INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Between 10 and 20% of total spore protein is degraded to amino acids in the first minutes of germination of spores of Bacillus species (30). The proteins degraded are a group of small, acid-soluble proteins (SASP) unique to the spore stage of the life cycle. The /-type SASP, which are coded for by a multigene family, are bound to the dormant spore's DNA and provide a significant component of spore resistance to heat, hydrogen peroxide, and UV radiation (28, 30, 31). The -type SASP, which are coded for by a single gene, are also in the spore core (the site of spore DNA) but are not bound to any macromolecule. SASP degradation at the beginning of spore germination both frees up the DNA for transcription and provides amino acids that are used for protein synthesis during subsequent development (22, 30).

SASP degradation during spore germination is initiated by one or two endoproteolytic cleavages catalyzed by a sequence specific germination protease termed GPR which is synthesized during sporulation at about the same time as its SASP substrates (30). GPR is synthesized as a protein of 46 kDa as measured by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE); this species, termed P₄₆, is inactive (13). Two hours later in sporulation, P₄₆ undergoes intramolecular autoprocessing which removes an N-terminal propeptide of 15 (Bacillus megaterium) or 16 (B. subtilis) amino acids; the new, fully active form of GPR is called P₄₁ (13, 24). Both P₄₆ and P₄₁ are tetramers, and only the tetrameric form of P₄₁ is active (12). The autoprocessing of P₄₆ to P₄₁ is triggered both in vivo and in vitro by low pH, dipicolinic acid, and dehydration (8). Although active in vitro, P₄₁ carries out no SASP degradation during sporulation because the conditions inside the developing and dormant spore (in particular the dehydration and mineralization) are not favorable for enzymatic action. However, in the first minutes of spore germination, the spore core rehydrates, allowing rapid attack of GPR on SASP. Synthesis of GPR in an inactive form and its autoprocessing at the correct time in sporulation are essential to generate a fully resistant spore; a strain with a GPR variant that autoprocesses P₄₆ to P₄₁ ~1 h earlier during sporulation produces spores that are more sensitive than wild-type spores to a variety of treatments, because the /-type SASP content of the mutant spores is reduced (6).

In addition to the autoprocessing reaction converting P₄₆ to P₄₁, the P₄₁ of B. megaterium also undergoes an additional autoprocessing to P₃₉ with loss of an additional seven N-terminal residues (8). The enzymatic activities of P₃₉ and P₄₁ are identical, and P₃₉ is not generated from P₄₁ of B. subtilis. Consequently, the generation of P₃₉ with B. megaterium GPR may not have any functional significance.

While the role played by GPR in SASP degradation as well as its autoprocessing are reasonably well understood, little is known of the structural differences between P₄₆ and P₄₁ and the nature of GPR's active site(s). Indeed, the precise class of protease to which GPR belongs has not been established. To obtain more information on these latter topics, we have analyzed the trypsin digestion products generated from P₄₆ and P₄₁ and have used a variety of techniques in an attempt to determine the class of protease to which GPR belongs. The latter data suggest that GPR is not a member of a previously described class of proteases.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacteria and plasmids used and isolation of DNA. The bacterial strains and plasmids used in this work are listed in Table 1. Escherichia coli strains were routinely grown at 37°C in 2× YT medium (5 g of NaCl, 16 g of tryptone, and 10 g of yeast extract per liter). Plasmid DNA was isolated from E. coli with the QIAprep Spin Miniprep kit (Qiagen).

