Mfn2/Hsc70 Complex Mediates the Formation of Mitochondria‐Lipid Droplets Membrane Contact and Regulates Myocardial Lipid Metabolism

Abstract The heart primarily derives its energy through lipid oxidation. In cardiomyocytes, lipids are stored in lipid droplets (LDs) and are utilized in mitochondria, although the structural and functional connections between these two organelles remain largely unknown. In this study, visible evidence have presented indicating that a complex is formed at the mitochondria‐LD membrane contact (MLC) site, involving mitochondrion‐localized Mfn2 and LD‐localized Hsc70. This complex serves to tether mitochondria to LDs, facilitating the transfer of fatty acids (FAs) from LDs to mitochondria for β‐oxidation. Reduction of Mfn2 induced by lipid overload inhibits MLC, hinders FA transfer, and results in lipid accumulation. Restoring Mfn2 reinstates MLC, alleviating myocardial lipotoxicity under lipid overload conditions both in‐vivo and in‐vitro. Additionally, prolonged lipid overload induces Mfn2 degradation through the ubiquitin‐proteasome pathway, following Mfn2 acetylation at the K243 site. This leads to the transition from adaptive lipid utilization to maladaptive lipotoxicity. The experimental findings are supported by clinical data from patients with obesity and age‐matched non‐obese individuals. These translational results make a significant contribution to the molecular understanding of MLC in the heart, and offer new insights into its role in myocardial lipotoxicity.


Echocardiography
Echocardiography was performed with a VEVO 3100 echocardiography system (VisualSonics Inc., Toronto, Canada) as previously described 1 .Generally, mice were anaesthetized with 2.5% isoflurane and anesthesia was maintained with 2% isoflurane.
In order to rule out the influence of heart rate variation on cardiac function, all mice subjected to echocardiography are anesthetized suitably to maintain a heart rate of 400-450 times per minute.A real time ECG monitoring equipment were used to detect mice heart rate while performing echocardiography.M-mode echocardiography was obtained to record the left ventricular systolic and diastolic motion profile.Doppler echocardiography was obtained to determine diastolic trans-mitral blood flow velocities for peak early (E) and late (A) fillings.All the echocardiographic images were analyzed using Vevo 3100 software as previously described 1 .

Biochemical and Anthropometric analysis in mice
Body weight and heart weight were measured in all mice.Low density lipoprotein (LDL)-cholesterol, high-density lipoprotein (HDL)-cholesterol and triglycerides were detected by clinical laboratory of Tangdu Hospital using a Chemray800 automatic biochemical analyzer (Rayto, Shenzhen, China) with standard protocol.Blood glucose level were measured with a sip-in sampling glucose meter (Glucometer 580, Yuwell, China).

Tissue harvesting
Mice were euthanized by CO2.After chest cavity was opened, the heart was cut out and perfused with PBS.For molecular analysis, the organ was placed immediately in ice cold PBS to remove blood.Then the organ was conserved in snap-frozen in liquid nitrogen.For histological analysis, tissue was fixed in 4% paraformaldehyde solution.
For transmission electron microscopy, heart samples from the left ventricular free wall were isolated and fixed with 2.5% glutaraldehyde in a 0.1 M phosphate buffer.

Transmission electron microscopy (TEM)
TEM analyze was performed as previously described 2 .All images were collected by a by a technician blinded to the treatment with one transmission electron microscope (JEM-1230, JEOL Ltd., Tokyo, Japan) at 300 kV.TEM images were analyzed by a technician blinded to the treatment using Image J software.The morphology of mitochondria and lipid droplet were analyzed as previously described 3 .The number of LDs combined with LDs diameter were used for the analysis of intramyocardial LDs accumulation.Percent of LDs contact with mitochondria and mean LDsmitochondria contact length were calculated to reflect mitochondria-LDs contact as previously described 3 .For calculation of percentage of LDs with direct contact to mitochondria, 6 visual field was randomly selected and the LDs number was recorded.
LDs which had direct contact with mitochondria refers to that LDs and mitochondria closely contacted with each other with no visible gaps were present.In contrast, segregated LDs with no visible connection to the mitochondria was excluded.Then, the percentage of LDs with direct contact to mitochondria was calculated.All calculation and measurements were taken independently by one experienced technician who was blinded to the experimental design.

Lipidomics
For lipidomics, heart sample were collected and conserved in liquid nitrogen until lipid extraction.Lipidomics experiments were performed as described previously 4,5 .
samples were thawed on ice, and metabolites were extracted from 20 µL of each sample using 120 µL of precooled 50% methanol buffer.Then the mixture of metabolites was vortexed for 1 min and incubated for 10 min at room temperature, and stored at -20°C overnight.The mixture was centrifugated at 4,000 g for 20 min, subsequently the supernatant was transferred to 96-well plates.The samples were stored at -80°C prior to the LC-MS analysis using a TripleTOF 5600 Plus highresolution tandem mass spectrometer (SCIEX, Warrington, UK) with both positive and negative ion modes.The acquired LC-MS data pretreatment was performed using XCMS software.The open access databases, KEGG and HMDB, were used to annotate the metabolites by matching the exact molecular mass data to those from the database within a threshold of 10 ppm.
For construction of adenovirus encoding mitochondrial-targeted red fluorescent protein (peptides), the mitochondrial targeting elements of cox8 (mitochondrial targeting sequence), a mitochondrial outer membrane protein, was fused with Cy3-labeled full-length or fragments of Mfn2 protein.

