A new non-canonical pathway of Gα_q protein regulating mitochondrial dynamics and bioenergetics

Abstract

Contrary to previous assumptions, G proteins do not permanently reside on the plasma membrane, but are constantly monitoring the cytoplasmic surfaces of the plasma membrane and endomembranes. Here, we report that the Gα_q and Gα₁₁ proteins locate at the mitochondria and play a role in a complex signaling pathway that regulates mitochondria dynamics. Our results provide evidence for the presence of the heteromeric G protein (Gα_q/11βγ) at the outer mitochondrial membrane and for Gα_q at the inner membrane. Both localizations are necessary to maintain the proper equilibrium between fusion and fission; which is achieved by altering the activity of mitofusin proteins, Drp1, OPA1 and the membrane potential at both the outer and inner mitochondrial membranes. As a result of the absence of Gα_q/11, there is a decrease in mitochondrial fusion rates and a decrease in overall respiratory capacity, ATP production and OXPHOS-dependent growth. These findings demonstrate that the presence of Gα_q proteins at the mitochondria serves a physiological function: stabilizing elongated mitochondria and regulating energy production in a Drp1 and Opa1 dependent mechanisms. This thereby links organelle dynamics and physiology.

Introduction

Heterotrimeric G proteins, consisting of an α subunit and a complex formed of β γ subunits, are well-established mediators of signal transduction pathways downstream from G protein-coupled receptors (GPCRs). For many years it was believed that G proteins perform their function at or close to the plasma membrane. Only recently did it become evident that G proteins can be localized at and signal to different endomembranes, including the endoplasmic reticulum (ER) and Golgi, and that their localization can be highly dynamic ¹. Recent findings have identified the mitochondria as a non-canonical localization for G proteins, including Gα₁₂ ², Gα_i ³ and Gβ2 ⁴. Moreover, recent reports confirm that some G protein-effectors or binding partners, such as MAPKs, Akt, GRK2 and PKC, are also present at the mitochondria; particularly at the outer mitochondrial membrane and in the intermembrane space ^{5, 6}, which suggests that this new localization of G proteins may be functionally important.

Of the different types of Gα, the Gα_q family members (including Gα_q, Gα₁₁, Gα₁₄ and Gα_15/16) ⁷ stimulate the β-isoform of phosphoinositide phospholipase C (PLC-β), which in turn increases inositol lipid (i.e., calcium/PKC) signaling ⁸. The members of the human Gq family, Gα₁₁, Gα₁₄ and Gα₁₆, share approximately 90%, 80% and 57% homology, respectively, of their amino acid sequence with Gα_q ⁷. Most downstream cellular responses result from enhanced calcium signaling, but growing evidence indicates that other events may account for some of the physiological roles of Gα_q family members ⁸. A growing list of scaffolding/adaptor proteins (caveolin-1 ⁹, EBP50/NHERF1 ¹⁰, CD9/CD81 ¹¹, Flotilin ¹², TRP1 ¹³), regulatory proteins (RGS ^{14, 15}), GRKs ^{16, 17}, effectors (RhoGEFs ¹⁸, Btk ¹⁹, PKCζ/ERK5 ²⁰) and activator proteins (Ric-8A ²¹, tubulin ²²) may help to explain some of the unexpected signaling pathways that they regulate. The importance of different subcellular localizations of Gα_q responses is still a matter of study.

Mitochondria are essential organelles enveloped by two close but opposed membranes. The outer membrane mediates exchange between the cytosol and intermembrane space, while the inner membrane delimits the matrix space and contains respiratory complexes for oxidative phosphorylation (OXPHOS) ²³. Mitochondria can be highly dynamic organelles that fuse and divide in response to environmental stimuli, developmental status, and the energy requirements of the cell ^24–26. These events are regulated by specific proteins involved in fission and fusion, and also in the maintenance of mitochondrial distribution ^{27, 28}. The most notable proteins involved in mitochondrial fission/fusion processes are: the dynamin-like protein DLP1/Drp1; the small helix-rich proteins Fis1 and Mff, linked to outer mitochondrial membrane fission. The dynamin-related GTPases, mitofusins (Mfn1/2), and optic atrophy 1 (OPA1), associated with the outer and inner membrane, respectively, mediate fusion of the membranes ^28–33.

The presence of signaling molecules at the mitochondria highlights the possibility of novel signaling pathways that control energy production. In the search for mitochondrial localized heterotrimeric G proteins, proteomic analysis together with fractionation and immunofluorescence analysis show that Gα_q and Gα₁₁ target mitochondria through their N-terminal sequence. Herein, we demonstrate that Gα_q proteins are necessary for maintenance of the proper balance between mitochondrial fusion and fission processes, and consequently for regulating the respiratory capacity of mitochondria.

Materials and Methods

Materials

pcDNA3-Gα_q and pcDNA3-Gα_q-R183C were as described elsewhere ⁷². pcDNAI-Gα_q-GFP was generously provided by C. Berlot (Yale University School of Medicine, USA). Gα_q-N-terminus (1–124 aas) in pEGFP was cloned from pcDNAI-Gα_q-GFP, and Gα_q-N-terminus-FLAG in pcDNA3 was amplified by PCR. The Gα_q-N-I25/26E mutant in pEGFP was amplified by PCR using pcDNA3-Gα_q-I25/26E ⁷² as a template. Mt-DsRed and mt-GFP were cloned from pWPXL-mt-DsRed ⁷³ and pWPXL-mt-GFP ⁷³, respectively. The ER marker was obtained from Clontech (USA). pcDNA3-Gβ₁-FLAG, Gβ₂-FLAG, Gβ₄-FLAG and Gγ₂-HA were obtained from Missouri S&T cDNA Resource (USA). HA-tagged Drp1 and Drp1-K38A in pcDNA3 were kindly provided by A. van der Bliek (University of California, LA, USA). PA-mitoGFP ⁷⁴ and Plasmid 23348 were purchased from Addgene (USA). The antibodies used were: Gα_q, Gα_q/11, Gβ and TOM20 (Santa Cruz Biotechnology); Gα_q internal, Porin-VDAC and Porin 31 HL (Calbiochem); SERCA2, Golgi [58K-9], GAPDH, Pan-cadherin, Mitofusin-1, Mitofusin-2 and LAMP1 (Abcam); Smac/diablo, Rab11, Drp1, COXI and OPA1 (BD Biosciences); Caveolin-1 (Zymed); complex II (anti-αFp70 kDa subunit) (Invitrogen); Hsp70 (BioReagents); FIS1 (BioVision); HA-tag (Roche); M2-Flag (Sigma Aldrich); complex I (anti-NDUFA9), complex III (anti-Core1) and complex V (anti-βATPase) (MitoSciences); and complex IV (anti-NDUFA4) (BioWorld Technology). Gα_q transgenic mice which over-expresses mouse Gα_q exclusively in the myocardium were a generous gift from G. W. Dorn II, Washington University at St Louis ³⁷.

Cell culture, lysis and immunoprecipitation

NIH3T3 cells were obtained from the American Type Culture Collection (ATCC) (Manassas, VA, USA). Human embryonic kidney cells (HEK293T) were from Invitrogen (Carlsbad, CA, USA). WT, and knockout Gα_q/11^−/− and Gα_12/13^−/− MEFs were provided by S. Offermanns, (University of Heidelberg, Germany). WT and knockout Mfn1^−/− and Mfn2^−/− MEFs were a gift from D.C. Chan (Division of Biology, California Institute of Technology, UA). HeLa cells stably expressing mt-DsRed are described elsewhere ⁷³. All cell types were grown in Dulbecco’s modified Eagle’s medium (DMEM) (Sigma, St. Louis, MO, USA), supplemented with 10% fetal bovine serum (FBS) (Invitrogen, GIBCO, USA). Cells were transiently transfected either with Metafectene Pro (Biontex, USA) or Fugene 6 (Roche, Switzerland) according to the manufacturer’s instructions.

