Reaction of Human Skin Chymotrypsin-like Proteinase Chymase with Plasma Proteinase Inhibitors*

The ability of plasma proteinase inhibitors to inac- tivate human chymase, a chymotrypsin-like proteinase stored within mast cell secretory granules, was inves- tigated. Incubation with plasma resulted in over 80% inhibition of chymase hydrolytic activity for small sub- strates, suggesting that inhibitors other than a2-mac-roglobulin were primarily responsible for chymase inactivation. Depletion of specific inhibitors from plasma by immunoadsorption using antisera against individual inhibitors established that al-antichymotrypsin (al-AC) and a,-proteinase inhibitor (al-PI) were re- sponsible for the inactivation. Characterization of the reaction between chymase and each inhibitor demon- strated in both cases the presence of two concurrent reactions proceeding at fixed relative rates. One reac- tion, which led to inhibitor inactivation, was about 3.5 and 4.0-fold faster than the other, which led to chymase inactivation. This was demonstrated in linear titrations of proteinase activity which exhibited endpoint stoichiometries of 4.5 (aI-AC) and 5.0 (al-PI) instead of unity, and SDS gels of reaction products which exhibited a banding pattern indicative of both an SDS-stable proteinase-inhibitor complex and two lower M, inhibitor degradation products which appear 0.45 M Tris-HC1 (pH 8.0), 1.8 M NaCl, 9% dimethyl sulfoxide. Bovine trypsin hydrolytic activity was determined using 1 mM benzoyl-Arg-NA in 0.05 M Tris-HCI (pH 8.0), 0.2 M NaCl, 9% dimethyl sulfoxide. Substrate hydrolysis was monitored at 410 nm using a Gilford 260 spectrophotometer. Purification of Proteinases and Inhibitors-Bovine pancreatic chymotrypsin chymase used in its specific of was difficult to detect in controls. Chymase migrates on SDS gels as a diffuse band at M, 30,000 (1,4,30). The diffuse banding pattern results from glycosylation (N. M. Schechter, unpublished observation). Samples corresponding to 1.6.3.8, and 48 pg of a1- AC were analyzed on lanes 1-6, 7-10, and 11 and 12, respectively.

The ability of plasma proteinase inhibitors to inactivate human chymase, a chymotrypsin-like proteinase stored within mast cell secretory granules, was investigated. Incubation with plasma resulted in over 80% inhibition of chymase hydrolytic activity for small substrates, suggesting that inhibitors other than a2-macroglobulin were primarily responsible for chymase inactivation. Depletion of specific inhibitors from plasma by immunoadsorption using antisera against individual inhibitors established that al-antichymotrypsin (al-AC) and a,-proteinase inhibitor (al-PI) were responsible for the inactivation. Characterization of the reaction between chymase and each inhibitor demonstrated in both cases the presence of two concurrent reactions proceeding at fixed relative rates. One reaction, which led to inhibitor inactivation, was about 3.5 and 4.0-fold faster than the other, which led to chymase inactivation. This was demonstrated in linear titrations of proteinase activity which exhibited endpoint stoichiometries of 4.5 (aI-AC) and 5.0 (al-PI) instead of unity, and SDS gels of reaction products which exhibited a banding pattern indicative of both an SDS-stable proteinase-inhibitor complex and two lower M, inhibitor degradation products which appear to have formed by hydrolysis within the reactive loop of each inhibitor.
At inhibitor concentrations approaching those in plasma where inhibitor to chymase concentration ratios were in far excess of 4.5 and 5.0, the rate of chymase inactivation by both serpin inhibitors appeared to follow pseudo-first order kinetics. The "apparent" second order rate constants of inactivation determined from these data were about 3000fold lower than the rate constants reported for human neutrophil cathepsin G and elastase with aI-AC and al-PI, respectively. This suggests that chymase would be inhibited about 650-fold more slowly than these proteinases when released into plasma. These studies demonstrate that although chymase is inactivated by serpin inhibitors of plasma, both inhibitors are better substrates for the proteinase than they are inhibitors. This finding along with the slow rates of inactivation indicates that regulation of human chymase activity may not be a primary function of plasma. Mast cells from human (1,2), dog (3, 4), and rat (5)(6)(7) contain substantial amounts of serine proteinases with chymotrypsin-like specificity (8). These proteinases, termed chymases, are constituents of cytoplasmic secretory granules (9)(10)(11) where they appear to be packaged in an active form (12)(13)(14) bound to heparin ( 5 , 9,10,14,15). They are released concomitant with histamine when mast cells degranulate in vitro (14,16,17). The localization of the chymases in secretory granules and their release coincident with degranulation suggests chymases function extracellularly. The biological role of these enzymes has not been established with certainty.
