Purification and characterization of kumamolysin, a novel thermostable pepstatin-insensitive carboxyl proteinase from Bacillus novosp. MN-32.

We have found a novel type of thermostable, pepstatin-insensitive carboxyl proteinase in the culture filtrate of Bacillus novosp. MN-32. The carboxyl proteinase, which was named kumamolysin, was purified about 8,300-fold by column chromatography including DEAE-Sepharose CL-6B, Sephadex G-100, and TSKgel DEAE-5PW. The purified kumamolysin gave a single band corresponding to 41 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The molecular mass of kumamolysin was estimated to be 40 kDa by gel filtration. The isoelectric point of kumamolysin was estimated to be pH 3.5 by isoelectric focusing. Kumamolysin has maximum proteolytic activity at 70 degrees C and at pH 3.0. Kumamolysin specifically hydrolyzed the Leu15-Tyr16 peptide bond in oxidized insulin B-chain (Km = 9.0 x 10(-5) M, Kcat = 71 s-1; at pH 3.0, 30 degrees C), and additional cleavage at Phe25-Tyr26 was detected at a considerably lower rate. Kumamolysin is insensitive to the known carboxyl proteinase inhibitors pepstatin, diazoacetyl-DL-norleucine methyl ester, and 1,2-epoxy-3-(p-nitrophenoxy)propane. Kumamolysin has no similarity to the thermostable acid protease thermopsin from Sulfolobus acidocaldarius (Lin, X.-L., and Tang, J. (1990) J. Biol. Chem. 265, 1490-1495). Thus, the substrate specificity, the inhibitor sensitivity, the molecular mass, and the thermostability all suggest that kumamolysin is a novel type of carboxyl proteinase.

Tracing the historical progress of the investigations on carboxyl proteinases, it is noteworthy that various specific compounds have served as inestimable tools in elucidating the detailed catalytic machinery of carboxyl proteinases, i.e. the development of the active site-directed reagents, diazoacetyl-DL-norleucine methyl ester (l), and 1,2-epoxy-3-(pnitrophenoxy)propane (2). Furthermore, the discovery of naturally occurring potent and specific inhibitors (pepstatin Ac (3) and pepstatin Isoval (4)) permitted the identification of catalytic residues by detailed biochemical or crystallographic approaches. It is well established that the catalytic sites of most carboxyl proteinases are composed of 2 active aspartic * This work was supported by a grant-in-aid for scientific research from the Ministry for Education, Science, and Culture, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
8 To whom correspondence and reprint requests should be addressed.
acid residues. Thus these enzymes have now been called aspartic proteinases. Most carboxyl proteinases, irrespective of their origin, are inactivated by diazoacetyl-DL-norleucine methyl ester, 1,2-epoxy-3-(p-nitrophenoxy)propane, and pepstatin. These inhibitors have been utilized as diagnostic reagents to deduce the nature of a newly isolated aspartic proteinase, and the inhibitor sensitivity is considered to be a criterion in enzyme nomenclature. Generally, the individual members of the carboxyl proteinase family share structural and functional similarity obviously different from serine type proteinases that form a strikingly diverse group, such as the trypsin, a-chymotrypsin, and elastase families. The distinctive susceptibility of individual carboxyl proteinase to the inhibitors may be reflected in the different nature of their active sites. The natural occurrence of inhibitor-insensitive carboxyl proteinases was established by the present authors in microbes: first from Scytalidium lignicolum in 1972 (5), and followed by Lentinus edodes (6), Ganoderma lucidum (7), Pseudomonas 101 (8), and Xanthomonm T-22 (9). Consequently, we have proposed that at least two types of carboxyl proteinase may be distinguished: pepsinlike enzymes, which are inhibited by pepstatin, diazoacetyl-DL-norleucine methyl ester, and 1,2-epoxy-3-(p-nitrophen-0xy)propane; Scytalidium-like enzymes, which are not inhibited by one or more of these three compounds (10). The occurrence of thermostable carboxyl proteinase is rare, and the thermostable enzyme, insensitive to these inhibitors, may be an attractive model for the investigation of structure and function in these classes of enzymes. In this paper, we describe the occurrence of a thermostable carboxyl proteinase in Bacillus novosp. MN-32, which is insensitive to pepstatin, diazoacetyl-DL-norleucine methyl ester, and 1,2-epoxy-3-(p-ni-trophen0xy)propane. A novel carboxyl proteinase, which we have named kumamolysin, has been isolated and characterized.

