Phosphorylation of Synaptic Vesicle Protein 2 Modulates Binding to Synaptotagmin*

Synaptic vesicle protein 2 (SV2) is a component of all synaptic vesicles that is required for normal neurotransmission. Here we report that in intact synaptic terminals SV2 is a phosphoprotein. Phosphopeptide mapping studies indicate that a major site of phosphorylation is located on the cytoplasmic amino terminus. SV2 is phosphorylated on serine and threonine but not on tyrosine residues, indicating that it is a substrate for serine/threonine kinases. Phosphopeptide mapping, in gel kinase assays, and surveys of kinase inhibitors suggest that casein kinase I is a primary SV2 kinase. The amino terminus of SV2 was previously shown to mediate its interaction with synaptotagmin, a calcium-binding protein also required for normal neurotransmission. Comparison of synaptotagmin binding with phosphorylated and unphosphorylated SV2 amino-terminal peptides reveals an increase in binding with phosphorylation. These results suggest that the affinity of SV2 for synaptotagmin is modulated by phosphorylation of SV2. Neurotransmitter secretion occurs via a tightly regulated membrane trafficking cycle localized to the presynaptic terminal. Many stages of this cycle, such as the targeting and docking washed, resuspended in HBS (142 m M NaCl, 2.4 m M KCl, 1 m M MgCl 2 , 0.1 m M EGTA, 10 m M Hepes, pH 7.5, 5 m M D -glucose), and then hypotonically lysed by diluting 1:9 with water, rehomogenizing, and incubating for 30 min on ice. Heavy membranes Mapping— phosphorylated in reactions containing approximately 1 mg/ml crude synaptic vesicle protein in 20 m M Hepes, pH 7.5, 140 m M KOAc, 10 m M magnesium acetate, and 33 n M [ g - 32 P]ATP. Reactions were incubated for 15 min at 30 °C and then quenched with the addition of chilled extraction buffer with a phosphatase inhibitor mixture (1 m M EDTA, 11 m M EGTA, 50 m M NaF, 10 m M sodium phosphate, 20 m M b -glycerol phosphate, 0.15 m M sodium van-adate). Reactions were extracted at 4 °C fro m 2 h toovernight. Immu- noprecipitations were performed as described above with either a monoclonal antibody directed against all SV2 isoforms or a polyclonal antibody specific to SV2A. Phosphorylated glutathione S -transferase (GST) fusion peptides containing the amino terminus of SV2A or SV2B (GST-AN or GST-BN) were generated in 100- m l reactions containing 2 m g of recombinant peptide and 1.0 mg/ml crude synaptic vesicle protein in 20 m M Hepes, pH 7.5, 140 m M KOAc, 10 m M Mg(CH 3 CO 2 ) 2 , and 33 n M [ g - 32 P]ATP. At the end of a 15-min incubation at 30 °C, insoluble pro- tein was removed by centrifugation for 1 min at 19,000 3 g . Reactions were quenched by the addition of 800 m l of phosphate-buffered saline containing 0.5% Tween 20 and phosphatase inhibitor mixture. Phos-pho-GST-AN and GST-BN were re-isolated by incubation with gluta- thione-agarose beads precoated with 0.1% gelatin and 0.1% bovine five of the 27-kDa GST protein was observed, whereas the 45-kDa GST-AN was phosphorylated. These results indicate that phosphorylation of GST-SV2 peptides reflects phosphorylation of the SV2 and not the GST portion of the fusion protein. histidine- tagged peptide corresponding to the amino terminus of SV2A ( His-AN ) and GST-synaptotagmin ( GST-syt ). Phosphorylated and unphosphoryl- ated His-AN were generated by incubation in the presence or absence of 500 units of recombinant casein kinase I and repurified. Binding of 1 m g of His-AN to 1 m g of either GST or GST-syt was assessed. His-AN binding was detected by immunoblot ( top ). Ponceau-stained GST and GST-synaptotagmin are shown below to demonstrate equal amounts across reactions. B , quantitation of His-AN and phospho-His-AN bind- ing to synaptotagmin. Enhanced chemiluminescence signals from SV2 and synaptotagmin Western analyses were quantitated using a Kodak IS440 imaging system. The SV2/synaptotagmin ratios were compared across samples. Data are from a representative experiment done in duplicate. These experiments suggest that phosphorylated SV2 binds synaptotagmin with . 2-fold higher affinity than nonphosphorylated SV2.


Synaptic vesicle protein 2 (SV2) is a component of all synaptic vesicles that is required for normal neurotransmission.
