The effect of Arg-atriopeptin III (ANP) on basal and stimulated (angiotensin II, acetylcholine and KCl depolarization) arginine vasopressin (AVP) release was characterized in the intact hypothalamo-neurohypophysial explant (HNS) and in isolated neurointermediate pituitary lobes (NIL). In initial experiments using 15-min incubation periods, ANP 10(-10) and 10(-9) M slightly inhibited basal AVP release in both NIL and HNS after a delay of at least 15 min. The most effective ANP concentration was 10(-10) M, and the inhibitory effect on AVP release was more marked in HNS (-52 +/- 5% of control compared to -29 +/- 8% for NIL). However, ANP 10(-10) M did not significantly attenuate KCl- or AII (10(-5) M)-stimulated AVP release from HNS after 15 min of exposure. When the incubation periods were increased to 30 min ANP 10(-10) and 10(-9) M significantly decreased AII-stimulated (10(-5) M) AVP release in a dose-dependent manner (p less than 0.05; p less than 0.01, respectively). The same concentrations of ANP did not significantly depress ACH-stimulated (10(-5) M) AVP release (p less than 0.1 for both concentrations). In summary, ANP generally inhibits AVP release in vitro by a slowly activated mechanism which appears to be specific for certain physiological stimuli. Although the site(s) of action cannot be absolutely localized to the ventral hypothalamus and/or the neurohypophysis, an effect in the hypothalamus seems very likely.