Discovery of high affinity and ultraselective delta opioid dipeptide antagonists composed of 2',6'-dimethyl-L-tyrosyl (Dmt) and 1,2,3,4-tetrahydroisoquinoline-3-carboxylic (Tic) served as the basis for the conformationally restricted diketopiperazine cyclo(Dmt-Tic) and related open chain analogues. These peptides primarily bind to delta opioid receptors: c(Dmt-Tic) displayed 30- to 50-fold higher delta affinity (Ki delta) than its diastereo-isomeric analogues and more than 4000-fold greater than its Tyr cognate; all of the c(Tyr-Tic) analogues were essentially inactive; c[(N-methyl)Dmt-Tic] lost 5-fold in Ki delta, while Ki mu increased 10-fold to yield a nonselective peptide; and the c(Dmt-Phe) series exhibited considerably reduced binding which indicated a synergism between Dmt and Tic in the binding mechanism. Whereas acetyl-Dmt-Tic linear peptides weakly interacted with opioid receptors, Ac-Dmt-Tic-NH2, exhibited better delta antagonist activity than c(Dmt-Tic) and greater delta receptor selectivity (Ki mu/Ki delta = 570). A three point attachment hypothesis for the interaction between c(Dmt-Tic) and the delta receptor was proposed: hydrophobicity imparted by the aromatic rings and the methyl groups of Dmt, hydrogen bonding through the tyramine hydroxyl group, and cation-pi interactions were suggested as contributing factors in binding the diketopiperazine in the receptor pocket. Although c(Dmt-Tic) exhibited a weak antagonist activity with mouse vas deferens, this diketopiperazine may provide a scaffolding for the formation of more potent antagonists for potential therapeutic applications.