INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial spore germination, defined as the irreversible loss of spore characteristics, is triggered by specific germinants and proceeds through a set of sequential steps. Spore germination is essential to allow spore outgrowth and the formation of a new vegetative cell; once triggered, it proceeds in the absence of germinants and germinant-stimulated metabolism. This fact indicates that spore germination is a process controlled by the sequential activation of a set of preexisting germination-related enzymes but not by protein synthesis (10, 26).

Among the key enzymes involved in the spore germination of Bacillus subtilis 168, Bacillus cereus IFO 13597, and Clostridium perfringens S40 are a group of cortex-lytic enzymes which degrade spore-specific cortex peptidoglycan. In the spores, at least two cortex hydrolases, spore cortex-lytic enzyme (SCLE) and cortical fragment-lytic enzyme (CFLE), are suggested to cooperatively function for cortex degradation. That is, cortex hydrolysis during germination is initiated by attack of SCLE on intact spore peptidoglycan, which likely leads to un-cross-linking of cortex peptidoglycan; this step is followed by further degradation of the polysaccharide moiety of SCLE-modified cortex peptidoglycan by CFLE (5, 6, 23, 24, 28, 29). Thus, SCLE and CFLE differ from each other in bond specificity and recognition of the morphology of the substrate. It is most likely that the in vivo activity of CFLE is regulated by its requirement for partially un-cross-linked spore cortex. On the other hand, SCLE, which acts on intact spores, needs some activation process for the expression of activity. The mechanism of activation is crucial to an understanding of bacterial spore germination.

SCLE of C. perfringens S40 is a mature form of SleC, which is synthesized at an early stage of sporulation as a precursor consisting of four domains: an N-terminal presequence (113 residues), an N-terminal prosequence (35 residues), mature enzyme (264 residues), and a C-terminal prosequence (25 residues) (24, 33, 40). During spore maturation, the N-terminal presequence and the C-terminal prosequence are sequentially processed; the resulting inactive proenzyme, with a mass of 35 kDa (termed proSCLE) and consisting of the N-terminal prosequence and a mature region which exists as a complex with the cleaved N-terminal prepeptide (termed the prepeptide-proSCLE complex) (33), is deposited on the outside of the cortex layer in the dormant spore (25). Proteolytic cleavage of the promature junction of proSCLE in the complex (the linkage between Val-149 and Val-150 of SleC) during germination generates active SCLE with a mass of 31 kDa (24, 33). The protease involved in the conversion of proSCLE to SCLE, denoted germination-specific protease (GSP), has been detected in germinated spores (40), but its enzymatic entity remains to be established.

A part of the nucleotide sequence of the gene, hereafter denoted cspC (C. perfringens serine protease C; see below), which is present just upstream of the 5' end of sleC has been reported (24). Comparison of the partial deduced amino acid sequence of the gene product with those registered in various databases suggested that the CspC sequence is homologous to that around the active center of serine proteases from Bacillus species. Bacterial structural genes are often organized into clusters that include genes coding for proteins whose functions are related (1, 17). This information raised the possibility that the cspC gene encodes a protease involved in the activation of SleC and prompted us to analyze the cspC gene in parallel with attempts to identify GSP. In this paper, we describe the characterization of GSP and the cloning of genes which are present upstream of the 5' end of sleC. GSP, a serine protease which specifically processes the N-terminal prosequence of proSCLE, was isolated as a fraction consisting of three species of proteins. Several pieces of evidence indicated that these proteins are products of three tandem genes, cspA, cspB, and cspC, which are positioned just upstream of the 5' end of sleC and encode subtilisin-like proteases with a triad active center. However, whether these proteases commit to the activation process for SleC as separate proteases or as a complex is not known at present.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and chemicals. Spores of C. perfringens S40 were used as the source of proSCLE and GSP. Chromosomal DNA was isolated from vegetative cells of the strain. The organism was cultured as described by Miyata et al. (23). Escherichia coli XL1-Blue (Stratagene Cloning Systems, La Jolla, Calif.) was used as a host for the screening library. Plasmids pUC118 and Bluescript II KS(+) (Stratagene) were used as cloning vectors. E. coli BL21(DE3) (Novagen, Inc., Madison, Wis.) and plasmid pET22b(+) (Novagen) were used as a host and as a vector, respectively, for the expression of recombinant protein. E. coli was routinely grown at 37°C in 2× YT medium (1.6% Bacto Tryptone, 1% Bacto Yeast Extract, 0.5% NaCl [pH 7.0]) with ampicillin added to 100 µg per ml for plasmid-carrying strains. Transformation was carried out according to standard protocols (13).

