Experimental observations showed that the analogs [D-Trp2]- and [D-Phe2]methionine enkephalin amide were weakly active in releasing GH from rat pituitary in vitro. These observations were used to design more active GH-releasing factors. Conformational energy calculations were carried out, and energetically favored conformations of these polypeptides were found. Structural similarities as well as structural differences between active and inactive analogs were examined, and new sequences were predicted. Progressively more active analogs were designed, then synthesized, and tested. This cycle of steps was repeated, each time using structural and chemical concepts as design guides, until a series of very active analogs resulted. The most active analog to date, Tyr-D-Trp-Ala-Trp-D-Phe-NH2, was shown to release GH in vitro at 10-30 ng/ml medium, which is approximately 10(3) times more active than the two starting enkephalin-based analogs. From the structure-activity data, a mechanism for binding at the receptors is formulated, and a comparison is made between the structural relationships of the GH-releasing peptide analogs and the GH inhibitor, somatostatin.