Electrostatic regulation of the cis- and trans-membrane interactions of synaptotagmin-1

Synaptotagmin-1 is a vesicular protein and Ca2+ sensor for Ca2+-dependent exocytosis. Ca2+ induces synaptotagmin-1 binding to its own vesicle membrane, called the cis-interaction, thus preventing the trans-interaction of synaptotagmin-1 to the plasma membrane. However, the electrostatic regulation of the cis- and trans-membrane interaction of synaptotagmin-1 was poorly understood in different Ca2+-buffering conditions. Here we provide an assay to monitor the cis- and trans-membrane interactions of synaptotagmin-1 by using native purified vesicles and the plasma membrane-mimicking liposomes (PM-liposomes). Both ATP and EGTA similarly reverse the cis-membrane interaction of synaptotagmin-1 in free [Ca2+] of 10–100 μM. High PIP2 concentrations in the PM-liposomes reduce the Hill coefficient of vesicle fusion and synaptotagmin-1 membrane binding; this observation suggests that local PIP2 concentrations control the Ca2+-cooperativity of synaptotagmin-1. Our data provide evidence that Ca2+ chelators, including EGTA and polyphosphate anions such as ATP, ADP, and AMP, electrostatically reverse the cis-interaction of synaptotagmin-1.

Protein purification. All SNARE and the C2AB domain of synaptotagmin-1 constructs based on rat sequences were expressed in E. coli strain BL21 (DE3) and purified by Ni 2+ -NTA affinity chromatography followed by ion-exchange chromatography as described elsewhere 10,20 . The stabilized Q-SNARE complex consists of syntaxin-1A (aa 183-288) and SNAP-25A (no cysteine, cysteines replaced by alanines) in a 1:1 ratio by the C-terminal VAMP-2 fragment (aa , and was purified as described earlier 25 . The C2AB domain of synaptotagmin-1 (aa 97-421) and soluble form of VAMP-2 lacking the transmembrane domain (VAMP-2 1-96 ) were purified using a Mono S column (GE Healthcare, Piscataway, NJ) as described previously 26 . The stabilized Q-SNARE complex was purified by Ni 2+ -NTA affinity chromatography followed by ion-exchange chromatography on a Mono Q column (GE Healthcare, Piscataway, NJ) in the presence of 50 mM n-octyl-β-d-glucoside (OG) 10 . The point mutated C2AB domain (S342C) was labelled with Alexa Fluor 488 C5 maleimide (C2AB A488 ) 26 .
Lipid composition of liposomes. All lipids were obtained from Avanti Polar lipids (Alabaster, AL).

Preparation of proteoliposomes.
Incorporation of the Q-SNARE complex into large unilamellar vesicles (LUVs) was achieved by OG-mediated reconstitution, called the direct method, i.e. incorporation of proteins into preformed liposomes 10,20 . Briefly, lipids dissolved in a 2:1 chloroform-methanol solvent were mixed according to lipid composition. The solvent was removed using a rotary evaporator to generate lipid film on a glass flask, then lipids were resuspended in 1.5 mL diethyl ether and 0.5 mL buffer containing 150 mM KCl and 20 mM HEPES/KOH pH 7.4. The suspension was sonicated on ice (3 × 45 s), then multilamellar vesicles were prepared by reverse-phase evaporation using a rotary evaporator as diethyl ether was removed. Multilamellar vesicles (0.5 mL) were extruded using polycarbonate membranes of pore size 100 nm (Avanti Polar lipids) to give uniformly-sized LUVs. After the preformed LUVs had been prepared, SNARE proteins were incorporated into them using OG, a mild non-ionic detergent, then the OG was removed by dialysis overnight in 1 L of buffer containing 150 mM KCl and 20 mM HEPES/KOH pH 7.4 together with 2 g SM-2 adsorbent beads. Proteoliposomes had protein-to-lipid molar ratio of 1:500.
Vesicle fusion assay. A FRET-based lipid-mixing assay was applied to monitor vesicle fusion in-vitro 10,20 .
