The DREAM complex functions as conserved master regulator of somatic DNA-repair capacities

Abstract

The DNA-repair capacity in somatic cells is limited compared with that in germ cells. It has remained unknown whether not only lesion-type-specific, but overall repair capacities could be improved. Here we show that the DREAM repressor complex curbs the DNA-repair capacities in somatic tissues of Caenorhabditis elegans. Mutations in the DREAM complex induce germline-like expression patterns of multiple mechanisms of DNA repair in the soma. Consequently, DREAM mutants confer resistance to a wide range of DNA-damage types during development and aging. Similarly, inhibition of the DREAM complex in human cells boosts DNA-repair gene expression and resistance to distinct DNA-damage types. DREAM inhibition leads to decreased DNA damage and prevents photoreceptor loss in progeroid Ercc1^-/- mice. We show that the DREAM complex transcriptionally represses essentially all DNA-repair systems and thus operates as a highly conserved master regulator of the somatic limitation of DNA-repair capacities.

Fig. 1. Mutations in genes encoding components of the DREAM complex confer resistance to UV-induced DNA damage during development and adulthood.

a, Sequences of the motifs found upon analysis of the promoters of the DDR genes using HOMER. BG, background; FC, fold change; E2F, E2F transcription factor; Zf, zinc finger domain. b, Result of the motif search for the CDE-CHR DREAM complex motif in the promoters of the DDR genes using HOMER. c,d, UV-irradiation assay during somatic development of WT, lin-52(n771), lin-35(n745), dpl-1(n2994), efl-1(se1) (c) and WT, lin-52(n771), dpl-1(n2994) and lin-52(n771); dpl-1(n2994) mutant worms (d). Y axis shows the percentage of the different larval stages, and x axis the UV dose applied in mJ/cm². Representative graph showing n = 3 biological replicates from 1 of at least 3 independent experiments. Data are shown as mean ± s.d. of each larval stage (L1–L4-A). For statistical analysis, a two-tailed t-test between the fraction of each larval stage of a mutant compared with WT in the same treatment condition was used, except for lin-52(n771); dpl-1(n2994), which was compared with lin-52(n771). e, Lifespan assay upon exposure of WT and DREAM-complex-mutant worms to UV-B (0 and 400 mJ/cm²). A log-rank test was performed to compare the lifespan of the DREAM mutants and WT worms in the same conditions. Top graphs, n = 122 (without UV), 150 (with UV) for WT; 121, 152 for lin-52; and 128, 154 for dpl-1 mutants. Bottom graphs, n = 138, 218 for WT; 162, 215 for efl-1 and 144, 201 for lin-35 mutants. Bar graphs show the percentage by which mean lifespan decreased for each strain irradiated with UV-B compared with the mock-treated worms of the same strain. f, UV-irradiation assay for germline development of WT, xpc-1(tm3886), lin-52(n771) and the double mutant, lin-52(n771); xpc-1(tm3886). Representative graphs of n = 3 biological replicates, from 1 of 3 independent experiments. The mean ± s.d. of eggs laid or the percentage hatched is shown. Two-tailed t-tests were performed to compare the number of eggs laid and hatched between the different strains within the same condition. For a and b, the P value over the background was calculated with a hypergeometric test, and the q value shows the Benjamini–Hochberg-adjusted P values. For c and d, P > 0.05, not shown; *P < 0.05, **P < 0.01, ***P < 0.001, **** P < 0.0001. Remaining comparisons and detailed P values are provided in Supplementary Table 13, including results from Fisher’s exact test.

Fig. 2. DREAM-complex mutants show enhanced repair of UV-induced DNA lesions and alleviate the UV sensitivity of csb-1 and xpc-1 mutant animals.