Site-directed mutagenesis. We used the megaprimer-based PCR method (26) to make mutations in B. megaterium gpr in which Ser²¹⁹ or Ser²²⁹ is replaced by Ala (Fig. 1A). The oligonucleotide primers used for the first round of PCR were A (5'-CAAGCTCTAATACGACTCAC-3'), B (5'-CATCCGGGGGCTGGGGTTGGAAAC-3'), and C (5'-GCGTAAAGAAATCGCTATGAAACACTTG GC-3'). Oligonucleotide A is in the T7 promoter of the pTZ19U portion of pPS1907 (Table 1); oligonucleotide B begins at nucleotide (nt) 753 and ends at nt 776 and oligonucleotide C begins at nt 779 and ends at nt 809 of the coding sequence of B. megaterium gpr. Oligonucleotides B and C are complementary to the gpr sequence except for the underlined nucleotides. In the first round of PCR, the reaction mixture (100 µl) contained 100 pmol of primer B or C and 100 pmol of primer A, 1.5 mM MgSO₄, 0.2 mM each deoxynucleoside triphosphate, 200 ng of template (plasmid pPS1907), and 10 µl of 10× Vent DNA polymerase buffer. Reaction mixtures were overlaid with mineral oil and heated for 8 min at 94°C, 1 U of Vent DNA polymerase was added, and samples were subjected to 35 cycles of PCR (1 min at 94°C, 1 min at 55°C, and 1 min at 72°C) and incubated for 10 min at 72°C at the end of the last cycle. The synthesized megaprimers were analyzed by electrophoresis on a 1.2% agarose gel to verify that they were of the expected sizes and then used for the second round of PCR. The second round of PCR was carried out in 50 µl containing ~50 ng of one of the megaprimers, 500 pmol of primer D (5'-GTAAAACGACGGCCAGTG-3'), which is complementary to the sequence of pTZ19U in pPS1907 downstream of the multiple cloning site, 1.5 mM MgCl₂, 0.2 mM each deoxynucleoside triphosphate, 200 ng of HindIII-linearized pPS1907 purified by agarose gel electrophoresis, and 5 µl of 10× Taq DNA polymerase buffer. Samples were treated as described above, 1 U of Taq DNA polymerase was added, and PCR was performed as described above. The final PCR products were subjected to agarose gel electrophoresis; fragments with the expected sizes were excised and ligated into plasmid pCRII (Invitrogen). The ligation mixture was used to transform E. coli INVF' cells (Invitrogen), and transformants were selected on 2×YT agar plates containing ampicillin (50 µg/ml) and 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (40 µg/ml). Plasmids were isolated from white colonies, and digestion with HincII and HindIII identified plasmids with the appropriate ~1.5-kb insert. DNA sequence analysis confirmed the presence of the desired mutations and that they were the only mutations present in the 211-bp ClaI-EagI fragment of B. megaterium gpr (Fig. 1A). This ClaI-EagI fragment was purified by agarose gel electrophoresis and ligated to ClaI-EagI-digested pPS1910 which had been purified to obtain the large fragment. The ligation mixture was used to transform E. coli UT481 to ampicillin resistance, and plasmid DNA from several clones was isolated and sequenced to ensure that gpr had acquired the desired mutation. E. coli UT481 strains carrying plasmids with the gpr S219A mutation and the gpr S229A mutation were called PS2577 and PS2578, respectively.

View larger version (31K): [in this window] [in a new window]   FIG. 1.   Locations of restriction sites in B. megaterium gpr and possible catalytic residues in B. megaterium GPR (A) and amino acid sequences of B. megaterium GPR variants (B). (A) The locations of restriction sites and codons for possible catalytic residues in B. megaterium gpr were taken from reference 30. The locations of codons for possible catalytic residues are given above the gene (open box), and restriction sites (C1, ClaI; E1, EagI; H2, HincII; H3, HindIII; NcI, NcoI; NrI, NruI) are indicated below the gene. N and C denote the N- and C-terminal coding regions, respectively. Note that the HincII and HindIII sites are outside the gpr gene's coding sequence. The arrow labeled P39 denotes the region in which P41 autoprocesses to P39. (B) Amino acid sequences (in one-letter code) of various forms of GPR. The sequences of P46 and P41 (note that this is P41 overexpressed in E. coli) are from references 6 and 24. The sequence of P39 is from this work, and those of the variants lacking 3 to 12 propeptide residues are from reference 17. The numbers above the sequences are the positions of lysine or arginine residues in P46; arrows labeled P41 and P39 denote the sites of cleavage generating P41 and P39, respectively.

