Fig 2: CLUH controls MiD49 and Mff protein levels through translational regulation in mammalian cells.a–d Immunofluorescence (IF) staining showing endogenous MiD49 and Mff protein levels in wild-type and CLUH KO cells, using anti-MiD49 and anti-Mff antibodies, respectively. e Quantification of IF signals of MiD49 and Mff, normalized to IF signals of a control mitochondrial protein TOM20 (images not shown) in the same cell, using Fiji/ImageJ. f, g Western blots showing Mff and MiD49 protein levels in response to CLUH overexpression or CLUH KO. VDAC was used as a loading control for mitochondrial mass, and actin was used as a housekeeping protein loading control. Western blot experiments were repeated independently 5 times, with representative images shown in (f) and statistical analysis shown in (g) (mean ± SEM, n = 5). h Quantitative PCR (qPCR) experiments showing Drp1, Mff, and MiD49 mRNA levels in response to CLUH overexpression or CLUH KO. Total RNAs extracted from HeLa cells of the indicated genotypes were used as templates for the synthesis of cDNAs, which were then used as templates for qPCR. Experiments were repeated in triplicate (mean ± SEM, n = 3). i, j RNA Immunoprecipitation (RIP) experiments show that CLUH binds mRNAs of MiD49 and Mff, but not mRNAs of control genes TOM20 or TBP (encoding the nuclear TATA-binding protein). Experiments were repeated in triplicate (mean ± SEM, n = 3), and representative images are shown in (j). k–n’ Puro-PLA was carried out using a mouse anti-puromycin antibody in combination with either a rabbit anti-MiD49 antibody (k–l’) or a rabbit anti-Mff antibody (m–n’) (red). Mitochondria: goat anti-Hsp60 antibody (green). Nucleus: DAPI (blue). o Quantification of relative Puro-PLA signals using Fiji/ImageJ. p Pearson’s coefficient indicates colocalization of Puro-PLA signals and mitochondria. a–d, k–n’ Experiments were repeated independently for 4 times, and representative images are shown (mean ± SEM, n = 4). In each experiment, 20–30 cells were analyzed for each genotype, and the average IF signal intensity (e, o) or Pearson’s coefficient (p) was calculated. Scale bars: 10 µM. e, i, o, p Two-sided Student’s t-test. g, h One-way ANOVA with post hoc Tukey’s HSD test, with P-values displayed in the graphs.

Fig 3: Models for how clu/CLUH promotes mitochondrial fission through regulating Drp1, and how functions of clu-drp1 interact with those of PINK1-parkin-mfn.a In wild-type cells, Drosophila Clueless and mammalian CLUH promote mitochondrial fission through regulating the recruitment of Drp1 from the cytosol to the OMM. In mammalian cells, this is accomplished by CLUH directly binding both mRNA and protein for Drp1 receptors Mff and MiD49. Moreover, CLUH regulates Mff and MiD49 protein levels through translational control. In Drosophila and mammalian cells, PINK1 and Parkin inhibit mitochondrial fusion by promoting degradation of Mfn proteins (Drosophila Marf, mammalian homologs Mfn1 and Mfn2). Together, the two pathways, clu-drp1 and PINK1-parkin-mfn, balance the opposing actions of mitochondrial fission and fusion and maintain mitochondrial integrity and tissue health. b Loss of clueless/CLUH results in decreased levels of Drp1 receptor proteins due to dysregulated translation. This further leads to decreased recruitment of Drp1 onto the OMM, reduced mitochondrial fission, mitochondrial dysfunction, and tissue damage.

