Nuclear Acetyl-CoA Production by ACLY Promotes Homologous Recombination

Abstract

While maintaining the integrity of the genome and sustaining bioenergetics are both fundamental functions of the cell, potential crosstalk between metabolic and DNA repair pathways is poorly understood. Since histone acetylation plays important roles in DNA repair and is sensitive to the availability of acetyl coenzyme A (acetyl-CoA), we investigated a role for metabolic regulation of histone acetylation during the DNA damage response. In this study, we report that nuclear ATP-citrate lyase (ACLY) is phosphorylated at S455 downstream of ataxia telangiectasia mutated (ATM) and AKT following DNA damage. ACLY facilitates histone acetylation at double-strand break (DSB) sites, impairing 53BP1 localization and enabling BRCA1 recruitment and DNA repair by homologous recombination. ACLY phosphorylation and nuclear localization are necessary for its role in promoting BRCA1 recruitment. Upon PARP inhibition, ACLY silencing promotes genomic instability and cell death. Thus, the spatial and temporal control of acetyl-CoA production by ACLY participates in the mechanism of DNA repair pathway choice.

Keywords: ATP-citrate lyase; DNA damage; DNA repair; acetyl-CoA; histone acetylation; homologous recombination; metabolism; non-homologous end joining; nucleus.

Figure 1. ACLY participates in recruitment of BRCA1 to sites of DNA double strand breaks (DSBs)

(A) After transfection with indicated siRNA SMARTpools, cells were treated +/− 10 Gy IR and harvested at 1 h or 6 h post-IR, along with unirradiated controls. Acid-extracted histones and corresponding cytoplasmic fractions were analyzed by Western blot. Quantification of three independent replicates is depicted and significance determined using paired t-test following normalization to siControl UT 1h, mean +/− SEM *, p<0.05. (B) Cells were treated as in (A); cytoplasmic and nuclear fractions were prepared and analyzed by Western blot. (C) Control and ACLY-silenced cells were treated with 2 Gy IR, and 4 h later immunofluorescence imaging was performed for BRCA1, 53BP1, or γH2AX, scale bar- 200 M. Representative experiment from n=2 per cell line and condition with quantification of 10 fields/experiment, mean +/− SEM, *, p<0.05, **, p<0.01. (D) Control and ACLY-silenced cells were treated with 2 Gy IR imaged for cyclin A and BRCA1 after 4 h, scale bar-200 M. Representative experiment from n=2. Quantification of BRCA1 foci in cyclin A+ cells in 10 fields, mean +/−SEM, *, p<0.05. See also Figure S1.

Figure 2. ACLY promotes histone acetylation at sites of DSBs

(A) U2OS reporter cell line transfected with indicated siRNA SMARTpools was treated with Shield-1 and 4-OHT for 5 h to induce DSBs by mCherry-Lac1-Fok1. Colocalization of BRCA1 or 53BP1 with Fok1 was imaged by IF, scale, 200 M. Representative experiment from n=3 each for 53BP1 and BRCA1. Quantitation of 15–20 fields from each sample is depicted, ****, p<0.0001, n.d.- not detected (lower left panel); A.U.- Arbitrary units. Re-drawn schematic of the reporter assay, previously published (Tang et al., 2013) is depicted (lower right panel) (B) BRCA1 colocalization with Fok1 was examined in 53BP1-deficient U2OS reporter cells. (See also Figure S2B.) n.s.- not significant; A.U.- Arbitrary units. (C) AcH4 ChIP using different primer sets for endogenous loci was performed in a U2OS-TRF1-Fok1 reporter cell line that induces DSBs within telomeres (Tang et al., 2013). DSBs were induced by treatment with 4-OHT for 5 h; mean +/− SEM for 5 h, *, p<0.05, **, p<0.01, n.s.- not significant, n.d.- not detected. See Figure S3 for primer locations. (D) HR assay performed in the DR-GFP U2OS reporter line (Pierce et al., 2001). Cells were transfected with individual siRNAs for 24 hours (siACLY A or siACLY B), followed by transfection of the endonuclease Sce-1 for an additional 48 h. GFP positivity, indicating HR efficiency, was analyzed by flow cytometer, mean +/− SEM, **, p<0.01; ***, p<0.001. (E) To assay HR, cells were transduced with EV or ACLY-expressing lentivirus. Cells were transfected with siRNA SMARTpools, followed by Sce-1 as in (D). Data is representative of 2 independent experiments. Mean +/− SEM is graphed, ***, p<0.001. (F) ACLY was silenced using two independent shRNAs in HeLa cells expressing TRF2^ΔB/ΔM. NHEJ was assessed by scoring end-to-end chromosomal fusions. Mean +/− SEM **, p<0.01, ***,p<0.001. See also Figure S2 and Figure S3.

