Synaptotagmin I is a synaptic vesicle protein that contains two C 2 domains and acts as a Ca 2+ sensor in neurotransmitter release. The Ca 2+-binding properties of the synaptotagmin I C 2A domain have been well characterized, but those of the C 2B domain are unclear. The C 2B domain was previously found to pull down synaptotagmin I from brain homogenates in a Ca 2+-dependent manner, leading to an attractive model whereby Ca 2+-dependent multimerization of synaptotagmin I via the C 2B domain participates in fusion pore formation. However, contradictory results have been described in studies of Ca 2+-dependent C2B domain dimerization, as well as in analyses of other C 2B domain interactions. To shed light on these issues, the C2B domain has now been studied using biophysical techniques. The recombinant C 2B domain expressed as a GST fusion protein and isolated by affinity chromatography contains tightly bound bacterial contaminants despite being electrophoretically pure. The contaminants bind to a polybasic sequence that has been previously implicated in several C 2B domain interactions, including Ca 2+-dependent dimerization. NMR experiments show that the pure recombinant C 2B domain binds Ca 2+ directly but does not dimerize upon Ca2+ binding. In contrast, a cytoplasmic fragment of native synaptotagmin I from brain homogenates, which includes the C2A and C2B domains, participates in a high molecular weight complex as a function of Ca2+. These results show that the recombinant C 2B domain of synaptotagmin I is a monomeric, autonomously folded Ca 2+-binding module and suggest that a potential function of synaptotagmin I multimerization in fusion pore formation does not involve a direct interaction between C 2B domains or requires a posttranslational modification. Synaptotagmins constitute a large family of neuronal membrane proteins ( 1, 2). Synaptotagmin I, the first member identified in this family, is a synaptic vesicle protein that is essential for fast Ca 2+-triggered neurotransmitter release ( 3), most likely acting as a Ca 2+ sensor in this process ( 4, 5). Synaptotagmin I contains an N-terminal intravesicular sequence, a transmembrane region, a linker sequence, and two C-terminal C2 domains referred to as the C 2A and C2B domains (Figure 1A). The synaptotagmin I C 2A domain has a â-sandwich structure and binds multiple Ca 2+ ions via a motif that includes five conserved aspartate residues at the tip of theâ-sandwich ( 6-8). The C2B domain from the Rab3 effector protein rabphilin, which shares substantial sequence homology with the synaptotagmin I C 2B domain (Figure 1B), also forms aâ-sandwich structure (see Figure 1C) and binds multiple Ca2+ ions via a similar motif ( 9). On the basis of this homology and its Ca 2+-dependent activities (see below), it has been hypothesized that the synaptotagmin I C 2B domain also binds Ca 2+ and contributes to the Ca 2+-sensing function of synaptotagmin I, but this hypothesis has not been directly tested. In addition, while both C 2 domains of synaptotagmin I have been implicated in several interactions that could mediate its function in neurotransmitter release (reviewed in refs 1 and 2), the interactions involving the C 2B domain have been controversial. Particular attention has attracted the involvement of the C2B domain in Ca2+-dependent synaptotagmin multimerization. Since synaptotagmin I forms Ca 2+-independent constitutive dimers via its transmembrane region ( 10-13), the observation that the recombinant C 2B domain immobilized as a GST1 fusion captured native synaptotagmin I from brain extracts in a Ca 2+-dependent manner ( 14, 15) led to an attractive model whereby Ca 2+-dependent self-association could result in a ring of synaptotagmins during fusion pore formation. Moreover, the observation that different synap† This work was supported by a grant from the Welch Foundation and by NIH Grant NS40944. * To whom correspondence should be addressed. Tel: 214-648-9026. Fax: 214-648-8673. E-mail: jose@arnie.swmed.edu. ‡ Department of Biochemistry. § Department of Pharmacology. | Department of Molecular Genetics. ⊥ Center for Basic Neuroscience. # Howard Hughes Medical Institute. 1 Abbreviations: ES-MS, electrospray -mass spectrometry; GST, glutathioneS-transferase; HSQC, heteronuclear single-quantum correlation; SNAP, soluble NSF attachment protein; SNARE, SNAP receptor; SV2, synaptic vesicle protein 2. 