A new connection of mRNP biogenesis and export with transcription-coupled repair

Published online 30 May 2007 Nucleic Acids Research, 2007, Vol. 35, No. 12 3893–3906 doi:10.1093/nar/gkm373 A new connection of mRNP biogenesis and export with transcription-coupled repair He ´le `ne Gaillard, Ralf Erik Wellinger and Andre ´s Aguilera* Departamento de Gene ´tica, Facultad de Biologı ´a, Universidad de Sevilla, and Departamento de Biologı ´a Molecular, CABIMER, CSIC-Universidad de Sevilla, Avenida Ame ´rico Vespucio s/n, 41092 Sevilla, Spain Received March 5, 2007; Revised and Accepted April 27, 2007 ABSTRACT Although DNA repair is faster in the transcribed strand of active genes, little is known about the possible contribution of mRNP biogenesis and export in transcription-coupled repair (TCR). Interestingly, mutants of THO, a transcription complex involved in maintenance of genome integrity, mRNP biogenesis and export, were recently found to be deficient in nucleotide excision repair. In this study we show by molecular DNA repair analysis, that Sub2-Yra1 and Thp1-Sac3, two main mRNA export complexes, are required for efficient TCR in yeast. Careful analysis revealed that THO mutants are also specifically affected in TCR. Ribozyme-mediated mRNA self-cleavage between two hot spots for UV damage showed that efficient TCR does not depend on the nascent mRNA, neither in wild-type nor in mutant cells. Along with severe UV damage-dependent loss in processivity, RNAPII was found binding to chromatin upon UV irradiation in THO mutants, suggesting that RNAPII remains stalled at DNA lesions. Furthermore, Def1, a factor responsible for the degradation of stalled RNAPII, appears essential for the viability of THO mutants subjected to DNA damage. Our results indicate that RNAPII is not proficient for TCR in mRNP biogenesis and export mutants, opening new perspectives on our knowledge of TCR in eukaryotic cells. INTRODUCTION Nucleotide excision repair (NER) is an evolutionarily conserved DNA repair pathway that deals with severely distorting DNA lesions including intrastrand crosslinks such as UV-induced pyrimidine dimers [reviewed in (1,2)]. Within NER two damage-sensing pathways are recognized: one for the entire genome, global genome repair (GGR), and one for the transcribed strand of active genes, transcription-coupled repair (TCR). In yeast, GGR requires Rad7, a protein carrying leucine-rich repeats, and Rad16, a member of the SWI2/SNF2 subfamily of putative helicases (3). These proteins presumably act in a complex that might be required in chromatin remodeling to facilitate damage recognition by Rad4/Rad23 [reviewed in (1,4)]. As ongoing transcription is required for TCR, damage recognition is likely done by the elongating RNA polymerase (RNAP) itself. RNAP arrests at injuries in the template strand triggering, likely via additional specific factors, the recruitment of the DNA repair machinery [reviewed in (5–7)]. Interestingly, TCR appears to be functional once a low and basal rate of transcription is achieved, beyond which there is no simple correlation between transcription and repair rates (8). In Escherichia coli, the stalled RNAP leads to the recruitment of the transcription-repair coupling factor (TCRF) Mfd, allowing for the release of RNAP and further recruitment of the repair factors (9,10). In eukaryotes the precise mechanism of TCR remains poorly understood. Mutations in proteins required for NER lead to severe disorders known as Xeroderma pigmentosum and Cockayne’s syndrome [for review see (11)]. One of these proteins, Cockayne syndrome B protein (CSB), and its yeast ortholog Rad26, share conserved functions (12,13) and represent putative eukaryotic TCRF candidates. CSB and Rad26 belong to the SWI2/SNF2 helicase superfamily. Although CSB has been shown to have DNA-dependent ATPase activity, an ATPase-deficient mutant partially restores CSB activity in vivo (14). The putative function of CSB as a TCRF has been substantiated by in vitro reconstitution of the TCR initiation steps, in which an elongating RNAPII arrested at a DNA lesion was shown to mediate an ATP-dependent incision of the damaged DNA only in the presence of CSB (15). XPG, one of the structure-specific DNA endonuclease responsible for the removal of an oligonucleotide containing the DNA lesion in NER, is another protein involved in TCR. Recent results imply a coordinated recognition of stalled RNAPII by XPG and CSB in TCR initiation in mammalian cells and suggest that TFIIH-dependent remodeling of stalled RNAPII without *To whom correspondence should be addressed. Tel: þ34-954-468-372; Fax: þ34-954-461-664; Email: [email protected] ß2007 The Author(s) This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/ by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. release may be sufficient to allow repair (16). In yeast, the Rpb9 subunit of RNAPII has also been shown to contribute to TCR (17,18). Alternatively, and analogous to the mRNA-dependent loading of termination factors in E. coli (19,20), it is also conceivable that the nascent mRNA, or proteins bound to it, may be required to load repair enzymes at stalled polymerases. On the other hand, RNAPII is subject to ubiquitylation and proteasome-mediated degradation in response to UV-generated DNA damage (21–24). It has been proposed that degradation of damage-stalled RNAPII complexes might be an alternative to TCR (25). Indeed, recent studies have shown that arrested RNAPII elongation complexes are the preferred substrate for ubiquitylation, which is dependent on the C-terminal repeat domain (CTD) of RNAPII and on the Def1 protein in yeast (26,27). In eukaryotic cells, export of nuclear mRNA to the cytoplasm requires correct RNA-processing and the association of a number of RNA-binding proteins to form export-competent ribonucleoprotein particles (mRNP) [for review see (28–30)]. Although there is growing evidence for transcription-coupled mRNA export, the physical nature of this coupling is not known. A connection between mRNP biogenesis and transcription is provided by THO, a conserved fourprotein complex composed of stoichiometric amounts of Tho2, Hpr1, Mft1 and Thp2 (31) that is recruited to active chromatin in vivo (32,33). Null mutation of any component of THO leads to similar phenotypes of transcription impairment and mRNA export defects, as well as to a strong transcription-associated hyperrecombination phenotype (31,32,34–36). Together with the mRNA export proteins Sub2/UAP56 and Yra1/Aly, THO forms a larger complex termed TREX in yeast and humans (32). However, even though deletion of any of the THO genes leads to complete depletion of the complex (37), THO remains stable in vivo in sub2 mutants (36), indicating that it forms a core complex independently of Sub2. Strikingly, mutants of the Mex67-Mtr2 mRNA export factor—a heterodimer that mediates the interaction of the mRNP with the nuclear pore complex (NPC)—show THO-like gene expression and recombination phenotypes (36). The idea of THO being functionally involved in mRNP biogenesis and export is further strengthened by the observation that mutants of the Thp1–Sac3 complex, which has been shown to function in mRNA export by docking the mRNP to specific nucleoporins at the nuclear pore entry (38,39) confer the same transcription, mRNA export, and hyper-recombination phenotypes as do THO/TREX mutations (40,41). Previously, we showed that null mutations of the HPR1 and THO2 genes confer defects in NER (42). With this precedent and considering that RNA-binding proteins are concomitantly assembled on the nascent mRNA to generate a stable and export-competent mRNP [reviewed in (30,43,44)], we studied whether TCR might be connected to mRNP biogenesis and export. We found that THO, Sub2-Yra1, Mex67-Mtr2 and Thp1-Sac3 are required for efficient TCR in yeast, thus linking mRNP biogenesis and export to TCR. Using a construct in which a self-cleaving Hammerhead ribozyme was cloned between two hot spots for UV damage, we demonstrate that TCR does not depend on the nascent mRNA, neither in wild-type nor in THO and Thp1-Sac3 deficient strains. Chromatin immunoprecipitation (ChIP) analyses revealed that, beside a severe UV damage-dependent loss in processivity, RNAPII is found to be bound to chromatin upon UV irradiation in THO mutants. Interestingly, Def1, a factor responsible for the removal of stalled RNAPII from a DNA template, is essential for the viability of THO