Synaptotagmin-7 Enhances Facilitation of Cav2.1 Calcium Channels

Visual Abstract


Introduction
Inward Ca 21 currents conducted by voltage-gated Ca 21 (Ca v ) channels couple action potentials and other depolarizing stimuli to many Ca 21 -dependent intracellular processes, including neurotransmission, hormone secretion, and muscle contraction (Zamponi et al., 2015). In presynaptic nerve terminals, Ca v 2.1, Ca v 2.2, and Ca v 2.3 channels conduct P/Q-type, N-type, and R-type Ca 21 currents that trigger rapid neurotransmission (for review, see Olivera et al., 1994;Zamponi et al., 2015;. However, only P/Q-type Ca 21 currents conducted by Ca v 2.1 channels can mediate short-term synaptic facilitation at the calyx of Held in mice (Inchauspe et al., 2004), pointing to a unique role of these Ca 21 channels in short-term synaptic plasticity.
In transfected nonneuronal cells, Ca 21 entry mediated by Ca v 2.1 channels causes calcium-dependent facilitation (CDF) and inactivation (CDI) during single depolarizations and in trains of repetitive depolarizing pulses (Lee et al., 1999(Lee et al., , 2000DeMaria et al., 2001;Catterall and Few, 2008;Christel and Lee, 2012;Ben-Johny and Yue, 2014). Both CDF and CDI of Ca v 2.1 channels are dependent on calmodulin (CaM; Lee et al., 1999Lee et al., , 2000DeMaria et al., 2001). CaM preassociates with the C-terminal domain of the pore-forming a1 subunit of Ca v 2.1 channels (Erickson et al., 2001). Following Ca 21 binding, CaM initially interacts with the nearby IQ-like motif (IM) and causes CDF, whereas further binding of Ca 21 /CaM to the more distal CaM-binding domain (CBD) induces CDI of Ca v 2.1 channels (DeMaria et al., 2001;Lee et al., 2003). Introducing the IM-AA mutation into the IQ-like motif of Ca v 2.1 impairs CDF and CDI, providing a tool to assess the significance of these processes in synaptic transmission and short-term synaptic plasticity (Zühlke et al., 1999;DeMaria et al., 2001;Lee et al., 2003).
CDF of Ca v 2.1 channels contributes significantly to short-term synaptic facilitation. Expression of Ca v 2.1 in cultured superior cervical ganglion neurons, whose endogenous Ca v 2.2 channels were specifically blocked by v -conotoxin GVIA, was sufficient to restore synaptic transmission and induce Ca 21 -dependent synaptic facilitation, which was impaired by introducing the IM-AA mutation in Ca v 2.1 channels (Mochida et al., 2003a(Mochida et al., , 2008. In mice in which Ca v 2.1 channels contained the IM-AA mutation, synaptic facilitation was substantially decreased at the neuromuscular junction as well as in hippocampal CA3-to-CA1 synapses and CA3-to-parvalbumin-expressing basket cell synapses (Nanou et al., 2016a(Nanou et al., ,b, 2018. These results support an important role for facilitation of Ca v 2.1 channels in short-term synaptic facilitation. In addition to Ca v 2.1 channels, the high-sensitivity Ca 21 sensor synaptotagmin-7 (Syt-7) has been proposed to support short-term synaptic facilitation by binding residual Ca 21 in the nerve terminal following the action potential, thereby increasing interaction with the SNARE complex and enhancing Ca 21 -dependent synaptic vesicle exocytosis (Jackman et al., 2016). Previous studies using Syt-7 KO mice have shown that Syt-7 is required for short-term plasticity in several types of synapses in the hippocampus, cerebral cortex, and cerebellum (Jackman et al., 2016;Turecek and Regehr, 2018). Because Ca v 2.1 channels and Syt-7 are located near each other in the active zones of nerve terminals (Müller et al., 2010), and Ca v 2.1 and Syt-7 are both implicated in synaptic facilitation, we have tested the hypothesis that these two proteins interact directly with each other and regulate Ca 21 entry through Ca v 2.1 channels. Our results reveal direct interactions of Syt-7 with Ca v 2.1 that enhance facilitation of the Ca v 2.1 Ca 21 current. These data suggest that interaction of Syt-7 with Ca v 2.1 channels may contribute to short-term synaptic facilitation.

Cell lines and transfection
Cells from tsA-201 cell line were maintained in DMEM (Invitrogen by Life Technologies) supplemented with 10% fetal bovine serum (Fisher Scientific), 1% glutamine (Sigma-Aldrich), 1% penicillin and streptomycin (Sigma-Aldrich). The cells were maintained at 37°C under 5% CO 2 . Cells were plated in 35-mm tissue culture dishes to achieve 70% confluency and then transfected using TransIT-LT1 transfection reagent (Mirus) with a total of 5-mg plasmid including: 2, 1.5, 1 mg of a 1A , b 2A , and a 2 d subunits composing the Ca v 2.1 channel and a ratio 3 ml of transfection reagent to 1 mg of cDNA plasmid. 0.22 mg eGFP was added to the plasmid mix to identify the transfected cells.

