Fig 2: MTA1 promotes the malignant phenotype of CRC via binding with ATP5A. A) The proliferation of MTA1‐OE and control HCT116 cells treated with 2.5 mm oligomycin. The values are the mean ± SD (****p < 0.0001), two‐way ANOVA, Tukey's multiple comparisons test, n = 2 replicate per group. B,C) The colony formation ability of MTA1‐OE and control HCT116 cells treated with 2 µm oligomycin for 10 days. The values are the mean ± SD (***p < 0.001; ns, not significant), two‐way ANOVA, Sidak's multiple comparisons test, n = 3 replicate/group. D,E) The invasion of MTA1‐OE and control HCT116 cells treated with 2 µm oligomycin for 48 h. The values are the mean ± SD (**p < 0.01; ns, not significant), two‐way ANOVA, Sidak's multiple comparisons test, n = 3 replicate/group. Scale bar: 100 µm. F) Western blotting was performed to verify that ATP5A was knocked down in MTA1‐OE and control HCT116 cells, and the expression of metastasis related markers in MTA1‐OE/ATP5A‐KD HCT116 cells. G–I) The proliferation (G) and invasion (H,I) of MTA1‐OE/ATP5A‐KD and control HCT116 cells. The cells were incubated in transwell plates for 48 h. The values are the mean ± SD (****p < 0.0001, ***p < 0.001; ns, not significant), two‐way ANOVA, Tukey's multiple comparisons test, n = 3 replicate/group. Scale bar: 200 µm. J,K) The ECAR (J) and OCR (K) of MTA1‐OE/ATP5A‐KD and control HCT116 cells. The values are the mean ± SD (****p < 0.0001, ***p<0.001, and *p<0.05), two‐way ANOVA, Tukey's multiple comparisons test. L) Representative image of the liver metastatic burden in BALB/c‐nu/nu mice injected with MTA1‐OE/ATP5A‐KD and control cells via the spleen tail. The group names are shown in (M); n = 3 mice/group. M) The liver weight of mice with liver metastasis shown in (L). The values are the mean ± SD (*p < 0.05; ns, not significant), one‐way ANOVA, Tukey's multiple comparisons test.

Fig 3: MTA1 interacts with ATP5A. A) Interaction network of 24 mitochondria‐associated proteins that bind MTA1 identified by LC–MS/MS. The shade of color indicates the number of proteins interacting with that protein. B) Schematic of the distribution of MTA1‐binding proteins, which are highlighted with different colors on ATP synthase. C) Simulation of the interaction between amino acids 670‐695 of MTA1 and ATP synthase in state 2. Binding sites are indicated by black boxes and magnified. MTA1 peptides are shown in orange (top) and pink (bottom). The ATP synthase α subunit is shown in green. The ATP synthase γ subunit is shown in purple. D) Average binding energy of the interaction between MTA1 peptides and ATP synthase subunits determined by simulated electron diffraction patterns (eV). E) Western blot analysis of MTA1 and ATP5A levels in mitochondria in MTA1‐KO cells and MTA1‐OE cells. F,G) Co‐IPs of MTA1 (F) and ATP5A (G) in HCT116 cells. H) Analysis of the binding between MTA1 and ATP5A in HCT8 cells by Co‐IP. I,J) The colocalization of MTA1 and ATP5A in mitochondria in HCT116 parent cells (I) and MTA1‐KO HCT116 cells (J) was visualized by immunofluorescence. Scale bar: 5 µm. K,L) Quantification of immunofluorescence. The values are the mean ± SD (**p < 0.01, *p < 0.05), Student's t‐test, n = 4 replicate/group.

Fig 4: MTA1 increases the efficacy of mTOR inhibitors both in vitro and in vivo. A) Schematic of the experiment. MTA1‐KO and control HCT116 cells were seeded in 384‐well plates (800 cells per well) 16 h before the experiment, and the cells were treated with 237 metabolism‐related anticancer drugs for 72 h (five dilutions per drug, two replicates). Cell viability was assessed by the CCK8 assay. B) Fold change in the IC50 of metabolism‐related anticancer drugs in MTA1‐KO cells versus control cells. The red dots are mTOR inhibitors, and the green dots are mTOR activators. C) Western blot analysis of the levels of MTA1, mTOR, p‐mTOR, AMPKα, and p‐AMPKα in MTA1‐KO cells, MTA1‐OE cells, and control HCT116 cells. D) IC50 curves of the mTOR inhibitors AZD8055, temsirolimus, zotarolimus, and Torin‐1 in MTA1‐KO and control HCT116 cells for individual pharmacological validation. The values are the mean ± SD. E) The viability of MTA1‐KO and control HCT116 cells treated with 5 µm temsirolimus. The values are the mean ± SD (****p < 0.0001), two‐way ANOVA, and Tukey's multiple comparisons test. F) Image of liver metastatic burden in the eight groups of BALB/c‐nu/nu mice injected with MTA1‐OE/ATP5A‐KD via the spleen tail. The mice were injected intraperitoneally with 100 µL of 8.8 mg mL−1 rapamycin every other day 2 weeks after spleen‐tail injection; n = 5 mice per group. G) The liver weight of mice with liver metastasis shown in (F). The values are the mean ± SD (****p < 0.0001, **p < 0.01; ns, not significant), one‐way ANOVA, and Tukey's multiple comparisons test. H) Representative HE staining images of the liver tumors of the mice in (F). Scale bar: 3 mm.

Fig 5: Location and structural modeling of the ATP5F1A missense variant p.Gly418Arg. (A) A three‐dimensional representation of the human ATPase F1 domain (PDB: 8H9I) is shown, with major subunits highlighted in the indicated colors. The residue Gly418 mutated in families B and C is situated at the α‐β intersubunit interaction space (ATP5F1A‐ATP5F1B). Moreover, Gly418 is in close proximity to the ADP binding site, as highlighted in the magnified view of the sequence containing this amino acid; ADP is shown in stick‐and‐ball representation (gray). (B) Model of ATPase and magnified view of the predicted p.Gly418Arg substitution in the α‐β communication space (PDB: 6ZPO). Substitution of the nonpolar Gly418 by a positively charged arginine may perturb functionally important subunit interactions and/or structural features that sustain binding (or nonbinding) behaviors. As illustrated by blue dashed lines, Arg418 may introduce new intermolecular bonding interactions (H‐bonds) involving the α‐ and β‐subunits. [Color figure can be viewed at wileyonlinelibrary.com]