Photoreactivation of UV-induced cyclobutane pyrimidine dimers in the MFA2 gene of Saccharomyces cerevisiae

Abstract

Photoreactive repair (PR) of cyclobutane pyrimidine dimers (CPDs) was mapped at nucleotide resolution in nucleotide excision repair (NER) proficient and deficient strains for the transcriptionally active and inactive MFA2 gene. Enhanced PR in the control region occurred in areas where no nucleosomes were present, particularly linker regions in the α mating type. The presence of excision plus transcriptional activation alleviated this preference, causing repair in the regions that were linker and core in the α mating type to be the same in this strain. Transcription had no effect on photoreactive repair in transcribed and downstream regions of MFA2, where similar rates were observed for specific CPDs in both strands. The presence of positioned nucleosomes in α mating types revealed slow repair in the nucleosome core, with faster repair occurring at the 3′ and 5′ edges. These data support the evidence that CPDs are repaired quicker in nucleosome-free regions and at edges of nucleosomes. CPDs in the linker regions are repaired more efficiently in the transcriptionally inactive strains, suggesting that nucleosome movement associated with transcription of MFA2 hampers PR irrespective of the strand. Proficient NER influenced PR in the TATA and Mcm1 binding sites by enhancing it, particularly when transcription was activated.

INTRODUCTION

Ultraviolet (UV) irradiation produces a variety of lesions in DNA. Two major types of UV-induced damage are cyclobutane pyrimidine dimers (CPDs) and pyrimidine (6–4) pyrimidone photoproducts [(6–4)PPs], which correspond to ∼75 and 25%, respectively, of the total UV-induced damage products (1).

The induction of UV-induced DNA damage in eukaryotic cells has a further level of complexity due to the packaging of DNA into chromatin. Chromatin has a basic unit of organisation, the nucleosome, which consists of DNA wrapped in two complete turns around an octamer of core histones (2–4). The nucleosome cores are joined together by shorter stretches of linker DNA of between 0 and 60 bp long, which vary in length in different organisms and tissues, and between individual nucleosomes (5).

CPDs induced at 254 nm form with no preference occurring between linker and nucleosomal DNA (6–9). However, the packaging of DNA by core histones dramatically influences the distribution of CPDs within nucleosomes (7). The CPDs are positioned where the DNA strand is farthest from the histone surface and a reduced yield occurs near the histone surface (6,9–11).

Photoreactivation (PR) is a unique pathway that specifically repairs CPDs by initiating the direct reversal of damage to restore the original pyrimidines. The reaction requires a single enzyme, photolyase (EC 4.1.99.3), and proceeds via a two-step process. Recognition of CPDs by photolyase is structure specific and the enzyme is not influenced by the nucleotide content surrounding the dimer. PR, however, is affected by the nucleotide content of the pyrimidine dimers and repair rates exhibit efficiencies T-T > T-C > C-C (12).

Photolyase has a paradoxical function involving both induction and interference with repair. Photolyase not only repairs UV damage by PR, but is also involved in stimulating nucleotide excision repair (NER) (13). In comparison with NER, photolyase in yeast has been reported to preferentially repair the non-transcribed strand (NTS) of active RNA polymerase II and III transcribed genes. Its action appears to be inhibited by stalled RNA polymerase on the transcribed strand (TS) (14–16). In yeast, two genes, PHR1 and PHR2, are involved in PR. PHR1 encodes a photolyase (17–19), whereas PHR2 may code for a regulatory protein for the PHR1 gene (20). PR can also be used as a tool for identifying the accessibility of lesions as photolyase repairs linker and open promoter regions easily, but has difficulty in core regions (21,22). Recently, the PR of CPDs on the NTS of an active gene was modulated as for positioned nucleosomes in an inactive gene (23). Thus, PR may be used to detect the position of nucleosomes on the NTS of active genes and where MNase sensitivity does not work.

Another major pathway involved in the removal of UV-induced DNA damage is NER. Eukaryotic NER is an evolutionary conserved pathway, requiring approximately 30 proteins that remove the damaged base as part of a larger oligonucleotide (24,25). Within NER, two sub-pathways operate to give a further level of complexity. Transcription coupled repair (TCR) deals with rapid and efficient removal of lesions that block transcription and is limited to the TS of active structural genes (26–29). The other pathway, known as global genomic repair (GGR) accomplishes the slower repair of damage from bulk DNA, including the NTS of active genes (1).

