texture, colour and 

flavour is deteriorated by convective dehydration 

(Lenart, 1996). A well-known process to achieve goodquality 

product is freeze drying, but it is an expensive 

method of food preservation. Therefore, there is a need 

for a techno-economic alternate process, which has low 

capital investment and offers a way to save highly 

perishable products and make them available for the 

regions away from production zones. Osmotic dehydration 

is one of these methods (Shi & Le Maguer, 2002). 

The osmotic dehydration is the process to remove 

some extent of water from cellular material such as 

fruits and vegetables prior to drying (Lenart, 1996) This 

gives rise to two major simultaneous counter-current 

mass transfer fluxes, namely water flow from the 

product to the surrounding solution and solute infusion 

into the product (Lewicki & Porzecka-Pawlak, 2005). 

There is a third flow of natural solutes such as sugars, 

organic acids, minerals, and salts leaching from the food 

into the solution (Waliszewski et al., 2002), which is 

quantitatively negligible, but may be important for the 

organoleptic and nutritional value of the product 

(Sablani et al., 2002). This pretreatment can also minimise 

drying colour losses (Nsonzi & Ramaswamy, 

1998), as well as reduce nutrient losses (Shi et al., 1999). 

The cellular membrane of arils exerts high resistance to 

doi:10.1111/j.1365-2621.2010.02328.x 


mass transfer and slows down the rate of osmotic 

dehydration (Erle & Schubert, 2001). Thus, it is imperative 

to utilise simple and inexpensive method to 

increase the rate and efficiency of the process (Falade 

& Adelakun, 2006). Freezing has been reported to 

enhance mass transfers during osmotic dehydration of 

Pumpkin, apple, and African Star Apple (Saurel et al., 

1994). Moreover, freezing of whole pomegranate fruit 

could be utilised to establish the long-term preservation 

of pomegranate arils prior to osmotic dehydration. 

The aim of this work was to optimise osmotic 

dehydration process of pomegranate arils as a function 

of sucrose concentration, osmotic temperature, and 

time, using response surface methodology with the 

purpose to achieve maximum possible water loss, solute 

gain, and retention of anthocyanins. 


Experimental design 

For the optimisation of osmotic dehydration process, 

the experiments were conducted according to Central 

Composite Rotatable Design (CCRD) (Khuri & Cornell, 

1987) with three variables at five levels each. The 

independent variables were process temperature, solute 

concentration, and duration of osmotic dehydration 

process. The low and high levels of the independent 

variables were 40 and 50 °C for temperature, 45 and 

55°Bx for sucrose solution concentration, 60 and 

100 min for duration of osmotic pretreatment, respectively 

(Ade-Omowaye et al., 2002). The fruit to solution 

ratio was kept 1:4 (w ⁄ w) during all the experiments to 

minimise problems related to the management of the 

osmotic solutions like reconcentration, microbial contamination, 

reutilisation, and discharge of the spent 

solution (Torreggiani & Bertolo, 2002). 

The relationship between levels of different coded and 

uncoded form of independent variables is given in 

Table 1. The experiments plan in coded and uncoded 

form of process variables along with results is as given in 

Table 2. The experiments were conducted randomly to 

minimise the effects of unexplained variability in the 

observed responses because of external factors. 

Table 1 The level of different process variables in coded and 

uncoded form for osmotic dehydration 

Independent 

variables Units 

Symbols Levels 

Actual Coded )1.68 )1 0 1 1.68 

Concentration °Bx X 1 x 1 41.59 45 50 55 58.41 

Temperature °C X2 x2 36.59 40 45 50 53.41 

Time min X 3 x 3 46.36 60 80 100 113.64 

Optimisation of osmotic dehydration by response surface methodology M. Mundada et al. 1733 

Preparation of sample 

For each experiment, fresh well-graded, whole pomegranates 

var. Kandhari were procured from local market 

of Sangrur, Punjab (India) and were washed and frozen 

at )18 °C for minimum time of 24 h. Before osmotic 

dehydration, frozen pomegranates were thawed at room 

temperature. Thawed pomegranates were immediately 

broken to separate arils. The freezing treatment before 

osmotic dehydration was given to arils to enhance mass 

transfer rate by increasing permeability of outer cellular 

layer (Falade & Igbeka, 2007). 

Osmotic dehydration 

For each experiment, the pomegranate arils of known 

weight were put in stainless steel containers having 

osmotic solution. The temperature of the osmotic 

solution was maintained by hot water bath agitating 

at the rate of 50 oscillations per min. Agitation was 

given during osmosis for reducing the mass transfer 

resistance at the surface of the fruit and for good 

mixing and close temperature control in osmotic 

medium (Chopra, 2001). The arils were removed from 

the container at the specified time and rinsed with 

fresh water to remove the excess solute adhered to the 

surface. The osmotically dehydrated arils were then 

spread on an absorbent paper to remove the free 

water present on the outer surface. Then out of the 

total osmotically dehydrated arils, about 15–20 g 

sample was put in the preweighed Petri dish for 

determination of dry matter by oven method. The 

remaining part of the sample was dried to final 

moisture content of 10% (wb) in hot air dryer at 

60 °C air temperature. The dried samples were packed 

in high density polyethylene (HDPE) (80 micron) bags 

and kept at ambient temperature for further analysis 

of anthocyanins content. 

Statistical analysis and optimisation 

The second-order polynomial equation was fitted to the 

experimental data of each dependent variable as given 

below 

Yk ¼ bk0 þ Xn 

bkixi þ 

i¼1 

Xn 

bkiix i¼1 

2 Xn 1 X 

i þ 

i¼1 

n 

bkijxixj ð1Þ 

j¼iþ1 

where Yk = response variable; Y1 = water loss (g) per 

100 g fresh arils); Y2 = solute gain (g) per 100 g fresh 

arils; Y3 = total anthocyanin content (mg) per 100 g 

of dehydrated arils; xi represent the coded independent 

variables (x1 = solution concentration, x2 = process 

duration, x3 = process temperature); where bko was 

the value of the fitted response at the centre point of 

the design, i.e. point (0,0,0), bki, bkii, and bkij were the 


1734 

Optimisation of osmotic dehydration by response surface methodology M. Mundada et al. 

Table 2 Central Composite Rotatable Design with experimental values of response variables 

Conc 

(°Bx) 

Time 

(min) 

Temp 

(°C) 

Conc 

(°Bx) 

Time 

(min) 

Temp 

(°C) 

linear, quadratic, and cross-product regression coefficients, 

respectively. 

The analysis of the experimental data was carried out 

to observe the significant effect of various process 

variables on the various responses. The b coefficient is 

that the magnitudes of these values help to compare the 

relative contribution of each independent variable in the 

prediction of the dependent variable. Higher the positive 

value of b of a parameter, higher would be the effect of 

that parameter and vice versa. 

The response surface and contour plots were generated 

for different interaction of any two independent 

variables, while holding the value of third variable as 

constant (at the central value). Such three-dimensional 

surfaces could give accurate geometrical representation 

and provide useful information about the behaviour of 

the system within the experimental design. The optimisation 

of the osmotic dehydration process was aimed at 

finding the levels of independent variables viz. osmotic 

solution concentration, temperature, and process duration, 

which would give maximum possible water loss, 

solute gain, and total anthocyanins. When the dehydrated 

product has to be rehydrated before final use like 

dehydrated vegetables, the optimisation of osmotic 

dehydration process is always aimed at minimum solute 

gain. But, in present case, the dehydrated product has to 

be utilised directly without rehydration. Therefore, the 

optimisation was aimed at the maximum solute gain 

during osmotic dehydration process. It will also help to 

make the product shelf stable at ambient conditions. 

Water loss 

(g per 100 g 

of fresh arils) 

Solute gain 

(g per 100 g 

of fresh arils) 

)1.00 )1.00 )1.00 45.00 40.00 60.00 16.19 1.76 68.32 

1.00 )1.00 )1.00 55.00 40.00 60.00 22.23 2.53 54.43 

)1.00 1.00 )1.00 45.00 50.00 60.00 22.46 1.91 63.05 

1.00 1.00 )1.00 55.00 50.00 60.00 28.80 2.83 47.26 

)1.00 )1.00 1.00 45.00 40.00 100.00 27.87 3.21 58.97 

1.00 )1.00 1.00 55.00 40.00 100.00 33.48 4.52 52.40 

)1.00 1.00 1.00 45.00 50.00 100.00 36.95 4.54 49.15 

1.00 1.00 1.00 55.00 50.00 100.00 42.80 5.52 46.27 

)1.68 0.00 0.00 41.59 45.00 80.00 22.73 2.14 62.48 

1.68 0.00 0.00 58.41 45.00 80.00 35.48 3.94 44.59 

0.00 )1.68 0.00 50.00 36.59 80.00 19.62 3.75 62.86 

0.00 1.68 0.00 50.00 53.41 80.00 31.33 4.81 46.78 

0.00 0.00 )1.68 50.00 45.00 46.36 19.25 1.55 66.03 

0.00 0.00 1.68 50.00 45.00 113.64 39.80 5.09 48.17 

0.00 0.00 0.00 50.00 45.00 80.00 23.52 2.61 54.20 

0.00 0.00 0.00 50.00 45.00 80.00 24.89 2.68 53.57 

0.00 0.00 0.00 50.00 45.00 80.00 23.97 2.76 53.25 

0.00 0.00 0.00 50.00 45.00 80.00 24.15 2.79 52.87 

0.00 0.00 0.00 50.00 45.00 80.00 24.64 2.85 53.69 

0.00 0.00 0.00 50.00 45.00 80.00 24.57 2.88 52.78 

Total anthocyanin 

content (mg per 100 g 

of dehydrated arils) 

Further, increase in sucrose concentration in product 

has also been reported to minimise anthocyanin losses 

during convective dehydration (Wrolstad et al., 1990). 

Response surface methodology was applied to the 

experimental data using a commercial statistical package, 

Design-Expert version 6.01 (Trial version; Statease 

Inc., Minneapolis, MN, USA). The same software was 

used for the generation of response surface plots, 

superimposition of contour plots, and optimisation of 

process variables. 

Mathematical calculations 

Water loss and solute gain during osmotic dehydration 

The water loss and solute gain during osmotic dehydration 

were calculated by the equations given by Ozen 

et al., 2002; Singh et al., 2007; 

Water loss ðgÞ per 100 g of fresh arils 

¼ ðWo WtÞþðSt 

Wo 

SoÞ 

100 

Solute gain ðgÞ per 100 g of fresh arils ¼ ðSt SoÞ 

Wo 

ð2Þ 

100 

ð3Þ 

Where Wo is the initial weight of arils (g), Wt is the 

weight of arils after osmotic dehydration for any time t 

(min), So is the initial weight of solids (dry matter) in the 


arils (g), and S t is the weight of solids (dry matter) of 

arils after osmotic dehydration for time t (min). 

Estimation of dry matter and moisture content 

The samples were oven-dried at 103 ± 2 °C with lids 

open until a constant weight loss (AOAC, 1965). 

Total anthocyanin content 

Total anthocyanin content was determined by procedure 

given by Ranganna, 1986, which involves extraction of 

anthocyanins with ethanolic HCl: 95% ethanol–1.5 N 

HCl (85:15) and measurement of colour at the wavelength 

of maximum absorption (535 nm). The total 

anthocyanin was calculated by use of molecular extinction 

coefficient values. 

The total anthocyanin content was measured by 

following formulae: 

Total OD per 100g of driedsample 

OD volume up of the extract total volume 

¼ 

ml of the extract used wt of sample taken 

100 

Total anthocyanin content in mg per 100g dried sample 

total OD per 100 g of sample 

¼ 

Extinction coefficient 

Where the extinction coefficient value is the average of 

anthocyanins in alcoholic media at 535 nm for 1% 

solution (10 mg mL )1 ) is equal to 982. Therefore, 

absorbance of solution containing 1 mg mL )1 is equal 

to 98.2. 


