The Calcium-binding Loops of the Tandem C2 Domains of Synaptotagmin VII Cooperatively Mediate Calcium-dependent Oligomerization*

Synaptotagmin VII (Syt VII), a proposed regulator for Ca2+-dependent exocytosis, showed a robust Ca2+-dependent oligomerization property via its two C2 domains (Fukuda, M., and Mikoshiba, K. (2001)J. Biol. Chem. 276, 27670–27676), but little is known about its structure or the critical residues directly involved in the oligomerization interface. In this study, site-directed mutagenesis and chimeric analysis between Syt I and Syt VII showed that three Asp residues in Ca2+-binding loop 1 or 3 (Asp-172, Asp-303, and Asp-357) are crucial to robust Ca2+-dependent oligomerization. Unlike Syt I, however, the polybasic sequence in the β4 strands of the C2 structures (so-called “C2 effector domain”) is not involved in the Ca2+-dependent oligomerization of Syt VII. The results also showed that the Ca2+-binding loops of the two C2 domains cooperatively mediate Syt VII oligomerization (i.e. the presence of redundant Ca2+-binding site(s)) as well as the importance of Ca2+-dependent oligomerization of Syt VII in Ca2+-regulated secretion. Expression of wild-type tandem C2 domains of Syt VII in PC12 cells inhibited Ca2+-dependent neuropeptide Y release, whereas mutant fragments lacking Ca2+-dependent oligomerization activity had no effect. Finally, rotary-shadowing electron microscopy showed that the Ca2+-dependent oligomer of Syt VII is “a large linear structure,” not an irregular aggregate. By contrast, in the absence of Ca2+ Syt VII molecules were observed to form a globular structure. Based on these results, we suggest that the linear Ca2+-dependent oligomer may be aligned at the fusion site between vesicles and plasma membrane and modulate Ca2+-regulated exocytosis by opening or dilating fusion pores.

Neuronal exocytosis is an extremely rapid process precisely controlled by Ca 2ϩ ions. Ca 2ϩ -binding proteins (so-called "Ca 2ϩ sensors") must be present in the synaptic vesicle to achieve fusion of synaptic vesicles with the presynaptic plasma membrane in response to the rapid increase in Ca 2ϩ ions (reviewed in Ref. 1). Genetic, structural, and biochemical studies during the past decade have indicated that synaptotagmin I (Syt I), 1 an abundant synaptic vesicle protein, is the most likely candidate for the major Ca 2ϩ sensor for neurotransmitter release in the central nervous system (Ref. 2; for reviews, see Refs. [3][4][5][6]. Syt I contains one transmembrane domain and two C2 Ca 2ϩbinding modules (the membrane-proximal C2A domain and membrane distal C2B domain) in the large cytoplasmic domain. The two C2 domains exhibit completely different biochemical properties in terms of phospholipid binding (3,7) and play distinct roles in synaptic vesicle trafficking (8 -10). Ca 2ϩ binding to the C2A domain stimulates interaction with negatively charged phospholipids (2,(11)(12)(13), whereas Ca 2ϩ binding to the C2B domain promotes clustering of the C2B domain in vitro (14 -20). Whereas these Ca 2ϩ -dependent actions of the C2 domains are now believed to be an essential component of synaptic vesicle exocytosis, how they drive membrane fusion in response to Ca 2ϩ remains unknown. The manner in which C2B oligomer is assembled (i.e. structure of Ca 2ϩ -dependent oligomer) on the synaptic vesicle is also unknown.
Syt has been found to comprise a large family of molecules both in vertebrates and in invertebrates (e.g. Drosophilia, Caenorhabditis elegans, and Arabidopsis thaliana) (Ref. 21 and references therein) and to share the same domain structure: an N-terminal single transmembrane domain, a spacer domain of varying length, two C2 domains, and a short carboxyl terminus (3)(4)(5)(6). However, the Ca 2ϩ -binding ability of the C2A domain differs among murine Syt isoforms, and some of them fail to bind Ca 2ϩ ions as a result of amino acid substitutions of key Asp/Glu residues responsible for Ca 2ϩ binding (13,(22)(23)(24). 2 The Ca 2ϩ -dependent oligomerization activity of the C2B domain and inositol polyphosphate binding activity also differs among the Syt isoforms (18,19,25,26), and these diversities among the C2 domains of the Syt isoforms suggest that the members of the Syt family probably have distinct functions in membrane trafficking. However, with the exception of Syt I, the exact roles of the Syt isoforms (e.g. Syt III and Syt VII) in membrane trafficking are still a matter of controversy (27)(28)(29)(30)(31).
