Genetic Analysis of Transcription-Associated Mutation in Saccharomyces cerevisiae

Corresponding author: Sue Jinks-Robertson, Department of Biology, Emory University, 1510 Clifton Rd., Atlanta, GA 30322., jinks{at}biology.emory.edu (E-mail)

Communicating editor: M. HAMPSEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

High levels of transcription are associated with elevated mutation rates in yeast, a phenomenon referred to as transcription-associated mutation (TAM). The transcription-associated increase in mutation rates was previously shown to be partially dependent on the Rev3p translesion bypass pathway, thus implicating DNA damage in TAM. In this study, we use reversion of a pGAL-driven lys2Bgl allele to further examine the genetic requirements of TAM. We find that TAM is increased by disruption of the nucleotide excision repair or recombination pathways. In contrast, elimination of base excision repair components has only modest effects on TAM. In addition to the genetic studies, the lys2Bgl reversion spectra of repair-proficient low and high transcription strains were obtained. In the low transcription spectrum, most of the frameshift events correspond to deletions of AT base pairs whereas in the high transcription strain, deletions of GC base pairs predominate. These results are discussed in terms of transcription and its role in DNA damage and repair.

GENE products required for cellular maintenance and cell division are critical for survival and are transcribed either constitutively or in a regulated manner during crucial time periods. Previous observations in the yeast Saccharomyces cerevisiae have shown that high levels of transcription through a gene increase its rate of recombination (VOELKEL-MEIMAN et al. 1987 ; STEWART and ROEDER 1989 ; THOMAS and ROTHSTEIN 1989 ; NEVO-CASPI and KUPIEC 1994 ). Paralleling these observations, we demonstrated that high levels of transcription increase both forward and reverse mutation rates at the yeast LYS2 locus (DATTA and JINKS-ROBERTSON 1995 ), and we refer to this phenomenon as transcription-associated mutation (TAM). In the pGAL-lys2Bgl reversion analyses, we observed that a galactose-inducible +1 frameshift allele of lys2 (lys2Bgl) reverted at a 30-fold higher rate under conditions of high transcription compared to low transcription conditions. Although initial studies on the genetic requirements of the TAM events revealed no dependence on the recombinational repair pathway, 70% of the revertants showed dependence on Rev3p, a component of DNA polymerase . This polymerase is a component of an error-prone repair pathway and exhibits translesion bypass activity (NELSON et al. 1996 ), thus implicating DNA damage as a causative factor in TAM.

The genetic link between DNA damage and TAM suggests either that highly transcribed DNA is more susceptible to damage or that highly transcribed DNA is less efficiently repaired than DNA transcribed at a low level. Transcribed DNA, for example, is known to be packaged into a more open chromatin conformation than nontranscribed and silent DNA (reviewed in KORNBERG and LORCH 1995 ), which might render highly transcribed DNA more accessible to endogenous DNA-damaging agents (WARTERS et al. 1987 ; BARTLETT et al. 1991 ; LJUNGMAN and HANAWALT 1992 ). Alternatively, the transient single-stranded nature of the nontranscribed strand of an active gene may make this strand more susceptible to damage. Such targeting of damage to the nontranscribed strand has been reported for the deamination of cytosine in active genes of Escherichia coli (BELETSKII and BHAGWAT 1996 ). In addition to facilitating DNA damage, the transcriptional machinery might interfere with DNA replication or the repair of replication errors (FOX et al. 1994 ; WIERDL et al. 1996 ). An alternative possibility is that the transcriptional machinery may sterically block access of other repair machineries to damage, thereby preventing efficient repair. Finally, prolonged pausing of the transcriptional machinery at natural pause sites in the DNA could result in an aberrant triggering of transcription-coupled repair, resulting in base misincorporation and mutation (HANAWALT 1994 ).

One approach to understanding the mechanistic basis of TAM is to examine the impact of individual DNA repair pathways on this process (for comprehensive reviews of DNA repair pathways, see FRIEDBERG et al. 1995 ; WOOD 1996 ). The major DNA damage repair pathways of yeast are the nucleotide excision repair (NER), base excision repair (BER), recombination repair, and translesion synthesis (error-prone) pathways. Here we focus on the NER and BER pathways. The NER pathway is primarily involved in the repair of large, bulky adducts in the DNA, including UV-induced pyrimidine dimers and 6-4 photoproducts. In S. cerevisiae (reviewed in SWEDER 1994 ), the repair process is initiated by recognition of the lesion by Rad14p, presumably within a large repairosome complex that shares many components with transcription factor TFIIH (HOEIJMAKERS et al. 1996 ). This repairosome complex is thought to either recognize and bind to damage or to scan the DNA for bulky lesions using the helicases within the repairosome. Once a lesion is recognized and bound, the latent endonucleolytic activities of Rad1/10p and Rad2p are activated and cleave the damaged strand ~25 nucleotides 5' and 5 nucleotides 3' of the lesion. The damaged oligonucleotide can then be removed by the 5' 3' or 3' 5' helicase activity of Rad3p or Ssl2p (Rad25p), respectively, followed by resynthesis of the removed segment by DNA polymerase and ligation by DNA ligase.

Transcription-coupled repair (TCR) is a subpathway of NER that results in preferential and rapid repair of damage on the transcribed strand of active DNA relative to the nontranscribed strand or the genome overall (reviewed in SELBY and SANCAR 1994 ). The process is restricted to RNA polymerase II-transcribed genes and appears to follow the same steps as NER, with the exception of damage recognition. During transcription, certain lesions in the transcribed strand can block RNA polymerase elongation, and the resulting stalled elongation complex is assumed to be recognized by the transcription-coupling repair factor Rad26p or Rad28p, the yeast homologs of the human Cockayne Syndrome B (CSB) and A (CSA) genes, respectively (VAN GOOL et al. 1994 ; BHATIA et al. 1996 ). Rad26p and Rad28p subsequently target the NER proteins described above to the site of the damage, thus triggering rapid repair of the transcribed strand. The elongation complex is assumed to translocate upstream of the lesion, allowing repair without dissociation from the DNA template, in a manner independent of transcription elongation factor TFIIS (VERHAGE et al. 1997 ).

In contrast to the requirement for Rad26p or Rad28p in TCR, repair of damage occurring on the nontranscribed strand or in silent DNA, as well as overall genome repair, shows dependence on the Rad7 and Rad16 proteins (VERHAGE et al. 1994 ), which complex together to form NEF4, an ATP-dependent DNA damage sensor (GUZDER et al. 1998 ). The NEF4 complex is believed to scan the DNA and help assemble NER components at sites of DNA damage. Rad7p has been shown to interact with Sir3p (PAETKAU et al. 1994 ), which helps package DNA into transcriptionally silent chromatin (reviewed in LAURENSON and RINE 1992 ), and this interaction may be used to allow access of the repair machinery to silent DNA. Additionally, deletion of SIR3 results in reduced UV sensitivity of a rad7 mutant (PAETKAU et al. 1994 ). Rad16p shows some homology to the helicase domains of Rad54p and Snf2p (BANG et al. 1992 ) and this feature may be used by Rad16p to remodel chromatin and thereby allow access to the NER machinery.

