A human mitofusin 2 mutation can cause mitophagic cardiomyopathy

  1. Antonietta Franco
  2. Jiajia Li
  3. Daniel P Kelly
  4. Ray E Hershberger
  5. Ali J Marian
  6. Renate M Lewis
  7. Moshi Song
  8. Xiawei Dang
  9. Alina D Schmidt
  10. Mary E Mathyer
  11. John R Edwards
  12. Cristina de Guzman Strong
  13. Gerald W Dorn  Is a corresponding author
  1. Department of Internal Medicine, Pharmacogenomics, Washington University School of Medicine, United States
  2. Department of Medicine, Cardiovascular Institute, Perelman School of Medicine, University of Pennsylvania, United States
  3. Department of Internal Medicine, Divisions of Human Genetics and Cardiovascular Medicine, Ohio State University, United States
  4. Center for Cardiovascular Genetic Research, University of Texas Health Science Center at Houston, United States
  5. Department of Neurology, Washington University School of Medicine, United States
  6. Department of Internal Medicine (Dermatology), Washington University School of Medicine, United States
7 figures, 3 tables and 1 additional file

Figures

Figure 1 with 1 supplement
Morphological effects of disease-related MFN2 mutants on mitofusin null and normal mitochondria.

(A) Hypothetical structures of human MFN2 in the basal/closed (left) and active/open (right) conformations. Domains are color-coded: green, GTPase; red, first heptad repeat (HR1); blue, second heptad repeat (HR2). Arrows indicate positions of human disease-linked MFN2 mutants of primary interest. (B) Sequencing pherogram from a hypertrophic cardiomyopathy subject showing heterozygous MFN2 R400Q mutation (arrow). (C) Immunoblot of wild-type (WT) and mutant MFN2s expressed in Mfn1/Mfn2 double knockout (DKO) murine embryonic fibroblasts (MEFs). In all studies, adenoviral β-gal is negative control; β-actin is loading control. (D) Intrinsic fusogenic activity of WT and mutant MFN2 expressed in MFN DKO cells, assayed as increase in mitochondrial aspect ratio compared to β-gal. (E) Same as (D) except MFNs were expressed in WT MEFs to reveal effects on endogenous MFN1 and MFN2. Unless otherwise stated, in all graphs black is WT MFN2, green is GTPase domain MFN2 mutants, red is HR1 domain MFN2 mutants. Numbers in bars are independent experiments for 6–30 cells per experiment. p-Values used ANOVA and Tukey’s test.

Figure 1—source data 1

gnomAD MFN2 coding variants.

Red font indicates variants found in CM.

https://cdn.elifesciences.org/articles/84235/elife-84235-fig1-data1-v1.xlsx
Figure 1—source data 2

Mitofusin 2 (75 kDa) and β-actin (45 kDa) expression level in MEFs Wt.

https://cdn.elifesciences.org/articles/84235/elife-84235-fig1-data2-v1.zip
Figure 1—figure supplement 1
Fusogenic activities of GTPase-defective, CMT2A-linked MFN2 R94Q, and laboratory-engineered K109A in Mfn1/Mfn2 double knockout (DKO) murine embryonic fibroblasts (MEFs) (left) and WT MEFs (right).

β-gal and WT MFN2 are from Figure 1. p-Values calculated using ANOVA and Tukey’s test.

Figure 2 with 1 supplement
Effects of disease-related MFN2 mutants on catalytic and respiratory function of mitofusin null and normal mitochondria.

(A) GTPase activity of human disease-linked MFN2 mutants expressed in mitofusin double knockout (DKO) murine embryonic fibroblasts (MEFs). (B) Loss of inner membrane polarization in WT MEF mitochondria induced by disease-related MFN2 mutants. To the right are representative MitoTracker Green/TMRE co-stained confocal images. (C) Seahorse studies of oxygen consumption rate (OCR) in MFN DKO MEFs expressing disease-linked MFN2 mutants. (Left) Average data from three experiments per condition. (Right) Group quantitative results for basal, ATP-linked and maximal respiration (pmol/min/20,000 cells). *p<0.05 vs. WT.p-Values for all studies used ANOVA and Tukey’s test.