Fig 4: CLUH directly binds MiD49 and Mff, which mediates the interactions between CLUH and Drp1.a–d Co-immunoprecipitation using lysates from HeLa cells transfected with the indicated constructs. The INPUT represents 5% (a) and 3% (c) of each total lysate, respectively. Both MiD49-Myc (a, b) and GFP-Mff (c, d) were co-immunoprecipitated with FLAG-CLUH by using a mouse anti-FLAG antibody. Western blots were probed with rabbit anti-FLAG, Myc, GFP, and actin antibodies. The levels of co-immunoprecipitated proteins were normalized to INPUT levels (mean ± SEM, n = 3 independent experiments, two-sided Student’s t-test). e In a cell-free system with recombinant GST- or His-fusions purified from E. coli, MiD49 pulls down both Drp1 and CLUH in vitro. f With purified Mff (isoform-e), CLUH, and Drp1 as INPUT, Mff pulls down both Drp1 and CLUH in vitro. e, f Gray dashed lines delineate the boundaries between non-adjacent lanes in the same gel. The whole gel images are shown in Supplementary Fig. 5b, c. Double asterisks indicate the position of a non-specific band. IB denotes Insert B domain. TM denotes transmembrane domain. All the above experiments were performed without the use of cross-linkers. g Drp1 does not pull down CLUH in vitro. e–g Experiments were repeated independently 3 times, and representative images are shown. h–i’ CLUH-Drp1-PLA (red) was carried out using rabbit anti-CLUH and mouse anti-Drp1 antibodies in HeLa cells transfected with control or MiD49 siRNA. Mitochondria: goat anti-Hsp60 antibody (green). Nucleus: DAPI (blue). j Quantification of CLUH-Drp1-PLA signals. k–m” FLAG-tagged Mff, MiD49 and CLUH fused to LAMP1 are visualized with a rabbit anti-FLAG antibody (red), and Drp1 is visualized with a mouse anti-Drp1 antibody (green). White arrowheads indicate lysosome-localized proteins. n Pearson’s coefficient indicates colocalization of the indicated LAMP1 fusion proteins and Drp1. h–i’, k–m” White dashed lines mark the boundaries of the cells and the nuclei. Scale bars: 10 µM. j, n Fiji/ImageJ was used for image analysis (mean ± SEM, n = 3 independent experiments, one-way ANOVA with post hoc Tukey’s HSD test). In each experiment, 20–30 cells were analyzed for each genotype and the average PLA signal intensities (j) or Pearson’s coefficient (n) was calculated.

Fig 5: CLUH regulates mitochondrial morphology in mammalian cells.a Generation of two CLUH knockout (KO) HeLa cell lines, CLUH-40 and CLUH-43, using the CRISPR–Cas9 system52,53. Positions of the deletions in the CLUH protein (red asterisk) and in the CLUH gene (within the second exon, red dashed lines) are shown. b Western blots using a rabbit anti-CLUH antibody confirm the absence of CLUH in both CLUH KO cell lines. Experiments were repeated independently 5 times, and representative images are shown. c–f Confocal microscopy images showing mitochondrial morphology in HeLa cells of the indicated genotypes under resting conditions. Yellow dashed lines mark the boundaries of cells. c’–f’ Mitochondrial morphology after nocodazole treatment. c”–f” Enlarged views of the boxed regions in (c’-f’). c–f” Mitochondria are labeled with a mouse anti-TOM20 antibody. g Mitochondrial morphology in (c–f”) is quantified using the mitochondrial morphology scoring assay, and one representative image for each category of cells is shown above the bar graph. h–j’ Confocal microscopy images showing mitochondrial morphology in response to CLUH RNAi or CLUH overexpression. Mitochondria are labeled by transfection of a plasmid expressing Mito-DsRed (red), and the nuclei are labeled with DAPI (blue). h’–j’ Enlarged views of the boxed regions in (h–j). k Mitochondrial morphology in (h–j’) is quantified using the scoring assay. l–o Expression of Drp1S637A (n), but not Drp1WT (o), rescues mitochondrial clustering in CLUH KO HeLa cells. Mitochondria are labeled with a mouse anti-TOM20 antibody. Yellow dashed lines mark the boundaries of cells. p Quantification of the percentage of cells showing clustered mitochondria for each indicated genotype. c–f”, h–j’, l–o Experiments were performed in triplicate, and representative images are shown. Scale bars: 10 µM. g, k, p In the mitochondrial morphology scoring assay, numbers of cells in each of the four categories were counted, with a total number of >200 cells in each experiment (mean ± SEM, n = 3 independent experiments). Statistical analysis was carried out using one-way ANOVA with post hoc Tukey’s HSD test, with P-values displayed in the graphs. P < 0.05: significantly different from the same morphology category in the control group.