Dynamin 1 mediates fission of endocytic synaptic vesicles in the brain and has two major splice variants, Dyn1xA and Dyn1xB, which are nearly identical apart from the extended C-terminal region of Dyn1xA. Despite a similar set of binding partners, only Dyn1xA is enriched at endocytic zones and accelerates vesicle fission during ultrafast endocytosis. Here, we report that Dyn1xA achieves this localization by preferentially binding to Endophilin A1 through a newly defined binding site within its long C-terminal tail extension. Endophilin A1 binds this site at higher affinity than the previously reported site, and the affinity is determined by amino acids within the Dyn1xA tail but outside the binding site. This interaction is regulated by the phosphorylation state of two serine residues specific to the Dyn1xA variant. Dyn1xA and Endophilin A1 colocalize in patches near the active zone, and mutations disrupting Endophilin A binding to the long tail cause Dyn1xA mislocalization and stalled endocytic pits on the plasma membrane during ultrafast endocytosis. Together, these data suggest that the specificity for ultrafast endocytosis is defined by the phosphorylation-regulated interaction of Endophilin A1 with the C-terminal extension of Dyn1xA.

Neurotransmitter is released from dedicated sites of synaptic vesicle fusion within a synapse. Following fusion, the vacated sites are replenished immediately by new vesicles for subsequent neurotransmission. These replacement vesicles are assumed to be located near release sites and used by chance. Here, we find that replacement vesicles are clustered around this region by Intersectin-1. Specifically, Intersectin-1 forms dynamic molecular condensates with Endophilin A1 near release sites and sequesters vesicles around this region. In the absence of Intersectin-1, vesicles within 20 nm of the plasma membrane are reduced, and consequently, vacated sites cannot be replenished rapidly, leading to depression of synaptic transmission. Similarly, mutations in Intersectin-1 that disrupt Endophilin A1 binding result in similar phenotypes. However, in the absence of Endophilin, this replacement pool of vesicles is available but cannot be accessed, suggesting that Endophilin A1 is needed to mobilize these vesicles. Thus, our work describes a distinct physical region within a synapse where replacement vesicles are harbored for release site replenishment.

Compensatory endocytosis keeps the surface area of secretory cells constant following exocytosis. At chemical synapses, clathrin-independent ultrafast endocytosis maintains such homeostasis. This endocytic pathway is temporally and spatially coupled to exocytosis, initiating within 50 ms at the region immediately next to where vesicles fuse: the active zone. How synaptic vesicle exocytosis induces ultrafast endocytosis is unknown. Here, we demonstrate that actin filaments are enriched in the region surrounding active zone at mouse hippocampal synapses and that the membrane area conservation due to this actin corral is necessary for exo-endocytic coupling. Simulations suggest that flattening of fused vesicles exerts lateral membrane pressure in the plasma membrane against the actin corral, resulting in rapid formation of endocytic pits at the border between the active zone and the surrounding actin-enriched region. Consistent with our simulations, ultrafast endocytosis does not initiate when actin organization is disrupted, either pharmacologically or by ablation of the actin-binding protein Epsin1. These data suggest that endocytosis is mechanically coupled to exocytosis at synapses.

As synaptic vesicles fuse, they must continually be replaced with new docked, fusion-competent vesicles to sustain neurotransmission. It has long been appreciated that vesicles are recruited to docking sites in an activity-dependent manner. However, once entering the sites, vesicles were thought to be stably docked, awaiting calcium signals. Based on recent data from electrophysiology, electron microscopy, biochemistry, and computer simulations, a picture emerges in which vesicles can rapidly and reversibly transit between docking and undocking during activity. This "transient docking" can account for many aspects of synaptic physiology. In this review, we cover recent evidence for transient docking, physiological processes at the synapse that it may support, and progress on the underlying mechanisms. We also discuss an open question: what determines for how long and whether vesicles stay docked, or eventually undock?

Nano-resolution fluorescence electron microscopy (nano-fEM) provides the precise localization of biomacromolecules within electron micrographs. Classically, electron microscopy has provided the highest possible cellular detail, boasting nanometer-scale resolution. However, while cellular ultrastructure is clearly defined, molecular identity is obscured even when electron dense tags in the form of antibodies or locally polymerized moieties are used. Fluorescence microscopy complements electron microscopy by providing significant molecular specificity. Further, super-resolution techniques surpass the diffraction limit and localize labelled proteins at ~20 nm resolution. However, sparse light-emitting points do little to provide the subcellular context of labeled molecules. In nano-fEM, fluorescently tagged biological samples are first high-pressure frozen and processed via freeze substitution for fixation and to preserve fluorescence. Afterwards, samples are embedded in hydrophilic resin, cut into ultrathin sections, and visualized by, for example, direct stochastic optical reconstruction microscopy (dSTORM) followed by transmission electron microscopy. Fluorescence and electron micrographs are correlated by use of fiduciary markers and post-processing. This approach also provides 3D information similar to Array Tomography by serial sectioning of ultrathin sections followed by super-resolution microscopy and electron microscopy of each section in a sequential manner, enabling 3D reconstruction of the z axis.