SNARE Proteins Synaptobrevin, SNAP-25, and Syntaxin Are Involved in Rapid and Slow Endocytosis at Synapses

Transcription

1 Cell Reports Report SNARE Proteins Synaptobrevin, SNAP-25, and Syntaxin Are Involved in Rapid and Slow Endocytosis at Synapses Jianhua Xu, 1,2,3 Fujun Luo, 1,3 Zhen Zhang, 1 Lei Xue, 1 Xin-Sheng Wu, 1 Hsueh-Cheng Chiang, 1 Wonchul Shin, 1 and Ling-Gang Wu 1, * 1 National Institute of Neurological Disorders and Stroke, 35 Convent Drive, Building 35, Room 2B-1012, Bethesda, MD 20892, USA 2 Institute of Molecular Medicine and Genetics and Department of Neurology, Georgia Health Sciences University, th Street, Augusta, GA 30912, USA 3 These authors contributed equally to this work *Correspondence: wul@ninds.nih.gov SUMMARY Rapid endocytosis, which takes only a few seconds, is widely observed in secretory cells. Although it is more efficient in recycling vesicles than in slow clathrin-mediated endocytosis, its underlying mechanism, thought to be clathrin independent, is largely unclear. Here, we report that cleavage of three SNARE proteins essential for exocytosis, including synaptobrevin, SNAP-25, and syntaxin, inhibited rapid endocytosis at the calyx of Held nerve terminal, suggesting the involvement of the three SNARE proteins in rapid endocytosis. These SNARE proteins were also involved in slow endocytosis. In addition, SNAP-25 and syntaxin facilitated vesicle mobilization to the readily releasable pool, most likely via their roles in endocytosis and/or exocytosis. We conclude that both rapid and slow endocytosis share the involvement of SNARE proteins. The dual role of three SNARE proteins in exo- and endocytosis suggests that SNARE proteins may be molecular substrates contributing to the exocytosis-endocytosis coupling, which maintains exocytosis in secretory cells. INTRODUCTION Vesicle endocytosis recycles exocytosed vesicles and thus maintains exocytosis in secretory cells (Royle and Lagnado, 2003). Endocytosis may be slow with a time constant (t) of s or rapid with a t of 1 3 s. Slow endocytosis, observed at most secretory cells examined (Royle and Lagnado, 2003), is mediated by the clathrin-dependent mechanism (Dittman and Ryan, 2009). Rapid endocytosis is observed at most cells where the membrane capacitance can be measured, such as retinal nerve terminals (von Gersdorff and Matthews, 1994), calyx of Held (Sun et al., 2002; Wu et al., 2005), hippocampal mossy fiber terminals (Hallermann et al., 2003), pituitary nerve terminals (Hsu and Jackson, 1996), auditory hair cells (Beutner et al., 2001), and nonneuronal secretory cells (Artalejo et al., 1995; He et al., 2008). Optical imaging also reveals rapid endocytosis at small synapses (Zhang et al., 2009), although this issue remains debated (Granseth et al., 2009; He and Wu, 2007). Since rapid endocytosis is triggered by calcium influx during intense activity (Artalejo et al., 1995; Neves et al., 2001; Beutner et al., 2001; Wu et al., 2005, 2009; but see von Gersdorff and Matthews, 1994), it may fulfill the need of faster vesicle recycling with a larger capacity when many vesicles are released by intense stimulation. It may also rapidly restore the normal membrane structure of secretory cells (Wu and Wu, 2009). Despite these important roles, its underlying molecular mechanism, which is clathrin independent (Artalejo et al., 1995; Jockusch et al., 2005), remains elusive. Here, we determined whether three SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) proteins essential for exocytosis, synaptobrevin, SNAP-25, and syntaxin (Sudhof, 2004) are involved in rapid and/or slow endocytosis at the calyx of Held nerve terminal. We found that all three SNARE proteins are involved in rapid and slow endocytosis, which may help to explain the tight coupling of exocytosis and endocytosis. RESULTS Rapid Endocytosis in Control At calyces of 7- to 10-day-old rats, we induced rapid endocytosis with ten pulses of 20 ms depolarization from 80 to +10 mv at 10 Hz (depol 20msX10 )(Wu et al., 2005, 2009). With boiled tetanus toxin (TeTx, 2 mm) in the pipette serving the control for TeTx application, depol 20msX10 induced a membrane capacitance (C m ) jump (DC m ) of 1,710 ± 325 ff, followed by a biexponential decay with time constants (t) of 1.6 ± 0.3 s (amplitude: 22% ± 4%) and 21.0 ± 2.2 s (n = 5, Figure 1A). The initial C m decay after the DC m (Rate decay ), measured from the trace between 0.5 and s after stimulation, was 237 ± 58 ff/s (n = 5). We did not measure the first 0.5 s after stimulation because it may contain C m artifacts with a t of 200 ms 1414 Cell Reports 3, , May 30, 2013 ª2013 The Authors