Preparation of spores and decoated spores. Spores, decoated spores, and the spore coat fraction of C. perfringens S40 were prepared by methods described previously (23, 24).

Assay of GSP activity. GSP activity was measured indirectly via the increase in SCLE activity after incubation of the prepeptide-proSCLE complex with GSP. The complex used as a substrate for GSP was obtained as described by Okamura et al. (33). A mixture containing 5 µl of the prepeptide-proSCLE complex (0.2 mg/ml), 134 to 115 µl of 40 mM potassium phosphate (pH 7.0), and 1 to 20 µl of GSP fractions in a total volume of 140 µl was incubated at 32°C for 3 min, and then 10 µl of decoated spore suspensions was added. This reaction mixture contained a large excess of substrates for both GSP and SCLE. The initial optical density at 600 nm (OD₆₀₀) of the mixture in a cell with a 1-mm light path was 0.18, and the decrease in OD₆₀₀ was monitored at 32°C to detect SCLE activity (20). One unit of activity was defined as a decrease in the OD₆₀₀ of 0.100 per min.

Preparation of GSP. Purification procedures were carried out at 20°C, unless noted otherwise, and GSP activity was monitored in each fractionation step as described above. Spores were germinated as described previously (23), and GSP was extracted by incubating germinated spores (4 g of packed weight) in 20 ml of 0.25 M KCl-50 mM potassium phosphate (pH 7.0) containing 0.2% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid (CHAPS) at 30°C for 2 h. After debris were removed by centrifugation (15,000 × g for 10 min at 4°C), the extract was heated at 52°C for 30 min to inactivate existing SCLE, and the supernatant (20 ml) was recovered by centrifugation (15,000 × g for 10 min at 4°C). The supernatant was concentrated threefold by using a UK-10 membrane filter (Advantec MFS, Inc., Tokyo, Japan), and 2 ml was applied to a Superose 12 column (2.0 by 31 cm; bead size, 20 to 40 µm; Pharmacia) which had been equilibrated with 0.1 M KCl-40 mM potassium phosphate (pH 7.0) containing 0.1% CHAPS. This procedure was carried out three times. Fractions containing GSP (total, 18 ml), which eluted near the voided volume of the column, were applied to a hydroxyapatite column (3.0 by 20 cm; fast flow; Wako Pure Chemicals) which had been equilibrated with 40 mM potassium phosphate (pH 7.0) containing 0.1% CHAPS. GSP did not bind to the column, while most proteins adsorbed to the column.

Effect of protease inhibitors and chemicals on GSP activity. All assays were carried out with 40 mM potassium phosphate (pH 7.0), except when we examined the effect of CaCl₂ and dipicolinic acid on activity, for which 40 mM Tris-HCl (pH 7.0) was used. Purified GSP (1.2 U; 50 µl in 0.15 M KCl-40 mM potassium phosphate [pH 7.0] containing 0.2% CHAPS or 0.15 M KCl-40 mM Tris-HCl [pH 7.0] containing 0.2% CHAPS) was mixed with 1.25 to 2.50 µl of protease inhibitors and chemicals of appropriate concentrations and the mixtures were incubated at 30°C for 1 h. Dimethyl sulfoxide was used to solubilize bestatin, E-64, pepstatin A, and PMSF, but the solubilizer included in the assay medium (maximum, 0.2%) had no effect on activity.