mutants subjected to DNA damage. Our results support a model in which mRNP biogenesis and export is required for efficient TCR by preventing the occurrence of defective RNAPII complexes, which may remain stalled at a DNA lesion. MATERIALS AND METHODS Strains and plasmids We used strains W303-1A, rad1::LEU2 (W839-5D, R. Rothstein), rad7::URA3 (MGSC97, 3), rad26::HIS3 (MGSC102, 13), hpr1::HIS3 (U674-1C), tho2::KAN (RK2-6D, 45), hpr1-101 (WH101-1A, 37), sub2::HIS3 pCM185-sub2-206 (TRP1, DLY33), sub2::HIS3 (Urasegregant from DLY23, 46), mex67-5 (WMC1-1A, 36), thp1::KAN (WFBE046, F. Fabre), which have been reported previously, and isogenic derivatives obtained by genetic crosses. The def1::HYGR strains were obtained by replacement of the DEF1 gene in rad7::URA3 hpr1::HIS3 diploids and subsequent tetrad dissection. For ChIP analyses, the GAL1 promoter fused to the 50-most 300 bp of the YLR454w open-reading frame was integrated at the YLR454w locus (32). Plasmids pHG001Rib þ and pHG001-rib m were obtained by insertion of a fragment containing 39 bp yeast T-tracts (DED1 promoter), 52 bp Hammerhead ribozyme sequence (47), the Eco47III-BssHII LacZ fragment, and 39 bp yeast T-tracts (DED1 promoter) into the Eco47III-BssHII sites of plasmid pRWE005 (48). UV survival curves Yeast cells were grown in YPD-rich medium to an OD 600 of 0.6. Plating, UV irradiation, and quantification were performed as described (42). All survival curves shown represent the average of at least three independent experiments. UV irradiation and repair Irradiation and repair was carried out as described (49) with minor modifications. Yeast cells were grown in 400 ml YPD-rich medium, or SG supplemented with the appropriate amino acids in the case of cells harboring the plasmid pHG001, to an OD 600 of 0.8, harvested, and resuspended in SD or SG to an OD 600 of 1.2. A 200 ml aliquot was irradiated with 230 J/m 2 UV light using germicidal lamps (Philips T UV 15 W). The medium was supplemented to YPD-rich or with the appropriate amino acids and the cells incubated at 308C in the dark for 3894 Nucleic Acids Research, 2007, Vol. 35, No. 12 recovery. Fifty milliliter aliquots were taken at the indicated repair times, chilled on ice and DNA purified using the described CTAB protocol (48). All steps from UV irradiation to DNA extraction were carried out in red light (Philips T LD 18 W RED). Geneand strand-specific DNA repair assays CPDs were mapped by indirect end-labeling and quantified as described (49). DNA was cut with appropriate restriction enzymes and aliquots were cut at CPDs with T4-endonuclease V (T4endoV, Epicentre) or mock treated. The DNA was electrophoresed in 1.3% alkaline agarose gels, blotted to Nylon membranes and hybridized with radioactively labeled strand-specific DNA probes. Strand-specific probes were generated by primer extension. Primers to generate DNA templates and probes that hybridize to the TSand NTS-strand, respectively, were: Rpb2-A: 50-TCTTGGAATAATAACTTCGCGGC-30; Rpb2-B: 50-GGTGGATGACAAGATACATGCC-30; pHG001-A: 50-ATTTTTGACACCAGACCAACTG-30; pHG001-B: 50-TCTGCCATTGTCAGACATGTAT-30. Membranes were analyzed and quantified with a PhosphorImager (Fuji FLA3000). The CPD content was calculated using the Poisson distribution, -ln(RF a /RF b ), where RF a and RF b represent the signal intensities of the intact restriction fragment of the T4endoVand mocktreated DNA, respectively. Region-specific damage was calculated as the signal of that region in the T4endoVtreated DNA divided by the signal of the whole lane. The corresponding signal of the mock-treated DNA was subtracted as background. The average of the initial damage generated with 230 J/m 2 was 0.3 CPD/kb. To allow direct comparison between different strains, repair curves were calculated as the fraction of CPDs removed versus repair time. The initial damage was set to 0% repair. Northern analyses RNA was extracted and northern analyses performed according to standard procedures. For RNA synthesis recovery analyses, filters were hybridized with a 324-bp long RPB2 fragment obtained by PCR using primers RPB2 A, RPB2 B. Northern blots were quantified using a Fuji FLA 3000 and normalized to the rRNA levels of each samples. For pHG001 ribozyme