YEAST

VOL. 8: 385-395

(1992)

Identification of RAD16, a Yeast Excision Repair Gene Homologous to the Recombinational Repair Gene RA0.54 and to the SNF2 Gene Involved in Transcriptional Activation DAVID SCHILD*t, BRIAN J. GLASSNERS, ROBERT K. MORTIMER*$, MARIAN CARLSON$ AND BREHON C. LAURENTP *Division of Cell and Molecular Biology, Lawrence Berkeley Laboratory, Berkeley, CA 94720, U.S.A. $Department of Molecular and Cell Biology, and Graduate Group in Biophysics, University of California, Berkeley, CA 94720, U.S.A. $Department of Genetics and Development, and Institute of Cancer Research, Columbia University College of Physicians and Surgeons, New York, N Y 10032, U.S.A. Received 20 January 1992; accepted 28 January 1992

The RAD54 gene of Saccharomyces cerevisiae is involved in the recombinational repair of DNA damage. The predicted amino acid sequence of the RAD54 protein shows significant homologies with the yeast SNF2 protein, which is required for the transcriptional activation of a number of diversely regulated genes. These proteins are 3 1% identical in a 492-amino acid region that includes presumed nucleotide and Mg2+binding sites. We noted previously that the SNF2 protein also shares homology with a partial open reading frame (ORF) that was reported with the sequence of an adjacent gene. This ORF also shares homology with the RAD54 protein. To test whether this ORF is involved in transcriptional activation or DNA repair, yeast strains deleted for part of it have been isolated. These strains do not show a Snf-like phenotype, but they are UV sensitive. This gene has been identified as RADI6, a gene involved in the excision repair of DNA damage. Analysis of the radl6 deletion mutations indicates that RADI6 encodes a nonessential function and is not absolutely required for excision repair. Outside the region of homology to RAD54 and SNFZ, the predicted RAD16 protein contains a novel cysteine-rich motif that may bind zinc and that has been found recently in eleven other proteins, including the yeast RAD18 protein. The homologies between RAD16, RAD54 and SNFZ are also shared by several additional, recently isolated yeast and Drosophila genes. KEY WORDS -DNA

repair genes; transcriptional activation; sequence homology; zinc fingers; potential helicases.

INTRODUCTION Genetic analysis of DNA repair genes in the yeast Saccharomyces cerevisiae has elucidated three major repair pathways for damage caused by UV and x-rays (for reviews see Friedberg, 1988; Game, 1983; Haynes and Kunz, 1981). Mutations in the RAD1,2.3.4,7,10,14and 16 genes are defective in the excision repair pathway, resulting in UV sensitivity. Mutations in the RAD5O through RAD57 genes are blocked in the recombinational repair pathway, which is necessary for the repair of x-ray-induced double-strand breaks. Mutations in RAD6 and RAD18, and a number of other genes, confer sensitivity to both UV and x-rays, and these genes are referred to as the post-replicative or error-prone tcorresponding author.

0749-503X/92/050385-11$05.50 0 1992 by John Wiley & Sons Ltd

repair group. _ _ _ Many of the genes in each of the three repair groups have been- cloned and sequenced (reviewed in Friedberg, 1988). Enzymatic activities for only two of these gene products have been reported: RAD6 encodes a ubiquitin-conjugating enzyme (Jentsch et al., 1987) and RAD3 encodes an ATPase-dependent DNA helicase (Harosh et al., 1989; Sung et al., 1987). There are two examples in which genes in the same yeast DNA repair group share regions of homology: RADI and RAD7 share tandemly repeated leucine-richmotifs, thought to be important as protein-protein interacting domains (Schneider and Schweiger, 1991), and RAD57 and RAD5l share a region of homology which includes a potential nucleotide binding site (Kans and Mortimer, 1991). We report here that the recombinational repair gene RAD54 shares homologies with

386 both the SNF2 gene, involved in transcriptional activation, and a partially sequenced, but generally uncharacterized gene (Fleig et al., 1986; Schultz and Carlson, 1987),which had previously been shown to be homolgous to SNF2 (Laurent et al., 1991). We have identified this third gene as the excision repair gene RAD16. The yeast SNF2 @ucrose Eon-fermenting) gene was originally identified by mutations which block the transcriptional derepression of the SUC2 gene encoding invertase (Neigeborn and Carlson, 1984). SNF2 is the same gene as SWI2 (K. Nasmyth, pers. comm.). Mutations in SWI2 block induction of the HO gene, responsible for homothallic mating-type switching. Mutations in the SNF2(S WI2) gene also block the expression of the yeast transposon Tyl, the IN01 gene encoding inositol synthetase, and several other diversely regulated yeast genes (Abrams et al., 1986;Happel et al., 1991;Stern et al., 1984). Studiesusing a promoter-bound LexA-SNF2 fusion protein have implicated SNF2 directly in activation of transcription in vivo (Laurent et al., 1991). RADI6 is a particularly interesting yeast DNA repair gene because it is one of only three yeast genes which are involved primarily in repairing nontranscribed regions. The radl6-2 mutation causes a much stronger defect in the repair of pyrimidine dimers in a transcriptionally inactive region (HMLa), as compared to MATa, a transcribed region; this phenotype is shared by the rad9 and rad24 mutations (Terleth et al., 1990). The human xeroderma pigmentosum complementation group C (XP-c) cells are likewise defective in the repair of non-transcribed regions of the genome (Kantor et al., 1990; Venema et al., 1991). The homologies observed between proteins involved in two different DNA repair pathways and in transcriptional activation suggest that these proteins share an enzymatic activity or structural function. The possibility that these genes encode helicases or proteins that alter chromatin structure is discussed. MATERIALS AND METHODS Yeast strains andgenetic methods Strains of S. cerevisiae used in these studies are listed in Table 1. All MCY strains are derivatives of S288C. A gel-isolated 2.3 kb PvuII fragment of pBLY22 was used to disrupt the chromosomal locus of diploid strain MCY1751 by the one-step gene replacement method (Rothstein, 1991), and

