To investigate the role of the nucleosome during repair of DNA damage in yeast, we screened for histone H2B mutants that were sensitive to UV irradiation. We have isolated a new mutant, htb1-3, that shows preferential sensitivity to UV-C. There is no detectable difference in bulk chromatin structure or in the number of UV-induced cis-syn cyclobutane pyrimidine dimers (CPD) between HTB1 and htb1-3 strains. These results suggest a specific effect of this histone H2B mutation in UV-induced DNA repair processes rather than a global effect on chromatin structure. We analyzed the UV sensitivity of double mutants that contained the htb1-3 mutation and mutations in genes from each of the three epistasis groups of RAD genes. The htb1-3 mutation enhanced UV-induced cell killing in rad1Δ and rad52Δ mutants but not in rad6Δ or rad18Δ mutants, which are defective in postreplicational DNA repair (PRR). When combined with other mutations that affect PRR, the histone mutation increased the UV sensitivity of strains with defects in either the error-prone (rev1Δ) or error-free (rad30Δ) branches of PRR, but did not enhance the UV sensitivity of a strain with a rad5Δ mutation. When combined with a ubc13Δ mutation, which is also epistatic with rad5Δ, the htb1-3 mutation enhanced UV-induced cell killing. These results suggest that histone H2B acts in a novel RAD5-dependent branch of PRR.
THE structure of chromatin is intimately linked to the function of the eukaryotic genome. The basic repeating unit of chromatin, the nucleosome, is assembled in a two-step process in which a tetramer of histones H3 and H4 is first deposited onto DNA, followed by the association of two H2A-H2B heterodimers (for review, see Wolffe 1998). Subsequent folding of nucleosome arrays then leads to multiple levels of chromatin compaction (Hayes and Hansen 2001). Chromatin is generally considered to present a barrier to processes that occur on DNA, and numerous studies have shown that events in transcription can be inhibited by the presence of nucleosomes. The repair of DNA damage also occurs in a chromatin context, but unlike transcription, the role of nucleosomes during repair processes remains less clear. However, a number of studies have revealed a mutual influence between chromatin structure and DNA repair, with the packaging of DNA into chromatin affecting both the acquisition as well as the repair of lesions induced by UV irradiation (for review, see Smerdon and Thoma 1998). Cis-syn cyclobutane pyrimidine dimers (CPDs) are formed around the dyad axis of the nucleosome and at sites where the minor groove of the DNA superhelix faces the histone octamer less frequently than in linker DNA (Liuet al. 2000). In addition, various steps in nucleotide excision repair (NER) are significantly inhibited by the presence of nucleosomes (Thoma 1999; Haraet al. 2000; Liu and Smerdon 2000).
During the process of transcriptional activation, chromatin is frequently remodeled and/or covalently modified through the activity of evolutionarily conserved remodeling factors (Kornberg and Lorch 1999; Travers 1999; Tyler and Kadonaga 1999). Recent studies suggest that activities similar to those used in transcription may also facilitate DNA repair in chromatin (for review, see Meijer and Smerdon 1999). In mammalian cells, both ATP-dependent nucleosome remodeling factors and histone-modifying enzymes have been associated with the increased accessibility of chromatin templates during NER (Brandet al. 2001; Uraet al. 2001). In Saccharomyces cerevisiae, mutations in the ATP-dependent nucleosome remodeling complex, Ino80, confer hypersensitivity to a wide range of DNA-damaging agents, suggesting that this complex plays a direct role in altering chromatin structure during DNA repair (Shenet al. 2000).
Besides nucleosome remodeling factors, two evolutionarily conserved chromatin assembly factors, antisilencing factor 1 (ASF1) and chromatin assembly factor 1 (CAF-I), have also been implicated in DNA repair. In yeast, mutations in ASF1 cause hypersensitivity to double-strand breaks (DSBs), but not to UV irradiation, while CAF-I mutations confer UV sensitivity in preference to other types of damage (Kaufmanet al. 1997; Game and Kaufman 1999; Emiliet al. 2001; Huet al. 2001). Both factors assemble acetylated forms of histones H3 and H4 into nucleosomes during DNA replication, raising the possibility that they perform a similar role on newly repaired DNA (Smith and Stillman 1991b; Verreaultet al. 1996). Consistent with this view, CAF-I is recruited onto DNA after UV irradiation of human cells and promotes extensive nucleosome assembly during the repair of a damaged template in vitro (Gaillardet al. 1996; Martiniet al. 1998). These data suggest that nucleosomes are disassembled in the vicinity of DNA damage.
