GDAP1 loss of function inhibits the mitochondrial pyruvate dehydrogenase complex by altering the actin cytoskeleton

Wolf, Christina; Pouya, Alireza; Bitar, Sara; Pfeiffer, Annika; Bueno, Diones; Rojas-Charry, Liliana; Arndt, Sabine; Gomez-Zepeda, David; Tenzer, Stefan; Bello, Federica Dal; Vianello, Caterina; Ritz, Sandra; Schwirz, Jonas; Dobrindt, Kristina; Peitz, Michael; Hanschmann, Eva-Maria; Mencke, Pauline; Boussaad, Ibrahim; Silies, Marion; Brüstle, Oliver; Giacomello, Marta; Krüger, Rejko; Methner, Axel

doi:10.1038/s42003-022-03487-6

Download PDF

Article
Open access
Published: 03 June 2022

GDAP1 loss of function inhibits the mitochondrial pyruvate dehydrogenase complex by altering the actin cytoskeleton

Communications Biology volume 5, Article number: 541 (2022) Cite this article

2970 Accesses
8 Citations
6 Altmetric
Metrics details

Subjects

Abstract

Charcot-Marie-Tooth (CMT) disease 4A is an autosomal-recessive polyneuropathy caused by mutations of ganglioside-induced differentiation-associated protein 1 (GDAP1), a putative glutathione transferase, which affects mitochondrial shape and alters cellular Ca²⁺ homeostasis. Here, we identify the underlying mechanism. We found that patient-derived motoneurons and GDAP1 knockdown SH-SY5Y cells display two phenotypes: more tubular mitochondria and a metabolism characterized by glutamine dependence and fewer cytosolic lipid droplets. GDAP1 interacts with the actin-depolymerizing protein Cofilin-1 and beta-tubulin in a redox-dependent manner, suggesting a role for actin signaling. Consistently, GDAP1 loss causes less F-actin close to mitochondria, which restricts mitochondrial localization of the fission factor dynamin-related protein 1, instigating tubularity. GDAP1 silencing also disrupts mitochondria-ER contact sites. These changes result in lower mitochondrial Ca²⁺ levels and inhibition of the pyruvate dehydrogenase complex, explaining the metabolic changes upon GDAP1 loss of function. Together, our findings reconcile GDAP1-associated phenotypes and implicate disrupted actin signaling in CMT4A pathophysiology.

Messenger RNA transport on lysosomal vesicles maintains axonal mitochondrial homeostasis and prevents axonal degeneration

Article Open access 10 April 2024

Molecular and cellular mechanisms of selective vulnerability in neurodegenerative diseases

Article 04 April 2024

Native-state proteomics of Parvalbumin interneurons identifies unique molecular signatures and vulnerabilities to early Alzheimer’s pathology

Article Open access 01 April 2024

Introduction

Charcot-Marie-Tooth (CMT) disease is the most frequently inherited peripheral neuropathy in humans and affects one in 2500 people. Clinically, this group of diseases can be distinguished by mode of inheritance, age of onset, and by electrophysiological characteristics that distinguish demyelinating and axonal forms. Mutations in the gene GDAP1 (ganglioside-induced differentiation-associated protein 1) cause various forms of CMT: the most frequent recessively inherited demyelinating subtype CMT4A¹, the axonal-recessive (AR)-CMT2², the intermediate-recessive subtype CMTRIA³, and the dominant subtype CMT2K⁴.

GDAP1 is located in the mitochondrial outer membrane facing the cytosol. It possesses two glutathione transferase (GST)-like domains^2,5,6 and a C-terminal hydrophobic anchor which crosses the outer mitochondrial membrane⁷. Whether GDAP1 is a catalytically active GST remains controversial^6,8,9,10, but it appears clear that GDAP1 can bind GST substrates^9,10.

One potential mechanism of GDAP1 action that has been connected to CMT4A disease is mitochondrial dynamics. Over-expression of GDAP1, but not over-expression of GDAP1 bearing recessive disease-causing mutations, results in more fragmented mitochondria, whereas a GDAP1 knockdown (KD) results in mitochondrial elongation¹¹. This GDAP1-mediated mitochondrial fragmentation depends on the activity of dynamin-related protein 1 (DRP1), the major mitochondrial fission factor^12,13. The mechanism behind this is unknown. DRP1 is a cytosolic protein which, when recruited to mitochondria, promotes fission by forming ring-like oligomers that constrict and divide mitochondria¹⁴. The recruitment of DRP1 to mitochondria depends on the presence of ER tubules¹⁵ and filamentous actin (F-actin)¹⁶ at the sites of constriction. The ER-protein inverted formin 2 (INF2) promotes actin polymerization which occurs before DRP1-driven constriction¹⁷. INF2-mediated actin polymerization also increases mitochondria-ER contact sites (MERCS)¹⁸. This further connects actin polymerization and GDAP1 function as mitochondrial and ER marker proteins colocalize less frequently in GDAP1 KD cells¹⁹. Interestingly, INF2 mutations also cause CMT disease (CMTDIE)²⁰, implying that actin polymerization is important for the survival of peripheral nervous system neurons, similar to GDAP1 function. Another protein at the interface of F-actin polymerization and DRP1 recruitment is Cofilin-1. Cofilin-1 binds to monomeric actin and F-actin and controls cytoskeletal dynamics mostly by actin depolymerization. Cofilin-1 deletion results in DRP1 accumulation at mitochondria and fragmentation²¹. In summary, defective actin polymerization affects the same mitochondrial fission pathway upstream of DRP1 and results in a clinical phenotype similar to GDAP1 mutation. Whether F-actin polymerization and its regulation play a role in CMT4A is still poorly understood.

Another process that is affected by loss of GDAP1 is cellular Ca²⁺ homeostasis. Neuronal GDAP1 KD reduces Ca²⁺ influx from the extracellular space that follows depletion of the ER Ca²⁺ stores, so-called store-operated Ca²⁺ entry (SOCE), possibly due to an impaired mitochondrial localization at subplasmalemmal microdomains¹⁹. GDAP1 KD also blunts the Ca²⁺-dependent increase of mitochondrial respiration upon SOCE²². Mitochondrial Ca²⁺ levels increase the activity of several enzymes of the tricarboxylic acid cycle (TCA) like pyruvate dehydrogenase (PDH)²³, the key enzyme of the pyruvate dehydrogenase complex (PDC) which catalyzes the conversion of pyruvate to acetyl-CoA and links glycolysis to the TCA. This connection between Ca²⁺ and TCA activity is thought to connect mitochondrial activity to ATP demand. Changes in the activity of the PDC have not been studied yet in CMT4A or in cells with perturbed GDAP1 expression but, interestingly, a pathogenic mutation of pyruvate dehydrogenase kinase isoenzyme 3 (PDK3) that inhibits the PDC also causes CMT disease, CMTX6²⁴. How GDAP1-mediated changes of the mitochondrial Ca²⁺ homeostasis are connected to its fission activity is still unclear.

In this study, we used motoneurons obtained from CMT4A-patient-derived induced-pluripotent stem cells and neuronal GDAP1 KD cells to study the pathophysiology of CMT4A. We found that GDAP1 interacts with actin-binding Cofilin-1. Loss of GDAP1 results in a reduction of F-actin fibers in mitochondrial proximity, which restricts DRP1 access to mitochondrial constriction sites and disrupts mitochondria-ER contact sites. This reduces mitochondrial Ca²⁺ levels and inhibits the PDC resulting in a rewired cellular metabolism characterized by glutamine dependence and increased consumption of fatty acids. Together, these findings implicate disrupted F-actin signaling in CMT4A pathophysiology.

Results

More tubular mitochondria and an increased mitochondrial membrane potential in CMT4A patient-derived neuronal cells and GDAP1 knockdown cells

To establish a model for CMT4A, we compared GDAP1 KD SH-SY5Y cells^19,25 with neuronal cells derived from CMT4A patients. Patient CMT#1 is a 25-year-old ambulant male with a compound heterozygosity (L239F/R273G) of mutations in the C-terminal GST domain of GDAP1 and patient CMT#2 is a 40-year-old wheelchair-bound male with a homozygous mutation of the intron 4 splice donor site (c.579 + 1G>A). This mutation causes skipping of exon 4 leading to a frameshift and a truncated protein lacking the C-terminal GST and the transmembrane domain of GDAP1 (Fig. 1a).

**Fig. 1: More tubular mitochondria and an increased mitochondrial membrane potential in CMT4A patient-derived neuronal cells and GDAP1 knockdown cells.**

We generated induced-pluripotent stem cells from fibroblast cell lines²⁶ from these patients using non-integrative expression of the Yamanaka factors and differentiated them to neuronal precursor cells (NPCs) (Fig. 1b) that express the neuronal marker protein β-III tubulin (Fig. 1c). In contrast to fibroblasts²⁶, control and CMT#1 NPCs express GDAP1 detectable by immunoblotting (Fig. 1d). NPCs from CMT#2 lacked GDAP1 expression (Fig. 1d). Because the polyclonal antiserum targets an antigen which should still be present in the patient, this is probably due to nonsense-mediated mRNA decay or degradation of the truncated protein.

We quantified mitochondrial shape and membrane potential (Δψm) using automated high-content confocal microscopy analysis of cells simultaneously stained with the fluorescent dyes mitotracker and tetramethylrhodamine methyl ester (TMRM). Patient-derived neuronal cells contained significantly more tubular mitochondria (Fig. 1e) with a more negative mitochondrial membrane potential (Fig. 1f). Applying the same methodology to GDAP1 KD cells (Fig. 1g), we found similar changes; more elongated mitochondria (Fig. 1h) and a significantly more negative membrane potential (Fig. 1i). The consistency between the phenotypes of patient-derived cells and GDAP1 KD cells suggests that GDAP1 KD cells are a suitable model to study CMT4A disease.

GDAP1 knockdown uncouples mitochondrial respiration from ATP generation

Mitochondria produce ATP by consuming oxygen and the energetic electron donors NADH and FADH₂ in a process called oxidative phosphorylation. NADH is provided by the TCA, PDH and β-oxidation. To assess the effect of GDAP1 KD on this all-important process, we measured mitochondrial oxygen consumption using high-resolution respirometry and found a higher mitochondrial routine respiration in GDAP1 KD cells (Fig. 2a). The maximal capacity of the electron transfer system (ETS), determined by titrating in the uncoupler FCCP, was however similar in both cell lines (Fig. 2a). Normalization of the respiratory states to the maximum ETS capacity, the so-called flux control ratio, revealed that the GDAP1 KD cells use a higher fraction of their maximal capacity for routine, leak and phosphorylating respiration (Fig. 2a′). An increased ratio between non-phosphorylating leak respiration (electron flow coupled to proton pumping to compensate for proton leaks) and ETS capacity suggests intrinsic uncoupling or dysfunction in GDAP1 KD cells. Quantification of ATP content using the ratiometric reporters BTeam targeted either to the cytosol or to the mitochondrial matrix²⁷ also demonstrated a reduced mitochondrial ATP generation in GDAP1 KD cells despite the increased Δψm and routine respiration (Fig. 2b). Taken together, these data suggest that GDAP1 KD results in the uncoupling of mitochondrial respiration and ATP generation via oxidative phosphorylation.

