IF1 plays a role in mediating metabolic reprogramming and is upregulated in many cancers contributing to Warburg phenotype (11). IF1 expression varies in different tissues and cell types (6) ( Figure 1A ). The phosphorylation level of c-Myc on the Thr-58 site appeared to be positively correlated with the expression level of IF1 ( Figure 1A ). In addition, the expression levels of SIRT3 and OSCP were both upregulated in some cell lines including HCT116 and MIA PaCa-2 cells ( Figure 1A ), which contributed to cell adaption to mitochondrial stress. The CRISPR/Cas9 technique was used to disrupt ATPIF1 gene in HCT116 cells ( Figure 1B ). The ablation of IF1 led to the declined expressions of c-Myc and PGC1α ( Figure 1C ), indicating that IF1 might mediate metabolic reprogramming through interplay with these transcription factors. The influence of ATPIF1 inactivation on metabolic phenotype was detected by Seahorse XF Analyzer. The ablation of IF1 led to a significant rise in both basal and maximal oxidative respiration ( Figures 1D, E ), as well as oligomycin-sensitive mitochondrial respiration ( Figure 1F ). The basal glycolytic metabolism was inhibited by ablation of IF1 ( Figures 1G, H ), suggesting that disruption of ATPIF1 gene favors metabolic switch to OXPHOS from glycolysis ( Figure 1I ). Interestingly, the ablation of IF1 caused significantly increased glycolytic capacity and glycolytic reserve ( Figures 1G, H ), indicating that IF1 plays a role in mediating glycolytic metabolism.

c-Myc drives metabolic reprogramming, which favors tumorigenesis and cancer cell survival (36). c-Myc enhances glycolytic gene expression and activation of c-Myc drives aerobic glycolysis, thus contributing to oncogenic metabolic state (36, 37). Phosphorylation sites Thr-58 and Ser-62 control c-Myc-dependent transactivation of gene expression and regulate c-Myc protein stability (38–40). Decreased Thr-58 and increased Ser-62 phosphorylation stabilize c-Myc protein, and ratios of Thr-58 and Ser-62 phosphorylation are altered in human cancer (39). Immunoprecipitation assay revealed that IF1 strongly bound to c-Myc ( Figure 2A ) and Thr-58 phosphorylated c-Myc ( Figures 2A, B ). To further investigate whether this binding could occur in mitochondria, mitochondria were purified for immunoprecipitation assay. IF1 bound to both c-Myc and Thr-58 phosphorylated c-Myc in mitochondria ( Figure 2C ). Overexpression of c-Myc accompanied by enhanced phosphorylation of c-Myc on the Thr-58 site ( Figure 2D ) did not affect oxidative metabolism ( Figures 2E–G ) but promoted glycolysis ( Figures 2H–J ). The ablation of IF1 abolished the stimulating effect of c-Myc on glycolytic activity ( Figures 2K–M ). The observations suggested that IF1 bound to c-Myc in mitochondria and then mediated the phosphorylation and activity of c-Myc.

Cancer cells are highly dependent on glucose to fuel their energy demand and proliferate, and c-Myc plays a key role in aerobic glycolysis. The effects of high glucose on cellular metabolism and the interaction between IF1 and c-Myc were then investigated. Exposure to high glucose promoted both oxidative and glycolytic activities ( Figures 3A–E ), while oxidative respiration was more influenced with a significant increase of basal OCR/basal ECAR ratio ( Figure 3F ). The maximal oxygen consumption rate induced by injection of FCCP and oligomycin sensitive respiration were almost identical in the absence or presence of glucose ( Figures 3A–C ), indicating that mitochondrial integrity and function were not affected by the pre-exposure to high glucose. The glycolytic capacity and reserve were significantly enhanced by high glucose ( Figures 3D, E ). High glucose enhanced the expressions of c-Myc and Thr-58 phosphorylated c-Myc, and their translocations to mitochondria ( Figure 3G ). Consistent with the observation of Western blotting analysis, immunofluorescence staining revealed that a fraction of c-Myc was localized to mitochondria upon high glucose stimulation ( Figure 3H ), which was required for the interaction with IF1 and Thr-58 phosphorylation of c-Myc ( Figure 3I ). These observations suggest that the localization of c-Myc to mitochondria and its interaction with IF1 might promote cellular metabolism and metabolic reprogramming.

