In vitro fusion of single synaptic and dense core vesicles reproduces key physiological properties

Abstract

Regulated exocytosis of synaptic vesicles is substantially faster than of endocrine dense core vesicles despite similar molecular machineries. The reasons for this difference are unknown and could be due to different regulatory proteins, different spatial arrangements, different vesicle sizes, or other factors. To address these questions, we take a reconstitution approach and compare regulated SNARE-mediated fusion of purified synaptic and dense core chromaffin and insulin vesicles using a single vesicle-supported membrane fusion assay. In all cases, Munc18 and complexin are required to restrict fusion in the absence of calcium. Calcium triggers fusion of all docked vesicles. Munc13 (C1C2MUN domain) is required for synaptic and enhanced insulin vesicle fusion, but not for chromaffin vesicles, correlating inversely with the presence of CAPS protein on purified vesicles. Striking disparities in calcium-triggered fusion rates are observed, increasing with curvature with time constants 0.23 s (synaptic vesicles), 3.3 s (chromaffin vesicles), and 9.1 s (insulin vesicles) and correlating with rate differences in cells.

Fig. 1

Characteristic fusion events of purified secretory vesicles with planar-supported membranes. The intensity traces of the peak intensity pixel of a single secretory vesicle docking and fusing with the planar-supported bilayer containing t-SNAREs are shown for a dense-core vesicles labeled with NPY-mRuby, b insulin vesicles labeled with C-peptide-GFP, and c synaptic vesicles labeled with acridine orange. The single events are representative examples with more example events shown in Supplementary Fig. 2. d Intensity traces from a double-labeled (acridine orange, red; NPY-Ruby, black) dense-core vesicle illustrates that during fusion, egress of acridine orange precedes NPY-Ruby by ~200 ms. Intensity traces from insulin vesicles double-labeled with e C-peptide-GFP and NPY-mRuby or with f NPY-mRuby and acridine orange show that C-peptide-GFP and NPY-mRuby have the same characteristic release behavior while acridine orange starts to be released about ~200 ms before the onset of release of the fluorescent proteins. g Schematic drawing of the theoretical intensity traces for single secretory vesicles labeled with a fluorescent protein (black) or acridine orange (red). Intensity traces are of a region of interest over time. I. Initially the labeled secretory vesicle is not within the TIRF field and no fluorescence is observed. II. The vesicle docks to the planar-supported membrane causing an immediate increase of fluorescence intensity in the TIRF field. III. After some time, the vesicle fuses with the planar membrane. The onset of fusion begins with the opening of a small fusion pore resulting in an increase in fluorescence of the small molecule dye acridine orange or a decrease in fluorescence of the fluorescent protein. The decrease in observed fluorescence is due to lateral diffusion of the fluorescent indicator away from the site of fusion. When fusion is observed with a fluorescent protein, the protein is first more slowly released through the initial narrow fusion pore until, at time point IV, the vesicle collapses into the planar membrane as observed by an increase in fluorescence due to the forward movement of the protein within the TIRF field following a continued decrease in intensity as the protein continues to leave the fusion site

Fig. 2

Purified secretory vesicles contain functional V-ATPase acidification of a synaptic vesicles, b dense-core vesicles, or c insulin vesicles. Bulk acridine orange fluorescence was monitored in the presence of the secretory vesicle and valinomycin (black line). Addition of ATP drives progressive self-quenching of the acridine orange fluorescence due to concentration into the interior of the secretory vesicle. The addition of CCCP dissipates the pH gradient causing an increase in acridine orange fluorescence. The presence of bafilomycin inhibits the proton pump preventing the acidification of the secretory vesicles (blue traces)

