Cloning a eukaryotic DNA glycosylase repair gene by the ...

4 downloads 5216 Views 1MB Size Report
Harvard School of Public Health, Charles A. Dana Laboratory of Toxicology, 665 Huntington Avenue, Boston, MA 02115. Communicated by ..... DNA glycosylase activity in cell-free extracts. 3-MeA (A) .... Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular. Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold.
Proc. Nati. Acad. Sci. USA Vol. 86, pp. 7961-7965, October 1989

Genetics

Cloning a eukaryotic DNA glycosylase repair gene by the suppression of a DNA repair defect in Escherichia coli (alkA tag/yeast DNA repair mutant/alkylation toxicity)

JIN CHEN, BRUCE DERFLER, AZMAT MASKATI, AND LEONA SAMSON* Harvard School of Public Health, Charles A. Dana Laboratory of Toxicology, 665 Huntington Avenue, Boston, MA 02115

Communicated by Bernard D. Davis, July 10, 1989

ABSTRACT If eukaryotic genes could protect bacteria with defects in DNA repair, this effect could be exploited for the isolation of eukaryotic DNA repair genes. We have thus cloned a DNA repair gene from Saccharomyces cerevisiae that directs the synthesis of a DNA glycosylase that specifically releases 3-methyladenine from alkylated DNA and in so doing protects alkylation-sensitive Escherichia coli from killing by methylating agents. The cloned yeast gene was then used to generate a mutant strain of S. cerevisiae that carries a defect in the glycosylase gene and is extremely sensitive to DNA methylation. This approach may allow the isolation of a large number of eukaryotic DNA repair genes.

MATERIALS AND METHODS Cells, Plasnids, and Enzymes. E. coli AB1157 (F- thr-1 leu-6 proA2 thi-1 argE lacYI galK ara-14 xyl-S mtl-i tsx-33 strA sup-37) was the strain used as the wild type for alkylation sensitivity. MV1932 and MV1902 (gifts from M. Volkert, University of Massachusetts, Worcester, MA) were alkylation-sensitive derivatives of AB1157 and are alkAl tag and alkAlO5::ApSG1 (13), respectively. DH5a, an rk- mk + E. coli strain, was purchased from BRL. S. cerevisiae strains were DBY745 (adel-100 ura3-52 leu2-3 leu2-112) and DBY747 (his3-AJ, leu2-3, leu2-112, trpl-289, ura3-52), gifts from Eric Eisenstadt (Office of Naval Research, Arlington, VA) and Louise Prakash (University of Rochester, Rochester, NY), respectively. The YEp13-borne S. cere-visiae genomic library (14) was purchased from American Type Culture Collection, and pUC19 was purchased from BRL. Plasmid DNAs were purified by alkaline lysis followed by banding in cesium chloride/ethidium bromide density gradients (15). Restriction endonucleases and T4 DNA ligase were from New England Biolabs. Screen for Alkylation-Resistant Transformants. The YEp13A yeast genomic library from American Type Culture Collection was in E. coli RR1 (rk- mk-); plasmid purified from this strain survived poorly when transfected into MV1932 (rk+ mk+); we therefore passed the library through the DH5a (rk- mk+) strain prior to transfection into E. coli MV1932. Ampicillin-resistant MV1932 transformants were individually toothpicked onto two Luria-Bertani (LB) ampicillin plates (15), one of which contained 0.01% MMS. (MMS was added to molten agar at about 480C.) The plates, each containing 50 toothpicked colonies, were inspected for MMSresistant transformants 12-18 hr later. We chose to screen rather than select for MMS-resistant transformants to avoid MMS-induced mutation of the clones. Survival Curves. Bacteria were grown in LB medium to 108 cells per ml, and yeast were grown in yeast extract/peptone/ dextrose (YPD) medium to 107 cells per ml. MMS, dimethyl sulfate (DMS), and N-methyl-N-nitrosourea (MNU) were added to the appropriate concentrations, and aliquots were removed from the cultures at the indicated times, diluted, and spread on LB or YPD plates to estimate cell survival. MNU at 4 mg/ml was dissolved fresh each time in 50% ethanol. MMS and DMS were stored at room temperature. Southern Blot Analysis. Bacterial, human, and yeast genomic DNAs were isolated by the procedures of Silhavy et al. (16), Maniatis et al. (15), and Winston et al. (17), respectively. Genomic DNA (2.5 ug of yeast, 5.0 pg of bacterial, and 10 Mg of human) was digested with EcoRI or HindIII; the products were electrophoresed in a 0.7% agarose gel, transferred to nitrocellulose filters, and probed with the 4-kilobase

