SUMMARY1. Pressure sensitive (P) neurons of the leech Hirudo medicinalis show both an inhibitory, Cl--dependent response and a depolarizing, cationic response to pipette application of serotonin (5-HT). Serotonergic Retzius (R) neurons in culture reform inhibitory, Cl--dependent synapses with P neurons but fail to elicit the extrasynaptic, depolarizing response to 5-HT. We have examined the localization of the selection of 5-HT responses by testing the sensitivity of P cell growth cones and neurites to 5-HT application.2. As measured by intracellular recording at the P cell soma, synaptic release of 5-HT from R cell processes activated only the Cl--dependent response in P cell neurites. Focal application of 5-HT from a micropipette depolarized uncontacted P cell growth cones and neurites. In contrast, processes from the same P cells that were contacted by R cells were rarely depolarized by 5-HT application unless the application pipette was moved along the neurites away from the sites of contact.3. The channels underlying the depolarizing response to 5-HT were identified in patch clamp recordings from P cell growth cones. These cation channels showed rare, brief openings in the absence of 5-HT. Application of 5-HT in the bath (outside the patch pipette) increased channel activity in uncontacted P cell growth cones but not in growth cones of the same P cells contacted by R cells.4. We conclude that the selection of transmitter responses during synapse formation was localized to discrete sites of contact between the synaptic partners.

1. We have examined whether tyrosine phosphorylation is required for synapse formation between identified neurons from the central nervous system of the leech in culture. 2. Within a few hours of contact with the cell body of the serotonergic Retzius neuron (R cell), the soma of the postsynaptic pressure-sensitive neuron (P cell), but not the R cell, could be labelled intracellularly with an antibody against phosphotyrosine residues. The labelling seemed specific for P cells contacted by R cells, as it was greatly reduced in pairs of either R or P cells and in single cells. Genistein (20,C4M) and lavendustin A (10 /sM), selective inhibitors of tyrosine kinases, blocked the labelling of contacted P cells, whereas their ineffective analogues (genistin and lavendustin B) had no effect on labelling. 3. R cell contact also induced the loss of an extrasynaptic, depolarizing response (due to modulation of cation channels) to serotonin (5-HT) in the P cell within a few days of juxtaposing cell bodies and within an hour of contact with growth cones. Treatment of the neurons with the tyrosine kinase inhibitors (but not the ineffective analogues) prevented the loss of the depolarizing response and of single cation channel modulation by 5-HT. 4. R cells formed inhibitory, CF-dependent synapses with P cells. Synapse formation was prevented by the tyrosine kinase inhibitors but not by their ineffective analogues. These compounds had no obvious effect on neurite outgrowth or cell adhesion. We conclude that tyrosine phosphorylation is a signal during the formation of this synapse.