NER of DNA damage in chromatin involves a biphasic mechanism. The first phase is an extremely rapid, non-uniform repair where the 5′ and 3′ ends of nucleosomal DNA and linker DNA are preferentially repaired relative to the central region of nucleosomal DNA (6,7,22,30–32). The enhanced incorporation of nucleotides into repair patches in human cells uniformly label DNA up to 60 bp from the 5′ end and ∼30 bp from the 3′ end, with an ∼50 bp central region being devoid of repair-incorporated nucleotides (33). The second phase is a contrasting slow, uniform repair, which may be linked to a slow reassociation of H1 histone, although this may not be the case with Saccharomyces cerevisiae where no H1 has been identified (6,7,30–32,34). Further evidence supporting nucleosome disruption and chromatin influence in repair is that NER requires an ∼100 bp region in order to remove the 27–29 bp patch containing the lesion (22).

Studies have indicated that an increased NER in the promoter region of the S.cerevisiae MFA2 gene is associated with the chromatin structure becoming more accessible (35). MFA2 provides an excellent model for studying the effects of transcription and chromatin structure on DNA repair in yeast (35–37). Its transcriptional regulation is well understood and can be controlled by investigating events in a or α mating type cells. In a cells the gene is active and the chromatin structure is opened, whilst in α cells the gene is silent and exists in a folded chromatin structure (38,39). Mcm1, a constitutive transcription factor, in conjunction with α1, α2 and a1 regulatory proteins, governs the transcriptional status of MFA2. α1 is required to activate transcription of α-specific genes, whilst α2 alone represses transcription of a-specific genes. Similarly, α2 in conjunction with the regulator a1 represses the transcription of haploid specific genes in a/α diploids.

In α cells, four nucleosomes have been identified within the control and coding regions of MFA2 in the α mating type (40), and are referred to as –2, –1, +1 and +2. The –2 and –1 nucleosomes are situated in the control region, and are located at approximate positions of –412 to –272 and –207 to –62, respectively. Nucleosome –1 covers the TATA box, whilst the Mcm1 binding site is positioned between these two nucleosomes (Fig. 1). There are also two nucleosomes situated within the transcribed region and are referred to as +1 and +2. These cover the fragment at positions –58 to +88 and +122 to +254, respectively. Between each nucleosome is a linker region of two or more base pairs, which are symbolised as L–1, L+1 and L+2 (Fig. 1). The α mating type strains studied possess positioned nucleosomes whereas in a mating types, the nucleosomes may not be in a consistently stable position and are not detected by MNase I footprinting (40). However, here the footprinting revealed differences in protein binding at the Mcm1 binding site between a and α cells, where MNase hypersensitive sites exist in α cells but not in a cells.

Figure 1.

The regions of the MFA2 gene. The Rsa1 restriction fragment contains the Mcm1 binding site at –251, the TATA box at –125 and the MFA2 coding region from –42 to +286. Only the NTS was analysed for the Rsa1/Bfa1 fragment. The ovals below the gene diagram depict the approximate positions and the classifications of the relative linker and nucleosomal regions. The numbers placed vertically represent the nucleotide position with respect to the MFA2 start point.

Using a 3′ end-labelling technique developed within our laboratory (35), we investigated the PR of CPDs in the MFA2 gene at the level of the nucleotide. The study involved the analysis of NER competent and NER deficient rad14Δ mutant strains belonging to both a and α mating type cell lines. Rad14 is a protein required at an early stage of NER and interacts with transcription factor TFIIH (41). The protein is homologous to xeroderma pigmentosum group A protein (42) which has a crucial role in both GGR and TCR (43). Thus, we were able to examine the influence of both the transcriptional status of MFA2 and the presence of NER complexes on the extent of PR observed with the control and coding regions of MFA2.

MATERIALS AND METHODS

Yeast strains

The strains used in this study were all haploid S.cerevisiae; W303a (RADa, ho, ade 2–1, trp 1–1, kan 1–100, leu 2–3, his 3–11,15, ura 3–3), W303α (RADα, as before), W303a rad 14Δ (a rad14Δ::URA3, as before) and W303α rad 14Δ (α rad14Δ::URA3, as before).

