Localization of three pathogenic mutations in SNAP25

A Schematic of the neuronal SNARE complex interacting with C2B domain of synaptotagmin-1 (Syt1; not to scale) via the primary interface. Position of the I67N mutation in the first SNARE domain of SNAP25 is depicted by an asterisk.

B Interaction site of the C2B domain of Syt1 and SNAP25. Syt1 interacts with SNAP25 both electrostatically (region I and II) and within the hydrophobic patch (HP patch)(Zhou et al., 2015).

C Position of the disease-linked mutations V48F (hydrophobic patch) and D166Y (region I) in the SNARE complex.

Pathogenic SNAP25 mutations compromise neuronal viability, but not synaptogenesis

A SNAP25 V48F and D166Y mutations are similarly expressed as the WT SNAP25 protein. EGFP-SNAP25 was overexpressed in neurons from CD1 (wildtype) mice; both endogenous and overexpressed SNAP25 are shown. Valosin-containing protein (VCP) was used as loading control.

B, C Quantification of EGFP-SNAP25 (B) and endogenous SNAP25 (C) from Western Blots (as in A). Displayed are the intensity of EGFP-SNAP25 or endogenous SNAP25 bands, divided by the intensity of VCP bands, normalized to the WT situation (n = 3 independent experiments). The expression level of mutants was indistinguishable from expressed WT protein (ANOVA).

D Representative images of control (WT) and mutant (V48F, D166Y) hippocampal neurons stained by dendritic (MAP2) and synaptic (vGlut1) markers. Displayed is MAP2 staining, representing the cell morphology, in inserts MAP2 staining is depicted in red and vGlut staining in cyan. The scale bar represents 50 µm.

E Number of synapses per neuron in WT and mutant cells.

F Total dendritic length of WT and mutant neurons.

G Cell viability represented as the number of neurons per glia island. ****p <0.0001, **p <0.01, *p<0.05, Brown-Forsythe ANOVA test with Dunnett’s multiple comparisons test.

–V48F and D166Y mutations increase miniature EPSC frequency

A,D,G Example traces of mEPSC release for WT, mutant and 1:1 co-expression of WT and mutant SNAP25, or (G) Syt1 WT and KO.

B, E The mEPSC frequencies were increased in both V48F and D166Y mutants and co-expressed conditions (V48F: n = 49, 47, 48 for WT, coexpressed and mutant conditions, respectively; D166Y: n = 54, 43, 50). ****p <0.0001, ***p <0.001, Brown-Forsythe ANOVA test with Dunnett’s multiple comparisons test.

C, F mEPSC amplitudes were on average increased by the V48F and D166Y mutations; this was significant for the V48F. *p <0.05, ANOVA with Dunnett’s multiple comparison test.

H, I Syt1 WT and KO data (Syt1: n = 28, 26 for the WT and KO condition). The mEPSC frequencies and amplitudes were increased and decreased in the KO, respectively. ****p <0.0001, Welch’s t-test, *p<0.05, unpaired t-test.

V48F and D166Y mutations reduce the amplitude of the eEPSC

A, E, I Example evoked excitatory post-synaptic currents (eEPSC) for WT, SNAP25 mutants and co-expressed WT/mutants, or (I) Syt1 WT and KO.

B, F, J eEPSC amplitude was decreased by both SNAP25 mutations (V48F: n = 50, 50, 45 for WT, co-expressed and mutant conditions, respectively; D166Y: n = 56, 35, 44) and by Syt1 KO (Syt1: n = 19, 26 for the WT and KO condition). SNAP25 mutations: ****p <0.0001, **p <0.01, Brown-Forsythe ANOVA test with Dunnett’s multiple comparisons test; Syt1: ****p<0.0001, Welch’s t-test.

C, G, K Overall evoked charge after a single depolarization (V48F: n = 50, 45, 50 for WT, mutant and co-expressed conditions, respectively; D166Y: 56, 44, 35; Syt1: 19, 20 for WT and KO). SNAP25: *p <0.05, Brown-Forsythe ANOVA with Dunnett’s multiple comparison test; Syt1: ****p<0.0001, Welch’s t-test.

