UV irradiation stimulates histone acetylation and chromatin remodeling at a repressed yeast locus Yachuan Yu, Yumin Teng, Hairong Liu, Simon H. Reed, and Raymond Waters* Department of Pathology, School of Medicine, Cardiff University, Cardiff CF14 4XN, United Kingdom Edited by Philip C. Hanawalt, Stanford University, Stanford, CA, and approved April 26, 2005 (received for review February 23, 2005)
Chromatin immunoprecipitation with anti-acetyl histone H3 (K9 and K14) and anti-acetyl histone H4 (K5, K8, K12, and K16) antibodies shows that Lys-9 and兾or Lys-14 of histone H3, but not the relevant sites of histone H4 in nucleosomes at the repressed MFA2 promoter, are hyperacetylated after UV irradiation. This level of histone hyperacetylation diminishes gradually as repair proceeds. Accompanying this, chromatin in the promoter becomes more accessible to restriction enzymes after UV irradiation and returns to the pre-UV state gradually. UV-related histone hyperacetylation and chromatin remodeling in the MFA2 promoter depend on Gcn5p and partially on Swi2p, respectively. Deletion of GCN5, but not of SWI2, impairs repair of DNA damage in the MFA2 promoter. The post-UV histone modifications and chromatin remodeling at the repressed MFA2 promoter do not activate MFA2 transcriptionally, nor do they require damage recognition by Rad4p or Rad14p. Furthermore, we show that UV irradiation triggers genome-wide histone hyperacetylation at both histone H3 and H4. These experiments indicate that chromatin at a yeast repressed locus undergoes active change after UV radiation treatment and that failure to achieve histone H3 hyperacetylation impairs the repair of DNA damage. nucleotide 兩 excision repair 兩 Saccharomyces cerevisiae
P
ackaging DNA into chromatin constrains the genome into the cell nucleus and plays important roles in DNA metabolism. The dynamics of chromatin are finely regulated to control DNA function in response to various stimuli. During activation of gene expression the binding of transcriptional activators and coactivators to promoters results in perturbation of repressive chromatin in the promoter. These events facilitate access of subsequent incoming transcriptional factors to DNA. The changes in repressive chromatin often include chromatin remodeling by various complexes, e.g., SWI兾SNF, and acetylation of histones by histone acetyltransferases, e.g., Gcn5p (1, 2). The regulatory roles of histone acetylation and chromatin remodeling, although well documented, are largely confined to transcription initiation. Their roles in other events such as nucleotide excision repair (NER) that also operates on a chromatin template are only beginning to be explored. NER is a DNA repair pathway with ⬎30 proteins involved in removing damage from naked DNA (3, 4). Studies have shown that the overall repair of DNA damage by NER is less efficient in reconstituted nucleosomes than in naked DNA (5–7), giving an indication that nucleosomes on damaged DNA inhibit efficient NER. In vivo, early studies with human cells showed that increasing the histone acetylation level overall in chromatin by inhibiting histone deacetylase activities enhances repair synthesis during NER (8–12). Chromatin rearrangement during repair synthesis in NER was also observed as newly repaired DNA in human fibroblasts is more sensitive to nuclease than bulk DNA in chromatin. The nascent repair patch lacks a canonical nucleosome DNase I footprinting, and this is significantly reduced in nucleosome core DNA. These features rapidly change to that of canonical nucleosomes (12, 13). Therefore, it was proposed that chromatin rearrangement occurs during NER (12, 14, 15). More recent studies on how chromatin structure impinges on 8650 – 8655 兩 PNAS 兩 June 14, 2005 兩 vol. 102 兩 no. 24
NER revealed a modulation of repair by nucleosomes such that slower repair occurs in the ‘‘internal protected region’’ of the nucleosomes and faster repair occurs in the linker DNA (16, 17). Previously, we observed a role for the histone acetyltransferase Gcn5p in the NER of UV radiation-induced cyclobutane pyrimidine dimers (CPDs) in the transcriptionally active Saccharomyces cerevisiae MFA2 gene (18). The deletion of GCN5 reduced the efficiency of CPD repair in MFA2 but much less so in RPB2, whereas no detectable defect was observed for repair of the genome overall. As a result, we reasoned that the Gcn5p-related changes in local chromatin structure around MFA2 modulate the functionality of NER. However, because MFA2 was transcriptionally active, we could not