Here we report that in intact synaptic terminals SV2 is a phosphoprotein. Phosphopeptide mapping studies indicate that a major site of phosphorylation is located on the cytoplasmic amino terminus. SV2 is phosphorylated on serine and threonine but not on tyrosine residues, indicating that it is a substrate for serine/threonine kinases. Phosphopeptide mapping, in gel kinase assays, and surveys of kinase inhibitors suggest that casein kinase I is a primary SV2 kinase. The amino terminus of SV2 was previously shown to mediate its interaction with synaptotagmin, a calcium-binding protein also required for normal neurotransmission. Comparison of synaptotagmin binding with phosphorylated and unphosphorylated SV2 amino-terminal peptides reveals an increase in binding with phosphorylation. These results suggest that the affinity of SV2 for synaptotagmin is modulated by phosphorylation of SV2.
Neurotransmitter secretion occurs via a tightly regulated membrane trafficking cycle localized to the presynaptic terminal. Many stages of this cycle, such as the targeting and docking of vesicles at active zone membranes, vesicle fusion, and endocytosis, are mediated by the formation of protein complexes (reviewed in Refs. 1 and 2). Protein phosphorylation, a ubiquitous mechanism of cellular regulation, plays an important role in the modulation of the synaptic vesicle cycle. Multiple examples of phosphorylation-mediated regulation of protein binding in the synapse have been reported. One well characterized example is the phosphorylation-dependent association of the actin-binding protein synapsin with synaptic vesicles, an interaction implicated in the maintenance of a reserve pool of vesicles (3,4). Likewise, proteins involved in the docking and fusion of vesicles, termed SNAREs, 1 are sub-strates of several protein kinases, and phosphorylation has been reported to alter their binding to other proteins and to each other (5)(6)(7)(8).
Synaptic vesicle protein 2 (SV2), a protein common to all neurons (9,10), has been implicated in the regulation of synaptic vesicle exocytosis (11,12). Three separately encoded isoforms, termed SV2A, SV2B, and SV2C, have been identified (13)(14)(15)(16)(17)(18). Loss of the most widely expressed isoform, SV2A, results in aberrant neurotransmission and death, indicating that SV2 is an essential protein (11,12). We previously reported that SV2A interacts with the synaptic vesicle protein synaptotagmin, a calcium-binding protein that is necessary for normal calcium-stimulated neurotransmission (19). The SV2synaptotagmin interaction is modulated by calcium, suggesting that it plays a role in the regulation of exocytosis.
The amino terminus of SV2, which mediates its interaction with synaptotagmin, contains substrate consensus sites of several protein kinases. In addition, SV2 was previously reported to be phosphorylated in vitro under conditions permissive to casein kinase I activity (20). In order to determine whether SV2 is phosphorylated in vivo and whether phosphorylation regulates its action, we examined SV2 phosphorylation in intact nerve terminals and in synaptic vesicles. We report that SV2 is a phosphoprotein in the synapse, that it is phosphorylated on its amino terminus, and that phosphorylation increases the binding of SV2 to synaptotagmin. These results suggest that the SV2-synaptotagmin interaction is regulated by changes in SV2 phosphorylation levels.

EXPERIMENTAL PROCEDURES
Materials-Radiolabeled orthophosphate and adenosine triphosphate were purchased from NEN Life Science Products, protein A-Sepharose was from Amersham Pharmacia Biotech, glutathione-agarose was from Sigma, and recombinant rat casein kinase I was from Calbiochem. Sequencing grade trypsin was purchased from Promega (Madison, WI), and Ni 2ϩ -NTA-agarose was from Qiagen. cAMP-dependent protein kinase was kindly provided by Dr. Brian Murphy (University of Washington, Seattle, WA).
Synaptosome and Synaptic Vesicle Isolation-Crude synaptosomes and synaptic vesicles were prepared as described by Huttner et al. (21). Briefly, fresh rat brains were homogenized in buffered sucrose (10 mM Hepes, pH 7.5, 0.3 M sucrose) with 10 strokes in a glass-Teflon homogenizer (0.004 -0.006-inch clearance). Homogenates were centrifuged at 1000 ϫ g to remove nuclei and intact cells. Crude synaptosomes were recovered from the remaining supernatant by centrifugation at 25,000 ϫ g for 13 min. The protein concentration of resuspended synaptosomes was determined with the Bio-Rad protein assay using bovine serum albumin as a standard. To obtain a crude synaptic vesicle preparation, synaptosomes were washed, resuspended in HBS (142 mM NaCl, 2.4 mM KCl, 1 mM MgCl 2 , 0.1 mM EGTA, 10 mM Hepes, pH 7.5, 5 mM D-glucose), and then hypotonically lysed by diluting 1:9 with water, rehomogenizing, and incubating for 30 min on ice. Heavy membranes were removed by centrifugation in a Beckman SW28 rotor at 15,000 rpm for 16 min. The resulting supernatant was centrifuged in the same rotor at 28,000 rpm for 4 h to isolate crude synaptic vesicles. Vesicles were resuspended in 1.1ϫ HK buffer (22 mM Hepes, pH 7.5, 154 mM KCH 3 CO 2 ).