The molecular genetic characterization of DNA repair functions in Escherichia coli has produced the view that mutagenesis by alkylating agents is avoided by the removal of lesions that mispair when replicated, and cell death is avoided by the removal of lesions that block DNA replication (1). Two DNA glycosylases, encoded by the alkA and tag genes, remove alkylated bases that block replication; 3-methyladenine (3-MeA) is repaired by both AlkA and Tag, and 3methylguanine (3-MeG), 02-methylthymine, and 02-methylcytosine are repaired by AlkA (2-4). After the removal of a damaged base by a DNA glycosylase, one of several apurinic/apyrimidinic endonucleases hydrolyzes the phosphodiester bond adjacent to the abasic deoxyribose, allowing DNA polymerase I, DNA ligase, and other enzymes to complete the repair process (5). E. coli strains deficient in the Tag and AlkA DNA glycosylases are extremely sensitive to killing by methylating agents such as methyl methanesulfonate (MMS) (1). The expression of certain prokaryotic DNA repair functions in mammalian cells can suppress the toxic effects of certain types of DNA damage (6-9). It thus seemed to us that the converse might be true, that eukaryotic DNA repair functions might be able to protect bacteria and that this might be a way of identifying eukaryotic DNA repair genes. Since DNA glycosylases can function in the absence of other proteins, we expected that a eukaryotic DNA glycosylase would be able to initiate the repair of damaged DNA in E. coli. Saccharomyces cerevisiae has been shown to contain 3-MeA DNA glycosylase activity (10), but none of the large number of alkylation-sensitive S. cerevisiae mutants has been assigned a deficiency in this or any other alkylation repair function (10-12). Here we describe the isolation of an S. cerevisiae gene that encodes a 3-MeA DNA glycosylase and can protect E. coli alkA tag mutants from alkylationinduced cell death. The cloned gene has been used to generate an alkylation-sensitive repair-deficient S. cerevisiae mutant.

Abbreviations: 3-MeA, 3-methyladenine; 3-MeG, 3-methylguanine; MMS, methyl methanesulfonate; MNU, N-methyl-N-nitrosourea; DMS, dimethyl sulfate. *To whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 7961

7962

Proc. Natl. Acad. Sci. USA 86 (1989)

Genetics: Chen et al.