LDCV fusion reactions were performed at 37 °C in 1 mL fusion buffer containing 120 mM K-glutamate, 20 mM K-acetate, 20 mM HEPES-KOH (pH 7.4), 1 mM MgCl 2 , and 3 mM ATP (Fig. 4b). Fusion buffer in Fig. 3a,b contains no ATP, but EGTA; 120 mM K-glutamate, 20 mM K-acetate, 20 mM HEPES-KOH (pH 7.4), 5 mM MgCl 2 , and 10 μM EGTA. ATP should be made freshly before all experiments, because it is easily destroyed by freezing and thawing. Free Ca 2+ concentration in the presence of Mg 2+ and ATP or EGTA was calibrated using the MaxChelator simulation program.
The PM-liposomes that contain NBD-DOPE and Rhodamine-DOPE as a donor and an acceptor dye, respectively, were incubated with LDCVs, thus leading to dequenching of donor fluorescence (NBD) as a result of lipid dilution with unlabelled vesicle membrane 10,20 (Fig. 2). Excitation wavelength was 495 nm and emission was measured at 520 nm. Anisotropy (r) was calculated using the formula r = (I VV − G × I VH )/(I VV + 2 × G × I VH ), where I VV indicates the fluorescence intensity with vertically polarized excitation and vertical polarization on the detected emission and I VH denotes the fluorescence intensity when using a vertical polarizer on the excitation and horizontal polarizer on the emission. G is a grating factor used as a correction for the instrument's differential transmission of the two orthogonal vector orientations. Lipid composition of the PM-liposomes (protein-free) was identical to those used in a fusion assay except labelled PE (45% PC, 15% PE, 10% PS, 25% Chol, 4% PI, and 1% PIP 2 ).  (Fig. 1b).

Ca
Negatively-charged oxygen atoms of ATP chelate divalent cations such as Mg 2+ , Ca 2+ , or Sr 2+27 . In the experiments, 5 mM ADP or 5 mM AMP chelated Ca 2+ , thereby reducing free [Ca 2+ ] from 122 to 57 μM and from 126 to 99 μM, respectively (Fig. 1c). Increasing the number of phosphate groups in Adenosine increases Ca 2+ affinity and lowers K d by increasing the number of Ca 2+ ions that are bound 27,33 . ATP, ADP, and AMP have distinct ranges of Ca 2+ -buffering capacity and distinct K d values 33 , so Ca 2+ -chelating effect is ATP > ADP > AMP (Fig. 1a-c). Altogether, the predictions of free [Ca 2+ ] in the complex buffer solutions including Mg 2+ , ATP and EGTA were confirmed using a fluorescent Ca 2+ indicator.
Ca 2+ -bound synaptotagmin-1 is inserted to native vesicle membranes such as synaptic vesicles and large dense-core vesicles (LDCVs) that contain anionic phospholipids 10 . However, ATP electrostatically prevents the cis-interaction of synaptotagmin-1, whereas the trans-interaction of synaptotagmin-1 to the plasma membrane remains active to mediate Ca 2+ -dependent vesicle fusion, because PIP 2 overcomes the inhibitory effect of ATP by increasing the membrane-binding affinity of the C2AB domain 10-12 . First, 1 mM Ca 2+ was applied to induce binding of the C2AB domain, then 100 μM EGTA was added ten times (arrows) to reverse this binding. A total dose of 800 μM EGTA disrupted C2AB binding to LDCV (red, d) and a final total dose of 1 mM EGTA reversed C2AB binding to liposomes (red, e). (f,g) LDCV fusion was increased by 1 mM Ca 2+ in the presence of 800 μM EGTA (f), but was not affected in the presence of 1 mM EGTA (g). www.nature.com/scientificreports/ We tested an assay that uses fluorescence anisotropy measurement to monitor the cis-and trans-membrane interaction of synaptotagmin-1 (Fig. 2). Direct measurement of the cis-and trans-membrane interaction of endogenous synaptotagmin-1 in native vesicle membranes is impossible, so we monitored the binding of an exogenously-added C2AB domain of synaptotagmin-1 (Syt 97-421 ), which was labelled with Alexa Fluor 488 at S342C (Fig. 2a). We took advantage of a single fluorescent labelling for anisotropy measurement to monitor the interaction of the C2AB domain with native vesicles or liposomes; the membrane-bound C2AB domain leads to increase of fluorescence anisotropy due to a reduction in the rotational mobility 10 (Fig. 2a,b). It is noted that our experiments using the cytoplasmic C2AB domain are intended to shed light on the cis-and trans-interactions, but the geometry is not truly being imitated.
