The structure–function relationship in the clostripain family of peptidases Nikolaos E. Labrou 1 and Daniel J. Rigden 2 1 Laboratory of Enzyme Technology, Department of Agricultural Biotechnology, Agricultural University of Athens, Greece; 2 School of Biological Sciences, University of Liverpool, UK In this study we investigate the active-site structure and the catalytic mechanism of clostripain by using a combination of three separate techniques: affinity labelling, site-directed mutagenesis and molecular modelling. A benzamidinyl- diazo dichlorotriazine dye (BDD) was shown to act as an efficient active site-directed affinity label for Clostridium histolyticum clostripain. The enzyme, upon incubation with BDD in 0.1 M Hepes/NaOH buffer pH 7.6, exhibits a time- dependent loss of activity. The rate of inactivation exhibits a nonlinear dependence on the BDD concentration, which can be described by reversible binding of dye to the enzyme prior to the irreversible reaction. The dissociation constant of the reversible formation of an enzyme–BDD complex is K D ¼ 74.6 ± 2.1 l M and the maximal rate constant of inactivation is k 3 ¼ 0.21Æmin )1 . Effective protection against inactivation by BDD is provided by the substrate N-benzoyl- L -arginine ethyl ester (BAEE). Cleavage of BDD-modified enzyme with trypsin and subsequent separation of peptides by reverse-phase HPLC gave only one modified peptide. Amino acid sequencing of the modified tryptic peptide revealed the target site of BDD reaction to be His176. Site- directed mutagenesis was used to study further the func- tional role of His176. The mutant His176Ala enzyme exhibited zero activity against BAEE. Together with previ- ous data, these results confirm that a catalytic dyad of His176 and Cys231 is responsible for cysteine peptidase activity in the C11 peptidase family. A molecular model of the catalytic domain of clostripain was constructed using a manually extended fold recognition-derived alignment with caspases. A rigorous iterative modelling scheme resulted in an objectively sound model which points to Asp229 as responsible for defining the strong substrate specificity for Arg at the P1 position. Two possible binding sites for the calcium required for auto-activation could be located. Database searches show that clostripain homologues are not confined to bacterial lineages and reveal an intriguing variety of domain architectures. Keywords: active site; affinity labelling; clostripain; mole- cular modelling; peptidase family C11. Clostripain (EC 3.4.22.8) is a cysteine endopeptidase with strict specificity for Arg–Xaa peptidyl bonds, isolated from the culture filtrate of the anaerobic bacterium Clostridium histolyticum [1]. It is heterodimeric enzyme composed of two polypeptide chains of molecular masses 43 000 kDa and 15 400 kDa, for the heavy and light chains, respectively [2]. The two chains are held together by strong noncovalent forces [1]. Both polypeptide chains of native clostripain are encoded by a single gene with an ORF of 1581 nucleotides encoding a polypeptide of 526 amino acid residues [2]. Heterologous expression experiments revealed that clostri- pain is synthesized as an inactive prepro-enzyme. In the presence of calcium ions, the core protein (amino acids 51–526) is able to catalyse the removal of the linker nonapeptide (residues 182–190) [3,4]. The enzyme is important both in sequence analysis and in enzymic peptide synthesis, as it accepts proline in the P1¢ position [5,6]. Study of the active site of clostripain, by using protein chemistry experiments, has shown that the Cys41 of the heavy chain (corresponding to Cys231 of the protein, as synthesized) is the catalytic sulfhydryl residue of the active site [7–9]. In addition, the inactivation of clostripain by diethylpyrocarbonate has suggested the involvement of one or more histidine residues in clostripain activity [7]. Never- theless, direct evidence for the involvement of a histidine residue in the catalytic mechanism of the enzyme has not yet been provided. In the MEROPS classification of proteinase sequences [10], clostripain is grouped into family C11. Although clostripain has no significant overall sequence similarity with other proteinase families, it has been placed in clan D, along with cysteine peptidase families C13 (legumains), C14 (caspases) and C25 (gingipains). Several criteria supported this grouping including shared sequence motifs, predicted secondary structure, strong specificity for the P1 position of the substrate peptide and immunity to inhibition by E-64 irreversible protease inhibitor [11]. Later support for the existence of structural homology between gingipains and caspases was provided by their common inhibition by the Correspondence to N. E. Labrou, Enzyme Technology Laboratory, Department of Agricultural Biotechnology, Agricultural University of Athens, Iera Odos 75, 11855 Athens, Greece. Fax: +30 210 5294308, Tel.: +30 210 5294308, E-mail: Lambrou@aua.gr Abbreviations:BAEE,N-benzoyl- L -arginine ethyl ester; BDD, benz- amidinyl-diazo dichlorotriazine dye; ChC, Clostridium histolyticum clostripain. Enzyme: clostripain (EC 3.4.22.8). (Received 31 October 2003, revised 26 December 2003, accepted 19 January 2004) Eur. J. Biochem. 271, 983–992 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04000.x baculovirus inhibitor p35 [12]. The separin family (peptidase family C50) has been added to clan D [13] and the composition, distribution and evolution of all these and other related families analysed through sequence compar- isons [14]. Reactive triazine dyes have been used successfully for the purification and resolution of many proteins by affinity chromatography and for affinity labelling of several enzymes and proteins [15–18]. We have previously estab- lished the use of reactive dichlorotriazine dye Vilmafix Blue A-R as a structural probe for labelling the NAD(H) binding site of formate dehydrogenase [16], malate dehydrogenase [17] and the oxaloacetate binding site of oxaloacetate decarboxylase [18]. In this study we describe the use of a reactive dichloro- triazine dye as an affinity label for clostripain and provide direct evidence by site-directed mutagenesis and molecular modelling studies that His176 is part of the catalytic dyad of clostripain. The molecular modelling, in conjunction with sequence analysis studies, indicates the P1 specificity deter- mining residue as Asp229 and locates possible calcium- binding sites involved in the auto-processing. Experimental procedures Materials N-benzoyl- L -arginine ethyl ester (BAEE), bovine pancreas trypsin (grade III, 10 800 UÆmg )1 )andC. histolyticum clo- stripain were from Sigma