Synaptotagmins: C2-Domain Proteins That Regulate Membrane Traffic

Synaptotagmins: C2-Domain Proteins That Regulate Membrane Traffic

Neuron, Vol. 17, 379–388, September, 1996, Copyright 1996 by Cell Press Synaptotagmins: C2-Domain Proteins That Regulate Membrane Traffic Thomas C. ...

3MB Sizes 14 Downloads 86 Views

Neuron, Vol. 17, 379–388, September, 1996, Copyright 1996 by Cell Press

Synaptotagmins: C2-Domain Proteins That Regulate Membrane Traffic Thomas C. Su¨dhof*† and Josep Rizo‡ *Howard Hughes Medical Institute † Department of Molecular Genetics ‡ Departments of Biochemistry and Pharmacology The University of Texas Southwestern Medical Center Dallas, Texas 75235

The synaptotagmins are a large family of membrane proteins consisting of at least nine genes that are expressed in brain and in other organs. Their salient features are a single transmembrane region and two copies of a Ca21 regulatory domain called the C 2-domain. C2domains were originally defined as sequence motifs in protein kinase C (Kikkawa et al., 1989). They have been identified in more than 50 proteins where they perform a variety of functions (Brose et al., 1995). In synaptotagmins, the C 2-domains mediate Ca21-dependent and Ca21-independent interactions with target molecules that may regulate membrane fusion and membrane budding reactions. In the present review, we first focus on synaptotagmin I and summarize current ideas about its role as an essential component of the fast Ca21-dependent machinery for neurotransmitter release. We then discuss the functional properties of other synaptotagmins for which there is much less information. Structure of Synaptotagmin I Synaptotagmin I was discovered as p65 using a monoclonal antibody that was raised against synaptic plasma membranes but found to react with a synaptic vesicle protein (Matthew et al., 1981). Purification, cloning, and biochemical mapping revealed that synaptotagmin I is an intrinsic membrane protein with a domain structure that is conserved in vertebrates and in invertebrates (Perin et al., 1990; Perin et al., 1991a, 1991b). Synaptotagmin I represents a type I membrane protein with a short intravesicular amino terminus (domain I in Figure 1) that is N-glycosylated followed by a single transmembrane region (domain II) that contains multiple palmitoylated cysteines (Chapman et al., 1996b). Most of the cytoplasmic sequences of synaptotagmin I are accounted for by two C2-domains (domains IV and VI) that are separated from the transmembrane rgion by a highly charged sequence (domain III), and are connected to each other by a short linker (domain V). At the carboxyl terminus, synaptotagmin I contains a short additional domain (VII) that is also conserved evolutionarily. Synaptotagmin I: Ca21 Sensor in Transmitter Release Ca21 influx into the presynaptic nerve terminal rapidly triggers synaptic vesicle exocytosis (200–500 microseconds). The Ca21 concentration dependence for release suggests that release requires the cooperative interactions of multiple Ca21-binding sites (3–6) (Dodge and Rahamimoff, 1967) and high concentrations of Ca21 (.100 mM) (Augustine et al., 1987; Llinas et al., 1995).

Review

Ca21 has several regulatory functions in nerve terminals. In addition to evoking fast synchronous transmitter release, Ca21 also triggers a second slower phase of asynchronous release. This second phase exhibits a Ca21 cooperativity similar to that of the first fast phase but has a higher Ca21 affinity and different cation specificity (Goda and Stevens, 1994). Furthermore, Ca21 mediates distinct types of short-term synaptic plasticity (Kamiya and Zucker, 1994) and is thought to have effects on other stages of the synaptic vesicle cycle. The finding that synaptotagmin I is a Ca21-binding protein suggested that it may function in a Ca21-regulated step of the synaptic vesicle cycle (Perin et al., 1990; Brose et al., 1992). The cation specificity of synaptotagmin I and its low Ca21 affinity were consistent with a role in triggering release but could not rule out other functions (Brose et al., 1992; Davletov and Su¨dhof, 1993). An important step in the functional analysis of synaptotagmin was provided by genetic studies in Caenorhabditis elegans, Drosophila, and Aplysia (Nonet et al., 1993; Littleton et al., 1993; DiAntonio et al., 1993; Martin et al., 1995) and by microinjection experiments in squid neurons and PC12 cells (Bommert et al., 1993; Elferink et al., 1993). Together, these results demonstrated that interference with synaptotagmin function inhibits evoked transmitter release but can in some instances even stimulate basal release, pointing toward a regulatory role in exocytosis. However, these studies also raised a number of questions. For example, in synaptotagmin mutants in Drosophila, Ca21-dependent neurotransmitter release is still maintained, suggesting that synaptotagmin is not an obligatory Ca21 sensor for release (DiAntonio et al., 1993). Some combinations of mutant alleles of synaptotagmin in Drosophila lead to changes in the Ca21 dependence of neurotransmitter release (Littleton et al., 1994). Similar changes are also observed with mutations in genes that probably have no direct function in neurotransmitter release, such as phosphodiesterase (Zhong and Wu, 1991), thereby complicating their interpretation. A further test of the function of synaptotagmin I was achieved by analysis of mice homozygous for a mutation in the synaptotagmin I gene (Geppert et al., 1994). The mutant mice develop normally until birth but die postnatally, revealing that synaptotagmin I is essential for life. Thus, in spite of the multiplicity of synaptotagmins and of other synaptic Ca21 sensors, synaptotagmin I is not redundant. Although synapses with apparently normal morphology form in the synaptotagmin I mutants, synaptic transmission is severely abnormal in that the fast phase of Ca21-dependent neurotransmitter release is strongly depressed (Figure 2). By contrast, there is no impairment of the slow asynchronous component of neurotransmitter release. Furthermore, neurotransmitter release is triggered normally by hypertonic sucrose or a-latrotoxin, agents that act by Ca21-independent mechanisms. The frequency of spontaneous miniature release events, the size of the pool of readily releasable synaptic vesicles, and the rate of replenishment of this pool are also indistinguishable between wild-type and

Neuron 380

Figure 1. Domain Structure of Synaptotagmins and Double C2-Domain Proteins The domains shared by synaptotagmin I with all synaptotagmins are shown on top. Those of the double C2-domain proteins that have similar carboxy-terminal domains but no transmembrane regions (rabphilin-3, B/K protein, doc2’s) are depicted on the bottom. The domains are numbered with roman numerals, and the range in the number of residues present in each domain in different proteins is shown below each domain. The positions of the transmembrane region (TMR) and the two C2-domains (C2 A and C2 B, respectively) are indicated.