Diagnostic checking of fitted models and response surfaces 

The results of second-order response surface model in 

the form of analysis of variance (anova) are given in 

Tables 3, 4, and 5. The results indicated that the fitted 

quadratic models accounted for more than 90% of the 

variation in the experimental data, which were highly 

significant (R 2 > 0.90). 

The magnitude of P values from Table 3 revealed that 

all linear and quadratic terms of process variables have 

significant effect at 5% level of significance (P < 0.05) 

on water loss during osmotic dehydration. Further, 

interaction of ‘temperature and time’ has significant 

effect on water loss. The model F-value is 198.56, which 

implies the model is significant. The relative magnitude 

of b values (Table 3) indicates the maximum positive 

effect of process duration (b = 6.2954) followed by 

osmotic solution temperature (b = 3.7292) and concentration 

(b = 3.3159) on water loss. The quadratic and 

interaction terms of all the process parameters have least 

effect on water loss as compared to the linear terms of 


Table 3 anova table showing the variables as a linear, quadratic, and 

interaction terms on water loss and coefficients for the prediction 

models 

Source df b Sum of squares F-value P-level 

Model 9 24.2707 108.9694 198.5582

1736 


Table 5 anova table showing the variables as a linear, quadratic, 

and interaction terms on total anthocyanin content variable and 

coefficient for the prediction model 

Source df b Sum of squares F-value P-level 

Model 9 53.4026 885.5648 39.6288

Solute gain(g)/100g of fresh arils 

(a) (b) 

4.62643 

3.83165 

3.03686 

2.24207 

1.44728 

100.00 

55.00 

90.00 

52.50 

80.00 

50.00 

70.00 

47.50 

60.00 45.00 

Time (min) 

Optimization of osmotic dehydration process 

Concentration (°B) 

A graphical multiresponse optimization technique was 

adopted to determine the workable optimum conditions 

for the osmotic dehydration of pomegranate arils. The 

contour plots for all responses were superimposed, and 

regions that best satisfy all the constraints were selected 

as optimum conditions. The main criterion for constraints 

optimization was maximum possible water loss, 

solute gain and total anthocyanin content. These constraints 

resulted in ‘feasible zone’ of the optimum 

solutions (shaded area in the superimposed contour 

plots). Superimposed contour plots having common 

superimposed area for all responses for osmotic dehydration 

are as shown in Fig. S1a, b. The points in the 

range 50–52.5°Bx of osmotic solution concentration, 

Solute gain (g)/100g of fresh arils 

5.04924 

4.24531 

3.44138 

2.63745 

1.83352 

100.00 

50.00 

90.00 

47.50 

80.00 

45.00 

70.00 

42.50 

60.00 40.00 

Time (min) 


Figure 2 Influence of process variables on solute gain (a) Sucrose concentration and time at 45 °C of osmotic solution temperature (b) 

temperature and time at 50°Bx of osmotic solution concentration. 

Total anthocyanin content 

(mg)/100gm of dehydrated arils 

(a) (b) 

63.19 

58.63 

54.06 

49.50 

44.94 

55.00 

52.50 

Concentration (°B) 

50.00 

47.50 

45.00 50.00 

47.50 


45.00 

42.50 


40.00 

Total anthocynain content 

(mg)/100g of dehydrated arils 

62.8423 

58.7514 

54.6606 

50.5698 

46.4789 

100.00 

50.00 

90.00 

47.50 

80.00 

45.00 

70.00 

42.50 

60.00 40.00 

Time (min) 


Figure 3 Influence of process variables on total anthocyanins (a) sucrose concentration and temperature for 80 min of process duration (b) 

temperature and time at 50°Bx of osmotic solution concentration. 

40–42.5 °C osmotic solution temperature and 98–100 

min process duration were found to be optimum for 

osmotic dehydration. 

To optimise the process conditions for osmotic dehydration 

process by numerical optimization technique, 

equal importance of ‘3’ was given to all the three process 

parameters (viz. osmotic solution concentration, process 

duration, and solution temperature). However, based on 

their relative contribution to quality of final product, the 

importance given to different responses was 4, 3, and 4 for 

water loss, solute gain, and anthocyanin content, respectively. 

The optimum operating conditions for concentration, 

temperature, and process duration were 55°Bx, 

40 °C, and 100 min, respectively. The optimum processing 

conditions were experimentally verified and proven to 

be adequately reproducible with ±0.1% deviation. 


1738 


Conclusion 

Response surface methodology was effective in optimising 

process parameters for the osmotic dehydration 

of pomegranate arils in osmotic aqueous solutions 

of sucrose having concentrations in the range 45– 

55°Bx, temperature 40–50 °C, and process duration 

60–100 min. The regression equations obtained in this 

study can be used for optimum conditions for desired 

responses within the range of conditions applied in this 

study. Graphical techniques, in connection with response 

surface methodology (RSM), aided in locating 

optimum operating conditions, which were experimentally 

verified and proven to be adequately reproducible. 

Optimum solution by numerical optimization obtained 

was 55°Bx osmotic solution concentration, 40 °C 

osmotic solution temperature, and 100 min of process 

duration to get maximum possible water loss, solute 

gain, and total anthocyanin retention. 

References 

Ade-Omowaye, B.I.O., Rastogi, N.K., Angersbach, A. & Knorr, D. 

(2002). Osmotic dehydration behavior of red paprika (Capsicum 

annuum L.). Journal of Food Science, 67, 1790–1796. 

Al-Maiman, S.A. & Ahmad, D. (2002). Changes in physical and 

chemical properties during pomegranate (Punica granatum L.) fruit 

maturation. Food Chemistry, 76, 437–441. 

AOAC. (1965). Official Methods of Analysis, 16th edn. Washington, 

DC: Association of Official Agricultural Chemists. 

Azoubel, P.M. & Murr, F.E.X. (2004). Mass transfer kinetics of 

osmotic dehydration of cherry tomato. Journal of Food Engineering, 

61, 291–295. 

Chopra, C.S. (2001). Osmo-vacuum drying of carrots: effect of salted 

syrup on drying behavior and product quality. Beverage & Food 

World, September, 15–17. 

Erle, U. & Schubert, H. (2001). Combined osmotic and microwave– 

vacuum dehydration of apples and strawberries. Journal of Food 


Falade, K.O. & Adelakun, T.O. (2006). Effect of pre-freezing and 

solutes on mass transfer during osmotic dehydration and colour of 

oven-dried african star apple during storage. International Journal of 


Falade, K.O. & Igbeka, J.C. (2007). Osmotic dehydration of tropical 

fruits and vegetables. Food Reviews International, 23, 373–405. 

Khuri, A.I. & Cornell, J.A. (1987). Response Surfaces Design and 

Analysis. New York: Marcel Dekker. 

Kingsly, A.R., Singh, B.D., Manikantan, M.R. & Jain, R.K. (2006). 

Moisture dependant physical properties of dried pomegranate seeds 

(Anardana). Journal of Food Engineering, 75, 492–496. 

Lenart, A. (1996). Osmotic-convective drying of fruits and vegetables: 

technology and application. Drying Technology, 18, 951– 

966. 

Lewicki, P.P. & Porzecka-Pawlak, R. (2005). Effect of osmotic 

dewatering on apple tissue structure. Journal of Food Engineering, 

66, 43–50. 

Madamba, P.S. & Lopez, R.I. (2002). Optimization of the osmotic 

dehydration of mango (Mangifera indica l.) slices. Drying Technology, 

20, 1227–1242. 

Negi, P.S., Jayaprakasha, G.K. & Jena, B.S. (2003). Antioxidant and 

antimutagenic activities of pomegranate peel extracts. Food Chemistry, 

80, 393–397. 

Nsonzi, F. & Ramaswamy, S. (1998). Quality evaluation of osmoconvective 

dried blueberries. Drying Technology, 16, 705–723. 

Ozen, B.F., Dock, L.L., Ozdemir, M. & Floros, J.D. (2002). Processing 

factors affecting the osmotic dehydration of diced green peppers. 

International Journal of Food Science and Technology, 37, 497–502. 

Panades, G., Fito, P., Aguiar, Y., Nunez de Villavicencio, M. & 

Acosta, V. (2006). Osmotic dehydration of guava: influence of 

operating parameters on process kinetics. Journal of Food Engineering, 

72, 383–389. 

Panagiotou, N.M., Karathanos, V.T. & Maroulis, Z.B. (1999). Effect 

of osmotic agent on osmotic dehydration of fruits. Drying Technology, 

17, 175–189. 

Ranganna, S. (1986). Handbook of Analysis and Quality Control for 

Fruits and Vegetable Products. New Delhi: Tata McGraw-Hill. 

Rosenblat, M. & Aviram, M. (2006). Antioxidative properties of 

pomegranate: in vitro studies. In: Pomegranates Ancient Roots to 

Modern Medicine (edited by N.P. Seeram, R.N. Schulman & D. 

Heber). Pp. 31–43. Boca Raton, FL: CRC Press. 

Sablani, S.S., Rahman, M.S. & Al-Sadeiri, D.S. (2002). Equilibrium 

distribution data for osmotic drying of apple cubes in sugar-water 

solution. Journal of Food Engineering, 52, 193–199. 

Saurel, R., Raoult-Wack, A.L., Rios, G. & Guilbert, S. (1994). Mass 

transfer phenomena during osmotic dehydration of apple I: fresh 

plant tissue. International Journal of Food Science and Technology, 

29, 531–542. 

Shi, J. & Le Maguer, M. (2002). Osmotic dehydration of foods: mass 

transfer modeling aspects. Food Reviews International, 18, 305–335. 

Shi, J., Le Maguer, M., Kakuda, Y., Liptay, A. & Niekamp, F. (1999). 

Lycopene degradation and isomerization in tomato dehydration. 

Food Research International, 32, 15–21. 

Singh, B., Kumar, A. & Gupta, A.K. (2007). Study of mass transfer 

kinetics and effective diffusivity during osmotic dehydration of 

carrot cubes. Journal of Food Engineering, 79, 471–480. 

Tonon, R.V., Brabet, C. & Hubinger, M.D. (2008). Influence of 

process conditions on the physicochemical properties of acai 

(Euterpe oleraceae Mart.) powder produced by spray drying. Journal 

of Food Engineering, 88, 411–418. 

Torreggiani, D. & Bertolo, G. (2002). The role of an osmotic step: 

combined processes to improve quality and control functional 

properties functional properties in fruits and vegetables. In: Engineering 

and Food for 21st Century (edited by W. Chanes, V. Gustavo 

& J.M. Aguilera). Pp. 1–19. Boca Raton: CRC Press. 

Vardin, H. & Fenerciog˘ lu, H. (2003). Study on the development of 

pomegranate juice processing technology: clarification of pomegranate 

juice. Nahrung, 47, 300–303. 

Waliszewski, K.N., Pardio, V.T. & Ramirez, M. (2002). Effect of 

EDTA on colour during osmotic dehydration of banana slices. 

Drying Technology, 20, 1291–1298. 

Wang, S.Y. (2007). Fruits with high antioxidant activity with 

functional foods. In: Functional Food Ingredients and Nutraceuticals: 

Processing Technologies (edited by J. Shi). Pp. 371–413. New York: 

Taylor and Francis group. 

Wrolstad, R.E., Wightman, J.D. & Durst, R.W. (1990). Glycosidase 

activitiy of enzyme preparations used in fruit juice processing. Food 

Technology, 48, 92. 



online version of this article 

Figure S1. (a) Overlaid contour plots of (a) osmotic 

solution concentration and temperature (b) time and 

osmotic solution temperature. 