Cloning of the csp genes. Chromosomal DNA from C. perfringens S40 was digested with EcoRI; fragments in the size range of 6.2 to 8.0 kb were separated by agarose gel electrophoresis, purified with a Geneclean II kit (Bio 101, Inc., La Jolla, Calif.), and inserted into pUC118. The ligation mixture was used to transform E. coli XL1-Blue. The csp genes were cloned from this library with a sleC probe obtained as a 2.2-kb HindIII fragment from pCP22H (24). The primers used for inverse PCR were 5'-TCACAATCTGGAGCTACTCC-3' and 5'-TGCTTGGATTCTTCAAAGAGA-3'.

Preparation of antiserum against recombinant protein r _1-78CspC. To prepare a polyclonal antibody against CspC, a recombinant CspC-His fusion was purified. The plasmid used to synthesize this protein was constructed by PCR amplification of a part of the cspC gene, encoding Ser-79 to Thr-582, from pCP70E using primers containing NdeI and XhoI restriction sites. The NdeI- and XhoI-digested PCR product was ligated into the NdeI and XhoI restriction sites of vector pET22b(+) to generate plasmid p1-78CspC, resulting in a sequence encoding processed CspC with a C-terminal six-histidine tag. Plasmid p1- 78CspC was introduced into E. coli BL21(DE3), and production of His-fused recombinant protein r_1-78CspC was induced with 1 mM isopropyl-1-thio--galactopyranoside. The resulting fusion protein obtained from an inclusion body was dissolved in 6 M urea and purified using His · Bind kits (Novagen) as recommended by the supplier. Polyclonal antibody against the purified fusion protein was raised in mice as described previously (24).

Nucleotide sequencing and analysis. Nucleotide sequencing was performed by the dideoxynucleotide chain termination method of Sanger et al. (35) using an ABI PRISM 310 automatic sequencer and an ABI PRISM BigDye Terminator cycle sequencing kit (Applied Biosystems, Foster City, Calif.). Nucleotide and amino acid sequence analyses and comparisons with DNA and protein sequences registered in databases (GenBank, EMBL, PIR, and SWISS-PROT) were performed with MacDNASIS software (Hitachi Software Engineering, Tokyo, Japan).

Other procedures. Protein concentrations were measured according to the methods of Bradford (4) and/or Groves et al. (12) using bovine serum albumin as a standard. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was carried out on slab gels using a Laemmli buffer system (19) at a constant current of 20 mA. Immunoblotting and immunoprecipitation were performed by use of anti-r_1-78CspC antiserum as described previously (24). Analysis of the N-terminal amino acid sequence was carried out with a protein sequencer (Procise cLC model 494; Applied Biosystems) according to the method of Matsudaira (22).

Nucleotide sequence accession number. The nucleotide sequence data reported in this paper will appear in the DDBJ, EMBL, and GenBank nucleotide sequence databases under accession number AB042154.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Purification of GSP. GSP was extracted from germinated spores in 0.25 M KCl-50 mM KH₂PO₄ (pH 7.0) containing 0.2% CHAPS. The presence of CHAPS (>0.05%) was needed for the stabilization of enzyme activity. Although most anion exchangers tested retained GSP even at 1 M KCl over the pH range of 6.5 to 8.5, the enzyme could be eluted from a DEAE-5PW column with 0.15 M KCl at pH 6.0, which is an acidic pH limit used to retain enzyme activity. The addition of dithiothreitol (0.2 mM) or EDTA (1 mM) or both in the extraction and purification steps had no effect on the activity. On this basis, the purification of GSP was performed under the conditions described in Materials and Methods.