synaptotagmin I knockout mice (Geppert et al., 1994; Y. Goda, M. Geppert, C. F. Stevens, and T. C. S., unpublished data). Thus, these mice suffer from a selective loss of only one of the two components of release, the fast Ca21-dependent exocytosis, without changes in other synaptic vesicle functions. In preparation for Ca21-triggered exocytosis, synaptic vesicles must first dock at the active zone and then undergo a priming reaction. The priming of the vesicles is thought to involve three membrane proteins, syntaxin, SNAP-25, and synaptobrevin/VAMP, but the molecular mechanisms underlying priming are unknown (reviewed by Martin, 1994; Su¨dhof, 1995). After priming, vesicles appear to exist in a meta-stable partially fused state, awaiting a Ca21 signal. It is likely that the spontaneous release of transmitters originates from the relative instability of primed vesicles. Hypertonic sucrose and a-latrotoxin probably act on these primed vesicles (Rosenmund and Stevens, 1996). The fact that the frequency of spontaneous release and the amount of release triggered by hypertonic sucrose or a-latrotoxin are normal in the mutant mice demonstrates that synaptotagmin I

is not essential for priming. By contrast, the impairment of fast Ca21-triggered release, the last step in exocytosis, implies that synaptotagmin I functions in this step. Thus, the most likely explanation for these data is that synaptotagmin I serves as an essential component of the Ca21 sensor for the final step in membrane fusion (Kelly, 1995). Alternatively, it is possible that synaptotagmin I functions not as the Ca21 sensor itself but as a link between the Ca21 sensor and Ca21 channels (Neher and Penner, 1994). To examine this question, it is essential to determine if synaptotagmin I has biochemical properties that are consistent with its being a Ca21 sensor. Ca21-Dependent Properties of Synaptotagmin I The following properties of synaptotagmin I support the idea that it forms part of the Ca21 sensor in fast exocytosis (summarized in Table 1): 1. Synaptotagmin I contains two C2-domains. Both C2-domains are stabilized by Ca21, suggesting that they both bind Ca21 (Davletov and Su¨dhof, 1994). The first C2domain (C2A-domain) binds two Ca21 ions with intrinsic affinities of z60 mM and of z300–400 mM, respectively Figure 2. Evoked Synaptic Responses in Wild-Type Mice and Synaptotagmin I Mutant Mice The traces show evoked synaptic currents recorded from cultured hippocampal neurons. Traces in (ii) are enlarged 4-fold vertically compared with (i) in order to illustrate the slow component of release. Figure was adapted from Geppert et al. (1994).

Review 381

Table 1. Molecules That Bind to Synaptotagmins Domain of Synaptotagmin

Synaptotagmin Isoforms Studied

Ca Phospholipids

C 2-A Domain C 2-A Domain

I All

Syntaxin

C 2-A Domain

All

Synaptotagmin

C 2-B Domain

I

60 mM and 400 mM I, II, III, V, VII 5 5–10 mM IV, VI, VIII, IX 5 no Differs between synaptotagmin isoforms (see Table 2) I: ≈250 mM Ca21

Zygin

C 2-B Domain

I

Yes

Munc13-1

Unknown

I

Yes

AP-2

C 2-B Domain

No

b-SNAP

C 2-B Domain

All except for II, IX I

Polyglutamatestring proteins Polyinositolphosphates Neurexins

C 2-B Domain

I

ND

C 2-B Domain

I, II

No

C-terminus

I

No

Molecule 21

Ca21-Dependence

No

Comments

Reference

Possible role in membrane fusion Syntaxin is involved in synaptic vesicle fusion

Shao et al., 1996 Li et al., 1995a Ullrich et al., 1994 Li et al., 1995a Kee and Scheller, 1996

May mediate Ca21-dependent Synaptotagmin polymerization Evolutionarily conserved ubiquitous trafficking protein May function in fusion May mediate recruitment of clathrin for endocytosis Closely related a-SNAP does not bind Probably binds to polylysine stretch on C 2-domain InsP 6 binds best; no regulation Includes high-affinity a-latrotoxin receptor

Sugita et al., 1996 Chapman et al., 1996 Sugita et al., submitted Betz, Su¨dhof, and Brose, unpublished data Zhang et al., 1994 Li et al., 1995a Schiavo et al., 1995 Sugita et al., submitted Fukuda et al., 1994 Hata et al., 1993 Perin, 1994

Abbreviations: ND, not determined.

(Shao et al., 1996; see also Figure 5). These affinities correlate with the Ca21 dependence of exocytosis (Heidelberger et al., 1994). The sequence and Ca21dependent interactions of the second C2 -domain (C2Bdomain) suggest that it also binds Ca21 with comparable affinities, thereby supplying synaptotagmin I with two independent Ca21-binding modules. 2. Native brain synaptotagmin I binds phospholipids in a Ca21-dependent manner (Brose et al., 1992). Phospholipid binding is mediated exclusively by the C 2Adomain (Davletov and Su¨dhof, 1993; Chapman and Jahn, 1994). Ca21 acts cooperatively with an EC50 of z5–10 mM; all negatively charged phospholipids tested bind independent of their headgroup (Davletov and Su¨dhof, 1993). No binding of phospholipids to the C2domains of native synaptotagmin I is observed without Ca21 (Li et al., 1995b). The Ca21 affinity of the C2 A-domain for phospholipid binding is higher than its intrinsic Ca21 affinity. This may result from the stabilization of the Ca21/C 2A-domain complex by the negatively charged phospholipid vesicles, a hypothesis that is supported by the finding that the Ca21 affinity for phospholipid binding increases with the percentage of negatively charged phospholipids in the liposomes (X. Shao, T. C. S., and J. R., unpublished data). Alternatively, occupation of only one Ca21-binding site may be sufficient for phospholipid binding. 3. Synaptotagmin I binds syntaxin, a protein with an essential function in exocytosis (reviewed by Bennett and Scheller, 1994). Binding is Ca21 dependent and mediated by the C2A-domain of synaptotagmin I. Half-maximal binding occurs at z250 mM Ca21 (Li et al., 1995a; Kee and Scheller, 1996), an affinity that parallels the intrinsic Ca21 affinity of the C2A-domain (Shao et al., 1996).