2 D A B C E G I H J F Figure 1. Block of Synaptobrevin Inhibits Endocytosis Induced by depol20msx10 (A) Sampled (upper, single trace) and averaged (lower) capacitance changes (Cm) induced by depol20msx10 (arrow) with a pipette containing boiled TeTx (2 mm, n = 5 calyces, black) or TeTx (2 mm, n = 9 calyces, red). Averaged Cm traces (lower left two traces) are also normalized to the same amplitude and superimposed (lower right). Traces were taken at 4 10 min after whole-cell break-in (applies to A F). The averaged Cm traces were plotted as mean ± SEM (applies to all averaged Cm traces in all figures). (B) Sampled calcium currents (ICa) induced by the first 20 ms depolarization during depol20msx10 with a pipette containing boiled TeTx (black) or TeTx (red). (C) Ratedecay_n (Rate), DCm, and QICa induced by depol20msx10 in the presence of boiled TeTx (2 mm, black, n = 5) or TeTx (2 mm, red, n = 9). Data were normalized to the mean of the boiled TeTx group. *p < 0.05; **p < Data are expressed as mean ± SEM (applies to all bar graphs in all figures). (D F) Similar to (A) (C), respectively, except that boiled TeTx and TeTx are replaced with Sybsp (2 mm, n = 6) and Sybp (2 mm, n = 8), respectively. (G J) Similar to (A), (C), (D), and (F), respectively, except that the data were obtained at 2 4 min after break-in. (Wu et al., 2005; Yamashita et al., 2005). For two reasons, rapid Cm decay beyond 0.5 s reflects endocytosis. First, botulinum toxin C (BoNT/C), which cleaves syntaxin (Sakaba et al., 2005), abolishes the DCm induced by depol20msx10 (Wu et al., 2005). Thus, the DCm reflects SNARE-mediated exocytosis. Subsequent rapid Cm decay must reflect retrieval of exocytosed vesicles. Second, rapid Cm decay is mediated by dynaminand calcium/calmodulin-dependent endocytosis because it is Cell Reports 3, , May 30, 2013 ª2013 The Authors 1415

3 inhibited by various dynamin inhibitors (Xu et al., 2008), calcium buffers, and calmodulin inhibitors (Wu et al., 2009). When the DC m was normalized to 1, the Rate decay measured from 0.5 to s after stimulation, now called Rate decay_n,was 0.14 ± 0.03/s (n = 5). Similar results were obtained in the absence or presence of other boiled toxins or scrambled peptides used in this study. The Rate decay or Rate decay_n reflected mostly the rapid component of endocytosis (Wu et al., 2005, 2009), as confirmed below. Based on the averaged capacitance trace in control (Figure 1A), the Rate decay of the rapid component of endocytosis, which can be theoretically calculated as the ratio between the amplitude and the t, was 235 ff/s (= 1,710 ff /1.6 s), whereas the Rate decay of the slow component of endocytosis was only 64 ff/s (= 1,710 ff /21 s). Thus, the rapid component represented 79% [= 235 / ( ]) of the overall Rate decay. In this study, Rate decay or Rate decay_n was measured directly from the trace, but not from theoretical calculation, because the block of endocytosis by toxins made it difficult for exponential fitting. Likewise, we did not use endocytosis t for statistics. Block of Synaptobrevin Inhibits Rapid Endocytosis We included in the pipette either tetanus toxin (TeTx, 2 mm) to cleave synaptobrevin or a peptide containing the N-terminal proline-rich domain of synaptobrevin (Syb p, 2 mm) that blocks the interaction between synaptobrevin and other exocytosis proteins (Cornille et al., 1995). The corresponding control was boiled TeTx (2 mm) or scrambled Syb p (Syb sp, 2 mm). At 4 10 min after whole-cell break-in, TeTx and Syb p significantly reduced the Rate decay_n induced by depol 20msX10 to 31% 32% and the DC m to 60% 67% of the corresponding control, but did not affect the calcium current amplitude or charge (Q ICa, Figures 1A 1F). Thus, reduction of Rate decay_n was not due to Q ICa reduction (Figures 1B, 1C, 1E, and 1F). Four sets of evidence indicate that the Rate decay_n reduction was not caused by the DC m reduction. First, since DCm was normalized to 1, Rate decay_n should be inversely proportional to C m decay t regardless of the DC m value (theoretically, Rate decay_n =1/t). Thus, Rate decay_n inversely reflects endocytosis t and