Nucleosome structure and repair of N-methylpurines in the GAL1-10 genes of Saccharomyces cerevisiae.

Nucleosome structure and repair of N-methylpurines were analyzed at nucleotide resolution in the divergent GAL1-10 genes of intact yeast cells, encompassing their common upstream-activating sequence. In glucose cultures where genes are repressed, nucleosomes with fixed positions exist in regions adjacent to the upstream-activating sequence, and the variability of nucleosome positioning sharply increases with increasing distance from this sequence. Galactose induction causes nucleosome disruption throughout the region analyzed, with those nucleosomes close to the upstream-activating sequence being most striking. In glucose cultures, a strong correlation between N-methylpurine repair and nucleosome positioning was seen in nucleosomes with fixed positions, where slow and fast repair occurred in nucleosome core and linker DNA, respectively. Galactose induction enhanced N-methylpurine repair in both strands of nucleosome core DNA, being most dramatic in the clearly disrupted, fixed nucleosomes. Furthermore, N-methylpurines are repaired primarily by the Mag1-initiated base excision repair pathway, and nucleotide excision repair contributes little to repair of these lesions. Finally, N-methylpurine repair is significantly affected by nearest-neighbor nucleotides, where fast and slow repair occurred in sites between pyrimidines and purines, respectively. These results indicate that nucleosome positioning and DNA sequence significantly modulate Mag1-initiated base excision repair in intact yeast cells.

Some DNA lesions that are repaired by nucleotide excision repair (NER) or certain base excision repair (BER) pathways are removed much faster in the transcribed strand (TS) than in the nontranscribed strand (NTS) of an active gene (5). In contrast, repair of NMPs does not appear to be coupled to transcription, since the TS and NTS have similar repair rates in the genes analyzed to date (6 -8). However, repair rates of NMPs vary dramatically at different sites in both the human PGK1 gene (7) and the yeast minichromosome YRpSO1 (8). One reason for this repair heterogeneity is the effect of nearest-neighbor nucleotides (8), in that slow repair occurs at NMPs between purines and fast repair occurs at NMPs between pyrimidines. It was proposed that different stabilities of base stacking between adjacent base pairs can affect flipping out of NMPs from the DNA helix during recognition and incision by DNA glycosylase (8).
In the nucleus of eukaryotic cells, DNA is packaged into a nucleoprotein complex known as chromatin (9). This complex provides the compaction and structural organization of DNA for processes such as replication, transcription, recombination, and repair. The fundamental subunits of chromatin are nucleosome cores, where 147 bp of DNA is wrapped around a histone octamer (10). DNA between two adjacent nucleosome cores is called linker DNA, which varies in length from about 20 to 90 bp in different organisms and tissues, or between individual nucleosomes in the same cell (11).
The effect of nucleosome structure on NMP repair is not understood. In rat liver cells, it was shown that the overall removal of NMPs occurred at a relatively uniform rate in different chromatin fractions (i.e. active chromatin, bulk genome and nuclear matrix) (12). On the other hand, within the yeast minichromosome YRpSO1, there was a mild correlation between repair rates and nucleosome positioning in regions of the inducible GAL1:URA3 fusion gene, but not in the constitutively expressed HIS3 gene on the same plasmid (8).
To further address questions on the effect of nucleosome structure on repair of NMPs in intact cells, we examined repair in the divergent yeast GAL1-10 genes, which share a common upstream-activating sequence (UAS). These genes are induced to very high levels of expression in galactose, but are completely repressed in glucose (13). Extensive studies have been done on isolated nuclei or chromatin to map the nucleosome structure in the GAL1-10 region and the effects of galactose induction (14 -17). However, it is possible that subtle changes in nucleosome structure of this region differ in whole cells, and were missed because of the procedure of nuclei and/or chromatin isolation. Thus, to examine the influence of nucleosome structure on repair of NMPs in intact cells, we developed a nucleosome mapping procedure using bleomycin (BLM), which avoids isolation of nuclei or chromatin. This basic glycopeptidederived antibiotic has been shown to preferentially cleave nucleosome linker DNA in isolated Chinese hamster nuclei (18), lysophosphatidylcholine-permeabilized human cells (19), and whole yeast cells (20). However, BLM has not been used to map nucleosome positions in specific sequences in whole cells, since highly specific cleavage in linker DNA has not been achieved. By using the mild nonionic detergent digitonin to efficiently permeabilize yeast cells, and rich medium to effectively stop BLM cleavage during DNA isolation, we were able to map nucleosome structure at nucleotide resolution in the GAL1-10 region in whole yeast cells and directly correlate repair of NMPs with nucleosome structure.
