In order to improve the in vivo protection of enkephalins from enzymatic degradation, a new series of inhibitors derived from kelatorphan [HONHCOCH2CH(CH2Ph)CONHCH(CH3)COOH], the first-described complete inhibitor of enkephalin metabolism, were designed by modification of the C-terminal amino acid. The progressive lengthening of the chain of this residue shows that a beta-alanine seems to be the best basic model for the conception of such types of compounds. On the other hand, the methylation of the amide bond, which is well accepted by aminopeptidase N (EC 3.4.11.2) and dipeptidylaminopeptidase, induced a significant loss of affinity for neutral endopeptidase -24.11. Starting from these data, compounds containing a variously substituted beta-alanine residue and corresponding to the general formula HONHCOCH2CH(CH2Ph)CONHCH(R1)CH(R2)COOH were synthesized. All these molecules inhibit neutral endopeptidase -24.11 and dipeptidylaminopeptidase in the nanomolar range, and those containing an aromatic chain (compound 7A, R1 = CH2Ph,R2 = H, and compound 8A, R1 = Ph, R2 = H) inhibit the biologically relevant aminopeptidase N, with IC50's around 10(-8) M. Intracerebroventricular injection in mice of these multienzyme inhibitors produced an efficient and naloxone-reversible analgesic response (hot plate test): compounds 7A and 8A were shown to be more potent than kelatorphan in increasing the jump latency time, in agreement with their in vitro properties, and these new compounds were found to increase the forepaw lick latency, a reflex considered as a typical morphine response.