Intra-myocardial injection of adeno-associated virus (AAV)
For intramyocardial injection of AAV, mice were anaesthetized using 2.5% isoflurane, intubated and maintained under anesthesia with 2% isoflurane during the surgical procedure.After the hearts were exposed, AAV was injected (using a 50 µl needle, Hamilton, 705RN, USA) into the left ventricle free wall (10 µl at each of four sites).
Total RNA was extracted with RNAisoPlus (Takara, Japan), cDNA was synthesized with a PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara), and quantitative RT-PCR was performed with SYBR® Premix Ex Taq™ II (Takara) following the manufactures' protocols.Primers used in this study are all listed in Table S1.

Histological Analysis
Oil Red O staining were performed following standard procedures as previously described 7 .All steps were performed on the machine with Ventana solutions.

Primary culture of neonatal rat cardiomyocytes
Primary cardiomyocytes were prepared from neonatal rat hearts as previously described 1 .Thereafter, primary cardiomyocytes were subjected to control medium (Con) or control medium containing palmitate (Pal) 8 .

Immunofluorescence microscopy
Cells grown on confocal dishes (NEST Biotechnology) were fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.01% Triton X-100 in PBS, and blocked with 1% BSA in PBS.Cells were stained overnight with primary anti-Mfn2

Fluorescent imaging of fatty acids (FAs), LDs and mitochondria
For fatty acid pulse-chase, cells were incubated with culture medium containing 1uM Bodipy 558/568 C12 (Red C12, Life Technology, USA) for 16 hours.Cells were then washed 3 times with HBSS buffer before confocal imaging.Mitochondria were labeled by 100nM Mitotracker-red or Mitotracker-green (Life Technology, USA).To view LDs, Bodipy 493/503 (Life Technology) was added to cells at 200 ng/ml immediately before confocal detection.All fluorescent images were acquired with a confocal laser-scanning microscope (Nikon A1 plus Confocal Microscope, Nikon, Japan).For FAs-LDs or mitochondria overlap coefficient analysis, ImagJ Colocalization Analysis software (https://imagej.net/imaging/colocalization-analysis)was used to calculate the overlap coefficient.For tethering between LDs and mitochondria analysis, the percentage of mitochondrial signal connected to LDs signal versus total LD signal was calculated.

FAs transfer pulse-chase assay
Neonatal cardiomyocytes were incubated with control medium containing 300μM palmitate and 1μm Bodipy 558/568 red C12 for overnight.Palmitate were employed to induce LDs biogenesis while red C12 were used to label fatty acids.Cells were then washed three times with PBS and chased every 5 minutes for 30 minutes in HBSS.

LD isolation
LDs were isolated from mice heart according to the methods reported by Zhang et al 9,10 .Specially formulated buffer A (25 mM tricine pH 7.6, 250 mM sucrose) and buffer B (20 mM HEPES, pH 7.4, 100 mM KCl and 2 mM MgCl2) were prepared before isolation.Mice heart sample were transferred to a tube containing Buffer A with 0.2mM PMSF.Then the sample was homogenized with a p polytron homogenizer for 30s with a SpeedMill Plus homogenizer (Analytik Jena) for 25 times.The homogenates were centrifuged at 3000g for 30 min to remove the debris.Supernatant with 3mL buffer B was loaded into an UltraClear centrifuge tube (Beckman Coulter).
The tube was centrifuged at 288 000 g for 54 min at 4°C.Then, a light-yellow band containing pure LDs (See supplementary figures) on the top of the tube was collected.

LC-MS/MS analysis
Liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis was performed using a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific) coupled with EASY-nLC 1200 UHPLC system (Thermo Fisher Scientific) at Novogene Genetics (Beijing, China) as previously described 6 .The 40 most abundant precursor ions from full MS scan were selected for analysis at a resolution of 15,000 (at 200 m/z).Raw files of mass spectrometry test were identified by MaxQuant (www.maxquant.org).Spectra data were searched refer to a UniProt mouse database (www.uniprot.org).

Proximity ligation assay (in situ PLA)
Proximity ligation assay was carried out following the instructions of Duolink In Situ Red Starter Kit Mouse/Rabbit (Sigma-Aldrich, DUO92101).Briefly, cells were fixed with 4% paraformaldehyde for 10 minutes at room temperature and then permeabilized for 5 minutes with 0.5% Triton X-100.The following steps were performed according to instructions of Duolink® Proximity Ligation Assay Kit.Cells were incubated with primary antibodies (Mfn2, 1:1000, Abcam; Hsc70, 1:1000, Abcam) at toom temperature for 1 hour.Images were captured with a Nikon A1R Scanning Laser Microscope.