We first monitored the cis-membrane interaction between the C2AB domain and the LDCV membranes (Fig. 2a). The presence of 1 mM Ca 2+ increased fluorescence anisotropy; this change indicates that the C2AB domains bind to LDCV membranes in a Ca 2+ -dependent manner. Five sequential applications of 1 mM ATP gradually decreased the anisotropy signal by chelating Ca 2+ ; this result suggests dissociation of the C2AB domain from LDCVs (Fig. 2a). 5 mM ATP in the presence of 1 mM Ca 2+ almost completely disrupted the cis-membrane interaction of the C2AB domain with the LDCV membranes (Fig. 2a); free [Ca 2+ ] in the presence of Mg 2+ , ATP and EGTA was calibrated using the MaxChelator simulation program and free [Ca 2+ ] was 351 µM in case of 5 mM ATP and 1 mM Ca 2+ (Table 1).
Next, we tested the trans-membrane interactions between the C2AB domain and the PM-liposomes; 10% PS, 4% PI, and 1% PIP 2 were included in the PM-liposomes (Fig. 2b). The C2AB domain of synaptotagmin-1 bound to liposomes in response to 1 mM Ca 2+ , and this trans-membrane interaction was reduced by ATP, 1 mM applied thirteen times sequentially (Fig. 2b). Free [Ca 2+ ] in different ATP concentrations was summarized in Table 1. Ca 2+ -dependent vesicle fusion is accelerated by the increase of the trans-interactions and the decrease of the cis-membrane interaction of synaptotagmin-1 10,20 , so we hypothesized that 5 mM ATP in the presence of 1 mM Ca 2+ is appropriate to observe Ca 2+ -dependent fusion (red in Fig. 1a,b).
To test this hypothesis and examine the effect of the cis-and trans-membrane interaction of synaptotagmin-1 on vesicle fusion, we applied a reconstitution system of vesicle fusion by using native LDCVs 10,20,24 . The PM-liposomes contain the stabilized Q-SNARE complex (syntaxin-1A and SNAP-25A in a 1:1 molar ratio 25 ). Indeed, 5 mM ATP in the presence of 1 mM Ca 2+ (i.e., 351 µM free [Ca 2+ ] according to the MaxChelator program (Table 1)) dramatically accelerated LDCV fusion, which was completely blocked by the soluble VAMP-2 (VAMP-2 1-96 ); this results indicates SNARE-dependent vesicle fusion (Fig. 2c). We have previously shown that 300-400 μM free [Ca 2+ ] in the absence of ATP fails to enhance vesicle fusion, but rather slightly inhibits fusion, because the cis-membrane interaction of the C2AB domain to native vesicle membranes becomes robust from 100 μM up to 3 mM 10 . ATP prevents this cis-membrane interaction by charge screening and competing with the vesicle membrane, thus allowing synaptotagmin-1 to interact in trans with the plasma membrane 10 .