Co. (St. Louis, MO, USA). The plasmid pKK223-3 was from Amersham Bioscience. All other molecular biology reagents were purchased from Promega. Synthesis and purification of benzamidinyl-diazo dichlorotriazine dye Synthesis of benzamidinyl-diazo dichlorotriazine (BDD) was as described previously [19]. Purification of BDD was achieved by preparative TLC on silica gel 60 plates, using the solvent system: MeOH/H 2 O/AcCN (2.5 : 2.5 : 5; v/v/v). Enzyme assays Clostripain assays were performed with a Hitachi U-2000 double-beam spectrophotometer carrying a thermostated cell holder (25 °C, 10-mm pathlength), according to a published method [20]. One unit of enzyme activity is defined as the amount that catalyses the conversion of 1 lmol of substrate (BAEE) to product per min. Enzyme activity calculations were performed using molar extinction coefficients of 1150 M )1 Æcm )1 at 253 nm. Determination of protein concentration Protein concentration was determined by the Lowry method [21] using crystalline BSA (fraction V) as standard. Enzyme inactivation studies Inactivation of clostripain was performed in an incubation mixture containing, in a total volume of 1 mL at 25 °C, 100 lmol Hepes/NaOH buffer pH 7.6, 0–148.6 nmol BDD, 1.2 units enzyme. The rate of inactivation was followed by periodically removing samples (10–50 lL) for assay of enzymatic activity. Initial rates of inactivation were deduced from plots of log (% of activity remaining) vs. time (min) for several dye concentrations and the slopes and intercepts of secondary double reciprocal plots were cal- culated by unweighted linear regression analysis. Inactivation studies of clostripain by BDD in the presence of substrate (BAEE) was performed in a total volume of 1mL(25°C) and the reaction mixture contained 100 m M Hepes/NaOH buffer pH 7.6, 16.9 nmol BDD, 1 m M or 5m M BAEE and 1.2 units clostripain. In order to calculate the pK a of the amino acid residue involved in the nucleophilic modification of C. histolyticum clostripain (ChC) by BDD, enzyme inactivation experi- ments were performed at various pH values (6.0–8.5). Inactivation was carried out in an incubation mixture containing, in a total volume of 1 mL at 25 °C: 100 lmol Mops/NaOH buffer pH 6–7, 23.1 nmol BDD, 1.2 units enzyme, or 100 lmol Hepes/NaOH buffer pH 7–8.5, 23.1 nmol BDD, 1.2 units enzyme. Data were analysed by the GRAFIT program (Erithacus Software Ltd). Stoichiometry of BDD binding to Ch C ChC(100lg) in 100 m M Hepes/NaOH buffer pH 7.6 was inactivated with 40.5 nmol BDD at 25 °C. The dye- inactivated enzyme was separated from the free dye by ultrafiltration (in an Amicon stirred cell 8050 carrying a Diaflo YM10 ultrafiltration membrane; cut-off 10 kDa) after extensive washing with distilled water. Further separ- ation was achieved by gel-filtration chromatography by applying the inactive dye–enzyme complex to a Sephadex G-25 column (9 cm · 1.6 cm) equilibrated with water, and collecting fractions (0.5 mL) at a flow rate of 10 mLÆh )1 . The solution with dye-inactivated ChC was then lyophilized and stored at )20 °C. The lyophilized ChC–BDD covalent complex was dissolved in 8 M urea, and the absorbance was determined spectrophotometrically at 387 nm using a molar extinction coefficient of 11.4 LÆcm )1 Æmmol )1 determined in 8 M urea. The protein concentration was determined by the method of Lowry [21]; no dye interference is observed in protein determinations. Tryptic digestion of the BDD-clostripain covalent complex and peptide purification using HPLC In order to covalently block the free -SH groups, before peptide purification, lyophilized BDD–clostripain covalent complex (100 lg) was dissolved in Hepes/NaOH buffer (0.1 M , pH 7.0, 1 mL) and was denatured by the addition of solid urea to yield 8 M solution. To the denatured enzyme N-ethyl-maleimide was added to a final concentration of 10 m M , and the solution incubated for 30 min at room temperature. The enzyme was then dialysed against 0.1 M ammonium bicarbonate buffer pH 8.3. The enzyme was digested by the addition of 10 lg trypsin. The digestion was allowed to continue overnight at 30 °C before the mixture was lyophilized and stored dry at )20 °C. Separation of the resulting peptides was achieved on a C18 reverse phase S5 ODS2 Spherisorb silica column (250 mm · 4.6 mm i.d.). 984 N. E. Labrou and D. J. Rigden (Eur. J. Biochem. 271) Ó FEBS 2004 Analysis was achieved by a H 2 O/acetonitrile linear gradient containing 0.1% trifluoroacetic acid (0–80% acetonitrile during 80 min) at a flow rate of 0.5 mLÆmin )1 . Fractions of 0.5 mL were collected. The eluents were monitored at both 220 nm and 387 nm. Cloning, expression, purification and site-directed mutagenesis of Ch C The gene encoding ChC was amplified by PCR from genomic DNA using oligonucleotide primers designed from the published gene sequence of ChC as follows [2]: the PCR reaction was carried out in a total volume of 100 lL containing 8 pmol of each primer (5¢-ATGAACA AAAATCAAAAAGTAACTATT-3¢ and 5¢-TTACCAT TGGTAATGATTAACTCCTCC-3¢), 100 ng template DNA, 0.2 m M of each dNTP, 10 mL 10· Pfu buffer and 1 U Pfu DNA polymerase. The PCR procedure comprised 30 cycles of 45 s at 95 °C, 1 min at 55 °Cand 2 min at 72 °C. A final extension time at 72 °Cfor 10 min was performed after the 30 cycles. The PCR products were run on a 1.2% (w/v) agarose gel and the product was excised, purified by adsorption to silica beads and ligated to the pKK223-3 expression vector, which was previously restricted with EcoRI and treated with T4 DNA polymerase. The resulting expression construct pChC was used to transform competent Escherichia coli JM105 cells. E. coli harbouring plasmid pChC were grown at 37 °C in 1 L Luria–Bertani medium containing 100 lgÆmL )1 ampicillin. The synthesis of clostripain was induced by the addition of 1 m M isoprophyl thio-b- D -galactoside when the absorbance at 600 nm was 0.6. Four hours after induction, cells (% 3g) were harvested by centrifugation at 4000 g for 15 min, resuspended in potassium phosphate buffer (50 m M ,pH7.5,9mL), sonicated, and centrifuged at 10 000 g for 20 min. The supernatant was collected and dialysed overnight against 2 L of activation buffer (50 m M Tris/HCl pH 6.0, 5 m M DTT). The dialysate was loaded onto a column of BDD– Sepharose, 1 mL [19] previously equilibrated with Mes/ NaOH buffer (20 m M , pH 6.0). Non-adsorbed protein was washed off with 10 mL equilibration buffer, followed by 10 mL Mes/NaOH buffer (20 m M , pH 6.0) containing 10 m M KCl. Bound ChC was eluted with equilibration buffer containing 1 mgÆmL )1 L-Arg. Collected fractions (1 mL) were assayed for ChC activity and protein. Site-directed mutagenesis was performed