BIOLOGICAL OVERVIEW

Synaptotagmin (Syt) is the major Ca2+ sensor for synaptic vesicle exocytosis. To characterize Ca2+-mediated synaptic vesicle fusion, Drosophila syt mutants deficient in specific interactions mediated by its two Ca2+ binding C2 domains were examined. In the absence of syt, synchronous release is abolished and a kinetically distinct delayed asynchronous release pathway is uncovered. Synapses containing only the C2A domain of synaptotagmin partially recover synchronous fusion, but have an abolished Ca2+ cooperativity. Mutants that disrupt Ca2+ sensing by the C2B domain have synchronous release with normal Ca2+ cooperativity, but with reduced release probability. These data suggest the Ca2+ cooperativity of neurotransmitter release is likely mediated through synaptotagmin-SNARE interactions, while phospholipid binding and oligomerization trigger rapid fusion with increased release probability. These results indicate that synaptotagmin is the major Ca2+ sensor for evoked release and functions to trigger synchronous fusion in response to Ca2+, while suppressing asynchronous release (Yoshihara, 2002).

Katz and colleagues established the hypothesis that Ca²⁺ influx into the presynaptic nerve terminal triggers neurotransmitter release (Katz, 1967). Current models for vesicle exocytosis propose that Ca²⁺ triggers fusion through activation of the SNARE complex. The SNARE complex (the term SNARE stands for SNAP receptors) is formed from the interaction of the synaptic vesicle protein synaptobrevin and the plasma membrane proteins syntaxin and SNAP-25 (see Syntaxin). Together these proteins assemble into a four-helix bundle that is sufficient to drive vesicle fusion in in vitro reconstitution experiments. However, vesicle fusion mediated through the assembly of reconstituted SNARE proteins is slow and Ca²⁺ independent. This sharply contrasts with synaptic transmission, where SNARE complex assembly and subsequent vesicle fusion is rapid and triggered by Ca²⁺ (Yoshihara, 2002 and references therein).

The identity of the Ca²⁺ sensor(s) that triggers vesicle fusion is still under investigation, but many studies point toward an essential function for synaptotagmin in coupling Ca²⁺ to SNARE-mediated fusion. Synaptotagmins form a large family of C2 domain-containing proteins with seven members in Drosophila and 19 members in mammals (Adolfsen, 2001; Craxton, 2001). Synaptotagmin is the most abundant Ca²⁺ binding protein present on synaptic vesicles and accounts for 7% of total vesicle protein (Perin, 1990; Chapman, 1994). Synaptotagmin contains two well-characterized Ca²⁺ binding motifs known as C2 domains. The C2 domain is an abundant motif present in over 100 human proteins and was initially found to encode a Ca²⁺-activated lipid binding domain in protein kinase C. Biochemical studies have demonstrated numerous Ca²⁺-dependent interactions mediated by synaptotagmin that suggest it may couple Ca²⁺ influx to vesicle fusion. Specifically, synaptotagmin binds the SNARE complex and individual t-SNAREs (syntaxin and SNAP-25) in a Ca²⁺-stimulated manner (Chapman, 1995, 1996; Davis, 1999; Schiavo, 1997; Gerona, 2000; Kee, 1996). Synaptotagmin also binds phospholipids in a Ca²⁺-dependent manner through lipid interactions with both C2 domains (Brose, 1992; Chapman, 1994; Davis, 1999; Südhof, 1996; Earles, 2001; Fernandez, 2001). In addition, synaptotagmin undergoes homo-oligomerization (Chapman, 1996; Osborne, 1999; Fukuda, 2000; Sugita, 1996; Littleton, 1999) via Ca²⁺-dependent activation of its C2B domain (Yoshihara, 2002).

Although synaptotagmin has been demonstrated to bind Ca²⁺ in vitro, it is still unknown if synaptotagmin is the major Ca²⁺ sensor for synaptic exocytosis, and if so, how it mediates Ca²⁺ activation of release. Several studies have demonstrated that synaptotagmin is required for normal synaptic transmission. Knockout mice lacking synaptotagmin have greatly reduced synchronous transmitter release following nerve stimulation (Geppert, 1994). Knockin mice with a mutated synaptotagmin that has a 2-fold reduction in Ca²⁺-dependent phospholipid binding by the C2A domain display a 50% reduction in evoked release (Fernández-Chacón, 2001). Drosophila synaptotagmin (syt) mutants also show reduced evoked neurotransmitter release (Littleton, 1993b, 1994; DiAntonio, 1994; Broadie, 1994; Mackler, 2002). However, mutants that affect SNARE proteins also disrupt evoked release, suggesting this phenotype alone is an insufficient indicator of a specific loss of Ca²⁺ sensing versus defects in other stages of vesicle trafficking. The fourth order Ca²⁺ cooperativity of release provides the steep relationship between Ca²⁺ and vesicle fusion, and is the best indicator for a specific role in Ca²⁺ sensing. However, alterations in the Ca²⁺ cooperativity of release have not been conclusively demonstrated for syt mutants. Finally, the relationship between synaptotagmin's biochemical interactions and its role in secretion is largely unknown. To address these questions and determine if synaptotagmin is the major Ca²⁺ sensor for synaptic exocytosis, an electrophysiological analysis was undertaken of Drosophila syt mutants that disrupt distinct functions of synaptotagmin (Yoshihara, 2002).

Three syt alleles (AD1, AD3, and AD4) have been biochemically characterized, and each has defects in specific molecular interactions (DiAntonio, 1994; Littleton, 2001; Fukuda, 2000). syt^AD4 is a null allele caused by an early stop codon that deletes the transmembrane and cytoplasmic domains of the protein and disrupts all of synaptotagmin's known interactions (DiAntonio, 1994). syt^AD1 has a premature stop codon that deletes the C2B domain and reduces Ca²⁺-dependent binding of synaptotagmin to SNAREs and Ca²⁺-dependent oligomerization, while preserving phospholipid binding by the C2A domain (Littleton, 2001; Davis, 1999). syt^AD3 encodes a Y364N change in C2B that does not abolish SNARE or phospholipid binding, but instead disrupts Ca²⁺-dependent conformational changes in C2B that are required for oligomerization of synaptotagmin (Littleton, 2001; Fukuda, 2000). These mutations allow the role of Ca²⁺-dependent phospholipid binding, Ca²⁺-dependent SNARE binding, and Ca²⁺-dependent oligomerization of synaptotagmin in vesicle fusion to be analyzed and separated. Each of these mutants alters exocytosis in a specific fashion, indicating multiple functions for synaptotagmin (Yoshihara, 2002).

These results provide convincing evidence that synaptotagmin is the major low-affinity Ca²⁺ sensor at synapses and integrates Ca²⁺ signals through several effector interactions. The results are summarized in a proposed model. The C2A domain alone can generate synchronous release with a small release probability, likely via Ca²⁺-dependent phospholipid binding by the C2A domain. Lipid binding may allow synaptic vesicles to intimately contact the plasma membrane and facilitate SNARE complex formation. In addition to lipid interactions, high-affinity SNARE binding by an intact C2A-C2B domain of synaptotagmin mediates the Ca²⁺ cooperativity of release, likely by initiating and stabilizing fusion pores formed from SNARE-synaptotagmin complexes. Cooperativity could be accounted for by multiple Ca²⁺ ions binding to one synaptotagmin monomer and triggering SNARE association (stoichiometric model), or by the requirement of multiple synaptotagmin-SNARE interactions per fusion event (stochastic model). Ca²⁺ binding by the C2B domain, which can promote oligomerization of synaptotagmin, dramatically increases release probability, but likely does not contribute to Ca²⁺ cooperativity. Ca²⁺-triggered oligomerization, which has been measured to occur on a submillisecond timescale (Davis, 1999), may be required to rapidly concentrate multiple synaptotagmin bound SNARE complexes into a single interface at the contact point of two lipid bilayers. The zippering together of multiple SNARE complexes, together with the lipid association mediated by the C2 domains, may be sufficient to account for the incredible speed with which fusion occurs in vivo in response to Ca²⁺ (Yoshihara, 2002).

In addition to providing strong evidence that synaptotagmin is the Ca²⁺ sensor for the rapid phase of synaptic vesicle fusion, these data support two kinetically and mechanistically distinct phases of release, consistent with reports from mammalian CNS synapses (Goda, 1994). Recordings from paired hippocampal neurons in culture (Goda, 1994) have revealed a fast component of release with a time constant of 5–10 ms, and a second asynchronous component with a time constant of 100–200 ms. Both of these phases of release have fourth order cooperativity and support the two-Ca²⁺ sensor model of exocytosis: a low-affinity Ca²⁺ sensor that activates the fast synchronous phase of release and a second distinct high-affinity Ca²⁺ sensor that triggers the slow asynchronous release mechanism. The similarities between mammalian CNS synapses and Drosophila NMJs argue that the mechanisms mediating Ca²⁺ cooperativity of release have been conserved across evolution. This work provides evidence that synaptotagmin is the only low-affinity Ca²⁺ sensor at the Drosophila NMJ and that it normally suppresses the second asynchronous phase of release, generating the high fidelity of normal synaptic transmission. In the complete absence of synaptotagmin (syt^AD4), the fast component of release is completely abolished and the second asynchronous phase of release is fully uncovered. The cooperativity of the residual delayed release in the null mutant is unchanged, but reflects the cooperativity of the asynchronous release mechanism. Although an increased variability in release latency was previously observed (Broadie, 1994), the presence of two distinct release mechanisms was not previously realized and the distinction between the cooperativity of fast release versus asynchronous release was not measured. The intact cooperativity of residual release previously reported in the null mutant (Broadie, 1994) reflects the cooperativity of the asynchronous release mechanism, since no synchronous release component remains in the null. In syt^AD1, both the synchronous and asynchronous phases coexist and can be kinetically separated, suggesting the AD1 mutant protein can trigger the fast phase of release (with a cooperativity of 1 instead of 4), but cannot fully suppress delayed release. The partial suppression of the asynchronous phase results in a decrease in the absolute number of delayed release events, but the population time constant for the remaining asynchronous events is unchanged. Both the high-affinity and low-affinity Ca²⁺ sensors that underlie these two phases of release likely impinge upon the basic SNARE fusion machinery, since mutations in syntaxin eliminate both components of release (Yoshihara, 2002).

Although this model is consistent with the data and the previous biochemical characterization of the synaptotagmin mutant proteins used in this study, the fact that other interactions also contribute to the defects observed in syt mutants cannot be ruled out. However, the finding that the cooperativity of neurotransmitter release (n) is abolished in syt^AD1 mutants (from 4 to 1) provides the best evidence to date that synaptotagmin functions as the Ca²⁺ sensor for fast exocytosis. Several other manipulations have also been reported to mildly reduce n, but none have been shown to have as severe an effect on Ca²⁺ cooperativity as syt mutants. Drosophila dunce mutants elevate cAMP levels and show a mild reduction in n to 2.4. It is possible that PKA phosphorylation patterns are altered in the dunce mutant and directly affect synaptotagmin-SNARE interactions. Alternatively, cAMP-dependent facilitation, which is also mediated by Ca²⁺, may alter additional Ca²⁺ effectors of release. Experimental decreases in the levels of the v-SNARE synaptobrevin and the t-SNARE syntaxin also cause a slight reduction in the Ca²⁺ cooperativity of release to 2.6. The reduction in SNARE proteins may cause abnormal cooperation between SNAREs and synaptotagmin during Ca²⁺ influx, thus leading to subtle reductions in cooperativity (Yoshihara, 2002 and references therein).

Given that synaptotagmin has been shown to stabilize the fusion pore during dense core vesicle fusion (Wang, 2001), it is proposed that synaptotagmin-SNARE interactions rapidly trigger the opening and stabilization of the fusion pore, while preventing fusion pore openings induced by lower Ca²⁺ concentration via the high-affinity sensor. A role for synaptotagmin-SNARE interactions in vesicle fusion is supported by voltammetry measurements from cracked PC12 cells using dominant-negative synaptotagmin probes that prevent synaptotagmin-SNARE binding (Earles, 2001). In addition, the Ca²⁺-dependent interaction of synaptotagmin with SNAP-25 has been shown to be essential for the triggering of vesicle fusion in PC12 cells (Zhang, 2002). Recent work on Drosophila synaptotagmin transgenic strains harboring mutants that alter the Ca²⁺ binding properties of the C2A and C2B domains have been reported. Disruptions of C2A Ca²⁺ binding cause a 50% decrease in evoked release from heat shock-induced mutant transgenics, but do not result in a null phenotype (Robinson, 2002). However, the mutant transgenes contain intact C2B Ca²⁺ binding ligands, suggesting C2B phospholipid binding may substitute for defects in C2A. Fernández-Chacón (2002) showed that C2A Ca²⁺ ligand mutations engineered into the intact C2A-C2B recombinant protein do not cause defective Ca²⁺-dependent phospholipid binding of the complete protein, confirming functional redundancy for lipid binding between the two C2 domains. Similar redundant lipid binding activities in synaptotagmin have been previously reported (Earles, 2001). Mutations in C2A that prevent Ca²⁺-dependent phospholipid binding by the intact C2A-C2B domain cause a 50% decrease in release probability, confirming an important role for C2A in lipid binding (Fernández-Chacón, 2001). Robinson (2002) also demonstrates that synaptotagmin IV, which has defective C2A phospholipid binding, can functionally substitute for synaptotagmin I. Interestingly, synaptotagmin IV maintains normal Ca²⁺-dependent interactions with the t-SNARE syntaxin (Littleton, 1999), suggesting Ca²⁺-dependent SNARE interactions may be important for synaptotagmin IV's ability to substitute for synaptotagmin I. Disruptions of the Ca²⁺ binding properties of the C2B domain result in a more severe defect in synaptic transmission, but with phenotypes that are more severe than the null mutant, indicating these transgenes function as dominant-negative inhibitors of vesicle fusion (Mackler, 2002). Although it is still unclear exactly how these dominant-negative phenotypes are manifested biochemically, TEM studies indicate an abundance of docked vesicles in these mutants, suggesting that Ca²⁺ binding to the C2B domain has a post-docking role in release. Although both the results of this study and those of Mackler (2002) support an essential post-docking role for synaptotagmin as the fast Ca²⁺ sensor, it cannot be ruled out that there may be subtle defects in endocytosis and vesicle docking. TEM analysis has revealed an overall decrease in the absolute numbers of synaptic vesicles in nerve terminals of null mutants, suggesting synaptotagmin may be required to maintain a completely wild-type pool of recycling vesicles (Reist, 1998). However, the TEM phenotype was similar between syt^AD3 and null mutants. Given the striking differences in physiology between the null and syt^AD3, the TEM phenotype might reflect a general decrease in overall nerve terminal function that leads to less recycling, but that does not reflect a specific role for synaptotagmin in the process. Indeed, physiological measurements to explore the asynchronous release component remaining in the syt nulls suggests that robust docking and endocytosis mechanisms still exist in the absence of synaptotagmin. Release via the high-affinity Ca²⁺ sensor is robust in syt nulls under conditions of sustained elevated Ca²⁺ levels and can continue at high frequencies for hours. This indicates the pool of vesicles that are cycling in the syt null can undergo endocytosis, docking, etc. at high frequency and high fidelity, and does not support the conclusion that this population of vesicles has major defects in docking or endocytosis (Yoshihara, 2002).