rule out changes related to transcription reinitiation from those for repair. Here, we examined the repair of CPDs at the repressed MFA2 and investigated the status of histone acetylation and chromatin accessibility before and after UV irradiation. After UV irradiation, we found that without disturbing the repressed transcriptional status of MFA2, local chromatin undergoes histone H3 hyperacetylation and chromatin remodeling. These post-UV modifications in the repressive chromatin are independent of DNA damage recognition by Rad4p or Rad14p, yet the H3 hyperacetylation facilitates efficient damage removal. The data suggest that localized histone acetylation and chromatin remodeling occur to ensure efficient repair of DNA damage in repressive chromatin in vivo. Materials and Methods Yeast Strains. The yeast strains used were PSY316 (MAT␣ or a
ade2–101 ura 3–52 leu 2–3,112 ⌬his 3–200 lys2 trp1) and PSY316 gcn5⌬␣ (as PSY316␣). RAD4, RAD14, and SWI2 deletion from PSY316␣ was achieved by replacing the corresponding genomic genes with URA3 markers. The hemagglutinin (HA)-tagged TATA box-binding protein (TBP) strain was provided by K. Struhl (19) (W303–1A, MATa ade2–1 trp1–1 leu 2–3 his 3–11,15 ura 3–3). The corresponding ␣ strain was obtained by matingtype switch. Yeast strains were grown in yeast complete medium (yeast extract兾peptone兾dextrose). Nucleotide Excision Repair. UV radiation treatment of cells, DNA
and RNA isolation, Northern blotting, detection of CPDs at nucleotide resolution, and quantitative analysis of repair were performed as described in refs. 18 and 20. Chromatin Immunoprecipitation (ChIP). ChIP was undertaken as
described in ref. 21. Formaldehyde cross-linking was carried out at room temperature for 15 min. The cell lysate was sonicated to yield fragmented DNA of 400–1,000 bp in length. A sheared chromatin solution from 6 ⫻ 107 cells was precipitated in a total volume of 1 ml with 5 l of anti-acetyl histone H3 (at K9 and This paper was submitted directly (Track II) to the PNAS office. Abbreviations: ChIP, chromatin immunoprecipitation; CPD, cyclobutane pyrimidine dimer; NER, nucleotide excision repair; TBP, TATA box-binding protein. *To whom correspondence should be addressed. E-mail:
[email protected]. © 2005 by The National Academy of Sciences of the USA
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0501458102
K14; Upstate Biotechnology), 5 l of anti-acetyl histone H4 (at K5, K8, K12, and K16; Upstate Biotechnology), and 25 l of anti-HA (Santa Cruz Biotechnology) antibodies. The immunoprecipitation of TBP-DNA with anti-HA was performed as described in ref. 19. Appropriate amounts of pUC18 (New England Biolabs) were added to the eluted materials and the inputs before DNA extraction to serve as an external control for the following PCR and quantification. Quantitative PCR. Quantitative PCR was performed in real time
by using iQ SYBR Green Supermix (Bio-Rad) and appropriately diluted DNA in the Bio-Rad iCycler. PCR was performed in triplicate for each sample, and melting curves were executed to ensure single PCR products. PCR of pUC18 was performed by using primers from New England Biolabs (nos. S1233s and S1224s). The amplification efficiency of each primer set on damaged DNA templates was adjusted as PCR was carried out on both input and immunoprecipitated DNA. The primer sequences were, for nucleosome ⫺2, forward 5⬘-CGTTTATGGTATA A AT TAGA A A AGT TA A AGC-3⬘ and reverse 5⬘GGCGTCCTATGCATGCACTTAA-3⬘, and for nucleosome ⫺1, forward 5⬘-TGCATGTCAGAGGA A A A AGA ACAAAG-3⬘ and reverse 5⬘-CGGATGAACGACAGAAGAAGTGG-3⬘. To detect the TBP occupancy, the above primer set for nucleosome ⫺1 was used, which gives a PCR product covering the MFA2 TATA box. The centromere sequence (CEN14) was used as an internal control for ChIP. The primers were as follows: forward 5⬘GGT TCTAGT TAGTCACGTGCAG-3⬘ and reverse 5⬘CGTTATTTTACTTTTCGGAAGACA-3⬘. The primers used to amplify the INO1 promoter were as follows: forward 5⬘-TCTTCATCCTTCTTTCCCAG-3⬘ and reverse 5⬘-AGGTGACGAAAGCTCCAATT-3⬘. Histone Purification and Western Blotting. Total histones were
Chromatin Accessibility Assay. Chromatin was prepared as described in ref. 23. After being gently washed once with corresponding restriction enzyme reaction buffer, chromatin from 4 ⫻ 108 cells was incubated with 300 units of RsaI (or BsaAI) for 2 h at 37°C. Purified DNA from the digest was subjected to a secondary digest by HaeIII and then resolved on 2% agarose gels. Probing of the DNA fragment of interest was performed as described in ref. 23.
