Reaction of Azapeptides with Human Leukocyte Elastase and Porcine Pancreatic Elastase NEW INHIBITORS AND ACTIVE SITE TITRANTS*

The reaction of a series of azapeptides with porcine pancreatic (PP) elastase and human leukocyte (HL) elastase has been studied and a series of new inhibitors and active site titrants were found for both PP elastase and HL elastase. Azapeptide p-nitrophenyl esters acylate both HL and PP elastase to form stable acyl- enzymes, which can be used for crystallographic studies. We have investigated the effect of a PI, P,, or P3 aza-amino acid residue on the reactivity of azapeptides with elastase. We have also studied the effect of chang- ing the nature of the PI‘ leaving group and other portions of azapeptide structure. N-Acetyl-L-alanyl-L- alanyl-a-azaalanine p-nitrophenyl ester, N-acetyl-L-alanyl-L-alanyl-a-azanorleucine p-nitrophenyl ester, and N-acetyl-L-alanyl-L-alanyl-a-azanorvaline p-ni-trophenyl ester are suitable titrants for either PP or HL elastase.

Since elastase inhibitors have considerable potential for use in therapy, synthetic inhibitors have been developed by ourselves and a number of other research groups (see Powers (1983) for a recent review). Some of these inhibitors have been shown to be effective in animal models of emphysema.
In the preceding paper (Gupton et ai., 1984), we have shown that azapeptides' are effective inhibitors of serine proteases, can be used as active site titrants, and can be used to generate stable acyl derivatives for crystallographic investigations. In * This investigation was supported by Grant HL 29307 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 "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
* 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-aza-alanine will be abbreviated Aala. Any peptide which contains an aza-amino acid residue will be referred to as an azapeptide in this paper. this paper, we have extended the use of azapeptides to both HL and PP elastase and report a series of new inhibitors and active site titrants for both elastases. In addition, we have investigated the effect of a PI, Pz, or P3 aza-amino acid residue3 on the reactivity of azapeptides with elastase. Finally, we have investigated the effect of changes in the nature of the PI' leaving group and in other portions of the azapeptide structure. These investigations extend our knowledge of the active site of elastase and the types of molecules which will bind and react with it.

MATERIALS AND METHODS
Porcine pancreatic elastase was obtained from Whatman or Worthington and used without further purification; its substrate, Boc-Ala-ONp, was purchased from Sigma. Human neutrophil leukocyte elastase was generously provided by Drs. R. Baugh and J. Travis of the University of Georgia and by Drs. A. Janoff and G. Feinstein of S.U.N.Y. at Stony Brook.
Synthetic procedures and the synthesis of many of the azapeptides utilized in this paper are reported in the supplementary section to the preceding paper (Gupton et al., 1984). The remainder are reported in a supplementary section to this paper.4 Reaction of Elastase with Azapeptides (Titration Procedure)-The reaction of porcine pancreatic and human leukocyte elastase with azapeptides was carried out in a solution which contained approximately a 50-fold excess of azapeptides over enzyme. Stock solutions of azapeptide in acetonitrile were prepared at a concentration of 0.5 mM. PP elastase stock solutions were made up in 1 mM HCI and had a concentration of -10-140 PM. HI. elastase was usually obtained as solutions in a 20-50 PM, pH 5.0-6.0 phosphate or acetate buffer containing -0.3-0.4 M NaCl. Exact enzyme concentrations were determined by absorbance at 280 nm (E;% = 20.2 for porcine pancreatic elastase (Shotton, 1970); E$% = 9.85 for human leukocyte elastase (Babul and Stellwagen, 1969)). Four buffers were prepared: pH 7.0, 0.1 M phosphate; pH 6.0, 0.1 M citrate; pH 5.0, 0.1 M citrate; pH 4.0, 0.1 M acetate. All reactions were performed at 25 "C. The reaction of elastase with azapeptides was carried by adding 50 PI of azapeptide stock solution 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 measured. The reaction was initiated by the addition of 25 pl of 1 mM HCI or water to the reference and 25 p1 of enzyme stock solution to the sample cuvette.
The burst of p-nitrophenol was used to obtain the concentration of active enzyme (345 nm, c = 6250 at pH 6.0). Assays were performed on a Beckman Model 25 or Model 35 spectrometer. The values of kcat reported are the averages of at least three separate determinations.