D. SCHILD ETAL.

a Ura' transformant heterozygous for the mutation, radl6-Al:: URA3, was identified by Southern blot hybridization analysis. MCY2048 is a radl6AI:: URA3 haploid segregant from MCY 1751. DST32 was derived directly fronm Y " 2 8 2 by gene replacement, and DST3 1 was derived similarly from YNN281. Strains LP3031-3A, X12-6B, X1410A, YNN281, YNN282 and DA2100 are from the Yeast Genetic Stock Center (Berkeley, CA). DA2100, a lys2-AI strain, was derived from SEY2102 (Barnes and Thorner, 1986). Since, for an unknown reason, DA2100 is rho-, we isolated a rho- derivative of SEY2102 (a gift of Jeremy Thorner) so that these strains would be as isogenic as possible. Plasmids pBLTY3 contains the 4.3-kb HindIII-BamHI fragment of pP2L4 (a gift of Carl Falco) in the URA3- and CEN6-containing plasmid pRS3 16 (Sikorski and Hieter, 1989). pBLTY4 contains the BamHI-SalI (polylinker site adjacent to HindIII) fragment of pBLTY3 cloned in YEp24. pBLY22 was constructed from pBLTY3 by replacing the 0.9 kb HpaI-BglII fragment with a 1.1 kb SmaI fragment containing the yeast URA3 gene. DNA sequencingand computer analysis The sequence of 1-1kb of the cloned RADI6 gene was determined for both strands by the method of Sanger et al. (1977) using Sequenase (U.S. Biochemical)and the universal primer (Amersham). Amino acid sequence homologies were determined using FASTA and LFASTA (Pearson, 1990), and the final alignment was obtained by visual inspection. The PAM250 matrix was used to determine conserved amino acids. The amino acid sequence reported in this paper is derived from the DNA sequence which has been deposited in the GenBank data base (accession no. M83553).

UV survival curves Cells were grown at 30°C to mid-log phase (OD,, of -0.4-0.8) in liquid YEPD media, except for strains containing replicating plasmids, which were grown in complete media lacking uracil, to maintain selection for the plasmids. Cells were sonicated to disperse clumps of cells, diluted in water and plated in triplicate, prior to irradiation with UV. Following irradiation, plates were incubated at 30°C in the dark for 4-6 days and then colonies were counted and survival was determined.

387

IDENTIFICATION OF RAD16

Table 1.

List of Saccharomyces cerevisiae strains used

Strain MCY1093 MCY1094 MCY 1751 MCY2048 MCY1997 MCY 1250 X2180-1 B X2 180 X14-10 XS1152 LP303 1-3A XS1158 XXI 160 DST3 1 DST32 YNN282 XS1169 XSI 171 SEY2102-p DA2 100 XSI 165 X 12-6B XSI 180

Genotype*

MATa his4-539 lys2-801 ura3-52 SUC2 MATa ura3-52 ade2-101 SUC2 MCY 1093 x MCY 1094 MATa radl6-Al::URAJ ura3-52 lys2-801 his4-539 SUC2 MATa snjZA2:: URA3 ura3-52 lys2-801 ade2-101 SUCZ MATa snfl-AI::HIS3 his3-A200 ura3-52 lys2-801 SUCZ MATa SUC2 gal2 MATa / M A Ta SUC2lSUC2 ga12/ga12 MATa radl6-1 ade2-1 X14-10x MCY2048 (radl6-l/radl6-Al) MATa radl6-1 leu2-3.-112 trp2 ura3-52 MCY2048 x LP3031-3A (radlb-l/radl6-Al) MATa IMATa radl6-Al:: URA3/radl6-Al::URA3 ura3-52/ura3-52lys2-801/lys2-801 leu21 -I his4-539/+ (radl6-Al/radl6-Al) M A Ta radlb-AI::URA3 ura3-52 ade2-1 his3A200 trpl-A lys2-801 M A Ta radl6-Al:: URA3 ura3-52 ade2-1 his3 A200 trpl-A lys2-801 M A Ta ura3-52 ade2-1 his3-A200 trpl-A lys2-801 DST31 x YNN282 (RADlblradl6-Al) DST31 x DST32 (radlb-Al/radl6-Al) MATa ura3-52 leu2-3,-112his4 suc2-A9 (rho-) MATa l y s 2 d I (rad16-A2)::URA3ura3-52 leu2-3,-112his4 suc2d9 (rho-) MATa MATa rad16-l/Iys2-A! (rad16-A2)::URA3ade2-l/+ MATa radl-1 ade2-1 M A Ta / M A Ta radl-llradl -I ade2-1 lade2-1 ade6/+

*The RAD16 genotypes of some of the diploid strains are given in parentheses.

RESULTS Homology between SNF2 and RAD54

The predicted RAD54 protein (Emery et al., 1991) shares significant homology with the recently published yeast SNF2 protein sequence (M. Goebl, pers. comm.; Laurent et al., 1991). The homology (Figure 1A) covers a region of 492 amino acids of RAD54 and 466 amino acids of SNF2, encompassing 55% of the RAD54 protein and -27% of the SNF2 protein. Approximately 31% of the amino acids are identical and 40% are conserved amino acid substitutions.

