The Involvement of Tyrosyl and Amino Groups in the Interaction of Trypsin and a Soybean Trypsin Inhibitor*

SUMMARY Modtication of the tyrosyl and amino groups of trypsin and soybean trypsin inhibitor (Kunitz) with N-acetylimidazole before and after their combination was used to assess the role of these amino acids in the binding process. Trypsin was inactivated by au excess of the inhibitor and the complex isolated by chromatography on Sephadex G-75, at pH 6.8. Prior to complex formation, trypsin and the inhibitor each possessed 4 reactive tyrosyl residues, but after complex formation only 4 of the expected 8 reacted with N-acetyl-imidazole. Trypsin and the inhibitor had four and three acetylatable amino groups, respectively, but only two of these were reactive in the complex. These results indicated that four tyrosyl and five amino groups had been prevented from reacting with N-acetylimidazole because of their involvement in the formation of the complex. Modified but fully active trypsin and inhibitor could be recovered from the acetylated complex by chromatography on sulfoethyl Sephadex at pH 2.6. Subsequent analysis of each of these isolated components revealed that 2 of the 4 shielded tyrosyl residues of the

From the Department of Biochemistry, College of Biological Sciences, University of Minnesota, Xt. Paul, Minnesota 55101 SUMMARY Modtication of the tyrosyl and amino groups of trypsin and soybean trypsin inhibitor (Kunitz) with N-acetylimidazole before and after their combination was used to assess the role of these amino acids in the binding process. Trypsin was inactivated by au excess of the inhibitor and the complex isolated by chromatography on Sephadex G-75, at pH 6.8. Prior to complex formation, trypsin and the inhibitor each possessed 4 reactive tyrosyl residues, but after complex formation only 4 of the expected 8 reacted with N-acetylimidazole.
Trypsin and the inhibitor had four and three acetylatable amino groups, respectively, but only two of these were reactive in the complex. These results indicated that four tyrosyl and five amino groups had been prevented from reacting with N-acetylimidazole because of their involvement in the formation of the complex. Modified but fully active trypsin and inhibitor could be recovered from the acetylated complex by chromatography on sulfoethyl Sephadex at pH 2.6. Subsequent analysis of each of these isolated components revealed that 2 of the 4 shielded tyrosyl residues of the complex had been derived from trypsin and 2 from the inhibitor.
Of the five shielded amino groups in the complex, three were contributed by trypsin and two by the inhibitor.
The stoichiometric combination of trypsin with naturally occurring inhibitors such as one found in soybeans, soybean trypsin inhibitor,' has provided the protein chemist with an excellent model system for studying protein-protein interaction.
* This research was supported by Grants GM-4614 and AM-13869 from the National Institutes of Health, United States Public Health Service, and GB-15385 from the National Science Foundation.
1 In view of the fact that several different species of trypsin inhibitors are known to be present in soybeans (l), it is important to point out that the particular species of inhibitor referred to in this paper as ST1 corresponds to the inhibitor originally isolated by Kunitz (2) and subsequently identified by chromatographic techniques as Fraction STI-A2 (3) or Fz (4).
Despite the intensive amount of study which this system has received (see review in Reference I), the precise manner in which these two macromolecules combine to form an inactive complex remains to be fully elucidated.
It has been shown that trypsin which has been inactivated by chemical modification of its active site is no longer capable of combining with ST12 (5-7). Complementing these observations are the reports that an arginine-isoleucine bond in soybean trypsin inhibitor is split by trypsin under certain well defined conditions (8,9) and that modification of the arginine residues of ST1 leads to a loss in inhibitory activity (10, 11). Thus, in many respects, the combination of trypsin with ST1 appears to resemble the interaction of trypsin with a substrate for which it is specific, although Haynes and Feeney (la), on the basis of studies involving other protease inhibitors, have questioned whether proteolysis is an obligatory feature of the inhibition process.
Other than an involvement of the catalytic site of trypsin and an arginine-isoleucine sequence in soybean trypsin inhibitor, little is known about the actual size and chemical features of the contact zone wherein these two macromolecules interact. There is some evidence to indicate, however, that some of the tyrosine and tryptophan residues of trypsin and ST1 may be located at or near the zone of contact (13,14).
More extensive characterization of the combining sites of trypsin and ST1 would undoubtedly lead to a better understanding of the nature of the forces (covalent bonds, ionic attraction, hydrophobic forces, hydrogen bonding, etc.) involved in the formation and stabilization of the trypsin-ST1 complex. This paper reports an initial attempt to obtain this kind of information by using the following sequence of operations: (a) treatment of the trypsin-ST1 complex with a group-specific reagent; peptides derived from the labeled regions of trypsin and trypsin inhibitor.