The BER pathway, in contrast to the NER pathway, recognizes a wide variety of small base damages, including lesions resulting from oxidation, alkylation, and deamination (reviewed in FRIEDBERG et al. 1995 ). In BER, a lesion is specifically recognized by its cognate N-glycosylase, which removes the damaged base by cleaving the N-glycosylic bond, leaving an apurinic or apyrimidinic (AP) site and releasing a free damaged base. The DNA backbone at the AP site is then nicked by Apn1p, the major AP endonuclease of S. cerevisiae (POPOFF et al. 1990 ), or by the AP lyase activity associated with some N-glycosylases (reviewed in KROKAN et al. 1997 ). Recently, a second AP endonuclease Pde1p, which may be encoded by APN2, has been identified and may also play a role in nicking AP sites (SANDER and RAMOTAR 1997 ; JOHNSON et al. 1998 ; BENNETT 1999 ). The resulting nick in the backbone is processed by a DNA deoxyribophosphodiesterase, a 3'-phosphodiesterase, or an exonuclease to regenerate 5'-PO₄ and 3'-OH ends, thus creating a small single-stranded gap in the DNA, which can then be filled in by DNA polymerase and sealed by ligase.

The effects of high transcription are not limited to the yeast LYS2 locus, as high levels of transcription have been shown to stimulate dinucleotide repeat instability in yeast (WIERDL et al. 1996 ). Furthermore, the phenomenon of TAM is not restricted to yeast, as transcription in E. coli increases mutation rates in a derepressed leu operon (WRIGHT et al. 1999 ) and increases C T transitions in an induced kanS-D94 allele (BELETSKII and BHAGWAT 1996 ). These transcription-associated transitions have been attributed to increased deamination of cytosine on the nontranscribed strand of active genes. Although hypermutation of the mouse immunoglobulin variable region sequences has been linked to transcription (reviewed in STORB et al. 1998 ), it has been reported that estrogen-induced transcription through the thymidine kinase locus in human cells decreases the mutation rate of this locus (LIPPERT et al. 1998 ). The reason for this discrepancy is unclear but may involve locus-specific factors such as DNA sequence, local chromatin structure, or other protein assemblies. In this study, we examine (1) TAM in yeast strains deficient in NER and BER, (2) the effects of endogenous reactive oxygen species on TAM, and (3) the lys2Bgl reversion spectra generated in wild-type high and low transcription backgrounds.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media and growth conditions:
Yeast strains were grown nonselectively in YEP medium (1% yeast extract, 2% bacto peptone; 2.5% agar for plates) supplemented with 2% glycerol and 2% ethanol (YEPGE); 2% galactose, 2% glycerol, and 2% ethanol (YEPGGE); 2% raffinose (YEPR); or 2% dextrose (YEPD). Lys⁺ revertants were identified on synthetic complete (SC) medium (SHERMAN et al. 1986 ) lacking lysine and containing either 2% galactose, 2% glycerol, and 2% ethanol (GGE-LYS); 2% galactose, 2% raffinose, 10% Oxyrase (Oxyrase, Inc., Mansfield, OH), 10 mg/liter ergosterol (Sigma, St. Louis), and 0.1% polyoxyethylenesorbitan monooleate (Tween 80, Sigma) (RGO-LYS); or 2% dextrose (SC-LYS). Can^r mutants were identified on SC dextrose medium lacking arginine and supplemented with 60 mg/liter canavanine with or without 10 mg/liter ergosterol and 0.1% Tween 80 (CAN-Ergo). Anaerobic cultures were grown in a BBL GasPak chamber, using H₂/CO₂ generator envelopes (Becton Dickinson, Sparks, MD). SC medium containing 1g/liter 5-fluoroorotic acid (5-FOA) was used to select Ura^- segregants (BOEKE et al. 1987 ), and SC medium containing 3% glycerol and 0.05% 2-deoxy-D-galactose (2DG, Sigma) was used to identify gal80 strains (PLATT 1984 ). LB medium (1% yeast extract, 0.5% bacto tryptone, 1% NaCl; 1.5% agar for plates), supplemented with 100 µg/ml ampicillin as appropriate, was used for growth of E. coli strains. Yeast and bacterial strains were grown at 30° and 37°, respectively.

Strain constructions:
Yeast transformations were carried out according to GIETZ and SCHIESTL 1995 . All strains used in this study are listed in Table 1 and are isogenic derivatives of SJR195. To grow strains anaerobically in the absence of dextrose or galactose, Suc2^- strains were converted to Suc2⁺ by transformation with EcoRI-digested pRB58 (CARLSON and BOTSTEIN 1982 ). The leu2-K, leu2-R, and trp11 alleles were introduced by two-step replacement using AflII-digested pJH188, KpnI-digested pJH189 (LICHTEN et al. 1987 ), or XhoI-digested p13 (obtained from P. Hieter), respectively.

Wild-type alleles were replaced with disruption alleles using the following DNAs: NcoI/SmaI-digested pSR244 (gal80::HIS3; DATTA and JINKS-ROBERTSON 1995 ); BamHI-digested JF1344 (gal80::hisG-URA3-hisG; QUIMBY et al. 1997 ); SalI/EcoRI-digested pR1.6 (rad1::hisG-URA3-hisG; SAPARBAEV et al. 1996 ), SalI-digested pWS521 (rad2::TRP1; SIEDE and FRIEDBERG 1992 ); EcoRI/SalI-digested pBRHSURA3 (rad52::URA3; KAYTOR and LIVINGSTON 1994 ); SalI/HindIII-digested pTZrad26::URA3 (obtained from A. J. van Gool); BglI-digested prad7::LEU2 (VERHAGE et al. 1994 ); BamHI/PvuI-digested pUB23 (rad16::URA3; BANG et al. 1992 ); HindIII-digested pMK310 (ung1::URA3; obtained from P. Burgers); EcoRI/BamHI-digested pSCP19A (apn1::HIS3; RAMOTAR et al. 1991 ); NcoI/NdeI-digested pLF298 (ntg1::LEU2; BARTON and KABACK 1994 ); XhoI/SacI-digested pGEM-ntg2::hisG-URA3-hisG (YOU et al. 1998 ); or HindIII-digested pUC-SOD1::URA3 (GRALLA and VALENTINE 1991 ). When the hisG-URA3-hisG cassette was used for disruption, Ura^- segregants were isolated on 5-FOA. Putative disruption of RAD1, RAD7, RAD16, or RAD2 was assayed by UV sensitivity (28 J/m² at 254 nm); disruption of RAD52 was assayed by sensitivity to methyl methanesulfonate (MMS; Kodak; 0.008–0.016% in YEPD); disruption of SOD1 was assayed by sensitivity to paraquat (methyl viologen; Sigma; 50 mM in YEPD). Strains carrying a disrupted GAL80 gene were screened for 2DG sensitivity. All disruptions were confirmed by Southern blot or PCR analysis.

