Actin dynamics and actin-nucleating mDia1/3 proteins facilitate SV endocytosis.

(A) Averaged normalized Synaptophysin-pHluorin (Syph-pH) fluorescence traces from transfected hippocampal neurons stimulated with 200 APs (40 Hz, 5s) at physiological temperature (37.5°C). Neurons were treated with 0.1% DMSO or JLY cocktail (containing 8 µM Jasplakinolide, 5 µM Latrunculin A and 10 µM Y-27632) as indicated. Data shown represent the mean ± SEM. N = 4 independent experiments from nDMSO = 23 videos; nJLY = 36videos.

(B) Endocytic decay constants (τ) of Synaptophysin-pHluorin traces in A: τDMSO = 29.1 ± 3.4 s; τJLY = 55.8 ± 7.2 s; p < 0.05, two-tailed student’s t-test. Data shown represent mean ± SEM.

(C) Averaged normalized bleach-corrected vGAT-CypHer fluorescence traces from hippocampal neurons treated with DMSO or JLY cocktail in response to 200 AP (40 Hz, 5s) stimulation. Data shown represent the mean ± SEM. N = 4 independent experiments from nDMSO = 23 videos; nJLY = 29 videos.

(D) Endocytic decay constants of vGAT-CypHer traces in C: τDMSO = 10.9 ± 0.7 s, τJLY = 24.6 ± 2.0 s; p < 0.001, two-tailed unpaired student’s t-test. Data shown represent the mean ± SEM.

(E) Averaged normalized Synaptophysin-pHluorin fluorescence traces from hippocampal neurons transfected with shRNA-encoding plasmids against no mammalian target (shCTR) or mDia1 (shmDia1) in response to 200 AP (40 Hz, 5s) stimulation. Neurons were co-transfected with mDia1-mCherry (mDia1-WT) or mCherry alone (shCTR & shmDia1) to exclude artefacts from overexpression. Data shown represent the mean ± SEM. N = 3 independent experiments from nshCTR = 28 videos, nshmDia1 = 21 videos, nshmDia1 + mDia1-WT = 21 videos.

(F) Endocytic decay constants of Synaptophysin-pHluorin traces in E: τshCTR = 29.7 ± 1.9 s; τshmDia1 = 64.7 ± 3.9 s; τshmDia1 + mDia1-WT = 30.6 ± 3.7 s; pshCTR vs shmDia1 < 0.001, pshmDia1 vs shmDia1 + mDia1-WT < 0.001, one-way ANOVA with Tukey’s post-test. Data shown represent mean ± SEM.

(G) Endocytic decay constants of averaged normalized vGLUT1-pHluorin fluorescence traces (Figure 1 – figure supplement 1B) from hippocampal neurons transduced with shCTR (τshCTR = 9.1 ± 0.8 s), shmDia1 (τshmDia1 = 14.3 ± 1.5 s) or shmDia1+3 (τshmDia1+3 = 16.4 ± 1.3 s) in response to 40 AP (20 Hz, 2s) stimulation (pshCTR vs shmDia1 < 0.05, pshCTR vs shmDia1+3 < 0.01, one-way ANOVA with Tukey’s post-test). Data shown represent mean ± SEM. N = 4 independent experiments from nshCTR = 17 videos; nshmDia1 = 19 videos; nshmDia1+3 = 18 videos.

(H) Endocytic decay constants of averaged normalized vGLUT1-pHluorin fluorescence traces (Figure 1 – figure supplement 1C) of neurons transduced with lentiviral vectors encoding shCTR (τshCTR = 13.6 ± 1.6 s), shmDia1 (τshmDia1 = 22.0 ± 3.2 s) or shmDia1+3 (τshmDia1+3 = 26.9 ± 3.6 s) in response to 80 AP (40 Hz, 2s) stimulation (pshCTR vs shmDia1+3 < 0.05, one-way ANOVA with Tukey’s post-test). Data shown represent mean ± SEM. N = 4 independent experiments from nshCTR = 12 videos, nshmDia1 = 15 videos, nshmDia1+3 = 18 videos.

(I) Averaged normalized bleach-corrected vGAT-CypHer fluorescence traces from hippocampal neurons transduced with shCTR or shmDia1+3 in response to 200 AP (40 Hz, 5s) stimulation. Data shown represent the mean ± SEM. N = 8 independent experiments from nshCTR = 37 videos, nshmDia1+3 = 35 videos.

(J) Endocytic decay constants of vGAT-CypHer traces in I: τshCTR = 14.1 ± 1.3 s; τshmDia1+3 = 27.3 ± 2.6 s; p < 0.001, two-tailed unpaired student’s t-test. Data shown represent the mean ± SEM.

(K) Averaged normalized vGLUT1-pHluorin fluorescence traces from transduced neurons in response to 80 AP (40 Hz, 2s) stimulation. Cells were treated with 0.1% DMSO or 10 µM mDia activator (IMM) in the imaging buffer. Data shown represent mean ± SEM. N = 3 independent experiments from nDMSO = 18 videos; nIMM = 16 videos.

(L) Endocytic decay constants of vGLUT1-pHluorin traces in K: τDMSO = 14.9 ± 0.8 s; τIMM= 9.8 ± 0.5 s; p < 0.05, two-tailed unpaired student’s t-test. Data shown represent mean ± SEM.

Loss of mDia1/3 causes an activity-dependent reduction of the SV pool.

(A,B) Representative synaptic electron micrographs from hippocampal neurons transduced with lentiviruses encoding shCTR (A) or shmDia1+3 (B). Postsynapse and postsynaptic cleft are colored in green and maroon, respectively. Scale bar, 250 nm.

