INTRODUCTION

Serpins are a family of proteins exhibiting diverse biological functions, most notable of which is the inhibition of serine proteinases. Inhibition requires interaction of the protease with a specific residue termed the reactive site or P1 residue as in substrate nomenclature (1). P1-S1 inhibitor-protease subsite interactions give serpins inhibition specificities reflecting the substrate preferences of serine proteases (2-4). The reactive site of the serpin is situated in a large solvent-exposed loop (P14-P5 or P9) of about 20-24 residues (5, 6) with hinge-like flexibility at both ends (7). This mobility implies that protease interaction with the reactive loop alone is not sufficient for inhibition and that processes subsequent to encounter are required to produce inactivation (7). Several serpin-protease reactions have been shown to demonstrate saturation kinetics, indicating at least a two-step mechanism of inhibition (8-14).

The processes responsible for protease inhibition may be chemical and conformational in nature (4, 15, 16). A feature characteristic of serpin inhibition is the formation of an inhibitory complex resistant to dissociation by SDS under heat and reducing conditions (2-4). Such stability implies a covalent bond between enzyme and inhibitor. This complex is proposed to arise from the cleavage of the reactive site peptide bond (P1-P1) and the trapping of the enzyme at the acyl-enzyme step of covalent catalysis by a conformational change(s) related to "strand insertion" (4, 16). Strand insertion is the embedding of a segment of the reactive loop into the body of the protein as a central strand of a six-stranded  sheet structure (2, 7). It was first observed as a spontaneous structural change accompanying proteolysis of the reactive loop that produced a thermodynamically more stable protein no longer active as an inhibitor (2, 7). Although studies with some proteases suggest that conformational change alone may be sufficient for inhibition and that SDS-stable complex formation is unimportant (13, 17, 18), several studies monitoring environmental changes of reactive loop residues during inhibition (19-22) and complex formation with anhydroserine proteases (16, 23) conclude that cleavage of the P1-P1 bond and acyl-enzyme formation are required for stable inhibition.

The requirement of a chemical step(s) associated with peptide bond hydrolysis in the mechanism of serpin inhibition suggests that the rate of inhibition (k_inh) may be critically dependent on the catalytic activity of the protease. The hydrolytic activity of serine proteinases is highly sensitive to pH. In general, a serine proteinase can demonstrate a bell-shaped pH profile with a maximum near neutral pH (24). Reduced activity at low pH is related to the protonation of a His residue (His⁵⁷ in chymotrypsin) critical to catalysis as part of the protease's catalytic triad; protonation affects both acylation and deacylation steps of substrate turnover (24-26). Reduced activity at high pH is related to disruption of a salt bridge between the -amino group of a terminal Ile residue (16 in chymotrypsin) and the carboxylate group of an Asp residue (194 in chymotrypsin) in the active site (24, 27, 28). This ionic bond is responsible for maintaining the active site in a functional conformation (29).

In the present study, the effect of pH on the inhibition of human chymase, a chymotrypsin-like protease, by the serpins ACT¹ and PI was investigated. These reactions are of special interest, because they include a concurrent hydrolytic reaction in which the reactive loop of the serpin is rapidly hydrolyzed in a substrate-like manner (30, 31). The presence of the hydrolytic and inhibitory pathways results in the formation of two products, an SDS-stable complex indicative of covalent bond formation and a hydrolyzed/inactivated serpin indicative of covalent bond hydrolysis. Due to consumption of a fixed proportion of inhibitor by the hydrolytic reaction, complete and virtually irreversible inhibition of chymase requires [I]₀/[E]₀ ratios > 1 (30, 31). The ratio or stoichiometry of inhibition (SI) for the reaction of chymase with ACT and PI at pH 8.0 is 4 and 6.0, respectively. Because the SI is a measure of the relative flux of inhibitor through hydrolytic and inhibitory pathways, analysis of the effect of pH on the SI may help to define the importance of chemical steps to the rates of both pathways. In an early study, low pH had contrasting effects on the SI for chymase-ACT and PI reactions (30). The SI for the chymase-ACT reaction decreased to near 1, while in the chymase-PI reaction it increased to 10.

EXPERIMENTAL PROCEDURES

PI from plasma and buffers were purchased from Calbiochem. SucAAPF-NA was purchased from Bachem. Bovine pancreatic chymotrypsin was obtained from Worthington.

Human chymase was purified as described previously (31, 32). Its concentration was estimated using the specific activity of 2.7 µmol of product (min nmol)¹ for the hydrolysis of SucAAPF-NA under standard conditions where the _410 nm of the product NA = 8800 cm¹ M¹. Standard conditions refer to a 1-ml reaction containing 1 mM substrate, 1.8 M NaCl, 0.45 M Tris-HCl, pH 8.0, 9.0% Me₂SO (30, 31). Chymotrypsin solutions were standardized using active site titrants as described previously (30, 31). rACT and variant rACTs were constructed, expressed, and purified as described previously (31, 33, 34). Purified enzymes and inhibitors were stored in solutions containing 0.01 M buffer at pH 6.8 or 8.0. pH adjustments in experiments were typically achieved by dilution into solutions of 0.1 M buffer.