Many serine proteinases which function in plasma or tissue during inflammation appear to be regulated by protein-proteinase inhibitors found in plasma (18,19). About 10% of the proteins in plasma are proteinase inhibitors (18). Most numerous in variety are those of the serpin superfamily which specifically inhibit serine class proteinases (18,19). Also found are the general proteinase inhibitor a2M1 and the Kunitz type inhibitor inter-a-trypsin inhibitor (18). Several of these inhibitors react with chymotrypsin-like proteinases. The high rate constant for the inactivation of cathepsin G ( k , = 5.1 X lo7 M" s-'), the major chymotrypsin-like proteinase of human neutrophils, by the serpin a,-AC, suggests that this inhibitor is responsible for regulating the activity of cathepsin G in vivo (20). In a similar fashion neutrophil elastase is believed to be regulated by the serpin al-PI (20). Lack of functional a,-PI has been correlated with production of lung tissue deterioration in emphysema (18,20).
Although the mature mast cell is usually found as a resident of tissue such as the dermis of skin (21), evidence obtained primarily from rodents indicates that mast cells are produced from stem cells in the bone marrow (22); therefore, like neutrophils, they may be considered part of the hematolymphoid system involved in host defense. This suggests that the mechanisms regulating neutrophil serine proteinases may also apply to mast cell serine proteinases, despite the location of these cells in tissues. When mast cell degranulate during the allergic response, for example, histamine is released from secretory granules. The action of histamine produces local increased vascular permeability (23) which may lead to an influx of plasma proteins, including inhibitors, into tissues. These inhibitors could then regulate chymase activity. The purpose of the present study was to determine whether human plasma proteinase inhibitors are effective inhibitors of human chymase, which is presumably also secreted from mast cells when they degranulate.

Materials-Proteinase inhibitors al-PI and chymostatin were from
Calbiochem. a,M, PMSF, active site titrants, bovine pancreatic proteinases, and peptide-NA substrates were from Sigma. Antisera against a2M (Sigma M-1893), a,-AC (DAKO A-022), and W-PI (DAKO A-012) were obtained from commercial sources. Standard proteins for gel analysis were from Bio-Rad, except that B chain of insulin ( M , = 3000) and aprotinin ( M , = 6500) were from Sigma.
Isolation of Human Plasma-Human blood was collected by venipuncture into containers having EDTA (7.5 mg/5 ml blood). After removal of cells by centrifugation, plasma was dialyzed against PBS (0.01 M NaP (pH 7.2) 0.15 M NaC1) and stored at -20 "C until used.
Depletion of Inhibitors from Plasma-Protein A-Sepharose was incubated with antisera against specific inhibitors to immobilize IgG. Saturating quantities of IgG were complexed to the matrix in 60-min incubations, and excess protein was removed by washing the gel with PBS. The amount of matrix required to adsorb out a specific inhibitor was based on the assumption that about 10% of the antisera's IgG was against the inhibitor. Adsorption of inhibitors from plasmas was performed in overnight incubations a t 4 "C. Supernatants were separated from gels by centrifugation, and removal of each inhibitor was confirmed by immunodiffusion or radial diffusion analysis. Consecutive removal of inhibitors was achieved by cycling the same plasma through protein A-Sepharose complexed to different antisera. Plasma concentrations were estimated by measurement of total protein using the method of Lowry (24).
Purification of Proteinases and Inhibitors-Bovine pancreatic chymotrypsin was further purified on phenylbutyl amine-CH-Sepharose. Elution was achieved with D-Trp-OEt. The purification insured that any non-chymotrypsin-like proteinases were removed from the preparation. Human chymase was purified from human skin as described previously (1,4). It has been reported that preparations of purified mast cells contain proteins cross-reactive with cathepsin G and elastase (25). In our purified chymase preparations, no reaction with anti-cathepsin G (antisera obtained commercially and as a gift from B. Wintroub, University of San Francisco) was detected by immunoprecipitation using proteinases radiolabeled with diisopropyl fluorophosphate (1) or by Western blotting (4). Purification of ul-AC was achieved using DNA cellulose as described by Tsuda et al. (26) and Abdullah et al. (27).