RESULTS
Production and Purification of Kumamolysin-Kumamolysin was an extracellular enzyme produced by a bacterium MN-32 strain at late stationary growth phase. From the taxonomic characterization, MN-32 was considered to be a new member of the genus Bacillus.
' Portions of the paper ("Experimental Procedures," part of "Results," Tables I and 11, and Figs. 1-5) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

349
Novel Thermostable Carboxyl Proteinase Kumamolysin was purified from the culture supernatant of B. novosp. MN-32 by the procedure described in the Miniprint Section. Typical elution profiles on preparative HPLCZ using a TSKgel DEAE-5PW column are shown in Fig. 1. The results of the purification procedure are summarized in Table I. The enzyme was purified 8300-fold with 14% recovery; 22 mg of lyophilized enzyme preparation was obtained from 120 liters of culture supernatant. The purified kumamolysin showed a single protein band on polyacrylamide gel electrophoresis (Fig. 2).
Some Characteristics of Kumamolysin-The effect of pH on the proteolytic activity and the stability of kumamolysin was investigated. The optimum pH for the proteolytic activity of kumamolysin on casein is pH 3.0 (incubation temperature: 70 "C). Below pH 1.0 and above pH 6.0, no activity was observed (Fig. 3A). As shown in Fig. 3B, kumamolysin was completely stable in the pH range 2.0-4.0 at 50 "C for 24 h, and in the pH range 1-10 at 5 "C for 3 days. The effect of temperature on the proteolytic activity and the stability of kumamolysin was examined. As shown in Fig. 4 A , the optimum temperature was 70 "C (incubation pH was 3.0). Kumamolysin was stable below 70 "C in a 10-min incubation, retaining 60% of the original activity at 80 "C for 10 min, but lost its proteolytic activity after 10 min at 90 "C ( Fig. 4B).
The molecular mass of kumamolysin, estimated by SDSpolyacrylamide gel electrophoresis under both native or reducing conditions, was 41 kDa (Fig. 5 B ) . Based on gel filtration by Sephadex G-100, the molecular mass of the enzyme was estimated to be 40 kDa (Fig. 5A).
The isoelectric point of the enzyme was determined by isoelectric focusing using carrier ampholine (pH 3-10). A single peak of proteinase activity and absorbance at 280 nm was detected at pH 3.5 (data not shown).
The UV absorption spectrum of a kumamolysin solution (50 mM sodium acetate buffer, pH 4.0) showed an absorption minimum at 250 nm, maximum a t 276 nm, and shoulders a t around 282 and 290 nm. The specific absorption coefficient, Ai:,,, 280 nm, the absorbance at 280 nm of a 1% solution of kumamolysin in a 1-cm optical path length, was 12.6. The amino acid composition of kumamolysin is shown in Table 11. Kumamolysin is not a glycoprotein as judged by the periodate-Schiff reagent or the phenol-sulfuric acid reagent. The absence of free thiol groups in kumamolysin, assayed with Ellman reagent, and the presence of 4 mol of halfcystine/molecule indicated the presence of two disulfide bridges/mol of kumamolysin.
Susceptibility of Kumamolysin to Various Proteinase Inhibitors-The effect of various proteinase inhibitors on the proteolytic activity of kumamolysin was determined. Kumamolysin was not inhibited by the tested inhibitors, including carboxyl proteinase inhibitors pepstatin, DAN, and EPNP; serine proteinase inhibitors PMSF, SSI, MAPI (inhibitor of a-chymotrypsin type and/or thiol proteinase), leupeptin (trypsin and/or thiol proteinase), and chymostatin (a-chymotrypsin and/or thiol proteinase); thiol proteinase inhibitors IAA, thiolstatin, and PCMB; and metallo proteinase inhibitors EDTA, o-phenanthroline, and talopeptin.
The optimum pH (as shown in Fig. 3A) of the proteolytic activity of kumamolysin indicated that the enzyme may be classified as belonging to the carboxyl proteinase group. The insensitivity of kumamolysin to the inhibitors specific for carboxyl proteinases was examined in detail. In the case of pepstatin, typical carboxyl proteinases (such as porcine pepsin) were completely inhibited by pepstatin in an equimolar dose response mode (enzyme:inhibitor = 1:1, molar ratio). As shown in Fig. 6A, kumamolysin was not inhibited by pepstatin a t a large molar excess (approximately 1300-fold molar excess). In the case of DAN, treatment of kumamolysin with a 210-fold molar excess of DAN did not affect the activity of kumamolysin (Fig. 6 B ) . And as shown in Fig. 6C, EPNP did not inhibit kumamolysin. In the positive control experiments, inhibition of porcine pepsin by these active-site-directed inhibitors was confirmed (shown as closed circles in Fig. 6, A-C ) .