Construction and expression of fusion proteins
Recombinant glutathione S-transferase (GST)-Syt-7a fusion proteins were synthesize from the expression plasmid in the vector pGEX-2T. His-fusion proteins containing the synprint site region from the intracellular loop between domain II and III of the P/Q-type Ca v 2.1 (synprint 724-981) or the equivalent synprint site from the Ltype Ca v 1.2 (680-800) used as a control, were expressed using the expression plasmid pET-28b. GST and His recombinant proteins were expressed in Escherichia coli BL26 cells, a protease-deficient strain (NEB). Fusion proteins were extracted by mild sonication (10 times 10 s with 1-min break) in lysis buffer containing: Tris 50 mM (pH 7.4), NaCl 150 mM, Na-deoxycholate 1%, NaF 10 mM, EDTA 1 mM, Triton X-100 1%, and glycerol 5%, supplemented with protease inhibitors Calpain I, Calpain II, and cOmplete protease inhibitor cocktail (Sigma-Aldrich). GST-Syt-7a, proteins were purified using glutathione Sepharose beads (Millipore Sigma) and eluted with 15 mM reduced glutathione (GSH) in 50 mM Tris (pH). His-synprints from Ca v 2.1 and Ca v 1.2 were purified by binding to Ni 21 -charged HisPur Ni-NTA Resin (ThermoFisher) and eluted with 250 and 500 mM imidazole in PBS. The amount of proteins used was standardized based on Coomassie Blue-stained SDS gels or estimated with a standard curve relating the intensity of the immunoblotting signal to the amount of a standard fusion protein applied.

Co-immunoprecipitation experiments
Immunoprecipitation experiments were performed using Dynabeads Protein G (Invitrogen) in TBS buffer with a Ca 21 -buffering system containing 50 mM Tris/HCl, 140 mM NaCl, 50 mM HEPES (pH 7.2), 5 mM N-(2-hydroxyethyl)ethylenediamine-N,N9,N9-triacetic acid (HydroxyEDTA), 0.3% Triton X-100 and different Ca 21 concentrations varying from 10 mM to 5 mM. The Ca 21 -buffering system Figure 1. Direct binding of Syt-7a to the synprint site of P/Q-type Ca v 2.1 channels. A, B, In brain lysates (A) and transfected tsA-201 cells (B), co-immunoprecipitation experiments show the binding of Syt-7a and a 1A subunit of Ca v 2.1 channels. C, Binding of Syt-7a (GST-Syt-7a) to the His-tagged synprint site (724-981) of a 1A subunit of P/Q-type Ca v 2.1 channel. GST-Syt-7a were immobilized on glutathione-Sepharose beads and incubated with His-Ca v 2.1 or His-Ca v 1.2 using different Ca 21 concentrations (10 mM to 1 mM). The binding experiments were performed in a Ca 21 buffering system containing: 5 mM N-hydroxyethyl ethylenediaminetriacetic, 50 mM HEPES (pH 7.2), and 150 mM NaCl. The free Ca 21 concentrations were estimated using the MAX CHELATOR software. The Ca 21 concentrations used were 10 mM, 50 mM, 100 mM, 300 mM, 500 mM, 700 mM, and 1 mM. After extensive washing, GST-Syt-7a bound to the beads were eluted with 15 mM reduced glutathione (GSH) in 50 mM Tris-HCl (pH 8) and proceed to electrophoresis and immunoblotting. The bound His-Ca v 2.1 synprint (724-981) proteins were detected with anti-His antibody. Because different Ca 21 concentrations were used in co-immunoprecipitation experiments, segments from different immunoblots were spliced together to show comparisons clearly. Those protein bands are delineated for clarification. The immunoblots presented here are representative of at least three experiments for each co-immunoprecipitation or immunoblot. A related co-immunoprecipitation experiment conducted with different experimental conditions is presented in Extended Data Figure 1-1. was used to produce free Ca 21 concentrations calculated using MAX CHELATOR software (UC Davis). Dynabeads were incubated with antibodies directed against Ca v 2.1 channels or Syt-7a for 1 h at 4°C. Then, whole brain lysates or transfected tsA cell lysates were added to the beads and incubated at 4°C under rotation overnight. Nonspecific proteins were washed three times with a washing buffer. Proteins attached to the beads were eluted using an elution buffer. Proteins were blotted with antibodies against Syt-7 (mouse monoclonal antibody N275/14, Product Number MABN665, Millipore Sigma) or Ca v 2.1 (rabbit polyclonal antibody catalog #ACC-001, Alomone Labs). The antibodies used for immunoblotting were titrated to assure that the concentration used was in the linear response range. The co-immunoprecipitation experiments and western blots have been repeated at least three times showing reproducible results.
Study of Syt-7 binding to the synprint site by affinity chromatography GST-Syt-7a proteins were bound to glutathione-Sepharose beads (Millipore Sigma) in TBS-Ca 21 buffer incubated at 4°C for 1 h under constant rotation. To remove unbound proteins, the mixture was washed two times with a washing buffer. Glutathione-Sepharose beads coupled with GST-fusion proteins were added to similar amount of purified His-Ca v 2.1 synprint (724-981) or His-Ca v 1.2 synprint (680-800). The mixture was incubated under constant rotation for 1 h at 4°C. The binding experiments were conducted in presence of TBS-Ca 21 buffering system with 0.1% Triton X-100. The beads were washed three times with washing buffer and bound complexes were eluted with 15 mM of reduced glutathione and 50 mM Tris-HCl (pH 8). Eluates were separated from beads by centrifugation at 10,000 Â g for 1 min and processed for 10-20% SDS/tricine gradient gel electrophoresis and immunoblotted with anti-His antibody.