according to the unique site elimination method described by Deng and Nickoloff [22]. The oligonucleotide primer sequence for the His176Ala mutation was as follows: 5¢-ATGGCT AAT GCAGGTGGTGCA-3¢ and the selection primer’s sequence was as follows: 5¢-GAATTC TCGTGGATCC GTCGACCT-3¢. This primer contains a mutation in a unique SmaI restriction site of the pChC vector. Altered nucleotides are shown underlined. The primers were phosphorylated before use with polynucleotide kinase. The expression construct pChC was used as template DNA in all mutagenesis reactions. All mutations were verified by DNA sequencing using the DyeDeoxy Terminator method. The mutant was expressed in E. coli and purified as described above for the wild-type enzyme. Bioinformatics Sequences homologous to clostripain were sought in the Genpept and Unfinished Microbial Genome databases at the NCBI using BLAST [23] and PSI - BLAST [24]. The resulting sequence set was aligned with T - COFFEE [25]. Jalview (http:// www.ebi.ac.uk/$michele/jalview) was used for alignment visualization, manipulation and the calculation of five maximally diverse representatives of the clostripain family. The limits of the common conserved region present in all clostripain homologues were determined by inspection of the alignment. This region, in diverse homologous sequences, was submitted for fold recognition experiments at the META - server [26]. The META -server unites most of the leading fold recognition methods and provides consensus predictions offering improved reliability. The most informative results in our case were provided by the FFAS 03 method [27], a sensitive sequence only based method which works by alignment of two profiles [27]. Secondary structure predictions were carried out using PSI - PRED [28]. The domain content of the portions, of varying lengths, flanking the common conserved region was analysed through searches at the PFAM [29] and SMART [30] databases, and through further PSI - BLAST and fold recognition experiments. Modelling of the common conserved region of clostripain was carried out with MODELLER 6 [31] using the structures of caspases 1 (PDB code 1bmq [32]), 3 (PDB code 1pau [33]); and 8 (PDB code 1jxq [34]), sharing 27–36% pairwise sequence identity over the region shown in Fig. 3, as templates. Despite these relatively low levels of sequence identity the regular secondary structure elements of the three templates superimpose extremely well; significant structural differences are confined to the connecting loops. Catalytic and specificity-determining residues superimpose very well. Use of multiple related templates is known to produce better models than use of a single one. The T - COFFEE alignment was used to transfer the fold recogni- tion alignment of the C. aurantiacus with caspases to clostripain itself. Default regimes of model refinement by energy minimization and simulated annealing were used. In regions in which all three templates superimposed well, information from each was incorporated into the modelling process. Where the templates differed the choice of which to use was based on local similarity in length and composition to the clostripain sequence. For the region of 20 residues neighbouring the site of caspase cleavage, the gingipain structure (PDB code 1cvr [35]) was used as template. Structural determination of gingipain showed that, despite a lack of significant sequence similarity with the caspases, the gingipain catalytic domain adopted the caspase-like fold [35]. The cleaved form of clostripain, lacking the internal nonapeptide was modelled. Given the low sequence simi- larity between target and templates, a rigorous iterative modelling scheme was used in which 20 models were constructed and analysed for each variant alignment. These models were analysed for stereochemical properties with PROCHECK [36] and for packing and solvent exposure characteristics with PROSA II [37]. Model regions corres- ponding to positive PROSA II profile peaks were treated as possibly resulting from misalignments. Alterations in align- ments were tested for these regions. When no further improvements were possible the final model was taken as Ó FEBS 2004 Clostripain family structure–function relationship (Eur. J. Biochem. 271) 985 that with the best PROSA II score. Protein structures were superimposed using LSQMAN [38] and visualized using O [39]. Structural figures were produced with PYMOL [40]. Secon- dary structure in experimental structures was defined with STRIDE [41]. Results and discussion Kinetics of reaction of BDD with clostripain Incubation of ChC with 5.65–148.6 l M BDD at pH 7.6 and 25 °C leads to a progressive loss of enzyme activity, as shown in Fig. 1A, whereas the control enzyme (in the absence of reagent) is stable under these conditions. The time-dependent inactivation follows pseudo-first order kin- etics over the first 10 min. The rate constant of inactivation (k obs ) exhibits a nonlinear dependence on the reagent concentration (Fig. 1B). This indicated that the reaction obeyed pseudo-first order saturation kinetics and was consistent with reversible binding of reagent prior to covalent modification according to [15–18]: E þ BDD  ! K D E:BDDÀ! k 3 E-BDD where, E represents the free enzyme; E:BDD is the reversible complex and E-BDD is the covalent product. The steady- state rate equation for the interaction is [15–18]: 1=k obs ¼ 1=k 3 þ K D =ðk 3 ½BDDÞ where K D is the apparent dissociation constant of the enzyme:BDD complex and k 3 is the maximum rate of inactivation at saturating concentration of the reagent. The rate constant was measured as shown in Fig. 1A. From the double reciprocal plot of 1/k obs vs.1/[BDD],shownin Fig. 1B a value of K D ¼ 74.6 ± 2.1 l M was estimated for the dissociation constant of a reversible clostripain:BDD complex. The observed maximum rate of inactivation at saturating concentration of the reagent was estimated 0.21 min )1 . Affinity labelling is a useful tool for the identification and probing of specific, catalytic and regulatory sites in purified enzymes and proteins. In the present study we demonstrate the usefulness of BDD as a structural probe for the argininyl-recognizing protease clostripain. The 1,3,5-triazine reactive scaffold is of special interest because of its synthetic accessibility, by taking advantage of the temperature- dependent successive displacement of the chloride atoms by different nucleophiles [42]. Other advantages of synthesis of triazine-based affinity labels are their high stability against biological and chemical degradation and their capacity to form hydrogen bonds with amino acid residues within the binding site due to the presence of