cleavage analyses, filters were hybridized with a 314-bp long LacZ fragment obtained by PCR using primers HG001-A and HG001-B. ChIP analyses Cells were grown and irradiated as described above. Forty milliliter aliquots were taken at the indicated repair times and cross-linked with a 1% formaldehyde solution for 15 min at RT. Glycine was added to a final concentration of 125 mM, and the cell pellets frozen in liquid nitrogen and kept at 808C. ChIP assays were performed as described (50). Monoclonal 8WG16 antibody (COVANCE) and protein A-Sepharose were used to immunoprecipitate RNAPII. The GFX purification system (GE Healthcare) was used for the last purification step. All samples were treated with 200 ng photolyase (TREVIGEN) for 30 min under photoreactivating light (Sylvania F15T8 BLB) prior to real-time quantitative PCR analysis. We used 20–30-bp oligonucleotides for the PCR amplification of two fragments of YLR454c (3–43 and 7621–7674) and the 9716–9863 intergenic region of chromosome V, which was used as non-transcribed control. Real-time quantitative PCR was performed using SYBR green dye in the 7500 Real Time PCR system (Applied Biosystems). Standard curves for all three pairs of primers were performed for each PCR analysis, all PCR reactions being performed in triplicate. The enrichment of each PCR amplification of interest was calculated as the ratio between the region-specific signal and the intergenic signal of the precipitated fractions normalized with respect to the corresponding ratios of the input fractions. At least three independent experiments were performed for each condition. Primer sequences are available upon request. RESULTS mRNA export-deficient cells are sensitive to UV in the absence of global genome repair To analyze whether defects in mRNP biogenesis and export result in impaired TCR, we studied mutants defective in both GGR and mRNA export. Isogenic sub2-206,mex67-5,rad7,sub2-206 rad7,mex67-5 rad7mutants were generated and survival after UV irradiation was determined (Figure 1, upper panel). Isogenic repair-proficient W303-1A, repair-deficient rad1and TCR-deficient rad26rad7strains were used as controls. The sub2-206 and mex65-7 single mutants show no increased UV sensitivity as compared to wild-type cells. However, upon UV irradiation viability of the sub2-206 rad7and mex67-5 rad7double mutants dropped below the levels of the rad7single mutant. Survival of the sub2-206 rad7strain was similar to survival of the rad26rad7strain, whereas mex67-5 rad7was less affected. Next, we analyzed mutants of the Thp1-Sac3 complex, which acts downstream of Mex67Mtr2 on the mRNP biogenesis and export route. Isogenic thp1,sac3,thp1rad7, and sac3rad7mutants were generated and UV survival was determined (Figure 1, middle panel). The thp1and sac3single mutants show no increased UV sensitivity as compared with wildtype cells. However, viability of the thp1rad7and sac3rad7double mutants was below the level of the rad7single mutant upon UV irradiation. Since THO mutants have been shown to be sensitive to UV irradiation in the absence of GGR (42), we performed UV survival curves of hpr1rad7and tho2rad7 strains for comparison of phenotype strength (Figure 1, lower panel). As expected, hpr1rad7and tho2rad7 survival were reduced below the levels of the rad7single mutant upon UV irradiation. UV sensitivity of sub2-206 rad7cells was stronger than hpr1rad7cells, and weaker than tho2rad7cells, whereas thp1rad7and hpr1rad7showed similar UV sensitivity, consistent with the individual phenotype of the single mutant in other assays (36,40,45,51). We have recently described the Nucleic Acids Research, 2007, Vol. 35, No. 12 3895 hpr1-101 mutant allele, which exhibits severe transcription defects but weak hyper-recombination (37). We tested the hpr1-101 allele for UV sensitivity in the absence of GGR to check whether the UV sensitivity of THO mutants was rather linked to their transcription deficiencies or to the formation of recombinogenic structures. Survival of hpr1-101 rad7and hpr1rad7cells were similar (Figure 1, lower panel), indicating that the observed UV sensitivity was linked to the transcription defects of THO mutants. Thus, because UV sensitivity in the absence of GGR is a phenotype mostly associated with TCR deficiencies, we decided to test whether Sub2-Yra1 and Thp1-Sac3 complexes are required for proficient TCR. TCR is impaired in cells defective in Sub2-Yra1, Thp1-Sac3 and THO Next, we analyzed removal of UV photoproducts in sub2and thp1cells at the molecular level. Isogenic repair-proficient wild-type, TCR-deficient rad26and GGR-deficient rad7were used as controls. Repair after UV irradiation was determined in a 4.4-kb restriction fragment containing the constitutively expressed RPB2 gene by alkaline electrophoresis and indirect end-labeling (Figure 2). As previously reported (3), rad7cells showed wild-type repair levels in the transcribed strand (TS) while repair of the non-transcribed strand (NTS) was strongly reduced. In rad26cells, repair of the TS was significantly reduced while repair of the NTS almost reached wild-type levels (13). In sub2and thp1cells, repair of the TS was severely impaired while repair of the NTS did not exhibit significant repair defects (Figure 2B). Thus, our results indicate that sub2and thp1show defects in TCR, but not in GGR. In a previous report, molecular analysis of DNA repair in hpr1and tho2cells indicated general defects in NER (42). In these studies, the UV doses used had produced extensive DNA damage, in contrast to the conditions used in this study, in which one repair event is sufficient to restore the intact DNA (about 1 CPD per restriction fragment). The observations that mutants of the Sub2-Yra1 and Thp1-Sac3 complexes, which act downstream of THO in mRNP export, are specifically affected in TS repair lead us to examine CPD removal in tho2 and hpr1cells, using our UV irradiation conditions (Figure 2A and C). Repair of the TS was significantly reduced in both strains. As observed by UV sensitivity assays in the absence of GGR (Figure 1), tho2cells were more strongly affected in TS repair than hpr1cells, reaching levels similar to rad26cells. Likewise, repair of hpr1and hpr1-101 cells were equally affected in DNA repair (data not shown). In the NTS, the repair levels of hpr1,tho2and rad26cells were similar, in contrast to previous results obtained with extensive DNA damage Figure 1. mRNA export deficient cells are sensitive to UV in the absence of global genome repair. UV sensitivity curves of sub2-206, mex67-5,thp1,sac3, hpr1-101, and rad7single and double mutants (plain symbols). Isogenic wild-type, NER-deficient rad1, and TCRand GGR-deficient and rad26rad7served as controls (empty symbols). Double mutants carrying the hpr1and tho2 mutations in combination with the rad7mutations were used as marker of phenotype strength. Data for the rad26control strain were taken from (42) (dash line). Average values from three independent experiments are plotted for each genotype. 3896 Nucleic Acids Research, 2007, Vol. 35, No. 12 Nucleic Acids Research, 2007, Vol. 35, No. 12 3897 (42), indicating that GGR is not significantly affected in hpr1and tho2cells in our conditions. In yeast, RNA synthesis is inhibited shortly after UV irradiation, probably due to the presence of CPDs in the TS of active genes. The ability of wild-type and mutant yeast cells to recover RNAPII synthesis in individual genes has been shown to mirror their strandspecific repair capacity (52). To gain additional information on the connection between the recovery of mRNA levels and TCR, kinetics of the RPB2 transcript levels were determined by northern analysis in rad26, tho2,sub2,thp1, and wild-type cells (Figure 3). A direct correlation between the RPB2 expression levels prior to UV irradiation and repair rates was not apparent, in agreement with previous work showing the absence of simple correlation between transcription and repair rates (8). However, upon UV irradiation, RPB2 transcript recovery appeared to be most efficient in rad26cells, while transcript recovery was clearly affected in all other mutant strains. This result points to THO, Sub2-Yra1 and Thp1-Sac3 behaving differently from Rad26 in response to UV damage, since they appear to undergo severe transcription impairment upon UV irradiation, in addition to their TCR deficiencies. Taken together, our results place THO, Sub2-Yra1, Mex67-Mtr2 and Thp1-Sac3 as new factors needed for efficient TCR. As these factors play a role at the interface between transcription and RNA export, it was important to determine whether the observed TCR defect was mediated by the RNAPII or by the nascent mRNA. TCR depends on RNAPII rather than on the nascent mRNA The molecular basis underlying the requirement of functional mRNP biogenesis and export factors for TCR might rely on the proper packaging of the nascent transcript or on their effect on transcription. In repairproficient cells, the nascent mRNA could mediate the TCR reaction in response to the transcriptional stalling occurring at DNA lesions. To test this possibility, we designed a construct containing two 39-bp long T-tract sequences inserted at different sites within the GAL1 promoter-driven LacZ ORF. Between the two T-tract sequences, 52 bp encoding either an active self-cleaving Hammerhead ribozyme or an inactive mutated form (47) were inserted (Figure 4A). Northern analysis confirmed a complete disappearance of the full length mRNA in the construct carrying the active ribozyme (Figure 4B), indicating that the nascent mRNA was efficiently cleaved between the T-tracts. We first assessed whether TCR efficiency depends on the integrity of the nascent mRNA in yeast wild-type cells by comparing repair rates in the T-tracts situated upstream (T1) and downstream (T2) of the ribozyme (Figure 4C and D). The initial damage was higher in Figure 2. Transcription coupled repair is impaired in mRNA export deficient cells. Southern blot analysis of representative experiments showing repair of a 4.4 kb (NsiI/PvuI) RPB2 fragment in W303 isogenic sub2,thp1,hpr1, tho2, rad26, rad7, and wt cells after UV irradiation (230 J/m 2 ). The initial damage averaged 0.28 0.05 CPD/Kb in the transcribed strand (TS, left) and 0.26 0.04 CPD/Kb in the non-transcribed strand (NTS, right). The remaining intact restriction fragment after treatment of damaged DNA with T4endoV (þUV, þT4endoV) corresponds to the fraction of undamaged DNA. Nonirradiated DNA (-UV) and DNA not treated with T4endoV (-T4endoV) were used as controls. The 90 min time point of tho2 cells was taken from a different gel derived from the same experiment (  ). As rad7 was merely used as a negative control, the analysis was restrained to the 60 and 120 min time points. (Band C) Graphical representation of the repair analysis as shown in A. Repair of TS (full symbols and lines) and NTS (empty symbols and dashed lines) are shown for each strain. The percentage of repair was determined from the signal intensities as described in ‘Materials and Methods’. Average values derived from at least two independent experiments are plotted. Figure 3. Delayed RNA synthesis recovery after UV irradiation in mRNA export deficient cells. (A) Northern blot analysis of RPB2 expression prior and after UV irradiation (230 J/m 2 ) in the indicated strains. Transcript levels prior to UV irradiation (pre UV) expressed in percent of the wild-type level are shown below the gels. (B) Graphical representation of RPB2 synthesis kinetics after UV irradiation as shown in A. Northern blots were quantified using a Fuji FLA 3000 and normalized to the rRNA levels of each samples. The transcripts levels during DNA damage recovery were normalized to the pre UV values for each strain. Average values derived from at least two independent experiments are plotted. 3898 Nucleic Acids Research, 2007, Vol. 35, No. 12 T2 than in T1, likely reflecting local differences in chromatin structure. Importantly, our results show no significant difference between repair of T1 and T2, neither in the active (Rib þ ) nor in the mutated (rib m ) ribozyme constructs, indicating that an intact and 50-capped nascent mRNA is not required for efficient TCR in wild-type cells. Nevertheless, in contrast to wild-type cells, the occurrence of sub-optimal mRNP might impede the process of TCR in mutants of THO/TREX and Thp1-Sac3. To test Figure 4. An intact nascent mRNA is not required for TCR in wild-type yeast cells. (A) Expression of a modified LacZ containing the self-cleaving Hammerhead ribozyme (Rib) between two 39-bp long T-tracts was placed under the control of the GAL1 promoter. The first (T1) and the second (T2) T-tracts