Protein overexpression and purification. B. megaterium P₄₆, P₄₁, and the GPR variants lacking 3, 6, 9, or 12 propeptide residues were purified, respectively, from E. coli JM83 carrying plasmid pPS740, pPS1910, pPS2339, pPS2340, or pPS2341 and from E. coli UT481 with plasmid pPS2342 as described previously (6, 17). All other strains for overexpression of GPR were grown at 37°C in 2×YT medium plus ampicillin (50 µg/ml). At an optical density at 600 nm of 0.9, isopropyl--D-thiogalactopyranoside was added to 1 mM, and the cells were harvested after 2 h of further growth at 37°C. GPR was routinely purified from these cells as described (6, 17).

Limited proteolytic digestion. GPR (0.5 to 1 mg/ml) was incubated at 37°C in 50 mM Tris-HCl (pH 7.4)-20% glycerol-5 mM CaCl₂ with tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin (Worthington) at a GPR/trypsin ratio of 1,000:1 (wt/wt). The inclusion of glycerol in the digestion buffer was essential to obtain stable functional products, as in the absence of glycerol, both P₄₆ and P₄₁ were rapidly degraded to small fragments. Reactions were stopped by addition of 2× electrophoresis sample buffer and boiling for 5 min, and aliquots were analyzed by SDS-PAGE on a 15% gel. Proteins were transferred to polyvinylidene difluoride membranes and stained with Coomassie blue; protein bands were analyzed by automated amino acid sequencing and matrix-assisted laser desorption-time-of-flight mass spectrometry (MALDI-TOF) as described elsewhere (11, 17).

GPR assay and autoprocessing. GPR was assayed as described previously (12). When trypsin-digested P₄₆ or P₄₁ was assayed, soybean trypsin inhibitor was added to digests in a 2:1 molar ratio with trypsin prior to GPR assays. When crude extracts were to be assayed, GPR was overexpressed as described above and the cells were harvested by centrifugation. The cells from ~50 ml of culture were resuspended in 2 ml of 50 mM Tris-HCl (pH 7.4)-5 mM CaCl₂-20% glycerol, disrupted by sonication, and centrifuged at 12,000 × g for 15 min. The protein concentration of the supernatant fluid was determined by the method of Lowry et al. (14).

Effect of protease inhibitors on GPR activity. Two different procedures were used to test the effects of protease inhibitors on the activity of purified P₄₁. To test the effect of 1,10-orthophenanthroline (Ophen), P₄₁ (0.2 mg/ml) was dialyzed overnight at 10°C against 50 mM Tris-HCl (pH 7.4)-5 mM CaCl₂-20% glycerol-2 mM Ophen prior to enzyme assays (1). To test the effects of 250 µM N-acetylimidazole (NAI), 1 mM 3,4-dichloroisocoumarin (DCI), 10 mM disopropylfluorophosphate (DFP), 10 mM 8-hydroxyquinoline-5-sulfonic acid (HQSA), 1 mM iodoacetic acid (IAA), 1 mM N-ethylmaleimide (NEM), 10 µM pepstatin, and 3 or 10 mM phenylmethylsulfonyl fluoride (PMSF), the inhibitors were routinely incubated in 50 mM Tris-HCl (pH 7.4)-5 mM CaCl₂-20% glycerol with 20 to 200 µg of purified P₄₁ per ml. Incubation was for 4 h at 23°C for NAI (the buffer used was 25 mM MOPS [pH 7.4]), 2 h at 4°C for DCI (with no glycerol in the incubation), 45 min at 37°C for DFP [50 mM 3-(cyclohexylamino)-1-propanesulfonic acid (pH 10) was used with this inhibitor in addition to Tris-HCl], 30 min at 37°C for HQSA, 1 h at 23°C for IAA, overnight at 23°C for NEM, 1 h at 23°C for pepstatin (with 20 µg of P₄₁ per ml), and 45 min at 23°C for PMSF. Aliquots of the incubation mixtures were assayed for GPR activity as described above both before and after incubation, and in parallel with P₄₁ samples incubated similarly but without inhibitors. In some cases the discontinuous colorimetric assay for GPR (12, 29) was used, as some inhibitors inhibit the aminopeptidase used in the GPR assay.