UV treatment and photoreactivation of S.cerevisiae

Cells were UV treated (125 J/m²) at a cell concentration of 2 × 10⁷ cells/ml in cold PBS as described (44). Immediately after treatment all cells were kept on ice in the dark, or subjected to PR. Individual samples were treated in 250 ml aliquots to a PR light produced by a halogen bulb (1000 W). The samples were kept in suspension by stirring and placed in a bath, at 20°C, containing a 5% copper sulphate solution (cupric sulphate pentahydrate; Sigma Ltd, St Louis, MO). A conical flask, containing the sample to be treated, was placed in the centre of a 23 cm diameter transparent Pyrex basin with ∼3 l of the copper sulphate solution, which acted as a light filter, allowing only 300–450 nm wavelengths through. After PR, each sample was placed immediately on ice, in the dark, and the DNA was isolated as described by Reed et al. (44).

Analysing CPDs in the MFA2 gene, sequencing of MFA2 containing fragments, probes/primers used for the analyses and the determination of DNA damage and repair

All techniques for the Rsa1 fragment, containing the MFA2 gene and its promoter regions, were undertaken as described by Teng et al. (35).

Analysing CPDs in the NTS Rsa1/Bfa1 fragment

Yeast DNA (30 µg) from each sample was double digested with 30 U of Bfa1 and 60 U of Rsa1 (New England Biolabs, Hitchin, UK) at 37°C for 1–2 h in a total reaction volume of 200 µl. The enzyme was removed by phenol/chloroform extraction and followed by a DNA precipitation at –20°C, with 0.1 vol 5 M NaCl and 2 vol ice-cold 100% ethanol. Complete digestion creates a 286 bp fragment of the MFA2 gene promoter region (–337 to –52, with respect to the MFA2 gene). The digested DNA was resuspended in 100 µl buffer (10 mM Tris–HCl pH 7.5, 1 mM EDTA). CPD-specific endonuclease (30 µl), obtained from Micrococcus luteus (ML), was added and the samples were incubated for 1 h at 37°C. The ML was removed by phenol/chloroform extraction. Biotinylated probe (1 pmol) was added to each sample. As only the induction and repair of CPDs in the NTS was to be analysed for this fragment (the TS is adequately analysed by the Rsa1 fragment), only one probe was required. The samples were denatured at 95°C for 5 min and then incubated at the annealing temperature (55°C) for 15 min to allow the probe to anneal to the NTS fragment. The Rsa1/Bfa1 fragment was separated from genomic DNA using Dynabeads. Attachment to Dynabeads, washes and labelling were performed as described previously (35).

Probes used in Rsa1/Bfa1 fragment analysis

In order to retrieve the required fragments, biotinylated probes were annealed to denatured DNA. The probe for the NTS fragment was 5′Biotin-gatagcttttttGTGATGTCAATGAACGGATGAACGACAGAA (sequences in lower case are overhang modifications).

Analyses and determination of DNA damage and repair

These were performed as described by Teng et al. (35).

RESULTS

In a and α mating type haploid cells we examined the incidence of CPDs in the MFA2 control and coding regions following UV with or without PR. This was undertaken in RAD cells and NER defective rad14 strains. Typical autoradiographs from experiments examining CPD repair in the TS and NTS of the Rsa1 and Bfa1 fragment from RAD cells are shown in Figures 2–4. By analysing CPDs in both the Rsa1/Bfa1 fragment and Rsa1 fragment we were able to examine the whole of the regulatory and coding sequences of MFA2. The quantitative data are presented as t_50% (the time taken for half of the CPDs at each site to be repaired) in Figures 5 and 6. They are the average of at least two independent experiments. Differences in t_50% are deemed significant when the standard errors on these values do not overlap. Although MFA2 is not transcriptionally active in α cells, we continue to define the two strands as TS and NTS.

Figure 2.

A typical autoradiograph depicting UV-induced CPDs in the TS of a RADa strain. CPD repair was examined in the Rsa1 restriction fragment, which contains the MFA2 transcribed region (green bar), the Mcm1 binding site (blue bar) and the TATA box (red bar). Nucleosome positions are symbolised by ovals on the right of each gel. Numbers on the right of the nucleosomes denote the nucleotide position in relation to the MFA2 transcription start point. Lane U, DNA from un-irradiated cells; lane 0, DNA from cells receiving 125 J/m² UV and extracted immediately; lanes 1–4, DNA from cells that received UV but were subjected to PR for 15, 30, 45 and 60 min; lanes 5–8, DNA from cells that received UV but were incubated for 15, 30, 45 and 60 min in the dark. Lanes T, A, C, and G denote the MFA2 sequence ladder. The numbers and letter on the left-hand side of the gel depict detected CPDs, their nucleotide position and sequence.

Figure 4.

A typical autoradiograph depicting UV-induced CPDs in the Bfa1 fragment of the NTS of a RADa strain. All details as for Figure 2, except CPD repair was examined in the Bfa1 restriction fragment.

Figure 5.

CPD repair in the control region of (A) rad14a, (B) rad14α, (C) RADa and (D) RADα strains. The autoradiographs were quantified with ImageQuant software (Molecular Dynamics). The time for 50% repair of CPDs (t_50%) at a given site was calculated. The t_50% for slowly or unrepaired CPDs (t_50% >60 min) is shown as > 60 min on the graph. Triangles in the upper half of the graph represent CPDs in TS and in the lower half CPDs in the NTS. Standard deviations are shown with error bars. The grey horizontal bar in the centre of the graph symbolises the Mcm1 binding site and the black square the TATA box. The oval shapes underneath the graphs indicate the positions of the respective nucleosomes. The dotted lines highlight the beginning and end of each nucleosome.

Figure 6.

CPD repair in the coding region of (A) rad14a, (B) rad14α, (C) RADa and (D) RADα strains. All details are as for Figure 5.

Control region (–338 to –42)

The induction of CPDs in the control region of the MFA2 gene is analysed by examining the Rsa1 fragment for the TS and the Rsa1/Bfa1 fragment for the NTS (Fig. 4). The Rsa1/Bfa1 fragment is required for the NTS as this enhances the resolution of the bands which are not well separated at the top of the gel containing the Rsa1 fragment. The positions of the respective linker and nucleosome regions depicted in Figure 1 and mentioned below were characterised by work done by Teng et al. (40).

Control linker regions. The control sequence (Fig. 5) contains two linker regions situated at nucleotide positions approximately –271 to –208 (between nucleosomes –2 and –1) and nucleotides –59 to –61 (between nucleosomes –1 and +1). The first linker region, referred to here as L–1, contains the Mcm1 binding site (base pairs –251 to –221) and has four groups of CPDs induced in the TS and eight in the NTS. Within the Mcm1 binding region, three groups of CPDs are induced within positions –234CCCTTT–229 in the TS and three groups of CPDs in the NTS at positions –245TTT–243, –241CCT–239 and –228TTT–226. Analysis of the three groups of CPDs in the TS show that for both the rad14a and α and the RADα, a t_50% > 60 min occurs. In the RADa strain the repair is quicker with the first and third group of CPDs having t_50% of 30 ± 5 and 42 ± 6 min, respectively. However, the CPD at the central position of –234CCCTTT–229 still has a t_50%  > 60 min. The CPDs induced at the Mcm1 binding site in the NTS exhibit similar results irrespective of transcriptional and NER status. CPDs at position –245TTT–243 exhibit a t_50% > 60 min in all four strains. The other CPDs, at position –241CCT–239, have repair rates of 24 ± 3, 14 ± 4, 13 ± 2 and 16 ± 4 min for rad14a and α and RADa and α, respectively, showing CPDs in rad14a are repaired significantly slower than the other three strains.

Six groups of CPDs are induced in the L–1 linker region, but are not in the Mcm1 binding site. One is situated in the TS at positions –216CCT–214 and five are in the NTS, three within position –269TTTTTT–264, one at –260TT–259 and the last at –211CC–210. Analysis of the TS reveals that the CPDs at position –216CCT–214 has t_50% > 60 min for both rad14 and RAD of the a mating type. This CPD has a t_50% of 31 ± 3 and 38 ± 6 min, respectively, in the rad14 and RAD of the α mating type. The CPDs induced on the NTS at positions –269TTTTTT–264 and –260TT–259 all show fast repair with all strains having a  t_50% ≤ 23 min. The CPD at position –211CC–210 has a t_50% > 60 min for rad14a and α and RADa strains, but the RADα has a t_50% of 19 ± 5 min.

The second linker region, referred to as L+1, is a small linker region containing very few nucleotides and has no groups of CPDs induced in either strand.

Control nucleosome regions. The control region contains two nucleosomes approximately covering nucleotides –412 to –272 and –207 to –61. For simplicity, these nucleosomes are referred to as –2 and –1, respectively. Within nucleosome –1 the TATA box region (–125 to –118) is situated, and a single group of CPDs is induced in the TS of this region at nucleotide position –120TTT–118. This group has a t_50% > 60 min in both mating types of the rad14 strain and also in the RADα strain. However, in the RADa strain a t_50% of 44 ± 4 min is observed.