(kb) Sph I fragment isolated from pUC-4.0 (see Fig. 3). The 4-kb fragment was 32P-labeled, and the final filter wash was at high stringency (30 mM NaCl/3 mM sodium citrate at 580C for 2 hr). Northern Blot Analysis. Yeast RNA isolation was modified from the procedure of Carlson and Botstein (18). Briefly, 108 midlogarithmic cells were washed in water and resuspended in 0.2 ml of 0.5 M NaCl/0.2 M Tris (pH 7.6)/0.01 M EDTA/1% SDS. Four-tenths gram of washed glass beads (0.4-0.5 mm) and 0.2 ml of phenol/chloroform/isoamyl alcohol, 25:24:1 (vol/vol), were added, and the mixture was vortexed vigorously for 2.5 min. After the further addition of 0.3 ml of buffer and 0.3 ml of phenol/chloroform/isoamyl alcohol, the aqueous phase was isolated, reextracted, and ethanol precipitated. Five micrograms of total RNA was electrophoresed in a formaldehyde/1% agarose gel (15), transferred to a nylon membrane, and probed with the 32P-labeled 4-kb Sph I or 2.1-kb EcoRI yeast genomic fragments (see Fig. 3). The blots were washed in 300 mM NaCl/30 mM sodium citrate at 650C for 1 hr and in 15 mM NaCl/1.5 mM sodium citrate at 250C for 1 hr. DNA Glycosylase Activity. Bacterial and yeast cell extracts were prepared as described (10, 19) in the buffer described by Nisson and Lawrence (10)-namely, 10 mM Tris'HCl (pH 7.4)/10 mM NaCl/5 mM MgCl2/0.5 mM CaCl2/0.2% Nonidet P-40/1 mM dithiothreitol. Various amounts of extract proteins were incubated for various periods with 37.5 ,ug of [3H]DMS-treated calf thymus DNA (1000 cpm/,g), prepared according to Samson and Linn (20). Reaction volumes ranged from 100 to 150 ,ul. The reaction was stopped by adding 0.1 volume of 2 M NaCl, 3 volumes of ice-cold ethanol, and incubating at -20°C for 15 min. The precipitated protein and nucleic acids were removed by centrifugation, and the ethanol supernatant was dried and resuspended in 20 ,ul of 0.1 M HCl. After the addition of alkylated base standards (7-MeG, 3-MeA, and 06-MeG), the mixture was spotted onto Whatman 3MM paper (2 X 40 cm) and separated by descending chromatography in 2-propanol/ammonium hydroxide/water, 7:1:2 (vol/vol), for 16-18 hr. After the standards were visualized under UV light, the dried chromatogram was cut into 1-cm strips; each strip was eluted into 1 ml of water (in a scintillation vial) for at least 2 hr and assayed for tritium after the addition of 10 ml of Scintiverse (Fisher). Labeled DNA substrate was incubated without extract for the appropriate periods, and the liberated 3-MeA and 7-MeG was subtracted from that released in the extract.

RESULTS Isolation of a Yeast Gene That Suppresses the Alkylation Sensitivity of E. coli aIkA tag Mutants. The extreme alkylation sensitivity of E. coli alkA tag mutants (1, 21) is shown in Fig. 1A and is reflected in the fact that nutrient agar plates containing 0.01% MMS support the growth of wild-type E. coli but not alkA tag mutants (data not shown). Our initial aim was to look for a yeast DNA fragment that would confer on E. coli alkA tag the ability to grow on 0.01% MMS plates. A YEp13 yeast genomic library (14) was introduced into E. coli alkA tag. Fifteen thousand transformants were individually streaked onto MMS plates and control plates; two transformants formed colonies on MMS plates, and the responsible plasmids (YEp13A and YEp13C) were isolated from the bacteria on the control plates. Judging from their restriction pattern, each plasmid proved to contain the same 6-kb insert, which hybridized to yeast DNA but not to bacterial or human DNA (Fig. 2A). Reintroduction of the purified YEp13A/C plasmid into alkA tag mutants protects against killing by MMS, DMS, and MNU (Fig. 1) but not against UV light or the UV-mimetic agent 4-nitroquinoline oxide (data not shown).

0.01D W 100 10

0.1

L

0.01 0

30 60

0

30 60

MINUTES FIG. 1. Alkylation-induced bacterial cell killing. The colonyforming ability of E. coli strains AB1157 (o), MV1932 (an alkA tag AB1157 derivative) (e), MV1932/YEp13A (A), MV1932/YEp13C (v), MV1902 (an aIkA::ApSG1 AB1157 derivative, see ref. 13) (*), MV1902/YEp13A (a), and MV1902/YEp13C (-) was measured after treatment with MMS at 0.05% (A), DMS at 0.005% (B), MNU at 200 pug/ml (C), and MMS at 0.034% (D) for the indicated times.

Our initial assumption was that the YEp13A/C gene product was functioning as a replacement for the AlkA and Tag DNA glycosylases. However, we had to exclude, formally, 1

2

3 4

23 9465 43-

5 6

7-5-l t

3

2

1

A

2

t -

e".