Metal analysis. GPR for metal analysis was purified and concentrated as described above and then dialyzed exhaustively against 10 mM Tris-HCl (pH 7.4)-10% glycerol-1 mM CaCl₂ in otherwise metal-free water (Milli-Q); this dialysis buffer had <2 µM Mg²⁺, Mn²⁺, Zn²⁺, Co²⁺, or Fe²⁺. Standard solutions were made with the dialysis buffer as the diluent, and metal analyses were carried out on a Leeman PS 1000 ICP atomic absorption spectrometer. The instrument was calibrated, and standards were checked to verify the calibration.

CD spectroscopy, thermal melting, and guanidine hydrochloride (GuHCl) denaturation. Circular dichroism (CD) spectra were recorded with an AVIV 62DS spectropolarimeter interfaced to a personal computer. Spectra were collected at 20°C for solutions containing 1.9 mg of P₄₁ or 1.6 mg of P₄₆ per ml in 10 mM Tris-HCl (pH 7.4)-5 mM CaCl₂-10% glycerol-2 mM EDTA. The glycerol and CaCl₂ were present to ensure protein stability. The CD spectra were measured every 0.5 nm with 2-s averaging per point and a 2-nm bandwidth; a 0.01-cm-path-length cell was used. Spectra were signal averaged by adding five scans, and the baseline was corrected by subtracting a spectrum for the buffer obtained under identical conditions. The temperature was maintained by a Lauda RS2 circulating water bath, and the cell temperature was measured with a thermosensor. The protein concentration in this experiment was determined from the UV absorption at 280 nm, using calculated molar absorption coefficients of 5,960 mol¹ cm¹ for P₄₁ and 7,450 mol¹ cm¹ for P₄₆ (14).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Proteolytic digestion of GPR. Previous work has indicated that the only covalent change in conversion of P₄₆ to P₄₁ is removal of 15 (B. megaterium) or 16 (B. subtilis) amino-terminal residues (24). To further probe the structures of both P₄₆ and P₄₁, we digested these proteins with trypsin and analyzed both the sequences and the activities of the digestion products. Trypsin rapidly converted P₄₆ to a 30-kDa species (termed P₃₀) (Fig. 2) that exhibited no further degradation after 24 h (data not shown). The amino-terminal sequence of P₃₀ was ELDLS, and its molecular mass measured by MALDI-TOF was 28,798.1 Da. These data indicate that P₃₀ is generated by trypsin cleavage after K³ and K²⁶⁸ (Fig. 1A), giving a polypeptide with a calculated mass of 28,756.9 Da, in good agreement with the value determined experimentally.

View larger version (36K): [in this window] [in a new window]   FIG. 2.   Digestion of P46 with trypsin. Purified P46 was incubated with trypsin as described in Materials and Methods. At various times, aliquots (~10 µg of protein) were withdrawn and subjected to SDS-PAGE (15% gel), and the proteins were stained with Coomassie blue. The GPR samples run in lanes 1 to 6 were digested for 0, 2, 5, 15, 30, and 60 min, respectively. Molecular weight markers run in parallel (not shown) indicated that the major band in lane 6 is ~30 kDa, and this band was analyzed for both molecular weight and N-terminal amino acid sequence.

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Autoprocessing of the stable P30 domain of trypsin-digested P46. P30 was generated and incubated under autoprocessing conditions as described in Materials and Methods; an aliquot (10 µg) was precipitated with trichloroacetic acid, resuspended in SDS sample buffer, and subjected to SDS-PAGE (12% gel). Lane 1, unprocessed P30; lane 2, autoprocessed P30. Arrows a and b denote the migration positions of P30 and the autoprocessed product, respectively.

View larger version (37K): [in this window] [in a new window]   FIG. 4.   Digestion of P41 with trypsin. Purified P41 was incubated with trypsin as described in Materials and Methods, and at various times aliquots (10 µg of protein) were subjected to SDS-PAGE (15% gel). The numbers below the lanes give the percentages of the initial activity of the enzyme assayed as described in Materials and Methods; the value for undigested P41 was set at 100%. The GPR samples in lanes 1 to 6 run were digested for 0, 2, 5, 15, 30, and 60 min, respectively. The two bands in lane 4 were taken for analysis of molecular weight and N-terminal protein sequence. Molecular weight markers run in parallel (not shown) indicated that the latter two bands ran at 29 and 27 kDa.

