Expression, refolding, and purification of recombinant human ...

Protein Expression and PuriWcation 39 (2005) 18–26 

www.elsevier.com/locate/yprep 

Expression, refolding, and puriWcation of recombinant 

human granzyme B 

Rikke H. Lorentsen a,b,¤ , Charlotte H. Fynbo a,b , Hans C. Thøgersen a,b , 

Michael Etzerodt a,b , Thor L. Holtet b 

a 

Department of Molecular Biology, University of Aarhus, Gustav Wieds Vej 10, DK-8000 Aarhus C, Denmark 

b 

Borean Pharma A/S, Science Park Aarhus, Gustav Wieds Vej 10, DK-8000 Aarhus C, Denmark 

Received 18 May 2004, and in revised form 14 August 2004 

Abstract 

Granzyme B (GrB) is a member of a family of serine proteases involved in cytotoxic T-lymphocyte-mediated killing of potentially 

harmful cells, where GrB induces apoptosis by cleavage of a limited number of substrates. To investigate the suitability of GrB as an 

enzyme for speciWc fusion protein cleavage, two derivatives of human GrB, one dependent on blood coagulation factor X a (FX a ) 

cleavage for activation and one engineered to be self-activating, were recombinantly expressed in Escherichia coli. Both derivatives 

contain a hexa-histidine aYnity tag fused to the C-terminus and expressed as inclusion bodies. These were isolated and solubilized in 

guanidiniumHCl, immobilized on a Ni 2+ –NTA agarose column, and refolded by application of a cyclic refolding protocol. The 

refolded pro-rGrB-H6 could be converted to a fully active form by cleavage with FX a or, for pro(IEPD)-rGrB-H6, by autocatalytic 

processing during the Wnal puriWcation step. A self-activating derivative in which the unpaired cysteine of human GrB was substituted 

with phenylalanine was also prepared. Both rGrB-H6 and the C228F mutant were found to be highly speciWc and eYcient processing 

enzymes for the cleavage of fusion proteins, as demonstrated by cleavage of fusion proteins containing the IEPD recognition 

sequence of GrB. 

© 2004 Elsevier Inc. All rights reserved. 

Keywords: Granzyme B 

Recombinant production of proteins with favourable 

properties, both for research and pharmaceutical applications, 

is frequently accomplished through production 

of fusion proteins, in which the target protein is linked to 

a fusion partner that may stimulate the expression, 

increase the stability, or facilitate aYnity puriWcation of 

the fusion protein. However, the fusion partner may 

introduce undesirable changes in the biological activity 

or increase the immunogenicity of the protein of interest 

and, therefore, it is necessary to separate the fusion 

partner from the protein of interest after expression and 

* Corresponding author. Fax: +45 86246850. 

E-mail address: rhl@boreanpharma.com (R.H. Lorentsen). 

puriWcation to get the target protein in authentic form. 

Presently, this processing is accomplished by designing 

fusion proteins with cleavage sites that allow speciWc 

enzymatic or chemical cleavage. Highly speciWc proteases 

isolated from mammalian sources, such as bovine blood 

coagulation factor X a (FX a ), are frequently used for 

enzymatic cleavage of fusion proteins. They are particularly 

suitable for that purpose, as they recognize a 

sequence of amino acid residues instead of only one residue, 

which decreases the probability of non-speciWc 

cleavage in a site not intended. However, isolation of 

these proteases is costly, and the risk of pathogenic contamination 

when cleaving proteins for therapeutic applications 

prohibits the use of these processing enzymes in 

recombinant production of pharmaceutical proteins. The 

1046-5928/$ - see front matter © 2004 Elsevier Inc. All rights reserved. 

doi:10.1016/j.pep.2004.08.017

R.H. Lorentsen et al. / Protein Expression and PuriWcation 39 (2005) 18–26 19 

objective of the investigations presented in this paper is 

to generate a protease, which can be produced on a large 

scale in a bacterial host and with advantage, be used for 

speciWc and eYcient processing of fusion proteins. 

Cytoplasmic granules of cytotoxic T-lymphocytes 

contain a family of serine proteases termed granzymes. 

In combination with the pore-forming protein, perforin, 

they play critical roles in events that culminate in the 

death of virus-infected and other potentially harmful 

cells [1]. Granzyme B (GrB) is one member of this protease 

family and is believed to induce apoptosis of abnormal 

cells by cleavage and activation of caspases, 

including caspase-3 and -8. In addition, GrB can initiate 

caspase-independent pathways of cell death by cleaving 

various other cellular substrates [2]. The inactive zymogen 

form of GrB is equipped with an N-terminal activation 

dipeptide that is cleaved oV by dipeptidyl peptidase 

I (DPPI) at the time of packaging into cytotoxic granules, 

providing storage of the single-chain active serine 

protease [1,3]. GrB exhibits the unique speciWcity among 

mammalian serine proteases to cleave substrate peptide 

bonds after aspartic acid residues and requires extended 

enzyme–substrate binding interactions for eYcient catalysis, 

consistent with its role as a regulative, rather than a 

degradative, enzyme [1,4]. By a combinatorial approach 

the tetrapeptide sequence IEPD has been identiWed as 

the preferred P4–P1 recognition motif of GrB for small 

peptide substrates, although optimal substrate recognition 

may involve features beyond this tetrapeptide 

sequence [5,6]. The crystal structure of rat and human 

GrB was recently determined and provides some rationale 

for the substrate speciWcity [7–9]. The three-dimensional 

structure is generally similar to that of the trypsin-family 

serine proteases, although profound diVerences in the S1 

site account for the strict requirement of GrB for aspartic 

acid in the P1 position. The high speciWcity by which 

GrB apparently cleaves a limited number of substrates 

suggests that it would be suitable as a processing 

enzyme, provided it could be produced in high yields by 

recombinant methods in prokaryotes or other microbiological 

host systems. Recombinant granzymes, including 

granzyme A, B, and K, have been produced in various 

prokaryotic and eukaryotic host systems as active 

enzymes or inactive zymogens for the purpose of structural, 

functional, and biochemical characterization 

[6,10–17]. Recently, production of recombinant human 

granzyme K from Escherichia coli inclusion bodies by 

subsequent refolding and activation was reported [18]. 

