13C and 1H NMR study of cellulose metabolism by Fibrobacter ...

Journal of Biotechnology 77 (2000) 37–47www.elsevier.com/locate/jbiotec13C and 1 H NMR study of cellulose metabolism byFibrobacter succinogenes S85Xavier Bibollet a , Nathalie Bosc b , Maria Matulova a,c , Anne-Marie Delort a, *,Genevieve Gaudet b , Evelyne Forano baLaboratoire de Synthèse, Electrosynthèse et Etude de Systèmes à Intérêt Biologique, UMR 6504 Uniersité Blaise Pascal-CNRS,63177 Aubière cedex, FrancebLaboratoire de Microbiologie, INRA, Centre de Recherches de Clermont-Ferrand-Theix, 63122 Saint-Genès-Champanelle, FrancecInstitute of Chemistry, Sloak Academy of Sciences, Dubraska cesta 9, 842 38 Bratislaa, Sloak RepublicReceived 8 February 1999; received in revised form 5 July 1999; accepted 6 July 1999AbstractFibrobacter succinogenes S85, a cellulolytic rumen bacterium, is very efficient in degrading lignocellulosic substratesand could be used to develop a biotechnological process for the treatment of wastes. In this work, the metabolism ofcellulose by F. succinogenes S85 was investigated using in vivo 13 C NMR and 13 C-filtered spin-echo difference 1 HNMR spectroscopy. The degradation of unlabelled cellulose synthesised by Acetobacter xylinum was studiedindirectly, in the presence of [1- 13 C]glucose, by estimating the isotopic dilution of the final bacterial fermentationproducts (glycogen, succinate, acetate). During the pre-incubation period of F. succinogenes cells with cellulose fibres,some cells (‘non-adherent’) did not attach to the solid material. Results for ‘adherent’ cells showed that about onefourth of the glucose units entering F. succinogenes metabolism originated from cellulose degradation. A huge reversalof succinate metabolism pathway and production of large amounts of unlabelled acetate which was observed duringincubation with glucose only, was found to be much decreased in the presence of solid substrate. The synthesis ofglucose 6-phophate was slightly increased in the presence of cellulose. Results clearly showed that ‘non-adherent’ cellswere able to metabolise glucose very efficiently; consequently the metabolic state of these cells was not responsible fortheir ‘non-adherence’ to cellulose fibre. © 2000 Elsevier Science B.V. All rights reserved.Keywords: Cellulose; 13 C and 1 H NMR; Fibrobacter; Metabolism; Adhesion; Rumen1. IntroductionAbbreiations: 13 C-FSED, 13 C-filtered spin-echo difference1H NMR spectroscopy.* Corresponding author. Fax: +33-4-73407717.E-mail address: amdelort@chimtp.univ-bpclermont.fr(A.-M. Delort)Microbial cellulases and hemicellulases arewidely used in different industrial activities, suchas in textile, detergent, brewery or wood-processing,and also in the treatment of domestic wastesand in biological treatment of fibrous feeds in the0168-1656/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.PII: S0168-1656(99)00206-0

38X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47non-ruminant livestock industry (Beguin andAubert, 1994; Selinger et al., 1996). However,these enzymes are not very efficient for the degradationof highly lignified plant cell walls becausecellulose and hemicelluloses are cross-linked tolignin which is very difficult to degrade, and protectscellulose and hemicelluloses against enzymatichydrolysis (Beguin and Aubert, 1994;Selinger et al., 1996).With the aim of developing a biotechnologicalprocess for the degradation of lignocellulosicresidues, we propose to take advantage of thehigh potential of Fibrobacter succinogenes, astrictly anaerobic bacterium from the rumen. Inthe rumen, this bacterium becomes predominantamong the cellulolytic bacteria when ruminantsare fed with a poor diet, i.e. one which ishighly lignified (Bryant and Burkey, 1953). Furthermore,it digests very efficiently the morecrystalline forms of cellulose such as straw (Dehority,1993). In this context, the cellulolytic activityof strain S85 was recently tested for itsability to degrade newspapers (Martin and Martin,1998).The enzymatic equipment of F. succinogenesexplains these specific performances. Firstly, thisbacterium degrades cellulose due to a very efficientcellulolytic system (Forsberg et al., 1993;Chesson and Forsberg, 1997). Cellulose isdepolymerised at the bacterial surface by differentcellulases and the released cellodextrins are hydrolysedto glucose and cellobiose in theperiplasm (Huang and Forsberg, 1987). Secondly,it produces ferulic acid and acetylxylan esterase,and arabinofuranosidase (McDermid et al., 1990)that are necessary to cleave the ester bonds linkinghemicelluloses to lignin, or to debranch