The kinetics of the reaction of a series of peptide chloromethyl ketones with subtilisin BPN' (EC 3.4.21.14) were investigated in order to correlate the rates of reaction in solution with the number of interactions which are observed in the cystallographic model of the inhibited enzyme (Robertus, J.D., Alden, R.A., Birktoft, J.J., Powers, J.C. and Wilcox, P.E. (1972) Biochemistry 11, 2439--2449). The second-order rate constant kobs/[I], measured at pH 7.0 in 9% 1,2-dimethoxyethane at 30 degress C, varied by a factor of 5900 from the slowest (Z-TrpCH2Cl) to the fastest inhibitor (Ac-Phe-Gly-Ala-LeuCH2Cl). Inhibitors with a P1 leucine or phenylalanine residue are equally effective. The increased reactivity of inhibitors containing alanine as the P2 residue is the result of a favorable contact between the methyl side chain of the alanyl residue and the S2 subsite of subtilisin. This result correlates nicely with the previously observed "secondary specificity" of subtilisin for substrates with alanine as the P2 residue. Teetrapeptide and tripeptide chloromethyl ketone inhibitors have kobs/[I] values of over 100-fold greater than those of most dipeptide and amino acid chloromethyl ketones. This again agress with the crystallographic model since tripeptide and tetrapeptide inhibitors could form a beta-sheet structure involving three or four hydrogen bonds with the enzyme while the others would form fewer hydrogen bonds. The S4 subsite of subtilisin exhibits a distinct preference for aromatic groups and our four most reactive inhibitors (Ac-Phe-Gly-Aca-LeuCH2Cl, Z-Gly-PheCH2Cl, Z-Gly-Gly-LeuCH2Cl, and Boc-Ala-Gly-PheCH2Cl) have either an aromatic (or a large hydrophobic) group as the P4 residue. The results demonstrate that the solution reactivity of peptide chloromethyl ketones can be explained on the basis of the crystal structures of chloromethyl ketone-inhibited substilisin derivatives and that substrate hydrolysis rates can be used to design effective chloromethyl ketone inhibitors for serine proteases.