Tau forms synaptic nano-biomolecular condensates controlling the dynamic clustering of recycling synaptic vesicles

Abstract

Neuronal communication relies on the release of neurotransmitters from various populations of synaptic vesicles. Despite displaying vastly different release probabilities and mobilities, the reserve and recycling pool of vesicles co-exist within a single cluster suggesting that small synaptic biomolecular condensates could regulate their nanoscale distribution. Here, we performed a large-scale activity-dependent phosphoproteome analysis of hippocampal neurons in vitro and identified Tau as a highly phosphorylated and disordered candidate protein. Single-molecule super-resolution microscopy revealed that Tau undergoes liquid-liquid phase separation to generate presynaptic nanoclusters whose density and number are regulated by activity. This activity-dependent diffusion process allows Tau to translocate into the presynapse where it forms biomolecular condensates, to selectively control the mobility of recycling vesicles. Tau, therefore, forms presynaptic nano-biomolecular condensates that regulate the nanoscale organization of synaptic vesicles in an activity-dependent manner.

Fig. 1. The recycling pool of SVs exhibit higher mobility than the total pool.

a SdTIM of VAMP2-pHluorin-bound At647N-GBP nanobodies in SVs, indicative of recycling pool SV mobility. (i) Epifluorescence image of a neuronal segment expressing VAMP2-pHluorin acquired before incubation with At647N-GBP. Inset (red outline) highlights a presynaptic compartment, shown at higher magnification in (ii). (iii) Fluorescence intensity, (iv) diffusion coefficient (the color bar represents log₁₀[μm²s⁻¹]) and (v) trajectory maps of recycling SVs. b SptPALM of vGLUT1-mEos2-containing vesicles, indicative of the total pool SV mobility. (i) Epifluorescence image of a neuronal segment expressing vGLUT1-mEos2. Inset (red outline) highlights a presynaptic compartment, shown at higher magnification in (ii). (iii) Fluorescence intensity, (iv) diffusion coefficient (the color bar represents log₁₀[μm²s^-1]) and (v) trajectory maps of total SVs. c, Average MSD of VAMP2-pHluorin/At647N-GBP trajectories (Recycling pool; black), and vGLUT1-mEos2 (Total pool; red) as a function of time. d Area under the MSD curve (AUC; µm² s). e Frequency distribution of the diffusion coefficients [D] shown in a semi-log plot. Grey dashed line indicates the threshold used to distinguish the immobile (Log₁₀[D] ≤ −1.6) from the mobile (Log₁₀[D] > −1.6) fraction of molecules. f Plot of the mobile fraction of molecules. g Three-state model of diffusive states inferred by vbSPT analysis of VAMP2-pHluorin/At647N-GBP trajectories. h, Three-state model of diffusive states inferred by vbSPT analysis of vGLUT1-mEos2 trajectories. Circles in (g) and (h) represent diffusive states, where (D) is the diffusion coefficient. Immobile state (State 1), confined state (State 2) and highly mobile state (State 3). The areas of the circles represent the average state occupancy (%) of SVs in their respective states. The arrows indicate the transition probability of an SV moving from one state to the other. i Example of a trajectory from a recycling SV undergoing stochastic switching between the three diffusive states inferred by vbSPT analysis. Data in (c–f) are displayed as mean ± SEM. Values were obtained from n = 15 neurons (Recycling pool), and n = 16 neurons (Total pool), from over 3 independent neuronal cultures. 131 presynapses were analyzed in (g) and 35 presynapses were analyzed in (h). Statistical comparisons in (d, f) were performed using unpaired two-tailed Student’s t-test with Welch’s correction. Source data are provided as a Source Data file.

Fig. 2. Phosphoproteome of stimulated mature hippocampal neurons.

Mouse hippocampal neurons were stimulated (high K⁺) or resting (low K⁺) for 5 min and the phosphoproteome was analyzed by mass spectrometry. a SynGO (biological processes) analysis of the top 500 differentially phosphorylated proteins. b Volcano plot of all phosphopeptides identified in stimulated versus resting mouse hippocampal neurons. X-axis is the log₂-ratio of high versus low phosphopeptide intensities and the Y-axis is the -log₁₀ transformed adjusted p-value for the log₂-ratios. Synaptosome phosphosites identified following global alignment of mouse and rat phosphosites is also superimposed (light blue indicates all mouse phosphopeptides; grey indicates phosphopeptides also identified in Rat; red indicates conserved Tau phosphopeptides; black indicates Tau sites not identified in rat). c Each presynaptic protein is plotted based on the percentage of disordered sequence versus the maximum change in upregulated phosphorylation following stimulation. Tau (red) was selected as a candidate for further investigation. d Log₂ fold change for phosphorylation sites from known presynaptic activity-dependent phosphoproteins, Tau, synapsin 1 (Sys) and dynamin 1 (Dnm1). Tau conserved residues in the human 2N4R isoform are also specified. The mass spectrometry proteomics data and MaxQuant output have been deposited to PRIDE (PXD020232 and 10.6019/PXD020232). The adjusted p-value in (b) is the result of a moderated two-tailed Student’s t-test corrected for multiple hypotheses using the Benjamini and Hochberg method (see Methods and Supplementary Data 1 for additional information). Source data are provided as a Source Data file.