Chymase titrations were in a solution of 1.0 M NaCl, 0.1 M buffer, 0.01% dodecyl maltoside, and chymotrypsin titrations were in the same solution, except 0.5 M NaCl. Buffers used to vary the pH were citrate, pH 4.5-5.0; MES, pH 5.5-6.5; MOPS, pH 6.5-7.5; Tris, pH 8.0-9.5; CHES, pH 9.0-10.5; and CAPS, pH 9.0-11.5. Reactions (25-50 µl total volume) containing a constant amount of protease (typically 0.2 or 0.4 µM) and increasing amounts of inhibitor were incubated at 25 °C until completion. Residual enzymatic activity was measured by assay under standard pH 8.0 conditions of aliquots removed from reactions. Assays involved a 50-200-fold dilution of the aliquot into the standard substrate solution followed by measurement of the rate of product formation over a 3-min period. Product accumulation was linear over this period (and longer periods) regardless of the original reaction pH. This indicates that the change in pH produced by dilution with standard assay medium did not alter the fraction of enzyme inhibited. The fraction of enzyme inhibited was obtained by comparing residual activities to that of a control incubation containing no inhibitor. Standard assay of chymase aliquots from control incubations always gave the expected activity for standard conditions, regardless of the pH and period of incubation (up to 24 h). Thus, incubation of chymase under the various pH conditions examined does not appear to alter the enzyme irreversibly. Serpins also were stable to these pH conditions. However, below pH 5.0 serpins lost all inhibitory activity rapidly.

Inhibition reactions evaluated by SDS-PAGE were typically performed in 25-40 µl of solution containing salt and buffer composition similar to that of titrations. Reactions were stopped by either 1) addition of phenylmethylsulfonyl fluoride to a final concentration of 3 mM, 2) immediate denaturation by SDS, or 3) trichloroacetic acid precipitation (10% final) aided by co-precipitation with RNA. Trichloroacetic acid precipitates were washed twice in 100% cold ethanol prior to further denaturation by SDS. Denaturation in SDS was accomplished by heating (90 °C) in a solution of 1% SDS and 15 mM dithiothreitol for 10 min. Prior to loading on 10% acrylamide gels (31), bromphenol blue in a glycerol solution was added to the denatured samples and the pH, if necessary, was adjusted to neutrality with NaOH or HCl. Protein bands were visualized by staining with Coomassie Brilliant Blue.

All reactions were performed at 25 °C. Second order inhibition rate constants (k_inh) were determined by several methods. The first method employed analysis of progress curves obtained using SucAAPF-NA as substrate and pseudo-first order [I]₀/[E]₀ ratios (35). Buffer, inhibitor, and substrate (S) were mixed in a cuvette (final conditions: 1 M NaCl, 0.1 M buffer, 9.0% Me₂SO), and reactions were initiated by addition of chymase. Product accumulation was continuously monitored in a spectrophotometer for 15-30 min at A_410 nm (A). k_obs, the observed first order inhibition rate constant in the presence of substrate, and other variables (A₀, v₀, v_s) were obtained by fitting the progress curves to Equation 1 using nonlinear least squares regression analysis. A₀ and v₀ are the zero time (t) values for substrate absorbance (A_410 nm) and velocity of substrate hydrolysis, respectively, and v_s is the steady state velocity at completion of the reaction. v_s always was close to zero in the reactions evaluated, and A₀ and v₀ were always near that expected from controls without inhibitor. k_inh, the apparent second order inhibition rate constant, was estimated using the relationship in Equation 2. Pseudo-first order conditions for each inhibitor were assumed to be when [I]₀  10[E]₀ × SI (30, 31). These conditions minimize the effect of [I]₀ loss resulting from the hydrolytic reaction.

(Eq. 1)

(Eq. 2)

Rate measurements also were made in the absence of substrate under pseudo-first order or second order ([E]₀ = [I]₀) conditions; incubations were in a solution identical to that used for titrations. Second order conditions were employed only for reactions with an SI close to 1. Loss of activity was monitored by periodic removal of samples from reactions followed by measurement of hydrolytic activity in standard or pH 10.0 (CHES instead of Tris) assay medium containing 3-4 mM substrate. pH 10 assay conditions take advantage of chymase's greater catalytic activity and lower K_m at this pH. Typically, aliquots were either removed from a single incubation at various times and diluted 10-fold with assay medium, or assay medium (450 µl) was added directly to a 50-µl reaction incubating in a cuvette. The last method was used for very fast reactions, and time courses of activity loss were developed from several identical reactions stopped at different times by addition of assay medium. Substrate concentrations (>K_m) and reaction dilutions were sufficiently high in assays to reduce the rate of inhibition to an inconsequential level. k_obs in pseudo-first order reactions was obtained by fitting the data to a single exponential function. Reactions at a 1:1 stoichiometry were analyzed by fitting the data to the relationship in Equation 3 using nonlinear least squares regression analysis. In Equation 3, v_t is the enzyme activity at any time, v₀ is the enzyme activity at time 0, and k_inh is the second order rate constant in terms of velocity and time units. Enzyme controls were used for experimental estimation of v₀, and fits always yielded values in close agreement with this value. To convert k_inh to molar units, the rate constant and v₀ obtained from fits were used to calculate the t1/2 for the reaction by Equation 3. This value was then used to calculate k_inh (M¹ s¹) according to the relationship k_inh = 1/(t_1/2 × [E]₀).