-

Analysis of a third yeast gene homologous to SNF2 and RAD54

Previously, we searched the GenBank sequence database for sequences homologous to the SNF2 protein and identified a 125-amino acid sequence encoded by the non-coding strand of the 3' untranslated region of the S. cerevisiae LYS2 gene (Fleig et al., 1986; Laurent et al., 1991). This partially

sequenced open reading frame (ORF), which terminates 128 nucleotides from LYS2, is 37% identical to residues 1132-1263 of SNF2, and is also 33% identical to residues 701-827 of RAD54. The entire region lying between LYS2 and SSN6 had been cloned previously (Barnes and Thorner, 1986; J. Schultz and M. Carlson, pers. comm.), and a 2.6-kb RNA had been identified (Schultz and Carlson, 1987). The direction of transcription is as shown in Figure 2 (J. Schultz and M. Carlson, pers. comm.) and is compatible with expression of the ORF. To determine whether the region of homology to SNF2 extends further towards the N terminus of the ORF, we determined additional nucleotide sequence (Figure IB). The most significant homology is confined to the C-terminal 170 residues, which are 34% identical to residues 1088-1263 of SNF2 and 3 1YOidentical to residues 654-824 of RAD54 (Figure IA). Located N-terminal to the region of homology is a cysteine-rich sequence similar to a conserved motif described recently (Freemont et al., 1991; Kakizuka et al., 1991) (Figure 1C).

388

D. SCHILD ETAL.

A. RAD54

327

SNFZ

784

RAD54 SNFZ

400 850

RAD54

472

-T

1 -T

1-

SNFZ

912

RAD54

545 985

SNF2

RAD54

614 1058 R A D l 6 (209) SNFZ

RAD54 SNFZ

-

:::"Ra:

a

-T

TV -

?I;(DILA

L KKDV E

Y [

c E

Y H IFVNLK ~ D KC ~ KMS ~ ~

E L N K L I S EV----KKVV G GSQPLRA G I V I&a H JYIR L F IG DP N N J M U L R G F N N P E M P m

N L L N DEFDD I D L E L P D D Y N M G km--FVmE---VmP IN-------iT

686 1117

RAD16 ( 2 5 3 ) -V

T

RAD54

759 SNFZ 1190 RAD16 (325)

RP

MSLSSC VD KQDVERLF S D N l R P P K N E N T I t a L D I D G K + l i G K F DNKS T I E E P E A L E L k DAE E E ANMIHAT NPDUAAISRLTPADOQF NN

B. (1) CSHVIMPHTN F F N H F M L K N I PKFGVEGPGL E S F N N I P T L L K N I M L R R T K V ERADDLGLPP R I V T V R R D F F

( 7 1 ) N E E E K D L Y R S L Y T D S K R K Y N S F V E E G V V L N N Y A N I F T L I T RMRPLADHPD L V L K R L N N F P G D D I G V V I U ( 1 4 1 ) -PI ( 2 1 1 ) LNMSGKUQSS (281) TIKYFMNNIQ (351) SIEARIIELP

E S W F C R TKIEALVEEL CEVFLVSLKA EKKANMIHAT

TC L-P YKLRSNKRTI KSIVFSPFTS GGVALNLCEA S P V F I L D P W U INPDEAAISR LTPADLQFLF

V cH I G L S I D L SPPALEVDLD S F K K Q S I V S O M L D L V E U R L K R A G F P T V K L P GSMSPTPRDE NPSVEUQSGD R V H R I G P Y R P V K I T R F C I E D NN*

C. Reportedmotif: ~ - ~ - ~ ~ , ~ , ~ ~ - C - X ~ 1 1 - 3 O ~ - C - X - H - X - ~ F , I , L ~ - C - X ~ 2 ~ - C - ~ I , L , M ~ - X ~ l O - l 8 ~ - C - P - X - C RAD16 motif: c-x-L -C-X(ll) -C-X-H-X-F -C-X(Z)-C-I -X(15) -C-P-X-C Figure 1. (A) Homologies between RADSI, SNFZ and RAD16. Identical amino acids are highlighted in black. Regions indicated by Roman numerals are potential nucleotide-binding(I) and Mg2+-binding(11) domains, as well as regionswhich could represent additional domains seen in helicases(see Figure 4). (B) Amino acid sequence of the C-terminal half of RAD16. Amino acids in the cysteine-richregion are underlined (see panel C), and the highlighted R amino acid is the beginning of the region aligned in panel A. From amino acid K at position 268 to the end of the open reading frame confirms the published sequence of this region (Fleig ef al., 1986). (C) Cysteine-richmotif in RAD16. The cysteine-richmotifin RAD16 is comparedto a published motif seen in 1 1 other proteins (Freemont ef al., 1991; Kakizuka ef al., 1991).

The sequencing of this gene was discontinued when it was learned that the entire gene had recently been completely sequenced independently by two groups (Mannhaupt et al., 1992; D. D. Bang and P. van de Putte, pers. comm.). Disruption of the gene adjacent to LYS2 reveals no functional similarity to SNF2

To determine whether the homologous gene has any functional relationship to SNF2, a diploid heterozygous for a disruption/deletion mutation, later designated radl6-Al:: URA3 (Figure 2), was constructed and subjected to tetrad analysis. The Ura+ segregants displayed no impairment in aerobic growth on glucose or glycerol, or anaerobic

growth on glucose, sucrose, raffinose, or galactose at 30°C or 37°C. Thus, disruption of this gene does not cause an Snf-like phenotype. We also tested whether the homologous gene in multicopy could compensate for the loss of SNFZ function. pBLTY4, a 2 pm-derived plasmid carrying the gene (Figure 2), was used to transform the snj2 null mutant MCY1250 (Abrams et al., 1986). The presence of pBLTY4 failed to suppress the raffinose and galactose growth defects of the snj2 mutant. To test whether the disruption in combination with snj2 might result in lethality or give rise to an altered phenotype, we crossed strain MCY 1997 (snj2-A2::URA3)to MCY2048 (radl6-Al::URA3). Tetrad analysis of the diploid yielded four viable

389

IDENTIFICATION OF RAD16

0.5kb

,

H

pBLTY3, pBLTY4

I H

pBLY22 \

B

I

Pv-

- I

(rad16-A1 ::URA3)

Pv

B

Figure 2. Restrictionmaps of the RAD16 gene and plasmids.The region ofchromosome I1 between the SSN6 and L YS2 genes is shown. Placement of the 5’ end of the RAD16 gene is approximate,based on the position of the stop codon of the open reading frame and the size of the RNA; probes prepared from the 0.9-kb HindIII-PvuII and 1. I-kb PvuII-SphI fragments hybridized to the 2.6-kb RAD16 RNA (J. Schultzand M. Carlson, pers. comm.). The sequenced portion ofR4D16 is denoted by a bar at the top of the figure. Plasmids are described in the text. Not all sites are shown for the SSN6 and LYS2 genes or for the plasmids. The fragment between the bold PvuII sites in pBLY22 was used to isolate in vivo disruptionsof the RAD26 gene. Restriction sites: B, BurnHI; Bg, BglII; C, CluI; H, HindIII; Hp, HpaI; Nc, NcoI; Pv, PvuII; R,EcoRI and Sp, SphI.

spores from each ascus. The double mutant segregants were indistinguishable from snj2 single mutant segregants with respect to anaerobic growth on raffinose and galactose and aerobic growth on glycerol. Thus, the disruption (radl6) mutation in no way exacerbates the phenotype of a snj2 mutant. Zdentzjication of the disruptedgene as RAD16 The homology with the RAD54 protein prompted us to test whether strains disrupted for the gene adjacent to L YS2 had a defect in DNA repair. These strains were UV sensitive (e.g. MCY2048 in Figure 3A), but not x-ray sensitive (data not shown), suggesting a defect in the excision repair pathway. Disruption of this gene does not result in the extreme UV sensitivity caused by mutations in some yeast excision repair genes, such as rad1,2,3 or 4 (e.g. X12-6B in Figure 3A), but the sensitivity is similar to that caused by mutations in several less well characterized excision repair genes, including radl6 (Cox and Parry, 1968; Prakash, 1977). As RAD16 has recently been mapped near LYS2 (J. C. Game, pers. comm.), we tested whether the gene adjacent to LYS2 was RAD16. The chromosomal disruption/deletion, radl6-Al:: URA3, failed to complement the radl6-1 allele; diploids (e.g. XS1152 and XS1158) constructed by mating the disruption strain with strains containing the radl6-1 allele were UV sensitive (e.g. XS1158 in Figure 3B), whereas similar diploids constructed by crossing the disruption and radl0-1 displayed wild-type UV

sensitivity (data not shown). The disruption and radl6-1 were also shown to segregate from each other in genetic crosses: when diploid strain XS1152 was sporulated, all 240 spores from 60 dissected tetrads were UV sensitive. As a control, diploids constructed by mating either the disruption or radl6-1 to wild-type strains were shown to segregate two UV-sensitive and two UV-resistant spores. Additionally, centromere- or 2-p-derived plasmids (pBLTY3 and pBLTY4 in Figure 2) contained the complete ORF adjacent to LYS2 complement the radl6-1 mutation for UV sensitivity (Figure 3A). Taken together, these data identify the disrupted gene as RAD16. Comparison of the UV survival curves of a radl6-1 haploid with a radl6-A1 haploid (Figure 3A) indicate that they have very similar survival. This suggests that the lower sensitivity of radl6-1, as compared to mutations in most other excision repair genes, is not due to leakiness of the allele. The difference in survival observed between the two radl6-A1 homozygous diploid strains (XS 1 160 and XS1171) (Figure 3B) suggests that modifiers affecting UV survival are present in at least one of these strains, and these modifiers may represent suppressors of the radl6 deletion. A lys2 deletion which deletes the C-terminal end of RADl6 results in UV sensitivity When Barnes and Thorner (1986) isolated a deletion of lys2, they also deleted the coding sequence

A d

100

cd 10 .A

3 k

1

3

cn 0.1

0.01

0.001 0

20

40

60

80

UV dose(J/m2)

B o X2180 (RAD/RAD) 0 XS1169 (RADlradl6-N) V XS1163 (radl6-llradl6-I) v XS1158

100 d

cd 10 ‘A

(r0d16-Ilradl6-N)

3 k

XS1165 (rad16-Il~ad16-d2) XS1160 (radl6-Al/radl6-AI) A XS1171 (radl 6-Al/radl6-N) A Xs1180

1

3

cn

(radl-Ilradl-1)

0.1

8 0.01

0.001 0

20

40

60

UV dose(J/m2)