Electrophysiological recording
Calcium current (I Ca ) or Barium current (I Ba ) were recorded at least 48 h after tsA-201 cell transfection using whole-cell configuration of the patch-clamp technique. Data acquisition was conducted using patch-clamp amplifier (HEKA Elektronik GmbH). Voltage-clamp protocols and facilitation protocols were applied, and data were acquired using Pulse (HEKA Elektronik GmbH). Currents were filtered at 5 kHz. Leak and capacitance transient currents were subtracted using a P/4 protocol.
Recording pipettes were pulled from borosilicate glass to achieve initial bath resistances of 1.5-3.0 MV and filled Figure 2. Syt-7a accelerates the onset of facilitation of Ca v 2.1 channels. Inset top, Pulse protocol. Currents recorded with 10 mM extracellular Ca 21 and 0.5 mM EGTA in the intracellular recording solution were elicited by test pulses to 110 mV before (P1) and 5 ms after (P2) 10-mV preconditioning prepulses of the indicated durations. Inset, Example traces from control and Syt-7a transfected tsA cells following P1 and P2 pulses. A, Effect of Syt-7a on facilitation as a function of prepulse duration. Facilitation was obtained by normalizing the peak current from P2 to that from P1. Single-exponential fits of the data are shown. B, in tsA-201 cells co-expressing Ca v 2.1 channel with Syt-7a, the slope is significantly increased compared with control cells. Data are represented as mean 6 SEM.
with an intrapipette solution containing (in mM): 120 Nmethyl-D-glucamine (NMDG), 60 HEPES, 1 MgCl 2 , 2 Mg-ATP, and 0.5 EGTA. The extracellular patch-clamp solution contained (in mM): 150 Tris, 1 MgCl 2 , and 10 CaCl 2 or BaCl 2 depending on the experimental protocols. The pH of both intrapipette and extracellular solutions was adjusted to 7.3 using methanesulfonic acid. tsA-201 cell membrane capacitance (Cm) varied from 15-25 pF and access resistance (Rs) varied from 8 to 20 MV. All averaged data represent the mean 6 SEM of at least 10 cells. For peak current measurement using current-voltage (I/V) curves, Ca v 2.1 P/Q-type current was generated using steps of depolarization from À80 to 160 mV every 10-mV step with a holding potential at À80 mV. A total of 10 mM CaCl 2 or BaCl 2 was used in the external patch-clamp solution. For facilitation protocol experiments, 10 mM of CaCl 2 were used in the external solution. Several facilitation protocols have been used to study the role of the three isoforms of Syt-7 (Syt-7a, Syt-7b , and Syt-7g ) in Ca 21 current facilitation. Paired-pulse facilitation protocols were evoked by applying two 1-s-spaced depolarizing pulses P1 and P2 from À80 to 110 mV. A preconditioning 50-ms depolarizing step from À80 to 110 mV was applied only 5 ms before P2. In order to study the voltage dependence of Ca v 2.1 channel activation, P1 and P2 were applied using variable voltages from À120 to 140 mV. P2 over P1 ratios were calculated and compared between transfected tsA-201 cells with and without Syt-7a, Syt-7b , or Syt-7g . The second protocol of paired-pulse facilitation was used to study the effect of changing voltages in the preconditioning pulse on P2. P1 and P2 pulses were maintained from À80 to 110 mV; however, the preconditioning pulse was applied with variable voltages from À120 to 140 mV. Finally, onset of facilitation was studied by increasing preconditioning pulse duration to 10 ms and measuring the ratios of P2 over P1.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 5 (GraphPad Software) and Origin Pro (OriginLab Inc.). All data are shown as the mean 6 SEM. The statistical details of the experiments can be found the results section and figure legends. A Student's t test was used to Figure 3. Effect of Syt-7a on prepulse facilitation of Ca v 2.1 at physiological Ca 21 levels. Inset top, Pulse protocol. Currents recorded with 2 mM extracellular Ca 21 and 0.5 mM EGTA in the intracellular recording solution were elicited by test pulses to 110 mV before (P1) and 5 ms after (P2) 10-mV conditioning prepulses of the indicated durations. Inset, Example traces from control and Syt-7a transfected tsA cells following P1 and P2 pulses. Main panel, Graph shows the effect of Syt-7a on facilitation as a function of prepulse duration. Facilitation was obtained by normalizing the peak current from P2 to that from P1. Single-exponential fits of the data are shown. Data are represented as mean 6 SEM. compare two sets of data. The significance was defined using a threshold of p = 0.05 throughout the study. Error bars indicate SEM. Sample sizes are described in the figure legends.

Syt-7 binds to the Ca v 2.1 channel in mouse brain
To determine whether the slow, high-affinity Ca 21 sensor Syt-7 binds to the presynaptic Ca v 2.1 channels in vivo, co-immunoprecipitation studies were performed on membrane preparations from mouse brain lysates. Ca v 2.1 channels extracted from these neuronal membranes were immunoprecipitated with specific anti-Ca v 2.1 antibodies, and the resulting complexes were probed with anti-Syt-7 antibody using the Dynabeads/Protein G co-immunoprecipitation protocol. The resulting immunoblots revealed Ca v 2.1/Syt-7a interaction with anti-Syt-7 when anti-Ca v 2.1 was used as the precipitating antibody (Fig. 1A, top). In a complementary experiment, Ca v 2.1 channels were immunoprecipitated with anti-Syt-7 antibodies and detected in immunoblots with anti-Ca v 2.1 antibodies (Fig. 1A, bottom). Co-immunoprecipitation experiments using a modified experimental protocol yielded comparable results (Extended Data Fig. 1-1). Together, these results show that Ca v 2.1 channels and Syt-7 are associated with each other in mouse brain membranes.