The titrations of PP and HL elastase were carried out at concentrations of 6.5-0.33 pM (92-5 pg in 550 pl) and 2.2-0.8 p M (31-11 pg in 550 PI), respectively. A concentration of 0.33 p M gives a hA of 0.002 which is not very reproducible with a noisy or low sensitivity spectrophotometer.
Kinetic Parameters for the Elastase-catalyzed Hydrolysis ofrlzapeptides-With those azapeptides that were substrates, enzymatic rate of hydrolysis was measured by adding 100 pl of the enzyme stock solution to a solution containing 2 ml of 0.1 M citrate buffer (pH 6.0) and 0.1 ml of azapeptide in acetonitrile. The increase in the absorbance at 345 nm was followed (eaG = 6250 at pH 6.0). The kinetic constants were determined from the initial rate of hydrolysis by the Lineweaver-Burk method and Cleland's iterative fitting method (Cleland, 1967;1979), and are based on duplicate rate determinations at six substrate concentrations. Correlation coefficients were greater than 0.98. The noncatalyzed hydrolysis rates of the azapeptides were measured similarly except that no enzyme was added.
Inhibition of Human Leukocyte Elastase with Tripeptide Azapeptides and Related Compounds-Inhibition of human leukocyte elastase by the azapeptide was carried out in solution which contained at least IO-fold excess of inhibitor over enzyme. Stock solutions of azapeptide in acetonitrile were prepared at a concentration of 0.07-14 mM. Enzyme stock solutions had a concentration of -7 pM. The reactions were carried out by mixing 50 p1 of inhibitor solution and 50 p1 of enzyme solution with 1 ml of buffer; 100-pl aliquots were removed from the reaction mixture at regular time intervals and the residual enzymatic activity was measured using the Boc-Ala-ONp spectrophotometric assay (Visser and Blout, 1972). The final concentration of inhibitors and enzyme in the reaction mixture is shown in Table V.

RESULTS
The kinetics of the reaction of serine proteases with azapeptides is described in the preceding paper (Gupton et al., 1984). The background hydrolysis rates at pH 6.0 of representative azapeptide p-nitrophenyl esters are shown in a table with the supplementary material. The hydrolysis rates for all of the azapeptides were quite similar except for two derivatives in which the Pl NH was substituted with a methyl group or replaced by a CH, group. These two derivatives had much slower rates than the others.
Reaction of Porcine Pancreatic Elastase with Azapeptides- Table I lists the results obtained upon reaction of 10 azapeptides with PP elastase. All of the azapeptides except for Ac-Ala-Ala-Agly-ONp and Ac-Ala-Ala-MeAala-ONp acylated the enzyme very rapidly and no attempt was made to measure the acylation rates. In addition to the compounds listed in the table, we checked Ac-Aphe-ONp, Ac-Ala-Aphe-ONp, and Ac-Ala-Aala-ONp at pH 5.0, 5.8, and 6.5 and observed no acylation of the enzyme. In most cases where an azapeptide acylated elastase, the acylation rate was greater than 0.20 s-', our limit of detection. One exception was Ac-Ala-Ala-Aleu-ONp which acylated elastase more slowly than the other azapeptide p-nitrophenyl esters. For Ac-Ala-Ala-Aala-ONp at pH 6.0, the concentration of active elastase obtained and the kcat were independent of azapeptide concentrations in the range of 0.01-0.10 mM which indicates that [SI > K M and that valid titrations were being obtained. Since the active site concentrations obtained with the other azapeptides were in agreement with those obtained with Ac-Ala-Ala-Aala-ONp, we believe that all the azapeptides are acylating elastase with 1:I stoichiometry based on the release of p-nitrophenol. The turnover rates varied with the nature of the P, aza-amino acid residue. Ac-Ala-Ala-Aval-ONp had the highest kcat value in this series, while Ac-Ala-Ala-Anle-ONp had the slowest. The turnover rate was pH-dependent and increased by a factor of 5 upon going from pH 6 to 7. The tetra-azapeptides Ac-Ala-Ala-Pro-Aala-ONp and Ac-Ala-Ala-Ala-Aala-ONp had higher kcat values than the corresponding triazapeptide Ac-Ala-Ala-Aala-ONp. We found the commercial elastase averaged 81% activity with respect to the AzW concentration.
