CSB-independent, XPC-dependent transcription-coupled repair in Drosophila

Significance We have discovered that Drosophila, which does not have the canonical TCR homologs, does nevertheless carry out TCR as efficiently as organisms that do. Furthermore, using the XR-seq and in vivo excision assay we have also shown that both global repair and TCR in Drosophila are dependent on the XPC protein and in that regard, Drosophila excision repair is more similar to the monocellular eukaryotic yeast repair system than it is to multicellular eukaryotes. Finally, we have generated genome-wide single nucleotide repair maps of Drosophila for CPDs, (6-4) photoproducts, and cisplatin-d(GpG) adducts that should be a useful source for investigators working on DNA damage, repair, and mutagenesis in Drosophila.

XPC j transcription-coupled repair j XR-seq N ucleotide excision repair (excision repair) is a universal DNA repair system in the biological world. It removes DNA damage by making dual incisions bracketing the lesion to generate oligonucleotides 12 to 13 nucleotides (nt) in length in prokaryotes and 24 to 32 nt in length in eukaryotes (1)(2)(3). While originally discovered as a repair system for ultraviolet (UV)-induced DNA damage, it was later shown to remove all bulky base lesions caused by carcinogens such as benzo(a)pyrene and chemotherapeutic drugs such as cisplatin. The mechanism of excision repair has been investigated in considerable detail in Escherichia coli, Saccharomyces cerevisiae, and humans. Genomics studies have revealed that bacteria have homologs of the excision repair genes of E. coli, and eukaryotes have excision repair genes homologous to yeast and human genes. However, only limited work has been done on excision repair in organisms other than E. coli, yeast, and humans. In particular, the lack of mechanistic repair studies on excision repair in Drosophila is striking because work with Drosophila has examined chemical-and radiation-induced mutagenesis, double-strand break repair, and recombination.
Recently, we developed a method for genome-wide high-resolution analysis of nucleotide excision repair, called XR sequencing (XR-seq) (4,5), which makes it possible to analyze excision repair in a wide range of organisms, including Drosophila, in which bulk biochemical approaches are difficult. This method has made it possible to examine with unprecedented sensitivity both pathways of excision repair, namely, global genomic repair, which repairs damage throughout the genome, and transcription-coupled repair (TCR), which repairs transcription-blocking damage in the template strand of genes at an accelerated rate (6). In a preliminary study, we used XR-seq with the Drosophila S2 cell line to demonstrate that these cells do perform TCR, in contrast to the generally held view that Drosophila lacks TCR, which was based on low-resolution repair assays and was seemingly supported by the fact that Drosophila lacks a CSB homolog, which is essential for TCR in humans and yeast (7)(8)(9)(10)(11)(12). This unexpected finding led us to investigate excision repair in Drosophila in more detail.
Here, we report single nucleotide resolution maps for excision repair of UV damage in Drosophila at different developmental phases: embryo, larva, pupa, and adult. Our data confirm and extend the finding of TCR in Drosophila S2 cells to the organismal level, and in response to cisplatin damage, and unexpectedly also show that in contrast to humans and Arabidopsis, in Drosophila both global repair and TCR are dependent on the XPC damage recognition protein. In this regard, Drosophila is more similar to budding and fission yeasts than to any other multicellular organisms tested.

Excision Repair of Different Damage Types in the Drosophila S2 Cell
Line. Previously, we reported that UV-induced cyclobutane pyrimidine dimers (CPDs) are repaired by TCR in S2 cells (7).

Significance
We have discovered that Drosophila, which does not have the canonical TCR homologs, does nevertheless carry out TCR as efficiently as organisms that do. Furthermore, using the XR-seq and in vivo excision assay we have also shown that both global repair and TCR in Drosophila are dependent on the XPC protein and in that regard, Drosophila excision repair is more similar to the monocellular eukaryotic yeast repair system than it is to multicellular eukaryotes. Finally, we have generated genome-wide single nucleotide repair maps of Drosophila for CPDs, (6-4) photoproducts, and cisplatin-d(GpG) adducts that should be a useful source for investigators working on DNA damage, repair, and mutagenesis in Drosophila.