Syt-7 binds to the synprint site of the Ca v 2.1 channel
Human embryonic kidney tsA-201 cells were transfected with the a 1A , b 2A , and a 2 d 1 subunits of Ca v 2.1 channels together with Syt-7a, the most abundant isoform of Syt-7 (Fukuda et al., 2002). A specific complex of Ca v 2.1 and Syt-7 was co-immunoprecipitated from lysates of tsA-201 cells transfected with Ca v 2.1a 1A subunit and Syt-7a, using either anti-Syt-7 or anti-Ca v 2.1 as the precipitating antibody (Fig. 1B). These results demonstrate a physical interaction between Syt-7a and Ca v 2.1 in intact cells expressing these proteins in vitro, and suggest that other neuron-specific proteins are not required for this protein-protein interaction.
To investigate which domain of the pore-forming a 1 subunit of Ca v 2.1 channels binds Syt-7a, in vitro binding experiments were performed using recombinant fusion proteins (Materials and Methods). Full-length Syt-7a protein Effect of Syt-7a on the decay from facilitation. The facilitation ratio was obtained by normalizing the peak current from P2 to that from P1 and was plotted against the interval between the conditioning prepulse and P2. Shown are results obtained with 50-ms conditioning prepulse. Graph shows the effect of Syt-7a on decay of facilitation as a function of interpulse duration. Data are represented as mean 6 SEM. was expressed as a GST-fusion protein, and the synprint site in Ca v 2.1 (724-981) was expressed as a His-fusion protein. As a control, the equivalent synprint site from the cardiac Ca 21 channel, Ca v 1.2 (680-800), was expressed as a His-fusion protein. GST-Syt-7a proteins were immobilized by binding to glutathione-Sepharose beads and incubated with a constant concentration of His-Ca v 2.1 synprint peptide (724-981) or His-Ca v 1.2 synprint peptide (680-800) using different free Ca 21 concentrations varying from 10 mM to 1 mM. After extensive washing, binding of His-Ca v 2.1 synprint (724-981) to GST-Syt-7a was revealed by immunoblot analysis using an anti-His antibody. As shown in Figure 1C, GST-Syt-7a bound to His-Ca v 2.1 (724-981) synprint in a Ca 21 -dependent manner in vitro, with binding first detected at 50 mM free Ca 21 concentration and increasing to a maximum at 500 mM Ca 21 . In contrast, the negative control peptide His-Ca v 1.2 (680-800) from the corresponding segment of cardiac Ca V 1.2 channels did not bind to GST-Syt-7a. These results demonstrate specific binding of Syt-7 to the synprint site from Ca v 2.1 channels in preference to the corresponding segment of the cardiac Ca v 1.2 channel.

Syt-7a increases the rate and extent of Ca 21 -dependent facilitation
Previous studies have shown the key role of Ca v 2.1 channels (Lee et al., 2000;Mochida et al., 2003aMochida et al., , 2008Inchauspe et al., 2004) and Syt-7 (Jackman et al., 2016;Turecek and Regehr, 2018) in synaptic facilitation, but it is not known whether functional interactions between these two proteins modulate paired-pulse facilitation of P/Qtype Ca 21 currents using pulse protocols similar to those in studies of short-term synaptic facilitation. In order to characterize the mechanism by which Syt-7a increases Ca 21 -dependent facilitation, the effects of Syt-7a on the onset and decay of facilitation were measured with 10 mM Ca 21 in the external solution to mimic the high local Ca 21 concentration near the intracellular mouth of Ca v 2.1 channels in nerve terminals during synaptic transmission. In a paired-pulse protocol, the rate of onset of facilitation was determined by plotting facilitation of I Ca as a function of prepulse duration (Dt; Fig. 2, inset). In cells expressing only Ca v 2.1 channels, the facilitation ratio increased with prepulse duration according to a single-exponential time course ( Fig. 2A, black). Facilitation ratio reached a plateau Figure 5. Syt-7a potentiates Ca v 2.1 facilitation in a paired-pulse protocol following change in prepulse voltage. Inset top, Pulse protocol shown represents paired pulse protocol. Ca 21 current was recorded using 10 mM Ca 21 and 0.5 mM EGTA in the external and internal solutions, respectively. Pulse 1 (P1; depolarization from À80 to 110 mV) elicits the first Ca 21 current. A second 5-ms pulse (P2) generating a second I Ca is applied 2 ms after a 50-ms conditioning prepulse with variable voltages (À40 to 60 mV). Inset bottom, Example traces from control and Syt-7a transfected tsA cells following P1 and P2 pulses. Main panel, Graph shows the effects of Syt-7a isoform on facilitation as a function of prepulse voltage. The facilitation ratio was obtained by normalizing the peak current from P2 to that from P1. Data are represented as mean 6 SEM.
Intracellular Ca 21 concentrations near presynaptic Ca 21 channels rise to nearly 100 mM during rapid stimulation (Berridge et al., 2000). To mimic that condition, we have used 10 mM extracellular Ca 21 in our standard experimental protocol to generate high Ca 21 influx. However, at the physiological level of extracellular Ca 21 , with 2 mM CaCl 2 the external recording solution, applying paired-pulse protocols revealed a significant acceleration of the onset and increase of the extent of Ca 21 -dependent facilitation in cells co-expressing Ca v 2.1 combined with Syt-7a compared with control Inset, Pulse protocol to study the voltage dependence of activation before (open circle or squares; P1) and after (closed circle or squares; P2) a depolarizing prepulse from À80 to 110 mV. Tail currents were measured by holding potential at À40 mV for 5 ms after test pulses (P1, P2) to variable voltages (À40 to 180 mV). Peak tail currents were normalized to the largest tail current measured during the nonfacilitated prepulses (P1) and plotted against the test pulse voltage. A, In control tsA cells, the protocol shows an increase in facilitation P2 normalized to P1. B, Syt-7a potentiated facilitation amplitude of Ca v 2.1 and induced a right shift in prepulse facilitation curve. C, Overlaying the two graphs in A, B shows the increase in amplitude of facilitation and the right shift in voltage dependency of activation. D, Difference in voltage shift in P1 and P2 between cells co-expressing Ca v 2.1 and Syt-7a and control cells. Data are represented as mean 6 SEM. (Fig. 3), as we observed with 10 mM external Ca 21 concentration.

