Plekhg5-regulated autophagy of synaptic vesicles reveals a pathogenic mechanism in motoneuron disease

Abstract

Autophagy-mediated degradation of synaptic components maintains synaptic homeostasis but also constitutes a mechanism of neurodegeneration. It is unclear how autophagy of synaptic vesicles and components of presynaptic active zones is regulated. Here, we show that Pleckstrin homology containing family member 5 (Plekhg5) modulates autophagy of synaptic vesicles in axon terminals of motoneurons via its function as a guanine exchange factor for Rab26, a small GTPase that specifically directs synaptic vesicles to preautophagosomal structures. Plekhg5 gene inactivation in mice results in a late-onset motoneuron disease, characterized by degeneration of axon terminals. Plekhg5-depleted cultured motoneurons show defective axon growth and impaired autophagy of synaptic vesicles, which can be rescued by constitutively active Rab26. These findings define a mechanism for regulating autophagy in neurons that specifically targets synaptic vesicles. Disruption of this mechanism may contribute to the pathophysiology of several forms of motoneuron disease.

Fig. 1

Plekhg5-deficient mice develop a motoneuron disease with late onset. a Survival analysis of Plekhg5 mutant mice. (Mean survival after 24 months: Plekhg5 ^+/+: 95%, n = 48; Plekhg5 ^+/−: 89%, n = 112; Plekhg5 ^−/−: 68%, n = 351; p = 0.0068; log-rank test) b Hind-limb clasping defects in 15-month-old Plekhg5-deficient mice. c Body weight measurements of Plekhg5-deficient and control mice of different ages (2−4 months, Plekhg5 ^+/+: n = 12, Plekhg5 ^−/−: n = 9; 9−15 months, Plekhg5 ^+/+: n = 19, Plekhg5 ^−/−: n = 42; 18−25 months, Plekhg5 ^+/+: n = 18, Plekhg5 ^−/−: n = 36; two-way ANOVA; Bonferroni post-test). d Number of motoneurons was counted in lumbar spinal cord sections (3 months, n = 5; 12 months, n = 5; 24 months, n = 3; unpaired t-test; two-tailed. e, f Grip strength measurements of fore- e and hindlimbs f. g Rotarod performance (9−15 months, Plekhg5 ^+/+: n = 9, Plekhg5 ^−/−: n = 10; 18−24 months, Plekhg5 ^+/+: n = 11, Plekhg5 ^−/−: n = 8; two-way ANOVA; Bonferroni post-test). h Nissl-stained motoneurons in spinal sections from Plekhg5 ^+/+ and Plekhg5 ^−/− mice. Scale bar: 40 µm. i−l Electrophysiological recordings of gastrocnemius and plantaris muscles (Plekhg5 ^+/+: n = 6, Plekhg5 ^−/−: n = 6; unpaired t-test; two-tailed). Latency i and amplitude j of muscle depolarization upon stimulation of the sciatic nerve. k, l Single motor unit potential (SMUP) k and motor unit number estimation (MUNE) l of the gastrocnemius muscle. m Succinate dehydrogenase (SDH) staining of the gastrocnemius muscle. Scale bar: 500 µm. n X-Gal staining of cross-sectioned spinal cord and gastrocnemius muscle. Scale bar: 20 µm. o Sciatic nerve conduction velocities (Plekhg5 ^+/+: n = 6, Plekhg5 ^−/−: n = 6; unpaired t-test; two-tailed) p, q Box- and whisker-plots show mechanical withdrawal thresholds p and heat withdrawal latencies (s) q of naive male Plekhg5 ^−/− and control littermates. Mechanical withdrawal thresholds did not differ between genotypes p. Plekhg5 ^−/− mice showed heat hypersensitivity as compared to control littermates (Plekhg5 ^+/+: n = 5, Plekhg5 ^−/−: n = 5; Mann−Whitney U-test)

Fig. 2

Degradation of neuromuscular junctions in Plekhg5-deficient mice. a, b NMJs within the gastrocnemius a and intercostal b muscles stained for BTX and synaptophysin. Scale bar: 10 µm. c, d Quantification of presynaptic area by evaluation of synaptophysin staining and measurement of the widest expansion of presynaptic sites (three animals per genotype with at least 15 NMJs analyzed per animal; mean ± SEM; two-way ANOVA; Bonferroni post-test)