GDAP1 knockdown shifts cellular metabolism towards glutaminolysis

To identify metabolic pathways in GDAP1 KD cells that compensate for the diminished capacity for oxidative phosphorylation, we quantified the expression levels of various metabolic enzymes in control and KD cells by immunoblotting. This revealed a significant higher protein levels of lactate dehydrogenase (LDHA) and glutamate dehydrogenase 1 (GLUD1) in GDAP1 KD cells (Fig. 2c, Supplementary Fig. 1). LDHA catalyzes the interconversion of pyruvate to lactate. GLUD1, in contrast, converts glutamate to α-ketoglutarate, the precursor of succinyl-CoA in the TCA. In line with the increased expression of GLUD1, GDAP1 KD cells consume significantly more glutamate, the substrate of GLUD1, and more glutamine which can be converted to glutamate by glutaminase (Fig. 2d). We also observed an attenuated decline in cytosolic ATP content in KD cells after treatment with the glucose antimetabolite 2-deoxyglucose (Fig. 2e) in line with a reduced use of glucose as the major fuel for the TCA cycle. The TCA metabolite upstream of α-ketoglutarate is citrate, which is produced by the transfer of acetyl-CoA to oxaloacetate. Acetyl-CoA is the product of the PDC in the mitochondrial matrix or of mitochondrial β-oxidation. We suspected a dysfunction of the TCA cycle in GDAP1 KD cells at the level of the PDC and therefore quantified the amount of lipid droplets as alternative sources for acetyl-CoA via β-oxidation in GDAP1 KD cells. We quantified the amount of lipid droplets by BODIPY C₁₂ staining and indeed found decreased cytosolic lipid droplet levels in GDAP1 KD cells, whereas lipid droplets associated with mitochondria were unchanged (Fig. 2f). Maximum oxygen consumption of GDAP1 KD cells indeed tended to be higher in the presence of the fatty acid oxidation substrate palmitate (Fig. 2g). Together, these results are in line with a high lipid catabolism that serves to replenish fatty acids and consequently acetyl-CoA levels in the TCA cycle. We posit that the increased demand of glutamine and fatty acids in GDAP1 KD probably serve as compensatory mechanisms, clearly pointing towards a dysfunction of the TCA cycle in GDAP1 KD cells at the level of the PDC (summarized in Fig. 2h).

Hyperphosphorylated pyruvate dehydrogenase and reduced mitochondrial Ca²⁺ levels in GDAP1 KD cells

Increased mitochondrial Ca²⁺ levels activate the PDC by stimulating PDH phosphatase²⁸. Ca²⁺ also activates isocitrate dehydrogenase which is upstream of α-ketoglutarate, the product of GLUD1, and α-ketoglutarate dehydrogenase²⁹. Based on the reported attenuated mitochondrial respiration upon SOCE in GDAP1 KD cells²², we suspected altered PDH phosphorylation levels driven by changes in mitochondrial Ca²⁺ content as the reason for PDC inhibition (see scheme in Fig. 3a). To test this idea, we probed phosphorylation of serine 293 of the E1 PDH subunit, which has been directly linked to PDC activity³⁰. Immunoblotting showed a significantly increased phosphorylation of PDH E1 serine 293 in GDAP1 KD cells as compared to total E1 PDH (Fig. 3b). This is in line with a decreased activity of the PDC. We then measured mitochondrial Ca²⁺ levels by imaging live cells stained with Rhod2-AM or expressing the genetically encoded mitochondrial Ca²⁺ sensor mito-CEPIA³¹ normalized to mito-FarRed. The Ca²⁺ sensor GEM-CEPIA1er targeted to the ER served as a control. Both methods revealed a reduction in steady-state mitochondrial Ca²⁺ levels in GDAP1 KD cells, whereas ER calcium levels were not affected (Fig. 3c). We conclude that the PDC malfunction in GDAP1 KD cells is likely caused by a reduction in mitochondrial Ca²⁺ levels resulting in an increased phosphorylation of PDH.

Patient-derived cells are similarly characterized by increased glutaminolysis, reduced lipid droplets and reduced mitochondrial Ca²⁺ levels

We next tested whether the findings of reduced mitochondrial Ca²⁺ levels and increased glutaminolysis also apply to patient-derived cells. We set out to specifically test this in motoneurons, a cell type affected by CMT4A. Motoneurons were differentiated from NPCs following established protocols (Fig. 4a) and motoneuronal identity was confirmed by immunostaining against the dendrite marker MAP2 and the motoneuronal neurofilament H marker antibody Smi32 (Fig. 4b). Control and patient-derived motoneurons stained similarly for these markers as shown by high-content imaging (Supplementary Fig. 2). These motoneurons express GDAP1 similar to NPCs with cells from patient CMT#2 lacking GDAP1 expression (Fig. 4c). High-content microscopy of MitoTracker-stained cells revealed an increased area occupied by mitochondria (Fig. 4d). Immunoblotting reproduced the increased expression of GLUD1 seen upon GDAP1 KD (Figs. 4e, 2c). Motoneurons differentiated from patient-derived cells also increased the consumption of the GLUD1-substrate glutamate (Figs. 4f, 2d) and its precursor glutamine. Furthermore, motoneurons showed a decrease in lipid droplets normalized to the nuclear area (Fig. 4g), indicative of increased fatty acid consumption. Together, these data further highlight the similarity between patient-derived cells and GDAP1 KD cells.

**Fig. 4: Patient-derived cells are similarly characterized by an anaplerotic state and reduced mitochondrial Ca²⁺ levels.**

Because motoneuronal cells showed a low transfection efficiency which made it difficult to identify single neurons in the cultures, we reverted to the patient-derived NPCs instead of fully differentiated motoneurons to assess the mitochondrial Ca²⁺ phenotype. We measured lower resting Ca²⁺ levels with Rhod2-AM and mito-CEPIA (Fig. 4h) in NPCs, similar to KD cells. Patient-derived NPCs also had increased levels of pPDH as compared to total PDH levels (Fig. 4i). We conclude that neuronal cells from patients with GDAP1 mutation not only have the same changes in mitochondrial shape and membrane potential as GDAP1 KD cells (Fig. 1) but also feature the same metabolic changes and alterations of Ca²⁺ levels. Together this suggests that GDAP1-mediated CMT4A is caused by inhibition of PDH activity resulting in an anaplerotic state.

GDAP1 knockdown reduces the number of contact sites between mitochondria and the endoplasmic reticulum

We suspected that changes in MERCS might underlie the changes in mitochondrial Ca²⁺ levels especially as a reduced colocalization of ER and mitochondrial markers hint to changes in MERCS, as shown previously¹⁹. MERCS are hot spots of interactions between the ER and mitochondria defined by a distance of mitochondrial and ER membranes between ~10 and ~50 nm³² and are important signaling hubs (reviewed in ref. ³³). Using transmission electron microscopy, we indeed found (i) a decrease in MERCS in GDAP1 KD cells defined as the percentage of mitochondrial perimeter covered by the ER, (ii) a decreased number of MERCS per mitochondrion, (iii) a decreased length of MERCS (Fig. 5a) and (iv) an increased width between the ER and mitochondria (Fig. 5a, b). As an additional readout to quantify the distance between mitochondria and the ER, we used a well-established assay based on Förster resonance emission transfer (FRET) between two fluorescent proteins targeted to the surface of mitochondria and the ER, both facing the cytosol^34,35. The lower FRET ratio (Fig. 5c) confirmed the data obtained by the morphometric analysis with an increased width between the organelles in GDAP1 KD cells. Comparing the amount of ER protein present in mitochondrial cell fractions using immunoblotting of the membrane ER protein Sec62 and the mitochondrial protein Cox4 also confirmed the reduction in contact sites (Fig. 5d). Such a reduction and increased distance should result in decreased Ca²⁺ transfer from the ER to the mitochondrial matrix. We studied this by triggering ER Ca²⁺ release with the cholinergic agent carbachol, which activates inositol trisphosphate receptors (IP3Rs) leading to Ca²⁺ release into the cytosol, generating Ca²⁺ microdomains at MERCS, which sustain rapid Ca²⁺ uptake by mitochondria^34,36. Surprisingly, we did not observe a difference in mitochondrial Ca²⁺ uptake measured by mitochondrially-targeted aequorin upon treatment with carbachol (Fig. 5e). The direct reducing effect of carbachol on PDH E1 phosphorylation was also similar in both cell lines while the baseline levels were significantly increased in GDAP1 KD cells (Fig. 5f). Apparently, upregulated MCU levels (Fig. 5g) and hyperpolarization (see Fig. 1I) can compensate for the increased distance between the organelles under conditions of high Ca²⁺ influx. The other components of the ER-mitochondrial Ca²⁺ uptake complex³⁷, the IP3R ER Ca²⁺ release channels, the molecular chaperone glucose-regulated protein 75 (Grp75, official name Heat Shock Protein Family A (Hsp70) Member (HSPA)), and the mitochondrial voltage-dependent anion channel 1 (VDAC1) at the outer mitochondrial membrane were however similarly expressed in GDAP1 KD and control cells (Fig. 5h). Taken together our findings support altered MERCS with an increased distance between the ER and mitochondria in GDAP1 KD cells.

Increased levels of proteins involved in β-oxidation in GDAP1 KD cells

To clarify how GDAP1 KD affects MERCS and mitochondrial Ca²⁺ levels, we next set out to identify proteins that are dysregulated in KD cells by label-free quantitative liquid chromatography coupled to mass spectrometry (LC-MS)³⁸. Comparing the proteome of control and GDAP1 KD cells identified 985 significantly dysregulated proteins out of 3914 proteins—456 proteins were found to be less abundant and 529 to be more abundant (Fig. 6a). Then, we submitted the list of upregulated proteins to STRING³⁹ and observed that the following KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways were over-represented with an false-detection rate below 2 × 10⁻⁴: protein processing in the ER, lysosome, regulation of the actin cytoskeleton, endocytosis, fatty acid degradation and citrate cycle (Fig. 6b). The analysis corroborated the increased levels of LDHA and GLUD1 observed by immunoblotting (Fig. 2c). It also demonstrated increased levels of proteins involved in β-oxidation (Fig. 6c) like carnitine palmitoyltransferase 1 (CPT1A) which catalyzes the transfer of the acyl group of long-chain fatty acid-CoA conjugates onto carnitine to transport the resulting acylcarnitine long-chain fatty acids inside the mitochondria. These results suggest that the decrease in lipid droplets observed in GDAP1 KD cells (Fig. 2f) indeed reflects a high lipid catabolism that serves to provide the TCA cycle with acetyl-CoA.

Fig. 6: LC-MS proteomics identifies proteins involved in β-oxidation as being more abundant in *GDAP1* KD cells and hints to alterations of the actin cytoskeleton mediated by redox-dependent interactions with GDAP1.