(Eq. 3)

SI values were determined from titrations of chymase hydrolytic activity by linear extrapolation of plots of activity loss versus [I]₀/[E]₀ ratio to zero activity. SI values and k_inh were used to calculate k_hyd, the apparent second order rate constant for hydrolysis, assuming the relationship described by Equation 4.

(Eq. 4)

Equation 4 assumes SI values > 1 arise from the presence of two available reaction pathways for the inhibitor: an inhibition pathway producing irreversible chymase inhibition (EI*) and a turnover or hydrolytic pathway producing a hydrolyzed, irreversibly inactivated inhibitor (I^c) and free enzyme (E). The relative flux of inhibitor (f_hyd/f_inh) proceeding through both pathways is SI-1. SI-1 also is equivalent to the partition ratio, which defines the number of catalytic turnovers per inactivation (36). Two minimal schemes that relate SI-1 to k_hyd/k_inh for chymase-serpin reactions are shown below (30, 31). In Scheme 1 the hydrolytic site is not at P1. SI-1 in this scheme is equivalent to the ratio of the rates of the reactions at the hydrolytic (k_hyd = k₃/K_PX, X  1) and inhibition (k_inh = k₂/K_P1) sites. In Scheme 2, both inhibition and hydrolysis progress through interaction at the reactive site. The partition ratio, k₃/k₂, in this mechanism also is equivalent to k_hyd/k_inh, as K_P1 is common to both pathways. Although these mechanisms are written assuming equilibrium, Equation 4 applies for steady state conditions as well.

pK values for pH-sensitive reaction parameters were obtained using Equations 5-7. Equation 5 relates the apparent value of a parameter (L_pH) at a given pH to that of the acid [L_EH] and basic [L_E] species derived from a single ionization. Equations 6 and 7 describe pH profiles exhibiting two different pK values; they are simplifications of the relationship involving three different protonated species (E₀ = EH₂ + EH + E) described by Fersht (26). Equation 6 assumes EH is the only active species, and Equation 7 assumes EH and E both have activity.

(Eq. 5)

(Eq. 6)

(Eq. 7)

Kinetic parameters were determined under reaction conditions corresponding to those used for rate constant and SI measurements. k_cat/K_m values observed for chymase hydrolysis of the substrate SucAAPF-NA were determined at [S]₀ 20-40 less than K_m by following the first order accumulation of product. In these measurements A_410 nm was continuously monitored in a thermostatted chamber (25 °C) until the reaction was nearly complete (0.5-1 h), and rate constants were obtained by fitting the data to a single exponential function. Individual k_cat and K_m values were determined from initial velocity versus [S]₀ plots, employing a nonlinear analysis to fit the data to the Michaelis-Menten rate equation. To obtain k_cat and K_m, [S]₀ was typically varied from 0.1 to 6 mM. _410 nm of NA was constant over the pH range of study.

The program Igor from Wavemetrics was used for fitting of data to the above equations.

RESULTS

A series of serpins exhibiting a range of SI values for inhibition of chymase at pH 8.0 was evaluated (Table I). The series is comprised of (i) rACT (recombinant wild type ACT or rACT-L358), (ii) point mutants of rACT containing a single substitution within the reactive loop at either P1 (numbers 2-4), P2 (number 5), or P3 (number 6), (iii) PI isolated from plasma, and (iv) ACT/PI loop chimeras (numbers 7 and 8) constructed by replacing a 6 or 9 residue sequence segment (P3-P3 and P6-P3, respectively) of the ACT reactive loop with the corresponding sequence of PI. The concentration of each inhibitor was determined by titration of a standardized amount of bovine chymotrypsin at pH 8.0 assuming 1:1 stoichiometry of interaction (SI = 1). Based on the A₂₈₀ of purified rACTs, all preparations were judged to have an SI near 1 and to be completely active. Confirming activity measurements, SDS-PAGE analysis of 1:1 reactions demonstrated nearly complete conversion of rACTs to SDS-stable chymotrypsin-inhibitor complex.

