- Email: [email protected]
phospholipid-binding and syntaxin-binding of native synaptotagmin I
Life Scien~, Vol. 61, No. 7, pp. 711-721, 1997 copyright Q 1997 J!.kvicr seicncc Inc. Printed in the USA. All rights resend w24-32osp7 517.al + .oo
PII soo24-3205(97)00536-5
ELSEVIER
Ca*+/PHOSPHOLIPID-BINDING BINDING OF NATIVE
Maurizio Racagnil.
Popolil-25
, Albert0
AND SYNTAXINSYNAPTOTAGMIN I
Venegonit
, Liliana
Buffat,
and Giorgio
‘Center of Neuropharmacology, Institute of Pharmacological Sciences, University of Milan, 2Department of Neurological Sciences, II School of Medicine, University of Naples Federico II (Italy). (Received in fmal form May 19, 1997)
Summary
Synaptotagmin, a synaptic vesicle protein endowed with multiple properties, is the putative calcium sensor in neuroexocytosis. Ca*+/phospholipid binding and syntaxin binding activity of synaptotagmin were previously investigated using recombinant fusion proteins. In phospholipid binding the EC50 for calcium obtained was different when fusion proteins containing one (C2A) or both (C2A+C2B) binding domains were used. It was alternatively proposed that one or both synaptotagmin binding domains are important for calcium-sensing and triggering of transmitter release. In this study the binding activity of native full-length synaptotagmin, immobilized on beads, was investigated. We found the kinetic parameters of Ca*+/phospholipid binding to be compatible with the role of calcium sensor for synaptotagmin (EC50 for calcium = 72 f 7 yM), with the two C2 domains supporting separate and complementary calcium sensing properties. The binding of native syntaxin to synaptotagmin was measurable in the absence of calcium, but was markedly stimulated (*.*-fold) in the presence of mM calcium. It may be speculated that the two domains have a synergistic action in fast synchronous transmitter release, whereas C2B domain alone may support slow asynchronous release, working as a high affinity calcium sensor. Key words: neurotransmitter
release, synaptic vesicle, synaptotagmin, Ca2’ /phosphoIipid
binding, syntaxin,
psychotropic drug
Regulation of neurotransmitter release from presynaptic nerve terminals involves a cascade of protein-protein interactions (1,2). Many of the proteins implicated in these multiple interactions have recently been identified (SNARE hypothesis). Correspondence to: Dr. Maurizio Popoli, Center of Neuropharmacology, Institute of Pharmacological Sciences (University of Milan), Via Balzaretti 9, 20133 Milan0 (Italy). Phone: +39 -2-26433379; fax: +39 -2-26433265
712
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
One of these proteins, synaptotagmin (SYT), an integral membrane protein of synaptic vesicles (SV) and secretory granules, was suggested to perform a central role in the exo-endocytotic cycle of SV (3), and was implicated in the mechanism of action of some psychotropic drugs (4,5). SYT sequence contains a long cytoplasmic tail bearing two repeats homologous of the calcium and phospholipid binding motif (C2) found in protein kinase C (6). Glutathione S-transferase (GST) fusion proteins containing the first (C2A) or both C2 domains (C2A+C2B) were shown to bind calcium and phospholipids with cooperativity for calcium (7). This finding, and the fact that knockout of SYT I gene abolishes the fast, synchronous component of transmitter release in hippocampal neurons (3), gave raise to the hypothesis that SYT is a low affinity calcium sensor mediating the triggering of calcium-dependent synchronous release. The EC50 for calcium of phospholipid binding was taken as a measure of how much the calcium-sensing property of SYT is compatible with the high calcium concentrations triggering release in presynaptic terminals. However, measurement of the EC50 for calcium of SYT obtained discordant results and it was suggested that the two C2 domains have different individual Ca2+dependence and may act synergistically in the binding (8). It was also suggested that the calcium-regulated interaction of SYT alternatively with phosphatidylinositol-3,4bisphosphate or phosphatidylinositol-3,4,5-trisphosphate may represent a bimodal switch between two different states of the calcium sensor (9). Recombinant SYT also binds syntaxin, a presynaptic membrane protein which has a functional interaction with voltage-gated calcium channels (10,ll). The physiological meaning of this interaction could be that of keeping in register calcium channel and calcium sensor (SYT) during release, but it was also proposed that it may be important in vesicle fusion. This binding is also calcium-dependent, with half-maximal binding (SYT I) occurring at calcium concentrations compatible with those reached in active zones following depolarization (200-300 uM, ref. 12). Moreover, it was recently suggested that the binding of syntaxin by SYT may also involve a synergistic action of the two C2 domains (13). It was also shown that the efficiency of the interaction SYTsyntaxin greatly increases when recombinant SYT is cleaved from GST, suggesting that GST-fusion proteins may have a folding different from native protein. Except for preliminary work (7,15), the interactions of SYT with Ca2+/phospholipids and syntaxin have not been investigated using purified, full-length protein. In view of the interesting but somewhat conflicting results obtained using recombinant proteins, especially when they are expressed as isolated functional domains, an investigation of native SYT binding properties may supply evidence helpful in a discussion of its role in binding and syntaxin binding of exocytosis in vivo. In this study Ca 2+/ phospholipid full-length SYT were investigated by using purified bovine brain protein immobilized on agarose beads. The results obtained here are compatible with the two C2 domains of SYT supporting complementary calcium sensing properties. It is hypothesized that the two domains may have a synergistic action in fast synchronous transmitter release, whereas C2B domain alone may support slow asynchronous release, working as a high affinity calcium sensor.
Methods Prebaration of subsvnaptosomal fractions from bovine brain. Synaptic vesicle- (LP2), synaptic membrane(LPl), and synaptic cytosol- (LS2) enriched fractions were prepared from bovine brain with minor modifications of a previously described
procedure processing.
713
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
(4). All fractions
were
aliquoted
and
stored
at -80°C
before
further
Purification of bovine brain synaptotagmin (SYT) immobilized on protein A-seaharose beads. Synaptotagmin (SYT) was purified from the LP2 fraction and immobilized on protein A-sepharose beads (Pharmacia), as previously described (5). SYT-beads, stored at 4oC, were used for Ca 2+-phospholipidand syntaxin-binding experiments within 1 week from preparation. SYT bound to the beads was quantitated for each individual preparation, by staining and densitometry of protein aliquots separated by SDS-PAGE. Bovine serum albumin was used as a standard. Preparation of ZH-labeled phosphatidylcholine-phosphatidylserine liposomes. Liposomes were prepared using phosphatidylcholine (PC, 10 mg/ml solutions in chloroform) and phosphatidylserine (PS, 10 mg/ml solutions in chloroform/methanol 955, Sigma). Liposomes with three different phospholipid contents were prepared: PC only; PC-PS (2.51); PC-PS (l:l), with 3H-PC (Amersham) added as a tracer. Dried phospholipids (1.75 mg total) were resuspended in 5 ml of Buffer A, 50 mM HEPES, 100 mM NaCI, pH 7.2 (prepared in calcium-free water, see below), by vigorous shaking, and then sonicated for 2 min (3 times) with a microprobe sonicator at 20 KHz. The liposomes were clarified by centrifugation for 10 min at 15,000 g to remove multilamellar aggregates, stored at 4oC, and used within 1 week.
IgG-LC-
31
rar
1 2 Fig.1 Purification of bovine brain synaptotagmin (SYT) immobilized on protein A-Sepharose beads. One pg of SYT separated by SDS-PAGE: 1. stained with Coomassie blue; 2. transferred to nitrocellulose membrane and incubated with monoclonal Ab 48. SYT, synaptotagmin; IgG-HC, IgG heavy chain; IgG-LC, IgG light chain. Mr standards in kilodaltons.