Cells were washed twice with ice-cold PBS prior to lysis in 700 µl of RIPA buffer A (0.3 M NaCl, 0.1% SDS, 50 mM Tris, pH 7.4, 0.5% deoxycholate, 1 mM Na₃VO₄, 10 mM NaF, 30 mM sodium pyrophosphate, 10 mM MgCl₂, 1% n-dodecyl β-D-maltoside, leupeptin 5 µg/ml, aprotinin 2 µg/ml, 1 mM PMSF) for 1 h at 4°C. Extracts were cleared by centrifugation at 13,000 rpm for 15 min at 4°C and the protein concentration was determined by Bradford analysis. For OPA1 immunoprecipitation, an anti-OPA1 antibody was incubated at 4°C overnight. Protein G-sepharose was added and incubated for 1 h, and then washed several times with RIPA buffer B (buffer A with 0.01% n-dodecyl β-D-maltoside and without protease inhibitors). The samples were resuspended in Laemmli buffer. Cells were visualized with either chemiluminescence (by film acquisition or LAS3000) or infrared detection (Odyssey System). Quantifications were performed with the software indicated in the figure legends.

Imaging

For immunofluorescence, cells seeded on coverslips were washed in PBS, fixed (4% formaldehyde) and permeabilized in PBS with 0.1% Triton X-100 and 0.05% sodium deoxycholate, before staining with primary antibodies and secondary Alexa Fluor antibodies (Invitrogen, CA, USA) in blocking solution (5% goat serum). Mitochondrial morphology was determined as described elsewhere ⁷⁵ and examined using a Nikon E600 microscope. Optical sections were acquired using a Leica TCS SP5 confocal system. Colocalization analyses were performed using LAS AF software (Leica Microsystems, Germany), Imaris colocalization module (Bitplane AG, Zurich, Switzerland) or ImageJ (National Institutes of Health, USA). Average mitochondrial area and length were quantified using the LAS AF software (Leica Microsystems, Germany). Scale bars of 10 µm and Z-stacks of 0.5 µm are shown (unless specified differently in the figure legends). For live imaging, cells were seeded on coverslips and transiently transfected as described above. For short-term live cell imaging, an UltraView ERS spinning disk confocal microscope (Perkin Elmer, USA) equipped with a 37°C incubation chamber with 5% CO2 was used. Z-stacks of the images were collected using a 100X NA 1.4 oil objective with a helium-neon laser at 543 nm. The Z-stacks were acquired continuously over 10 min in the spinning disk with 300 millisecond exposures. The images from the time-lapse imaging were processed using Volocity 3D image analysis software (Perkin Elmer, USA) and mounted as .MPEG4 files. For the electron microscopy, MEFs were plated in two 10 cm Ø plates for each sample and sent to the Electron Microscopy Platform (Scientific and Technological Centers, University of Barcelona, Spain) for sample preparation. Ultrathin sections (55 nm) were cut and mounted with 200 mesh cupper grid with supported film. Image acquisition was performed with a transmission electron microscope (JEOL-1010) coupled to Bioscan software (Gatan, UK).

Isolation of mitochondria

The mitochondrial isolation kit MITOISO2 (Sigma Aldrich) was utilized according to the manufacturer’s instructions. The final mitochondrial pellet was layered onto a Percoll density gradient (Sigma Aldrich), and resuspended as indicated by the manufacturer.

Crude mitochondrial isolation was performed as described elsewhere ⁷⁶. For trypsin/triton digestion of crude mitochondria, 200–800 µg/ml of trypsin was added to tube T (200/+, 400/++ and 800/+++) and to tube TT, 200 µg/ml of trypsin with 2% of triton X-100 was added. All the samples were incubated at 37°C, and after 15 min, 10% FBS was added to halt digestion. The samples were centrifuged at 13,000×g for 2 min and washed twice with incubation buffer before adding the SDS-loading buffer and analyzing by Western blot. Crude mitochondria fraction from mouse heart was performed with ProteoExtract Cytosol/Mitochondria Kit (Calbiochem), following the manufacturer’s instructions. Subfractions of mouse liver mitochondria were generated as described elsewhere ⁴¹.

Cellular treatments

To label mitochondria, cells were incubated with 300 nM MitoTracker^® CMXRos (Invitrogen) or were transfected with pcDNA3-mt-DsRed or pcDNA3-mt-GFP. The incubation of WT and Gα_q/11^−/− MEFs with carbonyl cyanide m-chlorophenylhydrazone (CCCP) (Sigma Aldrich) (10 µM) took place in supplemented DMEM at 37°C for 6 and 3 h, respectively.

Down-regulation of murine Gα_q/11 proteins

shRNA-mediated knockdown of Gα_q/11 was performed using specific Mission shRNA and nontargeting Mission shRNA negative control (Sigma Aldrich). Lentivirus particle production was developed following the manufacturer’s instructions in HEK293T cells and down-regulation was monitored by immunoblot analysis of cell lysates generated after 3 days of puromycin selection of infected cells.

Mass spectrometry analysis

Following mitochondrial Percoll gradient fractionation of NIH3T3 cells and SDS-PAGE, the gel was stained with Colloidal Coomassie Blue G250 (Sigma Aldrich) and the bands corresponding to 35 – 50 kDa were cut and sent to the PCB Proteomics Platform to proceed with mass spectrometry analysis. The related protocols can be found at http://www.pcb.ub.edu/homePCB/live/en/p1249.asp.

Mitochondrial fusion analysis

MEFs were transfected with mt-DsRed and mito-PAGFP. A cell was photobleached and photoactivated and then time lapse series of image stacks composed of 4 images (512×512) were taken every 4 s for 15 min. Intensity correlation analysis was performed against red (photobleached mt-DsRed) and green (photoactivated mito-PAGFP) using Volocity. The threshold was automatically performed ⁷⁷. The rates of fusion were analyzed using the overlap coefficient K2 (red dots containing green) and data were normalized according to the intensity of the first time point after photobleaching.

Growth Rates

Cells (5×10⁴) were plated in 24-well plates in 1 ml of the indicated medium and incubated at 37°C for up to 3 days. They were counted daily using a Neubauer chamber. The culture media used were: DMEM with 5 mM of either glucose or galactose, supplemented with 10% FBS.

Cellular ATP, oxygen consumption and membrane potential

The ATP Determination Kit (Invitrogen) was used following the manufacturer’s instructions. The oxygen consumption was determined in 4×10⁴ intact MEFs using a Seahorse Bioscience XF96 extracellular flux analyzer following the manufacturer’s instructions and using the materials provided. Protocol: 12 min of equilibration, followed by 3 measurements of 3 min, separated by mixing for 4 min. Uncoupled mitochondrial respiration was induced by injection of 1 µM CCCP. To stop the mitochondrial-dependent oxygen consumption, we utilized 1 µM Oligomycin.

To calculate the membrane potential, MEFs (5×10⁶) were treated with 100 nM of the fluorescent dye TMR for 30 min at 37°C, washed with PBS and resuspended with 400 µl of trypsin. To halt trypsin digestion, 1 ml of PBS containing 5% BSA was added. The cells were passed through a flow cytometer, MoFlo (Beckman Coulter) at 590 nm.