xylanes.Finally, several different xylanases andan -glucuronidase complete the cellulolytic system(Smith and Forsberg, 1991; Malburg et al.,1993).To develop a high-performance bioreactor,based on a general concept of ‘metabolic engineering’,F. succinogenes metabolism is a subjectof our study, particularly the regulation of carbonand nitrogen metabolism pathways. The measurementof metabolic fluxes is necessary to be able todirect bacterial metabolism towards the productionof biomass and enzymes of interest.NMR is a powerful tool to obtain qualitativeand quantitative data about bacterial metabolism;we have successfully applied this technique to thestudy of carbon and nitrogen metabolism in F.succinogenes. Glucose and cellobiose, final productsof cellulose degradation, are taken up andmetabolised by the cells into essentially succinate,acetate and a small amount of formate (Miller,1978; Gaudet et al., 1992; Matheron et al., 1996,1997, 1998a). These sugars are transported acrossthe cytoplasmic membrane through independenttransporters (Franklund and Glass, 1987; Maasand Glass, 1991). Matheron et al. (1996, 1998b)showed a simultaneous but different metabolismof glucose and cellobiose by Fibrobacter strains.When the extracellular sugar concentration ishigh, some of the carbon substrates are stored asglycogen (Gaudet et al., 1992; Matheron et al.,1998a) or are released as cellodextrins into theexternal medium (Wells et al., 1995; Matheron etal., 1996, 1998b). The quantitative determinationof metabolic fluxes showed the reversibility ofdifferent metabolic pathways in F. succinogenesS85 and also in other strains of this genus: reversibilityof glycolysis, reversibility of the succinatesynthesis pathway and futile cycling ofglycogen (Matheron et al., 1998a). We recentlyshowed that the presence of ammonia increasedthe reversal of succinate synthesis pathway whilethe other reversed routes remained unchanged(Matheron et al., 1999).Until now the metabolism of soluble substrates(glucose and cellobiose) has been studied. In thiswork the metabolism of cellulose was investigatedusing in vivo 13 C NMR and 13 C-filtered spinechodifference 1 H NMR spectroscopy. In particular,the contribution of endogenous glycogen,the reversal of the succinate pathwayand of the utilisation of exogenous substrates(glucose, cellulose) to the synthesis of finalmetabolites were quantified. Moreover, since duringformation of a biofilm on cellulose fibre somebacteria did not attach to the solid material, themetabolic state of ‘adherent’ and ‘non-adherent’cells was compared.

X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47 392. Materials and methods2.1. Bacterial strains and culture conditionsF. succinogenes S85 (ATCC 19169), the typestrain of this species, isolated from the bovinerumen, was grown at 38°C under anaerobic conditionson a synthetic medium containing 3gl −1cellobiose (Gaudet et al., 1992).Acetobacter xylinum (ATCC 23768) was grownunder aerobic conditions, at 30°C, in Roux flaskscontaining a Hoyer medium with 10 g l −1 ofglucose (Larpent and Larpent-Gourgaud, 1975).2.2. Synthesis of cellulose by A. xylinumPure cellulose was synthesised from cultures ofA. xylinum, and purified from the bacterial cultureas previously described (Bertocchi et al., 1997).From1lofculture 110 mg of dried pure cellulosewas obtained.2.3. Preparation of F. succinogenes S85 cells forNMR spectroscopy2.3.1. ‘Reference’ cellsFor in vivo experiments, F. succinogenes S85cells were prepared as described in Matheron etal. (1996). The cells harvested in the late log phasewere centrifuged (6000×g, 10 min, 4°C) and suspendedin a reduced 50 mM potassium phosphate–0.4%Na 2 CO 3 –0.05% cysteine–6.5 mM(NH 4 ) 2 SO 4 –buffer (pH 7.1). The cells at a finalconcentration of 4 mg protein ml −1 were incubatedwith different substrates depending on theexperiments.2.3.2. ‘Adherent’ and ‘non-adherent’ cellsF. succinogenes S85 cells were pre-incubated for30 min at 38°C in the reduced buffer (50 mMpotassium phosphate–0.4% Na 2 CO 3 –0.05% cysteine,6.5 mM (NH 4 ) 2 SO 4 , pH 7.1) in the presenceof 0.1g cellulose which was synthesised by A.xylinum as described above. The cell suspensionwas centrifuged for 4 min at 2000 rpm. The pelletconstituted the ‘adherent’ cells, while the cellsremaining in the supernatant were considered as‘non-adherent’ cells.