4. The C2B-domain of synaptotagmin I mediates the Ca21-dependent self-association of synaptotagmin I into multimers. This reaction also has a Ca21 dependence (z250 mM) similar to that of intrinsic Ca21 affinity of the C 2A-domain and of the exocytotic Ca21 sensor (Sugita et al., 1996; Chapman et al., 1996a). Since synaptotagmin I also forms Ca21-independent multimers via amino-terminal sequences (Brose et al., 1992), large Ca21-dependent multimeric structures could be formed during exocytosis. 5. Synaptotagmin I also binds in a Ca21-dependent manner to zygin I, a neuron-specific homolog of the unc76 gene of C. elegans (Sugita et al., submitted). Similar to synaptotagmins, zygins are present in neuron-specific and in ubiquitous isoforms in mammals and may represent general synaptotagmin effector molecules. Together, these studies establish that synaptotagmin I has the biochemical properties of a cooperative Ca21 sensor: it is a protein composed of homomultimers, with each monomer containing two independent Ca21-binding modules, each of which again binds two Ca21 ions. The two Ca21-binding modules are composed of homologous C2-domains but nevertheless bind distinct targets. The Ca21 affinity and cooperativity of synaptotagmin I correlate with those of the Ca21 sensor in exocytosis. An attractive hypothesis is that Ca21 simultaneously triggers interactions of the C 2A-domain of synaptotagmin I with phospholipids and with syntaxin (both of which are probably involved in membrane fusion), and of the C2B-domain with itself and possibly other molecules. According to this model, these interactions cooperate to execute the final step in the fusion reaction, with the fusion pore being either formed directly by polymerizing synaptotagmins or by part of an associated protein, such as the amino-terminal domain of synaptobrevin.

Neuron 382

Molecules That Bind to Synaptotagmin I in a Ca21-Independent Manner In addition to its Ca21-dependent binding activities, synaptotagmin I binds to several molecules in a Ca21-independent manner (Table 1). Ca21-independent ligands include AP-2, which recruits clathrin for endocytosis to membranes (Zhang et al., 1994; see section on endocytosis below), neurexins, which are neuron-specific cellsurface proteins (Hata et al., 1993; Perin, 1994), polyanions, such as polyinositolphosphates (Fukuda et al., 1994) or proteins containing polyglutamate strings (Sugita et al., submitted), and b-SNAP, which, like a-SNAP, recruits N-ethylmaleimide-sensitive factor (NSF) to membranes (Schiavo et al., 1995). The physiological roles of these interactions are unclear. a-latrotoxin induces massive neurotransmitter release and binds to neurexin Ia with high affinity. It would be tempting to speculate that a-latrotoxin causes neurotransmitter release by stimulating the interaction of neurexin Ia with synaptotagmin I. However, a-latrotoxin triggers transmitter release even in synaptotagmin I mutants, demonstrating that synaptotagmin I is not required for a-latrotoxin action (Geppert et al., 1994). Similarly, although injections of inositolphosphates inhibit release (Llinas et al., 1994), inositolphosphates bind to many proteins and are expected to have multiple effects. This complicates the assignment of their effects to synaptotagmins. Furthermore, the inositolphosphate that binds with highest affinity to synaptotagmins is IP6 , the most acidic compound, whose cellular concentrations are relatively high and not known to be regulated, making a signaling function unlikely. Finally, only b-SNAP, but not a-SNAP, binds to synaptotagmin I, a puzzling finding since a-SNAP and b-SNAP are structurally and functionally very similar (Whiteheart et al., 1993). Although it is more difficult to evaluate in a Ca21binding protein the functional significance of Ca21-independent interactions than of Ca21-dependent interactions, at least some of the Ca21-independent interactions will undoubtedly be important. The C2-domain is a compact domain, of which only one side is Ca21 regulated (Shao et al., 1996; Shao et al., submitted; see below). The other, Ca21-independent side of the domain could physiologically bind molecules in interactions that are functionally relevant but not Ca21 regulated. Thus, the C 2-domains could function as janus-faced modules with one Ca21-dependent and one Ca21-independent side.

The domain is composed of two b-sheets with four b-strands (labeled with roman numerals). The two Ca21 ions are shown in green. The regions of the C 2A-domain that exhibit Ca21-dependent chemical shift changes by NMR spectroscopy are depicted in orange, and regions with no Ca21-dependent chemical shift changes in blue. N and C indicate the amino and carboxyl terminus, respectively. The topology of the b-strands in the C2A-domain shown (topology A) differs from that of the C2-domain of phospholipase C-d1 (topology B). Topology B consists of a circular permutation of topology A, resulting in amino and carboxyl termini at the bottom of the domain (shown in parentheses) and a corresponding renumbering of the b-strands (indicated with roman numerals in parentheses next to the numbers for topology A).

C2-Domains as Ca21 Regulatory Modules The two C2-domains of synaptotagmin I form the executive center of the protein. They represent Ca21-binding modules that fold independently and have distinct physiological roles. To gain insight into how C2-domains function in synaptotagmin I and by what mechanism Ca21 regulates their activity, a structural understanding is required. The crystal structure of the C2A-domain from synaptotagmin I showed that it is composed of a compact b-sandwich formed by two b-sheets with three flexible loops on the top and the bottom (Sutton et al., 1995). Each of the b-sheets is comprised of four b-strands

(Figure 3). Sequence alignments of C2 -domains from synaptotagmins reveal that the b-strands are highly conserved, suggesting that the b-sandwich core is similar in all C2-domains (Figure 4). The loops have a more variable sequence and may confer binding specificities to the C2-domains. Although the crystal structure did not allow the definition of the mechanism of Ca21 binding to the C2 A-domain, analysis by nuclear magnetic resonance (NMR) spectroscopy showed that it contains a novel bipartite Ca21-binding motif (Figure 5). This motif, designated the C2-motif, is formed by the loops at the top of the molecule (Shao et al., 1996). Ca21 binding is mediated by five conserved aspartate residues located

Figure 3. Ribbon Diagram of the C2 A-Domain of Synaptotagmin I

Review 383

Figure 4. Sequence Alignment of the C2 Domains and Carboxy-Terminal Sequences of Synaptotagmins and Related Double C2 Domain Proteins The sequences of the C2A- and C2 B-domains (identified by [A] and [B] on left) of rat synaptotagmins I–VII and IX (RSI–RSIX), mouse synaptotagmin VIII (MSVIII), rat Doc2 proteins (RDoc15doc2b; RDoc25doc2a), rat rabphilin-3 (RRb), and rat B/K protein (B/K) are aligned with each other in separate groups. Negatively charged residues in the same positions as the aspartate residues that bind Ca21 in the C 2A-domain of synaptotagmin I are colored in red. Residues that are present in more than 50% of the sequences in both the C2 A- and C2B-domains are colored yellow, those that are present in more than 50% of only the C2 A-domain sequences in blue, and those present in more than 50% of only the C2B-domain sequences in green. Residues conserved in more than 50% of the carboxyterminal sequences are shown on an orange background. The locations of the b-strands (Figure 2) are indicated by the arrows above the sequence.

