Two-component latency distributions indicate two-step vesicular release at simple glutamatergic synapses

It is often assumed that only stably docked synaptic vesicles can fuse following presynaptic action potential stimulation. However, during action potential trains docking sites are increasingly depleted, raising the question of the source of synaptic vesicles during sustained release. We have recently developed methods to reliably measure release latencies during high frequency trains at single synapses between parallel fibers and molecular layer interneurons. The latency distribution exhibits a single fast component at train onset but contains both a fast and a slow component later in the train. The contribution of the slow component increases with stimulation frequency and with release probability and decreases when blocking the docking step with latrunculin. These results suggest that the slow component reflects sequential docking and release in immediate succession. The transition from fast to slow component, as well as a later transition to asynchronous release, appear as successive adaptations of the synapse to maintain fidelity at the expense of time accuracy.