A PCR-generated ogg1::kan disruption fragment was used to delete OGG1. Primers 5'-ATGTCTTATAAATTCGGCAAACTTGCCATTAATAAAAGTGAGCTATGTCTAGCAAATGTGcagctgaagcttcgtacg-3' (forward) and 5'-CTAATCTATTTTTGCTTCTTTGATGTGAAGATCAGACAATTCAACTTTCAGTTTCATTTGaggccactagtggatctg-3' (reverse) were used to amplify an ~1-kb disruption fragment using pFA6-kanMX2 (WACH et al. 1994 ) as template. The first 60 bases of each primer (uppercase) are complementary to OGG1, and the 3' ends (lowercase) are complementary to the kanamycin resistance cassette. Following transformation, cells were outgrown for 3 hr in YEPD before selective plating on YEPD plates containing 200 mg/liter Geneticin (Sigma). After 2 days incubation plates were replica plated onto fresh Geneticin-containing medium. Resistant colonies were confirmed for OGG1 disruption by PCR.

Measurement of reversion rates:
Reversion rates of lys2Bgl were determined by the method of the median (LEA and COULSON 1948 ). Independent 2-day-old colonies were inoculated into 5 ml of YEPGE liquid medium, with exceptions noted below, and grown nonselectively to 2 x 10⁸ cells/ml on a roller drum. Cells were harvested by centrifugation, washed once with sterile H₂O, and resuspended in 1 ml H₂O. Aliquots (100 µl) of appropriate dilutions were plated onto GGE-LYS to select Lys⁺ revertants, CAN to select forward mutations at the CAN1 locus, and onto YEPD to determine viable cell numbers. Can^r colonies were counted on day 2 and Lys⁺ colonies on day 3 after selective plating. The data from a minimum of 10 cultures were used for each rate determination. sod1 strains and the isogenic parents were grown nonselectively under aerobic conditions in YEPR liquid medium, washed as described above, plated selectively onto RGO-LYS and CAN-Ergo, and then incubated anaerobically. This modification is necessary as sod1 mutants must be grown anaerobically to phenotypically express the Lys⁺ or Can^r phenotype (CHANG and KOSMAN 1990 ; GRALLA and VALENTINE 1991 ).

To eliminate outlying cultures (in terms of viable cell numbers) for rate determinations, the median culture population for each strain was determined among all cultures. This median was used to center a twofold acceptable range [median ± (median/3)] for individual culture populations, and outliers were excluded from further analysis. Contingency chi-square analysis was used to determine whether rates were significantly different (WIERDL et al. 1996 ).

Isolation of revertants and DNA sequence analysis:
To isolate independent Lys⁺ revertants for DNA sequence analysis, 1-ml YEPGE cultures were grown as described above and a single aliquot was plated onto GGE-LYS. One revertant from each culture was purified and total genomic DNA was prepared from 2 ml of cells by glass bead lysis (HOFFMAN and WINSTON 1987 ). The lys2Bgl reversion window was amplified using primers 5'-CGGCTCGAGCGCTGATTAAATTACCCCAG-3' (forward, PGAL10X) and 5'-CTACCTCAGCTCGATGTG-3' (reverse, PCRLYSB) and sequenced as described previously (GREENE and JINKS-ROBERTSON 1997 ). High and low transcription reversion spectra were compared using the C++ version of the Adams and Skopek algorithm (CARIELLO et al. 1994 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

pGAL-lys2Bgl reversion system:
The genetic reporter used in this study is a chromosomally located lys2 allele that has an internal BglII site filled in, resulting in a +4 frameshift mutation (lys2Bgl). The lys2Bgl allele spontaneously reverts at a detectable frequency, and almost all reversion events are compensatory, second-site -1 frameshift events that restore the correct reading frame of the gene (GREENE and JINKS-ROBERTSON 1997 ). Transcriptional control of the lys2Bgl allele was achieved by placing it under the control of the highly inducible GAL1/10 promoter, which is regulated by the Gal4p activator and Gal80p repressor. Gal4p recognizes a sequence in the GAL1/10 promoter and in the presence of galactose serves as an activator of transcription. In the absence of galactose, Gal80p binds to Gal4p and masks the activation domain of Gal4p. Upon addition of galactose to the medium, the Gal80/4p interaction changes such that the Gal4p activation domain is uncovered and transcription can be activated (reviewed in LOHR et al. 1995 ). In the absence of galactose, transcription from pGAL1/10 occurs at a very low level in wild-type strains, while in gal80 strains transcription from pGAL1/10 occurs at a constitutively high rate. In the experiments reported here, we used isogenic Gal80⁺ and Gal80^- strains as low and high transcription strains, respectively. Glycerol and ethanol were used as carbon sources in all experiments, thus avoiding the growth lag and physiological changes associated with switching from glucose to galactose in GAL80 strains. The level of pGAL1/10 induction in gal80 strains ranges from 50-fold as measured by Northern analysis to 1000-fold as measured by ß-galactosidase activity assays (DATTA and JINKS-ROBERTSON 1995 ; N. J. MOREY and S. JINKS-ROBERTSON, unpublished observations).

Previous studies on the effect of transcription on the reversion of the lys2Bgl allele suggested a causative role for DNA damage. To further examine the role of DNA damage in TAM, we examined transcription-associated Lys⁺ rates in strains deficient for components of various DNA repair pathways. The logic behind this analysis is that strains deficient in the repair of lesions that accumulate under conditions of high transcription should show a synergistic increase in the reversion rate of the lys2Bgl allele. That is, relative to a repair-proficient, low transcription strain, the increase in lys2Bgl reversion observed in a repair-defective, high transcription strain should be greater than the sum of the increases observed in the wild-type, high transcription and repair-defective, low transcription strains. An additive effect of a repair defect and a high transcription state on lys2Bgl reversion would indicate no relationship between the relevant repair pathway and TAM.

