Reaction of Azapeptides with Chymotrypsin-like Enzymes NEW INHIBITORS AND ACTIVE SITE TITTRANTS FOR CHYMOTRYPSIN A,, SUBTILISIN BPN’, SUBTILISIN CARLSBERG, AND HUMAN LEUKOCYTE CATHEPSIN G*

A series of new azapeptide p-nitrophenyl esters containing a variety of PI aza-amino acid residues have been synthesized, and the reaction of these azapeptides with chymotrypsin A,, subtilisin BPN’, subtilisin Carlsberg, and human leukocyte cathepsin G at pH 47 has been studied. These azapeptides were found to be very useful as active site titrants and inhibitors of serine proteases with chymotrypsin-like specificity. Stable acyl derivatives of serine proteases are formed in the reaction with azapeptides and can be used for future crystallographic investigations. The effects of changing the nature of the PI’ leaving group (”ONp, “OPh, -OCH2CF3, -0Et) for these azapeptides was also investigated. N-Acetyl-L-alanyl-L-alanyl-a-aza-norleucinep-nitrophenyl ester can be used as an active site titrant for human leukocyte cathepsin G and N-acetyl-L-alanyl-a-azaphenylalanine p-nitrophenyl es- ter is a suitable titrant for chymotrypsin &, subtilisins, or cathepsin G. The aza-amino acid Ac-Aphe-ONp was obtained from Nutritional Biochemicals, Boc-N(CH3)NHz was prepared by the method of Jen- sen et al. (1968) and C6H5CH*NHNH2 by the method of Kost and Sagitullin (1959). Boc-Tyr-ONp was obtained from Bachem. All other amino acid derivatives, reagents, and solvents were analytical grade. Melting points are uncorrected. Thin layer chromatography was carried out with Merck Silica Gel G plates. Materials were detected using 12 vapor or FeC13-K3[Fe(CN)6] spray reagent (Ertel Horner, 1962). Column chromatography was performed with Merck Silica Gel 60 absorbent. Mass spectra were taken on an Hitachi Perkin-Elmer RMU-71 instrument and nmr spectra were taken on a Varian T-60 instrument. Deuterochloroform (CDCI,) and deuterodimethyl sulf- oxide (DMSO-D,) were the solvents used in obtaining nmr spectra. ir spectra were taken on a Perkin-Elmer 457 instrument. The ir spectra were taken in Nujol mulls. Assays were performed on Beck-man Model 25 or 35 spectrometers and a Radiometer automated pH- stat Model TTT11, and all spectrometer cells had a 1-cm path length.

Aza-amino acid residues are analogs of amino acids in which the a-methine group has been replaced by a nitrogen atom (Fig. 1). The substitution has a profound effect on the reactivity of aza-amino residues in simple derivatives or in peptides. Kurtz and Niemann (1961, a and b) were the first to study the reaction of an aza-amino acid derivative with a protease and showed that Ac-Aphe-OEt was a weak reversible inhibitor of chymotrypsin (KI = 20 mM).' Elmore and Smyth (1968) demonstrated that a more reactive ester derivative of an aza-amino acid (Ac-Aphe-ONp) could be utilized as an active site titrant for chymotrypsin. A polymer containing an Aphe-OPh ligand was then found to be useful for the covalent affinity purification of chymotrypsin (Barker et al., 1974). We prepared peptide 4-nitrophenyl esters containing a PI azaamino acid residue and showed that they were excellent inhibitors and titrants of chymotrypsin, subtilisin, and porcine pancreatic elastase and that considerable specificity could be obtained by altering the PI residue (Powers and * This investigation was supported by Grant HL29307 from the National Institutes of Health to the Georgia Institute of Technology. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence may be addressed. Aza-amino acid residues in which the a-methine of an amino acid residue is replaced by a nitrogen atom are abbreviated by placing an "A" before the standard three letter abbreviation for that amino acid. Thus, a-am-alanine will be abbreviated Aala. Any peptide which contains an aza-amino acid residue will be referred to as an azapeptide in this paper. The other abbreviations used are defined in the miniprint.  Carroll, 1975;.' Subsequently, two azaornithine phenyl ester derivative were shown to be inhibitors of trypsin and thrombin (Gray and Parker 1975;Gray et al., 1977). Inhibition of all of these serine proteases by the aza-amino acid phenyl or 4-nitrophenyl esters is believed to be due to the acylation of the active site serine residue forming an acylated enzyme. The nitrogen atom adjacent to the acyl carbonyl group gives a special stability to the acyl-enzymes, which are substantially less reactive toward deacylation than normal acyl-enzymes. Aza-amino acid derivatives without reactive leaving groups do not acylate serine proteases, but simply act as reversible inhibitors (Kurtz and Niemann, 1961a;Dorn et al., 1977).
Chymotrypsin A,, subtilisin BPN', subtilisin Carlsberg, and human leukocyte cathepsin G are serine proteases whose substrate specificity is directed toward PI amino acid residues with aromatic or large hydrophobic side chains. The active sites and substrate specificity of chymotrypsin (Blow 1971;Bender and Killheffer, 1973) and subtilisin BPN' and subtilisin Carlsberg (Kraut, 1971;Markland and Smith, 1971;Hunt and Ottesen, 1961) have been widely investigated. Studies on the reactivity of cathepsin G toward peptide 4-nitroanilide substrates (Nakajima et al., 1978), amide substrates (McRae et al., 1980), and peptide chloromethyl ketone inhibitors  have been reported.
In this paper, we report a study of the reaction of azapeptides containing a variety of PI aza-amino acid residues with chymotrypsin A,, subtilisin BPN' and subtilisin Carlsberg, and human leukocyte cathepsin G and have found the azapeptides to be useful as active site titrants and inhibitors and for generating stable acyl derivatives of serine proteases for crystallographic investigations. In addition, we have investigated the effects of changing the nature of the PI' leaving group. Our results demonstrate that azapeptides can be used as selective inhibitors and active site titrants of serine proteases.
research group of the Dept. of Biochemistry, University of Georgia. The aza-amino acid Ac-Aphe-ONp was obtained from Nutritional Biochemicals, Boc-N(CH3)NHz was prepared by the method of Jensen et al. (1968) and C6H5CH*NHNH2 by the method of Kost and Sagitullin (1959). Boc-Tyr-ONp was obtained from Bachem. All other amino acid derivatives, reagents, and solvents were analytical grade. Melting points are uncorrected. Thin layer chromatography was carried out with Merck Silica Gel G plates. Materials were detected using 12 vapor or FeC13-K3[Fe(CN)6] spray reagent (Ertel and Horner, 1962). Column chromatography was performed with Merck Silica Gel 60 absorbent. Mass spectra were taken on an Hitachi Perkin-Elmer RMU-71 instrument and nmr spectra were taken on a Varian T-60 instrument. Deuterochloroform (CDCI,) and deuterodimethyl sulfoxide (DMSO-D,) were the solvents used in obtaining nmr spectra. ir spectra were taken on a Perkin-Elmer 457 instrument. The ir spectra were taken in Nujol mulls. Assays were performed on Beckman Model 25 or 35 spectrometers and a Radiometer automated pHstat Model TTT11, and all spectrometer cells had a 1-cm path length. The syntheses of all new azapeptides used in this investigation are reported in a miniprint supplement to this paper or to the following paper (Powers et al., 1984)? Reaction of Enzymes with Azapeptides-The reaction of chymotrypsin, subtilisin BPN', and subtilisin Carlsburg with a series of azapeptides was carried out in solutions which contained approximately a 50-fold excess of azapeptide over enzyme. Stock solutions of the azapeptides in acetonitrile were prepared at a concentration of 5.0 mM and the use of acetonitrile was necessary to increase the solubility of the azapeptide in the reaction mixture. Enzyme stock solutions of chymotrypsin A, (1 mM HCI) and subtilisin had a concentration of approximately 100 p~. Enzyme concentration was quantitated using = 20.5 for chymotrypsin A (Wilcox, 1970), A% = 11.7 for subtilisin BPN', and Ai% = 9.6 for subtilisin Carlsburg (Ottesen and Svendsen, 1970). Chymotrypsin solutions were stored at 4 "C, while subtilisin solutions were used within 2 h. Four buffer solutions were prepared for the pH dependence studies: pH 7.0 (0.1 M phosphate), pH 6.0 (0.1 M citrate), pH 5.0 (0.1 M citrate), and pH 4.0 (0.1 M acetate). All reactions were performed at 25 "C and were carried out by mixing 100 pl of the azapeptide stock solution with 2.0 ml of the appropriate buffer in a cuvette. An identical reference sample was prepared and a base-line was recorded at 345 nm. The reaction was initiated by the addition of 100 pl of 1 mM HCI to the reference cuvette followed by the addition of 100 pl of the enzyme stock solution to the sample cuvette. The recorder was immediately started upon the addition of enzyme to the sample cell and the reaction rate was observed. Assay solutions were freshly prepared since the azapeptides underwent slow hydrolysis upon standing in buffered aqueous solutions. Some representative azapeptide hydrolysis rates are listed in the table in the supplementary section of the following paper (Powers et al., 1984).