Response surface optimisation for the extraction of phenolics and 

flavonoids from a pink guava puree industrial by-product 

Kin-Weng Kong, 1 Abdul Razak Ismail, 1 Seok-Tyug Tan, 1 Krishna Murthy Nagendra Prasad 1 & Amin Ismail 1,2 * 

1 Department of Nutrition and Dietetics, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, 

Malaysia 

2 Laboratory of Analysis and Authentication, Halal Products Research Institute, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia 


Summary Pink guava puree industry produces huge amount of by-products that have potential as sources for 

polyphenols. Response surface methodology was implemented to optimise the extraction conditions for 

phenolics (Y1) and flavonoids (Y2) from a by-product of the guava industry. A three-factor inscribed central 

composite design was employed to determine the effects of three independent variables, namely pH (X 1: 2–6), 

temperature (X 2: 40–60 °C) and time (X 3: 1–5 h), on the response variables. The corresponding predicted 

values for phenolics and flavonoids were 336.30 and 427.35 mg 100 g )1 , respectively. Predicted values for 

extraction rates of phenolics agreed well with experiment values; R 2 of 0.902. However, the model derived for 

flavonoids extraction was less reliable; R 2 of 0.983. Increase in time and temperature was found significant in 

increasing the extraction rate. The optimum conditions for extracting phenolics by ethanolic solvent 

occurred at a pH of 2 and 60 °C for a 5-h extraction. 

Keywords Extraction, flavonoids, phenolics, Psidium guajava by-products, response surface methodology. 

Introduction 

The food industry generates large amounts of byproducts 

from fruits and vegetables containing high 

amounts of polyphenolic antioxidants (Peschel et al., 

2006). The use of these by-products in the food industry 

has become widespread and is attractive because of their 

lower cost and availability in large quantities (Schieber 

et al., 2001). In coping with increased productivity, the 

transformation of polyphenol-rich by-products into 

functional ingredients is an effective and efficient way 

of managing these by-products. 

The polyphenols from these by-products must first be 

extracted with suitable solvents to obtain their useful 

effects. Besides, the extraction conditions must be 

optimised because the solvent concentration, extraction 

time, temperature, pressure, solid to liquid ratio and pH 

of the extraction solvent significantly influence the 

extraction efficiency (Prasad et al., 2009). The conventional 

one-factor approach for performing optimisation 

is very time consuming. In this method, the interactions 

among various factors are not taken into consideration, 

and hence, the chances of obtaining the optimum 


e-mail: amin@medic.upm.edu.my 

doi:10.1111/j.1365-2621.2010.02335.x 


conditions are slim (Liyana-Pathirana & Shahidi, 

2005). Response surface methodology (RSM) enables 

the outcome of many variables, as well as the interactions 

between them, to be evaluated. The use of RSM 

has been successful in optimising the extraction of 

phenolic compounds from food products such as apple 

pomace (Wijngaard & Brunton, 2010), wheat (Liyana- 

Pathirana & Shahidi, 2005), peanut skins (Ballard et al., 

2009) and grape seeds (Karvela et al., 2009). 

Pink guava puree production is accompanied by an 

important quantity of by-product residues accounted 

for 25% of the fruits’ total loading weight. Three types 

of by-products are derived during the crushing, refining 

and sieving stages namely guava refiner (12%), guava 

siever (8%) and guava decanter (5%), respectively 

(Kong & Ismail, 2010). The first fraction, guava refiner 

is the green-brownish fraction that mainly consists of 

peels and seeds. However, guava siever and guava 

decanter are mostly consisting the pulps with slightly red 

in colour. Previously, these by-products had been 

explored as a potential source for antioxidants (Amin 

& Mukhrizah, 2006). Our previous work also reported 

that guava decanter and guava refiner by-products can 

potentially serve as the functional sources for lycopene 

and polyphenolic antioxidants (Kong et al., 2010a,b). 

However, there is little data concerning the effects of

1740 

Phenolic extraction from guava by-product K.-W. Kong et al. 

different extraction parameters on the polyphenolic 

components from the refined pink guava by-product 

(guava refiner). Thus, the current study was carried out 

to extract these compounds using the RSM with a 

central composite design (CCD) to optimise the extraction 

conditions and obtain maximum phenolic and 

flavonoid compounds from the guava refiner. 


Chemicals 

Ethanol and sodium hydroxide (NaOH) were purchased 

from Fisher Scientific (Leicestershine, UK). The Folin- 

Ciocalteu reagent and citric acid were obtained from 

BDH Ltd. (Poole, England). Butylated hydroxytoluene 

(BHT), sodium bicarbonate (NaHCO3), sodium nitrate 

(NaNO2), aluminium chloride (AlCl3), gallic acid and 

rutin were purchased from Sigma Chemical Co. 

(St. Louis, MO, USA). 

Sample 

Refined pink guava (Psidium guavajava var. Sungkai 

Beaumont) puree industry by-product (guava refiner, 

particle size 1.2 mm) was conveniently sampled from 

Golden Hope Food and Beverages Sdn. Bhd., Perak, 

Malaysia. About 20 kg of sample was collected in frozen 

formed of 2.5 kg each and transported using the 

iceboxes to Laboratory of Nutrition, Universiti Putra 

Malaysia. Then, it was directly stored in )20 °C for 

5 weeks until further used. 

Extraction procedure 

Sample (1 g) was placed into a 15-mL falcon tube, and 

10 mL of acidified 80% aqueous ethanol containing 

citric acid (0.1%) was added. The pH was then 

adjusted to the desired level using 1 m citric acid or 

NaOH. The extraction was conducted using an orbital 

shaker (Heidolph Unimax 1010, Schwabach, Germany) 

at 200 rpm, using the time and temperature 

dictated by the experimental design. The optimisation 

procedure was designed based on a three-factor 

inscribed CCD consisting of extraction pH (2–6), 

temperature (40–60 °C) and time (1–5 h) using five 

levels of each variable (Table 1). The extract was 

then centrifuged (Universal 32R; Hettich Zentrifugen, 

Tuttlingen, Germany) at 900 · g at 4 °C, and the 

supernatant was collected to determine the phenolic 

and flavonoid contents. 

Determination of phenolic content 

The phenolic content was determined according to the 

method of Singleton & Rossi (1965). The sample extract 

Table 1 Levels of independent variables for the extraction process 

based on central composite design 

Independent variables 

Coded variable level 

1 0 )1 axial ()a) axial (+a) 

X 1: Acetic acid (pH) 5.22 4.00 2.78 2.00 6.00 

X2: Temperature (°C) 56.12 50.00 43.88 40.00 60.00 

X 3: Time (h) 4.22 3.00 1.78 1.00 5.00 

(0.2 mL) was mixed with 1.5 mL of Folin-Ciocalteu 

reagent and was allowed to stand at room temperature 

for 5 min. Then, 1.5 mL of sodium bicarbonate solution 

(0.566 m) was added to the mixture, and the absorbance 

was read at 725 nm using a UV-VIS spectrophotometer 

(UV-1601; Shimadzu Corporation, Victoria, Australia) 

after a 90-min reaction time. The standard calibration 

curve of gallic acid (0.02–0.10 mg mL )1 ) was plotted to 

calculate the results. Results were expressed in terms of 

fresh weight basis as mg gallic acid equivalents (GAE) 

per 100 g of the guava refiner. 

Determination of flavonoid content 

The flavonoid content was measured spectrophotometrically 

using the method of Zhishen et al. (1999). The 

sample extract (1 mL) was mixed with 4 mL deionised 

water. After 5 min, 0.3 mL of 5% NaNO2 was added to 

the mixture and was subsequently mixed with 0.3 mL of 

10% AlCl3. After 6 min, 2 mL of 1 m NaOH was 

added, and the mixture was diluted to a volume of 

10 mL using deionised water. A reagent blank was 

conducted by replacing the sample extract with deionised 

water. The absorbance of the mixture was 

measured at 510 nm by subtracting the absorbance of 

the blank. A standard calibration curve of rutin 

(0.02–0.10 mg mL )1 ) was plotted to calculate the 

results. The flavonoid content was expressed in terms 

of fresh weight as the mg rutin equivalent (RE) per 

100 g of guava refiner. 

Experimental design 

A three-factor inscribed CCD was developed to optimise 

the extraction process for the phenolics and flavonoids 

of the guava refiner. The experiment was also implemented 

to determine the effects of the three independent 

variables namely, pH (X1: 2–6), temperature (X2: 40– 

60 °C) and time (X3: 1–5 h), on the phenolic content 

(Y1) and the flavonoid contents (Y2). Twenty randomised 

experiments were assigned based on the secondorder 

CCD, and the centre points were repeated six 

times. The experimental design obtained from the CCD 

is presented in Table 2. 


Table 2 Experimental design for the extraction process from central 

composite design 

Extraction Blocks pH (X1) Temperature (°C) (X2) Time (h) (X3) 

1 1 5.22 43.88 4.22 

2 a 

1 4.00 50.00 3.00 

3 1 2.78 56.12 4.22 

4 1 5.22 56.12 1.78 

5 1 2.78 43.88 1.78 

6 a 

1 4.00 50.00 3.00 

7 3 4.00 40.00 3.00 

8 3 2.00 50.00 3.00 

9 a 

3 4.00 50.00 3.00 

10 3 4.00 50.00 1.00 

11 a 

3 4.00 50.00 3.00 

12 3 6.00 50.00 3.00 

13 3 4.00 50.00 5.00 

14 3 4.00 60.00 3.00 

15 2 5.22 43.88 1.78 

16 2 2.78 56.12 1.78 

17 2 5.22 56.12 4.22 

18 a 

2 4.00 50.00 3.00 

19 a 

2 4.00 50.00 3.00 

20 2 2.78 43.88 4.22 

a Centre points. 


The Minitab 14 statistical package (Minitab Inc., State 

College, PA, USA) was used to conduct the experimental 

design and the statistical analysis. Results for the 

contents of phenolics and flavonoids were expressed as 

means ± standard deviations. A response surface analysis 

and analysis of variance (anova) was employed to 

determine the regression coefficients, statistical significance 

of the model terms and to fit the mathematical 

models of the experimental data that aimed to optimise 

the overall region for both response variables. A secondorder 

polynomial model was applied to predict the 

response variables as given below: 

Y ¼ b0 þ b1X1 þ b2X2 þ b3X3 þ b 2 1X21 þ b22 X22 þ b 2 3X23 þ b1b2X1X2 

ð1Þ 

þ b1b3X1X3 þ b2b3X2X3 

where Y is the predicted dependent variable; b0 is a 

constant that fixed the response at the central point of 

the experiment; b1, b2 and b3 are the regression coefficients 

for the linear effect terms; b1 2 , b2 2 and b3 2 are the 

quadratic effect terms; and b1b2, b1b3 and b2b3 are the 

interaction effect terms, respectively. 

The adequacy of the model was predicted through the 

regression analysis (R 2 ) and the anova analysis 

(P < 0.05). The significance of the regression coefficients 

was analysed through a t-test, and non-significant coefficients 

were removed to obtain a reduced model. The 

relationship between the independent variables (X1, X2 

and X3) and the response variables (Y1 and Y2) was 

Phenolic extraction from guava by-product K.-W. Kong et al. 1741 

demonstrated by the response surface plots. The response 

optimiser was implemented for both graphical and 

numerical optimisations to attain the optimum conditions 

and predicted values for the response variables. 

Verification of the model 

Experimental data for the contents of phenolics and 

flavonoids were obtained accordingly to the recommended 

optimum conditions. Verification of the 

response surface model was conducted by comparing 

the experimental value obtained from an independent 

set of samples with the predicted value obtained from 

the optimised model. 


Sample and extraction parameters selection 

Guava refiner was selected because it is the largest 

fraction produced and exhibited the highest total 

phenolic content compared to other by-products of 

pink guava (Kong & Ismail, 2010; Kong et al., 2010a). 

This fraction consisted of seeds and the peels of pink 

guava fruits. Previous studies have shown that seeds and 

peels from fruits are rich in polyphenolic antioxidants 

such as grape seeds (Karvela et al., 2009), orange peels 

(Khan et al., 2009) and litchi peels (Prasad et al., 2009). 

Seeds and peels may have higher polyphenolic components 

than the fruit pulp. According to Jimeńez-Escrig 

et al. (2001), guava peels are twice as high in total 

phenols compared to the guava pulp. 