We wished to test the generality of this finding by analyzing the repair of other DNA lesions known to be processed by nucleotide excision repair. We chose  photoproducts [(6-4)PPs] and cisplatin-d(GpG) adducts for this purpose as these damages have been extensively studied in excision repair research. First, we wanted to find out whether PPs are repaired preferentially in Drosophila over CPDs as is the case in human excision repair (13)(14)(15)(16). Fig. 1 A and B show that in both S2 cells and in a normal human fibroblast cell line, NHF1, PPs are repaired more efficiently than CPDs. As we observed previously in S2 cells (7), the excision products (26 to 29 nt) are not readily degraded into smaller oligonucleotides, whereas we found that in the human cell lines, both photoproducts are degraded to smaller species rather rapidly (14,17). Apparently, the S2 cells either lack the potent nuclease(s) involved in degrading the excision products or the degrading nucleases are expressed at lower levels. We also note that because of different amounts of genomic DNA in S2 and NHF1 cells, the absolute levels of excision between the two cell lines cannot be compared based on the data in this figure which is meant to address the issue of preferential repair of the two photoproducts in human and Drosophila cell lines, respectively. Finally, in Fig. 1A, lane 9 we show that the Pt-d(GpG) adduct is also excised mainly in the form of 26 to 29 oligomers (mers), similar to UV photoproduct excision in Drosophila and similar to cisplatin repair in mouse (18) and human cells (19,20).
Next, we proceeded to characterize cisplatin adduct and UV photoproduct repair using the XR-seq procedure. In this assay (5), the excision products are isolated from cells, ligated to adaptors, purified with damage-specific antibodies [against cisplatin-, CPD-, or (6-4)PP-damaged DNA], repaired, amplified, and sequenced, and the sequenced reads are mapped to the genome (SI Appendix, Fig. S1). XR-seq was first done following treatment of S2 cells with cisplatin or UV light, and Fig.  1 C, Top shows, for each type of damage, the frequency of excision product reads as a function of their length. These results are consistent with the excision assay results in showing that the damages are excised mainly as 26 to 29 mers with a median of 28 nt (for cisplatin damage) or 27 nt (for UV damage). Fig. 1 C, Bottom shows the frequency of each base (y axis) at each position of the excised 28 mers (x axis) for cisplatin-d(GpG) adducts, CPDs, and (6-4)PPs. In each case, there is enrichment at positions 20 to 22 of G nucleotides (cisplatin) or pyrimidines (UV photoproducts). These sites are the presumptive location of damage in these excision products, and from these sites and the excision product length, the locations of incision sites may be determined: the 3 0 incision made during repair is located 6 nt from the platinated residues and the UV photoproducts, and the major 5 0 incision site is located 20 nt from the platinated residues and 19 nt from the photoproducts. These sites are consistent with prior studies of the dual incisions made by the excision nuclease in Drosophila and humans (7,19,20).

Effect of Transcription on Cisplatin Adduct and UV Photoproduct
Repair in S2 Cells. The reads generated by XR-seq were mapped to the genome and analysis of TCR was done by scaling reads obtained from all genes in Drosophila to a "unit gene," which thus represents the averaged repair in each strand of all Drosophila genes. This analysis of TCR revealed an interesting pattern: as shown in Fig. 1 D, Left, for cisplatin-d(GpG) adducts, the transcribed strand/nontranscribed strand (TS/NTS) repair ratio in the average gene body is ∼2, indicating strong TCR, considering that genes at all expression levels were included in the analysis and that the strong TCR in highly transcribed genes would be somewhat mitigated by weaker TCR in weakly transcribed genes. Nevertheless, Fig. 1 D, Left and SI Appendix, Fig. S2A convincingly show that cisplatin-d(GpG) is subject to TCR in S2 cells. Another feature noteworthy of cisplatin adduct repair is the "dip" of repair observed at transcription start sites (TSSs) and transcription end sites (TESs). These sites are known to be particularly A-T rich in the Drosophila genome and thus are expected to have lower than average frequency of the G-G sequence necessary for formation of the Ptd(GpG) adduct (21), as our own simulation results confirm. It is possible that transcription initiation and termination complexes also contribute to the dips in repair by impeding damage formation and repair. Fig. 1 D, Middle and SI Appendix, Fig.  S2A show a more modest TS/NTS ratio for CPD repair, comparable to what is seen in human NHF1 cells and is in agreement with our previous reports on S2 cells (4,7,19). We note that the relatively strong peak of CPD repair in both strands at the TES may arise from elevated CPD levels resulting from the nucleotide frequencies and sequences present (22,23). Also, looping of genes in Drosophila may occur to the extent that it enhances repair of the TES by bringing the transcription-repair factor TFIIH from its location of high density at the promoter in proximity to the TES (24,25). Finally, Fig. 1 D, Right and SI Appendix, Fig. S2A show that (6-4)PPs are not subject to substantial TCR at this resolution as is the case in human cells (4,19), even though in human cells TCR is mediated by CSB, CSA, and related proteins, whereas Drosophila lacks both CSB and CSA homologs (26)(27)(28)(29)(30)(31).
