The Interaction of Inhibitors of Proteolytic Enzymes with 3-Methylhistidine57=chymotrypsin*

SUMMARY The inhibition of proteolytic enzymes by protein inhibitors is accompanied by the formation of unusually stable complexes. The recognition of a specific substrate-like amino acid on the inhibitor is believed to be the initial event of the inhibitory process. In addition to the interactions involved in the binding of a good substrate, a variety of other noncovalent interactions are known to stabilize the complex. The formation of stable complexes between several inactive derivatives of proteolytic enzymes and a variety of protein inhibitors suggests strongly that the formation of any species resembling catalytic intermediates is unnecessary for inhibition. We have examined the interaction between several avian ovomucoids and cu-chymotrypsin in which histidine-57 has been converted to 3-methylhistidine-57. This derivative is easily prepared and can be isolated by affinity chromatography. Methylchymotrypsin retains unaltered its ability to bind specific substrates, but is essentially inactive. In spite of this loss of enzymatic activity, methylchymotrypsin forms strong complexes with several inhibitors. In addition, methylchymotrypsin which has been covalently linked to Sepharose is particularly useful for the isolation of protein inhibitors without the complications due to isolation of a mixture of partially

From the Department oj" Food Science and Technology, University of California at Davis, Davis, California 95616 SUMMARY The inhibition of proteolytic enzymes by protein inhibitors is accompanied by the formation of unusually stable complexes.
The recognition of a specific substrate-like amino acid on the inhibitor is believed to be the initial event of the inhibitory process.
In addition to the interactions involved in the binding of a good substrate, a variety of other noncovalent interactions are known to stabilize the complex. The formation of stable complexes between several inactive derivatives of proteolytic enzymes and a variety of protein inhibitors suggests strongly that the formation of any species resembling catalytic intermediates is unnecessary for inhibition.
We have examined the interaction between several avian ovomucoids and cu-chymotrypsin in which histidine-57 has been converted to 3-methylhistidine-57.
This derivative is easily prepared and can be isolated by affinity chromatography.
Methylchymotrypsin retains unaltered its ability to bind specific substrates, but is essentially inactive. In spite of this loss of enzymatic activity, methylchymotrypsin forms strong complexes with several inhibitors.
In addition, methylchymotrypsin which has been covalently linked to Sepharose is particularly useful for the isolation of protein inhibitors without the complications due to isolation of a mixture of partially cleaved forms of the inhibitor.
Because the inhibition of serine proteinases by protein inhibitors plays an important role in the control of biological processes involving proteolysis, it has been one of the most thoroughly studied protein-protein interactions. Dissociation constants for these complexes arc as low as lo-r3 (I) and values less than lop8 are common.
This remarkable stability has led several workers to suggest that a covalent bond between the two proteins, possibly a catalytic intermediate, is necessary for inhibition of the enzyme (2, 3 tered its ability to bind specific substrates and substrate analogs such as profiavin (7-9), N-acetyl-L-tyrosine ethyl ester (7), indole (10,11)) and N-acetyl-L-tyrosine p-methoxyanilide (8). In addition, methylchymotrypsin crystallizes isomorphously with the native enzyme and has been shown to exhibit only small changes in crystalline structure (11). Despite these similarities, however, methylchymotrypsin has only a small residual catalytic activity with rate constants reduced by 5 x lo3 to 2 X lo5 for specific substrates (7). The interaction of this derivative with protein inhibitors of the native enzyme helps to clarify the role of the catalytic ability of proteolytic enzymes in the formation of complexes with their protein inhibitors.