An extinction coefficient of 6250 M" cm" at 345 nm was used for p-nitrophenol at pH 6.0. The pK, of p-nitrophenol is 7.04 and 345 nm is near an isosbestic point of p-nitrophenol and this extinction should vary only slightly in the pH range of 4-7. However, the substrates have significant absorbance at 345 nm and, to avoid any systematic errors, the extinction coefficient should be measured with each substrate, buffer, and spectrometer utilized. The extinction coefficients for the following azapeptides were measured by complete hydrolysis of the azapeptide to products and the range of values obtained in the pH 4,5, 6, and 7 buffers are listed: Ac-Ala-Aala-ONp, 5920-6190; Ac-Ala-Ala-Aala-ONp, 5880-6580; Ac-Ala-Ala-Pro-Aala-ONp, 4890-5160; Ac-Aphe-ONp, 4920-5220; Ac-Ala-Aphe-ONp, The reaction of cathepsin G with azapeptides was carried out by adding 50 pl of the azapeptide in 1,2-dimethoxyethane to 1.0 ml of buffer. Aliquots (500 pl) of this solution were added to the sample and reference microcells and a base-line was established. The reaction was initiated by the addition of 50 p1 of 1 mM HCI to the reference 5340-5590; HONp, 5820-6320.
Portions of this paper are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 83M-2535, cite the authors, and include a check or money order for $3.60 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press. cell and 50 pl of enzyme to the sample cuvette. The exact protein concentration of cathepsin G was determined using E% = 6.64 (Baugh and Travis, 1976).
Determination of Deacylation Rates-The slow deacylation rates observed with chymotrypsin A, were determined using Ac-Tyr-OEt as a substrate (Wilcox, 1970;Cunningham and Brown, 1956) and one example of the procedure follows. A stock solution of Ac-Ala-Aphe-ONp in acetonitrile (5 mM) was prepared and used immediately. A stock solution of chymotrypsin A,was made up in 1 mM HC1 and had a concentration of approximately 100 p M (exact concentration determined by A m ) . The reaction was initiated by the addition of 100 pl of Ac-Ala-Aphe-ONp to 2.0 ml of the 0.10 M citrate buffer (pH 5.0). This was followed by the addition of 100 pl of the enzyme stock solution to the reaction mixture. The mixture was allowed to incubate for 5 min. A 200-pl portion of the reaction mixture was withdrawn and added to 2.0 ml of the appropriate buffer solution. At various time intervals, 100-pl fractions were withdrawn and assayed using the substrate Ac-Tyr-OEt with the pH-stat. The initial concentrations in the deacylation reaction mixture were as follows: enzyme, -0.5 pM; substrate, 23 pM; acetonitrile, 0.5% (v/v) in a total volume of 2.2 ml. Deacylation rates were then calculated using a least squares computer program. Correlation coefficients of better than 0.994 were obtained.
Inhibition of Cathepsin G with Azupeptlde Analogs-Inhibition of cathepsin G with a series of azapeptides was carried out in solution which contained at least a 65-fold excess of inhibitor over enzyme. Stock solutions of azapeptide in acetonitrile were prepared at a concentration 0.25-7 mM. The enzyme stock solution was made in 1 mM HC1 and had a concentration of -4 p~. The reactions were performed at 25 "C and were carried out by mixing 50 pl of inhibitor solution and 50 p1 of an enzyme solution with 1 ml of buffer. Aliquots (100 pl) were removed from the reaction mixture at regular time intervals and the residual enzymatic activity was measured using the Boc-Tyr-ONp spectrophotometric assay . The final concentrations of inhibitor and enzyme in the reaction mixture are shown in Table VI.

RESULTS
Synthesis of Azapeptides-A series of new azapeptides were prepared by three general methods (Gante, 1966;Dutta and Morley, 1975). Each synthesis required the synthesis of a 2substituted hydrazide (RCO-NHNHR', RC0,H = a blocked amino acid or a peptide acid) which was subsequently coupled withp-nitrophenyl chloroformate to produce the desired azapeptide p-nitrophenyl ester. The most convenient route, hydrazinolysis of an ester or coupling of a peptide acid with a substituted hydrazine, requires the availability of the appropriate hydrazine. The procedure was used in the synthesis of Ac-Ala-Aphe-ONp and Ac-Ala-Ala-Aala-ONp and requires the separation and identification of two isomeric hydrazides ( i e . the 1-substituted hydrazide Ac-Ala-Ala-NMeNHn and the 2-substituted hydrazide Ac-Ala-Ala-NHNHMe were obtained upon reaction of MeNHNH, with Ac-Ala-Ala-OBzl).