In this paper, we describe the production of recombinant 

human GrB, both as a derivative dependent on FX a 

cleavage for activation and a self-activating derivative. 

The activated recombinant enzyme can be produced in 

high yield, and is shown to cleave fusion proteins containing 

the IEPD recognition sequence with a high degree 

of speciWcity and eYciency, making it an ideal reagent for 

the processing of recombinant pharmaceutical products. 

Methods and materials 

Construction of the pT7-rGrB-H6, pT7-IEPD-rGrB-H6, 

and position 228 variants expression vectors 

The pT7-C-term-H6 vector was prepared by ligation 

of the annealed oligonucleotides 5-CATGGACGG 

AAGCTTGAATTCACATCACCATCACCATCACTA 

ACGC-3 and 5-AATTGCGTTAGTGATGGTGAT 

GGTGATGTGAATTCAAGCTTCCGTC-3 into the 

NcoI and EcoRI cloning site of the pT7 expression vector 

[19]. The DNA sequence encoding the mature human 

GrB (EC-nr.: 3.4.21.79) from Ile21 (Ile16 in chymotrypsinogen 

numbering) to Tyr247 was ampliWed by PCR 

from a mixture of cDNA isolated from human bone 

marrow, human leukocytes, human lymph nodes, and 

lymphoma (Raji) cells (Clontech Laboratories) using the 

following primers: 5CATGGGATCCATCGAGGGT 

AGGATCATCGGGGGACATGAG-3 and 5-GCGT 

GAATTCAGGTACCGTTTCATGGTTTTCTTTATC 

C-3. The PCR product was digested with BamHI and 

EcoRI, and ligated into the corresponding sites of the 

pT7-C-term-H6 expression vector, resulting in the pT7- 

rGrB-H6 expression vector. The pT7-IEPD-rGrB-H6 

expression vector was constructed by using the Quik- 

Change Site-directed Mutagenesis Kit (Stratagene) with 

pT7-rGrB-H6 as template and the following mutagenesis 

primers: 5-TCCATCGAGCCGGATATCATC 

GGGGGACATGAG-3 and 5-CCCCGATGATAT 

CCGGCTCGATGGATCCCATATG-3. Substitution of 

cysteine 228 (chymotrypsinogen numbering) in pT7- 

IEPD-rGrB-H6 with alanine, serine, threonine, valine, or 

phenylalanine was also performed by using the sitedirected 

mutagenesis kit and the mutagenesis primers 

5-TCCACGAGCCDCCACCAAAGTCTCAAG-3 

(DDA, G, or T) and 5-AGACTTTGGTGGHGGCTC 

GTGGAGGC-3 (H D A, C, or T) to create the pT7- 

IEPD-rGrB-H6-C228A, -C228S, and -C228T, respectively, 

or 5-TCCACGAGCCKTCACCAAAGTCTCA 

AG-3 (K D G or T) and 5-AGACTTTGGTGAMG 

GCTCGTGGAGGC-3 (M D A or C) to create pT7- 

IEPD-rGrB-H6-C228V and -C228F, respectively. All 

primers and oligonucleotides were purchased from 

DNA Technology A/S. All sequences and mutations 

were veriWed by nucleotide sequencing using the BigDye 

Terminator version 3.0 DNA sequencing Kit (Applied 

Biosystems). 

Expression and puriWcation of recombinant pro-rGrB-H6 

Single colonies of E. coli BL21 cells transformed with 

the pT7-rGrB-H6 expression vector were used to inoculate 

100 mL of 2£ TY medium containing 100 mg/L 

ampicillin. After overnight growth, 30 mL of this culture 

was transferred to 1 L of 2£ TY medium containing 

100 mg/L ampicillin and grown at 37 °C with shaking

20 R.H. Lorentsen et al. / Protein Expression and PuriWcation 39 (2005) 18–26 

until OD 600 D 0.7–0.8. Expression was induced by infection 

with bacteriophage λCE6 at a multiplicity of 

approximately 5, and incubation was continued for 

another 4 h before the cells were harvested by centrifugation. 

The cell pellet from 3 L of culture was resuspended 

in 100 mL of lysis buVer containing 50 mM Tris–HCl, 

pH 8.0, 25% (w/v) sucrose, 1 mM EDTA before addition 

of lysozyme (Sigma) to a Wnal concentration of 1 mg/mL 

and incubation at room temperature for 15 min. 200 mL 

detergent buVer containing 0.2 M NaCl, 1% (w/v) deoxycholic 

acid, 1% (w/v) nonidet P-40, 20 mM Tris–HCl, pH 

7.5, 2 mM EDTA was added and the preparation was 

gently homogenized before the inclusion bodies were 

collected by centrifugation. The inclusion bodies were 

washed at least three times with 50 mL of a buVer containing 

0.5% (w/v) Triton X-100, 1 mM EDTA before 

they were solubilized in a buVer containing 6 M guanidinium·HCl, 

50 mM Tris–HCl, pH 8.0, 100 mM dithiothreitol. 

The crude protein preparation was gel Wltrated into 

a buVer containing 8 M urea, 0.5 M NaCl, 50 mM Tris– 

HCl, pH 8.0, 5 mM 2-mercaptoethanol before it was 

batch-mixed with Ni 2+ –NTA agarose (Qiagen) for 5 min 

and packed into a column (2.6 £ 10 cm, Amersham Biosciences). 

The pro-rGrB-H6 fusion protein was puriWed 

from the majority of E. coli and λ-phage proteins by 

washing with one column volume of loading buVer containing 

8 M urea, 0.5 M NaCl, 50 mM Tris–HCl, pH 8.0, 

5 mM 2-mercaptoethanol followed by one-column volume 

of a buVer containing 6 M guanidinium·HCl, 

50 mM Tris–HCl, pH 8.0, 5 mM 2-mercaptoethanol and, 

Wnally, loading buVer until a stable baseline was reached. 