Immunoprecipitation of Phospho-SV2 from Synaptosomes-Synaptosomes (2 mg of protein/ml) were resuspended in labeling buffer (140 mM NaCl, 5 mM KCl, 5 mM NaHCO 3 , 1 mM MgCl 2 , 10 mM glucose, 10 mM Hepes, pH 7.5) containing 1 mCi [␥-32 P]PO 4 /ml for 1 h at 37°C. A separate reaction lacking 32 PO 4 was incubated for subsequent immunoblotting. Labeled synaptosomes were isolated by centrifugation at 19,000 ϫ g for 5 min and resuspended in extraction buffer (HKA (10 mM Hepes, pH 7.5, 140 mM KCH 3 CO 2 , 1 mM MgCl 2 , 0.1 mM EGTA) with 2% Triton X-100, 50 mM NaF. This was extracted overnight at 4°C. Insoluble material was removed by centrifugation at 19,000 ϫ g for 15 min. Extracts were precleared of proteins that bind Protein A nonspecifically by incubation with Protein A-Sepharose beads in 1ϫ HKA, 0.1% bovine serum albumin, and 0.1% gelatin. Extracts were then incubated with anti-SV2 monoclonal antibody (9) for 1 h at 4°C. Protein A-Sepharose beads were added, and the mixture was incubated for an additional 1 h at 4°C. Beads were washed twice in 1ϫ HKA, 0.1% gelatin and once in 1ϫ HKA and then incubated with SDS-PAGE loading buffer for at least 10 min at 65°C to elute bound proteins. Proteins were separated by SDS-PAGE. Gels were either dried and exposed to film or transferred to nitrocellulose for immunoblotting.
Phosphopeptide Mapping-SV2 was phosphorylated in reactions containing approximately 1 mg/ml crude synaptic vesicle protein in 20 mM Hepes, pH 7.5, 140 mM KOAc, 10 mM magnesium acetate, and 33 nM [␥-32 P]ATP. Reactions were incubated for 15 min at 30°C and then quenched with the addition of chilled extraction buffer with a phosphatase inhibitor mixture (1 mM EDTA, 11 mM EGTA, 50 mM NaF, 10 mM sodium phosphate, 20 mM ␤-glycerol phosphate, 0.15 mM sodium vanadate). Reactions were extracted at 4°C from 2 h to overnight. Immunoprecipitations were performed as described above with either a monoclonal antibody directed against all SV2 isoforms or a polyclonal antibody specific to SV2A. Phosphorylated glutathione S-transferase (GST) fusion peptides containing the amino terminus of SV2A or SV2B (GST-AN or GST-BN) were generated in 100-l reactions containing 2 g of recombinant peptide and 1.0 mg/ml crude synaptic vesicle protein in 20 mM Hepes, pH 7.5, 140 mM KOAc, 10 mM Mg(CH 3 CO 2 ) 2 , and 33 nM [␥-32 P]ATP. At the end of a 15-min incubation at 30°C, insoluble protein was removed by centrifugation for 1 min at 19,000 ϫ g. Reactions were quenched by the addition of 800 l of phosphate-buffered saline containing 0.5% Tween 20 and phosphatase inhibitor mixture. Phospho-GST-AN and GST-BN were re-isolated by incubation with glutathione-agarose beads precoated with 0.1% gelatin and 0.1% bovine serum albumin. Beads were washed five times with phosphate-buffered saline containing 0.5% Tween 20. Peptides were eluted with SDS-PAGE sample buffer and separated by SDS-PAGE. Gels were stained with Coomassie Blue and dried. Phosphoproteins were located by autoradiography, and bands of interest were excised and rehydrated in 25 mM ammonium bicarbonate containing 1.2-2.0 g of trypsin at 37°C for 18 h, during which additional 25 mM ammonium bicarbonate was added to keep gel slices submerged. The eluate was transferred to a new tube and dried while the gel slices were incubated with additional 25 mM ammonium bicarbonate for 2 h at 37°C. All eluates were pooled and dried. Peptides were washed five times with 100 -200 l of water and suspended in pH 4.72 buffer (5% n-butanol, 2.5% pyridine, 2.5% glacial acetic acid) containing a trace amount of phenol red. Samples were spotted onto cellulose plates and run 10 cm at 400 V. Plates were dried overnight and subjected to chromatography at 90°to the electrophoretic direction in 6% glacial acetic acid, 34% n-butanol, 26% pyridine, 34% water. Plates were dried and exposed to film.