Reaction of Human Leukocyte Elastase with Azapeptdes- Table I1 shows the results obtained upon reaction of a series of azapeptides with HL elastase. The azapeptides acylated the enzyme rapidly and stoichiometrically and a representative experiment with Ac-Ala-Ala-Aala-ONp is shown in Fig. 1. Two azapeptides, Ac-Ala-Ala-Agly-ONp and Ac-Ala-Ala-MeAala-ONp, did not react with the enzyme. In addition to the compounds listed in the table, we checked Ac-Aphe-ONp, Ac-Ala-Aphe-ONp, and Ac-Ala-Aala-ONp at pH 6.5 and observed no acylation of the enzyme. The acylation rates for the azapeptides with a P1 Aile, Aleu, or Anva were slightly slower than our detection limit of 0.20 s-', but we made no attempt to measure acylation rates. Concentration dependence studies with Ac-Ala-Ala-Aala-ONp demonstrated that both the active enzyme concentration and kcat were constant at an inhibitor concentration of 0.09-0.1 mM. The turnover rate kcat was dependent on the nature of the PI aza-amino acid residue and had the order Aval > Aile > Aala > Anva > Aleu,Anle. The azapeptides with straight chains at PI seemed to form more stable acyl-enzymes with lower turnover rates. The two tetraazapeptides, Ac-Ala-Ala-Pro-Aala-ONp and 2-Ala-Ala-Pro-Aala-ONp, had slightly slower turnover rates than the triazapeptide Ac-Ala-Ala-Ala-ONp. The average concentration of active HL elastase was 90% based on As with PP elastase, the acylation rates were faster than our detection limit. However, Dr. Antonio Baici, Kantonsspital Zurich, Switzerland, has measured the acylation rate using a stopped flow apparatus. With HL elastase and Ac-Ala-Ala-Anva-ONp at pH 6.0, he observed a pseudo-first order rate constant of 0.51 s" (half-life = 1.35 s). Hydrolysis of Triazapeptide Analogs by PP Elastase and HL Elastase- Table I11 shows the kinetic results obtained when each Ala residue of Ac-Ala-Ala-Ala-ONp is substituted with an aza-alanine residue in turn. Ac-Ala-Ala-Ala-ONp is a good substrate for HL elastase and substitution of the a-CH of either the P1, Pz, or P3 residue with a nitrogen reduced the kcat/&. The profound effect of the P, Aala residue is reflected in both kc,, and KM.
With PP elastase substitution of an Aala for Ala has a similar effect on kcat/& values. Again, the PI Aala had the most effect on kc, and K M , although there was less change in k J K M when compared with HL elastase. We investigated the reactivity of Ac-Ala-Ala-Aala-NA with PP elastase and found it to be a very poor substrate. No burst was observed. If this compound had given a burst with elastase, we intended to synthesize the 7-amino-4-methylcoumarin derivative. This would have increased the sensitivity of elastase titrations since 7-amino-4-methylcoumarin can be detected by fluorescence. We also tried to synthesize the 4methylumbelliferone derivative of Ac-Ala-Ala-Aala-for the same purpose, but found the compound to be too unstable toward hydrolysis.
Inhibition of HL Elastase with Azapeptides- Table IV shows the results of inhibition of several azapeptide esters with HL elastase. Both Ac-Ala-Ala-Anva-ONp and Ac-Ala-Ala-Anle-ONp acylated HL elastase and gave acyl-enzymes which turned over very slowly. Therefore, we investigated the corresponding derivatives with different leaving groups as potential inhibitors. All inhibited elastase much more slowly than thep-nitrophenyl esters. The azapeptides with the better   HONp.

K,>> [ I ] and that k,h/[A is equal to k,/KI under the exper-
imental conditions used.

DISCUSSION
Proteolysis of lung elastin by HL elastase and other proteases released from the granule fraction of human polymorphonuclear leukocytes is generally thought to cause the tissue destruction observed in pulmonary emphysema. Elastases are also found in the pancreas and in a variety of other tissues and probably play an important role in connective tissue turnover. The interactions of elastases with their natural inhibitors (aI-protease inhibitor and a2-macroglobulin) and with their natural substrates (elastin and other connective tissue proteins) are areas of active current investigation (Travis and Salvensen, 1983). Many of these studies require specific substrates, specific inhibitors, and active site-titrated enzyme. We began these studies to explore the use of azapeptides as titrants and inhibitors of elastase and have found these compounds to be exceptionally useful for active site mapping and other studies of the active sites of elastase.
Azapeptides have been studied by a number of investigators as inhibitors and active site titrants of chymotrypsin and trypsin (Kurtz and Neimann, 1961;Elmore and Smyth, 1968).