is independent of the DC m value. Second, there has been no report that an exocytosis decrease alone can prolong endocytosis or reduce Rate decay_n. Instead, decrease of exocytosis is accompanied by a linear decrease or no change of endocytosis t depending on whether the endocytic capacity is saturated or not (Wu and Betz, 1996; Sankaranarayanan and Ryan, 2000; Sun et al., 2002; Wu et al., 2005; Yamashita et al., 2005; Balaji et al., 2008). Accordingly, the DC m decrease itself may increase or not affect Rate decay_n, but not decrease it. The decrease of Rate decay_n is therefore a conservative estimate of endocytosis inhibition when DCm is decreased. Third, we recently showed that various strong stimuli induced a DC m (1,500 ff) similar to that induced by depol 20msX10 but different Rate decay (or Rate decay_n ) that varies by approximately eight times owing to the difference in calcium influx (Xue et al., 2012b). Various weaker stimuli induced a DC m of ff, a value similar to that induced by a 20 ms depolarization we used later, but approximately two to ten times of differences in the Rate decay or Rate decay_n (Wu et al., 2009; Xue et al., 2012b). Thus, DC m has minimal effect on Rate decay_n. Fourth, at 2 4 min after whole-cell break-in, TeTx and Syb p did not reduce the DC m, but the Rate decay_n to 57% ± 12% and 42% ± 7% of the corresponding control, respectively (n = 6 8, p < 0.05, Figures 1G 1J). Thus, the Rate decay_n decrease was not caused by the DC m decrease. Reduction of the DC m by TeTx (2 mm, 6 8 min dialysis) was smaller than an earlier report (Sakaba et al., 2005), likely because the earlier report used a higher concentration (5 mm) for a longer dialysis time (8 12 min). Consistent with this possibility, less block was observed when a lower concentration was dialyzed for only 5 6 min from the same lab (Hosoi et al., 2009). The decrease of Rate decay_n by TeTx and Syb p suggests that synaptobrevin is involved in endocytosis (Figure 1). However, if there are two vesicle populations, one more susceptible to TeTx and Syb p but better equipped with endocytosis machinery, the other in the opposite, the latter population would be left in the presence of TeTx or Syb p to mediate exocytosis and thus cause slower endocytosis. This scenario is highly unlikely for two reasons. First, it predicts that when TeTx or Syb p does not block exocytosis, endocytosis is not affected. In contrast, at 2 4 min after break-in, TeTx or Syb p did not reduce DC m, but reduced Rate decay_n (Figures 1G 1J). Second, the scenario predicts that at the time endocytosis finishes in control with two vesicle populations, endocytosis in the presence of blockers, in which only the slower population is left, should also finish. This prediction is incorrect. At 40 s after stimulation, C m returned to baseline in control, but did not decay much in the presence of TeTx or Syb p (Figures 1A, 1D, 1G, and 1I). Similar results were found for other toxins and slow endocytosis described later (e.g., Figures 2D, 4A, 4C, 4E, 4G, and 4I). We concluded that synaptobrevin is involved in endocytosis. Block of SNAP-25 or Syntaxin Inhibits Rapid Endocytosis We included botulinum neurotoxin E (BoNT/E, nm) or BoNT/A (1 mm) in the pipette to cleave SNAP-25 (Niemann et al., 1994). The corresponding control was boiled BoNT/E ( nm) or BoNT/A (1 mm). At 4 10 min after break-in, BoNT/E and BoNT/A reduced the Rate decay_n induced by depol 20msX10 to 38% 39% and the DC m to 28% 55% of control on average, but did not affect Q ICa (n = 6 10 for each group, Figures 2A 2F). These results suggest the involvement of SNAP-25 in rapid endocytosis. We included BoNT/C ( nm) in the pipette to cleave syntaxin 1 (Niemann et al., 1994). The corresponding control was boiled BoNT/C. Although BoNT/C has a weak effect on SNAP-25 (Niemann et al., 1994), it cleaves syntaxin rather than SNAP-25 within 10 min of dialysis at calyces (Sakaba et al., 2005). At 4 10 min after whole-cell break-in, BoNT/C reduced the Rate decay_n induced by depol 20msX10 to 60% and the DC m to 22% of control on average, but did not affect Q ICa (n = 7 14, Figures 3A 3C), suggesting the involvement of syntaxin in rapid endocytosis. Block of SNAP25 or Syntaxin Inhibits Replenishment of the Readily Releasable Pool Recent studies suggest that endocytosis may facilitate replenishment of the readily releasable pool (RRP) by clearance of fused vesicle membrane and proteins at the active zone of 1416 Cell Reports 3, , May 30, 2013 ª2013 The Authors

