小島 修一

Three new proteinaceous inhibitors of trypsin and subtilisin of the Streptomyces subtilisin inhibitor (SSI)-like (SIL) protein family were isolated and purified from culture media of Streptomyces strains; SIL5 from S. fradiae, SIL7 from S. ambofaciens and SIL12 from S. hygroscopicus. Their complete amino-acid sequences were determined by sequence analysis of the intact SIL proteins and peptides obtained by enzymatic digestion of S-pyridylethylated proteins. SIL7 showed high sequence similarity to other Arg-possessing SSI-family inhibitors at the P1 site. SIL12 is unique in having a two-residue insertion in the flexible loop region. Based on the amino-acid sequences of these inhibitors and other SSI-family inhibitors whose sequences have already been determined, the phylogenetic relationship of SSI-family inhibitors and Streptomyces strains was considered. Among about 110 amino-acid residues possessed by SSI-family inhibitors, 28 are completely conserved. The contribution of these conserved residues to the function and stability of the inhibitor molecules is discussed on the basis of the results obtained from mutational analysis of SSI and its crystal structure.

Amino acid sequences of protease inhibitors (Streptomyces subtilisin inhibitor-like proteins) widely distributed in Streptomyces were compared to clarify the taxonomic status of three strains of Streptomyces spp., S. coelicolor A3(2), S. lividans 66 and S. coelicolor Muller, which are closely related by conventional taxonomical procedures. The sequence comparison indicated that S. coelicolor A3(2) is distinct from the type strain S. coelicolor Muller, but belongs to the same taxon as S. lividans 66.

Streptomyces subtilisin inhibitor (SSI) contains three methionine residues in a subunit: two (at positions 73 and 70) in the crucial enzyme-recognition sites P1 and P4, respectively, and one (Met 103) in the hydrophobic core. The motions of the side chains of these three Met residues and the changes in mobility on binding with subtilisin were studied by deuterium NMR spectroscopy in solution and in crystalline and powder solids. For this purpose, the wild-type SSI was deuterium-labeled at the methyl groups of all three Met residues, and three artificial mutant proteins were labeled at only one specific Met methyl group each. In solution, for methionines 73 and 70, the effective correlation times were only 0.8-1.0 x 10(-10) s indicating that the two side chains on the surface fluctuate almost freely. On formation of a complex with subtilisin, however, these high mobilities were quenched, giving a correlation time of 1.1 x 10(-8) s for the side chains of methionines 70 and 73. The correlation time of Met 103, located in the hydrophobic core, was at least 1.0 x 10(-8) s in free SSI, showing that its side chain motion is highly restricted. The nature of the internal motions of the three Met side chains was examined in more detail by deuterium NMR spectroscopy of powder and crystalline samples. The spectral patterns of the powder samples depended critically on hydration: immediately after lyophilization, the side-chain motions of the three Met residues were nearly quenched. With gradual hydration to 0.20 gram of water per gram protein-water, the orientational fluctuation of the methyl axes of methionines 70 and 73 was selectively enhanced in both amplitude and frequency (to about 1 MHz) and, at nearly saturating hydration (0.60 gram of water per gram protein-water), became extremely high in amplitude and frequency (>10 MHz). In contrast, the polycrystalline wild-type SSI spectrum showed fine structures, reflecting characteristic motions of the Met side chains. The polycrystalline spectrum could be reproduced reasonably well by the same motion models and parameters used to simulate the powder spectrum at the final level of hydration, suggesting that the side-chain motions are similar in the fully hydrated powder and in crystals. Spin-lattice relaxation measurements gave evidence that, even in crystals, the methyl axes of all three Met residues undergo rapid motions with correlation times between 10(-8) and 10(-10) s, comparable to the correlation times in solution. Finally, in the hydrated stoichiometric complex of SSI with subtilisin BPN' in the solid state, large-amplitude motions are absent, but the side chains of methionines 70 and/or 73 are likely to have small-amplitude motions.

