A Compendium of Genetic Modifiers of Mitochondrial Dysfunction Reveals Intra-organelle Buffering - PubMed (original) (raw)

. 2019 Nov 14;179(5):1222-1238.e17.

doi: 10.1016/j.cell.2019.10.032.

Alejandro M Cuadros 1, Hardik Shah 2, Wendy H W Hung 1, Yang Li 2, Sharon H Kim 2, Daniel H F Rubin 2, Ryan H Boe 1, Sneha Rath 2, John K Eaton 1, Federica Piccioni 1, Amy Goodale 1, Zohra Kalani 1, John G Doench 1, David E Root 1, Stuart L Schreiber 3, Scott B Vafai 1, Vamsi K Mootha 4

Affiliations

A Compendium of Genetic Modifiers of Mitochondrial Dysfunction Reveals Intra-organelle Buffering

Tsz-Leung To et al. Cell. 2019.

Abstract

Mitochondrial dysfunction is associated with a spectrum of human conditions, ranging from rare, inborn errors of metabolism to the aging process. To identify pathways that modify mitochondrial dysfunction, we performed genome-wide CRISPR screens in the presence of small-molecule mitochondrial inhibitors. We report a compendium of chemical-genetic interactions involving 191 distinct genetic modifiers, including 38 that are synthetic sick/lethal and 63 that are suppressors. Genes involved in glycolysis (PFKP), pentose phosphate pathway (G6PD), and defense against lipid peroxidation (GPX4) scored high as synthetic sick/lethal. A surprisingly large fraction of suppressors are pathway intrinsic and encode mitochondrial proteins. A striking example of such "intra-organelle" buffering is the alleviation of a chemical defect in complex V by simultaneous inhibition of complex I, which benefits cells by rebalancing redox cofactors, increasing reductive carboxylation, and promoting glycolysis. Perhaps paradoxically, certain forms of mitochondrial dysfunction may best be buffered with "second site" inhibitors to the organelle.

Keywords: CRISPR screening; G6PD; GPX4; LARP1; complex I; genetic modifier; metformin; mitochondria; redox cofactors; reductive carboxylation.

Published by Elsevier Inc.

PubMed Disclaimer

Conflict of interest statement

DECLARATION OF INTERESTS

V.K.M. is a paid advisor to Raze Therapeutics, Janssen Pharmaceuticals, and 5AM Ventures, S.B.V. is currently an employee of Sanofi Pharmaceuticals. T.-L.T., S.B.V., and V.K.M. are listed as inventors on a provisional patent application submitted by the Broad Institute on technology described in this paper.

Figures

Figure 1.

Figure 1.. Genome-Wide CRISPR Screens to Identify Modifiers of Mitochondrial Dysfunction

A. Mitochondrial chemical inhibitors used in this study: piericidin (complex I), antimycin (complex III), oligomycin (complex V), antimycin+oligomycin (ΔΨm), ethidium bromide (mitochondrial DNA replication), chloramphenicol (mitochondrial translation), and metformin (complex I is among the proposed targets). B. Schematic overview of the genome-wide CRISPR screens in K562 cells. The following drug dosages were used: 0.1% DMSO, 10 mM metformin (Met), 10 nM piericidin (Pier), 100 nM antimycin (Anti), 10 nM oligomycin (Oligo), 10 nM antimycin + 10 nM oligomycin (AO), 100 ng/mL ethidium bromide (EtBr), and 10 μg/mL chloramphenicol (CAP). C. Growth curves for cumulative differences in growth under drug treatments. Growth curves for individual replicates over 15 days are shown. Gray arrows denote time points at which samples were harvested. D. Categories of genetic interactions and the number of genetic modifiers associated with each category. Scatter plots of Z-scores showing knockouts that are enriched or depleted in each drug vs. DMSO. Knockouts are scored by Δ_Z_Drug = (_Z_Drug−_Z_DMSO)/√2 (see Methods). Knockout enrichment (shown in red or magenta, Δ_Z_Drug > 4.8) and depletion (shown in blue, Δ_Z_Drug < −2.4) are used to define buffering and synthetic sick/lethal interactions, respectively. Among buffering interactions, suppressors (_Z_Drug > 2.4) are shown in red. Known genetic modifiers (vHL and ATPIF1 as genetic suppressors of OXPHOS inhibitors, and GOT1 loss as aggravator) are highlighted. The gray dotted lines represent the cutoffs for interactions.

Figure 2.