Q~phospholipid bindina of purified SYT-beads. Samples of SYT-beads (25-30 ~1 containing 600-800 ng SYT) were incubated in a total volume of 100 pl of Buffer A at various concentrations of free calcium (0.1 pM, luM, 20 pM, 60 PM, 200 pM, 1 mM, 6 mM), containing 3H-labeled liposomes (17.5 pg, 150,000-200,000 cpm). All solutions were prepared by using water for inorganic trace analysis (Fluka) and a 0.025 M standard calcium chloride solution (Sigma). This method was preferred to using Ca2+/EGTA buffers because in such buffers, calculated with standard software (MaxChelator 4.12 and Ligandy), free calcium concentrations above 0.2 yM resulted to
714
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, l!W7
be not reliable. Ba*+ and Sr*+ solutions were prepared as for calcium. The reaction was started by adding 3H-liposomes, carried on at room temperature with continuous mixing, and stopped after 15 min by adding 1 ml of cold Buffer A (of corresponding divalent cation concentration). The beads were pelleted by centrifugation and washed two more times. The bottom of the test tube containing the beads was cut and bound 31-1quantitated by liquid scintillation counting. Blank samples contained beads with bound Ab but no SYT, and were incubated in binding experiments at the same time as SYT-beads, at all free divalent cation concentrations. This amount of unspecific binding (8% average) was subtracted from all SYT-beads samples. Calculation of kinetic parameters and plotting of binding curves were carried out using GraphPad Prism 2 software. Bindina of native and recombinant syntaxin to SYT-beads. A detergent extract of crude synaptic membranes (LPl) was prepared by solubilizing LPl in buffer B (10mM Hepes, 150 mM NaCI, 0.1 mM EDTA, pH 7.4) containing 1% Triton X-100, at 4°C (2 mg protein/ml, final cont.). After removing insoluble material by centrifugation, the extract was diluted 4 times with buffer B. Samples of SYT-beads (0.6-0.8 pg SYT) were incubated with 1 ml of membrane extract, in the presence of either 2 mM EGTA or 1.5 mM CaCl2, for 2.5 h at 4°C. The beads were washed 3X with 1 ml of incubation buffer, and bound protein solubilized with SDS-electrophoresis buffer. After separation with SDS-PAGE and transfer to membrane, bound syntaxin was detected by using polyclonal Ab against syntaxin l A-l B and a peroxidase-conjugated second Ab (Southern Technology). For the binding of recombinant (4-265) syntaxin 1A the same amount of SYT-beads was incubated with 3 ug of syntaxin in buffer B containing 0.2% Triton X-100. After incubation the beads were processed as for native syntaxin. The blots were developed with standard ECL procedure (Amersham). Quantitation of bound syntaxin was carried out by using CCD camera images and image analysis software (Image 1.47, NIH) (4). Miscellaneous Procedures. SDS-PAGE and western blot were carried out as previously described (4). Protein content of subcellular fractions was measured with the BCA assay (Pierce). Hisg-tagged syntaxin 1A and polyclonal anti-syntaxin Ab were a kind gift of Dr. Theo Schafer (Friedrich Miescher Institute, Basel). Anti-SYT monoclonal Ab 48 was obtained from cultures of hybridoma 48 (16).
Results Purification of bovine brain svnaptotaamin (SYT) immobilized on protein A-Sepharose beads. In order to study the binding activity of native SYT the protein was purified by sequential incubation of a crude synaptic vesicle detergent extract with monoclonal Ab 48 (16) and protein A-Sepharose. MAb 48 recognizes the first identified and most abundant isoform of SYT in forebrain, SYT I (17). This MAb does not cross react with other SYT isoforms, as shown here (Fig. 1) and in other studies (9). Therefore we can exclude a contribution of other SYT isoforms to the binding activity measured here. As shown in Fig.1 SYT-beads contain virtually pure SYT (as determined by densitometry), together with heavy and light chain of the IgG. Yield of this procedure was 25-50 ng SYT/ ul of beads (dry weight). Ca2+/phospholipid bindina of native SYT-beads. In early studies of the protein binding activity it was found that SYT selectively binds amphipatic lipids containing negatively charged groups (glycosphingolipids and phospholipids), with highest affinity for acidic phospholipids (18,6,19). Successively the binding was investigated using GST-fusion protein containing the C2A binding site (7,20). C2A domain alone supported
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
715
Ca2+/phospholipid binding, but it was later suggested that the two C2 domains act synergistically in the interaction with the target membrane (8). The EC50 for calcium found in the latter study for the whole recombinant cytoplasmic tail (C2A + C2B) was 63 l.tM, a value higher than that found for the C2A alone (4-6 PM, ref. 7). This parameter is important, because it would be difficult to reconcile a low EC50 with the need of the calcium sensor to operate at the high calcium concentrations reached in presynaptic microdomains during exocytosis. We found an EC50 for calcium of 72 f 7 PM, by measuring the binding of 3H-labeled PC-PS (2.5:1)-liposomes to native SYT (Fig.2). The binding of native SYT showed cooperativity toward calcium, with an average Hill coefficient of 1.8 f 0.1 (cooperativity varied between 1.7 and 2.0 in eight different preparations; the curve showed in Fig. 2 has an Hill coefficient of 1.73). The high calcium concentration necessary for half-maximal binding we found is consistent with phospholipid binding of SYT being activated by high levels of calcium reached in active zones during depolarization.
6”““E =I 5000: ?Z 4 1 ooooI m
0
5000
, , ,,,,,r; 1o-7 1O-6
. 11111111 , I ,,,” 1o-4 1o-3 11l-2
1o-5
[Ca2+] Fig.2 Ca2+/phospholipid binding of native SYT-beads. Individual samples of SYT-beads incubated with phosphatidylcholine-phosphatidylserine (2.51) unilamellar liposomes, labeled with 3H-PC (200,000 cpm), in the presence of various concentrations of free calcium. Each point is the mean (k SEM) of six determinations. The binding curve shown has a cooperativity index (Hill coefficient) of 1.73 and an EC50 for calcium of 55.4 KM. Sensitivitv of the two C2 domains of native SYT to acidic phospholipid Iphosphatidylserine). A previous study reported that the two C2 domains bind phospholipids with different calcium sensitivity (8). It was found that recombinant C2A binds liposomes efficiently in the ~.LMto mM calcium concentration range and little or not at all around 1 PM calcium or lower. The binding activity of C2B instead appeared to be measurable over a large range of calcium concentrations, including concentrations I 1 PM, where C2A binding was undetectable (8). These results suggested that the binding at high calcium concentration is due to a combination of C2A and C2B binding, and that in the presence of 1 FM or lower the binding is due to C2B only. Therefore, in an attempt to dissect out the binding due respectively to (C2A+
716
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
C2B) and to C2B only, we investigated the binding of full-lenght native SYT at 200 uM and 1 uM free Ca2+. Binding reactions were performed using liposomes containing various percentages of acidic phospholipid (PS). As shown in Fig 3(a-b), when liposomes devoid of PS were used, the Ca2Vphospholipid binding of native SYT was negligible at both Ca2+ concentrations. Conversely, with PS-containing liposomes, the binding was significant and increased with increasing percentage of P.S. Hereafter, for sake of simplicity PC-PS(l:l)and (2.5:1)-liposomes are referred to as high PS- and low PS-liposomes. In the presence of 200 uM Ca2+ the ratio between binding to high PS- and to low PS-liposomes was 2.3 (Fig. 2b), while, when the binding was measured in the presence of 1 uM Ca2+, this ratio was equal to 4.5 (Fig. 2a). This result suggests that in native SYT C2B binding is more sensitive to changes in PS concentration.