Table I. Effect of pH on SI and kinh for the interaction of human chymase with rACT, PI, and rACT variants Two patterns of change termed pattern A and B were observed upon lowering the reaction pH. Pattern A inhibitors showed a decrease in SI and little change in the magnitude of kinh, while Pattern B inhibitors showed an increase in SI and a marked decrease in the magnitude of kinh. kinh was determined from inhibition progress curves obtained in the presence of substrate (condition +[S]) or profiles of activity loss obtained for incubations with no substrate (condition  [S]). kinh values obtained under [S] conditions were typically higher than +[S] conditions, thus comparisons of rate constants are made according to the method of measurement. We suspect that the difference in magnitude is due to the presence of 9% Me2SO in the +[S] measurements. Inhibitor SIa kinh × 104 pH 8.0 pH 5.5 Ratio 8.0/5.5 Conditions Ratio 8.0/5.5 +[S] ([S]) pH 8.0 pH 5.5  +[S]  [S] +[S]  [S] M1 s1 M1 s1 Pattern A   1. rACT 4.1 1.1 3.7 2.0 5.0 2.7b 2.5 0.7 (2.0)   2. rACT-L358W 1.5 0.9 1.7 20.0 80.0 13.0 40.0 1.5 (2.0)   3. rACT-L358M 2.0 1.2 1.7 4.4 2.1 2.1   4. rACT-L358F 6.5 3.0 2.2 1.8 6.8 2.4 8.0 0.8 (0.9)   5. rACT-L357V 1.7 1.3 1.3 2.2 2.0 1.1   6. rACT-L361V 4.0 1.3 3.1 1.6 2.0 0.8 Pattern Bc   7. rACT-P3-P3-PI 1.3 2.4 0.5 0.8 2.5 0.08 0.08 10 (30)   8. rACT-P6-P3-PI 2.0 6.7 0.3 0.2 0.01    (20)   9. PI 6.2 15.0 0.4 0.2 0.003    (70) a SI values reported are the minimum [I]0/[E]0 ratio required for complete inhibition of chymase enzymatic activity. b The insensitivity of kinh for pattern A inhibitors to pH indicates that these inhibitors are interacting with the low pH form of chymase. This raises a question concerning the adjustment for substrate competition of kobs determined under +[S] conditions at low pH (Equation 2). Data consistent with substrate competition was obtained by demonstrating a reciprocal relationship between kobs and the factor (1 + ([S]0/Km)) for a reaction at pH 5.5. Thus, low pH values reported are adjusted for the presence of substrate using the appropriate Km. c Titrations for reactions with low rate constants contained [E]0 of 400 nM and were monitored for 24 h. By this time titrations had stabilized showing good linearity between zero and at least 75% inhibition. Rate constant measurements for slow reactions contained up to 3.5 µM [I]0 so that they could be followed to virtual completion in a reasonable time.

The chymase-ACT and -PI reactions have SI values of 4 and 6.0 at pH 8.0. We have shown previously that ACT hydrolysis responsible for the high SI is not due to the presence of a nontarget protease contaminating our chymase preparations (31). This was most clearly shown by SDS-PAGE analysis of chymase inhibited by rACT at an [I]₀ in excess of SI × [E]₀. In this analysis the stability of complex and intact inhibitor can be assessed after complete inhibition of chymase activity. Such studies demonstrated the continued presence of intact inhibitor and chymase·rACT complex at times of incubation well after (24 h) completion of the reaction. More recently, we have demonstrated that ACT inhibition of recombinant human chymase exhibits the same SI value as that obtained with native human chymase (37). Similar studies with the same conclusion now have been demonstrated for PI (not shown). Thus, SI values greater than 1 are an intrinsic property of the interaction of chymase with both ACT and PI.

The results in Table I show that all variants effectively inhibit chymase at pH 8.0 and 5.5. SI values were determined from the end points of titrations of chymase hydrolytic activity with each inhibitor. Linear titrations consistent with irreversible inhibition were observed with all inhibitors at both pH conditions. Apparent second order rate constants of inhibition (k_inh) for each inhibitor were determined from plots relating the observed first order rate constant for inhibition to [I]₀ as shown for examples in Fig. 1. Comparison of SI and k_inh values obtained for each inhibitor at pH 8.0 and 5.5 revealed two patterns of change (Table I). These patterns, termed A and B, differ for the most part in the effect of pH on the direction of the SI change and the magnitude of k_inh. Pattern A behavior was observed with rACT and rACT point variants (inhibitors 1-6) and Pattern B was observed with PI and ACT/PI loop chimeras (inhibitors 7-9).

Lowering the pH from 8.0 to 5.5 in reactions of chymase with pattern A inhibitors generally produced a decrease in the SI to a value near 1 and only a modest change in the magnitude of k_inh. Two inhibitors, rACT and rACT-L358F, exhibiting high SI values at pH 8.0 were studied further to establish the relationship between SI-1 and pH (Fig. 2). SI-1 is plotted on the ordinate as this value defines the number of inhibitor molecules hydrolyzed per inactivation and corresponds to the ratio of rate constants for hydrolysis and inhibition, k_hyd/k_inh, as described by Equation 4. Plots of the data reveal a sigmoid pattern approaching an SI-1 of 0 at low pH and increasing to a maximum value at pH 8.0. Above pH 8.0 little change was observed. Given the apparent pH independence of k_inh (Table I), the reduction in SI-1 values for pattern A inhibitors below pH 8.0 is likely due to a diminished flux through the hydrolytic pathway (decreased k_hyd). Apparent pK values of 6.6 and 6.0 were estimated for the SI-1 change in the chymase-rACT and -rACT-L358F reactions, respectively. Both pK values are consistent with titration of a His residue and suggest that k_hyd exhibits pH characteristics associated with the hydrolysis of substrates over the acid pH range. The dissimilarity of the pK values indicate that the effect of pH on the hydrolytic properties of chymase alone does not fully explain the reductions in SI-1.

Plots of k_inh and k_hyd for the reaction of chymase with rACT between pH 5.5 and 10 are shown in Fig. 3. Similar results not shown were obtained for the chymase-rACT-L358F reaction. Comparison between the plots reveal the diverse effects of pH on k_inh and k_hyd below pH 8.0. Above pH 8.0, similar decreases in k_inh and k_hyd are shown. The absence of change in SI-1 over the high pH range, as shown in Fig. 2, requires that k_hyd estimated by Equation 4 decrease in an identical manner to k_inh. The parallel decreases in k_hyd and k_inh are consistent with a change in enzyme structure that makes the high pH form of chymase unreactive with ACT.

In a serpin-protease reaction with an SI > 1, consumption of the inhibitor by hydrolysis may influence the magnitude of k_inh by a factor of 1/SI² (34, 38). Therefore a change in SI could confound interpretation of pH effects on the inhibition pathway. Evaluating the product k_inh × SI is a way of normalizing inhibition rate constants for comparative purposes when the mechanism suggests such an adjustment is warranted. For the following reasons we have not reported adjusted values for pattern A inhibitors and do not believe that the SI change between pH 5.5 and 8.0 is the primary reason for the pH insensitivity of k_inh. First, k_inh values for pattern A inhibitors (Table I) with SI values close to 1 at pH 8.0 exhibited little change upon reducing the reaction pH, indicating that a large SI change is not necessary for k_inh to appear pH-independent. This observation is revealed more definitively in Fig. 4 where titration of the chymase-rACT-L358W reaction is shown. Among pattern A inhibitors this reaction exhibits the lowest SI and reveals little change in SI between pH 5.5 and 8.0. Although k_inh at pH 8.0 is about 2-fold higher than at pH 5.5, the titration shows little evidence for a specific pH effect on the inhibition rate constant.

Our second argument for dismissing SI as a factor influencing k_inh is based on mechanistic concerns. SI only will affect k_inh under certain mechanistic conditions (38). Two general schemes producing high SI values that apply to chymase inhibition are considered under "Experimental Procedures." SI has no affect on k_inh determined under pseudo-first order conditions when the hydrolytic and inhibitory reactions proceed through different sites of interaction as in Scheme 1. Nor does it affect k_inh when only a single site of interaction is evident as in Scheme 2 and the reaction proceeds under rapid equilibrium conditions.² SI will affect k_inh in Scheme 2 when the reaction is under steady state conditions or if a suicide mechanism of inhibition applies (4, 34, 38).

Recent evidence³ using MALD mass spectroscopy to analyze the products from a chymase-rACT reaction suggests that P2-P1 (Leu³⁵⁷-Leu³⁵⁸) is the major site of the hydrolytic pathway instead of P1-P1 (Leu³⁵⁸-Ser³⁵⁹). This result suggests that the high SI for this reaction is due primarily to a Scheme 1 mechanism. Supporting this MALD analysis, the interaction of chymase with the P2-variant rACT-L357V at pH 8.0 revealed a reduced SI of 1.7 compared with 4 for rACT (Table I). With respect to the chymase-rACT-L358F reaction, several techniques (31), including MALD mass spectroscopy,³ indicate that P1-P1 is the site of cleavage, consistent with Scheme 2. Taking into account the possible influence of SI on k_inh for this reaction by evaluating the effect of pH on the product of k_inh × SI also does not reveal a significant pH dependence for k_inh over the acid pH range. Thus, based on consideration of experimental evidence and mechanistic arguments, we conclude that the pH insensitivity of k_inh for the interaction of chymase with pattern A inhibitors is primarily a property of the inhibitory pathway and is not significantly related to a change in flux through the hydrolytic pathway.

SDS-PAGE of chymase-rACT reactions at pH 8.0 and 5.5 are shown in Fig. 5, A and B. The [I]₀/[E]₀ ratio of reactions was varied in a titration-like manner. Reactions exhibiting intact inhibitor (rACT) demonstrate completion of the titration ([I]₀ > SI × [E]₀). The intense 70-kDa bands appearing in pH 5.5 reactions at [I]₀/[E]₀ ratios close to and including 1 indicate quantitative formation of a covalent complex between chymase (30 kDa) and rACT (45 kDa) at low pH. These results are in agreement with an SI of near 1. Reactions at pH 5.5 were denatured in an SDS solution buffered at pH 5.5 to ensure that the sample was not exposed to high pH. In contrast to low pH, the major reaction product at pH 8.0 is cleaved inhibitor and the titration end point is above 4, consistent with the high SI of this reaction. Thus, covalent complex indicative of an acyl-enzyme product is formed at low pH in a quantitative manner.