The correct isomer is readily identifiable since the alkyl protons of the 1-substituted hydrazide (6 3.2 for the N-CH, of Ac-Ala-Ala-N(CH,)NH, and 6 4.62 for the N-CH, of Ac-Ala-N(CH2CsH5)NH2) appear further downfield in the nmr spectrum than those of the 2-substituted hydrazide (6 2.4 for the N-CH3 of Ac-Ala-Ala-NHNCH3 and 6 3.82 for the N-CHZ of Ac-Ala-NHNCH2C6H5). An alternate, although longer, unambiguous synthesis of Aala derivatives involves coupling of a blocked amino acid or peptide acid with NH2NMe-Boc followed by deblocking with CF8C02H to give only the correct hydrazide. Since few substituted hydrazines are commercially available, the majority of azapeptides reported in this paper  were synthesized by a route similar to that utilized by Elmore and Smyth (1968) for the preparation of Ac-Aphe-ONp and Kurtz and Niemann (1961b) for the synthesis of Ac-Aphe-OEt. In this procedure, a peptide hydrazide is converted to a hydrazone by reaction with an aldehyde or ketone. Reduction then gave a 2-alkylhydrazide. All new compounds were characterized by combustion analysis, infrared, nmr, mass spectra, and tlc. A particularly useful tool for the identification of these compounds was nmr. The nmr spectrum of each compound exhibited certain characteristic resonances which allowed l-and 2-substituted hydrazides to be readily distinguished as well as being able to tell precursors from products (Powers and Carroll 1975;Condon, 1972). Azapeptides as well as peptide and amino acid hydrazides and hydrazones were easily distinguished from other compounds on tlc by treatment of the plates with ferric chloride followed by potassium ferricyanide (Ertel and Horner, 1962). The appearance of a royal blue color indicated indicated the presence of these groups. Likewise, the p-nitrophenyl moiety was detected by treatment with base which produced a bright yellow color due to the p-nitrophenolate anion.
Kinetic Considerations-The kinetics of the reaction of serine proteases with substrate-related acylating agents are described by Equation 1. where E . S is the enzyme-substrate complex, E-S is the acylenzyme intermediate, and P' is p-nitrophenol in the case of azapeptide p-nitrophenyl esters. Deacylation results in regeneration of active enzyme. An example of this type of reaction is shown in Fig. 2 when the release of p-nitrophenol is followed. The biphasic curve represents the initial stoichiometric burst of p-nitrophenol followed by the steady state turnover of the acyl-enzyme. The equation of this curve is represented by Equation 2 (Kezdy and Kaiser, 1970;Bender et al., 1967;Bender et al., 1966).
As t becomes very large, Equation 2 reduces to Equation 3. These requirements are often met with reactive esters of azapeptides, since these compounds will acylate certain serine proteases rapidly to form stable acyl-enzymes. In many cases, the deacylation rates are so slow that they can only be measured by following the reappearance of enzyme activity after isolating the acyl-enzymes. The active enzyme concentration can be calculated from the release of p-nitrophenol ([El0 = ~A345/tHoN~, tHONp = 6250 at pH 6.0). The turnover rate kc,, is then (AA,,,/s)/AA,,,(burst). Two azapeptide pnitrophenyl esters (Ac-Aphe-ONp and Ac-Ala-Aphe-ONp) have previously been shown to exhibit the above kinetic behavior and have been used to titrate chymotrypsin A, and subtilisin BPN' (Elmore and Smyth, 1968;Powers and Carroll, 1975;. Reaction of Chymotrypsin with Azapeptides- Table I  . The enzyme was acylated by the two PI azaphenylalanine derivatives very rapidly over a pH range of 4-7 with no measurable turnover of the azapeptide on the time scale of the assay procedure. Furthermore, the initial acylation reaction proceeded with a 1:l stoichiometry with respect to enzyme concentration based on the release of p-nitrophenol. Azapeptides which contained aza-amino acid residues possessing long alkyl side chains (e.g. Aval, Anva, Aleu, and Aile) also acylated the enzyme stoichiometrically and showed no measurable turnover. In contrast, the shorter aza-alanine-containing peptides Ac-Ala-Aala-ONp and Ac-Ala-Ala-Aala-ONp would only react with the enzyme at higher pH values, while the longer aza-alanine peptide (Ac-Ala-Ala-Pro-Aala-ONp) reacted with the enzyme stoichiometrically over a pH range of 4-7 with no measurable turnover. The azapeptide Ac-Ala-Ala-Agly-ONp did not show any measurable reaction with chymotrypsin at either pH 6 or 7. The enzyme had an average activity of -89% with respect to the protein concentration determined from Azm.