Syt-7a induces a rapidly decaying form of Ca 21dependent facilitation
Syt-7a accelerates the onset of I Ca facilitation and increases facilitation amplitude at all potentials. However, the increased facilitation caused by Syt-7a decayed rapidly (t = 4.29 6 1.68 ms, p = 0.02, n = 7), compared with facilitation of Ca v 2.1 observed for tsA-201 cells in the absence of Syt-7a (t = 12.87 6 3.77 ms, n = 5; Fig. 4, inset, left). The facilitation ratio P2/P1 at the first interpulse duration point was significantly greater for Ca v 2.1-Syt-7a cells (facilitation ratio = 2.73 6 0.71, p = 0.004, n = 7) than for control cells (facilitation ratio = 1.36 6 0.08, n = 5; Fig. 4, inset, right). Although the facilitation ratio in the presence of Syt-7a decays rapidly, the integral of calcium current during the first 10 ms following stimulation is substantially increased (Fig. 4), illustrating the potential physiological significance of this increase in Ca v 2.1 channel activity.

Syt-7a increases voltage-dependent facilitation in paired-pulse protocols
In order to study the effect of Syt-7 on the voltage dependence of Ca v 2.1 activation and its consequences on facilitation, we measured facilitation using a paired-pulse protocol with variable stimulus potentials. In this protocol, facilitation induced by a 50-ms-long prepulse to a variable voltage (À40 to 160 mV) was measured by comparing I Ca elicited by a test pulse before (P1) and after (P2) the conditioning prepulse (Fig. 5, inset). In cells expressing Ca v 2.1 alone, paired-pulse facilitation increased to a maximum at a prepulse voltage of 120 mV and remained at a plateau until 160 mV (facilitation ratio at 20 mV =1.09 6 0.04, n = 8; Fig. 5). Co-expression of Ca v 2.1 with Syt-7a increased the maximum paired-pulse ratio to 1.2 6 0.05 (n = 8 at 140 mV, p , 0.01), approximately doubling the increase in Ca 21 current induced by paired-pulse facilitation in the absence of Syt-7.
We also expressed Ca v 2.1 channels without or with Syt-7 and measured the voltage dependence of activation Figure 7. Effect of Syt-7a prepulse facilitation of Ca v 2.1 channel at physiological levels. Inset, Voltage protocol. Currents recorded with 2 mM extracellular Ca 21 and 0.5 mM EGTA in the intracellular recording solution were elicited by test pulses to 110 mV before (P1) and 5 ms after (P2) 10-mV conditioning prepulses of the indicated durations. A-C, Effect of Syt-7a on facilitation as a function of prepulse voltage. Facilitation was obtained by normalizing the peak current from P2 to that from P1. Single-exponential fits of the data are shown. Data are represented as mean 6 SEM. of the resulting Ca 21 currents (Fig. 6). In this paired-pulse protocol described in Figure 6, inset, voltages were varied from À40 to 180 mV in both pulses P1 and P2 following a constant prepulse voltage of À80 to 110 mV before P2. Previous studies (Lee et al., 1999(Lee et al., , 2000 showed that this protocol induced facilitation of Ca v 2.1 channels. As shown in Figure 6A-C, Syt-7a significantly increased I Ca across the positive voltage range and increased maximum facilitation at potentials of 140 mV and higher in the presence of 10 mM Ca 21 . Syt-7a induced a significant ;5to 15-mV positive shift in the voltage dependence of Ca v 2.1 activation, as observed by comparing the half-activation voltage (V 50 ) at P1 (V 50 = 14.28 6 3.28, n = 7, p = 0.002) versus control cells (V 50 = 3.96 6 1.72, n = 19). During pulse P2, Syt-7a induced a significant ;3.6to 11-mV positive shift in the voltage dependence of Ca v 2.1 activation (V 50 = 7.96 6 2.3, n = 7, p = 0.02) versus control cells (V 50 = 0.7 6 1.4, n = 19; Fig. 6D). This Syt-7a effect was also observed at physiological Ca 21 levels, where both facilitation amplitude and the positive shift in voltage dependence of Ca v 2.1 activation were evident (Fig. 7). Together, these results show that Syt-7a induces strong facilitation of Ca 21 currents at membrane potentials in the range of the peak of action potentials (;0 to 140 mV).