4 A B D C E F Figure 2. Block of SNAP-25 Inhibits Endocytosis Induced by depol20msx10 (A C) Similar arrangements as Figures 1A 1C, respectively, except using boiled BoNT/E ( nm, n = 6) and BoNT/E ( nm, n = 10). Data were taken at 4 10 min after break-in (applies to A F). (D F) Similar to (A) (C), respectively, except using boiled BoNT/A (1 mm, n = 8) and BoNT/A (1 mm, n = 8). calyces (Wu et al., 2009; Hosoi et al., 2009). We noticed that BoNT/E, BoNT/A, and BoNT/C slowed the RRP replenishment. The DCm induced by each 20 ms depolarization during depol20msx10 reflects the rate of RRP replenishment, because each 20 ms depolarization depleted the RRP (Wu et al., 2009). When the DCm induced by the first 20 ms depolarization was normalized to the same amplitude, it was evident that the DCm induced by the subsequent nine depolarizing pulses during depol20msx10 decreased significantly in the presence of BoNT/ E, BoNT/A, or BoNT/C (Figure 3D). We did not observe a consistent decrease for TeTx or Sybp. However, a small decrease below our detection limit remains possible. These results suggest that SNAP-25 and syntaxin facilitate the RRP replenishment, likely via their roles in endocytosis and/or exocytosis. At 4 10 min after break-in, BoNT/A reduced the Ratedecay_n and the DCm induced by depol20ms to 35% ± 7% and 64% ± 11% (n = 6) of control (n = 4, Figures 4I 4J). BoNT/E caused a near full block of the DCm induced by depol20ms, making it difficult to measure the Ratedecay_n. Nevertheless, the slow component of endocytosis after depol20msx10 was largely blocked by BoNT/E, as evident at s after stimulation (Figure 2A). These results suggest the involvement of SNAP-25 in slow endocytosis. Similarly, BoNT/C nearly abolished the DCm induced by depol20ms. The slow component of endocytosis after depol20msx10 was largely blocked by BoNT/C, as evident at s after stimulation (Figure 3A), suggesting the involvement of syntaxin in slow endocytosis. DISCUSSION Block of Synaptobrevin, SNAP-25, or Syntaxin Inhibits Slow Endocytosis We induced slow endocytosis by a 20 ms depolarization ( 80 to +10 mv, depol20ms) (Wu et al., 2005, 2009). In control with 2 mm boiled TeTx in the pipette, depol20ms induced a DCm of 531 ± 61 ff, followed by a decay with a t of 12.4 ± 0.9 s and a Ratedecay of 40 ± 5 ff/s (n = 6, Figure 4A), as measured within the first 4 s after stimulation. The Ratedecay_n was 0.08 ± 0.01/s (n = 6). Similar results were obtained with different boiled toxins or scrambled peptides. At 4 10 min after whole-cell break-in, TeTx and Sybp significantly reduced the Ratedecay_n induced by depol20ms to 13% 42% and the DCm to 62% 80% of control, on average (n = 4 6 for each group, Figures 4A 4D). As described above, reduction of the DCm itself could not decrease Ratedecay_n. In addition, at 2 4 min after break-in, TeTx and Sybp did not significantly decrease the DCm, but reduced the Ratedecay_n to 51% of control (n = 4 5, Figures 4E 4H). These results suggest the involvement of synaptobrevin in slow endocytosis. We found that cleavage of synaptobrevin, SNAP-25, and syntaxin blocked rapid endocytosis (Figures 1, 2, and 3), suggesting the involvement of three SNARE proteins in rapid endocytosis, which is suggested to be clathrin independent (Artalejo et al., 1995; Jockusch et al., 2005). This finding provides mechanistic information for the poorly understood rapid endocytosis. We also found that these three SNARE proteins are involved in slow endocytosis (Figure 4), a clathrin-dependent form of endocytosis (Dittman and Ryan, 2009). Thus, two apparently different forms of endocytosis share a common mechanism the dual role of three SNARE proteins in exo- and endocytosis. Previously, three pioneering studies examined the involvement of SNARE proteins in endocytosis. One study insightfully implicates the involvement of synaptobrevin in rapid endocytosis at hippocampal synapses (Dea k et al., 2004). However, the technique used to detect rapid endocytosis is indirect and controversial (He and Wu, 2007; Zhang et al., 2009; Granseth Cell Reports 3, , May 30, 2013 ª2013 The Authors 1417