electron rich nitrogen sites [42]. Specificity of a protein chemical modification reaction can be indicated by the ability of substrate to protect against inactivation. The substrate was added to the incubation mixture at a concentration much higher than the known enzyme–ligand dissociation constant in order to assess its effect on the inactivation rates at pH 7.6 and 25 °C. For example, for BAEE the K m value is 0.235 m M [43]. Fig. 1C shows that the rate of enzyme inactivation by BDD decelerated in the presence of 1 or 5 m M BAEE. Fig. 1. Affinity labelling of ChC. (A) Time course for the inactivation of ChC by BDD. Inactivation was performed at pH 7.6 and 25 °C. No BDD (h); 5.66 l M (j); 11.32 l M (r); 16.97 l M (w); 37.0 l M (e); 148.6 l M (*). (B) Effect of BDD concentration on the observed rate of inactivation (k obs )ofChC expressed as a double-reciprocal plot. BDD, 5.66–148.6 l M . The slope and intercept of the secondary double- reciprocal plot were calculated by unweighted linear regression ana- lysis. Inset shows the structure of BDD. (C) Effect of substrate (BAEE) on the time course of inactivation of ChC by BDD (pH 7.6, 25 °C). No BDD (h); BDD, 16.97 l M (w); BDD, 16.97 l M in the presence of 1m M BAEE (r)or5m M BAEE (j). 986 N. E. Labrou and D. J. Rigden (Eur. J. Biochem. 271) Ó FEBS 2004 To determine the stoichiometry of dye binding, ChCwas completely inactivated by the dye and the dye–enzyme covalent complex was resolved from free dye by gel filtration on Sephadex G-25 and ultrafiltration. The molar ratio of [Dye]:[ChC active site] was determined by measuring the dye spectrophotometrically in urea solution, and the protein by the method of Lowry et al.[21].The molar ratio of dye to ChC active site was 1 : 1.1 ± 0.1, using a molecular weight 56 000, indicating a specific interaction between dye and protein. BDD exhibits several characteristics of an affinity label in its reaction with clostripain. It reacts stoichiometrically with the enzyme. Time- and dye concentration-dependent inac- tivation of clostripain by BDD is evident. The pseudo-first order kinetics obtained for clostripain inactivation indicates that the phenomenon occurs through the initial formation of a reversible Michaelis binary complex followed by subsequent formation of a covalent complex [16–18]. Protection against inactivation by BDD is provided by the synthetic substrate BAEE, indicating that the dye interacts with the enzyme at the substrate binding site. Isolation and analysis of peptides from clostripain modified by BDD Modified clostripain was subjected to trypsin digestion followed by fractionation by reverse-phase HPLC. Essen- tially, a single yellow peak (BDD-labelled peptide) eluted from the column. The yellow peak was freeze dried and subjected to amino acid analysis and sequencing. The overall recovery of modified peptide, based on the initial amount of modified enzyme was 22%. Automated Edman sequence analysis of the labelled peptide gave the sequence YVLIMAN-X-GGGAR, where X indicates that no phe- nylthiohydantoin derivative was detected in the cycle. By comparison with the amino acid sequence of clostripain, the X in the peptide was identified as His176, indicating that the site chain of His is the reactive group responsible for the nucleophilic attack on the diclorotriazine ring of the dye. Site directed mutagenesis and pH dependence of inactivation The wild-type enzyme and the mutant His176Ala were expressed in E. coli and characterized by steady-state kinetic analysis. Assay for clostripain activity of the purified mutant revealed that it was completely inactive. Thus both our chemical modification and site-directed mutagenesis data confirm the predictions made regarding clostripain’s cata- lytic site [9]. Our data provide the first direct evidence that catalysis by clostripain involves the Cys–His dyad almost ubiquitously involved in cysteine peptidase mechanisms [42,44]. The study of the effect of pH on enzyme inactivation allows the calculation of the pK a of the His176 side chain involved in the inactivation reaction. The rate of inactiva- tion exhibited a sigmoid-shaped pH-dependence indicating that the reaction depends strongly on the nucleophilicity of a deprotonated group. The pK a value measured from this curve was equal to 7.4 ± 0.2 (Fig. 2). This pK a value is higher than the expected value for the free amino acid but is in agreement with the expected value for a His interacting with a thiolate [45]. In the papain family, Cys25 and His159 form a thiolate–imidazolium ion pair in which the pK a values of the two residues are perturbed by approximately 4 units (Cys to pK a 4) and 2 units (His to pK a 8.5), respectively [45]. The absence of strong pK a perturbation, compared to that observed in papain, may be related to the greater separation of His and Cys in the caspase structures [46], and in the clostripain molecular model (see below). The greater separation would not allow for the degree of pK a pertur- bation observed in the papain family [47]. Clostripain homologues Previous searches for clostripain homologues and the current state of the PFAM database revealed only the presence of clostripain itself and three Thermotoga maritima homo- logues [12]. Our database searches using PSI - BLAST [24], in both GenBank and Unfinished Microbial Genome data- bases at the NCBI (http://www.ncbi.nlm.nih.gov/blast/), initially located, ignoring obviously partial sequences, 13 homologues in GenBank and three among unfinished microbial genome data. The species in which clostripain homologues were newly observed were C. perfringens, C. thermoceullum, C. tetani, Methanosarcina acetivorans, Chloroflexus aurantiacus, Geobacter metallireducens,and Ruminococcus albus. The observation, for the first time, of a clostripain homologue in the Archaea (M. acetivorans)is particularly interesting in view of the interest in under- standing the curious phyletic distributions of clostripains and related peptidase families [9,12]. Over the alignment section shown in Fig. 3, the archaebacterial homologue shares 16–27% sequence identity with the other clostripain family members. It contains all the possible functional residues discussed later. Alignment of these sequences enabled the location of a common conserved region presumably containing the catalytic domain. Of the three Thermotoga maritima sequences found, one (GenBank, 15643282) lacked a conserved N-terminal portion found in all the other Fig. 2. The pH dependence of clostripain inactivation by BDD at 25 °C. The reaction mixture contained 1.2 U enzyme, 22.1 l M BDD, and 100 m M (Mops/NaOH or Hepes/NaOH) buffer in pH values 6.0–8.5. Ó FEBS 2004 Clostripain family structure–function relationship (Eur. J. Biochem. 