Reaction of human chymase with reactive site variants of alpha 1-antichymotrypsin. Modulation of inhibitor versus substrate properties.

Inhibition of human chymase by alpha 1-antichymotrypsin produces 3.5 mol of degraded inhibitor for every mol of chymase inhibited, resulting in a stoichiometry of inhibition (SI) of 4.5. In the present study, the substrate versus inhibitor properties of this reaction were examined further using wild type and mutant recombinant antichymotrypsins (rACT). Titration of chymase hydrolytic activity with rACT-L358 (wild type) and reactive site (P1) variants of ACT, L358W, L358M, and L358F revealed that the SI was sensitive to P1 residue replacements. SI values increased in the order of Trp < Met < Leu < Phe where SI values were 1.5, 2, 4, and 7, respectively. Chymase inhibitor complex and cleaved inhibitor were demonstrated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis for all variants; the relative intensities of each band were consistent with SI values established by titration. NH2-terminal sequence analyses of the products formed in the reaction of chymase with rACT-L358F indicated that the P1-P1' bond was the primary site of cleavage resulting in the hydrolysis and inactivation of this variant. The apparent second-order rate constant for chymase inhibition (k'/[I]) by rACT also was affected by P1 substitution. k'/[I] values increased in an order opposite that obtained for SI values (Phe < Leu < Met < Trp). The reactive loop mutant (rACT-P3P3') produced by replacing the reactive site region of ACT (Thr356-Val361) with that of alpha 1-proteinase inhibitor (Ile356-Pro361) revealed a different reaction pattern. Although its SI was near 1, the value for k'/[I] was the lowest among variants. rACT-L358R, another P1 variant, did not inhibit chymase. These results are evaluated with respect to the substrate preferences of human chymase and with respect to partitioning schemes proposed to explain SI values greater than 1.

Inhibition of human chymase by al-antichymotrypsin produces 3.5 mol of degraded inhibitor for every mol of chymase inhibited, resulting in a stoichiometry of inhibition (SI) of 4.5. In the present study, the substrate versus inhibitor properties of this reaction were examined further using wild type and mutant recombinant antichymotrypsins (rACT). Titration of chymase hydrolytic activity with rACT-L358 (wild type) and reactive site (Pl) variants of ACT, L358W, L368M, and L358F revealed that the SI was sensitive to P1 residue replacements. SI values increased in the order of Trp < Met c Leu c Phe where SI values were 1.5, 2, 4, and 7 , respectively. Chymase inhibitor complex and cleaved inhibitor were demonstrated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis for all variants; the relative intensities of each band were consistent with SI values established by titration. NHa-terminal sequence analyses of the products formed in the reaction of chymase with rACT-L358F indicated that the PI-PI' bond was the primary site of cleavage resulting in the hydrolysis and inactivation of this variant. The apparent second-order rate constant for chymase inhibition (k'/[I]) by rACT also was affected by P1 substitution. k'/[I] values increased in an order opposite that obtained for SI values (Phe < Leu < Met < Trp). The reactive loop mutant (rACT-P3P3') produced by replacing the reactive site region of ACT (Thr36e-Va13e1) with that of al-proteinase inhibitor (Ile36e-Prose1 ) revealed a different reaction pattern. Although its SI was near 1, the value for k'/[I] was the lowest among variants. rACT-L358R, another P1 variant, did not inhibit chymase. These results are evaluated with respect to the substrate preferences of human chymase and with respect to partitioning schemes proposed to explain SI values greater than 1.
Human chymase from skin is a chymotrypsin-like proteinase that is stored within the secretory granules of mast cells (1)(2)(3). We have shown previously that this proteinase is inhibited by two human plasma inhibitors, ACT' and al-proteinase inhibitor (4). Both inhibitors are members of the serpin (serine proteinase inhibitor) protein family (5,6). Inhibition of chymase by both inhibitors appeared irreversible, showing the typical tight binding SDS-stable complexes characteristic of serpin-target proteinase interactions (5, 6).