Figure 3. DNA damage signaling does not stimulate ACLY-dependent global reprogramming of glucose metabolism

(A) Metabolite consumption/production over 6 hours in U2OS Cells treated +/− 5 Gy IR, mean +/− SEM of triplicates (B) A549 cells were treated with 5 Gy IR and ³H-2-deoxyglucose uptake monitored at indicated time points, mean +/− SEM of triplicates. Cpm- counts per minute. (C) A549 cells were treated with 5Gy IR and media replaced containing [U-¹³C]-glucose for 6 h. Enrichment of TCA cycle intermediates was examined using GC-MS, mean +/− SEM of triplicates. (D) A549 cells were treated with IR and then media replaced with that containing [U-¹³C]-glucose for 24 h. Palmitate enrichment was examined by fatty-acid methyl-ester analysis by GC-MS, mean +/− SEM of triplicates. (E) U2OS cells were transduced with indicated shRNAs targeting either control or ACLY, +/− 5 mM acetate, and treated with 5 Gy IR. Metabolites were extracted and analyzed by LC-MS at indicated time points. Each point on the heat map represents the mean of triplicate samples for each condition and time point. Full panel of metabolites and biological replicates are included in Supplementary Table 1. See also Figure S4.

Figure 4. Nuclear ACLY is phosphorylated following DNA damage in a cell cycle dependent manner

(A) HeLa cells subjected to 5 Gy IR were fractionated at indicated time points. Nuclear and cytoplasmic levels of indicated proteins determined by Western blot. (B) Subcutaneous B16 tumors were grown on both flanks of female C57Bl/6 mice. One tumor on each mouse was treated with 20 Gy IR, while the other tumor received no treatment. Both tumors were harvested in parallel and immunohistochemistry was performed for pACLY. The number of positive nuclei/field were scored for 20 fields per section, mean +/−SEM, ***, p<0.001; ****, p<0.0001. Representative images and quantification from one mouse per time point (0.5 and 1 h) is depicted. Reproducible data was observed from a total of n=3 mice for 0.5 h time point and a total of n=2 mice for the 1 h time point. (C) HeLa cells were pre-treated with an ATM inhibitor for 15 min, then exposed to IR, fractionated, and analyzed by Western blot. (D) A549 cells were pre-treated with an AKT inhibitor for 1 h, followed by IR treatment, fractionation, and western blotting of lysates. (E) HeLa cells were synchronized to S-phase using a double thymidine block and released into G2 and G1 at 4 and 10 h, respectively. 30 min prior to harvest in each cell cycle phase, cells were irradiated, fractionated, and lysates subjected to western blot analysis. Colored boxes around the time points correspond to the color of the PI profile representing a distinct stage of the cell cycle (right panel). See also Figure S5.

Figure 5. Phosphorylation and nuclear localization of ACLY promote BRCA1 recruitment to DSB sites

(A) EV or myc-tagged mACLY constructs (WT or H760A) were transfected in LN229 sgACLY (clone 3.8) cells, followed by irradiation with 2 Gy IR. Cells were fixed after 4 h and co-stained for myc-tag (representing ACLY) and BRCA1. Quantification on myc-tag+ cells (or all nuclei for EV controls) was performed from 25 different fields per sample using ImageJ, mean +/−SEM, **, p<0.01, ****, p<0.0001;, n.s.-not significant. Scale bar-200 M. (B) Experiment was conducted as in (A), with analysis of 53BP1 and myc-tag. (C) Experiment was conducted as in (A), using myc-tagged mACLY constructs (WT, S455A, or S455D). (D) ACLY null LN229 cells were co-transfected with empty vector and GFP, WT mACLY and GFP, or with mACLY-NES and GFP for 48 h, and fixed 4 h after 2Gy IR treatment. IF was performed for BRCA1; GFP and BRCA1 was imaged. Quantification was performed with at least 25 fields per sample using ImageJ scoring GFP positive cells, mean +/−SEM. ***, p<0.001, n.s.- not significant. See also Figure S6.

Figure 6. ACLY deficiency enhances PARP inhibitor-induced genomic instability and cytotoxicity

(A) U2OS cells were treated with olaparib at indicated doses for 72 h following ACLY silencing for 24 h. Viability was assessed using trypan-blue exclusion assay, mean +/− SEM, *, p<0.05. (B) Two independent U2OS reporter clones lacking 53BP1 (sg53BP1-clone 2A and clone 3D) were treated without (solid bars) or with (striped bars) olaparib at 30 M final concentration for 72 h following ACLY silencing for 24 h. Viability was assessed using trypan-blue exclusion assay. n.s.- not significant. (C) ACLY was silenced in HeLa cells, and cells were treated with olaparib for 24 h. Samples were collected following colcemid treatment for 1.5 h and metaphases were examined for chromosomal abnormalities. Quantification was performed manually on 70–80 total metaphases from 2 different experiments which were pooled together, mean +/− SEM; *, p<0.05, **, p<0.01. See also Figure S7.

Figure 7. Acetyl-CoA production by ATP-citrate lyase facilitates homologous recombination during S/G2 phases of the cell cycle

ACLY is phosphorylated at Ser455 within the nucleus in response to DNA damage in a cell cycle-dependent manner. This enables appropriate spatial and temporal control of acetyl-CoA production, enabling histone acetylation at DSB sites, BRCA1 recruitment, and DNA repair by HR.