Serine-129 phosphorylation of α-synuclein is an activity-dependent trigger for physiologic protein-protein interactions and synaptic function

Abstract

Phosphorylation of α-synuclein at the serine-129 site (α-syn Ser129P) is an established pathologic hallmark of synucleinopathies and a therapeutic target. In physiologic states, only a fraction of α-syn is phosphorylated at this site, and most studies have focused on the pathologic roles of this post-translational modification. We found that unlike wild-type (WT) α-syn, which is widely expressed throughout the brain, the overall pattern of α-syn Ser129P is restricted, suggesting intrinsic regulation. Surprisingly, preventing Ser129P blocked activity-dependent synaptic attenuation by α-syn-thought to reflect its normal function. Exploring mechanisms, we found that neuronal activity augments Ser129P, which is a trigger for protein-protein interactions that are necessary for mediating α-syn function at the synapse. AlphaFold2-driven modeling and membrane-binding simulations suggest a scenario where Ser129P induces conformational changes that facilitate interactions with binding partners. Our experiments offer a new conceptual platform for investigating the role of Ser129 in synucleinopathies, with implications for drug development.

Keywords: AlphaFold; CRISPR; alanine mutations; endogenous tagging; neuronal activity; pHluorin; proximity-ligation assay; serine phosphorylation; synapses; α-synuclein.

Figure 1:. Localization of α-syn Ser129P and effects of phosphorylation on synaptic targeting.

(A) Staining for α-syn Ser29P (left) or total α-syn (right) in mouse brains (representative). Unlike broad expression of total α-syn, Ser129P α-syn staining is restricted to subsets of brain regions. Section in A’ was pre-treated with phosphatases, which eliminated the Ser129P staining. See Supp. Fig. 1A–B for further characterization. Scale bar = 2 mm. (B) Hippocampus section: Note α-syn Ser129P staining in both synapses and nuclei – white arrowheads mark some pSer129+ nuclei (B’ zoomed image from B). Scale bar = 200 μm. (C) Cultured hippocampal neurons stained with antibodies to α-syn Ser129P and VAMP2 (to mark synapses). Note α-syn Ser129P is localized to nuclei and synapses (also see Supp. Fig. 1C). Scale bar = 10 μm. Quantitative analyses (see methods) showed that 50.95 +/− 3.32% (mean +/− SD) VAMP2-positive puncta were colocalized with α-syn Ser129P. Several Ser129P α-syn puncta were immediately adjacent to synapses or in processes resembling axons. (D) Formula for evaluating presynaptic targeting of transfected h-α-syn in α-syn −/− neurons. (E) Top: Serine phosphorylation sites in α-syn. Bottom: Representative images from synaptic targeting assays showing increased targeting of phospho-mimic (129D) α-syn. Scale bar = 10 μm. (F) Quantification of synaptic targeting for phospho-incompetent (129A) and phospho-mimic (129D) mutants (also see Supp. Fig. 2). Note targeting was substantially increased by mimicking phosphorylation at the 129-site, and this was independent of 87A phosphorylation status (~ 500–1000 boutons from 6–15 coverslips were analyzed for each condition; **p<0.001, one-way ANOVA followed by Kruskal-Wallis test). (G) Synaptic targeting of h-α-syn phospho-mutants relative to WT h-α-syn (horizontal dashed line marks targeting of WT h-α-syn). Note decreased presynaptic targeting of the familial PD mutant A30P and TsixK (mutant that disrupts membrane-binding) was augmented by the 129D (but not 129A) mutant, and as expected, the N-terminus was required for targeting (~ 600–900 boutons from 8–20 coverslips were analyzed for each condition; **p<0.001, Mann-Whitney nonparametric test). (H) Knock-in strategy to replace the last two exons (exon 5 and exon 6) and insert a fluorescent tag (oScarlet) at the C-terminus of endogenous mouse α-syn. Note that the A/D phospho-mutants can be inserted into exon 5 using this strategy (also see Supp. Fig. 3A–D and methods). (I) Representative image of cultured neurons expressing the knockin construct (Ser at 129-residue, more examples in Supp. Fig. 3E). (J) Exemplary images from cultured neurons expressing knock-in WT and S129A/D tagged to mScarlet, synaptic enrichment quantified in (K). Note augmented targeting of S129D α-syn (mean ± SEM ***p<0.0001 Mann-Whitney nonparametric test).

Figure 2:. Ser129P is required for α-syn-induced synaptic attenuation in pHluorin assays.

(A) Schematic of the pHluorin assay with experimental design on right. (B) Representative pre/post-stimulation images from visualizing vGLUT:pHl fluorescence in cultured mouse hippocampal neurons. Elapsed time in seconds is shown on the upper left (also see Supp. Movie 1). Scale bar = 5 μm. (C) Fluorescence fluctuations of vGLUT:pHl in presence of WT and phospho-mutant α-syn. WT-α-syn attenuates SV-recycling in pHluorin assays as previously reported, and Ser129 phospho-mimic mutant (S129D) further augments this suppression. (D) Preventing α-syn Ser129P (S129A) abolished the α-syn-induced synaptic attenuation. (E) Quantification of the recycling pool size in all pHluorin experiments (data from 12–15 coverslips, 30–50 boutons per coverslips and 4 independent cultures for each condition; ***p<0.0001 one-way ANOVA followed by a nonparametric Kruskal-Wallis; ns = not significant).