Isolation of respiratory complexes and supercomplexes – BN-PAGE

Mitochondria were isolated from cultured cell lines as described elsewhere ⁷⁸, with slight modifications ⁷⁶. Digitonin-solubilized mitochondrial proteins (50 µg) were separated on blue native gradient gels (3%-13% acrylamide).

Statistical analysis

Average mitochondrial surface area: Comparison between WT and knockout cells presented heterogeneous variance, so the Mann-Whitney non-parametric test was utilized. Comparisons between values found for shRNA presented normal distributions, which allowed the use of variance analysis, followed by Student’s t-test. Mitochondrial length: Values from WT and knockout cells presented heterogeneous variance, so the Mann-Whitney non-parametric test was performed; a heterogeneous variance distribution was also found in the shRNA groups, for these values the Kruskal-Wallis test was utilized. Mitochondrial morphology: The values were analyzed by Chi-squared test, only when fewer than five values were found was the G-test employed, followed by Fisher’s test. Mitochondrial fusion analysis by mt-PAGFP was performed using ANOVA. Mitochondrial membrane potential and O₂ consumption were analyzed by unpaired t-test; whereas total ATP content and Odyssey quantification of Supercomplex I+III+IV/Porin was by paired t-test. Growth rate differences were determined by one-way ANOVA followed by Tukey’s Test (p-values are given in the figure legends).

Results

Gα_q/11 proteins localize at mitochondria

In order to search for signaling proteins located at the mitochondria, a proteomic analysis of a mitochondrial fraction, obtained from a Percoll gradient, was performed on NIH3T3 cells. Only the gel bands around the overall molecular weight of G proteins (35–56 kDa) were chosen for sample analysis. Of 56 proteins (Table S1), 49 were either mitochondrial proteins included in MitoCarta ³⁴, or cited as putative mitochondrial-associated proteins; this validated our approach. Among them, both Gα_q and Gα₁₁ were recognized, as were the Gα_i2–3 and Gα_o1 subunits. Gα_i proteins had previously been reported to be located at the mitochondria ³. MitoProt analysis of the G proteins gave the highest scores for Gα_q and Gα₁₁ proteins: 30% and 37%, respectively (Fig. S1). Western blot analysis of the Percoll-gradient mitochondrial fraction agreed with the proteomic analysis (Fig. 1A). A band recognized by the common anti-Gα_q and anti-Gα₁₁ antibody was present (13%) in the fraction, as was the mitochondrial protein porin (38%). In order to obtain the purest mitochondria sample, only a fraction of total porin was recovered. A crucial finding was that no plasma membrane (P-Cadherin), ER (SERCA2) or golgi contamination was present. The fraction did, however, contain lysosomes (20%) and caveolin-1 (4%), a protein that binds to Gα_q ³⁵ and is also present in mitochondria ³⁶.

Fig. 1. Gα_q and Gα₁₁ are localized at the mitochondria.

(A) Distribution of endogenous Gα_q/11 and organelle markers in the mitochondrial (MT) and supernatant (S) fractions obtained by Percoll gradient of NIH3T3 cells. The MT fraction was resuspended in 1/5 of initial volume, and equal volumes of MT and S fractions were loaded in the gel and immunoblotted with: anti-Gα_q/11, Porin (mitochondria), SERCA2 (ER), Golgi, LAMP1 (lysosomes), Rab11 (ribosomes), P-cadherin (PM) and Caveolin-1 as a protein that binds to Gα_q ³⁵ and is present at the mitochondria ³⁶. Quantification was performed by Multi-Gauge. (B) Gα_q heart-specific transgenic mouse (Gα_q^+/+) show increased amount of Gα_q at the mitochondria. The mitochondrial (MT) and supernatant fraction (S) were immunoblotted with Gα_q/11 antibody. Quantification was performed by Alpha Ease FC. (C) Confocal micrographs of NIH3T3 cells (endogenous) immunostained with anti-Gα_q/11 or expressing Gα_q-GFP. Colocalization was performed by LAS AF. (D) Confocal micrographs of NIH3T3 cells incubated with mitotracker (red) and transfected with Gα_q-N-terminus (1–124 aa) GFP and Flag, immunostained with anti-Flag and mounted with DAPI. MEF cells transiently expressing Gα_q IE25/26AA mutants incubated with mitotracker (red), immunostained with anti-Gα_q/11 and mounted with DAPI. Line profile was generated by LAS AF.

The subcellular localization of Gα_q was analyzed in a transgenic mouse line containing 40 copies of the Gα_q gene expressed in the myocardium. Mitochondria fractionation showed that approximately 10% of total Gα_q is present in the mitochondrial fraction of the wild-type (WT) heart (Fig. 1B). Interestingly, the transgenic mice showed increased levels of Gα_q (55% of total) in the mitochondrial fraction (Fig. 1B). Again, these data indicate the presence of Gα_q at the mitochondria in heart tissue of wild type animals and increase amounts in the transgenic mouse line. Interestingly, these mice present dilated cardiomyopathy ³⁷ which has been reported to be associated with increased mitochondrial ROS production ³⁸.

On the other hand, immunofluorescence analysis of the endogenous proteins in NIH3T3 cells with anti-Gα_q/11 antibodies (Fig. 1C) revealed a punctuated cytoplasmic pattern coincident with the mitochondria. Expression of a functional Gα_q-GFP also showed considerable localization at the mitochondria in NIH3T3 (Fig. 1C). Taken together, these findings support the hypothesis that Gα_q proteins are located at the mitochondria.

The N-terminal region of Gα_q is necessary for mitochondrial targeting

Considering the possibility that Gα_q and/or Gα₁₁ were targeted to mitochondrial membranes, we searched for putative targeting sequences. A mitochondrial target prediction by Mitoprot analysis of different G alpha subunit sequences showed that mouse Gα_q and Gα₁₁ had 30 and 37% probability of being target to mitochondria, respectively, whereas Gα₁₂, Gα_i1–2 or Gα_o had lower probabilities (Fig. S1). Chimeric proteins were designed that contained the first 124 N-terminus amino acids or the C-terminus sequence of Gα_q fused to either GFP or Flag. The C-terminus sequence fused to GFP gave almost no detectable expression around any part of the cell. The Gα_q-protein N-terminus sequence flagged with either Flag or GFP showed mitochondrial localization (Fig. 1D), which suggests that the N-terminus of Gα_q is sufficient for the protein to located at the the mitochondria. The Gα_q N-terminus region contains both the S-palmitoyl cysteines (9–10 aa) required for plasma membrane binding, and also the contact sites for Gβγ interaction. Mutations of amino acids 25 and 26 (IE>AA) are reported to alter Gβγ binding and also to prevent correct palmitoylation of the Gα subunit ³⁹. The N-terminus-Gα_q-IE25/26AA mutant (Fig. 1D) failed to localize at the mitochondria, suggesting that the interaction with Gβγ and/or the state of palmitoylation of the protein is important for mitochondrial targeting. The fact that the mutated peptide was not present at the mitochondria rules out possible artifactual localization of the chimeric protein at the mitochondria.

The Gα_qβγ heterotrimer is localized at the outer membrane and the Ga_q subunit at the inner membrane

The previous results suggest that the Gβγ dimer could help to target the heterotrimer at the mitochondria. The expression of Gβ4γ2 alone or together with the Gα_q-GFP protein shows localization of both HA-tagged Gγ2 and Gα_q with mt-DsRed (Fig. 2A and Movies S1 and S2). Similar results were obtained by expressing Gβ1γ2 together with Gα_q-GFP (Fig. S2A). The Gβγ subunits are known to localize at the ER of cells ¹. To establish whether Gβγ localization corresponded to the ER, we compared the localization of Gβ4γ2 in HeLa cells expressing mt-DsRed and ER-DsRed. Significantly more Gβ4γ2 was observed at the mitochondria than at the ER (Fig. S2B). These results confirm the localization of the heterotrimer at the mitochondria.

Fig. 2. Gα_q and different Gβγ dimers localized at the mitochondria.

(A) Confocal micrographs of HeLa cells stably expressing mt-DsRed and Gβ4-Flag γ2-HA and/or Gα_q, immunostained with anti-Flag or HA and mounted with DAPI. (B) Mitochondrial fractions of NIH3T3 cells submitted to trypsin digestion in presence or absence of Triton X-100, immunoblotted with anti-Gα_q/11, OPA1 (inner membrane), Mfn2 (outer membrane), Porin (integral outer membrane) and Gβ-Pan (representing Gβγ dimer). (C) Diagram showing the likely actions of trypsin on proteins blotted in B. (D) Mouse liver mitochondrial sub-fractions shown the presence of Gα_q/11 at the inner membrane, immunoblotted with Gα_q/11 and markers: Porin (OM), Smac/Diablo (IMS), COXI (IM) and Hsp70 (M).

To examine the location of the G proteins within the mitochondrial subcompartments, we carried out trypsin digestion experiments on isolated mitochondria. Incubation of mitochondria with trypsin leads to the complete digestion of the outer membrane protein Mfn2 as well as that of the Gβ proteins (Fig. 2B), which is indicative of their outer membrane localization. The Gα_q/11 proteins were partially digested by trypsin, indicating that some protein was located together with Gβγ at the outer membrane facing the cytoplasm. However, a considerable amount of Gα_q/11 was protected from digestion, as also observed for the inner membrane protein OPA1 (see Fig. 2C for a diagram representing trypsin digestion). Sub-fractionation of mitochondrial membrane by established procedures ^{40, 41} yields fractions enriched in outer membrane (OM), intermembrane space (IMS), inner membrane (IM) and matrix (M). In adult mouse liver (Fig. 2D), we observed the presence of Gα_q/11 in the total mitochondrial fraction (TO) and associated with OM, and a relative enrichment in the IM fraction (also COXI enriched); this confirms the trypsin experimental results.

Taken together these results corroborate the presence of Gα_q/11 proteins in the mitochondria of different cells and tissues, and also suggest that those proteins are localized together with Gβγ at the outer membrane. A proportion of the Gα_q/11 subunits are protected from digestion, suggesting that they are localized inside the mitochondria.

Gα_q/11 deletion results in defects in mitochondrial morphology

As a first approach to determine whether Gα_q/11 plays a direct functional role when localized at the mitochondria, we compared mitochondria from Gα_q/11^−/− knockout (KO) and wild type murine embryonic fibroblasts (MEFs). Confocal microscopy of cells expressing the mitochondrial target peptide mt-DsRed showed marked differences in mitochondrial distribution and shape (Fig. 3A). Notably, mitochondria from Gα_q/11^−/− cells were compacted around the nucleus. These differences were also observed via in vivo time-lapse images of WT (Movie S3) and Gα_q/11^−/− (Movie S4) MEFs. Quantification of average mitochondrial surface area (Fig. 3B) showed that in Gα_q/11^−/− cells, the mitochondrial network is less distributed throughout the cell. Quantification of the mitochondrial length showed that the Gα_q/11^−/− mitochondria were also more fragmented than in WT cells (Fig. 3C). These results were corroborated using WT MEFs treated with shRNA against Gα_q (Fig. S3A). The Gα_q-down-regulated cells showed an increased compaction of their mitochondrial network and more fragmented mitochondria (Fig. 3D–F) relative to the scrambled shRNA. When Gα_q and Gα_q-GFP (Fig. S3B) were re-expressed in the Gα_q/11^−/− MEF cells, both proteins restored the normal mitochondrial morphology and resulted in a fused mitochondrial network. The Gα_q N-terminus (the first 124 aas) did not achieve this (Fig. S3B), demonstrating that the whole protein is required for recovery of the mitochondrial phenotype.

Fig. 3. MEF Gα_q/11 knockout (Gα_q/11^−/−) and shRNA depleted cells show alterations of the mitochondrial network and morphology.

(A) Representative confocal micrographs of MEF wild-type (WT) and Gα_q/11^−/− cells expressing the mitochondrial matrix-targeted mt-DsRed after 24 h of transfection. (B) Mitochondrial surface average area calculated by the Polygon ROI with LAS AF software. The mean value of different intensities inside the ROI is related to the distribution of the fluorochrome. Data represent mean ± s.d. (n=45). Mann-Whitney test was employed (*p<0.0001). Experiments were carried out as in A. (C) Mitochondria length was calculated by the polyline measurement. Data represent mean ± s.d. (n=25) for 10 ROIs each. Mann-Whitney test was utilized (*p<0.0001). Experiments were carried out as in A. (D) Representative confocal micrographs of MEF wild-type cells infected by a lentivirus containing two different sequences of shRNA (1 and 3) against Gα_q/11 and one control shRNA (shC1) expressing mt-DsRed protein after 24 h of transfection. (E) Experiments were carried out as in D and calculated as in B. Data represent mean ± s.d. (n=25). Student’s t-test was employed (*p=0.0063 and **p<0.001). (F) Experiments were carried out as in D and calculated as in C. Data represent mean ± s.d. (n=25). Kruskal-Wallis test was utilized (*p<0.0001). (G) TEM micrographs showing the mitochondrial ultrastructure of MEFs WT (1–2), Gα_q/11^−/− (4–9) and Gα_12/13^−/− (3) cells. Scale bars: 0.5 µm and 0.1 µm in 2. See also Figure S3C for more micrographs of Gα_12/13^−/− cells.

To corroborate the immunofluorescence results, the ultrastructure of the Gα_q/11^−/− mitochondria was examined by transmission electron microscopy (TEM). The TEM images revealed mitochondrial abnormalities (Fig. 3G, 4–9) compared with WT (Figure 3G1 and 3G2) or Gα_12/13^−/− mitochondria (Fig. 3G3 and S3C). The Gα_q/11^−/− mitochondria showed localized swelling accompanied by a constriction along the length of the mitochondria (Fig. 3G4 and 3G5). Almost no elongated mitochondria were found, thereby corroborating the immunofluorescence results. Remarkably, some mitochondria seem to be devoid of a cristae structure (Fig. 3G5–7) and the openings of cristae junctions appeared very narrow (Fig. 3G8). We also observed an increased number of autophagosomes containing mitochondrial remnants (Fig. 3G9), suggesting a greater turnover of damaged mitochondria through the process called mitophagy. Overall, these results suggest that Gα_q/11 proteins play an essential role in mitochondrial morphology

Gα_q-GDP state is needed to coordinate elongation of the mitochondrial network

The aforementioned changes in mitochondrial morphology associated with alterations in Gα_q expression could be the result of alterations in the processes of fission or fusion. To determine whether Gα_q/11 are involved in mitochondrial fusion, we transfected a photoactivated mito-PAGFP into WT (Movie S5) and Gα_q/11^−/− (Movie S6) MEFs. A constant increase in mitochondrial fusion events over time was detected in both cell types, but the rate of mitochondrial fusion in the Gα_q/11^−/− cells was significantly lower than that in WT MEFs (Fig. 4A). These results suggest that a lack of Gα_q/11 affects mitochondria fusion events over time.

Fig. 4. Impairment in mitochondria fusion in absence of Gα_q/11.

(A) MEFs transfected with mt-DsRed and mito-PAGFP. Mito-PAGFP was photoactivated, mt-DsRed was photobleached at t=0 s. Panels show the same cell at a range of time points. Scale bar: 75 µm. Data show mean ± s.e.m (n=5). ANOVA (*p<0.05, **p<0.01 and ***p<0.001) was employed. (B) Confocal micrographs of MEFs transfected with mt-GFP or Gα_q/mt-DsRED, immunostained with anti-Gα_q/11 antibody (right panel in green). Zoom shows only mitochondria. Scale bar: 25 µm. (C) Mitochondrial morphology was scored from B. Data represent mean ± s.d. (n=50) of three independent experiments. Chi-Square test was employed (*p<0.0001). (D) Confocal micrographs of MEFs in presence of Gβ2-Flag (green on left and purple on right panel)/mt-dsRED with or without Gα_q-GFP (green on right panel) mounted with DAPI (blue). Zoom shows only mitochondria.

To study the effect of Gα_q on fusion events further, we expressed Gα_q in Mfn2^−/− and Mfn1^−/− depleted MEFs. It is well documented that Mfn1 and Mfn2 are involved in mitochondrial fusion and, in their absence, mitochondria present a fragmented phenotype (Fig. 4B) ⁴². Gα_q expression in these cells induced a significant increase in the elongation of the Mfn2^−/− mitochondria (Fig. 4C), which indicates that Gα_q expression promotes fusion. As expected, Gα_q was unable to induce elongation in Mfn1^−/− cells, since the protein is essential for mitochondrial fusion, which corroborates previous findings ⁴². The activated form of Gα_q, Gα_q-R183C, did not produce the same effect (Fig. S4A). These results suggest that a GDP state of Gα_q is needed for this function.

In contrast, expression of Gβ2γ2 in Mfn2^−/− cells promoted mitochondrial bundle-like aggregation and perinuclear clustering (Fig. 4D) similar to those seen following mitofusin overexpression ⁴³. However, when Gα_q was expressed alone or together with the heterodimer, a hyperfused-mitochondrial network was observed in both WT and Mfn2^−/− cells (Fig. 4D). The fact that Gβγ did not phenocopy the effect of Gα_q suggests that these subunits play complementary roles at the mitochondria; this provides further evidence that Gα_q/11 regulates fusion and/or fission events at the mitochondria.

Gα_q stabilizes mitochondrial fusion, blocking fragmentation induced by Drp1 expression

Expression in cells of the dynamin-like protein, Drp1, induces strong mitochondria fragmentation (fission) in contrast to the expression of its mutant form Drp1K38A (Fig. 5A and S5). We used the fragmentation capacity of Drp1 as another approach to study the effect of Gα_q on mitochondrial fission. The fragmentation induced by Drp1 was significantly diminished upon Gα_q expression in WT cells (Fig. 5B and 5C). Mitochondria elongation was also augmented in Gα_q/11^−/− cells expressing Drp1K38A, which recovered the fragmented phenotype (Fig. S5). Gα_q acts as a potent inhibitor of mitochondrial fission, with its action depending on Drp1 at the mitochondria.

Fig. 5. Gα_q stabilizes mitochondrial fusion, blocking fragmentation induced by Drp1 or CCCP.

(A) Confocal micrographs of MEF wild-type cells transfected with mt-DsRed (grey) and Drp1-HA or Drp1-(K38A)-HA or treated with 10 µM CCCP (+) or DMSO (−) for 3h, overexpressing (+) or not (−) Gα_q, immunostained with anti-HA and anti-Gα_q/11 (not shown). (B-C) Mitochondrial morphology quantified as mentioned in Figure 4C. Chi-Square test was employed (*p<0.0001). Experiments were carried out as in A. Data represent mean ± s.d. (n=50) of three independent experiments. (D) MEF cells were transfected with pcDNA3 or pcDNA3-Gα_q and the day after incubated for 3h with 10 µM CCCP or DMSO. Lysates were immunoblotted with the indicated antibodies. (E) MEF WT and Gα_q/11^−/− cells immunoprecipitated for OPA1 isoforms and immunoblotted with OPA1 antibody, quantified by Line Profile of Odyssey System.

MEFs were treated with the uncoupling agent CCCP, which reduces mitochondrial fusion through dissipation of the mitochondrial membrane potential (ΔΨm) and OPA1 degradation. The fragmentation of mitochondria observed in the presence of CCCP (10 µM) decreased significantly in the presence of Gα_q (Fig. 5A, C and S5). Analysis of the OPA1 isoforms shows that CCCP could induce the degradation of OPA1, decreasing the higher isoforms (bands a and b) and increasing bands d and e (Fig. 5D), as expected. Consistently with this, cells expressing Gα_q present less decrease in bands b and e upon CCCP treatment. A change in the proportion of OPA1 bands was also observed in the untreated Gα_q and Gα₁₁ knock-out cells (Fig. 5E), which presented higher levels of bands b and e. We detected no major changes in the expression of the fusion and fission proteins (Fig. S5B) that could explain these effects. The results suggest that Gα_q and Gα₁₁ affect the proteolytic cleavage of OPA1 protein via either a direct or indirect effect, thus impinging on the mitochondria fission and cristae structure.

A lack of Gα_q/11 proteins leads to significant decreases in ΔΨm, overall respiratory capacity, ATP production and OXPHOS dependent growth

The results suggest that Gα_q and Gα₁₁ proteins are necessary to maintain a proper balance of fusion and fission, and also for maintaining crest morphology. At the inner mitochondrial membrane, a proton gradient that drives ATP production is created by the respiratory complexes passing electrons through the electron transport chain and giving rise to the ΔΨm. To establish whether Gα_q/11 proteins are necessary for normal mitochondrial bioenergetics, we first measured the intensity of tetramethyl rhodamine (TMR) fluorescence. We determined that the ΔΨm of Gα_q/11^−/− cells was significantly lower than that of WT cells (Fig. 6A). A recovery of the Gα_q/11^−/− cells through expression of Gα_q prevented this decrease in ΔΨm. In addition, we determined the endogenous ATP levels and the O₂ consumption rates (OCRs) in WT and Gα_q/11^−/− cells (Fig. 6B and 6C). Cells lacking Gα_q/11 have 15% less cellular ATP (Fig. 6B) and lower OCRs both under baseline conditions and when maximum respiratory capacity is activated by mitochondrial depolarization with CCCP (Fig. 6C).

Fig. 6. Gα_q/11 are needed for normal mitochondrial function and respiratory supercomplexes assembly.

(A) MEF Gα_q/11^−/− cells show decreased mitochondrial membrane potential (ΔΨm). + Gα_q indicates that cells were transfected with pcDNA3-Gα_q the day before analysis. Data represent mean ± s.d. (n=4 WT and n=6 KO). Unpaired t-test was employed (*p=0.0023). (B) Decreased content of cellular ATP is observed in Gα_q/11^−/− cells. Data represent mean ± s.d. (n=4) with normalized values respect to WT (%). Paired t-test (*p=0.0014) was utilized. (C) Gα_q/11^−/− cells present less oxygen consumption (OCRs) rate. After baseline measurements (normal mito) cells were depolarized (CCCP). Data represent mean ± s.e.m (n=18 WT and n=20 KO). Unpaired t-test (*p<0.0001) was employed. (D-H) Mitochondria from MEFs were lysed in digitonin and resolved by BN-PAGE then blotted to determine native complexes and supercomplexes: (D) Complex V (β-ATPase); (E) Complex I (detected by α-NDUFA9); (F) Complex III (α-Core1); (G) complex IV (NDUFA4) and (H) complex II (α-Fp). (I) Ratio of supercomplexes I+III+IV per Porin, quantified by Odyssey System utilizing experiments from E and F. Data represent mean ± s.d. (n=5 CI and n=3 CIII). Paired t-test (**p<0.0001, * p<0.01) was utilized. * denotes changes in supercomplex band intensity. (J) Cells growth in glucose (Glu) versus galactose (Gal) per days and doubling time (DT) ratio. Statistical significance was found by one-way ANOVA followed by Tukey´s test after day 2 in both conditions for KO compared to WT cells (p<0.05). Data represent mean ± s.e.m (n=6).

