Studies on protein semisynthesis. I. Formation of esters, hydrazides, and substituted hydrazides of peptides by the reverse reaction of trypsin.

An enzymatic method for converting tryptic peptides into their intermediates for the azide method of peptide synthesis was investigated by using model substrates, Bz-Arg, Bz-Gly-Arg, and Bz-Gly-Lys. The conditions favorite for the formation of the esters, hydrazides, and substituted hydrazides were deduced from the theoretical analysis on the pH and pKa dependence of the synthetic reactions and the pKa values of the substrates in various solvents. The theory based on the empirical relationship between the pKa of amines and the formation constant of the amides (Fersht, A.R. & Requena, Y. (1971) J. Am. Chem. Soc. 93, 3499-3504) indicates that the most favorite amine in the formation of amides is the one with a pKa value equal to the pKa of the carboxyl group and with such amines the amide formation takes place almost quantitatively even in aqueous solutions. Boc- and Cbz-hydrazine were found to satisfy the above condition approximately. The methyl esters of the substrates were formed in 30-35% yields in 50% aqueous methanol at pH near 4.5 by the reverse reaction of trypsin [EC 3.4.4.4]. Bz-Gly-Arg-NHNH2 was formed in a 34% yield in 2 M aqueous hydrazine and in 70-80% yields in 50% aqueous solutions of dioxane, DMF, and DMSO containing 1 M hydrazine at pH near 7. Bz-Gly-Arg-NHNH-Boc and Bz-Gly-Lys-NHNH-Boc were formed quite readily in 90-98% yields in H2O and in 50% aqueous solutions of dioxane, DMF, and DMSO containing 0.5 or 1 M Boc hydrazine at pH from 4 to 5. The Cbz-hydrazides of the peptides were also formed readily in 88-95% yields in the above mixed solvents containing 0.5 M Cbz-hydrazine. The formation constants of the hydrazides in H2O were in good agreement with those predicted by the empirical relationship. The formation constants in the form of log Ke in the mixed solvents calculated by taking the activity coefficients of the reactants as unity were 0.1 to 0.5 lower than the predicted values.