‘Adherent’ cells were suspended in the reducedbuffer (4 mg protein ml −1 ). ‘Non-adherent’ cellswere centrifuged again at 4000 rpm during 10min, and the pellet was suspended in the samebuffer (4 mg protein ml −1 ). These two types ofcell suspension were studied separately, they weretransferred to 10-mm NMR tubes, supplementedwith [1- 13 C]glucose, the utilisation of which wasfollowed by in vivo 13 C-NMR.2.4. NMR spectroscopy.NMR spectra were recorded on a BrukerAvance DSX 300 spectrometer operating at 75.46MHz for 13 C and at 300.13 MHz for 1 H. The 2 Hresonance of D 2 O (10%) was used to lock the fieldand for shimming.In vivo 13 C-NMR experiments in a 10-mmmultinuclear probe were performed at 38°C aspreviously described. In the Waltz-16 proton decoupledspectra, 360 scans were collected every4.5 min (acquisition time 0.256 s, relaxation time0.5 s) with a pulse length of 45°. An externalstandard of benzene ( 128.6) in a capillary wasused as an external reference for chemical shiftmeasurements and for normalisation of the valuesof metabolites integrals.1H NMR spectra were acquired using internaldeuterated 3-trimethylsilylpropionate sodium salt(TSP-d 4 )( 0.0) as a reference standard for chemicalshifts in a 5-mm inverse probe ( 1 H/ 13 C/ 15 N)with 13 C-filtered spin-echo difference ( 13 C FSED)pulse sequence (Matheron et al., 1998a):Preparation−(90°) H −/2−(180°) H /() X −/2−(90°) X −FID( 1 H)where the subscripts denote the nucleus experiencingthe pulse (H-proton; X- 13 C). The last carbonpulse is a purging pulse used to remove all spurioussignals originating from pulse imperfections.Spectra with =0 or 180° were acquired in subsequentscans and were stored independently in twoblocks of memory. Extensive phase cycling wasused to compensate quadrature detection artefacts(CYCLOPS) and 180° pulse imperfections (EX-ORCYCLE). In a preparation period, the solventresonance was presaturated by irradiation for 3.5s at 60 dB. The evolution interval was adjusted

40X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47according to the one-bond C,H coupling constantfor succinate and acetate (=1/ 1 J (C,H) =7.8 ms).After eight dummy scans, 128 scans were accumulatedinto 8K of memory with an acquisition timeof 3.5 s, a spectral width of 4000 Hz and arelaxation delay of 3.0 s.After in vivo 13 C-NMR experiments, sampleswere spun (15 000×g, 10 min, 4°C) to remove thecells and the supernatant was analysed by 13 C-FSED 1 H NMR.The determination of percentages of 13 C labellingof succinate and acetate is fully describedin Matheron et al. (1998a).2.5. Metabolite assaysProtein concentration was determined by theBradford method (Bradford, 1976), using bovineserum albumin as standard. Succinate, acetate,formate and glucose were assayed using aBoehringer kit. Glucose 6-phosphate was determinedas glucose except that hexokinase (ATP:D-hexose 6-phospho-transferase, EC2.7.1.1) wasomitted.For glycogen determination, cells were harvestedby centrifugation (15 000×g, 15 min, 4°C)and pellets were suspended in 0.25% SDS (sodiumdodecyl sulfate). The suspension was then dilutedten times in 50 mM potassium phosphate buffer(pH 4.5) and incubated with 80 mg ml −1 Rhizopusamyloglucosidase (1,4 -D-glucan glucohydrolase;EC 3.2.1.3, from Sigma) for 60 min at 55°C.Samples were centrifuged (15 000×g, 5 min) andglucose was assayed in the supernatant.As 13 C labelled cellulose is not commerciallyavailable, the degradation of cellulose was investigatedindirectly in the presence of [1- 13 C] labelledglucose; this strategy was already used in the caseof cellobiose (Matheron et al., 1996, 1998b). Thecontribution of 12 C-cellulose degradation duringthe simultaneous incubation with 100% exogenous[1- 13 C]glucose was evaluated by measuring theisotopic dilution of the final bacterial metabolites(glycogen, succinate, acetate). Indeed the participationof unlabelled glucose units in F. succinogenesmetabolism increases the ratio of 12 C/ 13 Cisotopomers (Fig. 1).Two experimental approaches were used toquantify this isotopic dilution: In vivo 13 C NMRexperiments connected to enzymatic assays of totalmetabolites and 1 H NMR experiments performedon the incubation medium.3.1.1. In io 13 C NMR experimentsFig. 2 shows in vivo 13 C NMR kinetics spectraof [1- 13 C]glucose utilisation by F. succinogenes S85(4 mg protein ml −1 ): ‘reference’ cells (A), cells‘adherent’ to 0.1 g A. xylinum unlabelled cellulose(B). In both cases, the 13 C NMR spectra recordedduring in vivo kinetics showed the same signals,2.6. Chemicals[1- 13 C]Glucose (99% labelled) was purchasedfrom Eurisotop (France). All enzymes and chemicalswere purchased from Sigma or Boehringer.3. Results and discussion3.1. NMR study of cellulose degradation by‘adherent’ cells in the presence of [1- 13 C]glucoseFig. 1. [1- 13 C]Glucose and non-labelled cellulose metabolismby F. succinogenes S85.