In eukaryotes, various types of DNA damage are repaired by specific mechanisms. On the basis of genetic epistasis analysis, the genes of S. cerevisiae that confer resistance to DNA-damaging agents have been assigned to three major groups (for reviews, see Friedberget al. 1995; Game 2000). The RAD3 group controls NER, which is responsible for the excision of UV-induced pyrimidine dimers or other bulky adducts. The RAD52 group repairs double-strand breaks induced by ionizing radiation and other kinds of damage and mediates homologous recombination. Genes in the third group, which have more complex roles and are less well understood, show an epistatic relationship with RAD6 (for review, see Kunzet al. 2000). The RAD6 group repairs or bypasses multiple forms of DNA lesions during or after DNA synthesis and contains genes whose products function in DNA replication and protein ubiquitylation (Lawrence and Christensen 1976; Prakash 1981; Sunget al. 1988; Liefshitzet al. 1998). Genetic epistasis studies have also placed CAF-I in the RAD6-dependent postreplication repair (PRR) pathway, suggesting a role for chromatin assembly in this pathway (Game and Kaufman 1999).
In this study, we focused on the role of histone H2B in the repair of UV-induced DNA damage. The notion that individual histones play specific roles in DNA damage repair is supported by the observation that double-strand breaks in both human and yeast cells induce the phosphorylation of the C terminus of H2A (the H2A variant H2A.X in humans and the major H2A-1/H2A-2 isoforms in yeast; Rogakouet al. 1998; Chenet al. 2000; Downset al. 2000; Paullet al. 2000). Modification of this histone occurs rapidly in regions surrounding DSBs in human cells, suggesting that it might help to disrupt chromatin at these sites either by directly altering H2A-DNA interactions or by recruiting chromatin remodeling factors (Rogakouet al. 1998; Paullet al. 2000). Its role in chromatin disruption is supported by the observation that a yeast strain containing a mutation that mimics the phosphorylated form of H2A shows extensive nucleosome instability (Downset al. 2000). Since H2B-DNA interactions play an important role in stabilizing the nucleosome (Luger et al. 1997a,b; Whiteet al. 2001), we reasoned that H2B might also play a specific role in the modulation of chromatin structure during DNA repair. We have isolated a new mutant of histone H2B, htb1-3, which shows sensitivity to UV irradiation in preference to other genotoxic agents. The UV sensitivity of this mutant does not result from an increase in the number of CPD lesions formed by UV-C or from a global defect in chromatin stability. Genetic epistasis analysis showed that the H2B mutation affected the RAD6/RAD18-dependent PRR pathway and specifically a novel, RAD5-dependent sub-branch of this pathway. RAD5 encodes a RING finger protein with homology to the SNF2 family of ATPases, which have known roles in nucleosome destabilization (Hirschhornet al. 1992; Johnsonet al. 1992; Pollard and Peterson 1998). Previous studies have linked the activity of Rad5p to chromatin (Ulrich and Jentsch 2000), and our results suggest a role for histone H2B in its activity.
MATERIALS AND METHODS
Yeast strains and media: The S. cerevisiae strains used in this study are listed in Table 1 and are isogenic to a W303 strain in which the rad5-535 allele had been corrected to wild type (obtained from H. Klein). Each strain was derived from JR5-2A, which carries the frameshift alleles htb1-1 and htb2-1 and plasmid YCp50-HTB1 (Recht and Osley 1999). RAD genes were disrupted in strain JR5-2A or EM1 by transformation with linear fragments isolated from plasmids that contained marked deletion constructs: pHU249, ubc13::HIS3 (SacI-ApaI); pTW033, rad9::HIS3 (NotI); pSH87, rad5::URA3 (HindIII-EcoRI); pR30-2, rad30::URA3 (EcoRI);pREV1.6, rev1Δ::URA3 (SphI); pR18.119,rad18Δ::LEU2 (HindIII-BamHI); pL962, rad1Δ::LEU2 (HindIII); pSM20, rad52Δ::LEU2 (BamHI); p46, rad6Δ::hisG::URA3::hisG (BamHI); pPK102, cac1Δ::hisG::URA3:: hisG (BamHI). The presence of the disruptions was confirmed by assaying for expected levels of sensitivity to UV irradiation and other genotoxic agents and in several cases by rescue with a plasmid carrying the wild-type allele. Standard protocols were followed for preparation of yeast media and transformation (Adamset al. 1997).
Plasmids: Plasmids YCp50-HTB1 and pRS314-HTB1 have been described (Recht and Osley 1999). Both plasmids carry the HTB1 open reading frame (ORF) as a BstEII-NotI fragment under control of the wild-type HTA1-HTB1 promoter. pRS314ΔNotI-HTB1 was derived from pRS314-HTB1 by removal of a NotI-BamHI restriction fragment from the polylinker in pRS314. pRS314-htb1-3 was generated by targeting mutations to the HTB1 ORF in pRS314-HTB1 in two steps using the QuikChange site-directed mutagenesis kit from Stratagene (La Jolla, CA). The following mutagenic primers were used: V47F, 5′-CTCTTCTTACATTTACAAATTTTTGAAGCAAACTCACCC-3′ and 5′-GGGTGAGTTTGCTTCAAAAATTTGTAAATGTAAGAAGAG-3′; and Y86F-N87S, 5′-CTAAATTGGCTGCGTTTAGCAAGAAGTCTACTATC-3′ and 5′-GATAGTAGACTTCTTGCTAAACGCAGCCAATTTAG-3′. Construction of Flag epitope-tagged pRS314-HTB1 has been described (Recht and Osley 1999). Flag epitope-tagged pRS314-htb1-3 was obtained by targeting the V47F, Y86F, N87S mutations to Flag-tagged HTB1 in pRS314. The presence of the mutations was confirmed by DNA sequence analysis. pRS324ΔNotI-HTB1 was obtained by inserting a NotI-BamHI fragment from pRS314-HTB1 into NotI-BamHI-digested plasmid pRS324ΔNotI.
