The calcium sensor synaptotagmin 7 is required for synaptic facilitation

Abstract

It has been known for more than 70 years that synaptic strength is dynamically regulated in a use-dependent manner. At synapses with a low initial release probability, closely spaced presynaptic action potentials can result in facilitation, a short-term form of enhancement in which each subsequent action potential evokes greater neurotransmitter release. Facilitation can enhance neurotransmitter release considerably and can profoundly influence information transfer across synapses, but the underlying mechanism remains a mystery. One proposed mechanism is that a specialized calcium sensor for facilitation transiently increases the probability of release, and this sensor is distinct from the fast sensors that mediate rapid neurotransmitter release. Yet such a sensor has never been identified, and its very existence has been disputed. Here we show that synaptotagmin 7 (Syt7) is a calcium sensor that is required for facilitation at several central synapses. In Syt7-knockout mice, facilitation is eliminated even though the initial probability of release and the presynaptic residual calcium signals are unaltered. Expression of wild-type Syt7 in presynaptic neurons restored facilitation, whereas expression of a mutated Syt7 with a calcium-insensitive C2A domain did not. By revealing the role of Syt7 in synaptic facilitation, these results resolve a longstanding debate about a widespread form of short-term plasticity, and will enable future studies that may lead to a deeper understanding of the functional importance of facilitation.

Extended Data Figure 1. Possible mechanisms for synaptic facilitation

It is established that calcium plays an important role in synaptic facilitation, and a number of mechanisms have been proposed that involve different aspects of calcium signaling. Here we discuss the calcium signals that evoke rapid vesicle fusion, and also those thought to be involved in facilitation (a), and 3 mechanisms of facilitation are presented schematically (b–d). a, To understand the mechanisms that have been proposed to account for facilitation, it is important to appreciate different aspects of presynaptic calcium signaling. Calcium signals are complex, but can be approximated by 2 components. An action potential opens calcium channels for less than a millisecond, and near open channels the calcium levels reach tens of micromolar. Release sites near calcium channels experience high local calcium levels (Ca_local) that are highly dependent on the distance from open calcium channels. Ca_local can be reduced by high concentrations of fast calcium buffers that rapidly bind calcium. In addition there is a residual calcium signal (Ca_res) that results from calcium equilibrating within presynaptic terminals, before calcium is gradually removed over tens to hundreds of milliseconds. The amplitude of Ca_res (and also total influx of Ca²⁺, Ca_influx) is determined by all of the calcium channels that open, not only those that produce Ca_local that drives release, and after initial equilibration Ca_res is roughly uniform throughout the presynaptic bouton. It is generally accepted that fast synaptic transmission is produced by calcium binding to syt1, syt2 or syt9 which have low-affinity binding sites, fast kinetics, and require the binding of multiple calcium ions^,. The time course of release follows the time course of calcium channel opening, but with a brief delay (< 1 ms). Ca_res after a single stimulus is much smaller than Ca_local. Typical fluorescence-based approaches to measure calcium readily detect Ca_res, but are insensitive to Ca_local which is too localized and short-lived to measure. Note the y-axis is logarithmic to show both Ca_local and Ca_res in (a), but not in (b–d). b, For one mechanism of facilitation a fast calcium buffer is present in presynaptic terminals that binds calcium and reduces Ca_local. Stimulating twice in rapid succession results in the same calcium influx for both stimuli. If there is no fast presynaptic buffer, the amplitudes of Ca_local and the EPSCs are the same for both stimuli (red traces). If a fast high-affinity buffer is present (black traces), it reduces the initial Ca_local and reduces the amplitude of the initial EPSC, but if enough calcium enters and binds to the buffer, it reduces its ability to buffer calcium. As a result the second stimulus produces larger Ca_local than the first, and the EPSC is facilitated. c, A second possible mechanism is that more calcium enters for the second stimulus, and as a result there is more neurotransmitter release. This could arise from a spike broadening, or from the modulation of calcium channels. It is possible that influx through all calcium channels in the presynaptic terminal would be increased, in which case both Ca_res and Ca_local would be increased. It is also possible that the only calcium channels that are modulated are the subset that produce Ca_local that triggers release, in which case Ca_res would not be significantly increased. d, Finally, it is possible that there is a specialized calcium sensor that produces facilitation that is distinct from syt1^,,. Previous studies have shown that such a sensor would need to be sensitive to Ca_res based on the observation that facilitation is altered at some synapses by manipulations that affect Ca_res without affecting Ca_local. According to this scheme, release is mediated by syt1 but calcium binding to a second sensor would increase p. The sensor is sufficiently slow that it does not influence release evoked by the first stimulus, but it able to influence release evoked by a second stimulus.