Nucleosome –1, covers nucleotides –207 to –61 and has 21 groups of CPDs induced in the TS and 9 in the NTS. For each nucleotide position analysed, irrespective of the strand, the t_50% for each group of CPDs in the α mating type is greater, or equal to, the corresponding point in the a mating type, except for three positions in the NTS (–146TT–145, –115TCT–113 and –85CTTCTTCT–78). The group at position –146TT–145 shows a significantly faster repair in the rad14α than any of the other strains, with a t_50% of 7 ± 2 min, whereas the RADa and α strains and the rad14a strain have similar t_50% values of 24 ± 1, 27 ± 8 and 25 ± 4 min, respectively. The CPDs at positions –115TCT–113 have a similar t_50% for the RADa and α strains with a t_50% of 30 ± 14 and 25 ± 3 min, respectively. The rad14 strains have identical t_50% of > 60 min irrespective of mating type. The group of CPDs at position –85CTTCTTCT–78 reveal t_50% > 60 min for both rad14a and RADa, whereas rad14α and RADα have t_50% of 22 ± 2 and 26 ± 9 min. Observing the t_50% within this nucleosome in the wild-type and rad14α mating types (where there are detectable positioned nucleosomes) reveals a pattern of quick repair on the 5′ and 3′ edges of the nucleosome, with the inner core exhibiting slow or limited repair.

No CPDs were detected in the TS of the –2 nucleosome region, whereas analysis of the NTS revealed seven groups of CPDs. The RADa and α strains repaired many of these CPDs at a faster rate than the corresponding CPDs in the rad14 strains. The rad14a and rad14α have three and five groups of CPDs, respectively, with t_50% > 60 min, whilst RADa has only one and RADα no CPDs with a t_50% > 60 min. Those groups with slow or limited repair are primarily situated within the central region of the nucleosome core between positions –330 and –298. However, the CPD at position –274, which is near the linker region, also shows slow repair. The other CPDs in this area are all repaired relatively quickly, irrespective of the strain analysed. The t_50% for these CPDs range between 7 ± 1 and 37 ± 20 min, and are mostly situated at the edges of the nucleosomal DNA.

Transcribed region (–42 to +286)

Transcribed linker region. Within the transcribed area (Fig. 6), one linker region (L+2) is situated between nucleosome +1 and +2 at approximate nucleotide positions +89 to +121. Here, two groups of CPDs are induced in the TS (+92CCCT+95 and +116TT+117) and three groups in the NTS, one at position +84CCTCTTCT+91 and two within +118TTTTT+122. The group at position +92CCCT+95 in the TS is repaired quite slowly, with t_50% > 60 min for all strains except in the rad14α strain which has a t_50% of 33 ± 14 min. The CPD at position +116TT+117 is also slowly repaired in both the rad14 strains, but has t_50% of 51 ± 3 and 45 ± 14 min for the RADa and RADα strains. In the NTS, the CPDs induced reveal greater t_50% for both the rad14a and RADa strains in comparison to the CPDs induced at the same position in the respective α mating type strains. The t_50% range between 12 ± 3 min and >60 min.

Transcribed nucleosome regions. Two nucleosomes, +1 and +2, cover nucleotides –59 to +88 and +122 to +254, respectively. Transcriptionally active a cells photoreactivate the CPDs faster within this region than in the corresponding α mating type for both strands of the RAD and rad14 strains, with a couple of exceptions. Towards the L+1 and L+2 linker region (positions; TS, –38CTC–36 and +79TTCCC+83; NTS, –35CT–34 and +154CCCTTTC+160) the α strain repairs the CPDs quicker. For example, the CPD in the NTS near the L+1 region (position –35CT–34) has a t_50% of 57 ± 3 and >60 min in the rad14a and RADa strains, but t_50% of 18 ± 4 and 38 ± 7 min in the rad14α and RADα strains. Likewise, the CPDs near the linker region L+2 (position +79TTCCC+83) have a t_50% of 37 ± 5 and 31 ± 4 min for the rad14a and RADa strains, with a t_50% of 18 ± 3 and 24 ± 6 min in the rad14α and RADα mating types.