The efficiency and effectiveness of the polyphenolic 

extraction process are generally manipulated by multiple 

variables such as extraction time, temperature, pH and 

solvent composition (Alothman et al., 2009). The lower, 

middle and upper levels of the three significant extraction 

parameters, namely pH (X1: 2–6), temperature (X2: 40–60 °C) and time (X3: 1–5 h) were subsequently 

selected. The experimental values, as well as the 

predicted values for the response variables, are shown 

in Table 3. The data attained was employed to predict 

an optimal set of extraction parameters for the production 

of guava refiner extract with high phenolic content 

(Y1) and flavonoid content (Y2). 

Fitting the response surface models 

The experimental values of phenolic content (Y1) and 

flavonoid content (Y2) were employed in a multiple 

regression analysis performed using response surface 

analysis to fit the second-order polynomial equations 

(Table 3). In this study, the values obtained experimentally 

for both response variables are near to the 

predicted values (Table 3). Regression coefficients, coefficients 

of determination (R 2 ), adjusted R 2 values, 


1742 


Table 3 Experimental and predicted data obtained for the dependent 

variables 

Extraction 

Phenolic content (Y1) 

(mg GAE per 100 g 

guava refiner) 

Experimental 

value 

probability values (P) and lack of fit values for both 

dependent variables are shown in Table 4. The quality 

of fit to the second-order polynomial models was 

confirmed based on the coefficients of determination 

(R 2 ), which were 0.815 and 0.967 for phenolic and 

flavonoid content, respectively. These results indicated 

that the models can significantly (P < 0.05) explain 

more than 80% (R 2 > 0.80) of the response variation. 

The ‘‘fitness’’ of the model was evaluated through the 

lack of fit test (P > 0.05), which indicated the adequacy 

of models to accurately predict the variation. 

Influence of extraction parameters on phenolic content 

The influence of three independent variables on the 

phenolic content was reported through the significant 

(P < 0.05) coefficient of the second-order polynomial 

regression equation. Solvent pH, extraction temperature 

and time showed significant first-order effects (linear 

term in X1, X2 and X3), second-order effects (quadratic 

term in X2 2 ) and interactive effects (interactive term in 

X1X3) for phenolic content (Y1). The predicted model 

obtained for Y1 was as follows: 

Y1 ¼ 171:795 35:527X1 þ 16:460X2 þ 27:710X3 

þ 27:345X 2 2 

Predicted 

value 

10:394X1X3 

Flavonoid content (Y 2) (mg 

RE per 100 g guava refiner) 

Experimental 

value 

Predicted 

value 

1 151.17 ± 5.01 150.82 294.02 ± 8.57 295.08 

2 a 

156.94 ± 5.21 171.80 269.70 ± 33.10 265.14 

3 265.00 ± 0.97 252.40 310.98 ± 22.69 306.65 

4 157.20 ± 2.08 151.66 251.96 ± 1.63 254.99 

5 179.71 ± 10.47 168.10 213.12 ± 15.21 205.19 

6 a 

165.86 ± 2.63 171.80 269.05 ± 13.63 265.14 

7 178.99 ± 2.45 182.68 233.56 ± 3.46 239.00 

8 185.59 ± 5.00 204.29 238.60 ± 3.65 250.64 

9 a 

171.49 ± 2.73 171.80 263.92 ± 5.19 265.14 

10 133.58 ± 3.27 143.64 230.34 ± 4.22 233.36 

11 a 

178.34 ± 16.03 171.80 266.48 ± 17.57 265.14 

12 136.96 ± 3.52 133.23 233.52 ± 9.10 225.52 

13 194.16 ± 3.07 199.06 299.50 ± 22.63 300.52 

14 204.32 ± 10.81 215.60 344.99 ± 15.05 343.60 

15 152.17 ± 4.55 154.79 184.46 ± 5.28 186.09 

16 190.19 ± 0.13 180.56 337.16 ± 8.02 333.40 

17 177.06 ± 1.74 178.69 289.75 ± 30.59 294.99 

18 a 

183.36 ± 9.86 171.80 265.57 ± 12.29 265.14 

19 a 

179.78 ± 3.37 171.80 257.48 ± 9.25 265.14 

20 213.38 ± 3.24 208.94 253.18 ± 34.66 247.44 

a Centre points. 

ð2Þ 

The response surface plots explained the relationships 

between the phenolic content and the three extraction 

parameters involved (Fig. 1a–c). According to the linear 

and quadratic coefficients in Equation 2, positive coefficients 

for X2 and X3 and negative for X1, indicating 

phenolic content was increased with the increase in 

extraction temperature (40–60 °C) and time (1–5 h), but 

reduced when the solvent pH (2–6) increased. This was 

agreed by Lapornik et al. (2005), who found that 

polyphenols yielded from grape, black and red currant 

by-products from an ethanol extraction solvent, increased 

when the extraction time increased. In addition, higher 

temperatures can improve the phenolic content through 

the increase in phenolic solubility, diffusion rate, mass 

transfer rate, extraction rate and reduced solvent viscosity 

and surface tension (Richter et al., 1996). Moreover, 

increased temperatures may enhance the concurrent 

decomposition of the compounds, including those mobilised 

at lower temperatures (Wettasinghe & Shahidi, 

1999). Thus, it will be helpful in increasing the extraction 

rate, as well as decreasing the extraction time (Cacace & 

Mazza, 2002). Conversely, a decrease in the polyphenol 

content may also occur at high temperatures because of 

the interference of compound stability caused by chemical 

and enzymatic degradation or reactions with other plant 

components that hinder their extraction efficiencies 

(Moure et al., 2001; Kiassos et al., 2009). 

The interaction coefficient showed that increase in 

extraction pH and time will cause the reduction in 

phenolic content. This was in agreement with Mylonaki 

et al. (2008) that a linear decline was found on phenolic 

content while pH value was increasing. In acidic 

condition, higher extractability for phenolic compounds 

was reported (Aliakbarian et al., 2009). A hydrolysis 

mechanism may occur to disintegrate the phenolic 

compounds that bound to their polymers or cell wall 

elements (Chirinos et al., 2007). Although previous 

study stated higher pH value could increase the polarity 

of polyphenols by enhancing the dissociation of the 

mainly acidic phenolic –OH groups and subsequently 

promote the solubility of polyphenols, but indirectly it 

will facilitate the oxidation process (Mylonaki et al., 

2008). Additionally, prolong extraction time may expose 

the phenolic compounds to oxidative degradation 

(Madhujith & Shahidi, 2006). 

Influence of extraction parameters on flavonoid contents 

The regression equation significantly (P < 0.05) explained 

the variation of flavonoids (Y2) in the function 

of different extraction parameters. For flavonoid extraction, 

the pH, temperature and time were found to be 

significant for three linear effects (X1, X2 and X3), two 

quadratic effects (X1 2 and X2 2 ) and three interactive 

effects (X1X2, X1X3 and X2X3). The predicted model 

obtained for Y2 is given below: 


Table 4 Regression coefficients, R 2 values, adjusted R 2 values and probability values for dependent variables 

Regression coefficient 

Phenolic content (Y1) 171.795 a 

Flavonoid content (Y2) 265.143 a 

R 2 

Constant Linear Quadratic Interaction 

b0 b1 b2 b3 b1 2 

)35.527 a 

)12.561 a 

16.460 a 

52.303 a 

27.710 a 

33.578 a 

b2 2 

)3.035 27.345 a 

)27.060 a 

26.155 a 

b3 2 

Y2 ¼ 265:143 12:561X1 þ 52:303X2 þ 33:578X3 

27:060X 2 1 þ 26:155X2 2 

þ 44:494X1X3 46:002X2X3 

b1b2 b1b3 b2b3 

)0.441 )10.394 )29.874 a 

1.798 )39.536 a 

44.494 a 

39:536X1X2 

20.669 

)46.002 a 

R 2 (adj) Regression (P value) Lack of fit 

0.902 0.815 0.001 0.204 

0.983 0.967 0.000 0.068 

a The regression coefficients showed a significant (P < 0.05) relationship. 

(a) 

Phenolic content 

(mg GAE/100 g guava refiner) 

(b) 



(c) 



250 

200 

150 

250 

200 

150 

2 

4 

50 

pH 

6 

40 


40 4.0 

50 

2.5 

60 

1.0 

Temperature (°C) Time (h) 

250 

200 

150 

2 

4 

pH 

6 

1.0 

4.0 

2.5 

Time (h) 

Figure 1 Response surface plot showing the effects of extraction (i) 

pH, (ii) temperature and (iii) time on the phenolic content of guava 

refiner. 

60 

5.5 

5.5 


ð3Þ 

Response surface plots explaining the relationships 

between flavonoid content and the three extraction 

parameters are shown in Fig. 2a–c. According to the 

linear and quadratic coefficients, the extraction temperature 

and time promoted the extraction of flavonoids, but 

solvent pH had an inverse relationship with flavonoid 

extraction (Equation 3). However, a positive interaction 

effect was observed for pH and time, where flavonoid 

content increased when both parameters increased. A 

previous study demonstrated that the extraction of 

flavonoids from mashua tubers was higher in less acidic 

solvents (Chirinos et al., 2007). Although higher yields of 

total flavonoids were found at pH values of 4.5 to 6, the 

flavonoid content was found to be consistent (Karvela 

et al., 2009). These polyphenols are easily oxidised at 

higher pH levels (Mylonaki et al., 2008). 

Interestingly, the interaction between pH and temperature 

and between temperature and time was 

inversely related to the flavonoid content. These results 

clearly indicated that the influences of pH and time on 

the flavonoid yield were dependent on temperature. 

Hence, additional time is not useful in extracting more 

flavonoid compounds because it will implicate the 

degradation of the compounds when a longer extraction 

time and a higher temperature are applied (Silva et al., 

2007). Although increase temperature may enhance the 

extraction but extreme temperatures will cause a reduction 

in the flavonoid content because of flavonoids are 

the thermo labile compounds that should be extracted at 

low temperatures of approximately 60 °C (Silva et al., 

2007). However, the time and pH were the important 

factors for monitoring the stability of these compounds. 

Verification of the models 

To obtain a guava refiner extract with high phenolic and 

flavonoid compounds, the optimal level of extraction 


1744 


(a) 

Flavonoid content 

(mg RE/100 g guava refiner) 

(b) 



(c) 



350 

300 

250 

200 

360 

300 

240 

180 

60 

5.5 

4.0 

2.5 

1.0 

40 

50 

Time (h) Temperature (°C) 

300 

250 

200 

150 

2 50 

4 

6 

40 

pH Temperature (°C) 

2 

pH 

4 

Time (h) 

Figure 2 Response surface plot showing the effects of extraction (i) 

pH, (ii) temperature and (iii) time on the flavonoid content of guava 

refiner. 

parameters was generated based on two single response 

variables and the two response variables together. Three 

optimal conditions were developed for the responses, 

which were as follows: a pH of 2 at 60 °C for 5 h for the 

phenolic content; a pH of 2 at 60 °C for 1 h for the 

flavonoids content; and a pH of 2 at 60 °C for 3.2 h for 

the responses combination. Under these optimum conditions, 

the predicted response values for phenolic and 

flavonoid contents were 336.30 mg GAE per 100 g and 

427.35 mg RE per 100 g, respectively. However, the 

predicted values obtained for the combination of both 

responses, the phenolic content and the flavonoid 

6 

1.0 

2.5 

4.0 

60 

5.5 

content were 265 mg GAE per 100 g and 363.91 mg 

RE per 100 g, respectively. 

The experiments were run in accordance with the 

recommended optimal conditions for two responses and 

the combined responses. The observed values for phenolic 

content and flavonoids were 324.84 ± 33.49 mg 

GAE per 100 g and 316.15 ± 12.77 mg RE per 100 g, 

respectively. The combination of both responses 

indicated that phenolic and flavonoid contents were 

238.27 ± 43.46 mg GAE per 100 g and 186.89 ± 

2.08 mg RE per 100 g, respectively. The response surface 

models were verified by comparing the observed values 

with the predicted values. No significant difference 

(P > 0.05) was found between the experimental and 

predicted values of the phenolic content, but a significant 

difference (P < 0.05) was observed between the experimental 

and predicted flavonoid content. However, in the 

combination of responses, only phenolic content did not 

exhibit a significant difference between the experimental 

and predicted values. Therefore, only the phenolic 

content model was preferred because of an insufficiency 

in the verified model for flavonoid content and the 

combined responses. Although the models generated for 

the flavonoid content and the combined responses were 

inadequate in predicting the expected values, the optimal 

conditions were similar in two parameters at a pH of 2 

and a temperature of 60 °C. 