BLM Cleavage of Chromatin and Naked DNA-Yeast cells were grown at 30°C in minimal medium containing 2% glucose or 2% galactose to late-log phase (OD 600 ϳ1.0). After washing twice with ice-cold 2% glucose (for glucose cultures) or 2% galactose (for galactose cultures), the cells were resuspended in 50 mM NaCl, 2 mM MgCl 2 , 0.02% glucose or galactose, to give a cell density of 2 ϫ 10 9 cells/ml. Digitonin (Sigma, 10% stock) and Fe(NH 4 ) 2 (SO 4 ) 2 (10 mM, freshly dissolved in H 2 O) were mixed with the cell suspension to give a final concentration of 0.05% and 50 M, respectively. BLM (Sigma, 20 units/ml stock) was then added to final concentrations of 0 -400 milliunits/ml, and the mixture was incubated at 30°C for 12 min. To stop the reaction, the cell suspension was mixed with 100 volumes of ice-cold YPD (1% yeast extract, 2% peptone, 2% glucose) or YPG (1% yeast extract, 2% peptone, 2% galactose) and pelleted by centrifugation. For the second wash, cells were resuspended in ice-cold 2% glucose or 2% galactose, mixed with 1:10 volume of a stock solution containing 10% yeast extract and 20% peptone, and pelleted by centrifugation. It has been shown that BLM associates with outer plasma membranes of mammalian cells (CV-1), and can cleave DNA after the cells are lysed, even in the presence of EDTA (21). Furthermore, BLM molecules can only be efficiently removed from mammalian cell membranes by trypsin treatment (21). We found BLM can be efficiently removed after treatment, simply by washing the cells with media containing yeast extract and peptone.
The naked DNA used in these experiments was a PCR fragment of the yeast GAL1-10 region. Reactions (50 l) contained 50 mM Tris-HCl (pH 8.0), 100 ng of a PCR product, 1 g of sonicated salmon sperm DNA, 1-5 M of freshly prepared Fe(NH 4 ) 2 (SO 4 ) 2 and 0 -5 milliunits/ml of BLM. After 12 min of incubation at 30°C, the DNA was quickly separated from the reaction mixture with the QIAquick Nucleotide Removal Kit (Qiagen).
DMS Treatment and Repair Incubation-Yeast cells were grown at 30°C in minimal medium containing 2% glucose or 2% galactose to late log phase (OD 600 ϳ1.0), and mixed with DMS (Sigma, undiluted solution) to give a final concentration of 0.03% (v/v). After 2 min incubation at room temperature, cells were washed twice with ice-cold 2% glucose (or 2% galactose) and resuspended in the same solutions containing 100 mM hydroxyurea, to prevent DNA replication during repair incubation (22). One-tenth volume of a solution containing 10% yeast extract and 20% peptone was added to the DMS-treated cultures. After different times of repair incubation at 30°C, an aliquot was removed and put on ice.
Isolation of Genomic DNA-After BLM treatment or repair incubation following DMS treatment, pellets of 2 ϫ 10 9 cells, were mixed with 2 ml of ice-cold nuclei isolation buffer (NIB: 50 mM Tris-HCl, 2 mM MgCl 2 , 150 mM NaCl, 17% glycerol, 0.5 mM spermine, 0.15 mM spermidine, pH8.0) and 2 ml of acid-washed glass beads (Sigma, 425-600 microns). The mixtures were vortexed for 30 s and kept on ice for 2-5 min. This procedure was repeated three times to completely break up the cells. The samples were mixed with 10 ml of 50 mM Tris-HCl, 400 mM NaCl, 2% SDS, 2 mM EDTA, pH 8.0, and incubated at 65°C for over 30 min. After cooling to room temperature, the samples were mixed with 8 ml of 5 M NaCl and left on ice overnight. The samples were then centrifuged at 4°C, and the supernatant was collected. The DNA was precipitated, treated with RNase A, and extracted with phenol/chloroform/isoamyl alcohol (25:24:1). After re-precipitation, the DNA was dissolved in H 2 O and stored at Ϫ20°C before using.
Mapping BLM Cleavage and NMP Sites-The sites of BLM cleavage and NMPs were mapped along a 1.9-kb region encompassing the common UAS and 5Ј-portion of the GAL1 and GAL10 genes. Five overlapping restriction fragments (four of these fragments are shown in Figs. 1 and 2) on each strand of the region were end-labeled using the procedure described previously (23,24), with slight modification. Briefly, about 2-3 g of genomic DNA was digested with restriction endonuclease(s) to release the fragments of interest. For NMP mapping, the restricted DNA was further cleaved at the NMP sites by incubating DNA in 1 M piperidine at 90°C for 30 min, and the piperidine removed by evaporation (8). Excess copies of a biotinylated oligonucleotide, which has a portion complementary to one end of the fragment to be labeled, were mixed with the sample. The T tract in the biotinylated oligonucleotides described formerly (23,24) was changed to short runs (4 -5) of Ts separated by Gs. This change ensured full-length incorporation of radioactive dAMPs opposite the Ts, and eliminated hybridization between the oligonucleotides and contaminating poly(A) tails of mRNA. The mixture was heated to 95°C for 5 min to denature the DNA and then cooled to an annealing temperature. The annealed fragments were attached to streptavidin magnetic beads (Dynal), and the other fragments were removed by washing the beads at the annealing temperature. The attached fragments were labeled using [␣-32 P]dATP (PerkinElmer Life Sciences) and non-radioactive dCTP, rather than [␣-32 P]dATP alone (23,24). The labeled fragments were resolved on sequencing gels and exposed to PhosphorImager screens (Molecular Dynamics).
Sequence ladders were generated from PCR products of the GAL1-10 fragments, using the rapid Maxam-Gilbert method (25). The ladders were labeled using the same procedure as that for the BLM or DMS treated DNA samples.
Quantitation of BLM Cleavage and NMPs-The BLM cleavage at individual sites is reflected by the band intensity on a gel. The doses of BLM used were relatively low, to ensure that cleavage by BLM followed single-hit kinetics for the majority of fragments analyzed. However, even under these conditions, a small portion of the fragments still have more than one cleavage, and a small portion of the signal at a site will not show up on a gel if the cleavage occurs upstream (relative to the labeled end) of the site. The effect of this "hidden signal" increases as a cleave site is more distant from the labeled 3Ј-end, which runs at the bottom of a gel. For a fragment of Ϸ400 nucleotides (i.e. the maximum length quantified on the sequencing gels), the hidden signal at a site close to the top of a gel can be as much as 30% of the total signal in a band. To correct for this, the following algorithm was used: Let A denote the total signal intensity in a gel lane, X the actual signal intensity (i.e. not reduced by upstream cleavages) at position N in a gel lane, and C the total signal intensity upstream of position N (the upstream sites run below position N in a gel lane). The proportion of signal intensity at position N is X/A and the proportion of signal intensity upstream of position N is C/A. Suppose BLM cleavage at specific sites is randomly distributed among different fragments (i.e. follow a Poisson distribution), then the probability of cleavages at position N and those upstream of position N on the same fragments is (X/A) ϫ (C/A). Therefore, the total number of cleavages at position N and those upstream of position N on the same fragment is A ϫ [(X/A) ϫ (C/A)], or (X ϫ C)/A. If B denotes the observed signal intensity at position N on the gel, then (X ϫ C)/A ϭ X Ϫ B and the actual signal intensity (X) at position N is The signal intensity at all pixels in a gel lane was measured by ImageQuaNT software (Molecular Dynamics) and the data transferred to Microsoft Excel. After the gel background signal was subtracted, the total signal intensities in different lanes of a gel were normalized to the same amount. The signal intensity at all pixels along a gel lane (including the control lanes not cleaved with BLM) was then corrected using the above described equation X ϭ (A ϫ B)/(A Ϫ C). The signal intensities in a lane containing the uncleaved sample were used as a baseline for other lanes that contain BLM cleaved samples. As the signal intensities along a gel lane of uncleaved sample fluctuate, the baseline was smoothed by local averaging of signal intensities (at continuous intervals of 100 -200 pixels) in the lane. The corrected signal intensities in a lane containing BLM cleaved sample were then corrected by subtraction of the smoothed baseline. The data was then imported to PeakFit (SPSS, Inc.) to fit and deconvolute peaks corresponding to the individual bands on the gels (24). Quantitation for NMPs followed the same procedure as that for BLM cleavage.

