The depressant action of adenosine on acetylcholine release at frog motor nerve terminals was studied by intracellular recording and Ca(2+)-imaging techniques. Adenosine (200 microm) quickly and reversibly decreased the amplitude and quantal content of end-plate potentials (EPPs) with no change in quantal size in a low-Ca(2+), high-Mg(2+) solution, and EPP amplitude in normal Ringer containing d-tubocurarine. Likewise, adenosine (200 microm) reduced miniature EPP (MEPP) frequency, but not amplitude, in a high-K(+) (6 mm) solution. Adenosine (40-200 microm), however, did not affect single or repetitive impulse(s)-induced rises in Ca(2+) in the nerve terminals or its basal level. Adenosine (100-200 microm) reduced the Ca(2+)-independent enhancement of MEPP frequency caused by hypertonicity. EPPs induced by tetanic stimulation (33 Hz) in Ringer with d-tubocurarine initially increased in amplitude within 10 stimuli and then declined to the minimum. Adenosine (200 microm) decreased EPP amplitude in the initial phase of the tetanus, but enhanced it in the middle phase, thus prolonging the decay of EPP amplitude. The total sum of these EPPs, reflecting the readily releasable pool of vesicles and its refilling, however, was not changed. The results suggest that adenosine inhibits a Ca(2+)-independent step of transmitter exocytosis at frog motor nerve terminals.