a, DNA-repair capacity assay in WT, xpa-1(ok698), lin-52(n771) and dpl-1(n2994) L1 worms. A representative slot blot of three independent experiments is shown. Samples labelled as ‘UV’ were collected right after UV irradiation; samples ‘UV + 24 h’ were collected 24 h after UV irradiation. Graphs show the mean ± s.d. of the improved or decreased repair of the mutants compared with that of WT worms. n = 3 biological replicates. A two-tailed t-test was used to compare the mutants’ repair with that of WT worms. b, Representative images of a focal plane of the anterior region of adult worms irradiated and stained with antibodies to CPDs and DAPI, collected right after irradiation (0 h) or after incubation for 60 h. Scale bar, 25 µm. c, Quantification of CPD nuclei signal intensity in the heads of adult worms irradiated and collected immediately (0 h, blue dots) or 60 h after irradiation (orange dots). The number of nuclei quantified was, at 0 h and 60 h, respectively, n = 1,469 and 1,479 for WT, n = 1.454 and 1.321 for lin-52(n771), from 5–7 heads per condition. The y axis shows the log₁₀-transformed intensity values of CPDs. Box midlines show the median, box limits show the top and bottom quartiles and whiskers extend to 1.5 × interquartile range (IQR). Two-way analysis of variance (ANOVA) between the strain (WT and lin-52) and the time component is shown. a.u., arbitrary units. d–f, UV-irradiation assay during somatic development of WT, lin-52(n771), csb-1(ok2335) and lin-52(n771); csb-1(ok2335) (d); WT, lin-52(n771), xpc-1(tm3886) and lin-52(n771); xpc-1(tm3886) (e); and WT, lin-52(n771), csb-1(ok2335); xpc-1(tm3886) and lin-52(n771); csb-1(ok2335); xpc-1(tm3886) (f). Y axis shows the percentage of the different larval stages, and x axis the UV dose applied in mJ/cm². Graphs are representative of n = 3 biological replicates from 1 of 3 independent experiments. Data are shown as mean ± s.d. of each larval stage. Results from two-tailed t-tests between the fraction of each larval stage of lin-52-mutated worms compared with WT, and each larval stage of lin-52-mutated NER-deficient worms compared with the NER-deficient control, are shown for the same treatment conditions. P > 0.05, not shown. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. For d–f, detailed P values and comparisons against WT worms are in Supplementary Table 13, including results from Fisher’s exact test to analyze the overall distribution of the larval stages. Source data

Fig. 3. The DREAM complex directly represses multiple DNA-damage-response genes that are normally enriched in the germline.

a,c, Differentially expressed genes (a) or proteins (c) in lin-52(n771) mutants (adjusted P < 0.05). DDR genes (or their products) are shown in orange, and the significantly changed DDR genes in common between a and c are labeled. b,d, GO enrichment analysis for the upregulated genes (b) or proteins (d) in lin-52(n771) mutants compared with WT (adjusted P < 0.05, two-sided Fisher’s exact test with FDR). Highly overlapping terms were removed for simplicity. Terms related to DNA-damage responses are shown in red. The dashed lines mark an adjusted P value of 0.05. The full list is in Supplementary Table 3 and Supplementary Table 4. e, FC of genes that were significantly changed (adjusted P < 0.05) in both the proteome and transcriptome of lin-52 mutants. f, qPCR analysis of DDR genes in lin-52(n771), dpl-1(n2994) and efl-1(se1) mutants. Data are shown as mean ± s.d., n = 3 biological replicates. Results from two-tailed t-tests are in Supplementary Table 13. g, Overlap between the DDR genes upregulated in lin-52(n771) and the genes involved in the main DNA-repair pathways. Overlap between repair pathways is not shown. Pathway genes were obtained from the GO database released on 8 October 2019. h, Overlap between the DDR genes that were upregulated in lin-52 compared with WT worms and two published transcriptome datasets on lin-35(n745)^,. i, GSEA of all DDR genes in the RNA-seq of lin-52 and the genes bound by DREAM, as described in ref. . Seventy-six out of the 211 DDR genes were found in both and were used for the analysis. DDR genes bound by DREAM that are upregulated in lin-52 mutants are shown in red; downregulated genes are shown in blue. NES, normalized enrichment score. j, Overlap between the upregulated DDR genes in lin-52 mutants and the genes that were found to be bound by the DREAM complex in the promoter area (41 in promoter area, 43 in total). This is a re-analysis of work in ref. . k,l, Overlap between all the DDR genes in C. elegans (k) and those upregulated in lin-52 mutants (l) and the genes that were enriched in the germline in ref. . m, GSEA of the RNA-seq of lin-52(n771) with the genes enriched in the germline. Results from two-sided Fisher’s exact test are shown for overlap analyses. All GSEA statistics were done as described in ref. .

Fig. 4. Mutations in the DREAM complex confer DNA-damage resistance against multiple damage types.