Comparison of GST and GST-SV2 Fusion Protein Phosphorylation-Two micrograms of GST or GST-AN fusion peptide were incubated for 30 min at 30°C in 50-l reactions containing 100 g of synaptic vesicle protein, 10mM Hepes, pH7.5, 140mM KCH 3 CO 2 , 0.1 mM EGTA, 20 mM magnesium acetate, and 20 Ci of [␥-32 P]ATP. Peptides were purified using glutathione-agarose and resolved by SDS-PAGE. Gels were dried and exposed to film.
Phosphoamino Acid Analysis-Phosphoamino acid analysis was carried out as described by Kamps and Sefton (22) on radiolabeled SV2 isolated by immunoprecipitation. Samples were resolved by thin layer chromatography in two dimensions. In the first dimension, pH 1.9 buffer was allowed to migrate 12 cm up the plate. The second dimension (run at 90°to the first) was resolved in pH 3.5 buffer (1% pyridine, 10% glacial acetic acid) for 6 cm. Standards were stained with 0.2% ninhydrin in acetone before the plates were subjected to autoradiography.
In Gel Kinase Assay-This assay was based on a protocol developed by Stratagene. 10% SDS-PAGE gels were poured with and without 0.375 mg/ml GST-AN added to the gel matrix. 0.5 g of crude synaptic vesicle protein was resolved in each gel lane. Gels were washed twice for 15 min each in buffer A (50 mM Hepes, pH 7.4, 5 mM 2-mercaptoethanol) with 20% isopropyl alcohol and equilibrated in buffer A for 1 h. Proteins were denatured by incubating gels in buffer A with 6 N guanidine-HCl for 2 h with two buffer changes. Proteins were renatured by two 8-h incubations in buffer A with 0.05% Tween 20. Gels were then equilibrated in kinase assay buffer (25 mM Hepes, pH 7.4, 10 mM MgCl 2 , 90 M sodium vanadate, 5 mM 2-mercaptoethanol) for 30 min. Proteins were phosphorylated in gel by incubating the gels in kinase assay buffer with 30 mM MgATP, 6 Ci/ml [␥-32 P]ATP for 1 h. Reactions were quenched by washing the gels five times for 20 min each in 5% (w/v) trichloroacetic acid, 1% (w/v) sodium pyrophosphate. Gels were dried and exposed to film.
Construction of His-tagged Fusion Protein-cDNAs encoding the amino-terminal region of SV2A (amino acids 1-183) were amplified by polymerase chain reaction and subcloned into the pTrcHisC expression vector (Invitrogen, San Diego, CA). Protein was produced in Escherichia coli strain BL26 and isolated by Ni 2ϩ -NTA-agarose chromatography.
CKI-7 Inhibition of SV2 Phosphorylation-75 g of crude synaptic vesicle protein was incubated for 15 min at 30°C in 50-l reactions containing 20 mM Hepes, pH 7.5, 140 mM potassium acetate, 10 mM magnesium chloride, 10 Ci of [␥-32 P]ATP, and 5 l of Me 2 SO with 0 -1 mM CKI-7. Proteins were then extracted by the addition of 18 volumes of 20 mM Hepes, pH 7.5, 140 mM potassium acetate, 50 mM sodium fluoride, and 2% Triton X-100 followed by incubation with gentle agitation for 2 h. Insoluble material was removed by centrifugation at 19,000 ϫ g for 15 min. SV2 was immunoprecipitated from the resulting supernatants. Immunoprecipitated protein was resolved by SDS-PAGE and subjected to autoradiography and Western analysis to measure incorporation of phosphate and the amount of SV2 precipitated, respectively.