Results Gcn5p Is Required for Efficient NER of UV-Induced CPDs in the Repressed MFA2 Promoter. After our previous studies on the tran-
scriptionally active MFA2 gene (18), we investigated the repair of UV-induced CPDs in the MFA2 promoter (HaeIII restriction fragment, ⫺516 to ⫹83) in wild-type and gcn5⌬ ␣ cells where MFA2 is repressed and associated with positioned nucleosomes (23), enabling us to examine the involvement of Gcn5p in NER in the absence of transcription. Fig. 1A depicts typical gels examining NER at nucleotide resolution after 150 J兾m2 UV irradiation. The quantitative results are reported as the time needed to remove 50% of the initial CPDs (T50%) at each site (Fig. 1B). The results show that the rates of CPD repair in this fragment were heterogeneous both in wild-type and the gcn5⌬ mutant cells. However, repair did not follow an easily distinguishable nucleosome-orientated profile of ‘‘faster toward the linker and slower in the central nucleosomal DNA’’ as described Yu et al.
Fig. 1. Repair of CPDs in the MFA2 promoter in ␣-mating type strains. (A) Gels depicting CPDs in the transcribed strand and nontranscribed strand of HaeIII restriction fragment (⫺516 to ⫹83) in the repressed MFA2 promoter in wild-type (WT) and the gcn5⌬ ␣ strains after 150 J兾m2 UV irradiation. Lane U, DNA from nonirradiated cells; lanes 0 –3, DNA from irradiated cells after 0 –3 h of repair. Alongside the gels are symbols representing nucleosome positions, MFA2 upstream activating sequences, and the start site of MFA2 transcription (arrow). Nucleotide positions are allocated in relation to the MFA2 start codon. (B) Time to remove 50% of the initial CPDs (T50%) at given sites. T50% of a single CPD or a group of CPDs with a similar repair rate was calculated (⬍3 h) or extrapolated (⬎3 h) as described in ref. 18. The T50% of slowly repaired CPDs (T50% ⱖ 6 h) was shown at the same level (ⱖ6 h) on the graph.
in some cases (16, 17). This finding is consistent with our previous studies on MFA2 (18, 20). Deletion of GCN5 resulted in a slower repair at almost all of the CPD sites, irrespective of their location (Fig. 1B). We summed the total damage signal in each lane and calculated the T50% of the total CPDs. The T50% of all CPDs in the HaeIII fragment for the strand that becomes transcribed in a mating type cells (referred to hereafter as the transcribed strand) was 2.7 ⫾ 0.2 h in wild-type and 3.7 ⫾ 0.2 h in the gcn5⌬ mutant, and that for the other strand (termed PNAS 兩 June 14, 2005 兩 vol. 102 兩 no. 24 兩 8651
GENETICS
purified as described in ref. 22. Acetylated histones were obtained from log-phase cells after further growing in yeast extract兾peptone兾dextrose supplemented with 20 mM of sodium butyrate for 4 h. Protein electrophoresis on 15% polyacrylamideSDS gel and Western blotting were performed according to the standard procedure.
Fig. 2. Post-UV histone acetylation. The level of acetylation is presented as the fold increase relative to that in unirradiated wild-type ␣ cells. At the top A, B, and C are the acetylated sites. PCR was performed to amplify a region of 135 bp in nucleosome ⫺2 and a region of 122 bp in nucleosome ⫺1 in the MFA2 promoter. These two close regions gave the same result. (A and B) The acetylation of histones H3 and H4 in the repressed MFA2 promoter before and after 150 J兾m2 UV irradiation. (C) The acetylation of histone H3 in the repressed MFA2 promoter of wild-type cells in response to 75 J兾m2 UV irradiation. (D) Total histone acetylation in response to 150 J兾m2 UV irradiation. (Top) Gels stained with Coomassie brilliant blue to indicate total histones. (Middle and Bottom) Western blots probed with either anti-acetyl histone H3 (K9 and K14; Upstate Biotechnology) or anti-acetyl histone H4 (K5, K8, K12, and K16; Upstate Biotechnology). Lane Ac⫹, positive control for the acetylated histones from sodium butyrate-treated cells (see Materials and Methods).