Figure 3:. Neuronal activity augments Ser129 phosphorylation.

(A) Experimental design to induce activity in cultured neurons by chemical (4-AP) or electrical (600AP) stimulation. (B) Western blots showing time-dependent increase in α-syn Ser129P after 4-AP treatment; quantified o right (n=6 for Ser129-P α-syn/total α-syn and n=3 for PLK2/total α-syn **p<0.001 unpaired Student’s t-test). (C) Top: Western blots showing increased Ser129P after 600AP/10Hz stimulation (resembling vGLUT:pHl assays, see Methods). Note that pre-incubated with 1 μM TTX or 50 nM PLK2 inhibitor BI2536 (putative kinase for Ser129P) for 15min blocked Ser129P (two biological replicates shown). Bottom: Quantification of the western blots. (D) Schematic for inducing neuronal activity in vivo by intraperitoneal 4-AP injections. (E) Western blots from 4-AP-injected mouse brains show increase in Ser29P α-syn without a change in total α-syn levels (three biological replicates shown); quantified on the right (n=3 mice, all data are means ± SEM - *p < 0.01, unpaired Student’s t-test). (F) Spatial distribution of increased Ser129P in 4-AP injected mouse brains, shown in representative sections stained with the Ser129P α-syn antibody (zoomed insets show hippocampal staining). Bar graph on the right shows quantification of intensities sampled from the cortex and hippocampus (*p<0.05, unpaired Student’s t-test, see methods for details).

Figure 4:. Ser129P regulates association of α-syn with functional binding partners VAMP2 and synapsin.

(A) Workflow for co-immunoprecipitation experiments in neuro2a cells. (B, C) Western blots from co-immunoprecipitation experiments show that both VAMP2 (B) and synapsin (C) co-immunoprecipitated with WT-h-α-syn but not phospho-incompetent (129A) h-α-syn. Mimicking Ser129P (S129D) augmented this interaction – quantified on the right (n=4 for myc α-syn/GL-VAMP2 and n=3 for myc α-syn/GL-Synapsin co-IP). All western-blot data quantified in (D, mean ± SEM ***p<0.0001, unpaired Student’s t-test). (E) Principle of our PLA assay and experimental design (neuron cartoon courtesy of Christophe Leterrier, Marseille). Note that a fluorescent signal is expected if transduced h-α-syn:mScarlet (WT or 129A/D) and endogenous mouse VAMP2 are <40 nm apart. (F) Representative images of fluorescent mScarlet and PLA signals. Note that neurons transduced with WT h-α-syn show punctate PLA signal at synapses (left column), while essentially no signal is seen with phospho-incompetent (S129A) h-α-syn (middle column). Increased PLA-signal is seen with S129D h-α-syn (right column); all data quantified in (G, mean ± SEM ****p<0.0001 one-way ANOVA followed by a nonparametric Kruskal-Wallis test). (H) Workflow for pulldown of GST-tagged WT/129A/129D h-α-syn after incubation with mouse brain lysates. Equivalent amounts of immobilized GST h-α-syn (or its phospho variants) were used. (I) Samples from GST-pulldown were analyzed by NuPAGE and immunoblotted with antibodies against VAMP2 (top panel) and synapsin (middle panel); two biological replicates are shown. Ponceau staining (bottom panel) shows equivalent loading of fusion proteins. Note that preventing Ser129P (S129A) blocked interaction with VAMP2 and synapsin; blots quantified below (n=4, mean ± SEM *p<0.01,**p<0.001,***p<0.0001, unpaired Student’s t-test). Also note that the GST-h-α-syn in our experiments appears to be innately phosphorylated at Ser129 (see results and further characterization in Supp. Fig. 5).

Figure 5:. Activity-dependent Ser129P triggers α-syn association with VAMP2 and synapsin in vivo.

(A) Workflow for on-bead dephosphorylation and phosphorylation. GST-tagged WT-h-α-syn (or S129A/D) were pretreated with alkaline phosphatase for 3h before pulldown, or with recombinant PLK2 overnight at 30 °C (see methods for details). (B, C) Samples from GST-pulldowns were analyzed by NuPAGE and immunoblotted with antibodies against α-syn Ser129P, VAMP2, and synapsin. Ponceau staining (bottom panels) show equivalent loading of fusion proteins. Note that α-syn dephosphorylation attenuates VAMP2 and synapsin interaction with α-syn, while Ser129P by PLK2 augments these associations (note that PLK2-induced phosphorylation leads to a substantial increase in phosphorylation that saturated the band marked with an asterisk). Quantification (n=3) of all pulldown experiments in bottom (B – mean ± SEM **p<0.001,***p<0.0001, unpaired Student’s t-test) and (D – mean ± SEM **p<0.001, unpaired Student’s t-test). (E) Workflow for chemically inducing activity (4-AP injections), followed by co-immunoprecipitation.. (F) Note that after 4-AP injections, a relatively higher fraction of Ser129P α-syn co-immunoprecipitated with VAMP2 (left) and synapsin (right), indicating that neuronal activity augmented these associations in vivo. (G) Quantification of western blots (n=4 and n=5 for Ser129P/VAMP2 and Ser129P/synapsin experiments, respectively). All data presented as mean ± SEM (*p < 0.01, **p < 0.001, ***p < 0.0001, t-test).