200
pM
Ed+
Fig.3 Ca2+/phospholipid binding of native SYT-beads: divalent cation- and The binding of 3H-labeledacidic phospholipid-dependence. phospholipid liposomes by SYT-beads was measured at 1 uM or 200 uM PhosphatidylcholineCa2+ (a-b), Ba2+ (c-d), Sr2+ (e-f). (PC) phosphatidylserine (2.5: 1; 1:l) or pure phosphatidylcholine liposomes, labeled with 3H-PC (150,000 cpm) were used. The data represent the mean f S.E.M. of four experiments in triplicate. Sensitivitv of the two C2 domains of SYT to barium and strontium. It was of interest measure native SYT binding activity as a function of Ba2+ and Sr2+ concentrations, because these divalent cations are able to some extent to substitute for Ca2+ triggering release (21). We measured the binding to high and low PS-liposomes, 200 uM and 1 uM Ba2+ or Sr2+, to distinguish the binding of (C2A + C2B) from that
to in at of
717
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
C2B only. At 200 uM Ba2+ (Fig. 3d) and Sr 2+ (Fig. 3f) the binding of low PS-liposomes was much lower than in the presence of 200 PM Ca2+ (3.1- and 4.5-fold lower respectively, compare with Fig. 3b). However, when high PS-liposomes were used this difference was reduced (1.6- and 1.3-fold). By contrast in the presence of 1PM Ba2+ (Fig. 3c) and Sr2+ (Fig. 3e) the binding was not significantly different from 1 uM Ca2+ (Fig. 3a), for both low and high PS-liposomes. Furthermore, at variance with what observed at 1 and 200 PM Ca 2+, the acidic phospholipid-dependence in the presence of Ba2+ and Sr 2+ did not change when using 1 yM or 200 PM divalent cations, suggesting that only C2B binds in the presence of these cations. No significant binding was detected in the absence of one of the three divalent cations or in the presence of Mg2+ (not shown). In all these cases the binding was equivalent to the blank. For this reason binding in the presence of 1 PM and 200 PM Mg2+ was not reported in Fig.3. These results would suggest that: 1. The binding of C2A in the presence of 200 uM divalent cation is poorly activated by Ba2+ and Sr2+, as compared to Ca2+, but this specificity for Ca2+ decreases with increasing concentration of PS; 2. The binding at 1 PM divalent cation (C2B) is not dependent primarily on Ca2+ and can be activated as well by low micromolar Ba2+ and Sr2+.
123456
7
8
910
mr*--
EGTA
Ca*+
LPl
NATIVE SY NT
EGTA
Ca*+
RECOMBINANT SYNT Fig.4
Binding of native or recombinant syntaxin to native SYT-beads. SYTbeads incubated with either a detergent-extract of crude synaptic membranes (LPI) or with recombinant (4-265) syntaxin lA, in the presence of 2 mM EGTA or 1.5 mM calcium. Bound syntaxin evidenced with western blot analysis (4). Binding of native syntaxin; lanes l-2: in EGTA; lanes 3-4: in calcium. Lanes 5-6: LPl detergent extract (12.5 pg protein). Binding of recombinant syntaxin; lanes 7-8: in EGTA; lanes 9-10: in calcium. The amount of SYT present was checked by staining of blots with Ponceau S (not shown). Higher molecular weight bands in lanes 710 represent traces of oligomeric forms of syntaxin, often seen in recombinant syntaxin.
718
Native Synaptotagmin Biadiag Activity
Vol. 61, No. 7,1997
Bindino of native or recombinant svntaxin to native SYT-beads. It was shown that calcium stimulates the interaction between GST-fusion SYT and native syntaxin (12). However, when the interaction was studied using both recombinant proteins, conflicting results were obtained, with calcium alternatively stimulating or inhibiting the binding (22,23). We studied the interaction of the two proteins by using native full-length SYT-beads and either native or recombinant syntaxin. In the first case SYT-beads were incubated with a detergent extract of crude synaptic membranes in the absence (2 mM EGTA) or presence of 1.5 mM calcium; bound protein, separated by SDS-PAGE and transferred to nitrocellulose membrane, was probed with anti-syntaxin polyclonal Ab (Fig. 4). In different experiments SYTbeads were incubated with recombinant syntaxin ?A. Binding of native syntaxin to native SYT occurred in the absence of calcium but was markedly stimulated by this divalent cation (Ca2+-dependent/Ca 2+-independent ratio = 2.16 , n= 5, p
Discussion Several lines of evidence point to synaptotagmin as to a multifunctional protein with a central role in the exo-endocytotic cycle of synaptic vesicles. Most interesting among the various properties of SYT are the Ca2Vphospholipid binding (related to its putative function as a calcium sensor) and the interaction with other proteins in the presynaptic core complex (e.g., with syntaxin and voltage-gated calcium channels)(24). The two binding activities were investigated in this study by using native full-length SYT immobilized on beads. The results obtained may be summarized as follows: 1. Full-length native SYT displays an EC50 for calcium (72 PM) compatible with the high calcium concentrations reached in presynaptic microdomains during transmitter release. The phospholipid binding shows a different dependence on divalent cations and acidic phospholipid content of the target membrane, when measured at 1 ~.LMor 200 uM divalent cation. A major component of the binding measured at 200 uM divalent cation (which, according to the studies on GST-fusion SYT, can be accounted for by C2A) is primarily dependent on the presence of calcium, being much lower in the presence of Ba2+ or Sr2+ (Fig.2). However the specificity for calcium decreases when acidic phospholipid content in the target membrane is increased. By contrast binding activity at 1 j.rM divalent cation (a concentration at which only C2B domain binds phospholipids) is sensitive to calcium but is activated as well by Ba2+ and Sr2+, at both liposome compositions tested. A similar difference has been observed in the calcium sensitivity of recombinant C2 domains (8) and it was speculated that C2B, for its higher sensitivity to Ca2+, may bind other divalent cations with similar efficiency (8). The present work, measuring for the first time C2B binding in the presence of Ba2+ and Sr2+, seems to confirm this hypothesis. With regard to sensitivity to acidic phospholipid (PS) content of the target membrane, the binding of native protein appears to behave quite differently at different calcium concentrations, again evidencing the different features of the C2 domains. The phospholipid binding measured at 200 uM Ca 2+ (C2A+C2B) is less dependent on PS content, with a 2.3-fold increase when changing from low to high PS-liposomes. The binding measured at 1 PM Ca2+ (C2B) has a higher dependence on PS content, with a 4.5fold increase when changing from low to high PS-liposomes. This result
Vol. 61, No. 7, 1997
Native Synaptotagmin Binding Activity
emphasizes the importance of the enrichment which may greatly condition SYT function (25).
of PS in membrane
719
patches,
a factor
It must be mentioned that this distinction of the properties of the two C2 domains in the full-length protein is necessarily an approximation, justified by the difficulty to obtain isolated binding domains from native protein. However, when made possible, an investigation of binding properties of isolated native C2 domains may also be exposed to the same criticism (25) addressed to studies using recombinant C2 domains (e.g., problems with correct folding). Although the use of native full-length protein is closer to the situation in vivo, we are aware that a problem of correct folding may also occur in this case (that is the folding may be different from when the protein is inserted in membranes). This is, however, a limitation of all in vitro studies using purified, immobilized proteins. 2. Native SYT binds both native syntaxin and recombinant syntaxin 1A. The binding of native syntaxin is markedly stimulated by mM calcium, whereas the binding of recombinant syntaxin at the conditions used here is calcium-independent. A possible reason for this discrepancy is that the recombinant syntaxin used (4-265) lacks the Cterminal hydrophobic transmembrane domain. In previous work calcium-dependent binding of full-lenght recombinant syntaxin to recombinant SYT was shown (22), but when soluble syntaxin (4-266) was used the binding was inhibited by calcium (23). However, when the same truncated syntaxin was immobilized on beads and tested with soluble SYT cytoplasmic domain the binding was calcium-dependent (23). Furthermore it is also possible that the lack of posttranslational modifications in recombinant syntaxin may induce an incorrect folding of the protein. However, our data obtained using native proteins emphasize the need for checking the recombinant protein-protein interactions, whenever possible, with native proteins. Our results demonstrate that in this last situation, again closer to the situation in vivo, the binding SYT-syntaxin is indeed stimulated by calcium. It was not investigated here whether SYT-bound syntaxin is complexed with other core complex proteins (e.g., synaptobrevin and SNAP-25). It is conceivable that at least part of this bound syntaxin is not present in monomeric form, and it would be interesting to test whether calcium modifies the binding of other proteins in this system (native protein-protein interaction). Two distinct components of transmitter release were shown in hippocampal neurons: a major, fast component which is inhibited by Sr2+, and a minor, slow component which is facilitated by substitution of Ca2+ with Sr2+ (21). It was proposed that the two components of release may be mediated by different calcium sensors: a low affinity sensor, promoting release when Ca2+ concentration in active zones is quickly raised, and a high affinity sensor, efficient in promoting release when Ca2+ concentration is much lower and the first sensor is not active. As the knock out of SYT I gene suppressed the fast component of release, without affecting the slow component, it was proposed that SYT I is the low affinity sensor, whereas another isoform of SYT (SYT III or IV) may be the high affinity sensor (3). In view of what reported by other authors (8,25) and of our own observations, it could be suggested that such a distinction is not strictly necessary, as SYT may be at the same time low and high affinity sensor, with the former activity due to synergistic action of the C2 domains and the latter localized to the C2B domain only. In hippocampus, where the effect of SYT I knockout was investigated, SYT III and IV are also expressed at a low level (26). The presence of these isoforms (or of other less abundant C2 domain-containing proteins) may not suffice to ensure an adequate amount of C2A (or better of C2A + C2B) to sustain fast release at high calcium concentrations, whereas even low levels of C2B, with its high sensitivity to calcium and capability to use other divalent cations, may be sufficient to sustain slow asynchronous release.
720
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
It was recently found that long-term treatment with certain psychotropic drugs, known to promote an increase in transmitter release in the hippocampus (5HT reuptake blockers) also increases the activity of presynaptic Ca2+/calmodulin-dependent protein kinase II (CaMKII) and Ca2+/calmodulin-dependent SYT phosphorylation (4,5). To our knowledge this was the first example of a modification in a protein regulating transmitter release induced by psychotropic drugs used in the therapy of psychiatric disorders. SYT was previously shown to be a substrate in synaptic vesicles in vitro for two vesicular protein kinases: (CaMKII) and casein kinase II (CKII) (27,28). An advantage of using native protein is that it always contains phosphorylation sites. It will be interesting to study whether phosphorylation of the calcium sensor changes its binding activity, and how this modification may be involved in the action of psychotropic drugs.
Acknowledgement This study was supported by a grant from Telethon Italy (no. 539) to M.P.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
T. SOLLNER, M.K. BENNETT, S.W. WHITEHEART, R.H. SCHELLER and J.E. ROTHMAN, Cell 75 409-418 (1993). M.K. BENNETT and R.H. SCHELLER, Annu. Rev. Biochem. 63 63-100 (1994). M. GEPPERT, Y. GODA, R.E. HAMMER, C. LI, T.W. ROSAHL, CF. STEVENS and T.C. SUDHOF, Cell 79 717-727 (1994). M. POPOLI, C. VOCATURO, J. PEREZ, E. SMERALDI and G. RACAGNI, Mol. Pharmacol. 48, 623-629 (1995). M. POPOLI, A. VENEGONI, C. VOCATURO, L. BUFFA, J. PEREZ, E. SMERALDI and G. RACAGNI, Mol. Pharmacol. 51 l-8 (1997). M.S. PERIN, V.A. FRIED, G.A. MIGNERY, R. JAHN and T.C. SUDHOF, Nature 345 260-263 (1990). B.A. DAVLETOV and T.C. SUDHOF, J. Biol. Chem. 268 26386-26390 (1993). C.K. DAMER and C.E. CREUTZ, J. Biol. Chem. 269 31115-31123 (1994). G. SCHIAVO, Q.M. GU, G.D. PRESTWICH, T.H. SOLLNER and J.E. ROTHMAN, Proc. Natl. Acad. Sci. USA 93 13327-13332 (1996). M.K. BENNETT, N. CALAKOS and R.H. SCHELLER, Science 257 255-259 (1992). I. BEZPROZVANNY, R.H. SCHELLER and R.W. TSIEN, Nature 378 623-626 (1995). C. LI, B. ULLRICH, J.Z. YOUNG, R.G.W. ANDERSON, N. BROSE and T.C. SUDHOF, Nature 375 594-599 (1995). E.R. CHAPMAN, S. AN, J.M. EDWARDSON and R. JAHN, J. Biol. Chem. 271 5844-5849 (1996). S. SUGITA, Y. HATA and T.C. SUDHOF, J. Biol. Chem. 271 1262-1265 (1996). M. POPOLI and R. PATERNO’, NeuroReport 3 177-180 (1992). W.D. MATTHEW, L. TSAVALER and L.F. REICHARDT, J. Cell. Biol. 91 257-269 (1981). B. WENDLAND, K.G. MILLER, JSCHILLING and R.H. SCHELLER, Neuron 6 993-1007 (1991). M., POPOLI and A. MENGANO, Neurochem. Res. 13 63-67 (1988). M. POPOLI, Neuroscience 54 323-328 (1993). C. LI, B.A. DAVLETOV and T.C. SUDHOF, J. Biol. Chem. 270 24898-24902 (1995).
Vol. 61, No. 7, 1997
21. 22. 23. 24. 25. 26. 27. 28.
Native Syoaptotagmin Biidiog Activity
Y. GODA and C.F. STEVENS, Proc. Natl. Acad. Sci. USA 91 12942-12946 (1994). E.R. CHAPMAN, P.I. HANSON, S. AN and R. JAHN, J. Biol. Chem. 270 23667-23671 (1995). Y. KEE and R.H. SCHELLER, J. Neurosci. 16 1975-1981 (1996). Z.H. SHENG, J. REllIG, M. TAKAHASHI and W.A. CATTERALL, Neuron 13 1303-1313 (1994). M. FUKUDA, T. KOJIMA and K. MIKOSHIBA, J. Biol. Chem. 271 8430-8434 (1996). B. ULLRICH, C. LI, J.Z. ZHANG, H. McMAHON, R.G.W. ANDERSON, M. GEPPERT and T.C. SUDHOF, Neuron 13 1281-1291 (1994). M. POPOLI, FEBS Lett. 317 85-88 (1993). M.K., BENNETT, K.G. MILLER and R.H. SCHELLER, J. Neurosci. 13 17011707 (1993).