hereafter the nontranscribed strand) was 3.3 ⫾ 0.2 h in wild type and 5.0 ⫾ 0.4 h in the gcn5⌬ mutant. Thus deletion of GCN5 impairs, but does not prevent, the NER of CPDs in the repressed MFA2 promoter. UV Irradiation Stimulates Histone Hyperacetylation: Gcn5p Is Required for the Post-UV Histone H3 Hyperacetylation in the Repressed MFA2 Promoter. We previously demonstrated that deletion of GCN5
does not impair the total cellular capability of NER (18). Gcn5p has been documented to possess a histone acetyltransferase activity with specificity on the Lys residues of histone tails (24, 25). Therefore, we monitored histone acetylation levels at the two nucleosomes in the repressed MFA2 promoter before and after UV irradiation (150 J兾m2) by ChIP with antibodies against acetylated histone H3 (at K9 and K14) and H4 (at K5, K8, K12, and K16). The histone acetylation at the active MFA2 promoter (in wild-type a cells) without UV radiation treatment was determined to indicate a state of histone hyperacetylation. We also examined the histone acetylation level (histones H3 and H4) in the centromere sequence, CEN14. The CEN14 sequence is packed in a compact chromatin structure termed the kinetochore, where UV-induced DNA damage is heavily resistant to NER (26). The histone acetylation level at CEN14 was constant before and after UV irradiation and served as an internal control for the examination of the histone acetylation level at MFA2. The repressed MFA2 exhibited hypoacetylation of histone H3 and H4 (nominally set as 1-fold), and the active MFA2 had a hyperacetylation of histone H3 (4.9 ⫾ 0.9-fold) but a basal acetylation of histone H4 (1.4 ⫾ 0.3-fold) at its promoter (Fig. 2 A and B). Immediately after UV irradiation (repair time of 0 h), neither the acetylation level on histone H3 nor that on 8652 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0501458102
histone H4 at the repressed MFA2 promoter showed any change. Whether Gcn5p was present did not affect the level of histone H3 and H4 acetylation at this stage. During the post-UV period, we observed a rapid increase of histone H3 acetylation starting at 10 min (5.1 ⫾ 1.0-fold) and reaching its peak level (9.9 ⫾ 0.4-fold) at 30 min in wild-type cells. As repair proceeds, the level of histone H3 acetylation decreased gradually to 8.6 ⫾ 0.7-fold at 1 h, 6.9 ⫾ 1.3-fold at 2 h, and 5.1 ⫾ 0.7-fold at 3 h. This UV-stimulated histone H3 acetylation is primarily Gcn5pdependent, because in the gcn5⌬ mutant histone H3 acetylation was only stimulated to 2.4 ⫾ 0.4-fold maximally 30 min after UV irradiation (Fig. 2 A). We did not observe the same increase after UV irradiation for histone H4 acetylation (Fig. 2B), and there was no difference in this response between wild type and the gcn5⌬ mutant. This result is not due to the antibody’s inability to detect an increase in H4 acetylation by ChIP, because the signal increases as reported (27) at the INO1 promoter when the histone deacetylase gene RPD3 is deleted (data not shown). UV-stimulated histone H3 hyperacetylation in the repressed MFA2 promoter depended on the dose of UV radiation (Fig. 2C). Thirty minutes after a lower UV radiation dose of 75 J兾m2, the acetylation of histone H3 in the MFA2 promoter was only stimulated to a level of 3.8 ⫾ 0.3-fold of the pre-UV level and dropped gradually to 1.7 ⫾ 0.4-fold after 3 h. UV radiation treatment also stimulated genome-wide histone hyperacetylation. Western blotting of purified histones (Fig. 2D) showed that both histone H3 and H4 are hyperacetylated after UV irradiation (150 J兾m2), and the acetylation returns to the pre-UV level gradually after UV radiation. This finding indicates that UV radiation also stimulates histone hyperacetylation at other loci in the genome. Deletion of GCN5 did not influence the UVYu et al.
Fig. 3. The accessibility of the RsaI site in chromatin in the repressed MFA2 promoter in response to 150 J兾m2 (A) and 400 J兾m2 (B) UV irradiation. Lane ⫺, naked DNA digested by HaeIII only; lane ⫹, naked DNA digested by both HaeIII and RsaI. The remainders are recovered DNAs after the sequential restriction enzyme digestion of chromatin (see Materials and Methods) from unirradiated (U) and UV-irradiated cells after various repair times in hours. Below the gels are the quantitative data presented as the percentage of chromatin being cut by RsaI.
stimulated genome-wide histone hyperacetylation significantly (Fig. 2D). This finding suggests Gcn5p is only responsible for the histone hyperacetylation in some domains (e.g., MFA2) in response to UV irradiation, similar to its pattern in transcription activation (28). A role for Gcn5p in the transcription of an obscure NER factor is unlikely because repair is not reduced in the genome overall, or at RBP2 (18), despite a modest UVinduced increase in histone acetylation that is not Gcn5pdependent at this latter gene (data not shown). Chromatin at the MFA2 Promoter Becomes More Accessible to Restriction Enzymes After UV. A molecular link has been reported
between chromatin remodeling and histone acetylation during gene activation (29, 30). Thus, we examined whether the chromatin in the repressed MFA2 promoter also undergoes remodeling after UV irradiation. We assessed the accessibility in chromatin of a restriction site (RsaI, ⫺337 relative to the start codon of MFA2) in the nucleosome core DNA at the MFA2 promoter. Before UV irradiation when nucleosomes are positioned (23), the RsaI recognition site was masked and only ⬇10% of the total chromatin was cut by RsaI (Fig. 3A, lane U). This site does not become more accessible immediately after 150 J兾m2 UV irradiation (Fig. 3A, lane 0). However, the accessibility of this site to RsaI increased gradually to ⬇40% by 2 h post-UV (Fig. 3A, lane 2). This level then dropped toward the pre-UV level (Fig. 3A, lane 3). Deletion of GCN5 had little effect on the increase of DNA accessibility in chromatin after UV irradiation Yu et al.