as well as the Rib sequence are indicated. (B) Northern analysis of LacZ mRNA isolated from plasmids containing either an active (Rib þ ) or a mutated form of Rib (rib m ). The apparent difference in signal intensity between the two constructs probably results from distinct transcript stability, unequal transfer efficiency of short and long RNA, as well as different hybridization efficiency due to the long transcript migrating very close to the abundant rRNA. (C) Southern blot analysis of representative experiments showing repair of a 2.2 kb (HpaI/EcoRI) LacZ fragment containing either the active Rib þ (LacZ-Rib þ ) or the mutated rib m (LacZ-rib m ) in W303 cells after UV irradiation (230 J/m 2 ). Description is as in Figure 2A, except that genomic DNA was probed for the TS of LacZ. The initial damage averaged 0.3 0.005 CPD/Kb. In the T-tracts, the initial damage corresponded to 0.54% (T1, Rib þ ), 1.08% (T2, Rib þ ), 0.79% (T1, rib m ) and 1.03% (T2, rib m ) of the respective total lane signal. (D) Graphical representation of the T-tracts repair analysis as shown in C. The percentage of repair was determined from the signal intensities as described in Materials and Methods. Average values derived from two independent experiments are plotted. Nucleic Acids Research, 2007, Vol. 35, No. 12 3899 this possibility, we used the ribozyme system to assess whether wild-type TCR can be restored in T2 after ribozyme self-cleavage of the nascent mRNA in tho2 and thp1cells (Figure 5). Repair in the full length LacZ fragment was clearly below repair levels achieved in wildtype cells, confirming the TCR deficiency of tho2cells. Comparison of repair efficiencies in T1 and T2 did not show any significant difference, neither in the active (Rib þ ) nor in the mutated (rib m ) ribozyme constructs. Similar results were obtained in thp1cells (data not shown). Thus, our results indicate that cleavage of the nascent mRNA does not restore TCR in THO and Thp1 mutants. Consequently, we assume that the key player in the organization of TCR is the RNAPII itself, or some associated factors, rather than the nascent mRNA. Def1 is required for removal of stalled RNAPII complexes in THO mutants Since we can rule out an active role of the nascent mRNA in TCR, it is conceivable that an intact RNAPII complex is sufficient to mediate proficient TCR. Recently, a protein called Def1 was shown to trigger ubiquitylation and degradation of RNAPII in response to UV damage as an alternative pathway to DNA repair (26,27,53). To study the possible requirement for Def1 in the removal of trapped RNAPII presumably present in THO mutants, hpr1def1double mutants were generated and analyzed. hpr1def1double mutants were viable, but very slow growing. Since all THO/TREX mutants grow poorly at 378C, we first investigated whether deletion of DEF1 might increase the temperature sensitivity (ts) phenotype of hpr1cells (Figure 6A). An additive effect that made cells inviable at 378C was observed in the double mutant as compared to the single mutant. Next, we performed UV survival curves in isogenic def1,hpr1 and def1hpr1mutants (Figure 6B). Viability of the def1hpr1double mutants was reduced below the levels of the def1and hpr1single mutants upon UV irradiation, indicating a synergistic effect of the two mutations on UV sensitivity. In the absence of GGR, def1has been shown to be highly sensitive to UV irradiation (53). Nevertheless, the UV sensitivity of def1 hpr1rad7cells was increased as compared to def1 Figure 5. Cleavage of the nascent mRNA does not restore TCR in cells defective in THO. (A) Southern blot analysis of strand-specific repair of a 2.2-kb LacZ fragment containing either the active (Rib þ ) or mutated Rib (rib m ) between two T-tracts in tho2cells after UV irradiation (230 J/m 2 ). The initial damage averaged 0.5 0.07 CPD/Kb. In the T-tracts, the initial damage corresponded to 1.05% (T1, Rib þ ), 2.51% (T1, Rib þ ), 1.55% (T1, rib m ) and 2.2% (T2, rib m ) of the respective total lane signal. Description is as in Figure 4C. (B) Graphical representation of the repair analysis as shown in A and in Figure 4D. Repair of the intact restriction fragment in wild-type and tho2 cells (left panel) and of the T-tracts in tho2 cells (right panel) are plotted. The percentage of repair was determined from the signal intensities as described in ‘Materials and Methods’. Average values derived from two independent experiments are plotted. 