in two loops (red in Figure 4), with three of the aspartates acting as bidentate ligands. Surprisingly, Ca21-binding produces only a minimal

Figure 5. Diagram of the Bipartite Ca21-Binding Motif of the C2ADomain of Synaptotagmin I (the C2-Motif) The relative positions of the two Ca21 ions and the five aspartate side chains participating in Ca21 binding are indicated. Dashed lines illustrate the coordination of the two Ca21 ions. Solid curved lines represent the protein backbone linking aspartate residues that are close in the sequence. Asp-172, Asp-230, and Asp-232 act as bidentate ligands, while Asp-178 and Asp-238 bind one Ca21 ion each (modified from Shao et al., 1996). The Asp-178 Ca21-binding site corresponds to the high affinity site and the Asp-238 site to the low affinity site.

conformational change in the top part of the C 2A-domain and no conformational change in other parts of the domain. However, it causes a major change in the electrostatic potential of the module. How then does Ca21 trigger binding of a target to the C2A-domain? Analysis by NMR spectroscopy showed that syntaxin binding to the C 2A-domain involves only the region of the Ca21-binding sites (Shao et al., submitted). The Ca21-binding sites of the C2A-domain are surrounded by a ring of positive charges. Mutations of these positive charges increase the Ca21 affinity of the C2A-domain but decrease its ability to bind syntaxin (Shao et al., submitted). The most plausible interpretation of these results is that the Ca21dependent change in electrostatic potential of the C2Adomain triggers syntaxin binding. According to this hypothesis, the C2A-domain acts as an electrostatic switch in which Ca21 induces a binding reaction through a change in the electrostatic potential, not through a conformational change. The mechanism of Ca21 regulation of the C2A-domain is very different from that of other intracellular Ca21 regulatory domains. Ca21-binding proteins such as calmodulin and troponin C contain multiple cooperative EF hands, each of which is formed by a contiguous sequence that independently binds Ca21. Ca21 binding causes a large conformational change, exposing a hydrophobic region that mediates target recognition (Gagne´ et al., 1995; Kuboniwa et al., 1995; Zhang et al., 1995). By contrast, recognition of syntaxin by the C2 Adomain involves local changes in electrostatic potential rather than a major conformational change. Thus, as an

Neuron 384

intracellular Ca21 regulatory domain, the C2A-domain is not only unique in that the Ca21-binding sites are formed by bidentate aspartate ligands on two loops that are widely separated in the primary sequence, but also in the mechanism by which Ca21 binding regulates its reactivity. These properties may be particularly useful for physiological reactions that require fast local reactions of Ca21 in order to regulate a cellular process, such as during transmitter release. How generally applicable is this model of Ca21 regulation to other C2-domains? In addition to the C2A-domain of synaptotagmin I, the three-dimensional structure of only one other C2 -domain, that from phospholipase C-d1, is known (Essen et al., 1996). The two C 2-domains contain almost identical b-sandwich structures, suggesting that the overall structure of C2-domains is very similar. Interestingly, however, the topology of the b-strands in the two C2-domains differs. In the C2Adomain from synaptotagmin I (Figure 3), the amino and carboxyl termini are located on the top of the module, with the carboxy-terminal b-strand (strand VIII) forming the right front edge of the b-sheet (topology A). The phospholipase C-d1 C2-domain (topology B) contains a circular permutation of the b-strands, connecting the amino- and carboxyl termini to form a loop, and disconnecting the loop formed by strands I and II so that the amino and carboxyl termini are now at the bottom of the molecule. These topologies impose different geometrical constraints onto domains preceding or following C2-domains, for example, if two C2-domains are present in tandem, but otherwise result in the same stable three-dimensional structure containing variable top and bottom loops. Sequence alignments suggest that most C2-domains have the C2 -motif, indicating similar Ca21-binding properties. This was confirmed in NMR experiments for the C 2-domain of protein kinase Cb (Shao et al., 1996). Although it is unknown if and how the C 2-domain from phospholipase C-d1 binds Ca21, based on the location of the aspartate residues in the top loops, it is likely that it also contains a functional C2-motif. However, some C 2-domains, including those of several synaptotagmins, contain amino acid substitutions in the Ca21-binding sites, which are expected to abolish Ca21 binding, particularly if they occur in the aspartate residues that coordinate two Ca21 ions (Figure 4). Thus, most C2-domains probably have a similar three-dimensional structure and bind Ca21 by similar bipartite binding sites but with distinct affinities, with some C2-domains containing substitutions in the Ca21-binding site that abolish Ca21 binding.

The Synaptotagmin Gene Family Nine synaptotagmins are currently known, making this an unusually large family of trafficking proteins (Table 2; Perin et al., 1990; Geppert et al., 1991; Wendland et al., 1991; Mizuta et al., 1994; Hilbush and Morgan, 1994; Ullrich et al., 1994; Craxton and Goedert, 1995; Hudson and Birnbaum, 1995; Li et al., 1995a). Two different synaptotagmins were simultaneously named synaptotagmin V. We propose to continue to call the one of Li et al. (1995a) synaptotagmin V because it is closer to synaptotagmins I and II than the other synaptotagmin