In the TS of the nucleosome +2 region, only three CPD groups are induced at positions +133TTCTCC+138, +141TTT+143 and +181TTT+183. Initially, the repair rates appear to be influenced by nucleosome positioning for each strain, and do not correspond with the repair rates observed for nucleosome +1 region, particularly in the a mating type strain. However, when comparing these CPDs with similar sites in the nucleosome +1 region, their repair rates do not deviate greatly. The first CPD in nucleosome +1 (–46TTT–48) has a t_50% of 20 ± 3 and 30 ± 3 min in the rad14a and RADa, respectively, whilst the first CPD in nuclosome +2 (+133TTCTCC+138) has t_50% of 26 ± 4 and 41 ± 3 min. The next two CPDs in nucleosome +2 show slow repair, which is similar to other CPDs found in the core region of nucleosome +1. The TS sequence that is covered by nucleosome +2 is a highly GA-rich region and so very few CPDs are induced. The analysis of this region is limited due to the low numbers of CPDs induced.

Downstream of the transcription region (+256 to +320). This region covers the last 30 transcribed nt and the area after the termination sequence (nucleotide +286). Within this region five groups of CPDs are induced in the NTS and only one in the TS. The CPDs in the TS, at position +318TCCTTT+323, are repaired quicker in the α versus the a mating type. Furthermore, the rad14 mutant repairs these CPDs quicker when compared with the RAD strain of the same mating type. In rad14a and α, the CPDs have a t_50% of 48 ± 5 and 31 ± 16 min, whereas the RADa and α strains have t_50% of >60 and 56 ± 1 min, respectively. The first group of CPDs, induced in the NTS of this region (+267CTTT+270), is repaired very quickly in all four strains. For the rad14a and α these CPDs have t_50% of 11 ± 2 and 10 ± 3 min and for the RADa and α strains t_50% of 10 ± 0 and 18 ± 5 min, respectively. The next two groups of CPDs, at positions +277TTTTTTC+283 and the beginning of +287TTTTTCTTTTCT+298 reveal similar repair rates in the rad14a, RADa and RADα, with the repair in rad14α being significantly quicker. The rad14a strain has a t_50% of 46 ± 7 and 45 ± 1 min, the RADa strain has a t_50% of 33 ± 1 and 46 ± 5 min and RADα has t_50% of 42 ± 5 and 52 ± 5 min. However, the repair in the rad14α strain reveals a t_50% of 27 ± 1 and 29 ± 0 min. A second group of CPDs is induced at the end of position +287TTTTTCTTTTCT+298 and is repaired at t_50% rates for all strains ranging between 15 ± 3 and 27 ± 6 min that were not significantly different. The last group of CPDs at position +305CCTTTCCCTCC+315 reveal significantly quicker repair in the a mating type of RAD and rad14 strains when compared with their corresponding α strain. RADa and RADα have a t_50% of 23 ± 5 and 48 ± 5 min, whilst the rad14a and the rad14α strains have a t_50% of 22 ± 5 and 45 ± 7 min.

DISCUSSION

Differential repair has been described for many organisms from Escherichia coli to humans, and has been considered a good strategy for cells to selectively repair transcription blocking damage in active genes over bulk DNA (29). In 1989, Terleth and colleagues first reported the existence of preferential NER in the yeast S.cerevisiae (45). They showed that the removal of CPDs from the active MATα gene was preferential to the removal from the inactive HMLα gene by ∼2–5-fold (45,46). This phenomenon has also been identified in the MFA2 gene, where enhanced NER of the TS of MFA2a occurs over the NTS of MFA2a and both strands of MFA2α (35). In contrast to NER, photolyase has been shown to preferentially repair the NTS of active genes. However, this is not seen in inactive genes where the repair rates are equal for both strands (14,15). Here we studied PR in the transcriptionally active and inactive MFA2 gene in NER proficient and deficient strains at the level of the nucleotide.

