Membrane penetration by synaptotagmin is required for coupling calcium binding to vesicle fusion in vivo

Abstract

The vesicle protein synaptotagmin I is the Ca(2+) sensor that triggers fast, synchronous release of neurotransmitter. Specifically, Ca(2+) binding by the C(2)B domain of synaptotagmin is required at intact synapses, yet the mechanism whereby Ca(2+) binding results in vesicle fusion remains controversial. Ca(2+)-dependent interactions between synaptotagmin and SNARE (soluble N-ethylmaleimide-sensitive fusion protein attachment receptor) complexes and/or anionic membranes are possible effector interactions. However, no effector-interaction mutations to date impact synaptic transmission as severely as mutation of the C(2)B Ca(2+)-binding motif, suggesting that these interactions are facilitatory rather than essential. Here we use Drosophila to show the functional role of a highly conserved, hydrophobic residue located at the tip of each of the two Ca(2+)-binding pockets of synaptotagmin. Mutation of this residue in the C(2)A domain (F286) resulted in a ∼50% decrease in evoked transmitter release at an intact synapse, again indicative of a facilitatory role. Mutation of this hydrophobic residue in the C(2)B domain (I420), on the other hand, blocked all locomotion, was embryonic lethal even in syt I heterozygotes, and resulted in less evoked transmitter release than that in syt(null) mutants, which is more severe than the phenotype of C(2)B Ca(2+)-binding mutants. Thus, mutation of a single, C(2)B hydrophobic residue required for Ca(2+)-dependent penetration of anionic membranes results in the most severe disruption of synaptotagmin function in vivo to date. Our results provide direct support for the hypothesis that plasma membrane penetration, specifically by the C(2)B domain of synaptotagmin, is the critical effector interaction for coupling Ca(2+) binding with vesicle fusion.

Figure 1.

Both the C₂A and C₂B domains of synaptotagmin have conserved hydrophobic residues at the tip of the Ca²⁺-binding pocket. A, Alignment of synaptotagmin I from rat and Drosophila. Bars indicate β-sheets, asterisks indicate the conserved hydrophobic residues mutated in this study, dots indicate Ca²⁺-binding residues, and open circles with a plus indicate conserved basic residues. Within the alignment, conserved residues are shown in gray and identical residues are boxed. B, Schematic representation of loops 1 and 3 in Drosophila that form the Ca²⁺-binding pockets of each C₂ domain. D₁–D₅ and E₅ indicate Ca²⁺-binding residues. F*, M, I*, and V indicate hydrophobic residues at the tip of the Ca²⁺-binding pockets.

Figure 2.

Evoked release is impaired in C₂A hydrophobic mutants. A, Representative voltage traces recorded from larval muscle fiber 6. Each top trace shows the mean of 30 consecutive evoked responses from the same muscle fiber. The bottom traces show 3 s of continuous spontaneous recordings. Line 3 shown for P[syt^A-FY]. B, Compared with P[syt^WT], the mean EJP amplitude of P[syt^A-FY], lines 3 and 5, and P[syt^A-FE] were significantly decreased (*p < 0.005; P[syt^WT], n = 24; P[syt^A-FY] line 3, n = 20; P[syt^A-FY] line 5, n = 14; P[syt^A-FE], n = 10). No difference was found between P[syt^A-FY] and P[syt^A-FE] (p > 0.9). C, Compared with P[syt^WT], the mean mEJP amplitude of P[syt^A-FY] (line 3 shown) and P[syt^A-FE] showed no significant change (p > 0.05).

Figure 3.

The C₂A hydrophobic mutations do not disrupt synaptotagmin expression or targeting. A, Representative Western blots from P[syt^A-FY] line 3, line 5, and P[syt^A-FE] and their P[syt^WT] controls. B, P[syt^A-FY] line 5 (n = 17) expressed less transgenic protein than P[syt^WT] (n = 28) and the other two C₂A mutants (P[syt^A-FY] line 3, n = 7; P[syt^A-FE], n = 11; p < 0.001). C, Transgenic synaptotagmin was localized to synaptic sites in all lines (line 5 shown for P[syt^A-FY]). Scale bar, 20 μm.

Figure 4.

syt^B-IE mutant embryos complete embryogenesis and exhibit gross morphology indistinguishable from wild-type embryos. Like controls, P[syt^B-IE] mutants were able to develop through stage 17, as demonstrated by the development of trachea (arrows).

Figure 5.

The C₂B hydrophobic mutation does not disrupt synaptic vesicle localization, synaptotagmin expression, or synaptotagmin targeting. A, A representative Western blot of embryos shows similar transgene expression levels for P[syt^WT] and P[syt^B-IE]. Blot was probed with an anti-actin antibody to confirm equal loading. B, A representative nerve terminal within the neuropil of a P[syt^B-IE] mutant ventral nerve cord exhibits abundant small clear vesicles (black arrow). C, Anti-CSP labeling was localized to the neuropil of the CNS in P[syt^B-IE] mutants demonstrating that synaptic vesicles are appropriately localized to synaptic regions. The white arrow indicates the neuropil of the ventral nerve cord; brain ganglia are out of the plane of focus. The mutant synaptotagmin was also correctly localized to the neuropil (D, white arrow) in the CNS, as well as to neuromuscular junctions (E, white arrow) in the peripheral nervous system. P[syt^B-IE] line 6 is shown. Scale bar: B, 180 nm; C, D, 40 μm; E, 5 μm.

Figure 6.