80

39 1

IDENTIFICATION OF RAD16

for the last nine amino acids of the adjacent gene that we have shown to be RAD16 (Figure 2). The UV sensitivity of DA2100, a strain containing the lys2A1 deletion, is similar to that of the rad16-A1 strain, while SEY2102, the isogenic strain from which DA2 100 was derived, displays wild-type UV sensitivity (Figure 3A). When DA2100 was crossed by a radl6-1 strain, the resulting diploid was UV sensitive (Figure 3B), indicating that lys2-Al strains are also radl6-A strains. This deletion has therefore been renamed lys2-Al(radl6-A2). By using the description of how the lys2 deletion was constructed (Barnes and Thorner, 1986), it was possible to reconstruct the predicted DNA sequence near the deleted end of rad16-A2; it appears that the deleted nine amino acids are replaced by a different 11 amino acids prior to a translation termination codon. Presumably either the terminal nine amino acids are important for function, or the missing transcription termination signal results in an unstable or inactive transcript. Genetic mapping of rad 16- 1 Initial data suggested only a loose linkage between lys2 and radl6-1 (J. C. Game, pers. comm.; Mortimer et al., 1989), but subsequent data showed a much tighter linkage (M. Bankmann, L. Prakash and S. Prakash, pers. comm.). Since our results indicated that the physical map position of RADI6 was immediately ajacent to LYS2, we retested the linkage of these two genes. Out of 36 tetrads analysed, 35 were parental ditype (showing radl6-1 and lys2 cosegregating), while one was tetratype, resulting from a recombination even between these genes. These data suggest that these genes are extremely closely linked (map distance is 1.4 cM), consistent with their physical map positions (Figure 2). In the single tetratype tetrad, lys2 segregated with the centromere-distal tyrl marker, which is most consistent with radl6 being located on the centromere-proximal side of lys2.

-

DISCUSSION The DNA sequences of the yeast recombinational repair gene RAD54 and the transcriptional activator

gene SNF2 have recently been published (Emery et al., 1991; Laurent et al., 1991), and the predicted amino acid sequences share a large region of homology. Another, previously uncharacterized gene located downstream of the LYS2 gene, also shares a similar region with both RAD54 and SNF2 (Figure 1A). Because of the homology of this gene to SNF2 and RAD54, we disrupted this gene and tested the disruption for Snf-like and Rad-like phenotypes. Although the disruption did not confer a Snf-like phenotype, strains disrupted for this gene were UV sensitive. We show here that this gene is RAD16. Independently, the RAD16 gene has also recently been cloned by direct complementation of a radl6 mutation (D. D. Bang and P. van de Putte, pers. comm.), and this gene is identical to the gene we have identified. The homologies between these three yeast genes are also shared by a number of other yeast and Drosophila genes. STHl @NF two h_omolog) (Laurent et al., 1992) and MOT1 bodifier of transcription) (Davis et al., 1992) both encode essential functions in yeast and both share significant homologies over the same region shared by RAD54 and SNF2. The yeast FUN30 Cunction Knknown now) gene also shares this region of homology (Clark et al., 1992). The brahma (brm) and lodestar genes of Drosophila also share homologies in the same region (Tamkun et al., 1992; Girdham and Glover, 1991); for brm the homology extends over the complete region shared by SNF2 and RAD54, but for lodestar the homology encompasses only about the first two-thirds of this region (RAD54 amino acids 330467). Mutations in brm suppress mutations in Antennapedia and other homeotic genes (Tamkun et al., 1992), and mutations in lodestar, a maternaleffect gene, cause chromosome tangling and breakage at anaphase (Girdham and Glover, 1991). All of these genes, including RAD16 (Mannhaupt et al., 1992), have consensus sequences for nucleotide and Mg2+ binding sites (Figure 4) (Davis et al., 1992; Girdham and Glover, 1991; Clark et al., 1992; Laurent et al., 1992; Tamkun et al., 1992), found frequently in proteins that bind and/or hydrolyze ATP (Gorbalenya and Koonin, 1990). The predicted RAD 16protein contains a cysteinerich motif that may bind Zn (Figure lC), located outside of the region of homology with RAD54 and

Figure 3. UV survival curves. (A) Haploid strains. (B) Diploid strains. The survival was determined at each dose by averaging the results from three or more separate experiments.Vertical bars indicate the standard error of the mean survival.The relevant genotype is shown under the strain name; complete genotypes are listed in Table 1.

392

D. SCHILD ET AL.

Domain I1 (Mg* binding)

Domain I (nucleotide binding) RADJ RAD54 SNFZ

34 338 785

vs

GGIS AYGC HLIG

885

SNFZ

+

CONSENSUS

G GKT

KVIWVH

CONSENSUS

++++DE

S

Domain V

Domain 111 RADJ RAD54 SNFZ

457 480 920

CONSENSUS

H D

I

SP

ION Y P

598 CSNGRGAI 721 [ E G Q H F I q r 1152 DSEYLC I SNFP RAD16 ( 2 8 8 ) N I Q C H - V m V L RADJ RAD54

V L

+

0

+TAT SGS

CONSENSUS

0

0

+++ T O + S

G O+

S

Domain VI RADJ RAD54

mtt[f

656 FDAMRHAA 753 SNFZ 1184 RAD16 ( 3 1 9 ) W m S V E W a S G b CONSENSUS

O

O

Q

G

R

-:-

@

R

H Figure 4. Potential helicase domains. Regions of the RAD54, SNF2 and RAD 16 proteins which share homologies with known helicases, such as RAD3 (Sung et al., 1987;Harosh et al., 1989).Domains I, 11,111, V and VI are similar to domains reported by Gorbalenya et al. (1989) and the consensus sequences are from their study. No regions homologous to domains IA and IV could be found (see Discussion). + represents hydrophobic amino acids (I, L, V, M, F. Y, W) and 0 represents charged or polar amino acids (S, T, D, E, N, Q, K, R). The number after each gene name refers to the positions of the domains. Identical amino acids are highlighted in black.