Bound pro-rGrB-H6 was either eluted in loading buVer 

supplemented with 10 mM EDTA to check the purity by 

SDS–PAGE or applied to refolding after washing with a 

buVer containing 8 M urea, 0.5 M NaCl, 50 mM Tris– 

HCl, pH 8.0, 3 mM reduced glutathione. 

Refolding and Wnal puriWcation of pro-rGrB-H6 

Over a 24-h period the column with bound pro-rGrB- 

H6 was applied to a cyclic refolding protocol described 

in [20]. In brief, each cycle started with 45 min of renaturation, 

followed by a step of denaturing and reducing 

conditions for 6 min, followed by a linear gradient of 

8 min from denaturation to renaturation and the start of 

the next cycle with 45 min of renaturation. The buVer 

applied for renaturation contained 0.5 M NaCl, 50 mM 

Tris–HCl, pH 8.0, 2 mM reduced glutathione, 0.2 mM 

oxidized glutathione, while the buVer applied in the 

denaturing phases was mixed from the renaturation 

buVer and a denaturation buVer containing 6 M urea, 

0.5 M NaCl, 50 mM Tris–HCl, pH 8.0, 3 mM reduced 

glutathione. The percentage of denaturation buVer was 

decreased in steps from 100% in cycle one to 35% in 

cycle 23 before the procedure was completed with 45 min 

of renaturation. The refolding was performed at a Xow 

rate of 2 mL/min and at 10 °C. After completion of the 

refolding protocol, the column was washed with twocolumn 

volumes of buVer containing 0.5 M NaCl, 

50 mM Tris–HCl, pH 8.0, before the refolded pro-rGrB- 

H6 was eluted in that buVer supplemented with 10 mM 

EDTA. The pooled sample was diluted with one volume 

of 50 mM Tris–HCl, pH 7.0, and the pH was adjusted to 

pH 7.0. by dropwise addition of 0.5 M HCl before the 

protein solution was loaded onto a SP Sepharose ionexchange 

column (1 £ 10 cm, Amersham Biosciences) 

pre-equilibrated in running buVer containing 250 mM 

NaCl, 50 mM Tris–HCl, pH 7.0. The protein was 

fractionated by elution with a linear gradient from 

250 mM to 1 M NaCl in 50 mM Tris–HCl, pH 7.0 at 

room temperature. Samples from the elution proWle were 

analyzed by non-reducing SDS–PAGE and the enzymatic 

activity was measured as described below. The 

peak fractions were pooled and the concentration of 

pro-rGrB-H6 was estimated by absorption at 280 nm 

and the assumption that A(1%) 280 for pro-rGrB-H6 is 

1g L ¡1 cm ¡1 . 

Activation of pro-rGrB-H6 

A sample of the pooled protein was directly used for 

activation processing by FX a . Approximately 0.15 mg of 

pro-rGrB-H6 in 1.5 mL was activated by the addition of 

1.5 μg of native bovine FX a (Borean Pharma A/S) and 

incubation at room temperature. Five-microliter aliquots 

were withdrawn after 0, 1, 2, 5, 24, 48, 75, and 

120 h of incubation and immediately used to measure the 

enzymatic activity as described below. The contribution 

from the FX a to the hydrolysis rate of the Ac-IEPDpNA 

substrate was measured and found to be zero. 

After 120 h of incubation, FX a was removed by dilution 

of the protein sample with two volumes of 50 mM 

Tris–HCl, pH 7.0, and loading onto an SP Sepharose 

ion-exchange column (1 £ 5 cm, Amersham Biosciences) 

pre-equilibrated in running buVer containing 250 mM 

NaCl, 50 mM Tris–HCl, pH 7.0. The column was 

intensively washed with the running buVer to remove 

the FX a before bound rGrB-H6 was eluted in one step 

with buVer containing 750 mM NaCl, 50 mM Tris–HCl, 

pH 7.0. 

Activity assay 

Enzymatic activity was measured as the initial rate of 

hydrolysis of the Ac-IEPD-pNA (Calbiochem) substrate 

by monitoring the increase of A 405 nm over time using a 

Varian Cary 50 Bio UV–Visible spectrophotometer. 

Five-microliter aliquots of enzyme were added to reaction 

mixture containing 400 μM substrate, 0.1 M Hepes, 

pH 7.75, in a total volume of 500 μL. A 100 mM substrate 

stock solution was prepared in 99.7% DMSO and 

stored at ¡20 °C.


Preparation of self-activating rGrB-H6 derivative 

Conditions of expression, refolding, and puriWcation 

of the pro(IEPD)-rGrB-H6 variant were identical to 

those described for the pro-rGrB-H6 enzyme. The enzymatic 

activity of a sample of the pooled refolded protein, 

withdrawn before the protein was loaded onto the SP 

Sepharose ion-exchange column and of the pooled protein 

after elution from the SP Sepharose ion-exchange 

column was measured as described above. The preparation 

procedure was repeated, and for complete autocatalytic 

activation, the ion-exchange column was washed 

with 250 mM NaCl, 50 mM Tris–HCl, pH 7.0, for 4 h, 

before the protein was eluted following the procedure 

described for pro-rGrB-H6. 

Preparation of rGrB-H6 variants in position 228 

Conditions of expression, refolding, and puriWcation 

of the pro(IEPD)-rGrB-H6-C228A, -C228S, -C228T, 

-C228V, and -C228F variants were identical to those 

described for the pro-rGrB-H6 enzyme. Refolding 

eYciency was estimated by comparison of the Wnal 

recovery. Activity assays of the mutants were performed 

on equivalent amounts of each mutant, determined by 

light absorption at 280 nm and by Bradford assay (Coomassie 

Plus Protein Assay Reagent Kit, Pierce Biotechnology) 

using bovine serum albumin as protein 

standard. Assays were performed at room temperature 

in 96-well microtiter plates with a reaction volume of 

250 μL. The Wnal substrate concentration was 400 μM 

and the Wnal concentration of each mutant was 150 nM 

in assay buVer containing 100 mM Hepes, pH 7.75. 

