The present study examined the ability of two analogs of human/rat corticotropin-releasing factor (h/rCRF), [Met18, Lys23, Glu27,29,40, Ala32,41, Leu33,36,38] h/rCRF9-41 (alpha-hel CRF9-41) and [D-Phe12, Nle21,38, C alpha MeLeu37] h/rCRF12-41 (D-Phe CRF12-41), to antagonize CRF- and stress-induced behavioral changes in rats. The potency and duration of action of D-Phe CRF12-41 in vivo was compared with that of alpha-hel CRF9-41, the most effective CRF antagonist studied to date. When administered i.c.v., both CRF antagonists dose-dependently reduced the locomotor activity induced by rat CRF (0.5 microgram/rat i.c.v.) and attenuated the social stress-induced anxiogenic-like effect, as measured by the elevated plus-maze. However D-Phe CRF12-41 was 5 times more potent and remained effective for a longer period (> 1.5 hr) than alpha-hel CRF9-41. High doses of alpha-hel CRF9-41 (25 micrograms/rat), by itself, exhibited weak agonist effects, which were not observed with D-Phe CRF12-41 at the same doses. These results demonstrate that a N-terminus-shortened CRF antagonist that encompasses the substitution D-Phe12 has significantly higher biological potency and an extended duration of action without intrinsic agonist effects in these in vivo systems. The availability of potent and effective CRF antagonists will provide valuable tools for exploring the functional role of brain CRF and may ultimately lead to an understanding of the biological basis of stress-related illnesses.