5 A B D C et al., 2009). A recent study shows that TeTx blocks slow endocytosis at calyces, suggesting the involvement of synaptobrevin in slow endocytosis (Hosoi et al., 2009). Knockout of SNAP-25 at cultured hippocampal synapses does not inhibit sucroseinduced FM dye uptake into synaptic vesicles, suggesting that SNAP-25 is not involved in endocytosis (Bronk et al., 2007). The present work extended over these previous studies in three aspects. First, by recording rapid endocytosis unequivocally with capacitance measurements (Wu et al., 2005; Xu et al., 2008; Xue et al., 2012b), we provided strong evidence showing that synaptobrevin is involved in rapid endocytosis. We consolidated the previous finding for the involvement of synaptobrevin in slow endocytosis by excluding the possibility that the block of endocytosis by TeTx is a side effect caused by reduction of exocytosis (Figures 1G, 1H, 4E, and 4F), which has not been addressed in the previous study (Hosoi et al., 2009). Second, we found that SNAP-25 was involved in both rapid and slow endocytosis in synapses. Unlike the previous study using sucrose to induce calcium-independent exo- and endocytosis (Bronk et al., 2007), we applied trains of depolarization to mimic physiological action potential stimuli (Wu et al., 2005, 2009; Xu et al., 2008), which induces calcium-dependent exo- and endocytosis. Third, we found that not only synaptobrevin and SNAP-25, but also syntaxin are involved in both rapid and slow endocytosis. More generally speaking, both vesicular- and membrane-targeted SNARE proteins are involved in endocytosis. Given that SNARE proteins mediate exocytosis at all nerve terminals and many nonneuronal secretory cells, the dual role of three SNARE proteins in exo- and endocytosis is likely to have wide implication in secretory cells. The dual role in exo- and endocytosis suggests that after mediating vesicle fusion, three SNARE proteins participate in the subsequent endocytosis, likely at the initiation step. Three SNARE proteins may thus be the molecular substrate underlying the tight coupling between exo- and endocytosis; that is, exocytosis is quickly followed by endocytosis with a similar amount a widely observed phenom1418 Cell Reports 3, , May 30, 2013 ª2013 The Authors Figure 3. Block of Syntaxin Inhibits Endocytosis Induced by depol20msx10 and Block of SNAP-25 or Syntaxin Slows the RRP Replenishment (A C) Similar arrangements as Figures 1A 1C, respectively, except using boiled BoNT/C ( nm, n = 7, black) and BoNT/C ( nm, n = 14, red). Data were taken at 4 10 min after break-in. (D) DCm induced by each 20 ms depolarization during depol20msx10 in the presence of BoNT/E (upper, red), BoNT/A (middle, red), BoNT/C (lower, red), and/or their corresponding control (boiled toxin, black). Traces are taken from the averaged traces in Figures 2A, 2D, and 3A, respectively, and the DCm induced by the first 20 ms depolarization was normalized to be the same. enon essential in recycling vesicles and maintaining the membrane homeostasis. The need of SNARE proteins in endocytosis may prevent futile endocytosis when the action potential-induced calcium influx, which triggers rapid, slow, and bulk endocytosis (Wu et al., 2009; Hosoi et al., 2009; Clayton and Cousin, 2009; but see von Gersdorff and Matthews, 1994; Leitz and Kavalali, 2011), fails to evoke SNARE-mediated exocytosis at nerve terminals with a low release probability. Thus, we suggest that calcium influx is not the only requirement for initiating endocytosis; SNARE proteins are also needed. Many studies provide clues as to how SNARE proteins are involved in endocytosis. For example, the N-terminal half of the SNARE motif of synaptobrevin binds to ANTH domain of endocytic adaptors AP180 and clathrin assembly lymphoid myeloid leukemia (CALM), both of which are involved in endocytosis (Koo et al., 2011; Miller et al., 2011). SNAP-25 binds to intersectin, an endocytic protein, as strong as its binding with syntaxin (Okamoto et al., 1999). This binding is significantly weakened by BoNT/A or BoNT/E that cleaves SNAP-25 (Okamoto et al., 1999), which might explain why BoNT/A and BoNT/E blocked endocytosis (Figures 2 and 4). Syntaxin may bind dynamin, a GTPase-mediating vesicle fission (Galas et al., 2000). It would be of great interest to understand how SNARE proteins participate in endocytosis in the future. We showed that in the presence of TeTx or BoNTs that cleave SNARE proteins, the remaining SNARE proteins support the remaining exocytosis, but