The C₂B hydrophobic mutation inhibits evoked transmitter release more severely than the syt^null mutation. A, Representative current traces recorded from embryonic muscle fiber 6. B, The mean EJC amplitude in transgenic control embryos was 1782 ± 224 pA (−/−; P[syt^WT], n = 8), in P[syt^B-IE] mutant embryos was 312.2 ± 62.6 pA (−/−; P[syt^B-IE], n = 8), and in syt^null mutants was 625.2 ± 89.0 pA (−/−, n = 11). Evoked release in P[syt^B-IE] was significantly less than in syt^null mutants (* vs **, p < 0.02).

Figure 7.

The syt^B-IE mutation abolishes Ca²⁺-dependent, C₂B interactions with membranes. A, C₂AB or C₂B domains with the syt^B-IE mutation are correctly folded. The CD spectra of the mutant domains (C₂AB-IE and C₂B-IE) did not show significant changes compared with wild type (C₂AB-WT and C₂B-WT, respectively). B, The syt^B-IE mutation abolishes the ability of the C₂B domain to bind negatively charged liposomes. Left, A representative co-sedimentation assay; right, graph of three experiments. In the absence of Ca²⁺, C₂AB and C₂B domains do not bind to liposomes containing 15% PS and thus remain in solution (left, −, WT and B-IE). In the presence of 1 mm Ca²⁺, increasing concentrations of liposomes bind increasing amounts of control C₂AB or C₂B domains (C₂AB-WT and C₂B-WT). The syt^B-IE mutation impairs this interaction with C₂AB (C₂AB-IE) but abolishes the interaction with the isolated C₂B domain (C₂B-IE).

Figure 8.

Direct, Ca²⁺-dependent C₂B interactions with t-SNAREs are intact in the syt^B-IE mutation. A, The syt^B-IE mutation does not inhibit Ca²⁺-dependent interactions with solubilized t-SNAREs. Left, A representative GST–syt pull-down assay; right, histogram of three experiments. The Ca²⁺-dependent increase in t-SNARE interactions (black bars relative to white bars) is not disrupted by the syt^B-IE mutation. Ca²⁺-independent t-SNARE interactions are decreased (white bars, C₂AB vs C₂AB I420E and C₂B vs C₂B I420E). B, The syt^B-IE mutation abolishes Ca²⁺-dependent C₂B interactions with membrane-embedded t-SNAREs. Left, A representative co-floatation assay; right, histogram of four experiments. The syt^B-IE mutation does not alter the amount of C₂B domain that binds to PS-free, t-SNARE vesicles (t-SNARE) in the absence of Ca²⁺ (−, left; EGTA, right). However, the syt^B-IE mutation abolishes the increase in t-SNARE vesicles binding induced by 1 mm Ca²⁺ (+, left; Ca²⁺, right; p < 0.05). C, The syt^B-IE mutation impairs C₂AB interactions with membrane-embedded t-SNAREs. Again, the syt^B-IE mutation does not alter the amount of C₂AB domain that binds to PS-free, t-SNARE vesicles in the absence of Ca²⁺ (EGTA, white bars). In the presence of 1 mm Ca²⁺ (Ca²⁺, black bars), the t-SNARE vesicles bind 70% of wild-type levels of C₂AB domain when the syt^B-IE mutation is present (p < 0.001; n = 6).

Figure 9.

Model of the role of Ca²⁺-dependent penetration of the anionic presynaptic membrane by synaptotagmin. The crystal structure of the core complex [Protein Data Bank (PDB) file 1SFC, containing syntaxin (red), SNAP-25 (synapse associated protein of 25 kD) (green), and VAMP (vesicle-associated membrane protein)/synaptobrevin (blue)], the nuclear magnetic resonance structures of the C₂A (PDB file 1BYN) and C₂B (PDB file 1K5W) domains of synaptotagmin (yellow), and Ca²⁺ (pink “+”) were rendered using PyMOL Molecular Graphics System (DeLano Scientific). The synaptic vesicle (SV), the presynaptic membrane (PM), the transmembrane domains, and the link between C₂A and C₂B were added in Adobe Photoshop. A, Docked synaptic vesicle with two synaptotagmin/SNARE complexes shown. B, syt/SNARE complex viewed end on from the site of synaptic vesicle/presynaptic membrane apposition (SNARE complex transmembrane domains in front of plane of section). Ca²⁺-independent priming between the C₂B polylysine motif (yellow, space-filled residues) and SNAP-25 [green, space-filled residues (Zhang et al., 2002; Rickman et al., 2004; Loewen et al., 2006b)] holds the C₂B Ca²⁺-binding site immediately adjacent to the SNARE complex and the C₂A Ca²⁺-binding site nearby. The negative charge of the Ca²⁺-binding pockets (cluster of − symbols) prevents interactions between the tips of the C₂ domains and the presynaptic membrane attributable to electrostatic repulsion. C, Ca²⁺ binding neutralizes the negative charge of the pockets resulting in a strong attraction of the negatively charged, phospholipid head groups of the presynaptic membrane by the bound Ca²⁺ (pink spheres +) and the basic residues at the tips of Ca²⁺-binding pockets (blue, space-filled residues +). Insertion of the hydrophobic residues at the tips of the C₂ domains (gray, space-filled residues) into the core of the presynaptic membrane and then triggers fusion by promoting a local Ca²⁺-dependent buckling of the plasma membrane (Martens et al., 2007; Hui et al., 2009). D, syt/SNARE complexes viewed from the presynaptic membrane. Multiple syt:SNARE complexes (colored as in A–C) mediate fusion (Jahn and Scheller, 2006) of a single synaptic vesicle (gray ring, not to scale). Ca²⁺-dependent membrane penetration by synaptotagmin (large gray ovals, arrows) pulls the presynaptic membrane toward the vesicle in the vicinity of the transmembrane regions of the SNARE complexes (small gray circles, arrowheads).