To gain a fuller understanding of the underlying nature of the mitochondrial electron transport chain alterations, we determined how the absence of Gα_q/11 alters the assembly of the respiratory complexes using BN-PAGE followed by immunoblotting for subunits of the respiratory chain (Fig. 6D–H). Gα_q/11^−/− cells had reduced amounts of dimer complex V (Fig. 6D). Moreover, significantly less of the supercomplexes I+III+IV (Fig. 6I) was observed. The critical reduction in larger complexes could be related to the narrow and disrupted cristae observed by TEM (Fig. 3G). Therefore, the absence of Gα_q/11 results in assembly disorganization of the respiratory supercomplexes which alters organelle energy production.

To determine whether the observed alterations in mitochondrial OXPHOS promoted by a lack of Gα_q/11 were functionally relevant, we studied the growth of the cells in a medium in which glucose was substituted by galactose; conditions that render the cells highly dependent on ATP produced by OXPHOS. Thus, the doubling time (DT) of both WT and Gα_q/11^−/− cells increases in galactose compared to that in glucose as a consequence of the adaptation to a more OXPHOS demanding medium (Fig. 6J). The increase in DT is higher in KO cells than in controls (1.44 vs. 1.25), which shows that Gα_q/11^−/− cells have more difficulties adapting to galactose. Altogether, these results demonstrate that Gα_q/11 cells are required not only for proper mitochondrial dynamics but also for OXPHOS function.

Discussion

Our studies provide insight into the mitochondrial location and functions of the G proteins, focusing on the Gα_q family of proteins, which are distinct from their capacity to signal through GPCRs at the plasma membrane. The results establish a previously unknown link between mitochondrial fission and fusion, energy metabolism and the Gα_q class of G proteins. The heterotrimeric G protein (Gα_q/11βγ) in its basal state (GDP form) is found at the outer mitochondrial membrane; whereas the alpha subunit (Gα_q/11) is alone inside the organelle. At the outer mitochondrial membrane, Gα_qβγ induces mitochondrial elongation dependent on the activity of mitofusin-1 and also reduces Drp1 induction of fission. The absence of Gα_q/11 affects OPA-1 isoforms, cristae structure, membrane potential and organelle bioenergetics, accompanied by a reduction in respiratory supercomplex assembly. The presence of Gα_q/11 proteins at the mitochondria serves as a new non-canonical pathway that stabilizes elongated mitochondria and regulates energy production, thereby linking organelle dynamics and physiology.

The role and location of G proteins at the mitochondria has not previously been explored in detail, despite the presence of these proteins in other endomembranes ^44–49. Nevertheless, there are recent reports of Gα_i, Gα₁₂ and Gβ₂ localization at mitochondria ^2–4. Adding to what has already been reported, Gα_o1, Gα₁₁, Gα_i2–3, Gβ₁, Gβ₄ and Gγ2 were found to be localized at the mitochondria in the work reported here. Thus, taken together, this amounts to strong support for the mitochondrial localization of heterotrimeric G proteins. Although the mitochondrial targeting signal of the Gα subunits is located in the N-terminus (124 aa) and this is sufficient to confer mitochondrial location, we show that the binding to Gβγ ⁵⁰ provide further support for this targeting. Taking into account that WD-propeller proteins cannot cross through the TOM complex ⁵¹ and that we and others ⁴ found Gβγ located at the outer membrane, it seems that only the Gα subunits can cross this membrane. At least for Gα_q, its dual location may be necessary for the coordination of the fusion of both the outer and the inner mitochondrial membrane. Both Gα₁₂ ² and Gα_q bind to the chaperon protein Hsp90 (unpublished results) and it is known that Gα₁₂ requires Hsp90 for mitochondrial targeting ². It is most likely that chaperone proteins are involved in the unfolding process necessary for crossing through the TOM complex ⁵². So far we have not detected any proteolytic cleavage of Gα_q. So, the mechanism of entry of the Gα subunits is still unclear and will require further research.

The mechanism of activation and the effectors downstream from the mitochondrial-located G proteins are still unknown. However, recent reports localized two different GPCRs at this organelle: the CB₁ receptor, found at the outer mitochondria membrane of neurons ⁵³; and a functional angiotensin system at the inner membrane ⁵⁴. Those authors demonstrate that Ang II type 1 and 2 receptors are present in this subcompartment of the mitochondria and that the activation of the mitochondrial angiotensin system is coupled to nitric oxide production which alters the respiratory capacity. These findings raise the possibility that GPCRs located at the mitochondria could couple to the G proteins also located at this organelle. The presence at the mitochondria of G protein-effectors or binding partners such as MAPKs, Akt, GRK2 and PKC ^{5, 6} also supports the signaling of G proteins at mitochondria. However, the fact that the Gβ₂ subunit ⁴ binds directly to an intrinsic mitochondrial protein such as mitofusin 1, raises the possibility that mitochondrial proteins may act as receptor/effectors for a a new non-canonical G protein effect.

Gα_q/11-depleted cells presented mitochondrial fragmentation and alterations in cristae; conversely, increases in Gα_q levels elongate mitochondria. These alterations are reminiscent of those found in cells depleted in Mfns1–2 ⁴² and OPA1 ⁵⁵. Meanwhile, expression of Mfns1–2 or Gβγ leads to the formation of clusters of mitochondria due to outer membrane fusion ^{4, 56}, but these clusters are not present when Gα_q is co-expressed. We propose a mechanism in which Gα_q and Gβγ are necessary for mitochondria fusion via the coordinated regulation of the Mfn1 protein. However, mitochondria elongation can also be the result of inhibition of fission. It is particularly interesting that Gα_q expression reduces Drp1 function. The Drp1 protein is recruited from the cytoplasm to the outer mitochondrial membrane where it forms assemblies and tubule constrictions around the organelle and, consequently, produces fission. Drp1 is posttranslationally modified by a variety of enzymes (phosphorylation, sumoylation, ubiquitination, S-nitrosylation), which highlights its complex regulation ⁵⁷. Our findings demonstrate that Gα_q regulate mitochondrial morphology and respiratory efficiency through a Mfn1 and Drp1-dependent mechanisms.