Our previous study showed that Syt VII has the strongest Ca 2ϩ -dependent self-oligomerization capacity in the Syt family (18,19) and that both the C2A and C2B domains of Syt VII function as a Ca 2ϩ -dependent oligomerization site (i.e. Syt VII has "two hands") (32). Although Syt VII has recently been proposed to regulate several Ca 2ϩ -dependent processes (i.e. lysosomal exocytosis, insulin secretion in pancreatic ␤-cells, and norepinephrine release in PC12 cells) (27,28,30,33,34), little is known about the functional involvement of the Ca 2ϩdependent self-oligomerization of Syt VII in these Ca 2ϩ -regulated events, the structure of the Ca 2ϩ -dependent oligomer, or the critical residues directly involved in the oligomerization interface. In this study, we attempt to identify the residues (or an oligomerization interface) critical to the Ca 2ϩ -dependent oligomerization of Syt VII and the functional relationship between the two C2 domains. We show by site-directed mutagenesis that, unlike Syt I, Ca 2ϩ -dependent homo-and heterooligomerization of Syt VII are cooperatively mediated by the Ca 2ϩ -binding loops of the two C2 domains, not by the putative C2 effector domains. Moreover, whereas expression of the wildtype cytoplasmic tandem C2 domains in PC12 cells inhibited Ca 2ϩ -dependent neuropeptide Y (NPY) release, mutant pro-teins incapable of Ca 2ϩ -dependent oligomerization had no effect on NPY release. Furthermore, we show by rotary-shadowing electron microscopy that the Ca 2ϩ -dependent Syt VII oligomer has a linear structure and is not a random aggregate. Based on these findings, we discuss how the Ca 2ϩ -dependent Syt oligomer regulates vesicular exocytosis.
NPY Release Assay-NPY cDNA was a kind gift of Dr. Wolfhard Almers (38). The addition of the C-terminal T7-GST tag (GGSGGTGG-MARMTGGQQMG ϩ GST; Gly linker underlined) to NPY was essentially performed by PCR as described above (39). The resulting NPY-T7-GST fragments were subcloned into the NotI site of the modified pShooter vector (Invitrogen) (named pShooter-NPY-T7-GST) as described previously (40). PC12 cells (6-cm dish) were cotransfected with 4 g each of pShooter-NPY-T7-GST and pEF-FLAG-Syt VII-cyto by using LipofectAMINE 2000 reagent (Invitrogen) according to the manufacturer's notes. Three days after transfection, cells were washed with prewarmed low KCl buffer (5.6 mM KCl, 145 mM NaCl, 2.2 mM CaCl 2 , 0.5 mM MgCl 2 , 5.6 mM glucose, and 15 mM HEPES-KOH, pH 7.4) and FIG. 1. Ca 2؉ selectively promotes synaptotagmin VII clustering. A, effect of divalent cations on Ca 2ϩ -dependent oligomerization of Syt VII. B, Ca 2ϩ -dependent oligomerization of Syt VII is sensitive to ionic strength. pEF-T7-Syt VII-cyto and pEF-FLAG-Syt VII-cyto were cotransfected into COS-7 cells. The proteins expressed were solubilized with 1% Triton X-100 and immunoprecipitated with anti-T7 tag antibody-conjugated agarose (IP) in the presence of the divalent cations indicated (1 mM in A) or in the presence of 1 mM Ca 2ϩ and the concentrations of NaCl indicated (in B) (32,35). Co-immunoprecipitated FLAG-Syts were first detected with HRP-conjugated anti-FLAG tag antibody (1:10,000 dilution; third panels). The same blots were stripped and reprobed with HRP-conjugated anti-T7 tag antibody to ensure loading of the same amounts of T7-Syt proteins (1:10,000 dilution; bottom panels). The top two panels indicate the total expressed T7-Syt VII-cyto and FLAG-Syt VII-cyto proteins ( 1 ⁄80 volume; input) used for immunoprecipitation. Note that Ca 2ϩ selectively promoted oligomerization of Syt VII (in A) and that it was highly sensitive to ionic strength (NaCl concentration). The positions of the molecular weight markers (ϫ 10 Ϫ3 ) are shown on the left. then stimulated with either low KCl buffer or high KCl buffer (56 mM KCl, 95 mM NaCl, 2.2 mM CaCl 2 , 0.5 mM MgCl 2 , 5.6 mM glucose, and 15 mM HEPES-KOH, pH 7.4) for 10 min at 37°C. Released NPY-T7-GST was recovered by incubation with glutathione-Sepharose beads and analyzed by immunoblotting with horseradish peroxidase (HRP)-conjugated anti-T7 tag antibody. The intensity of the immunoreactive bands on x-ray film was quantified as described previously (18) and normalized by total expressed NPY-T7-GST. Total cell lysates were obtained by incubation with a lysis buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, and 1% Nonidet P-40, and protease inhibitors). Under our experimental conditions, NPY-T7-GST was targeted to dense core vesicles (data not shown), and about 5% of total NPY-T7-GST was released only in a high KCl-dependent manner.
