A Mechanism Intrinsic to the Vesicle Fusion Machinery Determines Fast and Slow Transmitter Release at a Large CNS Synapse

EPSC deconvolution and presynaptic membrane capacitance both report a fast and a slow component of transmitter release. A1, Presynaptic Ca²⁺ currents. A2, EPSCs. A3, Transmitter release rates (dotted traces) and integrated release rates (continuous traces for a series of presynaptic depolarizations to 0 mV of different lengths). The experiment was first done in the presence of CTZ (black traces) and subsequently in the additional presence of 2 mm γ-DGG (red traces). The EPSCs in response to 4 and 16 ms depolarizations are also superimposed onto the EPSC evoked by a 64 ms depolarization (see asterisks and gray traces in A2). Note the slowing of the EPSC decay when longer presynaptic depolarizations were applied (see arrowhead in A2). B, Presynaptic membrane capacitance traces, for the same experiment as shown in A. The superimposed pink and gray bars indicate the time windows (30 ms) during which the poststimulus capacitance values were measured. C, Integrated release rate trace as obtained by EPSC deconvolution in response to a 64 ms depolarization. The trace was fitted by a double-exponential + line function (see green fit line), and the fast- and slow components of the fit are drawn separately (dotted and dashed line, respectively). D, Plot of the capacitance jumps (filled symbols) and cumulative transmitter release from EPSC deconvolution (open symbols) as a function of the length of the depolarization, in the absence (black symbols) or presence (red symbols) of 2 mm γ-DGG. This is the same cell as shown in A. Note that both the capacitance jumps and the cumulative release obtained from EPSC deconvolution indicate a fast and a slow component of release. Both datasets were fitted with double-exponential functions (continuous line, deconvolution; dotted line, capacitance). E, Average cumulative release normalized to the value at 8 ms (n = 4 cells). The capacitance jumps for the 32 and 64 ms depolarizations were significantly larger than the ones evoked by the 8 ms depolarization. (**p < 0.01), whereas the difference between 16 and 8 ms depolarizations was less significant (*p = 0.05).

Presynaptic Ca²⁺ uncaging evokes a fast and a slow release component similar as long presynaptic depolarizations. A, Three different Ca²⁺ stimuli were applied in the same presynaptic terminal: a Ca²⁺-uncaging stimulus that elevated [Ca²⁺]_i to 8 μm (black traces) and a short (8 ms) and a long (50 ms) presynaptic depolarization to 0 mV (blue and red traces, respectively). B, EPSCs induced by the presynaptic stimuli shown in A, with identical color code. In the right, the flash-evoked EPSC is shown, together with the spillover current estimated during the EPSC deconvolution analysis (see Materials and Methods). C, Release rates resulting from the deconvolution of the EPSCs shown in B. In the right, the release rate in response to the flash is shown again, together with a line fit (green line) that was used to extrapolate the peak release rate of the slow component (see Materials and Methods). D, Integrated release rate traces. The cumulative release rate trace obtained after Ca²⁺ uncaging (black trace) was fitted with a double-exponential function (right, green fit line underlying the data trace), giving fast and slow time constants (and amplitudes) of 3.2 ms (812 vesicles) and 28 ms (994 vesicles). The fast and slow exponential components of the fits are drawn separately (dashed black lines, right). The linear component of the fit is displayed as well (dotted line). Note the similarity of the cumulative release for the Ca²⁺-uncaging stimulus and for the long depolarization (black and red trace, respectively; left).

Submaximal release of fast releasable vesicles. A, Ca²⁺-uncaging stimuli of different intensities produced a series of presynaptic [Ca²⁺]_i elevations. B, EPSCs. C, Transmitter release rates. D, Cumulative release, all in response to the [Ca²⁺]_i elevations shown in A. Traces with matched colors refer to the same flash. The integrated release rate traces (D) were fitted with double-exponential or double-exponential + line functions, and the fast exponential components of these fits are shown (dashed lines). Note that Ca²⁺ steps of increasing amplitudes lead to a marked increase of the fast release component.

Differential regulation by Ca²⁺ of the number of fast and slowly released vesicles. A1, [Ca²⁺]_i was increased to different levels by Ca²⁺ uncaging (flash at arrowhead), and this was followed after 100 ms by a 50 ms depolarization to 0 mV (horizontal gray bar). [Ca²⁺]_i traces are shown in micromolar (top), as well as normalized to their peak values (bottom). Within the first 50 ms, [Ca²⁺]_i decayed to 88% of its peak value in this cell. A2, The corresponding EPSCs (top) and cumulative release rates (bottom), in the same color code as the traces shown in A1. The dotted lines in the bottom represent the fast components of double-exponential fits to the cumulative release traces. Note that the fast component markedly increased with Ca²⁺ steps of increasing amplitudes. The numbers in brackets indicate the temporal sequence of Ca²⁺-uncaging stimuli. B, The number of vesicles released in the fast and slow component, as well as the summed release (cross symbols), and the total release after the 50 ms depolarization is plotted as a function of [Ca²⁺]_i, for the same cell pair as shown in A. Note that the number of vesicles released in the fast component (filled squares) increases on the expense of the number of slowly released vesicles (open squares), whereas the total release after the flash and the 50 ms depolarization (round symbols) was nearly constant. C, The amplitude of the fast and slow release components were normalized to their values in a range of 10–15 μm [Ca²⁺]_i and plotted as a function of [Ca²⁺]_i. Each color represents data from one cell pair (n = 9 pairs); black symbols represent the same cell pair as shown in A and B. The fast component clearly increases with [Ca²⁺]_i steps of increasing amplitudes, whereas the slow component shows a tendency toward smaller amplitude, more than ∼7 μm [Ca²⁺]_i (top and bottom, respectively). D, The data shown in C were averaged for the indicated four ranges of [Ca²⁺]_i. This revealed a significant increase of the fast component with [Ca²⁺]_i (top) and a concomitant decrease of the slow component (middle). The bottom shows the summed release evoked by Ca²⁺ uncaging, obtained by summing the amplitude values of the fast and the slow exponential component. E, Plot of the total release after the 50 ms depolarizations, for the same four binned [Ca²⁺]_i ranges as shown in D.

Intracellular Ca²⁺ dependence of the fast and the slow component of transmitter release. Time constants of release (A), peak release rates (B), and amplitudes of the fast (C) and the slow (D) component of release are plotted against the amplitude of the [Ca²⁺]_i steps produced by Ca²⁺ uncaging. A shows the time constants estimated from the exponential fit analysis, in which various combinations of exponential and line functions were fitted to integrated release rate traces (see Materials and Methods). Each black symbol is chosen according to the fit function that described the data best (open squares, line; open circles, single exponential; open diamonds, single exponential + line). Note that, above ∼4 μm [Ca²⁺]_i, most cumulative release traces were best fitted by double-exponential or double-exponential + line functions, and the corresponding parameters obtained from the fast and the slow exponential component are shown as filled and open triangles, respectively. In A–D, the logarithmized datasets were fitted with line functions (see red and pink lines), giving slope values as indicated. The data in the range of 10–15 μm [Ca²⁺]_i are shown additionally as red average data points. The predictions of the models shown in E and F are also shown (symbols connected by lines, in colors). Predictions obtained from single-exponential fits of the simulated release traces are shown by open circles, and those obtained by double-exponential and double-exponential + line fits are shown by triangles. E, F, [Ca²⁺]_i waveforms (top) used for the model calculations of the cumulative release rates (bottom), according to a one-pool model (E) or a two-pool model (F). G, The allosteric model of Ca²⁺ binding and vesicle fusion used for the simulations. The state “F” represents the fused state. For the parameter set and for details on the simulations, see Materials and Methods.