Determination of Inhibitor and Proteinase Concentrations-The concentrations of chymotrypsin and trypsin were determined by active site titration. Commercial trypsin was quantified using pnitrophenyl ~"guanidinobenzoate as described by Chase and shaw (28). Chymotrypsin was quantified using the active site titrant 4-methylumbelliferyl p-trimethylammonium cinnamate chloride. Standardization of fluorescence changes was accomplished by quantifying fluorescence produced by the reaction of standardized trypsin with 4-methylumbelliferyl p-guanidinobenzoate. The concentrations of human chymase used in the present work were estimated from its specific activity, 2.7 pmol of product min"/nmol chymase measured under the standard assay conditions described above, using +410 = 8800 M" cm" to quantify p-NA produced. To obtain this specific activity, chymase concentrations in solutions of known hydrolytic activity were determined using the Bowman-Burk inhibitor, lima bean trypsin inhibitor, as a titrant (29). Two different concentrations of chymase were incubated with various amounts of lima bean trypsin inhibitor previously standardized using chymotrypsin. Loss of activity was linear to at least 70% inactivation, and chymase concentrations, 465 and 155 nM, determined from the data resulted in nearly identical specific activities. This specific activity was 40% lower than that previously reported by us using radioactive diisopropyl fluorophosphate to quantify chymase in less purified and less concentrated Titrations of Chymase with Purified Serpin Inhibitors-Titrations with chymase were performed similarly as described for chymotrypsin except that data were collected after longer periods of incubation.
Incubations of chymase with al-AC at pH 8 and 6', and ~I -P I a t pH 8.0 proceeded for 1 and 2 h. Fractional activity determined a t both time points were usually within experimental error (5%), and the average of the two data points is reported. Exceptions were for titrations of chymase with a,-PI at pH 6.0 where 2-and 4-h time points were used, and for the titration of 90 nM chymase with CYI-AC at pH 8.0, where only 90-min time points were taken. Chymase incubated without inhibitor lost less than 10% activity during these incubations. In 0.2 M NaCI, the addition of 0.005% Triton X-100 was required to prevent nonspecific loss of chymase activity.
SDS-Gel Electrophoresis-Discontinuous SDS-gel electrophoresis was performed as described previously (1, 4) using a Bio-Rad Mini-ProteinTM I1 dual slab cell and 10 or 15% acrylamide resolving gels (2.5 or 4.5% N,N'-methylene bisacrylamide). To denature, samples were diluted with SDS so that the final concentration of the denaturing solution was 1% SDS, 25 mM Tris-HCI (pH 6.8). Unless specified, they were incubated at 25 "C for 30 min prior to loading on gels. Gels intended to visualize small M , products were treated with a solution of 10% trichloroacetic acid, 5% sulfosalicylic acid prior to staining. Large samples used for sequencing were desaltedprior to denaturation by dialysis (al-PI) or by washing of lyophilized material in 80% ethanol (aI-AC). Apparent M , were calculated from plots of log M , of standard proteins uersus mobility using a second order polynomial to fit the data. Determination of Rate Constants and tH Inactiuation-Three different procedures were used depending on proteinase and inhibitor. 1) The reaction of chymase with al-PI was measured under pseudo first order conditions. Chymase concentration was 3.7 nM and initial inhibitor concentrations were varied so that I / E ratios ranged between 45 and 268. Each rate determination consisted of five data points, and each data point was obtained from an individual reaction set up in cuvettes containing 0.9 ml of 0.45 M Tris-HCI (pH 8.0), 2 M NaCl. At appropriate times, 0.1 ml of substrate was added and hydrolytic activity was measured immediately. Added substrate sufficiently stopped the inhibition reaction so that only a 3% underestimate of true residual activity would be observed in the worst case (32). Semi-logarithmic plots of fractional activity (In E/Eo) uersus time appeared linear for greater than 75-90% inactivation. The k,h8 at each I / E was determined from the slopes of regression lines according to the relationship: where E is residual activity of proteinase, Eo is initial activity, and t is time of incubation. 2) The reaction of chymase with al-AC was measured under pseudo first order conditions using the method of Petersen and Clemmensen (32). In this method, the progress of inactivation is monitored in the presence of substrate. Chymase concentration was 4.5 nM and I/E ratios ranged from 93 to 187.