In the present study Steps a to c have been successfully accomplished through the use of AcIm which selectively ncctglatrs the phenolic groups of tyrosyl residues and, to a lesser extent, the free amino groups of proteins (15). The data obtained permitted an assessment of the number of tyrosyl residues and amino groups which are involved in the site of interaction brtween trypsin and ST1 and the contribution made by each component. The solution was allowed to stand at 0" for 2 hours with occasional stirring.
A small amount of insoluble material which formed during this time was removed by centrifugation and the supernatant solution was applied to a column (4.5 x 103 cm) of Sephadex G-75 which had been equilibrated with the pH 6.8 imidazole buffer at 4". with the same buffer at a flow rate of 30 ml per hour and the clunte was collected in IO-ml fractions.
The cont4~nt~ of tubes corrcsl)onding to the trypsin-ST1 con~ples (Fig. I)  Acetylation with N-Acetylimidazole--Minor modifications of the procedure described by Simpson,Riordan,and Vallee (15) and Riordan, Wacker, and Vallee (18) were used to acetylate trypsin, STI, and the trypsin-ST1 complex with AcIm. A 120.fold molar excess of solid BcIm was added to the protein dissolved in 0.05 M imidazole buffer, pH 6.8, at room temperature. The pH was maintained at 6.8 by running the reaction in a pa-stat (Radiometer, Copenhagen, Denmark). When the acetylation of buried tyrosyl residues was desired, urea was included in the reaction mixture at a level which gave a final concentration of 8 M.
Aliquots of the reaction mixture were withdrawn at various periods of time and diluted with an equal volume of 0.2 M sodium acetate-O.05 M CaCl*, pH 4.4, which terminates the acetylation reaction and causes a rapid decomposition of excess N-acetylimidazole.
No subsequent hydrolysis of 0-or N-acetyl groups could be detected over a period of several months when these solutions were stored in a frozen state.
Zero time controls were obtained by following the same procedure in the absence of AcIm.

Dissociation of Trypsin-Soybean Trypsin Inhibitor
Complex-Dissociation of the nonacetylated trypsin-ST1 complex could be achieved by chromatography on sulfoethyl Sephadex C-25 at pH 2.6 with a salt gradient of 0 to 0.5 M NaCl as previously described for the purification of trypsin (16).
In the case of the acetylated complex, the ST1 component of the complex remained bound to the column under these conditions and could only be eluted by introducing 0.3 M sodium acetate buffer, pH 5.6, following the emergence of the trypsin component of the acetylated complex. Fig. 2 shows the chromatographic separation of trypsin and ST1 from the unmodified and acetylated complex. Based on activity measurements, tubes corresponding to trypsin and trypsin inhibitor from the acetylated and nonacetylated complexes were pooled, concentrated by ultrafiltration against 0.001 M HCl-0.05 M CaCl,, and kept frozen until used. Measurement of 0-Acetyl Groups-The number of 0-acetyl groups introduced onto the tyrosyl residues of the protein were calculated from the increase in absorbance at 278 nm following the addition of hydroxylamine (15). To 0.5 ml of the acetylated protein solution were added 2.5 ml of 0.

RESULTS
Acetylation of Tyrosyl Residues-The rates at which the tyrosyl residues of tryllsin, STI, and the trypsin-ST1 complex were acetylated in the presence and absence of urea are shown in Fig. 3. In the absence of urea the rate at which the tyrosyl residues of trypsin, STI, and the complex was acetylated was quite similar; 4 tyrosyl residues were acetylated in each case in about 15 min.
Thereafter the 0-ncetyl groups appear to undergo slow hydrolysis, as might be expected at pH 6.8 at which the acetylation was performed (18). In the presence of 8 M urea all of the tyrosine residues known to be present in trypsin and STI, 10 (25) and 4 (22), respectively, were acetyiated within 15 min. A total of 14 tyrosyl residues were ncetylated in the trypsin-ST1 complex, a value which represented the sum of the tyrosine residues of each of its components.
From these data it may be concluded that (a) 4 out of the 10 tyrosine residues of trypsin are reactive toward AcIm in the native protein, (b) all 4 of the tyrosine residues of ST1 are similarly reactive, and (c) once the complex is formed, 4 of these 8 reactive tyrosyl residues are no longer capable of reacting with N-acetylimidazole.
The fact that all of the tyrosine residues contained in the complex react with AcIm in 8 M urea may be explained by the observation that urea can effect the dissociation of the complex (26), in which case each component reacts with AcIm in an independent fashion. Acetylation of Amino Groups- Fig.  4 shows the rate at which the free amino groups of trypsin, STI, and the complex undergo acetylation by AcIm. The amino groups of trypsin and trypsin inhibitor were acetylated at somewhat different rates. The amino groups of trypsin were slowly acetylated, and only after 50 or 60 min did it appear that the reaction had subsided to the point at which approximately eight amino groups had been modified.