In Fig. 1 E and F we show XR-seq results in the form of browser screenshots of two representative housekeeping genes and quantitative plots. Several features of interest emerge. First, for both Pt-(GpG) and CPD there is clear TCR, whereas none is observed for (6-4)PPs, consistent with the results in Fig.  1D. Second, TCR of CPDs is particularly strong in the introns because in these particular genes the introns are A-T rich.
For our analyses of TCR, we exclude small and overlapping genes. Small genes are excluded due to the possibility of nonhomogeneous damage distribution in the two strands. In the case of overlapping genes, complications may arise. For example, a case in which the two genes are transcribed in opposite orientations is illustrated in SI Appendix, Fig. S2 B and C. XR-seq results show that in the region of overlap, likely due to relative transcription levels, TCR in the CG9821 gene predominates, resulting in more NTS than TS repair in CR43130.
In Vivo Model for Excision Repair in Drosophila. The above results and previous studies of excision repair in Drosophila employed cell lines cultured in vitro (32,33), and we were interested to know how the in vitro results extrapolated to the in vivo condition. Previous work showed that in other organisms, involvement of global and TCR are development dependent and repair may exhibit cell type-specific patterns (34)(35)(36). Drosophila is a complex organism with a life cycle that includes embryo (E2 [4 h]), larva (L3), pupa (P4), and adult (5 to 8 d old) developmental stages. We examined excision repair in vivo ( Fig. 2A) by irradiating organisms at each developmental stage with ultraviolet B (UVB) and then incubating them in the dark for 2 h at 25°C to allow excision repair to proceed in the absence of photoreactivation. Organisms at each stage were then pooled, lysed, and homogenized, and low-molecular-weight DNA was isolated and then immunoprecipitated with damage-specific antibodies following the Hirt protocol (5,14). Samples were then processed for excision assay or processed by optimized XR-seq procedures to generate libraries for sequencing (Fig. 2B).  Top shows frequency distribution profiles of excision products as percent of total reads (y axis) versus excision product length (x axis). In agreement with the excision assay results in A, reads of length 26 to 29 nt predominate. The peak read length is 28 nt for cisplatin and 27 nt for CPDs and (6-4)PPs. Bottom shows the 28-nt excision products, the relative frequency (y axis) of each nucleotide at each position (x axis). The 5 0 end of the excision products is located at position 1. Enrichment of G residues on the Left indicates the likely site of platination at G-G dinucleotides, and enrichment of pyrimidine residues indicates the likely site of formation of CPDs (Middle) and (6-4)PPs (Right). With respect to these damage sites, sites of dual incision are located 6 nt 3 0 from each adduct, and 20 nt 5 0 from the cisplatin adducts and 19 nt 5 0 from the CPDs and (6-4)PPs. (D) Repair of cisplatin and UV photoproducts in the TS (blue) and NTS (red). Excision product reads were mapped to the Drosophila genome and reads across all genes were scaled to a unit gene, which represents average repair in a Drosophila gene. Results are shown for cisplatin adducts (Left), CPDs (Middle), and (6-4)PPs (Right). Transcription-coupled repair is indicated by enhanced TS repair and is evident in cisplatin and CPD but not (6-4)PP repair. (E). Screenshots showing repair reads (y axis) across two representative housekeeping genes betaTub60D (Top) and ade3 (Bottom) and (F) quantitation of repair in each strand of the two genes. TCR is evident with cisplatin and CPDs but not (6-4)PPs. Small antisense transcripts are present in ade3 but not shown in E; apparently this antisense transcription is too weak to affect the overall TCR shown. Data points reflect means and SEs obtained from two experiments. Nucleotide frequency analyses show enrichment of dipyrimidines (T-T or T-C) 7 to 9 nt from the 3 0 end of the in vivo 28-mer excision products. As with Drosophila cells cultured in vitro (Fig. 1A), there is relatively little degradation of the excision product in vivo. These results demonstrate good agreement between repair in the in vitro and in vivo Drosophila models.