Figure 2—figure supplement 1
GTPase activity of, and inner membrane depolarization induced by, GTPase domain CMT2A-linked MFN2 R94Q and laboratory-engineered K109A expressed in Mfn1/Mfn2 double knockout (DKO) murine embryonic fibroblasts (MEFs) (left) and WT MEFs (right), respectively.

β-gal and WT MFN2 are from Figure 2. p-Values calculated using ANOVA and Tukey’s test.

Figure 3 with 1 supplement
MFN2 R400Q impairs conformational shifting and suppresses mitochondrial Parkin recruitment, but not mitochondrial motility in neuronal axons.

(A) Forster resonance energy transfer (FRET) studies of conformational switching in disease-linked MFN2 mutants expressed in MFN double knockout (DKO) murine embryonic fibroblasts (MEFs). Antagonist peptide normally induces closed conformation (increase in FRET) and agonist peptide open conformation (decrease in FRET). (B) Mitochondrial motility in cultured mouse dorsal root ganglion neuronal processes: representative kymographs are to the right (raw data above; motile mitochondrial emphasized below [red, antegrade; blue, retrograde]).Scale bar 10 μm, Data are n~3 5 DRGs for each independent experiment, statistic is one-way ANOVA. (C) mcParkin (red) recruitment to mitochondria in MFN DKO MEFs expressing MFN2 mutants before (baseline) and after FCCP (left, representative confocal micrographs; right, group quantitative data). p-Values calculated using ANOVA and Tukey’s test.

Figure 3—figure supplement 1
Effects of engineered synthetic MFN2 mutants on mitochondrial Parkin recruitment and mitochondrial motility in neuronal axons.

(A) Mitochondrial motility in cultured dorsal root ganglion (DRG) neuronal processes. Group quantitative data are to the left; representative kymographs are to the right. β-gal and WT MFN2 are from Figure 3. p-Values calculated using ANOVA and Tukey’s test. (B) Representative confocal micrograph showing Parkin localization (red) to mitochondria (green) after FCCP treatment. Arrows point to some co-localization events (yellow). Scale bar 10 μm, Data are n~3 5 DRGs for each independent experiment. (C) Parkin aggregation at mitochondria in MFN2 double knockout (DKO) murine embryonic fibroblasts (MEFs) in the absence (-) and presence (+) of FCCP; representative images to the right. MFN2 K109A is GTPase-defective; MFN2 EE constitutively binds Parkin; MFN2 AA cannot bind Parkin.

Figure 4 with 2 supplements
Perinatal cardiomyopathy in Mfn2 Q/Q400 mice.

(A, B) Mutant allele frequencies (A) and genotypes (B) from heterozygous crosses of Mfn2 T105M, M376V, and R400Q knock-in mice (CRISPR/Cas9 knock-in strategies are given in the figure supplement). Expected values are shown as dotted lines. Homozygous mice were absent for T105M and fewer than expected for R400Q. (C) Fetal and early postnatal lethality of homozygous R400Q knock-in mice (Q/Q400) reported as mice/litter at different time points. Controls (white) are C57/Bl6. Inset shows an E18.5 Q/Q400 litter having a degenerated fetus (arrowhead) and a live, non-viable fetus (arrow). (D) Hematoxylin and eosin-stained sections of left ventricular myocardium from P0 Q/Q400 mice with (left) and without (right) cardiomyocyte vacuolization. (E) Ultrastructural studies of Q/Q400 myocardium revealing myofibrillar degeneration (arrow), empty ‘ghost’ mitochondria (arrowheads) and mitochondrial fragmentation; wild-type control is shown for comparison on the right. *p<0.05 vs. expected (A, Fisher’s exact test) or control (C, Student’s t-test).

Figure 4—figure supplement 1
CRISPR engineering of disease-linked human MFN2 mutants into the mouse genome.

(A) Mfn2 R400Q; (B) Mfn2 T105M; and (C) Mfn2 M376V. For each mouse line, the left of the panel depicts the gene editing strategy, and to the right Sanger sequence of the resulting nucleotide changes. Note that no homozygous mice or embryos as early as E12.5 were identified for Mfn2 T105M in >40 breedings.

