Synaptic vesicle exocytosis-endocytosis at central synapses

Abstract

At nerve terminals, neurotransmitter release is mediated by exocytosis of synaptic vesicles at active zone. After exocytosis, vesicular components are efficiently retrieved by endocytosis. Tight coupling between synaptic vesicle exocytosis and endocytosis is critical for the maintenance of neurotransmission at central synapses. Recently, we have developed a new fluorescent pH reporter that permits us to examine exocytosis-endocytosis coupling at the level of individual synaptic vesicles at hippocampal synapses.¹ To our surprise, we observed that the tight coupling of exocytosis and endocytosis broke down at very low stimulation frequencies, resulting in the generation of two endocytic vesicles per single exocytic fusion event. As stimulation frequency increased, exocytosis-endocytic coupling was restored with one endocytic vesicle generated for each vesicle exocytosed. Further studies revealed that the dissimilar patterns of exocytosis-endocytic coupling at different stimulation frequencies were mediated by two pathways of endocytosis that are orchestrated during differential patterns of neuronal activity.¹ Here, we summarize our observations and further discuss their possible implications.

This article refers to:

At presynaptic nerve terminals, neurotransmission is mediated by the membrane trafficking of synaptic vesicles, a process coordinating the exocytosis of neurotransmitter and the endocytosis of vesicular components. To sustain high-frequency neurotransmitter release, synaptic vesicle proteins must be recaptured rapidly with high fidelity for regenerating release-competent vesicles. At the same time, the newly exocytosed lipid constituents must be quickly removed to prevent expansion of the presynaptic membrane. Two models of synaptic vesicle recycling have been proposed and debated for decades. In the “classical clathrin-mediated endocytosis” model, the vesicle fully collapses onto the plasma membrane before the subsequent endocytosis.2 In the “kiss-run” model, the vesicle briefly connects with the plasma membrane through a transient proteinaceous fusion pore without losing its protein identity.3

The recent study from our group1 aimed to examine exocytosis-endocytosis coupling at different levels of neuronal activity, as well as the possible involvement of the different endocytosis models in this process. Our most surprising observation was that single vesicle fusion at very low stimulation frequencies was followed by two distinct vesicle recoveries, with its protein components systematically distributed into two distinct daughter vesicles. Thus, the tight coupling of exocytosis-endocytosis breaks down during very low neuronal activity, with each endocytosed vesicle recovering half of the exocytosed protein components. Furthermore, our experiments revealed that the daughter vesicles were retrieved by two temporally distinct endocytosis pathways.

Although the time constant for the fast endocytosis (∼2–4 s) observed in our study might bear a resemblance to generally assumed fast time constant for the “kiss-run” mode, we found that protein components of the newly released vesicle were only partially recovered by the fast endocytosis. This observation implies that the “kiss-run” can not be responsible for this form of endocytosis. In addition, the fast endocytosis can be at least partially inhibited by siRNA against clathrin heavy chain, suggesting it is more likely a form of fast clathrin-mediated endocytosis rather than the “kiss-run”. Meanwhile, the slow endocytosis resembles the “classical clathrin-mediated pathway” in both the time course and the nature of partial retrieval of exocytosed proteins. Finally, the clathrin-dependence of both the fast and slow endocytosis in our observation is in agreement with recent reports in hippocampal synapses4 and Drosophila synapses.5

If newly endocytosed vesicles only contain half of protein components, then, further sorting process via intermediate compartments is necessary in order to restore the correct stoichiometry of protein components6 before next rounds of exocytosis. Indeed, several lines of evidence support the notion that endosome is involved in exocytosis-endocytosis cycles. First, Rab5, a GTPase involved in endosomal fusion reactions, is present in synaptic vesicles.6 Second, newly endocytosed vesicles isolated from synaptic terminals can fuse with each other, and with endosomes from PC12 and BHK cells.7 Third, at least a fraction of recently endocytosed vesicles are retrieved by an endocytic pathway mediated by AP3, an adaptor protein involved in the budding process from endosome.8

Interestingly, we found that, as stimulation frequency increased, the very tight coupling between exocytosis and endocytosis was re-established with the number of endocytosed vesicles began to match exocytosis. In the meantime, the slow endocytosis was accelerated by high stimulation and became the predominant pathway. Collectively, endocytosis was able to maintain a relatively steady rate under a certain range of stimulations. The predominant role of the slow pathway at a higher stimulation frequency in our study is in line with the prolonged endocytosis after stronger stimulations, which has been observed from a range of preparations, including ribbon synapses,9,10 frog neuromuscular junctions,11 hippocampal synapses12,13 and calyx of Held.14 In addition, the classical ultrastructural analysis performed by EM revealed substantial increase in coated vesicles and large cisternae in neuromuscular junctions after tetanic stimulation.2 A logical hypothesis for the activity-dependent switch between the fast and slow endocytosis is that fast endocytosis only has a limited capacity.15,16 Therefore, as the capacity of fast endocytosis was quickly saturated during high frequency stimulations, extra proteins must be diverted to slow endocytosis pathway to be retrieved. The main challenge in the future study is to identify the key molecular mechanisms regulating the contributions of the two endocytosis pathways during constantly changing neuronal activity. Based on the results from our study and from a number of recent studies,8,12,17,18 it is rational to propose that, by recruiting and/or regulating diverse adaptors and accessory proteins, the clathrin-based endocytic network can be adapted to different patterns of neuronal activity in a fine-tuning manner.

Addendum to: