Mitochondrial fusion is increased by the nuclear coactivator PGC-1beta

Abstract

Background: There is no evidence to date on whether transcriptional regulators are able to shift the balance between mitochondrial fusion and fission events through selective control of gene expression.

Methodology/principal findings: Here, we demonstrate that reduced mitochondrial size observed in knock-out mice for the transcriptional regulator PGC-1beta is associated with a selective reduction in Mitofusin 2 (Mfn2) expression, a mitochondrial fusion protein. This decrease in Mfn2 is specific since expression of the remaining components of mitochondrial fusion and fission machinery were not affected. Furthermore, PGC-1beta increases mitochondrial fusion and elongates mitochondrial tubules. This PGC-1beta-induced elongation specifically requires Mfn2 as this process is absent in Mfn2-ablated cells. Finally, we show that PGC-1beta increases Mfn2 promoter activity and transcription by coactivating the nuclear receptor Estrogen Related Receptor alpha (ERRalpha).

Conclusions/significance: Taken together, our data reveal a novel mechanism by which mammalian cells control mitochondrial fusion. In addition, we describe a novel role of PGC-1beta in mitochondrial physiology, namely the control of mitochondrial fusion mainly through Mfn2.

Figure 1. PGC-1β increases Mfn2 transcription through nuclear receptor ERRα.

(A) Total RNA was obtained from C2C12 myotubes transduced with LacZ (white bars) or PGC-1β (black bars) adenovirus at a MOI 100 from 3 independent experiments in duplicate. Mfn2 and β-Actin mRNA levels were measured by real-time PCR. Results are mean±SEM and are expressed relative to β-Actin values, *, statistical difference at p<0.001. (B) Transcriptional activity of −1982/+45 human Mfn2 promoter was determined by luciferase activity (corrected by renilla activity) in 10T1/2 and HeLa cells. Results are mean±SEM from 3 independent experiments performed in triplicate. Cells were transfected with 200 ng of an irrelevant vector (Basal, white bars) or PGC-1β expression vector (PGC-1β, black bars) and expressed relative to the basal group; * statistical difference at p<0.05. (C) Luciferase activity (corrected by renilla) from 10T1/2 cells transfected with a −459/−352 Mfn2 promoter fragment containing a mutation that disrupted box 2 (−459/−352 Box 2 Mut, black bars) or with the wild-type fragment (−459/−352, white bars), together with an irrelevant vector (Basal), ERRα, 200 ng of PGC-1β or ERRα+PGC-1β expression plasmids. Mean±SEM values from 3 independent transfection experiments performed in triplicate are shown. *, statistical difference compared with basal groups, p<0.05. #, statistical difference compared with the wild-type promoter fragment activity, p<0.05.

Figure 2. Effect of PGC-1β on the expression of Mfn2, other proteins involved in mitochondrial dynamics, Porin and ETC subunits.

Total lysates and mitochondrial fractions were obtained from C2C12 myotubes 48 h after adenoviral transduction, and the expression of proteins was analyzed by Western blot (40 µg). Representative films of a single experiment from four independent differentiation and transduction experiments are shown. (A) Total lysates and mitochondrial fractions of non-transduced myotubes (NT), transduced with LacZ or GFP control adenovirus at a MOI 100 (control adenovirus at a MOI 1 and 10 displayed no changes, data not shown) or with PGC-1β adenovirus at a MOI 1, 10 or 100 were probed with anti-Mfn2 and Porin antibodies. (B) Specific antibodies were used to detect mitochondrial dynamics proteins (Mfn1,OPA1, Drp1 and Fis1) and single subunits from each complex of the ETC (complex I, Ndufa9; complex II, Sdha; complex III, Uqcrc2; complex IV, Cox4 and complex V, ATP5a1) in mitochondrial fractions of non-transduced (NT) or transduced C2C12 myotubes at a MOI 100 with LacZ or PGC-1β adenovirus. (C) Quantification analysis of protein levels of mitochondrial dynamics components and the ETC subunits from complexes I to V detected by Western blot. Mean±SEM of Non- (white bars) , LacZ- (grey bars) or PGC-1β- (black bars) transduced C2C12 myotubes from n = 4 independent differentiation and transduction (MOI 100) experiments. *, statistical difference compared to LacZ transduction, p<0.05. #, statistical difference compared to Porin induction, p<0.05.