Apparent second-order rate constants obtained under pseudofirst-order conditions were 21,000 "' s" for ACT and 7,500 M" s-' for al-proteinase inhibitor (4). The magnitude of these values is on the low side for serpin-proteinase interactions, approximately 1,000 x lower than values reported for the inhibition of neutrophil cathepsin G by ACT and neutrophil elastase by al-proteinase inhibitor (7). The reaction of chymase with ACT and al-proteinase inhibitor also displayed an unusually high SI, a value empirically defined as mol of inhibitor required to inhibit 1 mol of chymase measured by titration. SI values of 4.5 and 5.0 were obtained for inhibition of chymase by ACT and al-proteinase inhibitor, respectively.
Further analysis of the interaction of chymase with each inhibitor revealed that the high SI for inhibition was caused by a competing reaction producing a hydrolyzed inactive inhibitor. SDS-PAGE and NH2-terminal sequence analyses indicated that the hydrolyzed inhibitor was generated by cleavage within the reactive loop (4); the cleavage site for ACT was provisionally located at the reactive site of the inhibitor ( L e~~~~-S e r~~' ) , whereas the cleavage site for a1proteinase inhibitor was located at a bond (Phe352-Le~353) several residues away from the reactive site (Met358). SI values obtained from titrations of chymase with either ACT or a1proteinase inhibitor did not vary as a function of the initial enzyme concentration, indicating that the fraction of total inhibitor converted to hydrolyzed inhibitor was an invariant property of each reaction (4).

Reaction of Human Chymase with Variant Antichymotrypsins 23627
is the reactive site of the inhibitor which is referred to as the P1-P1' site following the nomenclature of Schechter and Berger (8). We also investigate further the substrate uersus inhibitor properties of the reaction of chymase with ACT using recombinant ACT (rACT) variants. These studies provide additional support for the inclusion of a rapid partitioning scheme in the general equation for serpin-proteinase interactions to explain high SI values, as proposed by Cooperman et al. (9) in the preceding article.

EXPERIMENTAL PROCEDURES
Materials-Peptide-pNA substrates were purchased from Sigma or Bachem. Bovine chymotrypsin and trypsin were from Calbiochem or Sigma. TSK-heparin-5PW and Mono Q HPLC columns were from Supelco and Pharmacia LBK Biotechnology, Inc., respectively. Immobilon-P membranes (PVDF) were from Millipore. Heparin-Sepharose was from Pharmacia. The detergents dodecyl maltoside and Triton X-100 were obtained from Anatrace and Sigma, respectively.
rACT Variants-Expression of rACTs, production of reactive site, reactive loop (rACT-P3P3'), and cassette mutants of ACT, and purification of rACTs, were accomplished as described by Rubin et al. (10) and Kilpatrick et al. (11). rACT-Leu35&as, rACT-L358Fcas, and rACT-PBP3'cas refer to variants produced using a cassette construction (11). Insertion of the cassette into ACT cDNA required the creation of two restriction sites which resulted in three amino acid changes in the inhibitor. In one, Ala3''-Alaw (PlO-Pg) was changed to Gly-Thr, and in the other, Valrn (P10') was changed to Thr. Sites P9-P10 are located near reactive loop position P14, a site critical for serpin inhibitory function (12)(13)(14)(15). Mutations at site P10 of antithrombin I11 (Alam to Ser or Pro) and C1 inhibitor (Ala"' to Thr) have been shown to reduce the inhibitory activities of these serpins (13). The P10 Ala to Gly mutation in our cassette variants of ACT was a more conservative substitution than those affecting the activity of antithrombin 111. In addition, the P10 site of cY1-proteinase inhibitor is Gly, suggesting that this amino acid residue may be acceptable in the P10 position of ACT. These changes alone did not alter the inhibition properties of ACT toward chymotrypsin (11). rACTs also contained a short NH2-terminal extension of 8 (10) or 4 (in more recent constructions) amino acids. Inhibition of chymase by rACT-Leu3" variants containing both types of NHz-terminal constructions was identical with respect to SI and inhibition rate constants.
Purification of Human Chyme-Mast cell proteinases were extracted from skin and fractionated on a 500-ml heparin-Sepharose column (1,16,17). Chymase was eluted from the column in a single step with a solution of 2 M NaCl, 0.1 M MOPS (pH 6.8), and was further purified by affinity chromatography using soybean trypsin inhibitor-Sepharose (1,16,18). This material was used in most studies. For reasons still unclear, the fraction of chymase in skin extracts which binds to heparin-Sepharose is variable (18). In extracts in which 30-70% of the enzyme did not bind, an alternative purification method was used. Chymase not bound to heparin-Sepharose was precipitated by the addition of protamine chloride (1,18). After solubilization of chymase in high salt buffer, the preparation was fractionated on phenylbutylamine-Sepharose. Bound enzyme was eluted with a solution of 0.4 M NaCl, 0.01 M MOPS (pH 6.8), 0.2 M D-Trp-OMe. Chymase was then chromatographed on a 50-ml heparin-Sepharose column. The proteinase now bound to the column and was eluted in high salt buffer as described above. As far as we can detect, chymase purified by both procedures has identical catalytic, physical, and inhibition properties; therefore, distinctions between the two chymase preparations were not made in the text.
TSK-Heparin Chromatography-Additional fractionation of chymase was achieved on a TSK-heparin column eluted with an increasing NaCl concentration gradient. Chromatography was performed using a Perkin-Elmer-Cetus Instruments series 410 HPLC pump and series 235 diode array detector. Chymase isolated as described above was loaded onto the column in a solution of 0.4 M NaC1, 0.01 M MOPS (pH 6.8) and then eluted with a linear NaCl gradient ranging from 0.4 to 1.6 M NaCl. Fractions were assayed for chymotrypsin activity under standard conditions to obtain initial rates of hydrolysis, or they were allowed to incubate in substrate solutions for up to 24 h to detect minor contaminating proteinases. In the latter assays, 20pl samples from column fractions were incubated in 250 pl of assay buffer containing 2 mM Suc-A-A-P-F-pNA. Reactions were stopped by the addition of 0.5 ml of 5% acetic acid and were quantified spectrophotometrically.
Determination of Inhibitor and Proteinase Concentrations-The concentrations of rACTs were determined by titration of a standard amount of chymotrypsin, except for rACT-L358R, which was quantified using trypsin instead of chymotrypsin. Titrations were routinely performed by the addition of varying amounts of inhibitor to 50-100-p1 reactions containing approximately 200 nM proteinase, 0.1 M Tris-HCl (pH 8.0), 0.5 M NaCl, 0.01% dodecyl maltoside (25 "C). Residual activity was measured spectrophotometrically after 10-15-min incubations by dilution of an aliquot of the reaction mixture into 1 ml of assay buffer containing 1 mM Suc-A-A-P-F-pNA or benzyloxycarbonyl-GPA-pNA. Stock chymotrypsin and trypsin solutions were standardized with the active site titrants N-trans-cinnamoylimidazole (19) or p-nitrophenyl p'-guanidinonitrobenzoate (20). Chymase concentrations were determined under standard assay conditions (0.4 M Tris-HC1 (pH 8.0), 1.8 M NaCl, 9% MezSO, 1 mM Suc-A-A-P-F-pNA) assuming a specific activity of 2.7 pmol of product min"/nmol of chymase p-nitroaniline = 8,800 M-' cm"). This value was determined previously using titration with lima bean trypsin inhibitor to obtain the concentration of chymase (4).
Titrations and Time Course Studies-Reactions of chymase with rACTs were performed at 25 "C in 50-150 p1 of a solution containing 0.2 M Tris-HC1 (pH 8.0), 1 or 2 M NaCl, and 0.01% Triton x-100. After incubation with various amounts of inhibitor for appropriate times, residual activities were measured spectrophotometrically after dilution (25-50-fold) of a sample aliquot into 1 ml of standard assay buffer containing 1 mM substrate. Rates of substrate hydrolysis were constant over the 3-min period used to determine residual activities, indicating that chymase-ACT complexes were stable to dilution.
Inhibition After a further 20-min incubation, samples were concentrated, desalted as described above, dried under nitrogen, and finally resuspended in trifluoroacetic acid. A portion of the trifluoroacetic acid-solubilized sample was analyzed by automated Edman degradation, and another portion was subjected to quantitative amino acid analysis to confirm the protein concentration. rACT-L358Fcas reacted with chymase in the latter analyses was subjected to fractionation on a Mono Q column as an additional purification step.

Kinetics Constants for Peptide pNA Substrates-K, and k , values
for the hydrolysis of Suc-A-A-P-F-pNA by chymase in solutions containing 1 and 2 M NaCl were determined according to the Michaelis-Menten rate equation by nonlinear regression analysis of initial velocity versus substrate concentration data. Six to eight different substrate concentrations ranging from 0.25 to 3 mM were used in each experiment. Fits were all statistically significant with standard errors in K,,, and kat values of < 18 and lo%, respectively.

RESULTS
The Large SI for the Reaction of Human Chymase with ACT Is Not the Result of a Contaminating Proteinase or Poorly Folded Inhibitor-To confirm the purity of chymase preparations routinely used in reactions with rACTs, an enzyme preparation was subjected to additional fractionation on a TSK-heparin HPLC column. Only one peak of chymotrypsinlike activity was observed eluting from the column at 0.75 M NaC1, even when fractions were assayed for long time periods to allow for minor contaminating activities to be detected. The SI value for rACT and rACT-L358F inhibition of highly purified chymase was the same as that obtained with routinely purified chymase, indicating that our routine preparations were not contaminated with another proteinase capable of hydrolyzing rACTs.
In a second test of enzyme purity, chymase purified by routine procedures was reacted with rACT at [rACT],/[chymase], > SI, and then reactions were analyzed by SDS-PAGE to show that excess inhibitor was not subject to further hydrolysis by a possible contaminating proteinase. The presence of intact inhibitor in addition to degraded inhibitor at Fig. L4. Even after long incubation times intact inhibitor was observed on gels (Fig.  1B). Residual intact inhibitor was shown to be active in a parallel experiment in which chymotrypsin was added to the reaction after inhibition of chymase. SDS-PAGE analysis of this reaction mixture (not presented) showed the disappearance of the intact rACT band and the appearance of an additional high M, band corresponding to the chymotrypsin-rACT complex. These results also argue against the presence of a proteinase contaminant for which rACT is uniquely a substrate.
High SI values also could result from the presence of improperly folded inhibitor in our ACT preparations. Previous studies by us have shown that serum-ACT, as well as rACT and variant ACTs purified from bacterial extracts, have SI values near 1 in reactions with bovine pancreatic chymotrypsin or trypsin and that all ACTs behave as single species in experiments that determine their rate constants for inhibition of chymotrypsin or trypsin (10,11). This biochemical characterization indicates that serum ACT used previously (4) and the rACTs used in the present study were properly folded. The stoichiometries of rACTs used in the present work were determined by titration of each inhibitor with chymotrypsin or trypsin, and only the concentration of active inhibitor established from these titrations was used to calculate SI values for chymase inhibition.
The studies and arguments just described for ruling out artifactual possibilities strongly indicate that the large SI values observed for the inhibition of chymase by serum ACT and by rACTs are an intrinsic property of the interaction between these proteins.
Reaction of Chymase with Variant rACTs-Titrations for five reactive site (P1 site) variants denoted as rACTs L358 (wild type), L358M, L358F, L358W, and L358R, and a more extensively substituted variant, denoted as rACT-P3P3', are shown in Fig. 2. In the latter variant, the P3-P3' sequence of ACT was replaced by the corresponding sequence of crlproteinase inhibitor. The difference in the titrations demonstrate that the SI value was sensitive to mutation of the P1 site as well as to mutation of the P3-P3' sequence. The variant rACT-L358R did not inhibit chymase even at an [I]o/[E]o ratio of 20 (not shown in Fig. 2). SI values for each variant are reported in Table 1 along with SI values for three additional rACT variants that were made using a cassette construction (denoted as cas) as described under "Experimental Procedures." Examples of time courses for reactions show that the inhibition obtained at different [I]o/[E]o ratios was stable, even at the high ionic strength (1 M NaC1) conditions employed (Fig. 3). Residual activities in these studies were achieved within minutes and remained constant for at least 24 h. Consistent with the stability of residual activities, the SDS gel band representing the chymase-rACT complex remained intact for at least 24 h of incubation (Fig. 1B).
Apparent second-order rate constants (k'/[I]) for the inhibition of chymase by rACT variants are reported in Table I [E], ratios, we have shown previously that the reaction of chymase with serum ACT follows pseudo-first-order kinetics (4). Chymase activity loss in the presence of all rACT variants was first-order, proceeding to a complete loss of catalytic activity. Inhibitor concentration for these measurements did not exceed 300 nM. This concentration is not likely saturating for the reaction based on our previous study with serum ACT, which showed a linear relationship between k' and [I] up to an inhibitor concentration of 1 phi.
The results in Table I Fig. 4 A , lanes 4  and 8; Fig. 4B, lane 2; Fig. 4C, lane 2; and Fig. LA, lanes 2 and   3). The presence of minor bands is presumably due to the further degradation of inhibitor by free chymase. These same reactions also exhibit degradation of the complex band. This degradation did not result in the release of free enzyme, however, as demonstrated by the time course studies in Fig. 3.
Identification of the Site of Cleavage of rACT-L358F by Human Chyme-SDS-PAGE banding patterns showing major degradation products slightly smaller than rACTs sug- gest that inhibitors were cleaved by chymase within the reactive loop region (the reactive center is 40 residues from the COOH terminus). Small peptide products (-4 kDa) formed in the reaction of chymase with rACT-L358Fcas were identified on highly cross-linked SDS gels as described previously (4). Following electroblotting of these gels onto PVDF membranes, the small peptides were subjected to NHz-terminal sequence analysis. As shown in Table 11, experiment 1, the only peptide identified in the low M, band began with Ser369, the P1' position of ACT. Since rACT-L358Fcas has an SI of 8 (Table I), cleavage arising from its hydrolysis as a substrate should predominate over its hydrolytic product (P1'-COOH-terminal peptide) released upon SDS denaturation of proteinase-inhibitor complexes (23, 24). Thus, only finding peptide beginning with Ser369 is a result which indicates that inhibitor inactivation (hydrolysis as substrate) was produced by cleavage at the reactive site.

rACT-L358 vs Chymase
A similar analysis for the reaction of chymase with serum ACT, which has an SI of 4.5, was described previously (4). Although two apparent sites of cleavage were observed (Table  11, experiment 4), the peptide corresponding to cleavage at PI-P1' (Le~~"-Ser~~') was formed apparently in a higher yield than the peptide corresponding to cleavage at P3'-P4' (Leu3'l-Val3'*). Since Leu361 is located on the COOH-terminal side of Leu3", we suggest that P3' -P4' cleavage may have occurred following Pl-Pl' cleavage. This suggestion is supported by the results of Baumann et al. (25) who showed that the peptide Ser369-Le~36' was removed from ACT on prolonged incubation with chymotrypsin.
The identity of the cleavage site producing degraded rACT-L358Fcas was confirmed in an experiment where NH2-terminal sequence analysis was performed on the entire reaction mixture. The results obtained for two reactions performed at different [I]o/[E]o ratios (ratios = SI and 2SI) demonstrated only one NHz-terminal sequence starting at Ser3s9 (Table 11, experiments 2 and 3). Based on yields of phenylthiohydantoin derivatives, the estimated recoveries of peptide for the two reactions were roughly 90% and 50%, respectively. Recoveries were estimated assuming that 1) the coupling efficiency of protein to the sequencing support was 50%, and 2) yields of phenylthiohydantoin-Ala and phenylthiohydantoin-Leu in the second and third Edman degradation cycles were representative of peptide recoveries. These high recoveries indicate that the majority of the Ser3"-peptide identified in this analysis was a product generated by the hydrolysis of rACT-L358Fcas as a substrate. Amino acid sequences corresponding to the NH2 termini of rACT and chymase were not detected. The absence of the latter sequence was because of its low concentration. The absence of the former sequence is unclear. Although rACT is a bacterial product, other rACT variants prepared for analysis differently have demonstrated a free NH2 terminus.
For analyses of whole reaction mixtures, reactions of chymase with rACT-L358Fcas were performed at 4 "C to take advantage of the high SI (SI = 15-17) we observed at this lower temperature. The SI of rACT-L358F, the non-cassette analog of rACT-L358Fcas, also increased as the temperature was lowered (SI increased from 7 to E ) , and a similar effect of temperature has been reported for the reaction of C1 inhibitor will kallikrein (26). SDS-PAGE analyses of low temperature reactions confirmed the high SI and demonstrated that altering the temperature had no effect on the size of the hydrolyzed inhibitor product. The latter observation suggests that the location of the cleavage site producing hydrolyzed inhibitor was not changed by temperature.