Figure 2.. Genetic Modifiers of Distinct Modes of Mitochondrial Dysfunction

A. The 38 synthetic sick/lethal hits that score in ≥ 1 drug (Δ_Z_Drug < −2.4). B. The 63 suppressors that score in ≥ 1 drug (Δ_Z_Drug > 4.8, and_Z_Drug > 2.4) For (A) and (B), genes are further divided into two panels. The left panel contains genes that have interactions with multiple drugs, and is ordered by the number of drugs under which interactions occur. The right panel contains genes with only drug-specific interactions, and genes are listed by the specific drug and ordered by the strength of interaction. Gene-drug pairs that do not score are colored white (−2.4 ≤ D_Z_Drug≤ 4.8). C. Counts of modifiers that are implicated in intra-mitochondrial vs. extra-mitochondrial interactions. The expected fraction of “intra-mitochondrial” in each category is 6%, as only 6% of all nuclear genes encode mitochondrial proteins.

Figure 3.

Figure 3.. Losses of Genes Involved in Glycolysis or Pentose Phosphate Pathway are Synthetic Sick/Lethal with OXPHOS Dysfunction

A. Top categories of KEGG pathways among depleted knockouts by Gene Set Enrichment Analysis (GSEA). The 11,102 genes was ranked by Δ_Z_Drug in descending order and GSEA was run in the negative mode. Pathways are ordered by the maximum of −log10 FDR across 7 drugs. B. Growth phenotypes of K562, A375, or HT-29 cells with sgRNAs directed against the genes indicated under piericidin or antimycin treatment. Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). *p<0.05, **p<0.01 or ****p<0.0001 indicates two-tailed Student’s _t_-test p-value.

Figure 4.

Figure 4.. Aggravation of Oligomycin’s Toxicity by Loss of GPX4

A. Venn diagram showing the overlap between the 38 synthetic sick/lethal hits and the 100 most upregulated proteins in a previous proteomic profiling study in cells depleted of mtDNA (Bao et al., 2016). B. Growth phenotypes of wild-type cells under a combination of oligomycin and GPX4 inhibition by 10 μM JKE-1674 for K562 in spent medium, HAP1 in fresh medium, and HeLa in fresh medium. Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). C. Growth phenotypes of GPX4 KO cells in HAP1 under oligomycin treatment in fresh medium supplemented with ferroptosis inhibitors alpha-tocopherol (aTOC, 1 μM) or ferrostatin-1 (Fer-1, 1 μM), or pan-caspase inhibitor zVAD-fmk (1 μM). Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). D. Growth phenotypes of GPX4 KO cells in HAP1 under oligomycin with re-expression of, lGPX4, or lGPX4 supplemented with the catalytic site-specific GPX4 inhibitors JKE-1674 (10 μM) or ML 210 (10 μM) in fresh medium. Cell counts were performed 3 days post treatment (average +/− SEM, n = 2). E. Immunoblots for GPX4 and loading control in wild-type K562, HeLa and HAP1 cells treated with antimycin or oligomycin for 3 days. A representative experiment is shown. F. Mitochondrial GPX4 levels derived from a proteomics dataset mitochondrial proteome from the heart tissues of five conditional knockout mouse strains with OXPHOS dysfunction (Kuhl et al., 2017). The data shown represent the average +/− SEM (n≥3). In (B)-(D) and (F), *p<0.05, **p<0.01 or ***p<0.001 indicates two-tailed Student’s _t_-test p-value.

Figure 5.

Figure 5.. Loss of Cytosolic Protein LARP1 is Buffered under OXPHOS Inhibition

A. Scatter plots of Z-scores highlighting knockout of LARP1 in oligomycin (top panel) or ethidium bromide (bottom panel) vs. DMSO. The gray dotted lines represent the cutoffs for interactions. B. Oxygen consumption as assessed by the Seahorse Extracellular Flux (XFe96) Analyzer in K562 cells with sgRNAs directed against the genes indicated (average +/− SEM, n = 6). C. Cell death as assessed by Annexin V staining and flow cytometry in K562 cells cultured in glucose or galactose containing medium with sgRNAs directed against the genes indicated (average +/− SEM, n = 3). D. Heatmap of co-dependency based on hierarchical clustering of Pearson correlations between the mean log2-fold changes across sgRNAs for genes. Disjoint modules (demarcated with red triangles) were annotated by assigned GO term that was significantly enriched in that cluster (hypergeometric p ≤ 10−10). Black ticks represent genes that are listed in the MitoCarta2.0 database (Calvo et al., 2016). The red arrow denotes LARP1. E. Immunoblots for the specified mitochondrial proteins and LARP1 in K562 cells with sgRNAs directed against the genes indicated. A representative experiment is shown. F. Real time PCR-based measurement of mtDNA relative to nuclear DNA (nDNA) in K562 cells with sgRNAs directed against the genes indicated (average +/− SEM, n = 3). G. Taqman gene expression analysis of transcripts of OXPHOS components in K562 cells with sgRNAs directed against the genes indicated (average +/− SEM, n = 3). For (C) and (F), ns p>0.05 or ***<0.001 indicates two-tailed Student’s _t_-test p-value.