The amino groups of STI, on the other hand, were more rapidly acetylated, and the reaction was essentially complete in 40 min, at which time about four amino groups had been acetylated.
The trypsin-ST1 complex underwent acetylation in a fashion which seems to reflect the acetylation of the amino groups of ST1 in the early part of the reaction and the amino groups of trypsin in the latter stage.
Prior to acetylation it was found that only 12 of the 15 amino groups of trypsin and only 5 of the 12 amino groups of ST1 were reactive toward TNBS.
The nonacetylated complex was found to possess 12 TNBS-reactive amino groups, which means that   See text for details.
Each point is the average of at least two determinations.
a Values after 15 min of acetylation (see Fig. 3). 6 This value has been corrected for the partial hydrolysis of 0-acetyl groups that accompanies the isolation of ST1 from the acet.ylated complex (Fig. 2). c Sum of 0-acetyltyrosyl residues in trypsin and ST1 minus 0-acetyltyrosyl residues in complex. d 0-acetylt.yrosyl residues before complex formation minus 0-acetyltyrosyl residues after dissociation from acetylated complex.
e Number of tyrosyl residues, to the nearest integer, which are unreactive toward AcIm.  out of the total of 17 theoretical trinitrobenzene sulfonic acidreactive amino groups in the complex, five of these were no longer capable of reacting with TNBS after complex formation. If, from the acetylation data on Fig. 4, one plots the number of amino groups which escaped acetylation as a consequence of complex formation as a function of time, the curve shown in Fig. 5 may be constructed.
This plot reveals that the maximum number of amino groups which were shielded from reacting with AcIm is five, a value which was attained rather abruptly after I5 min of acetylation.
Longer periods of acetylation gave lower values for the number of protected amino groups, indicating that some of the masked amino groups may react more slowly with AcIm.
It may be reasonably concluded that the five amino groups which are masked during complex formation (as meas-to be refractory to acetylation provided that the time of acetylation is carefully controlled. Analysis time of the complex was limited to 15 min since the kinetic data had indicated that acetylation beyond this period favors deacetylation of 0-acetyl groups (Fig. 3) and, at the same time, permits the acetylation of masked amino groups in the complex (Fig. 5). The trypsin and trypsin inhibitor components of the acetylated complex were then isolated by chromatography (Fig. 2) and analyzed for 0-and N-acetyl groups. Data pertaining to tyrosine and amino groups are shown in Tables I and II, respectively, where a comparison is also made with the data obtained with trypsin and ST1 prior to complex formation.
Trypsin and ST1 derived from the acetylated complex were found to possess 2.1 and 2.0 0-acetyl groups, respectively, compared to 4 tyrosyl residues in each protein which could be acetylated prior to complex formation. Therefore of the 4 tyrosine residues shielded from reaction with AcIm in the complex, 2 are derived from trypsin and 2 from STI.
These conclusions are presented in a diagrammatic fashion in Fig. 6. Trypsin dissociated from the acetylated complex contained approximately one N-acetyl group compared to the four which were obtained in the absence of STI.
Therefore, three amino groups of trypsin are apparently prevented from reacting with   Table I). = Values after 15 min of acetylation (see Table II).
AcIm in the presence of STI. ST1 dissociated from the acetylated complex likewise had only one N-acetyl group which may be compared to the approximately three amino groups capable of being acetylated in the absence of trypsin. Hence, it may be Inhibitor Interaction Vol. 245,No. 19 concluded that two of the amino groups of ST1 escape acetylation when the latter is combined with trypsin. These conclusions are embodied in the diagram shown in Fig. 7.
Activity Measurements-Data pertaining to the activities of trypsin and ST1 before and after acetylation as well as after dissociation from the acetylated complex are presented in Table  III.
It is evident that the acetylation of the tyrosyl and amino groups of trypsin or ST1 as performed in these studies had no significant effect on their activities, neither did these proteins suffer any appreciable impairment in activity following their dissociation from the acetylated complex.

DISCUSSION
The extent to which the tyrosyl residues of trypsin may be modified and the changes in activity attendant upon such modification are highly dependent on the reagent and conditions which are used. Of the 10 tyrosine residues present in the trypsin molecule, spectrophotometric evidence indicates that 5 to 6 are exposed (27)(28)(29).