Im

1. ?8.: s

2.32.0 -

-

FIG. 2. Southern blot analysis of bacterial, yeast, and human DNA and Northern blot analysis of RNA isolated from S. cerevisiae. (A) Southern blot analysis. Five micrograms of bacterial DNA (lanes 1 and 2), 2.5 gg of yeast DNA (lanes 3 and 4), and 10 ,.g of human DNA (lanes 5 and 6) were digested with EcoRI (lanes 1, 3, and 5) or

HindIII (lanes 2, 4, and 6), and the resulting fragments were separated in a 0.7% agarose gel, transferred to nitrocellulose, and hybridized with the 32P-labeled 4-kb Sph I fragment from pUC-4.0 (see Fig. 3). (B) Northern blot analysis. In each lane, 5 Ug of total

RNA isolated from S. cerevisiae DBY745 was separated in a formaldehyde/1% agarose gel and transferred to a nylon membrane. Lane 1 was hybridized with the 32P-labeled 4.0-kb Sph I fragment from pUC-4.0 (see Fig. 3); lane 2 was hybridized with the 32P-labeled 2.1-kb EcoRI fragment isolated from pUC-2.1 (see Fig. 3); lane 3 was hybridized with a 1.6-kb yeast actin cDNA fragment. The labeled probes were of approximately the same specific activity.

Genetics: Chen et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

the possibility that YEp13A/C might encode a tRNA that suppresses the alkA or tag mutation or a protein that somehow prevents DNA alkylation. YEp13A/C is not functioning as a nonsense suppressor because it also conferred alkylation resistance to a nonsuppressible E. coli strain bearing a A insertion in the alkA gene (Fig. 1D) (13). Furthermore, it is not acting as a preventer of DNA alkylation because [3H]DMS delivered roughly the same number of alkyl groups to the DNA of E. coli alkA tag irrespective of the presence or absence of YEp13A/C; [3H]DMS (4.17 Ci/mmol; 1 Ci = 37 GBq) at 1.5 pug/ml for 5 min produced 47, 68, 59, and 70 cpm/,ug of DNA in, respectively, wild-type E. coli, alkA tag, and alkA tag carrying YEp13A or YEp13C. It seemed likely, therefore, that the YEp13A/C gene product is involved in repair. The following experiments showed that the cloned yeast genomic fragment encodes a 3-MeA DNA glycosylase. The Cloned Yeast Gene Encodes a 3-MeA DNA Glycosylase. We localized the active part of the clone to a 2.1-kb region of the YEp13A/C 6-kb insert (Fig. 3). A 4-kb Sph I fragment was subcloned from YEp13A into pUC19 to generate pUC-4.0; E. coli alkA tag carrying pUC-4.0 were resistant to MMS. From the 4-kb Sph I fragment we subcloned, into pUC19, two nonoverlapping EcoRI fragments of 1.5 and 2.1 kb, to generate pUC-1.5 and pUC-2.1, respectively; pUC-1.5 did not provide MMS resistance in E. coli alkA tag cells, but pUC-2.1 gave the same level of resistance as the original YEp13A recombinant plasmid. Northern blot analysis of RNA isolated from S. cerevisiae DBY745 showed that the 4-kb Sph I fragment and the 2.1-kb EcoRI fragment each hybridized to Sph I

Sph I

YEpI13A

4.0 EcoRI EcoRII

R

SphIfL

9Z SphI EcoRI

4.0 UC -

MMS resistant

a single transcript of similar size (namely, 1.8 kb; Fig. 2B); therefore, the 2.1-kb yeast genomic fragment probably directs the synthesis of a single mRNA. Compared to the level of actin mRNA, the 1.8-kb transcript appeared to be expressed at extremely low levels (Fig. 28). The level of the 1.8-kb mRNA was much higher in S. cerevisiae DBY747 cells harboring the YEp13A plasmid (data not shown). Further experiments showed that pUC-2.1 encodes a 3MeA DNA glycosylase. Fig. 4A shows that alkA tag! pUC-2.1 cells contained as much 3-MeA DNA glycosylase activity as wild-type E. coli and about 10 times more 3-MeA DNA glycosylase activity than alkA tag/pUC-1.5 cells. (Glycosylase levels were unaffected by the presence of pUC-1.5.) The base excised by alkA tag/pUC-2.1 extracts was identified as 3-MeA by using three chromatographic procedures (data not shown). All three strains showed the same level of a 7-MeG DNA glycosylase activity (Fig. 4B), and none showed any 3-MeG DNA glycosylase activity (data not

shown). The cloned yeast gene also directed the synthesis of a 3-MeA DNA glycosylase in yeast cells because the introduction of YEp13A into S. cerevisiae DBY745 caused a