Figure 6.

Figure 6.. Suppression of Oligomycin’s Toxicity by Loss of Complex I Activity

A. Scatter plot of the fitness of a gene knockout in oligomycin (_Z_Oligo) vs. its fitness in DMSO (_Z_DMSO). Gray dots denote all 11,102 genes in the analysis and the red dots denote 50 nuclear genes that encode structural subunits and assembly factors of complex I. B. Validation of NDUFA9 loss, piericidin, and metformin as suppressors of oligomycin in HAP1, 293T, HT-29, HeLa, and K562 cells. Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). C. Validation of piericidin as suppressors of oligomycin in K562 in the presence (+Uridine) or absence (-Uridine) of 50 mg/mL uridine supplementation. Dialyzed serum was used for -Uridine growth. Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). In (B) and (C), *p<0.05 or **p<0.01 indicates two-tailed Student’s_t_-test p-value. D. Oxygen consumption as assessed by the Seahorse Extracellular Flux (XFe96) Analyzer in K562 cells after 3 days of drug treatment (average +/− SEM, n = 12). E. Corrected extracellular acidification rate as assessed by the Seahorse Extracellular Flux (XFe96) Analyzer in K562 after 3 days of drug treatment (average +/− SEM, n = 12). F. Media lactate levels in K562 using targeted mass spectrometry after 3 days of drug treatment (average +/− SEM, n = 3). *p<0.05 indicates two-tailed Student’s _t_-test p-value. G. Volcano plot comparing piericidin+oligomycin to oligomycin from full-scan metabolomics in K562 using mass spectrometry after 3 days of drug treatment (average of n = 3). Highlighting in red indicates *p<0.05 two-tailed Student’s _t_-test p-value. H. Schematics for stable isotope labeling with [U-13C]glutamine to monitor 13C incorporation in key metabolites. I. Total metabolite pools and 13C labeling pattern of citrate, 2-hydroxyglutarate, and proline over the course of 8 hours upon tracer analysis with [U-13C]glutamine in K562 cells. Cells were pretreated in drugs in uridine-free medium 48 hours prior to [U-13C]glutamine supplementation (average +/− SEM, n = 3). *p<0.05 or **p<0.01 indicates two-tailed Student’s _t_-test p-value. J. Growth phenotypes of K562 cells with sgRNAs directed against_G6PD_ under oligomycin or piericidin+oligomycin treatment. Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). **p<0.01 indicates two-tailed Student’s_t_-test p-value. K. Growth phenotypes of cells stably expressing a vector vehicle or cytosolic TPNOX under the specified treatment in K562. Cell counts were performed 3 days post drug treatment (average +/− SEM, n = 3). ns p>0.05 or **p<0.01 indicates two-tailed Student’s_t_-test p-value. Top: immunoblots for FLAG and the TOMM20 loading control for cells expressing either the vector vehicle or cytosolic TPNOX.

Comment in

References

    1. Arroyo JD, Jourdain AA, Calvo SE, Ballarano CA, Doench JG, Root DE, and Mootha VK (2016). A Genome-wide CRISPR Death Screen Identifies Genes Essential for Oxidative Phosphorylation. Cell Metab 24, 875–885. - PMC - PubMed
    1. Balaban RS, Kantor HL, Katz LA, and Briggs RW (1986). Relation between work and phosphate metabolite in the in vivo paced mammalian heart. Science 232, 1121–1123. - PubMed
    1. Bao XR, Ong SE, Goldberger O, Peng J, Sharma R, Thompson DA, Vafai SB, Cox AG, Marutani E, Ichinose F, et al. (2016). Mitochondrial dysfunction remodels one-carbon metabolism in human cells. Elife 5. - PMC - PubMed
    1. Barzilai N, Crandall JP, Kritchevsky SB, and Espeland MA (2016). Metformin as a Tool to Target Aging. Cell Metab 23, 1060–1065. - PMC - PubMed
    1. Baughman JM, Nilsson R, Gohil VM, Arlow DH, Gauhar Z, and Mootha VK (2009). A computational screen for regulators of oxidative phosphorylation implicates SLIRP in mitochondrial RNA homeostasis. PLoS Genet 5, e1000590. - PMC - PubMed

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources