Mitochondrial double-stranded RNA homeostasis depends on cell-cycle progression

(A) Immunofluorescence of dsRNA, BrU, and DNA in the indicated fibroblast cell lines treated with BrU for 1 h and imaged by confocal microscopy. DAPI staining shown in blue (scale bar: 20 μm). (B) Number of mitochondrial dsRNA foci, BrU foci, and nucleoids per μm² of mitochondria determined by immunofluorescence and confocal microscopy. The mitochondrial network was imaged using anti-TOMM20 antibody. Box and whiskers plot represent the number of foci per μm² mitochondria for each cell. Whiskers represent minimum and maximum values. Boxes extend from the 25^th to the 75^th percentile with the median plotted in the middle. “+” indicates the mean value (N = 30 cells from two independent cultures). (C) Northern blot of mRNA transcripts of the coding and mirror regions of CYTB and ND5 in the fibroblast cell lines (top). Nuclear 7SL RNA was used as a loading control. Schematic representing the coding and mirror regions that were probed for CYTB and ND5 (bottom). Coding genes: green arrows. Coding regions for tRNAs: pink arrows. (D) Immunofluorescence of dsRNA in WT fibroblasts treated with the indicated siRNA for 48 h and imaged by confocal microscopy. The mitochondrial network was immunostained with anti-TOMM20. DAPI staining shown in blue (scale bar: 20 μm). (E) Immunofluorescence of dsRNA in WT, NME6 KO, and NME6 KO HeLa cells expressing NME6-MycFlag or NME6^H137N-MycFlag and imaged by confocal microscopy. DAPI staining shown in blue (scale bar: 20 μm). (E, F) Scatter plot of mean mt-dsRNA intensity per cell quantified from confocal images as shown in (E) (N = 100–120 cells from two independent cultures). The one-way ANOVA test was used to determine P-values compared with WT. P-value (WT versus NME6 KO) < 0.0001, P-value (WT versus NME6 KO + WT) = 0.0013, P-value (WT versus NME6 KO + H137N) < 0.0001. Horizontal lines indicate the mean value and error bars indicate the SD. (G) Immunofluorescence of dsRNA in WT and NME6 KO HeLa cells incubated with 100 μM nucleosides for 5 d and imaged by confocal microscopy. DAPI staining shown in blue (scale bar: 20 μm). (G, H) Scatter plot of mean mt-dsRNA intensity per cell quantified from confocal images as shown in (G) (N = 100–120 cells from two independent cultures). The Mann-Whitney t test was used to determine the P-value between NME6 KO versus NME6 KO + nuc. P-value < 0.0001. Horizontal lines indicate the mean value and error bars indicate the SD.

Source data are available for this figure.

(A) Immunofluorescence of dsRNA in WT HeLa and NME6 KO cells after double thymidine block (DTB) treatment and subsequent release for 2 or 4 h imaged by confocal microscopy. DAPI staining is shown in blue (scale bar: 20 μm). (A, B) Scatter plot of mean dsRNA intensity per cell quantified from confocal images as shown in (A) (N = 100–110 cells from two independent cultures). A one-way ANOVA test was used to determine P-values compared with untreated samples in each cell line. P-value (WT versus WT-DTB) < 0.0001, P-value (WT versus WT −2 h) < 0.0001, P-value (WT versus WT-4 h) = 0.008, P-value (KO versus KO-DTB) > 0.9999, P-value (KO versus KO-2 h) = 0.1011, P-value (KO versus KO-4 h) < 0.0001. Horizontal lines indicate the mean value and error bars indicate the SD. (C) Unsynchronised cell populations of WT HeLa and NME6 KO imaged by confocal microscopy. Cells in G1 (blue), S (magenta), and G2 (green) phases are delineated according to DAPI, EdU, and cyclin-A detection, respectively (scale bar: 20 μm). (C, D) Scatter plot of mt-dsRNA integrated intensity per cell quantified from confocal images as shown in (C) (N = 100–110 cells from two independent cultures). A one-way ANOVA test was used to determine P-values compared with WT and KO for cell-cycle phases in each cell line. P-value (WT G1 versus WT S) < 0.0001, P-value (WT G1 versus WT G2) < 0.0001, P-value (KO G1 versus KO S) < 0.0001, P-value (KO G1 versus KO G2) <0.0001. Bars indicate the mean and the error bars indicate the SD. (C, E) Scatter plot of the cytoplasmic area per cell quantified from confocal images as shown in (C) (N = 100–110 cells from two independent cultures). A one-way ANOVA test was used to determine P-values compared with WT and KO for cell-cycle phases in each cell line. P-value (WT G1 versus WT S) < 0.0001, P-value (WT G1 versus WT G2) < 0.0001, P-value (KO G1 versus KO S) < 0.0001, P-value (KO G1 versus KO G2) <0.0001. Bars indicate the mean and the error bars indicate the SD.