D, H, L Fractional contribution of the synchronous release component to the overall charge (V48F: n = 50, 50, 45 for WT, co-expressed and mutant conditions, respectively; D166Y: 56, 35, 44; Syt1: 19, 20 for WT and KO). SNAP25: ****p <0.0001, ***p <0.001, Brown-Forsythe ANOVA (V48F) or standard ANOVA (D166Y) with Dunnett’s multiple comparisons test; Syt1: ****p<0.0001, Welch’s t-test.

The apparent energy barrier for vesicle fusion is lowered by V48F and D166Y, but not by removing Syt1.

A, F, K Example traces for the WT, mutant and co-expressed condition. Each cell was stimulated by 0.25 M (in grey) and 0.5 M sucrose (in black or color).

B, G, L The charge released by 0.25 M sucrose (V48F: n = 28, 30, 29 for WT, co-expressed and mutant conditions, respectively; D166Y: n = 33, 30, 35; Syt1: n = 23, 18 for WT and KO). Syt1: p<0.05, Welch’s t-test.

C, H, M The charge released by 0.5 M sucrose (V48F: n = 28, 30, 29 for WT, co-expressed and mutant conditions, respectively; D166Y: n = 33, 30, 35; Syt1: n = 23, 26 for WT and KO). SNAP25: ****p <0.0001, **p <0.01, *p<0.05, Brown-Forsythe ANOVA with Dunnett’s multiple comparisons test; Syt1: p=0.0548, unpaired t-test.

D, I, N The ratio of 0.25 and 0.5 M sucrose pool (V48F: n = 28, 30, 29 for WT, co-expressed and mutant conditions, respectively; D166Y: n = 33, 30, 35; Syt1: n = 23, 18 for WT and KO). SNAP25: ****p <0.0001, **p <0.01, ANOVA with Dunnett’s multiple comparisons test; Syt1: *p < 0.05, unpaired t-test.

E, J, O Release probability calculated by dividing the charge of an eEPSC with the 0.5 M sucrose pool (V48F: n = 24, 25, 22 for WT, co-expressed and mutant conditions, respectively; D166Y: n = 33, 24, 30; Syt1, n = 16, 21 for WT and KO). SNAP25: ***p <0.001, ANOVA with Dunnett’s multiple comparisons test; Syt1: ****p<0.0001, unpaired t-test.

Dissection of the RRP reduction in V48F and D166Y mutations.

A One-pool model of the Readily Releasable Pool (RRP). k1 is the rate of priming (units vesicles/s), k-1 is the rate of depriming (s-1), kf is the rate of fusion (s-1).

B Estimation of the three parameters from the response to 0.5 M sucrose and a measurement of the spontaneous release rate.

C Variance-mean analysis in 50 ms intervals during the sucrose application allows determination of the corrected baseline by back-extrapolation of a regression line to the variance of the baseline.

D Normalized mEPSC frequency (kf) for V48F, D166Y and Syt1 KO. (V48F: n = 23, 24 for WT and mutant conditions, respectively; D166Y: n = 19, 19; Syt1: n = 23, 26). Brown-Forsythe ANOVA test with Dunnett’s multiple comparison test, testing the three mutant conditions against each other. ****p <0.0001, ***p <0.001.

E Normalized RRP size for WT and mutant conditions, with indications of the effect of the mutant-induced changes in k1, k-1, and kf on the RRP size.

Estimated parameters affecting the size of the RRP.

Displayed is mean ±SEM. Two-sample t-test or Welch’s t-test comparing mutant to WT: *p<0.05; ***p<0.001; ****p<0.0001, # non-significant (p=0.125), ¤ non-significant (p=0.210).

SNAP25 V48F and D166Y mutations change short-term plasticity towards facilitation.

A, E eEPSCs in response to 50 APs at 40 Hz recorded in 4 mM extracellular Ca2+ (V48F: 27, 17, 15 for WT, co-expressed and mutant conditions, respectively; D166Y: 27, 18, 16). Inserts: Normalized eEPSC amplitudes demonstrating facilitation of mutant conditions. ****p<0.0001; **p<0.01, Brown-Forsythe ANOVA with Dunnett’s multiple comparison test.

B, F Priming rate calculated as the slope of a linear fit to the cumulative evoked charges during the last part of stimulation (V48F: 27, 17, 15 for WT, co-expressed and mutant conditions, respectively; D166Y: 27, 18, 16). *p <0.05, ANOVA (V48F) or Brown-Forsythe ANOVA (D166Y) with Dunnett’s multiple comparisons test.

C, G RRP calculated by back-extrapolation of a linear fit to the cumulative evoked charges during the last part of stimulation (V48F: 27, 17, 15 for WT, co-expressed and mutant conditions, respectively; D166Y: 27, 18, 16). **p <0.01, *p <0.05, ANOVA (V48F) or Brown-Forsythe ANOVA (D166Y) with Dunnett’s multiple comparisons test.

D, H Release probability calculated as the charge of the first evoked response divided by the RRP obtained by back-extrapolation (V48F: 27, 17, 15 for WT, co-expressed and mutant conditions, respectively; D166Y: 27, 18, 16). ***p<0.001; **p <0.01, ANOVA (V48F) or Brown-Forsythe ANOVA (D166Y) with Dunnett’s multiple comparisons test.

Pathogenic SNAP25 mutations affect synaptotagmin-1 interaction and fusion rates in vitro.

A In the presence of SDS, SNAP25 I67N containing v-/t-SNARE complexes were more sensitive to temperature-dependent dissociation. Shown are mean ± SEM (n = 3) for SNARE-complexes including SNAP25 WT and the I67N, V48F and D166Y mutations.

B-C In vitro Syt1/VAMP2 SUVs docking to t-SNARE GUVs was significantly reduced by SNAP25 V48F, I67N and D166Y mutants either in absence (B) or presence (C) of PI(4,5)P2. Fusion was blocked by performing the assay on ice. **** p<0.0001; *** p<0.001, ANOVA with Dunnett’s multiple comparison test.

D-E In vitro lipid mixing assays of VAMP/Syt1 SUVs with t-SNARE GUVs containing SNAP25 V48F, I67N, or D166Y mutants showed impaired membrane fusion in the absence (left) or presence (right) of complexin-II. Fusion clamping in the presence of complexin was selectively reduced by V48F and D166Y. Bar diagrams show lipid mixing just before (pre) and after (post) Ca2+ addition and at the end of the reaction. Shown is mean ± SEM (n = 3). ****p<0.0001; **p <0.01, ANOVA with Dunnett’s multiple comparisons test, comparing each mutation to the corresponding WT condition.

The D166Y mutation increases binding to its SNARE partners.

A-C In vitro lipid mixing assays of VAMP/Syt1 SUVs with syntaxin-1A GUVs in the presence of soluble SNAP25. V48F, and D166Y mutants showed impaired fusion clamping in the absence (left) or presence (right) of complexin-II; I67N (red) showed impaired Ca2+-independent and Ca2+-triggered fusion. Bar diagrams show lipid mixing just before (pre) and after (post) Ca2+ addition and at the end of the reaction. Mean ± SEM (n = 3). ****p<0.0001; ***p<0.001; **p <0.01; *p <0.05, ANOVA with Dunnett’s multiple comparisons test, comparing each mutation to the corresponding WT condition.

D SNAP25 D166Y showed enhanced interactions with SUVs carrying reconstituted syntaxin-1A (Stx-1), VAMP2, Syt1/VAMP2 or a SUV-mixture containing Syntaxin-1A and VAMP2/Syt1 in co-flotation assays, whereas V48F displayed weaker increases in interactions with SUVs containing syntaxin-1A, or Syt1/VAMP2. Shown is mean ± SEM on a logarithmic scale. *** p<0.001, ** p<0.01, * p<0.05, two-tailed one-sample t-test comparing to 1.