Differential modulation of Ca v 2.1 by isoforms of Syt-7
Among the three Ca v 2 subfamily members, only the Ca v 2.1 channel supports short-term synaptic facilitation (Inchauspe et al., 2004); however, Syt-7 isoforms may have subtype-specific modulatory effects on Ca v 2.1. Three splice variants of Syt-7 exist in mouse and human: the major form Syt-7a and two minor forms, Syt-7b and Syt-7g (Fukuda et al., 2002). Syt-7b and Syt-7g contain additional 44 and 116 amino acids, respectively, in the connecting segment between their transmembrane domain and the cytoplasmic C2 Ca 21 -binding domain (Fukuda et al., 2002). Both the b and g isoforms of Syt-7 were bound to Ca v 2.1 channels in extracts of transfected tsA-201 cells to a similar extent as Syt-7a, as indicated by co-immunoprecipitation with anti-Ca v 2.1 antibodies and immunoblotting with isoform-specific anti-Syt-7 antibodies ( Fig. 8A-C, left). In complementary experiments, Ca v 2.1 was co-immunoprecipitated from transfected tsA-201 cells with antibodies against Syt-7b and Syt-7g ( Fig. 8A-C,  right). The consistent results in these two complementary immunoprecipitation protocols indicate that these protein interactions are specifically detected independent of the antibodies used for immunoblotting. Co-expression of Syt-7a, Syt-7b , and Syt-7g together with Ca v 2.1 channels did not have significant effects on the peak amplitude of either Ba 21 or Ca 21 currents in comparison to expression of Syt-7a alone (Extended Data Fig. 8-1). Together, these experiments indicate that all three Syt-7 isoforms bind to Ca v 2.1 channels in transfected cells without significantly altering their level of functional expression.
We investigated the effects of co-expression with Syt-7b and Syt-7g on facilitation of Ca v 2.1 channels with 10 mM CaCl 2 in the extracellular solution. Syt-7b increased the facilitation ratio of Ca v 2.1 channels (p , 0.01; Fig. 9A) subunit of Ca v 2.1 channels along with Syt-7a, b , and g , Ca v 2.1 channels were immunoprecipitated with an anti-Ca v 2.1 antibody and blotted with anti-Syt antibody. Right, Reverse experiment where the three isoforms of Syt-7 were immunoprecipitated with an anti-Syt antibody and blotted with anti-Ca v 2.1 antibody. Because different Ca 21 concentrations were used in co-immunoprecipitation experiments, segments from different immunoblots were spliced together to show specific comparisons. Those protein bands are delineated for clarification. The immunoblots presented here are representative of at least three experiments for each co-immunoprecipitation or immunoblot. Whole-cell voltage clamp experiments show that Ca v 2.1 expressed alone gives similar peak Ba 21 and Ca 21 currents as Ca v 2.1 1 Syt7-ab g (Extended Data Fig. 8-1). and accelerated the rate of facilitation, as demonstrated by the significant increase in the slope in cells co-transfected with Ca v 2.1 plus Syt-7b (slope = 0.02 6 0.004 ms À1 ; n = 15, p = 0.006) compared with control cells (slope = 0.009 6 0.001 ms À1 ; n = 20; Fig. 9A, inset). However, these effects were substantially smaller than with co-expression of Syt-7a (Fig. 2). Co-expression of Syt-7g also increased the peak level of facilitation (p , 0.01) to a lesser degree than Syt-7a (Fig. 9B), and it showed only a trend toward a significant increase in facilitation rate (slope = 0.03 6 0.01 ms À1 ; n = 9, p = 0.16, ns) compared with control cells (slope = 0.01 6 0.003 ms À1 ; n = 20; Fig. 9B, inset). To determine whether Syt-7b and Syt-7g can compete effectively with Syt-7a, we co-expressed all three Syt-7 isoforms and measured Ca 21 -dependent facilitation (Extended Data Fig. 9-1). We found that co-expression Syt-7b and Syt-7g together with Syt-7a reduced the strong increase in the rate and extent of facilitation observed with Syt-7a alone (Extended Data Fig. 9-1, red). These results are consistent with the conclusion that Syt-7b and Syt-7g effectively compete for occupancy of the synprint site and alter modulation of Ca v 2.1 by Syt-7a. Evidently, replacement of Syt-7a with either Syt-7b or Syt-7g at the synprint site would reduce facilitation of Ca v 2.1 channels.
In experiments testing the effect of a depolarizing prepulse on the voltage dependence of activation, neither Syt-7b ( Fig. 10A-C) nor Syt-7g (Fig. 10E-G) increased the maximum prepulse facilitation of Ca v 2.1 at positive membrane potentials in contrast to Syt-7a. Similarly, neither Syt-7b (Fig. 10A-D) nor Syt-7g (Fig. 10E-H) caused a significant shift in the voltage dependence of activation of Ca v 2.1 Figure 9. Syt-7b and Syt-7g differentially modulate facilitation of Ca v 2.1 channels. Inset, Pulse protocol. Currents recorded with 10 mM extracellular Ca 21 and 0.5 mM EGTA in the intracellular recording solution were elicited by test pulses to 110 mV before (P1) and 5 ms after (P2) 10-mV conditioning prepulses of the indicated durations. A, Left, Syt-7b increases the facilitation ratio with increasing prepulse duration. Right, Syt-7b accelerates the onset of facilitation as a function of prepulse duration. B, Left, Syt-7g increases the facilitation ratio with increasing prepulse duration. Right, Syt-7g does not accelerate the onset of facilitation as a function of prepulse duration. Facilitation was obtained by normalizing the peak current from P2 to that from P1. Single-exponential fits of the data are shown. Data are represented as mean 6 SEM additional experiments with different pulse protocols provide additional information on the effects of Syt-7b and Syt-7g on facilitation of Ca v 2.1 channels (Extended Data Fig. 9-1). channels, unlike Syt-7a (Fig. 10C,D). Co-expressing the three Syt-7 isoforms together caused a negative shift in the voltage dependence of activation following a depolarizing prepulse, in contrast to the positive shift in the voltage dependence of activation following a prepulse caused by coexpression of Syt-7a alone (Extended Data Fig. 10-1, red). All of these voltage-dependent activation curves are monophasic (Figs. 7, 10; Extended Data Fig. 10-1), consistent with stoichiometric binding of each Syt-7 isoform to Ca v 2.1 resulting in complete shifts of the activation curves. Together, these results suggest a dominant effect of Syt-7b and Syt-7g on the voltage dependence of activation in paired-pulse protocols in the presence of all three Syt-7 isoforms.
In addition to their differential effects on Ca 21 -dependent facilitation, co-expression of Syt-7b or Syt-7g also had different effects on Ca 21 -dependent inactivation of Ca v 2.1 channels compared with Syt-7a (Fig. 11). In the presence of 10 mM Ba 21 as the permeant extracellular cation, Ca v 2.1 channels activated rapidly and did not inactivate significantly in 200-ms depolarizing pulses when co-expressed with any of the Syt-7 isoforms (Fig.  11A). In contrast, in the presence of 10 mM Ca 21 as permeant ion, Ca v 2.1 channels inactivated with a time constant of ;600 ms through their Ca 21 /CaM-dependent inactivation mechanism (Fig. 11B, black). Strikingly, co-expression of Syt-7a substantially slowed Ca 21 -dependent inactivation (Fig. 11B, red), whereas co-expression of Syt-7b had a smaller effect (Fig. 11B, blue) and co-expression of Syt-7g had no effect on Ca 21 -dependent inactivation (Fig. 11B, green). These results indicate that replacement of Syt-7a with either Syt-7b or Syt-7g at the synprint site would decrease Ca 21 entry in single depolarizations by preventing the inhibition of Ca 21 -dependent inactivation of Ca v 2.1 channels induced by Syt-7a (Fig. 11), and at the same time would reduce prolonged Ca 21 entry by decreasing the enhanced facilitation of Ca v 2.1 channels caused by Syt-7a during repetitive depolarizations (Fig. 9). This parallel modulation of Ca 21 entry by single depolarizations plus trains of depolarizations would have a potent impact on synaptic transmission. Altogether, these results indicate that the two minor Syt-7 isoforms bind to Ca v 2.1 channels in cellular context without altering functional expression of Ca v 2.1. However, co-expression of Syt-7b and Syt-7g can partially reverse the functional effects of Syt-7a on facilitation (Extended Data Fig. 9-1), the voltage dependence of activation following a depolarizing prepulse (Extended Data Fig. 10-1), and the rate of Ca 21 -dependent inactivation (Fig. 11). The differential actions of the three isoforms of Syt-7 provide a rich panoply of modulatory effects on Ca v 2.1 channel activation, facilitation, and inactivation that would have a strong influence on synaptic transmission.