Synthesis
and Isolation of Xethylchymotrypsin-Treatment of cY-chymotrypsin with methyl p-nitrobenzene sulfonate under very mild conditions results in the rapid conversion of the enzyme to the 3-methylhistidine-57 derivative (6). To avoid the possibility of methylation at sites other than histidine-57, phenylmethanesulfonyl fluoride was added after only approximately 5Ooi, of the enzyme had been methylated.
Phenylmethanesulfonyl fluoride reacts rapidly with the remaining active enzyme to give the inactive phenylmethanesulfonyl derivative (18). This eliminates any complications from autodigestion of the enzyme during subsequent steps. Fig. 1  denatured protein, and any excess reactants by passing the mixture through a turkey ovomucoid-Sepharose affinity column. AS shown in Fig. ID, anhydrochgmotrypsin is also retained by the turkey ovomucoid column.
Since this affinity chromatography step is considerably more convenient than separation of the mixture by ion exchange chromatography (7, ll), mcthglchymotrypsin was prepared fresh daily. Methylchymotrypsin isolated in this fashion had less than 0.57, residual IWIG< activity. The formation and dissociation of the complex during isolation did not appear to reactivate the methylchymot,rypsin. Active chymotrypsin was also purified 011 this column immcdiately prior to its use. The presence of a small positive staining impurity with a higher mobility than ovomucoid at pH 8.6 was probably due to some trace ion eschange capacity of the coupled enzyme.

Interaction of Turkey
Ovomucoid with Native Chymotrypsin- The association constant for the complex between turkey ovomucoid and a-chymotrypsin was calculated to be 6.0 X lo8 hl-' from the departure from stoichiometric inhibition at equimolar concentrations of enzyme and inhibitor. The kinetics of association of cu-chymotrypsin with turkey ovomucoid was followed by the loss in IWEE activity after mixing the two proteins (Fig.  3). The association rate was 2.2 X lo4 h1-l s-l. An :irrhenius plot for the temperature dependence of the association yate constant (Fig. 4) showed the activation energy for the association to be 1 I .

