Article Figures & Data
Figures
- Figure 1. Identification of five individuals harbouring de novo pathogenic variants in DNM1L.
(A) Family pedigrees of DNM1L patients. Affected individuals are shown in black, squares represent males, circles represent females, triangles represent pregnancy not carried to term, and a diagonal line through the symbols indicates deceased subjects. (B) Schematic representation of known DRP1 variants and DRP1 protein domain organization: BSE (bundle signalling element), GTPase domain, stalk domain, variable domain (VD), and the GTPase effector domain (GED). Variants identified in this study are shown in black and previously reported variants are in grey. Partial amino acid sequence alignments of DRP1 showing evolutionary conservation across different species.
- Figure 2. In silico structural studies of DRP1 variants.
(A) Locations of pathogenic variants marked on the crystal structure of nucleotide-free DRP1 (PDB: 4BEJ). (B, C, D) Residue–residue interactions and spatial relationships of residues to neighbouring motifs or DRP1 monomers of the wild-type version of residues from (A) (CryoEM structure of DRP1 assembled and in complex with MID49, PDB: 5WP9). (B) Both G363 and G401 are α-helix capping residues found in close-proximity to each other between neighbouring DRP1 monomers. Substitution of either glycine to a charged aspartate (G363D) or polar serine (G401S) would induce unfavourable steric clashes with neighbouring residues and disrupt helix stability. (C) L230 is located within a small α helix between the G4 and G5 loop motifs, critical for nucleotide binding. Addition of another leucine to this helix may disrupt these motifs, impairing GTP binding. (D) The helix containing L230 is adjacent to the MID49 binding surface and the L230 duplication in this location may have negative effects on MID49 binding and recruitment of DRP1 to the mitochondria. (E) The residue R710G, located within the bundle signalling element domain, forms a salt bridge with E702. The R710G substitution would induce a loss of this salt bridge.
- Figure 3. The impact of DNM1L variants on mitochondrial network length and DRP1 mitochondrial co-localisation.
(A) Representative images of TMRM-stained mitochondrial network in paediatric (C1 and C2) and adult (C3 and C4) controls and DNM1L patient (P1–P4) fibroblasts. (B) Quantification of mean mitochondrial network length via MiNa using ImageJ n > 20 fields from two independent experiments, calculated by multiplying mean branch length and mean number of branches per network. Non-parametric one-way ANOVA and Dunn’s multiple comparisons using GraphPad Prism were used to calculate statistically significant differences between groups. (C) Mitochondrial network length using immunofluorescence analysis of fixed paediatric control (C1), adult control (C2), and DNM1L patient (P1–P4) fibroblasts labelled with TOM20 antibodies. The Columbus (PerkinElmer) software was used to quantify the hyperfusion of patient mitochondrial networks relative to controls and a minimum of 5,500 mitochondria were analysed for each case. The immunofluorescence labelling was performed three times. (D) Analysis of DRP1 co-localisation with the outer mitochondrial membrane protein TOM20 by immunofluorescence labelling of age-matched controls (C1: paediatric, C2: adult) and DNM1L patient (P1–P4) fibroblasts with anti-DRP1 (red puncta) and anti-TOM20 (in blue). DRP1 co-localisation with mitochondria was analysed in at least 32 cells per subject in two independent experimental sets. Pearson’s correlations between DRP1 puncta and TOM20 in each cell line are shown as box plots. One-way ANOVA with post hoc Tukey’s honest significant difference test was used to determine statistically significant differences (***P ≤ 0.001). Representative merged immunofluorescence images of fibroblasts stained with anti-TOM20 and anti-DRP1 antibodies are shown on the left.
- Figure 4. The effect of DNM1L variants on peroxisomal morphology and co-localisation of DRP1 with peroxisomes.
(A) Analysis of peroxisome length by immunofluorescence using a peroxisomal membrane marker (PMP70) in fixed age-matched controls (C1: paediatric, C2: adult) and DNM1L patient (P1–P4) fibroblasts. The Columbus (PerkinElmer) software was used to quantify the peroxisome length between patients and controls. The immunofluorescence labelling was performed three times and a minimum of 300 peroxisomes were analysed in each case. Statistically significant differences between groups were determined by a non-parametric one-way ANOVA (***P ≤ 0.001). Representative images of fixed cells stained for peroxisomes (PMP70) in control (C1 and C2) and DNM1L patient (P1–P4) fibroblasts are shown on the left. (B) Immunofluorescence analysis of DRP1 puncta (red) co-localising with peroxisomes (PMP70 in blue) in age-matched control (C1: paediatric, C2: adult) and DNM1L patient (P1–P4) fibroblasts. The analysis was performed on at least 32 cells from two independent experimental sets and mean values showing Pearson’s correlation between the proportion of DRP1 puncta and peroxisomal marker PMP70 are shown. Statistically significant differences were calculated via a one-way ANOVA with post-hoc Tukey’s honest significant difference test (***P ≤ 0.001). Representative merged immunofluorescence images of PMP70 and DRP1 stained cells are shown on the left.
- Figure S1. Enlarged mitochondrial nucleoids identified in P1 and P2 DNM1L patient fibroblasts.
(A) Enlarged mitochondrial nucleoids observed in P1 (p.Gly401Ser), P2 (p.Glys363Asp), P3 (p.Leu230dup), and P4 (p.Arg710Gly) DNM1L patient fibroblasts incubated with TMRM. Nucleoids are indicated by yellow arrows. (B) Representative merged images of paediatric control (C1 and C2) and P1 (p.Gly401Ser) and P2 (p.Gly363Asp) fibroblasts stained with TMRM (mitochondria, red) and PicoGreen (mtDNA, green). Arrows highlight enlarged mitochondria with accumulations of mitochondrial DNA nucleoids. (C) Analysis of mtDNA nucleoids stained with PicoGreen fluorescent dye in age-matched C1 (paediatric), C2 (adult) controls, and DNM1L patient (P1–P3) fibroblasts. The proportion of enlarged nucleoids >1.5 μm2 was calculated in the total nucleoid pool. The minimum number of cells analysed from a random field of view was n = 21 and the smallest number of nucleoids analysed was 3,400 (n = 3). Statistical differences between groups were determined by one-way ANOVA test (*P ≤ 0.05; P > 0.05 n.s. [not significant]). (D) Representative images of adult control (C2) and DNM1L P4 (p.Arg710Gly) fibroblasts stained with PicoGreen dye. Yellow arrows show enlarged mitochondrial DNA nucleoids in P4 cells, which also present with increased accumulation of lipofuscin granules.
- Figure S2. Steady-state levels of fission machinery proteins in DNM1L patient fibroblasts.
Immunoblotting analysis of paediatric (C1 and C2) and adult controls (C3 and C4) and DNM1L P1 (p.Gly401Ser), P2 (p.Glys363Asp), P3 (p.Leu230dup), and P4 (p.Arg710Gly) patient fibroblasts showing the steady-state levels of DRP1 and MID51 proteins. The nuclear-encoded SDHA (Complex II) and β-actin were used as loading controls. Data information: Representative blots of three independent SDS–PAGE experiments are shown. Densitometric quantification of Western blots showing the mean % of relative band intensities between DNM1L patients (P1–P4) and control samples (SDHA–loading control, CP–paediatric control, CA–adult control). The error bars represent SD (n = 3).
Source data are available for this figure.
Source Data for Figure S2[LSA-2021-01284_SdataFS2_FS5_FS7.pdf]
- Figure S3. Diagnostic quadruple immunofluorescent assay showing complex I–immunodeficient muscle fibres in DNM1L P2.
Mitochondrial respiratory chain expression profile plot showing COXI, NDUFB8, and porin protein levels in single muscle fibres from P2 (p.Gly363Asp). Each dot represents an individual muscle fibre, colour-coded according to its mitochondrial mass (very low: blue; low: light blue; normal: beige; high: orange; very high: red). Thin black dashed lines indicate the SD limits for the classification of fibres, lines next to the x and y axes indicate the levels of NDUFB8 and COXI, respectively (beige: normal; light beige: intermediate (+); light blue: intermediate (++); blue: deficient). Bold dashed lines indicate the mean expression level of normal fibres.
- Figure S4. Diagnostic mitochondrial respiratory chain complex activities in DNM1L patients.
(A, B) Diagnostic mitochondrial respiratory complexes activities (complexes I–IV) measured in control and DNM1L patient P2 (p.Gly363Asp), P3* (age 13 yr), P3** (age 16 yr) (p.Leu230dup), and P4 (p.Arg710Gly) muscle (A), and in control and P2 fibroblasts (B). Data information: mitochondrial respiratory chain enzyme activities were normalized to the activity of the mitochondrial matrix enzyme, citrate synthase. Mean enzyme activities of control muscle (n = 25) or fibroblasts (n = 8) are set to 100%, and error bars represent SD.
- Figure S5. DNM1L patient fibroblasts demonstrate OXPHOS deficiencies.