Figure 4—figure supplement 2
Quantitative mean data for transmission electron micrograph (TEM) studies of Mfn2 Q/Q400 mouse pups.

On the left are representative TEMs of myocardium. To the right are mean ± SEM data. Numbers are individual mice with >10 images per mouse; p-values by unpaired t-test.

Figure 5 with 3 supplements
Transcriptional and metabolomic profiling of Mfn2 Q/Q400 mouse hearts identifies metabolic abnormalities characteristic of mitophagy defects.

(A) Heat map of gene expression in individual Q400 and wild-type (WT) mouse hearts. (B) Heat map of anomalous Q400a mouse heart gene expression (top) and pie charts describing major KEGG functionally annotated pathway categories of up- (left) and downregulated (right) transcripts. (C) Individual heart mRNA levels for reactive oxygen species (ROS)-modulating enzymes. SOD, superoxide dismutase; GPX, glutathione peroxidase. Results shown are for the cardiac-expressed isoforms defined as WT CPM >100. Each point is an individual mouse heart. p-Values by ANOVA. (D) Metabolomics heat map showing unsupervised clustering of individual Q/Q400 and WT mouse hearts. (E–H) Relative expression of genes from indicated metabolic pathways. Blue is significantly decreased in MFN2 Q/Q 400a hearts; red is significantly increased; gray is no significant difference.

Figure 5—figure supplement 1
Expression of mitochondrial DNA-encoded respiratory genes (top) and mitochondrial biogenesis genes (bottom) in late fetal Mfn2 Q/Q400 mouse hearts.

Each point is an individual mouse heart. p-Values by ANOVA.

Figure 5—figure supplement 2
Expression of mitochondrial dynamics factor mRNAs in late fetal Mfn2 Q/Q400 mouse hearts.

Each point is an individual mouse heart. p-Values by ANOVA.

Figure 5—figure supplement 3
Metabolite levels in Mfn2 Q/Q400 mouse pup hearts.

Each point is an individual mouse heart. ‘Mutant’ indicates both abnormal and normal MFN2 Q400 mutants. p-Values from unpaired t-test.

Figure 6 with 2 supplements
MFN2 Q400 uniquely suppresses early and late mitophagy in cardiomyoblasts.

(A) Adenoviral expression of MFN2s in H9c2 cells. (B) Time courses of mitochondrial Parkin aggregation (circles), mitochondrial engulfment by autophagosomes (squares), and mitochondrial delivery to lysosomes (diamonds) in WT MFN2-expressing H9c2 cells after FCCP stimulation. Representative images are to the right. (C) Comparative time courses for mitochondrial Parkin aggregation (left), mitochondrial engulfment by autophagosomes (middle), and mitochondrial delivery to lysosomes (right) in MFN2 WT (black) and Q400 (red) expressing H9c2 cells. n = 3–4, *p<0.05 vs. pre-FCCP (time 0); #p<0.05 vs. WT at same time point (two-way ANOVA). (D, E) mcParkin aggregation 1 hr after FCCP treatment (D) and mitolysosome formation 8 hr after FCCP treatment (E) of H9c2 cells transduced with MFNs. Experimental n is shown in bars; stats by two-way ANOVA. Representative confocal images are shown in Figure 6—figure supplement 1.

Figure 6—source data 1

Mitofusin 2 (75kDa) and GAPDH (36 kDa) expression level in H9c2 cells.

https://cdn.elifesciences.org/articles/84235/elife-84235-fig6-data1-v1.zip
Figure 6—figure supplement 1
Representative mcParkin (A) and MitoQC (B) confocal images showing mitophagy provoked by FCCP in H9c2 cardiomyoblasts expressing WT and disease-linked MFN2 mutants.
Figure 6—figure supplement 2
Seahorse studies of oxygen consumption rate (OCR) in cultured H9c2 cardiomyoblasts expressing disease-linked MFN2 mutants.

(Lleft) Average data from 5 to 7 experiments per condition. (Right) Group quantitative results for basal, ATP-linked, and maximal respiration (pmol/min/20,000 cells). Statistical comparisons used ANOVA.

Mfn2 R400Q impairs reactive mitophagy and amplifies cardiomyocyte toxicity induced by doxorubicin in vitro and in vivo.

(A) Cell death curves provoked by increasing concentrations of doxorubicin (overnight treatment) in WT MFN2 (R400) and MFN2 Q400 transduced H9c2 cardiomyoblasts. Representative live-dead images corresponding to each group are to the right; scale bar 100 um. (B) TUNEL labeling of apoptotic H9c2 cardiomyoblasts treated with 10 uM doxorubicin overnight. Dotted line is basal level in absence of doxorubicin. (C) Reactive mitophagy measured as Parkin aggregation (left) and mitolysosome formation (right) in doxorubicin-treated H9c2 cardiomyoblasts. Dotted lines show basal levels in absence of doxorubicin. (D) Flow cytometric analysis of doxorubicin effects on mitochondrial-derived ROS in H9c2 cardiomyoblasts expressing WT MFN2 R400 or mutant MFN2 Q400. Representative (of three independent experiments) histograms are on the left; mean group data are on the right. *p<0.05 vs. same group without doxorubicin; #p<0.05 vs. WT cells treated identically (ANOVA). (E) Echocardiography of WT (R/R400), heterozygous (R/Q400), and homozygous (Q/Q400) knock-in mice 4 and 7 d after receiving doxorubicin (20 ug/g, intraperitoneal injection). LVEDD is left ventricular (LV) end diastolic dimension, which increases with dilated cardiomyopathy; FS is LV fractional shortening, which decreases in dilated cardiomyopathy; calculated LV mass increases in dilated cardiomyopathy. Representative m-mode echos are to the right. (F) Gravimetric heart weights 7 d after doxorubicin, indexed to tibial length. (G) MitoQC visualization of LV cardiomyocyte mitolysosomes. Group mean quantitative data on the left; representative confocal images on the right. Mitolysosomes are bright orange spots; scale bar is 10 um. (E–G): *p<0.05 vs. WT at the same time; $p=0.05–0.075 vs. WT at the same time; #p<0.05 vs. same genotype pre-doxorubicin; @p=0.05–0.075 vs. same genotype pre-doxorubicin; statistical comparisons used one- (F, G) or two-way (E) ANOVA.

Tables

Table 1
Results of genetic screening for MFN2 mutations in adult cardiomyopathy.

HCM, hypertrophic cardiomyopathy; DCM, dilated cardiomyopathy; MAF, minor allele frequency. MFN2 protein domains correspond to colored regions in Figure 1A.