View larger version (27K): [in this window] [in a new window]   FIG. 1.   SDS gel electrophoretic patterns of the purified active enzyme fraction (A) as well as the prepeptide-proSCLE complex and -, -, and -caseins treated or not treated with GSP (B). (A) An aliquot (1 ml) of the main peak fraction containing GSP activity which eluted from a DEAE-5PW column was dialyzed against H2O, dried by using of rotary evaporator, dissolved in 40 µl of 10 mM Tris-HCl (pH 8.0) containing 1% SDS and 1% 2-mercaptoethanol, and electrophoresed on a 0.1% SDS-10% polyacrylamide gel. (B) The prepeptide-proSCLE complex (lanes 1 and 2), -casein (lanes 3 and 4), -casein (lanes 5 and 6), and -casein (lanes 7 and 8) (5 to 10 µg each) were incubated with (odd-numbered lanes) or without (even-numbered lanes) GSP (1.5 U) in 40 mM potassium phosphate (pH 7.0) at 30°C for 4 h and electrophoresed on a 0.1% SDS-13.3% polyacrylamide gel. In lanes 1 and 2, 35- and 18-kDa bands are proSCLE and prepeptide, respectively. The standard proteins run were rabbit muscle phosphorylase b (97 kDa), bovine serum albumin (66 kDa), ovalbumin (45 kDa), bovine carbonic anhydrase (31 kDa), soybean trypsin inhibitor (22 kDa), and lysozyme (14 kDa).

Characteristics of GSP. GSP was active over a pH range of 5.8 to 8.5 in the presence of 0.1% CHAPS (at least 2 months at 4°C and pH 7.0) and was stable after heat treatment at 55°C for 40 min at pH 7.0. However, the enzyme lost its activity (>80%) within 72 h at 4°C in the absence of CHAPS. Detergents such as Triton X-100, nonaethyleneglycol n-dodecyl ether, octylglucoside, and deoxycholate, which are often used to stabilize membrane proteins but dissociate aggregated proteins more effectively than CHAPS (30), were ineffective in preventing the inactivation of GSP.

Nucleotide and deduced amino acid sequences of cspA, cspB, and cspC. An EcoRI genomic library from C. perfringens S40 was constructed in pUC118 and transformed into E. coli XL1-Blue. This library was screened with an sleC probe, and a positive clone carrying a 6.9-kb EcoRI fragment was isolated (pCP70E in Fig. 2). A 6.4-kb EcoRI-AccI region of this fragment was sequenced, revealing the presence of four open reading frames (ORFs); 5'-truncated cspA, cspB, and cspC and 3'-truncated sleC. The sequence of the 5' end of the cspA gene was determined by sequencing of an amplicon obtained by inverse PCR after HindIII chromosomal DNA digestion and religation. A 5,989-bp DNA sequence encompassing the complete cspA, cspB, and cspC genes was determined. Figure 3 shows the nucleotide sequence of nucleotides 1 to 4,648 and the corresponding predicted amino acid sequence. The sequence of nucleotides 4,643 to 5,989 determined in this study agreed completely with the sequence reported by Miyata et al. (24).

View larger version (11K): [in this window] [in a new window]   FIG. 2.   Physical map of a csp gene cluster of C. perfringens. The ORFs and direction of transcription are indicated by the arrows. The solid lines below the ORF arrows represent the inserts of recombinant plasmids pCP22H, from which the sleC probe was isolated, and pCP70E, from which the nucleotide sequence was derived. Restriction sites are indicated as A (AccI), E (EcoRI), Hc (HincII), Hd (HindIII), Nc (NcoI), Nd (NdeI), P (PvuII), and S (SacI). FPCR is a fragment of DNA cloned by reverse PCR. Sequence data for the HindIII-HindIII region marked by asterisks are given in Fig. 3.

View larger version (131K): [in this window] [in a new window]   FIG. 3.   Nucleotide sequence and deduced amino acid sequence of a 4,648-bp HindIII-HindIII fragment containing cspA, cspB, and partial cspC genes. The deduced amino acid sequences of cspA (nucleotides 171 to 1,907), cspB (nucleotides 1,923 to 3,620), and part of cspC (nucleotides 3,623 to 4,648) are given below the nucleotide sequences (See reference 24 for the 3'-end sequence of cspC and the whole sequence of sleC). The numbers for the nucleotides and amino acids are shown on the right. An inverted repeat is indicated by arrows. Putative ribosome binding sites (rbs) with moderate complementarity to the 3' end of the 16S rRNA of C. perfringens (11) are underlined. The asterisks indicate termination codons. The amino acid sequences of CspA, CspB, and CspC which agreed with the N-terminal sequences of GSP-57, GSP-52, and GSP-60, respectively, are underlined in bold.