Previously, we isolated a candidate for an endogenous target enzyme(s) of the Streptomyces subtilisin inhibitor (SSI), termed SAM-P20, from a non-SSI-producing mutant strain (S. Taguchi, A, Odaka, Y, Watanabe, and H. Momose, Appl, Environ, Microbiol. 61:180-186, 1995). In this study, in order to investigate the detailed enzymatic properties of this protease, an overproduction system of recombinant SAM-P20 was established in Streptomyces coelicolor with the SSI gene promoter, The recombinant SAM-P20 was purified by salting out and by two successive ion-exchange chromatographies to give a homogeneous band by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Partial peptide mapping and amino acid composition analysis revealed that the recombinant SAM-P20 was identical to natural SAM-P20, From the results for substrate specificity and inhibitor sensitivity, SAM-P20 could be categorized as a chymotrypsin-like protease with an arginine-cleavable activity, i,e, a serine protease with broad substrate specificity, For proteolytic activity, the optimal pH was 10.0 and the optimal temperature was shifted from 50 to 80 degrees C by the addition of 10 mM calcium ion, The strong stoichiometric inhibition of SAM-P20 activity by SSI dimer protein occurred in a subunit molar ratio of these two proteins of about 1, and an inhibitor constant of SSI toward SAM-P20 was estimated to be 8.0 x 10(-10) M. The complex formation of SAM-P20 and SSI was monitored by analytical gel filtration, and a complex composed of two molecules of SAM-P20 and one dimer molecule of SSI was detected, in addition to a complex of one molecule of SAM-P20 bound to one dimer molecule of SSI., The reactive site of SSI toward SAM-P20 was identified as Met-73-Val-74 by sequence analysis of the modified form of SSI, which was produced by the acidification of the complex of SSI and SAM-PZ0, This reactive site is the same as that toward an exogenous target enzyme, subtilisin BPN'.

The primary structure of transition protein 4 (TP4) from boar late spermatid nuclei was determined by automated Edman degradation of S-pyridylethylated protein and of peptides generated by cleavage with Staphylococcus aureus V8 protease, lysyl endopeptidase and CNBr. Boar TP4 is a basic protein consisting of a highly basic amino-terminal half (residues 1-73) and a less basic carboxy-terminal half (residues 74-138). The latter half includes a highly hydrophobic segment, a four-times tandemly repeated sequence, N(G)QNKR(K)X, and a carboxy-terminal segment containing Trp126. Ultraviolet absorption and CD spectra of TP4-rat-liver-nucleosome-core-DNA (double-stranded DNA) complexes suggest a TP4-induced local melting of DNA. Although at 1 mM NaCl TP4 brought about a slight stabilization of the DNA against thermal melting, a destabilization of the DNA was observed at 50 mM NaCl. From the results of quenching of tryptophan (Trp126) fluorescence of TP4 upon its binding to double-stranded and single-stranded boar liver nucleosome-core DNA at 50 mM NaCl, the apparent association constants for the binding of TP4 to double-stranded and single-stranded DNA were calculated to be 7.3x10(3) M(-1) and 4.1x10(3) M(-1), respectively. These results suggest that TP4, having different domain structures from TP1-3 and a higher affinity for double-stranded DNA, induces a local destabilization of DNA probably through the stacking of Trp126 with nucleic acid bases.

We determined the complete amino acid sequence of a novel subtilisin inhibitor, SIL15, which had been isolated from the culture supernatant of Streptomyces bikiniensis and shown to be a member of the Streptomyces subtilisin inhibitor (SSI)-like (SIL) protein family, and then identified its reactive site, SIL15 is composed of 113 amino acids and exists as a dimer, Compared with other SSI-family inhibitors, SIL15 was found to be unique in that it possesses a Gin residue at the P1 site of the reactive site and has two-residue insertions in two regions, one in the alpha(1)-helix and the other in the flexible loop region near the reactive site, Inhibition of subtilisin BPN' by SIL15 (inhibitor constant, 2.7X10(-11) M) was due to the presence of a Gin residue at the P1 site, which was well consistent with the results obtained for P1-site mutants of SSI and turkey ovomucoid domain 3.