Buffer composition was the same as described for al-PI. Reactions were started by the addition of proteinase to 1-ml cuvettes containing buffer, inhibitor, and 0.5 mM succinyl-Ala-Ala-Pro-Phe-NA, and absorbance changes at 410 nm were measured every 3 s for 5 min. Instantaneous velocities were estimated from the &,./12-s interval and velocities were determined a t 30-s intervals. In the absence of inhibitor, the rate of substrate hydrolysis by chymase was constant over the 5-min period. Semi-logarithmic plots of the instantaneous velocity uersus time appeared linear to over 75% inactivation (correlation coefficients 0.97-0.99), and k ' , the rate constant for proteinase inactivation in the presence of substrate, was determined from the slopes of regression lines through the data. kobr a t every I/E ratio was then determined according to the relationship (32) where S is the initial substrate concentration. K, for the peptide-NA substrate was 0.42 mM as determined by method of Lineweaver and Burk (33). 3) k. values for the reaction of chymotrypsin with a,-AC pH 6 conditions were obtained by dilution of 0.5 M NaP (pH 5.0) 10-fold in the assay. Because the dissociation constant of NaP is sensitive to buffer and salt concentration, the p H may have ranged between 5.5 and 6.
and a]-PI were determined under second order conditions as described by Beatty et al. (20).
Half-times ( tlh) of inactivation in plasma were calculated according to the relationship t H = o.693/z0 X kobsr where Io is the inhibitor concentration in plasma (34) and k is the apparent second order rate constant for chymase with each inhibitor.
NHz-terminal Sequence Analysis-Sequence analysis was performed at the protein chemistry facility at the Wistar Institute (Philadelphia). Products from reactions containing 1.4 nmol of a]-AC (14 @, Z/E = 5, pH 8.0 conditions) or 420 pmol of a,-PI inhibitor (7.5 FM, Z/E = 9, pH 6.0 conditions) were resolved on 15% SDS gels and then electroblotted onto PVDF membranes. Bands were visualized on the membrane by staining with Coomassie Blue. The band corresponding to a M, of 4000 was excised and sequenced directly from the membrane (35). Recoveries compared to estimates of products layered on gel were roughly 10%. NHp-terminal sequences were established from 20 cycles.

Inactivation of Human Chymase by Whole Plasma and
Identification of Inhibitors-The inactivation of human chymase by whole plasma is shown in Fig. 1. About 80% of the hydrolytic activity of chymase was inhibited by plasma when a peptide-NA substrate was used as a monitor of activity. The activity remaining at high plasma concentrations was not further reduced by the addition of lima bean trypsin inhibitor (0.2 mg/ml), a good inhibitor of chymase (see "Experimental Procedures"), suggesting this fraction of the proteinase may be complexed to a2M. Comparison of chymase inactivation as a function of plasma dilution with a comparable concentration of chymotrypsin indicates plasma was more efficient at inactivating chymotrypsin than chymase.
To determine the specific plasma proteinase inhibitors responsible for the inactivation of chymase, IgGs from antisera against specific inhibitors were bound to protein A-Sepharose and the resulting complexes were then used to selectively adsorb out inhibitors from plasma as described under "Experimental Procedures." Plasma was consecutively depleted of a2M, al-PI, and al-AC, and the dilution of each plasma needed to inactivate 50% chymase or chymotrypsin activity is shown in Table I. Removal of al-PI and al-AC reduced to zero the ability of plasma to inhibit chymase, demonstrating that they were the primary inhibitors of chymase in plasma. Inhibitory activity toward chymotrypsin still remained, however. Chymotrypsin has been shown to react with several other plasma inhibitors including a2-antiplasmin, heparin cofactor I1 and inter-a-trypsin inhibitor (18). To determine residual activity, samples were diluted to 1 ml in assay buffer containing substrate, and then immediately assayed for hydrolytic activity as described under "Experimental Procedures." Fractional activity is the quotient of the residual activity/activity of proteinase incubated in buffer. Plasma dilution is reported as a function of protein concentration. Protein concentration of undiluted plasma is 70 mg/ml. plasma protein or in incubation periods greater than 10 min. Our studies show that whole plasma is a more efficient inhibitor of chymotrypsin than of chymase. Part of the reason for this is that the range of inhibitors against chymase, primarily al-PI, and al-AC, was narrower than for chymotrypsin. Another reason, which will be subsequently described, is that both al-PI and al-AC are not highly efficient inactivators of chymase.