Extended Data Figure 2. Immunohistochemistry of syt7 expression at 4 different synapses

Fluorescent images of immunostaining for vGlut1 (top) and syt7 (bottom) in slices from WT and syt7 KO animals, showing (a) the stratum radiatum (SR) of hippocampal CA1 region, (b) the ventral thalamus, (c) mossy fibers (MF) in hippocampal CA3, and (d) the lateral and medial performant paths (LPP and MPP) in the outer molecular layer of the dentate gyrus. Notably, syt7 expression in WT animals was higher in the LPP, where synapses exhibit facilitation, compared to the MPP, where synapses exhibit depression. Scale bar, 50 μm. The presence of syt7 labeling in regions containing CA3→CA1 synapses, layer 6 to thalamus synapses, MF synapses and LPP→granule cell synapses that are also colabeled with antibodies to the presynaptic marker for glutamatergic synapses vGlut1, suggests that syt7 is located presynaptically at these synapses. It is, however, difficult to obtain sufficient resolution with confocal microscopy in brain slices to unambiguously establish that syt7 is located presynaptically at these synapses. Importantly, the Allen Brain atlas suggests that the presynaptic cells for these synapses contain mRNA for syt7. Lastly, immunoelectron microscopy revealed selective staining of presynaptic boutons in the CA1 region of the hippocampus.

Extended Data Figure 3. Immunohistochemistry of syt7 and calbindin expression at mossy fiber synapses

Fluorescent images of immunostaining for calbindin-D28k, which predominantly labels mossy fibers in the CA3 region of the hippocampus^, (top) and syt7 (bottom) in slices from WT and syt7 KO animals. Colocalization of syt7 and calbindin staining in WT animals provides further support for the expression of syt7 in mossy fiber terminals. Scale bar, 20 μm.

Extended Data Figure 4. Loss of facilitation in syt7 KO animals at multiple frequencies

Average normalized synaptic responses evoked by extracellular stimulation with trains at frequencies from 5–50 Hz at four synapses in slices from WT and syt7 KO animals. Enhancement during trains was eliminated for all synapses other than mossy fiber synapses, where significant enhancement was present by the 5th stimulus for 5 Hz and 10 Hz, the 3rd stimulus for 20 Hz and the 6th stimulus for 50 Hz (compared to 1 by a Wilcoxon signed rank test, P < 0.05). This indicates that another form of synaptic enhancement gradually builds during repetitive activation and is consistent with a specialized form of synaptic enhancement that has been described at mossy fiber synapses in which spike broadening gradually builds during repetitive activation and leads to increased calcium influx. The numbers of experiments are shown in Extended Data Table 1.

Extended Data Figure 5. Spontaneous release is similar in WT and syt7 KO animals

a, Representative sEPSCs recorded from voltage-clamped hippocampal CA1 cells in WT (black) and KO (red) animals. Vertical scale bars, 20 pA. b, Representative sEPSCs, averaged from >50 events recorded in WT and KO animals. Vertical scale bars, 10 pA. c–d, Average sEPSC (c) amplitude and (d) frequency in WT (N = 16) and syt7 KO animals (N = 18).

Extended Data Figure 6. MK801 blockade of NMDA receptor-mediated EPSCs reveals similar initial release probability in WT and KO synapses

a, Representative NMDA-EPSCs recorded in WT and KO animals before the application of MK801 (average of 10 traces) and after stimulation in the presence of MK801 (average response of 15–20^th stimuli). Vertical scale bars, 100 pA. b, Average NMDA-EPSCs recorded in the presence of MK801, normalized to the first stimulus. c, Half-decay times of NMDA-EPSC amplitudes. *P < 0.05, one-way ANOVA with Tukey’s post hoc test. N.S., not significant. The number of experiments is shown in Extended Data Table 2.

Extended Data Figure 7. Effect of virally-expressed syt7 WT and syt7 C2A* in WT animals

a–b, (Top) AAV was injected into the hippocampal CA3 region in WT animals to express ChR2 and (a) syt7 WT or (b) syt7 C2A*. (Bottom) Representative EPSCs and average paired-pulse ratios for responses evoked electrically and optically in WT slices with AAV-driven expression of (a) syt7 WT (electrical, N = 12; optical, N = 13) and (b) syt7 C2A* (electrical, N = 5; optical, N = 13). Vertical scale bars, 100 pA.

Extended Data Figure 8. Evidence suggests that syt7 does not produce facilitation by acting as a local calcium buffer at the CA3→CA1 synapse

This graph illustrates the general relationship between PPR and external calcium for synapses in which buffer saturation produces facilitation (green) and for facilitation observed at the CA3→CA1 synapse and many other synapses (black). It has been shown previously that the for buffer saturation mechanism (Extended Data Fig. 1B) the amplitude of facilitation is reduced when Ca_influx is reduced by lowering external calcium. This can be understood by considering that this form of facilitation is thought to require sufficient Ca_influx to saturate the endogenous buffer, and thereby reduce its ability to buffer calcium for subsequent stimuli. If Ca_influx is low then there is insufficient calcium entry to bind very much of the endogenous buffer, and little facilitation would result. In addition, as shown in Extended Data Figure 1 for a calcium buffer to produce facilitation it would need to buffer calcium sufficiently that it would reduce initial p. We have shown, however, that p is unaltered in syt7 knockouts. This is perhaps not surprising in light of the fact that syt7 is thought to be located on the plasma membrane, and in cases where this type of facilitation has been observed it is associated with high concentrations of a fast cytosolic buffer.