V (Craxton and Goedert, 1995; Hudson and Birnbaum, 1995), which we will refer to as synaptotagmin IX. All synaptotagmins have the same overall domain structure as synaptotagmin I (Figure 1). However, the sequences of their amino-terminal three domains (I–III) are not similar. Here, only synaptotagmins I and II are homologous to each other, and synaptotagmins III and VI are weakly related. By contrast, the sequences of the carboxy-terminal domains (IV–VI) are highly homologous to each other in all synaptotagmins, as shown in the alignment in Figure 4. The differences between the sequences of the C2A- and C2 B-domains are conserved between different synaptotagmins, indicating that the functional specializations between C2-domains are conserved. Only one synaptotagmin was studied in the invertebrates and probably corresponds to synaptotagmin I (Perin et al., 1991a; Bommert et al., 1993; Nonet et al., 1993; Martin et al., 1995); however, others are likely to be present. Open reading frames containing an amino-terminal transmembrane region and multiple C2domains are also found in yeast (Brose et al., 1995), suggesting that a functional equivalent of synaptotagmin may exist here as well. In addition to synaptotagmins, four other mammalian proteins are known that contain two C2-domains followed by a carboxyl terminus homologous to that of the synaptotagmins: rabphilin-3, which interacts with rab3’s, and three proteins of unknown function called DOC2a, DOC2b, and B/K protein (Figure 1; Shirataki et al., 1993; Orita et al., 1995; Sasakaguchi et al., 1995; Kwon et al., 1996). These proteins are homologous to synaptotagmins (Figure 4) but do not contain a transmembrane region. All synaptotagmins are highly enriched in brain. Synaptotagmins I and II, the most abundant synaptotagmins, are expressed in a complementary pattern: rostral brain areas primarily express synaptotagmin I and caudal brain areas synaptotagmin II, with some neurons containing both synaptotagmins. At least six synaptotagmins are widely expressed in non-neural tissues in addition to brain (III, IV, VI, VII, VIII, and IX; Li et al., 1995a; Hudson and Birnbaum, 1995; R. Fernandez-Chacon and T. C. S., unpublished data). In situ hybridizations showed that some synaptotagmins (III, IV, and VII) are uniformly distributed among neurons, whereas others (I, II, and VI) are differentially expressed (Ullrich et al., 1994; Ullrich and Su¨dhof, 1995; Marqueze et al., 1995; Vician et al., 1995). Synaptotagmins I, II, and III are enriched in synapses on vesicles. Synaptotagmin I is also present on large dense core granules and endocrine secretory vesicles (Matthew et al., 1981; Walch-Solimena et al., 1993; Jacobsson et al., 1994). The subcellular localization of the other synaptotagmins inside and outside of brain is unknown. Functions of Other Synaptotagmins Synaptotagmins I and II probably function similarly to primary Ca21 sensors in fast neurotransmitter release because their structures and properties are so similar and because they are distributed in a complementary pattern (Ullrich et al., 1994). Most of the other synaptotagmins, however, are likely to have distinct functions because many of these synaptotagmins (except for synaptotagmin II) are coexpressed with synaptotagmin I in

Review 385

Table 2. Synaptotagmin Isoforms: Distributions and Relative Properties of the First C2-Domains

Isoforms

Distribution

Ca21-Dependent Binding

Amino Acid Substitutions in Ca21-Binding Sitesa

Phospholipids

Syntaxin 200 mM Ca

Classb

Synaptotagmin I

Rostral brain reigons

No

5–10 mM Ca

Synaptotagmin II

Caudal brain regions

No

5–10 mM Ca21

200 mM Ca21

A

,1 mM Ca

21

B

21

21

A

Synaptotagmin III

Ubiquitous

No

5–10 mM Ca

Synaptotagmin IV

Ubiquitous

Yes

No c

No

C

Synaptotagmin V

Brain

No

5–10 mM Ca21

200 mM Ca21

A

Synaptotagmin VI

Ubiquitous with high levels in selected brain nuclei

No

No

No

C

Synaptotagmin VII

Ubiquitous

No

5–10 mM Ca21

,1 mM Ca21

B

Synaptotagmin VIII

Ubiquitous

Yes

No

No

C

Synaptotagmin IX (originally reported as distinct synaptotagmin V )

Ubiquitous

No

No

ND

C

21

Data were compiled from Ullrich et al. (1994), Li et al. (1995a), Hudson and Birnbaum (1995), and unpublished data. a Refers to the C2A-domain. See Figure 2 for an alignment. b Classes refer to Ca21-dependent binding activities of the C2A-domain: A 5 require high Ca21 concentrations for syntaxin binding; B 5 require low Ca21 concentrations for syntaxin binding; C 5 do not bind phospholipids and syntaxin in a Ca21-dependent manner. Class C is probably heterogenous since it includes synaptotagmins with amino acid substitutions or deletions in the Ca21-binding sites that are unlikely to bind Ca21 and synaptotagmins that may bind Ca21 but have other Ca21-dependent activities. c Ca21-dependent phospholipid binding to synaptotagmin IV was observed in one study (Fukuda et al., 1996) in contradiction to an earlier report (Ullrich et al., 1994). However, the Fukuda et al. (1996) study used ‘‘liposomes’’ from 100% phosphatidylserine, which are unstable, making it unclear what Ca21-dependent process was measured. The C2A-domain in synaptotagmin IV contains an aspartate to serine substitution at a site essential for ligating two Ca21 ions (Shao et al., 1996), making it highly unlikely that this site will bind Ca21. Abbreviations: ND, not determined.

the hippocampal neurons, which are functionally deficient in the knockout mice and therefore cannot substitute for the loss of synaptotagmin I function. What then are their functions? Comparisons of the binding properties of the C 2Adomains from different synaptotagmins reveals that they are strikingly different (Li et al., 1995a, 1995b; Hudson and Birnbaum, 1995; see Table I). Some synaptotagmins (I, II, III, V, and VII) bind phospholipids as a function of Ca21 with an EC50 of z5–10 mM, whereas others (IV, VI, VIII, and IX) do not. Those synaptotagmins that bind phospholipids also exhibit Ca21-dependent binding of syntaxin, but with distinct Ca21 dependencies: synaptotagmins I, II, and V require >200 mM free Ca21. By contrast, synaptotagmins III and VII bind syntaxin at low Ca21 concentrations (<1 mM). Thus, synaptotagmins can be classified into three groups based on their properties (Table 2): synaptotagmins that do not bind phospholipids or syntaxin in a Ca21-dependent manner (IV, VI, VIII, and IX), synaptotagmins that bind syntaxin at low Ca21 concentrations (III and VII), and synaptotagmins that require high Ca21 concentrations for syntaxin binding (I, II, and V). However, no direct information on Ca21 binding is available for synaptotagmins, except for synaptotagmin I. The lack of Ca21-dependent syntaxin and phospholipid binding for some synaptotagmins does not necessarily mean that these do not bind Ca21. In the absence of genetic data, the functional significance of the differences in the Ca21-dependent properties of synaptotagmins remains unclear. Nevertheless, it is striking that synaptotagmins with low Ca21 affinity (I, II, and V) are expressed almost only in brain. Conversely, synaptotagmins with no Ca21-dependent activities or