X lo2
Ac-Ala-Ala-Aala-ONp* 6.0 0.008 0.005 6.1 X 10' 0.002 0.003 1.5 X 103 Ac-Ala-Ala-Aala-NA' * Same conditions as a except that CH3CN was substituted for MeOH. e 0.1 M Tris buffer, pH 8.0, 25 "C, 4.5% (v/v) dimethyl sulfoxide, 3.55 p~ PP elastase. The reaction was followed at 410 nM.  The azapeptide binds to the enzyme with the side chain (n-butyl) of the PI Anle residue interacting with the S1 or primary substrate binding pocket of elastase. The acyl group is shown interacting with other portions of the extended substrate binding site of elastase. L represents a leaving group such as p-nitrophenol or trifluoroethanol. The azapeptide acylates the active site serine forming a carbazyl enzyme (right).
Previously, we showed that the azapeptidep-nitrophenol ester Ac-Ala-Ala-Aala-ONp could be used as an active site titrant of PP elastase and that 2-Ala-Ala-Pro-Aala-ONp acylated both PP and HL elastase rapidly with the formation of a stable acyl-enzyme (Powers and Carroll, 1975;Powers and Gupton, 1977). In the preceding paper (Gupton et aL, 1984), we studied the scope of the reaction of azapeptides with chymotrypsin, subtilisins BNP' and Carlsberg, and cathepsin G. In this paper, we have extended those studies to include PP and HL elastase.
The reaction scheme for an azapeptide with elastase is shown in Fig. 2. The azapeptide binds to the active site of elastase and then acylates the enzyme to form the acyl (car-bazy1)-enzyme shown. At this stage, a certain degree of substrate recognition is required. Neither elastase was acylated by the Aphe-containing peptides Ac-Aphe-ONp and Ac-Ala-Aphe-ONp nor by the short Aala-containing peptide Ac-Ala-Aala-ONp. This shows that both elastases recognize the side chain of the PI aza-amino acid residue and that an extended substrate structure is required for binding to elastase. The requirement of elastase for an extended peptide chain for recognition and binding is well known and has previously been observed with both substrates and chloromethyl ketone inhibitors. For example, dipeptide chloromethyl ketones inhibit PP elastase much more slowly than tripeptide or tetrapeptide chloromethyl ketones (Powers et al., 1977).
The NH group of the PI aza-amino acid residue is essential for acylation of elastase to occur. If this is substituted with a methyl group (Ac-Ala-Ala-MeAala-ONp), no reaction with either elastase occurs. In addition, replacement of this NH with a CH, gives a very poor inhibitor of HL elastase (Table  IV). In the binding of substrates and inhibitors to serine proteases, the NH of the PI residue forms a hydrogen bond with the backbone peptide bond carbonyl of residue 214. In the case of PP elastase, this has been observed crystallographically in the binding of a trifluoroacetyl dipeptide anilide to elastase, although in that case the peptide is binding in a reverse direction compared to normal substrates and most other inhibitors (Hughes et ai., 1982).
Stable Acyl-enzymes-In the previous paper, we discussed the electronic and steric effects that occur when an aza-amino acid residue is substituted for an amino acid in the PI site of a substrate. This substitution has a profound effect on the deacylation rate. In the case of PP elastase, the kat at pH 6.0 decreases by 11,000 upon going from Ac-Ala-Ala-Ala-ONp to Ac-Ala-Ala-Aala-ONp. In the case of HL elastase, a 17,000fold decrease is observed. Thus, azapeptides with PI azaamino acid residue give stable acyl derivatives with both elastases.
Considerable substrate specificity was exhibited in the deacylation step. At pH 6.0 in the series of azapeptides Ac-Ala-Ala-Aaa-ONp, the deacylation rate (kcat) varied by 37-fold with PP elastase and over 1900-fold with HL elastase as the PI aza-amino acid residue was changed. The order of reactivity with PP elastase was Aval > Aile > Aala > Anva > Aleu > Anle, with Ac-Ala-Ala-Anle forming the most stable acylenzyme with PP elastase. The following preference of PP elastase for substrates has been observed polypeptides, Ile > Val > Ala > Leu (Powers et al., 1977); tetrapeptide 4-nitroanilide substrates, Ala > Leu > Val, Ile no hydrolysis (Zimmerman and ; kcat values for Boc-Ala-Ala-AA-SBzl hydrolysis, Leu > Nva > Nle > Ala > Val > Ile. The order of reactivity with H1 elastase was Aval > Aile > Aala > Anva > Aleu = Anle. The following preference of HL elastase for substrates has been observed tetrapeptide 4-nitroanilide substrates, Val > Ile > Ala > Leu ; %at values for Boc Ala-Ala-AA-SBzl hydrolysis, Ala > Leu > Nva = Nle > Val > Ile.5 When we examine these preferences carefully, we detect an inversion in the general order of R. R. Cook, J. W. Harper, and J. C. Powers, unpublished results. by guest on March 24, 2020 http://www.jbc.org/ Downloaded from reactivity. When a particular P1 amino acid residue yields a more reactive substrate, the corresponding aza-amino acid residue in a peptide has a slower deacylation rate (kc,, ) and forms a more stable acyl-enzyme and vice versa.