The lower limit value of 1.8 x lo-, s" for the turnover rates was determined by preparing a series of lines of known slopes and superimposing each line on a previously recorded base-line. The line with the smallest slope value which had a preceptible deviation from the base-line was used to calculate the lower limit value for kcat. Since one purpose of the investigation was the preparation of stable acyl derivatives of chymotrypsin for crystallographic investigation, we investigated the acylation rates of chymotrypsin A, at pH values where most crystallographic studies are performed. At pH 5.8, Ac-Aphe-ONp, Ac-Ala-Aphe-ONp, and Z-Ala-Ala-Pro-Aala-ONp had k2 values >0.20 s-' , our limit of detection. At pH 5.0, the k, values for the first two azapeptides were also >0.20 s-', while that for Z-Ala-Ala-Pro-Aala-ONp was 0.056 s-'.  Reaction of Subtilisins with Azapeptides- Table I1 shows the results for the studies in which each of the 11 azapeptides were reacted with subtilisin BPN' under assay conditions identical with those used with chymotrypsin A,. The enzyme was acylated rapidly by all of the compounds with the exception of Ac-Ala-Ala-Agly-ONp which showed no release of pnitrophenol on the time scale of the experiment. The turnover rates showed substantial variance for the series of compounds. They were very slow for the reactions involving the two azaphenylalanine peptides. In contrast, the reactions involving the three aza-alanine peptides had very high turnover rates with Ac-Ala-Ala-Pro-Aala-ONp having the highest. The azapeptides possessing longer side chains on the PI aza-amino acid showed intermediate turnover rates. The initial burst of p-nitrophenol had a 1:l stoichiometry with respect to enzyme concentration and commercial subtilisin BPN' was found to be -82% active. The acylation rates (k2) for Ac-Aphe-ONp, Ac-Ala-Aphe-ONp, Ac-Ala-Aala-ONp, and Z-Ala-Ala-Pro-Aala-ONp were >0.20 s" at pH 5.8 and 5.0, although the rates at pH 5.0 for Ac-Aphe-ONp and Ac-Ala-Aala-ONp appeared to be slightly slower than the others. Table 111 shows the results for the experiments in which the azapeptides were reacted with subtilisin Carlsburg. The two azaphenylalanine peptides were again found to have the lowest turnover rates while the aza-alanine peptides were turned over very rapidly. The azapeptides containing larger alkyl side chains again had intermediate turnover rates. The   reactions had a 1:1 stoichiometry and the enzyme was found to be -78% active. Although this enzyme exhibited the same general trends as subtilisin BPN', a comparison of the turnover rates for the reaction between a given azapeptide and each of the two subtilisins shows subtilisin Carlsburg to consistently have more rapid deacylation rates than Subtilisin BPN'. The variance in the ratio of enzyme concentration calculated from the burst to that based on A m for these two enzymes may be the result of the rapid autolysis which these two enzymes undergo. This would result in a lowering of the "active" enzyme concentration without lowering the protein concentration.
Reaction of Cathepsin G with Azapeptides- Table 1V shows the results of the reactions between cathepsin G (several different preparations) and each of seven azapeptides. Azapeptides which contained Aphe, Aleu, Anle, and Aval acylated the enzyme very rapidly and with 1:l stoichiometry with respect to the enzyme concentration. With Ac-Ala-Aphe-ONp, the reactions had very small turnover rates and kat remained constant at inhibitor concentrations of 0.06-0.2 mM. The peptides with long aliphatic aza-amino acid (e.g. Aval, Anva, Aleu, Anle) acylate the enzyme with no measurable turnover rates. Ac-Ala-Ala-Aala-ONp does not react with cathepsin G at pH 6-7. Three different batches of cathepsin G were found to be -62%, 89%, and 45% active with respect to protein concentration determined from A~w .
Previous studies in this laboratory have shown that Ac-Ala-Aphe-ONp does not react with trypsin at pH 5.8 (Powers and Carroll, 1975), while Ac-Aphe-ONp has been reported to acylate both chymotrypsin A, and trypsin at pH 7.04, but with an acylation rate much slower for trypsin (Elmore and Smyth, 1968).