271) 987 homologues. Translation of the corresponding DNA revealed this portion lying upstream of the annotated start but failed to highlight any alternative start codons. This sequence was therefore not included in subsequent analysis as possibly representing an inactivated copy. Similarly, one of the four Chloroflexus aurantiacus sequences lacked both the catalytic Cys47 and His residues (this work) and, since our interest lay principally in understanding peptidase activity in the clostripain family, was not studied further. The appearance of inactivated copies of related peptidases in various evolutionary lineages appears common [12]. The set of clostripain homologues was remarkably diverse both in length and in composition. Considering only the identified common conserved region (correspond- ing to residues 56–446 in clostripain, see Fig. 3), no two sequences shared more than 56% sequence identity. The mean pairwise sequence identity among the 13 homologues in the common conserved region was just 21%. Only six positions were entirely conserved and another 10 were conserved in 12 of the 13 sequences (Fig. 3). In order to analyse the composition of the clostripain homologues outside the catalytic domain, searches were carried out in the PFAM [29] and SMART [30] domain databases and more distant domain matches sought for the remaining regions by fold recognition. The current PFAM database shows the presence of bacterial immunoglobulin- like domains (PFAM, PF02369; SMART, SM00634) in two T. maritima proteins but our searches revealed a much more diverse set of domain architectures in the family (Fig. 4). As well as the bacterial immunoglobulin-like domains members Fig. 3. Sequence alignment of five maximally diverse representatives of the clostripain homologue alignment with the three caspase templates used for model construction. GenBank identification numbers and abbreviated species names are shown for the clostripain homologues (399264 is clostripain itself), while PDB codes and enzyme names are provided for the templates. The predicted secondary structure for clostripain (obtained with PSIPRED [28] and clostripain numbering are shown above the alignment. The STRIDE [41] derived secondary structure of human caspase-1 and its numbering are shown beneath the alignment. Shaded regions indicate portions cleaved upon activation of clostripains or caspases, although cleavage has only been shown experimentally for clostripain, not for the homologues shown here. The boxed region indicates the single part of the clostripain molecular model obtained from the gingipain structure (see text for details). Bold italic face is used for the catalytic His and Cys residues. Bold face among the clostripains signifies conservation among at least 12 of the 13 sequences considered. Italic face is used to show portions of the clostripain sequence for which reliable modelling was not possible. The figure was made with ALSCRIPT [53]. Fig. 4. Schematic diagram of domain architectures present among clo- stripain homologues. Rectangles represent catalytic domains and other shapes the additional identified domains. Only the association of clo- stripain catalytic domains with bacterial immunoglobulin-like domains is visible in the current PFAM database [29]. Domains were identified through screening against PFAM and SMART [30], with the exception of the fibronectin type 3 domain in 15644337 which was identified by fold recognition. 988 N. E. Labrou and D. J. Rigden (Eur. J. Biochem. 271) Ó FEBS 2004 of the clostripain family contain forkhead domains (SMART, SM00240), fibronectin type 3 domains (PFAM, PF00041) and NHL domains (PFAM, PF00400). None of these domain entries gives more than a clue as to the physiological roles of the clostripain homologues but it is interesting to note that both forkhead and NHL domains are implicated in protein–protein interactions [48,49]. Simi- larly, both bacterial immunoglobulin-like domains and fibronectin type 3 domains are strongly associated with cell adhesion [50]. Most unexpectedly, one clostripain homo- logue from C. aurantiacus contains tandem peptidase catalytic domains (Fig. 4) with a peptidase M37 catalytic domain preceding the peptidase C11 domain and a fibro- nectin type 3 domain lying between the two. The picture that emerges is one in which clostripain itself, the only member of the peptidase C11 family to have been experi- mentally studied, is atypically simple in possessing the catalytic domain alone. The peptidase C11 family contains a large variety of domain architectures which probably reflect a range of physiological roles that deserve further study. Molecular modelling Existing data showed a distant evolutionary relationship between the clostripain family and other peptidase families [9,12]. The characterization of the clostripain mutant H176A and its specific chemical modification presented here provides further support for the hypothesis. In order to explore other aspects of the structure–function relationship of clostripain and its homologues, a molecular model would be invaluable. This would obviously require a reliable alignment of clostripain, or a homologue, with a known structure. Previous published alignments have covered only part of the conserved common region of the clostripain family [12], terminating shortly after the catalytic Cys and therefore not allowing for molecular modelling. We there- fore carried out fold recognition experiments in order to try to obtain an alignment that would enable the construction of a molecular model for clostripain. The recent availability of diverse clostripain homologues would facilitate fold recognition studies in two important ways: firstly by enabling the limits of the catalytic domain to be identified (thereby improving fold recognition accuracy); and sec- ondly, as fold recognition may sometimes be successful for one homologue but not for another, by providing several different distantly homologous sequences to serve as input for the fold recognition. Fold recognition experiments with several sequences corresponding to the common conserved region produced initially confusing results. Strongly significant results were obtained for the a/b hydrolase fold with the expected caspase-like fold scoring worse. Comparison of the clostri- pain sequences with the conserved characteristics of the a/b hydrolase fold, such as the so-called nucleophile-elbow [51], enabled it to be discarded as a possible fold for clostripains. In contrast, the alignments of clostripain sequences with caspases aligned both the Cys and His catalytic residues. Further examination of