To quantitatively confirm the UV adduct repair pattern over extended periods of time in female and male Drosophila adults, the immuno-slot blot method with damage-specific antibodies was used to measure the dynamic loss of the total amount of genomic DNA damage. Adults were irradiated with 1,200 J/m 2 of UVB, and as seen in Fig. 2 D and E, we found that in both genders, (6-4)PP repair is almost completed within 48 h, and about half of the CPDs are removed within 48 h, consistent with repair kinetics in other organisms, including humans (14).
Genome-Wide and TCR in Drosophila In Vivo. We began our analysis of genome-wide repair in vivo using adult flies, and we filtered the data to analyze genes identified by FlyBase RNA-sequencing (RNA-seq) data (37) as expressed reads per kilobase per million total reads (RPKM) > 10. Time courses showing TCR in these expressed genes, ranging from 15 min to 8 d, are shown in Fig. 3 A and B. Maximal TCR appears across a broad peak centered at ∼2 h (Fig. 3B). As reported previously, peak repair was also seen at an early time point (30 min) in S2 cells (7), (10-min, 30-min, and 8-h repair times were tested). Interestingly, Fig. 3 A and B show that following 2-d repair in vivo, NTS repair predominates as CPDs are depleted from the template strand by rapid TCR. At this 2-d time point, ∼40% of the CPDs have been removed from the genome (Fig. 2E). In S2 cells, NTS repair was also clearly predominant at 16 h post-UV, and the switch from TS to NTS repair appears to have started by 8 h post-UV (7).
We next examined TCR repair patterns in different developmental stages and genders (Fig. 3C). This analysis revealed TCR in vivo in each stage and in both genders, as seen in S2 cells. The developmental phase dataset was then reanalyzed without filtering out weakly transcribed and nontranscribed genes, to allow comparison with the unfiltered S2 data in Fig. 1 D, Middle. The results in SI Appendix, Fig. S4 show weaker TCR in vivo compared to S2 cells, likely due to differences in relative transcription levels. Interestingly, more dominant repair near the TES appeared across all the repair time points compared to the human repair profiles, probably due to uniform RNAP binding of highly expressed genes (38) or the cotranscriptional cleavage termination model in Drosophila (22).
Finally, in plots of (6-4)PP XR-seq data for Drosophila, four developmental phases and both genders show that within the resolution of our assay, (6-4)PPs are not subject to substantial TCR in vivo (SI Appendix, Fig. S5) as is the case in S2 cells.
Requirement for XPC in Drosophila Nucleotide Excision Repair. In all multicellular eukaryotes tested, including humans and Caenorhabditis elegans, six repair factors XPA, RPA, XPC, TFIIH, XPG, and XPF-ERCC1 are required for global repair (3,39). Interestingly, only five of the factors are required for TCR; XPC is dispensable (40). In fact, XPC mutants have been utilized to study exclusively TCR, in the absence of "background" global repair (4). To apply this approach to examine TCR in Drosophila, we used an available XPC mutant fly (G1) generated by chemical mutagenesis (31). Due to the possibility of off-target mutations, we also generated an XPC knockout (KO) by CRISPR-Cas9 genome engineering.