Figure 6:. Ser129P Is a trigger for protein-protein interactions.

(A) Interactome of synapse-related proteins in mouse brains preferentially associating with WT or phospho-mimic (129D) h-α-syn, compared to phospho-incompetent (129A) h-α-syn (see results for experimental details). GST-tagged h-α-syn (WT/129A/12D) was incubated with mouse brain lysates, and bound proteins were detected by mass-spectrometry. Note that a host of synaptic proteins preferentially associated with S129D α-syn. (B, C) List of some synaptic proteins preferentially associating with S129 and WT h-α-syn, compared to S129A. (D) Western-blot validation of selected candidates from the mass-spectrometry experiments above. Ponceau staining (bottom panel) shows equivalent loading of fusion proteins.

Figure 7:. Ser129P α-syn binds to a subset of synaptic vesicles and facilitates vesicle clustering.

(A) Workflow for pulldown of C-terminus of h-α-syn (residues 96–140) tagged to GST(WT/129A/129D) after incubation with intact SV-fractions (LP2) from mouse brain lysates. (B) Schematic for isolating SVs; see fractionation protocol in Supp. Fig. 5B and methods. (C) Western blots from the pellet (LP2, 15 μg) and supernatant (S4) were analyzed by western blots. Note that mouse α-syn is not present in these enriched fractions, eliminating potential confounding factors. (D) Western blots from GST h-α-syn C-terminus pulldown experiments. Samples were analyzed by NuPAGE and immunoblotted with antibodies against VAMP2 (top), synapsin (middle), and SV2 (bottom). Note that SVs robustly associate with WT (innately phosphorylated) or S129D α-syn 96–140. However, the binding of SVs enriched in VAMP2, synapsin, and the ubiquitous glycoprotein SV2 to the α-syn C-terminus was attenuated when Ser129 was rendered phospho-incompetent (S129A). These findings suggest that Ser129P triggers association of α-syn with a subset of SVs enriched in VAMP2 and synapsin (also see results and discussion). (E) Quantification of blots: VAMP2 (n=4), synapsin (n=4), and SV2 (n=4). All data presented as mean ± SEM (*p < 0.01,**p < 0.001, unpaired Student’s t test). (F) Top: Schematic of experimental design to evaluate WT or S129A/D h-α-syn effects on SV clustering. Bottom: Representative electron microscopy images of synapses. Note that while WT h-α-syn over-expression increased SV clustering, phospho-incompetent (S129A) h-α-syn decreased clustering, while phospho-mimic S129D h-α-syn further augmented SV clustering. Inter-vesicular distances are quantified in (G – mean ± SEM ***p<0.001 one-way ANOVA followed by a nonparametric Kruskal-Wallis test).

Figure 8:. Ser129-P mediated interactions may stabilize α-syn C-terminus and facilitate protein-protein interactions.

(A) Workflow for generating α-syn structures and molecular simulations. Illustration on the right shows overall principle of the membrane-binding simulations (see results and methods for more details). (B) Schematic showing boundary-definitions of membrane leaflets and protein residue depths used to calculate membrane penetrance of α-syn; viewed from the side. Note that the depths to which the alpha-helices can embed the membrane can be precisely calculated. (C) Results from membrane-binding simulations. Top: Near-final snapshots from simulations showing depth profiles of membrane-embedded unphosphorylated (left) phosphorylated (right) states of α-syn. Bottom: Quantitative residue-wise depth-profile (mean +/− SEM of each residue) shows membrane-embedding in the two states. Note that the C-terminus (magenta) appears folded in the phosphorylated state due to intramolecular electrostatic interactions. (D) Representative snapshots from early (0–50 ns) and late (150–200 ns) stages of the simulation show that in the phosphorylated state, five positively charged lysine residues within the 95–140 residues (60K, 80K, 96K, 97K, and 102K – marked in yellow) crowd around the Ser129-site (marked by an arrow in the zoomed images below). The distances between these five lysine residues and the Ser129-site across the simulations is quantified in (E). Note the relatively greater proximity of the five lysines to Ser129 in the phosphorylated state. (F) Schematic depicts a possible scenario where electrostatic interactions between the negatively charged S129-site and positive lysines along the C-terminus expose the α-syn protein-protein interaction site (residues 96–140), allowing functional associations with other synaptic proteins. Dashed line represents upper membrane leaflet.