Binding Studies-Approximately 20 g of His-AN fusion protein was phosphorylated in 450-l reactions containing 20 mM Hepes, pH 7.5, 10 mM MgCl 2 , 5 mM dithiothreitol, and 500 units of recombinant casein kinase I (Calbiochem). MgATP was added to 1 mM, and the incubation was carried out at 30°C for 1 h. A parallel reaction was carried out without kinase as a control. His-AN was purified from these reactions with Ni 2ϩ -NTA resin. 500 l of wash buffer (50 mM Na 2 HPO 4 , pH 8.0, 300 mM NaCl) and 500 l of 50% Ni 2ϩ -NTA resin were added to each reaction, after which they were incubated for 1 h at 4°C with agitation. Resin was washed twice with 500 l of wash buffer and once with 500 l of wash buffer with 20 mM imidazole. Peptide was eluted into 200 l of wash buffer containing 500 mM imidazole and quantified by comparison with bovine serum albumin standards in SDS-PAGE gels stained with Coomassie Blue. Equal amounts of phosphorylated or unphosphorylated His-AN were incubated with 0.5-1 g of GST or GST-synaptotagmin bound to 30 l of 50% glutathione-agarose beads in 1 ml of HEDTA buffer (50 mM Hepes, pH 7.5, 5 mM HEDTA, 95 mM KCH 3 CO 2 , 1% Triton) for 1 h at 4°C. Beads were washed three times with 1 ml of HEDTA buffer, and bound proteins were eluted with 10 l of 10 mM Tris, pH 8.0, 15 mM glutathione. Proteins were separated by SDS-PAGE. Gels were transferred to nitrocellulose and subjected to Western analysis.

SV2 Is a Phosphoprotein in Intact Nerve Terminals-
The amino acid sequences of all three SV2 proteins predict two major cytoplasmic domains, the amino terminus and the loop between the sixth and seventh transmembrane domains (13,14,16,17). These cytoplasmic domains contain the substrate consensus sequences (23,24) of several protein kinases reported to be present in the nerve terminal. These include casein kinase I (CKI), casein kinase II, protein kinase C, calcium/ calmodulin-dependent protein kinase II, and cAMP-dependent protein kinase (PKA). To determine whether SV2 is a phosphoprotein in situ, synaptosomes (intact synaptic terminals) were incubated with [ 32 P]orthophosphate in order to radiolabel synaptic nucleotide pools. After labeling, SV2 was extracted and isolated by immunoprecipitation using a monoclonal antibody that recognizes all known SV2 isoforms. Immunoprecipitated proteins were resolved by SDS-PAGE, and phosphoproteins were detected by autoradiography. A phosphoprotein of the appropriate size was isolated, suggesting that SV2 is a phosphoprotein in the synapse (Fig. 1A). An estimate of the stoichiometry of phosphorylation was precluded by the fact that SV2 does not label with standard protein staining techniques and therefore SV2 content in a sample cannot be measured. However, even though we were unable to estimate what the proportion of SV2 phosphorylated in nerve terminals, the observation that measurable levels of phospho-SV2 were obtained with a 30-min labeling incubation is consistent with a significant proportion of the protein cycling into a phosphorylated state.
To confirm that the immunoprecipitated protein was SV2, immunoprecipitates isolated from parallel, unlabeled incubations were subjected to Western analysis using a polyclonal antibody specific to SV2A, which is predicted to be the largest of the three SV2 isoforms. Anti-SV2A recognized a band that ran as a higher molecular weight subset of the radiolabeled material, consistent with the interpretation that multiple SV2 isoforms are phosphoproteins in the synapse. Radiolabeled phospho-SV2 was also generated in crude synaptic vesicle preparations incubated with [␥-32 P]ATP (Fig. 1B), indicating that an SV2 kinase is associated with synaptic vesicles. These studies demonstrate that SV2 is phosphorylated in intact nerve terminals and in synaptic vesicle preparations and suggest that SV2 is a phosphoprotein in vivo.
The Cytoplasmic Amino Terminus of SV2 Is Phosphorylated-SV2 contains two major cytoplasmic domains, the amino terminus, which is the most divergent region of the three isoforms, and a loop linking the sixth and seventh membrane domains. Because the amino terminus mediates the interaction of SV2A and SV2C with the synaptic vesicle protein synaptotagmin (19), 2 we were especially interested in determining whether it is phosphorylated. We therefore compared phosphopeptide maps of native SV2 with phosphopeptide maps of GST fusion proteins containing the cytoplasmic amino termi-nus of SV2A and SV2B (GST-AN and GST-BN, respectively) that had been phosphorylated with synaptic vesicle-associated kinases. Native SV2 was labeled in situ by incubation of synaptic vesicle preparations with [ 32 P]ATP and then isolated by immunoprecipitation using either a monoclonal antibody that recognizes all known SV2 isoforms or a polyclonal antibody specific to SV2A. Trypsin-digested peptide fragments were resolved by two-dimensional chromatography, and radiolabeled peptides were visualized by autoradiography (Fig. 2). Comparison of the peptide map total SV2 with the map of SV2A reveals that SV2A phosphopeptides make up a subset of those present in the total SV2 map. This is consistent with the interpretation that multiple isoforms of SV2 are phosphorylated. Phosphopeptide maps of GST-BN and GST-AN each contained a major phosphopeptide that migrated identically to phosphopeptides present in maps of native SV2. GST without attached SV2 was not phosphorylated (Fig. 3), indicating that the phosphopeptides observed on the maps reflect phosphorylation of the amino termini of SV2. These results indicate that the cytoplasmic amino terminus of SV2 is phosphorylated and suggest that synaptic vesicle kinase(s) phosphorylate the recombinant peptides at sites identical to those phosphorylated in the native protein. The presence of additional phosphopeptides in the maps of GST-AN and GST-BN maps may reflect additional sites of phosphorylation in the peptides. Alternatively, differential susceptibility to trypsin digestion could have produced differences in peptide migration patterns. The two phosphopeptides specific to the maps of total SV2 and SV2A (Fig. 3, small dark arrowheads) suggest sites of phosphorylation outside the amino terminus, perhaps in the cytoplasmic loop between the sixth and seventh membrane domains.