Hydrolysis rates of the Ac-IEPD-pNA substrate (from 

Calbiochem) were measured by monitoring the increase 

of A 405 nm over time using a Multiscan Ascent microplate 

reader (Thermo Labsystems). Experiments were performed 

four times, and the average of the determinations 

with standard errors is reported. 

Substrate kinetics 

Protein concentrations of enzyme stock solutions 

were determined by Bradford assay (Coomassie Plus 

Protein Assay Reagent Kit, Pierce Biotechnology) using 

bovine serum albumin as protein standard. Assays were 

carried out at room temperature in a total volume of 

500 μL using a Varian Cary 50 Bio UV–Visible spectrophotometer. 

The Wnal concentration of substrate ranged 

from 5 to 600 μM and the Wnal concentration of each 

enzyme was 20 nM in assay buVer containing 100 mM 

Hepes, pH 7.75, for the Ac-IEPD-pNA substrate and 

150 nM for the Ac-LEED-pNA and Ac-VEID-pNA substrates 

(all from Calbiochem). Initial rates were measured 

in duplicate for each substrate concentration and 

were averaged in each case. The kinetic constants were 

obtained from Lineweaver–Burk plots which had correlation 

coeYcients greater than 0.99. 

Cleavage of fusion proteins 

The pT7H6GrB-rTN123 expression vector was 

constructed by using the QuikChange Site-directed 

Mutagenesis Kit (Stratagene) with the primers 5-GGA 

TCCATCGAGCCTGACGGCGAGCCACCAACC-3 

and 5-GGCTCGCCGTCAGGCTCGATGGATCCG 

TGATGG-3 and pT7H6FX-rTN123, which has been 

described in [21], as template. Expression in E. coli, 

refolding, and puriWcation of H6GrB-rTN123 were performed 

as described previously for H6FX-rTN123 [21]. 

Approximately 50 μg of H6GrB-rTN123 in 200 μL of a 

buVer containing 100 mM Hepes, pH 7.4, was cleaved by 

the addition of 0.2 μg of rGrB-H6 wt and incubated at 

room temperature for 12 h. A similar cleavage in which 

the cleavage mixture was supplemented with 5 mM 

CaCl 2 was also performed. Samples were withdrawn and 

analyzed by non-reducing and reducing SDS–PAGE. 

H6UbiGrB-ApoAI was a kind gift from Jonas Graversen 

(Borean Pharma A/S) [22]. Approximately 450 mg 

of H6UbiGrB-ApoAI in a total volume of 500 mL of a 

buVer containing 25 mM NaCl, 25 mM Tris–HCl, pH 

8.0, was cleaved by the addition of 0.5 mg of rGrB-H6- 

C228F and incubated at room temperature overnight. 

Samples were withdrawn before and after cleavage and 

analyzed by non-reducing SDS–PAGE. 

The pT7H6UbiGrB-BPFI-0101 expression vector 

was constructed by subcloning a DNA fragment encoding 

a GrB recognition sequence immediately followed by 

the so-called HIV gp 41 HR2 region from HIV-1 strain 

NL4-3 (residues 625–672 of the env polyprotein) and the 

trimerization domain of human tetranectin (residues 1– 

54) into the pT7H6Ubi vector [23,24]. The H6UbiGrB- 

BPFI-0101 fusion protein was expressed in E. coli BL21 

cells following the protocol described for pro-rGrB-H6. 

Total protein was isolated by phenol extraction followed 

by ethanol precipitation and solubilization in 6 M guanidinium·HCl, 

50 mM Tris–HCl, pH 8.0, 5 mM 2-mercaptoethanol. 

The crude protein preparation was gel 

Wltrated into a buVer containing 8 M urea, 0.5 M NaCl, 

50 mM Tris–HCl, pH 8.0, 5 mM 2-mercaptoethanol, 

before puriWcation by Ni 2+ –NTA agarose (Qiagen) aYnity 

chromatography. After intensive washing of the column, 

bound H6UbiGrB-BPFI-0101 fusion protein was 

eluted in a buVer containing 8 M urea, 0.5 M NaCl, 

50 mM Tris–HCl, pH 8.0, 10 mM EDTA and gel Wltrated 

into 25 mM NaCl, 10 mM Tris–HCl, pH 7.5, prior to 

preparative cleavage. Approximately 500 mg of H6Ubi- 

GrB-BPFI-0101 in a total volume of 130 mL was cleaved 

by the addition of 0.25 mg of rGrB-H6-C228F and incubated 

at room temperature for 4 h. Samples were withdrawn 

before and after cleavage and analyzed by 

reducing SDS–PAGE.


Results and discussion 

Expression, refolding, and puriWcation of recombinant 

pro-rGrB-H6 

The pT7-rGrB-H6 expression vector was designed to 

express human GrB as a fusion protein with a fouramino 

acid recognition site for FX a immediately N-terminal 

to the Ile-1 (Ile-16 in chymotrypsinogen numbering) 

residue of the mature enzyme and a hexa-histidine (H6) 

aYnity tag fused to the C-terminus to facilitate puriWcation. 

Expression from this vector provides an inactive 

pro-rGrB-H6 fusion protein, which subsequently can be 

activated by FX a processing (Fig. 1). A high level of 

expression was achieved in the BL21 E. coli host in the 

form of insoluble inclusion bodies. Initially, the prorGrB-H6 

fusion protein was puriWed from cellular proteins 

by Ni 2+ –NTA agarose aYnity chromatography to 

yield highly pure pro-rGrB-H6 fusion protein (Fig. 2) 

with a molecular weight corresponding to the calculated 

molecular weight of 27.4 kDa for the non-activated 

recombinant fusion protein. Pro-rGrB-H6 fusion protein 

was directly applied to a column-based cyclic refolding 

procedure described in the Materials and methods 

section [20]. After completion of the refolding procedure, 

the pro-rGrB-H6 fusion protein was eluted from the column, 

and the correctly folded protein was separated 

from incorrectly folded derivatives by cation-exchange 

chromatography. The monomeric protein was eluted as 

one major peak in the A 280 elution proWle. Samples analyzed 

by non-reducing SDS–PAGE appear as a single 

distinct band (Fig. 2). The Wnal yield of puriWed 

monomeric pro-rGrB-H6 was approximately 4% after 

refolding. 