Analysis of P₄₆ and P₄₁ structures. The differences in sulfhydryl reactivity and trypsin sensitivity of P₄₆ and P₄₁ (7) indicate that these proteins differ somewhat in structure; this variance could reflect a difference in secondary structure, in tertiary structure, or in both. We used CD spectroscopy to assess the secondary structure content of P₄₁ and P₄₆. The two proteins had very similar far-UV CD spectra, including characteristics found in proteins with  helices (Fig. 5). Analysis of the spectra between 190 and 260 nm by the method of Chang et al. (3) indicated that P₄₁ contained 21% ± 5% -helix structure, 58% ± 5%  sheet, and 21% ± 5% random coil; for P₄₆, the corresponding values were 19% ± 5%, 51% ± 5%, and 28% ± 5%.

View larger version (19K): [in this window] [in a new window]   FIG. 5.   CD spectra of P41 and P46. The far-UV spectra of P41 (1.9 mg/ml) and P46 (1.6 mg/ml) were recorded and analyzed as described in Materials and Methods. Arrows 1 and 2 denote the spectra of P46 and P41, respectively.

View larger version (15K): [in this window] [in a new window]   FIG. 6.   Thermal unfolding of P46 and P41. The thermal unfolding of P46 and P41 was determined by analysis of CD at 220 nm as described in Materials and Methods. Curves 1 and 2 denote the melting points of P46 and P41, respectively.

Analysis of GPR residues essential for catalysis. The removal of ~30% of GPR residues without loss of enzyme activity or the capacity for autoprocessing indicates that the residues essential for GPR catalysis must reside in the remaining 70% of the protein. Comparison of the amino acid sequences of GPR from B. megaterium and B. subtilis has previously failed to reveal any consensus sequence for any known class of protease (32). This analysis can now be extended further (Fig. 7), as the sequence of GPR from Clostridium acetobutylicium has recently become available from Genome Therapeutics Corporation (Waltham, Mass.) through the Internet at www.cric.com. Analysis of these three protein sequences shows that the sequence in the P₄₆P₄₁ cleavage site is the same in all three proteins, but the clostridial protein has a much shorter propeptide. Again, there are no conserved signature consensus sequences for aspartic or cysteine proteases, metalloproteases, or serine proteases, the four main classes of proteases (2, 19-21). However, comparison of these three sequences does give information on residues that may or may not play a role in catalysis. Thus, the single cysteine residue conserved in the GPR of the two Bacillus species is replaced by a valine in the clostridial enzyme, although the latter protein contains a cysteine residue at position 147. There are also two serines that are conserved in the three GPRs; these residues are Ser²¹⁹ and Ser²²⁹ in B. megaterium GPR. Ser²⁷⁴ and Ser²⁷⁶ of B. megaterium GPR are also conserved in B. subtilis GPR but not in the C. acetobutylicum protein. While the latter enzyme has Ser²⁵⁸ in the same region which might substitute for either Ser²⁷⁴ or Ser²⁷⁶, these more carboxyl-terminal serines seem unlikely to be essential for catalysis, as they are probably removed by the trypsin digestion, generating P₃₀ and the analogous active forms generated from P₄₁.

View larger version (33K): [in this window] [in a new window]   FIG. 7.   Amino acid sequence alignment of GPRs from various species. The protein sequences were aligned by using the ClustalW program. Gaps introduced into the sequences are indicated by dashes. Positions at which identical residues are present in proteins from all three species are indicated by asterisks below the sequence. Numbers at the right indicate the last residue in that row for each protein. The labeled arrows denote the site cleaved in the generation of P41 (a), the site cleaved in generation of P39 (b), the single cysteine in GPR from Bacillus species (c), two serine residues conserved in all three proteins (d and e), the likely site of trypsin cleavage generating P30 from P46 (f), and serine residues conserved in the Bacillus proteins (g and h). Sources of the sequences: Bacillus megaterium (B. meg) (32), Bacillus subtilis (B. su) (32), and Clostridium acetobutilycum (C. acet) (the Internet at www.cric.com).