Fig. 3

Secretory vesicle docking and calcium-triggered fusion. a Docking of synaptic vesicles (green), dense-core vesicles (red), and insulin vesicles (blue depends on the presence of both target SNARE proteins syntaxin-1a and SNAP-25 in the planar-supported membrane in the absence of calcium (100 µM EDTA)). Syntaxin-1a (183–288) denotes a construct containing the transmembrane domain and the SNARE motif, whereas syntaxin-1a(1-288) is the full-length protein that also contains the regulatory H_abc domain. Secretory vesicle docking to planar-supported lipid bilayers containing different combinations of syntaxin-1a1-288, dSNAP-25, complexin-1, and Munc18. A summary of the statistics is shown in Supplementary Table 1. b Single-vesicle fusion events triggered by calcium from a docked state in the presence of syntaxin-1a1-288, dSNAP-25, complexin-1, Munc18, and the C1C2MUN domain of Munc13. Single events for synaptic vesicles, dense-core vesicles, and insulin vesicles are shown. Soluble Alexa647 was included in the buffer containing 100 μM calcium to measure calcium arrival (black trace) to determine delay times of fusion. Orange line is the precise time determined for calcium arrival. The traces represent the release of respective fluorescent content markers of acridine orange for synaptic vesicles (green), NPY-mRuby for dense-core vesicles (red), and C-peptide-GFP (blue) for insulin vesicles. Graphs with expanded times scales around the arrival time of the calcium buffer are shown in Supplementary Fig. 4 and more representative traces are shown in Supplementary Fig. 3. Error bars represent standard errors of the mean

Fig. 4

Munc13 dependence of secretory vesicle kinetics. Cumulative distribution functions for synaptic vesicles (a), dense-core vesicles (b), insulin vesicles (c), dense-core vesicles lacking endogenous CAPS-1/2 (previously described in ref. ) (d). All kinetics are triggered from a docked state in the presence of syntaxin-1a:dSNAP-25, Munc18, complexin-1 without (black), or with (purple) 0.2 μM of the C1C2MUN domain fragment of Munc13. Summary statistics for each condition are shown in Supplementary Table 2. e Bar graph for the total amount of fusion in the presence of 100 µM EDTA, 100 µM calcium, or 100 µM calcium in the presence of 0.2 µM C1C2MUN domain of Munc13 for each condition. Error bars represent standard errors of the mean

Fig. 5

Secretory vesicle kinetics have a vast range of fusion rates that depend on vesicle size. a Cumulative distribution functions of the delay times of fusion from the time of calcium arrival (Δt_Ca2+) for the three secretory vesicle types. Fusion of vesicles is triggered from a docked state in the presence of syntaxin-1a:dSNAP-25, Munc18, complexin-1, and the C1C2MUN domain of Munc13. Summary statistics for each condition are shown in Supplementary Table 2. b Rate constants of dense-core vesicles purified from PC12 cell lines depleted of endogenous synaptotagmins and overexpressing single isoforms of synaptotagmin-1, -7, or -9. Vesicles were docked in the presence of syntaxin-1a:dSNAP-25, Munc18, and complexin-1 and triggered to fuse with 100 µM calcium. Cumulative distribution functions for synaptotagmin knock-ins are shown in Supplementary Fig. 6. c Docking of all three secretory vesicles is inhibited by a soluble target membrane Q-SNARE complex (syntaxin-1a(191–253):SNAP-25) in the presence of Munc18 and complexin-1 and in the absence of calcium. d The rate constant of fusion of all three secretory vesicle types in the absence or presence of 0.5 µM syntaxin-1a(191–253):SNAP-25 is unchanged for calcium-triggered fusion with planar-supported membranes containing syntaxin-1a:dSNAP-25 in the presence of Munc18, complexin-1, and the C1C2MUN domain of Munc13. e Vesicle diameter distributions of purified secretory vesicles as measured by cryo-electron microscopy. Example images are shown on top with scale bars being 100 nm. f Fitted fusion rate constants (k (s⁻¹)) from a plotted as a function of the mean curvature (1/R), calculated from e. The fitted black line assumes that the activation energy depends linearly on the curvature of the fusing vesicles. g To-scale model showing the size differences of synaptic vesicles (green), dense-core vesicles (red), and insulin vesicles (blue) docked to a planar target membrane. Error bars represent standard errors of the mean