RESULTS
Nucleosome Structure in the GAL1-10 Genes-Wild type (Y452) yeast cells were grown in glucose or galactose medium, permeabilized with digitonin and treated with BLM. The time used for digitonin permeabilization and BLM treatment was 12 min, which is just long enough to see noticeable BLM cleavage in chromatin DNA. Under these conditions, different concentrations of BLM showed a linear response of band intensity (Figs. 1 and 2 and data not shown). Therefore, the cleavage by BLM followed single-hit kinetics for the majority of the chromatin DNA fragments.
Genomic DNA was isolated from the BLM-treated cells and cut with restriction enzyme to release the fragments of interest. A total of 5 overlapping restriction fragments were analyzed for each strand of the GAL1-10 region. The restricted fragments were strand-specifically end-labeled, resolved on DNA sequencing gels, and exposed to phosphorimager screens. For comparison, naked DNA of the same GAL1-10 region was also treated in parallel with the chromatin DNA. As can be seen from the gels in Figs. 1 and 2, naked DNA was cleaved at very low concentrations of BLM (5 milliunits/ml) and Fe 2ϩ (1 M), and no cleavage was observed in the presence of BLM or Fe 2ϩ alone (lanes 1 and 2 in Figs. 1 and 2; data not shown). Almost all pyrimidines 3Ј to guanines are cleaved by BLM (compare lanes 2 in each gel with the sequencing lanes CϩT and G) consistent with previous reports (26). Most adenines 3Ј to guanines are also cleaved, but generally to a lesser extent (Figs. 1 and 2, arrowheads to the right of gels). Furthermore, some sites of weak cleavage are also observed at pyrimidines 3Ј to adenines (Figs. 1 and 2, small horizontal bars on the right side of gels).
Much higher concentrations of BLM and Fe 2ϩ are needed for cleavage of chromatin DNA in permeabilized cells to reach the same level as that for naked DNA (Figs. 1 and 2). Furthermore, there was consistently more cleavage of chromatin DNA in glucose cultures than in galactose cultures, when the same concentration of BLM was applied (Figs. 1 and 2; data not shown). Most likely, BLM is hydrolyzed more rapidly in galactose cultures due to the induction of BLM hydrolase, which is encoded by the GAL6 gene (27).
To determine the protection level of a nucleosome to its DNA, the ratios of band intensities for BLM cleavage of naked DNA to chromatin DNA (in glucose and galactose cultures) were determined from scans of phosphorimages and peak deconvolution (24). This ratio was designated as 1.0 at a BLM cleavage site (nucleotide 1003 relative to the unique StuI site in the In galactose cultures, where GAL1-10 genes are induced, all the nucleosome core sequences in the GAL1-10 region analyzed are less protected from BLM cleavage (Figs. 1-3), indicating nucleosomes are disrupted upon galactose induction. This disruption is most striking in nucleosomes e and f, which are adjacent to the UAS, presumably due to their occupying the minimum number of positions in glucose cultures (Figs. 1-3).
Induction of NMPs in the GAL1-10 Genes-The same wild type (Y452) yeast cells as those used for nucleosome mapping were used for analyzing NMP induction and repair in the GAL1-10 region. The cells were treated with 0.03% DMS for 2 min to induce NMPs (see "Experimental Procedures"). After different times of repair incubation, genomic DNA was isolated, digested with restriction enzymes, and cleaved at NMPs by hot alkaline treatment. The cleaved fragments were strandspecifically end-labeled, resolved on DNA sequencing gels and exposed to phosphorimager screens. As can be seen from the gels in Figs. 4 and 5, the main type of NMPs induced is 7MeG (e.g. compare 0-h lanes with G lanes), while 3MeA is induced to a much lower extent (e.g. compare 0 h lanes with G and GA lanes). This pattern of NMP induction is similar to that seen with the yeast minichromosome YRpSO1 (8) and to that of other past reports (2).
In galactose cultures, a strong protection from DMS methylation is seen at Gs in the triplet sequences of CGG, AGG, or CGC located at both ends of the palindromic Gal4 binding sites (brackets in Figs. 4C and 5C; Gs marked with arrowheads on the right side of gels), consistent with a previous report (29). Meanwhile, an enhancement of methylation can also be seen in some sites of the UAS region, especially in the top strand (Figs. 4C and 5C, small horizontal bars on the right side of gels). Moreover, there is a difference between glucose and galactose cultures in NMP yield at several sites in the promoter regions of the two genes (Figs. 4B and 5B, stars on the right side of gels). On the other hand, NMP yields at most sites throughout the region are similar between the two cultures ( Figs. 4 and 5), indicating that the presence of nucleosomes does not markedly affect NMP induction. This agrees with a previous in vitro study showing that formation of nucleosomes does not significantly modulate NMP induction by DMS (30).