X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47 41Fig. 2. 13 C NMR spectra of in vivo kinetics of 32 mM [1- 13 C]glucose metabolism by F. succinogenes S85 cells: A, ‘reference’ cells.B, ‘adherent’ cells to 0.1 g Acetobacter xylinum unlabelled cellulose. [1- 13 C]Glucose was added at zero time to suspension of cells (4mg protein ml −1 ) in reduced buffer containing 6.5 mM (NH 4 ) 2 SO 4 and the proton decoupled 13 C-NMR spectra (360 scans) werecollected every 4.5 min. acet, acetate; Glc, glucose; Glyc, glycogen; succ, succinate.indicating the presence of the same metaboliteswith the same position of 13 C-labelling. Signals inthe spectra corresponded to two glucose anomers[1- 13 C ], (96.24 ppm) and [1- 13 C ], (92.41 ppm),to [1- 13 C]glycogen, (100.07 ppm) and [6-13C]glycogen, (61.12 ppm), to [2- 13 C]succinate(34.15 ppm) and to [2- 13 C]acetate (23.49 ppm).The [6- 13 C] resonance of glycogen resulted fromreversal of glycolysis after isomerisation at thetriose-phosphate level. All the observed labelled

42X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47positions are in agreement with the metabolicpathway of F. succinogenes S85 (Miller, 1978;Matheron et al., 1997).These results show that F. succinogenes cellsproduced the same metabolites when they wereincubated with [1- 13 C]glucose only, or with [1-13C]glucose and cellulose. It should be noted thatthe spectra collected in the presence A. xylinumcellulose were less resolved than recorded in thepresence of [1- 13 C]glucose alone.The relative integrals of [1- 13 C]glycogen and[2- 13 C]succinate measured during the incubationsof ‘reference’ cells or cells ‘adherent’ to unlabelledcellulose with 32 mM [1- 13 C]glucose are presentedin Fig. 3. In the last case, the incorporation of the13C-labelled C1 atom of [1- 13 C]glucose in [1-13C]glycogen and [2- 13 C]succinate was reducedcompared to the incubation with [1- 13 C]glucosealone. These results are consistent with the theoreticalisotopic dilution due to the presence ofunlabelled cellulose. To check this hypothesis, theconcentrations of succinate and acetate, glycogenand glucose 6-phosphate, which were present inthe medium at the end of the incubation, wereenzymatically assayed. The results presented inTable 1 show that the same amounts of succinateand glycogen were produced in the presence or inthe absence of cellulose and confirm that thedecrease of the integrals of [2- 13 C]succinate and[1- 13 C]glycogen resonances in in vivo 13 C NMRspectra measured during the incubations in thepresence of cellulose and [1- 13 C]glucose (Fig. 3)was due to an isotopic dilution and not the slowingdown of the metabolism. They show that F.succinogenes cells degrade cellulose even in thepresence of high concentrations of exogenousglucose.Enzymatic assays (Table 1) show that glucose6-phosphate concentration was increased from 2.1mM (in the presence of [1- 13 C]glucose alone) to3.3 mM (when cellulose was added). This featurewas also observed in cell suspensions incubated inthe presence of both [1- 13 C]glucose and cellobiose(Matheron et al., 1996, 1998b). In the last case,[1- 13 C] and [6- 13 C]glucose 6-phosphate signalscould be observed directly in 13 C NMR spectra;however, under our conditions these signals arenot visible (Fig. 2A, B) because glucose 6–phosphateconcentration is lower and also because theNMR spectral resolution is poor due to higherviscosity caused by the presence of non-solublecellulose.Fig. 3. Time-dependent [1- 13 C]glucose consumption (A) and changes of signal integrals of [2- 13 C]succinate (B) and [1- 13 C]glycogen(C) during incubation of 32 mM [1- 13 C]glucose: with ‘reference’ cells (), with ‘non-adherent cells’ () and with ‘adherent’ cells to0.1 g A. xylinum unlabelled cellulose (). Relative integrals were measured in the 13 C-NMR spectra, experimental conditions as inFig. 2. Glucose consumption was calculated by difference of the relative integrals of [1- 13 C]glucose at a given and at zero time.