not endocytosis. Here, we consider two mechanisms that may account for this observation. First, cleavage of SNARE proteins may not cause an all-or-none block of exocytosis. When SNAP-25 or synaptobrevin is cleaved at calyces, the release probability of RRP vesicles is reduced (Sakaba et al., 2005). However, all RRP vesicles can be released by ms depolarization, suggesting that fewer copies of SNARE proteins per vesicle could still support exocytosis (Sakaba et al., 2005). Indeed, two copies of synaptobrevin are sufficient to support exocytosis (Sinha et al., 2011). If endocytosis needs more copies of SNARE complexes than exocytosis, TeTx and BoNTs may reduce the SNARE protein number per

6 A B C D E F G H I J Figure 4. Block of Synaptobrevin and SNAP-25 Inhibits Slow Endocytosis Induced by depol 20ms (A) Sampled (single traces, left two panels) and averaged (right three panels) C m changes induced by depol 20ms with a pipette containing boiled TeTx (2 mm, black; n = 5) or TeTx (2 mm, red; n = 5). The averaged traces are also normalized and superimposed (right). Traces were taken at 4 10 min after break-in (applies to A D and I J). Scale bars apply to all C m traces in Figure 4 (except normalized traces). (B) Rate decay_n (Rate) and DC m induced by depol 20ms in the presence of boiled TeTx (2 mm, black, n = 5) or TeTx (2 mm, red, n = 5). Data were normalized to the mean of the boiled TeTx group. (C and D) Similar to (A) and (B), respectively, except using Syb sp (2 mm, n = 4, black) and Syb p (2 mm, n = 6, red). (E H) Similar to (A) (D), respectively, except that the data were taken at 2 4 min after break-in. (I J) Similar to (A) and (B), respectively, except using boiled BoNT/A (1 mm; n = 4, black) and BoNT/A (1 mm; n = 6, red). Data were taken at 4 10 min after break-in. vesicle and thus inhibit endocytosis, but not necessarily the DC m induced by depol 20ms or depol 20msX10 that depletes the RRP. Second, TeTx and BoNTs cleave only free SNAREs, but not the SNARE complex (Niemann et al., 1994). After mediating exocytosis, the SNARE complex is disassembled into individual SNARE proteins (Sudhof, 2004), which becomes accessible to TeTx and BoNTs. Cleavage of these individual SNARE proteins after exocytosis by TeTx and BoNTs may thus cause inhibition of endocytosis without affecting the preceding exocytosis. These two mechanisms may explain why TeTx, Syb p, or BoNTs inhibited endocytosis in the presence of exocytosis, including at early dialysis times when DC m was not reduced (Figures 1G 1J and 4E 4H). Our finding that block of SNAP-25 or syntaxin slowed down the RRP replenishment (Figure 3D) suggests the involvement of SNARE proteins in the RRP replenishment. This mechanism might be mediated via the involvement of SNARE proteins in endocytosis, because endocytosis may facilitate the RRP replenishment by clearance of the exocytosed vesicle membrane and proteins at the active zones (Kawasaki et al., 2000; Hosoi et al., 2009; Wu et al., 2009). Alternatively, it might be due to a direct role of SNARE proteins in docking and priming Cell Reports 3, , May 30, 2013 ª2013 The Authors 1419

7 of vesicles. Recent studies show that a pre-existing vesicle pool can be readily retrieved at hippocampal synapses (Fernández- Alfonso et al., 2006; Wienisch and Klingauf, 2006; Hua et al., 2011) and calyces (Xue et al., 2012a). Our finding that block of SNARE proteins may nearly abolish endocytosis (e.g., Figures 1A 1C, 4A, 4B, and 4I 4J) supports the possibility that SNARE proteins are involved in retrieving readily retrievable vesicles. It would be of interest to further explore this possibility in the future. EXPERIMENTAL PROCEDURES Animal care and use were carried out in accordance with US National Institutes of Health (NIH) guidelines and approved by the National Institute of Neurological Disorders and Stroke Animal Care and Use Committee. Slice Preparation, Capacitance Recordings, and Solutions Slice preparation and capacitance recordings were described previously (Wu et al., 2009). Briefly, parasagittal brainstem slices (200 mm thick) containing calyces of Held were prepared from 7- to 10-day-old male or female Wistar rats. Whole-cell capacitance measurements were made with the EPC-9 amplifier (HEKA, Lambrecht, Germany). The sinusoidal stimulus frequency was 1,000 Hz with a peak-to-peak voltage %60 mv. We pharmacologically