Figure 3. PGC-1β increases the rate of mitochondrial fusion and changes mitochondrial morphology.

(A) Representative Western blot of three independent experiments of C2C12 myoblasts transduced with LacZ or PGC-1β adenovirus at a MOI 200 during 48 h. 40 µg of protein from lysates were loaded and Mfn2, Mfn1, Opa1, Drp1, Fis1 and Porin were detected with specific antibodies. (B) Representative images from a duplicate of one of the same three independent transduction experiments performed in panel (A). Mitochondria were labelled with anti-Cox1 antibody detected with a secondary antibody conjugated to Alexa 594. Scale bar, 10 µm. (C) C2C12 myoblasts quantification and classification attending to mitochondrial tubule length (150–200 cells counted per transduction). Myoblasts with a mean mitochondrial tubule length <4 µm were considered “Short” and >4 µm were considered “Long”. Results are mean±SEM and expressed as percentage of total C2C12 myoblasts counted from three independent transduction experiments with LacZ (white bars) or PGC-1β (black bars) at MOI of 200. *, statistical difference at p<0.01. (D) Electron microscopy images of C2C12 myoblasts transduced with LacZ or PGC-1β adenovirus as in panel (A) were taken at 40.000x magnification . Scale bar 500 nm. (E) 4 hours after PEG, polykaryons of C2C12 myoblasts transduced as in (A) were counted attending to mtGFP (mitochondrial matrix- targeted green fluorescent protein) and mtRFP (mitochondrial matrix-targeted red fluorescent protein) mixing (detected as yellow mitochondria). Polykaryons classified as “Low” displayed <50% of mixed mtGFP and mtRFP (yellow mitochondria). Polykaryons classified as “High” displayed ≥50% of yellow mitochondria. Graphs show mean±SEM of percentage of total polykaryons counted from three independent cell fusion experiments and transduction experiments with LacZ (white bars) or PGC-1β (black bars) at MOI of 200. *, statistical difference at p = 0.02. (F) Representative images from one of three independent transduction and polyethylene glycol-mediated (PEG) cell fusion experiments. C2C12 myoblasts stably expressing mitochondrial matrix-targeted green fluorescent protein (mtGFP) or red fluorescent protein (mtRFP) were transduced with LacZ or PGC-1β adenovirus as in panel (A). Polykaryons were fixed 4 h after 30”–40” PEG treatment. During these 4 h, cells were incubated in growth medium with 100 µg/ml of cycloheximide to inhibit de novo synthesis of mtGFP and mtRFP. Degree of fusion is directly proportional to mixing of mtGFP and mtRFP (yellow mitochondria). Scale bar, 10 µm.

Figure 4. Mfn2 is required for PGC-1β changes in mitochondrial morphology.