X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47 43Table 1Enzymatic assays of residual glucose and metabolites produced at the end of the incubation of ‘reference’ cells of F. succinogenesS85 with [1- 13 C]glucose and of ‘adherent’ cells on A. xylinum cellulose with 32 mM [1- 13 C]glucoseSubstratesConcentrations (mM)Glucose Glucose 6-phosphate Glycogen a Succinate Acetate[1- 13 C]Glucose 1 2,1 3,2126[1- 13 C]Glucose+A. xylinum cellulose 5,2 3,3 2,8 13,7 2,8a Newly synthesised glycogen.It can be noticed that, as previously observed(Matheron et al., 1999), the presence of ammoniain the incubation with [1- 13 C]glucose increased thesynthesis of acetate, and thus decreased the succinate/acetateratio to 2 (Table 1) instead of 3,which is usually observed in the absence of ammonia.This phenomenon is no longer observed in thepresence of cellulose as the concentration of acetatewas decreased by a factor of two comparedto the incubation with glucose only (Table 1).3.1.2. 1 H NMR experiments1H NMR spectroscopy is a powerful tool toquantify 13 C/ 12 C ratios; recently we published theuse of a 13 C-filtered spin-echo difference ( 13 C-FSED) sequence (Matheron et al., 1998a, 1999) tomeasure very precisely (1% error) the 13 C enrichmentof C2 acetate and C2 succinate. Anexample of 1 H NMR experiments performed onan extract of F. succinogenes cells incubated with[1- 13 C]glucose is presented in Fig. 4. Fig. 4(A)shows a sub-spectrum resulting from the first scanof 13 C-FSED pulse sequence in which the 180°13C-pulse in the middle of the sequence was omitted(=0°). It resembles a classical 1H NMRspectrum. In Fig. 4(B) 13 C-linked proton signals( 13 C satellites which are split by one bond couplingconstant 1 J CH ) are inverted due to the 180° 13 Cpulse (=180°) in the second scan of the pulsesequence. By simple mathematical operations, e.g.subtraction and addition of both above mentionedspectra, the spectra of exclusively 12 C-linked(Fig. 4C) and 13 C-linked protons (Fig. 4D), respectively,were obtained. The integration of theirsignals provided quantitative data for a calculationof 13 C enrichment of C2 atoms of acetate andsuccinate.The values of 13 C labelling obtained for incubationsperformed with [1- 13 C]glucose and celluloseor [1- 13 C]glucose only are reported in Table 2. Aspreviously explained (Matheron et al., 1998a,Fig. 4. 1 H NMR spectra acquired with 13 C-filtered spin-echopulse sequence. Spectra A and B represent the output of theexperiment with and without 180° 13 C-pulse, respectively, inthe middle of the Spin-echo period. Spectra C and D wereobtained after addition or subtraction of the spectra A and B,respectively. Spectrum C corresponds to 12 C-linked, and spectrumD to 13 C-linked protons.

44X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–471999), the maximum percentage of labelling of C2succinate and C2 acetate should be 25 and 50%,respectively, if 100% labelled [1- 13 C]glucose wasconsumed. One [1- 13 C]glucose unit is cleaved intotwo triose-phosphates (Fig. 1) and thus twomolecules of acetate and succinate are produced,half of them being labelled (50%). In addition,succinate is a symmetrical molecule with two CH 2groups, thus only 25% of C2-succinate atoms are13C labelled. Consequently, the determination ofpercentage of C2-succinate and C2 acetate labellingallows to calculate the amount of succinate andacetate molecules produced.In this work, we found that when [1- 13 C]glucosewas the sole substrate for ‘reference’ cells, the extentof labelling of [2- 13 C]succinate was 22% (Table 2A).The labelling deficit (3% of one CH 2 of succinate,and thus 6% of the succinate molecule) correspondsto the degradation of endogenous glycogen due tothe futile cycling phenomenon (Matheron et al.,1999). This means that 12% of non-labelled glucoseunits are provided by non-labelled glycogen. In thecase of incubation of cells ‘adherent’ to non-labelledcellulose with [1- 13 C]glucose, the 13 C labellingof C2 succinate was 16% (Table 2A).Consequently, the labelling deficit (9% of CH 2 ) canbe explained by the contribution of endogenousglycogen degradation (3%) and by the degradationof A. xylinum cellulose (6%). Thus in this case 24%of glucose units come from the utilisation ofunlabelled cellulose by the bacteria.It was previously shown that when ‘reference’cells were incubated with [1- 13 C]glucose only thedeficit of the 13 C enrichment of C2 acetate representedboth the contribution of the glycogen degradationand