The influence of nucleosomes in preferential strand repair

In the MFA2 gene, for the average repair for all CPDs, we detect a slight preference occurring in the NTS of all strains, irrespective of their NER and transcriptional status. The average strand repair rates for the transcribed portion of the MFA2 fragment for each strain were: rad14a, 52 (TS) and 38 min (NTS); rad14α, 48 (TS) and 34 min (NTS); RADa, 47 (TS) and 37 min (NTS); and RADα, 49 (TS) and 41 min (NTS). Evidence has shown that RNA polymerase II stalled at a CPD on the transcribed strand prevents the CPD from being recognised by photolyase (47). However, the results shown here suggest that the presence of a stalled RNA polymerase at a CPD is not the only factor to influence a preference for the NTS. Inspection of the repair rates of CPDs in the MFA2 transcribed region at the level of the nucleotide indicates that the photoreactive repair rate is more likely to be due to changes in chromatin structure associated with the activation of MFA2. In the α cells there is a pattern of quick repair at the edges of the nucleosomes, with the inner core region exhibiting very slow or limited photoreactive repair by 60 min. This is particularly evident for the NTS of all the strains and, likewise, the TS of the α mating types of both rad14 and the RAD cells. The a mating type strains do not display this pattern in the TS of the area corresponding to that covered by nucleosome +1 in the α mating type. Here the t_50% values indicate a combination of quick and slow PR throughout this region. The NTS of the active MFA2 gene and both strands of the inactive gene exhibit similar preferential repair rates with CPDs at the edges of the nucleosome being repaired quicker than those within the core region. This occurs despite no positioned nucleosomes being detected by footprinting with MNase in these a cells (40). It suggests that the nucleosomes may be associated with the NTS of the active MFA2 gene such that differences in the MNase activity cannot be detected. Furthermore, the data imply that the PR of CPDs may well give a clear indication of nucleosomes in transcribed yeast genes as opposed to footprinting with MNase.

In contrast to Livingstone-Zatchej et al. (14), our results indicate that the inhibition of repair by nucleosomes is in fact stronger than inhibition of photorepair by stalled polymerase. The surprisingly long half-life they observed for the stalled polymerase, 50–60 min, may be due to analysing the repair of CPDs at the level of the gene rather than at the level of the nucleotide. The latter method should provide more detailed insight into the factors that affect repair rates. Additionally, it must be noted that the level of transcription from MFA2 is relatively low. It is quite feasible that effects due to a stalled polymerase are more evident in genes that are transcribed to a greater extent.

Efficient removal of CPDs from the promoter region

Most notable is the rapid repair in the control region of the active MFA2 gene, both in the presence and absence of NER. This links to work performed by Suter et al. (15), where the nuclease-sensitive promoter regions of two yeast minichromosomes (TRpTRURAP and YrpCS1) were repaired in 15 min (70–80% of CPDs). Our work reinforces the idea of photolyase’s role in regeneration of gene regulation by efficient removal of CPDs from promoter regions. The CPD group induced in the TATA box in the rad14a strain is repaired at an equal rate to that of the α mating types. The t_50% value of 60 min for the rad14a may reflect as yet undetected differences that would only come to light if longer PR times were employed. Recently, CPDs induced in the vicinity of the TATA box of the RNA polymerase II transcribed GAL10 gene were shown to have no resistance to photorepair, and can instigate multiple rounds of initiation by polymerase (14). The RNA polymerase III transcribed gene, SNR6, exhibited inhibition of PR in a NER proficient strain suggesting that proteins, such as the TATA binding protein (TBP)–TFIIB complex, inhibit repair by photolyase (45). Photolyase appears to be more efficient in CPD repair outside the TATA box, in inactive promoters and in open promoters that are not folded in nucleosomes (29). The differing levels of PR could also indicate differential stability of the RNA polymerase II and III initiation complexes. Similar patterns were observed for the CPDs induced in the TS of the Mcm1 binding site, where a t_50% > 60 min was revealed for both RAD and rad14 of the α mating types and the rad14a mating type. The RADa strain exhibited slightly quicker repair in the Mcm1 region with two out of three CPDs possessing a t_50% < 42 min. These data suggest that the binding of transcription factors to promoters of transcriptionally active genes enhances accessibility of photolyase within these regions, but that this phenomenon is dependent on a functional NER pathway. This is further revealed by the difference in repair rates in nucleosome –2 (Fig. 5). When NER is present the CPDs reveal an enhanced repair rate, which is not seen in the NER deficient strains. This phenomenon is observed irrespective of the transcriptional status. It is worth noting that 100 bp upstream from the first CPD in the MFA2 fragment is the 303 bp YNL 146W gene. Although its function is unknown, the repair rates observed here indicate an enhanced accessibility of photolyase in a functional NER background.

