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

Takafumi Miki,Gerardo Malagon,Erwin Neher,Alain Marty,Yukihiro Nakamura

doi:10.1038/s41467-018-06336-5

Abstract

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.

Highlights

It is often assumed that only stably docked synaptic vesicles can fuse following presynaptic action potential stimulation
A well-documented phenomenon involving a strong change in synaptic latency is ‘delayed release’, called ‘asynchronous release’, where release extends for 10 s of ms to 10 s of s following the end of an action potential (AP) train[26]
We determined the latency of individual synaptic vesicles (SVs) release events[31] (Fig. 1a; in this plot, latency is defined as the time difference between presynaptic stimulation and the onset of individual quantal EPSCs)

Summary

Introduction

It is often assumed that only stably docked synaptic vesicles can fuse following presynaptic action potential stimulation. Following action potential (AP) stimulation, the release of synaptic vesicles (SVs) occurs with a jitter that was estimated around 1 ms at the frog neuromuscular junction at room temperature[1]. This jitter reflects the short period of time when the presynaptic calcium concentration ([Ca2+]) transient in the vicinity of SVs is large enough to elicit release, as well as the subsequent fusion steps[2,3,4], and potentially contains key information on synaptic function. Another potentially relevant factor is the relation between release kinetics and the distance between SV docking site (DS) and voltage-gated Ca2+ channels (VGCCs)[18,19] If this distance varies among docked SVs, latencies gradually grow during trains due to an increasing participation of low release probability, slow SVs20 In spite of these similarities between facilitation and asynchronous release, it has been suggested that facilitation and asynchronous release each rely on a specific Ca2+ sensor (e.g., ref. 28), but so far no consensus has emerged concerning the molecular nature of these potential sensors[26]

Methods

Results

Conclusion