the reversal of the succinate pathway(Matheron et al., 1998a, 1999). In this work, thedeficit of labelling of C2 acetate was found to be22% (Table 2B): 6% came from glycogen degradationand thus the reversal of succinate pathwayscontributed to 16% of the isotopic dilution of C2acetate. As previously observed (Matheron et al.,1999), the contribution of this reversal was veryhigh in the presence of ammonia. When the cellswere incubated with [1- 13 C]glucose and 12 C-cellulose,three phenomena contributed to the finalisotopic dilution of C2 acetate: endogenous glycogendegradation (6%), degradation of cellulose(12%) and reversion of succinate synthesis pathway(3%). Thus, in the presence of cellulose, the contributionof this reversal to the isotopic dilution of C2acetate was much lower when compared to thatmeasured with [1- 13 C]glucose only (Table 1). In thecase of [1- 13 C]glucose incubation, 32% of acetatemolecules were produced by reversal of the succinatesynthesis pathway, while only 6% were providedby this route in the presence of cellulose.3.2. In io 13 C NMR study of [1- 13 C]glucosemetabolism by ‘non-adherent’ cellsAfter pre-incubation of the bacteria with cellulosefor 30 min, part of the cell suspension didnot-adhere to the solid substrate and remained inthe supernatant after centrifugation of the sampleat 2000 rpm during 4 min. The question is thusraised as to why these bacteria did not attach tocellulose? Roger et al. (1990) showed that theadhesion of F. succinogenes cells to cellulose wasdependent on the integrity of the bacterialmetabolic functions. This result prompted us toinvestigate the metabolic state of ‘non-adherent’cells. For this purpose, the metabolism of [1-13C]glucose by ‘non-adherent’ cells and by ‘reference’cells (which were not pre-incubated withcellulose), was monitored by in vivo 13 C NMR. Thetime course of [1- 13 C]glucose consumption and of[2- 13 C]succinate and [1- 13 C]glycogen productionsunder these two experimental conditions are reportedin Fig. 3(A, B and C), respectively. On thebasis of these data it is obvious that ‘non-adherent’cells metabolised glucose as efficiently as ‘reference’cells, thus their metabolic state is not responsiblefor the lack of adhesion. The lack of adhesion ofthese cells might be due to a saturation of accessibleadhesion sites on cellulose by the high concentrationof cells under experimental conditions.The integrity of the bacterial metabolic functionsof ‘non-adherent’ cells was confirmed by the usingof 13 C-FSED 1 H NMR experiments, previouslydeveloped for the study of ‘adherent’ and ‘reference’cells (Table 2). The extent of labelling of C2of acetate and succinate produced by ‘non-adherent’cells were 29 and 21%, respectively, and thusrather similar to that of ‘reference’ cells (28 and22%, respectively). From these data, the con-

Table 2Percentages of 13 C enrichment of C2 of succinate (A) and C2 of acetate (B) measured in 13 C-FSED 1 H NMR experiments performed on incubation media collectedat the end of the in vivo 13 C NMR kinetics of ‘non-adherent’ and ‘adherent’ cells to A. xylinum cellulose and F. succinogenes ‘reference’ cells; calculations of themetabolic pathway contribution (%) to the deficit of labelling (theoretical-measured % of 13 C enrichment)Substrates/cellsTheoreticalMeasuredDeficit oflabellingGlycogendegradationCellulosedegradationReversal of succinatesynthesis pathwayA % [2- 13 C]Succinate Metabolic pathway contribution to the deficit of labelling (%)[1- 13 C]Glucose+A. xylinum cellulose/‘adherent’ cells[1- 13 C]Glucose/‘reference’ cells2525162293336000[1- 13 C]Glucose/‘non-adherent’ cells 2521 44 00B % [2- 13 C]AcetateMetabolic pathway contribution to the deficit of labelling (%)[1- 13 C]Glucose+A. xylinum cellulose/ 50 29 21 6 12 3‘adherent’ cells[1- 13 C]Glucose/‘reference’ cells 5028 226 0 16[1- 13 C]Glucose/‘non-adherent’ cells 50 29 218013X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47 45