Casein Kinase I Is a Primary SV2 Kinase-Both serine/ threonine and tyrosine kinases have been identified in the nerve terminal (25)(26)(27). To identify the class of kinase that phosphorylates SV2, we performed phosphoamino acid analysis. Both acid (Fig. 4) and base (which favors the retention of phosphotyrosine (22)) hydrolysis methods were used. Both produced only phosphoserine and phosphothreonine residues, indicating that SV2 is phosphorylated by serine/threonine kinases.
We next sought to determine the molecular weight of the SV2 kinase(s) utilizing an in-gel kinase assay. With this procedure, proteins are resolved in gels containing substrate, after which they are incubated with radiolabeled ATP in reaction mixtures appropriate for kinase activity. Following the reaction, phosphorylated substrate co-localizes with kinases in the gel, and the location of active kinases can be detected by autoradiography. Synaptic vesicle proteins were resolved by SDS-PAGE in gels containing GST-AN or in control gels cast with no protein.
Following denaturation and renaturation procedures, the gels were incubated in reaction mixtures containing [␥-32 P]ATP, washed and exposed to film. In gels containing GST-AN, two distinct bands between 36 and 40 kDa were visualized by autoradiography (Fig. 5). The absence of bands in control gels indicated that the labeled bands correspond to phosphorylation of the GST-AN peptide and not to autophosphorylation. These results suggest that there are two synaptic vesicle-associated kinases that can phosphorylate the amino terminus of SV2A and that their catalytic subunits are approximately 36 and 40 kDa. Protein kinases known to be present in the nerve terminal with catalytic subunits of this size include CKI (20), PKA (28), and Ca 2ϩ /calmodulin-dependent protein kinase I (29,30). Of these, SV2 contains substrate consensus sequences for PKA and CKI, suggesting that one or both are the kinases responsible for SV2 phosphorylation in this assay, a possibility supported by the observation that purified CKI and PKA phosphorylate GST-AN peptide in vitro (not shown). Control immunoprecipitations done in the absence of antibody contain no phosphoprotein. An equivalent immunoprecipitate was immunoblotted for SV2A. The immunoblot reveals SV2A immunoreactivity to be a subset of the labeled material, suggesting that other SV2 isoforms are also phosphorylated. B, phosphorylation of SV2 in synaptic vesicles. Shown is an autoradiogram of total SV2 isolated from an extract of a crude synaptic vesicle preparation that had been incubated with [␥-32 P]ATP prior to extraction and immunoprecipitation.