A novel serine protease inhibitor SIL8, which was isolated from the culture medium of Streptomyces virginiae and shown to be a member of the Streptomyces subtilisin-inhibitor-like (SIL) inhibitor family by sequence analysis of its amino-terminal region [Taguchi, S., Kikuchi, H., Kojima, S., Kumagai, I., Nakase, T., Miura, K. and Momose, H. (1993) Biosci. Biotech. Biochem. 57, 522-524], is the first SIL inhibitor demonstrated to show marked inhibitory activity toward alpha-chymotrypsin, in addition to strong inhibitory activity toward subtilisin BPN', a common property of inhibitors of the Streptomyces subtilisin inhibitor (SSI) family. In this study, the complete amino acid sequence of SIL8 was determined from the sequence analysis of peptides obtained by specific cleavage at the reactive site and by enzymic digestion. SIL8 was shown to exist as a dimer protein, each subunit of which was composed of 111 amino acids, and to have less than 50% similarity with other SSI-family inhibitors, indicating its most distant relationship to other members of this family. Insertion of two residues was observed in the flexible loop region of SIL8, and amino acid replacements were found not only on the molecular surface but also in the beta-sheet and hydrophobic core, suggesting that packing rearrangements of the side chains may occur in these regions to maintain the tertiary and quaternary structures. The inhibitor constants K-i obtained using synthetic substrates are 92 pM for subtilisin BPN' and 11 nM for alpha-chymotrypsin. The P1 site was identified as methionine, which was in good agreement with the substrate specificity of alpha-chymotrypsin. SSI, which also possesses a methionine residue at the P1 site, inhibits alpha-chymotrypsin poorly (inhibitor constant, 4.0 mu M). Such a difference in the inhibitory properties of SIL8 and SSI toward alpha-chymotrypsin is discussed on the basis of the structures of the inhibitors.

The effect of an engineered disulfide bond between two identical subunits of a dimeric protein, Streptomyces subtilisin inhibitor, on the stability of the protein was studied by differential scanning calorimetry. The introduction of the linkage caused a large stabilization without changing the cooperativity of unfolding, with the denaturation temperature of a 2 mg/mL solution being increased by 14.3 degrees C to 95.0 degrees C at pH 9.5 and by 16.4 degrees C to 63.0 degrees C at pH 3.0. The stabilization was caused by a loss of denaturational entropy, i.e., -40 and -98 cal K-1 mol(-1) at pH 3.0 and 9.5, respectively, which more than compensated for the loss in the denaturational enthalpy.

We report here another example of renaturation of subtilisin BPN'(Sbtl) by using an immobilized preparation instead of applying a digestible mutant of Streptomyces subtilisin inhibitor (SSI), a proteinaceous inhibitor of Sbtl [M. Matsubara et al. (1994) FEBS Letters 342, 193-196]. The mature Sbtl was immobilized on agarose beads employing the amino group of the protein. After thorough washing, the immobilized Sbtl was subjected to denaturation in 6 M guanidine hydrochloride (GdnHCl) at pH 2.4 for 4 h, followed by renaturation in 2 M potassium acetate al pH 6.5 for 24 h. This denaturation/renaturation cycle was repeated five times. The recovered activity of the renatured immobilized Sbtl settled at a constant level after the third denaturation/renaturation cycle, demonstrating that almost 100% renaturation was attained by use of the immobilized Sbtl. This immobilized Sbtl preparation could well be utilized for the mechanistic study of protein folding. We then found that 2 M potassium acetate was superior to 2 M potassium chloride as a refolding medium and that the ability of SSI to induce the correct shape of the mature Sbtl was lacking in several refolding media in both thermodynamic and kinetic criteria. Thus the main cause for the increase of refolding yield of Sbtl by coexistence of SSI was prevention of the autolysis of Sbtl.

Protein proteinase inhibitors showing sequence homology with Streptomyces subtilisin inhibitor (SSI) have been found to be distributed widely in Streptomyces species, and accordingly have been named SSI-like (SIL) proteins. SIL1 from S. cacaoi was the first of these proteins to be isolated and to be given a serial number. To study the structure-function relationship of SIL proteins, we determined the primary structure of SIL1 and measured its inhibitory activities. It was found to be composed of 110 amino acids and to exist in dimer form. The amino-acid sequence of SIL1 was unique among other characterized SIL proteins in having a one-residue deletion in two regions and a three-residue insertion in the flexible loop region. Sequence comparison indicated that SIL1 was distantly related to other members of the SSI family, and that amino-acid replacements had occurred not only on the surface of the SIL1 molecule but also in the beta-sheet region. The reactive site of SIL1 was considered to be Arg(70)-Glu(71) from sequence alignment with other SSI-family inhibitors. SIL1 inhibited subtilisin BPN' strongly with an inhibitor constant (K-i) of 2.8.10(-11)M, like other members of the SSI family possessing an Arg residue at the P1 site. In contrast, SIL1 exhibited weak inhibition toward trypsin with a K-i value of 5.5.10(-8)M, possibly as a consequence of insertion of the three residues in the flexible loop region near the reactive site. This contrast seems to be due to the difference in the subsite structure of the two proteinases.