Determination of Deacylution Rates-The rates of deacylation for the reaction of chymotrypsin with Ac-Ala-Aphe-ONp were measured by employing the substrate Ac-Tyr-OEt to monitor the increase in enzymatic activity as excess azapeptide was removed by hydrolysis and dilution. Although the reaction mixture was diluted by a factor of 10, the relative ratios of azapeptide to enzyme were essentially identical with those used in the previous experiments and comparison with the previous results can be made. Table V shows the results. All the deacylation rates were slower than the lower limit (1.8 X s-') observed in the spectrophotometric assay.
The kinetics of inhibition was investigated over a range of inhibitor concentrations in order to treat the data according to Kitz and Wilson (1970). None of the inhibitors showed any variation in kobs over the concentration range investigated. This indicates that KI was much greater than the inhibitor concentrations utilized and that kob,/[II is probably equal to kB/KI under the experimental conditions utilized.

DISCUSSION
The goals of our research with azapeptides were 3-fold. First, we wished to prepare suitable stable acyl-enzyme derivatives of serine proteases for possible future crystallographic investigations. We were interested in developing new avenues of inhibiting serine proteases and finally we were interested in developing a new series of active site titrants for various serine proteases. It is evident that azapeptides are suitable for all three uses.
Several azapeptides containing a PI aza-amino acid residue have previously been shown to react stoichiometrically with serine proteases (Kurtz and Niemann, 1961, a and b;Elmore and Smyth, 1968;Powers and Carroll, 1975;. Although there is no direct evidence, it is almost certain that the azapeptides are acylating the active site serine of serine proteases to form a stable acyl-enzyme derivative. The acyl-enzyme mechanism for the serine protease-catalyzed hydrolysis of amides and esters has been demonstrated by many lines of evidence (Bender and Killheffer, 1973) including the crystallographic investigation of several stable acyl derivatives of serine proteases.
With a few exceptions, all of the azapeptide p-nitrophenyl esters containing a PI aza-amino acid residue investigated in this study acylated chymotrypsin, subtilisins BPN' and Carlsberg, and cathepsin G rapidly with the stoichiometric release of p-nitrophenol and the formation of a fairly stable acylenzyme derivative. Only limited specificity was observed in the acylation process itself. All four proteases have substrate specificity for amino acid residues with aromatic or large alkyl side chains and most of the azapeptides studied also had this feature a t PI. Thus, the only azapeptides which did not react with chymotrypsin were two Agly and Aala peptides and, even in this case, extending the peptide structure from Ac-Ala-Aala-ONp to Ac-Ala-Ala-Aala-ONp and Ac-Ala-Ala-Pro-Aala-ONp resulted in acylation of chymotrypsin especially a t higher pH values. At pH 5.0, the rate of acylation of chymotrypsin by the tetrapeptide Aala derivative was slower than those of the shorter Aphe derivatives. This is consistent with the preference of chymotrypsin for substrates with PI Phe residues over those with PI Ala residues. Cathepsin G also did not react with the Aala peptide Ac-Ala-Ala-Aala-ONp. However, the two subtilisins, which have broader substrate specificity than chymotrypsin, were acylated by all the azapeptides with the exception of Ac-Ala-Ala-Agly-ONp.
Once formed, the acyl derivatives of chymotrypsin and most of those formed with cathepsin G are extremely stable. Even at pH 7, Ac-Ala-Aphe-chymotrypsin has a half-life of over 5 days. This is 5-6 orders of magnitude more stable than the corresponding acyl derivative formed from substrates. For example, Ac-Aphe-chymotrypsin has a deacylation rate of 1.2 X loe4 s" at pH 7.0 and 37 "C (Barker et al., 1974) compared to 72 s-' for Ac-Phe-chymotrypsin at pH 7.0 and 25 "C (Zerner et al., 1964). The deacylation rate for Ac-Ala-Aphe-chymotrypsin at pH 7.0 and 25 "C which we found to be 0.9 X s-' is quite comparable to the value of 1.2 x observed with Ac-Aphe-chymotrypsin. The deacylation rate is pHdependent with the deacylation rate being 15-fold faster a t pH 7 compared to pH 5. The stability of the acyl derivatives formed from azapeptides and their close resemblance to true acyl derivatives makes azapeptide derivatives quite suitable for crystallographic investigations of acyl-enzyme intermediates of serine proteases.
The stability of acyl derivatives formed upon reaction of azapeptides with PI aza-amino acid residues with serine proteases can be attributed to two factors. The first is electronic. The acyl derivative formed upon acylation of the active site serine residue with an azapeptide is a carbazic acid ester or carbazate (RCO-NHN(R)-CO-0-serine) compared to the ester which is formed from normal amide or ester substrates (RCO-NHCH(R)-CO-0-serine). In the carbazate, the carbazate carbonyl carbon is much less electropositive than the ester due to the resonance effect of the adjacent nitrogen atom. Thus, it would be much less susceptible to nucleophilic attack by water which is required for deacylation to take place. Carbamates, which also share this same structural feature with carbazates, are well known to form stable acyl derivatives with serine proteases and indeed the crystal structure of carbamyl chymotrypsin (NH2-CO-0-serine) has been determined by x-ray crystallography (Robillard et al., 1972).
