Fig 2: SiRNA mediated knockdown of EXD2 in U2OS cells results in a more widely spread mitochondrial network. In order to understand possible effects of EXD2 knockdown on mitochondrial network behaviour we labelled cells with EdU to identify cells in S-phase, in order to get a handle on mosaicism resulting from cell-cycle mediated effects on mitochondrial network dynamics (see main text). For this, cells were incubated for 30 min with EdU following transfection and 48–60 hrs incubation with either a pool of non-targeting control siRNAs (Csi) or a pool of three EXD2 stealth siRNAs (EXD2si3x). Following labelling for EdU incorporation (green, gr), cells were further labelled with antibodies for Tomm20 (red, r) and EXD2 (white). Control siRNA treated cells show frequent and strong perinuclear clustering of the mitochondrial network in particular in EdU positive cells. In contrast, the EXD2si3x typically show a more distributed and sometimes more hyperfused network. Two fields of view are shown for each condition.

Fig 3: EXD2 is accessible to added antibody in immunofluorescence without mitochondrial lysis. Immunofluorescent detection following paraformaldehyde fixation requires mitochondrial lysis using for example Triton X100. In the absence of this lysis step mitochondrial matrix proteins and for example mtDNA are not detectable. Thus (panel a) results shows that in the absence of TX100 lysis, EXD2 is detected while neither MRPL12 nor mtDNA are detected by IF. With TX100 lysis, all three are detected. A similar experiment (panel b) shows that both Tomm20 and EXD2, but not mtDNA are detected in the absence of TX100 lysis. A high resolution 20 × 30 µM subsection of a cell using the EXD2 and MRPL12 antibodies illustrates that the EXD2 signal is often enveloping the MRPL12 signal (panel c: some examples are indicated by a white arrow in the merged image), further illustrating EXD2 its outer-membrane localization.

Fig 4: Loss of EXD2 Leads to Mitotic Abnormalities Associated with Under-Replicated DNA(A) Quantification of the HeLa WT and EXD2-/- anaphase or telophase cells showing DAPI-positive bridges (mean ± SEM, n = 3 independent experiments, chi-square). Scale bar, 10 µm.(B) Quantification of HeLa WT and EXD2-/- G1 cells with 53BP1 OPT domains in G1 cells (left panel). Quantification of the number of 53BP1 OPT domains per positive cell in HeLa WT and EXD2-/- cells (right panel) and representative images (mean ± SEM, n = 3 independent experiments, chi-square). Scale bar, 20 µm.(C) Quantification of HeLa WT and EXD2-/- cells showing MN and representative images. Phalloidin acts as a cytosolic marker (mean ± SEM, n = 3 independent experiments, chi-square). Scale bar, 20 µm.(D) Quantification of HeLa WT, EXD2-/-, and EXD2-/- cells complemented with either Flag-EXD2 WT or Flag-EXD2 nuclease dead (ND) mutant protein for anaphase or telophase cells showing DAPI-positive bridges (mean ± SEM, n = 3 independent experiments, chi-square).(E) Quantification of HeLa WT, EXD2-/-, and EXD2-/- cells complemented with either Flag-EXD2 WT or Flag-EXD2 nuclease dead (ND) mutant protein for G1 cells with 53BP1 OPT domains (mean ± SEM, n = 3 independent experiments, chi-square).(F) Quantification of HeLa WT, EXD2-/-, and EXD2-/- cells complemented with either Flag-EXD2 WT or Flag-EXD2 nuclease dead (ND) mutant protein for cells showing MN (mean ± SEM, n = 3 independent experiments, chi-square).

Fig 5: EXD2’s Nuclease Activity Is Required to Suppress Replication Fork Collapse(A) Quantification of the frequency of 53BP1 foci in HeLa WT and EXD2-/- S/G2 cells and representative images. Cyclin A (green) acts as a marker for S/G2 cells, DAPI acts as a nuclear stain (mean ± SEM, n = 3 independent experiments, Mann-Whitney). Scale bar, 10 µm.(B) Quantification of the frequency of chromosomal aberrations from mitotic spreads from HeLa WT and EXD2-/- cells (mean ± SEM, n = 75 metaphase spreads pooled from 3 independent experiments, t test).(C) Representative images of metaphase spreads from B). Arrows indicate chromatid breaks. Scale bar, 6.5 µm.(D) Boxplot of CldU tract length ratios of associated sister forks from HeLa WT, EXD2-/-, and EXD2-/- cells complemented with either Flag-EXD2 WT or Flag-EXD2 nuclease dead (ND) mutant protein (5–95 percentile, n = 60 sister fork pairs pooled from 3 independent experiments, Mann-Whitney).