W e examined cell-attached patches on principal cells of primary cultured, rabbit cortical collecting tubules. Under basal conditions, apical 9-pS Cl(-)-selective channels were observed in 9% of patches (11/126), and number of channels times open probability (NP0) was 0.56 +/- 0.21. The channel had a linear current-voltage relationship, reversal potential (Erev) near resting membrane potential, a P0 (0.30-0.70) that was independent of voltage, and complicated kinetics (i.e., bursting) at hyperpolarized potentials. NP0 and channel frequency were increased after 30 min of basolateral exposure to 0.5 microM PGE2 (18/56), 10 microM forskolin (23/36), or 0.5 mM dibutyryl cyclic adenosine monophosphate (cAMP) (25/41). Increases in NP0 appeared to be mediated primarily through an increase in the number of observed channels per patch (N), not changes in P0. After these cAMP-increasing maneuvers, N was inconsistent with a uniform distribution of channels in the apical membrane (P < 0.001), but rather the channels appeared to be clustered in pairs. Apical 0.5 microM PGE2 (12/91), apical or basolateral 0.5 microM PGF2 alpha (8/110), or 0.25 microM thapsigargin (releaser of intracellular Ca2+ stores) (7/73) did not increase NP0 or channel frequency. Conclusions: (a) 9-pS Cl- channels provide a conductive pathway for apical membrane Cl- transport across principal cells. (b) Channel activation by basolateral PGE2 is mediated via a cAMP-, but not a Ca(2+)-dependent mechanism. (c) Apical channels are clustered in pairs. (d) With its low baseline frequency and Erev near resting membrane potential, this channel would not contribute significantly to transcellular Cl- flux under basal conditions. (e) However, cAMP-producing agonists (i.e., PGE2, arginine vasopressin) would increase apical Cl- transport with the direction determined by the apical membrane potential.