Electron Microscopy-The specimens for electron microscopy were prepared by rotary shadowing as described previously (41). In brief, an aliquot of the purified Syt VII cytoplasmic fragment (about 50 -100 g/ml) was mixed with four volumes of 25 mM ammonium acetate containing 50% glycerol. The mixture was immediately sprayed onto the surface of freshly cleaved mica. Following rotary shadowing with Pt/C (elevation angle: 6 or 8°) and backing with pure carbon, replicas were floated off and were picked up onto copper grids. The images were  (43)(44)(45). The overall structure of the two C2 domains is quite similar, but an additional ␣-helix is present between the ␤7 and ␤8 strands in the C2B domain (45). Five Asp residues in the C2A domain (or in the C2B domain; parentheses) are thought to be crucial to Ca 2ϩ binding, by analogy with the C2A domain of Syt I (44). KKK indicates the putative C2 effector domain in the ␤4 strands (16,25). C, Ca 2ϩ -dependent self-oligomerization of Syt VII with the mutations in the C2 effector domains (AKQ and BKQ). pEF-T7-Syt VII-cyto and pEF-FLAG-Syt VII-cyto were cotransfected into COS-7 cells. The proteins expressed were solubilized with 1% Triton X-100 and immunoprecipitated with anti-T7 tag antibody-conjugated agarose (IP) in the presence or absence of 1 mM Ca 2ϩ as described previously (32,35). Co-immunoprecipitated FLAG-Syts were first detected with HRP-conjugated anti-FLAG tag antibody (1:10,000 dilution; third panel in C). The same blot was stripped and reprobed with HRP-conjugated anti-T7 tag antibody to ensure loading of the same amounts of T7-Syt proteins (1:10,000 dilution; bottom panel in C). The top two panels indicate the total expressed T7-Syt VII-cyto and FLAG-Syt VII-cyto proteins ( 1 ⁄80 volume; input in C) used for immunoprecipitation. Note that the mutations in the putative C2 effector domains (AKQ/BKQ) had virtually no effect.  (D172N, D303N, or D357N). B, Ca 2ϩ -dependent self-oligomerization of the tandem C2 domains of Syt VII with the mutations in the Ca 2ϩbinding loops (D172N and/or D303N). C, Ca 2ϩ -dependent hetero-oligomerization of T7-Syt VII-C2A mutant with FLAG-Syt VI-cyto. D, Ca 2ϩ -dependent hetero-oligomerization of T7-Syt VII-cyto mutants with FLAG-Syt VI-cyto. pEF-T7-Syts and pEF-FLAG-Syts were cotransfected into COS-7 cells. The proteins expressed were solubilized with 1% Triton X-100 and immunoprecipitated with anti-T7 tag antibody-conjugated agarose (IP) in the presence or absence of 1 mM Ca 2ϩ as described previously (32,35). Co-immunoprecipitated FLAG-Syts were first detected with HRP-conjugated anti-FLAG tag antibody (1:10,000 dilution; third panels in A-D). The same blots were stripped and reprobed with HRP-conjugated anti-T7 tag antibody to ensure loading of the same amounts of T7-Syt proteins (1:10,000 dilution; bottom panels in A-D). The top two panels indicate the total expressed T7-Syt and FLAG-Syt proteins ( 1 ⁄80 volume; input in A-D) used for immunoprecipitation. Note that the single mutation in the Ca 2ϩ -binding loops (D172N,  D303N, or D357N) is sufficient to abolish oligomerization of the single C2 domain but not self-oligomerization of the tandem C2 domains, whereas the double mutations in the Ca 2ϩ -binding loops (D172N/ D303N) completely abolished Ca 2ϩ -dependent homo-and heterooligomerization.

