Address correspondence and reprint requests to Yves Dunant, Université de Genève, Faculté de Médecine, Centre Médical Universitaire, 1211–Genève 4, Switzerland. E-mail: email@example.com
A low-affinity Ca2+/H+-antiport was described in the membrane of mammalian brain synaptic vesicles. Electrophysiological studies showed that this antiport contributes to the extreme brevity of excitation-release coupling in rapid synapses. Synaptotagmin-1, a vesicular protein interacting with membranes upon low-affinity Ca2+-binding, plays a major role in excitation-release coupling, by synchronizing calcium entry with fast neurotransmitter release. Here, we report that synaptotagmin-1 is necessary for expression of the vesicular Ca2+/H+-antiport. We measured Ca2+/H+-antiport activity in vesicles and granules of pheochromocytoma PC12 cells by three methods: (i) Ca2+-induced dissipation of the vesicular H+-gradient; (ii) bafilomycin-sensitive calcium accumulation and (iii) pH-jump-induced calcium accumulation. The results were congruent and highly significant: Ca2+/H+-antiport activity is detectable only in acidic organelles expressing functional synaptotagmin–1. In contrast, synaptotagmin-1-deficient cells – and cells where transgenically encoded synaptotagmin-1 was acutely photo-inactivated – were devoid of any Ca2+/H+-antiport activity. Therefore, in addition to its previously described functions, synaptotagmin-1 is involved in a rapid vesicular Ca2+ sequestration through a Ca2+/H+ antiport.
It has long been known that synaptic vesicles and other secretory granules take up Ca2+ by ATP-dependent mechanisms (Israël et al. 1980; Michaelson et al. 1980; Parekh 2008). Two distinct mechanisms contribute to transport calcium into vesicles: a high-affinity Ca2+-ATPase pump, and a low-affinity Ca2+/H+-antiport, which was extensively investigated using synaptic vesicles isolated from mammalian brain cortex. The vesicular Ca2+/H+-antiport is activated only at high Ca2+ concentrations (K0.5 = 217 μM; maximum velocity at ~ 500–600 μM). Its activity depends on the ΔpH across the vesicle membrane, and therefore on the vacuolar H+-transporting ATPase (V-ATPase) (Gonçalves et al. 1998, 1999b, 2000).
It is expected that the high-affinity Ca2+-pumps transport Ca2+ ions into vesicles (and into other organelles) when the synapse is in resting state, keeping the cytosolic Ca2+ concentration ([Ca2+]i) at the very low physiological level. As for the Ca2+/H+-antiport, it would operate only at brief instants following a presynaptic action potential, when [Ca2+]i abruptly raises to sub-millimolar levels in discrete spots (nano-domains) close to the inner mouth of voltage-operated calcium channels (VOCCs) (Llinas et al. 1992). Experiments designed for revealing the physiological role of the Ca2+/H+-antiport in rapid synaptic transmission was analysed using the Torpedo nerve electroplaque junction, a modified neuromuscular system. It was found that this antiport contributes to shorten the Ca2+ signal in nano-domains of presynaptic terminals, abbreviating the time course of transmitter release in individual impulses. Indeed, when the Ca2+/H+-antiport is inhibited, either directly by replacing external Ca2+ with Sr2+ or indirectly by using bafilomycin-1 (a specific V-ATPase inhibitor), the duration of transmitter release is prolonged in individual impulses, resulting in a post-synaptic potential lengthening (Dunant et al. 2009; Cordeiro et al. 2011).
Searching for a molecular counterpart of the vesicular Ca2+/H+-antiport activity, we suspected that synaptotagmins might be involved. Synaptotagmins form a family of proteins associated with synaptic vesicles. Synaptotagmin-1 (SYT-1) has a conserved N-terminal inside the vesicle and two C2 domains facing the cytosol, which are able to bind Ca2+ with low affinity (60 μM to 600 μM) (Takamori et al. 2006; Chapman 2008). SYT-1 also binds in a dynamic and Ca2+-dependent manner to the hetero-trimeric protein complex SNARE (VAMP/synaptobrevin + SNAP-25+ syntaxin), to other proteins (complexin, Munc-18, etc.), to VOCCs and to phospholipids. SYT-1 and the above proteins form between vesicles and the plasma membrane a multimolecular complex, which is important for efficient excitation-secretion coupling (Sorensen 2009). In most cases, SYT-1 deletion or inactivation disrupts the tight link between the Ca2+ spark and the synchronized emission of transmitter. However, quantal transmitter release can subsist or even be enhanced in the absence of SYT-1, especially in the form of desynchronized and spontaneous release. SYT-1 was therefore proposed as the Ca2+ sensor for the fast, synchronized release of transmitter, whereas another sensor was suspected to operate in the asynchronous release (Koh and Bellen 2003; Stevens and Sullivan 2003; Young and Neher 2009; Kochubey et al. 2011; Yao et al. 2011; Xu et al. 2012).
