RNA G-quadruplexes control mitochondria-localized mRNA translation and energy metabolism

Abstract

Cancer cells rely on mitochondria for their bioenergetic supply and macromolecule synthesis. Central to mitochondrial function is the regulation of mitochondrial protein synthesis, which primarily depends on the cytoplasmic translation of nuclear-encoded mitochondrial mRNAs whose protein products are imported into mitochondria. Despite the growing evidence that mitochondrial protein synthesis contributes to the onset and progression of cancer, and can thus offer new opportunities for cancer therapy, knowledge of the underlying molecular mechanisms remains limited. Here, we show that RNA G-quadruplexes (RG4s) regulate mitochondrial function by modulating cytoplasmic mRNA translation of nuclear-encoded mitochondrial proteins. Our data support a model whereby the RG4 folding dynamics, under the control of oncogenic signaling and modulated by small molecule ligands or RG4-binding proteins, modifies mitochondria-localized cytoplasmic protein synthesis. Ultimately, this impairs mitochondrial functions, affecting energy metabolism and consequently cancer cell proliferation.

Fig. 1. RG4 localization and function in mitochondria.

a RG4 detection in fixed cells using the BioCyTASQ probe (1 µM for 24 h) coupled with TOMM20 immunofluorescence in HeLa cells treated with DMSO or 20 µM of PhenDC3 for 24 h. Scale bar: 10 μm. Shown is a representative result from n = 3 independent experiments. b Graphs displaying the fluorescence intensity (arbitrary unit) in each channel over the distance depicted by the white arrow in (a). The correlation coefficient (R²) of the colocalized fluorescence intensities of BioCyTASQ with TOMM20 signal is plotted. Data are presented as mean values ± SEM for 5 or 6 different cells, P-value = 0.0024. c Quantification of the normalized Oxygen Consumption Rate (OCR) per cell, measured by a Seahorse assay, in HCT116 cells treated with DMSO or a dose scale of PhenDC3 for 16 h. Data are presented as mean values ± SEM of n = 3 independent experiments. d Quantification of the Mitochondrial Membrane Potential (MMP) using flow cytometry analysis of the fluorescent TMRE probe in HCT116 cells treated with DMSO or PhenDC3 for 16 h or 48 h and plotted relatively to the DMSO condition. Data are presented as mean values ± SEM of n = 3 independent experiments, P-value = 0.0135 and P-value = 0.0373 for 16 h and 48 h PhenDC3 treatment, respectively. e Quantification of the Oxygen Consumption Rate (OCR) linked to basal and maximal mitochondrial respiration, ATP turnover, proton leak, spare respiratory capacity, measured by a Seahorse assay, and of the proliferation in HCT116, HeLa and SKMEL28 cells treated with DMSO or a dose scale of PhenDC3 for 16 h. Data are presented as mean values ± SEM of n = 2 (for proliferation assay in HeLa) or n = 3 independent experiments. For all the panels, *P < 0.05, **P < 0.01, ***P < 0.001, ns: non-significant (two-sided paired t-test for OCR and two-way ANOVA for the proliferation). For (b, c, d, e) Source data are provided as a Source Data file indicating exact P-values for (e).

Fig. 2. RG4s are molecular determinants of OMM-localized translation.