MFN2 variantMFN2 domainPolyphen 2ClinVarHCM n = 286DCM1 n = 398DCM2 n = 281 probands (424 cases)CM MAF n = 965gnomAD MAF n = 141,456MAF CM vs. gnomAD (p-value)MAF CM vs. gnomAD (X2 stat)
S180RGTPaseBenignCMT2A1 hetSingleton1.591 E-05
V181MGTPaseProb damCMT2A1 family; 5 indivSingleton3.976 E-06
S263YGTPasePos dam1 hetSingleton---
R274QGTPaseBenignCMT2A1 family; 4 indivSingleton2.486 E-05
G298RGTPaseBenignCMT2A1 hetSingleton2.147 E-03
M393IHR1BenignLikely benign1 hetSingleton1.591 E-05
R400QHR1Prob damUncertain significance2 hets1 het1.55 E-037.423 E-05<0.0000189.595
Q413PHR1Benign1 family; 1 indivSingleton1.193 E-05
R468HinterPos damCMT2A1 het2 families; 5 indiv1.55 E-032.177 E-03NS0.3431
V705IHR2BenignLikely benign2 hets6 families; 15 indiv4.14 E-036.865 E-03NS2.0848
A716THR2Pos damCMT2A1 family; 1 indivSingleton1.308 E-04
Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Gene
(Mus musculus)
Mfn-2NCBI GeneGene ID: 170731MFN2
ENSMUSG00000029020
Gene
(human)
MFN-2NCBI GeneGene ID: 9927MFN2
ENSG00000116688
Genetic reagent (M. musculus)Mfn2 R400Q knock-in miceDornLabKnock-in mice generated with point mutation in position 400
Genetic reagent (M. musculus)Mfn2 T105M knock-in miceDornLabKnock-in mice generated with point mutation in position 105
Genetic reagent (M. musculus)Mfn2 M376V knock-in miceDornLabKnock-in mice generated with point mutation in position 376
Genetic reagent (M. musculus)C57BL/6J miceC57Bl/6The Jackson Laboratory:
000664
C57Bl/6
Cell line
H9c2
(rat myoblast)
CardiomyoblastATCCCRL-1446Rat embryonic cardiomyoblast
Cell line
Mfn1/Mfn2 null
(M. musculus)
Mfn1 and Mfn2 double knock out MEFsATCCCRL-2993Murine embryonic fibroblasts
Cell line
DRG neuronal cells
Adult mouse dorsal root ganglionFrom 8- to 12-week-old C57BL/6J using enzymatic dissociation.Franco et al., 2020
Cell line
mouse embryonic fibroblast
MEFs WTFrom E.14.5 pups from C57BL/6J using enzymatic dissociationhttps://pubmed.ncbi.nlm.nih.gov/18265203/
Transfected construct (human adenovirus type 5 dE1/E3)Adenovirus
Mito-Ds-Red2
SignagenCat# 12259
Transfected construct (human adenovirus type 5 dE1/E3)Adenovirus
Cre-recombinase
Vector BiolabsCat# 1794
Transfected construct (human adenovirus type5 dE1/E3)Ad-MFN2T105MDornLabFranco et al., 2022
Transfected construct (human adenovirus type 5 dE1/E3)Ad-MFN2R94QDornLabFranco et al., 2022
Transfected construct (human adenovirus type 5 dE1/E3)Ad-MFN2R400QDornLabHuman adenovirus type 5 (dE1/E3), with Mfn2R400Q
Transfected construct (human adenovirus type 5 dE1/E3)Ad-MFN2K109ADornLabRocha et al., 2018
Transfected construct (human adenovirus type 5 dE1/E3)Ad-MFN2M376ADornLabRocha et al., 2018
Transfected construct (human adenovirus type 5 dE1/E3)Ad-MFN2WTDornLabFranco et al., 2022
Transfected construct (human adenovirus type 5 dE1/E3)Ad-hFRETMFN22WTDornLabRocha et al., 2018
Transfected construct (human adenovirus type 5 dE1/E3)Ad-hFRETMFN2R94QDornLabFranco et al., 2022
Transfected construct (human adenovirus type 5 dE1/E3)Ad-hFRETMFN2K109ADornLabHuman adenovirus type 5 (dE1/E3), with Mfn2K109A
Transfected construct (human adenovirus type 5 dE1/E3)Ad-hFRETMFN2M376ADornLabFranco et al., 2022
Transfected construct (human adenovirus type 5 dE1/E3)Ad-hFRETMFN2T105MDornLabFranco et al., 2022
Transfected construct (human adenovirus type 5 dE1/E3)Ad-LC3-GFPDornLabHuman adenovirus type 5 (dE1/E3), with LC3 for autophagy
Transfected construct (human adenovirus type 5 dE1/E3)Ad-mCherry-ParkinGift from Dr. Åsa GustafssonN/A
Transfected construct (human adenovirus type 5 dE1/E3)Ad-mitoQCDornLabHuman adenovirus type 5 (dE1/E3), with mitoQC for mitophagy
AntibodyAnti-Mfn-2 (mouse monoclonal)AbcamCat# ab568891:1000
AntibodyAnti-COX-IV (rabbit polyclonal)AbcamCat# ab160561:1000
AntibodyAnti-GAPDH (mouse monoclonal)AbcamCat# ab822451:3000
