Thermodynamic criterion for the conformation of P-1 residues of substratesand of inhibitors in complexes with serine proteinases

Eglin c, turkey ovomucoid third domain, and bovine pancreatic trypsin inhib itor (Kunitz) are all standard mechanism, canonical protein inhibitors of s erine proteinases. Each of the three belongs to a different inhibitor famil y. Therefore, all three have the same canonical conformation in their combi ning loops but differ in their scaffoldings. Eglin c (Leu(45) at P-1) binds to chymotrypsin much better than its Ala(45) variant (the difference in st andard free energy changes on binding is -5.00 kcal/mol). Similarly, turkey ovomucoid third domain (Leu(18) at P-1) binds to chymotrypsin much better than its Ala's variant (the difference in standard free energy changes on b inding is -4.70 kcal/mol). As these two differences are within the +/-400 c al/mol bandwidth (expected from the experimental error), one can conclude t hat the system is additive. On the basis that isoenergetic is isostructural , we expect that within both the P-1 Ala pair and the P-1 Leu pair, the con formation of the inhibitor's P-1 side chain and of the enzyme's specificity pocket will be identical. This is confirmed, within the experimental error , by the available X-ray structures of complexes of bovine chymotrypsin A a lpha with eglin c (lacb) and with turkey ovomucoid third domain (1cho). A c omparison can also be made between the structures of P-1 (Lys(+))(15) of bo vine pancreatic trypsin inhibitor (Kunitz) (1mtn and 1cbw) and of the P-1 ( Lys(+))(18) variant of turkey ovomucoid third domain (1hja), both interacti ng with chymotrypsin. In this case, the conformation of the side chains is strikingly different. Bovine pancreatic trypsin inhibitor with (Lys(+))(15) at P-1 binds to chymotrypsin more strongly than its Ala(15) variant (the d ifference in standard free energy changes on binding is -1.90 kcal/mol). In contrast, turkey ovomucoid third domain variant with (Lys(+))(18) at P-1 b inds to chymotrypsin less strongly than its Ala(18) variant (the difference in standard free energies of association is 0.95 kcal/mol). In this case, P-1 Lys(+) is neither isostructural nor isoenergetic. Thus, a thermodynamic criterion for whether the conformation of a P-1 side chain in the complex matches that of an already determined one is at hand. Such a criterion may be useful in reducing the number of required X-ray crystallographic structu re determinations. More importantly, the criterion can be applied to situat ions where direct determination of the structure is extremely difficult. He re, we apply it to determine the conformation of the Lys(+) side chain in t he transition state complex of a substrate with chymotrypsin. On. the basis of k(cat)/K-M measurements, the difference in free energies of activation for Suc-AAPX-pna when X is Lys(+) and X is Ala is 1.29 kcal/mol. This is in good agreement with the corresponding difference for turkey ovomucoid thir d domain variants but in sharp contrast to the bovine pancreatic trypsin in hibitor (Kunitz) data. Therefore, we expect that in the transition state co mplex of this substrate with chymotrypsin, the P-1 Lys(+) side chain is dee ply inserted into the enzyme's specificity pocket as it is in the (Lys(+))( 18) turkey ovomucoid third domain complex with chymotrypsin.