Activation of PGC1α promotes mitochondrial oxidative respiration (25), and the c-Myc/PGC1α ratio is a main controller of metabolic phenotypes and plasticity in pancreatic cancer stem cells (27). IF1 is a mitochondrial specific protein (33). Immunofluorescence analysis showed that PGC1α could localize to both nucleus and mitochondria ( Figure 4A ). Immunoprecipitation revealed that IF1 could bind to PGC1α ( Figures 4B, C ). In the absence of IF1, c-Myc exhibited more significant enhancement on mitochondrial oxidative respiration and glycolytic metabolism than PGC1α ( Figures 4D–H ). However, the effect of c-Myc on glycolytic metabolism was more significant than that on oxidative metabolism, resulting in a significant decrease of OCR/ECAR ratio compared to the effect of PGC1α ( Figure 4I ). The presence of IF1 led to a dramatic decline of oxidative metabolism but did not influence the glycolytic metabolism ( Figures 4D–I ). IF1 might interact with PGC1α and destabilize PGC1α, resulting in a significant decrease of mitochondrial oxidative phosphorylation. The stability of c-Myc and PGC1α inhibited by IF1 may contribute to metabolic reprogramming.

The role of IF1 in c-Myc and PGC1α imported into mitochondria was further investigated. Inhibition of mitochondrial respiration or induction of mitochondrial dysfunction slightly reduced the expression levels of c-Myc and PGC1α ( Figure 5A ). The longer-term exposure to mitochondrial inhibitors induced a dramatic decrease in mitochondrial mass, which could be caused by mitophagy, resulting in a striking decline of c-Myc and PGC1α expressions ( Figure 5A ). Of note, mitochondrial stress induced by inhibitors of mitochondrial respiration, CoCl₂, or uncoupler FCCP promoted c-Myc and PGC1α import into mitochondria ( Figures 5B, D, F ). The ablation of IF1 prevented the mitochondrial import of c-Myc and PGC1α under mitochondrial stress, for instance, after antimycin A treatment ( Figures 5C, E, G ). These observations indicated that IF1 was required for the mitochondrial import of c-Myc and PGC1α under mitochondrial stress condition.

The combination with c-Myc stimulation of abnormal mitochondria and hypoxic microenvironment in tumors synergistically stimulates mitochondrial ROS production through the ineffective function of respiratory complex III under limiting oxygen conditions (19, 41). Chemical incubation cells with CoCl₂ induce cellular hypoxia, which triggers dephosphorylation of IF1 and inhibits both the synthetic and hydrolytic activities of F-ATP synthase (5). Hypoxic microenvironment and glucose deprivation facilitate proteolytic degradation of c-Myc and attenuate its function (36, 42). Hypoxia stimulates the expression of PGC1α and mitochondrial biogenesis in cardiac myocytes (43). Mild ROS intensity triggers nuclear reprogramming, which is involved in the metabolic adaption and activation of cellular survival (7).

Overexpression of IF1, c-Myc, or PGC1α induced cell death, which was prevented by CoCl₂ ( Figures 6A, F ). The ablation of IF1 ( Figure 6B ) abolished the protective function of CoCl₂ in IF1/c-Myc/PGC1α overexpression-induced cell death ( Figures 6C, F ). Knockout of IF1 had no effect on IF1/c-Myc/PGC1α overexpression-induced cell death, unless CoCl₂ was added ( Figures 6D–F ). These results suggested that IF1 was necessary for the protective role of hypoxia in c-Myc/PGC1α-induced cell death. Treatment of CoCl₂ dephosphorylated IF1 (5), which might favor the binding of IF1 to c-Myc/PGC1α and prevented their inducing effects on cell death. Overexpression of IF1 appeared to blunt the production of mitochondrial ROS ( Figure 6G ), which may be due to decreased OXPHOS activity and enhanced activity of the permeability transition pore (PTP) (33). Suppression of OXPHOS activity by IF1 overexpression may involve its interactions with c-Myc/PGC1α.