a, IR sensitivity dependent on HRR was tested in WT, lin-52(n771), HRR-deficient brc-1(tm1145); brd-1(dw1), NHEJ-deficient cku-70(tm1524), lin-52(n771); brc-1(tm1145); brd-1(dw1), lin-52(n771); brc-1(tm1145); brd-1(dw1); cku-70(tm1524) and lin-52(n771); cku-70(tm1524) worms. Data are shown as mean ± s.e.m., n = 4 independent experiments (n = 3 for lin-52; cku-70). Each independent experiment had three biological replicates. Two-tailed t-tests were used for statistical comparisons, and the most relevant comparisons are shown. b, IR sensitivity assay dependent on NHEJ repair in WT, lin-52(n771), NHEJ-deficient cku-70(tm1524) and lin-52(n771); cku-70(tm1524) worms. The graph is representative of n = 3 biological replicates in 1 of 3 independent experiments, each of which had 3 biological replicates. Data are shown as mean ± s.d. of each larval stage. Two-tailed t-tests were used to compare the fraction of the larval stages of lin-52 compared with WT, and lin-52; cku-70 compared with cku-70 (non-significant). c, Alkylation-damage assay of WT, lin-52(n771), alkylation-damage-sensitive polh-1(lf31) and double-mutant lin-52(n771); polh-1(lf31) worms. The graph is representative of n = 3 biological replicates in 1 of 3 independent experiments. Data are shown as mean ± s.d. of each larval stage. Two-tailed t-tests were used to compare the fraction of the larval stages of lin-52 with WT, and lin-52; polh-1 with polh-1. d, ICL assay of WT and lin-52 mutants upon cisplatin treatment. Because cisplatin was diluted in DMF, worms were also given the maximum dose of DMF that was given for the cisplatin treatments as additional control. The graph is representative of n = 3 biological replicates in 1 of 3 independent experiments. Data are shown as mean ± s.d. of each larval stage. Two-tailed t-tests were used to compare the fraction of the larval stages of lin-52 compared with WT. P > 0.05, not shown. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Detailed P values and more comparisons can be found in Supplementary Table 13, including results from Fisher’s exact test, which was used to analyze the distribution of the larval stages.

Fig. 5. Inhibition of DREAM using DYRK1A inhibitors confers DNA-damage resistance in human cells and decreases DNA damage and apoptosis in the retinas of Ercc1-deficient mice.

a, Overlap between the DNA-repair genes in humans (GO database released 1 January 2021) and the genes bound by DREAM that were described in ref. . A two-tailed Fisher’s exact test was used for statistical analysis. b, FC comparison of DREAM target genes whose expression levels are significantly changed upon harmine or INDY treatment (FDR < 0.01 in at least one of the datasets). DNA-repair genes are shown in orange. The upper-right quadrant of derepressed genes has a 2.39× over-enrichment. Detailed statistics are in Supplementary Table 13. c,d,f,g, Representative density plots of biological triplicates of U2OS cells, labeled with annexin V and 7-AAD, that were mock treated or that received harmine hydrochloride and/or UV (c), INDY and/or UV (d), harmine hydrochloride and/or MMS (f) or INDY and/or MMS (g). e,h, Percentage of apoptotic annexin-V-positive U2OS cells upon harmine (left) or INDY (right) treatment and UV (e) or MMS (h) treatment. Data are shown as mean ± s.d., n = 3 biological replicates. Two-tailed t-tests were used for statistical analysis between the populations under the same irradiation conditions. i, Representative images of WT and Ercc1^−/− retinas with TUNEL staining (green) and DAPI (blue). j, TUNEL-positive cells in the ONL from WT and Ercc1^−/− mice upon treatment with harmine. n = 3 (2 male/1 female), 5 (3 male/2 female), 7 (4 male/3 female) and 7 (4 male/3 female) mice from left to right. Data are shown as mean ± s.e.m. For statistical analysis, two-tailed t-tests were used to compare groups that received or did not receive harmine treatment. k, Representative images of WT and Ercc1^−/− retinas with γH2AX staining (red) and DAPI (blue). The intensity of the red channel has been equally increased in all images for visualization purposes. The INL of the retina is encircled by the dashed yellow line. l, γH2AX signal per nucleus in the INL of the retinas of WT and Ercc1^−/− mice upon treatment with harmine. γH2AX signal per nucleus from an image stack per mouse is shown. n = 904, 889, 1,123 and 841 total nuclei, from left to right, from 5 (3 male/2 female) 5(3 male/2 female), 6 (3 male/3 female) and 5 (3 male/2 female) imaged mice, respectively. Box midlines show the median, box limits show the top and bottom quartiles, and whiskers to 1.5 × interquartile range (IQR). For statistical analysis, two-tailed t-tests were used to compare groups that received or did not receive harmine treatment.

