Abstract
The synaptic protein interaction (synprint) site of the voltage-gated Ca2+ channel (VGCC) α1 subunit can interact with proteins involved in exocytosis, and it is therefore thought to be essential for exocytosis of synaptic vesicles. Here we report that the synprint site can also directly bind the μ subunit of AP-2, an adaptor protein for clathrin-mediated endocytosis, in competition with the synaptotagmin 1 (Syt 1) C2B domain. In brain lysates, the AP-2–synprint interaction occurred over a wide range of Ca2+ concentrations but was inhibited at high Ca2+ concentrations, in which Syt 1 interacted with synprint site. At the calyx of Held synapse in rat brainstem slices, direct presynaptic loading of the synprint fragment peptide blocked endocytic, but not exocytic, membrane capacitance changes. We propose that the VGCC synprint site is involved in synaptic vesicle endocytosis, rather than exocytosis, in the nerve terminal, via Ca2+-dependent interactions with AP-2 and Syt.
Introduction
In the nerve terminal, P/Q and N subtypes of voltage-gated Ca2+ channels (VGCCs) play pivotal roles in neurotransmitter release by providing pathways for Ca2+ entry (Luebke et al., 1993; Takahashi and Momiyama, 1993). VGCCs are also thought to play an additional role in synaptic vesicle exocytosis through direct interactions with syntaxin (Stx), SNAP-25, and synaptotagmin (Syt) at the so-called synprint site in a large intracellular loop between transmembrane domains II and III (LII–III) of the VGCC α1 subunit (Sheng et al., 1996; Catterall, 2000). This synprint hypothesis has been supported by the finding that loading a synprint fragment peptide into presynaptic neurons inhibits synaptic transmission (Mochida et al., 1996; Rettig et al., 1997). However, the synprint fragment can also block the interaction between Syt 1 and AP-2 (Chapman et al., 1998), and Syt can participate in vesicle endocytosis (Poskanzer et al., 2003; Llinás et al., 2004) as well as exocytosis (Geppert et al., 1994). Furthermore, synaptic transmission can eventually be blocked by vesicle depletion after sustained block of endocytosis (Koenig and Ikeda, 1989; Yamashita et al., 2005). Thus, it remains possible that VGCC synprint might also be involved in vesicle endocytosis. Indeed, in cultured secretory cells, overexpression of a synprint site-deleted VGCC mutant protein inhibits both exocytic and endocytic membrane capacitance changes (Harkins et al., 2004). While screening for proteins that interact with the VGCC synprint site, we found that AP-2μ subunit can bind to it directly. This interaction competed with the Syt–synprint interaction (Chapman et al., 1998) in a Ca2+ concentration-dependent manner. When loaded into the calyx of Held presynaptic terminal in acute slices of rat brainstem, the synprint fragment peptide nearly abolished endocytic changes of membrane capacitance but surprisingly had no effect on exocytic capacitance changes. These results taken together point toward an endocytic, rather than an exocytic, role for the VGCC synprint site in vesicle recycling.
Materials and Methods
All experiments were performed in accordance with the guidelines of the Physiological Society of Japan.
Materials.
cDNAs encoding rabbit N-type (GenBank accession number D14157), P/Q-type (GenBank accession number X57477), and L-type (GenBank accession number X15539) VGCCs were used for cloning deletion fragments. cDNAs encoding rat AP-2μ1 subunit (GenBank accession number M23674), Syt 1 (GenBank accession number X52772), and Stx 1a (GenBank accession number NM_053788) were obtained using reverse transcription-PCR from 7-d-old rat whole brains. The following vectors were used: pGEX4T-1 (GE Healthcare), pMAL–C2E (New England Biolabs), pCMV–Tag3 (Stratagene), and pEGFP–C (Clontech). The following antibodies were used: anti-CaV2.2 antibody (Alomone Labs); anti-β4 antibody (Kiyonaka et al., 2007); monoclonal antibody VD21, which recognizes all Ca2+ channel β subunits (Sakamoto and Campbell, 1991); anti-MBP antibody and anti-α-adaptin (clone 100/2) antibody (Sigma); anti-c-Myc antibody (9E10); anti-AP-2 antibody (AP-6; Calbiochem); anti-gamma adaptin (AP-1), anti-β-adaptin, and anti-AP-50 (AP-2μ) antibodies (BD Transduction Laboratories); anti-Syt antibody (SYA148; Nventa Biopharmaceuticals); anti-Stx1 antibody (MBL International); and anti-enhanced green fluorescent protein (EGFP) antibody (Invitrogen). Recombinant fusion proteins were purified according to the protocol of the manufacturer. Glutathione S-transferase (GST)-fusion and MBP-fusion proteins were dialyzed against 20 mm Tris, 1 mm EDTA, and 1 mm DTT, pH 7.5, at 4°C. To prepare rat brain lysate, whole brains of 7-d-old rats were homogenized in 20 mm Tris, 1 mm EGTA, 150 mm NaCl, 1% NP-40, and a protease inhibitor cocktail (Roche). Homogenates were then sonicated and centrifuged at 3000 × g for 10 min. Supernatants were collected and further centrifuged at 10,000 × g for 20 min to be used for the following experiments.
