A Vesicle Superpool Spans Multiple Presynaptic Terminals in Hippocampal Neurons

Summary Synapse-specific vesicle pools have been widely characterized at central terminals. Here, we demonstrate a vesicle pool that is not confined to a synapse but spans multiple terminals. Using fluorescence imaging, correlative electron microscopy, and modeling of vesicle dynamics, we show that some recycling pool vesicles at synapses form part of a larger vesicle “superpool.” The vesicles within this superpool are highly mobile and are rapidly exchanged between terminals (turnover: ∼4% of total pool/min), significantly changing vesicular composition at synapses over time. In acute hippocampal slices we show that the mobile vesicle pool is also a feature of native brain tissue. We also demonstrate that superpool vesicles are available to synapses during stimulation, providing an extension of the classical recycling pool. Experiments using focal BDNF application suggest the involvement of a local TrkB-receptor-dependent mechanism for synapse-specific regulation of presynaptic vesicle pools through control of vesicle release and capture to or from the extrasynaptic pool.


Supplemental Experimental Procedures Preparations
Dissociated hippocampal cultures were prepared from P0 rats as described previously (Darcy et al., 2006a;Morales et al., 2000) and used for experiments at 12-18 days in vitro (DIV). Neurons were transfected at DIV8-9, using a Ca 2+ phosphate protocol.

Labelling and imaging in cultured neurons
For cultured neurons, recycling synaptic vesicles were labelled using field stimulation (600 APs 10 Hz) in the presence of the dye (FM4-64 or a fixable analog of FM1-43, FM1-43FX, 10 μM, Molecular Probes, see below). After loading, cells were washed for 1 min in Advasep-7 (1 mM, Biotium Inc.) and subsequently rinsed with fresh bath solution for 10 min to remove surface dye. Plasmid encoding SypI-Dendra2 was constructed by replacing the GFP sequence in SypI-EGFP construct (Darcy et al., 2006a)  Quantification of fluorescence was carried out using ImageJ (NIH) on raw unfiltered images or after filtering (1 x 1 median filter) applied to the whole image.

FM-dye imaging in acute slices
Synapses were fluorescently labeled by washing slices into modified saline containing 40 mM KCl and 10 µM FM1-43 (45 s) following a preincubation in ACSF containing 10 µM FM1-43 (1 min). After loading, slices were rinsed again in the preincubation solution to allow completion of endocytosis. Surface FM-dye was removed by washing the slice into ACSF containing 1 mM Advasep-7, followed by continuous perfusion in fresh ACSF solution (30 min). Two-photon imaging was carried out with a Ti:sapphire pulsed laser (MaiTai, Spectra-Physics) tuned to 900 nm on an Olympus BX51WI microscope (60x objective, 0.9 NA; Olympus, Melville, NY). Synapses were imaged at a depth of approximately 50 -100 µm from the top surface of the slice in region CA1. Time lapse frames were taken every 20 s, each one generated from a maximal projection of five zsections (0.5 µm separation) in turn averaged from three separate images. For quantification of vesicle flux through synapses we generated a cumulative fluorescence change plot for synaptic terminals imaged in timelapse sequences, subtracting baseline imaging noise measured using a pollen grain imaged with the same acquisition settings.
All data was acquired using custom written software in Matlab 7.2 (Mathworks).
Quantification of fluorescence was carried out using ImageJ (NIH) on raw unfiltered images or after filtering (1 x 1 median filter) applied to the whole image.

Modeling
For the data shown in Figure S5, a total of 20 synaptic and 20 extrasynaptic compartments were simulated at ten second iterations. For simplicity, synaptic pools corresponded to the mobile fraction of the recycling pool, and non-mobile recycling vesicles and reserve pool vesicles were not simulated. For each iteration, vesicles were randomly moved: 1) from synaptic to the adjacent extrasynaptic pools, 2) from extrasynaptic to the adjacent synaptic pools and 3) between extrasynaptic pools. The direction of movement was randomized for each vesicle. The initial size of individual synaptic pools was 150 vesicles, and probabilities for 1 and 2 were taken from the experimentally measured rates of vesicle gain and loss. Before each experimental setting, the system was allowed to reach steady state, yielding an average synaptic pool size of 100, and an average extrasynaptic pool size of 50 vesicles. To estimate the rate of vesicle movement between extrasynaptic pools, the probability (p) of this event was gradually increased from 0 to 1, and the best fit to the data in Figure 2G selected (see Figure S5A, best fit was with p = 0.5 -0.6, average of 50 trials). This rate was then used to produce the data in Figure S5B,C. All simulations were performed in Matlab 7.2 (Mathworks).

Electrophysiology
For the data shown in Figure S5D,E, whole-cell recordings were established as described in Branco et. al (2008), and standard extracellular solution containing FM1-43 (10 µM), CNQX (20 µM) and APV (50 µM) was perfused for ~2 min. Dye loading was started after 30 s with 50 APs at 1 Hz in voltage-recording mode, and dye-free solution was re-perfused 30 s after the end of the stimulation.

Defining the number of synapses available to the superpool
Single neurons were FM-dye labelled via stimulation (50 APs, 1 Hz) through a patch pipette, which also contained Alexa 594. In this way synaptic loading was restricted to the target neuron and the axonal processes could be directly visualized at the same time to count the synapses.

Supplemental References
Branco, T., Staras, K., Darcy, K.J., and Goda, Y.       Figure 2G, for three vesicle exchange rates between extra-synaptic pools (coloured bands show mean ± SD, pink: p cross = 0.5, brown: p cross = 0.1, grey: p cross = 0). Note pink and brown bands overlap. Circles and error bars are experimentally measured data points for vesicle sharing from the EM experiments ( Figure 2G) normalized to the first synapse. (B) Lateral mobility of a single vesicle derived from a source synapse (50 trials) based on the experimentally-derived parameters. Over a period of 1 h, vesicles can reach and become integrated in any of the neighbouring 19 synapses. (C) Histograms revealing the change in vesicular composition over time for the mobile fraction of recycling vesicles. (Top) middle five synapses of a row of 20 terminals along an axon. (Bottom) plots of vesicles contributed from the synaptic terminals 1 to 20 for each of the five coloured synapses over time. At time 0 (top row), each synapse is filled only with its own vesicles (coloured bars) but over time, vesicles are contributed by synaptic neighbours (grey bars) until the composition of the synapse is substantially changed (bottom row). For clarity, each plot is normalized to the maximum vesicle count. (D) Estimating superpool size by counting synapses. Example image showing overlay of Alexa 594 dye fill (white) and FM-dye synaptic label (red) in a neuron used for these measurements. FM-dye labelling was achieved via stimulation (50 APs, 1 Hz) through a patch pipette, which also contained Alexa. In this way synaptic loading was restricted to the target neuron and the axonal processes could be directly visualized at the same time. Scale bar, 5 µm. (E) Regions of neuron from D (rectangles).