The second reason for the stability of the acyl derivatives formed from azapeptides is steric. When one examines the structures of the two acyl-enzyme derivatives of chymotrypsin whose structures have been examined crystallographically, it is evident that there are significant differences with the hypothetical structure of a true acyl-enzyme intermediate. In both carbamyl chymotrypsin and in indoleacryloyl chymotrypsin (Henderson, 1970), the plane of the acyl-enzyme carbonyl group is twisted away from that which would allow optimum attack by the water molecule which must be deprotonated by histidine-57 (Robillard et al., 1972). In the case of indoleacryloyl chymotrypsin, binding of the rigid indoleacryloyl moiety in the SI pocket of the enzyme forces the carbonyl out of proper alignment. In contrast, the carbamyl derivative has nothing forcing it to occupy any particular conformation and it simply adopts a conformation which allows it to participate in a favorable hydrogen bonding network.
The change in the PI residue of a peptidyl acyl-enzyme from an amino acid residue to an aza-amino acid residue would have considerable influence on the geometry of the carbonyl group of the acyl-enzyme. The a-carbon of an amino acid residue is tetrahedral, while the &-nitrogen atom of an aza-amino acid residue is probably trigonal and the N-CO-0 moiety of the carbazyl enzyme would prefer to be planar. If the side chain of the PI aza-amino acid is locked into the SI pocket of the enzyme, then the carbonyl group would undoubtly be twisted relative to the conformation which would be optimum for deacylation.
The SI pocket of chymotrypsin is a long narrow groove on the surface of the enzyme and it evidently rigidly locks the side chain of all'the azapeptides since none showed significant deacylstion rates. In contrast, subtilisin BPN' has a much larger SI pocket which resembles a depression rather than a pocket (Robertus et al., 1972) and this is reflected in the much less stringent substrate specificity of subtilisin when compared to chymotrypsin. The azapeptide deacylation rates observed with the subtilisins fit this model. With subtilisin BPN', the azapeptides with the PI Aphe have the slowest deacylation rates, the three with a PI Aala residue deacylated fastest, and the derivatives with other alkyl side chains have intermediate rates. The results with subtilisin Carlsberg were similar except that overall the deacylation rates were much higher for all the derivatives except the Aphe derivatives. It appears that the azapeptides with the larger side chains interact with the S1 pocket to twist the acyl carbonyl group into a conformation or conformations which are unsuitable for deacylation. As the side chain decreases in size, it can slide around more freely in the SI pocket and it becomes more likely that a proper deacylation conformation can be formed. Thus, deacylation rates increase as the size of the side chain decreases. The behavior of cathepsin G resembles that of chymotrypsin since all the acyl derivatives except one were stable.
Active Site Titration-Azapeptide p-nitrophenyl esters should be quite useful as active site titrants for serine proteases. Elmore and Smyth (1968) first showed that Ac-Aphe-ONp was a useful active site titrant for chymotrypsin and in this and the following paper (Powers et al., 1984) we have considerably expanded the number of reagents which are available for use with serine proteases of varing specificity.
There are several conditions which must be met for an azapeptide p-nitrophenyl ester to be a suitable active site titrant. These are: (a) The first is easily met by choice of reaction conditions. But the last two must be checked with each new enzyme or set of titration conditions. The second condition, k2 > k3, is easily met with most of the enzymes and azapeptides which we studied. In most cases, the acylation rates k, were greater than 0.2 s" and the deacylation rates k3 were slower than 1.8 X s-'. However, with subtilisin at pH 7.0, some of the acyl derivatives deacylated very rapidly and it is likely that this condition is not met. In these instances, it would be necessary simply to carry out the titrations at lower pH values to avoid this problem. The final condition for an accurate titration requires determination of KM or study of the concentration dependence of the titration. With chymotrypsin and the subtilisins, we did not carry out such studies. Instead, we simply checked our titration results using the more widely studied titrant 2-hydroxy-5-nitro-a-toluenesulfonic acid sultone (Kezdy and Kaiser, 1970) to demonstrate that the titration conditions were giving valid results. However, in the case of cathepsin G, a n enzyme for which no other titration procedure has been reported, we carried out concentration dependence studies at pH 6.0 with some of the azapeptides. As can be seen from the data in Table  IV, titration at low concentrations of some of the azapeptides resulted in enzyme concentrations which were too low. However, as the concentration of azapeptide was increased, the kc,, (e.g. Ac-Ala-Aphe-ONp) or the burst reached a plateau indicating that the [SI > K , condition was being met.