The rate of covalent complex formation relative to the rate of chymase inhibition by rACT at low pH is shown Fig. 5C. Evaluated is an incubation under second order reaction conditions ([E]₀ = [I]₀ = 1 µM), which was monitored for loss of activity and appearance of SDS-stable complex. Reactions analyzed by SDS-PAGE were stopped at the indicated times by the addition of trichloroacetic acid, and the loss of enzymatic activity was determined by measurement of residual catalytic activity. k_inh for activity loss was 25,000 M¹ s¹, consistent with values obtained from progress curves (Fig. 1). Inhibition was half complete in 40 s and 75% complete in 2 min. SDS gels of identical reactions stopped at 15 s and 2 min demonstrate a high molecular weight complex band accumulating at a rate in qualitative agreement with the rate of activity loss. Agreement between the rate of inhibition and complex formation also was observed for the chymase-rACT-L358W reaction at pH 5.0 measured under conditions where the half-life of activity loss was under 7 s. Thus, the rate of covalent complex formation and enzymatic activity loss appear indistinguishable at low pH under our experimental conditions.

The three inhibitors of this group are wild type PI and two PI/ACT loop chimeras (inhibitors 7-9 in Table I). Lowering the pH from 8.0 to 5.5 in reactions of chymase with these inhibitors produced an increase in SI and a marked decrease in the magnitude of k_inh (10- to >100-fold). k_inh for the chymase-PI reaction was not adjusted for SI, because previous work had located the hydrolytic site at P7-P6 (Phe³⁵²-Leu³⁵³), indicating that a Scheme 1 mechanism was responsible for the high SI (30). The increase in SI and the decrease in k_inh for reactions at pH 5.5 were not due to inhibitor denaturation as demonstrated by inhibition of chymotrypsin (Table II). Titrations of chymotrypsin with each pattern B inhibitor demonstrated an SI close to 1 at pH 8.0 and 5.5. Additionally incubation of inhibitors at pH 5.5 for 24 h did not significantly reduce their activity toward chymotrypsin. Furthermore, k_inh for the chymotrypsin-PI reaction did not significantly change under the two pH conditions, providing another example where k_inh is insensitive to low pH.

Table II. The effect of pH on the interaction of chymotrypsin with PI and PI/ACT loop chimeras Inhibitora SI kinh pH 8.0 pH 5.5 pH 8.0 pH 5.5  M1 s1 1. PIb 1.0 0.8 2  × 106 2  × 106 2. rACT-P3-P3-PIc 1.0 1.0 3  × 104 2  × 103 3. rACT-P6-P3-PIc 1.0 1.0 4  × 104 7  × 103 a Titrations and rate constant measurements were performed in a solution of 0.5 M NaCl, 0.01% dodecyl maltoside, with 0.1 M Tris-HCl, pH 8.0, or 0.1 M MES, pH 5.5, as buffer. b kinh was determined under second order conditions. c kinh was determined under pseudo-first order conditions in the absence of substrate.

Further analysis of chymase-PI and -rACT-P6P3PI reactions is shown in Figs. 6 and 7, where k_inh, SI-1, and k_hyd are reported as a function of pH. Both PI and rACT-P6P3PI reactions showed a sigmoidal decrease in k_inh as the pH was lowered from 8.0 to 5.5. An apparent pK of about 7 was estimated for the k_inh change observed in both reactions. Above pH 8.5, k_inh for the chymase-PI reaction showed a second phase of marked increase. Increased k_inh values for reactions above pH 8.5 may be reflective of inhibitor interaction with a catalytically enhanced form of chymase, as will be subsequently described. In contrast to k_inh, SI-1 progressively increased as the reaction pH was lowered. The abrupt increase in SI-1 for the chymase-PI reaction over the interval of pH 5.5 to 5.0 is most likely related to acid denaturation induced by protonation of PI carboxyl groups, while the nature of a similar increase from pH 6.0 to 5.5 in the chymase-rACT-P6P3PI reaction is unclear.

The effect of pH on the human chymase-rACT-P6P3PI reaction.

To evaluate the SI-1 change observed in both the chymase-PI and -rACT-P6P3PI reactions, k_hyd values were estimated by Equation 4. Ignoring the abrupt increases, k_hyd for both reactions progressively decreased as the pH was reduced from 8.0 to 5.5 (Figs. 6C and 7C), consistent with a substrate-like reaction. Decreases in both k_inh and k_hyd could produce the observed increases in SI-1 (k_hyd/k_inh) if the change in k_inh as a function of pH was proportionately greater than that in k_hyd. A slightly lower pK for k_hyd compared with k_inh could provide such a condition. The pK for the k_hyd decrease in the chymase-PI reaction was estimated to be 6.6. The dashed line through the SI-1 values for the chymase-PI reaction in Fig. 6B was calculated assuming pK values of 7.1 and 6.6 for k_inh and k_hyd, respectively, and a 5.5-fold higher value for k_hyd than k_inh at the upper end point of the titration. Dissimilarity of the apparent pK values for k_inh and k_hyd suggest that effects in addition to those on the catalytic His residue of chymase are influencing the sites of interaction mediating hydrolysis and inhibition (Scheme 1).