(C) Average number of SVs per μm2 in boutons from shCTR (92.2 ± 2.5) and shmDia1+3 (81.4 ± 2.9; p < 0.0001, Mann-Whitney test) treated neurons. Data shown represent mean ± SEM. N = 3 independent experiments from nshCTR = 326 synapses, nshmDia1+3 = 321 synapses.

(D) Average number of SVs per μm2 in synaptic boutons from hippocampal neurons transduced with lentiviruses encoding shmDia1+3 and treated with 0.1 % Vehicle (10 µM NaOAc; 83.2 ± 2.9) or 1 µM Tetrodotoxin (TTX; 93.8 ± 3.1; p < 0.01, Mann-Whitney test) for 36 h before chemical fixation. Data shown represent mean ± SEM from two independent experiments and nVehicle = 225 synapses, nTTX = 221 synapses. Representative synaptic electron micrographs are shown in figure 2 – figure supplement 1D, E. The dotted line represents the average SV numbers/ μm2 in shCTR boutons treated with DMSO from the same experiments as a reference (see figure 2 – figure supplement 1A-C).

(E,F) Representative electron micrographs of synapses in hippocampal neurons from WT (E) or mDia1 KO mice (F). Postsynapse and postsynaptic cleft are colored in green and maroon, respectively. Scale bar, 250nm.

(G) Average number of SVs per μm2 in WT (117.6 ± 5.3) and mDia1 KO (84.6 ± 5.2; p < 0.0001, Mann-Whitney test) boutons. Data shown represent the mean ± SEM from nWT = 103, nKO= 96 synapses.

mDia1 associates with endocytic proteins and localizes to presynaptic sites.

(A) Schematic representation of functional domains of mDia1. Rho-binding domain (RBD), Diaphanous inhibitory domain (DID), Dimerization domain (DD), Coiled coil domain (CC), Formin homology domain 1 (FH1), Formin homology domain 2 (FH2), Diaphanous autoinhibitory domain (DAD). The unstructured N-terminus (first 60 amino acids) contains three basic stretches and was truncated in the ΔN mutant.

(B) Endocytic decay constants of Synaptophysin-pHluorin traces (see Figure 2-supplement 1C) from hippocampal neurons transfected with shRNAmiR against no mammalian target (shCTRmiR) or mDia1 (shmDia1miR) in response to 200 AP (40 Hz, 5s) stimulation. For rescue experiments, neurons were co-transfected with plasmids encoding mDia1-WT-mCherry (τshmDia1miR + mDia1-WT = 20.0 ± 0.8 s), mDia1-ΔN-mCherry (τshmDia1miR + mDia1-ΔN = 34.5 ± 2.9 s) or mCherry alone (τshCTRmiR = 21.8 ± 1.1 s, τshmDia1miR = 30.4 ± 1.9 s) to exclude artefacts from overexpression (pshCTRmiR vs shmDia1miR < 0.05; pshmDia1miR vs shmDia1miR + mDia1-WT < 0.01; pshmDia1miR + mDia1-WT vs shmDia1miR + mDia1-ΔN < 0.01, one-way ANOVA with Tukey’s post-test). Data shown represent mean ± SEM. N = 5 independent experiments from nshCTRmiR = 41 videos, nshmDia1miR = 51 videos, nshmDia1miR + mDia1-WT = 35 videos, nshmDia1miR + mDia1-ΔN = 37 videos.

(C) Representative three-channel time-gated STED image of a synapse from hippocampal cultures treated with 0.1% DMSO for 10 min before fixation and immunostained for Bassoon (presynaptic marker, magenta), mDia1 (cyan) and Homer1 (postsynaptic marker, green). Scale bar, 250 nm.

(D) Averaged normalized line profiles for synaptic distribution of mDia1 and Homer1 relative to Bassoon (Maximum set to 0 nm). Data represent mean ± SEM. N = 3 independent experiments from n = 235 synapses.

(E) Volcano plot of proteins associating with synaptic mDia1 analyzed by label-free proteomics of anti-mDia1 versus control (CTR) immunoprecipitates from detergent-extracted mouse synaptosomes (P2’ fraction). The logarithmic ratios of protein intensities are plotted against negative logarithmic p-values derived from two-tailed student’s t-test. N = 3 independent experiments. Each dot represents one protein/gene. Selected cytoskeletal hits include: Actin, Myosin IIB (MyoIIB) and Rac1. Selected endocytic hits include Amphiphysin (p < 0.05), Dynamin1, Endophilin-A1, PACSIN1, PACSIN2 (p < 0.05) and Synaptojanin1.

(F) Endogenous immunoprecipitation of mDia1 from detergent-extracted mouse synaptosomes (P2’ fraction) using mDia1-specific antibodies. Immunoprecipitates were analyzed by immunoblotting for mDia1, Dynamin1 and β-Actin.

(G) Representative three-channel time-gated STED image of a synapse from hippocampal cultures, treated with 80 μM Dynasore for 10 min before fixation and immunostained for Bassoon (magenta), mDia1 (cyan) and Homer1 (green). Images C and G stem from the same experiment and brightness and contrast of all three channels were set to the same values. Scale bar, 250 nm.

(H) Presynaptic mDia1 levels in synapses treated with 0.1% DMSO (100 ± 7.3) or 80 µM Dynasore (145.8 ± 9.3; p = 0.0001; one sample Wilcoxon test) for 10 min in response to 200 AP (40 Hz, 5s) stimulation. Absolute line profiles of mDia1 overlapping with Bassoon (presynapse) distribution were integrated. Data shown are normalized to DMSO (set to 100) and expressed as mean ± SEM. N = 3 independent experiments from nDMSO = 92 synapses, nDynasore = 135 synapses. (I,J) Representative three-channel time-gated STED images of synapses from hippocampal cultures transduced with wildtype Dynamin1 (WT; I) or GTPase-deficient Dynamin1 (K44A; J) in response to 200 AP (40 Hz, 5s) stimulation. Cells were immunostained for Bassoon (magenta), mDia1 (cyan) and Homer1 (green). Scale bar, 250 nm.

(K) Presynaptic mDia1 levels in synapses from hippocampal neurons transduced with wildtype Dynamin1 (WT; 100 ± 6.2) or GTPase-deficient Dynamin1 (K44A; 142.9 ± 8.3, p < 0.0001, one sample Wilcoxon test) in response to 200 AP (40 Hz, 5s) stimulation. Line profiles of mDia1 overlapping with Bassoon distribution were integrated. Data shown are normalized to Dynamin1-WT (set to 100) and expressed as mean ± SEM. N = 2 independent experiments from nWT = 43 synapses, nK44A = 51 synapses.

mDia facilitates SV endocytosis by regulating presynaptic F-actin.

(A) Representative three-channel time-gated STED image of a synapse from hippocampal cultures transduced with shCTR, fixed and immunostained for Bassoon (presynaptic marker, magenta), F-Actin (cyan) and Homer1 (postsynaptic marker, green). Scale bar, 250 nm.

(B) Averaged normalized line profiles for synaptic distribution of F-Actin and Homer1 relative to Bassoon (Maximum set to 0 nm). Data are expressed as mean ± SEM. N = 4 independent experiments from n = 154 synapses.

(C) Presynaptic F-Actin levels in synapses treated with 0.1% DMSO (Figure 4 – figure supplement 1A; 100 ± 4.8) or 80 µM Dynasore (Figure 4 – figure supplement 1B; 134.7 ± 6.8; p = 0.001, one sample Wilcoxon test) for 10 min before fixation. Cells were immunostained for Bassoon (magenta), F-Actin (cyan), and Homer1 (green). Absolute line profiles of F-Actin overlapping with Bassoon (presynapse) distribution were integrated. Data shown are normalized to DMSO (set to 100) and expressed as mean ± SEM. N = 3 independent experiments from nDMSO = 207 synapses, nDynasore = 211 synapses.

(D) Endocytic decay constants of vGLUT1-pHluorin traces from hippocampal neurons transduced with lentiviral particles encoding shCTR (τshCTR = 20.7 ± 0.9 s) or shmDia1 (τshmDia1 = 26.4 ± 2.0 s) in response to 200 AP (40 Hz, 5s) stimulation. For rescue experiments, neurons were co-transduced with lentiviruses encoding mDia1-WT-SNAP (τshmDia1 + mDia1-WT = 16.1 ± 1.9 s) or mDia1-K994A-SNAP (τshmDia1 + mDia1-K994A = 29.0 ± 1.9 s) (pshmDia1 vs shmDia1 + mDia1-WT < 0.01; pshmDia1 + mDia1-WT vs shmDia1 + mDia1-K994A < 0.001, one-way ANOVA with Tukey’s post-test). Data are expressed as mean ± SEM. N = 6 independent experiments from nshCTR = 21 videos; nshmDia1 = 21 videos, nshmDia1 + mDia1-WT = 16 videos, nshmDia1 + mDia1-K994A = 19 videos.

(E) Representative three-channel time-gated STED image of two synapses from hippocampal cultures, transduced with shmDia1+3, fixed and immunostained for Bassoon (magenta), F-Actin (cyan) and Homer1 (green). Synapses from A and E stem from the same experiment and brightness and contrast of all three channels were set to the same values. Scale bar, 250 nm.

(F) Presynaptic F-Actin levels in synapses from hippocampal cultures transduced with shCTR (A; 100 ± 6.4) or shmDia1+3 (E; 58.1 ± 2.9; p < 0.001, one sample Wilcoxon test). Line profiles of F-Actin overlapping with Bassoon (presynapse) distribution were integrated. Data shown are normalized to shCTR (set to 100) and expressed as mean ± SEM. N = 4 independent experiments from nshCTR = 155 synapses, nshmDia1+3 = 158 synapses.

(G,H) Representative confocal and two-channel time-gated STED images of endogenous β-Actin (cyan) in vGLUT1 (magenta) positive boutons from hippocampal neurons transduced with lentiviruses encoding shCTR (G) or shmDia1+3 (H). Synapses from G and H stem from the same experiment and brightness and contrast of the β-Actin channel were set to the same values. Scale bar, 250 nm.

(I) Analysis of presynaptic endogenous β-Actin levels in vGLUT1 positive boutons from shCTR (G; 100 ± 6.3) and shmDia1+3 (H; 47.7 ± 4.3; p < 0.0001 one sample Wilcoxon test) transduced neurons. β-Actin STED mean intensity was measured using confocal vGLUT1 signal as a mask. Data shown are normalized to shCTR (set to 100) and expressed as mean ± SEM from two independent experiments and nshCTR = 67 synapses, nshmDia1+3 = 53 synapses.