Oligomerization of Synaptotagmin VII Is Selectively Stimulated by Ca 2ϩ and Mediated by Electrostatic Interaction-If
oligomerization of Syt VII via the two C2 domains is crucial to the Ca 2ϩ -regulated events previously described (27,28,30,33,34), oligomerization of Syt VII must be selectively stimulated by Ca 2ϩ ions and not by other divalent cations. To confirm this, we first investigated the effect of divalent cations (1 mM) on self-oligomerization of Syt VII. The T7-and FLAG-tagged Syt VII cytoplasmic domains (Syt VII-cyto) were coexpressed in COS-7 cells, and the association between these two proteins was evaluated by the immunoprecipitation method (35,42) in the presence of various divalent cations (1 mM). As expected, self-oligomerization of Syt VII was selectively promoted by Ca 2ϩ ions, whereas Sr 2ϩ and Ba 2ϩ ions only marginally activated self-oligomerization. 1 mM Mg 2ϩ ions had no effect at all (Fig. 1A, third panel). This divalent cation selectivity was quite similar to that of Syt I described previously (15,16). Since the oligomerization of Syt VII was highly sensitive to ionic strength (failing to occur above 750 mM of NaCl), the Syt VII Ca 2ϩ -de-pendent oligomerization is most likely to be mediated by electrostatic interaction rather than hydrophobic interaction (Fig.  1B, third panel).
Mutational Analysis of the Ca 2ϩ -binding Loops and the Putative C2 Effector Domains of Synaptotagmin VII-Site-directed mutagenesis was performed to define the oligomerization interface of the Syt VII C2 domains (Fig. 2, A and B). We initially focused on the Lys cluster in the putative C2 effector domains, which is located in the ␤4 strand (Lys-183, Lys-184, and Lys-186 in the C2A domain and Lys-320, Lys-321, and Lys-325 in the C2B domain, corresponding to the "C2B effector domain" of Syt I) of the C2 ␤-sandwich structure (16,25,(43)(44)(45)(46), because neutralization of these basic residues (Lys-to-Gln or Lys-to-Ala substitution) completely abrogated the Ca 2ϩ -dependent self-oligomerization of Syts I and II (16,20). However, neutralization of all three basic residues in the C2A domain (named the AKQ mutation) and C2B domain (BKQ mutation) had no effect on the Ca 2ϩ -dependent self-oligomerization activity (Fig. 2C, third panel), although the AKQ mutation slightly increased the Ca 2ϩ -independent self-oligomerization activity.
We next focused on the Asp residues in putative Ca 2ϩ -binding loop 1 (Asp-172 in the C2A domain and Asp-303 in the C2B domain) and loop 3 (Asp-357 in the C2B domain) (43)(44)(45). The single Ca 2ϩ -binding loop mutations (D172N, D303N, or  D357N) of the isolated C2 domain were sufficient to abolish the Ca 2ϩ -dependent self-oligomerization activity (Fig. 3A, third  panel). To our surprise, however, the single Ca 2ϩ -binding loop mutations (D172N or D303N) of the full Syt VII cytoplasmic domain resulted in normal Ca 2ϩ -dependent self-oligomerization activity, whereas the double Ca 2ϩ -binding loop mutations (D172N and D303N) completely abrogated the Ca 2ϩ -dependent self-oligomerization activity (Fig. 3B, third panel). Similar results were obtained in regard to Ca 2ϩ -dependent hetero-oligomerization of Syt VII with Syt VI (compare Fig. 3, C and D,  third panels); the single D172N mutation was neutral in regard FIG. 4. Ca 2؉ -dependent oligomerization property of the C2A chimera between Syt VII and Syt I. A, schematic representation of the C2A chimera between Syt I and Syt VII (named Syt VII(S1A)). The transmembrane domain (TM), Syt I C2A domain, and Syt VII C2B domain are represented by the open box, hatched box, and shaded box, respectively. B, Ca 2ϩ -dependent self-oligomerization of Syt VII(S1A) mutants. pEF-T7-Syt VII(S1A)-cyto and pEF-FLAG-Syt VII(S1A)-cyto were cotransfected into COS-7 cells. The proteins expressed were solubilized with 1% Triton X-100 and immunoprecipitated with anti-T7 tag antibody-conjugated agarose (IP) in the presence or absence