Plasmid's construction and transfection.
Full length Mfn2 and specific-sequence Mfn2 were cloned to a Flag-tagged destination vectors for different detections.Point mutation for Mfn2 was conducted by site-directed mutagenesis.Plasmid transfection in 293T cell line were conducted with Lipofectamine 3000 Transfection Reagent (Invitrogen) according to the manufacturer's instructions.

Study population
Obese patients (BMI≧30) and non-obese age-matched controls were continuously Table S2 Mass spectrum analysis of Mfn2-interacting proteins.

Table S3
Mass spectrum analysis of LD proteins.

Figure S2
Figure S2 Mfn2 protein level were determined and MLC were detected by labeling mitochondria (red) and LDs (green) respectively in palmitate-treated cardiomyocytes.A&B, Cardiomyocytes were treated with palmitate at different concentrations, then LDs were labeled with Bodipy 493/503 (green) and cell viability was examined by CCK8 assay.C, Cardiomyocytes were treated with palmitate, then LDs were labeled with Bodipy 493/503 (green) and mitochondria was labeled with Mito-tracker Red (red) at time points indicated.D, LDs area per cell (Mean± SEM, n=5 independent experiments, 30 cells were quantified per group, **P<0.01).E, Percentage of LDs with direct contact to mitochondria was measured (Mean± SEM, n=5 independent experiments, 30 cells were quantified per group, **P<0.01).F&G, Mfn2 protein level were determined by Western-bolt at different time point of palmitate treating.

Figure S3 ,
Figure S3, A Red C12 pulse-chase assay were employed to further determine the traffic of FAs from LDs to mitochondria.A, Schematic representation of the red C12 pulse-chase assay to visualize FAs transport from LDs to mitochondria.B, Red C12 Pulse-chase assay showing the intracellular transport of FAs.Degraded LDs during detection are visible as bright spots (green arrow) and mitochondria are visible as faint lines (white dashed lines).

Figure S4
Figure S4 Mfn2 overexpression promoted lipid oxidation in mitochondria.A, Oxygen consumption rate (OCR) in the presence or absence of etomoxir.B, Basal respiration due to utilization of exogenous FAs.C, Maximal respiration due to utilization of exogenous FAs.Data presented as mean ± SEM, n=4 independent experiments.Differences are significant for **P<0.01.

Figure S6
Figure S6 Lipid droplet were perfectly isolated and interaction of GRP78 and Mfn2 were detected.A, Lipid droplets were identified by Western-blot analysis.VDAC1, marker of mitochondria; Calreticulin, marker of ER; β-actin, marker of

Figure S8 ,
Figure S8, Schematic diagrams of conserved domains shared by Mfn1 and Mfn2.

Figure
Figure S9 Mfn2-driven mitochondrial lipid oxidation were blunted by Hsc70 knockdown.A, Oxygen consumption rate (OCR) in the presence or absence of etomoxir.B, Basal respiration due to utilization of exogenous FAs.C, Maximal respiration due to utilization of exogenous FAs.Data presented as mean ± SEM, n=4

Figure
Figure S10 Ad-MTS-649-693aa failed in inducing mitophagy and mitochondria-SR tethering in the cardiomyocytes.A, Mitochondria and SR were labeled with red and green (Calreticulin) fluorescence respectively, and the relative co-efficient was analyzed.B, Mitophagy was determined by LC3 puncta (green) colocalized with mitochondria (red).

Figure
Figure S11 Cardiac-specific knockout of Mfn2 in Mfn2 CKO mice was confirmed by western-bolt and RT-PCR analysis.A, Representative blot images of Mfn2 protein level in different tissue from Mfn2 fl/fl and Mfn2 CKO mice (n= 4 mice per group).B, mRNA level of gene involved in lipolysis (PPARα, ATGL, CPT-1b, CPT-2) and lipid uptake (CD36) were determined by RT-PCR (n= 8 mice per group).Data presented as Mean ± SEM.Differences are significant for **P<0.01.ns, no significance.

Figure S12
Figure S12 Mass spectrometry (MS)-based quantification of cardiac lipid species across the different experimental groups.A, Heatmap of metabolic alterations organized by lipid class.B-E, Quantification showing the relative levels of lipids in 4 four groups indicated.Data presented as Mean ± SEM, n=3 mice per group.

Figure S13 .
Figure S13.Cardiac cardiolipin level was determined in heart of Mfn2CKO mice.A-B, Heatmap of cardiolipin and quantification showing the relative levels of cardiolipin in 4 four groups indicated.

Figure
Figure S14 Cardiac-specific transgene of Mfn2 was confirmed by western-bolt and RT-PCR analysis.A, Representative blot images of Mfn2 protein level in different tissue from NTG and Mfn2 TG mice (n=4 mice per group).B, mRNA level of gene involved in lipolysis (PPARα, ATGL, CPT-1b, CPT-2) and lipid uptake (CD36) were determined by RT-PCR (n=8 mice per group).Data presented as Mean ± SEM.

Figure S17
Figure S17 Mfn2 protein level and mRNA level were quantified in hearts of HFD-fed mice at different time points.Data presented as mean ± SEM, n=6 mice per group.Differences are significant for **P<0.01,*P<0.05.