Substrate Specificity of Kumamolysin-The cleavage site of oxidized insulin B-chain by kumamolysin was determined. Reversed-phase HPLC analysis of the final reaction products from the oxidized insulin B-chain following kumamolysin hydrolysis indicated the formation of only four products. Under the optimum reaction conditions (70 "C, pH 3.0), the hydrolysis rate was so fast that the reaction process could not be followed. Even at lower incubation temperature (37 "C), the substrate was completely hydrolyzed within 20 s at a mass ratio of substrate:enzyme of 200:l. Accordingly, the time course of the reaction was followed under suboptimum conditions: incubation temperature, 20 "C; substrate:enzyme ration 100,OOO:l (by mass ratio); pH 3.0. As shown in Fig. 7A, two peptides, designated 2 and 3, were predominant during the initial incubation time. After prolonged incubation times, product 3 was gradually hydrolyzed to yield products 1 and 4 ( Fig. 7 B ) . After the disappearance of the oxidized insulin Bchain peak, no additional hydrolysis products were detected in further incubations. These peptides were isolated by preparative HPLC, and their structures were deduced by amino acid analysis, NH2terminal amino acid identification, and FAB-MS. As shown in Table 111, the corresponding sequence in the original substrate was elucidated as follows; product 1 (residues 26 to 30), product 2 (1-15), product 3 (16-30), product 4 (16-25). From these results, the main cleavage site was deduced to be Led5-Tyr'', and PheZ5-Tyrz6 bond was very slowly attacked. The scissile bond in the substrate was shown to be unique, compared to various carboxyl proteinases hitherto reported.
The predominant cleavage at the Le~'~-Tyr'' peptide bond was studied kinetically. The hydrolysis rate was estimated by measuring peptides 2 and 3 by reversed-phase HPLC (external standard method). From the Lineweaver-Burk plot, K , was 9.0 X M, and Kc,, was 71 s-', at 30 "C, pH 3.0   Matching sequence in TyrZ6-Pbel-Tyr"-T y Pthe substrate Ala3" Leu" Ala3" PheZ5 "Analysis was performed on a 24-h hydrolyzate, using constant boiling HCI a t 110 "C.

Glu-His-Leu-CySO:,H-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-
ND, not determined. Further Evidence for Identification of Kumamolysin as a Carboxyl Proteinase-To clarify the assignment of kumamolysin to the carboxyl proteinase group, the following two independent experiments were performed. First, the pH dependence of the kumamolysin-catalyzed cleavage of oxidized insulin B-chain (hydrolysis at Leu''-Tyr"j bond) was investigated kinetically. As shown in Fig. 8, the pH profile of the ratio K J K , of kumamolysin is bell-shaped, which has been taken to indicate the presence of two catalytically important prototropic groups, having pK, values of 1.97 and 3.47. Second, to investigate the participation of carboxyl groups in the catalytic site of kumamolysin, binding of zinc(I1)-PAD, an active-site probe for carboxyl proteinases, was studied spectrophotometrically. The zinc(I1)-PAD complex had an absorption maximum at around 530 nm (PAD: 2.91 X

M;
ZnSO4: 2.00 X M, in 50 mM sodium acetate buffer, pH 5.0). An addition of kumamolysin (4.65 X M ) to the zinc(I1)-PAD complex solution increased the absorbance ( b i 6 3 0 = 0.032). This spectral change indicated the formation of a ternary complex, the zinc(I1)-PAD-kumamolysin. In the case of carboxyl proteinases, such as porcine pepsin, the zinc(I1)-PAD complex was considered to bind the two carboxyl groups of the catalytic site. Addition of pepstatin to the zinc(I1)-PAD-kumamolysin solution caused no spectral changes. In the positive control experiment using pepstatinsensitive carboxyl proteinase (such as pepsin), a distinct spectral change caused by addition of pepstatin was observed, approaching the initial absorption spectrum of the zinc(I1)-PAD complex. The same spectral change was observed by the addition of a 16-fold molar excess amount of a substrate (oxidized insulin B-chain) to the zinc(I1)-PAD-kumamolysin solution.