7-fold increase in 3-MeA DNA glycosylase levels (Fig. 4C); interestingly, extra glycosylase activity in wild-type DBY745 did not provide any additional resistance to MMS (data not shown). Since the major purine substrate for this yeast DNA glycosylase appears to be 3-MeA, we propose that its gene be called the yeast MAG gene (3-MeA Glycosylase). However, it remains to be determined whether the Mag DNA glycosylase also removes alkylated pyrimidines. Generation of a Yeast DNA Alkylation Repair Mutant. The cloned MAG gene of S. cerevisiae DBY747 was disrupted in vitro by the insertion of the URA3 gene. A 0.77-kb Bgl II fragment from the middle of the cloned 2.1-kb EcoRI fragment (carrying the MAG gene) was replaced with a 3.8-kb Bgl II/BamHI fragment carrying a functional URA3 gene (22). This construction generated pUC-5.1 (Fig. SA) in which the URA3 gene is flanked on each side by about 0.65 kb of DNA from the cloned 2.1-kb fragment. We transformed DBY747 with the 5.1-kb EcoRI fragment isolated from pUC-5.1 (bearing the disrupted MAG gene) and selected URA+ transformants. Southern analysis of one transformant confirmed that the MAG gene had been disrupted; the 0.77-kb fragment from the middle of the MAG gene hybridized to a 2.5-kb EcoRI fragment from the DBY747 genome but did not hybridize to any fragment from the genome of the URA+ transformant (Fig. 5B, lanes 1 and 2). Hybridization with the MAG gene plus flanking sequences (isolated from pUC-4.0) demonstrated a disruption of this 2.5-kb EcoRI fragment in the URA+ transformant (Fig. SB, lanes 3 and 4). However, the disruption did not generate a new 5.5-kb fragment that would 1600 -A

EcoRI

oRI

7963

B

6C

EcoRIcoRI 1200

4C

pUC-1.5

pUC-2.1

& 8002z

400

MMS sensitive

MMS resistant

FIG. 3. Subcloning of a 2.1-kb yeast genomic DNA fragment that

confers alkylation resistance. A 4.02-kb Sph I fragment was subcloned from YEp13A into pUC19 to yield pUC-4.0, which provided MV1932 with MMS resistance. Two nonoverlapping EcoRI fragments (1.5 and 2.1 kb) isolated from pUC-4.0 were subcloned into pUC19 to generate pUC-1.5 and pUC-2.1. pUC-1.5 did not provide MV1932 with MMS resistance; pUC-2.1 provided MV1932 with the same level of resistance as did YEp13A. MMS cell killing curves were as in Fig. 1.

0 0 30 60

120 30 60 MINUTES

120

25 100

300 PROTEIN 4ig)

FIG. 4. DNA glycosylase activity in cell-free extracts. 3-MeA (A) and 7-MeG (B) released during incubation at 25°C by 25 Ig of bacterial cell extract proteins from AB1157 (o), MV1932 alkA tag/pUC-2.1 (A), and MV1932 alkA tag/pUC-1.5 (A). (C) 3-MeA released in 60 min at 25°C by various amounts of cell extract protein from S. cerevisiae DBY745 (e) and S. cerevisiae DBY745/YEp13A (o).

7964

Proc. Natl. Acad. Sci. USA 86 (1989)

Genetics: Chen et al.