721
PII soo24-3205(97)00536-5
ELSEVIER
Ca*+/PHOSPHOLIPID-BINDING BINDING OF NATIVE
Maurizio Racagnil.
Popolil-25
, Albert0
AND SYNTAXINSYNAPTOTAGMIN I
Venegonit
, Liliana
Buffat,
and Giorgio
‘Center of Neuropharmacology, Institute of Pharmacological Sciences, University of Milan, 2Department of Neurological Sciences, II School of Medicine, University of Naples Federico II (Italy). (Received in fmal form May 19, 1997)
Summary
Synaptotagmin, a synaptic vesicle protein endowed with multiple properties, is the putative calcium sensor in neuroexocytosis. Ca*+/phospholipid binding and syntaxin binding activity of synaptotagmin were previously investigated using recombinant fusion proteins. In phospholipid binding the EC50 for calcium obtained was different when fusion proteins containing one (C2A) or both (C2A+C2B) binding domains were used. It was alternatively proposed that one or both synaptotagmin binding domains are important for calcium-sensing and triggering of transmitter release. In this study the binding activity of native full-length synaptotagmin, immobilized on beads, was investigated. We found the kinetic parameters of Ca*+/phospholipid binding to be compatible with the role of calcium sensor for synaptotagmin (EC50 for calcium = 72 f 7 yM), with the two C2 domains supporting separate and complementary calcium sensing properties. The binding of native syntaxin to synaptotagmin was measurable in the absence of calcium, but was markedly stimulated (*.*-fold) in the presence of mM calcium. It may be speculated that the two domains have a synergistic action in fast synchronous transmitter release, whereas C2B domain alone may support slow asynchronous release, working as a high affinity calcium sensor. Key words: neurotransmitter
release, synaptic vesicle, synaptotagmin, Ca2’ /phosphoIipid
binding, syntaxin,
psychotropic drug
Regulation of neurotransmitter release from presynaptic nerve terminals involves a cascade of protein-protein interactions (1,2). Many of the proteins implicated in these multiple interactions have recently been identified (SNARE hypothesis). Correspondence to: Dr. Maurizio Popoli, Center of Neuropharmacology, Institute of Pharmacological Sciences (University of Milan), Via Balzaretti 9, 20133 Milan0 (Italy). Phone: +39 -2-26433379; fax: +39 -2-26433265
712
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
One of these proteins, synaptotagmin (SYT), an integral membrane protein of synaptic vesicles (SV) and secretory granules, was suggested to perform a central role in the exo-endocytotic cycle of SV (3), and was implicated in the mechanism of action of some psychotropic drugs (4,5). SYT sequence contains a long cytoplasmic tail bearing two repeats homologous of the calcium and phospholipid binding motif (C2) found in protein kinase C (6). Glutathione S-transferase (GST) fusion proteins containing the first (C2A) or both C2 domains (C2A+C2B) were shown to bind calcium and phospholipids with cooperativity for calcium (7). This finding, and the fact that knockout of SYT I gene abolishes the fast, synchronous component of transmitter release in hippocampal neurons (3), gave raise to the hypothesis that SYT is a low affinity calcium sensor mediating the triggering of calcium-dependent synchronous release. The EC50 for calcium of phospholipid binding was taken as a measure of how much the calcium-sensing property of SYT is compatible with the high calcium concentrations triggering release in presynaptic terminals. However, measurement of the EC50 for calcium of SYT obtained discordant results and it was suggested that the two C2 domains have different individual Ca2+dependence and may act synergistically in the binding (8). It was also suggested that the calcium-regulated interaction of SYT alternatively with phosphatidylinositol-3,4bisphosphate or phosphatidylinositol-3,4,5-trisphosphate may represent a bimodal switch between two different states of the calcium sensor (9). Recombinant SYT also binds syntaxin, a presynaptic membrane protein which has a functional interaction with voltage-gated calcium channels (10,ll). The physiological meaning of this interaction could be that of keeping in register calcium channel and calcium sensor (SYT) during release, but it was also proposed that it may be important in vesicle fusion. This binding is also calcium-dependent, with half-maximal binding (SYT I) occurring at calcium concentrations compatible with those reached in active zones following depolarization (200-300 uM, ref. 12). Moreover, it was recently suggested that the binding of syntaxin by SYT may also involve a synergistic action of the two C2 domains (13). It was also shown that the efficiency of the interaction SYTsyntaxin greatly increases when recombinant SYT is cleaved from GST, suggesting that GST-fusion proteins may have a folding different from native protein. Except for preliminary work (7,15), the interactions of SYT with Ca2+/phospholipids and syntaxin have not been investigated using purified, full-length protein. In view of the interesting but somewhat conflicting results obtained using recombinant proteins, especially when they are expressed as isolated functional domains, an investigation of native SYT binding properties may supply evidence helpful in a discussion of its role in binding and syntaxin binding of exocytosis in vivo. In this study Ca 2+/ phospholipid full-length SYT were investigated by using purified bovine brain protein immobilized on agarose beads. The results obtained here are compatible with the two C2 domains of SYT supporting complementary calcium sensing properties. It is hypothesized that the two domains may have a synergistic action in fast synchronous transmitter release, whereas C2B domain alone may support slow asynchronous release, working as a high affinity calcium sensor.
Methods Prebaration of subsvnaptosomal fractions from bovine brain. Synaptic vesicle- (LP2), synaptic membrane(LPl), and synaptic cytosol- (LS2) enriched fractions were prepared from bovine brain with minor modifications of a previously described
procedure processing.
713
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
(4). All fractions
were
aliquoted
and
stored
at -80°C
before
further
Purification of bovine brain synaptotagmin (SYT) immobilized on protein A-seaharose beads. Synaptotagmin (SYT) was purified from the LP2 fraction and immobilized on protein A-sepharose beads (Pharmacia), as previously described (5). SYT-beads, stored at 4oC, were used for Ca 2+-phospholipidand syntaxin-binding experiments within 1 week from preparation. SYT bound to the beads was quantitated for each individual preparation, by staining and densitometry of protein aliquots separated by SDS-PAGE. Bovine serum albumin was used as a standard. Preparation of ZH-labeled phosphatidylcholine-phosphatidylserine liposomes. Liposomes were prepared using phosphatidylcholine (PC, 10 mg/ml solutions in chloroform) and phosphatidylserine (PS, 10 mg/ml solutions in chloroform/methanol 955, Sigma). Liposomes with three different phospholipid contents were prepared: PC only; PC-PS (2.51); PC-PS (l:l), with 3H-PC (Amersham) added as a tracer. Dried phospholipids (1.75 mg total) were resuspended in 5 ml of Buffer A, 50 mM HEPES, 100 mM NaCI, pH 7.2 (prepared in calcium-free water, see below), by vigorous shaking, and then sonicated for 2 min (3 times) with a microprobe sonicator at 20 KHz. The liposomes were clarified by centrifugation for 10 min at 15,000 g to remove multilamellar aggregates, stored at 4oC, and used within 1 week.
IgG-LC-
31
rar
1 2 Fig.1 Purification of bovine brain synaptotagmin (SYT) immobilized on protein A-Sepharose beads. One pg of SYT separated by SDS-PAGE: 1. stained with Coomassie blue; 2. transferred to nitrocellulose membrane and incubated with monoclonal Ab 48. SYT, synaptotagmin; IgG-HC, IgG heavy chain; IgG-LC, IgG light chain. Mr standards in kilodaltons.