Conclusion 

The RSM was successfully implemented for the optimisation 

of phenolic compounds from pink guava puree 

by-product. The most efficient set of extraction condition 

was a pH of 2 and a temperature of 60 °C for 5 h to 

achieve the maximum extraction of phenolic content 

from the refined pink guava by-product. In terms of 

flavonoids, a more selective study is needed to study the 

specific extraction conditions for flavonoid components. 

Further studies on other higher technologies to assist 

extraction and increase the recovery of polyphenols 

from this guava refinery by-product are also warranted. 


The authors would like to thank from the Ministry of 

Science, Technology and Innovation of Malaysia (Project 

No. 03-01-04-SF0011) for financial support, the 

Golden Hope Food and Beverages (M) Sdn. Bhd. 

Malaysia for providing the samples, and the Universiti 

Putra Malaysia for providing the laboratory facilities. 

References 

Aliakbarian, B., Dehghani, F. & Perego, P. (2009). The effect of citric 

acid on the phenolic contents of olive oil. Food Chemistry, 116, 617– 

623. 


Alothman, M., Bhat, R. & Karim, A.A. (2009). Antioxidant 

capacity and phenolic content of selected tropical fruits from 

Malaysia, extracted with different solvents. Food Chemistry, 115, 

785–788. 

Amin, I. & Mukhrizah, O. (2006). Antioxidant capacity of methanolic 

and water extracts prepared from food-processing by-products. 

Journal of the Science of Food and Agriculture, 86, 778–784. 

Ballard, T.S., Mallikarjunan, P., Zhou, K. & O’keefe, S.F. (2009). 

Optimizing the extraction of phenolic antioxidants from peanut 

skins using response surface methodology. Journal of Agricultural 


Cacace, J.E. & Mazza, G. (2002). Extraction of anthocyanins and 

other phenolics from black currants with sulfured water. Journal of 


Chirinos, R., Rogez, H., Campos, D., Pedreschi, R. & Larondelle, Y. 

(2007). Optimization of extraction conditions of antioxidant 

phenolic compounds from mashua (Tropaeolum tuberosum Ruíz 

& Pavón) tubers. Separation and Purification Technology, 55, 217– 

225. 

Jiménez-Escrig, A., Rincón, M., Pulido, R. & Saura-Calixto, F. (2001). 

Guava fruit (Psidium guajava L.) as a new source of antioxidant 

dietary fiber. Journal of Agricultural and Food Chemistry, 49, 5489– 

5493. 

Karvela, E., Makris, D.P., Kalogeropoulos, N., Karathanos, V.T. & 

Kefalas, P. (2009). Factorial design optimisation of grape (Vitis 

vinifera) seed polyphenol extraction. European Food Research and 

Technology, 229, 731–742. 

Khan, M.K., Abert-Vian, M., Fabiano-Tixier, A.-S., Dangles, O. & 

Chemat, F. (2009). Ultrasound-assisted extraction of polyphenols 

(flavanone glycosides) from orange (Citrus sinensis L.) peel. Food 

Chemistry, 119, 851–858. 

Kiassos, E., Mylonaki, S., Makris, D.P. & Kefalas, P. (2009). 

Implementation of response surface methodology to optimise 

extraction of onion (Allium cepa) solid waste phenolics. Innovative 

Food Science and Emerging Technologies, 10, 246–252. 

Kong, K.W. & Ismail, A. (2010). Lycopene content and lipophilic 

antioxidant capacity of by-products from Psidium guajava fruits 

produced during puree production industry. Food and Bioproducts 

Processing, doi:10.1016/j.fbp.2010.02.004 (Article in press) 

Kong, K.W., Emmy, H.K.I., Azizah, O., Amin, I. & Tan, C.P. (2010a). 

Effect of steam blanching on lycopene and total phenolics in pink 

guava puree industry by-products. International Food Research 

Journal, 17, 461–468. 

Kong, K.W., Ismail, A., Tan, C.P. & Rajab, N.F. (2010b). Optimization 

of oven drying conditions for lycopene content and lipophilic 

antioxidant capacity in a by-product of the pink guava puree 

industry using response surface methodology. LWT-Food Science 



Lapornik, B., Prosˇ ek, M. & Wondra, A.G. (2005). Comparison of 

extracts prepared from plant by-products using different solvents 

and extraction time. Journal of Food Engineering, 71, 214–222. 

Liyana-Pathirana, C. & Shahidi, F. (2005). Optimization of extraction 

of phenolic compounds from wheat using response surface methodology. 


Madhujith, T. & Shahidi, F. (2006). Optimization of the extraction of 

antioxidative constituents of six barley cultivars and their antioxidant 

properties. Journal of Agricultural and Food Chemistry, 54, 

8048–8057. 

Moure, A., Cruz, J.M., Franco, D. et al. (2001). Natural antioxidants 

from residual sources. Food Chemistry, 72, 145–171. 

Mylonaki, S., Kiassos, E., Makris, D.P. & Kefalas, P. (2008). 

Optimisation of the extraction of olive (Olea europaea) leaf 

phenolics using water ⁄ ethanol-based solvent systems and response 

surface methodology. Analytical and Bioanalytical Chemistry, 392, 

977–985. 

Peschel, W., Sánchez-Rabaneda, F., Diekmann, W. et al. (2006). An 

industrial approach in the search of natural antioxidants from 

vegetable and fruit wastes. Food Chemistry, 97, 137–150. 

Prasad, N.K., Yang, B., Zhao, M., Wang, B., Chen, F. & Jiang, Y. 

(2009). Effects of high pressure treatment on the extraction yield, 

phenolic content and antioxidant activity of litchi (Litchi chinensis 

Sonn.) fruit pericarp. International Journal of Food Science and 


Richter, B.E., Jones, B.A., Ezzell, J.L., Porter, N.L., Avdalovic, N. & 

Pohl, C. (1996). Accelerated solvent extraction: a technique for 

sample preparation. Analytical Chemistry, 68, 1033–1039. 

Schieber, A., Stintzing, F.C. & Carle, R. (2001). By-products of plant 

food processing as a source of functional compounds-recent 

developments. Trends in Food Science and Technology, 12, 401–413. 

Silva, E.M., Rogez, H. & Larondelle, Y. (2007). Optimization of 

extraction of phenolics from Inga edulis leaves using response 

surface methodology. Separation and Purification Technology, 55, 

381–387. 

Singleton, V.L. & Rossi, J.A. (1965). Colorimetry of total phenolics 

with phosphomolybdic-phosphotungstic acid reagents. American 

Journal of Enology and Viticulture, 16, 144–158. 

Wettasinghe, M. & Shahidi, F. (1999). Evening primrose meal: a 

source of natural antioxidants and scavenger of hydrogen peroxide 

and oxygen-derived free radicals. Journal of Agricultural and Food 

Chemistry, 47, 1801–1812. 

Wijngaard, H.H. & Brunton, N. (2010). The optimisation of solidliquid 

extraction of antioxidants from apple pomace by response 

surface methodology. Journal of Food Engineering, 96, 134–140. 

Zhishen, J., Mengcheng, T. & Jianming, W. (1999). The determination 

of flavonoid contents in mulberry and their scavenging effects on 

superoxide radicals. Food Chemistry, 64, 555–559. 


1746 


Effect of tigernut (Cyperus esculentus) flour addition on the quality 

of wheat-based cake 

Chiemela Enyinnaya Chinma, 1* Joseph Oneh Abu 2 & Yusuf Aisha Abubakar 1 

1 Department of Food Science and Nutrition, Federal University of Technology, Minna, Nigeria 

2 Department of Food Science and Technology, University of Agriculture, Makurdi, Nigeria 

(Received 12 March 2010; Accepted in revised form 2 June 2010) 

Summary The effect of substituting tigernut flour for wheat flour on the proximate, mineral and pasting properties of 

the resultant blends and cake quality were studied. The proximate composition of flour blends increased with 

increasing level of tigernut. Protein increased from 22.30 to 26.93% and fat from 4.17 to 7.21% resulting in 

an increase in energy value from 342.09 to 390.93 kcal. The pasting properties of the flour blends were 

affected significantly (P £ 0.05) by tigernut substitution. Pasting peak time and temperature decreased with 

increasing level of tigernut. Mineral elements such as iron and calcium increased from 3.13 to 4.19 and 54.01 

to 56.41 (mg 100 g )1 ) respectively, with tigernut substitution. The weight and volume of cakes increased with 

tigernut level while batter density and volume index decreased. Acceptable cakes can be made with up to 

30% tigernut flour substitution. Such composite cakes may help in reducing protein energy and 

micronutrient deficiencies. 

Keywords Mineral composition pasting, physical and sensory properties, proximate, tigernut flour, wheat flour. 

Introduction 

The current global economic meltdown has partly 

resulted in an upsurge in the number of well-informed 

consumers who put into consideration the health and 

nutritional benefits of food products they consume. This 

has contributed to the stimulation of research into 

alternative food crops with functional or health benefits 

to be incorporated as composite flour in wheat-based 

bakery products. 

Tigernut (Cyperus esculentus) is a tuber crop that 

belongs to the family Cyperaceae, which is cultivated 

throughout the world and is widely found in the 

northern parts of Nigeria. It has been found to contain 

appreciable quantities of the fatty acids; myristic acid, 

oleic acid, linoleic acid (Eteshola & Oraedu, 1996). 

Tigernuts have been reported as helping in the prevention 

of heart attacks and thrombosis by enhancing 

blood circulation. In addition, tigernuts are believed to 

assist in reducing the risk of colon cancer (Anonymous, 

2005). 

Tigernuts are rich in energy content (starch, fat, 

sugars and protein), mineral (phosphorus, potassium) 

and vitamins E and C (Anonymous, 2005). According to 

Umerie & Enebeli (1997), tigernut is valued for its high 

dietary fibre content. Dietary fibre plays an important 

*Correspondent: E-mail: chinmachiemela@yahoo.com 


role in the human health because of the prevention, 

reduction and treatment of some disease such as 

diverticular disease, colorectal cancer, diabetes, obesity 

or cardiovascular disease (Escudero & Gonza´lez, 2006). 

However, tigernut is toxicologically safe for human 

consumption, and products such as starch, ice cream, 

milk and kunnu (a non-alcoholic beverage consumed in 

Nigeria) have been prepared using tigernut seed (Umerie 

& Obi, 1997; Onovo & Ogaraku, 2007; Belewu & 

Abodurin, 2008). 

Several flour samples such as African yam bean, 

banana and chickpea have been incorporated into wheat 

flour for cake preparation to provide better overall 

essential amino acid balance, combat the world protein 

calorie malnutrition problem and micronutrient deficiency, 

enhance sensory property and reduce total 

dependence on imported wheat flour (Go´mez et al., 

2008; Alozie et al., 2009; Eke et al., 2009). There is no 

research work on the effect of incorporating tigernut 

flour in wheat-based cakes. Considering the welldocumented 

health benefits of tigernut, substitution of 

wheat flour with tigernut flour for cake preparation 

might increase the nutritional status of the populace, 

reduce over-dependence on imported wheat flour 

because wheat cannot be grown in commercial quantity 

in Nigeria because of the country’s climatic condition, 

save foreign exchange and increase the utilisation of 

tigernuts in developing countries like Nigeria. 

doi:10.1111/j.1365-2621.2010.02334.x 


The objectives of this study were to determine the 

effect of substitution of tigernut flour for wheat flour on 

proximate composition and pasting properties of wheat 

and tigernut flour blends, and to determine the chemical, 

physical and sensory properties of cakes prepared from 

their blends. 


Materials 

Three kilograms of freshly harvested tigernut seeds and 

three kilograms of wheat flour (Golden Penny) and 

other ingredients were purchased from Minna Central 

Market, Minna, Nigeria. 