GDAP1 interacts with proteins of the actin cytoskeleton in a redox-dependent manner

We then used label-free proteomics to identify GDAP1-interacting proteins to get a more complete picture of how GDAP1 loss of function affects cellular metabolism. We transduced primary cortical cultures from mice expressing the biotin ligase BirA⁴⁰ with adeno-associated viruses expressing GDAP1 tagged with an Avi-tag, a specific substrate for BirA⁴¹. Only Avi-HA-tagged GDAP1 but not the HA-tagged control was biotinylated (Fig. 6d). By treating the mitochondrial preparations with GSH or GSSG as described⁴², we further aimed to identify proteins whose interaction is affected by their specific redox state. This was driven by the fact that GDAP1 can bind potential GST substrates^9,10 and the assumption that such an interaction might depend on the local redox environment. A total of 268 proteins were pulled down from Avi-HA but not or less from HA-GDAP1-transduced cultures. 176 of these proteins showed a statistically different abundance in preparations treated with GSH compared to vehicle and 106 in preparations treated with GSSG. For 72 proteins the interaction with GDAP1 was altered only by GSH and not by GSSG treatment (Fig. 6e). Interestingly, the most strongly and most significantly regulated proteins can be linked to the cytoskeleton like tubulins (Fig. 6f, green), crucial components of the cytoskeleton that can serve as a scaffold for mitochondrial transport⁴³. This is consistent with regulation of the actin cytoskeleton being one of the top hits in the KEGG analysis of proteins upregulated in GDAP1 KD cells (Fig. 6b, Supplementary fig. 3). In addition, an interaction between GDAP1 and β-tubulin TUBB has been reported in a yeast-two-hybrid experiment⁴⁴. This led us to further concentrate our subsequent analysis on a potential misregulation in actin dynamics in CMT4A-disease models.

GDAP1 knockdown reduces Cofilin-1 abundance at mitochondria

The actin-regulatory protein Cofilin-1 (Fig. 6g) specifically caught our attention because it affects mitochondrial shape by inducing depolymerization of actin filaments in mitochondrial proximity which restricts access of DRP1 to mitochondria²¹. Its antagonist is the ER-anchored INF2, which also causes CMT disease when mutated²⁰. We further explored the identified interaction between GDAP1 and Cofilin-1, and hypothesized that GDAP1 alters actin abundance, polymerization or both at mitochondrial sites by interacting with Cofilin-1. The actin-binding ability of Cofilin-1 is inhibited by phosphorylation of serine 3 which is controlled by the presence of intramolecular disulfide bridges^45,46 or protein glutathionylation⁴⁷ and could thus be a target of GDAP1’s GST activity. Alternatively, GDAP1 could just recruit Cofilin-1 to the mitochondrial surface in a redox-dependent manner. We studied Cofilin-1 abundance and intracellular localization and phosphorylation in GDAP1 KD and control cells. GDAP1 KD increased the total abundance of Cofilin-1 but did not affect phosphorylation (Fig. 7a). However, when we compared the protein levels in mitochondrial fractions, we observed the opposite, a significant reduction in GDAP1 KD cells (Fig. 7b). To further test for the amount and localization of Cofilin-1 close to mitochondria, we used HSPA9, a mitochondrial matrix protein, as a marker as it was not regulated in GDAP1 KD cells (Fig. 5h) and plays a role in ER-mitochondrial Ca²⁺ transfer³⁷. Confocal microscopy demonstrated an increase in the area covered by mitochondria (Fig. 7c and d) in line with the more tubular network in GDAP1 KD cells (Fig. 1h) and a reduction of mitochondrial Cofilin-1 compared to total Cofilin-1 in GDAP1 KD cells (Fig. 7e), in line with the fractionation assays. In addition, the distance of Cofilin-1 spots from mitochondria was significantly increased (Fig. 7f) and the colocalization of the surfaces of Cofilin-1 and HSPA9-expressing structures was reduced (Fig. 7g). These results imply that GDAP1 controls the presence of Cofilin-1 in proximity to the mitochondrial surface, possibly at MERCS.

**Fig. 7: Reduced presence of Cofilin-1 at mitochondria.**

GDAP1 KD reduces F-actin fibers at mitochondrial surfaces and restricts access of DRP1 to the mitochondria resulting in less DRP1 and MFF

Because Cofilin-1 is an actin-binding protein, we suspected less actin in the proximity of mitochondria in GDAP1 KD cells. We quantitated F-actin levels and its colocalization with mitochondria in living cells by transiently transfecting control and GDAP1 KD cells with GFP-tagged F-tractin and labeling of mitochondria with mitotracker red. F-tractin does not perturb actin rearrangement^48,49,50 and was previously used to study the effects of INF2 perturbation on mitochondrial dynamics^18,51. Our data revealed a significant reduction of surface colocalization between F-actin and mitochondria in GDAP1 KD cells (Fig. 8a) in line with our hypothesis. As expected from previous work¹⁷, this severely reduced DRP1 localization in mitochondrial fractions (Fig. 8b). The decreased colocalization of DRP1 with mitochondria was also evident when we overexpressed GFP-tagged DRP1 in control and GDAP1 KD cells (Fig. 8c). Interestingly, DRP1 also covered less space in GDAP1 KD cells (Fig. 8d). To rule out that this is caused by a reduced transfection efficiency, we quantified total endogenous DRP1. This also revealed a reduced abundance of DRP1 and its receptor at the mitochondrial surface, mitochondrial fission factor (MFF) (Fig. 8e). In summary, these findings imply that in the absence of GDAP1, F-actin fibers are less present at the mitochondrial surface which restricts the access of ER tubules and DRP1 to sites of mitochondrial constriction and decreased levels of DRP1 and its receptor MFF resulting in a dysfunction of mitochondrial dynamics (Fig. 8f).

Discussion

In this work, we used neuronal GDAP1 KD in human SH-SY5Y neuroblastoma cells and patient-derived cells to study the pathophysiology of CMT4A. We found that GDAP1 interacts with the actin-interacting protein Cofilin-1 and that loss of GDAP1 results in a reduction of F-actin fibers in mitochondrial proximity. This limits the access of ER tubules to mitochondria which connects two processes: it impedes DRP1 recruitment to mitochondrial constriction sites resulting in more tubular mitochondria and it disrupts mitochondria-ER contact sites causing reduced mitochondrial Ca²⁺ levels. The reduced mitochondrial Ca²⁺ levels inhibit the PDC and result in a rewired cellular metabolism characterized by dependence on glutamine and fatty acids to compensate for an impaired TCA cycle. We therefore conclude that the reduction in F-actin presence at mitochondria caused by GDAP1 loss of function represents the probable cause of autosomal-recessive CMT4A.

The reduced mitochondrial Ca²⁺ levels in GDAP1 KD and GDAP1-loss-of-function NPCs are difficult to understand because over-expression of the MCU was shown previously to increase the matrix Ca²⁺ concentration⁵². Other, unchanged components of the ER-mitochondrial complex that regulate Ca²⁺ transfer, such as the ER release channels, the mitochondrial VDAC1 channels and the linking protein HSPA9 (Grp75)³⁷ (Fig. 5h) might account for this. In addition, the amount and stoichiometry of the MCU-interacting proteins MICU1 and MICU1, which are regulatory subunits of the large protein complex which mediates mitochondrial Ca²⁺ uptake, also affect the matrix Ca²⁺ concentration⁵² and were not further studied here as it is difficult to fathom how the outer mitochondrial membrane protein GDAP1 could affect or modify these proteins. Despite the reduction in MERCs in GDAP1 KD cells, agonist-induced increases in mitochondrial Ca²⁺ levels were unchanged. Such Ca²⁺ transients are believed to instigate the increase in PDC activity^28,29. We therefore have to conclude that the reported blunting of the Ca²⁺-dependent increase of mitochondrial respiration upon SOCE²² also affects PDC activity. In line with this, it was recently shown that in mouse Gdap1−/− motoneurons, glutamate treatment results in a decreased decay of the Ca²⁺ signal and a reduction in respiration⁵³. The authors attributed this to defects in mitochondrial movement resulting in a lack of correct positioning of mitochondria at sites where an intense stimulation of ATP production by oxidative phosphorylation is required like the plasma membrane. Interestingly, the actin cytoskeleton is essential for short-distance mitochondrial movements and for immobilization of mitochondria at the actin cortex, a specialized layer of cytoplasmic proteins on the inner face of the cell membrane (reviewed in⁵⁴) suggesting that the changes described here could also underlie these defects in mitochondrial positioning. Alternatively, other yet unknown mechanisms exist that link the actin cytoskeleton to steady-state mitochondrial Ca²⁺ levels.

We found that GDAP1 interacts with Cofilin-1 in a redox-dependent manner and restricts its presence at the mitochondrial surface. This was shown by confocal microscopy using HSPA9 as a marker protein. HSPA9 (Grp75) is recognized to be enriched in MERCS. An enrichment of Cofilin-1 in MERCs has not been reported yet and we reproduced the reduced abundance of Cofilin-1 in GDAP1 KD cells in fractions containing crude mitochondria (Fig. 7b). We therefore conclude that the reduction of Cofilin-1 at mitochondrial surfaces is not restricted to MERCS. Very recently, Cofilin-1 was shown to affect mitochondrial shape and function⁵⁵. Theoretically, GDAP1 could affect Cofilin-1 function through its still unresolved potential GST-like enzymatic activity^6,8,9,10 because Cofilin-1 contains four potential GST target cysteine residues at the positions 39, 80, 139 and 147. Cysteines 39 and 80 are buried inside the protein while cysteines 139 and 147 are located on the surface of the protein⁵⁶. Redox-mediated modifications of these cysteine residues clearly affect the function of Cofilin-1. Treatment with hydrogen peroxide leads to the formation of an intramolecular disulfide bond resulting in a conformational change that prevents phosphorylation and thereby actin polymerization⁴⁵. Moreover, intermolecular disulfides and oligomeric forms of Cofilin-1 have been described. Monomeric Cofilin-1 possesses severing activity, whereas the dimeric and oligomeric forms have actin-bundling activity⁵⁷. Interestingly, only monomeric Cofilin-1 is phosphorylated⁵⁸ which represents the most-studied post-translational modification of Cofilin-1 to date. Cofilin-1 can be post-translationally modified by phosphorylation of serine 3. Increased phosphorylation inhibits the actin-binding ability of Cofilin-1⁵⁹; dephosphorylated Cofilin-1 preferentially localizes to mitochondria, whereas a mutation mimicking the phosphorylated protein prevents translocation from the cytosol to mitochondria⁶⁰. We found no changes in Cofilin-1 phosphorylation in whole cell lysates and no phosphorylated Cofilin-1 in mitochondrial fractions. Upon oxidation of all four cysteine residues and dephosphorylation at serine 3, Cofilin-1 loses its affinity for actin, translocates from the cytosol to mitochondria and induces apoptosis^60,61. Oxidation of methionine 115 apparently also prevents its actin depolymerization activity and induces forced mitochondrial translocation and apoptosis⁶². In addition, Cofilin-1 can be glutathionylated, which represents the presumed enzymatic activity of GDAP1. Glutathionylation of Cofilin-1 was demonstrated in lymphocytes treated with the thiol-oxidizing reagent diamide⁶³ and in cells of the rat nucleus accumbens during cued cocaine seeking in the absence of cocaine⁴⁷. Interestingly, glutathionylation reduces cofilin-1-dependent depolymerization of F-actin, implying regulatory functions in cell signaling⁴⁷. It remains to be clarified whether GDAP1 affects the redox state and function of Cofilin-1 by specific glutathionylation.