Electronic effects must make a relatively minor contribution to the deacylation rates since the hydrolysis rates at pH 6.0 for the two peptides Ac-Ala-Ala-Ala-ONp and Ac-Ala-Ala-Aala-ONp are very similar. These hydrolysis rates, to a first approximation, can be considered to be measures of the relative susceptibility of an aza-amino acid and an amino acid residue to nucleophilic attack. Since electronic effects have little effect on the deacylation rates, we believe that the major effect is steric. As we discussed in the preceding paper, substitution of a N atom for the a-CH in the P, amino acid residue will result in a change in the geometry of the azapeptide carbonyl group relative to the ideal geometry for the deacylation reaction. This would especially be the case when a well defined interaction took place between the side chain of the Pl aza-amino acid residue and the Sl pocket of the serine protease. We would expect the PI amino acid residues in the best substrates to have the strongest interaction with the SI pocket of elastase. This would twist the carbonyl group of the corresponding azapeptide out of a good alignment for deacylation and the most stable acyl derivatives would thus be formed. This in general was what we observed.
The deacylation rates observed with both elastases were considerably faster than those observed with chymotrypsin and cathepsin G. We believe this is caused by a strong PI-S1 interaction in chymotrypsin and cathepsin G which prevents the carbonyl group of the PI aza-amino acid from occupying the conformation necessary for deacylation.
Since the S , binding pocket in elastase is not as deep and the side chains of good elastase substrates are not as long as they are in the case with chymotrypsin and cathepsin G and their substrates, the SI-P1 interaction is less effective at preventing the PI azaamino acid from occasionally occupying a suitable conformation for deacylation. Thus, the deacylation rates are higher with elastase. When comparing the two elastases, HL elastase forms more stable acyl-enzymes than does PP elastase. Again this is consistent with the fact that HL elastase has a larger SI binding pocket as evidenced by its acceptance of substrates with longer PI alkyl chains than PP elastase. In comparison to the subtilisin, the elastase deacylations were in the same range. However, the most stable derivatives at pH 6.0 formed from the elastases were more stable than the most stable subtilisin acyl-enzymes.
Recently, a series of heterocyclic derivatives such as N-acyl saccharins and N-arylbenzisothiazolinone 1,l-dioxides have been reported to he acylating inhibitors of HL and PP elastase, cathepsin G, and chymotrypsin (Zimmerman et al., 1980;Ashe et al., 1981). These heterocyclic compounds form acylenzymes which have considerable variability in stability. Seemingly small changes in structure can have considerable influence on the deacylation rates. We believe this is another example where the directionality and strength of the P1-S1 interaction impart a twist to the acyl carbonyl and determine the ease with which the various derivatives deacylate.
Active Site Titration-Azapeptide p-nitrophenyl esters are extremely useful for active site titration of both HL and PP elastase. Standardized elastase solutions are essential for measuring the stoichiometry of interaction of elastase with its natural protein inhibitors such as al-protease inhibitor, for studying the interaction of elastase with its natural substrate elastin and other connective tissue proteins, and for calculating accurate kcat values for synthetic peptide substrates.
Prior to the development of azapeptide p-nitrophenyl esters, the only reported titrants for elastase were nonspecific inhibitors such as diethyl p-nitrophenyl phosphate (Bender et al., 1966) and radiolabeled diisopropyl phosphofluoridate. These compounds are characterized by low specificity and react with numerous other serine proteases. The nitrophenyl ester acylates elastase slowly and must be used at high pH values where the background hydrolysis of the titrant makes accurate determination of a burst difficult. In contrast, many of the azapeptide p-nitrophenyl esters which we have studied acylate elastase within the time interval required for mixing the solution and placing the cell into the spectrometer (>go% complete in 9 s, k2 > 0.2 s-') and can be used at pH values from 4-7 where turnover of the titrant does not interfere with the titration results.