(J) Endocytic decay constants of vGLUT1-pHluorin traces (Figure 4 – figure supplement 1I) for neurons transduced with shCTR or shmDia1+3 in response to 40 AP (20 Hz, 2s) stimulation. Neurons were pre-incubated with 0.1 % DMSO or 1 µM Jasplakinolide (Jasp) for 30 min in the media before imaging (τshCTR + DMSO = 13.4 ± 1.0 s, τshCTR + Jasp = 15.0 ± 2.2 s, τshmDia1+3 + DMSO = 25.0 ± 2.7 s, τshmDia1+3 + Jasp = 15.6 ± 2.4 s; pshCTR vs shmDia1+3 < 0.01; pshmDia1+3 + DMSO vs shmDia1+3 + Jasp < 0.05, one-way ANOVA with Tukey’s post-test). Data are expressed as mean ± SEM. N = 6 independent experiments from nshCTR + DMSO = 32 videos, nshmDia1+3 + DMSO = 35 videos, nshCTR + Jasp = 33 videos; nshmDia1+3 + Jasp = 34 videos).

RhoA/B facilitate presynaptic endocytosis and are regulated by mDia1/3.

(A) Schematic representation of activation of mDia1 by RhoA-GTP and positive feedback loop of mDia1 on RhoA-GTP levels through GEF stimulation.

(B) Representative three-channel time-gated STED image of a synapses from hippocampal cultures, fixed and immunostained for Bassoon (magenta), RhoA (cyan) and Homer1 (green). Scale bar, 250 nm.

(C) Averaged normalized line profiles for synaptic distribution of RhoA and Homer1 relative to Bassoon (Maximum set to 0 nm). Data are expressed as mean ± SEM. N = 5 independent experiments from n = 230 synapses.

(D) Endocytic decay constants of averaged normalized Synaptophysin-pHluorin fluorescence traces (figure5 – figure supplement 1A) in response to 200 AP (40 Hz, 5s) stimulation. Neurons were transfected with the annotated combinations of plasmids encoding wild-type (WT) or dominant-negative (DN, T19N mutation) RhoA and RhoB (τRhoA-WT + RhoB-WT = 18.4 ± 0.7 s, τRhoA-WT + RhoB-DN = 16.0 ± 1.0 s, τRhoA-DN + RhoB-WT = 19.8 ± 2.4 s, τRhoA-DN + RhoB-DN = 30.1 ± 1.0 s; pRhoA-WT + RhoB-WT vs RhoA-DN + RhoB-DN < 0.01, one-way ANOVA with Tukey’s post-test). Data shown represent mean ± SEM. N = 3 independent experiments from nRhoA-WT + RhoB-WT = 21 videos, nRhoA-DN + RhoB-WT = 31 videos, nRhoA-WT + RhoB-DN = 23 videos, nRhoA-DN + RhoB-DN = 22 videos.

(E) Analysis of RhoA activity by RhoA-GTP pulldown (PD) from whole-cell lysates (input) of mouse hippocampal neurons expressing shCTR or shmDia1+3 using immobilized Rhotekin as a bait. Samples were analyzed by immunoblotting for mDia1, mDia3, RhoA and Tubulin using specific antibodies. Input, 10% of material used for the pulldown. Contrast of pulldown and input blots was seperatly adjusted for visualisation purposes.

(F) Densitometric quantification of RhoA-GTP normalized to total RhoA levels (input) in lysates from neurons transduced with shCTR or shmDia1+3 (0.7 ± 0.0, p < 0.001, one sample t-test) from immunoblots exemplified in E. Values for shCTR were set to 1. Data are expressed as mean ± SEM from N = 3 independent experiments.

mDia1/3-Rho and Rac1 signaling facilitates presynaptic endocytosis.

(A) Schematic of the interplay between RhoA and Rac1 signaling via GTPase activating proteins (GAPs) common for RhoA and Rac1.

(B) Analysis of Rac1 activity by Rac1-GTP pulldown (PD) from whole-cell lysates (input) of mouse hippocampal neurons expressing shCTR or shmDia1+3 utilizing immobilized PAK as a bait. Samples were analyzed by immunoblotting for mDia1, mDia3, Rac1 and Tubulin using specific antibodies. Input, 10% of material used for the pulldown. Contrast of pulldown and input blots was seperatly adjusted for visualisation purposes.

(C) Densitometric quantification of Rac1-GTP normalized to total Rac1 levels (input) in lysates from neurons transduced with shCTR or shmDia1+3 (2.2 ± 0.2; p < 0.05, one sample t-test) from immunoblots exemplified in (B). Values for shCTR were set to 1. Data are expressed as mean ± SEM from N = 3 independent experiments.

(D) Representative three-channel time-gated STED image of a synapses from hippocampal cultures, fixed and immunostained for Bassoon (magenta), Rac1 (cyan) and Homer1 (green). Scale bar, 250 nm.

(E) Averaged normalized line profiles for synaptic distribution of Rac1 and Homer1 relative to Bassoon (Maximum set to 0 nm). Data represent mean ± SEM. N = 3 independent experiments from n = 79 synapses.

(F) Averaged normalized vGAT-CypHer fluorescence traces for neurons transduced with shCTR or shmDia1+3 in response to 200 AP (40 Hz, 5 s) stimulation. Cells were acutely treated with 0.1 % DMSO or 10 µM Rac1 Inhibitor (EHT 1864) in the imaging buffer. Data shown represent the mean ± SEM. N = 8 independent experiments from nshCTR + DMSO = 46 videos, nshmDia1+3 + DMSO = 45 videos, nshCTR + EHT 1864 = 42 videos, nshmDia1+3 + EHT 1864 = 43 videos.

