Synaptotagmin 1 oligomerization via the juxtamembrane linker regulates spontaneous and evoked neurotransmitter release

Abstract

Synaptotagmin 1 (syt1) is a Ca²⁺ sensor that regulates synaptic vesicle exocytosis. Cell-based experiments suggest that syt1 functions as a multimer; however, biochemical and electron microscopy studies have yielded contradictory findings regarding putative self-association. Here, we performed dynamic light scattering on syt1 in solution, followed by electron microscopy, and we used atomic force microscopy to study syt1 self-association on supported lipid bilayers under aqueous conditions. Ring-like multimers were clearly observed. Multimerization was enhanced by Ca²⁺ and required anionic phospholipids. Large ring-like structures (∼180 nm) were reduced to smaller rings (∼30 nm) upon neutralization of a cluster of juxtamembrane lysine residues; further substitution of residues in the second C2-domain completely abolished self-association. When expressed in neurons, syt1 mutants with graded reductions in self-association activity exhibited concomitant reductions in 1) clamping spontaneous release and 2) triggering and synchronizing evoked release. Thus, the juxtamembrane linker of syt1 plays a crucial role in exocytosis by mediating multimerization.

Keywords: exocytosis; neurotransmission; oligomerization; synaptotagmin.

Fig. 1.

The complete cytoplasmic domain of syt1 forms large oligomers in solution upon addition of 6:0 PS. (A) Structure of full-length syt1 embedded in a SV membrane with relevant residue annotations. Structures of the C2 domains were derived from the Research Collaboratory for Structural Bioinformatics Protein Data Bank (PDB) 5t0r (C2A) and PDB 2yoa (C2B); other linker segments, derived from PDB 5w5c, were added. The syt1 oligomerization mutations tested in this study are color-coded as follows: K326,327A, magenta; F349A, orange; R398,399Q, yellow; and the polylysine juxtamembrane region is emphasized in blue. The amino acid sequence (residues 80 through 95) of the syt1 juxtamembrane region immediately following the transmembrane domain is shown below, with lysine residues shown in blue. (B) Cartoon depiction of the DLS assay. (C) Representative results of the DLS analysis showing the diameter of 2 µM WT and mutant syt1(80-421) with and without the addition of 100 µM 6:0 PS. Average diameters for each condition, performed in triplicate, are shown in Table 1. (D) Representative EM images of WT and Juxta K mutant syt1(80-421) samples that were analyzed by DLS with and without the addition of 6:0 PS. (Scale bar: 200 nm.)

Fig. 2.

AFM imaging of WT and mutant syt1(80-421) on supported lipid bilayers. (A) Cartoon depiction of the AFM experimental conditions. (B) Representative AFM topographical images of 1 µM WT syt1(80-421) on supported lipid bilayers (72% DOPC, 25% DOPS, 3% PIP₂) in the presence (1 mM) and absence (0.5 mM EGTA) of Ca²⁺. (C) Representative AFM topographical images of the syt1(80-421) mutants under the same conditions as the WT samples. The protein concentration was 1 µM in all cases except for the R398,399Q sample in which a concentration of 300 nM was used to better visualize ring formation. Pixel size: left column, 9.8 nm/px and right column, 2.0 nm/px. (D) Ring diameters for WT syt1(80-421) (green), K326,327A (purple), F349A (orange), and R398,399Q (yellow) in the absence of Ca²⁺ and of Juxta K (blue) in the presence of Ca²⁺. For the Juxta K sample in Ca²⁺, 41.6% of the observed structures were rings, while the remainder were particles. Additional C2B mutations in the Juxta K background abolished ring formation, so these samples were excluded from ring analysis. **** denotes P values < 0.0001 for Juxta K compared to all other conditions. Other statistical comparisons were not significantly different (ns). For WT, F349A, K326,327A, R398,399Q, and Juxta K, the number of rings measured were the following: 26, 13, 9, 10, and 5 for each condition from 3, 2, 2, 2, and 2 sets of analyses using independent SLBs, respectively. Error bars represent SEM.

Fig. 3.

WT and mutant syt1 constructs are efficiently targeted to SVs. (A) Representative anti-syt1 immunoblots of WT, syt1 KO (generated using a Cre virus), and KO neurons expressing WT syt1, Juxta K mutant syt1, F349A mutant syt1, and the double Juxta K + F349A mutant syt1 in mouse hippocampal neurons at 15 DIV. (B) Representative super-resolution fluorescent immunocytochemistry images from mouse hippocampal neurons at 21 DIV. Images of WT, syt1 KO (+Cre), and KO neurons expressing WT, Juxta K mutant, F349A mutant, and the double Juxta K + F349A mutant syt1 variants stained with anti-synaptophysin (magenta) and anti-syt1 (yellow) antibodies; in the last column, these signals are merged. (Scale bar, 10 µm.) (C) Bar graph of the Pearson’s correlation coefficient statistic R. Values were obtained using regions of interest (ROIs) from whole fields of view and JaCoP for ImageJ . Plot of the median Pearson’s R ± 95% Cl; n = 15 for each condition from three trials. Line of idenity at y = 0 (dashed line). These data were not normally distributed; **** denotes P value < 0.0001 between WT and syt1 KO; no other conditions were statistically significantly different from WT using Kruskal–Wallis test with Dunn’s correction. Full statistics are included in the figure source data. (D) Partially cropped, averaged, and normalized vGlut1-pHluorin traces from each indicated condition obtained using wide-field fluorescence, imaged once a second for 120 s. A 20-Hz field stimulation began at t = 10 s (200 action potentials) in 14 DIV hippocampal mouse neurons. (E) vGlut1-pHluorin 10 to 90% peak rise times were plotted. Values are median with 95% CI; 500 to 700 ROIs (n) were analyzed from three separate trials. These data were not normally distributed; **** denotes P value < 0.0001 between WT and syt1 KO, and no other conditions were statistically significantly different from WT using Kruskal–Wallis test with Dunn’s correction. (F) vGlut1-pHluorin decay times represented as time constants determined from fitting data to a single exponential function. Values are median with 95% CI; 500 to 700 ROIs (n) measured from three separate trials. These data were not normally distributed; **** denotes P value < 0.0001 between WT and syt1 KO and WT and F349A mutant rescue; no other conditions were statistically significantly different (ns) from WT using Kruskal–Wallis test with Dunn’s correction.

