The enkephalins are degraded in vitro by three types of metallopeptidases including a dipeptidylaminopeptidase (DAP) which releases Tyr-Gly. In order to test the physiological significance of this enzyme in enkephalin metabolism, a membrane-bound DAP from porcine brain was purified. The structural characteristics of the active site of this enzyme were studied using several enkephalin-related fragments as substrates and various peptides as inhibitors. The active site possesses an anionic moiety able to interact with the ammonium group of the substrate and four hydrophobic subsites S2, S1, S1', S2' surrounding the metal-containing catalytic site. This porcine brain enzyme shows the same characteristics as a partially purified DAP from rat brain. Potent inhibitors of both enzymes were obtained with compounds interacting with either the S1', S2', or S1, S2 parts of the active site and containing a hydroxamate of an N-acyl-N-hydroxy amino group as the metal-chelating agent. The compounds synthesized to bind the S1', S2' subsites, such as HN(OH)--CO--CH2--CH(CH2 phi)-CONH--CH(CH2 phi)--COOH, behave as highly potent and mixed inhibitors of both DAP (Ki = 2.5 nM) and enkephalinase (Ki = 0.3 nM). In contrast, Tyr-Phe-NHOH, designed to interact with S2,S1 subsites through its positively charged Tyr as the P2 component, is a highly potent (Ki = 9 nM) and selective DAP inhibitor. The ability of Tyr-Phe-NHOH to protect, in vivo, endogenous enkephalins was evaluated using the hot plate test with mice. The intracerebroventricular injection of Tyr-Phe-NHOH alone (0.17 mumol, 60 micrograms) does not significantly modify the jump latency time as compared to the control. The effect of the association of Tyr-Phe-NHOH (0.17 mumol, 60 micrograms), bestatin (0.16 mumol, 50 micrograms) and thiorphan (0.20 mumol, 50 micrograms) is not statistically different from that produced by bestatin (0.16 mumol) plus thiorphan (0.20 mumol). Likewise, the effect of bestatin (50 micrograms) or thiorphan (25 micrograms) is not significantly modified by intracerebroventricular co-administration of increasing concentrations of Tyr-Phe-NHOH. These results seem to indicate that the DAP does not play a major role in enkephalin metabolism, at least at the supraspinal level.