Mutagenesis of HTB1: To obtain UV-sensitive htb1 alleles, we adapted a method that is based on the low fidelity of Taq DNA polymerase (Hirschhornet al. 1995). The HTB1 ORF was amplified from pRS314-HTB1 using primers to generate a PCR product extending from 116 bp upstream of the HTB1 ATG (5′-CTCAGATGGTCAGATTTATTA-3′) to 147 bp downstream of the HTB1 termination codon (5′-ATTTCGAGAACACAATTTTAC-3′). PCR reactions were performed with 100 ng of plasmid DNA in the presence of 0.25 mm MnCl2 and 7.5 mm MgCl2 to generate an average of five to seven base changes per HTB1 ORF. The PCR products were cotransformed into strain JR5-2A [Ycp50-HTB1] along with gel-purified plasmid pRS314-HTB1 that had been digested with BstEII-NotI to release the HTB1 ORF. Following gap repair in vivo (Hirschhornet al. 1995), Trp+ Ura+ transformants were selected. The transformants were then patched onto 5-fluoroorotic acid plates to identify viable cells that had lost plasmid YCp50-HTB1. Trp+ Ura− cells were screened for hypersensitivity to UV irradiation by exposure to 100 J/m2 of UV-C.
Measurement of 2μ plasmid DNA topoisomers: DNA was isolated from 10-ml YPD cultures grown to midlog as described (Morse 1999). Total DNA was electrophoresed through a 0.7% agarose gel in Tris-phosphate buffer (0.09 m Tris-phosphate, 0.002 m EDTA) containing 10 μg/ml of chloroquine at 50 V for 27 hr at 4°. DNA was transferred to Hybond C Extra membrane (Amersham, Piscataway, NJ), and plasmid DNA topoisomers were detected by hybridization to a 687-bp AvaI-XbaI fragment of 2μ DNA that was labeled by random priming. DNA topoisomer distributions were quantified and analyzed using Quantification One software from Bio-Rad (Richmond, CA).
Micrococcal nuclease digestion of chromatin: Nuclei were isolated according to the method of Bernardi et al. (1991). Spheroplasts were prepared from 500 ml of cells grown to midlog phase in YPD following treatment with 0.5 mg/ml Zymolyase in 1 m sorbitol. Spheroplasts were lysed in 18% Ficoll, 20 mm potassium phosphate, 1 mm MgCl2, 0.25 mm EGTA, 0.25 mm EDTA pH 6.8, 1 mm phenylmethylsulfonyl fluoride (PMSF) and resuspended in 7 ml of 10 mm Tris-HCl pH 8, 150 mm NaCl, 5 mm KCl, 1 mm EDTA, and 1 mm PMSF. Aliquots of isolated nuclei were immediately subjected to micrococcal nuclease (MNase) digestion with indicated amounts of MNase in the presence of 5 mm CaCl2 for 5 min at 37°. DNA was purified by proteinase K treatment, phenol-chloroform extraction, and ethanol precipitation. DNA was subjected to electrophoresis through a 1.5% agarose gel in Tris-borate buffer (0.09 m Tris-borate, 0.001 m EDTA). The MNase profile of total genomic DNA was visualized by ethidium bromide staining, and DNA was then transferred to Hybond C Extra membrane. The MNase profiles of the SUC2 and HIS3 loci were measured by hybridization to a SUC2 probe fragment obtained by PCR (+133 to +770) or to a 1.8-kb HIS3 BamHI fragment that contained the HIS3 ORF + 5′ flanking sequences, both labeled by random priming.
Measurement of CPD lesions: Twenty-milliliter cultures of HTB1 and htb1-3 strains were grown to midlog phase in YPD, washed with water, and resuspended in 30 ml of phosphate-buffered saline (PBS). Three 8.5-cm diameter petri dishes containing 10 ml of the cell suspension were kept on ice and irradiated with 0, 30, or 150 J/m2 of UV-C. Six milliliters of UV-irradiated or control cells were then incubated in 50% ethanol plus 12 mm EDTA. DNA was extracted and purified as described (Adamset al. 1997) and quantified using a DyNA Quant fluorometer (Pharmacia Biotech, Piscataway, NJ). For each sample, aliquots of 10 and 15 ng were adjusted to a final volume of 400 μl and treated as described (Bio-Dot user manual, Bio-Rad), and 200 μl were then slot-blotted in duplicate onto a 0.2-μm nitrocellulose membrane (PROTRAN, BA83; Schleicher & Schuell, Keene, NH). To detect CPDs, the blot was blocked overnight in 0.5% nonfat milk in TBS-Tween (136 mm NaCl, 2.7 mm KCl, 25 mm Tris-Cl pH 8.0 plus 0.1% Tween 20) and then incubated with anti-CPD antibodies (1:250 dilution of clone KTM53; Kamiya Biomedical, Thousand Oaks, CA) in 0.5% TBS, 0.02% Tween 20 for 1 hr at 37° (Perdizet al. 2000). Enhanced chemiluminescence was used to detect antigen-antibody complexes according to the manufacturer's directions (Amersham). The same blot was also hybridized to a probe obtained by random priming of bulk genomic DNA extracted from strain EM1. The number of CPD lesions was normalized to the amount of DNA present on the filters using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).