Biochemistry.
In affinity column chromatography, GST-fusion proteins were immobilized on glutathione Sepharose 4B (GSH) beads (GE Healthcare) and packed into a column. Brain lysates prepared from 7-d-old Wistar rats were passed through a filter (0.45 μm), and the filtrated lysates were passed through a GST column before they were applied to sample columns. After washing the sample columns with a buffer solution (20 mm Tris, 1 mm EDTA, 150 mm NaCl, and 1% NP-40), binding proteins were eluted by a 500 mm NaCl containing buffer solution. Eluted samples were treated with SDS sample buffer (60 mm Tris, 2% SDS, 10% glycerol, and 250 mm 2-mercaptoethanol) and submitted to silver staining or immunoblot analysis. In vitro binding assays were performed using GST-fusion proteins immobilized onto GSH-beads in separate tubes. The immobilized beads were incubated with MBP-fusion proteins in buffer with 1 mm DTT and 0.2 mg/ml BSA at 4°C for 2 h. After washing the beads four times with the buffer, they were suspended in SDS buffer. The bead-bound proteins were subjected to immunoblot analysis with anti-MBP antibody. Ca2+ buffers of different concentrations were prepared using 1 mm EGTA according to the online protocol (http://www.stanford.edu/∼cpatton/maxc.html). For urea denature of brain lysate (Fig. 1 C), brain lysates (6 mg/ml) were incubated with 7.5 m urea for 15 min at room temperature and then diluted fivefold into PBS containing 1% Triton X-100. Beads were washed five times with 4 vol of buffer. For immunoprecipitation assay (Fig. 1 F), anti-AP-2 antibody (4 μg) was immobilized on protein A agarose beads and then incubated with 800 μl aliquots of the heparin-purified samples (1 mg of proteins) for 4 h at 4°C. Native VGCC complexes were partially purified from C57BL/6 mouse whole brain and submitted to immunoprecipitation assays as reported previously (Kiyonaka et al., 2007). To disrupt the native AP-2–VGCC association, partially purified VGCC complexes were incubated with 450 nm GST–N-synprint (718-963 aa) at 4°C for 8 h. Mass spectrometry analysis was performed as reported previously (Nishimura et al., 2006).
Electrophysiology.
Whole-cell recordings and membrane capacitance (C m) measurements were performed at room temperature (22–25°C) at the calyx of Held presynaptic terminals in auditory brainstem slices (200 μm thick) containing the medial nucleus of trapezoid body prepared from 7- to 8-d-old Wistar rats as described previously (Yamashita et al., 2005). The extracellular solution contained 115 mm NaCl, 10 mm tetraethylammonium Cl, 2.5 mm KCl, 1 mm MgCl2, 2 mm CaCl2, 26 mm NaHCO3, 1.25 mm NaH2PO4, 10 mm glucose, 0.5 mm ascorbic acid, 3 mm myo-inositol, 2 mm sodium pyruvate, 1 μm tetrodotoxin, 0.5 mm 4-aminopyridine, 10 μm bicuculline methiodide, and 0.5 μm strychnine hydrochloride (pH 7.4 when bubbled with 95% O2 and 5% CO2). Patch pipettes (6–8 MΩ) had a series resistance of 11–27 MΩ, which was compensated by up to 75% for its final value to be 6.7–7.0 MΩ. Pipette solution contained the following (in mm): 109 CsCl, 40 HEPES, 0.5 EGTA, 1 MgCl2, 12 Na2-phosphocreatine, 3 Mg-ATP, and 0.3 Na-GTP (pH 7.3–7.4 adjusted with CsOH). GST–N-synprint peptide or GST–LII–III peptide was included in the pipette solution, together with trehalose (0.2–1 mm) to minimize protein aggregations (Singer and Lindquist, 1998). These proteins (5 μm) were infused into presynaptic terminals, by gently applying blows of positive pressure, while the series resistance was maintained below 12 MΩ, 10 min before the start of recordings. The C m analysis was made as described previously (Yamashita et al., 2005). Briefly, C m changes within 200 ms after depolarizing pulse were excluded from analysis to avoid contaminations of conductance-dependent capacitance artifacts. The ΔC m amplitude was measured as a difference between the baseline and at 200–250 ms after depolarization. For C m > 40 s, regression line, obtained from baseline 0.005–10 s before stimulation, was subtracted (Yamashita et al., 2005).