DISCUSSION
In this paper, we have described the isolation and the characterization of a novel type of carboxyl proteinase from the culture filtrate of Bacillus novosp. MN-32, newly isolated from hot-spring water of the Aso volcano region (Kumamoto, Japan). This enzyme, named kumamolysin, is a new member of the pepstatin-insensitive carboxyl proteinase group, which has been reported by the present authors (10). However, it exhibits several characteristics quite different with respect to thermostability, substrate specificity, and inhibitor sensitivity.
Thermostability-Carboxyl proteinases reported up to the present, including those of animal and microbial origin, are relatively thermolabile. Generally, in the case of microbial carboxyl proteinases, about 50% inactivation of enzyme activity was observed after 10-15-min incubation at 60-65 "C within the stable pH region. The upper limit of stable temperature was as follows; 50 "C for the carboxyl proteinase of Aspergillus saitoi (conditions: pH 3.0, 15 min) (24); 55 "C for Cladosporium sp. (pH 2.5, 10 min) (25); 47 "C for Lentinus edodes (pH 4.0, 30 min) (26); 45 "C for the A enzyme of Scytalidium lignicolum (pH 4.0, 15 min) (27). A relatively thermostable carboxyl proteinase was reported in Penicillium duponti; 65 "C at pH 4.5 for 15 min (28). At the time when we found kumamolysin in 1988, kumamolysin was the only thermostable carboxyl proteinase reported. However, recently, there was a report of an extremely thermostable carboxyl proteinase from a thermophilic archaebacterium (29).
Substrate Specificity-Generally, carboxyl proteinases show a broad substrate specificity, which has been characterized by a preference for aromatic or bulky amino acid residues at the cleavage site (-Pl-Pl'-). Hydrolysis by kumamolysin of the oxidized insulin B-chain was restricted to only two peptide bonds, Leul6-Tyr" and Phe25-Tyr?6. The hydrolysis of the former peptide bond was exceptionally fast (Kcat = 71 s-', at suboptimum temperature (30 "C)). This specificity is one of the unique characteristics of kumamolysin not yet reported in the microbial carboxyl proteinase group.
Inhibitor Sensitiuity-The proteolytic activity of kumamolysin was not inhibited by all the tested proteinase inhibitors, including serine, metallo, thiol, and carboxyl proteinase inhibitors. This insensitivity to the inhibitors was quite unique, when compared to the hitherto reported carboxyl proteinases, the exceptions being the carboxyl proteinase from mesophilic microorganisms, L. edodes (26), S. lignicolum A enzyme (lo), and G. lucidum (7), which were also not inhibited by pepstatin, DAN, and EPNP, as reported by us previously. Accordingly, kumamolysin is the fourth case of the natural occurrence of a unique inhibitor-insensitive carboxyl proteinase.
As described in the introductory statement, we propose that carboxyl proteinases may be classified into two groups. Group I: this is the well known typical proteinases (now classified as aspartyl proteinases); e.g. pepsin, cathepsin D, and carboxyl proteinases of various microorganisms. These are inhibited by specific inhibitors of carboxyl proteinases, such as pepstatin, EPNP, and DAN. Group 11: this group of enzymes is insensitive to these inhibitors, as described above. It has been well established that the active site of pepsin is composed of 2 aspartic acid residues (Asp3' and Aspz1'). Their carboxyl side chains are sufficiently close to each other that they may share a proton (oxygen-oxygen distance are assumed to be 2.9 A). The former aspartic residue is susceptible to esterification by DAN in the presence of copper; the latter residue is reactive to EPNP. Irrespective of origin, these two catalytic residues are retained in group I enzymes in a close geometric arrangement. The insensitivity of kumamolysin to the two activesite-directed reagents, DAN and EPNP, suggests striking differences in the active site. The participation of the carboxyl groups in the catalytic action of kumamolysin was deduced from the following experimental results. The pH dependence of the kinetic constant for kumamolysin was examined by using oxidized insulin B-chain as substrate (cleavage at Leu"-Tyr" bond). The pH profile of the Kcat/K,,, ratio (as shown in Fig. 8) indicated two ionizing group with pK. values of 1.97 and 3.47. These values were comparable to that of S. lignicolum acid proteinase A (pKel = 1.53; pKe2 = 3.77) (30), which is a member of group I1 enzyme.