A

25C00 rl^

A

20 5< 4±N~j

icx 15

00

MMS resistant

o

L

00

0.1

0.5

1.0

PROTEIN (mg) iHI)

MMS sensitive

B 1

2

3

4

23 9.4 6.7 4.3 2.32.0

-

FIG. 5. Disruption of the yeast MAG gene. (A) A 0.77-kb Bgl II fragment from the middle of the pUC-2.1 insert was replaced with a 3.8-kb BgI II/BamHI fragment carrying the URA3 gene (23) to generate pUC-5.1. (B) Southern blot analysis. DNA (2-3 pg) from DBY747 (lanes 1 and 3) and its URA' transformant (lanes 2 and 4) were digested with EcoRI and probed with the 32P-labeled 0.77-kb Bgl II fragment from pUC-2.1 (lanes 1 and 2) or the 4-kb Sph I fragment from pUC-4.0 (lanes 3 and 4).

be expected from a simple gene transplacement; instead a more complex rearrangement must have occurred to generate a new band of over 23 kb (Fig. SB, lane 4). Thus, the MAG gene was not disrupted by a simple gene transplacement but rather by a more complex rearrangement. Cell extracts from the URA+ transformant contained undetectable 3-MeA DNA glycosylase activity (Fig. 6A), confirming the mag- genotype; these cells became extremely sensitive to killing by MMS (Fig. 6B). The introduction of YEp13A into the URA+ mag- strain produced a glycosylase activity of about 10 times the level in wild-type DBY747 cells (Fig. 6A); however, despite the overproduction of the Mag DNA glycosylase, MMS resistance was only restored to the wild-type level (Fig. 6B).

DISCUSSION The isolation of yeast genes and mammalian cDNAs by the functional complementation of E. coli mutants is not unprecedented. However, all the genes that have been cloned this way were involved in intermediary metabolism and could be identified because they suppressed nutritional auxotrophies.

15 30

60

MINUTES

FIG. 6. Phenotypic analysis of DBY747 bearing a disrupted MAG (A) 3-MeA released in 2 hr at 25TC by various amounts of cell extract proteins from wild-type DBY747 (e), the URA+ transformant of DBY747 transformed with the 5.1-kb EcoRI fragment isolated from pUC-5.1 (o), and this URA' transformant carrying YEp13A (A). (B) MMS (0.32%)-induced cell killing of the same yeast strains as in A. gene.

For example, the yeast imidazole glycerol phosphate hydrolase gene (his3) (23), the murine adenosine deaminase cDNA (24), and the human purine-nucleoside phosphorylase cDNA (25) were cloned by the suppression of the appropriate auxotrophic mutants of E. coli. We have now shown that it is feasible to clone eukaryotic DNA repair genes by virtue of their ability to rescue repair-deficient mutants of E. coli from the effects of DNA damage and have used this procedure to isolate an S. cerevisiae 3-MeA DNA glycosylase gene. Genes whose products protect yeast against alkylation damage have been indicated by the rad (radiation sensitive) and cdc (cell division cycle) mutants, and several mutants have been directly isolated as alkylation sensitive (10-12). Of 37 alkylation-sensitive yeast mutants, 13 have been mapped, and, of these, 8 have been cloned and sequenced (11, 26). Partial sequencing of 1250 base pairs of the MAG gene (J.C. and B.D., unpublished data) has not revealed any significant homology with these 8 genes or with any other genes whose sequences have been entered in the GenBank data base (release no. 60, 1989). We find that, as in E. coli, the repair of 3-MeA DNA lesions contributes to alkylation resistance in S. cerevisiae, since a deficiency in the Mag 3-MeA DNA glycosylase results in alkylation sensitivity. Furthermore, it appears that wild-type yeast cells express saturating levels of the Mag enzyme because the provision of up to 10 times the wild-type level of Mag does not provide any additional resistance to alkylating agents. Similarly, the overexpression of the yeast Mag enzyme does not provide wild-type E. coli with extra resistance to DNA alkylation (data not shown). In summary, we have shown that it is feasible to clone eukaryotic DNA repair genes by virtue of their ability to rescue repair-deficient mutants of E. coli from the effects of DNA damage. This approach may perhaps be limited to the isolation of genes whose products are not required to complex with other proteins, but this limitation may not be absolute because it appears that organisms spanning enormous evolutionary distances employ similar DNA repair proteins to protect their genomes from DNA damage. We believe that this general approach may allow the isolation of a large number of eukaryotic DNA repair genes. We thank John Cairns and Bruce Demple for critical discussions. This awork was supported by American Cancer Society Research Grant NP448 and National Institute of Environmental Health Science Grant 1-PO-ES03926. L.S. was supported by an American Cancer Society Faculty Research Award. J.C. was supported by a Dana Foundation Training Program for Graduate Study in Toxicology.