Modulation of NMP Repair by Nearest-Neighbor Nucleotides-The repair rates of NMPs at different sites were dramatically different (Ͼ30-fold at sites) (Figs. 4 -6). As can be seen from the gels in Figs. 4 and 5, most NMP sites located between pyrimidines (C or T) were repaired much faster than NMPs between purines (A or G), and those located between a purine and a pyrimidine were repaired at intermediate rates. In order to assess the generality of these observations, the percentages of 7MeGs remaining following different times of repair incubation in the same contexts of nearest-neighbor nucleotides (i.e. between purines, pyrimidines, and between a purine and a pyrimidine) were averaged. As can be seen from Fig. 7, the statistical data confirm these observations. As limited 3MeA sites were available for analysis, the effect of the nearest-neighbor nucleotides on 3MeA repair cannot be analyzed by this method.
Modulation of NMP Repair by Nucleosome Structure in the GAL1-10 Genes-In glucose cultures, the strongest correlation between nucleosome positioning and NMP repair can be seen in the two nucleosomes (e and f) with the most fixed positions (Figs. 4 -6). In these regions, slow repair occurs in the nucleosome core DNA and faster repair takes place in nucleosome linker or nucleosome-free DNA (i.e. the UAS region). This correlation sharply fades off in the nucleosomes that are more distant from the UAS, in agreement with the observations that the variability of nucleosome positioning sharply increases with distance from the UAS. This profile of correlation can be seen more clearly after the individual times required for repairing 50% of the NMPs (T1 ⁄2 ) are smoothed (by locally averaging the values in continuous 40-nucleotide intervals) (Fig.  6B). However, with the marked effects of nearest-neighbor nucleotides on NMP repair (see above) superimposed on the effects of nucleosome structure, the nucleosome effect is masked and more difficult to discern in the regions distant from the UAS.
Galactose induction causes enhancement of repair in both strands of the nucleosome core DNA (Figs. 4 -6). Indeed, increases of as much as 8-fold occur in the core regions of nucleosomes e and f, which are disrupted most dramatically (Fig. 6C).
Interestingly, this galactose-enhancement of NMP repair also fades off with distance from the UAS (Fig. 6C), correlating well with the nucleosome positioning and disruption trends (compare Figs. 3B and 6C).
The Role of NER in NMP Repair-It has been shown that mag1 mutants that are defective in NER (rad1 or rad2 mutants) are extremely sensitive to MMS-induced killing, and the effects of these mutations are synergistic (31). This suggests that NER may provide an alternative pathway for repair of NMPs. To assess the contribution of NER, repair of NMPs was analyzed in different genomic regions of isogenic wild type, mag1, rad1, and mag1 rad1 cells. As examples, NMP repair in regions of the GAL1 gene and the constitutively expressed RPB2 gene, which encodes the second largest subunit of RNA polymerase II, is shown in Fig. 8. As can be seen, essentially normal repair occurred in rad1 cells. In contrast, deletion of the MAG1 gene almost completely abolishes repair of NMPs, although residual repair can be seen if the RAD1 gene is present in the mag1 cells (Fig. 8). Repair analysis in other genomic regions shows the same trends (data not shown), indicating NER plays little, if any, role in the repair of NMPs in S. cerevisiae. DISCUSSION We have mapped nucleosome structure and repair of Nmethylpurines in whole yeast cells. The major nucleosome positions observed in this report agree well with past reports on isolated nuclei or chromatin (15)(16)(17). However, in the present study, nucleosomes that occupy fixed positions in the GAL1-10 promoter region were only observed in the regions adjacent to the UAS in glucose cultures, and the variability of positions sharply increases with increasing distance from the UAS (Figs. 1-3). This observation fits well with the finding that the binding of Y factor to a short sequence that overlaps the Gal4 binding site II of the UAS serves as a nucleosome positioning boundary (15). In contrast, mapping with isolated nuclei or chromatin showed longer arrays of precisely positioned nucleosomes in the entire GAL1-10 region (15)(16)(17). This difference may reflect a selection for lowest energy nucleosome positions in chromatin during nuclei isolation, compared with more dynamic features of nucleosomes that are distant from the UAS in intact cells. This notion agrees well with the NMP repair data (Figs. 4 -6), as well as results of nucleotide excision repair of ultraviolet light induced cyclobutane pyrimidine dimers. 