Ovomucoids are commonly present in bird egg white and exhibit inhibitory activity toward various serine proteases. To investigate the structure-function relationship of ovomucoid domain 3, we established a secretory expression system for the chicken ovomucoid domain 3 (OMCHI3)-encoding gene in Escherichia coli by ligating it downstream from the tac promoter and signal peptide of E. coli alkaline phosphatase. E. coli JM105 was transformed with the resulting plasmid and induced with 1 mM isopropyl-beta-D-thiogalactopyranoside (IPTG). The mature OMCHI3 was detected in the culture supernatant, and was purified to homogeneity by three-step chromatography. Amino-acid sequence analysis showed that processing by the signal peptidase was carried out exactly at the expected site. Measurements of circular dichroism spectra and inhibitory activity indicated that OMCHI3 was produced in the properly folded form. Furthermore, site-specific replacement of the Ala residue at the P1 site with Met or Lys resulted in acquisition of inhibitory activity toward chymotrypsin or trypsin, respectively, indicating that the P1 site is the predominant determinant for inhibitory specificity.

The pro-sequences of proteases have been considered to be required for the refolding of denatured proteases. However, here we report achievement of almost complete restoration of enzymatic activity of subtilisin BPN' in the absence of its pro-sequence. The presence of 2 M potassium acetate in the folding medium enhanced the refolding efficiency of guanidine hydrochloride (GdnHC1)-denatured subtilisin BPN' by up to 28%, and other organic salts were also found to be useful, suggesting that general contribution of the bulky hydrophobic moieties of the salts to the formation of a favorable environment required for folding. This finding will provide new insights into the folding mechanisms not only of proteases but also of various other proteins. Almost complete restoration of enzymatic activity of denatured subtilisin in the organic salt solution was accomplished by further addition of mutated Streptomyces subtilisin inhibitor (SSP), which had been converted to a digestible temporary inhibitor by removal of the disulfide bridge near the reactive site.

In order to improve a natural enzyme so as to fit industrial purposes, we have applied experimental evolution techniques comprised of successive in vitro random mutagenesis and efficient screening systems. Subtilisin BPN', a useful alkaline serine protease, was used as the model enzyme, and the gene was cloned to an Escherichia coli host-vector system. Primary mutants with reduced activities of below 80% of that of the wild type were first derived by hydroxylamine mutagenesis directly applied to subtilisin gene DNA, followed by screening of clear-zone non-forming transformant colonies cultured at room temperature on plates containing skim-milk. Then, secondary mutants were derived from each primary mutant by the same mutagenic procedure, but screened by detecting transformant colonies incubated at 10 degrees C with clear zones that were greater in size than that of the wild type. One such secondary mutant, 12-12, derived from a primary mutant with 80% activity, was found to gain 150% activity (k(cat)/K-m value) of the wild-type when the mutant subtilisin gene was subcloned to a Bacillus subtilis host-vector system, expressed to form secretory mutant enzyme in the medium, and the activity measured using N-succinyl-L-Ala-L-Ala-L-Pro-L-Phe-p-nitroanilide as the substrate. When N-succinyl-L-Ala-L-Ala-L-Pro-L-Leu-p-nitroanilide was used, 180% activity was gained. Genetic analysis revealed that the primary and secondary mutations corresponded to D197N and G131D, respectively. The activity variations found in these mutant subtilisins were discussed in terms of Ca2+-binding ability. The thermostability was also found to be related to the activity.

Three novel proteinaceous inhibitors of serine proteases which had been identified as Streptomyces subtilisin inhibitor-like (SIL) inhibitors were isolated from culture supernatant of Streptomyces; SIL2 from Streptomyces parvulus, SIL3 from Streptomyces coelicolor and SIL4 from Streptomyces lavendulae. They exhibited not only strong inhibitory activity toward subtilisin BPN' but also less strong inhibition of trypsin. Their primary sequences were determined by sequence analysis of peptides obtained by specific cleavage at the reactive site and subsequent proteolytic digestion. Each inhibitor consisted of about 110 amino acids, and was considered to form a dimer. The reactive site of the inhibitors was identified as Arg-Glu for SIL2 and SIL3, and Lys-Leu for SIL4, from sequence analysis of modified forms of the inhibitors produced from the inhibitor-subtilisin complex under acidic conditions. The presence of an arginine/lysine residue at the P1 site was in agreement with their trypsin-inhibition property. Sequence comparison with other members of the Streptomyces subtilisin inhibitor family revealed that amino acid replacements in the three isolated SIL inhibitors were frequently localized on the surface region, and many of the amino acid residues in beta-sheets and the hydrophobic core were highly conserved. Values of the inhibitor constant (K-i) toward subtilisin BPN' and trypsin were also measured, and the differences were discussed on the basis of the determined structures of the inhibitors.