the alignments revealed the reasons behind the unexpected results. Firstly, the cleavage of caspases shortly after the catalytic Cys has led to their structures being deposited with the PDB with different chain names for the cleaved N- and C-terminal portions. The two pieces are therefore considered as separate chains by the fold recognition algorithms and the clostripain– caspase alignments covered only the caspase regions prior to the cleavage point. The complete alignments would presumably have scored much better. Secondly, among the several insertions of clostripains relative to caspases is a very large one towards the C terminus (Fig. 4), predicted to contain four a-helices which, by chance, aligned with some members of the a/b hydrolase superfamily containing a similarly placed all-a excursion to the main fold. The best alignment of a clostripain homologue with a caspase structure (caspase-9; PDB code 1jxq [34]); was produced for the C. aurantiacus homologue with GenBank 22972276 by the FFAS 03 method [27] and given a highly significant score of )7.6. Using this incomplete alignment of clostripain with the caspases as a base, the alignment was manually extended through matching of caspase secondary structure with clostripain predicted secondary structure (Fig. 3). At certain key points, residue conservation could be used to improve confidence in the correctness of the alignment. For example, the caspases have a serine conserved at position 332, whose side chain forms hydrogen bonds with both the carboxyl oxygen and the nitrogen atoms of the backbone of the residue preceding the catalytic Cys. The conservation of this interaction is suggestive of its Fig. 5. The final model of the clostripain catalytic domain. The ribbon is coloured according to secondary structure and key residues shown using a stick representation and labelled. Residues at the catalytic site are shown in larger face, residues of possible calcium binding sites (see text for details) in smaller face. The model is of the cleaved clostripain lacking the internal nonapeptide. The final residue of the resultant a-chain, Arg181, and the first residue of the b-chain, Ala191, are also labelled (italics). The magenta colouring towards the bottom of the figure marks the position of the large unmodelled insertion in clostri- pain compared to caspases towards the C terminus. Ó FEBS 2004 Clostripain family structure–function relationship (Eur. J. Biochem. 271) 989 importance so it was reassuring that a serine, conserved with one exception among clostripain homologues (numbered 257 in clostripain itself), could be aligned with this position (Fig. 3). Similarly conserved caspase Trp340, lining the catalytic site could be aligned with a conserved aromatic residue in the set of clostripain homologues. The very C-terminal portion of the caspase structure, around residue 400, forms a key part of the domain structure and adopts an extended conformation which is not defined as b-structure due to the absence of the necessary hydrogen bonds. It was aligned with a predicted b-strand in the clostripain family. This defined a very large insertion in clostripains relative to the caspases which was not amenable to modelling. However, the absence of significant sequence conservation and variable length of the region were not suggestive of functional importance. The match between predicted clostripain secondary structure and actual caspase secon- dary structure of the final alignment is very good (Fig. 3). Nevertheless this region must be considered less reliable than other portions of the model. Only one of the putative functional residues discussed below is located in this region. With the most complete caspase–clostripain alignment available, a process of iterative model building was carried out using as templates the highest resolution structures available of caspases 1, 3 and 9 as described in Experimental procedures. Over the modelled portion of clostripain the templates shared 12–16% sequence identity with clostripain. A particular problem was encountered for the clostripain region near to the cleaved portion of the caspases. In all the caspase structures, cleavage results in the segments pre- ceding and following the site of cleavage adopting highly extended conformations with no contacts to the compact domain structure. In contrast, a predicted helix is present in the corresponding, uncleaved portion of the clostripains. For this region only, the corresponding part of gingipain, whose structure also indicates distant homology to the caspases [34], was used (Fig. 3). Structural similarity between gingipain and caspases is particularly strong for the catalytic site residues. The cleavage of clostripain with loss of internal peptide was included in the model (Fig. 3). During the iterative modelling scheme, several alignment changes were found to result in improved models, as judged by PROSA II [37] analysis resulting in the final alignment shown in Fig. 3. Although the final model (Fig. 5) lacked several insertions, too large to model effectively, it scored )6.24 by PROSA II , corresponding to a near-optimal pG value [52] of 0.99. This result confirms the correctness of the fold used as template for modelling and is suggestive of largely accurate alignment [52]. Eighty-six per cent of residues occupied core regions of the Ramachandran plot in Fig. 6. Determinants of P1 substrate specificity in (A) clostripain (specific for Arg) (B) caspase (specific for Asp) and (C) gingipain (specific for Arg). The same colouring by secondary structure is used in all panels. Key residues are shown as ball-and-stick and coloured pink (catalytic) or light grey (specificity-determin- ing, experimentally determined for caspase and gingipain, predicted for clostripain). The caspase and gingipain structures shown (1bmq [32] and 1cvr [35], respectively) both contain inhibitors bound at the catalytic site and cov- alently attached to the catalytic Cys residues which are shown as cyan sticks. Portions of the caspase and gingipain structures lying outside the common conserved structural core are coloured grey. 990 N. E. Labrou and D. J. Rigden (Eur. J. Biochem. 