High UV damage and low repair, but not cytosine deamination, stimulate mutation hotspots at ETS binding sites in melanoma

Abstract

Noncoding mutation hotspots have been identified in melanoma and many of them occur at the binding sites of E26 transformation-specific (ETS) proteins; however, their formation mechanism and functional impacts are not fully understood. Here, we used UV (Ultraviolet) damage sequencing data and analyzed cyclobutane pyrimidine dimer (CPD) formation, DNA repair, and CPD deamination in human cells at single-nucleotide resolution. Our data show prominent CPD hotspots immediately after UV irradiation at ETS binding sites, particularly at sites with a conserved TTCCGG motif, which correlate with mutation hotspots identified in cutaneous melanoma. Additionally, CPDs are repaired slower at ETS binding sites than in flanking DNA. Cytosine deamination in CPDs to uracil is suggested as an important step for UV mutagenesis. However, we found that CPD deamination is significantly suppressed at ETS binding sites, particularly for the CPD hotspot on the 5' side of the ETS motif, arguing against a role for CPD deamination in promoting ETS-associated UV mutations. Finally, we analyzed a subset of frequently mutated promoters, including the ribosomal protein genes RPL13A and RPS20, and found that mutations in the ETS motif can significantly reduce the promoter activity. Thus, our data identify high UV damage and low repair, but not CPD deamination, as the main mechanism for ETS-associated mutations in melanoma and uncover important roles of often-overlooked mutation hotspots in perturbing gene transcription.

Keywords: CPD-seq 2.0; ETS; NER; mutagenesis.

Fig. 1.

Mutation analysis in gene promoters. (A) Average density of somatic mutations of 140 cutaneous tumors along transcribed human genes. Genes were aligned by the TSS and presented in the transcriptional direction (e.g., promoter on the left and coding on the right side of the TSS). Mutation density was shown as the number of mutations per tumor per gene in 10-bp non-overlapping moving windows. (B) Mutation spectra in the mutation peak between −200 bp and +100 bp of the TSS. The pie chart shows the percentage of each type of mutation. (C) The top 10 most frequently mutated promoters in the tumor cohort. The left column indicates the number of tumors that have the specific mutation. The second left column indicates the genomic location of the mutation site. The next two columns present the linked gene and the sequence context of the mutation site. Ribosomal proteins gene names are shown in bold and the TTCCGG motif is shown in gray in the sequence context. (D) Left: the average mutation density in 968 TTCCGG genes. The expected mutation density was calculated based on the mutation probability of each trinucleotide. Right: the average mutation density in 9,864 non-TTCCGG genes. These genes have at least one UV mutation in the peak region in the cohort, but none of them is associated with the TTCCGG motif. (E) Comparison of mutation density in the peak region between TTCCGG and Non-TTCCGG genes. Each dot represents the average mutation density of 10 bp in the peak region. (F) ETS binding to the promoter of TTCCGG and Non-TTCCGG genes. The density of ETS ChIP-seq peaks (e.g., peak per gene) was plotted relative to the TSS in 10 bp moving windows.

Fig. 2.

CPD-seq 2.0 analysis of mCPD formation and repair. (A) The schematic of CPD-seq 2.0. The Top panel shows the experimental setup. The Lower panel details the steps for CPD-seq 2.0 library preparation. T = T indicates CPD damage; “OH” indicates a free 3′OH group; “dd” indicates dideoxy (3′H). (B) Counts of dinucleotides associated with each CPD-seq read. The two nucleotides immediately upstream of the 5′ end of Read 1 on the opposite strand were collected and counted. CPDs are expected to occur at dipyrimidines such as TT, TC, CC, and CT. mCPD: mutagenic CPD. (C) Average mCPD (reads per gene) in 968 TTCCGG genes in 20 bp moving windows. From Top to Bottom shows mCPDs at 0, 6, and 24 h, respectively. (D) Average mCPDs in 9,864 Non-TTCCGG genes for different time points. (E) Fraction of remaining mCPDs after 6 and 24 h repair normalized to the initial damage at 0 h for TTCCGG (Left) and Non-TTCCCGG genes (Right).

Fig. 3.

Mutation and UV damage signatures in TTCCGG gene promoters. (A) The 968 TTCCGG genes were aligned by the TTCCGG motif to present melanoma mutations and mCPDs relative to the motif site. The center (position 0) is the midpoint of the TTCCGG motif. (B) Heatmaps of mutations and mCPDs for each TTCCGG gene. Genes were aligned by the motif sequence and sorted based on the mutation frequency in the melanoma cohort (e.g., Top shows the most frequently mutated promoters). (C) Mutations and mCPDs in the core TTCCGG and flanking DNA (21 bp in total). Similar to panel (A), the TTCCGG promoters are aligned by the conserved motif. Mutation and damage are shown for each single nucleotide position. The consensus sequence in the core motif and flanking DNA is generated with WebLogo (28).

Fig. 4.

CPD deamination in TTCCGG promoters and at ETS binding sites. (A) Cytosine deamination (DA) in control (i.e., no UV) and UV-treated cells (48 h). The DA data was obtained from the published circle-damage-seq study (29) and the average DA reads (reads per gene) are plotted for TTCCGG genes aligned by the conserved motif sequence. The panel on the Right shows DA rate, which is DA normalized by mCPDs. (B) Average DA reads (UV treated, 48 h) in the core TTCCGG motif and flanking DNA at each position. (C) The ETS binding sites mapped by ChIP-seq were aligned by the conserved motif sequence and the average DA reads were plotted at each single nucleotide position. (D) A subset of TTCCGG promoters in which the −3 and −4 positions are two consecutive Cs were aligned and the average mCPDs, melanoma mutations, and DA reads were plotted for each position. (E) Comparison of DA rate (DA/mCPDs) between −3/−4 and 0/+1. The rate at −3/−4 is set at 1.0 and the relative fold change at 0/+1 is shown. (F) Model showing two discernible mCPD hotspots at ETS binding sites at −3/−4 and 0/+1 positions, respectively. Only the hotspot at 0/+1 is deaminated to form uracil. Deamination at the −3/−4 hotspot is suppressed by ETS binding; however, both damage hotspots are converted to C>T mutations in melanoma.

Fig. 5.

ETS binding modulates H₂O distribution near the two damage hotspots. (A) Schematic of spontaneous cytosine deamination in a CPD. The amino group of cytosine can be hydrolyzed to form uracil. (B) Surface view of the ETS–DNA complex structure showing a cluster of charged amino acids (red) in close contact with the 0/+1 bases (blue), which forms a pocket to trap H₂O molecules. The protein–DNA complex structure is from the published study (34) and the DNA sequence used in the complex is shown on the Top. (C) Zoom-in view of charged amino acids near the 0/+1 bases. (D) In contrast to the 0/+1 bases, no charged amino acids are in the vicinity of the −3/−4 bases. (E) A representative depiction of H₂O distribution near 0/+1 and −3 bases. The −4 position is an A (adenine) in the complex structure (see panel B) and is not shown in the simulation data. (F) The number of hydrogen bonds formed between H₂O and different bases.

Fig. 6.

Mutation hotspots reduce promoter activities. (A) Two single mutations at position −3 (mutation 1; MT1) and 0 (MT2) of RPL13A promoter were introduced into the fire luciferase plasmid. The WT and double mutation (MT3) plasmids were also generated and expressed in melanoma A375 cells. The firefly luciferase signal was normalized by the co-transfected renilla luciferase and fold change in each mutant plasmid relative to WT is plotted. (B) In vitro ETS1 binding levels at ETS sites in the WT and MT1 RPL13A promoter sequences, relative to the overall distribution of ETS1 binding levels at 12,619 putative sites in the human genome (35). The x-axis shows the log-transformed binding intensity measurements from gcPBM experiments (GEO Series GSE97794) normalized to the [0,1] interval, as in ref. . The gcPBM data was binned into 112 bins with width of 0.01. The y-axis shows the number of ETS1 sites in each bin. The vertical, dashed lines show the predicted ETS1 binding levels for the sites in each promoter sequence, according to an iMADS model trained on gcPBM data (35). (C) Analysis of WT and mutant RPS20 promoter using the luciferase system. (D) Similar to panel B, but for the two ETS1 binding sites (site A and site B) identified in the RPS20 promoter. (E) Mutation in CDC20 promoter and its impact on luciferase activity. (F) Similar to panel B, but for the ETS1 site in the WT and MT CDC20 promoter. *P < 0.05, **P < 0.005, ***P < 0.0005.