View larger version (53K): [in this window] [in a new window]   FIG. 4.   Sequence alignment (A) and dendrogram analysis (B) of Csp proteins and several subtilisin-like serine proteases. (A) Only the sequences around three canonical residues (asterisks) and an oxyanion binding site (number sign) of the proteases are aligned. Identical residues in more than five proteins are indicated by black boxes. Bold underlining indicates the N terminus of mature or putative mature enzymes. Amino acid residue numbers given in parentheses indicate the first and last residues of amino acid alignments represented. The proteases are as follows: Apy, hyperthermostable protease from Aquifex pyrophilus (7); Aqua, mesothermophile alkaline protease from Thermus aquaticus YT-1 (18); Carls, subtilisin Carlsberg from Bacillus licheniformis (15); ISP, intracellular alkaline protease from Thermoactinomyces sp. strain HS682 (39); PrtP, cell envelope-located protease from Lactobacillus paracasei (14); and Pyro, hyperthermostable protease pyrolysin from Pyrococcus furiosus (41). (B) Dendrogram of Csp proteins and proteases listed above, in which branch lengths are in inverse proportion to the degree of sequence similarity. The tree was constructed using the programs ClustalX and NJPLOT (38).

GSP-57, GSP-52, and GSP-60 are processed forms of CspA, CspB, and CspC, respectively. The 16 amino acid residues starting from Glu-78 of CspA, the 35 residues starting from Ala-97 of CspB, and the 31 residues starting from Ser-79 of CspC corresponded to the N-terminal sequences of GSP-57, GSP-52, and GSP-60, respectively. The results unequivocally indicate that GSP-52, GSP-57, and GSP-60 are processed forms of CspA, CspB, and CspC, respectively. The data suggest that the Csp proteases are produced in a form possessing propeptides, with the promature junctions in CspA, CspB, and CspC being similar to each other (Fig. 3). Thus, the predicted molecular weights of the processed GSP proteins are 51,442 for GSP-52, 55,458 for GSP-57, and 55,585 for GSP-60. Percent identities among the putative mature protease sequences are 34% between GSP-52 and GSP-57, 54% between GSP-57 and GSP-60, and 39% between GSP-52 and GSP-60.

View larger version (32K): [in this window] [in a new window]   FIG. 5.   Immunological detection of CspC-related proteins in dormant spores. The spore coat fraction was separated from decoated spores as described in Materials and Methods. The decoated spores (1 g of wet weight) were disrupted with a bead beater in a 20-ml tube containing 10 ml H2O and 10 g of glass beads (diameter, 0.1 mm). The supernatant obtained by centrifugation (15,000 × g for 5 min at 4°C) was used as an extract of disrupted decoated spores. GSP obtained by DEAE-5PW column chromatography, the extract obtained from germinated spores by incubation in CHAPS solution, the spore coat fraction, and the extract of disrupted decoated spores were subjected to 0.1% SDS-10% PAGE followed by immunoblot analysis with anti-r1-78CspC antiserum. Lanes 1, GSP obtained by DEAE-5PW column chromatography (~5 µg of protein); 2 and 4, spore coat fraction (~100 µg of protein for lane 2 and ~70 µg of protein for lane 4); 3, extract from germinated spores (~100 µg of protein); 5, extract of disrupted decoated spores (~100 µg of protein). Samples for lanes 1 and 2 were electrophoresed on a separate gel different from that used for lanes 3 to 5. The standard proteins were the same as those used for Fig. 1.

