Global genome nucleotide excision repair is organized into domains that promote efficient DNA repair in chromatin

Abstract

The rates at which lesions are removed by DNA repair can vary widely throughout the genome, with important implications for genomic stability. To study this, we measured the distribution of nucleotide excision repair (NER) rates for UV-induced lesions throughout the budding yeast genome. By plotting these repair rates in relation to genes and their associated flanking sequences, we reveal that, in normal cells, genomic repair rates display a distinctive pattern, suggesting that DNA repair is highly organized within the genome. Furthermore, by comparing genome-wide DNA repair rates in wild-type cells and cells defective in the global genome-NER (GG-NER) subpathway, we establish how this alters the distribution of NER rates throughout the genome. We also examined the genomic locations of GG-NER factor binding to chromatin before and after UV irradiation, revealing that GG-NER is organized and initiated from specific genomic locations. At these sites, chromatin occupancy of the histone acetyl-transferase Gcn5 is controlled by the GG-NER complex, which regulates histone H3 acetylation and chromatin structure, thereby promoting efficient DNA repair of UV-induced lesions. Chromatin remodeling during the GG-NER process is therefore organized into these genomic domains. Importantly, loss of Gcn5 significantly alters the genomic distribution of NER rates; this has implications for the effects of chromatin modifiers on the distribution of mutations that arise throughout the genome.

Figure 1.

Genome-wide UV-induced DNA repair is organized around gene structure. (A) A linear genome plot of a section of Chromosome 14 showing 3D-DIP-Chip results from wild-type cells. The black line shows the mean (n = 3) CPD level observed immediately after UV irradiation (100 J/m², shading highlights the SEM). Gray dots indicate the positions of microarray probes. Yellow arrows indicate ORF positions and their direction of transcription. CPD levels are plotted as arbitrary units on the y-axis. (B) CPD repair rates displayed in a linear genome plot. The black line shows the mean of CPD levels 120 min post-UV (n = 2) subtracted from the mean at 0 min post-UV shown in A. Annotations are as described in A. (C) Relative rates of CPD repair around ORF structures. Solid lines show the mean of CPD repair rates in wild-type (n = 3, black line) and rad16Δ cells (n = 2, red line). Shaded areas indicate the SD, with CPD levels plotted as arbitrary units on the y-axis.

Figure 2.

GG-NER is organized from Abf1 binding sites, and Abf1 occupancy does not change significantly in response to UV. (A) The positions of Abf1 binding relative to ORFs. Abf1 binding levels at the ∼3800 detected binding sites are shown. Each binding site is represented by a single data point, with the overall relative amount of binding throughout the region shown above. (B) ChIP-chip data for Abf1 binding. Data are shown for unirradiated (black, circle), 0 min post-UV (dark gray, diamond), and 30 min post-UV (light gray, square) cells. Solid lines show the means of three data sets per time point. (C) As in B, plotted around ORF structure. (D) Relative CPD repair rates around Abf1 binding sites. The data depicted in Figure 1C are used here to plot the relative rates of CPD removal around Abf1 binding sites in wild-type (black) and rad16Δ cells (red). Solid lines show mean CPD repair rates in wild-type (n = 3, black line) and rad16Δ cells (n = 2, red line). The shaded areas show the SEM and SD, with CPD levels plotted as arbitrary units on the y-axis.

Figure 3.

The colocalization of the GG-NER factor Rad7 in chromatin at Abf1 binding sites and its redistribution in response to UV irradiation. (A) Rad7 binding data around detected Abf1 binding sites in the absence of UV (black) and 15 min post-UV (gray). Solid lines show the means of three data sets, and shaded areas show the SEM. (B) As in A, plotted around ORF structure.

Figure 4.

Rad16 associates with chromatin surrounding Abf1 binding sites and is redistributed in response to UV similar to Rad7. (A) Rad16 binding data around Abf1 binding sites for unirradiated (black) and 30 min post-UV (gray) cells. Solid lines show the means of three data sets, and shaded areas show the SEM. (B) As in A, plotted around ORF structure.

Figure 5.

The activity of both the ATPase and RING domain of Rad16 determine its chromatin occupancy before and after UV irradiation. (A) Representation of the linear structure of Rad16. The amino acids targeted by the point mutations introduced in the ATPase (K216A) and RING domains (C552A, H554A) are highlighted. (B–E) Composite plots of Rad16 chromatin occupancy in the mutants described. Mutated Rad16 binding data around Abf1 binding sites and ORF structures in the absence of UV irradiation (dark blue) and 15 min after UV irradiation (green) are shown here. The Rad16 RING mutant binding data (dashed lines) are shown in B (around Abf1 binding sites) and C (around ORFs). The binding data for Rad16 ATPase domain mutant (dotted lines) are shown in D (around Abf1 binding sites) and E (around ORFs). Lines show the means of three data sets per condition, and shaded areas show the SEM.

Figure 6.

Gcn5 is recruited to Abf1 binding sites and ORFs in response to UV in a Rad16-dependent manner. (A) Gcn5 binding data in wild-type cells around Abf1 binding sites for unirradiated (black; circle highlight), 0 min post-UV (dark gray; diamond highlight), 15 min post-UV (mid-gray; square highlight), and 60 min post-UV (light gray; triangle highlight) cells. Solid lines show means (n = 3, 3, 2, and 3, respectively), and shaded areas show the SEM (SD for n = 2). (B) Gcn5 binding data in rad16Δ cells around Abf1 binding sites for unirradiated (red; circle highlight), 0 min post-UV (dark pink; diamond highlight), 15 min post-UV (mid-pink; square highlight), and 60 min post-UV (light pink; triangle highlight) cells. Solid lines show the means of two data sets per time point, and shaded areas show the SD. (C) As in A, plotted around ORF structure (see Fig. 1C). (D) As in B, plotted around ORF structure (see Fig. 1C).

Figure 7.

Histone H3 acetylation levels in response to UV irradiation in wild-type and rad16Δ cells depend on the GG-NER complex. (A) Histone H3 acetylation in wild-type (n = 5, black/gray) and rad16Δ (n = 3, red/pink) cells in response to UV irradiation around Abf1 binding sites. The hatched areas define the genomic regions of GG-NER-dependent UV-induced histone H3 acetylation. Solid lines show the mean, and shaded areas show the SEM. (B) As in A, plotted around ORF structure.

Figure 8.

The GG-NER pathway coordinates lesion removal by controlling UV-induced histone H3 acetylation in genomic domains around Abf1 binding sites. (A) Rad16-dependent repair (purple line) and UV-induced H3Ac (orange line) are shown here. The shading highlights the domain where these processes are controlled by the GG-NER complex, initiated from sites of Abf1 binding. (B) Relative rates of CPD removal around ORF structures in wild-type (n = 3, black) and gcn5Δ (n = 2, green) cells. Solid lines show the mean of relative CPD repair rates levels, with the shaded areas highlighting the SEM or SD, respectively. CPD levels are plotted as arbitrary units on the y-axis.

Figure 9.

Model to illustrate how GG-NER is organized in the yeast genome. (Top panel) In undamaged cells, the GG-NER complex is located at multiple Abf1 binding sites predominantly in the promoter regions of genes. This occupancy is dependent on the RING domain of the Rad16 protein. The enrichment of GG-NER-independent basal levels of Gcn5 can be detected at these sites. (Middle panel) In response to UV irradiation, the GG-NER complex dissociates from the Abf1 component at Abf1 binding sites. This process depends on the activity of the ATPase domain in Rad16. Concomitantly, the HAT Gcn5 is recruited onto the chromatin with its increased levels and distribution dependent on the Rad7-Rad16 GG-NER complex. (Bottom panel) During this process, histone H3 acetylation is increased over a domain defined by the redistribution of the Rad7-Rad16 proteins from Abf1 binding sites. This mechanism drives the chromatin remodeling necessary for the efficient repair of UV damage.