Extended Data Fig. 1. Mutations of components of the DREAM complex but not of other synMuv B or NuRD components confer resistance to UV-induced DNA damage during development.

a, UV-irradiation assay during somatic development of WT, dpl-1(n3643) and lin-53(n3368) I/hT2 [bli-4(e937) let-?(q782) qIs48]. Larval stages were determined 48 h post-irradiation. Homozygous lin-53(n3368) were identified by inspection with a fluorescence microscope (non-GFP population). Representative graph showing n = three biological replicates from one out of at least three independent experiments. Mean +/− SD of each larval stage. Two-tailed t-test between the fraction of each larval stage of a mutant compared to WT in the same treatment condition. b, UV-irradiation assay during somatic development of WT, lin-52(n771), efl-1 (se1) and lin-52(n771); efl-1 (se1) mutants. Representative graph showing n = three biological replicates from one out of three independent experiments. Mean +/− SD of each larval stage. Two-tailed t-test between the fraction of each larval stage of a mutant compared to WT in the same treatment condition for the single mutants, and the comparison of the double mutant and lin-52(n771) are shown. c, UV-irradiation assay during somatic development of WT, hpl-2(tm1489), lin-13(ok838), lin-15(n765), met-2(n4256) and let-418(n3536) mutants. Representative graph showing n = three biological replicates from one out of at least two independent experiments. Mean +/− SD of each larval stage. Two-tailed t-test between the fraction of each larval stage of a mutant compared to WT in the same treatment condition. If P > 0.05, statistics not shown. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Detailed P values and Fisher’s exact tests to compare the population distribution can be found in the Supplementary Table 13.

Extended Data Fig. 2. lin-52 mutant worms show a slight starvation sensitivity and reduced fecundity.

a, Starved L1 larvae of WT and lin-52(n771) were plated in seeded plates each day and allowed to recover from starvation. The mean +/− SD percentage of worms surviving the starvation period compared to the total of worms per plate is presented, n = three biological replicates per day for each strain. The P value of the strain variable from two-way ANOVA is shown. b, Average of hatching eggs laid by adult lin-52(n771) and WT worms per day (left) or during their reproductive lifespan (right). n = 12 individual worms per strain were used, mean +/− SD is shown. The P value of the strain variable from two-way ANOVA is shown (left) or unpaired two-tailed t-test (right) between the strain.

Extended Data Fig. 3. lin-52 mutant worms retained motility upon DNA damage induction.

lin-52(n771) and WT day 1 adult worms were irradiated or mock-treated with UV-B and their motility was measured after 72 h. Maximum speed per individual is presented (mean +/− SD in red). Two-tailed Mann-Whitney test was used. From left to right, n = 20, 23, 29, 32.

Extended Data Fig. 4. lin-52 mutant worms show improved UV damage repair capacity without having additional replication events compared to WT.

a, Representative slot blot out of three independent experiments labelled with antibodies against 6-4PPs and SYBR^TM Gold for DNA staining. DNA samples were collected from irradiated worms (WT and lin-52(n771)) right after or 24 h after 60 mJ·cm² UV-B irradiation. Graph show the percentage of damage repair of WT and lin-52 mutant worms after 24 h compared to 0 h (n = 3, mean +/− SEM). Two-tailed t-test was used to compare the repair with WT worms. *P < 0.05. Uncropped blot in Source Data ED Fig. 4. b, Mock-treated (left) or UV-B irradiated (right) WT and lin-52(n771) L1-stage worms were EdU labelled and imaged at different timepoints. Representative graphs of two independent experiments. The number of replication events per individual worm are presented (average with SD). Unpaired t-test with Welch’s correction was applied. No significant differences were found. From left to right, n = 6, 6, 7, 7, 6, 6 // 6, 6, 6, 7, 10, 8 worms. c, Day 1 adult lin-52 mutant and WT worms were fed with EdU containing bacteria for three days. Worms were cut open and intestines and germlines were analyzed using confocal microscopy to find EdU positive nuclei. n = 152 intestinal cells of WT and lin-52 mutants were counted. All observed germlines were EdU positive. Scale bar, 75 µm. Source data

Extended Data Fig. 5. Germline DNA repair is similar between WT and lin-52 mutant worms and mutant lin-52 alleviates the UV sensitivity of csa-1 but not of xpa-1 mutants.