Changes in SI among Variants Compared to the Substrate Specificity of Human Chymase-Kinetic constants in Table  I11 for Suc-V-P-X-pNA substrates were reported by Powers et al. (3), and those for angiotensin I analogs were reported by Kinoshita et al. (27). Human chymase cleaves angiotensin I at the Phes-Hisg peptide bond, thus for this substrate position 8 corresponds to the P1 position (28, 29). As shown in Table 111, the substrate preference (kcat and kCat/Km values) of chymase for a series of P1-substitutedpeptide-pNA substrates followed an order (Phe > Leu > Met -Trp), which parallels the substrate preference of chymase for rACTs, as shown by the SI values in Table I. Kinetic constants for P1-substituted angiotensin I analogs did not follow this order, however. For this series of substrates, the catalytic preference of chymase was Trp -Phe > Leu.

DISCUSSION
Inhibition of proteinases by serpins may occur with SI values greater than 1 as reviewed by Cooperman et al. (9) in the preceding article. The high SI is the result of a concurrent reaction producing a hydrolyzed form of the inhibitor which is inactive. Results presented in this study and a previous study (4) demonstrate that the interaction of human chymase with ACT is another example of this type of inhibition. The dual outcomes of the chymase-ACT interaction were evidenced by titrations that demonstrated SI values greater than 1 and SDS-PAGE analyses of reactions that demonstrated the presence of both a hydrolyzed inhibitor and a stable chymase-ACT complex (Fig. 1).
NH2-terminal analysis of the reaction products from inhibitors with high SI values (serum-ACT and rACT-L358F) identified the cleavage site resulting in inhibitor inactivation as the P1-P1' bond (Table 11). The absence of a minor sequence in these analyses indicating a second proteolytic site within the reactive loop suggests that the P1 site also was  responsible for chymase inhibition, as expected from inhibition studies with other proteinases (5,6). This result was different from that obtained for the interaction of chymase with another serpin, a1-proteinase inhibitor, which under certain reaction conditions exhibited an SI of 9 (4). In this case, NH2-terminal analysis demonstrated a minor peptide product corresponding to cleavage at PI-P1' and a major peptide product corresponding to cleavage at P6-P7, thereby indicating that the hydrolytic site was different from the reactive site. A second observation supporting the P1 site as the chymase inhibitory site was that replacement of Leu358 with Arg produced an rACT variant active toward trypsin and chymotrypsin but not chymase. Since human chymase does not readily hydrolyze ester or peptide substrates with Arg at P1, the lack of inhibition by rACT-L358R agrees with the expectation of reduced interaction at the P1 site. In contrast, chymase inhibition was observed with all other P1 variants of rACT having residues recognized by chymase in substrates.
Other serpin-proteinase pairs that demonstrate reactive sitebond cleavage concurrent with proteinase inhibition are thrombin-anthithrombin and kallikrein-C1 inhibitor (26, 31, 32). Analysis of the interaction of chymase with reactive site variants of ACT (Table I)

I1
Zdentificatwn of chyrnase cleavage sites in rACTs by NH2-terrninal sequence analysis of reaction products Experiments 1 and 4 are of low M, peptide products (Mr between 3,000 and 6,000) obtained after resolving reactions on SDS gels and electroblotting proteins to PVDF membranes. Experiments 2 and 3 are of whole reaction mixtures. The native sequence of ACT from P1' to  Two sequences, (a) and (b), were deduced from data. The results, but not the actual data, were reported previously (4).

P15' is S-A-L-V-E-T-R-T-I-V-R-F-N-R-P. Values in parentheses are net vields reDorted in Dmol.
e No residue was identified in this cycle because of a high background. Based on the sequence following this cycle it is assumed that serine (a) and valine (b) are the true NH2-terrninal residues of these peptides.  value was about 10-fold lower than rACT-L358W, which had an SI close to 1, and 4-fold lower than rACT-L358M, which had an SI of 2. The order of k'/[I] values for three P1 variants was Met > Leu > Phe (total range was 2-fold). This order was opposite that observed for chymotrypsin (total range was 5-10-fold) with the same variants (10,11). The variation of rACT inhibitor properties as a function of P1 substitution may be rationalized by either of two schemes.