of 1 mM Ca 2ϩ as described previously (32,35). Co-immunoprecipitated FLAG-Syts were first detected with HRP-conjugated anti-FLAG tag antibody (1:10,000 dilution; third panel in B). The same blot was stripped and reprobed with HRP-conjugated anti-T7 tag antibody to ensure loading of the same amounts of T7-Syt proteins (1:10,000 dilution; bottom panel in B). The top two panels indicate the total expressed T7-Syt VII(S1A)cyto and FLAG-Syt VII(S1A)-cyto proteins ( 1 ⁄80 volume; input in B) used for immunoprecipitation. to the interaction between Syt VII-cyto and Syt VI-cyto ( Fig.  3D) but completely abrogated the interaction between Syt VII-C2A and Syt VI-cyto (Fig. 3C). In addition, the double Ca 2ϩbinding loop mutant (D172N and D303N) was incapable of interacting with Syt VI-cyto even in the presence of Ca 2ϩ (Fig.  3D). This result markedly contrasts with those of Syt I, because neutralization of the corresponding acidic (Asp) residues of Syt I in the Ca 2ϩ -binding loops enhanced the Ca 2ϩ -independent self-oligomerization activity (47). These findings strongly indicated that the fundamental mechanism of Ca 2ϩ -dependent self-oligomerization by Syt VII and Syt I is different, at least in terms of oligomerization interface, and we hypothesized that the two C2 domains of Syt VII cooperatively mediate Ca 2ϩ -dependent self-oligomerization (i.e. existence of a redundant Ca 2ϩ -binding site).
Ca 2ϩ -binding Loops of Two C2 Domains Cooperatively Mediate Ca 2ϩ -dependent Oligomerization of Synaptotagmin VII-If this hypothesis were true, pairing of the C2A domain and C2B domain should be critical for Ca 2ϩ -dependent selfoligomerization of Syt VII, and to test this, we prepared a chimera between Syt VII and Syt I that contains the Syt I C2A domain (named Syt VII(S1A); see Fig. 4A). The Syt VII(S1A)cyto protein showed weaker Ca 2ϩ -dependent self-oligomerization activity than that of the wild-type Syt VII protein (Fig. 4B,  third panel), probably as a result of the loss of one hand (i.e. the C2A domain of Syt VII), because the Syt I C2A domain did not show Ca 2ϩ -dependent self-oligomerization activity (15,16). It is noteworthy that the Syt VII(S1A)(D303N)-cyto protein showed Ca 2ϩ -independent self-oligomerization activity, the same as the Syt I Ca 2ϩ -binding mutation (47) (Fig. 4B, third   panel). Thus, pairing of the C2A and C2B domains is crucial to efficient Ca 2ϩ -dependent self-oligomerization of Syt VII.
Effect of Mutations in the Ca 2ϩ -binding Loops on Ca 2ϩ -dependent NPY Release in PC12 Cells-We then investigated the involvement of Ca 2ϩ -dependent oligomerization of Syt VII in Ca 2ϩ -dependent NPY-T7-GST release by expression of wildtype or mutant (D172N/D303N) cytoplasmic fragments of Syt VII in PC12 cells. Expression of the wild-type Syt VII-cyto in PC12 cells resulted in about 50% inhibition of Ca 2ϩ -induced exocytosis, whereas expression of Syt VII-cyto(D172N/D303N) completely reversed the inhibitory effect (Fig. 5). Ca 2ϩ -independent NPY-T7-GST release was unaltered by the expression of recombinant proteins (data not shown). These results suggested a critical function of Ca 2ϩ -dependent oligomerization of the C2 domains in Ca 2ϩ -induced exocytosis.
Structure of Ca 2ϩ -dependent Oligomerization of Synaptotagmin VII Visualized by Rotary-shadowing Electron Microscopy-Finally, we attempted to visualize the structure of the Ca 2ϩ -dependent oligomer of Syt VII by rotary-shadowing electron microscopy. Since the bacterial recombinant Syt VII-cyto proteins were difficult to prepare due to inclusion bodies (data not shown), we used recombinant proteins from mammalian cultured cells for electron microscopy. The recombinant cytoplasmic domains of Syt VII fused to GST (Fig. 6A) were expressed in COS-7 cells and were affinity-purified as described under "Experimental Procedures." The purity of the recombinant Syt VII on the SDS-polyacrylamide gel was always greater than 95% (Fig. 6B).