HEK293T, HCT116, HeLa, and MIA PaCa-2 cells were supplied by Cell Bank, Chinese Academy of Sciences. Cells were cultured in DMEM medium (high glucose, pyruvate, Gibco, C11995500BT) supplemented with 10% fetal bovine serum (FBS). The primary antibodies are as follows: ATPIF1 polyclonal antibody (12067-1-AP, Proteintech); PGC1α monoclonal antibody (66369-1-Ig, Proteintech); HDAC1 polyclonal antibody (10197-1-AP), ATP5A1 polyclonal antibody (14676-1-AP, Proteintech); ATPB polyclonal antibody (17247-1-AP, Proteintech); ATP5O polyclonal antibody (10994-1-AP, Proteintech); SirT3 (C73E3) rabbit mAb (#2627, Cell Signaling Technology); p-c-Myc (Thr 58) rabbit polyclonal antibody (sc-135647, Santa Cruz Biotechnology); Myc/c-Myc mouse monoclonal antibody (sc-40, Santa Cruz Biotechnology); and Anti-TOMM20 antibody (db50, diagbio).

The ATPIF1 knockout cell line was generated using the CRISPR/Cas9 technique (63). Briefly, the specific single guide RNA (sgRNA) was constructed into the lentiviral expression vector for Cas9 and sgRNA. The lentiCRISPR vector was linearized by the BsmBI restriction enzyme (New England Biolabs). The sequences of sgRNA are as follows: ATPIF1 sgRNA-1: 5′-CACCGCAGTGACGGCGTTGGCGGCG-3′ (forward), 3′-AAACCGCCGCCAACGCCGTCACTGC-5′ (reverse); ATPIF1 sgRNA-2: 5′-CACCGTCCAGCAGCAATGGCAGTGA-3′ (forward), 3′-AAACTCACTGCCATTGCTGCTGGAC-5′ (reverse); ATPIF1 sgRNA-3: 5′-CACCGGGCTTGGCGTGTGGGGCGTG-3′ (forward), 3′-AAACCACGCCCCACACGCCAAGCCC-5′ (reverse).

Cells were collected by centrifugation at 600 × g for 5 min at 4°C, and lysed in RIPA buffer supplemented with protease inhibitors on ice for 30 min. Supernatant was harvested after centrifugation at 12,000 × g for 30 min. Protein concentration was determined by BCA assay. The protein solution was supplemented with SDS sample loading buffer followed by SDS-PAGE gel electrophoreses. Proteins were transferred to nitrocellulose membranes followed by membrane blocking with 5% (w/v) nonfat dry milk. The blotted membranes were incubated with primary antibodies overnight at 4°C. The bands were visualized with enhanced chemiluminescence.

OXPHOS and glycolysis activities were detected with an Agilent Seahorse XFe24 Analyzer. HCT116 and HeLa cells were seeded in XF24 microplates at 2×10⁴ cells/well in DMEM medium supplemented with 10% FBS. Cells were incubated at 37°C in a 5% CO₂ humidified incubator for 48 h, or 24 h later, cells were transfected with plasmids and then incubated for 24 h. The culture medium was replaced with Seahorse XF Base Medium (without Phenol Red) supplement with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose. Cells were incubated at 37°C for 45 min to allow temperature and pH equilibration. Following the baseline measurement, 2 µM oligomycin, 0.25 µM FCCP, 1 µM rotenone plus antimycin A, and 50 mM 2-deoxyglucose (2-DG) were sequentially added to each well. The oxygen consumption rate (OCR) values were subtracted for the rotenone and antimycin-insensitive respiration. The extracellular acidification rate (ECAR) values were subtracted for the 2-DG insensitive glycolysis. The values of OCR and ECAR where indicated were normalized for µg of protein determined by BCA assay.

HEK293T or HeLa cells were collected by centrifugation at 600 × g for 5 min and washed once with prechilled PBS buffer. Cell pellet was resuspended in isolation buffer containing 250 mM sucrose, 10 mM Tris-HCl, and 0.1 mM EGTA, pH 7.4, and homogenized with a Potter homogenizer. The supernatant was collected after centrifugation at 600 × g for 5 min. The mitochondrial pellet was harvested and washed once with isolation buffer by centrifugation at 8,000 × g for 15 min. BCA assay was used to determine mitochondrial protein concentration.

cDNA clones of ATPIF1, c-Myc containing XhoI and EcoRI restriction sites were obtained by a PCR-based method using the whole cDNA of HEK293T as template. The primers used for PCR are as follows: for ATPIF1: 5′-ATGCGAATTCATGGCAGTGACGGCGTTGGC-3′ (forward), 3′-ATGCCTCGAGATCATCATGTTTTAGCATTT-5′ (reverse); for c-Myc: 5′-CAGTGTGCTGGAATTCTGGATTTTTTTCGGGTAGTGGA-3′ (forward), 3′-CGTAGGGGTACTCGACGCACAAGAGTTCCGTAGC-5′ (reverse). The C terminus HA-tagged pCDNA3.1 vector was linearized by XhoI and EcoRI restriction enzymes (New England Biolabs). The PCR products were cloned into the linearized vector using a ligation high kit (Toyobo). The inserted genes were verified by DNA sequencing.