Reaction with al-AC-The reaction of serpins with serine proteinases usually results in the irreversible inhibition of proteinase through the formation of a stable 1:l complex which has no hydrolytic activity (18,19). The linear titration of chymase shown in Fig. 2 is consistent with this mechanism of inhibition. After 30 min, the reaction was complete and the extent of inhibition remained unchanged for up to 72 h. When the titration data were plotted as a function of the ratio of the inhibitor to proteinase concentration ( I / E ) and extrapolated t o complete inhibition, a stoichiometry of 4.5 instead of Reaction of Human Chymase with Plasma Proteinase Inhibitors 21311 unity was determined (Fig. 3, filled symbols). The stoichiometry of reaction remained constant over a 14-fold variation in proteinase concentrations (92 nM to 1.3 pM), Strongly indicating an irreversible mechanism of inactivation (Fig. 3).
The high stoichiometry of reactions was not due to the presence of contaminating protein or denatured inhibitor because initial al-AC concentrations were based on titrations with chymotrypsin. To determine whether the excess inhibitor required to fully inactivate chymase was still active after the reaction, al-AC was reacted with chymase at an I / E of 4.5 and 10.0, and then samples from these reactions were used to titrate chymotrypsin activity. The chymase treated w-AC samples lost virtually all (at Z/E of 4.5) and about half (at Z/E of 10) of their reactivity toward chymotrypsin. The lack of activity toward chymotrypsin suggests chymase was degrading the inhibitor during the process of its inhibition. The reactivity of chymase with a,-AC was also measured at pH 6.0 where chymase is about 75% less active toward synthetic substrates than at pH 8.0. Titrations performed at this p H showed an endpoint stoichiometry near 1:l (Fig. 3, open symbols) and reaction kinetics performed at I / E ratio of 4.5 shifted toward a pseudo first order loss of activity (Fig. 4) as expected for a reaction in the presence of excess inhibitor. The observations at low pH indicate that all the active inhibitor in the purified preparation of al-AC was capable of inhibiting chymase and support the hypothesis that the high stoichiometry observed at pH 8.0 involves inhibitor degradation. The reaction of chymase with q-AC was also effected by ionic strength (Fig. 3, inset). The amount of inhibitor needed to fully inactivate chymase increased slightly as the ionic strength was lowered.
SDS-gel analysis of the products produced in the reaction of chymase with a,-AC are shown in  . Inset is a semi-ln plot of data obtained at pH 6.0. Reactions were performed in 0.5 ml and at various times 0.1-ml samples were removed and diluted 10fold with normal assay buffer containing substrate. Fractional activity ( F A ) was determined as described in Fig. 1. products: a minor product migrating with a higher M, (80,000-84,000) than the native inhibitor (65,000) and a major product migrating with a lower apparent M , (52,000). This pattern differs from those observed with chymotrypsin at Z/E ratios of 1 and I (Fig. 5, lanes 2 and 3 ) which show primarily one major product migrating at M, 84,000. Production of both the high and low M , reaction products was prevented by PMSF and chymostatin (Fig. 5, lanes 5-7), suggesting that chymase mediated both reactions and that active enzyme was required.
The higher M, band is consistent with a 1:l SDS-stable complex between the inhibitor and proteinase (M, of chymase is 30,000 on SDS gels). These complexes are characteristic of the interaction of serpins with proteinases and indicate irreversible inactivation (18,19). The lower M , band, on the other hand, is indicative of a hydrolyzed inhibitor. Its staining intensity on SDS gels relative to the 84-kDa band is consistent with the results of titrations, and it does not appear on SDS gels of reaction products produced at pH 6.0 (Fig. 5, lane 1 0 ) where titration endpoints are near 1:1. The contrasting banding patterns observed at pH 8.0 and 6.0 strongly indicate that the hydrolytic product represented by the band at M, 52,000 was not the result of complex perturbation during SDS-gel electrophoresis. This was further confirmed by altering denaturing conditions. Denaturation of reaction products in SDS without heating or by heating to 90 "C for 2 or 10 min did not change the banding pattern, and neither did the addition of PMSF before SDS denaturation (compare Fig. 5,  lanes 4 and 8 ) .