SNF2. The amino acid sequence of the RAD16 motif matches the conserved residues of the consensus sequence motif described recently for 1 1 other proteins (Freemont et al., 1991; Kakizuka et al., 1991), including RADI8 (involved in yeast DNA repair), RAG-I (a human V(D)J recombinationactivating gene), the herpes simplex virus IEI 10 gene, the varicella-zoster virus gene 61, the baculovirus CG30 gene, ret (a human transforming gene), rpt-1 (a human interleukin 2 receptor regulator gene), RING1, PML (a human putative transcription factor), the Z gene of lymphocytic choriomeningitus virus, and the mouse Mel-18 gene expressed in melanoma cells. The functions of these proteins, where known, bear no obvious relationships to each other, except that many are expected to involve DNA binding. The RAD18 gene, involved in the repair of both UV- and x-ray-induced DNA damage, is the only previously described Saccharomyces cerevisiae gene encoding this motif. The presence of this motif in both the RAD16 and RAD18 proteins suggests that in yeast this region may be involved in recognition and/or binding to damaged DNA.

We have recently identified a more conventional potential zinc finger motif in the RAD54 protein downstream of the region homologus to SNF2 and RAD16. The sequence of the RAD54 protein in this region (amino acids 832-869) is: CETHETYHCKRCNAQGKQL KPAPAMLYGDATTWNHLNH. Human poly(ADP-ribose) polymerase contains two zinc fingers (C-K-X-C-X(28 or 30)-H-X-X-C), and although both zinc fingers have been implicated in the binding of this enzyme to single-stranded nicks, only one appears necessary for binding to DNA double-stranded breaks (Ikejima et al., 1990). Since RAD.54 has been implicated primarily in the repair of double-strand breaks, it is possible that its potential zinc finger may be involved in the recognition and/or binding to double-strand breaks. A subset of proteins with consensus sequences for nucleotide and Mg2+ binding sites has been shown to have DNA, RNA or DNA-RNA helicase activities. Two superfamilies of helicases, each including

393

IDENTIFICATION OF RAD16

both DNA and RNA helicases, have been described and members of each family share certain motifs, in addition to the nucleotide and Mg2+ binding motifs (Gorbalenya et al., 1989; Hodgman, 1988). Both the SNF2 and RAD54 proteins have regions that fit reasonably well with five of the previously described helicase motifs in the second superfamily (Gorbalenya et al., 1989) (Figure 4), although the fit to the consensus sequence of domain V is not very good for either of these genes or for RAD3. No region homologous to domains IA or IV could be found, although a region was found (Figure 1A) which shares some similarities to domain IV of the first superfamily of helicases (Hodgman, 1988). The part of RAD16 which we sequenced encodes only regions homologous to domains V and VI, but the complete sequence of RAD16 does encode regions homologous to domains I, I1 and I11 as well (Mannhaupt et al., 1992). Neither SNF2 nor RAD54 shares significant homologies, outside of these potential helicase domains, with known yeast helicases, such as RAD3 (Sung et al., 1987; Harosh et al., 1989) or PIFl (Lahaye et al., 1991), or with helicases from other organisms (Gorbalenya et al., 1989; Hodgman, 1988). The SNF2-related genes may encode a new family of helicases, but there are currently no biochemical data to support a helicase activity. If these genes do encode helicases, it might explain why genes affecting different aspects of DNA and RNA metabolism, such as DNA repair and transcription, share so much homology. DNA repair in some organisms has been shown to require helicase activity (reviewed by Matson and KaiserRogers, 1990), and helicase activity may also be involved in transcription initiation (Buratowski et al., 1991). There are indirect experimental data from several laboratories which suggest that both SNF2 and RAD16 may play a role in altering chromatin structure. Suppressors of snfz(swi2) deletion mutations have been isolated and several appear to be in chromatin-associated proteins. SZNI, mutations in which suppress a swi2 deletion, encodes a high mobility group protein, homologous to the mammalian HMGl protein (Kruger and Herskowitz, 1991). In addition, deletions of the yeast HTA1HTBl region, which encodes one of the two copies of both histone H2A and H2B, act as suppressors of the snfz deletion (J. Hirschhorn, S. Brown, C. Clark and F. Winston, pers. comm.). These bypasssuppressors of SNF2 function suggest that SNF2 may play a role in altering chromatin prior to or during transcription. It has been suggested that the

RAD16 protein may also be involved in changing chromatin structure, since the radl6-1 mutation affects repair primarily in non-transcribed regions (Terleth et al., 1990). Even if these proteins are involved in changing chromatin structure to allow accessibility for transcription and DNA repair, they may still encode helicases and possibly even helicases whose actions result in altered chromatin. To explain the observed homologies between SNF2 and two DNA repair genes, we favor models in which related enzymatic activities are required for both transcription and DNA repair. We cannot rule out, however, that these proteins primarily affect transcription, and only indirectly affect DNA repair in some cases. Although neither rad54 nor radl6 deletions exhibits a Snf-like phenotype, they may affect specifically the transcription of genes involved only in DNA repair. One would have to argue additionally that in radl6 mutant strains the level of repair proteins is sufficient for repair in transcribed but not non-transcribed regions. Although the significance of the observed homologies is not yet understood, the experiments presented provide additional information about the RADl6 gene. Since deletions of this gene are viable, it encodes a non-essential function. Although most of the excision repair genes are non-essential, the RAD3 gene is essential (Higgins et al., 1983; Naumovski and Friedberg, 1983). In addition, UV survival curves of the radl6 deletion strains indicate that this gene is not completely required for excision repair, since they are considerably less sensitive than strains with mutations or deletions in the RAD1,2,3,4 or 10 genes (Friedberg, 1988; Haynes and Kunz, 1981). The partial UV resistance we observe in radl6 deletion strains is consistent with the ability of such strains to repair UV damage in transcribed regions (Terleth et al., 1990). ACKNOWLEDGEMENTS Most of all we wish to thank Mark Goebl for pointing out the homologies to us, since much of this work resulted from his observation. We are grateful to Michelle Treitel and Reggie Case for assistance. We also thank John Game, Jennifer Davis, Jeremy Thorner, Hannah Klein and George Basile for useful discussions. We are also grateful to Horst Feldmann and a number of researchers for allowing us to cite their unpublished results. This work was supported by grants from the National Institutes of Health (GM30990 to R.K.M. and GM34095 to M.C.) and the Office of Health