Cells were seeded in a 6-well plate and cultured overnight. Plasmid DNA transfection was achieved by a polyethylenimine (PEI)-mediated transfection method. For one well, a mixture of 2 µg of plasmids, 6 µl of PEI, and 200 µl of free DMEM medium was incubated at room temperature for 15 min, and added to the six-well plate. After 24 h, proteins were extracted for immunoprecipitation or immunoblotting.

Cells were solubilized in CoIP lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, 10% glycerophosphate, and a cocktail of proteinase inhibitors. The supernatant was collected after a centrifugation at 12,000 × g for 20 min at 4°C. Cell lysate was incubated with anti-HA Agarose Beads (Abmart), or incubated with Protein A/G Plus Agarose Beads (Santa Cruz) supplemented with ATPIF1 antibody at 4°C overnight. The beads were washed three times with PBS buffer supplemented with 0.1% NP-40, and boiled in 2× loading buffer for 10 min. Immunoblot analysis was then performed to detect the proteins that were pulled down.

Cells were seeded in a 12-well dish on sterile 18-mm glass coverslips, cultured overnight, and rinsed three times with PBS for 5 min each. The fixation step was initiated with the addition of 4% paraformaldehyde and incubated at room temperature for 15 min. Aspirated fixative and rinsed three times with PBS for 5 min each. Cells were permeabilized with 0.1% Triton X-100, incubated at room temperature for 10 min, aspirated with Triton X-100, and rinsed three times with PBS for 5 min each. Cells were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature; aspirated with blocking solution; incubated with 1:100 dilution primary antibodies against IF1 (Proteintech, 12067-1-AP), PGC1α (Proteintech, 66369-1-Ig), c-Myc (Santa Cruz Biotechnology, sc-40), and TOM20 (Proteintech, 11802-1-AP) at 4°C overnight; rinsed three times with PBST buffer (PBS with 0.1% Tween 20) for 5 min each; incubated with DAPI and 1:200 dilution of secondary antibodies (goat anti-mouse IgG Alexa Fluor 488 and goat anti-rabbit IgG Alexa Fluor 647) in 5% BSA at room temperature for 1 h; and rinsed three times with PBST buffer for 5 min each. Fluorescence images were acquired with confocal microscopy (Leica STELLARIS 5).

Cells were seeded in a 96-well plate at 1×10⁴ cells/well in DMEM medium supplemented with 10% FBS, and incubated at 37°C in a 5% CO₂ humidified incubator overnight. Cells were then transfected with empty C terminus HA-tagged pCDNA3.1 vector or ATPIF1 expressing vector. Cells were cultured for another 24 h to allow the expression of inserted gene. The culture medium was replaced with 5 µM MitoSOX working solution in Hanks’ Balanced Salt Solution (HBSS), incubated at 37°C for 10 min, and rinsed three times with HBSS. Fluorescence images were acquired with a Th4 200 inverted fluorescence microscope (Olympus).

Cells were seeded in 96-well plates at 5×10³ cells/well in DMEM medium supplemented with 10% FBS and incubated at 37°C in a 5% CO₂ humidified incubator overnight. Cells were treated with 0.2 mM CoCl₂, and transfected with empty C terminus HA-tagged pCDNA3.1 vector or ATPIF1 expressing vector. Cells were maintained at 37°C in a 5% CO₂ humidified incubator for 24 h. One hundred microliters of freshly prepared MTT solution (5 mg/ml) was added to each well and incubated at 37°C for 4 h. Then, 100 µl of 10% SDS was added to each well, and OD value at 570 nm was measured with an Infinite 200 PRO plate reader (Tecan).

All the data were presented as mean ± SD of at least three independent experiments, or as representative blots and images. p-values indicated in the figures are calculated with GraphPad. One-way ANOVA was used to analyze the difference between more than two groups. Two-way ANOVA was used to determine the interaction between two independent variables. ^* p < 0.05 was considered statistically significant.