Further evidence for the hydrolysis of al-AC by chymase was obtained by identification of a second cleavage product of much lower apparent M, (about 4000) on a more highly cross-linked gel (Fig. 5 , lune 1 2 ) . Production of this fragment by hydrolysis from either terminal region of a,-AC would help to account for the production of the 52-kDa product. The origin of the low M, peptide as a COOH-terminal fragment was indicated by elucidation of its NHz-terminal sequence (20 cycles). Sequence analysis was performed on peptide electroblotted from the gel onto PVDF membranes as described under "Experimental Procedures." Two NHz-terminal sequences were found corresponding to hydrolysis between residues L e~~" -S e r~'~, the P,-P; site of the inhibitor, and Leu361-VaP6* (P4-P;).3 Based on recoveries of phenylthiohydantoin derivatives, both peptides were present in a 2:l ratio. The  10% (lanes 1-10) and 15% gels (lanes 11 and 12) were prepared as described under "Experimental Procedures." Reactions analyzed in lanes [1][2][3][4][5][6] proceeded for 30 min at 25 "C, while those in lanes [7][8][9][10][11][12] proceeded for 60 min at 25 "C. Controls where proteinases were inhibited with PMSF (2 mM for 15 min) prior to reaction are shown in lanes 5, 6, and 12, and a control where a reaction was performed in the presence of 100 pg/ml chymostatin is shown in lane 7. All reactions containing chymase were performed at pH 8.0 in 0.1 M Tris-HC1,0.8-1.0 M NaCI, except for reactions analyzed in lanes 9 and 10 which were performed at pH 6.0 in 0.05 M Nap, 0.8 M NaCI. Products analyzed in lanes [1][2][3][4][5][6] were denatured in SDS without heat or reducing agent as described under "Experimental Procedures," whereas in lanes 7-12, PMSF (2 mM for 15 min) was added at the conclusion of the incubation to ensure that any possible remaining proteinase activity was inhibited prior to denaturation in SDS. Addition of reducing agent or heating for 2 or 10 min (lanes [7][8][9][10][11][12] had no effect on the banding pattern of products produced by chymase. Except for lane 12  identity of these amino termini along with the limited number of cleavage products on SDS gels indicate that the reaction resulting in inhibitor degradation was a limited process occurring at and near the Pl-P; site of the inhibitor. Finding hydrolysis of the Pl-P; peptide bond was surprising considering the mechanism of proteinase inhibition by serpins (18,19). Bjork et al. (36,37) have also reported hydrolysis at the Pl-P; peptide bond in proteolytically modified antithrombin generated by thrombin in a process occurring concomitant with thrombin inactivation. Further work is necessary to understand the mechanism producing these hydrolytic products.
The results from these studies suggest that the reaction of human chymase with al-AC involves two pathways: one leading to degradation of the inhibitor, and the other one leading to inactivation of the proteinase by complex formation. The similarity in the titration data over a range of different proteinase concentrations indicates that the rates of reaction over these pathways are linked in a fixed ratio which is independent of inhibitor and proteinase concentrations. The rate of inhibitor degradation was 3.5-fold faster than proteinase inactivation. Thus at pH 8.0, al-AC is a better substrate for chymase than it is an inhibitor.
Reaction of Chymase with a,-PI"Similar studies were performed to characterize the reaction of chymase with al-PI. Titration of chymase with al-PI exhibited a linear relationship having an endpoint of 5.0 instead of unity (Fig. 6, solid  symbols). This suggests that al-PI, like al-AC, is both a substrate and an inhibitor of chymase and that these reactions may occur by a two-pathway mechanism as observed for al-AC. SDS-gel analysis of the reaction products on 10% gels (Fig. 7, lanes 5 and 8) showed the formation of a higher (62,000) and a lower M , product (47,000) than native inhibitor (52,000). Again, these products are indicative of proteinase inactivation and inhibitor degradation. The lower M , product was not observed in similar studies with chymotrypsin ( Fig.   (1 ,-PI 2 and 3). The higher M, product (62,000) obtained after the reaction with chymase and chymotrypsin had a significantly lower M , than expected for a complex between al-PI and either proteinase. The reason for this is unclear. However, formation of all reaction products was inhibited by PMSF and chymostatin (Fig. 7, lanes 4, 6, and 7) indicating that active chymase was required to form both products.

7, lanes
The reaction of al-PI with chymase was also dependent on pH and ionic strength. However, with this inhibitor, lowering the pH had the opposite effect of that observed with al-AC; i.e. more inhibitor degradation relative to proteinase inhibition was observed at pH 6.0 (Fig. 6, open circles)  Reaction conditions are as described in Fig. 5; products analyzed in lanes 1-8 were obtained under pH 8.0 conditions and products in lanes 9 and 10 were obtained under pH 6.0 conditions. All incubations proceeded for 60 min. Controls where chymase was inhibited with PMSF prior to reaction are shown in lanes 1, 4, and 9, and controls where 100 and 10 pg/ml chymostatin were added to the reaction are shown in hnes 6 and 7, respectively. Products analyzed in lanes 1-5 were denatured as described under "Experimental Procedures," whereas to the products analyzed in lanes 6-10, PMSF was added to the conclusion of reactions and samples were heated for 2 min at 90 "C. Protein corresponding to 1.5, 8.0, and 9.5 pg of aI-PI were layered on gels in lanes 1-5, 6-8, and 9 and 10, respectively.