394 and Environmental Research, Office of Energy Research, U.S. Department of Energy, under Contract DE-AC03-76SF00098, and a Faculty Research Award from the American Cancer Society (to M.C.). B.C.L. was supported by National Research Service Award Postdoctoral Fellowship F32 GMll933. REFERENCES Abrams, E., Neigeborn, L. and Carlson, M. (1986). Molecular analysis of SNF2 and SNFS, genes required for expression of glucose-repressible genes in Saccharomyces cerevisiae. Mol. Cell. Biol. 6,3643-365 1. Barnes, D. A. and Thorner, J. (1986). Genetic manipulation of Saccharomyces cerevisiae by use of the LYS2 gene. Mol. Cell. Biol. 6,2828-2838. Buratowski, S., Sopta, M., Greenblatt, J. and Sharp, P. A. (1991). RNA polymerase 11-associated proteins are required for a DNA conformation change in the transcription initiation complex. Proc. Natl. Acad. Sci. USA 88,7509-7513. Clark, M. W., Zhong, W. W., Keng, T., Storms, R. K., Barton, A., Kaback, D. B. and Bussey, H. (1992). Identification of a S. cerevisiae homolog of the SNF2 transcriptional regulator in the DNA sequence of an 8.6 kb region in the LTE1-CYS1 interval on the left arm of chromosome I. Yeast 8, 133-145. Cox, B. S. and Parry, J. M. (1968). The isolation, genetics, and survival characteristics of ultraviolet light-sensitive mutants in yeast. Mutation Res. 6,37-55. Davis, J. L., Kunisawa, R. and Thorner, J. (1992). A presumptive helicase (MOT1 gene product) affects gene expression and is required for viability in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol. 12, 1879-1 892. Emery, H. S., Schild, D., Kellogg, D. E. and Mortimer, R. K. (1991). Sequence of RAD54, a Saccharomyces cerevisiae gene involved in recombination and repair. Gene 104,103-106. Fleig, U. N., Pridmore, R. D. and Philippsen, P. (1986). Construction of LYS2 cartridges for use in genetic manipulation of Saccharomyces cerevisiae. Gene 46, 237-245. Freemont, P. S., Hanson, I. M. and Trowsdale, J. (1991). A novel cysteine-rich sequence motif. Cell64,483-484. Friedberg, E. C. (1988). Deoxyribonucleic acid repair in the yeast Saccharomyces cerevisiae. Microbiol. Rev. 52, 70-102. Game, J. C. (1983). Radiation-sensitive mutants and repair in yeast. In Spencer, J. F. T., Spencer, D. M. and Smith, A. R. W. (Eds), Yeast Genetics, Fundamental and Applied Aspects. Springer-Verlag, New York, pp. 109-137. Girdham, C. H. and Glover, D. M. (1991). Chromosome tangling and breakage at anaphase result from mutations in lodestar, a Drosophilu gene encoding a