In addition to examining lys2Bgl reversion in repair-deficient strains, we also measured the forward mutation rate at the CAN1 locus. CAN1 encodes arginine permease, which transports the toxic arginine analog L-canavanine as well as arginine (HOFFMANN 1985 ). Mutation of CAN1 results in resistance to canavanine (WHELAN et al. 1979 ), and forward mutations in CAN1 can be selected on medium containing canavanine and lacking arginine. The CAN1 locus therefore provides an easy system for measurement of forward mutation rate in different genetic backgrounds.

Role of NER in the etiology of TAM:
Rad1/10p and Rad2p are the endonucleases that nick the phosphodiester backbone of DNA on either side of a damaged nucleotide(s) to initiate repair (reviewed in FRIEDBERG et al. 1995 ). As shown in Table 2, low transcription strains lacking a functional Rad1p have an ~3-fold increase in the reversion rate of lys2Bgl compared to a wild-type, NER-proficient strain. A high level of transcription alone stimulates the lys2Bgl reversion rate ~9-fold in a wild-type background. Relative to the low transcription, repair-proficient strain, a synergistic 64-fold increase in the reversion rate is observed when Rad1p is eliminated and lys2Bgl is highly transcribed. This synergism is specific to the highly transcribed lys2Bgl allele, as Can^r rates are similarly elevated in both the low transcription, rad1, and high transcription, rad1 backgrounds. In a Rad2p-deficient strain, there is no significant increase in the Lys⁺ rate compared to that in a wild-type, low transcription strain, but a synergistic 30-fold increase in the Lys⁺ rate is seen under high transcription conditions. As observed in the rad1 strain, the effect of rad2 is specific to the highly transcribed allele.

Previous studies in recombination-deficient (rad52) strains suggested a role of recombination in the repair of lesions associated with high levels of transcription (DATTA and JINKS-ROBERTSON 1995 ). Given the role of Rad1p in some recombination pathways (IVANOV and HABER 1995 ) and the different effects of deletion of RAD1 vs. RAD2, we examined lys2Bgl reversion rates and forward mutation at CAN1 in rad1 rad52 and rad2 rad52 double mutants. The synergism seen between the single rad1 or rad2 and rad52 in lys2Bgl reversion and CAN1 forward mutation suggests that NER and recombination are competing for the repair of lesions. With regard to TAM, the effects of the rad1 rad52 mutations and high transcription are much greater than additive. It should be noted that in the double mutant strains, which can no longer perform NER, single-stranded annealing, or traditional recombination, the rad1 rad52 strain consistently exhibits a higher lys2Bgl reversion rate and Can^r rate than the rad2 rad52 strain.

NER-associated strand bias in TAM:
The strand bias for repair of DNA damage via NER prompted us to examine whether there is a possible strand bias in the repair of transcription-associated damage. Mutant strains deficient for NER of either the transcribed (rad26 mutant) or nontranscribed (rad7 and rad16 mutants) strand of active genes were examined to detect a possible strand bias in the accumulation of DNA damage. An enhanced rate of TAM in a rad26 strain would be expected if damage is introduced preferentially on the transcribed strand, whereas an enhanced rate of TAM would be expected in rad7 and rad16 strains if damage accumulates primarily on the nontranscribed strand. As shown in Table 2, elimination of neither Rad26p nor Rad7p impacts TAM, but a statistically significant increase specific to TAM is evident in the rad16 strain.

Role of BER in the etiology of TAM:
To investigate the possible role of nonbulky base damages in TAM, strains deficient for various components of the BER pathway were constructed and examined for reversion of lys2Bgl under low and high transcription conditions (Table 3). Cytosine deamination results in uracil and removal of uracil from DNA is initiated by the Ung1p glycosylase in yeast (IMPELLIZZERI et al. 1991 ). Although the low transcription lys2Bgl reversion rate is not elevated in an ung1 mutant, the high transcription reversion rate is elevated approximately twofold. In both low and high transcription backgrounds a modest mutator effect is evident at the CAN1 locus.

To examine a broad spectrum of potential base damages, strains deficient in enzymes involved in processing AP sites were constructed. Strains carrying a deletion of APN1 do not exhibit an increase in lys2Bgl reversion under low transcription conditions but do exhibit a twofold increase in reversion of this marker when it is highly transcribed. No effect on Can^r rates is evident in the apn1 strains. As genetic data indicate that the AP lyase activity of the Ntg proteins can compete with Apn1p for the repair of abasic sites in vivo (SWANSON et al. 1999 ), the analysis of the BER pathway was extended by examining ntg1, ntg2, and ntg1 ntg2 double mutants. No significant change in the reversion rate of the low or high transcription lys2Bgl allele or in forward mutation rates at CAN1 is seen in any of these backgrounds (Table 3 and data not shown). When an ntg1 ntg2 apn1 triple mutant is examined, however, an increase in the lys2Bgl reversion rate and the Can^r forward mutation rate is seen in both low and high transcription strains. Although the low transcription lys2Bgl reversion rate in the ntg1 ntg2 apn1 triple mutant is significantly elevated relative to the apn1 or ntg1 ntg2 mutant, the high transcription reversion rate in the triple mutant is similar to that in an apn1 single mutant under high transcription conditions.

Finally, to address the possible role of oxidative base damage in the etiology of TAM, strains defective in the recognition and removal of various oxidized bases were examined. The N-glycosylase activities of Ntg1p and Ntg2p are involved in the recognition and excision of oxidized pyrimidines (EIDE et al. 1996 ; AUGERI et al. 1997 ; YOU et al. 1998 ) while Ogg1p is involved in the repair of oxidized purines, particularly 8-oxo-guanine (VAN DER KEMP et al. 1996 ). Mutants deficient for Ntg1p, Ntg2p, Ogg1p, or all three proteins were constructed and the reversion of lys2Bgl was examined. As shown in Table 3, there is a small, but significant, decrease in the reversion rate in both low and high transcription strains in an ogg1 background (0.6-fold in both instances), and yet a significant mutator effect is seen at the CAN1 locus (8- to 9-fold). A similar decrease in the lys2Bgl reversion rate in the ntg1 ntg2 ogg1 triple mutant is observed under high transcription conditions, as well as an ~2-fold decrease in the Can^r rates of both low and high transcription backgrounds.