To determine whether PKA and/or CKI are responsible for the phosphorylation of SV2 in situ, we examined the effects of kinase-specific activators and inhibitors on the phosphorylation of native SV2 in synaptosomes and synaptic vesicle preparations. To test the possibility that PKA is an SV2 kinase, we examined the effects of the PKA activator 8-bromo-cAMP on SV2 phosphorylation in both synaptosome and synaptic vesicle preparations. In neither case did 8-bromo-cAMP increase SV2 phosphorylation (data not shown). To determine whether CKI is an SV2 kinase, the effect of the CKI inhibitor CKI-7 (31) was tested. Incubation of crude synaptic vesicle preparations with increasing concentrations of CKI-7 led to a decrease in phosphorylation of native SV2 with an IC 50 of approximately 10 M (Fig. 6), a concentration reported to be specific for CKI (31). These results indicate that CKI is a primary SV2 kinase. This interpretation is supported by the presence of an identical phosphopeptide in the phosphopeptide maps of GST-AN phosphorylated by recombinant CKI and GST-AN phosphorylated by synaptic vesicle-associated kinases (Fig. 2). In addition, we have observed that the inositol lipid phosphatidylinositol 4,5bisphosphate (PIP 2 ) inhibits phosphorylation of GST-AN by crude synaptic vesicle preparations (not shown). This lipid was FIG. 2. Phosphopeptide maps of total SV2, SV2A, and recombinant peptides corresponding to the amino termini of SV2A and SV2B suggest that SV2 is phosphorylated on its cytoplasmic amino terminus. Shown are phosphopeptide maps of total native SV2 isolated with a monoclonal antibody that recognizes isoforms A-C (total SV2), native SV2A (SV2A), GST-SV2B amino terminus peptide phosphorylated by synaptic vesicle kinase(s) (BN), GST-SV2A amino terminus peptide phosphorylated by synaptic vesicle kinase(s) (AN), and GST-SV2A amino terminus peptide phosphorylated by recombinant casein kinase I (AN ϩ CKI). Phosphorylated peptides present on more than one map are denoted by matching arrowheads. Phosphopeptide maps of GST-BN and GST-AN each contained a major phosphopeptide that migrated identically to phosphopeptides present in maps of native SV2. The white arrowhead marks a peptide that migrated identically in maps of total native SV2, native SV2A, GST-AN, and GST-AN phosphorylated by recombinant casein kinase I. The wide black arrowhead marks a peptide that migrated identically in maps of total native SV2 and GST-BN. These results suggest that the cytoplasmic amino termini of these SV2 isoforms are sites of phosphorylation and that the amino terminus of SV2A is phosphorylated by casein kinase I.

FIG. 3. GST is not phosphorylated by synaptic vesicle protein kinases.
To determine whether the GST moiety of GST fusion proteins is a substrate for synaptic vesicle-associated kinases, the phosphorylation of GST and GST-SV2A amino terminus peptide (GST-AN) by synaptic vesicle-associated protein kinases was compared. Shown is an autoradiogram of 2 g of GST or GST-SV2A (GST-AN) incubated with 100 g of synaptic vesicle protein and [␥-32 P]ATP. Peptides were reisolated and resolved by SDS-PAGE, and phosphopeptides were visualized by autoradiography. No phosphorylation of the 27-kDa GST protein was observed, whereas the 45-kDa GST-AN was phosphorylated. These results indicate that phosphorylation of GST-SV2 peptides reflects phosphorylation of the SV2 and not the GST portion of the fusion protein.
FIG. 4. SV2 is phosphorylated on serine and threonine residues. SV2 immunoprecipitated from radiolabeled synaptosomes was subjected to phosphoamino acid analysis as described under "Experimental Procedures." Shown are samples generated by acid hydrolysis. Radiolabeled phosphoserine (S) and phosphothreonine (T), but not phosphotyrosine (Y), were obtained from phospho-SV2, suggesting that SV2 is phosphorylated by serine/threonine kinase(s).
reported by Gross et al. (20) to inhibit the activity of a synaptic vesicle-associated CKI. Finally, we have observed that SV2 phosphorylation is stimulated by excess magnesium (not shown), a characteristic of CKI (32). Together these results suggest that CKI acts as an SV2 kinase in vivo.
Phosphorylation of SV2A Stimulates Binding to Synaptotagmin-The amino terminus of SV2A mediates its interaction with synaptotagmin. The finding that this region of SV2 is phosphorylated suggested that the SV2-synaptotagmin interaction might be regulated by phosphorylation. To test this hypothesis, we analyzed the effects of phosphorylation on the binding of recombinant His 6 -SV2A amino terminus peptides (His-AN) to recombinant peptides corresponding to the cytoplasmic domain of synaptotagmin (GST-syt). For these experiments, His-AN peptide was incubated in an initial reaction in the absence or presence of recombinant CKI, after which peptide was reisolated and used in binding studies with GST-syt. Phosphorylation by CKI was confirmed in parallel reactions (data not shown). Unphosphorylated His-AN bound GST-syt, as reported previously (19). However, at identical concentrations, phospho-His-AN demonstrated significantly increased binding, suggesting that the affinity of the interaction is higher when SV2 is phosphorylated (Fig. 7A, top). Increased binding was not due to differences in the amount of GST-syt present in the binding incubation as shown by Ponceau staining of the immunoblot (Fig. 7A, bottom). Quantitation revealed a greater than 2-fold stimulation in binding (Fig. 7B). Consistent with these findings, we have observed that SV2 expressed in fibroblasts is not significantly phosphorylated and that it binds recombinant synaptotagmin much more weakly than SV2 isolated from brain. 3 These results indicate that phosphorylation of SV2 modulates its binding to synaptotagmin and suggest that the SV2-synaptotagmin interaction may be regulated in vivo through changes in SV2 phosphorylation. DISCUSSION The synaptic vesicle protein SV2 is a membrane glycoprotein common to all synaptic vesicles and is essential for normal 3 A. Schivell and S. Bajjalieh, unpublished observations. and GST-synaptotagmin (GST-syt). Phosphorylated and unphosphorylated His-AN were generated by incubation in the presence or absence of 500 units of recombinant casein kinase I and repurified. Binding of 1 g of His-AN to 1 g of either GST or GST-syt was assessed. His-AN binding was detected by immunoblot (top). Ponceau-stained GST and GST-synaptotagmin are shown below to demonstrate equal amounts across reactions. B, quantitation of His-AN and phospho-His-AN binding to synaptotagmin. Enhanced chemiluminescence signals from SV2 and synaptotagmin Western analyses were quantitated using a Kodak IS440 imaging system. The SV2/synaptotagmin ratios were compared across samples. Data are from a representative experiment done in duplicate. These experiments suggest that phosphorylated SV2 binds synaptotagmin with Ͼ2-fold higher affinity than nonphosphorylated SV2. neurotransmission. We have shown here that SV2 is a phosphoprotein in situ, that it is phosphorylated on its cytoplasmic amino terminus, and that phosphorylation modulates SV2 binding to synaptotagmin. These results suggest that phosphorylation regulates SV2 action in vivo.