(A) Representative immunocytochemistry images (performed as in Fig. 3B) from one of three independent cotransfection experiments with an empty vector (Control) or PGC-1β- encoding vector (PGC-1β) and GFP (ratio GFP:PGC-1β 1:10) of wild-type mouse embryonic fibroblasts (MEFs). Transfected cells were visualized by GFP fluorescence (data not shown). Scale bar, 10 µm. (B) Wild type mouse embryonic fibroblasts (MEF) quantification and classification attending to mitochondrial tubule length (50-75 transfected cells counted per independent experiment). Wild type MEFs with a mean mitochondrial tubule length <4 µm were considered “Short” and >4 µm were considered “Long”. “Perinuclear” MEFs displayed most mitochondria surrounding the nucleus (see Figure S4) and almost no “Fragmented” cells (with tiny spherical mitochondria, see Figure S4) were observed in transfected wild type MEFs. Results are mean±SEM and expressed as percentage of total transfected wild type MEFs counted from three independent cotransfection experiments with an empty vector (white bars) or PGC-1β (black bars) with GFP (empty vector/PGC-1β : GFP ratio 10:1). Transfected cells were observed by GFP fluorescence (data not shown). *, statistical difference at p<0.02. (C) Representative immunocytochemistry images (performed as in Fig. 3B) from one of three independent cotransfection experiments with an empty vector (Control) or PGC-1β- encoding vector (PGC-1β) and GFP (ratio GFP:PGC-1β 1:10) of Mfn2 −/−mouse embryonic fibroblasts (MEFs). Transfected cells were visualized by GFP fluorescence (data not shown). Scale bar, 10 µm. (D) Mfn2 −/− mouse embryonic fibroblasts (MEF) quantification and classification attending to mitochondrial tubule length (100–250 transfected cells counted per independent experiment). Mfn2 −/− MEFs with a mean mitochondrial tubule length <4 µm were considered “Short” and >4 µm were considered “Long”. “Perinuclear” MEFs displayed most mitochondria surrounding the nucleus (see Figure S4) and a significant percentage of “Fragmented” cells (with tiny spherical mitochondria, see Figure S4) were observed in transfected Mfn2 −/− MEFs. Results are mean±SEM and expressed as percentage of total transfected Mfn2 −/− MEFs counted from three independent cotransfection experiments with an empty vector (white bars) or PGC-1β (black bars) with GFP (empty vector/PGC-1β : GFP ratio 10:1). Transfected cells were observed by GFP fluorescence (data not shown).

Figure 5. PGC-1β KO mice show a specific decrease in Mfn2 expression in skeletal muscle.

Gastrocnemius muscles from 4- or 8-month-old wild-type (WT, black bars) or PGC-1β KO male mice (KO, white bars) were used to obtain total RNA and protein. (A) Representative image from a Western blot with specific detection of Mfn2 and Porin in muscle lysates. Graph represents mean±SEM of Mfn2 levels relative to Porin values (n = 6 mice per group). *, statistical difference at p<0.01. (B) Abundance of proteins involved in mitochondrial dynamics (Mfn1, OPA1, Drp1 and Fis1) in total lysates. Graph represents mean±SEM of protein levels relative to Porin values (n = 6 mice per group). (C) Real-time PCR analysis of Mfn2 from WT and KO mice. Graph represents mean±SEM and data are expressed as values of Mfn2 relative to 36B4 mRNA levels (n = 8 mice per group). *, statistical difference at p<0.01. (D) Abundance of the ETC subunits in mitochondrial fractions (n = 6 mice per group). Subunits of complexes I to V of the ETC were detected with specific antibodies. Graph represents mean±SEM and data are expressed as values relative to Porin expression. *, statistical difference at p<0.05.

Figure 6. Hepatic Mfn2 levels are diminished in PGC-1β KO mice.

Livers from 4- or 8- month-old wild-type (WT, black bars) and PGC-1β KO male mice (KO, white bars) were used to obtain total RNA and protein. (A) Representative image from a Western blot with specific detection of Mfn2 and Porin in liver lysates from WT or KO mice. Graph represents mean±SEM of Mfn2 levels related to Porin (n = 6 mice per group). *, statistical difference at p<0.01. (B) Abundance of proteins involved in mitochondrial dynamics in total lysates (Mfn1, OPA1, Drp1 and Fis1). Graph represents mean±SEM of protein levels relative to Porin values (n = 6 mice per group). (C) Real-time PCR analysis of Mfn2 from WT and KO mice. Graph represents mean±SEM and data are expressed as values of Mfn2 relative to 36B4 mRNA levels (n = 8 mice per group). *, statistical difference at p<0.01. (D) Abundance of the ETC subunits in mitochondrial fractions (n = 6 mice per group). Single subunits of complexes I to V of the ETC were detected with specific antibodies. Graph represents mean±SEM and data are expressed as values relative to Porin expression. *, statistical difference at p<0.05.