A gene encoding a novel FMN-binding protein from Desulfovibrio vulgaris (Miyazaki F) was cloned, and its expression system was constructed in Escherichia coli. The 1.4-kilobase pair DNA fragment isolated from D. vulgaris (Miyazaki F) by double digestion with KpnI and SmaI was found to express a protein binding FMN as a prosthetic group under control of the lac promoter in E. coli. This DNA fragment contained several putative open reading frames. The partial amino acid sequence of the polypeptide portion of the purified FMN-binding protein and its tryptic peptides were completely consistent with those deduced from the nucleotide sequence of the third open reading frame in the cloned SmaI-SmaI fragment of D. vulgaris (Miyazaki F) DNA, which may include promoter and regulatory sequences. The nucleotide sequence of FMN-binding protein indicated that the protein is composed of 122 amino acids including an initiator Met residue and lacks a signal peptide for secretion. The main redox potential of the FMN-binding protein was measured as -325 mV using direct current cyclic and differential pulse voltammetric techniques and an electroreflectance method, suggesting that this FMN-binding protein functions as a redox protein like other FMN-binding; proteins. Immunoblot analysis of the whole proteins from D. vulgaris (Miyazaki F) clearly indicated that this protein is expressed in this bacteria. However, the protein was found to have a primary structure distinct from those of other FWN-binding proteins and to be the smallest FMN-binding protein yet reported.

The tertiary structure of proteinaceous protease inhibitors is considered to be maintained by various interactions in the molecule that prevent degradation by protease. In this study, the Arg(29) of Streptomyces subtilisin inhibitor (SSI) forming a salt bridge with the carboxyl group of carboxyl-terminal Phe(113) was replaced with Ala, Met or Lys by cassette mutagenesis to clarify the role of Arg(29) in the function of SSI. The inhibitory activity of each mutated SSI decreased with increasing incubation time after mixing with subtilisin, indicating that the SSI was changed into a temporary inhibitor upon mutation. This decrease was shown by SDS polyacrylamide gel electrophoresis to be due to cooperative degradation of the mutated SSI by subtilisin. In addition, the denaturation temperature of the Ala or Met mutant was decreased by ten degrees and that of the Lys mutant by 1.5 degrees, suggesting that the destabilization of SSI may be related to its temporary inhibition. Thus, interaction in the protease inhibitor molecule for maintaining the tertiary structure, such as that of Arg(29) in SSI, was shown to be required for the inhibitory action.

Three novel proteinaceous inhibitors of serine proteases which had been identified as Streptomyces subtilisin inhibitor‐like (SIL) inhibitors were isolated from culture supernatant of Streptomyces SIL2 from Streptomyces parvulus, SIL3 from Streptomyces coelicolor and SIL4 from Streptomyces lavendulae. They exhibited not only strong inhibitory activity toward subtilisin BPN' but also less strong inhibition of trypsin. Their primary sequences were determined by sequence analysis of peptides obtained by specific cleavage at the reactive site and subsequent proteolytic digestion. Each inhibitor consisted of about 110 amino acids, and was considered to form a dimer. The reactive site of the inhibitors was identified as Arg‐Glu for SIL2 and SIL3, and Lys‐Leu for SIL4, from sequence analysis of modified forms of the inhibitors produced from the inhibitor‐subtilisin complex under acidic conditions. The presence of an arginine/lysine residue at the P1 site was in agreement with their trypsin‐inhibition property. Sequence comparison with other members of the Streptomyces subtilisin inhibitor family revealed that amino acid replacements in the three isolated SIL inhibitors were frequently localized on the surface region, and many of the amino acid residues in β‐sheets and the hydrophobic core were highly conserved. Values of the inhibitor constant (Ki) toward subtilisin BPN' and trypsin were also measured, and the differences were discussed on the basis of the determined structures of the inhibitors. Copyright © 1994, Wiley Blackwell. All rights reserved