Q~phospholipid bindina of purified SYT-beads. Samples of SYT-beads (25-30 ~1 containing 600-800 ng SYT) were incubated in a total volume of 100 pl of Buffer A at various concentrations of free calcium (0.1 pM, luM, 20 pM, 60 PM, 200 pM, 1 mM, 6 mM), containing 3H-labeled liposomes (17.5 pg, 150,000-200,000 cpm). All solutions were prepared by using water for inorganic trace analysis (Fluka) and a 0.025 M standard calcium chloride solution (Sigma). This method was preferred to using Ca2+/EGTA buffers because in such buffers, calculated with standard software (MaxChelator 4.12 and Ligandy), free calcium concentrations above 0.2 yM resulted to
714
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, l!W7
be not reliable. Ba*+ and Sr*+ solutions were prepared as for calcium. The reaction was started by adding 3H-liposomes, carried on at room temperature with continuous mixing, and stopped after 15 min by adding 1 ml of cold Buffer A (of corresponding divalent cation concentration). The beads were pelleted by centrifugation and washed two more times. The bottom of the test tube containing the beads was cut and bound 31-1quantitated by liquid scintillation counting. Blank samples contained beads with bound Ab but no SYT, and were incubated in binding experiments at the same time as SYT-beads, at all free divalent cation concentrations. This amount of unspecific binding (8% average) was subtracted from all SYT-beads samples. Calculation of kinetic parameters and plotting of binding curves were carried out using GraphPad Prism 2 software. Bindina of native and recombinant syntaxin to SYT-beads. A detergent extract of crude synaptic membranes (LPl) was prepared by solubilizing LPl in buffer B (10mM Hepes, 150 mM NaCI, 0.1 mM EDTA, pH 7.4) containing 1% Triton X-100, at 4°C (2 mg protein/ml, final cont.). After removing insoluble material by centrifugation, the extract was diluted 4 times with buffer B. Samples of SYT-beads (0.6-0.8 pg SYT) were incubated with 1 ml of membrane extract, in the presence of either 2 mM EGTA or 1.5 mM CaCl2, for 2.5 h at 4°C. The beads were washed 3X with 1 ml of incubation buffer, and bound protein solubilized with SDS-electrophoresis buffer. After separation with SDS-PAGE and transfer to membrane, bound syntaxin was detected by using polyclonal Ab against syntaxin l A-l B and a peroxidase-conjugated second Ab (Southern Technology). For the binding of recombinant (4-265) syntaxin 1A the same amount of SYT-beads was incubated with 3 ug of syntaxin in buffer B containing 0.2% Triton X-100. After incubation the beads were processed as for native syntaxin. The blots were developed with standard ECL procedure (Amersham). Quantitation of bound syntaxin was carried out by using CCD camera images and image analysis software (Image 1.47, NIH) (4). Miscellaneous Procedures. SDS-PAGE and western blot were carried out as previously described (4). Protein content of subcellular fractions was measured with the BCA assay (Pierce). Hisg-tagged syntaxin 1A and polyclonal anti-syntaxin Ab were a kind gift of Dr. Theo Schafer (Friedrich Miescher Institute, Basel). Anti-SYT monoclonal Ab 48 was obtained from cultures of hybridoma 48 (16).
Results Purification of bovine brain svnaptotaamin (SYT) immobilized on protein A-Sepharose beads. In order to study the binding activity of native SYT the protein was purified by sequential incubation of a crude synaptic vesicle detergent extract with monoclonal Ab 48 (16) and protein A-Sepharose. MAb 48 recognizes the first identified and most abundant isoform of SYT in forebrain, SYT I (17). This MAb does not cross react with other SYT isoforms, as shown here (Fig. 1) and in other studies (9). Therefore we can exclude a contribution of other SYT isoforms to the binding activity measured here. As shown in Fig.1 SYT-beads contain virtually pure SYT (as determined by densitometry), together with heavy and light chain of the IgG. Yield of this procedure was 25-50 ng SYT/ ul of beads (dry weight). Ca2+/phospholipid bindina of native SYT-beads. In early studies of the protein binding activity it was found that SYT selectively binds amphipatic lipids containing negatively charged groups (glycosphingolipids and phospholipids), with highest affinity for acidic phospholipids (18,6,19). Successively the binding was investigated using GST-fusion protein containing the C2A binding site (7,20). C2A domain alone supported
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
715
Ca2+/phospholipid binding, but it was later suggested that the two C2 domains act synergistically in the interaction with the target membrane (8). The EC50 for calcium found in the latter study for the whole recombinant cytoplasmic tail (C2A + C2B) was 63 l.tM, a value higher than that found for the C2A alone (4-6 PM, ref. 7). This parameter is important, because it would be difficult to reconcile a low EC50 with the need of the calcium sensor to operate at the high calcium concentrations reached in presynaptic microdomains during exocytosis. We found an EC50 for calcium of 72 f 7 PM, by measuring the binding of 3H-labeled PC-PS (2.5:1)-liposomes to native SYT (Fig.2). The binding of native SYT showed cooperativity toward calcium, with an average Hill coefficient of 1.8 f 0.1 (cooperativity varied between 1.7 and 2.0 in eight different preparations; the curve showed in Fig. 2 has an Hill coefficient of 1.73). The high calcium concentration necessary for half-maximal binding we found is consistent with phospholipid binding of SYT being activated by high levels of calcium reached in active zones during depolarization.
6”““E =I 5000: ?Z 4 1 ooooI m
0
5000
, , ,,,,,r; 1o-7 1O-6
. 11111111 , I ,,,” 1o-4 1o-3 11l-2
1o-5
[Ca2+] Fig.2 Ca2+/phospholipid binding of native SYT-beads. Individual samples of SYT-beads incubated with phosphatidylcholine-phosphatidylserine (2.51) unilamellar liposomes, labeled with 3H-PC (200,000 cpm), in the presence of various concentrations of free calcium. Each point is the mean (k SEM) of six determinations. The binding curve shown has a cooperativity index (Hill coefficient) of 1.73 and an EC50 for calcium of 55.4 KM. Sensitivitv of the two C2 domains of native SYT to acidic phospholipid Iphosphatidylserine). A previous study reported that the two C2 domains bind phospholipids with different calcium sensitivity (8). It was found that recombinant C2A binds liposomes efficiently in the ~.LMto mM calcium concentration range and little or not at all around 1 PM calcium or lower. The binding activity of C2B instead appeared to be measurable over a large range of calcium concentrations, including concentrations I 1 PM, where C2A binding was undetectable (8). These results suggested that the binding at high calcium concentration is due to a combination of C2A and C2B binding, and that in the presence of 1 FM or lower the binding is due to C2B only. Therefore, in an attempt to dissect out the binding due respectively to (C2A+
716
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
C2B) and to C2B only, we investigated the binding of full-lenght native SYT at 200 uM and 1 uM free Ca2+. Binding reactions were performed using liposomes containing various percentages of acidic phospholipid (PS). As shown in Fig 3(a-b), when liposomes devoid of PS were used, the Ca2Vphospholipid binding of native SYT was negligible at both Ca2+ concentrations. Conversely, with PS-containing liposomes, the binding was significant and increased with increasing percentage of P.S. Hereafter, for sake of simplicity PC-PS(l:l)and (2.5:1)-liposomes are referred to as high PS- and low PS-liposomes. In the presence of 200 uM Ca2+ the ratio between binding to high PS- and to low PS-liposomes was 2.3 (Fig. 2b), while, when the binding was measured in the presence of 1 uM Ca2+, this ratio was equal to 4.5 (Fig. 2a). This result suggests that in native SYT C2B binding is more sensitive to changes in PS concentration.
200
pM
Ed+
Fig.3 Ca2+/phospholipid binding of native SYT-beads: divalent cation- and The binding of 3H-labeledacidic phospholipid-dependence. phospholipid liposomes by SYT-beads was measured at 1 uM or 200 uM PhosphatidylcholineCa2+ (a-b), Ba2+ (c-d), Sr2+ (e-f). (PC) phosphatidylserine (2.5: 1; 1:l) or pure phosphatidylcholine liposomes, labeled with 3H-PC (150,000 cpm) were used. The data represent the mean f S.E.M. of four experiments in triplicate. Sensitivitv of the two C2 domains of SYT to barium and strontium. It was of interest measure native SYT binding activity as a function of Ba2+ and Sr2+ concentrations, because these divalent cations are able to some extent to substitute for Ca2+ triggering release (21). We measured the binding to high and low PS-liposomes, 200 uM and 1 uM Ba2+ or Sr2+, to distinguish the binding of (C2A + C2B) from that
to in at of
717
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
C2B only. At 200 uM Ba2+ (Fig. 3d) and Sr 2+ (Fig. 3f) the binding of low PS-liposomes was much lower than in the presence of 200 PM Ca2+ (3.1- and 4.5-fold lower respectively, compare with Fig. 3b). However, when high PS-liposomes were used this difference was reduced (1.6- and 1.3-fold). By contrast in the presence of 1PM Ba2+ (Fig. 3c) and Sr2+ (Fig. 3e) the binding was not significantly different from 1 uM Ca2+ (Fig. 3a), for both low and high PS-liposomes. Furthermore, at variance with what observed at 1 and 200 PM Ca 2+, the acidic phospholipid-dependence in the presence of Ba2+ and Sr 2+ did not change when using 1 yM or 200 PM divalent cations, suggesting that only C2B binds in the presence of these cations. No significant binding was detected in the absence of one of the three divalent cations or in the presence of Mg2+ (not shown). In all these cases the binding was equivalent to the blank. For this reason binding in the presence of 1 PM and 200 PM Mg2+ was not reported in Fig.3. These results would suggest that: 1. The binding of C2A in the presence of 200 uM divalent cation is poorly activated by Ba2+ and Sr2+, as compared to Ca2+, but this specificity for Ca2+ decreases with increasing concentration of PS; 2. The binding at 1 PM divalent cation (C2B) is not dependent primarily on Ca2+ and can be activated as well by low micromolar Ba2+ and Sr2+.
123456
7
8
910
mr*--
EGTA
Ca*+
LPl
NATIVE SY NT
EGTA
Ca*+
RECOMBINANT SYNT Fig.4
Binding of native or recombinant syntaxin to native SYT-beads. SYTbeads incubated with either a detergent-extract of crude synaptic membranes (LPI) or with recombinant (4-265) syntaxin lA, in the presence of 2 mM EGTA or 1.5 mM calcium. Bound syntaxin evidenced with western blot analysis (4). Binding of native syntaxin; lanes l-2: in EGTA; lanes 3-4: in calcium. Lanes 5-6: LPl detergent extract (12.5 pg protein). Binding of recombinant syntaxin; lanes 7-8: in EGTA; lanes 9-10: in calcium. The amount of SYT present was checked by staining of blots with Ponceau S (not shown). Higher molecular weight bands in lanes 710 represent traces of oligomeric forms of syntaxin, often seen in recombinant syntaxin.
718
Native Synaptotagmin Biadiag Activity
Vol. 61, No. 7,1997
Bindino of native or recombinant svntaxin to native SYT-beads. It was shown that calcium stimulates the interaction between GST-fusion SYT and native syntaxin (12). However, when the interaction was studied using both recombinant proteins, conflicting results were obtained, with calcium alternatively stimulating or inhibiting the binding (22,23). We studied the interaction of the two proteins by using native full-length SYT-beads and either native or recombinant syntaxin. In the first case SYT-beads were incubated with a detergent extract of crude synaptic membranes in the absence (2 mM EGTA) or presence of 1.5 mM calcium; bound protein, separated by SDS-PAGE and transferred to nitrocellulose membrane, was probed with anti-syntaxin polyclonal Ab (Fig. 4). In different experiments SYTbeads were incubated with recombinant syntaxin ?A. Binding of native syntaxin to native SYT occurred in the absence of calcium but was markedly stimulated by this divalent cation (Ca2+-dependent/Ca 2+-independent ratio = 2.16 , n= 5, p
Discussion Several lines of evidence point to synaptotagmin as to a multifunctional protein with a central role in the exo-endocytotic cycle of synaptic vesicles. Most interesting among the various properties of SYT are the Ca2Vphospholipid binding (related to its putative function as a calcium sensor) and the interaction with other proteins in the presynaptic core complex (e.g., with syntaxin and voltage-gated calcium channels)(24). The two binding activities were investigated in this study by using native full-length SYT immobilized on beads. The results obtained may be summarized as follows: 1. Full-length native SYT displays an EC50 for calcium (72 PM) compatible with the high calcium concentrations reached in presynaptic microdomains during transmitter release. The phospholipid binding shows a different dependence on divalent cations and acidic phospholipid content of the target membrane, when measured at 1 ~.LMor 200 uM divalent cation. A major component of the binding measured at 200 uM divalent cation (which, according to the studies on GST-fusion SYT, can be accounted for by C2A) is primarily dependent on the presence of calcium, being much lower in the presence of Ba2+ or Sr2+ (Fig.2). However the specificity for calcium decreases when acidic phospholipid content in the target membrane is increased. By contrast binding activity at 1 j.rM divalent cation (a concentration at which only C2B domain binds phospholipids) is sensitive to calcium but is activated as well by Ba2+ and Sr2+, at both liposome compositions tested. A similar difference has been observed in the calcium sensitivity of recombinant C2 domains (8) and it was speculated that C2B, for its higher sensitivity to Ca2+, may bind other divalent cations with similar efficiency (8). The present work, measuring for the first time C2B binding in the presence of Ba2+ and Sr2+, seems to confirm this hypothesis. With regard to sensitivity to acidic phospholipid (PS) content of the target membrane, the binding of native protein appears to behave quite differently at different calcium concentrations, again evidencing the different features of the C2 domains. The phospholipid binding measured at 200 uM Ca 2+ (C2A+C2B) is less dependent on PS content, with a 2.3-fold increase when changing from low to high PS-liposomes. The binding measured at 1 PM Ca2+ (C2B) has a higher dependence on PS content, with a 4.5fold increase when changing from low to high PS-liposomes. This result
Vol. 61, No. 7, 1997
Native Synaptotagmin Binding Activity
emphasizes the importance of the enrichment which may greatly condition SYT function (25).