In the absence of Ca 2ϩ , we observed only "globular structures," probably corresponding to the monomeric form of Syt VII molecules (arrowheads in the middle panel in Fig. 7). In the presence of Ca 2ϩ , however, "linear structures" of various length were observed in addition to the globular structures (Fig. 7, left  panel and insets), indicating that Syt VII forms large Ca 2ϩ -dependent oligomers (or sometimes polymers), consistent with the results of our previous gel filtration analysis (32). These linear structures did not have branches and probably corresponded to the assembly of the globular structures observed in the absence of Ca 2ϩ (Fig. 7, left panel). By contrast, no such linear structures were observed in the Syt VII(D172N/D303N) mutant specimens even in the presence of Ca 2ϩ (globular structure, arrowheads in the right panel of Fig. 7), consistent with the results of immunoprecipitation described above (Fig. 3B). These results together with the results of immunoprecipitation (Figs. 2-4 and Ref. 18) and gel filtration analyses (32) indicated that the Ca 2ϩ -induced Syt VII oligomer is a large linear structure and not a random aggregate. DISCUSSION In our previous study, we showed that a single C2 domain of Syt VII is sufficient for Ca 2ϩ -dependent homomultimerization and hetero-oligomerization with other Syt isoforms (18,19,32), but the functional relationship between the two C2 domains and the structure of the Syt VII oligomer had never been determined. In this paper, we have presented several lines of evidence indicating that the two C2 domains of Syt VII are not functionally independent and that the Ca 2ϩ -binding loops of the two C2 domains cooperatively mediate Ca 2ϩ -dependent oligomerization. First, mutation of the single Asp residue (D172N in the C2A domain or D303N in the C2B domain) in Ca 2ϩ -binding loop 1 abrogated the Ca 2ϩ -dependent self-oligomerization of the "isolated C2 domain" but was neutral in regard to the Ca 2ϩ -dependent self-oligomerization of the "tandem C2 domains" (Figs. 2 and 3). The tandem C2 domains with the double mutation (D172N/D303N), however, did not show Ca 2ϩ -dependent self-oligomerization, suggesting the presence of redundant Ca 2ϩ -binding site(s). Second, chimeric analysis between Syt I and Syt VII showed that pairing of the C2A and C2B domains is an important factor for efficient Ca 2ϩ -dependent oligomerization of Syt VII (Fig. 4). Based on these results, together with the recent crystallographic data showing that the Ca 2ϩ -binding regions of the two C2 domains of Syt III face each other (45), we propose that the Ca 2ϩ -binding loops of the two C2 domains directly contribute to formation of the oligomerization interface; Ca 2ϩ binding to the Asp residues in the loops of the C2 structures changes the electrostatic charges around the loop domains, which may directly form the oligomerization interface between two C2 domains. This model is completely different from the model of Syt I (or II), which also showed Ca 2ϩ -dependent oligomerization mediated by the C2B effector domain but not the Ca 2ϩ -binding loops themselves (16,20,47). The structure of the Ca 2ϩ -dependent oligomer of Syt I is now under investigation in our laboratory, and it will be interesting to determine whether the Ca 2ϩ -dependent Syt I oligomer exhibits the same linear structure.
We also demonstrated the physiological importance of the Ca 2ϩ -dependent oligomerization of Syt VII by using the dominant negative approach (28,30,47). When the wild-type tandem C2 fragment was expressed in PC12 cells, Ca 2ϩ -dependent NPY release was significantly inhibited, most likely by competing endogenous Syt proteins. By contrast, the mutants lacking Ca 2ϩ -dependent oligomerization activity had no effect on NPY release (Fig. 5).
Although the exact role of the Ca 2ϩ -dependent oligomer of Syt VII in vesicular exocytosis is unknown, based on the structure of the Ca 2ϩ -dependent oligomer of Syt VII (i.e. the linear structure) visualized by rotary-shadowing electron microscopy (Fig. 7), it may be involved in dilation or opening of fusion pores by aligning to form a straight line at the fusion site between the vesicles and plasma membrane (48). Consistent with this hypothesis, Syt I and IV have recently been shown to modulate fusion pore kinetics in regulated exocytosis of PC12 cells (49). Further work is necessary to examine the effect of wild-type or mutant (D172N/D303N) expression on fusion pore kinetics to clarify the relationship between Syt VII oligomerization and fusion pore formation.
In summary, site-directed mutagenesis and chimeric analysis in this study demonstrated that the Ca 2ϩ -binding loops of the two C2 domains cooperatively mediate both Ca 2ϩ -dependent oligomerization of Syt VII and Ca 2ϩ -dependent NPY release. We have also shown that the Ca 2ϩ -dependent Syt VII oligomer is a linear structure, not an irregular random aggregate.