Preparation tigernut flour 

Brown tigernut seeds were sorted, winnowed and 

washed using cold tap water, dried at 60 °C in an 

air-draft oven (Gallenkamp 300 plus series, England) 

and then ground into flour using attrition mill (Globe P 

44, China). The flour samples were passed through a 

0.45 mm mesh size sieve. It was then packaged in an air 

tight polyethylene bag and stored in a plastic container 

with lid and then stored in a freezer at )18 °C from 

where samples were taken for analysis. 

Formulation of blends 

Wheat and tigernut flours were mixed at different 

proportions (100: 0%; 90 : 10%; 80 : 20%; 70 : 30%; 

60 : 40% and 50 : 50%) where 100% wheat flour 

served as control. A Kenwood mixer was used for 

mixing samples at speed 6 for 5 min to achieve uniform 

mixing. 

Proportion of ingredients 

The proportion of ingredients used consists of flour 

(100 g), sugar (62.5 g), margarine (62.5 g), (47.9 g), 

baking powder (5.7 g) and vanilla essence (three drops) 

as described by Akubor (2004). 

Preparation of cake 

The method of Akubor (2004) was adopted for the 

preparation of cake. The margarine and sugar were 

creamed manually for 2 min in a bowl until soft and 

fluffy. The egg was beaten for 3 min, added to the 

mixture and mixed manually for 3 min. Flour samples 

from various composite blends were separately sieved, 

and baking powder was then added and mixed lightly by 

hand until soft dough was formed. The dough was 

transferred to a greased baking pan and baked in a preheated 

oven at 200 °C for 30 min. 

Effect of tigernut (Cyperus esculentus) flour on wheat-based cake C. E. Chinma et al. 1747 

Chemical analysis 

Moisture, protein, crude fibre, fat, ash, carbohydrate, 

energy and mineral contents of flour blends and baked 

cake samples were determined according to AOAC 

(1995). 

Determination of pasting properties using rapid visco 

analyzer 

Pasting parameters was determined using rapid visco 

analyzer (RVA) (Newport Scientific Pty Ltd., Warriewood 

NSW 2102, Australia). A 2.5 g of starch samples 

was weighed into a dried empty canister; then 25 mL of 

distilled water was dispensed into the canister containing 

the sample. The suspension was thoroughly mixed, and 

the canister was fitted into the RVA. Each suspension 

was kept at 50 °C for 1 min and then heated up to 95 °C 

at 12.2 °C min )1 and held for 2.5 min at 95 °C. It was 

then cooled to 50 °C at 11.8 °C min ) and kept for 2 min 

at 50 °C. 

Determination of physical properties of cakes 

Batter density was determined with a measuring cylinder 

and expressed as the relation between the weight of 

batter and the same volume of distilled water. Volume 

of cake was determined by seed displacement method as 

described by AACC (2000). Volume index of cake 

samples was measured according to AACC (2000). In 

this method, cake is cut vertically through the centre and 

the heights of the cake sample were measured at three 

different points (B, C and D) along the cross-sectioned 

cakes using the template. According to this method, 

volume index was determined by the following formula: 

Volume index ¼ B þ C þ D 

where C is the height of the cake at the center point, and 

B and D are the heights of the cake at the points 2.5 cm 

away from the center towards the left and right sides of 

the cake, respectively. Weight of cake was determined by 

weight measurement using the electronic digital balance. 


A trained twenty-member panel consisting of students 

and members of Food Science option, Department of 

Animal Production Federal University of Technology, 

Minna, Nigeria was selected based on their experience 

and familiarity with cake for the sensory evaluation. 

Cake samples prepared from each flour blend were 

presented in coded white plastic plates. The order of 

presentation of samples to the panel was randomised. 

Tap water was provided to rinse the mouth between 

evaluations. The panelists were instructed to evaluate 

the coded samples for appearance, crust colour, crumb 


1748 

Effect of tigernut (Cyperus esculentus) flour on wheat-based cake C. E. Chinma et al. 

colour, crumb grain, texture, aroma and overall acceptability. 

Each sensory attribute was rated on a 9-point 

Hedonic scale (1 = disliked extremely while 9 = liked 

extremely). 


Data were analysed by analysis of variance (Steel & 

Torrie, 1980). The difference between mean values was 

determined by least significant difference test. Significance 

was accepted at 5% probability level. 


Proximate composition of flour blends 

The result of the proximate composition of flour blends 

presented in Table 1 shows that substitution of tigernut 

flour for wheat flour in the blends had no significant 

(P ‡ 0.05) effect on the moisture content of the composite 

blends, which were generally within the safe level 

of 10% recommended by Standards Organisation of 

Nigeria (SON, 1988) for flour samples. The protein 

content of the composite blends was slightly higher than 

that of wheat flour owing to the higher protein of 

tigernut flour (11.05%) compared to wheat flour 

(10.05%). In addition, fat, crude fibre and ash content 

Table 1 Proximate composition of wheat and tigernut flour blends 

of the flour blends increased significantly with increasing 

level of tigernut flour substitution because of the 

addition effect of tigernut flour with higher fat, crude 

fibre and ash content than 100% wheat flour. On the 

other hand, substitution of wheat flour by tigernut flour 

caused a reduction in carbohydrate content of composite 

blends owing to the dilution effect of tigernut flour as 

wheat flour had higher carbohydrate content than 

tigernut flour. 

Pasting properties of flour blends 

The pasting properties of wheat and tigernut flour 

blends are presented in Table 2. The RVA properties 

of flour samples showed that wheat-tigernut flour 

blends had the highest peak viscosity, trough, breakdown, 

final viscosity and setback value than wheat 

flour. There were significant (P £ 0.05) differences in 

peak viscosity, trough, breakdown, final viscosity, 

setback and pasting temperature among flour blends. 

In addition, peak time and peak temperature decreased 

with increasing level of tigernut flour substitution 

possibly indicating a lesser energy requirement owing 

to the shortened time and temperature required for 

spontaneous increase in viscosity during the processing 

of the flour blends when compared to wheat flour 

alone. 

Flour blends 

Wheat:Tigernut Moisture (%) Protein (%) Fat (%) Crude fibre (%) Ash (%) Carbohydrate (%) 

100:0 (control) 10.05 a ± 0.16 10.19 c ± 0.21 1.37 e ± 0.05 0.94 d ± 0.42 1.03 c ± 0.07 76.42 a ± 1.08 

0:100 8.33 c ± 0.18 11.25 a ± 0.56 6.04 a ± 0.85 5.10 a ± 0.04 3.17 a ± 0.11 66.11 f ± 1.75 

90:10 8.60 bc ± 0.46 10.43 bc ± 0.02 3.13 d ± 0.30 1.14 d ± 0.03 1.78 b ± 0.12 74.92 b ± 0.84 

80:20 8.84 bc ± 0.70 10.59 b ± 0.19 3.35 d ± 0.53 2.03 c ± 0.11 1.94 ab ± 0.34 73.25 c ± 0.51 

70:30 8.72 bc ± 0.63 10.77 a ± 0.34 3.82 c ± 0.89 2.99 bc ± 0.5 2.42 ab ± 0.25 71.35 d ± 1.09 

60:40 9.18 b ± 0.15 10.85 a ± 0.09 4.46 b ± 1.01 3.30 a ± 0.01 2.65 ab ± 0.13 69.56 e ± 0.97 

50:50 9.31 b ± 0.03 10.91 a ± 0.60 4.58 b ± 0.45 3.74 b ± 0.50 2.80 a ± 0.08 68.66 e ± 1.43 

Values are means and standard deviation of three determinations (n = 3). 

Values followed by different superscript in a column are significantly (P £ 0.05) different from each other. 

Table 2 Pasting properties of wheat and tigernut flour blends 

Blend ratio 

Wheat:Tigernut 

Peak viscosity 

(RVU) 

Trough 

(RVU) 

Breakdown 

(RVU) 

Final viscosity 

(RVU) 

Set back 

(RVU) 

Peak time 

(min) 

Pasting temperature 

(°C) 

100:0 (control) 57.10 f ± 0.91 33.47 f ± 0.56 23.63 f ± 0.28 81.23 f ± 0.41 47.76 f ± 0.70 5.81 a ± 0.00 89.1 a ± 0.11 

90:10 125.24 e ± 1.22 80.14 e ± 0.34 45.10 e ± 0.61 130.53 e ± 1.07 50.39 e ± 0.51 5.50 a ± 0.00 86.30 b ± 0.32 

80.20 141.30 d ± 0.73 83.25 d ± 0.49 58.05 d ± 0.87 138.68 d ± 0.55 55.43 d ± 0.97 5.42 a ± 0.00 85.05 c ± 0.05 

70:30 167.92 c ± 0.69 103.97 c ± 1.51 63.95 c ± 1.03 164.53 c ± 0.91 60.56 c ± 1.14 5.33 ab ± 0.00 86.69 b ± 0.40 

60:40 182.50 b ± 1.25 110.46 b ± 0.97 72.04 b ± 0.56 184.76 b ± 1.37 74.30 b ± 0.89 5.25 b ± 0.01 83.01 d ± 0.63 

50:50 231.63 a ± 2.09 145.29 a ± 1.15 86.34 a ± 0.71 230.37 a ± 1.61 85.08 a ± 0.72 5.17 b ± 0.00 82.94 e ± 0.91 




Peak viscosity is the ability of starch to swell freely 

before their physical breakdown (Sanni et al., 2004). 

Peak viscosity of the flours increased with increasing 

substitution of tigernut flour in the blends. This may be 

possibly because of the low amylose and phosphorus 

contents of tigernut flours as Zaidul et al. (2007) and 

Blennow et al. (2001) found high peak viscosity to be 

associated with low amylose and low phosphorus 

contents in flour and starch samples. In addition, high 

protein and fat contents in food samples are known to 

influence pasting properties (Eliasson et al., 1981). In 

this study, however, the increases recorded in protein 

and fat contents with increasing levels of tigernut 

substitutions did not seem to influence trough, breakdown, 

final viscosity and setback value. 

The setback values for the flours in general were low 

although the values of composite blends were higher 

than 100% wheat flour. According to Go´mez et al. 

(2008), a low setback value is an indication of softer 

crumb. This implies that such composite blends when 

used in bakery should produce products with soft 

crumb, a desirable attributes in cakes. 

Pasting temperature was observed to be higher in 

wheat flour than composite blends. The high pasting 

temperature of wheat flour when compared to wheattigernut 

flour blends indicates the presence of starch that 

is highly resistant to swelling. Pasting temperature in the 

range 81.00–84.05 °C has been previously reported for 

instant yam-breadfruit blends (Adebowale et al., 2008). 

The results of pasting properties obtained in this study 

are in agreement with the results of Zaidul et al. (2007) 

and Adebowale et al. (2008). 

Proximate composition of cakes 

Table 3 shows the proximate composition of cakes 

processed from tigernut and wheat flour blends. The 

addition of tigernut flour to wheat flour increased 

the contents of protein from (22.30–26.93%), fat 

(4.17–7.21%), crude fibre (3.12–5.79%), ash (1.94–3.81%) 

and energy value (342.09–390.93 Kcal) while moisture 

and carbohydrate contents decreased with the addi- 

tion of tigernut flour from (14.63–11.20%) and 

(54.58–48.67%), respectively. Moisture in foods facilitates 

various reactions, which determine the shelf-life 

and microbial susceptibility of food products. The 

higher moisture content of cakes prepared from composite 

blends than 100% wheat flour may be attributed 

to high water-binding properties of wheat flour than 

tigernut flour. 

Increased protein, fat, crude fibre, ash and energy 

value of cake samples prepared from composite blends 

as observed also in flour blends may be because of the 

addition effect of tigernut flour. Increase in protein 

content with increased tigernut flour in cakes was as a 

result of the higher protein content of tigernut flour. In 

addition, the increased protein content of the cakes 

when compare to the composite flours could have been 

because of the added ingredients such as egg during cake 

preparation. The moisture, ash and protein contents of 

cake samples obtained in this study was comparable to 

the values of 14.00–14.40% (moisture) and 3.70–5.50% 

(ash) reported by Alozie et al. (2009) for wheat- African 

yam bean cake as well as 11.00–24.70% (protein) 

reported by Akubor (2004) for wheat-cowpea cake. 