Analyses of metal ions in GPR. Previous work has shown that Ca²⁺ is required for P₄₁ stability, although Ca²⁺ is likely not required for catalytic activity (29). However, incubation of P₄₁ as described in Materials and Methods with chelators such as Ophen and HQSA, which have a low affinity for Ca²⁺ but a high affinity for other divalent cations (i.e., Zn²⁺), gave no inhibition of GPR activity (data not shown). These results suggest that GPR is not a metalloprotease, which is consistent with the absence of the classical HEXXH motif of metalloproteases from GPR (21, 32). However, it is still possible that GPR is an unusual metalloprotease with a very tightly bound divalent cation. To test this point directly, we dialyzed concentrated GPR solutions against metal-free buffer (except for the stabilizing Ca²⁺ ions) and analyzed the protein for metal ions by atomic absorption spectroscopy. Analysis of both P₄₁ and P₄₆ showed that Cd²⁺, Co²⁺, Cu²⁺, Fe²⁺, Mg²⁺, and Mn²⁺ were present at <0.03 mol/mol of enzyme subunits, while Zn²⁺ was present at <0.05 mol/mol of enzyme subunits (data not shown). These data effectively rule out the possibility that GPR is a metalloprotease.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Proteases include enzymes that have been classified into four familiesaspartic and cysteine proteases, metalloproteases, and serine proteasesas well as a number of unclassified enzymes (2, 18-21). The data in this communication clearly show that GPR is not a cysteine protease, metalloprotease, or serine protease. Presumably, the inhibition of GPR by PMSF was due to reaction with some amino acid residue other than a serine or with a serine residue essential for protein stability but not catalysis. One catalytic residue other than serine that might react with PMSF is threonine, and recently the catalytic residue in the proteasome has been identified as an N-terminal threonine residue (27). While P₄₁ does not have an N-terminal threonine residue, there are nine threonine residues that are conserved in the three known GPR sequences (Fig. 7). Thus, it is formally possible that GPR could utilize a threonine residue for catalysis. However, the proteasome is inhibited by DCI (4), and this inhibitor had no effect on P₄₁. While the latter finding suggests that GPR is not a threonine protease, it remains possible that the highly specific P₄₁ is a threonine protease; further mutagenesis will be required to test this possibility.

The facts that GPR's pH optimum is ~8 (28) and that the enzyme is not inhibited by pepstatin also suggest that GPR is not an aspartic protease, as these enzymes are generally active at much lower pH values and are inhibited by pepstatin (20). However, it is formally possible that GPR is an aspartic protease that is active at neutral pH, possibly because of the environment around the active-site aspartates, as has been suggested for several aspartic proteases (5, 25), and perhaps the high specificity of GPR's active site precludes inhibition by pepstatin. Indeed, while GPRs lack the conserved motifs found in many aspartic proteases, the sequences of GPR from the three species examined do contain seven conserved aspartate residues.

Examination of the GPR sequences for other conserved residues that might be nucleophiles in GPR catalysis reveals one conserved tyrosine, one conserved histidine, one conserved lysine, and five conserved arginines. Although none of these amino acids have been reported to be the direct catalytic residues in proteases, several (e.g., histidine and lysine) participate in catalysis, in particular in serine proteases (19). The lack of inhibition of GPR by NAI suggests that a tyrosine residue does not participate in catalysis (22), but we have as yet no direct proof of the involvement (or lack of involvement) of the other residues. Again, further mutagensis will be required to examine the role of these other residues and to definitively identify the catalytic residues in GPR.

Recently, a substrate-specific protease has been identified in the obligate intracellular pathogen Chlamydia trachomatis (9). This protease, termed EUO, appears specific for the histone-like proteins which cause chromatin condensation late in this organism's life cycle, and in this regard EUO displays some similarity to GPR in its biological action. However, EUO differs tremendously from GPR in size, sequence, and inhibition by both pepstatin and a serine protease inhibitor (9).