Role of endogenous oxygen radicals in TAM:
The possible role of oxidative damage in TAM was further investigated by examining the reversion rate of lys2Bgl in strains lacking the cytosolic Cu, Zn superoxide dismutase enzyme Sod1p, which catalyzes the dismutation of the highly reactive superoxide anion to hydrogen peroxide (FRIDOVICH 1978 ). A synergistic increase in the reversion rate of lys2Bgl would be predicted in a sod1 mutant under high transcription conditions if an elevated level of oxidative damage is involved in the mechanism of TAM. The selective plating for these experiments was done under anaerobic conditions as, regardless of genotype, sod1 strains are phenotypically Lys^- and Can^s when grown aerobically. Furthermore, as glycerol and ethanol cannot be used as carbon sources under anaerobic conditions, raffinose was substituted as the noninducing, nonrepressing carbon source for nonselective and selective growth. We note that these growth conditions result in a different lys2Bgl reversion rate compared to that obtained in the NER and BER experiments (Table 4). In both low and high transcription backgrounds, the sod1 strain shows a 10- to 15-fold mutator effect at the CAN1 locus (Table 4). Under conditions of low transcription, the sod1 strain shows an ~5-fold increase in the reversion of lys2Bgl, resulting in a rate comparable to the high transcription reversion rate in a wild-type background. Under high transcription conditions, the sod1 strain does not exhibit a significant increase in the lys2Bgl reversion rate relative to the SOD1 strain.

View this table: In this window In a new window   Table 4. Relative mutation rates in a sod1 strain

Reversion spectra analysis:
To examine the molecular nature of reversion events, independent Lys⁺ revertants from repair-proficient strains under low and high transcription conditions were isolated, and the lys2Bgl reversion window was sequenced. The low and high transcription spectra thus obtained are shown in Fig 1 and summarized in Table 5. Reversion events in both spectra are found throughout the window, and as observed previously, homopolymer runs >3N are hotspots for frameshift events (GREENE and JINKS-ROBERTSON 1997 ). Comparison of the low and high transcription spectra by the Adams and Skopek algorithm (CARIELLO et al. 1994 ) reveals them to be significantly different (P < 10^-7). Although approximately one-half of the events occur in homopolymer runs, the single base-pair deletions preferentially occur in the 6A run under low transcription conditions, but in the 4C run under high transcription conditions. This deletion bias pertains not only to the homopolymer runs, but also to single base-pair deletions in noniterated sequences. Under low transcription conditions 60% of the mutations are AT deletions and 24% are GC deletions, whereas 22% are AT deletions and 74% are GC deletions under high transcription conditions (Table 5). Using these percentages and the mutation rates, we estimate that transcription increases the rate of GC deletions 28-fold, but increases the rate of AT deletions only 3.4-fold.

View larger version (15K): In this window In a new window Download PPT slide   Figure 1. lys2Bgl reversion spectra in wild-type (A) low and (B) high transcription strains (n = 50). The reversion window is diagrammed at the top, with the +4 insertion that created the lys2Bgl allele underlined and in boldface type. Each single-base deletion is indicated below the sequences; A and T deletions are represented by a and G and C deletions are represented by a . Insertions are indicated above the sequences.

View this table: In this window In a new window   Table 5. lys2Bgl reversion spectra

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study, we have examined the roles of DNA damage repair pathways in TAM avoidance and report lys2Bgl reversion spectra under low vs. high transcription conditions. Although the lys2Bgl frameshift reversion system is a convenient assay for studying the genetic requirements of TAM, the system is limited in that only those mutations that restore the correct reading frame of the LYS2 gene are recovered. Results obtained in wild-type and mutant strains thus reflect only a limited subset of all the possible mutations that can occur within the highly transcribed gene, as the target is small and frameshifts generally account for only 10% of all mutations (KUNZ et al. 1998 ). This ascertainment bias must be considered when interpreting the data.

NER is a major DNA damage repair pathway that generally removes large, bulky, helix-distorting lesions (reviewed in FRIEDBERG et al. 1995 ). The Rad1/10 and Rad2 proteins are crucial for the incision steps of NER, and elimination of either completely blocks the pathway. Simultaneous deletion of RAD1 and high levels of transcription stimulated lys2Bgl reversion 64-fold, an increase that represents a greater-than-additive effect. This result suggests that NER is involved in repairing the damage responsible for TAM, and it also eliminates gratuitous TCR as a causative factor in TAM. If gratuitous TCR was responsible for a substantial proportion of the transcription-associated lys2Bgl reversions, then the rate of reversion would be expected to decrease, not increase, upon elimination of the NER pathway. Surprisingly, the effect of deleting RAD2 was consistently less than that of deleting RAD1. This observation suggests either that a redundant Rad2p function exists in S. cerevisiae or that the additional role of Rad1p is derived from other, non-NER pathways, such as recombination (IVANOV and HABER 1995 ). Combining the NER deficiencies with a deletion of RAD52, which eliminates the yeast recombination pathway, resulted in synergistic increases in the reversion rate of lys2Bgl. This result implies that recombination is likely competing with NER for repair of the lesions that accumulate under conditions of high transcription.

TCR is an important subpathway of NER and allows preferential repair of the template strand of actively transcribed genes (reviewed in FRIEDBERG et al. 1995 ). The preferential deamination of cytosine on the nontranscribed strand of active genes in E. coli (BELETSKII and BHAGWAT 1996 ) prompted us to search for a possible strand bias in yeast TAM using rad26, rad7, and rad16 mutants. Rad26p has been implicated as the major transcription-coupling repair factor involved in preferential repair of the transcribed strand of DNA in yeast (VAN GOOL et al. 1994 ; BHATIA et al. 1996 ; VERHAGE et al. 1996 ), while Rad7p and Rad16p are involved in the repair of the genome overall as well as the nontranscribed strand of active genes (VERHAGE et al. 1994 ). If a strand bias exists, we would expect to see an increase in lys2Bgl reversion only in mutants deficient in repair of the strand accumulating damage during transcription. Although no change in the high transcription reversion rate was observed in a rad26 or rad7 background, the rad16 mutant exhibited a 3-fold increase in the high transcription reversion rate (31-fold vs. 9-fold). These observations suggest that there may be a bias toward damage accumulating on the nontranscribed strand under the conditions of high transcription, but further experiments are needed to fully answer this question. For example, if there is a transcription-associated bias in the strand that accumulates mutations, then one might expect the mutation spectrum to change if the direction of pGAL-driven transcription through the lys2Bgl locus was reversed. Although other studies have found no difference between the phenotypes of rad7, rad16, and double mutant strains (VERHAGE et al. 1994 ; REED et al. 1996 ; SCOTT and WATERS 1997 ; TIJSTERMAN et al. 1997 ), the subtle difference between Rad7p and Rad16p uncovered in the lys2Bgl reversion system indicates separate repair functions of these two proteins under some conditions.