Although SV2 possesses the substrate consensus sites of several protein kinases, results from phosphopeptide mapping, in gel kinase assays, and kinase inhibitor studies suggest that a primary SV2 kinase is casein kinase I. The casein kinase I family contains at least seven members (33). The ␣ isoform, which is present on synaptic vesicles, was previously suggested to be an SV2 kinase, because incubation of synaptic vesicles in reactions permissive to CKI activity resulted in the phosphorylation of SV2 (20). Our studies support this interpretation and suggest that CKI is the primary SV2 kinase.
The CKI substrate consensus site requires negatively charged amino acids 1-2 positions upstream. In some cases, this is provided by a phosphoamino acid (33). Therefore, for some substrates, CKI activity is indirectly regulated by other protein kinases. Although several of the potential CKI sites in SV2 require upstream phosphorylation, we were unable to influence the phosphorylation of SV2 by activators of PKA or protein kinase C or by potassium depolarization of radiolabeled synaptosomes (data not shown), suggesting that CKI phosphorylation of SV2 is regulated independently of other kinases.
The isoform of CKI present at the synapse is negatively regulated by the lipid PIP 2 (20), and PIP 2 inhibited SV2 phosphorylation in crude synaptic vesicle preparations (data not shown). It is interesting to note that generation of PIP 2 has been linked to the priming of regulated secretory vesicles (34). It is reasonable to speculate that one of the consequences of vesicle priming is decreased SV2 phosphorylation, which in turn would result in decreased affinity of SV2 for synaptotagmin. However, whether the half-life of higher PIP 2 concentrations is long enough to significantly affect SV2 phosphorylation levels is currently unknown.
The role of the SV2-synaptotagmin interaction, like the action of both proteins, remains unknown. SV2 binds the second calcium binding (C2B) domain of synaptotagmin, a region of the protein essential to its function (35). The C2B domain has been reported to interact with multiple synaptic molecules (38 -43). Interestingly, all of these interactions, including the interaction with SV2, map to the same region of the C2B domain (39,44). 2 It is therefore possible that SV2 regulates the availability of synaptotagmin for its other binding partners. Consistent with this hypothesis, a small peptide corresponding to a short region of the SV2 amino terminus was reported to increase synaptotagmin binding to the adaptor protein AP2 (45). Whether this effect is via direct interaction of SV2 with synaptotagmin or with AP2 is unknown, but it suggests a role for the SV2-synaptotagmin interaction in regulating synaptotagmin's interactions. Alternatively, the SV2-synaptotagmin interaction may regulate SV2's ability to act as a transporter, a function suggested b SV2's structural similarity to small molecule transporters. Precedent for this hypothesis is provided by the observation that the C2B domain of synaptotagmin can confer hemicolinium-3 sensitivity to endogenous choline transporters in oocytes (46). Finally, although we have demonstrated that phosphorylation affects SV2 binding to synaptotagmin, phosphorylation may also directly regulate SV2 transport activity or membrane trafficking. This interpretation is consistent with the observation that SV2B, which does not bind synaptotagmin, is also a phosphoprotein.
The cycle of synaptic vesicle exo-and endocytosis requires the carefully orchestrated formation and dissociation of protein complexes. Regulation of binding affinity through phosphoryl-ation is likely to mediate changes in secretory efficiency that contribute to synaptic plasticity. The results reported here suggest that SV2 function is regulated by phosphorylation and therefore may serve as a site of modulation in the synapse.