Klotz et al. (31)
pointed out from the change in the visible absorption spectrum that the azo dye PAD which could not be bound to pepsin by itself, could be bound in the presence of certain transition metal ions such as zinc(I1). Based on these results, we (32) showed that all the tested acid (carboxyl) proteinases also exhibited the same spectral changes as pepsin at pH 5. However, lysozyme (hen egg white), amylases (bacterial a-amylase and fungal glucoamylase), and neutral or alkaline proteinases such as a-chymotrypsin or subtilisin did no show any such a spectral change at pH 5. The spectral change of zinc(I1)-PAD caused by carboxyl proteinases is very similar to that by oxalic acid. It is reasonable to consider that the spectral change of zinc(I1)-PAD on reaction with oxalic acid is due to the formation of the mixed complex, as shown in Scheme I, in which the two carboxyl moieties of oxalic acid are located in a close geometric arrangement. The distance between the two oxygen atoms of the carboxyl moiety, which could form a ternary c!mplex with the zinc(I1)-PAD reagent, was estimated to be 3 A (33). This value is comparable to the distance of two active carboxyl groups (Asp3' and Aspz1') in pepsin, assumed from the crystallographic investigations. The spectral change of zinc(I1)-PAD, caused by the addition of carboxyl proteinases (such as pepsin), is the same as that observed with oxalic acid. An identical spectral change was observed for all the tested carboxyl proteinases, even in the case of group I1 carboxyl proteinases. It is quite reasonable to consider that the zinc(I1)-PAD reagent as an excellent and specific active site probe of carboxyl proteinases, suggesting the presence of two adjacent catalytic carboxylate groups in ',Zn,+*' . . .

NLN
N(CH:,), SCHEME I Novel Thermostable Carboxyl Proteinase 353 the active site (33) in a similar geometric arrangement as oxalic acid. As described under "Results," the presence of two adjacent carboxyl groups in kumamolysin was deduced from the zinc(I1)-PAD experiment. The difference in absorbance due to the zinc(I1)-PAD-kumamolysin ternary complex formation was 0.032 at 530 nm (kumamolysin, 4.65 X

M).
This value was considered to be significant, compared to the following data (34): 0.058 for pepsin (2.46 X M); 0.043 for carboxyl proteinase A2 from s. lignicolum (2.17 X M); 0.136 for carboxyl proteinase B from S. lignicolum (2.13 X M), the latter two enzymes are members of group 11. The value of the absorbance increase was characteristic of the individual enzymes, which may depend on differences in the active site of each enzyme. The absorbance increase with kumamolysin was not observed at 37 "C (data not shown). When the temperature was raised to 50 "C, the absorbance increase was observed. As described under "Results," the optimum temperature for kumamolysin was 70 "C, using casein as substrate. The absorbance increase with kumamolysin ( A A 5 3 0 = 0.034; E = 4.65 X M), observed at suboptimum temperature (50 "C), may be rather small compared to other carboxyl proteinases.
In the above-mentioned zinc(I1)-PAD experiment, the possibility of participation by cysteine and/or histidine residues in the active site of kumamolysin may be anticipated. (About cysteine: in the presence of H F ion, PAD was used as a titrant for sulfhydryl groups in protein (35).) Kumamolysin contains 4 cysteine residues, as shown in Table 11. These residues were deduced to form two disulfide bridges in kumamolysin, estimated by Ellman reagent titration in the reductive or nonreductive condition. The activity of kumamolysin was not inhibited by thiol proteinase inhibitors such as PCMB, IAA, and thiolstatin. Accordingly, the participation of cysteine residues in the spectral change of the zinc(I1)-PAD reagent was considered to be improbable. (About histidine: from the two independent experimental results, 1) activity-pH profile (optimum pH of kumamolysin was pH 3.0); 2) pH-KCat/Km profile (pKel = 1.97; pKe2 = 3.47), the allocation of histidine residue to the active site was considered to be improbable.) Addition of the substrate (oxidized insulin B-chain) to the zinc(I1)-PAD-kumamolysin complex revealed that bound zinc(I1)-PAD was liberated from the ternary complex. Accordingly, it is reasonable to assume that the location of the two reactive carboxyl groups in kumamolysin are at the active site. We also reported the presence of two carboxyl groups in the active site of S. lignicolum carboxyl proteinase (A and B enzyme) by the same spectrophotometric method, using angiotensin I as the substrate (34) and the pH dependence of the ratio K,.,/Km (30). Together with these results, this assumption was confirmed from another line of evidence. A naturally occurring carboxyl proteinase inhibitor, pepstatin, which is a transition-state analog inhibitor, had no effect in liberating the dye probe from the zinc(I1)-PAD-kumamolysin complex. Pepstain showed no inhibitory effect on kumamolysin, and interaction withlor binding to kumamolysin was not observed by a difference spectrophotometric experiment (data not shown). Inability of pepstatin to liberate the dye probe from the active site of kumamolysin was considered to be consistent with the lack of inhibitory effect. Finally, we concluded that kumamolysin is apparently related to the carboxyl proteinase group 11.