with a high Ca21 affinity are expressed ubiquitously (III, IV, and VI–IX). This raises the possibility that synaptotagmins I, II, and V are specialized for synaptic vesicle exocytosis requiring high Ca21 concentrations. In contrast, the other synaptotagmins may be involved in more general functions performed at low concentrations of Ca21, such as constitutive exocytosis or recycling of vesicles, or intracellular budding or fusion reactions. The most concrete clues for the function of a synaptotagmin besides synaptotagmins I and II exist for synaptotagmin III. Synaptotagmin III, but not synaptotagmin I, can utilize Sr21 instead of Ca21 to support syntaxin binding (Li et al., 1995b). The slow component of neurotransmitter release that is still intact in the synaptotagmin I knockout mice (Geppert et al., 1994) is stimulated by Sr21, but the fast component is not (Goda and Stevens, 1994), suggesting that synaptotagmin III may mediate the slow component of neurotransmitter release. However, this hypothesis requires genetic testing, which might also provide insight into the enigmatic physiological significance of the slow component. Synaptotagmin Function in Endocytosis Following exocytosis, synaptic vesicles are endocytosed rapidly, probably via clathrin-coated pits (Miller and Heuser, 1984). Recruitment of clathrin for endocytosis is mediated by a protein complex called AP-2 that binds to sites on the inside of the plasma membrane. AP2-binding sites are very abundant in synaptic vesicles, probably in order to allow quicker endocytosis. Analysis of the AP-2-binding sites on synaptic vesicles surprisingly revealed that they were largely comprised of synaptotagmins, raising the possibility that synaptotagmins

Neuron 386

may function in endocytosis as well as in exocytosis (Zhang et al., 1994). The C2 B-domains of all synaptotagmins tested bind AP-2 with high affinity independent of Ca21, suggesting that all synaptotagmins could be endocytotic AP-2 receptors (Li et al., 1995a). Since several synaptotagmin isoforms are expressed ubiquitously, these could potentially serve as AP-2 receptors for endocytosis in all cells. No defect in endocytosis was observed in synaptotagmin I knockout mice, possibly because of the presence of the other synaptotagmins. In synaptotagmin mutants in C. elegans, however, there is a depletion of synaptic vesicles consistent with an endocytic defect (Jorgensen et al., 1995). In addition, injections of antibodies to the C 2B-domain inhibit vesicle recycling in squid, providing further support for a role of synaptotagmins in endocytosis (Fukuda et al., 1995). A function for synaptotagmin I in endocytosis would be very attractive. It would provide an economical mechanism to couple the last, Ca21-dependent step in exocytosis (which involves synaptotagmin I) to rapid endocytosis. By extension, other synaptotagmins could have similar functions in membrane fusion and membrane budding reactions in other trafficking steps, such as the exo- and endocytosis at the plasma membrane of fibroblasts.

Perspectives Although many tantalizing clues exist, it is still too early to assign definite functions to synaptotagmins as a gene family. Only for synaptotagmins I and II is there a specific idea of what they do in vivo, and even here it is not clear if they are involved only in exocytosis or also in endocytosis. Furthermore, the mode of action of synaptotagmin I in membrane fusion is obscure, partly because so little is known about the molecular events during fusion in general. A bewildering complexity of molecules interact with synaptotagmins, most of which specifically bind to either the C2A-domain or the C2 Bdomain. This large number of interacting molecules may not be totally unrealistic if one considers the likely complexity of membrane traffic. There is no reason why membrane fusion should be less complicated than, for example, transcription, which involves the sequential regulated assembly of large protein complexes composed of 30 or more subunits. Our understanding of the biology of synaptotagmins would be greatly enhanced if we could address three major questions. First, what is the exact molecular process that allows synaptotagmin I to couple Ca21 sensing to neurotransmitter release? To address this question, a better understanding of membrane fusion itself and the role of phospholipids will be needed. Second, what are the functions of the other synaptotagmins? To answer this question, it will first be necessary to determine the subcellular localizations and points of action of the different synaptotagmins. Third, which of the binding proteins for synaptotagmins are functionally important, and in what order do they act? Answering these questions will require an in vivo mutational analysis of protein–protein interactions. Addressing these questions should be feasible in coming years and not only provide

insight into the nature of neurotransmitter release but also of the regulation of membrane fusion in general. Acknowledgments We wish to thank Drs. M. S. Brown, J. L. Goldstein, Y. Goda, R. Jahn, E. Kandel, E. Liebe, and C. F. Stevens for invaluable advice and discussions. The work described in this review was partially supported by the National Institutes of Health, Human Frontier Science Project, and W. M. Keck Foundation. References Augustine, G.J., Charlton, M.P., and Smith, S.J. (1987). Calcium action in synaptic transmitter release. Annu. Rev. Neurosci. 10, 633– 693. Bennett, M.K., and Scheller, R.H. (1994). A molecular description of synaptic vesicle membrane trafficking. Annu. Rev. Biochem. 63, 63–100. Bommert, K., Charlton, M.P., DeBello, W.M., Chin, G.J., Betz, H., and Augustine, G.J. (1993). Inhibition of neurotransmitter release by C2-domain peptides implicates synaptotagmin in exocytosis. Nature 363, 163–165. Brose, N., Petrenko, A.G., Su¨dhof, T.C., and Jahn, R. (1992). Synaptotagmin: a Ca21 sensor on the synaptic vesicle surface. Science 256, 1021–1025. Brose, N., Hofmann, K., Hata, Y., and Su¨dhof, T.C. (1995). Mammalian homologues of C. elegans unc-13 gene define novel family of C2-domain proteins. J. Biol. Chem. 270, 25273–25280. Chapman, E.R., and Jahn, R. (1994). Calcium-dependent interaction of the cytoplasmic region of synaptotagmin with membranes. J. Biol. Chem. 269, 5735–5741. Chapman, E.R., An, S., Edwardson, J.M., and Jahn, R. (1996a). A novel function for the second C2-domain of synaptotagmin. J. Biol. Chem. 271, 5844–5849. Chapman, E.R., Brose, N., Johnston, P.A., Su¨dhof, T.C., and Jahn, R. (1996b). Fatty acylation of synaptotagmin I in PC-12 cells and in isolated nerve terminals. Biochem. Biophys. Res. Commun., in press. Craxton, M., and Goedert, M. (1995). Synaptotagmin V: a novel synaptotagmin isoform expressed in rat brain. FEBS Lett. 361, 196–200. Davletov, B.A., and Su¨dhof, T.C. (1993). A single C2 -domain from synaptotagmin I is sufficient for high affinity Ca21/phospholipidbinding. J. Biol. Chem. 268, 26386–26390. Davletov, B.A., and Su¨dhof, T.C. (1994). Ca21-dependent conformational change in synaptotagmin I. J. Biol. Chem. 269, 28547–28550. DiAntonio, A., Parfitt, K.D., and Schwarz, T.L. (1993). Synaptic transmission persists in synaptotagmin mutants of Drosophila. Cell 73, 1281–1290. Dodge, F.A., and Rahamimoff, R. (1967). Co-operative action of calcium ions in transmitter release at the neuromuscular junction. J. Physiol. 193, 419–432. Elferink, L.A., Peterson, M.R., and Scheller, R.H. (1993). A role for synaptotagmin (p65) in regulated exocytosis. Cell 72, 153–159. Essen, L.-O., Perisic, O., Cheung, R., Katan, M., and Williams, R.L. (1996). Crystal structure of a mammalian phosphoinositide-specific phospholipase Cd. Nature 380, 595–602. Fukuda, M., Aruga, J., Niinobe, M., Aimoto, S., and Mikoshiba, K. (1994). Inositol 1,3,4,5-tetrakisphosphate binding to the C2 B domain of IP4 BP/synaptotagmin II. J. Biol. Chem. 269, 29206–29211. Fukuda, M., Moreira, J.E., Lewis, F.M., Sugimori, M., Niinobe, M., Mikoshiba, K., and Llinas, R. (1995). Role of the C2B-domain of synaptotagmin in vesicular release and recycling as determined by specific antibody injection into the squid giant synapse preterminal. Proc. Natl. Acad. Sci. USA 92, 10708–10712. Fukuda, M., Kojima, T., and Mikoshiba, K. (1996). Phospholipid composition dependence of Ca21-dependent phospholipid binding to the C2 A-domain of synaptotagmin IV. J. Biol. Chem. 271, 8430–8434.