observed with al-AC (Fig. 6, inset). SDS gels of reaction products produced at pH 6.0 are shown in Fig. 7 (lane 10). The gel was highly cross-linked and shows the presence of a small peptide with a M , of about 4000. After electroblotting of the protein onto PVDF membranes, NH2terminal sequence analysis was performed on the small peptide. Two NH2-terminal sequences were obtained indicating that these peptides were produced by hydrolysis at Phe"'-Leu353 (P7-P6) and Met3"-Ser3", the PI-P; site of the inhibitor.
In contrast to that observed with aI-AC, recoveries of phenylthiohydantoin derivatives indicate the former hydrolytic site and not the PI-P; site predominated. It was found in a 151 ratio over the second hydrolytic site. These data also indicate that al-PI is a better substrate for chymase than it is an inhibitor. Estimated Rates of Chymase Inactivation in Plusma-The rate of chymase inactivation by al-AC and a1-PI at I/E values substantially greater than those required for complete inactivation (i.e. >>4.5 and 5.0, respectively) appeared to be pseudo first order over a range of inhibitor concentrations (Fig. 8). The highest concentration of inhibitor used, about 1 PM, is approaching the values of these inhibitors in plasma   "apparentn rate constants of inactivation obtained from the slopes in Fig. 8 are reported in Table 11. Using these values and the concentration of each inhibitor in plasma, the halftime of inactivation for chymase by each inhibitor was estimated (Table 11). Because the concentration of al-PI is about 10-fold higher than aI-AC, it would appear to be %fold better physiological inhibitor. Based on the combined concentrations of both inhibitors, the tt,., for chymase in plasma would be 1.3 s. This is about 650-fold slower than estimated for neutrophil cathepsin G, which reacts most rapidly with a1-AC. The time of inactivation of chymase may even be somewhat slower than this because these studies were performed in 2.0 M NaCl and for both inhibitors the rate of inhibitor hydrolysis to proteinase inactivation increased as the ionic strength was lowered.
The second -order rate constant was also determined for the reaction of chymase with a,-AC at pH 6.0, where inhibitor degradation was barely detectable (Table 11). This value was lower than observed for chymotrypsin at this pH, suggesting that even with the degradation pathway nearly eliminated, al-AC is still not an efficient inhibitor of chymase.

DISCUSSION
Even though mature mast cells are found almost exclusively in tissue such as the dermis of skin, current evidence indicates that these cells originate from precursor cells within the bone marrow. As a cell of hemopoietic origin, it might be consistent that serine proteinases secreted from mature mast cells during degranulation are regulated by plasma proteinase inhibitors in the same fashion as human neutrophiI cathepsin G and elastase. The release of histamine and other vasoactive components from mast cells during degranulation may produce a rapid influx of plasma components into tissue so that released proteinases would be quickly inhibited before they can diffuse from the site of secretion.
Our study demonstrates that plasma components irreversibly inhibit human chymase, a chymotrypsin-like proteinase found within mast cell secretory granules. The serpins al-AC and al-PI were identified as the primary inhibitors and both were shown to react with chymase forming products characteristic of the SDS-stable proteinase-inhibitor endproducts associated with serpin inhibition. The titration data with whole plasma also suggest that a2M may inhibit chymase as well. However, only 20% of the inactivation could be attributed to aZM, indicating it is not the primary inhibitor of chymase in whole plasma. Removal of the above three inhibitors from plasma virtually eliminated activity against chymase but did not totally eliminate activity against chymotrypsin. This indicates a more limited variety of inhibitors in plasma for chymase than might be expected for a chymotrypsin-like proteinase.
Studies on the interaction of chymase with the individual serpin inhibitors demonstrated that the reaction of this proteinase with al-AC and al-PI was not as highly efficient as reported for neutrophil serine proteinases. Part of the reason for the inefficiency was that both inhibitors were good substrates for chymase. Under conditions approaching physiological ionic strength and pH, the rate of al-AC and al-PI degradation by chymase was 3.5 and 4.0-fold faster than its rate of inhibition, thus requiring a 4.5 and 5.0 stoichiometric ratio of each serpin to inactivate chymase. Actual rates of chymase inhibition for concentrations of serpins present in plasma were estimated from apparent second-order rate constants determined at excessive I / E ratios where the concentration of inhibitor was approaching that observed in normal plasma. The tC,> inactivation of chymase in plasma was 5.4 and 1.8 s with al-AC and aI-PI, respectively. These values indicate chymase would be inhibited 650-fold more slowly than cathepsin G if released into plasma (20). These values are also slower than those calculated for chymotrypsin, a pancreatic proteinase that is not found in plasma.