271) Ó FEBS 2004 the final model. There were no Ramachandran-disallowed residues and just two located in generously allowed zones. Model analysis With the good objective quality of the final model estab- lished, it was used to address issues of the structure–function relationship in the clostripain family. The first question was the mechanism by which clostripains specify a strong preference for Arg at the P1 position of the substrate. Examination of the alignment (Fig. 3) alone reveals several conserved acidic residues, any one of which could be responsible for substrate specificity. However, examination of conserved residues (Fig. 3) in the context of the model (Figs5and6),andcomparisonofthemodelwithcaspase and gingipain crystal structures (where specificity-determin- ing residues are understood; Fig. 6) led to a clear answer. Residue Asp229 (clostripain numbering) is totally conserved and well positioned to interact with substrate Arg residues at position P1 (Fig. 6A). Even taking into account the possibility of local alignment errors, no other conserved acidic residue could be responsible. Interestingly, Asp229 is structurally positioned differently to the specificity-deter- mining Arg residues in caspases (Fig. 6B) and the Asp163 in gingipain (Fig. 6C). However, the totally conserved caspase Gln residue corresponding to Asp229 (numbered 283 in caspase-9; Fig. 3) does interact with the P1 side chain of the substrate (e.g. [33]). This provides strong additional support for our assignment of Asp229 as specificity determinant. Clostripain is known to undergo a calcium-dependent auto-activation process [1–4]. Although the details are not well understood, and it is not known if all members of the family will behave similarly in this regard, this implies the existence of a calcium-binding site on clostripain. Exam- ination of the final model revealed two suggestively positioned possibilities (Fig. 5), one positioned near the site of cleavage, the other near to the catalytic site. The first contains Glu212 and Glu237, both conserved in 12 of the 13 homologues, along with Asp215 found only in clostri- pain itself. The second site contains three acidic residues not conserved between clostripain sequences ) Glu110, Asp114 and Asp269. The residues of the first site lie within or near the central portion of the alignment which contains the catalytic dyad. Here the alignment of clostripain and the templates is particularly clear so that model quality should be good. Each site could be relevant to calcium- dependent auto-activation, the first through an effect on the site of cleavage, the second through a direct influence on the catalytic site, but the determination of which site is truly occupied will require further experiments. Since it is not known if all clostripain homologues undergo this auto- activation the conservation of the first possibility within the family does not conclusively indicate it as the likely calcium-binding site. Conclusions In this study we investigate the structure–activity relation- ship of clostripain, and its homologues in the peptidase C11 family, by affinity labelling, site-directed mutagenesis and molecular modelling. A catalytic dyad of His176 and Cys231 is definitively shown to be responsible for cysteine peptidase activity in the C11 peptidase family. However, the lack of strong perturbation of the pK a value of His176 is consistent with the two catalytic residues lying further apart than they do in papain, as indeed observed in the distantly homologous caspases. Molecular modelling revealed the likely source of clostripain substrate specificity and possible sites of binding for the calcium required for auto-activation, thus providing attractive targets for further study by site- directed mutagenesis. The domain structures of peptidase family C11 members are surprisingly diverse. Further study of the family may be facilitated by the dye based labelling of thekindusedinthiswork. References 1. Mitchell, W.M. & Harrington, W.F. (1968) Purification and properties of clostridiopeptidase B (Clostripain). J. Biol. Chem. 243, 4683–4692. 2. Dargatz, H., Diefenthal, T., Witte, V., Reipen, G. & von Wett- stein, D. (1993) The heterodimeric protease clostripain from Clostridium histolyticum is encoded by a single gene. Mol. Genl Genet. 240, 140–145. 3. Witte, V., Wolf, N. & Dargatz, H. (1996) Clostripain linker deletion variants yield active enzyme in Escherichia coli:apossible function of the linker peptide as intramolecular inhibitor of clos- tripain automaturation. Curr. Microbiol. 33, 281–286. 4.Witte,V.,Wolf,N.,Diefenthal,T.,Reipen,G.&Dargatz,H. (1994) Heterologous expression of the clostripain gene from Clostridium histolyticum. Escherichia coli and Bacillus subtilis: maturation of the clostripain precursor is coupled with self-acti- vation. Microbiology 40, 1175–1182. 5. Gunther, R., Stein, A. & Bordusa, F. (2000) Investigations on the enzyme specificity of clostripain. A new efficient biocatalyst for the synthesis of peptide isosteres. J. Org Chem. 65, 1672–1679. 6. Ullmann, D. & Jakubke, H.D. (1994) The specificity of clostripain from Clostridium histolyticum. Mapping the S’ subsites via acyl transfer to amino acid amides and peptides. Eur. J. Biochem. 223, 865–872. 7. Kembhavi, A.A., Buttle, D.J., Rauber, P. & Barrett, A.J. (1991) (1991) Clostripain: characterization of the active site. FEBS Lett. 3, 277–280. 8. Gilles, A.M. & Keil, B. (1984) Evidence for an active-center cysteine in the SH-proteinase alpha-clostripain through use of a-N-tosyl- L -lysine chloromethyl ketone. FEBS Lett. 173, 58–62. 9. Gilles, A.M., De Wolf, A. & Keil, B. (1983) Amino-acid sequences of the active-site sulfhydryl peptide and other thiol peptides from the cysteine proteinase alpha-clostripain. Eur. J. Biochem. 130, 473–479. 10. Rawlings, N.D., O’Brien, E. & Barrett, A.J. (2002) MEROPS: the protease database. Nucleic Acids Res. 30, 343–346. 11. Chen, J.M., Rawlings, N.D., Stevens, R.A. & Barrett, A.J. (1998) Identification of the active site of legumain links it to caspases, clostripain and gingipains in a new clan of cysteine endopeptidases. FEBS Lett. 441, 361–365. 12. Snipas, S.J., Stennicke, H.R. & Riedl, S., Potempa, J., Travis, J., Barrett, A.J. & Salvesen, G.S. (2001) Inhibition of distant caspase homologues by natural caspase inhibitors. Biochem. J. 357, 575–580. 13. Uhlmann, F., Wernic, D., Poupart, M.A., Koonin, E.V. & Nas- myth, K. (2001) Cleavage of cohesin by the CD clan protease separin triggers anaphase in yeast. Cell 103, 375–386. 14. Aravind, L. & Koonin, E.V. (2002) Classification of the caspase- hemoglobinase fold. detection of new families and implications for the origin of the eukaryotic separins. Proteins 46, 355–367. Ó FEBS 2004 Clostripain family structure–function relationship (Eur. J. Biochem. 271) 991 15. Small, D.A., Lowe, C.R., Atkinson, T. & Bruton, C.J. (1982) Affinity labelling of enzymes with triazine dyes. Isolation of a peptide in the catalytic domain of horse-liver alcohol dehydro- genase using Procion blue MX-R. as a structural probe. Eur. J. Biochem. 128, 119–123. 16. Labrou, N.E. & Clonis, Y.D. (1995) The interaction of Candida boidinii formate dehydrogenase with a new family of chimeric biomimetic dye-ligands. Arch. Biochem. Biophys. 