Review 387

Gagne´, S.M., Tsuda, S., Li, M.X., Smillie, L.B., and Sykes, B.D. (1995). Structures of the troponin C regulatory domains in the apo and calcium-saturated states. Nature Struct. Biol. 2, 784–789. Geppert, M., Archer, B.T., III, and Su¨dhof, T.C. (1991). Synaptotagmin II. A novel differentially distributed form of synaptotagmin. J. Biol. Chem. 266, 13548–13552. Geppert, M., Goda, Y., Hammer, R.E., Li, C., Rosahl, T.W., Stevens, C.F., and Su¨dhof, T.C. (1994). Synaptotagmin I: a major Ca21 sensor for transmitter release at a central synapse. Cell 79, 717–727. Goda, Y., and Stevens, C.F. (1994). Two components of transmitter release at a central synapse. Proc. Natl. Acad. Sci. USA 91, 12942– 12946. Hata, Y., Davletov, B., Petrenko, A.G., Jahn, R., and Su¨dhof, T.C. (1993). Interaction of synaptotagmin with the cytoplasmic domains of neurexins. Neuron 10, 307–315. Heidelberger, R., Heinemann, C., Neher, E., and Matthews, G. (1994). Calcium dependence of the rate of exocytosis in a synaptic terminal. Nature 371, 513–515. Hilbush, B.S., and Morgan, J.I. (1994). A third synaptotagmin gene, Syt3, in the mouse. Proc. Natl. Acad. Sci. USA 91, 8195–8199. Hudson, A.W., and Birnbaum, M.J. (1995). Identification of a nonneuronal isoform of synaptotagmin. Proc. Natl. Acad. Sci. USA 92, 5895– 5899. Jacobsson, G., Bean, A.J., Scheller, R.H., Juntti-Berggren, L., Deeney, J.T., Berggren, P.O., and Meister, B. (1994). Identification of synaptic proteins and their isoform mRNAs in compartments of pancreatic endocrine cells. Proc. Natl. Acad. Sci. USA 91, 12487– 12491. Jorgensen, E., Hartwied, E., Schuske, K., Nonet, M.L., Jin, Y., and Horvitz, H.R. (1995). Defective recycling of synaptic vesicles in synaptotagmin mutants of C. elegans. Nature 378, 196–199. Kamiya, H., and Zucker, R.S. (1994). Residual Ca21 and short-term synaptic plasticity. Nature 371, 603–606. Kee, Y., and Scheller, R.H. (1996). Localization of synaptotagminbinding domains on syntaxin. J. Neurosci. 16, 1975–1981. Kelly, R.B. (1995). Neural transmission: synaptotagmin is just a calcium sensor. Curr. Biol. 5, 257–259. Kikkawa, U., Kishimoto, A., and Nishizuka, Y. (1989). The protein kinase C family: heterogeneity and its implications. Annu. Rev. Biochem. 58, 31–44. Kuboniwa, H., Tjandra, N., Crzesiek, S., Ren, H., Klee, C.B., and Bax, A. (1995). Solution structure of calcium-free calmodulin. Nature Struct. Biol. 2, 768–776. Kwon, O.-J., Gainer, H., Wray, W., and Chin, H. (1996). Identification of a novel protein containing two C2-domains selectively expressed in the rat brain and kidney. FEBS Lett. 378, 135–139. Li, C., Ullrich, B., Zhang, J.Z., Anderson, R.G. W., Brose, N., and Su¨dhof, T.C. (1995a). Ca21 dependent and Ca21 independent activities of neural and nonneural synaptotagmins. Nature 375, 594–599. Li, C., Davletov, B.A., and Su¨dhof, T.C. (1995b). Distinct Ca21 and Sr21 binding properties of synaptotagmins: definition of candidate Ca21 sensors for the fast and slow components of neurotransmitter release. J. Biol. Chem. 270, 24898–24902.