2 Our data strongly suggest that nucleosomes in intact cells inhibit NMP repair. First of all, in nucleosomes with fixed positions (especially e and f), repair is much slower in the core DNA sequences than in linker DNA (Figs. 4 -6). Secondly, induction of transcription enhanced NMP repair, with the regions close to the UAS (where nucleosome disruption is most dramatic) being most striking (Fig. 6C). We note that, it is unlikely that this enhancement is caused by a direct coupling between transcription and repair, as no strand bias for repair is observed in these regions . This enhancement, however, correlates well with nucleosome positioning and disruption profiles (compare Figs. 3B and 6C). As multiple factors seem to influence repair of NMPs, nucleosome affected NMP repair may be obscured, especially in the more dynamic nucleosomes in the cell. In addition to the influence of nearest neighbor nucleotides (discussed below), the binding of nonhistone proteins may also affect NMP repair. Indeed, repair of NMPs in the UAS region was faster in glucose cultures than in galactose cultures, where the nonhistone protein Gal4 is bound to the UAS region (Figs. 4C, 5C, and 6). These considerations may explain why little or no correlation between NMP repair and nucleosome positioning is seen in regions distant from the UAS. In a previous report on NMP repair in the yeast minichromosome YRpSO1 (8), we observed a mild correlation between repair rates and nucleosome positioning in some regions of the GAL1:URA3 fusion gene, but not in the HIS3 gene (8). Presumably, these observations reflect differences in dynamics of nucleosome positions in these genes on the minichromosome.
NER and Mag1-initiated BER are synergistic in response to MMS-induced DNA lesions (31), indicating a subset of these lesions may be repaired by both pathways. Furthermore, in vitro experiments with purified human 3-methyladenine-DNA glycosylase (MPG protein), which is the counterpart of yeast Mag1, show that MPG interacts with the human homologue of Rad23 (hHR23) (32). Importantly, this interaction elevates the rate of MPG-catalyzed excision from hypoxanthine-containing substrates (32). Our results with mag1 and rad1 mutant cells suggest that, in S. cerevisiae, repair of NMPs is accomplished primarily by the Mag1-initiated BER pathway, and that NER contributes very little to the repair of these lesions. We also analyzed NMP repair in mutants lacking Rad7, Rad16 and Rad23 proteins, each of which is a component of the NER pathway (33). None of these mutants showed a detectable deficiency in repair of NMPs (data not shown). Thus, it is possible that the substrate shared by NER and Mag1 initiated BER is not NMPs, but a rare DNA lesion that cannot be detected by our technique. We note, however, that a very small amount of repair of NMPs does occur in mag1 cells if Rad1 is present (Fig. 8). This residual repair may be sufficient to cause the observed synergy between NER and the Mag1 initiated BER for MMS-induced lesions.
A number of organisms have a strong backup pathway for repairing NMPs. In E. coli, the AlkA and Tag proteins can initiate repair of these lesions, even though the substrate specificity differs for the two enzymes (34,35). In Schizosaccharomyces pombe, NMPs may be repaired primarily through the NER pathway, rather than a BER pathway (36). In mammalian cells, a pathway may exist that repairs 7MeG in the absence of the ordinarily used Aag DNA glycosylase (37). However, deletion of the MAG1 gene in S. cerevisiae cells almost completely abolishes NMP repair ( Fig. 8 and data not shown), and no NMP repair can be seen in mag1 rad1 double deletion cells. This indicates that S. cerevisiae may lack a strong back up pathway for repairing NMPs.
Finally, as observed with the yeast minichromosome YRpSO1 (8), there was a significant correlation between the nearest neighbor nucleotides of NMP sites and the repair of NMPs in the yeast genomic GAL1-10 region. These data suggest that the same repair machinery is used for repairing NMPs in genomic and minichromosome DNA. This finding is similar to that of Sweder and Hanawalt (38), who studied repair of ultraviolet light-induced cyclobutane pyrimidine dimers in genomic and minichromosome DNA of yeast.