(G) Endocytic decay constants of vGAT-CypHer traces in F: τshCTR + DMSO = 14.7 ± 0.9 s, τshmDia1+3 + DMSO = 27.5 ± 2.3 s, τshCTR + EHT 1864 = 30.3 ± 6.7 s, τshmDia1+3 + EHT 1864 = 41.0 ± 4.3 s; pshCTR + DMSO vs shmDia1+3 + DMSO < 0.05, pshCTR + DMSO vs shmDia1+3 + EHT 1864 < 0.0001, Kruskal-Wallis test with Dunn’s post-test. Data represent mean ± SEM.

(H) Endocytic decay constants of Synaptophysin-pHluorin traces (Figure 6 –figure supplement 1K) of neurons transduced with shCTR (τshCTR = 12.0 ± 0.7 s) or shmDia1+3 (τshmDia1+3 = 22.7 ± 2.0 s) and transfected with constitutively active Rac1 (Rac1-CA; Q61L variant; τshCTR + Rac1-CA = 13.6 ± 1.2 s, τshmDia1+3 + Rac1-CA = 13.3 ± 1.4 s) or dominant negative Rac1 (Rac1-DN; T17N variant; τshCTR + Rac1-DN = 27.8 ± 1.3 s, τshmDia1+3 + Rac1-DN= 33.4 ± 1.6 s) in response to 200 AP (40 Hz, 5s) stimulation (pshCTR vs shmDia1+3 < 0.01; pshCTR vs shCTR + Rac1-DN < 0.0001, pshCTR vs shmDia1+3 + Rac1-DN < 0.01, pshmDia1+3 vs shmDia1+3 + Rac1-DN < 0.01, one-way ANOVA with Tukey’s post-test). Data are expressed as mean ± SEM. N = 3 independent experiments from nshCTR = 12 videos, nshmDia1+3 = 23 videos; nshCTR + Rac1-CA = 10 videos, nshmDia1+3 + Rac1-CA = 14 videos, nshCTR + Rac1-DN = 9 videos; nshmDia1+3 + Rac1-DN = 13 videos.

Defects in presynaptic ultrastructure induced by interference with mDia1/3-Rho and Rac1 signaling.

(A,B) Representative synaptic electron micrographs from hippocampal neurons transduced with lentiviruses encoding shCTR (A) or shmDia1+3 (B) and treated with 0.1 % DMSO before fixation. Invaginations and ELVs are colored in blue and yellow, while postsynapse and synaptic cleft are colored in green and maroon, respectively. Scale bar, 250 nm.

(C) Average number of invaginations per μm2 in shCTR (0.5 ± 0.1) and shmDia1+3 (0.9 ± 0.1; p < 0.0001, Mann-Whitney test) boutons. Data represent mean ± SEM. N = 3 independent experiments from nshCTR = 326 synapses, nshmDia1+3 = 323 synapses.

(D) Average number of ELVs per μm2 in shCTR (1.4 ± 0.1) and shmDia1+3 (2.7 ± 0.2; p < 0.0001, Mann-Whitney test) boutons. Data represent mean ± SEM. N = 3 independent experiments from nshCTR = 326 synapses, nshmDia1+3 = 323 synapses.

(E,F) Representative synaptic electron micrographs from neurons transduced with shCTR (A) or shmDia1/3 (B) and treated with 10 µM Rac1 Inhibitor (EHT 1864) for 2 h before fixation. Invaginations and ELVs are colored in blue and yellow, while postsynapse and synaptic cleft are colored in green and maroon, respectively. Scale bar, 250nm.

(G) Average number of invaginations per μm2 in shCTR and shmDia1+3 boutons treated with 0.1 % DMSO (0.8 ± 0.1 for shCTR; 1.2 ± 0.1 for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + DMSO < 0.01) or 10 µM EHT 1864 (1.8 ± 0.1 for shCTR, pshCTR + DMSO vs shCTR + EHT 1864 < 0.0001; 1.9 ± 0.2 for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + EHT 1864 < 0.0001, pshmDia1+3 + DMSO vs shmDia1+3 + EHT 1864 < 0.05, Kruskal-Wallis test with Dunn’s post-test) for 2 h before fixation. Data represent mean ± SEM from nshCTR + DMSO = 144 synapses, nshmDia1+3 + DMSO = 143 synapses, nshCTR + EHT 1864 = 136 synapses, nshmDia1+3 + EHT 1864 = 153 synapses.

(H) Average number of ELVs per μm2 in shCTR and shmDia1+3 boutons treated with 0.1 % DMSO (2.2 ± 0.2 for shCTR; 3.3 ± 0.3 for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + DMSO < 0.05) or 10 µM EHT 1864 (2.6 ± 0.3 for shCTR; 4.3 ± 0.4 for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + EHT 1864 < 0.001, pshCTR + EHT 1864 vs shmDia1+3 + EHT 1864 < 0.01, Kruskal-Wallis test with Dunn’s post-test) for 2 h before fixation. Data represent mean ± SEM from nshCTR + DMSO = 144 synapses, nshmDia1+3 + DMSO = 143 synapses, nshCTR + EHT 1864 = 136 synapses, nshmDia1+3 + EHT 1864 = 153 synapses.

Role of formins and mDia1/3 in SV endocytosis.

(A) Analysis of knockdown efficiency of lentiviral particles carrying shRNA against no mammalian target (shCTR) or mDia1 and mDia3 (shmDia1+3) in mouse hippocampal cultures harvested 12 days after transduction. Protein abundance of mDia1, mDia3 and Tubulin were immunoblotted with specific antibodies.