The BER pathway repairs frequently occurring base damages resulting from oxidation, alkylation, deamination, and depurination (reviewed in FRIEDBERG et al. 1995 ). Unlike the lesions recognized by NER, the lesions removed by BER generally do not cause significant helix distortion. Uracil in DNA arises via cytosine deamination and is excised by the yeast Ung1p glycosylase (IMPELLIZZERI et al. 1991 ), leaving an abasic (AP) site that is assumed to be processed by Apn1p, which nicks the DNA backbone immediately 5' of the AP site. Because a specific increase in C T transitions occurs in Ung1p-deficient strains (IMPELLIZZERI et al. 1991 ), one would not expect the rate of frameshift mutations to be altered in an ung1 mutant. Although no change in the lys2Bgl reversion rate was observed in ung1 or apn1 low transcription strains, a twofold increase in the reversion rate was observed for the highly transcribed lys2Bgl allele in each of these repair-deficient backgrounds. These data suggest that cytosine deamination to uracil, and possibly other types of base damage, may contribute to TAM. In the absence of Ung1p, for example, unrepaired GU mismatches in the DNA might be processed by a pathway that can potentially generate frameshift mutations, resulting in the increased reversion of the lys2Bgl allele. The assumed Apn1p dependence of AP site processing resulting from uracil removal and the indistinguishable phenotypes of the ung1 and apn1 mutants prevent us from assessing possible effects of base damages other than cytosine deamination in the apn1 mutant.

AP sites are created by the action of N-glycosylases, but bases can also be lost spontaneously from both double- and single-stranded DNA, with purines being liberated much more frequently than pyrimidines (reviewed in LINDAHL 1993 ). Recent data indicate that the AP lyase activities of Ntg1p and Ntg2p can compete with Apn1p for the processing of abasic sites in vivo (SWANSON et al. 1999 ). An assessment of AP sites in the generation of TAM using the ntg1 ntg2 apn1 triple mutant revealed a 4.4-fold increase in the reversion rate of lys2Bgl under low transcription conditions, compared to the lack of an increase in the ntg1 ntg2 double or apn1 single mutants. Under high transcription conditions, the elevation of lys2Bgl reversion in the ntg1 ntg2 apn1 triple mutant was approximately equal to that in the apn1 mutant (22-fold vs. 17-fold). While this elevation is clearly not synergistic, an additive effect cannot be excluded.

Oxidative damage as a potential causative factor in TAM was examined as this is generally assumed to be the major source of in vivo base damage. Ogg1p is the S. cerevisiae homolog of the E. coli Fpg protein (VAN DER KEMP et al. 1996 ) that releases oxidized purines, while Ntg1p and Ntg2p act primarily on oxidized pyrimidines (EIDE et al. 1996 ; AUGERI et al. 1997 ; YOU et al. 1998 ). Elimination of Ogg1p decreased the lys2Bgl reversion rate approximately twofold under both low and high transcription conditions but resulted in a significant mutator effect at CAN1, which is assumed to represent primarily base substitution mutations. Elimination of Ntg1p and Ntg2p had no effect on lys2Bgl reversion, and the ntg1 ntg2 ogg1 triple mutant resembled the ogg1 single mutant. The decrease in lys2Bgl reversion in ogg1 mutants suggests that the damage processing, and not the initial lesion, gives rise to frameshift mutations. Frameshift mutations presumably cannot be caused by simple miscoding at oxidized bases and therefore would be dependent on removal of the initial lesion. It is interesting to note that Ogg1p, unlike Ung1p examined above, has an associated AP-lyase activity, which specifically nicks the backbone 3' of the AP site. The difference in the ung1 vs. ogg1 phenotypes raises the interesting possibility that the role of a particular type of damage or its cognate DNA repair enzyme in TAM may be determined by the mode of AP site incision and subsequent DNA resynthesis reaction.

In addition to mutationally altering the repair of oxidative lesions, we also utilized sod1 strains to assess the potential role of oxidative damage in TAM. sod1 strains are deficient in the cytosolic Cu, Zn superoxide dismutase and are presumed to have increased intracellular pools of reactive oxygen species (reviewed in FRIDOVICH 1978 ; FEE 1982 ). In a Sod1p-deficient strain, there was a significant increase in the reversion rate of lys2Bgl only under low transcription conditions. Although the sod1 results do not exclude an involvement of oxidative damage in TAM, they do indicate that it is not the major lesion responsible for TAM.

DNA sequence analysis of the lys2Bgl reversion events revealed a clear difference in the low and high transcription frameshift spectra. Under low transcription conditions AT base pairs were deleted more often than GC base pairs, while under conditions of high transcription, a clear bias for deletion of GC base pairs was seen. The reason for this bias is not clear, although a preferential deletion of GC base pairs has been seen in E. coli cells treated with singlet oxygen (VAN DEN AKKER et al. 1994 ) and GC base pairs are targeted during transcription-associated somatic hypermutation in the immune system (BACHL and WABL 1996 ). Due to the nature of the lys2Bgl frameshift assay, only compensatory frameshift mutations within the defined reversion window can be recovered, thereby limiting the types of events detected. An analysis of transcription-associated forward mutations at the LYS2 locus would likely be more informative, but given the size of the target locus a meaningful spectrum would be very difficult to obtain. It should be noted that the high transcription spectrum reported here does not resemble the lys2Bgl reversion spectrum in mismatch-repair defective yeast cells (GREENE and JINKS-ROBERTSON 1997 ), where >90% of frameshift events are in homopolymer runs and no GC vs. AT deletion bias is evident. This suggests that the high level of transcription does not significantly impact mismatch repair, although analyses of poly(GT) stability suggested that transcription may interfere with both mismatch repair and polymerase fidelity (WIERDL et al. 1996 ).

Genetic analyses have revealed competition among the NER, translesion synthesis, recombination, and the BER pathways for repair of alkylation damage (XIAO et al. 1996 ; JOHNSON et al. 1998 ; XIAO and CHOW 1998 ), as well as competition among the NER, BER, recombination, and translesion bypass pathways for the repair of spontaneous and induced damage, presumably abasic sites (SWANSON et al. 1999 ). Furthermore, it was previously shown that 70% of the transcription-associated lys2Bgl reversion events are dependent on the Rev3p translesion bypass pathway (DATTA and JINKS-ROBERTSON 1995 ). In the transcription-associated lys2Bgl reversion assay BER deficiencies gave relatively weak phenotypes compared to the NER and recombination deficiencies. This observation can be explained in three ways, none of which is mutually exclusive. First, the effect of BER in the prevention of TAM may be underestimated because of the frameshift-specific assay used. The BER pathway normally does not produce frameshift mutations because of the very limited DNA synthesis that occurs following removal of the lesion and processing of the AP site. Second, the damage associated with high levels of transcription may not be appreciably repaired via BER. Finally, transcription may direct the associated damage into other, non-BER pathways. With regard to this last possibility, it should be noted that the majority of lesions repaired by base excision repair do not block RNA polymerase but simply result in miscoding by RNA polymerase. This is unlike the lesions typically recognized by NER, which do cause RNA polymerase to stop until repair of the lesion is completed. The high level of transcriptional activity could effectively preclude the BER enzymes from recognizing their cognate lesions, allowing the lesion either to be repaired by the NER, recombination, or error-prone repair pathways or to be fixed during the next round of DNA replication. Thus, if BER is eliminated in addition to having the lys2Bgl allele highly transcribed, one would not expect to see a further increase in TAM.