One involves a branching mechanism described by Cooperman et al. (9) in the preceding article. A compressed version of that scheme sufficient to account for the results presented in this paper is shown below as Scheme 1. Similar branching schemes to this Scheme 1 have been proposed by Bjork et al.
(33) and Patston et al. (26). In this model, I binds reversibly to E to form a complex E-I, which is then partitioned along two pathways: one leading to the formation of EI*, a stable enzyme-inhibitor complex, and the other leading to cleavage of the inhibitor as a substrate producing free E and Is, a reactive site-cleaved inhibitor. EI* formation for serpin-proteinase reactions have been proposed to involve a structural rearrangement within the inhibitor induced by binding to the enzyme (13,33). If, for a fraction of its encounters with the inhibitor, a proteinase is capable of hydrolyzing the P1-P1' bond before the conformational change is complete, the SI for the reaction will be greater than 1 (33). The second scheme (Scheme 2) is based on the studies of Declerck et al. (34) which suggest that a single serpin may have interconvertible inhibitor (Ia) and substrate (Ib) conformations. The value SI would then depend on the relative rates of chymase interaction with each conformer and the position of the conformer equilibrium. The order of SI values (Phe > Leu > Met > Trp) as a function of P1 substitution qualitatively paralleled the substrate preference of chymase (Phe > Leu > Met -W) toward peptide-pNA substrates as measured by $.t or %at/Km values (Table 111). This correlation is consistent with the notion that the substrate pathway in either Scheme 1 or Scheme 2 is more sensitive to P1 substitution than is the inhibitor pathway and suggests a parallelism for the effects of P1 substitution on SI values and on the rate determining step(s) for peptide-pNA hydrolysis. Similar agreement, however, was not observed for chymase catalysis of PI-substituted angiotensin I analogs where hat and kat/Km values were reported to vary in the order of Phe -Trp > Leu (27). The difference observed for the two simple sets of substrates might reflect differences in the nature of the rate-determining steps for chymase catalyzed hydrolysis of each substrate type (35).
Partitioning occurs as two first-order reactions in Scheme 1 and as two second-order reactions in Scheme 2. AS a result, the naive expectation is that in Scheme 1, SI-1 values (rate of the substrate pathway reaction relative to the rate of the inhibitor pathway) may parallel kcat values, whereas in Scheme 2 they may parallel kat/Km values. Quantitatively, relative SI-1 values (Phe:Leu:Met:Trp) for P1 variants agree better with kat than k&Km values of peptide-pNA substrates; however, the results do not permit a clear choice between Schemes 1 and 2.
The result of P1 substitutions on k'/[I] values showed that these values decreased as SI values for corresponding variants increased (Table I). Although this result for P1 variants is a possible consequence of the branch points in Scheme 1 (the partitioning of E .I by reaction steps with rate constants k2 and k3) and Scheme 2 (1a:Ib equilibrium), the low k'/[I] and low SI values obtained for the rACT-P3P3' variant demonstrate that these two parameters do not necessarily vary inversely. rACT-P3P3' is a more extensively substituted variant compared with P1 variants. The greater degree of change to the reactive loop of this rACT is likely an factor influencing its behavior.
The order of k'/[I] values for inhibition of chymase by Met, Leu, and Phe variants was opposite that observed for inhibition of chymotrypsin (Phe > Leu > Met) by the same variants (10, l l ) , despite both proteases having the same general substrate preferences (3,36). In contrast to chymase, the SI value for the interaction of chymotrypsin with each of these variants is close to 1 (10,11), suggesting that suppression of the substrate pathway may account for the observed difference in the order of inhibition rate constants.
In summary, the high SI observed for the interaction chymase with ACT and its wide variation in mutants allow for the analysis of serpin-proteinase interactions in a manner not available for other target proteinases. Our results, however, do not provide definitive evidence permitting a clear choice between the branched or conformer model to explain SI values greater than 1. Nevertheless these studies demonstrate the importance of including a rapid partitioning scheme in the general equation for serpin-proteinase interactions and the need for further defining the mechanism producing high SI values.