46X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47tributions of the different metabolic pathways tothe isotopic dilution of C2 acetate and C2 succinatewere calculated: 16% of glucose units enteringthe glycolytic pathway are provided by thedegradation of endogenous glycogen; 26% of acetatemolecules are synthesised after reversal ofthe succinate metabolic pathway. These contributionsare in the same range as those calculated for‘reference’ cells (12% for glycogen futile cycle,32% for reversal of the succinate pathway).In addition, these experiments show that ‘nonadherent’cells could survive perfectly during halfan hour pre-incubation in the absence of addedsoluble substrate. We also found that an increaseof the incubation time of up to 1 h led to the sameresult (not shown). This might indicate intracellularglycogen utilisation by the ‘non-adherent’ cellsduring the pre-incubation time. However, theglycogen/protein ratio was the same in ‘reference’cells, that were not incubated, and in ‘non-adherent’cells after the 30-min incubation, indicatingthat ‘non-adherent’ cells did not deplete theirglycogen. This result rather suggests a nutritionalinteraction between ‘adherent’ and ‘non-adherent’cells during the pre-incubation period: ‘adherent’cells degrade cellulose and produce cellodextrinsin the incubation medium that could be utilised inturn by ‘non-adherent’ cells. Such an interactionwas evidenced when F. succinogenes and Streptococcusbois (a non-cellulolytic ruminal bacterium)were co-cultured with cellulose as the solecarbon source (Wells et al., 1995). We have alsopreviously shown the synthesis and excretion ofcellodextrins by Fibrobacter strains (Matheron etal., 1996, 1998b) during simultaneous metabolismof glucose and cellobiose. The same phenomenoncould occur during metabolism of cellulose.4. ConclusionsIn this work 1 H and 13 C NMR experimentswere carried out with the aim of collecting newinformation about the metabolism of the rumencellulolytic bacterial strain, F. succinogenes S85.First we studied the degradation of cellulose by‘adherent’ cells in the presence of [1- 13 C]glucose.Under these conditions about one fourth of theglucose molecules entering F. succinogenesmetabolism was shown to come from cellulosedegradation. Using 13 C-FSED 1 H NMR, we werealso able to quantify other metabolic contributionsto the synthesis of final products: endogenousglycogen degradation provided about 14%of glucose units entering the glycolytic pathwaywhile only 6% of acetate molecules were formedafter reversal of the succinate pathway. Theseresults showed that the huge reversal of the succinatemetabolism pathway, observed when onlyglucose was metabolised, and which producedlarge amounts of unlabelled acetate, was muchdecreased in the presence of solid substrate. Anotherdifference concerned the synthesis of glucose6-phophate, which was slightly increased inthe presence of cellulose, as previously observedfor cellobiose.Secondly, we studied the glucose metabolism of‘non-adherent’ cells; in vivo 13 C NMR and 1 HNMR analyses clearly showed that the bacteriabehaved similarly to ‘reference’ cells. Consequentlythe metabolic state of these cells was notresponsible for their ‘non-adherence’ to cellulosefibre. The active metabolic states of these bacteriasuggested a metabolic interaction between planktonicand adherent cells.Further investigation is now necessary in orderto understand the phenomena of adhesion. Theformation of a biofilm is of interest as it is the firststep for the development of a biotechnologicalprocess, efficient in degrading lignocellulosicwastes.In the future, the degradation of 13 C enrichedpure cellulose (produced from A. xylinum culturesand plants) will be investigated by 1 H and 13 CNMR using the methodologies developed in thiswork. This will allow direct monitoring of themetabolism of the predominant rumen cellulolyticspecies F. succinogenes under conditions closer toreal bioreactors, and particularly of cells adherentto more complex substrates.AcknowledgementsThis work was supported by the PIRGP-BIO(Génie des Procédés, Biotechnologie) Programme

X. Bibollet et al. / Journal of Biotechnology 77 (2000) 37–47 47of the Centre National de la Recherche Scientifique.X. Bibollet is grateful to the Centre Nationalde la Recherche Scientifique and theRégion Auvergne for a doctoral fellowship. M.Matulova is a visiting professor at the UniversityBlaise Pascal. We thank N. Veyret for technicalassistance in this project. We acknowledge D.Aitken for reading the manuscript.ReferencesBeguin, P., Aubert, J.P., 1994. The biological degradation ofcellulose. FEMS Microbial. Rev. 13, 25–58.Bertocchi, C., Delneri, D., Signore, S., Weng, Z., Bruschi,C.V., 1997. Characterization of microbial cellulose from ahigh producing mutagenized Acetobacter pasteurianusstrain. Biochim. Biophys. Acta 1336, 211–217.Bradford, M.M., 1976. A rapid sensitive method for thequantification of microgram quantities of protein, utilisingthe principle of protein-dye binding. Anal. Biochem. 