Mitochondria fusion involves the coordination of both the outer and inner membrane ^{33, 57}. Our results suggest that Gα_q/11 can also impinge on the inner membrane and crest morphology. Gα_q/11 effects could be explained by the altered processing of OPA1. Post-transcriptional regulation of OPA1 is rather complex, added to constitutive processing by different proteases (i-AAA, m-AAA and Parl) ⁵⁸, which generate long and short OPA1 isoforms, all of which are required for mitochondrial fusion. L- and S-forms of OPA1 (a-b and c-d-e, respectively) form oligomers and keep cristae junctions tightly closed, thereby preventing cytosolic release of cytochrome c ⁵⁹. Under conditions of mitochondrial membrane depolarization (or CCCP treatment), ATP deficiency or apoptosis, the L-OPA1 isoforms undergo inducible cleavage by OMA1 which generates S-OPA1 forms (d and e), resulting in mitochondria fragmentation. It is interesting to note that Gα_q/11-depleted cells that have reduced ATP due to lower membrane potential and fragmentation, show an increase in L-forms (band b) and also in S-forms (band e). The increase in band e could be related to the lower membrane potential and is in agreement with the fact that mitochondria are fragmented. The higher band b, however, can only be a consequence of Gα_q/11 altering OPA1 processing. We consider that the mechanism of action of Gα_q/11 does not directly inhibit OMA1 function, since Gα_q/11-depleted cells present an increase in band e. Therefore, we propose that Gα_q/11 affects the processing of OPA1 isoforms by altering membrane potential and inner membrane fusion processes.

Our data also indicate that Gα_q/11 are required to maintain the proper organization of the respiratory complexes. Elongated mitochondria have higher levels of the dimeric form of ATPase, associated with increased efficiency in ATP production ⁶⁰. The absence of Gα_q/11 reduces not only this dimeric form (CV) but it also provokes an important reduction in supercomplexes containing complex I ^{61, 62}. A primary consequence of the alteration in mitochondrial elongation and cristae organization by the absence of Gα_q/11, was a decrease in energy transfer. These results confirm previous findings linking mitochondrial morphology and energy fluxes ⁶³. OPA1 regulates cristae shape, organization and dynamics ^{59, 64} and has been shown to interact directly with OXPHOS complexes I, II and III but not IV ⁶⁵. More recently, the lack of OPA1 or its processing by specific proteases that induces cristae remodeling was shown to consequently reduce the amount of both respiratory supercomplexes containing complex I and complex I-dependent respiration ⁶⁶.

Under normal conditions, mitochondrial fusion and fission take place at a balanced rate and thus a relatively constant tubular morphology is maintained. However, perturbation of the fission/fusion balance has been found to be associated with numerous human diseases ^{33, 67}. Our findings raise the possibility that the non-canonical mitochondrial-function of Gα_q may also account for some of the known functions of Gα_q otherwise attributed to other pathways. Particularly interesting is the fact that Gα_q signaling is essential for cardiomyocyte hypertrophy and proliferation during development ⁶⁸. However, it is also linked to hypertrophy of the adult myocardium and subsequent heart failure ⁶⁹ where mitochondria need to sustain a high-energy demand ⁷⁰ to avoid cell death under pathological conditions ⁷¹. Therefore, understanding the functional role of Gα_q at the mitochondria may open up new approaches to therapeutic treatments for cardiac diseases as well as other diseases.