SI-1 for the chymase-PI reaction evaluated at pH 9.5 was 1.5. Although the lowest SI so far seen for this reaction, k_hyd was higher than that observed at pH 8.0 and 8.5. Thus, k_hyd appears to be enhanced by high pH in parallel with k_inh, although the effect of pH on k_inh appears to be proportionately greater.

SDS-PAGE of chymase-PI reactions at pH 7.0 and 9.5 demonstrated covalent complex formation (data not shown). The banding pattern of the pH 7.0 reaction was consistent with the high SI of 7 observed from titrations of enzymatic activity. The banding pattern of the pH 9.5 reaction revealed a somewhat lighter intensity complex band than expected for an SI of 2.5. The intensity of the complex band increased when reactions were stopped by addition of trichloroacetic acid and denaturation in SDS at neutral pH instead of SDS at pH 9.5. We therefore attribute the weak complex band in samples denatured at pH 9.5 to the more rapid hydrolysis of an acyl bond under alkaline conditions. Thus, complex formation appears to accurately mirror SI values in reactions performed at high as well as low pH.

The effect of pH on chymase catalyzed hydrolysis of the substrate SucAAPF-NA is shown in Fig. 8. Upon reduction of the pH from 8.0 to 5.0, k_cat/K_m and k_cat decreased in a sigmoidal manner approaching zero at low pH, while K_m remained relatively constant. The apparent pK value for both decreases was 7.0, consistent with titration of the catalytic triad His residue. These results indicate that chymase exhibits a low pH form that is catalytically inactive. Above pH 8.0, k_cat/K_m and k_cat markedly increased and K_m markedly decreased. Changes were progressive until about pH 11.0. At pH 11.5 chymase lost activity rapidly when assayed under conditions of [S]₀ << K_m. The pattern of change above pH 8.0 is unusual for a serine proteinase for it suggests that chymase exhibits a high pH form with enhanced catalytic efficiency. Activity loss above pH 11.0 is probably related to the disruption of the Ile-Asp salt bridge.

Scheme 3 depicts the pH properties of chymase hydrolysis of SucAAPF-NA. It shows low (H₂E), neutral (HE), and high pH (E) forms of chymase interacting with substrate to form Michaelis complexes with both HES and ES proceeding to product. Product formation from ES would account the increase in k_cat at alkaline pH. Formation of H₂ES from H₂E is depicted because chymase inhibition measured by progress curves at low pH was sensitive to [S]₀ present in the reaction (see Table I, legend). The competition between inhibitor and substrate implied by this result indicates that substrate is able to combine with the low pH form of chymase.

DISCUSSION

A body of evidence indicates that inhibition of proteases by serpins involves cleavage of the reactive site peptide bond (P1-P1) and formation of a stabilized, acyl-enzyme, inhibitory complex. SDS-PAGE of serpin-proteinase reactions with 1:1 inhibition stoichiometries demonstrate both a high molecular weight complex consistent with covalent linkage between protease and serpin and excision of a 4-5 kDa COOH-terminal peptide consistent with cleavage of the reactive site peptide bond (39-41). Kinetic studies using serpins chemically labeled in the reactive loop show environmental changes of the P1 residue consistent with cleavage and separation of the residues forming the P1-P1 peptide bond (19-22). Dependence of inhibition on cleavage of the P1-P1 bond may suggest that the magnitude of k_inh is dependent on the hydrolytic activity of the protease and that serpin-protease interactions may demonstrate a pH profile similar to that found for hydrolysis of substrates (k_cat/K_m and k_cat). A low and high pH dependence for k_inh similar to that observed for hydrolysis of substrates has been reported for the inhibition of thrombin by antithrombin and protease nexin 1 (13).

Contrary to these expectations and observations, the different pH dependencies determined for the chymase-rACT and -PI reactions show that the value of k_inh may not always vary in parallel with the hydrolytic activity of the enzyme. While the pH profile for k_inh of the chymase-PI reaction followed that of k_cat/K_m and k_cat for hydrolysis of a peptide substrate, the chymase-rACT reaction did not. In fact, the acid pH insensitivity of k_inh for the chymase-rACT reaction reveals a reaction where the rate of inhibition appears independent of enzyme hydrolytic activity. Chymase-serpin reactions exhibit an SI greater than 1 due to consumption of inhibitor by a concurrent substrate-like reaction. Although the pH dependence for SI in the chymase-rACT and -PI reactions also differed, k_hyd, the estimated rate constant for the substrate-like reaction, always appeared to exhibit substrate-like pH behavior regardless of the mechanism (Scheme 1 or 2) producing the high SI.