Photolyase, but not NER, is inhibited by the TBP and associated factors bound to a (6–4)PPs-damaged TATA box (48) and, therefore, this could also be the case for CPD-damaged TATA boxes. DNA footprinting data revealed a preferential MNase sensitivity in the TATA box in a versus α cells that results from a differing organisation of chromatin, whereas an MNase hypersensitivity exists in the Mcm1 binding site in α cells but not a cells (40). In this region, Mcm1 binds alone to its binding site in a cells, whereas in α cells it requires α2 factor for binding suggesting that the different proteins binding to this region may influence the differing MNase sensitivities (40). This could also be the case for the contrary repair rates in this linker region, and the faster repair seen in the RADa may demonstrate that the presence of Mcm1 alone in the a mating type strains facilitates easy access of photolyase. A CPD induced on the TS, at position –205TTCCTT–200, was repaired extremely fast. This was located directly after the Mcm1 binding site, and it is possible that, similar to the Mcm1 binding site, certain protein–DNA interactions influence repair in this region. In contrast to the TS, the NTS of this region appears to be repaired at a similar rate irrespective of mating type or transcriptional activity, indicating the binding, or lack of, regulatory and transcriptional proteins to this site has no effect on photoreactive repair.

Nucleosome effects in the promoter region

The rest of the control region shows that for CPDs induced in the TS of both the RAD and rad14 strains, repair is faster in the a mating type than the α mating type within the nucleosomal regions (Fig. 5). However, there are a small number of CPDs that are actually repaired quicker in the α mating type of the RAD and rad14 strains. These CPDs appear to be present at the edges, either 60 bp 5′ or 30 bp 3′, of the nucleosome region. The 5′ and 3′ ends of core DNA and linker DNA are preferentially repaired by NER relative to the central region of core DNA in human fibroblasts and the yeast URA3 gene (6,22). The enhanced incorporation of nucleotides into repair patches in human fibroblasts uniformly label DNA up to 60 bp from the 5′ end and ∼30 bp from the 3′ end, with an ∼50 base central region being devoid of repair-incorporated nucleotides (33). The TS and NTS of the α mating type strains and the NTS of the a-mating type exhibit a ‘nucleosome curve’ with the outer edges having CPDs with short t_50% values (<26 min) whilst the inner core have CPDs t_50% > 60 min. Studies have shown that NER in the NTS is faster for dimers removed from linker DNA and towards the 5′ end of the positioned nucleosome, whilst slow repair correlates to the internal protected core region. Tijsterman et al. (49) showed that in the URA3 gene ‘slow spots’ of CPD repair coincided with cores of nucleosomes and were interspersed with regions that were quite efficiently repaired. Repair heterogeneity, therefore, can reflect modulation of NER and PR by positioned nucleosomes in the NTS (22,49). Contrarily, we see repair on the TS is far less heterogeneous and shows no correlation with chromatin structure in the a mating type. Apparently, transcription overrides chromatin modulation of NER in the TS (22,49). Here we report a similar effect on photoreversal of CPDs in the TS of the transcriptionally active MFA2 gene.

Photoreactivation as a molecular tool

Within our laboratory, studies indicated an increased preferential NER in the promoter region of the MFA2 gene occurs when the chromatin is ‘open’ in a cells, and this is absent in α cells where the chromatin is closed (35). PR can also be used as a tool for identifying the accessibility of lesions. Photolyase repairs linker, open promoter regions and origins of replication easily, but has difficulty in core regions (1,21,22,50). Rapid phase PR appears to involve the removal of CPDs from the transcribed strand of the actively transcribed NER proficient strain, which can be observed by the low number of CPDs with a t_50% > 60 min. The removal of CPDs occurs at approximately equal rates from linker and DNA at the core ends (6,51). Similarly to other work (35), we observed that the induction of CPDs in relation to the transcriptional status of MFA2 showed no evidence for changes in the pattern or frequency of CPDs in the control or coding regions. The data presented here reveal that, akin to NER, PR displays enhanced repair of CPDs in linker regions and at the edges of nucleosomes. The lack of NER in the rad14 strains revealed little effect on CPD repair by PR irrespective of mating type and transcriptional activity. This is in contrast to studies by Livingstone-Zatchej and colleagues who presented evidence for preferential photoreactive repair in the NTS of only the active URA3 and HIS3 and transcriptionally regulated GAL10 genes (14). Here it should be borne in mind that this difference could reflect a reduced extent of transcription for MFA2 compared with the URA3, HIS3 and GAL10 genes, and that different trends between different genes should not be construed as contradictory. Similar to work on other genes (15), the data presented here demonstrate a direct role of chromatin in modulating the PR of CPDs.

Figure 3.

A typical autoradiograph depicting UV-induced CPDs in the Rsa1 fragment of the NTS of a RADa strain. All details are as for Figure 2.