(B) Averaged normalized vGLUT1-pHluorin fluorescence traces from stimulated (40 APs; 20 Hz, 2 s) hippocampal neurons transduced with lentiviruses encoding shCTR, shmDia1 or both shmDia1 and shmDia3 combined (shmDia1+3). Data represent mean ± SEM. N = 4 independent experiments from nshCTR = 17 videos, nshmDia1 = 19 videos, nshmDia1+3 = 18 videos. The corresponding endocytic decay constants are shown in figure 1G.

(C) Averaged normalized vGLUT1-pHluorin fluorescence traces for neurons transduced with shCTR, shmDia1 or shmDia1+3 in response to 80 AP stimulation (40 Hz, 2 s). Data represent mean ± SEM. N = 4 independent experiments from nshCTR = 12 videos; nshmDia1 = 15 videos; nshmDia1+3 = 18 videos. Corresponding endocytic decay constants are shown in figure 1H.

(D) Maxima of background-corrected vGLUT1-pHluorin fluorescence traces (surface normalized) for neurons transduced with shCTR (1.9 ± 0.1) or shmDia1+3 (1.9 ± 0.1) in response to 40 AP stimulation (20 Hz, 2 s). Data represent mean ± SEM. N = 22 independent experiments from nshCTR = 105 videos and nshmDia1+3 = 128 videos.

(E) Schematic representation of the activation of mDia1. Binding of RhoA-GTP to the RBD domain (green) or application of mDia1 activator (IMM) competes with the intramolecular interaction of the N-terminal DID (yellow) with the C-terminal DAD (red) domain (see figure 3A for domain structure). The release of autoinhibition leads to dimerization of mDia formins in solution.

SV depletion in mDia1/3-depleted neurons is activity dependent.

(A,B) Representative synaptic electron micrographs from neurons transduced with shCTR and treated with 0.1% Vehicle (A) or 1 µM TTX (B) for 36 h before fixation. Postsynapse and postsynaptic cleft are colored in green and maroon, respectively. Scale bar, 250 nm.

(C) Average number of SV per μm2 in synaptic boutons of neurons transduced with shCTR and treated with 0.1 % Vehicle (A; 90.7 ± 3.1) or 1 µM TTX (B; 98.4 ± 3.3) for 36 h before chemical fixation. Data shown represent the mean ± SEM from two independent experiments and nVehicle = 180, nTTX = 204 synapses. The dotted line represents the average SV numbers/μm2 in shmDia1+3 boutons treated with DMSO from the same experiments as a reference.

(D,E) Representative synaptic electron micrographs from neurons transduced with shmDia1+3 and treated with 0.1% Vehicle (D) or 1 µM TTX (E) for 36 h before fixation. Postsynapse and postsynaptic cleft are colored in green and maroon, respectively. Scale bar, 250nm. Average number of SV per μm2 in boutons is shown in figure 2D.

mDia1 binds membranes and localizes to presynaptic endocytic sites.

(A) Membrane levels of mDia1-WT-mCherry versus mDia1-ΔN-mCherry proteins overexpressed in HEK293T cells. Membrane and cytosolic cellular fractions were isolated by ultracentrifugation and analyzed by immunoblotting with specific antibodies (LAMP1) and in-gel fluorescence of mCherry tags.

(B) Densitometric quantification of mDia1-WT versus mDia1-ΔN (0.6 ± 0.1; p < 0.05, one sample t-test) membrane-associated protein levels. Data shown are normalized to mDia1-WT (set to 1) and expressed as mean ± SEM. Representative immunoblot is shown in A. N = 5 independent experiments.

(C) Averaged normalized Synaptophysin-pHluorin fluorescence from stimulated (200 APs, 40 Hz, 5s) hippocampal neurons transfected with shCTRmiR or shmDia1miR. For rescue experiments, neurons were co-transfected with plasmids encoding mDia1-WT-mCherry, mDia1-ΔN-mCherry or mCherry alone (shCTRmiR & shmDia1miR). Endocytic decay constants are shown in figure 3B.

(D) Full volcano plot of proteins from figure 3E associating with synaptic mDia1 analyzed by label-free proteomics of anti-mDia1 versus CTR immunoprecipitates from detergent-extracted mouse synaptosomes (P2’ fraction). The cyan dot shows the specific enrichment of mDia1 as the bait protein of the immunoprecipitation (p < 0.001, two-tailed student’s t-test). N = 3 independent experiments.

(E) Endogenous co-immunoprecipitation of Myosin IIB by mDia1 from detergent-extracted mouse synaptosomes (P2’ fraction). Samples were analyzed by immunoblotting using specific antibodies against mDia1, Myosin IIB and β-Actin.

(F) Representative three-channel time-gated STED image of a synapse from hippocampal cultures fixed and immunostained for Bassoon (presynaptic marker, magenta), Myosin IIB (cyan) and Homer1 (postsynaptic marker, green). Scale bar, 250 nm.

(G) Averaged normalized line profiles for synaptic distribution of Myosin IIB and Homer1 relative to Bassoon (Maximum set to 0 nm). Data are expressed as mean ± SEM. N = 3 independent experiments from n = 267 synapses.

mDia1 regulates presynaptic actin and SV endocytosis.

(A,B) Representative three-channel time-gated STED images of synapses from hippocampal cultures treated with 0.1% DMSO (A) or 80 µM Dynasore (B) for 10 min. Cells were fixed and stained for Bassoon (magenta), F-Actin (cyan) and Homer1 (green). Synapses from A and B stem from the same experiment and brightness and contrast of all three channels were set to the same values. Scale bar, 250 nm. Corresponding analysis of presynaptic F-Actin levels is shown in figure 4C.

