The structure of carboxypeptidase A. 8. Atomic interpretation at 0.2 nm resolution, a new study of the complex of glycyl-L-tyrosine with CPA, and mechanistic deductions.

Abstract

Bovine pancreatic carboxypeptidase Aa (CPA),J the subject of these studies, is a zinc containing enzyme of molecular weight 34 600, which catalyses the hydrolysis of polypeptides and esters at the C-terminal peptide or ester bond. Experiments to date have shown that in order to be hydrolysed, the substrate must contain a C-terminal residue in the l configuration, with the carboxyl group free and α to the peptide or ester bond which is to be cleaved. In addition, the reaction is favoured if the C-terminal residue is aromatic. Our crystallographic studies of CPA have yielded electron density maps of the native enzyme at 0.6 nm (Lipscomb, Coppola, Hartsuck, Ludwig, Muirhead, Searl & Steitz 1966), 0.28 nm (Ludwig, Hartsuck, Steitz, Muirhead, Coppola, Reeke & Lipscomb 1967) and 0.20 nm resolution (Reeke, Hartsuck, Ludwig, Quiocho, Steitz & Lipscomb 1967; Lipscomb, Hartsuck, Reeke, Quiocho, Bethge, Ludwig, Steitz, Muirhead & Coppola 1968). Concurrently, a study of the binding of a number of substrates and inhibitors at 0.6 nm resolution was under way (Steitz, Ludwig, Quiocho & Lipscomb 1967). Subsequently, the most promising of these complexes, that of glycyl-L-tyrosine with CPA, was carried to atomic resolution (Reeke et al. 1967; Lipscomb et al. 1968). Even though chemical sequence is not available for several of the binding and catalytic groups of the enzyme, we have been able to deduce the identity of the binding residue Arg-145 and the catalytic residue Glu-270 (Reeke et al. 1967), to describe the mode of binding of Gly-Tyr, to extrapolate these conclusions to the binding of polypeptides, and to propose a mechanism, with certain ambiguities, for the action of the enzyme (Lipscomb et al. 1968). We have now completed the detailed atomic interpretation of the 0.20 nm electron density map, supplying residue identifications where the primary sequence is not available, and have subjected the resulting atomic coordinates to a model building procedure (Diamond 1966) which forces them to conform to standard bond distances and angles. The improved coordinates have been entered in a structure factor calculation, which gave a standard crystallographic R factor of 0.44.