Three novel proteinaceous inhibitors, which had been identified as "Streptomyces subtilisin inhibitor-like (SIL) proteins" and exhibited trypsin inhibition in addition to strong inhibition toward subtilisin BPN', were purified from the culture broth of three Streptomyces strains: SIL10 from S. thermotolerans, SIL13 from S.galbus, and SIL14 from S.azureus. Their primary structures were determined by sequence analysis of intact SIL inhibitors and peptides obtained by enzymatic digestions of S-pyridylethylated SIL inhibitors. These inhibitors were composed of about 110 amino acids and existed as dimer proteins. The reactive site was identified as Lys-Gln for all three inhibitors by sequence analysis of their modified forms in which the reactive-site peptide bond was specifically cleaved by subtilisin BPN' under acidic conditions. Thus, their inhibition toward trypsin and subtilisin BPN' was due to the presence of a Lys residue at the P1 site. Inhibitor constants toward subtilisin BPN' and trypsin were also determined. These inhibitors showed relatively high sequence homology to other SSI-family inhibitors possessing a Lys residue at the P1 site, with amino acid replacements on their molecular surface. © 1994 BY THE JOURNAL OF BIOCHEMISTRY.

Streptomyces subtilisin inhibitor-like proteins were found to be distributed widely in streptomycetes by using the combination of the convenient, newly developed plate assay system and an established liquid culture assay. Almost all the strains formerly categorized as Streptoverticillium species produced proteins that exhibited inhibitory activity against both subtilisin BPN' and trypsin. N-terminal regions of three purified proteins showed high structural similarity to those of other previously reported SIL inhibitors.

An ultraviolet absorption difference spectrum characteristic of the ionization change of a tyrosyl residue was observed on the binding of subtilisin BPN' with Streptomyces subtilisin inhibitor (SSI) at alkaline pH. This difference spectrum was considered to be induced by a pK(a) shift (from 9.7 to greater than or equal to 11.5) of a tyrosyl residue of subtilisin BPN' in the interaction with carboxyls of SSI [Inouye et al. (1979) J. Biochem. 85, 1115-1126]. In the present paper, the tyrosyl residue in subtilisin BPN' and the carboxyls in SSI were identified by analyzing the difference spectrum using mutants of subtilisin BPN' and SSI: naturally occurring mutants and those prepared by site-directed and cassette mutagenesis. The difference spectrum disappeared on the binding of a mutant subtilisin BPN' of which Tyr104 was replaced by Phe (S-BPN'Y104F) and SSI at pH 9.8. The magnitude of the absorption difference was much smaller when subtilisin BPN' was bound with a mutant SSI of which both Glu67 and Asp68 were replaced by Gly than with the wild-type SSI. These lines of evidence indicated that the difference spectrum was caused by Tyr104 of subtilisin BPN' interacting with Glu67 and Asp68 of SSI. The binding of subtilisin BPN' and SSI is accompanied by an increase of tryptophan fluorescence, which is pH-dependent in the range of pH 7-11 [Uehara et al. (1978) J. Biochem. 84, 1195-1202]. In the present study, this pH-dependence of the fluorescence diminished when SSI bound with S-BPN'Y104F. This suggested that the fluorescence increase was due to Trp106 of subtilisin BPN' and was influenced by the ionization of Tyr104.

kinetic analysis was performed on the interaction between subtilisin BPN' and recombinant species of a proteinaceous proteinase inhibitor, Streptomyces subtilisin inhibitor (SSI), of which the P, site amino acid residue, Met73, was replaced by site-directed mutagenesis. The inhibitor constant, K(i), was determined from the residual enzyme activity by using a peptide substrate. The rate constant of binding, k(on), and the rate constant of dissociation, k(off), were determined from a progress curve of the substrate hydrolysis in the presence of the inhibitor by using newly derived equations. A recombinant SSI in which Met73 was replaced by Ile showed an affinity (1/K(i)) toward subtilisin BPN' of only about 7% of that of the wild-type SSI, and the kinetic analysis revealed that the increase of k(off) was responsible for this difference. The affinity of other SSI mutants in which Met73 was replaced by Glu or Asp decreased significantly as pH became increasingly alkaline. The decrease in the affinity of these recombinants was due to the decrease of k(on) rather than the increase of k(off). Stopped-flow studies revealed that the binding reaction was reconcilable with a two-step mechanism, and the kinetic parameters for each step were obtained for the binding of the enzyme and recombinant SSIs.