Results
Direct interaction of VGCC synprint with AP-2μ
To identify proteins that interact with VGCC synprint site, we applied juvenile rat brain tissue lysate to an affinity column with GST-fused N-type VGCC synprint fragment (N-synprint). Proteins with the relative molecular masses of 110 and 100 kDa specifically bound to N-synprint (Fig. 1 A). Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry analysis indicated that these proteins were the AP-2αA and AP-2αC components of the clathrin adaptor AP-2 complex. Affinity columns showed binding of AP-2 complex to N-synprint, as well as to the synprint fragment of P/Q-type VGCC (P/Q-synprint) but not to the II–III loop domain of L-type VGCC (L–LII–III) (Fig. 1 B). In contrast, AP-1, a related adaptor protein, showed no binding to these fragments. Pull-down assays using urea-denatured brain lysate (Haucke et al., 2000) showed that N-synprint bound to the AP-2μ subunit but not to the AP-2α or AP-2β subunits (Fig. 1 C). MBP-fused recombinant AP-2μ showed direct binding to N- or P/Q-synprint, as well as to Syt 1 C2B domain (Fig. 1 D). The binding of AP-2μ to N-synprint was concentration dependent, with its K d being estimated, by scatchard analysis, to be 30.8 ± 5.6 nm (n = 3) (Fig. 1 E). Thus, AP-2μ subunit directly binds to the VGCC synprint site in vitro. In purified mouse brain lysates (Kiyonaka et al., 2007), native AP-2 could coimmunoprecipitate with CaV2.2 or VGCC β-subunit complexes (Fig. 1 F) but not with CaV1.2 complex (supplemental Fig. S1, available at www.jneurosci.org as supplemental material). These interactions could be blocked by addition of GST–N-synprint (Fig. 1 F). These results suggest that native AP-2 and VGCCs in the brain interact with each other.
Competitive binding of AP-2 and Syt 1 to VGCC
We next searched for the binding regions of AP-2μ and N-synprint in pull-down assays. MBP-tagged AP-2μ showed binding to the N-synprint amino acid fragments 718–859, 832–963, and 718–963 but not to 718–820 (Fig. 2 A). These regions are in common with those involved in the binding of N-synprint to Syt 1 and Stx 1 (Fig. 2 A). Among AP-2μ deletion mutants, AP-2μ 283–394 showed binding to N-synprint (Fig. 2 B). This fragment also showed bindings to Syt 1 C2B, as reported previously (Haucke et al., 2000). However, mutating AP-2μ tyrosine-344 and lysine-354 to alanine (Y344A and K354A) impaired the binding of AP-2μ 123–435 to N- and P/Q-synprint (Fig. 2 C), as well as to Syt 1 C2B (Haucke et al., 2000), albeit to a lesser extent (Fig. 2 C). In binding assay, Syt 1 competed off the AP-2μ–N-synprint binding and vice versa (Fig. 2 D). It has been reported that mutations of Syt1 C2B domain at lysine-326 and alanine-327 abolishes bindings between AP-2, Syt1 C2B, and the synprint site (Chapman et al., 1998). Thus, AP-2μ, Syt 1 C2B, and the synprint site share common binding regions for their mutual interactions. In contrast, no competition was observed between Stx 1 and AP-2μ for binding to the synprint site (Fig. 2 E) (supplemental Fig. S2, available at www.jneurosci.org as supplemental material).