At pH 6.0, the degradation of al-AC by chymase was reduced so that the ha of the inhibitor with the proteinase could be measured in the nearly complete absence of the degradation pathway. A k, of 1.6 X lo4 M" s-' was determined.
This value was about 6-fold lower than observed for chymotrypsin at the same pH, suggesting that the interaction of chymase with a,-AC is slow even in the absence of the inhibitor degradation reaction. In the case of al-PI, lowering the pH increased the rate of inhibitor hydrolysis relative to proteinase inactivation. A stoichiometry of 9 was required to fully inactivate chymase. These observations at low pH along with those a t higher pH indicate that plasma would not efficiently inhibit chymase within the pH range of 6-8.
The slow apparent rate constants of inactivation suggest chymase may have sufficient time to inactivate both inhibitors through multiple nonspecific cleavages. However, the Iinearity and constancy of the titration curves over a variety of chymase concentrations, SDS-gel analyses of reaction products showing a limited number of hydrolytic products, and sequence analysis of cleavage products indicating that both inhibitors are inactivated by hydrolysis within the reactive loop, suggest that the inhibitor degradation reaction is specific and possibly linked in some fixed manner to the proteinase inactivation reaction. Reactions of serpins with proteinases that exhibit linear titrations with endpoints greater than unity, and that show the concurrent production of inactive inhibitor-proteinase complexes as well as inactivated inhibitor have been reported for the reaction of antithrombin I11 with thrombin, factor Xa, and IXa (37,381, and for the reaction of al-AC with human pancreatic elastase I1 (38), porcine pancreatic elastase (391, cathepsin G (40), and bovine chymotrypsin (41). Except for the reaction of al-AC with porcine elastase which exhibited a titration endpoint of 5.5 (39), most endpoints were at an I / E ratio less than 2. The endpoints of 4.5 and 5.0 (somewhat higher in low salt) observed for the titration of chymase with a,-AC and al-PI, respectively, suggest that this enzyme is one of least suited proteinases for inhibition by these serpins.
Even though the serpin inhibitors al-AC and al-PI irreversibly inactivate chymase, the slow apparent rates of chymase inactivation compared to other proteinases, and the observation that these inhibitors are better substrates for chymase than they are inhibitors strongly suggests that the regulation of chymase may not be a primary function of plasma. Tryptase, another human mast cell serine proteinase, is not inhibited by plasma (42,43). Its activity appears to depend on a self-denaturation process which is regulated by interaction of the enzyme with heparin (44). Thus released serine proteinases from human mast cells may not be regulated in the same fashion as proteinases released from neutrophils. Possibly, these proteinases have extracellular roles that require their slow inactivation or that there are tissue inhibitors that rapidly inactivate these proteinases. Rat mast cells contain an inhibitor, termed trypstatin, which presumably regulates rat tryptase and possibly rat chymase activity (45), and bovine mast cells are the major source of bovine pancreatic trypsin inhibitor (46). Inhibitors comparable to these have not been so far identified in human mast cells.
The findings described here may not be extended to mast cell chymases of other species. Travis et al. (47) have shown that a chymotrypsin-like proteinase isolated from dog mastocytoma tissue was inactivated by human al-AC. SDS-gel analysis of the products produced in a reaction with excess inhibitor did not contain a noticeable amount of a lower M , product indicative of degraded inhibitor. In preliminary studies, we have titrated purified dog chymase isolated from normal dog skin with human aI-AC (4) and have found an endpoint of 1.6, which is much lower than that obtained for human chymase. This indicates that al-AC may be more effective against dog chymase than human chymase. Caution should be taken when extrapolating mechanisms of regulation to different species, however, because proteinases as well as inhibitors may have evolved different functions. An example of this has been described for a plasma inhibitor. Comparison of cDNA structure between mouse contrapsin, a trypsin inhibitor, and human al-AC, a chymotrypsin inhibitor, indicate both are evolutionary counterparts that have developed different specificities by divergence within the reactive loop region (48).