Pairing-specific, activity-dependent presynaptic facilitation at Aplysia sensory-motor neuron synapses in isolated...

Synapses made by Aplysia sensory neurons onto motor- and interneuron followers in the intact nervous system exhibit an associative form of synaptic facilitation that is thought to contribute to classical conditioning of the animal's gill and siphon withdrawal reflex (Hawkins et al., 1983; Walters and Byrne, 1983). Here we demonstrate that a similar associative facilitation can be induced between individual sensory and motor neurons isolated in culture. Pairing tetanic stimulation with either of two facilitatory transmitters, 5-HT or small cardioactive peptide, considerably prolongs facilitation compared to either tetanus or transmitter alone. When corrected for the depression that occurs simply in response to low-frequency testing, the facilitation produced by one pairing trial does not decay for more than 20 min after training. This facilitation requires the temporal pairing (0.5 sec forward interstimulus interval) of the two stimuli, tetanus and 5-HT. Delivering the same two stimuli in an unpaired fashion (1 min forward interval) fails to produce the long-lasting effect. Measurements of spontaneous transmitter release during either paired or unpaired training reveal no changes in unitary mEPSP or mEPSC ("mini") amplitude, indicating that the facilitation involves a presynaptic mechanism. While both forms of training dramatically increase the initial frequency of spontaneous release, mini frequency does not remain elevated as long as the evoked EPSP following paired training, nor does paired training specifically enhance spontaneous release frequency. Pairing-specific facilitation was not blocked by the protein kinase C inhibitor H7. In contrast, the same training procedure produced pairing-specific increases of sensory neuron excitability and action potential width, suggesting that cAMP-mediated processes are involved in the paired effect. Although Ca2+ influx is necessary for the associative effect (Abrams, 1985), we find that the facilitation does not require influx through L-type voltage-gated Ca2+ channels, since the effect was not blocked by the dihydropyridine antagonist nitrendipine. Together, these findings indicate that the mechanism underlying associative, activity-dependent facilitation is intrinsic to the sensory neuron synapse, that it is presynaptically mediated by processes unique to evoked synaptic transmission, and that it appears to involve a pairing-specific broadening of the presynaptic action potential, allowing enhanced Ca2+ influx through the dihydropyridine-insensitive channels responsible for release.