Phosphinic peptides, the first potent inhibitors of astacin, behave as extremely slow-binding inhibitors

A series of phosphinic pseudo-peptides varying in length and composition have been designed as inhibitors of the crayfish zinc endopeptidase astacin, the prototype of the astacin family and of the metzincin superfamily of metalloproteinases. The most efficient phosphinic peptide, fluorenylmethyloxycarbonyl-Pro-Lys-PhePsi(PO2CH2)Ala-P ro-Leu-Val, binds to astacin with a Ki value of 42 nM, which is about three orders of magnitude below the corresponding values for previously used hydroxamic acid derivatives. However, the rate constants for association (kon = 96.8 M-1.s-1) and dissociation (koff = 4.1 x 10(-6) s-1) are evidence for the extremely slow binding behaviour of this compound. N-terminally or C-terminally truncated phosphinic analogues of this parent molecule are much less potent, indicating a critical role of the peptide size on the potency. In particular, omission of the N-terminal proline residue leads to a 40-fold increase in Ki which is mostly due to a 75-fold higher koff value. These findings are consistent with the previously solved crystal structure of astacin complexed with one of the phosphinic peptides, benzyloxycarbonyl-Pro-Lys-PhePsi(PO2CH2)Ala-Pro-O-methyl, Ki = 14 microM [Grams, Dive, Yiotakis, Yiallouros, Vassiliou, Zwilling, Bode and Stöcker (1996) Nature Struct. Biol. 3, 671-675]. This structure also reveals that the phosphinic group binds to the active site as a transition-state analogue. The extremely slow binding behaviour of the phosphinic peptides is discussed in the light of the conformational changes involving a unique 'tyrosine switch' in the structure of astacin upon inhibitor binding. The phosphinic peptides may provide a rational basis for the design of drugs directed towards other members of the astacin family which, like bone morphogenetic protein 1 (BMP1; i.e. the procollagen C-proteinase), have become targets of pharmacological research.