Activation 

The puriWed refolded pro-rGrB-H6 was activated by 

FX a processing. The progress of activation was monitored 

by assaying the enzymatic activity of samples collected 

at diVerent times (Fig. 3). After 120 h of 

incubation, the activation was found to be almost 

complete, and FX a was removed by cation-exchange 

chromatography. Apparently, the refolded monomeric 

pro-rGrB-H6 could be quantitatively converted into the 

active rGrB-H6 enzyme with a Wnal yield of approximately 

1 mg/L of culture. 

Preparation of the self-activating rGrB-H6 derivative 

Fig. 1. Schematic representation of the two GrB fusion protein constructs. 

Seven residues fused to the N-terminus of the amino acid 

sequence of mature granzyme B constitute the pro sequence, which is 

removed upon activation. The hexa-histidine (H6) aYnity tag fused to 

the C-terminal facilitates puriWcation. Pro-rGrB-H6 is dependent on 

FX a processing for activation, whereas pro(IEPD)-rGrB-H6 is engineered 

to be self-activating. 

A self-activating GrB derivative was generated by 

mutation of the FX a recognition sequence to the recognition 

sequence of GrB itself. The resulting pT7-IEPDrGrB-H6 

vector provides expression of the inactive 

pro(IEPD)-rGrB-H6 fusion protein of which activation 

is not dependent on the addition of an external activator 

(Fig. 1). Expression and refolding eYciency of the 

pro(IEPD)-rGrB-H6 derivative was similar to the pro- 

Fig. 2. SDS–PAGE analysis of pro-rGrB-H6 fusion protein. Lane M, 

molecular weight markers; lane 1, pro-rGrB-H6 fusion protein puri- 

Wed from solubilized inclusion bodies by Ni 2+ –NTA agarose aYnity 

chromatography; lane 2; pro-rGrB-H6 fusion protein after refolding 

and cation-exchange chromatography. 

Fig. 3. Progress curve for activation of pro-rGrB-H6 by blood coagulation 

factor X a (FX a ) processing. Samples were withdrawn at the indicated 

times, and the enzymatic activity was measured as the hydrolysis 

rate of the Ac-IEPD-pNA substrate by monitoring the increase of 

A 405 nm over time as described in the Methods and materials section.


rGrB-H6 enzyme. No detectable enzymatic activity was 

measured after completion of the refolding. However, 

after puriWcation by ion exchange chromatography, the 

protein was found to be enzymatically active, suggesting 

that activation occurs while the refolded protein is concentrated 

by immobilization on the ion-exchange column. 

The pro-rGrB-H6 with the FX a activation 

sequence had no detectable enzymatic activity at this 

stage in the puriWcation procedure, clearly indicating 

that the observed activity of the self-activating derivative 

is a result of autocatalytic activation. The preparation 

procedure was repeated and 4 h of incubation at room 

temperature on the ion-exchange column, was found to 

accomplish complete autocatalytic activation (data not 

shown). 

Preparation of rGrB-H6 variants in position 228 

Human GrB contains an unpaired cysteine in position 

228 in contrast to phenylalanine in the corresponding 

position of mouse and rat GrB [8,9]. The diVerence 

results in an extended S1 subsite of the human enzyme, 

occupied by three water molecules in the crystal structure 

of the complex between human GrB and the Ac- 

IEPD-CHO inhibitor [8]. In order to eliminate possible 

Fig. 4. Enzymatic activities of diVerent rGrB variants in position 228. 

Enzymatic activity of each of the diVerent variants was measured as 

the hydrolysis rate of the Ac-IEPD-pNA substrate by monitoring the 

increase of A 405 nm over time as described in Methods and materials. 

Hydrolysis rates were normalized to protein concentration and standard 

error bars of four experiments are shown. 

complications caused by the unpaired cysteine, mutants 

were generated in which this residue was substituted 

with alanine, serine, threonine, valine, or phenylalanine 

by site-directed mutation of the self-activating construct, 

and properties of refolding and enzymatic activity were 

evaluated for each mutant. Expression levels of the single-residue 

mutants were similar to the pro-rGrB-H6 

level. However, the refolding eYciency diVered by up to 

90% relative to that of the wild type, and one mutant 

(C228S) was not analyzed further because of a very low 

refolding eYciency. The hydrolysis rate of the Ac-IEPDpNA 

substrate was measured for the other mutants and 

compared to the wild type as shown in Fig. 4. The 

C228A and C228F mutants were found to hydrolyze the 

Ac-IEPD-pNA substrate at essentially the same rate as 

the rGrB-H6, whereas the C228T and C228V mutants 

possessed slightly lower enzymatic activities. Furthermore, 

the C228F mutant was found to refold slightly 

better than the pro(IEPD)-rGrB-H6 and this mutant 

was further investigated to evaluate its suitability as a 

processing enzyme. 

Substrate kinetics 

To quantify the catalytic eYciency of our recombinant 

GrBs and evaluate possible speciWcity changes 

caused by the mutation of cysteine to phenylalanine, we 

determined kinetic (Michaelis–Menten) constants for 

hydrolysis of the Ac-IEPD-pNA, Ac-LEED-pNA, and 

Ac-VEID-pNA substrates by rGrB-H6 and rGrB-H6- 

C228F with protein concentrations determined by Bradford 

assay. The kinetic values are summarized in Table 1. 

