N-terminally shortened analogs of the 27-amino-acid and 38-amino-acid forms of the pituitary-adenylate-cyclase-activating neuropeptide, PACAP(1-27) and PACAP(1-38), were synthesized by a solid-phase method. Systematic deletion of the first 13 amino acids of both PACAP was tested by evaluating their ability to occupy the specific and selective PACAP receptor of human neuroblastoma NB-OK-1 cell membranes and to stimulate adenylate cyclase or, when inactive per se, to inhibit PACAP-stimulated adenylate cyclase activity. For each peptide, the Kact (concentration required for half-maximal adenylate cyclase activation) or Ki [concentration required to shift the dose/response curve of PACAP(1-27) twofold to the right] was in good agreement with the corresponding IC50 [concentration inhibiting 50% of 125I-[AcHis1]PACAP(1-27) binding to membranes], suggesting interaction with the same homogeneous class of adenylate cyclase-coupled receptors. The deletion of the two first amino acids (His1 and Ser2) sufficed to decrease the affinity for receptors and to suppress the capacity to activate adenylate cyclase. The shorter fragments 3-27 and 3-38, 4-27 and 4-38, 5-27 and 5-38, 6-27 and 6-38, 7-27 and 7-38, 8-27 and 8-38, and 9-27 and 9-38 were all competitive antagonists of PACAP(1-27)-stimulated activity with the N-terminally shortened PACAP(1-38) derivatives being 4-30-fold more potent than the equivalent PACAP(1-27) derivatives. In this group PACAP(6-38) was the most potent antagonist (Ki 1.5 nM). Surprisingly, the N-terminally shorter fragments 10-27 and 10-38, 11-27 and 11-38, 12-27 and 12-38, 13-27 and 13-38, and 14-27 and 14-38 were again able to stimulate adenylate cyclase, the smallest fragments, PACAP(14-27) and PACAP(14-38), being the most potent and efficient (Kact 2 microM and 0.1 microM, respectively). In this group of agonists, PACAP(1-38) derivatives deleted at the N-terminus were also more potent than the equivalent PACAP(1-27) derivatives.