Fig. 3

Structural alterations and accumulation of synaptic vesicle proteins in Plekhg5-deficient mice. a–c Neuromuscular junction of Plekhg5 ^+/+ a and Plekhg5 ^−/− b, c mice. # labels empty inclusions. MF myofibers. Scale bar: 1 µm. d Membrane fragments in nerve terminals of Plekhg5-deficient mice. Scale bar: 300 nm. e Double membrane fragment forming an inclusion. Arrow points to single membrane. Arrowheads point to double membrane structures. ^# labels inclusions. Scale bar: 300 nm. f, g Synaptic vesicles in Plekhg5-deficient mice g appear frequently enlarged, in contrast to synaptic vesicles in wild-type mice f, which appear smaller and more uniform in size. Asterisks label enlarged synaptic vesicles. Scale bar: 100 nm. h Quantification of synaptic vesicle diameter. (Synaptic vesicles of five NMJs were analyzed per genotype. Mean ± SEM; two-way ANOVA; Bonferroni post-test). i Expression of several synaptic vesicle markers in sciatic nerve lysates of four mice per genotype. j Quantification of western blot analysis. Mean ± SEM; two-way ANOVA; Bonferroni post-test). Images have been cropped for presentation. Full size images are presented in Supplementary Fig. 7

Fig. 4

Loss of Plekhg5 impairs axonal integrity. a Swellings in distal motor axons. MF, muscle fiber. Scale bar: 2 µm. b, c High magnification micrographs of axons from wild-type b and Plekhg5-deficient c mice with altered cytoskeleton organization in Plekhg5 ^−/− mice. Scale bar: 250 nm. d Semi-thin cross-section of sciatic nerve from Plekhg5-deficient mice showing an axonal swelling. Scale bar: 10 µm. e High magnification micrograph of an axonal swelling from sciatic nerve. Scale bar: 1 µm. Inset, Scale bar: 2 µm. f–i Longitudinal- f, g and cross- h, i sections of lumbar spinal cord showing axonal swellings within the white matter of Plekhg5 ^−/− mice. f, h Scale bar: 10 µm. g, i Fine structure of axonal swellings in spinal cord white matter of Plekhg5-deficient mice. ^# labels axon swellings; asterisks label axons with unaltered morphology. g Scale bar: 2 µm. h Scale bar: 10 µm. i Scale bar: 500 nm. j Quantification of axonal swellings in spinal cord semi-thin cross-sections. Three animals per genotype were analyzed. Each data point represents the mean of 10 sections from individual animals with a distance of 100 µm between each section

Fig. 5

Plekhg5 regulates biogenesis of autophagosomes. a Motoneurons of Plekhg5 ^+/+ and Plekhg5 ^−/− mice were transduced with mRFP-GFP-LC3-expressing lentiviruses and cultured for seven days. At day seven cells were treated with Bafilomycin A1 for four hours or left untreated and the number of autophagosomes (mRFP⁺-GFP⁺-LC3) and autolysosomes (mRFP⁺-GFP⁻-LC3) was determined in the soma b and axon c (three independent experiments with 15 cells analyzed in each experiment; mean ± SEM; two-way ANOVA; Bonferroni post-test). Soma, scale bar: 5 µm. Axon, scale bar: 20 µm. Axon (blow up), scale bar: 5 µm. d Representative kymographs of autophagosome motility. Cultured motoneurons were transduced with mRFP-GFP-LC3 and the movement of GFP positive punctae in axons was monitored for 20 min. e Proportion of retrogradely, anterogradely or stationary/bidirectionally moving autophagosomes (mean ± SEM; n = 10 cells; two-way ANOVA; Bonferroni post-test). f Number of retrogradely moving autophagosomes per minute (mean ± SEM; n = 10 cells; Studentʼs t-test; one-tailed). g Western blot analysis of LC3 expression in neurosphere-derived cortical neurons from control and Plekhg5-deficient mice. Cells were treated with 400 nM Bafilomycin A1 for 4 h or left untreated. LE low exposure, HE high exposure. h Quantification of LC3 western blots. (Mean ± SEM; n = 3; mean ± SEM; one-way ANOVA). i Expression of LC3 in spinal cord lysates of three mice per genotype. j Quantification of LC3 Western blots from spinal cord extracts. (mean ± SEM; n = 4; mean ± SEM; Studentʼs t-test; two-tailed). k NSC34 cells were transfected with Plekhg5-YFP. Seventy-two hours after transfection, cells were starved for 4 h in HBSS or left untreated. The expression of Plekhg5-YFP was analyzed by western blot. l Motoneurons were cultured for seven days in motoneuron-media. At day seven cells were cultured for 4 h with different media each deprived of specific components. Subsequently, motoneurons were lysed and the expression of endogenous Plekhg5 was analyzed by western blot. Asterisks label bands that do not change upon nutrient deprivation. Arrowheads point to bands that change upon nutrient deprivation. m NSC34 cells were transfected with Flag-Plekhg5. Seventy-two hours after transfection, cells were lysed and Flag-Plekhg5 was pulled-down with PI(3)P-coated beads. Western blot images have been cropped for presentation. Full size images are presented in Supplementary Fig. 7