5854 Biochemistry2001,40, 5854-5860 10.1021/bi010340c CCC: $20.00 © 2001 American Chemical Society Published on Web 04/25/2001 totagmin isoforms can form heterodimers ( 16-22) has led to models whereby heterooligomerization of different combinations of synaptotagmins may lead to Ca 2+ sensors with distinct Ca2+ sensitivities, although at least some of these combinations may not be relevant since not all synaptotagmins are localized on synaptic vesicles ( 23). Furthermore, it is unclear whether the synaptotagmin I C 2B domain actually dimerizes as a function of Ca 2+. Thus, direct Ca 2+-dependent dimerization of the recombinant C 2B domain and the double C2 domain region has been reported in one study ( 21) but disputed in another ( 24). The significance of other interactions described for the synaptotagmin I C2B domain is also unclear. The C2B domain was found to bind in a Ca 2+-independent manner to the clathrin adaptor protein AP-2 ( 5), Ca2+ channels ( 26), and inositol polyphosphates ( 27). However, all of these interactions were found to involve the same site of the C 2B domain, a polybasic region that was also implicated in dimerization, and some of them can be reproduced with incomplete C2B domain fragments ( 16, 28) that are unlikely to be properly folded. In addition, interactions of the C 2B domain withâ-SNAP and with SV2 were observed in some studies ( 29, 30) but not in others ( 16, 21). It is possible that these contradictory results may have arisen in part from the common use of immobilized GST -C2B domain fusions that are isolated by affinity chromatography without further purification. In view of the central function of synaptotagmin I in Ca 2+triggered exocytosis and the potential importance of Ca 2+dependent synaptotagmin multimerization, we have now investigated whether the C 2B domain is in fact a Ca 2+binding module and whether it dimerizes as a function of Ca2+. We find that GST-C2B domain fusion proteins produced by standard approaches are highly contaminated by nonproteinaceous impurities. We have developed a procedure to obtain a properly folded, pure C 2B domain and show that the C2B domain does bind Ca 2+ but remains monomeric upon Ca 2+ binding. These results suggest that the C2B domain cooperates with the C 2A domain in the Ca 2+ sensor function of synaptotagmin I via Ca 2+-dependent interactions with target molecules other than the C 2B domain itself. EXPERIMENTAL PROCEDURES Protein Expression, Purification, and Spectroscopy . Since with an Asp374fGly substitution the amino acid sequence of the rat synaptotagmin I C 2 domains is identical to the bovine sequence, DNA fragments encoding the rat C 2B domain (residues 271 -421) and double C 2 domain region (residues 140 -421) were obtained by PCR amplification from bovine synaptotagmin I cDNA using customary designed primers (we use the rat sequence numbers to avoid confusions in the literature). The PCR products were subcloned into the pGEX-KG vector ( 31) and expressed in Escherichia coliBL21 using LB medium or minimal medium containing NH4Cl as the sole nitrogen source for 15Nlabeling. Cell pellets were resuspended in PBS, passed three times through an EmulsiFlex-C5 cell disrupter (Avestin) at 14 000 psi and spun at 28000 g for 30 min. The supernatants were then incubated with glutathione -agarose (1 mL/L of culture). For standard preparations, the resin was extensively washed with PBS until there was no detectable UV absorption in the eluate, equilibrated with cleavage buffer (50 mM Tris, pH 8.0, 0.2 M NaCl, 2.5 mM CaCl 2), and cleaved with FIGURE 1: Domain diagram of synaptotagmin I and structural model of its C 2B domain. (A) Domain diagram with the different regions of the synaptotagmin I sequence indicated below (Iv, intravesicular sequence; TM, transmembrane region) and their approximate boundaries indicated above. The two C 2 domains are represented by red boxes. (B) Sequence alignment of the rat synaptotagmin I (RSI) C 2A and 2B domains and the rat rabphilin (rRb) C 2B domain. Residues involved in Ca 2+ binding to the synaptotagmin I C 2A domain (8) and the rabphilin C2B domain (9), as well as those predicted to bind Ca 2+ in the synaptotagmin I C 2B domain based on the observed homology, are shown with a dark background. Other background color coding highlights conserved residues according to their location in the threedimensional structures of C 2 domains, which consist of a â-sandwich with loops emerging at the top and the bottom ( 6-9): blue,â-strands; yellow, top loops; green, bottom loops. A specific feature of C 2B domains, not present in C 2A domains, is anR-helix at the bottom of the domain (conserved residues highlighted in red) ( 9). Gly374 of the synaptotagmin I C 2B domain, which was an aspartate in the original rat sequence, is indicated with an asterisk, and the polybasic region is indicated by a solid bar. (C) Ribbon diagram of the structure of the rabphilin C2B domain (9) that we use as a model for the structure of the synaptotagmin I C 2B domain. The positions of residues exhibiting 1H-15N HSQC cross-peak shifts due to binding of contaminants to the synaptotagmin I C 2B domain (see Figure 5) are colored in blue (for basic residues) or cyan (for other residues). The position of K327 is labeled to indicate the location of strand 4, which contains the polybasic sequence involved in binding to the contaminants. Nand C-termini are indicated by N and C, respectively. The ribbon diagram was prepared with the program Molmol ( 39). Accelerated Publications Biochemistry, V