Discussion
Syt-7 modulates Ca v 2.1 channels through binding to the synprint site In presynaptic nerve terminals, Ca v 2.1 and Ca v 2.2 channels associate with SNARE proteins and a large number of other presynaptic proteins (Khanna et al., 2007;Müller et al., 2010;. SNARE proteins interact with the synprint site located in the intracellular loop between domains II and III (Sheng et al., 1994Rettig et al., 1996), which is thought to play an important role in the incorporation of Ca v 2.1 channels into the synaptic vesicle fusion machinery and regulation of their function (Mochida et al., 2003b;Szabo et al., 2006). Ca 21 influx through Ca v 2.1 channels is a crucial step in triggering Ca 21 -dependent exocytosis of neurotransmitter vesicles (Olivera et al., 1994;Dunlap et al., 1995). Previous studies showed that the fast Ca 21 sensor Syt-1 binds to the synprint site of both Ca v 2.1 and Ca v 2.2 channels (Sakurai et al., 1996;Sheng et al., 1996;Charvin et al., 1997;Kim and Catterall, 1997). These protein interactions are likely to modulate the rapid, synchronous component of neurotransmitter release mediated by Syt-1 (Bacaj et al., 2013).
In contrast to these extensive studies of SNARE proteins and the fast Ca 21 sensor Syt-1, the slow, high-affinity Ca 21 sensor Syt-7 is unique in mediating synaptic facilitation (Jackman et al., 2016;Nanou et al., 2016a;Turecek and Regehr, 2018) and asynchronous release (Bacaj et al., 2013;Turecek and Regehr, 2018), but its interactions with Ca v 2.1 channels had not previously been investigated. Our results presented here show that Syt-7 binds to the synprint site of Ca v 2.1 channels in vivo in mouse brain membranes, in vitro in transfected cells, and in solution in protein-interaction experiments. Unexpectedly, in contrast to Syt-1, our results provide evidence that Syt-7 modulates Ca2 + -dependent facilitation and inactivation (CDI) of Ca v 2.1 channels, which are implicated in short-term forms of synaptic plasticity, including synaptic facilitation and the rapid phase of synaptic depression (Lee et al., 2000(Lee et al., , 2002Mochida et al., 2008;Nanou et al., 2016aNanou et al., , b, 2018. Direct interaction of Syt-7 and Ca v 2.1 as shown here may contribute to short-term synaptic facilitation, in which both of these interacting protein partners are thought to play essential roles.