AntibodyAnti-β-actin (unconjugated monoclonal)ProteintechCat# 66009-11:3000
AntibodyGoat anti-rabbit IgGThermo FisherCat# 314601:3000
AntibodyPeroxidase-conjugated anti-mouse IgGCell SignalingCat# 7076S1:3000
Sequence-based reagentmMFN2R400Q knock-in mouse (forward)DornLabKnock-in mice generated with point mutation in position 400GGCATGTATGTGTAGGTCAGAG
Sequence-based reagentmMFN2 R400Q knock-in mouse (reverse)DornLabKnock-in mice generated with point mutation in position 400CCCAGCTCCTCTGATTTGA
Sequence-based reagentmMFN2 R400Q knock-in mouse (sequencing)DornLabKnock-in mice generated with point mutation in position 400CAGGTCTCCTTCCACCTTTAC
Sequence-based reagentmMFN2 T105M knock-in mouse (forward)DornLabKnock-in mice generated with point mutation in position 105TGTTTACTTTGGAAGTAGGCAGTCT
Sequence-based reagentmMFN2 T105M knock-in mouse (reverse)DornLabKnock-in mice generated with point mutation in position 105TTGTTCTTGGTGTCCCACTCTGA
Sequence-based reagentmMFN2 T105M knock-in mouse (sequencing)DornLabKnock-in mice generated with point mutation in position 105TCGATGCTTAATGAGTGCTGCTGG
Sequence-based reagentmMFN2 M376V knock-in mouse (forward)DornLabKnock-in mice generated with point mutation in position 376GTCCGGGCCAAGCAGATTGCAGAGGCCGTTCGTCTCATCATGGATTCCCTGCACATCGCAGCTCAGGAGCAGCGGTGAGA
Sequence-based reagentmMFN2 M376V knock-in mouse (reverse)DornLabKnock-in mice generated with point mutation in position 376GGTAGTAAAGAGCCTTTCTAGCTGAT
Sequence-based reagentmMFN2 M376V knock-in mouse (sequencing)DornLabKnock-in mice generated with point mutation in position 376TTTCGAGAGGCAGTTTGAGGTAA
Commercial assay or kitGTPase-Glo AssayPromegaCat# V7681
Chemical compound, drugMitofusin agonist
TAT-MP-1
Thermo FisherFranco A et al., Nature 2016
Chemical compound, drugMitofusin antagonist
TAT-MP-1
Thermo FisherFranco A et al., Nature 2016
Chemical compound, drugL-GlutamineGibcoCat# 25030-149
Chemical compound, drugGoat serumJackson ImmunoResearchCat# 005-000121
Chemical compound, drugDoxorubicinSigma44583
Chemical compound, drugGlutaraldehydeElectron Microscopy ScienceCat# 16216
Chemical compound, drugMitoTracker GreenThermo FisherCat# M7514
Chemical compound, drugFCCPThermo FisherCat# C2920
Chemical compound, drugCalcein AMThermo FisherCat# C3100MP
Chemical compound, drugHematoxylin-eosinSigmaCat#: GHS116 & HT110116
Chemical compound, drugTUNELPromegaCat#: G3250
Chemical compound, drugHoechstThermo FisherCat# H3570
Chemical compound, drugMitoTracker OrangeThermo FisherCat# M7510
Chemical compound, drugTetramethylrhodamine ethyl esterThermo FisherCat# T-669
Chemical compound, drugMitoSOX RedThermo FisherCat# M36008
Software, algorithmImageJA. Schneiderhttps://imagej.net/Sholl_Analysis
Software, algorithmPartek Flowhttps://www.partek.com/partek-flow/N/A
Software, algorithmFlowJo 10 softwarehttps://www.flowjo.com/solutions/flowjo/downloads/N/A
Author response table 1
MFN2 mutantGTPase activityConformation changeFusionPolarizationRespirationMitophagyMotilityKI phenotype
T105MInactiveNormalDominant suppressorDominant suppressorDominant suppressorNormalDominant suppressorCMT2A (het)
Emb let (homo)
M376A/VNormalImpairedFunctional nullNormalFunctional nullNormalNormalNormal
R400QNormalImpairedDominant suppressorNormalFunctional nullDominant suppressorNormalCardiomyopathy (homo)

Additional files

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Antonietta Franco
  2. Jiajia Li
  3. Daniel P Kelly
  4. Ray E Hershberger
  5. Ali J Marian
  6. Renate M Lewis
  7. Moshi Song
  8. Xiawei Dang
  9. Alina D Schmidt
  10. Mary E Mathyer
  11. John R Edwards
  12. Cristina de Guzman Strong
  13. Gerald W Dorn
(2023)
A human mitofusin 2 mutation can cause mitophagic cardiomyopathy
eLife 12:e84235.
https://doi.org/10.7554/eLife.84235