The most useful titrant for any particular titration would be dependent on the needs of the individual investigation. For titrations of cathepsin G in our laboratory, we frequently use Ac-Ala-Ala-Anle-ONp a t concentrations of -0.20 mM at pH 6.0 or 7.0. We average three or more replicates to obtain the final enzyme concentration. Many of the other azapeptide derivatives are equally suitable, but this titrant is easily synthesized and is also suitable for titration of human leukocyte elastase (Powers et al., 1984). If specificity was an important consideration, then it would be better to use Ac-Aphe-ONp or Ac-Ala-Aphe-ONp. Ac-Aphe-ONp will react with both chymotrypsin and trypsin at pH 7.04 (Elmore and Smyth, 1968), while we have found that Ac-Ala-Aphe-ONp reacts with chymotrypsin A, and Ax, subtilisins BPN' and Carlsberg, and cathepsin G but does not react with trypsin, porcine pancreatic elastase, and human leukocyte elastase (Powers and Carroll, 1975;Powers et al., 1984). These reagents would be useful if it were necessary to determine the concentration of one enzyme in the presence of small amounts of other contaminating serine proteases. Similarly, Ac-Ala-Aala-ONp could be used to titrate subtilisins in the presence of chymotrypsin or cathepsin G, although it is not likely that this would be necessary in many instances.
Inhibitors-Finally, we examined azapeptides as potential inhibitors of serine proteases. Azapeptides which acylate serine proteases rapidly to form stable acyl-enzymes are in essence good inhibitors. With chymotrypsin and cathepsin G, these criteria are met with almost all the azapeptide p-nitrophenyl esters. In contrast, none of the azapeptides can be considered to be suitable inhibitors for the two subtilisins. The most stable acyl derivatives formed from subtilisin had half-lives of less than 6 min at pH 7.0.
The nature of the leaving group (Pl') has a considerable influence on the rate at which a peptide will react with a serine protease. For example, in the series of 4-nitroanilide substrates Ac-Ala-Ala-Pro-AA-NA, Zimmerman and Ashe (1977) showed that cathepsin G would only hydrolyze the AA = Phe and Leu derivatives, while Ala, Val, and Ile were untouched. In contrast, with a set of more reactive thiobenzyl ester substrates Boc-Ala-Ala-AA-SBzl, we have observed that the AA = Phe, Nle, Leu, and Nva derivatives were hydrolyzed by cathepsin G. The azapeptide p-nitrophenyl esters are in the same high reactivity category and show very little specificity in their reaction with any particular serine protease. One way of increasing the specificity would.be to change the nature of the leaving group in the azapeptide structure. The use of p-nitrophenol allows the reaction to be followed quite readily, but is not an optimum choice if specificity is the major goal.
The results in Table  VI show that cathepsin G can be inhibited by azapeptides with a variety of leaving groups. In the series of compounds which we examined, the order of reactivity is -0Np > -0Ph > -OCH2CF3 > -0Et. It is clear that the ethyl ester is just at the borderline in terms of reactivity. In a favorable case, Ac-Ala-Ala-Anle-OEt, acylation of the enzyme and irreversible inhibition occurs, although quite slowly. In a less favorable situation (Ac-Ala-Ala-Anva-OEt), no irreversible inhibition occurs. A similar situation has been observed with chymotrypsin. Ac-Aphe-ONp has been shown to acylate chymotrypsin stoichiometrically (Elmore and Smyth, 1968). The corresponding ethyl ester Ac-Aphe-OEt was initially believed to a competitive inhibitor with KI = 20 mM at pH 7.9 and 25 "C (Kurtz and Niemann, 1961, a and b). However, Barker et al. (1974) later showed that chymotrypsin A, slowly lost its activity when incubated with a large excess of Ac-Aphe-OEt at pH 7.0 and 37 "C. If one wishes to use azapeptides as inhibitors, it is likely that trifluoroethyl esters, phenyl esters, or something similar would be the most suitable. These derivatives would be more reactive than ethyl esters, but would probably exhibit more specificity than the p-nitrophenyl esters.
In conclusion, we have reported a new series of azapeptides which are useful as active site titrants and inhibitors of serine proteases with chymotrypsin-like specificity. In the following paper we show that these compounds can also be utilized with elastase (Powers et ai., 1984).