The values for hydrolysis of the Ac-IEPD-pNA substrate 

were found to be within the range of those 

reported for recombinant rat GrB [6]. Apparently, substitution 

of Cys228 with Phe results in a decreased K m 

value for this substrate; however, as long as the k cat value 

is not inXuenced, the eVect on substrate-binding speciWcity 

may not be detrimental for rGrB-H6-C228F to be 

used as a processing enzyme. This is also conWrmed from 

the kinetic values for hydrolysis of the Ac-LEED-pNA 

and Ac-VEID-pNA substrates. Although the two variants 

show diVerences in both k cat and K m values, rGrB- 

H6-C228F has only slightly higher substrate speciWcities. 

Table 1 

Kinetic constants for hydrolysis of peptide-pNA substrates with diVerent tetrapeptide sequences by recombinant GrB 

Enzyme Substrate k cat (s ¡1 ) K m (10 ¡6 M) k cat /K m (10 4 M ¡1 s ¡1 ) 

rGrB-H6 IEPD 5.0 § 0.45 67 § 6 7.5§ 0.95 

LEED 0.15 § 0.086 117 § 69 0.13 § 0.104 

VEID 0.30 § 0.112 88 § 32 0.34 § 0.180 

rGrB-H6 C228F IEPD 4.9 § 0.26 27 § 2 18.0 § 1.44 

LEED 0.22 § 0.0.14 82 § 53 0.27 § 0.240 

VEID 0.72 § 0.150 170 § 35 0.42 § 0.125 

rat rGrB a IEPD 4.2 § 0.05 57 § 4 6.7§ 0.32 

a Taken from [6].


Cleavage of fusion proteins 

Functional properties of both rGrB-H6 and rGrB- 

H6-C228F for cleavage of diVerent fusion proteins, all 

containing a GrB recognition sequence, were measured 

to evaluate whether GrB is suitable to be used as a processing 

enzyme. The H6GrB-rTN123 fusion protein, 

which should be a substrate for GrB, was generated by 

substitution of the FX a recognition sequence in the 

H6FX-rTN123 fusion protein with the IEPD recognition 

sequence of GrB (Fig. 5A) [21]. Additionally, an 

AQPD (residues 142–145) sequence, which could represent 

an internal GrB cleavage site, was identiWed in the 

sequence of tetranectin. The sequence is located in a loop 

structure and includes the residues Q143 and D145, 

which are directly involved as ligands for calcium-binding 

in tetranectin [23]. Binding of Ca 2+ to the calciumbinding 

sites of tetranectin, results in a rigid loop 

structure that probably makes the AQPD sequence inaccessible 

to GrB cleavage, whereas the absence of calcium 

will loosen the structure and make the AQPD sequence 

accessible for GrB cleavage. Cleavage after the AQPD 

sequence will result in a two-chain disulWde-bonded 

structure, probably possessing decreased mobility in a 

non-reducing SDS–PAGE gel. Cleavage of the H6GrBrTN123 

fusion protein was performed both with and 

without CaCl 2 present in the cleavage mixture, and the 

diVerent cleavage patterns were evaluated by SDS– 

PAGE analysis (Fig. 6). Initially, the fusion protein was 

present as a single band. After partial processing with 

rGrB-H6 in the absence of CaCl 2 , the fusion protein was 

cleaved to give three products (Fig. 6A), whereas the 

processing performed in the presence of CaCl 2 generated 

Fig. 5. Schematic representation of the diVerent fusion proteins 

designed to be substrates for GrB. (A) H6GrB-rTN123 consists of an 

N-terminal hexa-histidine (H6) aYnity tag followed by the IEPD recognition 

sequence of GrB preceding the sequence of human tetranectin. 

(B) H6UbiGrB-ApoAI consists of an N-terminal H6 aYnity tag 

followed by the sequence of human ubiquitin (Ubi) and the IEPD recognition 

sequence of GrB immediately preceding the apolipoprotein 

AI (ApoAI) sequence. (C) H6UbiGrB-BPFI-0101 consists of an N-terminal 

H6 aYnity tag followed by the sequence of human ubiquitin 

(Ubi) and the IEPD recognition sequence of GrB immediately preceding 

the HR2 region from HIV-1 and the trimerization domain of 

human tetranectin (HR2-Trip). 

Fig. 6. SDS–PAGE analysis of the cleavage of the H6GrB-rTN123 

fusion protein by rGrB-H6. (A) Non-reducing SDS–PAGE; lane M, 

molecular weight markers; lane 1, non-cleaved H6GrB-rTN123; lane 

2, H6GrB-rTN123 cleaved in a buVer without CaCl 2 . (B) Non-reducing 

SDS–PAGE; lane 1, non-cleaved H6GrB-rTN123; lane 2, H6GrBrTN123 

cleaved in the presence of CaCl 2 ; lane M, molecular weight 

markers; arrows indicate from above the intact fusion protein and 

rTN123. (C) Reducing SDS–PAGE; lane M, molecular weight markers; 

lane 1, non-cleaved H6GrB-rTN123; lane 2, H6GrB-rTN123 

cleaved in a buVer without CaCl 2 ; arrows indicate, from above, the 

intact fusion protein, rTN123, H6GrB-rTN123 liberated from the 

»4 kDa C-terminal cleavage product, and rTN123 liberated from the 

»4 kDa C-terminal cleavage product. 

only a single product (Fig. 6B). The upper band of each 

of the two double bands obtained from the cleavage in 

the absence of CaCl 2 (Fig. 6A) represents H6GrBrTN123 

and rTN123, respectively, cleaved after the 

internal AQPD sequence, conWrmed by reducing SDS– 

PAGE analysis of the samples (Fig. 6C), where the three 

products correspond to rTN123, H6GrB-rTN123 liberated 

from the » 4 kDa C-terminal cleavage product, and 

rTN123 liberated from the »4 kDa C-terminal cleavage 

product, respectively. The single product obtained from 

the cleavage in the presence of CaCl 2 corresponds to 

rTN123 from which only the H6 fusion partner has been 

cleaved oV. The same cleavage patterns were observed 

when using the rGrB-H6-C228F in a similar experiment 

(data not shown). These results clearly demonstrate that 

GrB cleaves fusion protein substrates after Asp residues 

preceded by the sequence identiWed as the recognition 

motif of GrB, but when the motif is located in a rigid 

structure, as is the case when the calcium-binding sites of 

tetranectin are occupied, the cleavage does not occur. 