We have unexpectedly found that Streptomyces subtilisin inhibitor (SSI) and some other similar serine protease inhibitors produced by Streptomycetes strongly inhibit Streptomyces griseus metallo-endopeptidase II (SGMP II) [Kajiwara, K. et al. (1991) J. Biochem. 110, 350-354]. In order to elucidate the mode of their unusual interaction with SGMP II in more detail, we prepared twelve kinds of SSI analogues, in which one or two amino acid residues in the peptide segment from Thr64 to Val74 of wild-type SSI had been replaced or deleted by site-directed mutagenesis, and determined the dissociation constants of their complexes with SGMP II. Six analogues among them showed dissociation constants one order of magnitude lower than that of the wild type. Three had higher values. The results suggest that at least some residues in this segment are interacting with SGMP II in the complex. We also prepared an SSI mutant in which the disulfide bridge between Cys71 and Cys101 had been eliminated by replacing the two Cys residues with Ser residues. This mutated SSI inhibited SGMP II as strongly as the wild-type SSI did. While peptide bonds in the wild-type molecule did not suffer from the hydrolytic action of SGMP II except those at the amino-terminal fragile portion, the Pro72-Met73 bond of the mutant was specifically cleaved by the enzyme. This peptide bond, therefore, seems to play the role of the reactive site in the interaction of SSI with SGMP II.

We attempted to screen a series of Streptomyces subtilisin inhibitor-like (SIL) proteins among several Streptomyces strains by using a highly sensitive assay system established by us. Of six randomly tested strains, four were found to produce SIL inhibitors as their major secreted proteins, suggesting that they might be distributed in a high frequency among this genus. Three inhibitors exhibited inhibition of both subtilisin BPN' and trypsin. Comparison of the amino terminal sequences of these isolated proteins with those of other reported SIL inhibitors revealed that the beta1- and beta2-sheets in SSI were highly conserved.

A proteinaceous protease inhibitor was isolated from the culture broth of Streptomyces lividans 66 by a series of purification steps (salting out by ammonium sulfate, ion-exchange chromatography on DEAE-cellulose, hydrophobic chromatography on Phenyl-Sepharose, and gel-filtration on Sephacryl S-200), and was named S. lividans protease inhibitor (SLPI). The purified SLPI existed in a dimeric form consisting of two identical subunits, each of which was composed of 107 amino acids. SLPI exhibited strong inhibitory activity toward subtilisin BPN'. These features were similar to those of protein protease inhibitors produced by other Streptomyces (SSI family inhibitor). In addition, SLPI was capable of inhibiting trypsin with an inhibitor constant (K(i)) of about 10(-9) M. The primary structure of SLPI and location of two disulfide bridges were homologous to those of the other serine protease inhibitors of Streptomyces. The reactive site of SLPI was found to be Arg67-Glu68 from the sequence analysis of cleaved SLPI which was produced by acidification of subtilisin-SLPI complex. An Arg residue at the P1 site was consistent with the trypsin-inhibitory property of SLPI. Sequence comparison with other members of the SSI family revealed that amino acid replacements in SLPI were mainly localized on the surface of the SLPI molecule, and many of the amino acid residues in beta-sheets and hydrophobic core were well conserved.

Proteinase specificity of a proteinaceous inhibitor of subtilisin (SSI; Streptomyces subtilisin inhibitor) can be altered so as to strongly inhibit trypsin simply by replacing P1 methionine with lysine (with or without concomitant change of the P4 residue) through site-directed mutagenesis. Now the crystal structure of one such engineered SSI (P1 methionine converted to lysine and P4 methionine converted to glycine) complexed with bovine trypsin has been solved at 2.6 angstrom resolution and refined to a crystallographic R factor of 0.173. Comparing this structure with the previously established structure of the native SSI complexed with subtilisin BPN', it was found that (i) P1 lysine of the mutant SSI is accommodated in the S1 pocket of trypsin as usual, and (ii) upon complex formation, considerable conformation change occurs to the reactive site loop of the mutant SSI. Thus, in this case, flexibility of the reactive site loop seems important for successfully changing the proteinase specificity through mere replacement of the P1 residue.

