Fig 2: LC-MS proteomics identifies proteins involved in ß-oxidation as being more abundant in GDAP1 KD cells and hints to alterations of the actin cytoskeleton mediated by redox-dependent interactions with GDAP1.a Number of proteins dysregulated in GDAP1 KD cells identified by label-free proteomics from three independent experiments, measured in quadruplicates, green less abundant, magenta more abundant. Statistical significance was determined by Student’s t test with Bonferroni correction. b KEGG pathways significantly altered in proteins more abundant in GDAP1 KD cells. c Section of a STRING analysis of proteins more abundant in GDAP1 KD cells corroborating altered carbon metabolism (blue) and fatty acid degradation (green) in these cells. Open dots correspond to more abundant interacting proteins that do not fall directly into these categories. Proteins in bold were found to be more abundant independently using immunoblotting. d Schematic illustration showing the biotinylation of the AviTag fused to GDAP1 protein and subsequent pulldown and enrichment of biotinylated GDAP1 protein via streptavidin-labeled magnetic beads. Immunoblot of cell lysates prior to pulldown of biotinylated GDAP1 stained against GDAP1 and Biotin with an infrared-labeled streptavidin (SA) dye shows an overlap of biotinylation and GDAP1 protein. Streptavidin pulldown was performed with primary neuronal cultures from BirA-expressing E16 mouse embryos. e Results from label-free quantitative proteomics performed with GSH or GSSG added to the preparation, (f) shows the 72 proteins with an altered expression after addition of GSH but not GSSG. g TOP3 quantification of Cofilin-1 abundance and increased Cofilin-1 abundance upon GSH, but not GSSG addition indicated an interaction with GDAP1 in a redox-dependent manner. Statistical significance in g was determined by one-way ANOVA and the Tukey test, ***p < 0.0001.

Fig 3: GDAP1 knockdown reduces the number of contact sites between mitochondria and the endoplasmic reticulum.a Transmission electron microscopy of knockdown (KD) and control (Ctrl) cells. The indicated parameters were quantified using ImageJ by a blinded investigator. M, mitochondrion. b Histogram showing the distribution of MERCS’ widths, and an increased distance in KD cells between the ER and mitochondria. c Proximity of ER and mitochondria was measured using a FRET-FEMP sensor comprising ER CFP-Sac1 and mitochondrial YFP-Akap1. Proximity leads to high intensity of YFP-FRET-emission (410-430/520-560 ex/em). 100 nM rapamycine was added to achieve the closest possible distance, and the measurement of FRET ratio was calculated as (FRETmax-FRETbasal)/FRETbasal. d Quantification of ER membrane (Sec62) and mitochondrial (Cox4) proteins in mitochondrial fractions. e Representative curve for the aequorin Ca2+ measurement and quantification of mitochondrial Ca2+ levels after addition of 200 µM carbachol (CCH), which releases Ca2+ from the ER by activation of a G-protein-coupled receptor. f) Immunoblot of Ctrl and KD cells lysed 10 min after 200 µM CCH addition to quantify the PDH phosphorylation (pPDH/(PDHtotal) after Ca2+ release into the MERCS and mitochondria. g, h Immunoblot of total cell lysates of Ctrl and KD cells showing increased expression levels of MCU but not pan-IP3R, HSPA9 or VDAC1, actin served as loading control, size is indicated. Data in (a) were obtained from 11 (Ctrl) and 9 (KD) cells, in (b) from 4 independent experiments in triplicate or quintuplicate with a range of 162–818 cells per experiment, and in (d) from 4 independent experiments. Data in (e) from 5 independent experiments performed in triplicates. Data points corresponding to the example blots are highlighted. Statistical variation is shown as Tukey boxplots or scatter plots with the indication of mean ± SEM and significance was calculated using the non-parametric Mann–Whitney test, *p < 0.05, **p < 0.001, ***p < 0.0001.

Fig 4: Patient-derived cells are similarly characterized by an anaplerotic state and reduced mitochondrial Ca2+ levels.a Differentiation protocol to obtain motoneurons (MN) from induced-pluripotent stem cells (iPSCs) and (b) confirmatory immunocytochemistry by staining against dendrite marker MAP2 and motoneuronal neurofilament Smi32. c Immunoblot of total MN cell lysates for validation of GDAP1 expression, actin served as loading control, size is indicated. d Automated high-content microscopy analysis of MitoTracker-stained mitochondria revealed an increased total mitochondrial mass per well. e Immunoblot of total MN cell lysates demonstrated increased glutamate dehydrogenase (GLUD1, arrowhead) in patient-derived MNs, actin served as loading control. f Reduced glutamate levels in CMT4A patient-derived MN determined by a luminescence-based assay indicating increased glutamine consumption compared to Ctrl. g Automated high-content microscopy analysis of BODIPY-positive spots normalized to the nuclei area in MN revealed a reduced number of lipid droplets in CMT4A patient-derived cells. h Reduced relative mt[Ca2+] levels of NPCs determined by confocal microscopy with Rhod2-AM and mito-CEPIA3. i Immunoblots and quantification demonstrating increased phosphorylation of PDH E1 serine 293, normalized to total PDH E1 levels and control cells. Data in (d–g) were from 3 independent differentiations and data in (d, g) were obtained from n > 40,000 cells in quadruplicate, experiments in (f) were performed in duplicate. Data in (h) were from 20–22 Rhod2-AM-stained cells and 22–33 cells mitoCEPIA3 transfected from three independent experiments. Data in (i) were from 6 different passages. Data points corresponding to the example blots are highlighted. Statistical variation is shown as Tukey boxplots in (d, e) or scatter plots in (i) with mean ± SEM indicated. Significance was calculated using the non-parametric Kruskal–Wallis test, *p < 0.05, **p < 0.001, ***p < 0.0001.

Fig 5: GDAP1 KD reduces F-actin fibers at mitochondrial surfaces and restricts access of DRP1 to the mitochondria resulting in less DRP1 and MFF.a Live cell imaging of GFP-F-tractin transfected cells. Mitochondria were stained with MitoTracker and images analyzed for contact area using Imaris. b Immunoblot of DRP1 in cytosolic and mitochondrial fractions demonstrating a reduced DRP1 abundance in mitochondrial fractions. TOM20 and actin served as loading controls, size is indicated. c, d Live cell imaging of GFP-DRP1 transfected cells. Mitochondria were stained with MitoTracker and images analyzed for (c) colocalization and (d) area. e Immunoblot of total DRP1 and MFF demonstrating a reduced abundance of both proteins. Actin served as loading control, size is indicated. f Scheme depicting our findings. Cof1, Cofilin-1. In the absence of GDAP1 mitochondrial F-actin is reduced resulting in an increased distance between ER and mitochondria, reduced mitochondrial Ca2+ levels, and a reduced presence of DRP1 at mitochondrial constriction sites. Data from (a) are from n = 19 cells per cell line from three independent experiments performed in triplicates. Data from (c) are from n = 10 and (d) from n = 16 cells per cell line from three independent experiments. Data points corresponding to the examples are highlighted. Statistical variation is shown as scatter plots with the indication of mean ± SEM and significance calculated using the non-parametric Mann–Whitney test, *p < 0.05.