a Immunoprecipitation (IP) of in cellulo RNA-protein complexes using the BG4 antibody or control IgG, followed by RT–qPCR analysis. Fold changes are indicated. Data represent n = 2 or n = 3 independent experiments in duplicate when 4 or 6 dots are plotted, respectively. b Western blot analysis in HCT116 cells treated with PhenDC3 for 48 h. c Fold enrichment of protein/mRNA levels from (b). d AKAP1 mRNA/protein localization analysed by RNAscope and immunofluorescence (IF). (e) Quantification from (d). P-value = 0.0297 or 0.0425 for PhenDC3 or Puro treatment, respectively (two-sided paired t-test). f RNAscope analysis of AKAP1 and MTND5 mRNA localization as in (d). g Quantification from (f). h Proximity Ligation Assay coupled with puromycin Labeling (PLA-Puro) of AKAP1 or IkB, combined with TOMM20 IF. Graphs displaying the fluorescence intensity in each channel over the distance are depicted. i Quantification from (h). P-value = 0.0125 for puro-AKAP1 after PhenDC3 treatment. j Ratio of Renilla/Firefly luciferase activities (Rluc/Fluc) after transfection of RLuc-GALNT2 mRNA reporters (depicted in Supplementary Fig. 3j) containing the RG4 unmodified (WT), mutated (Mut) or 7-deaza-modified (7dG) for n = 4 independent experiments. k Quantification of the colocalization between TOMM20 and PLA-Puro dots from in vitro transcribed Rluc-GALNT2 reporter mRNA containing the RG4 WT, Mut or 7dG from images in Supplementary Fig. 2i. P-value = 0.0243 for the Puro-Rluc. In (a, c, j), data represent mean ± SEM. In (e, g, i, k), boxes extend from 25th to 75th percentiles; middle line is the median, and “+” is the mean; the whiskers show upper and lower extremes. For (a, b, c, e, g, i, j, k) Source Data are provided as a Source Data file indicating exact P-values for (a, c, j). In (b, c), the blot shown is representative of n = 3 independent experiments. In (d, f, h) a representative field from n = 3 independent experiments is shown. In (e, g, i, k), data represent n = 3 independent experiments and each dot is a cell. In (a, c, g, j, k), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns non-significant (two-sided paired t-test).

Fig. 3. hnRNP U binds nuclear-encoded RG4-containing nuclear-encoded mitochondrial mRNAs.

a Subcellular fractionation of HCT116 cells, followed by western blot analysis. Lamin A: nuclear marker, Tubulin: cytoplasmic marker associated with microsomes. Nuclear (N), perinuclear (PN), microsomal (M), and cytosolic fractions (C). Shown is a representative result from n = 3 independent experiments. b Proportion of hnRNP U binding sites in 5’UTR, CDS, and 3’UTR and percentage of mitochondrial mRNAs (mito) bound by hnRNP U in each region. c −log10(P-value) of the enrichment analysis of hnRNP U targets. P-value computed from Fisher exact test. d Immunoprecipitation of cellulo RNA-protein complexes (RIP) using the hnRNP U antibody, followed by RT-qPCR analysis. Fold changes are indicated in italics. e RNA affinity chromatography using the G3A2 and NRAS RG4s native (WT) or 7-deaza-modified (7dG), followed by western blot analysis. Shown is a representative result from n = 6 and n = 2 independent experiments for the G3A2 and NRAS RG4s, respectively. f Proportion of hnRNP U binding sites in 5’UTR, CDS, and 3’UTR containing experimental and predicted RG4s. g Density of hnRNP U binding sites around RG4s located in mRNAs (all mRNA), 5’UTRs, CDSs, and 3’UTRs. At distance 0 and until +30 −80: P = 0.004 for CDS, P = 0 for 3’UTR, P = 1 for 5’UTR (Bootstrap test with 1000 sets of random sites). h RIP using the BG4 antibody, followed by RT–qPCR analysis. Fold changes are indicated in italic. USP1: positive control. i RG4 detection after control (Ctr) or hnRNP U siRNAs. Scale bar: 20 μm. Shown is a representative result from n = 4 independent experiments. j Quantification of signal intensities from (i), each dot is a cell, P-value = 0.000181 (two-sided paired t-test). For (a, d, e, h, j) source data are provided as a Source Data file indicating exact P-values for (d) and (h). When indicated, *P < 0.05, **P < 0.01, ***P < 0.001, ns: non-significant (two-sided paired t-test). In (d, h), data are presented as mean values ± SEM of n = 2 or n = 3 independent experiments in duplicate when 4 dots or 6 dots are plotted, respectively.