The present work shows clearly that only the N-terminal two-thirds of GPR is required for both catalytic activity of P₄₁ and autoprocessing of P₄₆ to P₄₁. This is shown not only by the activity and autoprocessing of the 27- to 30-kDa forms generated from P₄₁ and P₄₆ by trypsin digestion but also by the absence of 30 of the C-terminal residues in the Bacillus GPRs from the clostridial enzyme. However, the C-terminal third of the molecule is needed to impart stability at least to P₄₁. It is striking that the C-terminal regions of both P₄₆ and P₄₁ are removed by trypsin cleavage at the same bond (K²⁶⁸). The trypsin sensitivity of the bond between K²⁶⁸ and E²⁶⁹ suggests that these residues are a part of a structure, possibly a loop, that is exposed to the solvent. The loss of the 102 C-terminal residues from P₄₆ without loss of the autoprocessing activity of P₃₀ further suggests that these C-terminal residues form a domain (or domains) which is distinct from the catalytic domain of the enzyme, although the C-terminal domain may be important in stabilizing the intact enzyme. The resistance of P₃₀ to further trypsin digestion is in contrast to the lability of the 27- and 29-kDa forms generated from P₄₁ by trypsin. The degradation and instability of the latter species, the lack of reactivity of the SH group in P₃₀ (in contrast to the SH group in P₄₁), the much greater trypsin sensitivity of P₄₁ than of P₄₆, the lower melting temperature for P₄₁ than for P₄₆, and the greater susceptibility of P₄₁ than of P₄₆ to unfolding by GuHCl are further evidence for a significant difference in tertiary structure between P₄₆ and P₄₁, with the latter presumably having a less compact structure which is less stable and more accessible to DTNB and trypsin. Such a difference in the structure and stability of zymogen and active enzyme has been seen with a number of other proteases, as the zymogen's propeptide not only helps maintain the enzyme in an inactive state but also assists in the compact folding and stabilization of the protein (15). However, this difference in tertiary structure between these two forms of GPR is not reflected in any significant difference in secondary structure content.

The only major difference between P₃₀ and the larger of the two active forms of P₄₁ generated by trypsin is the presence of most of the propeptide residues in P₃₀. Previous work has shown that removal of 3, 6, or 9 propeptide residues from P₄₆ does not result in an active enzyme and does not cause an increase in the reactivity of the enzyme's SH group, while removal of 12 residues gives an active enzyme with a more reactive SH group (17). Our new data largely reinforce these earlier results, as (i) the clostridial GPR has only five propeptide residues (analogous to the 2-10 variant of B. megaterium GPR), which presumably are sufficient to maintain the enzyme as a zymogen; and (ii) trypsin digestion of the GPR variants lacking 3 or 6 propeptide residues gave results similar to those obtained upon digestion of P₄₆, while trypsin digestion of the variant lacking 12 propeptide residues was like that of P₄₁. In contrast, the 2-10 variant has essentially no enzymatic activity (17), as is the case for P₄₆, but trypsin digestion of the 2-10 variant gave results similar to those with the active P₄₁. While we cannot explain this discrepancy completely, we note that there is an arginine three residues after the N terminus of the 2-10 variant (Fig. 1B). Possibly the propeptide region in the 2-10 region is unstable enough that trypsin can cleave at this arginine residue, generating an enzyme that is identical to the 2-13 variant that is rapidly degraded by trypsin.

While these results provide further support for the importance of the propeptide in determining GPR structure, they also suggest that most of the propeptide is unlikely to be freely accessible to the solvent, as it is only in the 2-10 variant that there may have been significant internal trypsin cleavage in the propeptide. However, the N-terminal region of the propeptide must be solvent accessible, as trypsin does cleave adjacent to K³. This result, as well as the relative stability of the 2-7 variant to trypsin and a propeptide of only five residues in the clostridial GPR, indicates that only a few propeptide residues are needed to maintain GPR in the zymogen conformation. Since there is no large conformational change in the P₄₆P₄₁ conversion, a clear challenge for future work will be to elucidate the precise nature of this conformational change.