Fig. 3. Tau controls the mobility of the recycling pool of SVs.

a, b (i) Epifluorescence image of VAMP2-pHluorin expressed in WT neurons (a) or Tau KO neurons (b) incubated with At647N-GBP nanobodies. Insets (red outline) on a presynapses are shown at higher magnification in (ii). (iii) Intensity maps of recycling SVs. c, e Average MSDs of recycling SVs (c) or total SVs (e) from WT (black) or Tau KO (red) neurons. d, f Plots of the area under the MSD curves (AUC). g (i) Epifluorescence image VAMP2-pHluorin expressed in Tau KO neurons incubated with At647N-GBP nanobodies. Empty mEos2 was expressed as a control. Inset (red outline) on a presynapse is shown at higher magnification in (ii) (trajectory map of SVs). h (i) Epifluorescence image of VAMP2-pHluorin expressed in Tau KO neurons where Tau-mEos2 was re-expressed, and incubated with At647N-GBP nanobodies. Inset (red outline) on a presynapse is shown at higher magnification in (ii) (trajectory map of SVs). i Average MSD of recycling SVs from neurons expressing mEos2 (black) or Tau-mEos2 (red). j The corresponding area under the MSD curves (AUC). k Average MSD of recycling SVs from Tau WT neurons treated with control vehicle (Veh, DMSO; black), okadaic acid (1 μM, 15 min, OkAc; red), or staurosporine (1 μM, 15 min, Staur; blue). l The corresponding area under the MSD curves (AUC). m Average MSD of recycling SVs from Tau KO neurons treated with control vehicle (Veh, DMSO; black), okadaic acid (OkAc; red), or staurosporine (Staur; blue). n The corresponding area under the MSD curves (AUC). Data are displayed as mean ± SEM. Values were obtained from n = 13 and 16 neurons in (c, d), n = 17 and 18 neurons in (e, f), n = 21 and 31 neurons in (i, j), n = 16, 10 and 10 neurons in (k, l), and n = 11, 12 and 18 neurons in (m, n). Data was obtained from >2 independent neuronal cultures. Statistical comparisons were performed using the unpaired two-tailed Student’s t-test in (d) and (f) and the unpaired two-tailed Mann-Whitney U test in (j), and using the one-way ANOVA test followed by Dunnett post hoc test comparing the groups to the control (Veh) in (l) and (n). Source data are provided as a Source Data file.

Fig. 4. Tau forms differently regulated presynaptic and axonal nanoclusters.

a Representative Tau-mEos2 trajectories generated using segNASTIC showing nanoclusters at the presynapse and in the axonal compartment. Color-coding of the clusters represents their appearance in time across the acquisition (16,000 frames, 320 s). The insets (red outlines) of the (i) axonal or (ii) presynaptic compartment are shown at a higher magnification. b (i) 3D (X, Y, Time) plot of Tau-mEos2 trajectories from the axonal compartment. (ii) 3D (X, Y, Time) plot of Tau-mEos2 trajectories from the presynaptic compartment. Squares in X represent 200 nm; squares in Y represent 100 nm in (i) or 200 nm in (ii); squares in Time represent 50 s. c–h Comparison of cluster metrics from the axonal (Ax) and the presynaptic (Ps) compartments. c Average MSD of clustered Tau-mEos2 trajectories represented as the area under the curve (AUC). d Average density of Tau-mEos2 clusters per μm². e Average cluster membership (trajectories/cluster). f Average apparent lifetime of Tau-mEos2 clusters. g Average Tau-mEos2 cluster area. h Average density of Tau-mEos2 trajectories per cluster area. Data are displayed as mean ± SEM. Values were obtained from n = 17 synapses in (c) or 13 synapses in (d) to (h), and n = 15 axons in (c) or 37 axons in (d) to (h). Data was obtained from ≥2 independent neuronal cultures. Statistical comparisons were performed using the unpaired two-tailed Student’s t-test with Welch’s correction. Source data are provided as a Source Data file.

Fig. 5. Tau mobility is sensitive to 1,6-HD.

a, b Representative Tau-mEos2 trajectories generated using segNASTIC showing instantaneous diffusion coefficients (the color bar represents log₁₀[μm²s^-1]) in an (a) synaptic region and (b) an axonal region, determined by VAMP2-pHluorin intensity (inset), of neurons treated with control vehicle (DMSO) or with 1,6-HD (2.5%). c Average MSD of Tau-mEos2 presynaptic trajectories from neurons treated with control vehicle or with 1,6-HD. d Corresponding areas under the MSD curves (AUC). e Average MSD of Tau-mEos2 axonal trajectories from neurons treated with control vehicle or with 1,6-HD. f Corresponding areas under the MSD curves (AUC). g Representative segNASTIC trajectories showing instantaneous diffusion coefficients from Tau-mEos2, and Tau LLPS mutants Tau^{I277P / I308P}-mEos2 and ΔTau74-mEos2. h, j Average MSD of the trajectories of mEos2-tagged Tau WT and LLPS Tau mutants in (h) presynapses and (j) axons. i, k Corresponding areas under the curve (AUC), in the (i) presynapses and (k) axonal compartment. Data in (c) to (f) and (h) to (k) are displayed as mean ± SEM. Data in (c, d) was obtained from n = 9 synapses from 9 different neurons. Data in (e, f) was obtained from n = 7 and 5 axons from different neurons. Data in (h) to (k) was obtained from n = 14 presynapses from 14 neurons and 18 axons from 18 neurons in Tau^WT, n = 11 synapses and axons from 11 neurons in Tau^{I277P / I308P}, and n = 17 synapses from 17 neurons and 12 axons from 12 neurons in ΔTau74. Data was obtained from 2 independent neuronal cultures. Statistical comparisons in (d) and (f) were performed using the unpaired two-tailed Student’s t-test with Welch’s correction in (d) and (f), or using the one-way ANOVA test in (i) and (k). Source data are provided as a Source Data file.

Fig. 6. Tau nanoscale organization is regulated by synaptic activity.

a (i) Representative epifluorescence image of a WT hippocampal neuronal segment expressing VAMP2-pHluorin and Tau-mEos2, used for imaging under resting conditions (Low K⁺) and after stimulation (High K⁺). (ii) The inset of the presynaptic compartment in resting condition is shown at a higher magnification. (iii) Trajectory map of Tau-mEos2 in the unstimulated presynaptic compartment. (iv) The inset of the presynaptic compartment during stimulation is shown at a higher magnification. (v) Trajectory map of Tau-mEos2 in the stimulated presynaptic compartment. b Plot showing the total number of Tau-mEos2 trajectories within the presynapse, in resting conditions (black) and after stimulation (red). c Plot showing the total number of Tau-mEos2 trajectories within the axon, in resting conditions (black) and after stimulation (red). d–g Average MSD of total Tau-mEos2 trajectories and the corresponding area under the curve (AUC) in resting conditions (black) and after stimulation (red), in the presynapse (d, e) and in the axonal compartment (f, g). h Trajectory map of (i) recycling SVs and (ii) Tau-mEos2 in the unstimulated presynaptic compartment, obtained by dual-color single-tracking super-resolution imaging. (iii) Merged image. i Trajectory map of (i) recycling SVs and (ii) Tau-mEos2 in the stimulated presynaptic compartment, obtained by dual-color single-tracking super-resolution imaging. (iii) Merged image. j, k Representative Tau-mEos2 trajectories generated using segNASTIC showing instantaneous diffusion coefficients (the color bar represents log₁₀[μm²s^-1]) in a presynaptic region, determined by VAMP2-pHluorin intensity (inset), of a neuron that was unstimulated (j) and subsequently stimulated (k). l Percentage of Tau-mEos2 trajectories that are clustered. m Average cluster density of Tau-mEos2. n Average MSD of clustered Tau-mEos2 trajectories represented as the area under the curves (AUC). o Average apparent lifetime of Tau-mEos2 clusters. Data are displayed as mean ± SEM. Values were obtained from n = 29 synapses in (b, e), n = 23 axons in (c, g), and n = 15 synapses in (l, m, n, o). Data was obtained from ≥2 independent neuronal cultures. Statistical comparisons were performed using the two-tailed Student’s paired t-test. Source data are provided as a Source Data file.

Fig. 7. Graphical representation of Tau nanoscale organization in axons and nerve terminals and its role in controlling the clustering the recycling pool of SVs.

a, b Graphical representation of SVs in WT (a) and Tau KO (b) nerve terminal neurons. Tau KO impacts the nanoscale organization of recycling SVs within the presynaptic bouton, resulting in an increase in their mobility. c, d Graphical representation of Tau nano-biomolecular condensates in nerve terminals in resting (a) and stimulated (b) conditions. Under resting conditions, the axonal segment contains larger Tau nanoclusters, whereas synaptic boutons harbor a more dense population of smaller Tau nano-biomolecular condensates. Neuronal stimulation increases the number of Tau molecules detected at the presynapse and promotes a re-organization of Tau nano-biomolecular condensates, leading to more numerous, smaller Tau nano-biomolecular condensates. Created with BioRender.com.