The eYciency of cleavage by GrB was evaluated for 

processing a relatively large amount of the H6UbiGrB- 

ApoAI fusion protein. This is apolipoprotein AI 

(ApoAI) designed with an N-terminal fusion partner 

consisting of a H6 aYnity tag followed by the sequence 

of human ubiquitin (Ubi) and the IEPD recognition 

sequence of GrB immediately preceding the ApoAI 

sequence (Fig. 5B) [22,24]. Approximately 450 mg of


H6UbiGrB-ApoAI in a total volume of 500 mL was 

cleaved with rGrB-H6-C228F at an enzyme to substrate 

ratio of approximately 1 : 1000. After overnight incubation, 

the cleavage performance was evaluated by SDS– 

PAGE analysis (Fig. 7), and the fusion protein was 

found to be entirely cleaved into a single product corresponding 

to ApoAI. The fusion partner also appears in 

the SDS–PAGE gel, but no non-speciWc cleavage products 

are observed. The eYciency of cleavage by GrB was 

further evaluated from a large-scale processing of the 

H6UbiGrB-BPFI-0101 fusion protein, which consists of 

the H6UbiGrB fusion partner described above followed 

by the HR2 region from HIV-1 and the trimerization 

domain of human tetranectin (Fig. 5C) [23]. Approximately 

500 mg of H6UbiGrB-BPFI-0101 in a total volume 

of 130 mL was cleaved with rGrB-H6-C228F at an 

enzyme to substrate ratio of approximately 1:2000. After 

4 h of incubation, the cleavage performance was evaluated 

by reducing SDS–PAGE analysis (Fig. 8), and the 

cleavage was found to be nearly complete. Both the 

HR2-Trip product and the fusion partner appear in the 

gel without any non-speciWc cleavage products. 

Overall, both rGrB-H6 and rGrB-H6-C228F were 

found to cleave the fusion proteins with a high degree of 

speciWcity and eYciency. The presence of the short C-terminal 

sequence containing the H6 aYnity tag does not 

seem to interfere with the activity of our recombinant 

enzymes when compared to recombinant rat GrB. The 

unpaired cysteine at position 228 of human GrB was 

changed to phenylalanine and, apparently, this substitution 

only caused a slight diVerence in the tetrapeptide 

substrate selectivity. The recombinant protease cleaves 

the synthetic tetrapeptide substrate Ac-IEPD-pNA with 

high catalytic eYciency and was found to cleave fusion 

Fig. 8. SDS–PAGE analysis of the cleavage of the H6UbiGrB-BPFI- 

0101 fusion protein by rGrB-H6-C228F. Lane M, molecular weight 

markers; lane 1, non-cleaved H6UbiGrB-BPFI-0101; lane 2, after 4 h 

incubation with rGrB-H6-C228F in a 1:2000 ratio of enzyme to substrate; 

arrows indicate, from above, the intact fusion protein, HR2- 

Trip, and the fusion partner. 

proteins containing the IEPD recognition sequence of 

GrB very fast and speciWcally at the anticipated site. No 

non-speciWc cleavage of the investigated fusion protein 

by the C228F mutant was observed. Previously reported 

studies indicate that GrB also has a preference for certain 

residues C-terminal to the scissile bond [6,25]. 

Although these residues may have an impact on substrate 

speciWcity and catalytic eYciency for cleavage of 

protein substrates, we did not observe any problems in 

cleaving fusion proteins comprising prime-side 

sequences reported to be non-optimal, but restraints on 

these residues will be further investigated. 

In summary, the described procedure yields high 

amounts of a pure, active protease with high speciWc 

activity, good stability, and enzymatic properties that 

make it suitable for use in pharmacological production. 

The complete puriWcation procedure includes in vitro 

refolding of protein obtained from inclusion bodies followed 

by activation during the Wnal puriWcation step. 

About 0.5–1 mg of the refolded, activated protease could 

be obtained from 1L of culture, and the refolding step 

was found to provide the limit on the recovery level. 

However, we believe that the refolding eYciency can be 

considerably improved by rational optimization. Further 

characterization of stability and functional properties of 

the recombinant enzyme will be reported elsewhere. 

References 

Fig. 7. SDS–PAGE analysis of the cleavage of the H6UbiGrB-ApoAI 

fusion protein by rGrB-H6-C228F. Lane M, molecular weight markers; 

lane 1, non-cleaved H6UbiGrB-ApoAI; lane 2, after overnight 

incubation with rGrB-H6-C228F in a 1:1000 ratio of enzyme to substrate; 

arrows indicate, from above, the intact fusion protein, Apo-AI, 

and the fusion partner. 

[1] J.A. Trapani, Granzymes: a family of lymphocyte granule serine 

proteases, Genome Biol. 2 (2001) 3014.3011-3014.3017 (reviews). 

[2] F. Andrade, L.A. Casciola-Rosen, A. Rosen, Granzyme B-induced 

cell death, Acta Haematol. 111 (2004) 28–41. 

[3] C.T. Pham, T.J. Ley, Dipeptidyl peptidase I is required for the processing 

and activation of granzymes A and B in vivo, Proc. Natl. 

Acad. Sci. USA 96 (1999) 8627–8632. 

[4] M. Poe, J.T. Blake, D.A. Boulton, M. Gammon, N.H. Sigal, J.K. 

Wu, H.J. Zweerink, Human cytotoxic lymphocyte granzyme B. Its 

puriWcation from granules and the characterization of substrate 

and inhibitor speciWcity, J. Biol. Chem. 266 (1991) 98–103.


[5] N.A. Thornberry, T.A. Rano, E.P. Peterson, D.M. Rasper, T. Timkey, 

M. Garcia-Calvo, V.M. Houtzager, P.A. Nordstrom, S. Roy, 

J.P. Vaillancourt, K.T. Chapman, D.W. Nicholson, A combinatorial 

approach deWnes speciWcities of members of the caspase family 

and granzyme B. Functional relationships established for key 

mediators of apoptosis, J. Biol. Chem. 272 (1997) 17907–17911. 

