1. After blocking K+ currents with 10 mM-tetraethylammonium (TEA) or TEA plus 250-mu-M-3,4-diaminopyridine (3,4-DAP), motor nerve terminal Ca2+ currents were recorded using focal extracellular electrodes. Two transmitters released from the terminal, ATP and acetylcholine (ACh), were then applied, and the effects on the nerve teminal Ca2+ current were measured. 2. ATP (50-mu-M) reduced the Ca2+ current by 34%, but this action is prevented when hydrolysis to adenosine is blocked by alpha, beta-methyladenosine 5'-diphosphate (200-mu-M). Thus, inhibition by ATP presumably occurs subsequent to ATP hydrolysis to adenosine. 3. Adenosine (50-mu-M) inhibited the terminal Ca2+ current by 29%. This was mimicked by the adenosine analogue L-phenylisopropyl adenosine (L-PIA) and blocked by theophylline (100-mu-M), which antagonizes adenosine receptors at micromolar concentrations. 4. ACh (100-mu-M) or the anticholinesterase methane sulphonyl fluoride (MSF; 1 mM) also depressed the terminal Ca2+ current. This response was mimicked by muscarine (100-mu-M) and antagonized by atropine (100-mu-M) or pirenzipine (4-mu-M), which is generally specific for M1 receptors. 5. Addition of Ba2+, which blocks adenosine-mediated K+ currents, had no effect on the inhibitory effects of adenosine or ACh; similarly, neither adenosine nor ACh in the bath affected K+ current records obtained after blocking all inward currents with 10 mM-Co2+ and focal application of tetrodotoxin. 6. Incubation of the muscle for 4 h in pertussis toxin (10(-5) g ml-1) eliminated both adenosine- and ACh-induced inhibition of the terminal Ca2+ current. This result indicates the possible involvement of a G protein in the transduction of the feedback pathway. 7. Neither cyclic AMP analogues, the adenylate cyclase activator forskolin (10-mu-M), the phorbol ester phorbol 12-myristate 13-acetate (PMA; 3-mu-M) nor the diacyglycerol analogue 1,2-oleoylacetylglycerol (OAG;3-mu-M) had any effect on adenosine- or ACh-induced depression of the terminal Ca2+ current. Therefore, pathways involving these particular second messengers are most probably not involved. 8. The effects of adenosine and ACh are non-additive. 9. These results indicate that ATP and ACh, which are released during exocytosis, may inhibit their own release through attenuation of the terminal Ca2+ current via autoreceptors coupled to a G protein.