isolated Ca 2+ currents with a bath solution (22 C 24 C) containing (in mm): 105 NaCl, 20 TEA-Cl, 2.5 KCl, 1 MgCl 2, 2 CaCl 2, 25 NaHCO 3, 1.25 NaH 2 PO 4, 25 glucose, 0.4 ascorbic acid, 3 myo-inositol, 2 sodium pyruvate, tetrodotoxin (TTX), 0.1 3,4-diaminopyridine, mosm (ph 7.4) when bubbled with 95% O 2 and 5% CO 2. The pipette contained (in mm): 125 Cs-gluconate, 20 CsCl, 4 MgATP, 10 Na 2 -phosphocreatine, 0.3 GTP, 10 HEPES, 0.05 BAPTA, mosm (ph 7.2), adjusted with CsOH. Synaptobrevin peptide (SATAATVPPAAPAGEFFPPAPPPNLT) and scrambled synaptobrevin peptide (FPTAPAPASNPALPFTGPTAPAVEAP) were purchased from 21st Century Biochemicals (Marlboro, MA). Data analysis The statistical test was t test. Means are presented as ± SEM. Rate decay and Rate decay_n were measured during s after depol 20msX10 and during s after depol 20ms. For comparison of the Rate decay_n in the presence of a toxin or a boiled toxin, data were normalized to the mean Rate decay_n for the boiled-toxin group. C m baseline drift rate was usually less than 5% 10% of the Rate decay, and thus did not significantly affect the Rate decay measurement. If the drift was >20%, which was infrequent, data were discarded. LICENSING INFORMATION This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ACKNOWLEDGMENTS We thank Dr. Peter Wen for reading of the manuscript. This work was supported by the National Institute of Neurological Disorders and Stroke Intramural Research Program. Received: July 27, 2012 Revised: November 12, 2012 Accepted: March 8, 2013 Published: May 2, 2013 REFERENCES Artalejo, C.R., Henley, J.R., McNiven, M.A., and Palfrey, H.C. (1995). Rapid endocytosis coupled to exocytosis in adrenal chromaffin cells involves Ca2+, GTP, and dynamin but not clathrin. Proc. Natl. Acad. Sci. USA 92, Balaji, J., Armbruster, M., and Ryan, T.A. (2008). Calcium control of endocytic capacity at a CNS synapse. J. Neurosci. 28, Beutner, D., Voets, T., Neher, E., and Moser, T. (2001). Calcium dependence of exocytosis and endocytosis at the cochlear inner hair cell afferent synapse. Neuron 29, Bronk, P., Deák, F., Wilson, M.C., Liu, X., Südhof, T.C., and Kavalali, E.T. (2007). Differential effects of SNAP-25 deletion on Ca2+ -dependent and Ca2+ -independent neurotransmission. J. Neurophysiol. 98, Clayton, E.L., and Cousin, M.A. (2009). The molecular physiology of activity-dependent bulk endocytosis of synaptic vesicles. J. Neurochem. 111, Cornille, F., Deloye, F., Fournié-Zaluski, M.C., Roques, B.P., and Poulain, B. (1995). Inhibition of neurotransmitter release by synthetic proline-rich peptides shows that the N-terminal domain of vesicle-associated membrane protein/ synaptobrevin is critical for neuro-exocytosis. J. Biol. Chem. 270, Deák, F., Schoch, S., Liu, X., Südhof, T.C., and Kavalali, E.T. (2004). Synaptobrevin is essential for fast synaptic-vesicle endocytosis. Nat. Cell Biol. 6, Dittman, J., and Ryan, T.A. (2009). Molecular circuitry of endocytosis at nerve terminals. Annu. Rev. Cell Dev. Biol. 25, Fernández-Alfonso, T., Kwan, R., and Ryan, T.A. (2006). Synaptic vesicles interchange their membrane proteins with a large surface reservoir during recycling. Neuron 51, Galas, M.C., Chasserot-Golaz, S., Dirrig-Grosch, S., and Bader, M.F. (2000). Presence of dynamin syntaxin complexes associated with secretory granules in adrenal chromaffin cells. J. Neurochem. 75, Granseth, B., Odermatt, B., Royle, S.J., and Lagnado, L. (2009). Comment on The dynamic control of kiss-and-run and vesicular reuse probed with single nanoparticles. Science 325, 1499, author reply Hallermann, S., Pawlu, C., Jonas, P., and Heckmann, M. (2003). A large pool of releasable vesicles in a cortical glutamatergic synapse. Proc. Natl. Acad. Sci. USA 100, He, L., and Wu, L.G. (2007). The debate on the kiss-and-run fusion at synapses. Trends Neurosci. 30, He, Z., Fan, J., Kang, L., Lu, J., Xue, Y., Xu, P., Xu, T., and Chen, L. (2008). Ca2+ triggers a novel clathrin-independent but actin-dependent fast endocytosis in pancreatic beta cells. Traffic 9, Hosoi, N., Holt, M., and Sakaba, T. (2009). Calcium dependence of exo- and endocytotic coupling at a glutamatergic synapse. Neuron 63, Hsu, S.-F., and Jackson, M.B. (1996). Rapid exocytosis and endocytosis in nerve terminals of the rat posterior pituitary. J. Physiol. 