[6] J.L. Harris, E.P. Peterson, D. Hudig, N.A. Thornberry, C.S. Craik, 

DeWnition and redesign of the extended substrate speciWcity of 

granzyme B, J. Biol. Chem. 273 (1998) 27364–27373. 

[7] S.M. Waugh, J.L. Harris, R. Fletterick, C.S. Craik, The structure 

of the pro-apoptotic protease granzyme B reveals the molecular 

determinants of its speciWcity, Nat. Struct. Biol. 7 (2000) 762–765. 

[8] J. Rotonda, M. Garcia-Calvo, H.G. Bull, W.M. Geissler, B.M. 

McKeever, C.A. Willoughby, N.A. Thornberry, J.W. Becker, The 

three-dimensional structure of human granzyme B compared to 

caspase-3, key mediators of cell death with cleavage speciWcity for 

aspartic acid in P1, Chem. Biol. 8 (2001) 357–368. 

[9] E. Estebanez-Perpina, P. Fuentes-Prior, D. Belorgey, M. Braun, R. 

Kiefersauer, K. Maskos, R. Huber, H. Rubin, W. Bode, Crystal 

structure of the caspase activator human granzyme B, a proteinase 

highly speciWc for an Asp-P1 residue, Biol. Chem. 381 (2000) 

1203–1214. 

[10] J.A. Kummer, A.M. Kamp, F. Citarella, A.J. Horrevoets, C.E. 

Hack, Expression of human recombinant granzyme A zymogen 

and its activation by the cysteine proteinase cathepsin C, J. Biol. 

Chem. 271 (1996) 9281–9286. 

[11] P.J. Beresford, C.M. Kam, J.C. Powers, J. Lieberman, Recombinant 

human granzyme A binds to two putative HLA-associated 

proteins and cleaves one of them, Proc. Natl. Acad. Sci. USA 94 

(1997) 9285–9290. 

[12] J. Sun, C.H. Bird, M.S. Buzza, K.E. McKee, J.C. Whisstock, P.I. 

Bird, Expression and puriWcation of recombinant human granzyme 

B from Pichia pastoris, Biochem. Biophys. Res. Commun. 

261 (1999) 251–255. 

[13] Z. Xia, C.M. Kam, C. Huang, J.C. Powers, R.J. Mandle, R.L. Stevens, 

J. Lieberman, Expression and puriWcation of enzymatically 

active recombinant granzyme B in a baculovirus system, Biochem. 

Biophys. Res. Commun. 243 (1998) 384–389. 

[14] C.T. Pham, D.A. Thomas, J.D. Mercer, T.J. Ley, Production of 

fully active recombinant murine granzyme B in yeast, J. Biol. 

Chem. 273 (1998) 1629–1633. 

[15] M.J. Smyth, M.J. McGuire, K.Y. Thia, Expression of recombinant 

human granzyme B. A processing and activation role for dipeptidyl 

peptidase I, J. Immunol. 154 (1995) 6299–6305. 

[16] A. Caputo, M.N. James, J.C. Powers, D. Hudig, R.C. Bleackley, 

Conversion of the substrate speciWcity of mouse proteinase granzyme 

B, Nat. Struct. Biol. 1 (1994) 364–367. 

[17] L.M. Babe, S. Yoast, M. Dreyer, B.F. Schmidt, Heterologous 

expression of human granzyme K in Bacillus subtilis and characterization 

of its hydrolytic activity in vitro, Biotechnol. Appl. Biochem. 

27 (1998) 117–124. 

[18] E. Wilharm, M.A. Parry, R. Friebel, H. Tschesche, G. Matschiner, 

C.P. SommerhoV, D.E. Jenne, Generation of catalytically active 

granzyme K from Escherichia coli inclusion bodies and identiWcation 

of eYcient granzyme K inhibitors in human plasma, J. Biol. 

Chem. 274 (1999) 27331–27337. 

[19] J.H. Christensen, P.K. Hansen, O. Lillelund, H.C. Thøgersen, 

Sequence-speciWc binding of the N-terminal three-Wnger fragment 

of Xenopus transcription factor IIIA to the internal control region 

of a 5S RNA gene, FEBS Lett. 281 (1991) 181–184. 

[20] H.C. Thøgersen, M. Etzerodt, T.L. Holtet, Improved method for 

the refolding of proteins, International Patent Application 1994, 

WO 94/18227. 

[21] T.L. Holtet, J.H. Graversen, I. Clemmensen, H.C. Thøgersen, M. 

Etzerodt, Tetranectin, a trimeric plasminogen-binding C-type lectin, 

Protein Sci. 6 (1997) 1511–1515. 

[22] J. Graversen, S. Moestrup, Apolipoprotein conjugates, International 

Patent Application 2002, WO 02/038609. 

[23] B.B. Nielsen, J.S. Kastrup, H. Rasmussen, T.L. Holtet, J.H. 

Graversen, M. Etzerodt, H.C. Thøgersen, I.K. Larsen, Crystal 

structure of tetranectin, a trimeric plasminogen-binding 

protein with an alpha-helical coiled coil, FEBS Lett. 412 (1997) 

388–396. 

[24] L. Ellgaard, T.L. Holtet, P.R. Nielsen, M. Etzerodt, J. Gliemann, 

H.C. Thogersen, Dissection of the domain architecture of the 

alpha2macroglobulin-receptor-associated protein, Eur. J. Biochem. 

244 (1997) 544–551. 

[25] J. Sun, J.C. Whisstock, P. Harriott, B. Walker, A. Novak, P.E. 

Thompson, A.I. Smith, P.I. Bird, Importance of the P4 residue in 

human granzyme B inhibitors and substrates revealed by scanning 

mutagenesis of the proteinase inhibitor 9 reactive center loop, J. 

Biol. Chem. 276 (2001) 15177–15184.

Expression, refolding, and purification of recombinant human ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?