Mutations in other proteins that target the same pathway also cause CMT disease. Mutated PDK3 also inhibits the PDC by hyperphosphorylation of the PDH E1α subunit similar to GDAP1 loss of function. This causes CMTX6²⁴. Loss of function of the actin-polymerizing protein INF2 alters mitochondrial shape by changing DRP1 access to mitochondria similar to GDAP1 KD¹⁷. INF2 is also implicated in the abundance of MERCS and changes in mitochondrial Ca²⁺ dynamics¹⁸ which we also observed in GDAP1 loss-of-function cells. Mutations in INF2 also cause CMT disease, CMTDIE²⁰. Mutations in Dynamin-2 (DNM2), a ubiquitously expressed large GTPase that interacts tightly with the actin and microtubule network^64,65 also cause CMT disease, dominant intermediate CMT (CMTDIB)⁶⁶ and autosomal-dominant CMT2M⁶⁷. Similar to GDAP1, DNM2 works in concert with DRP1 to orchestrate mitochondrial constriction events that result in division⁶⁸. DNM2 therefore directly interferes with the same processes that we found to be affected by GDAP1 loss of function. Finally mutations in BSCL2 (Bernardinelli-Seip Congenital Lipodystrophy Type 2) also known as seipin cause a distal hereditary motor neuronopathy (HMN5C) which can present with features of axonal CMT2⁶⁹. BSCL2 is an ER protein crucial in the formation of lipid droplets^70,71. BSCL2 also interacts with the adaptor protein YWHAB (14-3-3-beta) which then recruits cofilin-1 to remodel the actin cytoskeleton for adipocyte differentiation⁷².

In summary, our results shed light on the pathophysiology of CMT4A and highlight the importance of GDAP1 and actin for mitochondrial function. The fact that mutations in diverse proteins that all interact with the overarching mechanisms described here—metabolic and mitochondrial remodeling caused by changes in the the interaction of the actin cytoskeleton with mitochondria—cause CMT disease strengthens the importance of this mechanism for the health of the peripheral nervous system.

Methods

Generation of iPSCs

iPSC lines from two CMT4A patients and one healthy donor were generated with Sendai virus (CytoTune, DNAVEC Corporation) coding for POU5F1, SOX2, KLF4 and MYC. All subjects gave written informed consent prior and the study was approved by the local ethics committee at the Universities of Warsaw (patient #1), Düsseldorf (patient #2) and Bonn (healthy donor). In brief, Sendai virus-infected primary fibroblasts were immediately centrifuged for 45 min at 32 °C with 1500 g (spinfection) and cultivated in Advanced DMEM containing 5% fetal calf serum (FCS) and 1% GlutaMAX™ (all from Life Technologies). On the following day the virus-containing medium was replaced with fresh culture medium. Five days post infection (d5), transduced fibroblasts were trypsinized and seeded onto mouse feeder-coated dishes in DMEM/F-12 containing 10% KnockOut™ Serum Replacement, 1% nonessential amino acids (NEAA), 1% GlutaMAX™, 1% pyruvate, 0.1 mM β-mercaptoethanol and 10 ng/ml basic fibroblast growth factor (bFGF) (all from Life Technologies). Medium was changed every other day until clonal iPSC colonies were manually picked and adapted to feeder-free culture conditions. Several clonal lines were subjected to SNP genotyping in order to identify iPSC clones with normal karyotype.

SNP analysis of iPSC lines

Genomic DNA was prepared using the DNeasy Blood & Tissue Kit (Qiagen). Whole-genome single nucleotide polymorphism (SNP) genotyping was performed at the Institute of Human Genetics at the University of Bonn. Genomic DNA at a concentration of 60 ng/µL was used for whole-genome amplification. Afterwards, the amplified DNA was fragmented and hybridized to sequence-specific oligomers bound to beads on an Illumina HumanOmniExpress-12v1.0 chip. Data were analyzed using Illumina Bead Studio.

Differentiation of hiPSCs into neural cells

Induced-pluripotent stem cells were grown using mTeSR™1 medium (Stemcell Tech., 05850) on matrigel (Corning, 354277). Human iPS cells were detached from matrigel using ReLeSR^TM (Stemcell Tech., 05872), resuspended in medium consisting of Dulbecco’s modified Eagle’s medium/ F12 supplemented with 10% knockout serum (Invitrogen, 10828-028), 1% N2 supplement (Gibco, 17502-048), 0.05% B27 (Gibco, 17504-044), 20 ng/ml epidermal growth factor (Sigma, E9644) and 10 ng/ml basic fibroblast growth factor (Gibco, PHG0024), and plated within a low-attachment petri dish to induce embryoid body (EB) formation for 4 days. The EBs were plated on polyornithine (Sigma, P3655)/laminin (Sigma, L2020)-coated dishes for additional 6 days to induce the formation of neural rosettes. Neural rosettes were then manually removed, dissociated with accutase (Stemcell Tech., 07920), plated on poly-L-ornithine/laminin-coated dishes, and then treated with 3 µM retinoic acid for 7 days. The medium was changed daily and cultures were passaged weekly by accutase and plated on matrigel-coated dishes in the above-mentioned neural medium.

Small molecule differentiation of hiPSCs into motoneurons

iPSCs differentiation was performed by the addition of small molecules adapted from a previously described protocol⁷³. Briefly, for neuronal induction, iPSCs were seeded as colonies resuspended in E8 medium 2 days prior differentiation. To start the differentiation, medium was changed to N2B27 medium (50% Neurobasal medium/ 50% DMEM-F12 medium with 1:200 N2 supplement, 1:100 B27 supplement without vitamin A and 1% penicillin, streptomycin and L-glutamine respectively) supplemented with the small molecules 10 µM SB-431542 (SB), 1 µM dorsomorphin (DM), 3 µM CHIR 99021 and 0.5 µM PMA for four days. Subsequently, SB and DM were replaced by 150 µM Ascorbic Acid (AA), and cells were fed daily until epithelium-like structures emerged. Neural epithelial structures were picked, dissociated mechanically and plated on 12-well plates, which were coated with Matrigel (1:100 in DMEM/F12) overnight. After around five passages the smNPC cells reached a high purity and were further kept in culture on Matrigel-coated plates and N2B27 medium supplemented with CHIR, PMS and AA. Detaching of the cells for passaging was performed with accutase. Starting from smNPC passage 13, differentiation to MN could be initiated. N2B27 medium with 1 µM PMA was added 3 days after passaging. After two more days, 1 µM retinoic acid (RA) and 1 µM PMA were supplemented to the medium for the following 8 days, until culturing in maturation medium began, consisting of N2B27 medium with BDNF, GDNF and dbcAMP for two more weeks.

Immunoblotting

Denatured total cellular protein samples were separated on SDS polyacrylamide gels 4-15% Mini-PROTEAN® TGX Stain-Free™ gels (Bio-Rad) and transferred onto a nitrocellulose membrane using the Trans-Blot® Turbo™ Transfer System (Bio-Rad). Membranes were blocked with 3% (w/v) milk powder in PBS-T or TBS-T (1x PBS or TBS, 0.05% (v/v) Tween 20) for 1 h at room temperature (RT). Chameleon Duo Pre-stained Protein Ladder (Li-Cor Biosciences) was used as a molecular weight standard. Primary antibodies were anti-actin mAB (clone C4, 1:4000; Millipore MAB1501), anti-Cofilin-1 mAB (clone D3F9, 1:1000, Cell Signaling, 5175), anti-p-Cofilin-1 mAB (Ser3, clone 77G2, 1:1000, Cell Signaling 3313), anti-DRP1 mAB (clone 4E11B11, 1:1000, Cell Signaling 14647), anti-GDAP1 (1:750, Sigma HPA014266), anti-G6PD mAB (clone D5D2, 1:1000, Cell Signaling 12263), anti-GAPDH mAB (clone 14C10, 1:2000, Cell Signaling 2118), anti-GLUD1 mAB (clone D9F7P, 1:1000, Cell Signaling 12793), anti-HK1 mAB (clone C35C4, 1:1000; Cell Signaling 2024), anti-HK2 mAB (clone C64G5, 1:1000, Cell Signaling 2687), anti-HSPA9 (clone N52A/42, 1:1000, UC Davis/NIH NeuroMab Facility, Davis, USA, 73-127), anti-LDHA mAb (clone C4B5, 1:1000, Cell Signaling 3582), anti-MCU (1:1000, Sigma HPA016480), anti-MFF (1:1000, Proteintech 17090-1-AP), anti-MFN2 mAB (clone M03, 1:500; Abnova H00009927-M03), anti-Nestin mAB (clone rat-401, 1:1000, Merck chemicals MAB353), anti-PDH mAb (E1α, clone C54G1, 1:1000, Cell Signaling 3205), anti-PDH mAB (E2,E3bp, clone 13G2AE2BH5, 1:1000, Abcam, ab110333) anti-p-PDH (Ser293, 1:1000, Cell Signaling 31866), anti-PKM1/2 mAb (clone C103A3, 1:1000, Cell Signaling 3582), anti-TOM20 (1:1000, Sigma HPA011562), anti-βIII-Tubulin mAB (clone TuJ-1, 1:1000, R&D Systems MAB1195), anti-VDAC1 (clone 20B12AF2, 1:1000, Abcam ab14734). The membranes were incubated with the primary antibodies overnight at 4 °C. For visualization, membranes were incubated with an infrared fluorescence IRDye® 680RD Streptavidin for biotinylation staining or IRDye 800-conjugated anti-mouse, 800-conjugated anti-rabbit, or 680-conjugated anti-mouse IgG secondary antibody (1:15,000; Licor), for 1 h at RT and detected with the Odyssey Infrared Imaging System (Licor). Western Blots were analyzed with the Image Studio Lite Software (Li-Cor Biosciences).

Immunocytochemistry

Neuronal cells were grown on matrigel-coated µ-Slide 8 Well, ibiTreat (Ibidi, 80826), fixed with 4% paraformaldehyde (PFA) (CarlROTH, 3105.2) and permeabilized by 0.25% (v/v) Triton X-100 in PBS. Unspecific binding of antibodies was blocked with 1X Roti®-Block (CarlROTH, A151.4) for 30 min. Primary antibodies anti-Nestin (1:100, Ebioscience, 14-9843-82), anti-MAP2 (1:500, Synaptic Systems, 188004), anti-Smi32 (Clone SMI32, 1:1000, Biolegend 801702), anti-beta-III tubulin (1:1000, R&D Systems, MAB1195) were treated in 1X Roti®-Block at 4 °C overnight. The cells were washed three times with PBS and incubated with the fluorescent conjugated secondary antibody (Molecular Probes, A-11001) in 1X Roti®-Block for 1 h at RT. Subsequently, three PBS washing steps were done. Nuclei were stained with 300 nM 4′,6-diamidino-2-phenylindole (DAPI). Pictures were taken with a Leica TCS SP5 inverted confocal microscope with a 63x/NA1.4 oil immersion objective, a BX51 Fluorescence microscope (Olympus) with a 20x objective or an Opera PhenixTM spinning disc high-content screening microscope (Perkin Elmer, USA) equipped with two 16-bit sCMOS cameras and a 40x, 1.1 NA water immersion objective. For Cofilin-1 evaluations, SH-SY5Y cells were grown on 12 mm glass cover slides. Primary antibodies were anti-Cofilin-1 mAB (clone D3F9, 1:100, Cell Signaling, 5175), anti-GRP75 mAB (clone N52A/42, 1:20, Neuromab 75-127). Pictures were taken with a Leica TCS SP8 inverted confocal microscope with a 63x/1.4 NA oil immersion objective in z-stacks. Images were analyzed in Imaris 9.5.1 (Bitplane). For motoneurons analysis, images were imported into marcopolo columbus database and underwent batch analysis that determined the fluorescence intensity of each marker. The signal intensities of Smi32 and MAP2 were normalized to β-tubulinIII each.

Confocal microscopy and image analysis

High-content light microscopy analysis was conducted with the Opera Phenix^TM spinning disc microscope. Fluorescence (Ex/Em) for Hoechst, MitoTrackerTM Green FM and tetramethylrhodamine methyl ester perchlorate (TMRM) was measured at 405/435-480, 488/500-550 and 561/570-630 nm ex/em, respectively. BODIPY™ 558/568 C12 was added to the cells in parallel to MitoTracker in a final concentration of 1 µM for 15 min and subsequently, cells were incubated with normal growth medium. The images were analyzed with the software packages Harmony (version 4.5) and Columbus (version 2.7.1) containing the PhenoLOGICTM machine learning plugin (Perkin Elmer). Cells were selected via nuclear staining with HOECHST, but the cytosol region was defined using the mitochondrial stain. Within this selection, mitochondria and LDs were defined as isolated populations and analyzed respectively for their area and intensity. For mitochondria-associated LDs, the intensity of BODIPY^TM 558/568 C12 staining was determined within the mitochondrial area. The complete image analysis sequences are available upon request. Live-cell imaging of actin filament analysis was conducted two days after transfection of GFP-F-tractin (gift of Henry N. Higgs, Dartmouth University) and together with staining of mitochondria with 100 nM MitoTracker™ Red CMXRos. Cells were imaged using the Leica TCS SP8 inverted confocal microscope with a 63x/1.4 NA oil immersion objective in z-stacks. Cells were captured at 1024×1024 with sequential image capturing. Step size was set at 0.25 µm. Images from confocal microscopy for Cofilin-1, HSPA9 and actin filament quantification were analyzed in Imaris 9.5.1 (Bitplane). To improve the signal-to-noise ratio, iterative deconvolution was performed (Huygens Essential 20.10). The software then created 3D surfaces of mitochondria and Cofilin-1 spots to measure the surface area and to determine the shortest distance from the Cofilin-1 surface to the mitochondrial surface. Colocalizing Cofilin-1/actin with a distance ≤ 0 µm was duplicated to a new surface and taken for the Imaris surface-surface contact XTension to quantify the contact area, defined as the percentage of mitochondrial surface contacting Cofilin-1. Similarly, 3D surfaces of Actin and the mitochondria were created. The software computed the contact area between the two surfaces as well as the shortest distance between the actin and mitochondria surfaces. Live-cell imaging of DRP1 surface analysis was conducted two days after transfection of pcDNA3-GFP-DRP1 (gift of Prof. Dr. Culmsee, Philipps-Universität Marburg) and together with staining of mitochondria with 100 nM MitoTracker™ RedCMXRos. Images were captured in a similar manner using the Leica TCS SP8 inverted confocal microscope. Images were then imported to and analyzed in Imaris 9.7.2 (Bitplane) where 3D surfaces of Actin and the mitochondria were created. The software computed the contact area between the two surfaces as well as colocalization XTension to quantify the colocalized voxels between the Drp1 surface and the mitochondrial surface.

Cell culture and generation of stable cell lines

SH-SY5Y cell lines were grown in DMEM/F12 medium (Gibco) supplemented with 10% (v/v) fetal calf serum (FCS; Thermo Scientific), 2 mM L-glutamine (Gibco), 1 x MEM non-essential amino acids, 100 U/ml penicillin and 100 µg/ml streptomycin (Gibco) in a humidified incubator at 37 °C, 5% CO₂ and 95% air. The SH-SY5Y cell lines pLKO-NT (control) and G4 (knockdown) were a kind gift of David Pla-Martin and Francesc Palau¹⁹ and were grown in growth medium containing 2 µg/ml puromycin (InvivoGen).

Measurement of mitochondrial oxygen consumption

Mitochondrial respiration and oxygen consumption were analyzed using the Oxygraph-2k (Oroboros Instruments). Cells in suspension at a density of 1.5–2.0 × 10⁶ cells/ml were analyzed under continuous stirring at 750 rpm and 37 °C. All chemicals with the exception of palmitate-BSA were purchased from Sigma. Palmitate-BSA was purchased from Biomol. In a phosphorylation-control-protocol, the routine respiration of cells in their general growth medium was measured following the addition of 2 µg/ml oligomycin to inhibit the ATP synthase and measure leak respiration. By titration of the protonophore carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone in 0.5 µM steps the respiration was stimulated up to a maximum oxygen flow and the electron transfer system capacity was determined. By the addition of 0.5 µM rotenone and 2.5 µM antimycin A the respiration was inhibited and the non-mitochondrial residual oxygen consumption was measured. In 2 s intervals, the oxygen concentration and the oxygen flow per cells were recorded using the DatLab software 5.1 (Oroboros Instruments). All measurements were performed after daily calibration of the polarographic oxygen sensors and using instrumental background correction. The measured respiratory states were analyzed after correction with ROX to compare only mitochondrial oxygen consumption.

ATP measurements

Cytosolic and mitochondrial ATP levels were quantified as described²⁷. Cells were transiently transfected with plasmids carrying the bioluminescence energy transfer (BRET)-based ATP biosensor BTeam without targeting signal sequence (for cyto-ATP determination) or targeted to mitochondria (for mito-ATP determination) using TurboFectin reagent (OriGene). The plasmids were a kind gift of Hiroshi Imamura, University of Kyoto. 48 h later, the cells were incubated for 30 min in phenol red-free medium supplemented with 30 μM NanoLuciferase (NLuc) inhibitor to avoid disturbance from the BTeam released from dead cells. Afterwards, NLuc substrate (Promega) was added to the medium and the plate incubated for 20 min. Subsequently, luminescent emissions from the cells were measured at 37 °C at 520/560 nm (Yellow Fluorescent Protein (YFP) emission) and at 430/470 nm (NLuc emission) using a Tecan Spark® Multimode Microplate Reader.

Glutamine and glutamate measurements

The glutamine-glutamate-glo assay (Promega, # J8021) was performed according to manufacturer’s instructions to determine intracellular glutamine and glutamate concentrations. Briefly, 20,000 cells of the SH-SY5Y cell line and 100,000 motoneurons were plated in triplicates per experiment on white 96-well plates (Greiner, # 655083) two or seven days prior to the experiment, respectively. On the day of the experiment, cells were washed twice with PBS and 30 µl of PBS as well as 15 µl of 0.3 HCl solution was added to the cells and mixed for 5 min. 15 µl of 450 mM Tris solution, pH 8.0 was added and incubated for further 60 sec. From each lysate, 25 µl was transferred into a new white 96-well plate for I) glutamine plus glutamate and II) glutamate only measurement. Glutaminase was only added to the first set of wells and incubated for 30 min at RT. The detection reagent composed of Luciferin detection solution, reductase, reductase substrate, GLUD1 and NAD was added to all wells and incubated for 60 min at RT before luminescence detection in a Tecan Infinite 200 pro plate reader. Concentrations were calculated with glutamine and glutamate standard curves (0.78 to 50 µM) and a blank was included to remove any assay background signal. Glutamate levels were calculated by subtracting the glutamate-only signal from the signal containing glutamate and glutamine levels together.

Electron microscopy and analysis

Cells were pelleted by centrifugation at 1300 rpm for 3 min and fixed in 3% glutaraldehyde overnight. Following several rinses in 0.2 M sodium cacodylate buffer (pH 7.3), the samples were postfixed in 1% osmium tetroxide in cacodylate buffer for 2 h, dehydrated through an ascending series of ethanol concentrations and embedded in resin with propylene oxide as an intermediary. Semi-thin (0.65 µm) sections for light microscopy and ultrathin (50 nm) sections for electron microscopy were cut on a Leica Ultracut UCT ultramicrotome. Semi-thin sections were stained with methylene blue. Ultrathin sections were stained with an alcoholic solution of 1% uranyl acetate and lead citrate in sodium hydroxide and examined with a Zeiss EM-910 transmission electron microscope. For morphometric and quantitative analysis, representative cells were photographed at a magnification of 10,000 and 18,000. Analyses were done with ImageJ. MERCs were defined by a distance of mitochondrial and ER membranes between ~10 and ~50 nm. The frequency distribution of the MERCs width was calculated using GraphPad.

Biochemical analysis of mitochondria-ER contact sites

Cells were detached using trypsin/ETDA and resuspended in 1–2 ml of a buffer containing 0.32 M sucrose, 10 mM Tris-HCl, 1 mM EDTA and protease (Roche Diagnostics, 04693124001) and phosphatase (Roche Diagnostics, 04906845001) inhibitors. Cells were disrupted by a nitrogen decompression instrument (Parr Instrument Company, 4639) and centrifuged (2000 g, 10 min). The supernatant was transferred to a new microtube, centrifuged at 10,000 g for 10 min and the cell pellet resuspended in 1 ml of the same buffer containing the same components as above except 0.5 M sucrose. Finally, mitochondria were sedimented by centrifugation at 10,000 g for 10 min and resolved in an appropriate amount of RIPA buffer for the subsequent experiments.

FRET-based FEMP probe to quantify mitochondria-ER contact sites

MERCS were quantified with a FRET-based sensor indicating the proximity between the ER and mitochondria. The plasmid encodes for a YFP-linked outer mitochondrial membrane protein Akap1 and a CFP-conjugated ER-protein Sac1, as well as a fused FKBP and FRB domain respectively and is available from Marta Giacomello, University of Padova. These domains can form heterodimers upon rapamycin treatment. The specific localization of these proteins is ensured by the introduction of a self-cleavable Tav2A sequence^34,35. The FEMP plasmid (FRET-based ER-mitochondria probe) was transfected with GenJet. 48 h later, images were taken with the Perkin Elmer Operetta High-Content Imaging System acquiring the CFP- (410-430/460-500 ex/em), YFP- (490-510/520-560 ex/em) and YFP_FRET-emission (410-430/520-560 ex/em) with a ×40 water objective for determination of the basal distances between the organelles. Subsequently, the cells were treated with 100 nM rapamycine for 15 min for FKBP-FRB dimerization induction and to reach a maximum of YFP_FRET signal. Cells were fixed for another 20 min with 1% PFA and imaging was performed again with equal microscopy settings. For the analysis, the Harmony software was used. First, the cells were identified using the YFP channel. Within each cell and region of interest (ROI), the intensities of the three acquired channels were calculated, including background subtraction. FRET basal and FRET max were calculated as: (FYFP-FRETcell-FYFP-FRET_background)/ (FCFPcell-FCFP_background); FRET Ratio was calculated as (FRET_max–FRET_basal)/FRET_basal.