494, Hua, Y., Sinha, R., Thiel, C.S., Schmidt, R., Hüve, J., Martens, H., Hell, S.W., Egner, A., and Klingauf, J. (2011). A readily retrievable pool of synaptic vesicles. Nat. Neurosci. 14, Jockusch, W.J., Praefcke, G.J., McMahon, H.T., and Lagnado, L. (2005). Clathrin-dependent and clathrin-independent retrieval of synaptic vesicles in retinal bipolar cells. Neuron 46, Kawasaki, F., Hazen, M., and Ordway, R.W. (2000). Fast synaptic fatigue in shibire mutants reveals a rapid requirement for dynamin in synaptic vesicle membrane trafficking. Nat. Neurosci. 3, Koo, S.J., Markovic, S., Puchkov, D., Mahrenholz, C.C., Beceren-Braun, F., Maritzen, T., Dernedde, J., Volkmer, R., Oschkinat, H., and Haucke, V. (2011). SNARE motif-mediated sorting of synaptobrevin by the endocytic adaptors clathrin assembly lymphoid myeloid leukemia (CALM) and AP180 at synapses. Proc. Natl. Acad. Sci. USA 108, Leitz, J., and Kavalali, E.T. (2011). Ca 2+ influx slows single synaptic vesicle endocytosis. J. Neurosci. 31, Cell Reports 3, , May 30, 2013 ª2013 The Authors

8 Miller, S.E., Sahlender, D.A., Graham, S.C., Höning, S., Robinson, M.S., Peden, A.A., and Owen, D.J. (2011). The molecular basis for the endocytosis of small R-SNAREs by the clathrin adaptor CALM. Cell 147, Neves, G., Gomis, A., and Lagnado, L. (2001). Calcium influx selects the fast mode of endocytosis in the synaptic terminal of retinal bipolar cells. Proc. Natl. Acad. Sci. USA 98, Niemann, H., Blasi, J., and Jahn, R. (1994). Clostridial neurotoxins: new tools for dissecting exocytosis. Trends Cell Biol. 4, Okamoto, M., Schoch, S., and Südhof, T.C. (1999). EHSH1/intersectin, a protein that contains EH and SH3 domains and binds to dynamin and SNAP-25. A protein connection between exocytosis and endocytosis? J. Biol. Chem. 274, Royle, S.J., and Lagnado, L. (2003). Endocytosis at the synaptic terminal. J. Physiol. 553, Sakaba, T., Stein, A., Jahn, R., and Neher, E. (2005). Distinct kinetic changes in neurotransmitter release after SNARE protein cleavage. Science 309, Sankaranarayanan, S., and Ryan, T.A. (2000). Real-time measurements of vesicle-snare recycling in synapses of the central nervous system. Nat. Cell Biol. 2, Sinha, R., Ahmed, S., Jahn, R., and Klingauf, J. (2011). Two synaptobrevin molecules are sufficient for vesicle fusion in central nervous system synapses. Proc. Natl. Acad. Sci. USA 108, Sudhof, T.C. (2004). The synaptic vesicle cycle. Annu. Rev. Neurosci. 27, Sun, J.Y., Wu, X.S., and Wu, L.G. (2002). Single and multiple vesicle fusion induce different rates of endocytosis at a central synapse. Nature 417, von Gersdorff, H., and Matthews, G. (1994). Inhibition of endocytosis by elevated internal calcium in a synaptic terminal. Nature 370, Wienisch, M., and Klingauf, J. (2006). Vesicular proteins exocytosed and subsequently retrieved by compensatory endocytosis are nonidentical. Nat. Neurosci. 9, Wu, L.G., and Betz, W.J. (1996). Nerve activity but not intracellular calcium determines the time course of endocytosis at the frog neuromuscular junction. Neuron 17, Wu, X.S., and Wu, L.G. (2009). Rapid endocytosis does not recycle vesicles within the readily releasable pool. J. Neurosci. 29, Wu, W., Xu, J., Wu, X.S., and Wu, L.G. (2005). Activity-dependent acceleration of endocytosis at a central synapse. J. Neurosci. 25, Wu, X.S., McNeil, B.D., Xu, J., Fan, J., Xue, L., Melicoff, E., Adachi, R., Bai, L., and Wu, L.G. (2009). Ca(2+) and calmodulin initiate all forms of endocytosis during depolarization at a nerve terminal. Nat. Neurosci. 12, Xu, J., McNeil, B., Wu, W., Nees, D., Bai, L., and Wu, L.G. (2008). GTP-independent rapid and slow endocytosis at a central synapse. Nat. Neurosci. 11, Xue, L., McNeil, B.D., Wu, X.S., Luo, F., He, L., and Wu, L.G. (2012a). A membrane pool retrieved via endocytosis overshoot at nerve terminals: a study of its retrieval mechanism and role. J. Neurosci. 32, Xue, L., Zhang, Z., McNeil, B.D., Luo, F., Wu, X.S., Sheng, J., Shin, W., and Wu, L.G. (2012b). Voltage-dependent calcium channels at the plasma membrane, but not vesicular channels, couple exocytosis to endocytosis. Cell Rep. 1, Yamashita, T., Hige, T., and Takahashi, T. (2005). Vesicle endocytosis requires dynamin-dependent GTP hydrolysis at a fast CNS synapse. Science 307, Zhang, Q., Li, Y., and Tsien, R.W. (2009). The dynamic control of kiss-andrun and vesicular reuse probed with single nanoparticles. Science 323, Cell Reports 3, , May 30, 2013 ª2013 The Authors 1421