Polyphosphates such as ATP reverse an inactivating cis-interaction of synaptotagmin-1 by an electrostatic effect (Fig. 2a-c). Next, we tested whether other Ca 2+ chelators, e.g., EGTA, can have a similar inhibitory effect on the cis-membrane interaction. Anisotropy measurement was performed to monitor the cis-and trans-membrane interaction of the C2AB domain (Fig. 2a,b). EGTA was applied 10 times (100 μM each in the presence of 1 mM Ca 2+ ) to reverse the cis-interaction of the C2AB domain to LDCVs (Fig. 2d). Application of 800 μM EGTA dramatically disrupted the cis-interaction in the presence of total 1 mM Ca 2+ (red in Fig. 2d) Fig. 2e), whereas 1 mM EGTA significantly disrupted both the cis-and trans-membrane interactions of the C2AB domain (Fig. 2d,e); free [Ca 2+ ] was 12 μM (Table 1).
Anisotropy measurement is useful to find a Ca 2+ -buffering condition to observe Ca 2+ -dependent vesicle fusion, where the cis-membrane interaction is prevented and the trans-interaction remains active. The presence of 800 μM EGTA with 1 mM Ca 2+ (200 μM free [Ca 2+ ], Table 1) significantly reversed the cis-interaction (Fig. 2d), but had a minor effect on the trans-interaction (Fig. 2e). Indeed, 800 μM EGTA with 1 mM Ca 2+ reproduced Ca 2+ -dependent LDCV fusion (Fig. 2f). 1 mM EGTA with 1 mM Ca 2+ (12 μM free [Ca 2+ ], Table 1) failed to accelerate fusion, because the trans-interaction of the C2AB domain was dramatically disrupted by 1 mM EGTA (red in Fig. 2e); it is mainly because of low free [Ca 2+ ]. Taken together, we established an anisotropy assay to monitor the cis-and trans-membrane interaction of synaptotagmin-1 by using native LDCVs and the PM-liposomes. Our data suggest that Ca 2+ chelators such as EGTA, in addition to polyphosphates such as ATP, can prevent the cis-membrane interaction of synaptotagmin-1 by the electrostatic effect in a certain range of free [Ca 2+ ].  20 . We examined whether EGTA reproduces the biphasic regulation of Ca 2+ on LDCV fusion (Fig. 3a,b). Instead of ATP, 10 μM EGTA was included in fusion buffer and free [Ca 2+ ] was calculated using the MaxChelator program. As expected, biphasic regulation of Ca 2+ on LDCV fusion was observed, where Ca 2+ -dependent fusion progressively increased until [Ca 2+ ] = ~ 100 μM, and gradually decreased at [Ca 2+ ] from 300 μM to 1 mM (Fig. 3a,b). Biphasic regulation of Ca 2+ on LDCV fusion is mediated by two different mechanisms: (1) millimolar range of [Ca 2+ ] decreases the trans-interaction of synaptotagmin-1 by shielding PIP 2 and (2) sub-millimolar range of [Ca 2+ ] above 300 μM increases the cis-interaction of synaptotagmin-1 to its own vesicle membrane 20 . To further confirm the cis-interaction at higher [Ca 2+ ], we performed anisotropy measurement (Fig. 2a,d) to study the Ca 2+ dose-response of the cis-interaction of synaptotagmin-1 in the presence of EGTA instead of ATP (Fig. 3c). Indeed, the cis-membrane interaction of the C2AB domain gradually increased from 300 μM [Ca 2+ ] and remained robust at millimolar [Ca 2+ ] (Fig. 3c). Note that ATP and EGTA give rise to different kinetics of the Ca 2+ dose-response curves of vesicle fusion and the cis-interaction of synaptotagmin-1 20 , because ATP effectively buffers free [Ca 2+ ] in the range of 10-500 μM, but EGTA cannot efficiently buffer free [Ca 2+ ] in this range. PIP 2 concentration regulates Ca 2+ cooperativity of synaptotagmin-1. Synaptotagmin-1 binds to anionic phospholipids by electrostatic interaction and the Ca 2+ -binding loops of the C2 domains are inserted to anionic phospholipids in a Ca 2+ -dependent manner; aspartate residues of the Ca 2+ -binding loops in the C2-domains together with anionic membrane lipids coordinate Ca 2+ -ions 21,23,36 . PIP 2 enhances Ca 2+ -sensitivity of synaptotagmin-1 by interacting with the polybasic patch in the C2B domain 10,21 . Ca 2+ -cooperativity of synaptotagmin-1 varies among cell types, with the Hill coefficients ranging from ~ 2 to ~ 5. We tested that PIP 2 also regulates Ca 2+ -cooperativity of synaptotagmin-1 for membrane binding (Fig. 4a, Table 2) and vesicle fusion (Fig. 4b, Table 2). Increases of PIP 2 concentration from 1 to 5% in the PM-liposomes shifted Ca 2+ titration curves for membrane binding to the left side; this change indicates increased Ca 2+ sensitivity, but reduced Ca 2+ cooperativity (Fig. 4a, Table 2).