316, 169–178. 17. Labrou, N.E., Eliopoulos, E. & Clonis, Y.D. (1996) Dye-affinity labelling of bovine heart mitochondrial malate dehydrogenase and study of the NADH-binding site. Biochem. J. 315, 687–693. 18. Labrou, N.E. & Clonis, Y.D. (1995) Oxaloacetate decarboxylase. on the mode of interaction with substrate-mimetic affinity ligands. Arch. Biochem. Biophys. 321, 61–70. 19. Clonis, Y.D., Stead, C.V. & Lowe, C.R. (1987) Novel cationic triazine dyes in protein purification. Biotechnol. Bioengin. 30, 621– 627. 20. Porter, W.H., Cunningham, L.W. & Mitchell, W.M. (1971) Stu- dies on the active site of clostripain. The specific inactivation by the chloromethyl ketone derived from a-N-tosyl- L -lysine. J. Biol. Chem. 246, 7675–7682. 21. Lowry, O.H., Rosebrough, N.J., Farr, A.L. & Randall, R.S. (1951) Protein Measurement with the folin phenol reagent. J. Biol. Chem. 193, 265–275. 22. Deng, W.P. & Nickoloff, J.A. (1992) Site-directed mutagenesis of virtually any plasmid by eliminating a unique site. Anal. Biochem. 200, 81–90. 23. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol. 215, 403– 410. 24. Altschul, S.F., Madden, T.L., Scha ¨ ffer, A.A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D.J. (1997) Gapped BLAST and PSI- BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402. 25. Notredame, C., Higgins, D.G. & Heringa, J. (2000) T-Coffee: a novel method for fast and accurate multiple sequence alignment. J. Mol. Biol. 302, 205–217. 26. Bujnicki, J.M., Elofsson, A., Fischer, D. & Rychlewski, L. (2001) Structure prediction meta server. Bioinformatics 17, 750–751. 27. Rychlewski, L., Jaroszewski, L., Li, W. & Godzik, A. (2000) Comparison of sequence profiles. Strategies for structural predic- tions using sequence information. Protein Sci. 9, 232–241. 28. Jones, D.T. (1999) Protein secondary structure prediction based on position-specific scoring matrices. J. Mol. Biol. 292, 195–202. 29. Bateman, A., Birney, E., Cerruti, L., Durbin, R., Etwiller, L., Eddy, S.R., Griffiths-Jones, S., Howe, K.L., Marshall, M. & Sonnhammer, E.L. (2002) The PFAM protein families database. Nucleic Acids Res. 30, 276–280. 30. Letunic, I., Goodstadt, L., Dickens, N.J., Doerks, T., Schultz, J., Mott, R., Ciccarelli, F., Copley, R.R., Ponting, C.P. & Bork, P. (2002) Recent improvements to the SMART domain-based sequence annotation resource. Nucleic Acids Res. 30, 242–244. 31. Sali, A. & Blundell, T.L. (1993) Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779–815. 32. Okamoto, Y., Anan, H., Nakai, E., Morihira, K., Yonetoku, Y., Kurihara,H.,Sakashita,H.,Terai,Y.,Takeuchi,M.,Shibanuma, T. & Isomura, Y. (1999) Peptide based interleukin-1 beta con- verting enzyme (ICE) inhibitors: synthesis, structure activity relationships and crystallographic study of the ICE-inhibitor complex. Chem. Pharm. Bull. (Tokyo). 47, 11–21. 33. Rotonda, J., Nicholson, D.W., Fazil, K.M., Gallant, M., Gareau, Y., Labelle, M., Peterson, E.P., Rasper, D.M., Ruel, R., Vaillancourt, J.P., Thornberry, N.A. & Becker, J.W. (1996) The three-dimensional structure of apopain/CPP32, a key mediator of apoptosis. Nat. Struct. Biol. 3, 619–625. 34. Renatus, M., Stennicke, H.R., Scott, F.L., Liddington, R.C. & Salvesen, G.S. (2001) Dimer formation drives the activation of the cell death protease caspase 9. Proc. Natl Acad. Sci. USA 98, 14250–14255. 35. Eichinger,A.,Beisel,H.G.,Jacob,U.,Huber,R.,Medrano,F.J., Banbula, A., Potempa, J., Travis, J. & Bode, W. (1999) Crystal structure of gingipain R.: an Arg-specific bacterial cysteine proteinase with a caspase-like fold. EMBO J. 18, 5453– 5462. 36. Laskowski, R., MacArthur, M., Moss, D. & Thornton, J. (1993) PROCHECK: a program to check stereochemical quality of protein structures. J. Appl. Crystallog. 26, 283–290. 37. Sippl, M.J. (1993) Recognition of errors in three-dimensional structures of proteins. Proteins 17, 355–362. 38. Kleywegt, G.J. (1996) Use of non-crystallographic symmetry in protein structure refinement. Acta Cryst. D52, 842–857. 39. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. (1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta. Cryst. A47, 110–119. 40. DeLano, W.L. (2002) The PyMOL Molecular Graphics System on World Wide Web http://www.pymol.org. 41. Frishman, D. & Argos, P. (1995) Knowledge-based protein sec- ondary structure assignment. Proteins. 23, 566–579. 42. Labrou, N.E. (1999) Affinity labeling of oxaloacetate decarboxy- lase by novel dichlorotriazine linked alpha-ketoacids. J. Protein Chem. 18, 729–733. 43. Gilles, A.M., Imhoff, J.M. & Keil, B. (1979) alpha-Clostripain. Chemical characterization, activity, and thiol content of the highly active form of clostripain. J. Biol. Chem. 254, 1462–1468. 44. Barrett, A.J. & Rawlings, N.D. (2001) Evolutionary lines of cysteine peptidases. Biol. Chem. 382, 727–733. 45. Theodorou, L.G., Lymperopoulos, K., Bieth, J.G. & Papamichael, E.M. (2001) Insight into the catalysis of hydrolysis of four newly synthesized substrates by papain: a proton inventory study. Biochemistry 40, 3996–4004. 46. Stennicke, H.R. & Salvesen, G.S. (1999) Catalytic properties of the caspases. Cell Death Differ. 6, 1054–1059. 47. Wolthers, B.C. (1969) Kinetics of inhibition of papain by TLCK and TPCK in the presence of BAEE as substrate. FEBS Lett. 2, 143–145. 48. Durocher, D., Henckel, J., Fersht, A.R. & Jackson, S.P. (1999) The FHA domain is a modular phosphopeptide recognition motif. Mol. Cell. 4, 387–394. 49. Slack, F.J. & Ruvkun, G. (1998) A novel repeat domain that is often associated with RING finger and B-box motifs. Trends Biochem. Sci. 23, 474–475. 50. Kelly, G., Prasannan, S., Daniell, S., Fleming, K., Frankel, G., Dougan, G., Connerton, I. & Matthews, S. (1999) Structure of the cell-adhesion fragment of intimin from enteropathogenic Escher- ichia coli. Nat. Struct. Biol. 6, 313–318. 51. Nardini, M. & Dijkstra, B.W. (1999) Alpha/beta hydrolase fold enzymes: the family keeps growing. Curr. Opin. Struct. Biol. 9, 732–737. 52. Sanchez, R. & Sali, A. (1998) Large scale protein structure mod- eling of the Saccharomyces cerevisiae genome. Proc. Natl Acad. Sci. USA 95, 13597–13602. 53. Barton, G.J. (1993) ALSCRIPT, a tool to format multiple sequence alignments. Prot. Eng. 6, 37–40. 992 N. E. Labrou and D. J. Rigden (Eur. J. Biochem. 271) Ó FEBS 2004 . for clostripain, not for the homologues shown here. The boxed region indicates the single part of the clostripain molecular model obtained from the gingipain. details) in smaller face. The model is of the cleaved clostripain lacking the internal nonapeptide. The final residue of the resultant a-chain, Arg181, and the