Location of GSP-60 in dormant spores. The coat fraction was separated from decoated spores as described previously (24). Anti-r_1-78CspC antiserum was used to detect GSP-60 in the spore fractions. As shown in Fig. 5, the antiserum cross-reacted only with a 60-kDa polypeptide in the extract from germinated spores incubated in CHAPS solution (lane 3) as well as purified GSP (lane 1). These findings unequivocally show that the antiserum specifically recognizes GSP-60. GSP-60 was detected in the coat fraction from dormant spores (Fig. 5, lanes 2 and 4) but not in the extract from decoated spores (lane 5), suggesting a peripheral location of GSP-60 in dormant spores. GSP activity was not detected in the coat fraction under reduced alkaline pH, nor was it regained by the addition of CHAPS after dialysis against neutral buffer.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The present results provide direct evidence demonstrating the involvement of GSP in the process of activation of SleC during the germination of C. perfringens spores. It has been confirmed that the germination protease of B. subtilis and Bacillus megaterium is involved in degrading small acid-soluble proteins during spore germination (32, 34). However, germination protease is not responsible for a germination-triggering reaction, in the sense that the enzyme hydrolyzes the so-called storage proteins of spores. Although the involvement of trypsin-like proteolytic activity in the triggering reaction has been suggested by the arrest of germination of B. cereus and B. subtilis spores in the presence of trypsin inhibitors (2, 3), several of the inhibitors used are rather hydrophobic and relatively high concentrations of the inhibitors (on the order of millimolar or more) are needed to interrupt germination. It should be noted that hydrophobic compounds such as alcohols, fatty acids, and surfactants in such concentrations are known to exert an inhibitory effect on spore germination (31, 42, 43). In this regard, the results obtained by using protease inhibitors in vivo, without consideration of possible inhibitory effects on germination due to the hydrophobic nature itself, should be viewed with some caution, since neither the relevant proteases nor their genes have been characterized. Moreover, it was reported that the inactive proform of a germination-specific cortex-lytic enzyme from spores of B. megaterium KM is processed to release the active enzyme (9), but the processing enzyme has not been identified.

GSP activity eluted from a DEAE-5PW column as a single peak which consisted of three protein components, GSP-52, GSP-57, and GSP-60. There was no disulfide linkage among these proteins. We have detected active GSP as a large molecular fraction, as suggested by its elution near the void volume in Superose 12 column chromatography. A comparison of the N-terminal amino acid sequences of the isolated proteins with the deduced amino acid sequences of Csp proteins unequivocally demonstrated that GSP-57, GSP-52, and GSP-60 are hydrolyzed forms of CspA, CspB, and CspC, respectively. Anti-r_1-78CspC antiserum recognized only GSP-60. It is known that genes for enzymes with related activities are often organized into clusters in bacteria; indeed, tandem cspA, cspB, and cspC genes were aligned just upstream of the 5' end of the sleC gene. The failure of immunoprecipitation of active GSP by anti-r_1-78CspC antiserum may imply that epitopes are masked in the multiprotein complex. The percentage of charged amino acids (Asp, Glu, Arg, and Lys) in each GSP protein is over 20%. It has been shown that salt bridges between charged amino acids are a major factor stabilizing protein structure (7). It is likely, therefore, that intermolecular electrostatic interactions contribute to the formation of the active multiprotein enzyme with a relatively high thermal stability. At present, however, whether all three of the proteins in the active enzyme fraction or only one of the proteins is involved in the process of activation of proSCLE is unknown. In addition to the functional characteristics of each GSP protein, knowledge of the gross conformation and compositional stoichiometry of the active GSP molecule obtained here is needed to understand the mechanism of GSP function.

Isolated GSP failed to digest different caseins. Furthermore, the enzyme attacked only the Val-Val linkage of the promature junction of proSCLE (residues 149 and 150 of SleC) and not neighbor Val-Val bonds (residues 155 and 156 and residues 161 and 162 of SleC), which lie within a prolyl cluster region between Pro-131 and Pro-167 that contains 10 proline residues (24). GSP is not responsible for processing of the N-terminal prepeptide and the C-terminal propeptide of SleC during sporulation (33). It appears that the processing of the extended sequences of SleC during sporulation and germination of C. perfringens spores is tightly regulated by proteases which have a rigorous substrate specificity and are specific to the time of development of the spores.