Genetics: Chen et al. 1. Lindahl, T., Sedgwick, B., Sekiguchi, M. & Nakabeppu, Y. (1988) Annu. Rev. Biochem. 57, 133-158. 2. Karran, P., Lindahl, T., Ofsteng, I., Evensen, G. & Seeberg, E. (1980) J. Mol. Biol. 140, 101-127. 3. Evensen, G. & Seeberg, E. (1982) Nature (London) 296, 773-775. 4. Karran, P., Hjelmgrem, T. & Lindahl, T. (1982) Nature (London).296, 770-773. 5. Friedberg, E. C. (1985) in DNA Repair (Freeman, New York), pp. 141-211. 6. Valerie, K., de Reil, J. K. & Henderson, E. E. (1985) Proc. Nati. Acad. Sci. USA 82, 7656-7660. 7. Samson, L., Derfler, B. & Waldstein, E. (1986) Proc. Nati. Acad. Sci. USA 83, 5607-5610. 8. Brennand, J. & Margison, G. P. (1986) Proc. Nati. Acad. Sci. USA 83, 6292-6296. 9. Kataoka, H., Hall, J. & Karran, P. (1986) EMBO J. 5, 31953200. 10. Nisson, P. E. & Lawrence, C. W. (1986) Mutat. Res. 165, 129-137. 11. Friedberg, E. C. (1988) Microbiol. Rev. 52, 70-102. 12. Prakash, L. & Prakash, S. (1977) Genetics 86, 33-35. 13. Volkert, M. R., Nguyen, D. C. & Beard, K. C. (1986) Genetics 112, 11-26. 14. Nasmyth, K. A. & Reed, S. I. (1980) Proc. NatI. Acad. Sci. USA 77, 2119-2123.

Proc. Natl. Acad. Sci. USA 86 (1989)

7965

15. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 16. Silhavy, T. J., Berman, M. L. & Enquist, L. W. (1984) Experiments with Gene Fusions (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 17. Winston, F., Chumley, F. & Fink, G. R. (1983) Methods Enzymol. 101, 211-227. 18. Carlson, M. & Botstein, D. (1982) Cell 28, 145-154. 19. Rebeck, G. W., Coons, S., Carroll, P. & Samson, L. (1988) Proc. Natl. Acad. Sci. USA 85, 3039-3043. 20. Samson, L. & Linn, S. (1987) Carcinogenesis 8, 227-230. 21. Clarke, N. D., Kvaal, M. & Seeberg, E. (1984) Mol. Gen. Genet. 197, 368-372. 22. Alani, E., Cao, L. & Kleckner, N. (1987) Genetics 116, 541-545. 23. Struhl, K. & Davis, R. W. (1977) Proc. Natl. Acad. Sci. USA 74, 5255-5259. 24. Yeung, C.-Y., Ingolia, D. E., Roth, D. B., Shoemaker, C., Al-Ubaidi, M. R., Yen, J.-Y., Ching, C., Bobonis, C., Kaukman, R. J. & Kellems, R. E. (1985) J. Biol. Chem. 260, 1029910307. 25. Goddard, J. M., Caput, D., Williams, S. R. & Martin, D. W. (1983) Proc. Natl. Acad. Sci. USA 80, 4281-4285. 26. Jones, J. S., Weber, S. & Prakash, L. (1988) Nucleic Acids Res. 16, 7119-7131.