Fig. 4. hnRNP U protein interactors and sub-cellular localization.

a Subcellular fractionation of HCT116 cells to obtain Total (T), Microsomal (M), and Mitochondrial (Mi) fractions, followed by western blot analysis. Shown is a representative result from n = 8 independent experiments for hnRNP U and DDX3X, n = 5 independent experiments for RPS6, OXPHOS complex and GRSF1, n = 6 independent experiments for TOMM20 and DHX30, n = 2 independent experiments for PRMT1 and DDX5. b Protease protection assay after mitochondria purification of HCT116 cells followed by western blot analysis of the indicated proteins. TX: Triton-X. PK: Proteinase K. Shown is a representative result from n = 2 independent experiments. c RNA affinity chromatography using the G3A2 sequences either native (WT) or 7-deaza-modified (7dG) and HCT116 OMM (Outer Mitochondrial Membrane) extracts, followed by western blot analysis. Shown is a representative result from n = 4 independent experiments for hnRNP U and GRSF1 and n = 3 independent experiments for DDX3X. d Distribution of hnRNP U partners identified by immunoprecipitation (IP) cytoplasmic HCT116 cell extracts, followed by mass spectrometry (IP-MS, n = 4 independent experiments). P = 1.58e-46, P = 5.01e-46, and P = 1.58e-22 for the mitochondrial proteins, the proteins involved in the cytoplasmic translation, and the RG4-BPs, respectively (Fisher test). For (a, b, c), source data are provided as a Source Data file.

Fig. 5. hnRNP U is involved in translation regulation of RG4-containing nuclear-encoded mitochondrial mRNAs.

a Polysome profile of HCT116 cells DMSO or puromycin (Puro) treated, followed by western blot analysis. RPS6: positive control. Shown is a representative result from n = 2 independent experiments. b RT-qPCR analysis from non-polysomal (NP), light (LP) and heavy (HP) polysomal fractions from Supplementary Fig. 7c. *P < 0.05, **P < 0.01, ***P < 0.001, ns: non-significant (two-sided paired t-test). c In vitro translation of Rluc-GALNT2 reporter mRNAs (depicted in Supplementary Fig. 3j) containing the RG4 unmodified (WT), or mutated (Mut) and of the control Rluc reporter mRNA with or without recombinant hnRNP U. d RNAscope analysis of AKAP1 mRNA localization and immunofluorescence (IF) analysis of AKAP1 protein localization in HeLa cells treated with control (Ctr) or hnRNP U siRNAs. Scale bar: 20 μm. e Quantification from images in (d). Data represent n = 3 independent experiments and each dot is a cell, P-value = 0.02486 (two-sided paired t-test). (f) RNAscope and IF analysis as in (d) in HeLa cells treated 20 µM PhenDC3 for 24 h or 300 nM Torin 1 for 2 h. Scale bar: 10 μm. g Quantification from images in (f). Data represent n = 3 and n = 2 independent experiments for the Puro, Torin 1 and PhenDC3 treatments, respectively, each dot is a cell. h RNA affinity chromatography using RG4 sequences either native (WT) or 7-deaza-modified (7dG) and cytoplasmic extracts of HCT116 treated with DMSO or 300 nM Torin 1 for 3 h, followed by western blot analysis. i Quantification of the protein levels on the RG4 WT from (h). P-value = 0.02336, ns non-significant (two-sided paired t-test). For (a, b, c, e, g, h), Source Data is provided as a Source Data file indicating exact P-values for (b, c, g). Data are presented as mean values ± SEM of n = 3 for (b), (c), n = 4 for (i) independent experiments. For (e, g), the box extends from 25th to 75th percentiles; middle line is the median and “+” is the mean; the whiskers show upper and lower extremes. For (d, f), shown is a single representative field over n = 3 independent experiments.

Fig. 6. hnRNP U modulates OXPHOS expression and function.