(C,D) Representative three-channel time-gated STED images of synapses from hippocampal cultures transduced with Dynamin1-WT (C) or Dynamin1-K44A (D). Cells were fixed and stained for Bassoon (magenta), F-Actin (cyan) and Homer1 (green). Synapses from C and D stem from the same experiment and brightness and contrast of all three channels were set to the same values. Scale bar, 250 nm.

(E) Presynaptic F-Actin levels in synapses from neurons transduced with Dynamin1-WT (C; 100 ± 5.9) or Dynamin1-K44A (D; 119.8 ± 6.2, p < 0.01, one sample t-test). Absolute line profiles of F-Actin overlapping with Bassoon (presynapse) distribution were integrated. Data shown are normalized to WT (set to 100) and expressed as mean ± SEM. nWT = 54 synapses, nK44A = 49 synapses.

(F) Averaged normalized vGLUT1-pHluorin fluorescence traces for neurons transduced with shCTR or shmDia1 in response to 200 AP (40 Hz, 5s) stimulation. For rescue purposes, cells were co-transduced with mDia1-WT-SNAP or mDia1-K994A-SNAP. Data are expressed as mean ± SEM. N = 6 independent experiments from nshCTR = 21 videos; nshmDia1 = 21 videos; nshmDia1 + mDia1-WT = 16 videos; nshmDia1 + mDia1-K994A = 19 videos. Corresponding endocytic decay constants are shown in figure 4D.

(G) Representative STED images of endogenous β-Actin in vGLUT1 positive synapses in hippocampal neurons transduced with shCTR or shmDia1+3 and treated with 0.1 % DMSO or 1 µM Jasplakinolide for 45 min. Neurons were co-transfected with pOrange-GFP-β-Actin knockin and vGLUT1-mCherry plasmids before fixation and immunostaining. Scale bar, 2.5 µm.

(H) Averaged normalized vGLUT1-pHluorin fluorescence traces for neurons transduced with shCTR or shmDia1+3 in response to 40 AP (20 Hz, 2s) stimulation. Neurons were pre-incubated with 0.1 % DMSO or 1 µM Jasplakinolide (Jasp) for 30 min in the cell media before imaging. Data are expressed as mean ± SEM. N = 6 independent experiments from nshCTR + DMSO = 32 videos, nshmDia1+3 + DMSO = 35 videos, nshCTR + Jasp = 33 videos; nshmDia1+3 + Jasp = 34 videos. The corresponding endocytic decay constants are shown in figure 4I.

RhoA/B regulate SV endocytosis.

(A) Averaged normalized Synaptophysin-pHluorin fluorescence traces from stimulated (200 APs; 40 Hz, 5s) hippocampal neurons transfected with plasmids encoding the indicated combinations of WT or DN RhoA and RhoB variants. Data represent mean ± SEM. N = 3 independent experiments from nRhoA-WT + RhoB-WT = 21 videos, nRhoA-DN + RhoB-WT = 31 videos, nRhoA-WT + RhoB-DN = 23 videos, nRhoA-DN + RhoB-DN = 22 videos. Endocytic decay constants are shown in figure 5D.

(B) Averaged normalized Synaptophysin-pHluorin fluorescence traces from stimulated (200 APs; 40 Hz, 5s) hippocampal neurons transfected with shRNA against no mammalian target (shCTR) or against RhoA and RhoB (shRhoA+B). Data represent mean ± SEM. N = 3 independent experiments from nshCTR = 27 videos, nshRhoA+B = 25 videos.

Cooperative action of mDia1/3 and Rac1 pathways in presynaptic endocytosis.

(A) Analysis of Rac1 activity by Rac1-GTP pulldown (PD) from whole-cell lysates (input) of mouse hippocampal cultures upon inhibition of Rho activity utilizing immobilized PAK as bait. Cells were treated with 0.1 % DMSO or 10 µM Rho Inhibitor (Rhosin) for 2 h before harvest. Samples were analyzed by immunoblotting for Rac1 and Tubulin using specific antibodies. Input, 10% of material used for the pulldown. Contrast of pulldown and input blots was seperatly adjusted for visualisation purposes.

(B,C) Representative three-channel time-gated STED images of synapses from hippocampal cultures treated with 0.1 % DMSO or 10 µM Rac1 Inhibitor (EHT 1864) for 2 h. Cells were fixed and stained for Bassoon (magenta), F-Actin (cyan) and Homer1 (green). Scale bar, 250 nm.

(D) Presynaptic F-Actin levels in synapses of neurons treated with 0.1 % DMSO (100 ± 8.5) or 10 µM Rac1 Inhibitor (EHT 1864; 58.6 ± 6.5; p < 0.0001, one sample Wilcoxon test) for 2 h. Line profiles of F-Actin overlapping with Bassoon (presynapse) distribution were integrated. Data shown are normalized to DMSO (set to 100) and expressed as mean ± SEM. nDMSO = 30, nEHT 1864 = 46 from two independent experiments.

(E) Averaged normalized Synaptophysin-pHluorin fluorescence traces from stimulated (200 APs; 40 Hz, 5s) hippocampal neurons transduced with lentiviruses encoding shCTR or shmDia1+3 and transfected with plasmids for expression of constitutively-active Rac1 (Rac1-CA; Q61L variant) or dominant-negative Rac1 (Rac1-DN; T17N variant). Data represent mean ± SEM. N = 3 independent experiments from nshCTR = 12 videos, nshmDia1+3 = 23 videos, nshCTR + Rac1-CA = 10 videos, nshmDia1+3 + Rac1-CA = 14 videos, nshCTR + Rac1-DN = 9 videos; nshmDia1+3 + Rac1-DN = 13 videos. The corresponding endocytic decay constants are shown in figure 6H.