Extraction of active GSP from germinated spores was enhanced by the addition of CHAPS, a nondenaturing detergent (30). Immunochemical studies using anti-r_1-78CspC antiserum revealed that GSP-60 is liberated from dormant spores by treatment of the spores with 0.1 M H₃BO₃ (pH 10.0) containing 1% 2-mercaptoethanol. This treatment also detaches from spores proSCLE and CFLE, which have been shown to be located on the exterior of the cortex layer (25). These observations suggest the localization of GSP to a hydrophobic and peripheral region of dormant spores, such as the outer membrane, which lies underneath the spore coat. Close localization of cortex degradation-related enzymes is pertinent to the pursuit of a cascade of cortex hydrolytic reactions during germination.

The results suggested that Csp proteins belong to a family of subtilisin-like proteases, based on DNA sequence analysis. This family of proteases is characterized by an N-terminal extended prosequence which functions as the intramolecular chaperone to assist correct folding of the protease domain (8, 36). It is likely, therefore, that the N-terminal extended sequences of Csp proteins also act as intramolecular chaperones. Unlike the situation for all known Bacillus exoproteases, which are synthesized as preproenzymes carrying both a signal peptide and a propeptide (37), there is no amino acid alignment showing the characteristics of a signal peptide-like sequence in the N-terminal extensions of the Csp proteins. In the cortex-lytic enzyme of Bacillus species, SleB, which is produced in the forespore compartment of sporulating cells, a signal sequence is needed for translocation across the forespore inner membrane to deposit forespore-synthesized proteins on the exterior side of the cortex (27). Furthermore, together with our observation that CspC is synthesized at stages II to III of sporulation (unpublished results), it seems that the Csp proteins are produced in the mother cell compartment of sporulating cells. When GSP proteins are activated in the cell by processing of the prosequences and the mechanism of processing remain to be studied.

GSP proteins possess the following biochemical properties. The heat stabilization effect of added Ca²⁺ on protease activity, as observed for subtilisins (21), was defective in GSP. Each of the proteins that copurified with the GSP activity had four cysteine residues, one of which was conserved (Cys-519 of CspA, Cys-504 of CspB, and Cys-524 of CspC). Since GSP was inactivated by the addition of sulfhydryl reagents, at least one of the residues might be a free cysteine which is a functionally essential sulfhydryl group. The predicted pI values were 5.02 for GSP-52, 5.53 for GSP-57, and 4.85 for GSP-60. The acidic nature might account for the observed strong adsorption of the enzyme to anion exchangers.

Although not completed, the nucleotide sequences of genomic DNAs of Clostridium acetobutylicum and Clostridium difficile are now available from Genome Therapeutics Corporation (http://www.cric.com/sequence center) and The Sanger Centre (http://www.sanger.ac.uk/Profect/C difficile/), respectively. Computer research using these databases suggested the presence of sleC and csp homologs in both genomes. Tandem alignment of the csp and sleC genes observed in C. perfringens S40 appears to be conserved in the C. acetobutylicum genome. The csp homolog of C. acetobutylicum (termed csp_Ca), encoding a putative serine protease, is found upstream of the 5' end of the sleC homolog, and the product of csp_Ca has double catalytic triads. It seems that the csp_Ca gene product is a bifunctional protein with two subtilisin-like protease domains or, alternatively, might undergo proteolytic processing after expression to generate two proteases. The third csp gene is missing in C. acetobutylicum. Unlike the genome organization of both C. perfringens and C. acetobutylicum, the C. difficile sleC homolog is found at a distance from two csp homologs. Although these genome sequences are incomplete, it seems that SleC- and Csp-like proteins are conserved among the clostridia.