In addition to a role for synaptotagmin in triggering fast vesicle fusion, it is also observed that synaptotagmin suppresses delayed release during sustained Ca²⁺ elevation. It is hypothesized that the positive role of synaptotagmin as the Ca²⁺ sensor promoting the fast release phase is mechanistically distinct from its role in clamping asynchronous fusion. The best evidence for this model is that the AD3 mutant has a completely intact suppression of delayed release, yet is still defective in synaptotagmin's positive role in promoting fusion. syt^AD3 mutants also have less release evoked by salines containing high K⁺ or Ca²⁺ ionophores than the null mutant syt^AD4. The data indicate that in syt^AD4, where there is no promoting or clamping function of synaptotagmin remaining, there is increased release due to the unrestricted activity of the high-affinity Ca²⁺ sensor. In syt^AD3, the suppression of asynchronous release is still intact and thus there is less release promoted through the high-affinity sensor than in syt^AD4. The remaining release events in syt^AD3 can be attributed to synaptotagmin's triggering of synchronous fusion, which is greatly reduced in this mutant. The data do not allow the differentiation of whether synaptotagmin's suppression of asynchronous release is a Ca²⁺-dependent or -independent function. Both asynchronous release and facilitation are postulated to be mediated by the high-affinity Ca²⁺ sensor since they share a similar Ca²⁺ dependence (Kamiya, 1994). The identity of the high-affinity Ca²⁺ sensor responsible for delayed release is unknown, but may be another member of the large synaptotagmin family or a distinct class of Ca²⁺ binding proteins. Though facilitation is important for short-term synaptic plasticity, asynchronous release is extremely deleterious for normal synaptic transmission. The suppression of asynchronous release by synaptotagmin is critical for providing the temporal resolution of synaptic transmission. These properties make Synaptotagmin an efficient Ca²⁺ sensor that generates the high signal-to-noise ratio of neurotransmitter release that is required for brain function (Yoshihara, 2002).

GENE STRUCTURE

cDNA clone length - 2180

Bases in 5' UTR - 359

Exons - 11

Bases in 3' UTR - 398

PROTEIN STRUCTURE

Amino Acids - 474

Structural Domains

Synaptotagmin (p65) is an abundant synaptic vesicle protein that contains two copies of a sequence that is homologous to the regulatory region of protein kinase C. Full length cDNAs encoding human and Drosophila synaptotagmins were characterized to study its structural and functional conservation in evolution. The deduced amino acid sequences for human and rat synaptotagmins show 97% identity, whereas Drosophila and rat synaptotagmins are only 57% identical but exhibit a selective conservation of the two internal repeats that are homologous to the regulatory region of protein kinase C (78% invariant residues in all three species). The two internal repeats of synaptotagmin are only slightly more homologous to each other than to protein kinase C, and the differences between the repeats are conserved in evolution, suggesting that they might not be functionally equivalent. The cytoplasmic domains of human and Drosophila synaptotagmins produced as recombinant proteins in Escherichia coli specifically bind phosphatidylserine similar to rat synaptotagmin. They also hemagglutinate trypsinized erythrocytes at nanomolar concentrations. Hemagglutination is inhibited both by negatively charged phospholipids and by a recombinant fragment from rat synaptotagmin that contains only a single copy of the two internal repeats. Together these results demonstrate that synaptotagmin is highly conserved in evolution compatible with a function in the trafficking of synaptic vesicles at the active zone. The similarity of the phospholipid binding properties of the cytoplasmic domains of rat, human, and Drosophila synaptotagmins and the selective conservation of the sequences that are homologous to protein kinase C suggest that these are instrumental in phospholipid binding. The human gene for synaptotagmin was mapped by Southern blot analysis of DNA from somatic cell hybrids to chromosome 12 region cen-q21, and the Drosophila gene by in situ hybridization to 23B (Perin, 1991).

date revised: 18 February 2003

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.