The results presented here suggest that transcription-associated mutation in yeast is likely the result of both bulky helix distorting lesions and small base damages. Although the genetic analyses clearly implicate DNA damage in TAM, we are unable to distinguish between increased levels of DNA damage vs. interference with the DNA repair machinery under conditions of high transcription. Regardless of the precise mechanism, those genes that are highly expressed or induced under certain environmental stresses may prove to be more susceptible to mutation, and mutation rates may fluctuate in a cell-cycle- or tissue-specific manner.

	ACKNOWLEDGMENTS

We thank the fellow members of the Jinks-Robertson lab for their helpful comments on the manuscript. S. Jinks-Robertson was supported by National Institutes of Health (N.I.H.) grant GM-38464. N. J. Morey was partially supported by an N.I.H. predoctoral training grant T32 GM-08490-04 and C. N. Greene was partially supported by the Graduate Division of Biological and Biomedical Sciences of Emory University.

Manuscript received June 30, 1999; Accepted for publication September 16, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AUGERI, L., Y. M. LEE, A. B. BARTON, and P. W. DOETSCH, 1997  Purification, characterization, gene cloning, and expression of Saccharomyces cerevisiae redoxyendonuclease, a homolog of Escherichia coli endonuclease III. Biochemistry 36:721-729[Medline].

BACHL, J. and M. WABL, 1996  An immunoglobulin mutator that targets G·C base pairs. Proc. Natl. Acad. Sci. USA 93:851-855[Abstract/Free Full Text].

BANG, D. D., R. VERHAGE, N. GOOSEN, J. BROUWER, and P. VAN DE PUTTE, 1992  Molecular cloning of RAD16, a gene involved in differential repair in Saccharomyces cerevisiae.. Nucleic Acids Res. 20:3925-3931[Abstract/Free Full Text].

BARTLETT, J. D., D. A. SCICCHITANO, and S. H. ROBISON, 1991  Two expressed human genes sustain slightly more DNA damage after alkylating agent treatment than an inactive gene. Mutat. Res. 255:247-256[Medline].

BARTON, A. B. and D. B. KABACK, 1994  Molecular cloning of chromosome I DNA from Saccharomyces cerevisiae: analysis of the genes in the FUN38-MAK16-SPO7 region. J. Bacteriol. 176:1872-1880[Abstract/Free Full Text].

BELETSKII, A. and A. S. BHAGWAT, 1996  Transcription-induced mutations: increase in C to T mutations in the nontranscribed strand during transcription in Escherichia coli.. Proc. Natl. Acad. Sci. USA 93:13919-13924[Abstract/Free Full Text].

BENNETT, R. A., 1999  The Saccharomyces cerevisiae ETH1 gene, an inducible homolog of exonuclease III that provides resistance to DNA-damaging agents and limits spontaneous mutagenesis. Mol. Cell. Biol. 19:1800-1809[Abstract/Free Full Text].

BHATIA, P. K., R. A. VERHAGE, J. BROUWER, and E. C. FRIEDBERG, 1996  Molecular cloning and characterization of Saccharomyces cerevisiae RAD28, the yeast homolog of the human Cockayne syndrome A (CSA) gene. J. Bacteriol. 178:5977-5988[Abstract/Free Full Text].

BOEKE, J. D., J. TRUEHEART, G. NATSOULIS, and G. R. FINK, 1987  5-Fluoroorotic acid as a selective agent in yeast molecular genetics. Methods Enzymol. 154:164-175[Medline].

CARIELLO, N. F., W. W. PIEGORSCH, W. T. ADAMS, and T. R. SKOPEK, 1994  Computer program for the analysis of mutational spectra: application to p53 mutations. Carcinogenesis 15:2281-2285[Abstract/Free Full Text].

CARLSON, M. and D. BOTSTEIN, 1982  Two differentially regulated mRNAs with different 5' ends encode secreted and intracellular forms of yeast invertase. Cell 28:145-154[Medline].

CHANG, E. C. and D. J. KOSMAN, 1990  O₂-dependent methionine auxotrophy in Cu, Zn superoxide dismutase-deficient mutants of Saccharomyces cerevisiae.. J. Bacteriol. 172:1840-1845[Abstract/Free Full Text].

DATTA, A. and S. JINKS-ROBERTSON, 1995  Association of increased spontaneous mutation rates with high levels of transcription in yeast. Science 268:1616-1619[Abstract/Free Full Text].

EIDE, L., M. BJORAS, M. PIROVANO, I. ALSETH, and K. G. BERDAL et al., 1996  Base excision of oxidative purine and pyrimidine DNA damage in Saccharomyces cerevisiae by a DNA glycosylase with sequence similarity to endonuclease III from Escherichia coli.. Proc. Natl. Acad. Sci. USA 93:10735-10740[Abstract/Free Full Text].

FEE, J. A., 1982  Is superoxide important in oxygen poisoning? Trends Biochem. Sci. 7:84-86.

FOX, M. S., J. P. RADICELLA, and K. YAMAMOTO, 1994  Some features of base pair mismatch repair and its role in the formation of genetic recombinants. Experientia 50:253-260[Medline].

FRIDOVICH, I., 1978  The biology of oxygen radicals. Science 201:875-880[Abstract/Free Full Text].

FRIEDBERG, E. C., G. C. WALKER and W. SIEDE, 1995 DNA Repair and Mutagenesis. American Society for Microbiology, Washington, DC.

GIETZ, R. D. and R. H. SCHIESTL, 1995  Transforming yeast with DNA. Methods Mol. Cell. Biol. 5:255-269.

GRALLA, E. B. and J. S. VALENTINE, 1991  Null mutants of Saccharomyces cerevisiae Cu, Zn superoxide dismutase: characterization and spontaneous mutation rates. J. Bacteriol. 173:5918-5920[Abstract/Free Full Text].

GREENE, C. N. and S. JINKS-ROBERTSON, 1997  Frameshift intermediates in homopolymer runs are removed efficiently by yeast mismatch repair proteins. Mol. Cell. Biol. 17:2844-2850[Abstract].

GUZDER, S. N., P. SUNG, L. PRAKASH, and S. PRAKASH, 1998  The DNA-dependent ATPase activity of yeast nucleotide excision repair factor 4 and its role in DNA damage recognition. J. Biol. Chem. 273:6292-6296[Abstract/Free Full Text].