Fig. 6

Depletion of Plekhg5 results in axon growth defects and degeneration of axon terminals in vitro. a Motoneurons were transduced with sh-control or sh-Plekhg5 #D lentiviruses or left untreated. Scale bar: 100 µm. b Seven days after sh-RNA transduction, knockdown of Plekhg5 reduced axon length, whereas dendrite length and number of dendrites were not affected. Each data point represents the mean of one individual experiment with at least 20 cells analyzed. One-way ANOVA, Bonferroni post-test. c–e Motoneurons were cultured for seven days and stained for synaptophysin, Rab26 and F-actin and imaged by confocal c and SIM microscopy e. c Scale bar: 10 µm. d Quantification of axon terminal size (n = 15 cells; unpaired t-test; two-tailed). e Motoneurons were cultured for 7 days and stained for synaptophysin, Rab26 and F-actin and imaged by SIM microscopy. Axons of Plekhg5 ^−/− motoneurons display synaptophysin positive swellings. Scale bar: 10 µm. Scale bar (blow up): 2 µm

Fig. 7

Plekhg5 functions as a GEF for Rab26. a Neurosphere-derived cortical neurons were transduced with EGFP-Rab26 and labeled with P³² for 4 h. EGFP-Rab26 was immunoprecipitated and GTP- or GDP-bound Rab26 was separated by TLC. b Densitometric quantification of GTP/GDP ratios (each data point represents one independent experiment. Mean ± SEM; unpaired t-test; two-tailed). c–h Exchange activity of Plekhg5 on different GTPases. c Scheme of the DH-PH tandem domain, which was purified and used for the biochemical assays. Exchange activity of Plekhg5 on Rab5 d, Rab27b e, Rab33b f, and Rab26 g was measured by monitoring the fluorescent increase of Mant-GppNHp upon binding to Rab proteins. g The initial GTP exchange rate of Rab26 was evaluated using the first 10 s of each time course. (Mean ± SEM; Studentʼs t-test; two-tailed)

Fig. 8

Expression of constitutively active Rab26 rescues axonal growth and autophagy defects in Plekhg5-deficient cells. a Representative images of GFP-Rab26-WT or GFP-Rab26-QL positive structures in axons of control cells and Plekhg5-deficient cells. Scale bar: 10 µm. b Morphology of GFP, GFP-Rab26-WT or GFP-Rab26-QL expressing motoneurons cultured for 7 days. Scale bar: 100 µm. c, d Size and number of axonal EGFP-Rab26 structures. e Axon length of GFP, GFP-Rab26-WT, or GFP-Rab26-QL expressing motoneurons isolated from Plekhg5 ^+/+ and Plekhg5 ^−/− mice (each data point represents one individual experiment with 15 cells analyzed in each experiment; mean ± SEM; two-way ANOVA). f Scheme of lentiviral vectors for simultaneous expression of RFP-GFP-LC3 and Flag-Rab26-WT or Flag-Rab26-QL, respectively. g Western blot analysis of Flag-Rab26-WT and Flag-Rab26-QL expression. Images have been cropped for presentation. Full size images are presented in Supplementary Fig. 7. h Motoneurons of Plekhg5 ^+/+ and Plekhg5 ^−/− mice expressing mRFP-GFP-LC3 and Flag-Rab26-WT or Flag-Rab26-QL were cultured for 7 days and the number of mRFP-GFP-LC3 positive structures was analyzed. Scale bar: 10 µm. i Number of autophagosomes and autolysosomes upon expression of Flag-Rab26-WT and Flag-Rab26-QL (three independent experiments with 10 cells analyzed in each experiment; mean ± SEM; two-way ANOVA; Bonferroni post-test)

Fig. 9

Plekhg5 depletion in SOD1 G93A motoneurons results in elevated ER-stress. a Expression of HSP70, HSP90, Calreticulin, and Calnexin in spinal cord lysates from three animals per genotype. b Quantification of western blot shown in a (each data point represents expression levels of one animal; unpaired t-test; two-tailed). c Expression of IRE1α and Chop1 in spinal cord lysates from three animals per genotype. d Quantification of western blot shown in c (each data point represents expression levels of one animal; unpaired t-test; two-tailed). e SOD1 G93A and non-transgenic motoneurons were depleted of Plekhg5 and several ER-stress markers were examined after 7 days in culture. f Quantification of western blots shown in e (each data point represents one individual experiment; mean ± SEM; unpaired t-test; two-tailed). g Survival of SOD1 G93A motoneurons decreased upon knockdown of Plekhg5 using two independent sh-RNA constructs (each data point represents the % of motoneuron-survival from one individual embryo. At least 50 motoneurons were evaluated from one embryo; mean ± SEM; two-way ANOVA; Bonferroni post-test). Images have been cropped for presentation. Full size images are presented in Supplementary Fig. 7