Fig. 4.

Lysine residues in the syt1 juxtamembrane linker regulate evoked neurotransmitter release. (A) Representative iGluSnFR traces from one field of view after a single field stimulus (indicated by a black arrow) imaged at 100 Hz. The plots show the iGluSnFR responses from multiple individual regions of interest (ROIs) (black) and the average responses (green). Note that the KO, KO + Juxta K, and KO + Juxta K, F349A conditions resulted in fewer responses after stimulation. (B) Histograms of iGluSnFR (ΔF/F0) peaks plotted using 10-ms bins. The timing of the stimulus is indicated with a black arrow. Peaks were binned over the entire 1.5 s of recording. Conditions were color-coded as follows: WT (black), syt1 KO (gray), KO + WT (green), KO + Juxta K (blue), KO + F349A (orange), and KO + Juxta K, F349A (magenta). The histograms include combined data from three independent trials. The y-axis zoom-in histograms of iGluSnFR (ΔF/F0) peaks are shown as Insets to emphasize the presence of asynchronous release. (C) Average traces of iGluSnFR ΔF/F0 from a single field stimulus, indicated by a black arrow, imaged at 100 Hz. Note that the 1-ms stimulus was applied in the last millisecond of the frame that is indicated by a black bracket; this causes a small portion of the rise in signal to be captured within that frame. Conditions are labeled in the figure, and the same color scheme is used throughout. Briefly, WT (black), syt1 KO (gray), WT syt1 rescue (green), Juxta K mutant rescue (blue), F349A mutant rescue (orange), and the double Juxta K + F349A mutant rescue (magenta). Values are means with 95% CI error bars; the number of ROIs analyzed were WT (1,155), syt1 KO (212), WT rescue (748), Juxta K rescue (547), F349A rescue (1,128), and double-mutant rescue (324), collected from three separate trials. (D) Average peak iGluSnFR responses 10-ms post-stimulus from each condition. Quantification of data in C. Values are mean ± SEM from three independent trials. These data were normally distributed, and only WT rescue and F349A mutant rescue were statistically similar (ns); all other groups are different using ordinary ANOVA with Holm–Sidak’s correction. All comparisons and statistical analysis are provided in the figure source data.

Fig. 5.

Lysine residues in the syt1 juxtamembrane linker regulate synchronized neurotransmitter release. (A) Cumulative frequency distribution of glutamate peaks throughout imaging, analyzed from Fig. 4; these data were not normally distributed. Using the Kruskal–Wallis test with Dunn’s correction, no difference (ns) was detected between WT, WT rescue, and F349A mutant rescue. All other comparisons were significantly different with P < 0.05. Full statistics are included in the figure source data. (B) Synchronous fraction of each condition quantified. Synchronous release defined as percentage of iGluSnFR ΔF/F0 peaks within 10 ms following a single stimulus (data are from Fig. 4 C and D) from an entire field of view. Values are mean ± SEM from three independent trials. These data were normally distributed, with 8 to 17 fields of view for each group; **** denotes P value < 0.0001 between WT and labeled conditions using ANOVA with Holm–Sidak’s correction. All comparisons and statistical analysis are provided in the figure source data.

Fig. 6.

Lysine residues in the syt1 juxtamembrane linker regulate spontaneous neurotransmitter release. (A) Representative traces of mIPSCs from syt1 KO neurons (gray) or syt1 KO–expressing WT rescue (green), Juxta K mutant rescue (blue), F349 mutant rescue (orange), and Juxta K + F349A mutant rescue (magenta) syt1. (B) Quantification of mIPSC frequencies from neurons expressing the various syt1 constructs. mIPSC frequency was 2.7 ± 0.07 Hz in syt1 KO neurons (mean ± SEM; n = 17), 1.5 ± 0.10 Hz in WT syt1 rescued cells (n = 27), 2.2 ± 0.10 Hz in Juxta K rescued cells (n = 17), 1.9 ± 0.14 Hz in F349A rescued cells (n = 21), and 2.4 ± 0.0.08 Hz in Juxta K, F349A rescued cells (n = 16). (C) The quantification of mIPSC amplitudes after expression of the various syt1 constructs described in A. No statistical differences (ns) were observed. * denotes P values < 0.05; ** denotes P values < 0.01, and **** denotes P value < 0.0001 determined by ANOVA using Tukey’s multiple comparisons test. All comparisons and statistical analyses are provided in the figure source data.