Mitochondrial Ca²⁺ measurement

Dye: Cells were plated (20,000 cells/cm²) in 8-well µ-slides ibiTreat (ibidi, 80826) and treated with 5 µM of Rhod2-AM (Molecular Probes, R1245MP) in culture medium without FBS for 60 min at 37 °C the next day. Rhod2-AM fluorescent signals were analyzed at (Ex/Em) 549/578 nm wavelengths using a Leica SP5 confocal microscope and analyzed by ImageJ. Genetically encoded reporter: Cells were transfected with the genetically encoded reporters and analyzed two days later in an inverted TCS-SP5 confocal microscope (Leica) with appropriate excitation/emission wavelengths as reported for CEPIA3mt (Addgene, 58219) and GEM-CEPIA1er (Addgene, 58217)³¹. The CEPIA3mt construct was co-transfected and normalized to mito-TurboFarRed.

Aequorin Ca²⁺ measurements

For Ca²⁺ measurements, the biosensor aequorin (AEQ) was used, which is a 22 kDa calcium-binding photoprotein isolated from jellyfish Aequorea Victoria. The plasmid is available from Marta Giacomello, University of Padova. SH-SY5Y cells were grown on 12 mm glass coverslips to a confluence of 40–50% and transfected with cytosolic or mitochondria-targeted AEQ (cytAEQ/mtAEQ) using GenJet. On the day of the experiment, the cells were treated with 5 µM coelenterazine-N-AM (Santa Cruz Biotechnology sc-205904) in basic saline solution (BS, 135 mM NaCl₂, 5 mM KCl, 0.4 mM KH₂PO₄, 1 mM MgSO₄ x 7 H₂O, 20 mM HEPES, 0.1% (w/v) glucose, pH 7.4 adjustment at 37 °C with NaOH 10 N) containing 1 mM CaCl₂ for 2 h at 37 °C and 5% CO₂. Coelenterazine served as a substrate for AEQ. Cover slides were then placed in the luminometer with constant buffer perfusion with BS supplemented with I) 1 mM Ca²⁺ (30 s) II) 200 µM EGTA (30 s) III) 200 µM Carbachol (CCH) + 200 µM EGTA (120 s) IV) 100 µM Digitonin + 5 mM CaCl₂ (220 s).

Precipitation of biotinylated Avi-GDAP1

Primary cortical neuron cultures were prepared from embryos (E16) from the transgenic mouse line Gt(ROSA)26Sortm1(birA)Mejr (ROSA26-BirA) of a C57BL/6 N background. In this mouse strain, the biotin ligase BirA was inserted into the gene locus of the ROSA26 promoter⁴⁰. Cortical neurons were cultured on poly-d-lysine coated plates (0.05 mg/ml) in Neurobasal medium (NBM, Life Technologies, 21103049) supplemented with 2% (v/v) B-27 supplement (Life Technologies, 17504-044), 1% (v/v) L-glutamine (Sigma-Aldrich, G1251) and 100 U/ml penicillin and 100 µg/ml streptomycin (Sigma-Aldrich, P0781). Medium change was performed at day 1 and day 4 after isolation. For transduction of the primary neurons, viruses were added in a volume that equaled 6 × 10⁷−8 × 10⁷ copies/µl per well of a 6 well plate containing 4 ml of NBM and neurons on day 4 after isolation. Neurons were cultured for further 7 days. Cells were harvested, washed with PBS and lysed in RIPA buffer supplemented with protease inhibitors. The samples were centrifuged at 21,000 g for 30 min at 4 °C, the proteins concentration was determined via BC-Assay and 20 µg protein lysate was removed as input-control. Protein lysates were either directly incubated with Dynabeads™ MyOne™ Streptavidin T1 according to the manufacturer’s protocol or taken for GSH or GSSG incubation (5 mM GSH or GSSG in 10 mM HEPES pH 7.4, 35 mM sucrose, 40 mM KCl, 0.25 mM EGTA, 2 mM Mg(CH₃COO)₂, 0.5 mM GTP, 1 mM ATP (K⁺), 5 mM Na-succinate, 0.08 mM ADP, 2 mM K₂HPO₄ for 30 min at 37 °C). Dynabeads™ MyOne™ Streptavidin T1 were incubated for 45 min at RT on a shaker. Dynabeads were washed three times with RIPA buffer and once with PBS and stored at −80 °C until further processing for LC-MS.

Protein elution, lysis and digestion

Bound proteins were eluted from Dynabeads in 10 mM Tris pH 8.0, 2% SDS, 1 mM Biotin at 80 °C. Whole cells were lysed in 5 µl of 10% SDS at 95 °C for 5 min. Protein samples were digested using the SP3 (“Single-Pot Solid-Phase-Enhanced Sample Preparation”) protocol⁷⁴ with modifications⁷⁵. Proteins were reduced by adding 5 µl of 200 mM Dithiothreitol (DTT) per 100 µl lysate (45 °C, 30 min). Free cysteines were subsequently alkylated by adding 10 µl 100 mM Iodoacetamide (IAA) per 100 µl lysate (RT, 30 min, in the dark). Subsequently, remaining IAA was quenched by adding 10 µl 200 mM DTT per 100 µl lysate. Magnetic carboxylate modified particles Beads (SpeedBeads, Sigma) were used for protein clean-up and acetonitrile (ACN), in a final concentration of 70%, was added to the samples to induce the binding of the proteins to the beads by hydrophilic interactions (18 min RT). By incubating the bead-protein mixture on a magnetic stand for 2 min, the sample was bound to the magnet and the supernatant removed, followed by two washing steps with 70% ethanol (EtOH), addition of 180 µl ACN, incubation for 15 s and removal of the solvent. Finally, 5 µl digest buffer (50 mM ammonium bicarbonate, 1:25 w/w trypsin:protein ratio) was added to the air-dried bead-protein mixtures and incubated overnight at 37 °C. To purify peptides after digestion, ACN was added to a final concentration of 95%. After another washing step. the beads were resuspended in 10 µl 2% DMSO (in water), put into an ultrasonic bath for 1 min and then shortly centrifuged. 10 µl of the resulting supernatant was mixed with 5 µl 100 fmol/µl Enolase digest (Waters Corporation) and acidified with 5 µl 1% formic acid (FA).

LC-MS analysis

Liquid chromatography (LC) of tryptic peptides was performed on a NanoAQUITY UPLC system (Waters Corporation) equipped with 75 µM × 250 mm HSS-T3 C18 column (Waters corporation). Mobile phase A was 0.1% (v/v) formic acid (FA) and 3% (v/v) DMSO in water. Mobile phase B was 0.1% (v/v) FA and 3% (v/v) DMSO in ACN. Peptides were separated running a gradient from 5 to 40% (v/v) mobile phase B at a flow rate of 300 nL/ min over 90 min. The column was heated to 55 °C. MS analysis of eluting peptides was performed by ion-mobility enhanced data-independent acquisition (UDMS^E). Precursor ion information was collected in low-energy MS mode at a constant collision energy of 4 eV. Fragment ion information was obtained in the elevated energy scan applying drift-time specific collision energies. The spectral acquisition time in each mode was 0.7 s with a 0.05 s interscan delay resulting in an overall cycle time of 1.5 s for the acquisition of one cycle of low and elevated energy data. [Glu1]-fibrinopeptide was used as lock mass at 100 fmol/μL and sampled every 30 s into the mass spectrometer via the reference sprayer of the NanoLockSpray source. All samples were analyzed in three technical replicates.

Data processing and label-free protein quantification

UDMS^E data processing and database search was performed using ProteinLynx Global Server (PLGS, ver. 3.0.2, Waters Corporation). The resulting proteins were searched against the UniProt proteome database (species: Mus musculus, UniProtK-Swissprot release 2019_05, 17.051 entries; Homo sapiens, UniProtK-Swissprot release 2019_10, 20367 entries) supplemented with a list of common contaminants. The database search was specified by trypsin as enzyme for digestion and peptides with up to two missed cleavages were included. Carbamidomethyl cysteine was set as fixed modification and oxidized methionine as variable modification. False discovery rate assessment for peptide and protein identification was done using the target-decoy strategy by searching a reverse database and was set to 0.01 for database search in PLGS. Retention time alignment, exact mass retention time (EMRT), as well as normalization and filtering were performed in ISOQuant ver.1.8. By using TOP3 quantification, absolute in-sample amounts of proteins were calculated.

Statistics and reproducibility

Normal distribution was tested using the D’Agostino-Pearson omnibus normality test. Statistical significance was then verified using appropriate parametric (Student’s t test or ANOVA) or non-parametric tests (Mann–Whitney and Kruskal–Wallis tests) followed by multiple comparison tests as indicated. The Wilcoxon signed rank test was used when normalization to 100% was necessary as indicated. Statistical analysis of mass spectrometry data was performed using two-tailed, paired t-tests and subsequent Bonferroni correction. Here, a corrected p < 0.01 was considered significant for the biotinylated Avi-GDAP1 pulldown experiment and p < 0.05 for whole cell lysates. In all other data a p < 0.05 was considered to be statistically significant. Pathway over-representation analysis was performed using the STRING database with default parameters³⁹. KEGG pathway visualization was performed using the R package clusterProfiler⁷⁶.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository⁷⁷ with the data set identifiers: <PXD024555> for the biotinylated Avi-GDAP1 coprecipitation experiment; and <PXD028460> for the GDAP1 KD label-free quantification experiment. Plasmids used are mentioned in the Methods section and can be requested from those that generated them upon reasonable request. Supplementary Data 1 contains all raw data and uncropped immunoblots.