DNA damage sensitivity assays: Exponentially growing cultures of wild-type and mutant strains were grown in YPD to ~6 × 106 cells/ml. UV survival was measured after spreading appropriate dilutions of the cultures in duplicate on YPD plates and then subjecting the plates to specific UV doses at 254 nm (Spectroline XX-15G lamp; Spectronics, Westbury, NY). Plates were incubated in the dark for 3 days at 30° and the number of colonies was counted. Each experiment presented here was repeated at least three times, and data from single representative experiments are shown. For each experiment, data points are the average of four determinations, and error bars represent the range of values (minimum and maximum) obtained in that experiment.
Bleomycin sensitivity was measured by spotting 3 μl of 10-fold serial dilutions of cells onto YPD plates containing 6 milliunits/ml bleomycin, followed by incubation for 2 days at 30°. Sensitivity to gamma irradiation was measured by harvesting 5 ml of culture, resuspending the cells in 1 ml of ice-cold PBS in a 1.5-ml Eppendorf tube, and then irradiating with 150 Gy of gamma irradiation using a 137Cs source (Mark 1, model 68; J. L. Shepherd & Associates, San Fernando, CA). After irradiation, cells were resuspended in 5 ml of water and 3 μl of 10-fold serial dilutions were spotted onto YPD plates. The plates were incubated at 30° for 2 days before counting survivors.
UV-induced mutagenesis: Exponentially growing cells were grown in YPD to ~6–10 × 106 cells/ml, washed, and resuspended in distilled water at a density of 5–10 × 107 cells/ml. Cells were plated in duplicate on YPD plates to determine viable cell number and on synthetic media lacking adenine or tryptophan to measure reversion of ade2-1 or trp1-1. After UV irradiation at doses of 2, 10, and 50 J/m2, the plates were incubated in the dark at 30°. Survival was measured after 3 days on YPD plates and reversion frequency was measured after 6 days on synthetic media.
Isolation of a new class of H2B mutants that are sensitive to UV irradiation: In S. cerevisiae, histone H2B is encoded by two unlinked genes, HTB1 and HTB2, which together are essential for cell viability (for review, see Osley 1991). We used a strain that carried frame-shift mutations in the genomic copies of both HTB genes and whose viability was maintained by a URA3/CEN plasmid that contained the HTB1 gene as the only source of H2B in the cell (Recht and Osley 1999). Using a PCR-based mutagenesis procedure that reduced the fidelity of Taq DNA polymerase (Hirschhornet al. 1995), we generated random mutations in the HTB1 ORF in vitro and screened in vivo for viable mutants that were hypersensitive to killing by UV-C. From ~700 colonies, we identified four recessive htb1 mutants that showed a similar hypersensitivity to UV irradiation (Figure 1 and data not shown). The UV sensitivity was moderate in comparison to the sensitivity exhibited by many of the known rad mutants.
DNA sequence analysis revealed that multiple mutations were present in the H2B ORF of each UV-sensitive mutant. A comparison of the amino acid changes did not identify residues that were commonly mutated in all four htb1 alleles, and we therefore focused on one mutant, htb1-16, which encoded five altered amino acids (V47F, N66D, Y86F, N87S, and S127P). None of the five mutations on its own was sufficient to confer UV sensitivity (data not shown), and only combination of the mutations V47F, Y86F, and N87S recapitulated the phenotype of the original htb1-16 mutant (Figure 1 and data not shown). Our subsequent studies utilized the mutant that contained these three changes, which we named htb1-3. Besides exhibiting UV sensitivity, this mutant, like htb1-16, also showed poor growth on YPD plates at 16° (data not shown).
We compared the UV sensitivity of strains carrying the htb1-3 mutation on either a CEN or multicopy plasmid. In both cases, similar levels of cell survival were seen after UV irradiation (Figure 2). These results suggested that the UV sensitivity of the htb1-3 mutant was not caused by a reduction in the cellular levels of histone H2B. In support of this conclusion, Western blot analysis of Flag epitope-tagged H2B isolated from a wild-type strain and the htb1-3 mutant revealed no significant differences in the levels of this histone (data not shown).
Chromatin structure in the htb1-3 mutant: Many forms of DNA lesions are preferentially targeted to linker DNA between nucleosomes, reflecting the greater accessibility of these regions to DNA damage (Kuo and Hsu 1978; Conconiet al. 1984; and for reviews see Smerdon and Thoma 1998; Thoma 1999). Thus, it was possible that the increased UV sensitivity of the htb1-3 mutant was due to a global disruption in chromatin structure that effectively increased the number of sites where lesions could occur. To test this possibility, we first examined the pattern of MNase digestion of bulk chromatin in strains carrying either a wild-type HTB1 gene or the htb1-3 allele. MNase preferentially cleaves chromatin in the linker regions between nucleosomes and thereby provides an indication of both the presence and spacing of nucleosomes in the chromatin fiber. The two strains showed no significant differences in the pattern of MNase digestion of bulk chromatin or of two specific loci, HIS3 and SUC2 (Figure 3). Next, we examined the ability of nucleosomes to supercoil DNA in vivo by analyzing the distribution of DNA topoisomers from the endogenous 2μ plasmid. Each time a nucleosome is assembled onto a closed circular DNA molecule, a single superhelical turn is introduced (Worcelet al. 1981). If the htb1-3 mutation led to nucleosome instability or loss, then we would expect to see a shift in the distribution of plasmid DNA topoisomers as analyzed by chloroquineagarose gel electrophoresis. Again, no apparent differences were detected between the two strains (Figure 4). Together, the results suggest that the htb1-3 mutation does not grossly disrupt bulk chromatin structure, although we cannot exclude the possibility that the mutation causes locus-specific alterations of chromatin.
Induction of cyclobutane pyrimidine dimers is not increased in the htb1-3 mutant: UV irradiation induces various kinds of lesions, the most abundant being CPDs (for review, see Friedberget al. 1995). To examine the effect of the htb1-3 allele on CPD formation, we measured the number of lesions induced in bulk genomic DNA immediately after UV irradiation. Using a Western blot analysis with monoclonal antibodies specific for CPDs (Perdizet al. 2000), we detected a UV-dependent increase in the number of CPDs in both HTB1 and htb1-3 strains (Figure 5). The anti-CPD blots were scanned and the results were normalized to the amount of DNA present on the filters (materials and methods). The results from three independent experiments indicated that there was no apparent difference in the number of CPDs present in wild-type and mutant strains (Figure 5 and data not shown). This suggests that the htb1-3 mutation does not significantly influence induction of CPDs by UV-C.
Sensitivity of the htb1-3 mutant to other genotoxic agents: Genotoxic agents other than UV irradiation induce additional forms of DNA lesions and lead to repair by alternate pathways. We therefore tested the survival of the htb1-3 mutant after exposure to ionizing irradiation, methyl methanesulfonate (MMS), or bleomycin (Figure 6 and data not shown). Each of these agents produces a broad spectrum of DNA damage that includes base damage, single-strand breaks (SSBs), and DSBs, lesions that are generally repaired by the RAD52-dependent recombinational repair pathway (Resnick and Martin 1976; and for review see Nickoloff and Hoekstra 1998). Both htb1-3 haploid and diploid cells behaved like wild-type cells and were resistant to a dose of gamma irradiation that caused >99.9% lethality in a rad9Δ mutant, which is hypersensitive to a wide range of DNA-damaging agents (Figure 6, top). Similar results were observed after incubation of both htb1-3 strains on plates that contained 0.02% MMS (data not shown). These results indicated that the htb1-3 mutant was not defective in DSB repair. Notably, the htb1-3 mutant was hypersensitive to bleomycin compared to the wild-type strain (Figure 6, bottom).
Epistasis analysis of the htb1-3 UV-sensitive phenotype: Many forms of DNA damage are repaired in yeast through the action of genes falling in three broad epistasis groups. To determine if the htb1-3 mutation affected the function of one of these three groups, we disrupted a gene from each group in both HTB1 and htb1-3 cells and compared the UV sensitivity of the double mutants to the corresponding single mutants. We first examined the relationship between htb1-3 and rad1Δ, which is defective in the endonuclease that incises DNA on the 5′ side of lesions during NER (Klein 1988; Schiestl and Prakash 1988; Friedberget al. 1995; Nickoloff and Hoekstra 1998). Exposure to UV-C enhanced the killing of a rad1Δ htb1-3 mutant compared to a rad1Δ mutant (Figure 7A), indicating that the htb1-3 mutation affects a pathway other than NER.
The RAD52 epistasis group is involved in recombinational repair of double-strand breaks induced by agents such as ionizing radiation as well as by high doses of UV-C (Resnick and Martin 1976; Mortensenet al. 1996; Nickoloff and Hoekstra 1998). We examined the epistatis relationship between rad52Δ and htb1-3 at doses of UV that give significant killing of rad52Δ cells. A rad52Δ htb1-3 double mutant showed sensitivity to high doses of UV irradiation greater than that of either a rad52Δ (Figure 7B) or htb1-3 (Figure 7D) single mutant. Together with the observation that the htb1-3 mutant was not hypersensitive to gamma irradiation, this result supports the view that the mutations in H2B do not affect the major pathway of recombinational repair.
The third epistasis group contains a heterogeneous collection of genes involved in PRR. On the basis of double mutant analysis, the most upstream gene in the PRR group is RAD6. RAD6 encodes a multifunctional ubiquitin-conjugating enzyme that targets unknown substrates during the repair of many different types of DNA damage, including damage induced by UV-C (Monteloneet al. 1981; Reynoldset al. 1985; Jentschet al. 1987; Sunget al. 1988). A rad6Δ htb1-3 mutant showed the same sensitivity to UV-C as a rad6Δ mutant (Figure 7C), indicating that histone H2B is a member of the RAD6 epistasis group. It is a formal possibility that we were unable to detect the effect of the H2B mutations because of the extreme UV sensitivity of rad6Δ mutants. However, we consider this unlikely because a rad1Δ mutant is as UV sensitive as a rad6Δ mutant, and we were able to detect enhanced killing of a rad1Δ htb1-3 mutant compared to a rad1Δ strain (compare Figure 7A and 7C). The assignment of H2B to the RAD6 epistasis group is further supported by the finding that a rad18Δ htb1-3 mutant was no more UV sensitive than a rad18Δ mutant (Figure 7C). RAD18 and RAD6 encode gene products that appear to function upstream of all other PRR functions. Rad18p has been shown to interact directly with Rad6p both in vivo and in vitro and has been proposed to target Rad6p to sites of DNA damage (Cassier-Chauvat and Fabre 1991; Baillyet al. 1994; Ulrich and Jentsch 2000). Together, the results indicate that histone H2B plays a role in RAD6/RAD18-dependent PRR.
The htb1-3 mutation affects a RAD5-dependent repair pathway: The PRR pathway is perhaps the least understood of the major DNA repair pathways. At least two subpathways of PRR can be distinguished on the basis of whether mutations are generated during the repair process itself (Xiaoet al. 2000). The error-prone PRR pathway includes the products of the REV1, REV3, and REV7 genes, which perform mutagenic translesion synthesis (Lawrence and Christensen 1976; Lawrence et al. 1984, 1985; Larimeret al. 1989; Bayntonet al. 1999). The error-free PRR pathway has less effect on damage-induced mutagenesis and has been proposed to contain two different sub-branches, defined by the RAD5 and RAD30 genes (Johnson et al. 1992, 1994, 1999; McDonaldet al. 1997; Roushet al. 1998). To determine if the htb1-3 mutation affected the error-prone or error-free PRR pathway, we examined the UV sensitivity of rev1Δ htb1-3, rad5Δ htb1-3, and rad30Δ htb1-3 double mutants (Figure 8). Like the htb1-3 mutant, rad30Δ and rev1Δ mutants are moderately sensitive to UV. However, both rev1Δ htb1-3 (Figure 8B) and rad30Δ htb1-3 (Figure 8C) double mutants exhibited a synergistic reduction in cell survival after UV irradiation. In contrast, the htb1-3 mutation did not enhance UV killing in a rad5Δ mutant (Figure 8A) or in a rad5Δ rad30Δ double mutant, which shows extreme sensitivity to UV (data not shown). Together, the results suggest that rad5Δ is epistatic to htb1-3, thus placing HTB1 in a RAD5-dependent branch of PRR.
The UBC13 and MMS2 genes, whose products form a ubiquitin-conjugating complex, also act in a RAD5-dependent branch of PRR (Broomfieldet al. 1998; Hofmann and Pickart 1999; Bruskyet al. 2000; Ulrich and Jentsch 2000). Rad5p is proposed to act upstream of the Ubc13p/Mms2p complex because a rad5Δ mutation is epistatic to both the ubc13Δ and mms2Δ mutations and because Rad5p is required to recruit the Ubc13p/Mms2p complex to chromatin after DNA damage (Ulrich and Jentsch 2000). To determine if the H2B mutations affected this particular activity of Rad5p, we measured the UV sensitivity of a ubc13Δ htb1-3 mutant (Figure 8D). Cell survival was reduced in the double mutant compared to either single mutant, suggesting that Rad5p-H2B and Rad5p-Ubc13p-Mms2p represent distinct sub-branches of RAD5-dependent PRR (see discussion).
Effect of the htb1-3 allele on UV-induced mutagenesis: UV-induced DNA damage that is not repaired by NER is bypassed by DNA polymerases that function in either a predominantly error-prone (Rev3/Rev7) or error-free (Rad30) mode (Johnsonet al. 1992; Lawrence and Hinkle 1996; McDonaldet al. 1997; Roushet al. 1998). Thus, mutations in REV1, REV3, or REV7 cause a marked reduction in the levels of UV-induced mutations, while mutation of RAD30 enhances, reduces, or has little effect on UV mutagenesis, depending on the locus examined (McDonaldet al. 1997; Roushet al. 1998; Rajpalet al. 2000). To determine if the htb1-3 allele affected UV mutagenesis, we measured reversion of the ade2-1 and trp1-1 alleles after exposure to different doses of UV-C (Figure 9). The frequency of Ade+ and Trp+ revertants in the histone mutant was similar to that of a wild-type strain. Thus the htb1-3 allele does not appear to significantly impair UV-induced mutagenesis.
UV sensitivity of a cac1Δ htb1-3 double mutant: CAF-I is the only other chromatin-associated factor known to participate specifically in the repair of UV-induced DNA damage. This evolutionarily conserved factor deposits histone H3-H4 tetramers onto DNA during both replication and nucleotide excision repair synthesis (Smith and Stillman 1991a,b; Kaufmanet al. 1995; Gaillardet al. 1996; Kamakakaet al. 1996; Martiniet al. 1998). CAF-I is not essential for cell viability in yeast, but deletion of any one of the three genes that encode its sub-units confers moderate sensitivity to UV irradiation (Enomotoet al. 1997; Kaufmanet al. 1997; Game and Kaufman 1999). Like H2B, CAF-I acts in the error-free pathway of RAD6-dependent PRR, but unlike H2B, it appears to function independently of RAD5 (Game and Kaufman 1999). To determine if H2B plays a role with CAF-I after UV irradiation, we measured survival of a cac1Δ htb1-3 double mutant following exposure to UV-C (Figure 7D). The double mutant showed reduced survival after UV irradiation compared to either a cac1Δ or htb1-3 mutant, suggesting that H2B and CAF-I act in different branches of PRR.
htb1-3, a UV-sensitive mutant of histone H2B: This study describes the characterization of a new mutant of histone H2B that exhibits hypersensitivity to UV irradiation, showing for the first time a link between a specific nucleosome constituent and the repair of UV-induced DNA damage. The UV sensitivity of this mutant results from a combination of three mutations in two structural domains—loop 1 (L1) and loop 2 (L2)—that are common to all four core histones. Both L1 and L2 are involved in the binding of DNA on the surface of the histone octamer (Lugeret al. 1997a; Whiteet al. 2001), and mutations in these domains of histones H3 and H4 confer nucleosome instability (Kurumizaka and Wolffe 1997; Wechseret al. 1997). L1 of yeast histone H2B is also predicted to participate in internucleosomal interactions that might contribute to chromatin compaction (Whiteet al. 2001). Since the presence of nucleosomes influences the formation of UV-induced lesions (Liuet al. 2000; Uraet al. 2001), we thought it likely that the H2B mutations might increase the overall accessibility of chromatin to DNA-damaging agents. However, we found no evidence that nucleosomes were less stable in the htb1-3 mutant or that the formation of CPD lesions was increased in this strain. This suggests that the particular amino acid changes in L1 and L2 of H2B do not lead to global chromatin opening, although we cannot exclude the possibility that they cause local changes in chromatin structure that influence lesion formation at discrete sites. Excision of CPDs and 6-4 photoproducts by the NER pathway is also strongly inhibited by the presence of nucleosomes (Liu and Smerdon 2000; Uraet al. 2001). Thus, a lack of nucleosome mobility in the htb1-3 mutant could interfere with the repair process itself. However, if this were the case, we would have expected the htb1-3 mutation to be epistatic with mutations in NER, which was not observed. It therefore appears that the htb1-3 allele does not significantly affect either the formation or the excision of UV-induced lesions on a global level.
H2B is in a novel RAD5-dependent branch of RAD6/RAD18-dependent postreplication repair: Genetic epistasis studies showed that the htb1-3 allele affects the RAD6/RAD18-dependent PRR pathway, which corrects DNA lesions through both error-free and error-prone functions. Moreover, a rad5Δ mutation was epistatic to the htb1-3 mutation (Figure 8A), placing H2B in a RAD5-dependent DNA repair pathway. The role of RAD5 in PRR is unclear, although recent genetic and biochemical studies suggest that one function of Rad5p is to recruit the Ubc13p/Mms2p ubiquitin-conjugating complex to chromatin (Ulrich and Jentsch 2000). This and other genetic evidence have led to the proposal that Rad5p-Ubc13p-Mms2p represents a distinct branch of PRR (Ulrich and Jentsch 2000). Since H2B is a major chromatin constituent, we anticipated that it might also act in this branch. However, the htb1-3 and ubc13Δ mutations were additive for UV sensitivity (Figure 8D), suggesting that H2B is involved in a novel RAD5-dependent branch of PRR that is separate from the Ubc13p-Mms2p branch (Figure 10).
H2B and chromatin assembly factor I act in distinct repair pathways: CAF-I was previously shown to play a role in the repair of UV-induced lesions (Kaufmanet al. 1997; Game and Kaufman 1999), and in this study we show that histone H2B also helps to protect against damage induced by UV. While both H2B and CAF-I act in PRR, it is likely that they perform different functions in this pathway. First, a cac1Δ htb1-3 mutant was more sensitive to UV than either single mutant, indicating that CAF-I and H2B act in different branches of PRR. Second, CAF-I mutations enhance the UV sensitivity of mutations in all the major PRR genes except RAD6 and RAD18 (Game and Kaufman 1999), whereas the htb1-3 mutation is epistatic with mutations in RAD5 as well as in RAD6 and RAD18. Third, CAF-I interacts exclusively with histones H3 and H4 in its assembly function (Smith and Stillman 1991b).
Distinguishing additive from synergistic effects of combining mutations in DNA repair factors: The observation of an additive effect when two null mutations are combined suggests that the genes involved affect different pathways acting on different lesions, whereas a synergistic effect suggests that the genes either belong in the same pathway or belong in different pathways that work on the same lesion (Haynes and Kunz 1981). Previous work described a synergistic increase in UV sensitivity when the rad5Δ mutation was combined with either the rev1Δ or rad30Δ mutation (Johnsonet al. 1992; McDonaldet al. 1997; Xiaoet al. 2000). On the basis of this observation, RAD5 and RAD30 have been proposed to be involved in the same pathway or in two overlapping pathways (McDonaldet al. 1997). mms2Δ and ubc13Δ mutations show an epistatic relationship with rad5Δ in response to UV irradiation but an additive increase in UV sensitivity when combined with the rad30Δ mutation (Ulrich and Jentsch 2000). These results suggest that Rad30p and the Ubc13p-Mms2p complex are involved in two completely different subbranches of the PRR pathway and also support the idea that Rad5p plays a dual role during the repair of UV-induced lesions (Cejkaet al. 2001).
In this study, we show that (1) like rad5Δ, htb1-3 shows a synergistic increase in UV sensitivity when combined with rad30Δ; (2) like rad30Δ but in contrast to rad5Δ, htb1-3 shows an additive increase in UV sensitivity when combined with ubc13Δ; and (3) an htb1-3 rad30Δ rad5Δ triple mutant is not more UV sensitive than a rad5Δ rad30Δ double mutant (data not shown). Because htb1-3 is a hypomorphic allele rather than a null mutation, the distinction between synergistic and additive effects has to be interpreted with caution. However, taken together, these results support the idea that H2B is involved with Rad5p in a UV-induced DNA repair pathway that is independent of the Ubc13p-Mms2p complex but related to Rad30p.
Effects of htb1-3 on UV-induced mutagenesis: Earlier analysis of spontaneous and UV-induced mutagenesis at various loci performed in rad5Δ or rad30Δ single mutants did not specifically place RAD5 and RAD30 in the error-prone or error-free sub-branches of the PRR pathway. Instead, the combination of results obtained from double and triple mutants suggested a role for these two genes principally in the error-free subbranch of the PRR pathway, with a minor role in the errorprone pathway (Johnsonet al. 1992; McDonaldet al. 1997; Roushet al. 1998). In this work, we show that the htb1-3 mutation alone did not significantly alter UV-induced reversion at two different loci and did not increase the reversion frequency when combined with the rad5Δ mutation (Figure 9 and data not shown). This is in contrast to what has been observed for the rad5Δ rad30Δ double mutant, which exhibits an elevated frequency of UV-induced mutation (McDonaldet al. 1997). These results reinforce the idea that H2B and Rad5p could be involved together in a specific subbranch of UV-induced DNA repair. However, we cannot determine whether the htb1-3 mutation affects the error-free or error-prone pathway or both pathways.
What is the relationship between Rad5p and H2B? Rad5p contains several structural motifs that have been implicated in a number of biological processes. A RING finger motif in the C-terminal half of the protein has been shown to be necessary for its interaction with Ubc13p (Ulrich and Jentsch 2000). Rad5p also contains an ATPase domain with homology to the Swi2p/Snf2p family of ATPases (Hirschhornet al. 1992; Johnsonet al. 1992; Schildet al. 1992; Cairnset al. 1996; Duet al. 1998; Papoulaset al. 1998; Pollard and Peterson 1998). Several proteins that contain this ATPase domain have direct roles in destabilizing nucleosomes, a function that is important for transcriptional activation (Muchardt and Yaniv 1999; Peterson and Workman 2000). If Rad5p is also able to remodel nucleosomes, this activity might be important for some aspects of PRR. We suggest that H2B might play a role in chromatin remodeling by Rad5p, perhaps to facilitate the chromatin association or activity of both mutagenic and error-free repair polymerases.
Frédéric Baudat, Paul Kaufman, Hannah Klein, Jac Nickoloff, Louise Prakash, Rodney Rothstein, Lorraine Symington, Helle Ulrich, and Ted Weinert are gratefully acknowledged for their generous gifts of plasmids or for advice. This work was supported by National Institutes of Health grants GM40118 (to M.A.O.) and GM58673 (to S.K.), Human Frontiers Science Program grant RG0254 (to M.A.O.), and a fellowship from the Association pour la Recherche contre le Cancer to E.M.
Communicating editor: L. S. Symington
- Received September 18, 2001.
- Accepted January 16, 2002.
- Copyright © 2002 by the Genetics Society of America