At least 4 of these exposed tyrosyl residues can be nitrated (30)(31)(32) or can react with cyanuric fluoride (33) without appreciable loss in activity toward synthetic substrates. 011 the other hand, reports on the effect of acetylating the tyrosyl residues with AcIm on the activity of trypsin have been highly variable.
As little as 1.7 to as high as 6.7 tyrosyl residues have been reported to be modified by Nacetylimidazole, with activities ranging from completely inactive to superactive, depending on the experimental conditions (34-39) .3 The conditions that we used were such that 4 tyrosyl residues of trypsin could be acetylated with AcIm with no significant effect on activity.
In this respect, our results agree with reports that 4 tyrosine residues of trypsin could be acetylated with AcIm (36) or nitrated with tetranitromethane (32) with very little effect on activity.
The fact that four of the amino groups of trypsin could also be acetylated by AcIm without affecting its activity was not unexpected since all of the e-amino groups of trypsin can be acetylated with acetic anhydride without loss in activity (41). Although the 4 tyrosine residues of ST1 titrate normally (23), they appear to react somewhat sluggishly with certain chemical modifying reagents. Thus only 1 to 2 tyrosines react with cyanuric fluoride (42), and 2 tyrosine residues can be iodinated without loss in activity (43). Only 1.5 (44) to 3 (42) tyrosine residues of ST1 have been reported to be reactive toward AcIm, values which are lower than the value of 4 reported here.
It shown that a maximum number of 4 tyrosine residues are acetylated at the end of 15 min, followed by a slow decline (decetylation) to a value of 3 at the end of 1 hour.
This would suggest that one of the 0-acetyl groups of acetylated ST1 is rather unstable, and its presence would be overlooked if the time of acetylation were prolonged. The acetylation of the 4 tyrosine residues of ST1 and three of its amino groups did not interfere with its ability to inhibit trypsin.
Other workers (10,11,20,43) have also shown that the activity of ST1 toward trypsin is unaffected by extensive modification of its amino groups. Most studies designed to elucidate the amino acid residues involved in the interaction of trypsin and ST1 have used chemical modification of certain amino acids prior to the formation of the complex (5-7, 11, 43, 45, 46). Any conclusion regarding the nature of the combining sites based on this approach must take into account the possibility that chemically induced ronformational changes might indirectly prevent the interaction of these two proteins.
Since there is no assurance that chemical modification will be limited to those residues which are directly in volved in the combining sites, conformational changes, leading to an inability of the two proteins to combine with rach other, could occur in regions of the molecule quite removed from the actual site of contact. Conversely, the failure of chemical modification to affect the ability of trypsin and ST1 to interact does not necessarily preclude the possibility that groups which have been chemically modified might still be located at or near the zone of contact.
Witness the fact that although acetylation of t,he tyrosine and amino groups of ST1 and trypsin does not affect their ability to interact, nevertheless, as reported here, some of these same residues are evidently located at or near the site of interaction.
How does one explain this apparent paradox? It would appear that chemical modification of an amino acid residue located at or near the zone of contact need not interfere with the formation of whatever bonds are involved in the binding process.
Once these bonds have been allowed to form, however, these same residues may be located in an environment which renders them sterically inaccessible to the modifying agent. Some of the uncertainties described in the preceding paragraph can be largely obviated by comparing the reactivity of certain amino acid residues before and after formation of the trypsinST1 complex.
Such a study was made many years ago by Kunitz (47) who measured the amino groups of trypsin and ST1 by form01 titration before and after complexation. From his data it may be calculated that 5.8 amino groups were no longer detected after the two proteins had combined; this value is not too far removed from the value of 5 reported here.
Physical measurements, including spectral changes, fluorescence, and optical rotation, have implicated tryptophan and tyrosine as being located within the confines of the combining site (13,48). When trypsin was combined with STI, 2 residues of tryptophan and 3 residues of tyrosine were shielded from reaction with N-bromosuccinimide and iodine, respectively (14). We conclude from our data, based on the acetylation of tyrosine residues by XcIm, that 4 tyrosine residues are involved in the site of contact. This discrepancy of 1 residue may be due to the difference in the reactivity of tyrosine residues toward iodine and AcIm.
Since tyrosine residues and amino groups are common to both trypsin and STI, the foregoing studies do not serve to distinguish those shielded residues which originate from try@ from those which are derived from STI.
The present study is believed to represent the first successful attempt to answer this 1 Melolm,FriE,and Slirm (49) have recently described experiments in which complexes of the pancreatic inhibitor with trypsin and chymotrypsin were nitrated with tetranitromethane and the nitrated inhibitor isolated from both complexes. The same 2 tyrosine residues were fonnd to be nitrated in both cases, and these were identical with the tyrosine residnes nitrated in the free inhibitor. 3.