Next, we observed that Ca 2+ -cooperativity of synaptotagmin-1 for vesicle fusion was also reduced by increasing PIP 2 concentration, correlating with the Ca 2+ -cooperativity of synaptotagmin-1 for membrane binding. The Ca 2+ dose-response curve for LDCV fusion was shifted leftward as PIP 2 concentration was increased in the PMliposomes (Fig. 4b, Table 2). Taken together, high PIP 2 concentration increases the sensitivity of synaptotagmin-1 to Ca 2+ , but lowers Ca 2+ cooperativity. These changes imply that increasing the negative electrostatic potential in the plasma membranes attracts Ca 2+ -bound synaptotagmin-1 with low Ca 2+ cooperativity, in which the total numbers of Ca 2+ ions coordinated to one synaptotagmin-1 might be reduced to 2-3 (see section "Discussion").

Discussion
The cis-binding of synaptotagmin-1 occurs in native vesicles such as LDCVs and synaptic vesicles, and inactivates Ca 2+ -dependent vesicle fusion by preventing the trans-interaction of synaptotagmin-1. Independent groups have confirmed that ATP at physiological concentrations disrupts such cis-interaction of synaptotagmin-1 11,12,37 . www.nature.com/scientificreports/ Here we show that Ca 2+ chelators, including EGTA and polyphosphate anions such as ATP, ADP, and AMP, electrostatically reverse the cis-interaction of synaptotagmin-1. We propose that Ca 2+ chelators compete with vesicle membranes that contain anionic phospholipids in binding to Ca 2+ and disrupt the cis-interaction of synaptotagmin-1 by charge screening 10 . However, PIP 2 overcomes this inhibitory effect of ATP, because PIP 2 dramatically enhances the Ca 2+ -binding affinity of synaptotagmin-1 21,38 ; this high Ca 2+ affinity of the C2AB domain to PIP 2 -containing membranes is not affected by ATP 10 . EGTA and 1,2-bis(o-aminophenoxy)ethane-N,N,N0,N0-tetraacetic acid (BAPTA) are well-known and reliable Ca 2+ buffers in the range of 10 nM-1 μM [Ca 2+ ] at the typical intracellular pH of 7.2 33,35 . Given that EGTA and BAPTA have a K d of 67 nM and 192 nM [Ca 2+ ] at pH 7, respectively, and have a higher affinity for Ca 2+ than for Mg 2+35 , both EGTA and BAPTA effectively buffer free [Ca 2+ ] only at concentrations < 1 μM 33,39 , which is close to intracellular free [Ca 2+ ]. However, EGTA is sensitively dependent on pH 35 , and BAPTA family has a strong dependence on ionic strength 40 ; importantly, because EGTA and BAPTA have nanomolar-level K d , they poorly buffer free [Ca 2+ ] in the range of 10-500 μM. In contrast, ATP has K d 230 µM 27 and is an excellent buffer for free [Ca 2+ ] in the range of 10-500 μM 33 .