(A) hTERT fibroblasts were treated with basic FGF (10 ng/ml), aphidicolin (Aph; 6 μM), or a combination of both. Cell proliferation was measured by confluency (N = 3 independent cultures; points indicate mean and error bars represent SD). (B) Immunofluorescence of dsRNA in hTERT fibroblasts treated with FGF/Aph for 48 h and BrU for 1 h and imaged by confocal microscopy. The mitochondrial network was labelled with MitoTracker DeepRed (MitoT; magenta). DAPI staining shown in blue (scale bar: 20 μm). (C) Scatter plot of mean BrU intensity per cell quantified from confocal images as shown in (Fig S4A) (N = 92–111 cells from two independent cultures). A one-way ANOVA test was used to determine P-values compared with DMSO. P-value (DMSO versus Aph) = 0.0130, P-value (DMSO versus FGF) < 0.0001, P-value (DMSO versus FGF + Aph) < 0.0001. Horizontal lines indicate the mean value and error bars indicate the SD. (D) Scatter plot of mean mt-dsRNA intensity per cell quantified from confocal images as shown in (B) (N = 89–111 cells from two independent cultures). A one-way ANOVA test was used to determine P-values compared to DMSO. P-value (DMSO versus Aph) = 0.2653, P-value (DMSO versus FGF) < 0.0001, P-value (DMSO versus FGF + Aph) = 0.0735. Horizontal lines indicate the mean value and error bars indicate the SD. (E) Heatmap showing relative RNA expression of heavy and light mitochondrial transcripts from hTERT fibroblasts treated with FGF/Aph for 48 h as measured by strand-specific RT-qPCR. Fold changes versus mean DMSO values for each transcript are shown. C_q values were normalised against GAPDH (N = 3 independent cultures). One-way ANOVA analysis was performed for each transcript among the four conditions and the resulting P-value is indicated below. (F) mtDNA levels of hTERT fibroblasts treated with FGF/Aph for 48 h as measured by qPCR of the indicated mtDNA regions, MT-RNR1 and MT-CYTB. C_q values were normalised against nuclear ACTB (N = 3 independent cultures). Bars indicate the mean and the error bars represent the SD. Individual data points are shown as grey circles. (G) Immunoblot analysis of mitochondrial and MRG-associated proteins in whole-cell lysates obtained from hTERT fibroblasts treated with FGF/Aph for 48 h. Samples were blotted on different membranes which are indicated by the dotted lines.

Source data are available for this figure.

(A) Immunofluorescence of mt-dsRNA with anti-J2 and anti-9D5 antibodies in U2OS cells imaged by confocal microscopy. The mitochondrial network was labelled with MitoTracker (MitoT) DeepRed (scale bar: 20 μm). The region indicated by a white dotted box on the confocal image was re-imaged with greater resolution with the mitochondrial area outlined (scale bar: 5 μm). The mitochondrial region was straightened as a line scan with the intensity of J2 and 9D5 plotted below. (B) Immunofluorescence of mt-dsRNA with anti-J2 and anti-9D5 antibodies in U2OS cells treated with either IMT1 (10 μM for 3 h) or RNAse III after fixation. The mitochondrial network was labelled with MitoTracker (MitoT) DeepRed (scale bar: 20 μm; zoom scale bar: 5 μm). (C) An FFPE section of human colorectal adenocarcinoma was segmented into normal (blue), dysplastic (red), and stroma (yellow, hatched) regions. Subsequent sections of the sample were incubated with anti-9D5 with and without RNase III treatment or anti-ATP5A and visualised with DAB staining (brown). Haematoxylin was used as a counterstain (scale bar 0.5 mm). (D) An FFPE section of normal human lung tissue (top row; scale bar 1 mm) and human lung adenocarcinoma (bottom row; scale bar 2 mm) was segmented into normal (blue), adenocarcinoma (red), necrotic (pink), and stroma (yellow, hatched) regions. Subsequent sections of the sample were incubated with anti-9D5 with and without RNase III treatment or anti-ATP5A and visualised with DAB staining (brown). Haematoxylin was used as a counterstain.

(A) Lung adenocarcinoma tissue microarrays (TMAs) consisting of three biopsy cores from 80 patients each were used for multiplex staining of cytokeratin, dsRNA, ATP5A, and DAPI. The immunostaining was used to annotate and train a deep learning algorithm to identify regions of interest (ROIs) as “tumour,” “stroma,” and “necrosis” shown in pink, yellow, and blue overlays, respectively. Individual cells were segmented within the ROIs. (A, B) A representative core from lung adenocarcinoma TMA showing the ROI overlay as annotated in (A). Individual staining with anti-cytokeratin, anti-dsRNA, anti-ATP5A, and DAPI shown in grayscale (scale bar: 200 μm). (C) Representative core from lung adenocarcinoma TMA showing staining with anti-dsRNA (green), anti-ATP5A (magenta), and DAPI (blue) (scale bar: 200 μm). Regions of tumour epithelia are indicated by white boxes which are zoomed in on the right (scale bar: 5 μm). (D) Box and whiskers plot of median cytoplasmic dsRNA intensity in stroma and tumour regions per core. Whiskers represent minimum and maximum values. Boxes extend from the 25^th to the 75^th percentile with the median plotted in the middle (N = 223 cores detected with stroma regions, N = 221 cores detected with tumour regions). The Mann-Whitney t test was used to determine the P-value between stroma versus tumour; P-value < 0.0001 (E) Density plot of the distribution of cytoplasmic dsRNA intensity per segmented cell in stroma and tumour ROIs (N = 686,546 cells detected within stroma ROIs, N = 761,319 cells detected within tumour ROIs).