Unlike trypsin-like serine proteases having only one conspicuous binding pocket in the active site, subtilisin BPN' has two such pockets, the S1 and S4 pockets, which accommodate the P1 and P4 residues of ligands (after Schechter and Berger notation) respectively. Using computer graphics, the geometrical nature of the two pockets was carefully examined and strategies for site-directed mutagenesis studies were set up against a protein SSI (Streptomyces subtilisin inhibitor), which is a strong proteinaceous inhibitor (or a substrate analogue) of subtilisin BPN'. It was decided to convert the P1 residue, methionine 73, into lysine (M73K) with or without additional conversion of the P4 residue, methionine 70, into glycine (M70G). The crystal structures of the two complexes of subtilisin BPN', one with the single mutant SSI (M73K) and the other with the double mutant SSI (M73K, M70G) were solved showing that (i) small 'electrostatic induced-fit movement' occurs in the S1 pocket upon introducing the terminal plus charge of the lysine side chain, and (ii) large 'mechanical induced-fit movement' occurs in the S4 pocket upon reducing the size of the P4 side chain from methionine to glycine. In both (i) and (ii), the induced-fit movement occurred in a concerted fashion involving both the enzyme and 'substrate' amino acid residues. The term 'substrate-assisted stabilization' was coined to stress the cooperative nature of the induced-fit movements.

Just one amino acid substitution (Trp86 replaced by His), which is more than 30 angstrom away from the reactive site, changed the inhibitor, Streptomyces subtilisin inhibitor (SSI), into a temporary inhibitor without a change in the inhibition constant. When the inhibitor was in excess of subtilisin BPN', the wild-type SSI was stable under protease attack, while the mutant inhibitor was hydrolyzed to peptide fragments in an all-or-none manner. The mechanism of this temporary inhibition induced by the amino acid substitution was studied on the basis of structural, thermodynamic, and kinetic data obtained by a combined use of NMR, hydrogen-deuterium exchange, differential scanning calorimetry, and gel filtration HPLC. The mutation did not induce major structural changes, and in particular, the structure of the enzyme-binding region was virtually unaffected. The denaturation temperature of SSI, however, was decreased by 10 deg upon mutation, although it still remained a thermostable protein with a denaturation temperature of 73-degrees-C. Furthermore, the activation enthalpy for denaturation was reduced dramatically, to half that of the wild type. When the mutated SSI is present in excess of the enzyme, the proteolysis followed first-order reaction kinetics with respect to the total concentration of the mutated SSI molecules present. From these combined results, we conclude that the proteolysis proceeds not through the native form of the inhibitor in the inhibitor-enzyme complex but through the denatured (unfolded) form of the inhibitor whose fraction is increased by the mutation. This conclusion states that the necessary condition for being a serine protease inhibitor lies not only in the design of the reactive site structure that is highly resistant to protease attack but also in the suppression of such structural fluctuation that brings about cooperative denaturation. In contrast, when the protease existed in excess of the mutated inhibitor, the proteolysis reaction was accelerated by more than 2 orders of magnitude. Furthermore, the reaction occurred even in the wild-type SSI at a comparable rate as in the mutated protein. This indicates that in the enzyme excess case another, more efficient digestion mechanism involving fluctuation within the native manifold of the inhibitor dominates.

It has been shown that the P1 site (the center of the reactive site) of protease inhibitors corresponds to the specificity of the cognate protease, and consequently specificity of Streptomyces subtilisin inhibitor (SSI) can be altered by substitution of a single amino acid at the P1 site. In this paper, to investigate whether similar correlation between inhibitory activity of mutated SSI and substrate preference of protease is observed for subtilisin BPN', which has broad substrate specificity, a complete set of mutants of SSI at the reaction site P1 (position 73) was constructed by cassette and site-directed mutagenesis and their inhibitory activities toward subtilisin BPN' were measured. Mutated SSIs which have a polar (Ser, Thr, Gln, Asn), basic (Lys, Arg), or aromatic amino acid (Tyr, Phe, Trp, His), or Ala or Leu, at the P1 site showed almost the same strong inhibitory activity toward subtilisin as the wild type (Met) SSI. However, the inhibitory activity of SSI variants with an acidic (Glu, Asp), or a beta-branched aliphatic amino acid (Val, Ile), or Gly or Pro, at P1 was decreased. The values of the inhibitor constant (K(i)) of mutated SSIs toward subtilisin BPN' were consistent with the substrate preference of subtilisin BPN'. A linear correlation was observed between log(1/K(i)) of mutated SSIs and log(1/K(m)) of synthetic substrates. These results demonstrate that the inhibitory activities of P1 site mutants of SSI are linearly related to the substrate preference of subtilisin BPN', and indicate that the binding mode of the inhibitors with the protease may be similar to that of substrates, as in the case of trypsin and chymotrypsin. On the other hand, the Cys73 mutant showed only temporary inhibition after mixing of the SSI and subtilisin BPN', which may be due to irregular disulfide bridge formation in the SSI molecule.