of PS in membrane
719
patches,
a factor
It must be mentioned that this distinction of the properties of the two C2 domains in the full-length protein is necessarily an approximation, justified by the difficulty to obtain isolated binding domains from native protein. However, when made possible, an investigation of binding properties of isolated native C2 domains may also be exposed to the same criticism (25) addressed to studies using recombinant C2 domains (e.g., problems with correct folding). Although the use of native full-length protein is closer to the situation in vivo, we are aware that a problem of correct folding may also occur in this case (that is the folding may be different from when the protein is inserted in membranes). This is, however, a limitation of all in vitro studies using purified, immobilized proteins. 2. Native SYT binds both native syntaxin and recombinant syntaxin 1A. The binding of native syntaxin is markedly stimulated by mM calcium, whereas the binding of recombinant syntaxin at the conditions used here is calcium-independent. A possible reason for this discrepancy is that the recombinant syntaxin used (4-265) lacks the Cterminal hydrophobic transmembrane domain. In previous work calcium-dependent binding of full-lenght recombinant syntaxin to recombinant SYT was shown (22), but when soluble syntaxin (4-266) was used the binding was inhibited by calcium (23). However, when the same truncated syntaxin was immobilized on beads and tested with soluble SYT cytoplasmic domain the binding was calcium-dependent (23). Furthermore it is also possible that the lack of posttranslational modifications in recombinant syntaxin may induce an incorrect folding of the protein. However, our data obtained using native proteins emphasize the need for checking the recombinant protein-protein interactions, whenever possible, with native proteins. Our results demonstrate that in this last situation, again closer to the situation in vivo, the binding SYT-syntaxin is indeed stimulated by calcium. It was not investigated here whether SYT-bound syntaxin is complexed with other core complex proteins (e.g., synaptobrevin and SNAP-25). It is conceivable that at least part of this bound syntaxin is not present in monomeric form, and it would be interesting to test whether calcium modifies the binding of other proteins in this system (native protein-protein interaction). Two distinct components of transmitter release were shown in hippocampal neurons: a major, fast component which is inhibited by Sr2+, and a minor, slow component which is facilitated by substitution of Ca2+ with Sr2+ (21). It was proposed that the two components of release may be mediated by different calcium sensors: a low affinity sensor, promoting release when Ca2+ concentration in active zones is quickly raised, and a high affinity sensor, efficient in promoting release when Ca2+ concentration is much lower and the first sensor is not active. As the knock out of SYT I gene suppressed the fast component of release, without affecting the slow component, it was proposed that SYT I is the low affinity sensor, whereas another isoform of SYT (SYT III or IV) may be the high affinity sensor (3). In view of what reported by other authors (8,25) and of our own observations, it could be suggested that such a distinction is not strictly necessary, as SYT may be at the same time low and high affinity sensor, with the former activity due to synergistic action of the C2 domains and the latter localized to the C2B domain only. In hippocampus, where the effect of SYT I knockout was investigated, SYT III and IV are also expressed at a low level (26). The presence of these isoforms (or of other less abundant C2 domain-containing proteins) may not suffice to ensure an adequate amount of C2A (or better of C2A + C2B) to sustain fast release at high calcium concentrations, whereas even low levels of C2B, with its high sensitivity to calcium and capability to use other divalent cations, may be sufficient to sustain slow asynchronous release.
720
Native Synaptotagmin Binding Activity
Vol. 61, No. 7, 1997
It was recently found that long-term treatment with certain psychotropic drugs, known to promote an increase in transmitter release in the hippocampus (5HT reuptake blockers) also increases the activity of presynaptic Ca2+/calmodulin-dependent protein kinase II (CaMKII) and Ca2+/calmodulin-dependent SYT phosphorylation (4,5). To our knowledge this was the first example of a modification in a protein regulating transmitter release induced by psychotropic drugs used in the therapy of psychiatric disorders. SYT was previously shown to be a substrate in synaptic vesicles in vitro for two vesicular protein kinases: (CaMKII) and casein kinase II (CKII) (27,28). An advantage of using native protein is that it always contains phosphorylation sites. It will be interesting to study whether phosphorylation of the calcium sensor changes its binding activity, and how this modification may be involved in the action of psychotropic drugs.
Acknowledgement This study was supported by a grant from Telethon Italy (no. 539) to M.P.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
T. SOLLNER, M.K. BENNETT, S.W. WHITEHEART, R.H. SCHELLER and J.E. ROTHMAN, Cell 75 409-418 (1993). M.K. BENNETT and R.H. SCHELLER, Annu. Rev. Biochem. 63 63-100 (1994). M. GEPPERT, Y. GODA, R.E. HAMMER, C. LI, T.W. ROSAHL, CF. STEVENS and T.C. SUDHOF, Cell 79 717-727 (1994). M. POPOLI, C. VOCATURO, J. PEREZ, E. SMERALDI and G. RACAGNI, Mol. Pharmacol. 48, 623-629 (1995). M. POPOLI, A. VENEGONI, C. VOCATURO, L. BUFFA, J. PEREZ, E. SMERALDI and G. RACAGNI, Mol. Pharmacol. 51 l-8 (1997). M.S. PERIN, V.A. FRIED, G.A. MIGNERY, R. JAHN and T.C. SUDHOF, Nature 345 260-263 (1990). B.A. DAVLETOV and T.C. SUDHOF, J. Biol. Chem. 268 26386-26390 (1993). C.K. DAMER and C.E. CREUTZ, J. Biol. Chem. 269 31115-31123 (1994). G. SCHIAVO, Q.M. GU, G.D. PRESTWICH, T.H. SOLLNER and J.E. ROTHMAN, Proc. Natl. Acad. Sci. USA 93 13327-13332 (1996). M.K. BENNETT, N. CALAKOS and R.H. SCHELLER, Science 257 255-259 (1992). I. BEZPROZVANNY, R.H. SCHELLER and R.W. TSIEN, Nature 378 623-626 (1995). C. LI, B. ULLRICH, J.Z. YOUNG, R.G.W. ANDERSON, N. BROSE and T.C. SUDHOF, Nature 375 594-599 (1995). E.R. CHAPMAN, S. AN, J.M. EDWARDSON and R. JAHN, J. Biol. Chem. 271 5844-5849 (1996). S. SUGITA, Y. HATA and T.C. SUDHOF, J. Biol. Chem. 271 1262-1265 (1996). M. POPOLI and R. PATERNO’, NeuroReport 3 177-180 (1992). W.D. MATTHEW, L. TSAVALER and L.F. REICHARDT, J. Cell. Biol. 91 257-269 (1981). B. WENDLAND, K.G. MILLER, JSCHILLING and R.H. SCHELLER, Neuron 6 993-1007 (1991). M., POPOLI and A. MENGANO, Neurochem. Res. 13 63-67 (1988). M. POPOLI, Neuroscience 54 323-328 (1993). C. LI, B.A. DAVLETOV and T.C. SUDHOF, J. Biol. Chem. 270 24898-24902 (1995).
Vol. 61, No. 7, 1997
21. 22. 23. 24. 25. 26. 27. 28.
Native Syoaptotagmin Biidiog Activity
Y. GODA and C.F. STEVENS, Proc. Natl. Acad. Sci. USA 91 12942-12946 (1994). E.R. CHAPMAN, P.I. HANSON, S. AN and R. JAHN, J. Biol. Chem. 270 23667-23671 (1995). Y. KEE and R.H. SCHELLER, J. Neurosci. 16 1975-1981 (1996). Z.H. SHENG, J. REllIG, M. TAKAHASHI and W.A. CATTERALL, Neuron 13 1303-1313 (1994). M. FUKUDA, T. KOJIMA and K. MIKOSHIBA, J. Biol. Chem. 271 8430-8434 (1996). B. ULLRICH, C. LI, J.Z. ZHANG, H. McMAHON, R.G.W. ANDERSON, M. GEPPERT and T.C. SUDHOF, Neuron 13 1281-1291 (1994). M. POPOLI, FEBS Lett. 317 85-88 (1993). M.K., BENNETT, K.G. MILLER and R.H. SCHELLER, J. Neurosci. 13 17011707 (1993).
721