The I67N mutation inhibits spontaneous and evoked release.

A SNAP25 I67N is similarly expressed as the WT SNAP25 protein. EGFP-SNAP25 was overexpressed in neurons from CD1 (wildtype) mice; both endogenous and overexpressed SNAP25 are shown. Valosin-containing protein (VCP) was used as the loading control.

B, C Quantification of EGFP-SNAP25 (B) and endogenous SNAP25 (C) from Western Blots (as in A). Displayed are the intensity of EGFP-SNAP25 or endogenous SNAP25 bands, divided by the intensity of VCP bands, normalized to the WT situation (n = 3 independent experiments). The expression level of the I67N mutant was indistinguishable from WT protein (ANOVA).

D Cell viability represented as the number of neurons per glial islet. ****p <0.0001, Brown-Forsythe ANOVA test with Dunnett’s multiple comparisons test.

E Representative image of mutant (I67N) hippocampal neurons stained for the dendritic marker MAP2 and the synaptic markers vGlut1. Displayed is MAP2 staining, representing the cell morphology, in inserts MAP2 staining is depicted in red and vGlut staining in cyan. The scale bar represents 50 µm.

F Number of synapses per neuron in WT and mutant cells. The WT data are the same as in Fig. 2D-E because these experiments were carried out in parallel. The difference was tested using ANOVA between all conditions, which was non-significant.

G Total dendritic length of WT and mutant neurons.

H Example traces of mEPSC release for WT, mutant (I67N) and 1:1 co-expression of WT and SNAP25 mutant.

I The mini frequency was decreased in both I67N mutant and the WT+I67N combination (I67N: n = 39, 36, 30 for WT, coexpressed and mutant). ****p <0.0001, ***p <0.001, Kruskal-Wallis with Dunn’s multiple comparisons.

J mEPSC amplitudes were unchanged by the I67N mutation.

K Example evoked excitatory post-synaptic currents (eEPSC) for WT, mutant (I67N) and co-expressed WT and mutant.

L eEPSC amplitude was decreased by the I67N mutations (I67N: n = 39, 37, 30 for WT, co-expressed and mutant conditions, respectively). SNAP25 mutations: ****p <0.0001, **p <0.01, Brown-Forsythe ANOVA test with Dunnett’s multiple comparisons test.

M Overall evoked charge after a single depolarization (I67N: 24, 10, 0 for WT, co-expressed and mutant conditions, respectively).

N Fractional contribution of the synchronous release component to the overall charge (I67N: 24, 10, 0 for WT, co-expressed and mutant conditions, respectively).

The I67N mutation has normal RRP size, but increased energy barrier for fusion.

A, E Example traces for the WT, mutant and co-expressed condition. Each cell was stimulated by 0.25 M (A, in grey) and 0.5 M sucrose (A, in color) or 0.375 M sucrose (E, in grey) and 0.75 M (E, in color).

B, F The charge released by 0.25 M sucrose (B, I67N: n = 21, 15, 8 for WT, co-expressed and mutant conditions, respectively) or 0.375 M sucrose (F, I67N: n = 12, 16, 18; a few cells were stimulated with 0.35 M sucrose – shown with open symbols). B, **p<0.01; *p<0.05, Kruskal-Wallis test with Dunn’s multiple comparison test; F, p=0.0339 Brown Forsythe ANOVA test; Dunnett’s multiple comparison test, p=0.0571.

C, G The charge released by 0.5 M sucrose (C, I67N: n = 21, 15, 8 for WT, co-expressed and mutant conditions, respectively), or 0.75 M sucrose (G, I67N: n = 13, 16, 18). C, **p<0.01, *p<0.05, Brown-Forsythe ANOVA with Dunnett’s multiple comparisons test.

D, H The ratio of the 0.25 M and 0.5 M sucrose pool (D, I67N: n = 21, 15, 8 for WT, coexpressed and mutant conditions, respectively), or the ratio of 0.375 and 0.75 M sucrose pool (H, n = 13, 16, 18). D, * p<0.05, Kruskal-Wallis test with Dunn’s multiple comparisons test. H, **** p<0.0001; *** p<0.001, ANOVA with Dunnett’s multiple comparisons test.