Syt-7 isoforms differentially enhance facilitation of Ca v 2.1 channels
In cells expressing Ca v 2.1 channels, we consistently observed Ca 21 -dependent facilitation of the Ca 21 current, as reported previously (Lee et al., 1999(Lee et al., , 2000(Lee et al., , 2003. In the presence of Syt-7a, both the rate and extent of facilitation of Ca v 2.1 channels were increased, and the rate of decay of facilitation was also accelerated. These results suggest that expression of Syt-7a in presynaptic terminals in vivo would enhance Ca 21 -dependent facilitation and sharpen the time-dependent peak of facilitation of Ca v 2.1 channels. Syt-7b and Syt-7g also bound to the synprint site. However, compared with Syt-7a, Syt-7b , and Syt-7g had lesser effects on facilitation in response to voltage steps and did not shift the voltage dependence of prepulse facilitation. Differential expression of these Syt-7 isoforms could confer cell-specific regulation via interactions with Ca v 2.1 channels and other regulatory targets. Syt-7 isoforms differentially modulate inactivation of Ca v 2.1 channels In our depolarizing step protocols, none of the Syt-7 isoforms had any significant effect on the peak amplitude of Ba 21 current. However, our data show that Syt-7a significantly increased Ca 21 -dependent inactivation of the Ca 21 channel, which would oppose facilitation. Syt-7b and Syt-7g had lesser effects. The combination of increased Figure 11. Syt-7 isoforms differentially modulate Ca 21 -dependent inactivation of Ca v 2.1 channels. Ca v 2.1 currents were elicited by depolarizing from a holding potential of À80 mV to a test potential of 110 mV. A, Time courses (200 ms) of I Ba with 10 mM Ba 21 as a permeant cation. B, Time courses (1000 ms) of I Ca with 10 mM Ca 21 as permeant ion. Syt-7a and Syt-7b significantly slowed inactivation of the Ca v 2.1 channel in the presence of 10 mM Ca 21 , whereas Syt-7g had no effect. Data are represented as mean 6 SEM. facilitation followed by increased inactivation induced by Syt-7a would have the overall effect of sharpening the peak of the presynaptic calcium current to allow effective facilitation of repetitive rounds of neurotransmitter release. Syt-7b and Syt-7g would bind to the synprint site of Ca v 2.1 channels but induce lesser functional effects.

Comparison with regulation by CaM-like calcium sensor proteins
Our work characterizes an unexpected form of regulation of P/Q-type current conducted by Ca v 2.1 channels by the high affinity Ca 21 sensor Syt-7 through a direct interaction with the synprint site. Interaction of Ca v 2.1 with Syt-7 may enhance facilitation of presynaptic Ca 21 current and thereby play a role in triggering activation of the Ca 21 -dependent exocytosis machinery, including the SNARE proteins. These effects would be dependent on the isoform of Syt-7 that is expressed in different cells and synapses. In previous experiments, CaM has been shown to regulate Ca v 2.1 channel activity, inducing increased facilitation and increased Ca 21 -dependent inactivation, dependent on the local Ca 21 concentration (Lee et al., 1999(Lee et al., , 2000DeMaria et al., 2001). Our results further show that Ca 21 -dependent inactivation of Ca v 2.1 channels is modulated by Syt-7 in an isoform-dependent manner. In presynaptic nerve terminals, these changes in both the Ca 21 entry in response to single action potentials plus trains of action potentials would substantially alter the encoding properties of synaptic transmission.
Other neuronal Ca 21 sensor proteins related to CaM are expressed in the central nervous system, including Ca 21 binding protein-1 (CaBP-1), visinin-like protein-2 (VILIP-2), and neuronal Ca 21 sensor-1 (NCS-1). These Ca 21 sensor proteins displace CaM from the C-terminal domain of Ca v 2.1 and modify short-term synaptic facilitation and rapid synaptic depression . It will be interesting to further investigate how these two distinct regulatory mechanisms mediated by Syt-7 and Ca 21 sensor proteins converge on the Ca v 2.1 channel on the millisecond time frame of short-term synaptic plasticity.
In conclusion, our work characterizes a novel form of regulation of P/Q-type Ca v 2.1 channels by the high affinity Ca 21 sensor Syt-7 through direct interaction with the synprint site. Ca v 2.1/Syt-7 interaction potentiates facilitation of Ca 21 current and may play a role in triggering Ca 21 -dependent exocytosis along with other SNARE proteins. Syt-7 also modulates Ca 21 /CaM-dependent inactivation. Understanding the mechanism by which Syt-7 isoforms enhance facilitation and modulate inactivation of Ca v 2.1 channels in presynaptic terminals is a first step toward deciphering the complete picture of the role played by Syt-7 in the brain.