Considering that both display low-affinity reactions with Ca2+, and both seem involved in rapid presynaptic excitation-release coupling, we addressed the question of a possible link between SYT-1 and the vesicular Ca2+/H+-antiport. To this end, we measured Ca2+/H+-antiport activity in a vesicle-rich fraction prepared from rat PC12 cells. Pheochromocytoma PC12 cells provide a fundamental model for investigating the molecular tenets of neurosecretion. PC12 cells secrete catecholamines, acetylcholine, ATP and other compounds; they are particularly rich in both electro-lucent and dense-core vesicles (Kasai et al. 1999; Zhang et al. 2011), which can establish a robust, bafilomycin-sensitive H+- gradient (Bloc et al. 1999). Of great importance for this study is the fact that a battery of SYT-1-positive and SYT-1-negative PC12 populations is available (Fig. 1). We started with the spontaneous variants selected by Shoji-Kasai et al. (1992), and then with transfected negative variants to obtain clones where SYT-1 is either present, or acutely and specifically inactivated.
Materials and methods
Ethical and general considerations
The experiments were carried out at the University of Geneva, Switzerland. They did not involve living animals or human beings. Concerning other ethical requirements, this research was conducted in conformity with J. Neurochemistry authors’ guidelines. The concentration of free Ca2+ in solution in the presence of indicated chelators was calculated using WINmaxC 2.05 (C. Patton, Stanford University). Statistical significance of the results was assessed using unpaired Student's t-test.
PC12 cell variants and transfected clones
PC12 cell variants (Shoji-Kasai et al. 1992) were kindly provided by Yoko Shoji-Kasai and Masami Takahashi (Mitsubishi Kagaku Institute of Life Sciences, MITILS, 11 Minamiooya, Machida, Tokyo 194-8511, Japan). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Eurobio, Courtaboeuf, France) supplemented with glutamine (0.2 mM), 5% foetal calf serum, 5% horse serum and penicillin–streptomycin (5 units/L; Invitrogen, Life Technologies, Zug, Switzerland). The cultures were carried out at 37°C in a 5% CO2; 95% air mixture saturated with water in 175 cm2 plastic bottles (Falcon, Fort Worth, TX, USA). For morphological experiments, cultures were grown in monolayer on plastic coverslips (Thermanox; Nunc, Ocala, FL, USA).
Two synaptotagmin-1 genes were engineered for transfection of PC12 cell variant devoid of endogenous SYT-1 (F7(−/−)). Constructs were similar to those described earlier by Marek and Davis (2002), but using the mouse sequence instead of Drosophila. For that the pMH4-SYN-P65-I-EGFP plasmid (corresponding to mouse NM_009306 sequence) was used to amplify (F: AAA ATG GTG AGT GCC AGT CGT; R: TTA CTT CTT GAC AGC CAG CAT; Rtag: TTA AGC ACG AGC ACA TTC ACG ACA AGC TTC ACG AGC AGC TTC AGC CTT CTT GAC AGC CAG CAT) synaptotagmin-1 cDNA without or with the FlAsH-binding tetracysteine motif (AEAAAREACCRECCARA) (Griffin et al. 1998) and subcloned in pcDNA3.1. The constructs were sequenced. These mouse cDNA showed 96.32% homology to rat synaptotagmin sequence and 99.76% homology was found between the two proteins. Cells were transfected with Lipofectamine™ 2000 (Invitrogen), according to the manufacturer's guidelines. After the formation of liposome/DNA complexes they were delivered into cell cultures during ~ 4 h. After this period, cells were cultured in normal growth medium for two more days before being assayed.