Interaction of Methylchymotrypsin with Avian Ovomucoids-
The stabilities of complexes between avian ovomucoids and methylchymotrypsin were examined using equilibrium competitive binding assays. As shown in Fig. 5B, a mixture of equilimolar concentrations (4.7 X lo-' M) of active chymotrypsin, methylchymotrypsin, and turkey ovomucoid reached an equilibrium in which almost 20% of the active enzyme was found free. Fig. 5A shows a similar displacement except a lo-fold excess of the inactive enzyme was used. In this case 33% of the active enzyme was found free at equilibrium. This corresponds to a value of Kinactive/Kactive of 0.017. Similar results for duck and golden pheasant ovomucoids resulted in K. lnao+,ive/Kactive values of 0.014 and 0.011, respectively. The initial very rapid phase of the displacements (Fig. 5A) should be a second order process involving complex formation between free inhibitor and added enzyme. Subsequent to this rapid reaction, displacement of enzyme from a complex should be limited by the first order dissociation of the complex. This was confirmed as shown in Fig. 6 by reasonably linear first order plots. The dissociation rate constant for the active enzyme-turkey ovomucoid complex from this data is 8.9 x lo+ s-l and for the complex with methylchymotrypsin is 1.7 x 1OP. The fact that the calculated equilibrium constant (k,/kd) is smaller than the observed equilibrium constant by more than a factor of 2 (Table I) suggests that even when EEt = 10 Etot the experiments do not eliminate the contributions to the observed kd made by k,. The appropriate experimental conditions for calculation of /cd for the inactive enzyme-inhibitor complex are even more difficult to approximate and the observed value of 1.7 X 10P4 s-I may easily be much larger than the true dissociation rate constant.
This means that the rate constant for the dissociation of the complex between turkey ovomucoid and methylchymotrypsin differs from the rate for the active enzyme by no more than an order of magnitude and perhaps much less. The agreement between the two association constants (Table I) would also be better if the observed kd for the methylchymotrypsin complex more closely approximated the true kd. The use of metl~ylcl~yn~otrypsin as an affinity adsorbent has several distinct advantages over other possibilities. Native chymotrypsin is frequently used in the isolation of protein inhibitors but its use is complicated by the formation of a mixture of cleaved and native forms of the inhibitor.
The use of active chymotrypsin is particularly undesirable in the isolation of socalled temporary inhibitors because they are not completely resistant to proteolysis by proteinases.
Anhydrochymotrypsin has also been used as an affinity adsorbent but its use is limited by the fact that it is difficult to isolate in large quantities.
Methylchymotrypsin, however, is easily prepared in large amounts and the problem of undesirable proteolysis occurring during isolation is virtually eliminated.
The fact that turkey, ringnecked pheasant, and duck ovomucoids could be isolated on a methylchymotrypsin-Sepharose resin suggests that no major conformational changes occur when histidine-57 is mcthylated, and, in addition, that the catalytic activity of the enzyme is not absolutely necessary for the formation of a complex with these inhibitors.
Second order rate constants for the formation of complexes between proteinases and protein inhibitors at neutral pH values are usually very large (lo6 to 10' M-I 0) (19). These values are close to the association rate constants for small substrates (20) and approach the theoretical limit for the diffusion-controlled collision between macromolecules (21). As we have shown, however, turkey ovomucoid has an unusually low rate of association with chymotrypsin (2.2 x IO4 M-~ s-l). The activation energy for the association is 11.3 Cal per mol, which is higher than would be expected if a simple diffusion step were rate-limiting, but reveals little in addition about the nature of the rate-limiting step. Two explanations are possible. The formation of a complex may be limited by either a conformational change involving rearrangements of amino acid side chains of one or both of the proteins (22), or the formation of a structure resembling a catalytic intermediate (an acyl enzyme or a tetrahedral intermediate).
The latter proposal has been suggested as a general requirement for the inhibition of proteinases by protein inhibitors (3). However, our studies with methylchymotrypsin, in addition to work on other inactive enzyme derivatives, show any scheme involving the obligatory formation of any structure resembling the usual intermediates of catalysis to be very unlikely.
Methylchymotrypsin forms a very stable complex with turkey ovomucoid (K = 1.0 x 10' M-') and it does so with an association rate differing only by a factor of 13 from the association rate with the native enzyme (see Table I). Dissociation of a methylchymotrypsin-turkey ovomucoid complex occurs at a rate slightly faster than dissociation of a chymotrypsin-turkey ovomucoid complex.
Since the rate at which methylchymotrypsin is either acylated or deacylated by specific substrates is reduced by factors of 5 x lo3 to 2 x lo5 (7), it is clear that the formation of a complex with turkey ovomucoid does not involve either of these processes. When the active unmodified enzyme forms a complex with a protein inhibitor it is certainly possible for structures resembling the intermediates of peptide bond hydrolysis to form. The cleavage and resynthesis of reactive site peptide bonds in many inhibitors is clear evidence of this possibility.
However, the interaction of turkey ovomucoid with methylchymotrypsin strongly suggests that formation of any such intermediates is not a rate-hmiting part of the inhibitory process and does not contribute substantially to the stability of the complex when it is formed.
The formation of stable complexes between several inhibitors and methylchymotrypsin agrees well with previous observations of complex formation between inhibitors and other inactive proteinase derivatives. Inactive enzymes with added bulky groups in the active site, such as TPCK-chymotrypsin, form weak, but nevertheless observable, complexes with inhibitors of the native enzyme (5). Xlodifications which result in loss of activity but only slight increases in bulk and no apparent conformational changes, such as methylchymotrypsin, show only small losses in affinity for protein inhibitors.
And finally, modifications which result in no added groups in the active site, such as the anhydro derivatives of trypsin and chymotrypsin, can form complexes with inhibitors of equal or even greater stability than analogous complexes with the native enzyme (23, 24). Ako el al. have recently reported that lima bean inhibitor forms a complex with anhydrochymotrypsin which is more stable by 3 Cal than the complex with the native enzyme (23). Since anhydrochymotrypsin is unable to form any structures resembling catalytic intermediates, it is clear that formation of these intermediates must be incidental to the mechanism of inhibition. The strong interactions between proteinases and their protein inhibitors, therefore, must be due to the summation of a large number of noncovalent interactions. This is consistent with the large entropic contributions to complex formation (25) and the presence of such interactions has been confirmed in two cases by crystallographic examination of proteinase-protein inhibitor complexes (26,27). In the complex between trypsin and bovine pancreatic trypsin inhibitor, more than 200 individual Van der Waals contacts and several hydrogen bonds are formed between the two proteins (27). The additive effects of such individually weak forces are more than sufficient to account for the stability of the complex without invoking any more elaborate mechanistic proposals.