Figure 1. Facilitation is absent in syt7 KO mice

a–d, Representative traces (top) and average paired-pulse ratios (PPR) at different interstimulus intervals (Δt) (bottom) recorded in slices prepared from WT (black) and syt7 KO animals (red). Postsynaptic responses were recorded using whole-cell voltage clamp from (a) hippocampal CA1 pyramidal cells, and (b) thalamic relay cells. fEPSPs were recorded from (c) hippocampal-mossy-fiber to CA3 synapses, and (d) lateral-performant-path synapses in the dentate gyrus. Vertical scale bars, 100 pA (a,b) and 100 μV (c,d). e–h, Synaptic responses to 20 Hz trains from the same preparations as a–d (top), normalized amplitudes during 20 Hz trains (middle), and normalized responses to the 10^th stimulus as a function of stimulus frequency (bottom). Peak PPR was significantly different for WT and syt7 KO mice at all synapses, as was response₁₀/response₁ for 5 to 50 Hz trains (P < 0.01, Student’s t-test). Data in this and subsequent figures represent mean ± SEM. Number of experiments shown in Extended Data Table 1.

Figure 2. Facilitation is altered in syt7 KO animals despite similar presynaptic Ca²⁺ signals

a, Presynaptic Ca_res evoked by a single stimulus recorded from Schaffer collateral fibers loaded with a low-affinity Ca²⁺ indicator (left), and Ca_res half-decay times (right). N.S., not significant. b, Ca_res signals recorded with low-affinity indicator evoked by 1 or 2 stimuli (left). The ratio of the increase in Ca_res evoked by the first (ΔF₁) and second (ΔF₂) stimuli (right). c, Ca_res signals recorded with high-affinity indicator evoked by 1 or 2 stimuli. d, Average EPSC amplitudes for CA3→CA1 synapses recorded in different external Ca²⁺ (Ca_e) concentrations, normalized to the amplitude in 2 mM Ca_e. e, EPSCs recorded in different Ca_e. Vertical scale bars, 50, 100, 200 and 300 pA in 0.5, 1, 2 and 3 mM Ca_e respectively. f, PPR for interstimulus interval of 20 ms recorded in different Ca_e. In 0.5 mM Ca²⁺ PPR in KOs (1.24 ± 0.12) was not significantly different from 1 (P = 0.084, Wilcoxon signed rank test). Number of experiments is shown in Extended Data Table 2.

Figure 3. Change in the initial probability of release does not underlie the absence of facilitation in syt7 KO mice

a, Extracellular recordings of presynaptic fiber volley and fEPSP evoked by the indicated stimulus intensities. Scale bar, 200 μV. b, fEPSP slope plotted against fiber volley amplitude, for 20–100 μA stimulation. c, fEPSPs recorded in 1 and 3 mM Ca_e. Scale bar, 100 μV. d, Average ratio of the fEPSP to the fiber volley in different Ca_e. e, (Top) Initial release probability was measured by stimulating Schaffer collaterals every 10 seconds while recording NMDA-fEPSPs before and after MK801 bath application. (Middle) Traces averaged from 10 trials before (dark traces), and trials 10–15 after MK801 application (light traces). (Bottom) Average NMDA-fEPSPs amplitudes evoked in the presence of MK801. f, Same as in (e) but with 3 stimuli at 50 Hz every 30 seconds. First response to trains is shown. g, Half-decay times of NMDA-fEPSP amplitudes in the presence of MK801. *P < 0.05, one-way ANOVA with Tukey’s post hoc test. N.S., not significant. Number of experiments shown in Extended Data Table 2.

Figure 4. Viral expression of syt7 restores facilitation at Schaffer collateral synapses

a–d, (Top) Fluorescence images of ChR2-YFP and syt7 immunostaining in the CA1 region after AAV injection into CA3 to express the indicated proteins in WT animals (a) or syt7 KO animals (b–d). Stratum pyramidale (PY) and stratum radiatum (SR). Scale bar, 100 μm. (Bottom) EPSCs and PPRs for responses evoked electrically (open symbols) and optically (blue symbols). In a and b only ChR2-YFP was expressed, in c both ChR2-YFP and syt7(WT) were expressed (separated by a P2A cleavage peptide) and in d ChR2-YFP and Ca²⁺-insensitive syt7C2A* were expressed. e–f, Summary of PPRs for 50 ms interstimulus interval. Asterisks denote significant difference from responses evoked electrically in uninjected WT animals (e), or optically in WT animals expressing ChR2 alone (f). *P < 0.05, one-way ANOVA with Tukey’s post hoc test. Number of experiments shown on bar graphs.