a Western blot analysis in HCT116 cells treated with control (Ctr), and 2 different hnRNP U (#1, #2) siRNAs for 72 h. The blot shown is a representative result from n > 4 independent experiments. b Quantification of the protein levels in (a) normalized to β-Actin and plotted relatively to si Ctr condition. Data are presented as mean values ± SEM of n = 4 independent experiments, *P < 0.05, **P < 0.01, ***P < 0.001, ns: non-significant (two-way ANOVA). c Quantification of the Oxygen Consumption Rate (OCR) per cell, measured by a Seahorse assay, in HCT116 cells treated with control (si Ctr), hnRNP U (si hnRNP U), DDX3X (si DDX3X) or a combination of hnRNP U and DDX3X (si hnRNP U + si DDX3X) siRNAs for 48 h. Data are presented as mean values ± SEM of n = 3 independent experiments. d Quantification of the Oxygen Consumption Rate (OCR) linked to basal and maximal mitochondrial respiration, ATP turnover, proton leak, and spare respiratory capacity, measured by a Seahorse assay, in HCT116 treated with si Ctr, si hnRNP U, si DDX3X or a combination of si hnRNP U + si DDX3X for 48 h. Data are presented as mean values ± SEM of n = 3 independent experiments, *P < 0.05, **P < 0.01 (two-sided paired t-test). e Quantification of the ECAR linked to non-glycolytic acidification, glycolytic capacity and glycolytic reserve, measured by a Seahorse assay (glycolysis stress test), in HCT116 treated with si Ctr and si hnRNP U for 48 h. Data are presented as mean values ± SEM of n = 3 independent experiments, P-value = 0.0316 for the non-glycolytic acidification (one-sided paired t-test). f Proliferation assay in HCT116 cells treated with si Ctr, si hnRNP U#1, si hnRNP U#2 for 48 h. Data are presented as mean values ± SEM of n = 3 independent experiments, *P < 0.05, **P < 0.01 (two-sided paired t-test). For all the panels, Source Data is provided as a Source Data file indicating exact P-values for (b, d, f).

Fig. 7. hnRNP U cooperates with DDX3X and GRSF1 to regulate RG4-dependent translation.

a Immunoprecipitation (IP) with the Flag antibody of protein from cytoplasmic cell extracts (CE) of HCT116 cells transfected with flag-tagged unmodified hnRNP U (WT), hnRNP U RBD deleted (ΔRBD), or empty vector (e.v), followed by western blot analysis. b RNA affinity chromatography using the RG4 sequences either native (WT), 7-deaza-modified (7dG) or mutated (Mut) and CE of HCT116 cells transfected with plasmids encoding for hnRNP U-Flag WT or ΔRBD, followed by western blot analysis. c IP with the DDX3X antibody of cellulo RNA-protein complexes (RIP) in CE of HCT116 treated with control (Ctr) or hnRNP U (U) siRNAs, followed by RT-qPCR analysis. Fold changes are indicated in italic. *P < 0.05, **P < 0.01, ***P < 0.001 (one-sided paired t-test). d RNA affinity chromatography using the RG4 sequences either WT, 7dG, and CE of HCT116 treated control (si Ctr) or DDX3X (si DDX3X) siRNAs, followed by western blot analysis. e IP with the GRSF1 antibody of CE of HCT116 treated or not with benzonase, followed by western blot analysis. Shown is a representative result from n = 3 independent experiments. f Western blot analysis in HCT116 cells treated with si Ctr, si hnRNP U, or si DDX3X for 48 h and quantification of the protein levels normalized to β-Actin. P-value = 0.0067 for hnRNP U protein in si hnRNP U condition (two-sided paired t-test). g Western blot analysis in HCT116 cells treated with si Ctr, si hnRNP U, si DDX3X or both si hnRNP U + si DDX3X for 72 h. h Quantification of the protein levels in (g) normalized to β-Actin. *P < 0.05, **P < 0.01, (two-sided paired t-test). In (a, b, c, d, f, g, h), source data are provided as a Source Data file indicating exact P-values for (c, h). Data are presented as mean values ± SEM of n = 3 for (f, h), of n = 4 in duplicates for (c) independent experiments. In (a, b, d, f, g), the blot shown is a representative result from n = 3 independent experiments. When indicated, ns non-significant.

Fig. 8. Model for the role of hnRNP U-RG4 interactions in regulating OMM-localized mRNA translation of nuclear-encoded mitochondrial mRNAs.

Following the “bind, unfold, lock” model previously proposed, hnRNP U in complex with DDX3X unfolds RG4s and indirectly recruits GRSF1 onto G-rich sequences. This molecular mechanism, which is regulated by mTOR, activates OMM-localized mRNA translation of nuclear-encoded mitochondrial mRNAs carrying RG4s in their 3’UTR, resulting in mitochondrial protein expression. Ultimately, this modulates mitochondrial functions, increasing energy metabolism and consequently cancer cell proliferation.