With the XPC mutant and knockout flies, we first conducted UV survival assays. Survival was monitored by counting the living flies every 3 or 4 d following different doses of UVB irradiation. The results with no UV and 4,800 J/m 2 UVB (Fig. 4A) and several other doses (SI Appendix, Fig. S6) show that the XPC mutant and XPC knockout flies are quite sensitive compared to wild type. Next, we used the immuno-slot blot assay to measure loss of CPDs and (6-4)PPs following irradiation. Loss of (6-4)PP repair observed in the knockout flies was expected, since in other organisms, (6-4)PPs are repaired primarily by the XPC-dependent global repair pathway. Interestingly, CPD repair was also undetectable in XPC knockout flies, while clearly evident in wild type (Fig. 4 B and C). We then conducted excision assays using adults and embryos of the mutant and knockout strains. Surprisingly, and consistent with the slot blot results, we failed to detect any CPD or (6-4)PP excision product (Fig. 4 D and E and SI Appendix, Fig. S7). Finally, we conducted XR-seq using XPC adult knockout flies. Fig. 4F shows the frequency distribution profiles for each nucleotide at each position of the sequenced 28 mers. Clearly, the enrichment for pyrimidines at positions 21 to 22 seen in wild-type flies is absent in the knockouts, indicating that the 28 mers in the mutant are background fragments generated by shearing during DNA processing. Thus, in Drosophila, both global and TCR are dependent on XPC as is the case in S. cerevisiae and Schizosaccharomyces pombe (41)(42)(43)(44).
Tests for a Novel Protein with CSB Function in Drosophila. The presence of TCR in Drosophila, which lacks the CS homologs, prompted an investigation into a possible alternative coupling factor. By computational analysis we identified Arip4 and the Domino helicase, based upon homologies with human CSB, and by mass spectrometry we identified Domino, RecQ5, Mit2, lds, and Brm proteins as interacting with stalled RNAPII. Mutations in each candidate were made; however, individual mutations had no effect on TCR (SI Appendix, Fig. S8). Thus it appears that TCR in Drosophila applies a different molecular mechanism or that there might be a translocase not yet identified that might promote TCR in Drosophila.

Discussion
We initiated investigations of excision repair in Drosophila melanogaster to address the question of TCR in this organism. We previously reported that Drosophila S2 cells do perform TCR even though they lack the canonical TCR factors CSA and CSB (7). The present study confirms and extends our earlier finding to include the intact Drosophila organism in various stages of development, and both genders. Cisplatin-d(GpG) adducts are also shown in this work to be substrates for TCR in Drosophila, as they are in mice (18,45).
Overall, we find that D. melanogaster demonstrates considerable similarity in excision repair with other multicellular eukaryotes, including humans, mice, and C. elegans. Similarities include the existence of both global and transcription-coupled repair pathways, and relative rates of repair of different types of damage. The helix-distorting (6-4)PP damage is readily recognized and repaired by global repair in these organisms, and (6-4)PPs undergo global repair faster than TCR (1,3,16). The rate of TCR is presumably limited by the rate of transcription; TCR initiates when elongating RNAPII is blocked by damage in the template strand. TCR initially is the more rapid pathway for repair of CPDs and cisplatin adducts, because these damages are poor substrates for global repair (15,16). Similar to the other multicellular eukaryotes, Drosophila exhibits a positive association between the level of transcription and TCR, and the sites of 5 0 and 3 0 incision made by the excision nucleases are the same among the multicellular eukaryotes.
To study Drosophila TCR in the absence of global repair, we employed XPC mutant flies, since in all other multicellular eukaryotes studied, XPC is required for global repair but is dispensable for TCR (5,40). Our surprising finding that unlike other multicellular eukaryotes, Drosophila requires XPC for TCR is of exceptional interest. The requirement for XPC makes Drosophila similar to the two yeast species S. cerevisiae and S. pombe (46), the only other organisms known to require XPC for TCR. However, in contrast to Drosophila, yeast do possess orthologs of the CSA and CSB genes (46), which makes Drosophila unique in the components required for TCR.
This investigation raises the question of what the role is of XPC in TCR in Drosophila. The components necessary for TCR in eukaryotes have been identified and studied in considerable detail (9)(10)(11)(12); however, the mechanism of eukaryotic TCR is not well defined for the lack of an in vitro TCR system in eukaryotes. In global repair, damage recognition and assembly of preincision complexes have been shown to be cooperative, multienzyme, multistep processes in which XPA, RPA, and XPC participate in damage recognition (15,16,47,48). In addition, intermediate structures in both global repair and Fig. 4. XPC ortholog mus210 is required for repair in Drosophila. (A) Survival of wild type (WT, W 1118 ), mus210 mutant (XPC G1 ), and mus210 knockout (XPC KO ) adult flies without and with 4,800 J/m 2 UVB exposure. Group data were analyzed by two-way ANOVA (Tukey's multiple comparison test for more than two groups by using GraphPad Prism 8 software) and expressed as means ± SEM, n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 were considered to be statistically significant. XPC G1 males exhibited poor survival in the absence of UV and were not characterized due to this complication. (B) Representative slot blot showing CPD and (6-4)PP formation and repair in wild-type and XPC KO female flies. The Left shows anti-CPD and anti-(6-4)PP immunoreactivity, and anti-ssDNA reactivity is shown to the Right. (C) Plot of slot blot results. The CPD signal was normalized to the amount of ssDNA detected in each slot. Group data were analyzed by two-way ANOVA ( Sid ak's multiple comparisons test for two groups by using GraphPad Prism 8 software) and expressed as means ± SEM, n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 were considered to be statistically significant.  transcription exist in which the DNA substrate is melted or unwound in the vicinity of the substrate/damage (13,49). Thus in Drosophila, there are several steps in TCR in which XPC may be required; in addition to assisting RNAPII in damage recognition, XPC may be required for assembly of preincision complexes with appropriate repair protein partners, or generation and maintenance of the appropriate DNA structure in the context of a stalled RNAPII (50). Following these preincision events, incisions are made in the damaged strand by the XPG and XPF/ERCC1 nucleases. Additional investigation is needed to characterize the role of XPC in Drosophila TCR.
Fly Stocks. Drosophila stocks were kept at 25°C and fed with standard cornmeal medium. The Yw fly line was kindly provided by Mark Peifer, University of North Carolina at Chapel Hill, Chapel Hill, NC. W 1118 and XPC (mus210) G1 mutant fly lines were from the J.S. laboratory (31). The XPC G1 mutant files were from the W 1118 background.
CRISPR-Cas9 Deletion of XPC via Scarless Allele Replacement. The endogenous XPC gene was deleted and replaced with dsRed using CRISPR-Cas9 genome engineering and scarless allele replacement as illustrated in SI Appendix, Fig. S9 and described previously (51). A plasmid containing dsRed under the control of the eye-expressing 3xP3 promoter and flanked with DNA homologous to 5 0 and 3 0 XPC flanking sequences (pGEM ΔXPC), and another plasmid containing 5 0 and 3 0 XPC gRNAs (pCFD4 XPC gRNA), were simultaneously injected into Drosophila embryos expressing Cas9 in their germline stem cells under control of the nanos promoter (Genetivision). Upon eclosion of these embryos, male progeny were screened for expression of dsRed (and likely deletion of XPC) in their eyes using a fluorescent microscope. These males were then mated to Pin/CyO females to isolate and balance the suspected deletion. Deletions were then further screened via genomic extraction, PCR, and sequencing of parental flies used to establish the XPC deletion stock to confirm the deletion. All parents contained both the correct 5 0 and 3 0 flanking genomic sequence, indicating that XPC was successfully deleted and replaced with dsRed.
Adult Fly Survival Assay Following UVB Treatment. W 1118 , XPC G1 , and XPC KO mutant flies were counted and equal numbers of each were separated into individual vials. A total of 50 flies of each sex were irradiated with doses of UVB ranging from 0 to 9,600 J/m 2 . Flies were returned to their respective vials, and the number of surviving flies was recorded every 3 or 4 d. Flies were transferred to new tubes weekly. Survival was plotted to analyze sensitivity as a function of mutation status and sex. Three independent biological replicates were performed.
Excision Assay and XR-Seq. For in vitro excision and XR-seq assays of Drosophila, S2 cells were plated in 150-mm tissue culture plates and grown to about 80% confluence. For ultraviolet C (UVC) treatment, one 150-mm plate was irradiated with 20 J/m 2 UVC after medium was removed and then fresh medium was added. For cisplatin treatment, 3.3 mM cisplatin in 0.9% NaCl was added to the medium to dilute to 200 μM. Following predetermined repair times, plates were placed on ice and cells were harvested by scraping, washed with ice-cold PBS, pelleted, and transferred to Eppendorf tubes. The cells were then resuspended in 320 μl Tris-EDTA (TE). After the addition of 40 μL 10% sodium dodecyl sulfate (SDS), the samples were incubated at room temperature for 20 min. After incubation, 100 μL 5 M NaCl was added to the samples. The mixtures were inverted to mix and they were incubated at 4°C overnight.
In vivo excision and XR-seq assays were performed with Drosophila embryos, larva, pupa, and adults. For these experiments, Drosophila adults were isolated with fly food containing 5% sucrose and 6% agar for 2 h during which time they laid eggs. After the adults were removed, embryos were collected and irradiated with predetermined doses of UVB and then were allowed to repair in the dark for predetermined times. Larvae and pupae were transferred to 100-mm plates and irradiated with UVB and allowed to repair in the dark. After repair, samples (embryo and larva) were collected with ice-cold phosphate-buffered saline (PBS) and washed several times. A total of 320 μL TE pH 8.0 was added to each sample and samples were ground. After grinding, 40 μL 10% SDS was added to each sample and samples were incubated at 70°C for 30 min. After incubation, 100 μL 5 M NaCl was added to each sample and they were incubated at 4°C overnight.
Drosophila adults were collected on a CO 2 panel for UVB irradiation. Flies were then returned to their vials and allowed to repair in the dark for predetermined times. After their respective repair, pupa and adults were transferred to 15-mL Falcon tubes and frozen with liquid nitrogen. While frozen, they were ground in liquid nitrogen and the powder was transferred to individual Eppendorf tubes. Then, 320 μL TE pH 8.0 and 40 μL 10% SDS were added, and samples were incubated at 70°C for 30 min. After incubation, 100 μL 5 M NaCl was added to the samples and they were incubated at 4°C overnight.
After overnight incubation, samples from each developmental phase were centrifuged at high speed at 4°C for 1 h. Supernatants were processed for excision or XR-seq assay as described previously (4,14). Briefly, supernatants were incubated with 5 μL RNase A and then 5 μL Proteinase K, purified, and then immunoprecipitated with anti-CPD, anti-(6-4)PP, or anti-cisplatin modified DNA antibodies. Immunoprecipitates were then either 3 0 end (cordycepin) or 5 0 end (γ-[ 32 P]ATP) radiolabeled, or 3 0 end biotin labeled, and excision products were separated on 10 to 11% sequencing gels, or, for XR-seq assay, immunoprecipitates were ligated to the adaptors, purified by a second immunoprecipitation, and DNA damage was reversed (using photoreactivation for UV photoproducts, and NaCN treatment for cisplatin damage). Libraries were generated by PCR and then sequenced.
Bioinformatics Analysis. At least 6 million unique mapped reads were obtained for each sample. Obtained reads were aligned onto the D. melanogaster dm6 genome assembly. Genomic distributions of the XR-seq reads were visualized using the Integrative Genomics Viewer (IGV) as described previously (7). For plotting average repair profiles as a unit gene, we chose genes that did not overlap with any other gene and that were at least 1 kb long. With these criteria, the total number of selected genes was 6,218. Regarding developmental and sex-specific expression, expressed (RPKM > 10) genes were determined according to FlyBase RNA-seq data (37). Unit genes used to show TCR were formed by dividing each gene evenly into 100 bins from the TSS to the TES and adding 2-kb (25 bins) flanking regions to upstream of the TSS and downstream of the TES. RPKM for each bin was determined and the average RPKM of bins