Synaptotagmin-1 is a low-affinity Ca 2+ sonsor; 10-100 μM [Ca 2+ ] exponentially induce synaptotagmin-1 binding to membrane that contain PS and PIP 2 with K d ~ 50 μM 21,26 . Therefore, ATP is an appropriate and better Ca 2+ buffer than EGTA or BAPTA to study the synaptotagmin-1 activity to bind membrane and trigger vesicle fusion. Indeed, we oberserved that ATP and EGTA result in different kinetics of the Ca 2+ dose-response curves of vesicle fusion and of the cis-interaction of synaptotagmin-1 10,20 (Fig. 3b,c), because ATP has a different Ca 2+ -buffering capacity than EGTA.  www.nature.com/scientificreports/ The K d of low-affinity Ca 2+ indicator dyes can vary depending on ionic strength and is changed by anions such as ATP 41 ; e.g., the K d of low-affinity Ca 2+ indicator dyes is increased by ATP and slightly decreased by excess Mg 2+ .  35 . We confirmed the MaxChelator-based predictions using a Fluo-5N fluorescent Ca 2+ indicator (Fig. 1b).
Both the C2A and C2B domains of synaptotagmin-1 have highly cooperative Ca 2+ -dependent binding to membranes that contain anionic phospholipids 26,[42][43][44][45] . Furthermore, synaptotagmin-1 contains a polybasic region within the C2B domain that binds to PIP 2 in an Ca 2+ -independent manner 46,47 and enhances Ca 2+ sensitivity of synaptotagmin-1 membrane binding 21 and exocytosis 48 . The C2AB domain has five possible Ca 2+ -binding sites 22,23 ; negatively charged oxygen atom from acidic aspartate residues in the C2AB domain and negatively charged oxygen atom from anionic phospholipids provide complete coordination sites for Ca 2+23,36 . Ca 2+ cooperativity of the C2AB domain seems reasonable when the Hill coefficient is ~ 4 to 5, but what regulates Ca 2+ cooperativity remains poorly understood, e.g., low Hill coefficient (n, 2-3) in neuroendocrine cells such as pituitary melanotrophs (n, 2.5) 18 and chromaffin cells (n, 1.8) 19 , but high Hill coefficient in synapses including calyx-of-Held synapses (n, 4.2) 13-15 , neuromuscular junctions (n, 3.8) 16 , and bipolar cells (n, 4) 17 . We overserved that increasing PIP 2 concentration reduces the Hill coefficient, which represents Ca 2+ cooperativity (Fig. 4). Our data support that local PIP 2 concentration might control Ca 2+ cooperativity by allosterically-stabilized dual binding of synaptotagmin-1 to Ca 2+ and PIP 2 38 . In this study, we investigate the electrostatic regulation of C2AB binding to vesicle membrane and the PMliposomes. We have previously observed that Ca 2+ -independent interactions of the C2AB domain with the PM-liposomes containing anionic phospholipids (10% PS/1% PIP 2 ) is significantly disrupted in the presence of physiological concentration of ATP/Mg 2+ , but this Ca 2+ -independent interaction remains strong when the PM-liposomes contain high PIP 2 (10% PS/5% PIP 2 ), suggesting that high PIP 2 concentrations are required for Ca 2+ -independent binding of the C2AB domain in physiological ionic strength 20 . Here, we have used 10% PS/1% PIP 2 in the PM-liposomes to selectively examine the Ca 2+ -dependent membrane interaction and binding of the C2AB domain. However, in the pre-fusion state for vesicle docking and priming, the C2AB domain of synaptotagmin-1 is most likely bound to the plasma membrane through the PIP 2 -interacting polybasic region of the C2B domain 20 or the SNARE complex 49 in a Ca 2+ -independent manner. Ca 2+ can induce a re-orientation of the C2AB domain on the plasma membrane by changing the binding mode with the SNARE complex 49 or PIP 2 45 . This change in orientation may act as a switch to trigger synaptotagmin-1-dependent vesicle fusion in neurons and neuroendocrine cells. Our results do not rule out the possibility for Ca 2+ -independent interactions of synaptotagmin-1 with the SNARE complex despite extremely weak interaction 49 and it remains a topic of further study to include Ca 2+ -independent interactions of synaptotagmin-1 in our system for physiological relevance.

Data availability
The datasets generated during the current study are available from the corresponding author on reasonable requests. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.