Mineral composition of cakes 

Table 3 Proximate composition of cake samples prepared from wheat and tigernut flour blends 

Blends 


Moisture 

(%) 

Protein 

(%) 

Fat 

(%) 

The result of mineral compositions of cakes prepared 

from wheat and tigernut flour blends is presented 

in Table 4. Addition of tigernut flour to wheat flour 

increased the contents of copper (0.10–0.13 mg 100 g )1 ), 

manganese (12.30–64.77 mg 100 g )1 ), iron (3.13–4.19 mg 

100 g )1 ), sodium (286.34–335.45 mg 100 g )1 ), calcium 

(54.01–56.41 mg 100 g )1 ) in cakes while phosphorus 

contents in cakes decreased with the addition of tigernut 

flour (308.01–300.00 mg 100 g )1 ). Increase in mineral 

contents of cakes as a result of tigernut flour addition to 

wheat flour could be attributed to the relatively higher 

mineral contents of the former. The importance of these 

minerals in human system cannot be over-emphasised. 

For instance, calcium helps in bone development and its 

deficiency can lead to improper development of bone in 

growing children leading to various deformities of the 

Crude fibre 

(%) 

Ash 

(%) 

Carbohydrate 

(%) 

Energy value 

(Kcal) 

100:0 (control) 14.63 b ± 0.17 22.30 c ± 0.32 4.17 d ± 0.01 3.12 c ± 0.18 1.94 c ± 0.01 53.84 a ± 0.54 342.09 c ± 1.93 

90:10 11.50 a ± 0.30 24.96 b ± 0.16 4.39 c ± 0.67 3.96 c ± 0.02 2.22 c ± 0.01 52.97 b ± 1.27 351.23 b ± 0.55 

80:20 11.66 a ± 0.24 25.15 b ± 0.11 5.07 b ± 0.09 4.43 b ± 0.15 2.95 b ± 0.05 50.74 c ± 0.94 349.19 c ± 2.80 

70:30 11.79 a ± 0.06 25.74 b ± 0.52 5.92 a ± 0.62 4.81 ab ± 0.43 3.07 b ± 0.01 48.67 d ± 0.73 350.92 a ± 1.12 

60:40 11.20 a ± 0.33 26.49 a ± 0.21 6.34 a ± 0.81 5.35 a ± 0.22 3.57 a ± 0.03 52.95 e ± 1.14 374.82 a ± 2.19 

50:50 11.84 a ± 0.10 26.93 a ± 0.58 7.21 a ± 0.05 5.79 a ± 0.03 3.81 a ± 0.01 54.58 f ± 2.01 390.93 b ± 0.86 





1750 


Table 4 Mineral contents of cakes prepared from wheat and tigernut flour blends 

Blends 


Cu 

(mg 100 g )1 ) 

Mn 

(mg 100 g )1 ) 

Fe 

(mg 100 g )1 ) 

skeletal system (Gahlawat & Sehgel, 1993). Calcium 

contributes towards bone and teeth formation, muscle 

contraction and blood clotting (Igoe, 1989). Iron is 

necessary for the prevention of anaemia, (Igoe, 1989). 

Zinc functions as a nutrient and dietary supplement and 

is believed to be necessary for nucleic acid metabolism, 

protein synthesis and cell growth (Igoe, 1989). Phosphorus 

is an important mineral associated with calcium in 

bone. The calcium ⁄ phosphorus ratio in bone is 2:1. The 

imbalance of these minerals may lead to bone disorder 

known as osteoporosis, which is a major public health 

problem, as it may cause serious complication leading to 

incapacitation and requiring costly medical care (Spencer 

& Karmer, 1988). Manganese functions as a nutrient and 

dietary supplement. It is needed in a very small amount 

(Robinson, 1973). 

The values of manganese, iron, phosphorus, calcium 

and sodium values obtained in this study were 

higher than manganese (14.60–16.30 mg 100 g )1 ), iron 

(0.61–2.21 mg 100 g )1 ), phosphorus (119.50–300.00 mg 

100 g )1 ), calcium (8.50–18.30 mg 100 g )1 ) and sodium 

(246.00–306.00 mg 100 g )1 ) reported by Eke et al. 

(2008) for some Nigerian cakes. The result of mineral 

contents of cakes indicates that cakes prepared from 

wheat-tigernut flour blends contain appreciable levels of 

minerals and could contribute to mineral intake of 

consumers. 

Physical properties of batter and cakes 

The physical properties of batter and cakes prepared 

from wheat-tigernut flour blend are presented in 

Table 5. The weight (31.25–33.58 g) and volume 

(232.60–242.43 cm 3 ) of tigernut flour supplemented cake 

increased, whereas batter density and volume index of 

cakes decreased from (0.98–0.84) and 107.69–86.35) 

with increasing level of tigernut flour. Increase in weight 

and volume of cakes with increasing the level of tigernut 

flour in wheat flour may be attributed to low batter 

density or increased bulk density of flour blends because 

of high bulk density of tigernut flour. This is in line with 

the earlier report of Chinma et al. (2007) that weight 

Mg 

(mg 100 g )1 ) 

Table 5 Physical properties of batter and cakes 

Blend 

Wheat: 

Tigernut 

Batter 

density 

Na 

(mg 100 g )1 ) 

Cake 

weight 

(g) 

and volume of snack products depends on bulk density 

of the flour blends in the product. Also, increased water 

retention in cakes as a result of increasing level of 

tigernut flour substitution may be responsible for such 

increase in weight of cake samples while high peak 

viscosity in composite blends could be responsible for 

high gas retention and high expansion of the product 

resulting to an increase in cake volume. Such cakes that 

are heavier with large volume attributes will be desirable 

by consumers because cake products with increased 

volume and weight are preferred. The decrease in batter 

density may be because of the fact that increasing the 

level of tigernut flour decreased air incorporation in the 

batter. Similar observation has been reported by Go´mez 

et al. (2008) for wheat-chickpea cake. 

Sensory properties of cakes 

Ca 

(mg 100 g )1 ) 

Cake 

volume 

(cm 3 ) 

P 

(mg 100 g )1 ) 

100:0 (control) 0.10 b ± 0.17 12.30 c ± 0.32 3.13 c ± 0.01 60.09 f ± 0.18 286.34 f ± 0.01 54.01 c ± 0.54 308.01 a ± 1.93 

90:10 0.10 a ± 0.30 19.51 b ± 0.16 3.22 b ± 0.05 83.65 e ± 0.02 302.11 e ± 0.01 55.19 b ± 1.27 303.07 d ± 0.55 

80:20 0.12 a ± 0.24 27.75 b ± 0.11 3.47 b ± 0.01 104.09 d ± 0.15 316.05 d ± 0.05 55.42 b ± 0.94 305.90 c ± 2.80 

70:30 0..14 a ± 0.06 42.04 b ± 0.52 3.68 b ± 0.12 118.82 c ± 0.43 323.10 c ± 0.01 55.80 b ± 0.73 300.00 a ± 1.12 

60:40 0.14 a ± 0.33 55.50 a ± 0.21 3.92 a ± 0.10 135.13 b ± 0.22 335.45 a ± 0.03 55.97 b ± 1.14 306.40 b ± 2.19 

50:50 0.13 a ± 0.10 64.77 a ± 0.58 4.19 a ± 0.05 152.46 a ± 0.03 328.81 b ± 0.01 56.41 a ± 2.01 303.67 d ± 0.86 



Cake 

volume 

index 

100:0 

(control) 

0.98 a ± 0.01 31.25 c ± 0.01 232.60 e ± 0.34 107.69 a ± 0.05 

90:10 0.96 a ± 0.01 31.79 bc ± 0.14 236.35 d ± 0.09 94.52 b ± 0.01 

80:20 0.93 a ± 0.00 32.25 b ± 0.05 240.17 c ± 1.15 86.55 d ± 0.01 

70:30 0.90 a ± 0.00 32.60 b ± 0.01 241.84 b ± 0.78 88.01 c ± 0.05 

60:40 0.87 a ± 0.05 32.41 b ± 0.36 241.79 b ± 0.05 86.35 d ± 0.01 

50:50 0.84 a ± 0.01 33.58 a ± 0.01 242.43 a ± 1.64 88.84 c ± 0.03 


Values followed by different superscript in a column are significantly 

(P £ 0.05) different from each other. 

Table 6 shows the sensory properties of cakes. Crust 

colour of cake samples increased with increasing level of 

tigernut flour substitution. There was no significant 

(P ‡ 0.05) difference in crust colour between control 

sample cake and cakes prepared from composite blend 

from 20% tigernut flour substitution. The crust colour 

of cakes is usually a result of the Maillard reactions 

between sugars and amino acids, and the caramelisation 


Table 6 Sensory properties of cakes prepared from wheat and tigernut 

flour blends 

Blend 

Wheat: 

Tigernut 

Crust 

colour 

100:0 (control) 7.5 a 

90:10 5.9 bc 

80:20 6.1 b 

70:30 6.4 b 

60:40 6.9 b 

50:50 7.1 a 

Crumb 

colour 

7.2 a 

6.8 a 

7.1 a 

7.0 a 

6.9 a 

7.0 a 

Crumb 

grain Texture Aroma 

6.7 a 

6.5 a 

6.0 a 

6.1 a 

5.7 b 

5.4 b 

6.0 a 

5.9 a 

6.2 a 

6.2 a 

6.0 a 

6.0 a 

7.0 a 

5.9 b 

6.8 a 

7.3 a 

7.3 a 

7.0 a 

Overall 

acceptability 

7.4 a 

7.0 a 

7.1 a 

7.2 a 

7.2 a 

7.4 a 

Mean value of twenty-member panelist. 

Values followed by different superscript in a column are significantly 

(P £ 0.05) different from each other. 

process of sugars during the baking process. Increasing 

crust and crumb colour of cake observed as the level of 

tigernut flour substitution in wheat flour increased may 

be attributed to the high protein content of the 

composite flours compared to the wheat flour. 

Crumb colour scores showed no significant (P ‡ 0.05) 

difference among samples. The crumb colour score 

increased slightly with increasing level of tigernut flour 

substitution. Crumb of baked products such as cakes 

and bread are not exposed to temperatures above 

100 °C which implies that Maillard or browning reaction 

may not take place at such temperatures. On this 

basis, it means that the slight increase in crumb colour in 

composite cakes with increasing tigernut substitution 

may be mainly because of the colour of the tigernut 

interactions of added ingredients. The crumb grains of 

100% wheat cake compared favourably with composite 

cakes at up to 30% tigernut flour substitution. Significant 

difference (P £ 0.05) was recorded in crumb grain 

of cakes prepared with up to 40% and 50% tigernut 

flour substituted flour. However, these blends showed 

significant (P £ 0.05) difference in crumb grain with 

other cake samples. Crumb grain score of 100% wheat 

cake was higher than cakes prepared from composite 

flour blends and decreased as the level of tigernut flour 

substitution in wheat flour increased. This could be 

attributed to high particle size of tigernut flour than 

wheat flour, which became pronounced as the level 

(from 40%) of tigernut substitution increases in the 

blend. 

Texture of cake improved with tigernut flour substitution. 

The increase in texture scores for cakes with 

increasing level of tigernut flour substitution in wheat 

flour could be because of the higher crude fibre content 

of tigernut flour. 

The aroma of cakes increased with increasing level of 

tigernut flour. Food aroma is usually associated with the 

interaction of flavour compounds when food products 

are subjected to high temperatures. 


There were no significant (P ‡ 0.05) differences in 

overall acceptability among samples. However, when 

other sensory parameters are put into consideration, 

cakes prepared from 70:30 (wheat:tigernut) flour blend 

showed higher scores in all the sensory parameters 

evaluated by panelists Overall, 70:30 (wheat: tigernut) 

flour blend enjoyed high sensory score in all sensory 

parameters evaluated and compared favourably with 

100% wheat cake. 