HANAWALT, P. C., 1994  Transcription-coupled repair and human disease. Science 266:1957-1958[Free Full Text].

HOEIJMAKERS, J. H., J. M. EGLY, and W. VERMEULEN, 1996  TFIIH: a key component in multiple DNA transactions. Curr. Opin. Genet. Dev. 6:26-33[Medline].

HOFFMAN, C. S. and F. WINSTON, 1987  A ten-minute DNA preparation from yeast efficiently releases autonomous plasmids for transformation of Escherichia coli.. Gene 57:267-272[Medline].

HOFFMANN, W., 1985  Molecular characterization of the CAN1 locus in Saccharomyces cerevisiae. A transmembrane protein without N-terminal hydrophobic signal sequence. J. Biol. Chem. 260:11831-11837[Abstract/Free Full Text].

IMPELLIZZERI, K. J., B. ANDERSON, and P. M. BURGERS, 1991  The spectrum of spontaneous mutations in a Saccharomyces cerevisiae uracil-DNA-glycosylase mutant limits the function of this enzyme to cytosine deamination repair. J. Bacteriol. 173:6807-6810[Abstract/Free Full Text].

IVANOV, E. L. and J. E. HABER, 1995  RAD1 and RAD10, but not other excision repair genes, are required for double-strand break-induced recombination in Saccharomyces cerevisiae.. Mol. Cell. Biol. 15:2245-2251[Abstract].

JOHNSON, R. E., C. A. TORRES-RAMOS, T. IZUMI, S. MITRA, and S. PRAKASH et al., 1998  Identification of APN2, the Saccharomyces cerevisiae homolog of the major human AP endonuclease HAP1, and its role in the repair of abasic sites. Genes Dev. 12:3137-3143[Abstract/Free Full Text].

KAYTOR, M. D. and D. M. LIVINGSTON, 1994  Saccharomyces cerevisiae rad52 alleles temperature-sensitive for the repair of DNA double-strand breaks. Genetics 137:933-944[Abstract].

KORNBERG, R. D. and Y. LORCH, 1995  Interplay between chromatin structure and transcription. Curr. Opin. Cell. Biol. 7:371-375[Medline].

KROKAN, H. E., R. STANDAL, and G. SLUPPHAUG, 1997  DNA glycosylases in the base excision repair of DNA. Biochem. J. 325:1-16.

KUNZ, B. A., K. RAMACHANDRAN, and E. J. VONARX, 1998  DNA sequence analysis of spontaneous mutagenesis in Saccharomyces cerevisiae.. Genetics 148:1491-1505[Abstract/Free Full Text].

LAURENSON, P. and J. RINE, 1992  Silencers, silencing, and heritable transcriptional states. Microbiol. Rev. 56:543-560[Abstract/Free Full Text].

LEA, D. E. and C. A. COULSON, 1948  The distribution of the numbers of mutants in bacterial populations. J. Genet. 49:264-284.

LICHTEN, M., R. H. BORTS, and J. E. HABER, 1987  Meiotic gene conversion and crossing over between dispersed homologous sequences occurs frequently in Saccharomyces cerevisiae.. Genetics 115:233-246[Abstract/Free Full Text].

LINDAHL, T., 1993  Instability and decay of the primary structure of DNA. Nature 362:709-715[Medline].

LIPPERT, M. J., Q. CHEN, and H. L. LIBER, 1998  Increased transcription decreases the spontaneous mutation rate at the thymidine kinase locus in human cells. Mutat. Res. 401:1-10[Medline].

LJUNGMAN, M. and P. C. HANAWALT, 1992  Efficient protection against oxidative DNA damage in chromatin. Mol. Carcinogen. 5:264-269[Medline].

LOHR, D., P. VENKOV, and J. ZLATANOVA, 1995  Transcriptional regulation in the yeast GAL gene family: a complex genetic network. FASEB J. 9:777-787[Abstract].

NELSON, J. R., C. W. LAWRENCE, and D. C. HINKLE, 1996  Thymine-thymine dimer bypass by yeast DNA polymerase . Science 272:1646-1649[Abstract].

NEVO-CASPI, Y. and M. KUPIEC, 1994  Transcriptional induction of Ty recombination in yeast. Proc. Natl. Acad. Sci. USA 91:12711-12715[Abstract/Free Full Text].

PAETKAU, D. W., J. A. RIESE, W. S. MACMORRAN, R. A. WOODS, and R. D. GIETZ, 1994  Interaction of the yeast Rad7 and Sir3 proteins: implications for DNA repair and chromatin structure. Genes Dev. 8:2035-2045[Abstract/Free Full Text].

PLATT, T., 1984  Toxicity of 2-deoxygalactose to Saccharomyces cerevisiae cells constitutively synthesizing galactose-metabolizing enzymes. Mol. Cell. Biol. 4:994-996[Abstract/Free Full Text].

POPOFF, S. C., A. I. SPIRA, A. W. JOHNSON, and B. DEMPLE, 1990  Yeast structural gene (APN1) for the major apurinic endonuclease: homology to Escherichia coli endonuclease IV. Proc. Natl. Acad. Sci. USA 87:4193-4197[Abstract/Free Full Text].

QUIMBY, B. B., A. ALANO, S. ALMASHANU, A. M. DESANDRO, and T. M. COWAN et al., 1997  Characterization of two mutations associated with epimerase-deficiency galactosemia, by use of a yeast expression system for human UDP-galactose-4-epimerase. Am. J. Hum. Genet. 61:590-598[Medline].

RAMOTAR, D., S. C. POPOFF, E. B. GRALLA, and B. DEMPLE, 1991  Cellular role of yeast Apn1 apurinic endonuclease/3'-diesterase: repair of oxidative and alkylation DNA damage and control of spontaneous mutation. Mol. Cell. Biol. 11:4537-4544[Abstract/Free Full Text].

REED, S. H., S. BOITEUX, and R. WATERS, 1996  UV-induced endonuclease III-sensitive sites at the mating type loci in Saccharomyces cerevisiae are repaired by nucleotide excision repair: RAD7 and RAD16 are not required for their removal from HML.. Mol. Gen. Genet. 250:505-514[Medline].

SANDER, M. and D. RAMOTAR, 1997  Partial purification of Pde1 from Saccharomyces cerevisiae: enzymatic redundancy for the repair of 3'-terminal DNA lesions and abasic sites in yeast. Biochemistry 36:6100-6106[Medline].

SAPARBAEV, M., L. PRAKASH, and S. PRAKASH, 1996  Requirement of mismatch repair genes MSH2 and MSH3 in the RAD1-RAD10 pathway of mitotic recombination in Saccharomyces cerevisiae.. Genetics