I eEPSCs in response to 50 APs at 40 Hz recorded in 2 mM extracellular Ca2+ (I67N: n = 23, 16, 20 for WT, co-expressed and mutant conditions, respectively). Inserts: Normalized eEPSC amplitudes demonstrating facilitation of mutant conditions.

J Normalized eEPSC amplitudes in response to 50 APs at 40 Hz recorded in 2 mM extracellular Ca2+.

K Paired pulse ratio at interstimulus interval 25 ms (I67N: n = 24, 14, 17 for WT, co-expressed and mutant conditions, respectively). *p<0.05, ANOVA with Dunnett’s multiple comparison test.

Adding positive surface charges to the SNARE complex partly compensate for the I67N mutation.

A Example traces of mEPSC release for WT, I67N/E183K/S187K/T190K/E194K (‘I67N/4K’) and E183K/S187K/T190K/E194K (‘4K’) SNAP25. Data from the 4K mutation was obtained in a separate experiment and is shown for comparison, but statistical tests with 4K mutation data were not carried out.

B The mini frequencies for the WT and I67N/4K are not significantly different; data from the 4K mutation is shown for comparison (n = 19, 25, 13 for WT, I67N/4K and 4K, respectively).

C Mini amplitudes remain unaffected by I67N/4K mutation.

D-E eEPSC examples (D) and amplitudes (E) for WT and I67N/4K; 4K is shown for comparison (n = 19, 25, 13 for WT, I67N/4K and 4K, respectively). ****p<0.0001 Mann-Whitney test.

F Electrostatic triggering model (blue line; Ruiter et al., 2019) refitted to WT spontaneous and evoked data points (black points). Fitted parameters: rate 0.00029 s-1 at zero (0) charge (Z); fraction f=0.030; the maximum rate was fixed at 6000 s-1. WT (black points), I67N, I67N/4K (red points): means of log-transformed data. The charge-values (Z, horizontal axis) for I67N and I67N/4K were found by interpolation in the model; the two spontaneous points (I67N, I67N/4K) are separated by 5.6 charges. For evoked release, rates were found by deconvolution and normalizing to RRP0.5 (Ruiter et al., 2019). The Z-values for evoked release were found by interpolation in the model; the two mutants (I67N, I67N/4K) are separated by 5.9 charges.

Energy landscapes.

The energy landscapes of WT and mutants were calculated as explained in Materials and Methods and displayed to scale. Energy landscapes for D166Y (A), V48F (B) and Syt1 KO (C) are shown at rest and are characterized by a higher priming barrier (“loss-of-function” phenotype), a destabilized RRP and a lower fusion barrier (“gain-of-function” phenotype). The I67N (D) is characterized by a higher fusion barrier (“loss-of-function” phenotype). The relative increase in the fusion barrier by the I67N mutation is higher during stimulation than at rest. Dotted lines represent energy levels for which less is known.

Kinetic parameters of evoked EPSCs.

A eEPSC (black trace), and integrated eEPSC (after multiplication with -1, red trace) with double exponential fit (green trace).

B Zoom-in of eEPSC (black trace), and integrated eEPSC (after multiplication with -1, red trace) with double exponential fit (green trace). Equations for integration and double exponential function used for fit is given.

Kinetic parameters of eEPSCs.

A, E, I Synchronous release components (A, V48F: n = 50, 50, 45 for WT, co-expressed and mutant conditions, respectively; E, D166Y: n = 56, 35, 44; I, Syt1: 19, 20 for WT and KO). A: * p < 0.05, Welch’s ANOVA with Dunnett’s multiple comparison test, E: *** p <0.001, Brown-Forsythe ANOVA with Dunnett’s multiple comparison test, I: **** p < 0.0001, Welch’s unpaired t-test.