Conclusions 

It can be inferred from this study that substitution of 

tigernut flour in wheat flour for cake preparation is 

possible. The proximate, mineral and pasting properties 

of wheat flour improved as a result of tigernut flour 

substitution. Cakes prepared from such composite flours 

could help in reducing protein energy and micronutrient 

deficiency prevalent in developing countries such as 

Nigeria. 

References 

AACC. (2000). Approved methods of the American Association of 

Cereal Chemists, method 10 91, 10th edn. St. Paul, MN: American 

Association of Cereal Chemists, Inc. 

Adebowale, A.A., Sanni, S.A. & Oladapo, F.O. (2008). Chemical, 

functional and sensory properties of instant yam-bread fruit flour. 

Nigerian Food Journal, 26, 2–12. 

Akubor, P.I. (2004). Protein contents, physical and sensory properties 

of Nigerian snack food (cake, Chin-chin and puff-puff) prepared 

from cowpea-wheat flour blends. International Journal of Food 


Alozie, Y.A., Udofia, S., Lawal, O. & Ani, I.F. (2009). Nutrient 

composition and sensory properties of cakes made from wheat and 

African yam bean flour blends. Journal of Food Technology, 7, 115– 

118. 

Anonymous (2005). Tigernut and health. Available at: http://www. 

tigernut.com/salud.html (last accessed 4th June 2009). 

AOAC (1995). Official Methods of Analysis, 16th edn. Washington 

DC.: Association of official analytical chemist. 

Belewu, M.A. & Abodurin, O.A. (2008). Preparation of kunu from 

unexploited rich source tigernut. Pakistan Journal of Nutrition, 7, 

109–111. 

Blennow, A., Bay-Smidt, A.M. & Bauer, R. (2001). Amylopectin 

aggregation as a function of starch phosphate content studied by 

size exclusion chromatography and on-line refractive index and light 

scattering. International Journal of Biological Macromolecules, 28, 

409–420. 

Chinma, C.E., Ingbian, E.K. & Akpapunam, M.A. (2007). Processing 

and acceptability of fried cassava balls (‘‘Akara-akpu’’) supplemented 

with melon and soyabean flour. Journal of Food Processing 

and Preservation, 31, 143–156. 

Eke, J., Achinewhu, S.C. & Sanni, L. (2008). Nutritional and 

sensory qualities of some Nigerian cakes. Nigerian Food Journal, 

26, 12–17. 

Eke, J., Sanni, S.A. & Owuno, F. (2009). Proximate and sensory 

properties of banana cakes. Nigerian Food Journal, 27, 102–106. 

Eliasson, A.C., Carlson, T.L.-G., Larsson, K. & Miezis, Y. (1981). 

Some effects of starch lipids on the thermal and rheological 

properties of wheat starch. Starch ⁄ Starke, 33, 130–134. 

Escudero, E. & González, P. (2006). La fibra diete¢tica. Nutricio¢ 

Hospitalaria, 21, 61–72. 


1752 


Eteshola, E. & Oraedu, A.C.I. (1996). Fatty acids composition of 

Tigernut tubers (Cyperus esculentus L.), baobab seeds (Adansonia 

digitata L.) and their mixtures. Journal of the American Oil Chemists 

Society, 73, 255–267. 

Gahlawat, P. & Sehgel, S. (1993). The influence of roasting and 

malting on the total and extractable mineral contents of human 

weaning mixtures prepared from Indian raw materials. Food 

Chemistry, 46, 253–256. 

Go´mez, M., Oliete, B., Rosell, C.M., Pando, V. & Fernández, E. 

(2008). Studies on cake quality made of wheat-chickpea flour blends. 

LWT-Food Science and Technology, 41, 1701–1709. 

Igoe, R.S. (1989). Dictionary of Food Ingredients. New York: Vas Nost 

and Reinhoed. 

Onovo, J.C. & Ogaraku, A.O. (2007). Studies on some microorganisms 

associated with exposed tigernut milk. Journal of Biological Science, 

7, 1548–1550. 

Robinson, C.H. (1973). Fundamentals of Normal Nutrition, 2nd edn. 

New York: London Collier Macmillan Publishers. 

Sanni, L.O., Kosoko, S.B., Adebowale, A.A. & Adeoye, R.J. (2004). 

The influence of palm oil and chemical modification on the pasting 

and sensory properties of fufu flour. International Journal of Food 

properties, 7, 229–237. 

Spencer, H. & Karmer, L. (1988). Calcium, phosphorus and fluoride. 

In: Trace Minerals in Foods (edited by K.T. Smith). Pp. 40–46. New 

York and Basel: Marcel and Dekker, Inc. 

Standard Organization of Nigeria (SON) (1988). Nigerian Industrial 

standard for gari. 188–189. 

Steel, R.D.G. & Torrie, J.H. (1980). Principle and Procedures of 

Statistics. A Biometrical Approach, 2nd edn. P. 623. New York: 

McGraw Hill Co. 

Umerie, S.C. & Enebeli, J.N. (1997). Malt caramel from the nuts of 

Cyperus esculentus. Journal of Bioresource Technology, 52, 215– 

216. 

Umerie, S.C. & Obi, N.A.N. (1997). Short communication: isolation 

and characterization of starch from Cyperus esculentus tubers. 

Journal of Bioresource Technology, 62, 63–65. 

Zaidul, I.S.M., Yamauchi, H., Kim, S.J., Hashimoto, N. & Noda, T. 

(2007). RVA study of mixtures of wheat flour and potato starches 

with different phosphorus content. Food Chemistry, 102, 1105– 

1111. 



Book review 

The obesity epidemic and its management 

By T. Maguire, D. Haslam 

London, UK: Pharmaceutical Press, 2010. 

ISBN-978-0-85369-786-2; Price: £29.99. 

The book, entitled The Obesity Epidemic and its Management 

(264 pp) is aimed at primary healthcare 

professionals and provides a basic understanding of 

obesity, explaining why it is a major public health 

problem. This understanding will assist primary healthcare 

workers in supporting their patients and clients to 

make healthier choices in their daily lives and modify 

their lifestyles. The book covers all aspects of the obesity 

and the components of its management. It is broadly 

divided into two parts; Part 1 states the problem and 

Part 2 considers some solutions. 

Part 1 discusses the background and epidemiology 

of obesity. This section approaches obesity as a public 

health problem and addresses the different aspects of 

the problem in six chapters. Chapter 1 introduces the 

problem using current epidemiological data and 

explains why it has become such a major public 

health crisis over the last 30 years. Chapter 2 considers 

the definitions of obesity and the measurements 

used to define the problem. The use of anthropometrical 

and clinical measurements at different age groups 

is discussed to show how effective these measurements 

are in practice. Chapter 3 explains the medical and 

social impacts of obesity and its burden using the 

robust data. Since obesity is accepted as an independent 

risk factor for major diseases such as coronary 

heart disease, type 2 diabetes mellitus and cancers, the 

chapter covers the explanations about all these medical 

conditions. Chapter 4 discusses the complex 

relationship between nutrition, exercise, food politics 

and public health. The chapter gives a short history of 

food policy and politics of food. Food labelling is 

suggested as an effective solution in making healthy 

choices. In Chapter 5, the metabolic basis of nutrition 

is explained to provide a background to understanding 

of what constitutes a healthy diet. Carbohydrates, 

proteins, fats and cholesterol are described, and the 

metabolic processes of these nutrients are given 

briefly. The science of appetite regulation, especially 

with regard to obesity is discussed in Chapter 6. The 

discussion covers physiological and genetic factors 

which can influence hunger, appetite and satiety. The 

role of FTO gene is explained as a part of genetic 

factors. The physiological factors cover strong and 

weak signals of meal initiation, maintenance of eating 

and meal ending. The roles of endocannabinoid 

doi:10.1111/j.1365-2621.2010.02287.x 

Ó 2010 The Author. Journal compilation Ó 2010 Institute of Food Science and Technology 

system, cholecystokinin and leptin are discussed in 

detail. 

Part 2 focuses on the approaches to the management 

of obesity, including various psychological, 

pharmacological and surgical interventions, and examines 

evidence of their effectiveness. Chapter 7 is the 

first chapter of Part 2 and addresses the developing 

public-health-based approach to support people who 

are already overweight or obese, and to stop normalweight 

individuals becoming overweight or obese 

in the future. The chapter discusses public health 

campaigns as a part of the management and gives 

‘Five-a-day’ and ‘Change4Life’ as examples. In addition 

to campaigns, it also enlightens the spectrum of 

services offered that address the obesity problem. This 

includes obesity management services and primary 

care based programmes. The topic of Chapter 8 is 

behavioural modification techniques. The chapter 

describes the role of behavioural interventions in 

weight management and explains effective techniques 

which are used in one-to-one intervention or in a 

group context. Chapter 9 focuses on physical activity 

and exercise as a part of intervention and explains the 

exercise patterns and types of activities in detail. The 

chapter also includes practical tips for avoiding injury 

and keeping up the exercise. The topics of Chapter 10 

are nutrition and diet. The chapter considers the key 

components of a healthy diet in a general manner, 

and explains starvation diets, food energy, vitamins 

and minerals. It also provides the links of websites 

which gives nutritional guidance for pharmacists. The 

pharmacological interventions are covered in Chapter 

11. The chapter firstly explains who should use 

pharmacotherapy and then discusses the anti-obesity 

agents and other pharmacological agents. The two 

currently licensed anti-obesity agents, orlistat and 

sibutramine, and rimonabant are discussed in detail 

using the NICE guidance. Metformin, topiramate, 

tesofensine and lorcaserin are also given under the 

title of other pharmacological agents. Chapter 12 

covers the surgical management of obesity and 

describes the surgical procedures intended to induce 

weight loss. Firstly, the patients who are eligible for 

bariatric surgery are described, and then the procedures, 

gastric banding, gastric bypass surgery, biliopancreatic 

diversion and sleeve gastrectomy, are 

discussed. The chapter also includes the role of 

primary care and pharmacy for surgery and explains 

the process after surgery. Chapter 13 covers fad diets 

and considers the facts behind the failure of many 

commercial diets to support sustainable weight loss.

1754 

Book review 

Atkins Diet, South Beach Diet, Ornish Diet, Zone 

Diet and F-Plan Diet are some of the diets discussed 

in this chapter. Besides fad diets, the chapter describes 

the ideal diet and explains the importance of the diet 

quality. Chapter 14, ‘Over-the-counter slimming aids’, 

covers the dietary slimming aids. Firstly, historical 

background of the use of these aids is given, and 

evidence for the efficiency of slimming aids is discussed. 

Ephedra sinica, chitosan, hoodia gordanii, 

chromium picolinate, L-carnitine, green tea, conjuagated 

linoleic acid, hydroxycitric acid, herbal diuretics, 

lecithin, guarana and yerba mate, hydroxmethylbutyrate, 

5-hydroxytryptophan, pyruvate and orlistat are 

discussed in detail. The chapter points the huge 

market for a range of over-the-counter slimming aids 

sold in the US and in Europe, mainly through 

pharmacies and health stores, and also draws attention 

on the role of pharmacists in the market. 

In summary, the book is a very useful resource for 

healthcare professionals providing a basic understanding 

of the background and management of obesity. It is 

well-structured and well-written by two knowledgeable 

and experienced authors. It offers an approachable and 

easy reading introduction to all aspects of the obesity. 

The Obesity Epidemic and its Management can be 

recommended as a source book for intermediate classes 

in clinical and public health nutrition as well as a 

reference for researchers and academicians. 

Dr Zehra Buyuktuncer 

Lecturer in Nutrition and Dietetics, 

Department of Biological Sciences, 

University of Chester, Parkgate Road, 

Chester CH1 4BJ, UK 

E-mail: zehra.buyuktuncer@chester.ac.uk 

International Journal of Food Science and Technology 2010 Ó 2010 The Author. Journal compilation Ó 2010 Institute of Food Science and Technology

Psychrophilic and psychrotrophic clostridia: sporulation and ...

Create successful ePaper yourself

Delete template?

Save as template?