Marqueze, B., Boudier, J.A., Mizuta, M., Inagaki, N., Seino, S., and Seagar, M. (1995). Cellular localization of synaptotagmin I, II, and III mRNAs in the central nervous system and pituitary and adrenal glands of the rat. J. Neurosci. 15, 4906–4917. Martin, T.F. (1994). The molecular machinery for fast and slow neurosecretion. Curr. Opin. Neurobiol. 4, 626–632. Martin, K.C., Hu, Y., Armitage, B.A., Siegelbaum, S.A., and Kandel, E.R. (1995). Evidence for synaptotagmin as an inhibitory clamp on synaptic vesicle release in Aplysia neurons. Proc. Natl. Acad. Sci. USA 92, 11307–11311. Matthew, W.D., Tsaveler, L., and Reichardt, L.F. (1981). Identification of a synaptic vesicle-specific membrane protein with a wide distribution in neuronal and neurosecretory tissue. J. Cell Biol. 91, 257–269. Miller, T.M., and Heuser, J.E. (1984). Endocytosis of synaptic vesicle membrane at the frog neuromuscular junction. J. Cell Biol. 98, 685–698. Mizuta, M., Inagaki, N., Nemoto, Y., Matsukura, S., Takahashi, M., and Seino, S. (1994). Synaptotagmin III is a novel isoform of rat synaptotagmin expressed in endocrine and neuronal cells. J. Biol. Chem. 269, 11675–11678. Neher, E., and Penner, R. (1994). Mice sans synaptotagmin. Nature 372, 316–317. Nonet, M.L., Grundahl, K., Meyer, B.J., and Rand, J.B. (1993). Synaptic function is impaired but not eliminated in C. elegans mutants lacking synaptotagmin. Cell 73, 1291–1305. Orita, S., Sasaki, T., Naito, A., Komuro, R., Ohtsuka, T., Maeda, M., Suzuki, H., Igarashi, H., and Takai, Y. (1995). Doc2: a novel brain protein having two repeated C2-like domains. Biochem. Biophys. Res. Commun. 206, 439–448. Perin, M.S. (1994). The COOH terminus of synaptotagmin mediates interaction with the neurexins. J. Biol. Chem. 269, 8576–8581. Perin, M.S., Fried, V.A., Mignery, G.A., Jahn, R., and Su¨dhof, T.C. (1990). Phospholipid binding by a synaptic vesicle protein homologous to the regulatory region of protein kinase C. Nature 345, 260–261. Perin, M.S., Johnston, P.A., O¨zcelik, T., Jahn, R., Francke, U., and Su¨dhof, T.C. (1991a). Structural and functional conservation of synaptotagmin (p65) in Drosophila and humans. J. Biol. Chem. 266, 615–622. Perin, M.S., Brose, N., Jahn, R., and Su¨ dhof, T.C. (1991b). Domain structure of synaptotagmin (p65). J. Biol. Chem. 266, 623–629. Rosenmund, C., and Stevens, C.F. (1996). Definition of the readily releasable pool of vesicles at hippocampal synapses. Neuron 16, 1197–1207. Sasakaguchi, G., Orita, S., Maeda, M., Igarashi, H., and Takai, Y. (1995). Molecular cloning of an isoform of doc2 having two C 2-like domains. Biochem. Biophys. Res. Commun. 217, 1053–1061. Schiavo, G., Gmachl, N.J., Stenbeck, G., Sollner, T.H., and Rothman, J.E. (1995). A possible docking and fusion particle for synaptic transmission. Nature 378, 733–736. Shao, X., Davletov, B., Sutton, R.B., Su¨dhof, T.C., and Rizo, J.R. (1996). A bipartite Ca21-binding motif in C2 domains of synaptotagmin and protein kinase C. Science 273, 248–251.

Littleton, J.T., Stern, M., Schulze, K., Perin, M., and Bellen, H.J. (1993). Mutational analysis of Drosophila synaptotagmin demonstrates its essential role in Ca21-activated neurotransmitter release. Cell 74, 1125–1134.

Shirataki, H., Kaibuchi, K., Sakoda, T., Kishida, S., Yamaguchi, T., Wada, K., Miyazaki, M., and Takai, Y. (1993). Rabphilin-3A, a putative target protein for smg p25A/rab3A p25 small GTP-binding protein related to synaptotagmin. Mol. Cell. Biol. 13, 2061–2068.

Littleton, J.T., Stern, M., Perin, M., and Bellen, H.J. (1994). Calcium dependence of neurotransmitter release and rate of spontaneous vesicle fusions are altered in Drosophila synaptotagmin mutants. Proc. Natl. Acad. Sci. USA 91, 10888–10892.

Su¨dhof, T.C. (1995). The synaptic vesicle cycle: a cascade of protein–protein interactions. Nature 375, 645–653. Sugita, S., Hata, Y., and Su¨dhof, T.C. (1996). Distinct Ca21-dependent properties of the first and second C2-domains of synaptotagmin I. J. Biol. Chem. 271, 1262–1265.

Llinas, R., Sugimori, M., Land, E.J., Morita, M., Fukuda, M., Niinobe, M., and Mikoshiba, K. (1994). The inositol high-polyphosphate series blocks synaptic transmission by preventing vesicular fusion: a squid giant synapse study. Proc. Natl. Acad. Sci. USA 91, 12990–12993.

Sutton, B., Davletov, B.A., Berghuis, A.M., Su¨dhof, T.C., and Sprang, S.R. (1995). Structure of the first C2-domain of synaptotagmin I: a novel Ca21/ phospholipid binding fold. Cell 80, 929–938.

Llinas, R., Sugimori, M., and Silver, R.B. (1995). The concept of calcium concentration microdomains in synaptic transmission. Neuropharmacology 34, 1443–1451.

Ullrich, U., and Su¨dhof, T.C. (1995). Differential distributions of novel synaptotagmins: comparison to synapsins. Neuropharmacology 34, 1371–1377.

Neuron 388

Ullrich, U., Li, C., Zhang, J.Z., McMahon, H., Anderson, R.G.W., Geppert, M., and Su¨dhof, T.C. (1994). Functional properties of multiple synaptotagmins in brain. Neuron 13, 1281–1291. Vician, L., Lim, I.K., Ferguson, G., Tocco, G., Baudry, M., and Herschman, H.R. (1995). Synaptotagmin IV is an immediate early gene induced by depolarization in PC12 cells and in brain. Proc. Natl. Acad. Sci. USA 92, 2164–2168. Walch-Solimena, C., Takei, K., Marek, K., Midyett, K., Su¨dhof, T.C., DeCamilli, P., Jahn, R. (1993). Synaptotagmin: a membrane constituent of neuropeptide-containing large dense-core vesicles. J. Neurosci. 13, 3895–3903. Wendland, B., Miller, K.G., Schilling, J., and Scheller, R.H. (1991). Differential expression of the p65 gene family. Neuron 6, 993–1007 Whiteheart, S.W., Griff, I.C., Brunner, M., Clary, D.O., Mayer, T., Buhrow, S.A., and Rothman, J.E. (1993). SNAP family of NSF attachment proteins includes a brain-specific isoform. Nature 362, 353–355. Zhang, J.Z., Davletov, B.A., Su¨dhof, T.C., and Anderson, R.G.W. (1994). Synaptotagmin I is a high affinity receptor for clathrin AP2: implications for membrane recycling. Cell 78, 751–760. Zhang, M., Tanaka, T., and Ikura, M. (1995). Calcium-induced conformational transition revealed by the solution structure of apo calmodulin. Nature Struct. Biol. 2, 758–767. Zhong, Y., and Wu, C.-F. (1991). Altered synaptic plasticity in Drosophila memory mutants with a defective cyclic AMP cascade. Science 251, 198–201.