References

Baxter, R. V. et al. Ganglioside-induced differentiation-associated protein-1 is mutant in Charcot-Marie-Tooth disease type 4A/8q21. Nat. Genet. 30, 21–22 (2002).
Article CAS PubMed Google Scholar
Cuesta, A. et al. The gene encoding ganglioside-induced differentiation-associated protein 1 is mutated in axonal Charcot-Marie-Tooth type 4A disease. Nat. Genet. 30, 22–25 (2002).
Article CAS PubMed Google Scholar
Senderek, J. et al. Mutations in the ganglioside-induced differentiation-associated protein-1 (GDAP1) gene in intermediate type autosomal recessive Charcot-Marie-Tooth neuropathy. Brain 126, 642–649 (2003).
Article PubMed Google Scholar
Chung, K. W. et al. A novel GDAP1 Q218E mutation in autosomal dominant Charcot-Marie-Tooth disease. J. Hum. Genet. 53, 360–364 (2008).
Article CAS PubMed Google Scholar
Marco, A., Cuesta, A., Pedrola, L., Palau, F. & Marín, I. Evolutionary and structural analyses of GDAP1, involved in Charcot-Marie-Tooth disease, characterize a novel class of glutathione transferase-related genes. Mol. Biol. Evol. 21, 176–187 (2004).
Article CAS PubMed Google Scholar
Shield, A. J., Murray, T. P. & Board, P. G. Functional characterisation of ganglioside-induced differentiation-associated protein 1 as a glutathione transferase. Biochem. Biophys. Res. Commun. 347, 859–866 (2006).
Article CAS PubMed Google Scholar
Wagner, K. M., Rüegg, M., Niemann, A. & Suter, U. Targeting and function of the mitochondrial fission factor GDAP1 are dependent on its tail-anchor. PLoS ONE 4, e5160 (2009).
Article PubMed PubMed Central CAS Google Scholar
Pedrola, L. et al. GDAP1, the protein causing Charcot-Marie-Tooth disease type 4A, is expressed in neurons and is associated with mitochondria. Hum. Mol. Genet. 14, 1087–1094 (2005).
Article CAS PubMed Google Scholar
Huber, N. et al. Glutathione-conjugating and membrane-remodeling activity of GDAP1 relies on amphipathic C-terminal domain. Sci. Rep. 6, 36930 (2016).
Article CAS PubMed PubMed Central Google Scholar
Googins, M. R. et al. Structural and functional divergence of GDAP1 from the glutathione S-transferase superfamily. FASEB J. 34, 7192–7207 (2020).
Article CAS PubMed Google Scholar
Niemann, A., Ruegg, M., La Padula, V., Schenone, A. & Suter, U. Ganglioside-induced differentiation associated protein 1 is a regulator of the mitochondrial network: new implications for Charcot-Marie-Tooth disease. J. Cell Biol. 170, 1067–1078 (2005).
Article CAS PubMed PubMed Central Google Scholar
Huber, N., Guimaraes, S., Schrader, M., Suter, U. & Niemann, A. Charcot-Marie-Tooth disease-associated mutants of GDAP1 dissociate its roles in peroxisomal and mitochondrial fission. EMBO Rep. 14, 545–552 (2013).
Article CAS PubMed PubMed Central Google Scholar
Niemann, A., Wagner, K. M., Ruegg, M. & Suter, U. GDAP1 mutations differ in their effects on mitochondrial dynamics and apoptosis depending on the mode of inheritance. Neurobiol. Dis. 36, 509–520 (2009).
Article CAS PubMed Google Scholar
Ingerman, E. et al. Dnm1 forms spirals that are structurally tailored to fit mitochondria. J. Cell Biol. 170, 1021–1027 (2005).
Article CAS PubMed PubMed Central Google Scholar
Friedman, J. R. et al. ER tubules mark sites of mitochondrial division. Science 334, 358–362 (2011).
Article CAS PubMed PubMed Central Google Scholar
De Vos, K. J., Allan, V. J., Grierson, A. J. & Sheetz, M. P. Mitochondrial function and actin regulate dynamin-related protein 1-dependent mitochondrial fission. Curr. Biol. 15, 678–683 (2005).
Article PubMed CAS Google Scholar
Korobova, F., Ramabhadran, V. & Higgs, H. N. An actin-dependent step in mitochondrial fission mediated by the ER-associated formin INF2. Science 339, 464–467 (2013).
Article CAS PubMed Google Scholar
Chakrabarti, R. et al. INF2-mediated actin polymerization at the ER stimulates mitochondrial calcium uptake, inner membrane constriction, and division. J. Cell Biol. 217, 251–268 (2018).
Article CAS PubMed PubMed Central Google Scholar
Pla-Martín, D. et al. Silencing of the Charcot-Marie-Tooth disease-associated gene GDAP1 induces abnormal mitochondrial distribution and affects Ca2+ homeostasis by reducing store-operated Ca2+ entry. Neurobiol. Dis. 55, 140–151 (2013).
Article PubMed CAS Google Scholar
Boyer, O. et al. INF2 mutations in Charcot-Marie-Tooth disease with glomerulopathy. N. Engl. J. Med. 365, 2377–2388 (2011).
Article CAS PubMed Google Scholar
Rehklau, K. et al. Cofilin1-dependent actin dynamics control DRP1-mediated mitochondrial fission. Cell Death Dis. 8, e3063 (2017).
Article PubMed PubMed Central Google Scholar
González-Sánchez, P. et al. CMT-linked loss-of-function mutations in GDAP1 impair store-operated Ca2+ entry-stimulated respiration. Sci. Rep. 7, 42993 (2017).
Article PubMed PubMed Central CAS Google Scholar
Chance, B. The energy-linked reaction of calcium with mitochondria. J. Biol. Chem. 240, 2729–2748 (1965).
Article CAS PubMed Google Scholar
Kennerson, M. L. et al. A new locus for X-linked dominant Charcot-Marie-Tooth disease (CMTX6) is caused by mutations in the pyruvate dehydrogenase kinase isoenzyme 3 (PDK3) gene. Hum. Mol. Genet. 22, 1404–1416 (2013).
Article CAS PubMed PubMed Central Google Scholar
Pla-Martín, D. et al. Junctophilin-1 is a modifier gene of GDAP1-related Charcot-Marie-Tooth disease. Hum. Mol. Genet. 24, 213–229 (2015).
Article PubMed CAS Google Scholar
Noack, R. et al. Charcot-Marie-Tooth disease CMT4A: GDAP1 increases cellular glutathione and the mitochondrial membrane potential. Hum. Mol. Genet. 21, 150–162 (2012).
Article PubMed CAS Google Scholar
Yoshida, T., Kakizuka, A. & Imamura, H. BTeam, a novel BRET-based biosensor for the accurate quantification of ATP concentration within living cells. Sci. Rep. 6, 39618 (2016).
Article CAS PubMed PubMed Central Google Scholar
Denton, R. M., Randle, P. J. & Martin, B. R. Stimulation by calcium ions of pyruvate dehydrogenase phosphate phosphatase. Biochem. J. 128, 161–163 (1972).
Article CAS PubMed PubMed Central Google Scholar
Denton, R. M., Richards, D. A. & Chin, J. G. Calcium ions and the regulation of NAD+-linked isocitrate dehydrogenase from the mitochondria of rat heart and other tissues. Biochem. J. 176, 899–906 (1978).
Article CAS PubMed PubMed Central Google Scholar
Zhao, Y. et al. Site-directed mutagenesis of phosphorylation sites of the branched chain alpha-ketoacid dehydrogenase complex. J. Biol. Chem. 269, 18583–18587 (1994).
Article CAS PubMed Google Scholar
Suzuki, J. et al. Imaging intraorganellar Ca2+ at subcellular resolution using CEPIA. Nat. Commun. 5, 4153 (2014).
Article CAS PubMed Google Scholar
Giacomello, M. & Pellegrini, L. The coming of age of the mitochondria-ER contact: a matter of thickness. Cell Death Differ. 23, 1417–1427 (2016).
Article CAS PubMed PubMed Central Google Scholar
Marchi, S., Patergnani, S. & Pinton, P. The endoplasmic reticulum-mitochondria connection: one touch, multiple functions. Biochim. Biophys. Acta 1837, 461–469 (2014).
Article CAS PubMed Google Scholar
Csordás, G. et al. Imaging interorganelle contacts and local calcium dynamics at the ER-mitochondrial interface. Mol. Cell 39, 121–132 (2010).
Article PubMed PubMed Central CAS Google Scholar
Naon, D. et al. Critical reappraisal confirms that Mitofusin 2 is an endoplasmic reticulum-mitochondria tether. Proc. Natl Acad. Sci. USA 113, 11249–11254 (2016).
Article CAS PubMed PubMed Central Google Scholar
Giacomello, M. et al. Ca2+ hot spots on the mitochondrial surface are generated by Ca2+ mobilization from stores, but not by activation of store-operated Ca2+ channels. Mol. Cell 38, 280–290 (2010).
Article CAS PubMed Google Scholar
Szabadkai, G. et al. Chaperone-mediated coupling of endoplasmic reticulum and mitochondrial Ca2+ channels. J. Cell Biol. 175, 901–911 (2006).
Article CAS PubMed PubMed Central Google Scholar
Tenzer, S. et al. Proteome-wide characterization of the RNA-binding protein RALY-interactome using the in vivo-biotinylation-pulldown-quant (iBioPQ) approach. J. Proteome Res. 12, 2869–2884 (2013).
Article CAS PubMed Google Scholar
Szklarczyk, D. et al. The STRING database in 2021: customizable protein–protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 49, D605–D612 (2020).
Article PubMed Central CAS Google Scholar
Driegen, S. et al. A generic tool for biotinylation of tagged proteins in transgenic mice. Transgenic Res. 14, 477–482 (2005).
Article CAS PubMed Google Scholar
Cull, M. G. & Schatz, P. J. Biotinylation of proteins in vivo and in vitro using small peptide tags. Methods Enzymol. 326, 430–440 (2000).
Article CAS PubMed Google Scholar
Shutt, T., Geoffrion, M., Milne, R. & McBride, H. M. The intracellular redox state is a core determinant of mitochondrial fusion. EMBO Rep. 13, 909–915 (2012).
Article CAS PubMed PubMed Central Google Scholar
Bittermann, E. et al. Differential requirements of tubulin genes in mammalian forebrain development. PLoS Genet. 15, e1008243 (2019).
Article CAS PubMed PubMed Central Google Scholar
Estela, A., Pla-Martín, D., Sánchez-Piris, M., Sesaki, H. & Palau, F. Charcot-Marie-Tooth-related gene GDAP1 complements cell cycle delay at G2/M phase in Saccharomyces cerevisiae fis1 gene-defective cells. J. Biol. Chem. 286, 36777–36786 (2011).
Article CAS PubMed PubMed Central Google Scholar
Klemke, M. et al. Oxidation of cofilin mediates T cell hyporesponsiveness under oxidative stress conditions. Immunity 29, 404–413 (2008).
Article CAS PubMed Google Scholar
Gellert, M., Hanschmann, E.-M., Lepka, K., Berndt, C. & Lillig, C. H. Redox regulation of cytoskeletal dynamics during differentiation and de-differentiation. Biochim. Biophys. Acta 1850, 1575–1587 (2015).
Article CAS PubMed Google Scholar
Kruyer, A., Ball, L. E., Townsend, D. M., Kalivas, P. W. & Uys, J. D. Post-translational S-glutathionylation of cofilin increases actin cycling during cocaine seeking. PLoS ONE 14, e0223037 (2019).
Article CAS PubMed PubMed Central Google Scholar
Melak, M., Plessner, M. & Grosse, R. Actin visualization at a glance. J. Cell Sci. 130, 525–530 (2017).
Article PubMed Google Scholar
Bunnell, T. M., Burbach, B. J., Shimizu, Y. & Ervasti, J. M. β-Actin specifically controls cell growth, migration, and the G-actin pool. Mol. Biol. Cell 22, 4047–4058 (2011).
Article CAS PubMed PubMed Central Google Scholar
McGough, A., Pope, B., Chiu, W. & Weeds, A. Cofilin changes the twist of F-actin: implications for actin filament dynamics and cellular function. J. Cell Biol. 138, 771–781 (1997).
Article CAS PubMed PubMed Central Google Scholar
Ji, W.-K., Hatch, A. L., Merrill, R. A., Strack, S. & Higgs, H. N. Actin filaments target the oligomeric maturation of the dynamin GTPase Drp1 to mitochondrial fission sites. Elife 4, e11553 (2015).
Article PubMed PubMed Central Google Scholar
Patron, M. et al. MICU1 and MICU2 finely tune the mitochondrial Ca2+ uniporter by exerting opposite effects on MCU activity. Mol. Cell 53, 726–737 (2014).
Article CAS PubMed PubMed Central Google Scholar
Civera-Tregón, A. et al. Mitochondria and calcium defects correlate with axonal dysfunction in GDAP1-related Charcot-Marie-Tooth mouse model. Neurobiol. Dis. 152, 105300 (2021).
Boldogh, I. R. & Pon, L. A. Interactions of mitochondria with the actin cytoskeleton. Biochim. Biophys. Acta 1763, 450–462 (2006).
Article CAS PubMed Google Scholar
Hoffmann, L. et al. Cofilin1 oxidation links oxidative distress to mitochondrial demise and neuronal cell death. Cell Death Dis. 12, 953 (2021).
Article CAS PubMed PubMed Central Google Scholar
Pope, B. J., Zierler-Gould, K. M., Kühne, R., Weeds, A. G. & Ball, L. J. Solution structure of human cofilin: actin binding, pH sensitivity, and relationship to actin-depolymerizing factor. J. Biol. Chem. 279, 4840–4848 (2004).
Article CAS PubMed Google Scholar
Pfannstiel, J. et al. Human cofilin forms oligomers exhibiting actin bundling activity. J. Biol. Chem. 276, 49476–49484 (2001).
Article CAS PubMed Google Scholar
Goyal, P. et al. Cofilin oligomer formation occurs in vivo and is regulated by cofilin phosphorylation. PLoS ONE 8, e71769 (2013).
Article CAS PubMed PubMed Central Google Scholar
Siudeja, K., Grzeschik, N. A., Rana, A., de Jong, J. & Sibon, O. C. M. Cofilin/Twinstar phosphorylation levels increase in response to impaired coenzyme a metabolism. PLoS ONE 7, e43145 (2012).
Article CAS PubMed PubMed Central Google Scholar
Chua, B. T. et al. Mitochondrial translocation of cofilin is an early step in apoptosis induction. Nat. Cell Biol. 5, 1083–1089 (2003).
Article CAS PubMed Google Scholar
Klamt, F. et al. Oxidant-induced apoptosis is mediated by oxidation of the actin-regulatory protein cofilin. Nat. Cell Biol. 11, 1241–1246 (2009).
Article CAS PubMed PubMed Central Google Scholar
Luo, S., Uehara, H. & Shacter, E. Taurine chloramine-induced inactivation of cofilin protein through methionine oxidation. Free Radic. Biol. Med. 75, 84–94 (2014).
Article CAS PubMed Google Scholar
Fratelli, M. et al. Identification by redox proteomics of glutathionylated proteins in oxidatively stressed human T lymphocytes. Proc. Natl Acad. Sci. USA 99, 3505–3510 (2002).
Article CAS PubMed PubMed Central Google Scholar
Orth, J. D., Krueger, E. W., Cao, H. & McNiven, M. A. The large GTPase dynamin regulates actin comet formation and movement in living cells. Proc. Natl Acad. Sci. USA 99, 167–172 (2002).
Article CAS PubMed PubMed Central Google Scholar
Lee, E. & De Camilli, P. Dynamin at actin tails. Proc. Natl Acad. Sci. USA 99, 161–166 (2002).
Article CAS PubMed PubMed Central Google Scholar
Züchner, S. et al. Mutations in the pleckstrin homology domain of dynamin 2 cause dominant intermediate Charcot-Marie-Tooth disease. Nat. Genet. 37, 289–294 (2005).
Article PubMed CAS Google Scholar
Fabrizi, G. M. et al. Two novel mutations in dynamin-2 cause axonal Charcot-Marie-Tooth disease. Neurology 69, 291–295 (2007).
Article CAS PubMed Google Scholar
Lee, J. E., Westrate, L. M., Wu, H., Page, C. & Voeltz, G. K. Multiple dynamin family members collaborate to drive mitochondrial division. Nature 540, 139–143 (2016).
Article CAS PubMed PubMed Central Google Scholar
Luigetti, M. et al. Seipin S90L mutation in an Italian family with CMT2/dHMN and pyramidal signs. Muscle Nerve 42, 448–451 (2010).
Article CAS PubMed Google Scholar
Boutet, E. et al. Seipin deficiency alters fatty acid Delta9 desaturation and lipid droplet formation in Berardinelli-Seip congenital lipodystrophy. Biochimie 91, 796–803 (2009).
Article CAS PubMed Google Scholar
Tian, Y. et al. Tissue-autonomous function of Drosophila seipin in preventing ectopic lipid droplet formation. PLoS Genet. 7, e1001364 (2011).
Article CAS PubMed PubMed Central Google Scholar
Yang, W. et al. BSCL2/seipin regulates adipogenesis through actin cytoskeleton remodelling. Hum. Mol. Genet. 23, 502–513 (2014).
Article CAS PubMed Google Scholar
Reinhardt, P. et al. Derivation and expansion using only small molecules of human neural progenitors for neurodegenerative disease modeling. PLoS ONE 8, e59252 (2013).
Article CAS PubMed PubMed Central Google Scholar
Hughes, C. S. et al. Ultrasensitive proteome analysis using paramagnetic bead technology. Mol. Syst. Biol. 10, 757 (2014).
Article PubMed CAS Google Scholar
Sielaff, M. et al. Evaluation of FASP, SP3, and iST Protocols for Proteomic Sample Preparation in the Low Microgram Range. J. Proteome Res. 16, 4060–4072 (2017).
Article CAS PubMed Google Scholar
Wu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innov. J. 2, 100141 (2021).
Google Scholar
Vizcaíno, J. A. et al. The PRoteomics IDEntifications (PRIDE) database and associated tools: status in 2013. Nucleic Acids Res. 41, D1063–D1069 (2013).
Article PubMed CAS Google Scholar