One of the unique characteristics of kumamolysin is its substrate specificity. In general, irrespective of enzyme origin and inhibitor susceptibility, the substrate specificity of Carboxyl proteinases was shown to be relatively broad compared to serine type proteinases. For example, hydrolysis of oxidized insulin B-chain by porcine pepsin occurs is largely at His''-Leu'', Va1'2-Glu'3, Ala'4-Leu'5, TyrI6-Led7, and PheZ4-Phez5 (see Ref. 36). The cleavage sites of various carboxyl proteinases, determined with the same substrate, was shown to be in two main regions which were located around the sequence A1a'4-Le~'5-Tyr'6-Le~17 and PheZ4-Phez5, where amino acids with large hydrophobic side chains were clustered. However, additional cleavage sites were characteristic of individual enzymes, as most carboxyl proteinases showed multiple attack sites, usually more than five, in the oxidized insulin B-chain. Nevertheless, as shown in Fig. 7 and Table 111, the cleavage sites of kumamolysin were restricted to only two peptide bonds, Le~'~-Tyr'' and PheZ5-Tyrz6. This unique substrate specificity closely resembled those of the aspartic proteinase from pig intestinal mucosa (37). However, they apparently differ in their susceptibility to pepstatin, DAN, and E P N P i.e. pig intestinal mucosa aspartic proteinase was inhibited by these inhibitors. The absence of kinetic investigations on the hydrolysis reaction of various aspartic proteinases using oxidized insulin B-chain prevents a direct comparison with that of kumamolysin. However, the rate constant of kumamolysincatalyzed hydrolysis of the Le~'~-Tyr'' bond was remarkably high (Kc,, = 71 s-I), even though the reaction was done at suboptimum temperature (30 "C). Detailed experiments on the substrate specificity of kumamolysin are now in progress. The kinetic constant of newly synthesized peptides containing the sequence around the scissile bond in the oxidized insulin B-chain, was evaluated; H-Val-Glu-Ala-Leu-Tyr-Leu-OH was an equally favorable substrate (hydrolysis point: -Leu-Tyr-), but deletion of the NHz-terminal valine residue (corresponding to P,) lowered the rate constant to one-hundredth of that for the parent peptide, and further deletion of the NHzterminal glutamic acid residue (corresponding to P3) resulted in no hydrolysis (results not shown). These preliminary experiments indicate the active-site cleft of kumamolysin is a rather expanded one.
Even in the case of group I aspartic proteinases, some differences in inhibition rate by pepstatin, acyl-Val-Val-Sta'-Ala-Sta-OH, are always observed with individual enzymes. This difference may be explained by the energy of interaction, derived from the Sta' residue to the 2 catalytic aspartic acid residues and probable interaction to the binding sites (38). The complete insensitivity of kumamolysin to pepstatin probably also indicates the different nature of the subsites in the active-site cleft of kumamolysin as compared to pepsin. Favorable substrates for various aspartic proteinases (39), such as H-Lys-Pro-Ala-Lys-Phe-Nph-Arg-Leu-OH, are extremely poor substrates for kumamolysin (results not shown). Therefore, it is likely that the subsite length and/or the nature of the active-site cleft in kumamolysin are quite different from various other aspartic proteinases.
Recently, a thermostable "acid proteinase," named thermopsin, was reported (29). However, kumamolysin is obviously unrelated to thermopsin based on inhibitor sensitivity, substrate specificity, amino acid composition (as shown in Table II), absence of carbohydrate, and other characteristics.
Kumamolysin is the fourth case of the natural occurrence of carboxyl proteinases of a bacterial origin; i.e. kumamolysin, thermopsin (29), and two pepstatin-insensitive carboxyl proteinases from mesophilic bacteria (Xanthomonas sp. and Pseudomonas sp.), the latter two enzymes were reported by previously the present authors (8,9). These microorganisms are the most ancient organism that produce carboxyl proteinases. Comparative investigations of these bacterial enzymes may serve as a beginning in the understanding of carboxyl