(Fig. 3A). Because the fraction of cells bearing CPDs in the region around the RsaI site is limited after a UV radiation dose of 150 J兾m2, we applied a higher UV radiation dose of 400 J兾m2 and then examined the accessibility of the same site in chromatin. The result (Fig. 3B) shows the same trends, but the extent of accessibility increased compared with after 150 J兾m2 UV irradiation. The DNA in chromatin became most accessible 2.5 h after UV irradiation in both wild-type and gcn5⌬ cells, with 65% of the chromatin being cut by RsaI before returning to the pre-UV state. Fig. 3B also shows that Swi2p contributed to the increase of DNA accessibility in chromatin in the MFA2 promoter after UV irradiation. The swi2⌬ mutant had 30% of the DNA in chromatin cut by RsaI, and this level remained unchanged up to 6 h after UV irradiation. However, the swi2⌬ mutant is not more UV-radiation-sensitive than wild type (data not shown), and deletion of SWI2 does not affect the repair of CPDs in the MFA2 promoter significantly. After 150 J兾m2 UV irradiation, the T50% of all CPDs in the HaeIII fragment in the swi2⌬ mutant was on average 2.6 ⫾ 0.2 for the transcribed strand and 3.2 ⫾ 0.3 for the nontranscribed strand, with no discernable differences at specific CPDs within the fragment (gel images not shown). This repair rate is similar to that in wild type (see above). Nevertheless, we cannot exclude a role for Swi2p in NER at MFA2 because we have not dramatically reduced chromatin accessibility in the swi2⌬ mutant, so SWI兾SNF factors may yet be linked to NER at this locus. We also tested the accessibility of another site (BsaAI) in nucleosome ⫺1 after UV irradiation (400 J兾m2) in wild-type cells. The result indicated this site also became more accessible after UV irradiation and returned toward the pre-UV level during repair (data not shown). Post-UV Chromatin Modifications in the Repressed MFA2 Promoter Do Not Trigger Transcriptional Activation of MFA2. Alteration of chro-
matin around the promoters of many genes is generally accompanied with changes of their transcriptional activity. To test whether the post-UV chromatin modifications (histone H3 hyperacetylation and chromatin remodeling) in the repressive MFA2 promoter are associated with the activation of MFA2 transcription, we monitored MFA2 mRNA before and after UV PNAS 兩 June 14, 2005 兩 vol. 102 兩 no. 24 兩 8653
GENETICS
Fig. 4. Transcription of MFA2 after UV radiation. (A) Northern blots of MFA2 mRNA. Lane A, RNA from a mating type cells where MFA2 is transcribed; lane U: RNA from unirradiated ␣ cells; lanes 0, 0.5, 1, and 2, RNA from UV-irradiated (150 J兾m2) cells after various repair times as indicated in hours. (B) The occupancy of TBP at the MFA2 promoter. Lane A (active), binding of TBP at the MFA2 promoter when MFA2 is active; lane U, occupancy of TBP at the repressed MFA2 promoter in ␣ cells; lanes 0, 1, and 2, after UV following the repair times in hours.
reductions in the histone H3 acetylation level and the accessibility of the RsaI site during post-UV time. This finding indicates that the post-UV chromatin modifications in the repressed MFA2 promoter do not occur solely when Rad4p or Rad14p is recruited for NER. However, the restoration of chromatin to its pre-UV state requires these proteins.
Fig. 5. Histone H3 acetylation and accessibility of the RsaI site in chromatin at the repressed MFA2 promoter in the rad4⌬ and rad14⌬ ␣ mutants. (A) Histone H3 acetylation (K9 and K14) in response to 150 J兾m2 UV irradiation. The histone H3 acetylation level is presented as the fold increase relative to that before UV irradiation (U). The remaining samples are from UV-treated cells after repair from 0 –3 h. (B) Accessibility of DNA in chromatin to RsaI in response to 400 J兾m2 UV irradiation. The accessibility of DNA in chromatin is represented by the percentage of chromatin sensitive to RsaI at the MFA2 promoter. Lane ⫺, naked DNA digested by HaeIII only; lane ⫹, naked DNA digested by both HaeIII and RsaI; lane U, chromatin sample with no UV; lanes 0 – 6, chromatin samples from cells receiving 400 J兾m2 UV irradiation after various repair times in hours.
irradiation. No increase occurred after UV irradiation as shown by Northern blotting (Fig. 4A). We also examined the occupancy of TBP at this promoter before and after UV irradiation. MFA2’s transcription requires TBP as evidenced by its 20-fold increase in binding when the gene is active (Fig. 4B). Furthermore, mutation in the MFA2 TATA box abolished TBP binding and 98% of detectable transcription (data not shown). The results (Fig. 4B) indicate that the occupancy of TBP at the repressed MFA2 promoter remained unchanged up to 2 h after 150 J兾m2 UV irradiation. This finding indicates that although the MFA2 promoter is modified so that its chromatin becomes more accessible to restriction after UV irradiation, these changes do not promote the recruitment of TBP to initiate transcription. UV-Stimulated Chromatin Modifications Are Independent of Rad4p or Rad14p. Next, we questioned whether these post-UV chromatin
modifications occur in the repressed MFA2 promoter before the recognition of damage for NER or whether they happen only in steps during NER. Rad4p and Rad14p are proposed to be involved in the early damage recognition steps during NER (31, 32). Fig. 5 shows that in the NER-defective rad4⌬ and rad14⌬ mutants, histone H3 was still hyperacetylated, and after UV irradiation the RsaI site in chromatin also became more accessible. The extent of both is lower than that in wild-type cells: Histone H3 acetylation was stimulated to ⬇6-fold of the pre-UV level, and ⬇50% of the DNA in chromatin was cut by RsaI maximally in these mutants, whereas in wild type these levels are ⬇10-fold and 65%, respectively. Nevertheless, extensive levels of both events are induced by UV irradiation in these NERdeficient cells. Neither the rad4⌬ nor the rad14⌬ exhibited any 8654 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0501458102
Discussion Efficient repair of DNA damage is vital for maintaining the integrity of the genome (33). The NER mechanism on naked DNA has been well elucidated (3, 4). However, the packaging of DNA into chromatin in eukaryotic cells provides a considerably different template for NER. Central to this is the question of whether NER can operate on nucleosomal DNA efficiently without modifying and兾or disrupting this closely contacted histone DNA complex. In vitro studies indicated that NER is slow or nonexistent on the nucleosomal surface (34, 35). However, numerous in vivo studies have observed repair of different forms of DNA damage by NER in transcriptionally inactive loci where chromatin generally maintains a repressive state (20, 36). This apparent discrepancy between the efficient repair of DNA damage in repressive chromatin and the physical barrier of repressive chromatin itself leads to a model of ‘‘unfolding– access–refolding’’ during NER (12, 14–15). Our studies here provide evidence supporting this. The modifications in repressive chromatin after UV irradiation include both histone hyperacetylation and the DNA in chromatin becoming more accessible. Failure to achieve the former considerably reduced the efficiency of CPD repair within MFA2 chromatin. However, a 50% reduction in restriction access in the swi2⌬ strain did not reduce NER significantly. The remodeling events examined at MFA2 do not require Rad4p or Rad14p, considered to be involved in the early steps in NER on naked DNA, and which are essential for NER. The actual state of histone acetylation at a specific locus is a balanced result of histone acetylation and deacetylation. Upon removal of repressors, reversal of associated deacetylation to a basal status of histone acetylation occurs at 5–8 min and finishes by 12–15 min (37). An elevated histone H3 acetylation level is observed 10 min after UV irradiation at the MFA2 promoter, and its peak is observed at 30 min. Comparing this with the repair of CPDs in this region, it is evident that histone H3 hyperacetylation is switched on soon after UV irradiation. During the transcriptional activation of genes, the Gcn5-mediated histone acetylation and SWI兾SNF-mediated chromatin remodeling have been found to be linked (29, 30). Our data indicate that both processes occur at the repressive MFA2 promoter in response to UV irradiation, and histone H3 hyperacetylation happens before chromatin remodeling (Figs. 2 and 3). However, deletion of GCN5 only affects post-UV histone H3 hyperacetylation but has little effect on the accessibility of chromatin in the MFA2 promoter after UV irradiation (Fig. 3 A and B). This finding suggests either that the mechanistic link of these two events under gene activation may not likewise exist in the cellular response after UV irradiation or that other forms of modifications at histone tails may compensate for the loss of Gcn5prelated histone H3 acetylation to anchor and stabilize chromatin remodeling complexes for their possible roles in NER. Our data also indicate that the post-UV chromatin modifications in the MFA2 promoter relate to the frequency of local DNA damage. First, despite the fact that all cells receive UV radiation at the doses used, both histone H3 acetylation and the DNA accessibility in chromatin in the MFA2 promoter depend on the dose of UV radiation (Fig. 3 B and C). Second, after the post-UV increase in histone acetylation and chromatin accessibility, their tendency to return to the pre-UV states correlates well with the repair progress. This occurs at times when much of the genome still contains damage. Rubbi and Milner (38) reported UVYu et al.
stimulated H3 acetylation in human cells at the level of the whole genome, and they proposed that it is global and is needed for all GGR, whereas here we examined local events and identified a specific acetyltransferase with a role limited to part of the genome. The exact roles of the post-UV chromatin modifications remain to be elucidated. First, Gcn5p-mediated H3 hyperacetylation is not absolutely required for NER because the repair of CPDs still occurs, albeit at a slower rate, at the MFA2 promoter in the gcn5⌬ mutant. However, there is some increase in histone H3 acetylation after UV irradiation in this mutant. Studies employing multiple histone acetyltransferase mutants may indicate whether histone H3 hyperacetylation itself is necessary for NER. Second, these modifications are unlikely to simply overcome the physical barrier of chromatin because deletion of GCN5 does not affect the extent and speed of chromatin becoming accessible to restriction, yet it affects the efficiency of NER. It is possible that the Gcn5p-mediated histone acetylation operates at a different level compared with SWI兾SNF-related chromatin-remodeling events. Third, with chromatin being modified extensively at the MFA2 promoter, one transcriptional factor (TBP) is still excluded from this site. This indicates that modified histones might either be forming an epigenetic code as proposed during gene regulation (39) or are part of a signaling pathway between proteins (40), both of which might be specific for NER. Finally, the role of Swi2p and its associated chromatinremodeling function in NER is still not clear. In vitro studies showed that SWI兾SNF is able to shift nucleosome away from
UV-damaged DNA and that SWI兾SNF stimulates CPD removal by UV photolyase (41). However, it was reported to only stimulate the NER of certain forms of DNA damage, i.e., 6-4 photoproduct, but not CPDs (42). Here, we demonstrated that Swi2p contributes to the chromatin remodeling at the MFA2 promoter after UV irradiation, but the swi2⌬ mutant, unlike gcn5⌬, is not sensitive to UV irradiation, and it repairs CPDs at the MFA2 promoter as efficiently as wild type. It is possible that efficient NER of CPDs can still occur because chromatin accessibility at MFA2 in the swi2⌬ strain is only reduced to ⬇50% of wild type. Studies as to whether other SWI兾SNF factors also contribute to the NER at MFA2 are required, and it is possible that multiple mutations may be needed to reduce NER activity. Recent evidence indicates that histone modifications and chromatin remodeling occur in other DNA repair events. Phosphorylation of histone H2A (43) and hyperacetylation of histone H4 (44) occur during repair of DNA double-strand breaks. Our studies indicate that repressive chromatin undergoes modifications after UV irradiation, and the chromatin-modifying factors identified are known to also be involved in transcription. A leading question therefore remains: How do cells link individual processes (e.g., transcription and various DNA repair pathways) to chromatin modification and remodeling via specific channels so as to ensure the necessity of one biological function as required without interfering with the others?
1. Peterson, C. L. & Workman, J. L. (2000) Curr. Opin. Genet. Dev. 10, 187–192. 2. Berger, S. L. (2002) Curr. Opin. Genet. Dev. 12, 142–148. 3. de Laat, W. L., Jaspers, N. G. & Hoeijmakers, J. H. (1999) Genes Dev. 13, 768–785. 4. Reed, S. H. & Waters, R. (2003) in Nature Encyclopaedia of the Human Genome, ed. Cooper, D. N. (Nature Publishing Group, London), Vol. 2, pp. 148–154. 5. Wang, Z.G., Wu, X. H. & Friedberg, E. C. (1991) J. Biol. Chem. 266, 22472–22478. 6. Sugasawa, K., Masutani, C. & Hanaoka, F. (1993) J. Biol. Chem. 268, 9098–9104. 7. Ura, K., Araki, M., Saeki, H., Masutani, C., Ito, T., Iwai, S., Mizukoshi, T., Kaneda, Y. & Hanaoka, F. (2001) EMBO J. 20, 2004–2014. 8. Smerdon, M. J., Lan, S. Y., Calza, R. E. & Reeves, R. (1982) J. Biol. Chem. 257, 13441–13447. 9. Dresler, S. L. (1985) Biochemistry 24, 6861–6869. 10. Smith, P. J. (1986) Carcinogenesis 7, 423–429. 11. Ramanathan, B. & Smerdon, M. J. (1989) J. Biol. Chem. 264, 11026–11034. 12. Smerdon, M. J. & Conconi, A. (1999) Prog. Nucleic Acid Res. Mol. Biol. 62, 227–255. 13. Smerdon, M. J. & Lieberman, M. W. (1978) Proc. Natl. Acad. Sci. USA 75, 4238–4241. 14. Meijer, M. & Smerdon, M. J. (1999) BioEssays 21, 596–603. 15. Green, C. M. & Almouzni, G. (2002) EMBO Rep. 3, 28–33. 16. Wellinger, R. E. & Thoma, F. (1997) EMBO J. 16, 5046–5056. 17. Tijsterman, M., de Pril, R., Tasseron-de Jong, J. G. & Brouwer, J. (1999) Mol. Cell. Biol. 19, 934–940. 18. Teng, Y., Yu, Y. & Waters, R. (2002) J. Mol. Biol. 316, 489–499. 19. Kuras, L. & Struhl, K. (1999) Nature 399, 609–613. 20. Teng, Y., Li, S., Waters, R. & Reed, S. H. (1997) J. Mol. Biol. 267, 324–337. 21. Kuo, M. H. & Allis, C. D. (1999) Methods 19, 425–433. 22. Edmondson, D. G., Smith, M. M. & Roth, S. Y. (1996) Genes Dev. 10, 1247–1259. 23. Teng, Y., Yu, S. & Waters, R. (2001) Nucleic Acids Res. 29, E64–E64.
24. Grant, P. A., Duggan, L., Cote, J., Roberts, S. M., Brownell, J. E., Candau, R., Ohba, R., Owen-Hughes, T., Allis, C. D., Winston, F., et al. (1997) Genes Dev. 11, 1640–1650. 25. Grant, P. A., Eberharter, A., John, S., Cook, R. G., Turner, B. M. & Workman, J. L. (1999) J. Biol. Chem. 274, 5895–5900. 26. Capiaghi, C., Ho, T. V. & Thoma, F. (2004) Mol. Cell. Biol. 24, 6907–6918. 27. Suka, N., Suka, Y., Carmen, A. A., Wu, J. & Grunstein, M. (2001) Mol. Cell 8, 473–479. 28. Holstege, F. C. P., Jennings, E. G., Wyrick, J. J., Lee, T. I., Hengartner, C. J., Green, M. R., Golub, T. R., Lander, E. S. & Young, R. A. (1998) Cell 95, 717–728. 29. Hassan, A. H., Neely, K. E. & Workman, J. L. (2001) Cell 104, 817–827. 30. Neely, K. E. & Workman, J. L. (2002) Mol. Genet. Metab. 76, 1–5. 31. Guzder, S. N., Sung, P., Prakash, L. & Prakash, S. (1996) J. Biol. Chem. 271, 8903–8910. 32. Jansen, L. E. T., Verhage, R. A. & Brouwer, J. A. (1998) J. Biol. Chem. 273, 33111–33114. 33. Friedberg, E. C. (2001) Nat. Rev. Cancer 1, 22–33. 34. Thoma, F. (1999) EMBO J. 18, 6585–6598. 35. Ura, K. & Hayes, J. J. (2002) Eur. J. Biochem. 269, 2288–2293. 36. Verhage, R., Zeeman, A. M., de Groot, N., Gleig, F., Bang, D. D., van de Putte, P. & Brouwer, J. (1994) Mol. Cell. Biol. 14, 6135–6142. 37. Katan-Khaykovich, Y. & Struhl, K. (2002) Genes Dev. 16, 743–752. 38. Rubbi, C. P. & Milner, J. (2003) EMBO J. 22, 975–986. 39. Jenuwein, T. & Allis, C. D. (2001) Science 293, 1074–1080. 40. Schreiber, S. L. & Bernstein, B. E. (2002) Cell 111, 771–778. 41. Gaillard, H., Fitzgerald, D. J., Smith, C. L., Peterson, C. L., Richmond, T. J. & Thoma, F. (2003) J. Biol. Chem. 278, 17655–17663. 42. Hara, R. & Sancar, A. (2003) Mol. Cell. Biol. 23, 4121–4125. 43. Downs, J. A., Lowndes, N. F. & Jackson, S. P. (2000) Nature 408, 1001–1004. 44. Bird, A. W., Yu, D. Y., Pray-Grant, M. G., Qiu, Q., Harmon, K. E., Megee, P. C., Grant, P. A., Smith, M. M. & Christman, M. F. (2002) Nature 419, 411–415.
Yu et al.
PNAS 兩 June 14, 2005 兩 vol. 102 兩 no. 24 兩 8655
GENETICS
We thank colleagues in the Pathology Department for their intellectual inspiration and technical assistance. This work was supported by a Medical Research Council program award (to R.W.).