3900 Nucleic Acids Research, 2007, Vol. 35, No. 12 rad7cells (Figure 6B and C). These genetic interactions between DEF1 and HPR1 indicate that Def1 is important for the viability of THO mutants subjected to stress and DNA damage. Given the exacerbated ts and UV sensitivity phenotypes of hpr1def1double mutant (Figure 6) and the reduced transcription processivity of THO mutants (54), we determined the kinetics of RNAPII distribution after UV damage in hpr1,def1,def1hpr1and wild-type cells. For this purpose, the levels of RNAPII at distal positions within the large (8 kb) YLR454w gene driven by the GAL1 promoter were analyzed by ChIPs with antibody 8WG16 directed against RNAPII (Figure 7). In wild-type cells, we observed a drop in the overall amount of RNAPII after UV irradiation that was accompanied by a significant loss of processivity along the transcribed unit (Figure 7A). Recovery of both the amount of RNAPII and processivity occurred at similar rates and were nearly complete 90 min after UV irradiation. In hpr1cells, a drop in the overall amount of RNAPII was observed after induction of UV lesions (Figure 7B). Both the drop in RNAPII density and its recovery were comparable to wild-type cells. However, the loss of processivity observed in hpr1cells was much stronger than in wild-type cells and its recovery very slow, 50% of the polymerases being lost between the 50and the 30-end of the gene 90 min after UV irradiation. Therefore, the amount of RNAPII loaded on a transcribed unit is not the limiting factor for TCR in THO mutants. Importantly, the recovery of RNAPII association toward the end of the YLR454w coding region (30-end) with increasing repair time is consistent with the repair rates observed previously (Figure 2). While the amount of RNAPII loaded on the YLR454w gene was strongly decreased after UV irradiation in wildtype and hpr1cells, only a weak drop in the amount of RNAPII was observed in def1cells (Figure 7C), probably reflecting the defects in RNAPII degradation of this mutant strain. Recovery of RNAPII on the 30-end of the gene was slow as compared to wild-type, with kinetics similar to those of hpr1, indicating a DNA damage-dependent processivity defect. In hpr1def1 cells (Figure 7D), a weak drop in the amount of RNAPII loaded on the gene was observed after UV irradiation. The amount of RNAPII localized on the 50-end of the gene was similar to the levels found in def1cells, suggesting that the hpr1mutant did not significantly alter the accumulation of RNAPII observed in the def1mutant. Noteworthy, the amount of RNAPII associated with the 50-end of the gene was repeatedly higher at 60 min than at 90 min after UV irradiation in both strains. This observation was also made in hpr1strains but never in wild-type cells. Thus, in mutant strains in which TCR and/or RNAPII degradation is impeded, RNAPII tends to accumulate on the template with a peak 60 min after UV irradiation, this effect being stronger in def1mutants which fail to ubiquitylate and degrade the stalled RNAPII. In hpr1def1cells, recovery of RNAPII at the 30-end of the gene after UV irradiation was slow, reflecting a DNA damage-dependent loss of processivity. Taken together, our data indicate that RNAPII is indeed associated with transcribed genes after UV irradiation in THO mutants, even if repair of the lesions and transcription recovery are impeded. However, no accumulation of RNAPII was detected at the 50of Figure 6. Synergistic increase of temperature and UV sensitivity phenotypes in def1 hpr1 double mutants. (A) Growth of isogenic W303 yeast strains carrying single and double combinations of the hpr1 and def1 mutations at 308C (left panel) and at 378C (right panel). (Band C) UV sensitivity curves of isogenic hpr1,def1, and rad7 single, double and triple mutants. Isogenic wild-type and NER-deficient rad1strain were used as controls. Note that the UV dose used in panel C was lower than in panel B. Description is as in Figure 1. Nucleic Acids Research, 2007, Vol. 35, No. 12 3901