The possible effects of SI on the rate of inhibition was ruled out as the cause for different pH dependencies of k_inh, as discussed under "Results." We speculate that the diverse effects of pH on k_inh and SI are the result of the unequal abilities of each inhibitor to interact with the various pH titratable forms of chymase. As shown by analysis of substrate hydrolysis kinetics, chymase exhibited three pH-dependent forms: a low pH form, which was catalytically inactive; a neutral pH form, which was nominally active; and a high pH form, which was "catalytically enhanced." The chymase-PI reaction exhibited a good correspondence of k_inh and k_hyd with chymase catalytic activity (k_cat/K_m, k_cat) over the entire pH range of analysis (compare Fig. 6 with 8). Such behavior would be consistent with the ability of PI to react with the neutral and high pH forms of chymase as both an inhibitor and substrate, and the inability of PI to react as either an inhibitor or substrate with the low pH form of chymase. The pH dependence of chymase-rACT reaction, in contrast, would be consistent with an almost opposite set of interactions. The insensitivity of k_inh to acid pH suggests that rACT has the ability to react with low and neutral pH forms of chymase as an inhibitor. Though active as an inhibitor at low pH, reductions estimated for k_hyd from the SI decrease indicate that rACT is not a chymase substrate at low pH. The identical decreases in k_inh and k_hyd over the high pH range (Fig. 3) suggest that in contrast to PI rACT lacks the ability to interact with the high pH form of chymase as either an inhibitor or substrate.

For this interpretation of the results to apply some aspect of reactive loop structure must differ between inhibitors to allow for the different interaction patterns at low and high pH. In addition some feature of the rACT-chymase inhibition reaction must allow for the different pH dependencies of k_inh and k_hyd over the low pH range. Supporting a structural distinction, rACT/PI loop chimeras, at least over the low pH range, exhibited pH dependencies more like PI than ACT, while pattern A inhibitors, which contained only single amino acid substitutions in the reactive loop, behaved more like rACT.

Several mechanisms that may explain the acid pH insensitivity for the chymase-ACT reaction are discussed below using Scheme 4 as a working model for the reaction. Scheme 4 shows ACT interacting with the neutral (HE) and low (H₂E) pH forms of chymase, and it describes the presence of two inhibitory steps (1 and 2) subsequent to Michaelis complex (HEI and H₂EI) formation. HE-I* signifies a the tight covalent complex observed under all pH conditions by SDS-PAGE. The inability of the high pH form of chymase (E) to interact with ACT accounts for its depiction as an unreactive form. Hydrolytic steps affecting the SI are not considered because the chemical steps determining turnover of ACT as a substrate did not appear to affect k_inh significantly. According to Scheme 4, k_inh would be pH independent if step 1 is noncovalent and rate-limiting or if the products of step 1 are relatively stable. Evidence for a relatively stable noncovalent inhibitory intermediate in serpin-protease reactions has been suggested from studies with certain anhydroproteases (42, 43), and evidence for a noncovalent rate-limiting step has been provided by Stone and Le Bonniec (16). Evaluation of the chymase-ACT and -rACT-L358W reactions at low pH by SDS-PAGE failed to demonstrate a noncovalent complex as a product or reaction intermediate, however. In both reactions, production of SDS-stable complexes occurred at a rate qualitatively similar to that of inhibition. Covalent complex formation would appear pH-insensitive if step 2, enzyme-acylation, is more rapid than the breakdown of step 1 products at neutral and low pH. Our methodology could not have detected a noncovalent inhibitory complex if the rate constant for step 2 was faster than 0.1 s¹. A rate constant for SDS complex formation of about 10 s¹ has been measured for the interaction of chymotrypsin with rACT (44).

The low pH insensitivity of k_inh also can be rationalized in the context of an alternative model providing for only covalent inhibitory intermediates as proposed by Gettins et al. (4). In this model enzyme acylation is considered as a series of steps linked to loop insertion. According to Scheme 4, k_inh would be pH-independent if step 1, the pH-sensitive formation of the tetrahedral intermediate, is fast and step 2, the collapse of the tetrahedral intermediate to acyl-enzyme modulated by loop insertion, is slow. Two further mechanisms that do not require a pH insensitive rate-limiting step also are possible. The first alternative presumes that formation of the tetrahedral intermediate for enzyme acylation is rate-limiting (Scheme 4 without step 1), but kinetically invisible because of a dramatic shift in the apparent pK for k_inh toward very low pH. This may occur if the serpin is acting like a "sticky" substrate and is capable of forming HEI from H₂E, as pointed out by Stone and Hermans (13). The second alternative presumes low pH is affecting the ionization and/or structural state of ACT making it a much more efficient inhibitor (I^super) of catalytically active chymase (EH) than ACT at neutral pH.

This study defines the effect of pH on the interaction of human chymase with serpins and provides evidence for a rate-limiting step of inhibition, which does not depend on the hydrolytic efficiency of the enzyme. It also establishes the pH characteristics of human chymase catalysis and provides evidence for distinct high and low pH enzyme forms with active site structures that can at least discriminate between the reactive site loops of ACT and PI. Further understanding of these reaction and structural properties will require a more complete characterization of the rate-limiting steps for chymase-serpin reactions, as well as a more complete analysis of serpin and chymase structural properties as a function of pH. Recent production of recombinant human chymase in high yield⁴ (unpublished from this laboratory) will help in this regard.