Download references

Acknowledgements

The excellent support by the IMB Core Facility Microscopy is gratefully acknowledged. This work was supported by grants of the Deutsche Forschungsgemeinschaft to A.M. (CRC1080 project A10) and to A.M. and E.M.H. and S.T. (DFG ME1922 16-1; Ha 8334/2-2; TE599/6-1) within the priority program 1710. This work was supported by an Excellence Grant of the Luxembourg National Research Fund (FNR) within the PEARL program (FNR; FNR/P13/6682797 to R.K.). Work of R.K. is further supported by a grant from the FNR within the National Centre of Excellence in Research on Parkinson’s disease (NCER-PD). We thank the Disease Modelling and Screening Platform of the Luxembourg Institute of Systems Biomedicine and the Luxembourg Institute of Health for their support on high-content imaging of motoneurons. We thank Marion Silies for excellent and exhaustive proofreading.

Funding

Open Access funding enabled and organized by Projekt DEAL.

Author information

Authors and Affiliations

Institute of Molecular Medicine, University Medical Center Mainz, Mainz, Germany
Christina Wolf, Alireza Pouya, Sara Bitar, Annika Pfeiffer, Diones Bueno, Liliana Rojas-Charry, Federica Dal Bello, Caterina Vianello & Axel Methner
Institute for Immunology, University Medical Center Mainz, Mainz, Germany
Sabine Arndt, David Gomez-Zepeda & Stefan Tenzer
Department of Biology, University of Padova, Padova, Italy
Federica Dal Bello, Caterina Vianello & Marta Giacomello
Department of Biomedical Sciences, University of Padova, Padova, Italy
Federica Dal Bello, Caterina Vianello & Marta Giacomello
Institute of Molecular Biology (IMB) gGmbH, Mainz, Germany
Sandra Ritz & Jonas Schwirz
Institute of Reconstructive Neurobiology, University of Bonn Medical Faculty and University Hospital Bonn, Bonn, Germany
Kristina Dobrindt, Michael Peitz & Oliver Brüstle
Cell Programming Core Facility, University of Bonn School of Medicine, Bonn, Germany
Michael Peitz
Department of Neurology, Medical Faculty, Heinrich-Heine University, Düsseldorf, Germany
Eva-Maria Hanschmann
Translational Neuroscience, Luxembourg Centre for Systems Biomedicine (LCSB), University of Luxembourg, Belvaux, Luxembourg
Pauline Mencke, Ibrahim Boussaad & Rejko Krüger
Disease Modeling and Screening Platform (DMSP), Luxembourg Centre of Systems Biomedicine (Biomedicine), University of Luxembourg and Luxembourg Institute of Health (LIH), Belvaux, Luxembourg
Ibrahim Boussaad
Institute of Development Biology and Neurobiology, Johannes Gutenberg-University Mainz, Mainz, Germany
Marion Silies
Transversal Translational Medicine, Luxembourg Institute of Health (LIH), Strassen, Luxembourg
Rejko Krüger

Authors

Christina Wolf
View author publications
You can also search for this author in PubMed Google Scholar
Alireza Pouya
View author publications
You can also search for this author in PubMed Google Scholar
Sara Bitar
View author publications
You can also search for this author in PubMed Google Scholar
Annika Pfeiffer
View author publications
You can also search for this author in PubMed Google Scholar
Diones Bueno
View author publications
You can also search for this author in PubMed Google Scholar
Liliana Rojas-Charry
View author publications
You can also search for this author in PubMed Google Scholar
Sabine Arndt
View author publications
You can also search for this author in PubMed Google Scholar
David Gomez-Zepeda
View author publications
You can also search for this author in PubMed Google Scholar
Stefan Tenzer
View author publications
You can also search for this author in PubMed Google Scholar
Federica Dal Bello
View author publications
You can also search for this author in PubMed Google Scholar
Caterina Vianello
View author publications
You can also search for this author in PubMed Google Scholar
Sandra Ritz
View author publications
You can also search for this author in PubMed Google Scholar
Jonas Schwirz
View author publications
You can also search for this author in PubMed Google Scholar
Kristina Dobrindt
View author publications
You can also search for this author in PubMed Google Scholar
Michael Peitz
View author publications
You can also search for this author in PubMed Google Scholar
Eva-Maria Hanschmann
View author publications
You can also search for this author in PubMed Google Scholar
Pauline Mencke
View author publications
You can also search for this author in PubMed Google Scholar
Ibrahim Boussaad
View author publications
You can also search for this author in PubMed Google Scholar
Marion Silies
View author publications
You can also search for this author in PubMed Google Scholar
Oliver Brüstle
View author publications
You can also search for this author in PubMed Google Scholar
Marta Giacomello
View author publications
You can also search for this author in PubMed Google Scholar
Rejko Krüger
View author publications
You can also search for this author in PubMed Google Scholar
Axel Methner
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

C.W., A.P., S.B., AnnP, D.B., P.M., F.d.B., C.V., I.B., D.G., P.M. carried out experiments; K.D., M.P., and O.B. generated the iPS cells; J.S. and S.R. helped establishing high-content microscopy; S.A., D.G., and S.T. analyzed quantitative proteomic analyses; M.G. supervised and analyzed the FEMP and aequorin measurements; R.K. supervised and analyzed the experiments with human motoneurons; C.W., M.S., and E.M.H. analyzed data and helped writing the manuscript; A.M. devised experiments, analyzed data and wrote the manuscript.

Corresponding author

Correspondence to Axel Methner.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

This manuscript has been previously reviewed at another Nature Portfolio journal. The manuscript was considered suitable for publication without further review at Communications Biology. Primary Handling Editors: Eve Rogers.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1

Reporting Summary

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Wolf, C., Pouya, A., Bitar, S. et al. GDAP1 loss of function inhibits the mitochondrial pyruvate dehydrogenase complex by altering the actin cytoskeleton. Commun Biol 5, 541 (2022). https://doi.org/10.1038/s42003-022-03487-6

Download citation

Received: 24 March 2022
Accepted: 11 May 2022
Published: 03 June 2022
DOI: https://doi.org/10.1038/s42003-022-03487-6

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.