D. SCHILD ETAL.

putative nucleoside triphosphate-binding protein. Genes andDev. 5,1786-1799. Gorbalenya, A. E. and Koonin, E. V. (1990). Superfamily of UvrA-related NTP-binding proteins: implications for rational classification of recombination/repair systems. J . Mol. Biol. 213,583-591. Gorbalenya, A. E., Koonin, E. V., Donchenko, A. P. and Blinov, V. M. (1989). Two related superfamilies of putative helicases involved in replication, recombination, repair and expression of DNA and RNA genomes. Nucl. Acids Res. 17,4713-4730. Happel, A. M., Swanson, M. S. and Winston, F. (1991). The SNF2, SNFS and SNF6 genes are required for Ty transcription in Saccharomyces cerevisiae. Genetics 128,69-77. Harosh, I., Naumovski, L. and Friedberg, E. C. (1989). Purification and characterization of Rad3 ATPase/ DNA helicase from Saccharomyces cerevisiae. J. Biol. Chem. 264,20532-20539. Haynes, R. H. and Kunz, B. A. (1981). DNA repair and mutagenesis in yeast. In Strathern, J. N., Jones, E. W. and Broach, J. R. (Eds), The Molecular Biology of the Yeast Saccharomyces: Life Cycle and Inheritance. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, pp. 371-414. Higgins, D. R., Prakash, S., Reynolds, P., Polakowska, R., Weber, S. and Prakash, L. (1983). Isolation and characterization of the RAD3 gene of Saccharomyces cerevisiae and inviability of rad3 deletion mutants. Proc. Natl. Acad. Sci. USA 80,5680-5684. Hodgman, T. C. (1988). A new superfamily of replicative proteins. Nature 333,22-23, 578 (erratum). Ikejima, M., Noguchi, S., Yamashita, R., Ogura, T., Sugimura, T., Gill, D. M. and Miwa, M. (1990). The zinc fingers of human poly(ADP-ribose) polymerase are differentially required for the recognition of DNA breaks and nicks and the consequent enzyme activation. J. Biol. Chem. 265,21907-21913. Jentsch, S., McGrath, J. P. and Varshavsky, A. (1987). The yeast DNA repair gene RAD6 encodes a ubiquitinconjugating enzyme. Nature 329,131-134. Kakizuka, A., Miller, Jr., W. H., Umesono, K., Warrell, Jr., R. P., Frankel, S. R., Murty, V. V. V. S., Dmitrovsky, E. and Evans, R. M. (1991). Chromosomal translocation t( 15;17)in human acute promyelocytic leukemia fuses RARa with a novel putative transcription factor, PML. Cell 66,663-674. Kans, J. A. and Mortimer, R. K. (1991). Nucleotide sequence of the RADS7 gene of Saccharomyces cerevisiae. Gene 105, 139-140. Kantor, G. J., Barsalou, L. S. and Hanawalt, P. C. (1990). Selective repair of specific chromatin domains in UV-irradiated cells from xeroderma pigmentosum complementationgroupC. Mutation Res. 235,171-180. Kruger, W. and Herskowitz, I. (1991). A negative regulator of HO transcriptions, SIN1 (SPT2), is a nonspecific DNA-binding protein related to HMGl. MoI. Cell. Biol. 1 I, 4 I 35-4 I 46.

IDENTIFICATION OF RAD16

Lahaye, A., Stahl, H., Thines-Sempoux, D. and Foury, F. (1991). PIFl: a DNA helicase in yeast mitochondria. EMBO J . 10,997-1007. Laurent, B. C., Treitel, M. A. and Carlson, M. (1991). Functional interdependence of the yeast SNF2, SNFS, and SNF6 proteins in transcriptional activation. Proc. Natl. Acad. Sci. USA 88,2687-2691. Laurent, B. C., Yang, X. and Carlson, M. (1992). An essential Saccharomyces cerevisiae gene homologous to SNF2 encodes a helicase-related protein in a new family. Mol. Cell. Biol. 12, 1893-1902. Mannhaupt, G., Stucka, R., Ehnle, S., Vetter, I. and Feldmann, H. (1992). Molecular analysis of yeast chromosome I1 between C D M l and L YS2: the excision repair gene RAD16 located in this region belongs to a novel group of double-finger protein. Yeast, submitted. Matson, S. W. and Kaiser-Rogers, K. A. (1990). DNA helicases. Annu. Rev. Biochem. 59,289-329. Mortimer, R. K., Schild, D., Contopoulou, C. R. and Kans, J. A. (1989). Genetic map of Saccharomyces cerevisiae, edition 10. Yeast 5,321-403. Naumovski, L. and Friedberg, E. C. (1983). A DNA repair gene required for the incision of damaged DNA is essential for viability in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 80,4818-4821. Neigeborn, L. and Carlson, M. (1984). Genes affecting the regulation of SUC2 gene expression by glucose repression in Saccharomyces cerevisiae. Genetics 108, 845-858. Pearson, W. R. (1990). Rapid and sensitive sequence comparison with FASTP and FASTA. Methods Enzymol. 183,63-98. Prakash, L. (1977). Defective thymine dimer excision in radiation-sensitive mutants radlO and radl6 of Saccharomyces cerevisiae. Molec. Gen. Genet. 152, 125-128. Rothstein, R. J. (1991). Targeting, disruption, replacement, and allele rescue: integrative DNA

395 transformation in yeast. Methods Enzymol. 194, 281-301. Sanger, F., Nicklen, S. and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74,5463-5467. Schneider, R. and Schweiger, M. (1991). The yeast DNA repair proteins RADl and RAD7 share similar putative functional domains. FEBS Lett. 283,203-206. Schultz, J. and Carlson, M. (1987). Molecular analysis of SSN6, a gene functionally related to the SNFI protein kinase of Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 3637-3645. Sikorski, R. S. and Hieter, P. (1989). A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122, 19-27. Stern, M. J., Jensen, R. E. and Herskowitz, I. (1984). Five SWI genes are required for expression of the HO gene in yeast. J. Mol. Biol. 178,853-868. Sung, P., Prakash, L., Matson, S. W. and Prakash, S. (1987). RAD3 protein of Saccharomyces cerevisiae is a DNA helicase. Proc. Natl. Acad. Sci. USA 84, 8951-8955. Tamkun, J. W., Deuring, R., Scott, M. P., Kissinger, M., Pattatucci, A. M., Kaufman, T. C. and Kennison, J. A. (1992). brahma: a regulator of Drosophila homeotic genes structurally related to the yeast transcriptional activator SNF2/SWI2. Cell 68,561-572. Terleth, C., Schenk, P., Poot, R., Brouwer, J. and Van De Putte, P. (1990). Differential repair of UV damage in rad mutants of Saccharomyces cerevisiae: a possible function of G2 arrest upon UV irradiation. Mol. Cell. Biol. 10,4678-4684. Venema, J., Van Hoffen, A., Karcagi, V., Natarajan, A. T., Van Zeeland, A. A. and Mullenders, L. H. F. (1991). Xeroderma pigmentosum complementation group C cells remove pyrimidine dimers selectively from the transcribed strand of active genes. Mol. Cell. Biol. 11,4128-4134.