Ca2+ dependence of VGCC/AP-2/Syt 1 interaction
The interaction between the VGCC synprint site and Stx is Ca2+ dependent (Sheng et al., 1996). To test whether the interactions between N-synprint and AP-2 or Syt 1 were also Ca2+ dependent, we performed pull-down assays at different Ca2+ concentrations ([Ca2+]). AP-2μ showed binding to N-synprint at a wide range of [Ca2+]. However, the binding decreased as [Ca2+] was elevated and became undetectable at 2 mm [Ca2+] (Fig. 3 A). In contrast, Syt 1 showed binding to N-synprint only at >100 μm [Ca2+] (Fig. 3 A). When Ca2+ .was replaced by Sr2+ or Ba2+, N-synprint still showed binding to AP-2 but no longer to Syt 1 (Fig. 3 B), and no concentration dependence was observed for Sr2+ or Ba2+ for the AP-2–synprint interaction. In the absence of brain lysate, however, binding of recombinant AP-2μ to N-synprint showed no Ca2+ dependence (Fig. 3 C). Given that the synprint–Syt 1 interaction is strengthened by Ca2+-dependent oligomerization of Syt 1 at 50 μm to 1 mm [Ca2+] (Chapman et al., 1998; Wu et al., 2003; Grass et al., 2004), the apparent Ca2+ dependence of the AP-2μ–synprint interaction in brain lysate likely arose from competition between AP-2μ and Syt 1 for binding to the synprint site (Fig. 3 C). However, it cannot be ruled out that an additional factor contained in brain lysate might confer the Ca2+ dependence.
Involvement of VGCC synprint in endocytosis
An N-type VGCC synprint site fragment (N-synprint) blocks synaptic transmission when injected into cultured presynaptic neurons (Mochida et al., 1996; Rettig et al., 1997). This effect could, in principle, be caused by a block of exocytosis, endocytosis, or a combined impairment of these processes. This peptide fragment could also inhibit multiple intermolecular interactions, in which VGCC synprint and its interacting proteins are involved. To reexamine the effect of this peptide, we loaded N-synprint fragment (832–963 aa, 5 μm) into the calyx of Held presynaptic terminals and tested whether exocytosis and/or endocytosis of synaptic vesicles were affected, using C m measurements (Sun and Wu, 2001; Yamashita et al., 2005). To minimize secondary effects caused by changes in vesicle recycling (Yamashita et al., 2005), we restricted our analysis of C m changes (ΔC m) to the first to fifth events induced by presynaptic Ca2+ currents (I pCa) 10–20 min after peptide loading. In this condition, the N-synprint fragment slightly increased exocytic ΔC m, concomitantly with an increase in I pCa amplitude (Fig. 4 A,B). In addition, N-synprint fragment loading markedly blocked endocytic C m decay (Fig. 4 C). Loading of GST–L–LII–III fragment had no such effects. The facilitatory effect of synprint fragment on I pCa (Fig. 4 A,B) is consistent with the idea that interaction of Stx with the synprint site regulates I pCa (Stanley and Mirotznik, 1997). Consistent with this interpretation, readjustment of the I pCa amplitude in N-synprint fragment-loaded terminals back to the control level abolished the exocytic ΔC m increase. Thus, the target of the N-synprint fragment at this nerve terminal was I pCa and vesicle endocytosis but not vesicle exocytosis.
Discussion
In the nerve terminal, synaptic vesicles are docked on an active zone of 200–300 nm in diameter, in which VGCCs are thought to be inserted within a distance of 20–200 nm from synaptic vesicles (Neher, 1998; Meinrenken et al., 2002). Ca2+ entry through 1–60 VGCCs (Stanley, 1993; Borst and Sakmann, 1996) can trigger single vesicle fusion for exocytic release of neurotransmitter. After fusion, a variety of endocytic proteins are assembled for clathrin-mediated vesicle endocytosis, in which the adaptor protein AP-2 plays an essential role (Schmid, 1997). Our present results suggest that VGCC synprint sites anchor a fraction of AP-2 complex, which is likely linked to the plasma membrane via interactions with phosphatidylinositol 4,5-bisphosphate (McPherson et al., 2008). This synprint-AP-2 interaction occurs for both N- and P/Q-type VGCCs but is not entirely universal because VGCCs in invertebrates such as Drosophila and Caenorhabditis elegans lack synprint sites (Littleton and Ganetzky, 2000; Spafford et al., 2003). At Drosophila neuromuscular junctions, the AP-2 complex forms a network surrounding the active zone (González-Gaitán and Jäckle, 1997), in which clathrin-mediated vesicle endocytosis occurs (Roos and Kelly, 1999). At the amphibian neuromuscular junction, clathrin-mediated vesicle endocytosis occurs in the area close to active zone, supporting the existence of local exo-endocytic cycling pool of vesicles (Teng and Wilkinson, 2000). It remains to be seen where exactly VGCCs are localized in the active zone. With respect to the coupling role of VGCCs for exocytosis and endocytosis of synaptic vesicles, it might be most efficient if they are localized at the periphery of the active zone.