Several of the azapeptide p-nitrophenyl esters make suitable titrants for either PP or HL elastase. In our laboratory, we have frequently used Ac-Ala-Ala-Aala-ONp, Ac-Ala-Ala-Anle-ONp, and Ac-Ala-Ala-Anva-ONp for such titrations. Ac-Ala-Ala-Aala-ONp does not react with cathepsin G at pH 6.0, nor with chymotrypsin at pH 4.0 and 5.0, although it does react at pH 6.0 and 7.0. If specificity is important or if the elastase contains traces of contaminating serine proteases then this would be the titrant of choice. However, we have tended to use Ac-Ala-Ala-Anle-ONp or Ac-Ala-Ala-Anva-ONp more frequently. These two azapeptides are in the group with the lowest deacylation rates with both HL and PP elastase, are easier to synthesize in large amounts than the Aala derivative, and can also be used to titrate cathepsin G.
Most titrations are carried out at pH 6.0 to minimize background hydrolysis, although this effect is cancelled by having titrant in both the sample and the reference cells and titrations can be carried out over a much wider pH range without difficulty. To minimize the use of precious enzymes such as HL elastase, we recommend the use of microcells. All three azapeptides have now been used in the laboratories of several other independent investigators with satisfactory results. Azapeptides As Inhibitors-Those azapeptides which form stable acyl-enzymes can be considered to be good inhibitors of elastase. In the case of PP elastase, none of the azapeptide p-nitrophenyl esters formed acyl-enzymes which had any significant life at pH 6.0 or 7.0, although Ac-Ala-Ala-Anle-ONp formed the most stable derivative. In the case of HL elastase, this same compound formed an acyl-enzyme which appears to be stable at pH 6.0 and 7.0. In addition, Ac-Ala-Ala-Aleu-ONp formed a stable derivative with HL elastase at pH 6.0. One disadvantage with p-nitrophenyl esters is their instability to hydrolysis and the possibility that they are too reactive as acylating agents. Therefore, we tested azapeptides with less reactive leaving groups (Table IV). Azapeptide phenyl or trifluoroethyl esters of Ac-Ala-Ala-Anva-or Ac-Ala-Ala-Anle-were inhibitors of HL elastase, although the trifluoroethyl esters were quite poor. Ethyl esters did not inhibit irreversibly at all. They were not tested as reversible inhibitors and it is likely that they would bind competitively to HL elastase (Dorn et al., 1977). The prospects for using azapeptides in uiuo for the treatment of emphysema are probably only fair due to their hydrolytic instability and the possibility of acylating other nucleophilic groups in proteins.
Conclusion-Azapeptides have been shown to be extremely useful reagents for the study of elastase and other serine proteases. Azapeptide p-nitrophenyl esters acylate both HL and PP elastase to form stable acyl-enzymes. Some of these derivatives appear to have sufficient stability for crystallographic studies, especially at low temperature. Although relatively little has appeared to date, there are currently active investigations on the structures of small molecule binding to PP elastase and a number of attempts are being made to obtain crystals of H L elastase. Azapeptide derivatives would be worth considering for such studies.
Substitution of an aza-amino acid residue into the structure of a peptide would be an excellent way of stabilizing the peptide toward protease digestion in vivo (Dutta and Giles, 1976) with the possibility of only slight changes in its interaction with its enzyme or receptor. For example, the k,,,/K, for the PP elastase-catalyzed hydrolysis of Ac-Ala-Ala-Aala-NA (1.7 M-' s-') is a t least 700-1000-fold lower than that for Ac-Ala-Ala-Ala-NA (kCat/KM = 745-2160 M-' s-l at pH 8.0 depending on the organic solvent and buffer used (Bieth and Wermuth, 1973)), while there is almost no change in KM (6.6 mM (Kasafirek et al., 1974)), the major effect being in kcat. The results in Table I11 also show that substitution of an azaamino acid residue in the Pz or P3 site of a substrate substantially reduces kcat/KM. Of course, the effects are even more pronounced at P1, but it appears that substitution in regions of peptides adjacent to cleaved bonds can be an effective way of reducing the proteolysis of peptides in vivo.
Finally, azapeptides can be used as active site titrants and inhibitors for elastase and other serine proteases.