Accession	Coverage	# PSMs	# Peptides	# AAs	MW [kDa]	calc. pI	Score	Description
P05202	79.77	291	43	430	47.4	9.00	886.47	Aspartate aminotransferase, mitochondrial OS=Mus musculus GN=Got2 PE=1 SV=1 - [AATM_MOUSE]
Q9QYA2	78.95	60	20	361	37.9	7.74	203.62	Mitochondrial import receptor subunit TOM40 homolog OS=Mus musculus GN=Tomm40 PE=1 SV=3 - [TOM40_MOUSE]
Q99JB2	80.17	54	29	353	38.4	8.87	178.95	Stomatin-like protein 2 OS=Mus musculus GN=Stoml2 PE=1 SV=1 - [STML2_MOUSE]
P05064	83.52	58	30	364	39.3	8.09	170.75	Fructose-bisphosphate aldolase A OS=Mus musculus GN=Aldoa PE=1 SV=2 - [ALDOA_MOUSE]
P08752	57.75	46	26	355	40.4	5.45	140.70	Guanine nucleotide-binding protein G(i) subunit alpha-2 OS=Mus musculus GN=Gnai2 PE=1 SV=4 - [GNAI2_MOUSE]
Q9DC51	72.32	50	28	354	40.5	5.69	138.03	Guanine nucleotide-binding protein G(k) subunit alpha OS=Mus musculus GN=Gnai3 PE=1 SV=3 - [GNAI3_MOUSE]
Q9ER88	68.29	48	25	391	44.7	8.94	128.17	28S ribosomal protein S29, mitochondrial OS=Mus musculus GN=Dap3 PE=2 SV=1 - [RT29_MOUSE]
Q9D6R2	53.83	46	25	366	39.6	6.73	117.94	Isocitrate dehydrogenase [NAD] subunit alpha, mitochondrial OS=Mus musculus GN=Idh3a PE=1 SV=1 - [IDH3A_MOUSE]
Q9DBL1	49.77	34	21	432	47.8	7.87	115.01	Short/branched chain specific acyl-CoA dehydrogenase, mitochondrial OS=Mus musculus GN=Acadsb PE=1 SV=1 - [ACDSB_MOUSE]
Q07417	59.22	34	24	412	44.9	8.79	106.53	Short-chain specific acyl-CoA dehydrogenase, mitochondrial OS=Mus musculus GN=Acads PE=2 SV=1 - [ACADS_MOUSE]
Q9CXW2	63.23	37	23	359	41.2	8.56	91.61	28S ribosomal protein S22, mitochondrial OS=Mus musculus GN=Mrps22 PE=2 SV=1 - [RT22_MOUSE]
Q99K85	59.73	28	19	370	40.4	8.03	87.01	Phosphoserine aminotransferase OS=Mus musculus GN=Psat1 PE=1 SV=1 - [SERC_MOUSE]
Q99KV1	39.39	25	16	358	40.5	6.32	79.85	DnaJ homolog subfamily B member 11 OS=Mus musculus GN=Dnajb11 PE=1 SV=1 - [DJB11_MOUSE]
Q8BGC4	48.54	22	15	377	40.5	7.42	77.43	Zinc-binding alcohol dehydrogenase domain-containing protein 2 OS=Mus musculus GN=Zadh2 PE=2 SV=1 - [ZADH2_MOUSE]
O35855	44.53	24	16	393	44.1	8.29	72.26	Branched-chain-amino-acid aminotransferase, mitochondrial OS=Mus musculus GN=Bcat2 PE=2 SV=2 - [BCAT2_MOUSE]
Q8QZS1	55.32	21	17	385	43.0	8.06	71.64	3-hydroxyisobutyryl-CoA hydrolase, mitochondrial OS=Mus musculus GN=Hibch PE=1 SV=1 - [HIBCH_MOUSE]
Q920E5	43.63	23	15	353	40.6	5.66	69.44	Farnesyl pyrophosphate synthase OS=Mus musculus GN=Fdps PE=2 SV=1 - [FPPS_MOUSE]
Q921H8	45.99	24	15	424	43.9	8.44	68.46	3-ketoacyl-CoA thiolase A, peroxisomal OS=Mus musculus GN=Acaa1a PE=2 SV=1 - [THIKA_MOUSE]
Q91ZE0	46.08	23	15	421	49.6	8.25	67.29	Trimethyllysine dioxygenase, mitochondrial OS=Mus musculus GN=Tmlhe PE=2 SV=2 - [TMLH_MOUSE]
P18872	44.63	22	14	354	40.1	5.53	63.74	Guanine nucleotide-binding protein G(o) subunit alpha OS=Mus musculus GN=Gnao1 PE=1 SV=3 - [GNAO_MOUSE]
Q99N87	38.66	23	18	432	48.2	10.14	61.98	28S ribosomal protein S5, mitochondrial OS=Mus musculus GN=Mrps5 PE=2 SV=1 - [RT05_MOUSE]
Q99LC3	48.45	23	17	355	40.6	7.78	60.83	NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 10, mitochondrial OS=Mus musculus GN=Ndufa10 PE=1 SV=1
Q9WUR2	41.69	15	12	391	43.2	8.92	58.98	Peroxisomal 3,2-trans-enoyl-CoA isomerase OS=Mus musculus GN=Peci PE=1 SV=2 - [PECI_MOUSE]
P21278	38.16	21	15	359	42.0	5.97	57.17	Guanine nucleotide-binding protein subunit alpha-11 OS=Mus musculus GN=Gna11 PE=1 SV=1 - [GNA11_MOUSE]
Q9Z1G3	45.03	20	16	382	43.9	7.46	56.37	V-type proton ATPase subunit C 1 OS=Mus musculus GN=Atp6v1c1 PE=1 SV=4 - [VATC1_MOUSE]
Q8R3F5	36.75	18	12	381	41.9	8.10	54.81	Malonyl-CoA-acyl carrier protein transacylase, mitochondrial OS=Mus musculus GN=Mcat PE=2 SV=3 - [FABD_MOUSE]
P22315	49.29	20	18	420	47.1	8.91	53.50	Ferrochelatase, mitochondrial OS=Mus musculus GN=Fech PE=1 SV=2 - [HEMH_MOUSE]
Q8BGA9	27.25	19	11	433	48.2	9.61	50.72	Mitochondrial inner membrane protein OXA1L OS=Mus musculus GN=Oxa1l PE=2 SV=1 - [OXA1L_MOUSE]
P21279	31.20	17	13	359	42.1	5.68	48.91	Guanine nucleotide-binding protein G(q) subunit alpha OS=Mus musculus GN=Gnaq PE=1 SV=4 - [GNAQ_MOUSE]
Q8K2M0	40.53	20	15	380	45.0	8.10	44.52	39S ribosomal protein L38, mitochondrial OS=Mus musculus GN=Mrpl38 PE=2 SV=2 - [RM38_MOUSE]
Q8QZT1	37.03	16	13	424	44.8	8.51	43.10	Acetyl-CoA acetyltransferase, mitochondrial OS=Mus musculus GN=Acat1 PE=1 SV=1 - [THIL_MOUSE]
Q91V12	34.91	16	13	381	42.5	8.68	41.86	Cytosolic acyl coenzyme A thioester hydrolase OS=Mus musculus GN=Acot7 PE=1 SV=2 - [BACH_MOUSE]
P08249	40.24	12	10	338	35.6	8.68	41.13	Malate dehydrogenase, mitochondrial OS=Mus musculus GN=Mdh2 PE=1 SV=3 - [MDHM_MOUSE]
Q99M04	46.11	15	13	373	41.9	8.88	40.91	Lipoyl synthase, mitochondrial OS=Mus musculus GN=Lias PE=2 SV=1 - [LIAS_MOUSE]
P10605	35.99	12	11	339	37.3	5.91	39.06	Cathepsin B OS=Mus musculus GN=Ctsb PE=1 SV=2 - [CATB_MOUSE]
P56480	29.49	11	9	529	56.3	5.34	38.90	ATP synthase subunit beta, mitochondrial OS=Mus musculus GN=Atp5b PE=1 SV=2 - [ATPB_MOUSE]
Q9D8V0	36.24	14	10	378	41.7	6.04	38.79	Minor histocompatibility antigen H13 OS=Mus musculus GN=Hm13 PE=1 SV=1 - [HM13_MOUSE]
O89017	20.69	11	7	435	49.3	6.39	36.39	Legumain OS=Mus musculus GN=Lgmn PE=1 SV=1 - [LGMN_MOUSE]
Q8JZM0	44.06	13	12	345	38.9	9.47	35.31	Dimethyladenosine transferase 1, mitochondrial OS=Mus musculus GN=Tfb1m PE=2 SV=1 - [TFB1M_MOUSE]
O09174	41.21	12	11	381	41.7	7.40	34.86	Alpha-methylacyl-CoA racemase OS=Mus musculus GN=Amacr PE=1 SV=3 - [AMACR_MOUSE]
Q9CR16	27.03	11	9	370	40.7	7.43	34.80	Peptidyl-prolyl cis-trans isomerase D OS=Mus musculus GN=Ppid PE=1 SV=3 - [PPID_MOUSE]
Q9D7B6	28.57	14	12	413	45.0	8.13	34.16	Isobutyryl-CoA dehydrogenase, mitochondrial OS=Mus musculus GN=Acad8 PE=2 SV=2 - [ACAD8_MOUSE]
Q3URS9	34.24	11	9	406	45.1	8.09	33.59	Coiled-coil domain-containing protein 51 OS=Mus musculus GN=Ccdc51 PE=2 SV=1 - [CCD51_MOUSE]
Q924D0	27.78	10	9	396	43.3	9.20	33.23	Reticulon-4-interacting protein 1, mitochondrial OS=Mus musculus GN=Rtn4ip1 PE=1 SV=2 - [RT4I1_MOUSE]
Q99JY4	34.04	15	11	376	42.2	8.41	32.73	TraB domain-containing protein OS=Mus musculus GN=Trabd PE=2 SV=1 - [TRABD_MOUSE]
Q8VCM4	26.01	10	9	373	42.1	8.48	32.70	Lipoyltransferase 1, mitochondrial OS=Mus musculus GN=Lipt1 PE=2 SV=1 - [LIPT_MOUSE]
O35435	53.42	13	13	395	42.7	9.55	32.01	Dihydroorotate dehydrogenase, mitochondrial OS=Mus musculus GN=Dhodh PE=2 SV=2 - [PYRD_MOUSE]
Q91WK2	26.99	9	8	352	39.8	6.67	29.85	Eukaryotic translation initiation factor 3 subunit H OS=Mus musculus GN=Eif3h PE=1 SV=1 - [EIF3H_MOUSE]
P63085	35.47	10	8	358	41.2	6.98	28.75	Mitogen-activated protein kinase 1 OS=Mus musculus GN=Mapk1 PE=1 SV=3 - [MK01_MOUSE]
Q8BVU5	36.00	14	12	350	38.6	6.76	28.43	ADP-ribose pyrophosphatase, mitochondrial OS=Mus musculus GN=Nudt9 PE=2 SV=1 - [NUDT9_MOUSE]
Q8R0Z5	26.92	8	8	364	39.3	8.44	28.15	Mitoferrin-2 OS=Mus musculus GN=Slc25a28 PE=2 SV=1 - [MFRN2_MOUSE]
Q9CZ57	32.02	12	11	381	42.8	8.25	27.90	Putative methyltransferase NSUN4 OS=Mus musculus GN=Nsun4 PE=2 SV=1 - [NSUN4_MOUSE]
Q99M87	21.25	12	9	480	52.4	9.22	27.51	DnaJ homolog subfamily A member 3, mitochondrial OS=Mus musculus GN=Dnaja3 PE=1 SV=1 - [DNJA3_MOUSE]
Q8R0G7	22.16	6	5	528	56.7	7.15	27.12	Protein spinster homolog 1 OS=Mus musculus GN=Spns1 PE=2 SV=1 - [SPNS1_MOUSE]
Q8CAY6	31.23	9	7	397	41.3	7.50	26.23	Acetyl-CoA acetyltransferase, cytosolic OS=Mus musculus GN=Acat2 PE=1 SV=2 - [THIC_MOUSE]
Q791T5	31.36	12	11	389	41.5	9.32	26.19	Mitochondrial carrier homolog 1 OS=Mus musculus GN=Mtch1 PE=1 SV=1 - [MTCH1_MOUSE]