Clathrin coating of vesicles is initiated by the interaction of AP-2 with Syt (Schmid, 1997). Their interactions are supposedly promoted by stonin, which binds to both AP-2 and Syt (Diril et al., 2006). Given that the VGCC synprint site can bind to Syt only at high Ca2+ concentrations (Fig. 3), it may be speculated that VGCCs sequentially interact with AP-2 and Syt during transient increases in Ca2+ concentration in and around the active zone. These interactions may promote the dynamic assembly of AP-2 and Syt. In this regard, Ca2+-dependent binding of the VGCC C terminal to endophilin, an endocytic protein that interacts with dynamin, has been suggested previously to promote vesicle endocytosis (Chen et al., 2003).
The importance of Ca2+ in slow vesicle endocytosis is variable between preparations. At immature calyces of Held, it has been reported recently that slow vesicle endocytosis, lasting tens of seconds, requires a high concentration of Ca2+, which can only be attained within a Ca2+ microdomain (Hosoi et al., 2009). Given that the Ca2+ concentration in the microdomain returns to the baseline within 10 ms (Neher, 1998), Ca2+ must play only a triggering role for the subsequent slow steps toward vesicle endocytosis. In our present study, intraterminal loading of the synprint fragment blocked endocytic capacitance changes but spared an initial recovery phase lasting several seconds (Fig. 4). A similar blocking effect, which spares the initial phase of endocytosis, has been observed for dynamin 1 proline-rich domain peptide (Yamashita et al., 2005). Thus, the initial endocytic component might be independent of clathrin and/or dynamin. In HeLaM cells, clathrin-mediated endocytosis can still be observed after ablating AP-2 proteins down to undetectable levels (Motley et al., 2003). Thus, also at the calyx of Held, AP-2-independent clathrin-mediated endocytosis might underlie the initial phase of endocytosis.
At resting Ca2+ concentrations, VGCC synprint sites can directly bind to AP-2μ (Figs. 1, 3), whereas at higher Ca2+ concentrations, they can bind to Stx (Sheng et al., 1996) and Syt 1 (Fig. 3) (Grass et al., 2004). Syt 1 and AP-2 also bind to each other (Zhang et al., 1994; Haucke et al., 2000; Grass et al., 2004). Thus, the synprint fragment peptide loaded into nerve terminals must interfere with the triplet interactions between the VGCC synprint site, AP-2μ, and Syt 1 C2B. Syt 1 C2B is thought to regulate the endocytic rate (Poskanzer et al., 2006). Genetic ablation of Syt 1 or its C2B domain slows the endocytic rate at cultured hippocampal synapses (Nicholson-Tomishima and Ryan, 2004) and at Drosophila neuromuscular junction (Poskanzer et al., 2006), supporting the endocytic role of Syt1 (Poskanzer et al., 2003; Llinás et al., 2004). Thus, the blocking effect of synprint fragment peptide loading on endocytosis observed at the calyx of Held (Fig. 4) can, at least in part, be explained by its inhibitory effect on the Syt–AP-2 interaction (Haucke et al., 2000). However, compared with mild blocking effects on endocytosis reported for Syt 1-null mutations (Nicholson-Tomishima and Ryan, 2004; Poskanzer et al., 2006), much stronger blocking effects on endocytosis are observed for synprint fragment loaded into the calyx terminal (Fig. 4). This suggests that interactions between the VGCC synprint site, AP-2, and Syt may additionally be involved in vesicle endocytosis (Fig. 3).
Despite the fact that N-synprint can bind to the exocytic proteins Stx and Syt (Fig. 2) (Sheng et al., 1996, 1997), the N-synprint fragment peptide had no direct inhibitory effect on exocytosis at the calyx of Held in acute brainstem slices of immature rats (Fig. 4). Because the peptide blocked endocytosis, it is suggested that previously reported block of transmitter release by N-synprint peptide at cultured synapses (Mochida et al., 1996; Rettig et al., 1997) might be a secondary effect caused by the arrest of vesicle endocytosis and recycling (Koenig and Ikeda, 1989; Yamashita et al., 2005). However, our results cannot preclude the possibility that the VGCC synprint site might become involved in vesicle exocytosis at mature synapses.
Footnotes
-
This study was supported by a Grant-in-Aid for Specially Promoted Research from the Ministry of Education, Culture, Sports, Science, and Technology. We thank Masayoshi Mishina, Takao Shimizu, and Kentaro Shiraki for technical advice, Idy Ko and Masakuni Yagi for technical assistance, and Ervin Johnson, Shigeo Takamori, and Kohji Takei for comments.
- Correspondence should be addressed to Tomoyuki Takahashi, Department of Neurophysiology, Doshisha University Faculty of Life and Medical Sciences, Kyoto 610-0394, Japan. ttakahas{at}oist.jp