B, F, J Asynchronous release components. B: ** p< 0.01; * p<0.05, Brown-Forsythe ANOVA with Dunnett’s multiple comparison test.

C, G, K Fast time constants. C: **** p< 0.0001; ** p<0.01, Brown-Forsythe ANOVA with Dunnett’s multiple comparison test, G: **** p< 0.0001, Brown-Forsythe ANOVA with Dunnett’s multiple comparison test, K: **** p < 0.0001, Welch’s unpaired t-test.

D, H, L Slow time constants.

Effect of sucrose stimulation on estimates of k1 and k-1.

The figure shows the effect of the fold-increase in fusion rate (N) induced by sucrose (abscissa) on the estimates of k1 (A, C) or k-1 (B, D) using Eqs. 3, 4 and 5b and the values estimated for D166Y (A, B), V48F (C, D) and WT (Table 1). Previous data showed that 0.5 M sucrose increases the fusion rate by a factor ∼5000 (Schotten et al., 2015). The plots show that the estimates in Table 1 are not strongly affected by small changes in the effect of sucrose upon kf.

Train stimulations of V48F and D166Y in 2 mM Ca2+.

A, D eEPSCs in response to 50 APs at 40 Hz recorded in 2 mM extracellular Ca2+ (V48F: 25, 18, 24 for WT, co-expressed and mutant conditions, respectively; D166Y: 23, 15, 15). Inserts: Normalized eEPSC amplitudes of first five stimulations. * p < 0.05, *** p < 0.001, one-way ANOVA with Dunnett’s multiple comparison test.

B, E Priming rate calculated by as the slope of a linear fit to the cumulative evoked charges during the last part of stimulation (V48F: 24, 18, 24 for WT, co-expressed and mutant conditions, respectively; D166Y: 23, 15, 15).

C, F RRP calculated by back-extrapolation of a linear fit to the cumulative evoked charges during the last part of stimulation (V48F: 24, 18, 24 for WT, co-expressed and mutant conditions, respectively; D166Y: 23, 15, 15).

Cumulative charges of V48F and D166Y trains in 4 mM Ca2+.

A, B Cumulative charges obtained by integrating eEPSCs during 40 Hz trains. The slope of the linear part of the curve reports on the priming rate, which is reduced by the mutations. The back extrapolation of the linear fit to zero time reports on the RRPev, the part of the RRP which APs draw on, which is also reduced by mutation (V48F: n= 27, 17, 15 for WT, co-expressed and mutant conditions, respectively; D166Y: 27, 18, 16).

Floatation assay.

Example Coomassie and silver stained gels demonstrating binding of SNAP25 WT and mutants to different populations of SUVs: Syntaxin-1 (Stx-1) and VAMP2/Syt1, Syntaxin-1 (Stx-1), VAMP2/Syt1, VAMP2, or Syt1 SUVs. Note increased binding of D166Y and V48F to most SUV populations, strongest for D166Y (quantified data in Fig. 9D).

Molecular dynamics simulations of mutants.

A Alignment of helices across the three systems (WT, V48F, and D166Y), reveals close correspondence. The structures displayed represent the most prevalent configurations from the dominant cluster observed during simulations.

B Stability evaluation (RMSD) of the two helices across the three systems relative to WT’s average structure during their simulations.

C A detailed view of the region displaying residue pairs 48-52 and 162-166 on the structures. D, E Computed electrostatic (Coulomb) and van der Waals (Lennard-Jonson, LJ) interactions for residue pairs 48-52 (D) and 162-166 (E) calculated in 200 ns blocks within the 800 ns trajectory (see Materials and Methods). The bar plots represent the means calculated using the block averaging method, while each block’s average is depicted as a dot alongside. The error bars capture the standard error of the mean, premised on treating each block as an independent measure (i.e., n=4). Notably, for D166Y (panel E, blue bar), the interaction energy is considerably more negative, indicating a stronger interaction compared to WT. * p < 0.05, ** p < 0.01, **** p < 0.0001, one-way ANOVA with Tukey HSD post-hoc tests.