On the basis of the minimal octapeptide sequence of the renin substrate, a series of peptides was synthesized containing (3S,4S)-4-amino-3-hydroxy-6-methylheptanoic acid (statine) or (3S,4S)-4-amino-3-hydroxy-5-phenylpentanoic acid (AHPPA) at the P1P1' position. Some of these peptides also contained Nin-formyltryptophan at the P5, P3, or P3' position. Renin-inhibitory potency varied over a wide range (from inactive to IC50 = 3 nM). Potency was reduced by at least 10-fold when the peptide was shortened by two residues at either the amino or carboxy terminus. The AHPPA-containing inhibitors were several-fold less potent than the statine-containing inhibitors. Analysis of models for the three-dimensional structure of inhibitors at the active site of human renin suggests that the diminished potency of the AHPPA peptides in comparison with the statine-containing peptides was caused by a shift in the peptide backbone due to a steric conflict between the phenyl ring of the AHPPA residue and the S1 subsite. The importance of the side chain and the 3(S)-hydroxyl group of the statine residue was demonstrated by substituting 5-aminovaleric acid for a dipeptide unit at the P1P1' position, which resulted in a peptide devoid of renin-inhibitory activity. Substitutions of other basic amino acids for histidine at the P2 position caused a great loss in potency, possibly due to disruption of a hydrogen bond as suggested by molecular modeling. Studies on the plasma renins of four nonhuman species suggest that the isoleucine-histidine segment at the P2'P3' position is important to defining the human specificity of the substrate. This work suggests a number of properties important to the design of potent renin inhibitors, and demonstrates the usefulness of three-dimensional models in the interpretation of structure-activity data.