(F) Maxima of background-corrected Synaptophysin-pHluorin fluorescence traces (surface normalized maximum values of traces shown in E) from stimulated (200 APs; 40 Hz, 5s) hippocampal neurons transduced with lentiviruses encoding shCTR (Fmax/F0 = 1.3 ± 0.0) or shmDia1+3 (Fmax/F0 = 1.5 ± 0.0) and transfected with plasmics encoding CA (Fmax/F0 shCTR + Rac1-CA = 1.4 ± 0.2; Fmax/F0 shmDia1+3 + Rac1-CA = 1.5 ± 0.1) or DN versions ((Fmax/F0 shCTR + Rac1-DN = 1.2 ± 0.1; Fmax/F0 shmDia1+3 + Rac1-DN = 1.3 ± 0.1) of Rac1. Data represent mean ± SEM.

(G) Representative three-channel time-gated STED image of a synapses from hippocampal mouse cultures, fixed and immunostained for Bassoon (magenta), Cdc42 (cyan) and Homer1 (green). Scale bar, 250 nm.

(H) Averaged normalized line profiles for synaptic distribution of Cdc42 and Homer1 relative to Bassoon (Maximum set to 0 nm). Data are expressed as mean ± SEM (N = 3; n = 96 synapses).

(I) Densitometric quantification of Cdc42-GTP normalized to total Cdc42 levels in lysates from shmDia1+3 neurons (2.7 ± 0.6; p < 0.05, one sample t-test).Values for shCTR were set to 1. Data are expressed as mean ± SEM from N = 3 independent experiments.

(J) Averaged normalized vGAT-CypHer fluorescence traces for neurons transduced with shCTR or shmDia1+3 in response to 200 AP (40 Hz, 5s) stimulation. Cells were acutely treated with 0.1 % DMSO or 10 µM Cdc42 Inhibitor (ML141) in the imaging buffer. Data shown represent the mean ± SEM. N = 6 independent experiments from nshCTR + DMSO = 31 videos, nshmDia1+3 + DMSO = 33 videos, nshmDia1+3 + ML141 = 32 videos.

(K) Endocytic decay constants of vGAT-CypHer traces in I: τshCTR + DMSO = 15.6 ± 1.0 s, τshmDia1+3 + DMSO = 28.0 ± 3.1 s, τshCTR + ML141 = 17.6 ± 1.6 s, τshmDia1+3 + ML141 = 33.1 ± 7.7 s; pshCTR + DMSO vs shmDia1+3 + DMSO < 0.01, Kruskal-Wallis test with Dunn’s post-test. Data shown represent the mean ± SEM. N = 6 independent experiments from nshCTR + DMSO = 31 videos, nshmDia1+3 + DMSO = 33 videos, nshCTR + ML141 = 29 videos, nshmDia1+3 + ML141 = 32 videos.

mDia1/3 and Rac1 cooperatively regulate the SV cycle and presynaptic ultrastructure.

(A) Average number of invaginations per μm2 in WT (0.1 ± 0.1) and mDia1 KO (0.4 ± 0.1; p < 0.01, Mann-Whitney test) boutons. Data shown represent the mean ± SEM from nWT = 103 synapses, nKO = 96 synapses.

(B) Average number of ELVs per μm2 in WT (1.3 ± 0.2) and mDia1KO (3.1 ± 0.5; p < 0.001, Mann-Whitney test) boutons. Data shown represent the mean ± SEM from nWT = 103 synapses, nKO = 96 synapses.

(C) Average invagination length in shCTR and shmDia1+3 boutons treated with 0.1 % DMSO (97.6 ± 4.5 nm for shCTR; 136.8 ± 6.0 nm for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + DMSO < 0.0001) or 10 µM EHT 1864 (130.9 ± 4.5 nm for shCTR, pshCTR + DMSO vs shCTR + EHT 1864 < 0.001; 143.1 ± 4.9 nm for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + EHT 1864 < 0.0001, Kruskal-Wallis test with Dunn’s post-test) for 2 h before chemical fixation. Data represent mean ± SEM from nshCTR + DMSO = 77 invaginations, nshmDia1+3 + DMSO = 141 invaginations, nshCTR + EHT 1864 = 176 invaginations, nshmDia1+3 + EHT 1864 = 189 invaginations.

(D) Average invagination width in shCTR and shmDia1+3 boutons treated with 0.1 % DMSO (124.5 ± 5.6 nm for shCTR; 184.1 ± 6.6 nm for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + DMSO < 0.0001) or 10 µM EHT 1864 (179.0 ± 5.8 nm for shCTR, pshCTR + DMSO vs shCTR + EHT 1864 < 0.0001; 191.0 ± 5.4 nm for shmDia1+3, pshCTR + DMSO vs shmDia1+3 + EHT 1864 < 0.0001, Kruskal-Wallis test with Dunn’s post-test) for 2 h before chemical fixation. Data represent mean ± SEM from nshCTR + DMSO = 77 invaginations, nshmDia1+3 + DMSO = 141 invaginations, nshCTR + EHT 1864 = 176 invaginations, nshmDia1+3 + EHT 1864 = 189 invaginations.