72,248–254.Bryant, M.P., Burkey, L.A., 1953. Numbers and some ofpredominant groups of bacteria in the rumen of cows feddifferent rations. J. Dairy Sci. 36, 218–224.Chesson, A., Forsberg, C.W., 1997. Polysaccharide degradationby the rumen microorganisms. In: Hobson, P.N.,Stewart, C.S. (Eds.), The Rumen Microbial Ecosystem,2nd ed. Blackie Academic, London, pp. 329–381.Dehority, B.A., 1993. Microbial ecology of cell wall fermentation.In: Jung, H.G., Buxton, D.R., Hatfield, R.D., Ralph,J. (Eds.), Forage Cell Wall Structure and Digestibility.American Society for Agronomy-Crop Science Society ofAmerica-Soil Science Society of America, Madison, WI,pp. 425–453.Forsberg, C.W., Gong, J., Malburg, L.M. Jr., Zhu, H., Iyo,A., Cheng, K.J., Krell, P.J., Philipps, J.P., 1993. Cellulasesand hemicellulases of Fibrobacter succinogenes and theirroles in fibre digestion. In: Shimida, K., Hoshino, S.,Ohmiya, K., Sakka, K., Kobayashi, Y., Karita, S. (Eds.),Genetics, Biochemistry and Ecology of LignocelluloseDegradation, Uni-Publishers Co., Tokyo, Japan, pp. 125–136.Franklund, C.V., Glass, T.L., 1987. Glucose uptake by thecellulolytic ruminal anaerobe Bacteroides succinogenes. J.Bacteriol. 169, 500–506.Gaudet, G., Forano, E., Dauphin, G., Delort, A.M., 1992.Futile cycling of glycogen in Fibrobacter succinogenes asshown by in situ 1 H-NMR and 13 C-NMR investigation.Eur. J. Biochem. 207, 155–162.Huang, J., Forsberg, C.W., 1987. Isolation of a cellodextrinasefrom Bacteroides succinogenes. Appl. Environ. Microbiol.53, 1034–1041.Larpent, J.-P., Larpent-Gourgaud, M., 1975. Memento techniquede microbiologie. In: Perfectionnement scientifique ettechnique (Ed.), Technique de documentation (Lavoisier)Publisher, Paris.Maas, L.K., Glass, T.L., 1991. Cellobiose uptake by thecellulolytic ruminal anaerobe Fibrobacter (Bacteroides) succinogenes.Can. J. Microbiol. 37, 141–147.Malburg, L.M. Jr, Smith, D.C., Schellhorn, H.E., Forsberg,C.W., 1993. Fibrobacter succinogenes S85 has multiplexylanase genes. J. Appl. Bacteriol. 75, 564–573.Martin, S.A., Martin, J.A., 1998. Paper digestion by thecellulolytic ruminal bacterium Fibrobacter succinogenes.Curr. Microbiol. 37, 431–432.Matheron, C., Delort, A.M., Gaudet, G., Forano, E., 1996.Simultaneous but differential metabolism of glucose andcellobiose in Fibrobacter succinogenes cells, studied by invivo 13 C NMR. Can. J. Microbiol. 42, 1091–1099.Matheron, C., Delort, A.M., Gaudet, G., Forano, E., 1997.Re-investigation of glucose metabolism in Fibrobacter succinogenes,using NMR spectroscopy and enzymatic assays.Evidence for pentose phosphates phosphoketolase andpyruvate formate lyase activities. Biochim. Biophys. Acta1355, 50–60.Matheron, C., Delort, A.M., Gaudet, G., Forano, E., Liptaj,T., 1998a.13C- and 1 H-NMR study of glycogen futilecycling in strains of the genus Fibrobacter. Appl. Environ.Microbiol. 64, 74–81.Matheron, C., Delort, A.M., Gaudet, G., Forano, E., 1998b.In vivo 13 C NMR study of glucose and cellobiosemetabolism by four cellulolytic strains of the genus Fibrobacter.Biodegradation 9, 451–461.Matheron, C., Delort, A.M., Gaudet, G., Liptaj, T., Forano,E., 1999. Interaction between carbon and nitrogenmetabolism in Fibrobacter succinogenes S85. A 1 H and 13 CNMR study. Appl. Environ. Microbiol. 65, 1941–1948.McDermid, K.P., McKenzie, C.R., Forsberg, C.W., 1990.Esterase activities of Fibrobacter succinogenes subsp. succinogenesS85. Appl. Environ. Microbiol. 56, 127–132.Miller, T.L., 1978. The pathway of formation of acetate andsuccinate from pyruvate by Bacteroides succinogenes. Arch.Microbiol. 117, 145–152.Roger, V., Fonty, G., Komisarczuk-Bony, S., Gouet, P., 1990.Effects of physicochemical factors on the adhesion tocellulose Avicel of the ruminal bacteria Ruminococcusflaefaciens and Fibrobacter succinogenes subsp. succinogenes.Appl. Environ. Microbiol. 56, 3081–3087.Selinger, L.B., Forsberg, C.W., Cheng, K.J., 1996. The rumen:a unique source of enzymes for enhancing livestock production.Anaerobe 2, 263–284.Smith, D.C., Forsberg, C.W., 1991. -Glucuronidase andother hemicellulase activities in Fibrobacter succinogenesS85 grown on crystalline cellulose or ball-milled barleystraw. Appl. Environ. Microbiol. 57, 3552–3557.Wells, J.E., Russel, J.B., Shi, Y., Weimer, P.J., 1995. Cellodextrinefflux by the cellulolytic ruminal bacterium Fibrobactersuccinogenes and its potential role in the growth of non-adherentbacteria. Appl. Environ. Microbiol. 61, 1757–1762.