Immunoblotting analysis of whole cell lysates from paediatric (C1 and C2) and adult controls (C3 and C4) and DNM1L P1 (p.Gly401Ser), P2 (p.Glys363Asp), P3 (p.Leu230dup), and P4 (p.Arg710Gly) fibroblasts showing a decrease in the levels of multiple OXPHOS subunits. The nuclear-encoded SDHA (Complex II) and GAPDH were used as loading controls. Data information: representative blots of three independent SDS–PAGE experiments are shown and densitometric quantification of Western blots are showing the mean % of relative band intensities between DNM1L patients (P1–P4) and control samples (GAPDH: loading control, CP: paediatric control, CA: adult control). The error bars represent SD (n = 3).
Source data are available for this figure.
Source Data for Figure S5[LSA-2021-01284_SdataFS2_FS5_FS7.pdf]
- Figure S6. Recombinant DRP1 WT and variants are well-folded.
Circular dichroism analysis of the far UV spectra of 0.05 mg/ml WT Drp1 (purple trace), Drp1 G363D (green trace), Drp1 G401S (orange trace), and Drp1 R710G (magenta trace) collected at 25°C, converted to mean residue ellipticity, and scaled at 260 nm to achieve a baseline of 0.
- Figure 5. Clinically identified DNM1L variants alter GTPase activity.
(A) Substrate kinetics of recombinant wild-type DRP1 (WT) (1 μM) and genetic variants. DRP1 GTPase activity was measured using an enzyme coupled assay monitoring NADH depletion, which is subsequently converted to activity (min−1). Data from three independent experiments were globally fit to a Michaelis–Menten model. Residuals of the fit are shown. (B, C, D) Distribution of K0.5, (C) kcat, and (D) kcat/K0.5 parameters from GTPase activity measurements. Reported values were obtained by globally fitting DRP1 GTPase activity measurements (n = 3) to a Michaelis–Menten model. The resulting values are reported in Table 2. K0.5 differences between WT and each variant significant to ***P < 0.05. kcat differences between WT and G363D, G363D and R710G, G363D and G401S, and R710G and G401S significant to ***P < 0.05. kcat/K0.5 differences between WT and both G363D and G401S, as well as between R710G and both G363D and G401S significant to ***P < 0.05.
- Figure 6. Patient DRP1 variants alter DRP1 assembly-state and melting temperature.
(A) SEC-MALS analysis of WT DRP1 (purple trace), DRP1 G363D (green trace), DRP1 G401S (orange trace), and DRP1 R710G (magenta trace) to assess for differences in multimeric distributions. Overlay of normalized differential refractive index of all protein samples (200 μg, 2.0 mg/ml) with peaks corresponding to monomeric, dimeric, and tetrameric oligomer species labelled as determined by predicted molecular masses of each multimeric species. Data normalized and scaled to allow for easier comparison because of slight differences in amount of protein loaded onto the column. (B) Melt curves of WT DRP1 and patient variants. Thermafluor analysis of protein unfolding of WT DRP1 (5.0 μM) and three patient variants (G363D, G401S, and R710G) either alone (black, dotted line), in the presence of 500 μM GDP (dark grey, dashed line), or GMP-PNP (light grey, solid line). Data plotted as the first derivative of the fluorescence signal with respect to time. (C, D) Tm values determined from the temperature corresponding to the maximum fluorescence value in the absence of and presence of 500 μM GDP or GMP-PNP. (C) Thermafluor analysis of the first protein unfolding event reported as the melting temperature (Tm) of WT DRP1 (5.0 μM) and three patient variants either alone, or in the presence of 500 μM GDP or GMP-PNP. (D) Thermafluor analysis of the second protein unfolding event reported as the melting temperature (Tm). Only WT and R710G shown as they are the only two constructs with a prominent second unfolding event. Data are representative of two independent experiments, each with three technical replicates. ***P < 0.00001. Differences between Tm values of all constructs alone in comparison with constructs with 500 μM GDP or 500 μM GMP-PNP significant to P < 0.003. Tm values of all constructs with 500 μM GDP in comparison to 500 μM GMP-PNP are significant to P < 0.03 except for R710G with 500 μM GDP in comparison to R710G with 500 μM GMP-PNP where P < 0.0003.
- Figure S7. Analysis of DRP1 oligomers in DNM1L patient fibroblasts.
(Left) Immunoblot analysis of total cell lysates isolated from age-matched control (C1 and C2: peadiatric; C3 and C4: adult) and DNM1L patient (P1 [p.Gly401Ser], P2 [p.Glys363Asp], P3 [p.Leu230dup], and P4 [p.Arg710Gly]) fibroblasts treated with DMSO only (−BMH) or the cross-linking reagent (+BMH) showing DRP1 higher order oligomers, DRP1 dimers (**), and DRP1 monomers (*). (Right) The immunoblot analysis was repeated in control (C4) and DNM1L P4 (p.Arg710Gly) fibroblasts showing DRP1 complexes and Ponceau staining as indicated above. In (Left) and (Right) equal amounts of total cell lysates (50 μg) were separated on a gradient (3–8%) Tris acetate gel. Data information: analysis for P1–P3 was performed once (Left) and a representative Western blot for control and P4 (p.Arg710Gly) is shown in (Right) with densitometric quantification of relative band intensities (%) for monomeric DRP1 between control and P4. SDHA (n = 3) and HSP60 (n = 2) were used as loading controls and the error bars represent SD.
Source data are available for this figure.
Source Data for Figure S7[LSA-2021-01284_SdataFS2_FS5_FS7.pdf]
Tables
ID DNM1L variants Clinical features Muscle biopsy and laboratory findings cDNA (NM_012062.5) Protein (NP_036192.2) Age-at-onset Clinical course Consanguinity; country of origin Clinical features and relevant biochemical findings Diagnostic muscle biopsy findings Diagnostic biochemical findings Patient 1a female c.1201G>A, p.(Gly401Ser) de novo heterozygous 8 mo Died, 10 mo No; UK Seizures, developmental delay, microcephaly, sudden deterioration in feeding and breathing, brain MRI normal, ECG and echocardiogram abnormal, end-stage dilated cardiomyopathy with previous signs of hypertrophic cardiomyopathy, raised 3-MGA type IV, and plasma lactate 7.0 mmol/l (normal range 0.7–2.1 mmol/l) Hyperfused and enlarged mitochondria, abnormal mitochondrial morphology with low cristae density on TEM Low complex IV ratio of 0.010 (0.014–0.034) in muscle Patient 2a,b female c.1088G>A, p.(Gly363Asp) de novo heterozygous Birth Died, 13 mo No; UK Seizures, growth failure, developmental delay, failure to thrive, microcephaly, micrognathia, infantile spasms, hypotonia, brain MRI abnormal, electroencephalogram abnormal—hypsarrthythmia, echocardiogram showed mild left ventricular hypertrophy, CSF lactate 4.6–7.0 mmol/l (normal range 0.7–2.1 mmol/l) n.d. Complex I–immunodeficient muscle fibres (IHC) and low complex I and II respiratory chain complex activities in muscle; low complex I activities in fibroblasts Patient 3c female c.687_689dupATT, p.(Leu230dup) de novo heterozygous 6 yr Died, 20 yr No; UK, Caucasian Learning difficulties, epilepsy, ataxia, dystonia, myoclonus and peripheral neuropathy, blood and CSF lactate normal, glucose concentrations normal, urine organic acid and plasma amino acid analysis normal Muscle electron microscopy and skin histology were not conclusive, but mainly normal Complexes I–IV normal in the 1st muscle biopsy. 2nd muscle biopsy 3 yr later showed decreased complex I and IV activity Patient 4a male c.2128A>G, p.(Arg710Gly) de novo heterozygous 3 yr Died, 17 yr No; UK Chronic inflammatory demyelinating polyneuropathy, extra-pyramidal movement disorder, epilepsy, optic atrophy, fatigue, and episodic regression of developmental skills precipitated by infection n.d. Mitochondrial respiratory chain activities (complexes I–IV) in muscle normal Patient 5d male c.1201G>A, p.(Gly401Ser) de novo heterozygous 33 mo Alive, 3 yr No; UK Caucasian Early onset epileptic encephalopathy, global developmental delay, hypotonia, nystagmus, dyskinesia, lactate and pyruvate concentrations in the CSF normal, plasma amino acids, urinary amino acids, organic acids and urine sialic acid normal n.d. n.d. - Table 2.
Reported kinetic values among DRP1 variants. Kinetic parameters (K0.5, Vmax, kcat, and kcat/K0.5) were computed for DRP1 WT and each clinical variant.
K0.5 ± SD (μM) Vmax ± SD (μM/min) kcat (min−1) kcat (min−1)/K0.5 (μM) WT 201 ± 51 0.24 ± 0.01 0.24 × 10−6 1.2 × 10−9 G363D 79 ± 11 0.58 ± 0.02 0.58 × 10−6 7.3 × 10−9 G401S 55 ± 9 0.36 ± 0.011 0.36 × 10−6 6.5 × 10−9 R710G 96 ± 18 0.10 ± 0.004 0.10 × 10−6 1.0 × 10−9
Table S1. Clinical, molecular genetics, biochemical and cellular findings in individuals with DNM1L variants. [LSA-2021-01284_TableS1.xlsx]
Supplemental Data 1.
DNM1L patients clinical case reports.[LSA-2021-01284_Supplemental_Data_1.docx]