a, Quantification of CPD nuclei signal intensity in the germline of adult worms irradiated and collected immediately or 60 h after irradiation. The number of nuclei quantified was n = 962, 462, 479, 494 from left to right from 3–5 germlines per strain and condition. The y axis shows the CPD intensity normalized to nuclear DAPI with subtracted background. Box depicts the median with top and bottom quartiles, whiskers to 1.5 IQR. 2-way ANOVA analysis shows a negligible effect of the genotype. b, UV-irradiation assay during somatic development of WT, lin-52(n771), csa-1(tm4539) and lin-52(n771); csa-1(tm4539) mutants. Representative graph showing n = three biological replicates from one out of three independent experiments. Mean +/− SD of each larval stage. Two-tailed t-tests between the fraction of each larval stage of a lin-52; csa-1 compared to csa-1 and lin-52 compared to WT in the same treatment condition are presented. c, UV-irradiation assay during somatic development of WT, lin-52(n771), xpa-1(ok698) and the lin-52(n771); xpa-1(ok698) mutants. Representative graph showing n = three biological replicates from one out of two independent experiments. Mean +/− SD of each larval stage. Two-tailed t-tests between the fraction of each larval stage of a lin-52; xpa-1 compared to xpa-1 and lin-52 compared to WT in the same treatment condition are presented. If P > 0.05, statistics not shown. *P < 0.05, **P < 0.01, ***P < 0.001. Detailed P values and all comparisons against WT and Fisher’s exact tests to compare the population distribution can be found in the Supplementary Table 13.

Extended Data Fig. 6. DREAM binds DDR genes and DREAM´s mediated DDR repression is not occurring in the germline.

a, GSEA of all the DDR genes in the RNA-seq of lin-52(n771) with the genes which promoters are bound by DREAM upon analysis of reference. 63 out of the 211 DDR genes are found in both and used for the analysis. In red, DDR genes bound by DREAM and upregulated, in blue, downregulated. b, Overlap between the genes bound by DREAM and all the DDR genes (Supplementary Table 1) (11 not found). Re-analysis of reference . Overlap P value calculated by performing Fisher’s exact test. c, Overlap between a dataset of germline-enriched genes and all the genes significantly up-regulated in lin-52 mutants compared to WT worms (p-adj < 0.05) (left) and the genes found down-regulated in lin-52 mutants compared to WT worms (p-adj < 0.05) (right). Overlap P value was calculated by Fisher’s exact test. d, GSEA of the DDR genes within a microarray of lin-54 mutant embryos from reference. 30 out of the 211 DDR genes are found in the lin-54 embryo dataset consisting of 977 genes and used for the analysis. In red, DDR genes upregulated in lin-54 mutant embryos, in blue, downregulated. e, GSEA of the DDR genes within a microarray of lin-54 mutant germlines from reference. 7 out of the 211 DDR genes are found in the lin-54 germline dataset consisting of 328 genes and used for the analysis. In red, DDR genes upregulated in lin-54 mutant germlines, in blue, downregulated. Two-sided Fisher’s exact test is shown for overlap analysis. GSEA statistics done as described in.

Extended Data Fig. 7. DREAM complex mutants confer DSB damage resistance and can partially alleviate the sensitivity of DNA repair deficient strains.

a, HRR-dependent IR sensitivity assay in WT and efl-1(se1) mutant. Graph shows the mean of 8 independent experiments, with SEM, each of them with three biological replicates. Two-tailed t-tests between the fraction of surviving embryos within the treatment are presented. b, NHEJ-dependent IR sensitivity assay in WT, lin-52(n771), xpa-1(ok698), polh-1(lf31) and the double mutants lin-52(n771); xpa-1(ok698) and lin-52(n771); polh-1(lf31). Representative graph of one out of three independent experiments, each with n = 3 biological replicates. Mean +/− SD of each larval stage. Two-tailed t-tests between the fraction of the larval stages of lin-52 compared to WT, lin-52; xpa-1 compared to xpa-1 and lin-52; polh-1 compared to polh-1 are presented (showed statistics indicated with upper bracket). c, Lifespan assay upon IR (control and 1000 Gy treatment) of WT, lin-52(n771) and efl-1(se1) mutant worms. log-rang (Mantel-Cox) test was performed to compare the lifespan of the DREAM mutants and WT in the same conditions. Without IR and with IR respectively, n = 62 and 118 for WT, 63 and 116 for lin-52(n771) and 85 and 100 for efl-1(se1). For b, P > 0.05, statistics not shown. *P < 0.05, **P < 0.01. The precise P values and all comparisons to WT worms and Fisher’s exact tests to compare the population distribution can be found in the Supplementary Table 13.

Extended Data Fig. 8. Analysis of DREAM recognition sites in promoters of genes significantly upregulated upon harmine hydrochloride or INDY treatment in U2OS cells.

The 4 previously reported promoter motifs bound by DREAM corresponding to E2F, NRF2, CREB and n-MYC consensus motifs from top to bottom, were searched in the gene-sets of significantly (FDR < 0.01) upregulated genes by either harmine or INDY. All four motifs in both treatments show a highly significant overrepresentation. The P value was calculated with two-sided hypergeometric tests and the q value shows the Benjamini–Hochberg adjusted P values. BG, background; FC, fold change.