Immunofluorescence micrographs of PC12 cells with antibodies against SYT-1 and VAMP-1 (synaptobrevin-1)
Cells were cultured on collagen-coated 35 mm plastic Petri dishes (Nunc) and fixed with 4% (w/v) paraformaldehyde. They were incubated overnight in bovine serum albumin (0.3%)/Tween (0.1%)-containing phosphate buffer saline (PBS) (blocking) solution with mouse MAb5200 anti-SYT-1 antibody diluted 1 : 500 and with Synaptic Systems rabbit anti-VAMP-1 antibody diluted 1 : 500. They were further incubated with fluorescein-conjugated Alexa Fluor 488 anti-mouse IgG (SYT-1 in green; Invitrogen, Sparks, MD, USA) and rhodamine-conjugated Alexa Fluor 555 anti-rabbit IgG (VAMP-1 in red; from Life Tecnologies, Zug, Switzerland) and observed in Zeiss LSM-510 (Carl Zeiss AG, Feldbach, Switzerland) confocal microscope (Fig. 1).
Post-nuclear supernatants (PNS) preparations from PC12 cell clones
Crude PNS were prepared from PC12 cells as described (Bloc et al. 1999). PC12 cells were detached from cultured flasks by washing in PBS composed of the following: 137 mM NaCl, 2.7 mM KCl, 10 mM Na/Na2 Phosphate, pH 7.4 at 37°C (no need for trypsin). Cells were centrifuged at 800 g for 5 min and re-suspended (~ 2.5 mL for every 106 cells) very gently in ice-cold (4°C temperature kept until the end of the procedure) homogenization buffer (HB) composed of 250 mM sucrose, 3 mM imidazole, pH 7.4 and centrifuged again at 800 g. Cells were re-suspended in HB and counted in a Neubauer haemocytometer. They were centrifuged at 1000 g for 10 min. The cell pellet was diluted (~ 0.5 mL for every 106 cells) in HB containing protease inhibitors (10 μM leupeptin and 1 μM pepstatin A) and homogenized by 5–10 passages through a 22-gauge needle. The suspension was monitored by phase contrast microscopy. The homogenate was centrifuged at 2000 g for 15 min and the post nuclear supernatant was collected and stored at −80°C until usage. A small aliquot was taken for protein quantification by the BCA protein assay (Pierce, Rockford, IL, USA).
Cell labelling and fluorescence-assisted light inactivation with FlAsH: FlAsH-FALI
Labelling PC12 cells with the biarsenical derivative of fluorescein, FlAsH, was done by combining the PNS protocol with FlAsH cell loading protocol (Bloc et al. 1999; Marek and Davis 2002). FlAsH reagent was acquired from Invitrogen under the name of Lumio™ Green. It is not fluorescent until it binds the tetracysteine motif at which time it becomes highly fluorescent. The motif consists of Cys-Cys-Xaa-Xaa-Cys-Cys where Cys equals cysteine and Xaa equals any amino acid other than cysteine. This motif is rarely seen in naturally occurring proteins, allowing specific fluorescence labelling of recombinant proteins fused to the tetracysteine motif (Tag). Cells of the transfected clone (F7 + SytTag) are expected to bind FlAsH with high affinity (Gaietta et al. 2002) and guarantee nearly 100% labelling (Marek and Davis 2002). FlAsH is supplied pre-complexed to (1,2-ethanedithiol) EDT2 that solubilizes and stabilizes the molecule. It is membrane permeable, and readily enters cells. Cells were labelled at 20°C either in suspension (after harvesting with PBS as above) or attached to plastic coverslips. All steps occurred under protection from light until the assays. FlAsH loading medium contained the following: 137 mM NaCl, 2.7 mM KCl, 2 mM MgCl2, 0.5 mM CaCl2, 1 mM Na-pyruvate, 2.5 mM glucose, 10 mM Na/Na2 Phosphate, pH 7.4 supplemented with 1 μM FlAsH- EDT2 and 15 μM EDT2 and lasted 20 min. Cells were then washed (in case of cells in suspension, they were centrifuged at 800 g followed by re-suspension) in loading medium supplemented with 250 μM EDT2 to remove non-specific FlAsH binding and incubated under slight agitation for 10 min. After two additional washing steps in the loading medium alone, they were either visualized (Fig. 1) under a fluorescence microscope with a FITC (fluorescein) filter (excitation at 488 nm and emission at 528 nm), or proceeded to obtain PNS preparations as described above.
FlAsH-FALI of synaptotagmin I was performed by adapting the method described by Marek and Davis (2002) to sub-cellular suspensions. Briefly, PNS suspensions from cells of the F7 + SytTag clone, labelled with FlAsH were kept at 4°C while being exposed for 1 min to UV-light from a 200 W HBO lamp. After being ‘flashed’, the samples were assayed within 5 min according to the three protocols used for Ca2+/H+-antiport assessment.
Ca2+/H+-antiport activity measured as Ca2+-induced H+-gradient dissipation in rat brain isolated synaptic vesicles and PC12 cells PNS
Synaptic vesicles were isolated from the brain of young male rats (Wistar; Zootechnie, University of Geneva) as previously described (Hell et al. 1988; Gonçalves et al. 2000). ATP-dependent proton transport was measured in synaptic vesicles by following the fluorescence quenching of acridine orange (Gonçalves et al. 1998) in a Perkin-Elmer spectrofluorimeter LS-50B (Waltham, MA, USA) with excitation and emission wavelengths of 495 and 530 nm respectively. The assay (see Fig. 2a) was carried out at 30°C in the following: 150 mM KCl, 2 mM MgCl2, 60 mM sucrose, 10 mM Tris-HCl, pH 8.5, and containing 50 μM EGTA, 3 μM acridine orange and the synaptic vesicle sample (0.6 mg protein/mL). After addition of 0.5 mM Mg-ATP, H+-pumping into the vesicles was followed as quenching of the AO fluorescence signal. When the steady state was reached, addition of 500 μM free Ca2+ caused partial dissipation of the vesicle H+-gradient. Finally, full dissipation was achieved (and the gradient measured) using the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP; 2 μM final concentration). Calcium-induced H+-gradient dissipation was expressed as fractional quench (%) calculated 30 s after Ca2+ addition or at equivalent time for controls (no Ca2+ addition), as reported elsewhere (Gonçalves et al. 1999b).
ATP-dependent proton transport was measured similarly on PC12 cell PNS (Bloc et al. 1999) (Fig. 2b–e). The assay was carried out at 30°C in the following: 130 mM KCl, 2 mM MgCl2, 20 mM MOPS/Tris, pH 7.4 and containing 50 μM EGTA and 3 μM acridine orange. PNS (300 μg/mL) from different cell lines were added to this medium, before addition of 1 mM Mg-ATP to start H+-pumping into the acidic organelles. Partial gradient dissipation was assayed in PNS from SYT1 positive or negative cell clones by addition of 500 μM free Ca2+.
Ca2+/H+-antiport activity measured as bafilomycin-sensitive calcium accumulation
Ca2+ accumulation by acidic organelles in PNS was measured by rapid filtration and scintillation counting. A PNS sample (0.3 mg protein/mL) was incubated at 30°C in 850 μL of reaction medium containing 130 mM KCl, 2 mM MgCl2, 20 mM MOPS/Tris, pH 8.5 and 50 μM EGTA. Radiolabelled 45calcium was from ARC (Saint Louis, MO, USA). In the ATP-dependent Ca2+ uptake assay, PNS (0.3 mg protein/mL) suspension was allowed to equilibrate for 6 min in reaction medium containing 10 μM Na-orthovanadate. Then, 1 mM Mg-ATP was added and a pH gradient was allowed to form for 2 min before 45Ca2+ (550 μM free; 0.5 mCi/mmol calcium) addition. Rapid filtration took place 2 min after 45Ca2+ addition, by placing the 800 μl aliquot in Millipore (Millipore, Billerica, MA, USA) filter HAWP (0.45 mm) under vacuum. The filtration cycle included pre-washing of filters with 1.5-mL ice-cold reaction medium without MgCl2 followed by filtration of the sample and washing with 3 mL more of the same medium. The radioactivity retained in the filters was measured by liquid scintillation spectrometry (Beckman, Brea, CA, USA).
Ca2+/H+-antiport activity measured as pH-jump-induced calcium accumulation
In another set of experiments, Ca2+ uptake was driven by the pre-established pH gradient (pH jump) between the acidic vesicular lumen of the organelles in PNS (prepared at pH 7.4) and the reaction medium (pH 8.5). Ca2+ uptake assays started by adding PNS suspensions (0.3 mg protein/mL) to reaction medium already containing 10 μM Na-orthovanadate and 45Ca2+ (550 μM free; 0.5 mCi/mmol calcium) and were stopped 2 min after by rapid filtration (as above).
Ionomycin-induced 45Ca2+ accumulation in acidic organelles of PC12 cells
45Ca2+-accumulation assays were carried out by adding PNS (0.3 mg protein/mL) to ATP-free saline medium containing 2 μM ionomycin free acid, 45Ca2+ (550 μM free; 0.5 mCi/mmol calcium) and 10 μM Na-orthovanadate (+ionomycin). Controls were carried out by allowing PNS to equilibrate for 6 min in ATP-free saline medium before 45Ca2+ addition (−ionomycin). Rapid filtration took place 2 min after.
SYT-1-positive and SYT-1-negative PC12 cell populations
The starting cell lines in the present investigation were two PC12 spontaneous variants selected by Shoji-Kasai et al. (1992), namely G11(+/+) (SYT-1 positive) and F7(−/−) (SYT-1 negative). Figure 1 shows that the G11(+/+) variant expresses SYT with a distribution very similar to that of VAMP (another typical presynaptic protein). In contrast, SYT-1 immunolabelling was totally absent in F7(−/−) cells, which nevertheless expressed as much VAMP as G11(+/+) cells.
We then transfected F7(−/−) cells (with no detectable SYT-1 protein) with the full-length Syt-1 gene, providing a SYT-1-positive clone (F7 + Syt). As a result, a proportion of the F7 + Syt population displayed intense SYT immunolabelling, which apparently filled the whole cell interior. A very similar picture was obtained with F7(−/−) cells transfected with the Syt-1 gene tagged with a tetracysteine epitope that specifically binds membrane permeable 4′,5′-bis(1,3,2-dithioarsolan-2-yl)fluorescein-(1,2-ethanedithiol)2 (FlAsH-EDT2). These F7 + SytTag were particularly useful for our purpose since SYT-1 in these cells could be acutely and specifically damaged by the process of fluorophore-assisted light inactivation (FALI), with exquisite (< 4 nm) spatial resolution (Marek and Davis 2002). To complete the panel, we prepared a negative ‘mock’ control by transfecting F7(−/−) cells with an empty expression vector.
Ca2+/H+-antiport activity measured in synaptic vesicles isolated from the rat brain
The vesicular Ca2+/H+-antiport was described and fully characterized on synaptic vesicles isolated from sheep cerebral cortex (Gonçalves et al. 1998, 2000). We found it useful to show that it is also active in rat brain vesicles (Fig. 2a). A sample of isolated vesicle fraction was introduced to a saline medium containing acridine orange (AO), a fluorophore taken up by acidic organelles, that reports transmembrane H+-gradients. Application of Mg-ATP caused a rapid decrease in the AO fluorescence intensity, revealing formation of vesicular H+-gradient. When the steady-state was reached, application of 500 μM Ca2+ (final concentration; grey trace) provoked partial dissipation of the H+-gradient. At the end of the run, full dissipation was achieved using the protonophore (CCCP). In control samples (black trace), no Ca2+ was added and no dissipation occurred.
SYT1 requirement for vesicular Ca2+-dependent H+-gradient dissipation in PC12 cells
In a first approach (Fig. 2b–e), we used the same technique for measuring Ca2+/H+-antiport activity in SYT positive and SYT-negative cell populations. A PNS sample of G11(+/+) cells (which fully express SYT1) was tested first (Fig. 2b). Like with rat brain vesicles, application of 500 μM Ca2+ (grey trace) provoked partial dissipation of the gradient. At the end of the run, full dissipation was achieved using the protonophore CCCP. In control samples (black trace), no Ca2+ was added and no dissipation took place.
In control experiments, we checked that bafilomycin-A1, like CCCP, fully dissipates the H+-gradient. No further dissipation could be achieved by subsequent addition of either CCCP or 500 μM Ca2+ (not illustrated). Therefore, the vesicular H+-gradient is built by the action of the V-ATPase in this preparation, and Ca2+ application acts by promoting gradient dissipation.
When the same protocol was applied to F7(−/−) cells, CaCl2 application did not induce any H+-gradient dissipation (Fig. 2c). Instead, we sometimes observed a slight increase in the gradient. This might be due to direct stimulation of the V-ATPase by the added Ca2+ or by an increase in the Cl− gradient (Moriyama and Nelson 1987). Importantly, transfection of the tagged Syt1 full-length gene to F7(−/−) cells (F7 + SytTag) rescued the dissipation as long as these cells were kept under obscurity (Fig. 2d). However, UV-light irradiation of the F7 + SytTag cell clone, which selectively damages SYT1, fully abolished the Ca2+-induced H+-gradient dissipation (Fig. 2e).
Summarized in Fig. 3a, the results obtained with this approach are clear-cut. Significant Ca2+-induced dissipation of the H+-gradient occurs with G11(+/+), with F7 + Syt, and with the non-irradiated F7 + SytTag populations. In contrast, no dissipation could be detected in F7(−/−) cells, in the UV-irradiated F7 + SytTag cells, or in ‘mock’ controls. In summary, Ca2+/H+-antiport activity, measured as Ca2+-induced H+-gradient dissipation, is detectable only in cells containing functional SYT1.
SYT-1 requirement for bafilomycin-sensitive calcium accumulation in vesicles
In a second approach, we assessed Ca2+/H+-antiport activity by measuring 45Ca accumulation into organelles in the presence or absence of bafilomycin-A1 (Fig. 3b). When calcium ions are allowed to enter vesicles, they rapidly bind to the internal gel matrix (Reigada et al. 2003), and therefore they accumulate. By selectively blocking V-ATPase, bafilomycin-A1 prevents H+--gradient formation. Thus, only the bafilomycin-sensitive part of accumulation is attributable to Ca2+/H+ exchange in vesicles.
After equilibration in pH 8.5 saline, PC12 samples were given Mg-ATP, which (in the absence of bafilomycin) allowed formation of the pH gradient. We then added 45Ca2+ (0.5 mCi/mmol; final free concentration 550 μM) for 2 min. After rapid filtration and rinsing, 45Ca2+ was measured on the filters. In PNS prepared from the three clones expressing an active SYT-1, a substantial amount of 45Ca accumulated into organelles (Fig. 3b; grey histograms) over the value found in the corresponding bafilomycin-A1-treated preparations (black histograms). Such H+-gradient dependent calcium accumulation did not occur in the three clones where SYT-1 was either absent or inactivated by UV-light. Therefore, bafilomycin-sensitive 45Ca accumulation into PC12 cell acidic organelles takes place exclusively in the presence of functional SYT-1.
SYT-1 requirement for pH-jump-induced calcium accumulation
In a third approach, we assessed calcium accumulation in organelles following a ‘pH-jump’ (Fig. 3c). PC12 cell samples (prepared as usual in a pH 7.4 medium) were added into the ATP-free pH 8.5 saline, which already contained 45Ca2+ and 40Ca2+ (concentration as before). Energized by the pH difference, Ca2+ will accumulate into acidic organelles only if a Ca2+/H+-exchanger is available (condition named ‘+ pH-jump’; grey histograms). For corresponding controls, we pre-incubated PNS in ATP-free saline at pH 8.5 before adding Ca2+, a condition which annihilates the H+-gradient (‘− pH-jump’; black histograms). Again, G11(+/+), F7 + Syt and non-irradiated F7 + SytTag cells significantly accumulated calcium over the non-specific level in response to the pH-jump. Strikingly, the amount of calcium accumulation observed with these clones upon the pH-jump was equivalent to the ATP-driven, bafilomycin-dependent accumulation (cf. Fig. 3b). As for the SYT-1-deficient or UV-light inactivated cells, they did not show any significant 45Ca uptake attributable to Ca2+/H+-exchange in response to the pH-jump.
Non-specific calcium accumulation following ionophore treatment
A Ca2+ ionophore such as ionomycin should induce Ca2+ entry in acidic vesicles independently from any Ca2+/H+-antiport activity. PC12 samples were added to an ATP-free medium containing 45Ca2+ and 40Ca2+ (concentrations as before), with or without 2 μM ionomycin. As expected, ionomycin provoked 45Ca retention in all tested clones to a similar extent (Fig. 4). Thus, when calcium is allowed to enter vesicles through a process other than Ca2+/H+-antiport, accumulation is no longer dependent on the presence of SYT-1.
Vesicular Ca2+/H+-antiport and calcium clearance from cells or nerve terminals
The vesicular Ca2+/H+-antiport was described using synaptic vesicles isolated from the sheep brain cortex. Its activity depends on the H+ gradient across the vesicle membrane. Proton-dependent vesicular Ca2+ uptake through this antiport is inhibited by Zn2+ and Cd2+ (which are transported by the antiport) and by Sr2+ (500 μM; which apparently, is not transported) (Gonçalves et al. 1998, 1999a, b, 2000). In this study, we gave evidence that the Ca2+/H+-antiport is also active in synaptic vesicles isolated from the rat brain. In addition, Ca2+/H+-antiport activity was also recorded in intact (sealed) synaptosomes prepared from mossy fibres of rat hippocampus (Bancila et al. 2009).
Knowing that the Ca2+/H+-antiport is activated in the presence of high [Ca2+]i we suspected that it could play a specific role in rapid neurotransmission. Indeed, [Ca2+]i was shown to raise momently at a very high level in restricted spots (nanodomains) of presynaptic active zones (Llinas et al. 1992; Yazejian et al. 2000). Logically, inactivation of the Ca2+/H+-antiport should enhance or prolong the ‘Ca2+ spark’, which takes place in active zones upon action potential arrival. As a consequence, the phasic transmitter release will be prolonged, causing a lengthening of the post-synaptic potential. It is exactly what we observed when testing fast transmission of isolated impulses in Torpedo nerve electroplaque synapses. From these electrophysiological experiments, it was concluded that the Ca2+/H+-antiport role is to shorten the duration of individual Ca2+-transients, allowing the synapse to fire at higher frequencies (Cordeiro et al. 2011).
The vesicular Ca2+/H+-antiport is only one of the processes working towards buffering [Ca2+]i in secreting cells or nerve terminals (Neher 1998; Castonguay and Robitaille 2001; Parekh 2008; Desai-Shah and Cooper 2009). As shown in experiments in situ, a brief tetanic stimulation provokes a transient calcium accumulation in synaptic vesicles (Parducz et al. 1987, 1994; Buchs et al. 1994). Within vesicles, Ca2+ binds to a gel matrix in exchange for neurotransmitter and ATP (Reigada et al. 2003). Calcium is later expelled, most probably by exocytosis (Parducz et al. 1994), a process which is expected to prevent Ca2+ depletion in the synaptic cleft in the case of intense synapse stimulation (Borst and Sakmann 1999).
Synaptotagmin-1 and the Ca2+/H+-antiport in PC12 cells
Here, we looked for an appropriate preparation to gain insight into the molecular counterpart of the vesicular Ca2+/H+-antiport antiport. Using SYT1-positive and SYT1-negative PC12 cell populations, we found a clear relationship between the presence of functional SYT-1 and the expression of vesicular Ca2+/H+-antiport activity. Congruent and significant results were obtained by measuring activity with three independent methods. Strikingly, the Ca2+/H+-antiport, which was not functioning in the starting SYT-1-deficient F7(−/−) variant, was rescued upon transfection of the full-length Syt-1 gene; it was then abolished by the FlAsH-FALI procedure which acutely and specifically inactivates SYT-1 (Marek and Davis 2002). Interestingly, when SYT-1 was over-expressed by transfection its distribution inside PC12 cells seemed different from that in G11(+/+) variants. Nevertheless, acute and specific inactivation of transfected SYT-1 by the FlAsH-FALI procedure completely abolished Ca2+/H+-antiport activity. Therefore, the antiport activity depends indeed on SYT-1, irrespective of its protein density or apparent cell localization.
We used PC12 cells as a well-characterized secretory model where we could unequivocally test vesicular Ca2+/H+ antiport activity in the presence or the absence of SYT-1, as well as to be able to asses Ca2+/H+ antiport activity after acutely inactivating this protein. PC12 cells exhibit one of the most robust, bafilomycin sensitive, proton gradient across acidic compartments (Bloc et al. 1999). Furthermore, using SYT-1−/− PC12 cells instead of KO animals allows the usage of FlAsH-FALI by simple transfection of SYT-1−/− cells, without having to generate transgenic animals bearing the tetracysteine motif that binds FlAsH into SYT-1 null animals, which are not viable to adulthood. Furthermore, FlAsH-FALI of PC12 cells transfected with Syt1 clearly ascertained SYT-1 as a molecular tenant of Ca2+/H+ antiport, discarding possible effects concerning compensatory expression of other synaptotagmins in a SYT-1 null environment. In addition, we used PC12 clones where both catecholamine and ATP secretion as well as membrane accretion was tested in SYT-1(+/+) and SYT-1(−/−) cells. Both parameters were increased in SYT-1(−/−) cells (Shoji-Kasai et al. 1992). Also both parameters increase with intracellular Ca2+, pointing-out for an effect on calcium homeostasis that we now confirm. All these features pointed-out for using PC12 cells instead of KO animals to molecularly ascertain the vesicular Ca2+/H+ antiport.
We have successively reported (i) a full description of the vesicular Ca2+/H+ antiport in isolated mammalian brain synaptic vesicles (Gonçalves et al. 1998), (ii) a clear analysis of its functional role in the Torpedo electric organ (Cordeiro et al. 2011) and (iii) an insight into its molecular requirement in PC12 cells (present work). It would certainly be fascinating to elucidate whether and how the vesicular antiport operates in other synaptic or secreting systems, such as transgenic mice, invertebrate synapses, and chromaffin cells. This would, however, exceed the objective of the present work, which was to pinpoint a key molecule involved in a Ca2+-regulatory mechanism putatively regulating a plethora of secreting systems.
Could the physiological changes reported after SYT-1 alteration or inactivation be attributed to vesicular Ca2+/H+-antiport impairment?
The PC12 variants used in the present work provide a first response in that direction. We recall that SYT-1-deficient F7(−/−) cells release more dopamine and ATP than the SYT-1-expressing G11(+/+) cells (Shoji-Kasai et al. 1992). This may be expected from preparations where Ca2+ clearance is perturbed: [Ca2+]i remaining longer at a super-threshold level after stimulation will enhanced transmitter release. Transmitter release was also reported to be enhanced in certain SYT-deficient synapses (Stevens and Sullivan 2003; Hua et al. 2007; Young and Neher 2009; Xu et al. 2012). We may also recall that Sr2+ ions (which inhibit Ca2+/H+ antiport) increase the duration of transmitter release in individual impulses (Xu-Friedman and Regehr 2000; Cordeiro et al. 2011).
However, the picture is more complex for two reasons. First, SYT-1 deletion or alterations provoke a variety of different functional changes according to different synapses and experimental conditions (Koh and Bellen 2003; Chapman 2008; Kochubey et al. 2011; Xu et al. 2012). After an initial increase in Ca2+-dependent neurotransmitter release, inhibition of Ca2+/H+ antiport results in fading of transmission, particularly when synapses are stimulated repetitively (Cordeiro et al. 2011; Wu and Cooper 2012), an effect probably involving desensitization of the release mechanism through mediatophores, consecutive to steady increase in sub-threshold [Ca2+]i (Adams et al. 1985; Israël et al. 1987). Such ‘fading’ of transmission is also observed after SYT-1 inactivation. It seems prominent at drosophila neuromuscular junction, or at CNS glutamatergic and other synapses (Marek and Davis 2002; Burgalossi et al. 2010).
These complex and apparently contradictory data may arise from the complexity of protein-membrane interactions taking place in presynaptic active zones, as described below.
SYT-1 as a multifaceted coordinator of a presynaptic complex involving VOCCs, phospholipids, mediatophore, SNARE and other proteins at presynaptic active zones
The active zone is defined as a restricted sector of presynaptic axon endings – or en passant boutons – facing the post-synaptic area (Couteaux and Pécot-Dechavassine 1970). Active zones are characterized by a differentiation of the presynaptic membrane, by the proximity of synaptic vesicles and by the accumulation of several specific proteins, which include SYT-1, SNAREs, VOCCs and mediatophores (a proteolipid complex similar to the V0 sector of V-ATPase which is directly involved in neurotransmitter release and membrane fusion; (Dunant and Israël 2000; Peters et al. 2001; Dunant et al. 2009). The above-mentioned proteins and local membrane phospholipids are linked by multiple molecular interactions in active zones (Galli et al. 1996; Shiff et al. 1996; Davis et al. 1999; Zhong et al. 1999; Chapman 2008). They form therefore a multimolecular complex, so we can anticipate that each of them is involved in more than one of presynaptic processes. We would like to recall that alteration of other presynaptic proteins, particularly the SNAREs, causes functional changes often resembling those provoked by SYT-1 inactivation (Koh and Bellen 2003; Stevens and Sullivan 2003; Hua et al. 2007; Xu et al. 2007; Chapman 2008; Young and Neher 2009).
The present results support a clear link between synaptotagmin-1 and vesicular Ca2+/H+-antiport activity. It will be fascinating to investigate whether other proteins, like SNAREs, might also be linked to the vesicular Ca2+/H+-antiport activity. Another exciting avenue will be to understand how the disruption of this presynaptic complex may affect different secretory processes, from ultra-rapid synapses to the slow releasing systems with their large granules.
Study supported by a Swiss FNRS grant N°31-057135.99., and the European Commission project Lipidiet (QLK1-CT-2002-00172) to Y.D., and by the Portuguese Foundation for Science and Technology (FCT: SFRH / BD / 6403 / 2001; FSE-POPH-QREN) to J.M.C. We thank Graeme Davis for initial guidance on the FALI technique, Shoji-Kasai and Masami Takahashi for providing G11 and F7 PC12 clones, Dominique Muller and Alain Bloc, for criticism and suggestions, as well as Jean-Pierre Andrivet, Françoise Loctin, Agostino Massiero and Lorena Jourdain for excellent technical assistance. The authors declare no conflict of interest.