Molecular Characterization of Mutant Alleles of the DNA Repair/ Basal Transcription Factor Haywire/ERCC3 in Drosophila

Molecular Characterization of Mutant Alleles of the DNA Repair/ Basal Transcription Factor haywire/ERCC3 in Drosophila

Leslie C. Mounkes*,†,1 and Margaret T. Fuller† *Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, Colorado 80309 and †Departments of Developmental Biology and Genetics, Stanford University School of Medicine, Stanford, California 94309

ABSTRACT The haywire gene of Drosophila encodes a putative helicase essential for transcription and nucleotide excision repair. A haywire allele encoding a dominant acting poison product, lethal alleles, and viable but UV-sensitive alleles isolated as revertants of the dominant acting poison allele were molecularly character- ized. Sequence analysis of lethal haywire alleles revealed the importance of the nucleotide-binding domain, suggesting an essential role for ATPase activity. The viable hay nc2 allele, which encodes a poison product, has a single amino acid change in conserved helicase domain VI. This mutation results in accumulation of a 68-kD polypeptide that is much more abundant than the wild-type haywire protein.

HE haywire locus of Drosophila encodes the fly ho- ily of DNA/RNA helicases (Gorbalenya et al. 1989; Tmolog of ERCC3, a human gene associated with the Linder et al. 1989). Two of these, the predicted nucleo- DNA repair deficiency disease xeroderma pigmentosum tide- and magnesium-binding domains, are responsible group B (XP-B) (Mounkes et al. 1992). Mutations in for the essential ATPase activity of the SSL2/RAD25 the Saccharomyces cerevisiae homolog SSL2/RAD25 are de- (Park et al. 1992) and ERCC3 (Ma et al. 1994a) products. fective in overall nucleotide excision and transcription- Helicase activity is associated with the TFIIH transcrip- coupled repair (Sweder and Hanawalt 1994). In addition complex (Schaeffer et al. 1993; Drapkin et al. tion, these proteins appear to play a role in transcrip- 1994), and helicase activity of the ERCC3 protein has tion. The human ERCC3 (Schaeffer et al. 1993) and been demonstrated in vitro (Ma et al. 1994b). Helicase S. cerevisiae SSL2/RAD25 (Feaver et al. 1993) gene prod- activity of ERCC3 requires the remaining conserved heli- ucts are found to be associated with the basal transcrip- case domains, which have been shown by mutational tion factor TFIIH. Purified TFIIH complements the ex- analysis to function in excision repair of UV-sensitive cision repair defect of lysates from cells lacking ERCC3 cell cultures (Ma et al. 1994a). function (Drapkin et al. 1994). Consistent with a func- The Drosophila ERCC3 homolog haywire was originally tion in transcription, SSL2/RAD25 is an essential gene identified through an unusual mutant allele, hay nc2, that (Park et al. 1992), and a temperature-sensitive allele acted as a dominant enhancer of mutations in the testis- of SSL2/RAD25 inhibits transcription at the restrictive specific ␤-tubulin B2t. Males heterozygous for either temperature (Qui et al. 1993). Different alleles of SSL2/ hay nc2 or B2t null and wild type are fertile at 25Њ. However, RAD25 show defects that can separate the roles of the flies transheterozygous for one copy of hay nc2 and one protein in excision repair and transcription (Guzder copy of B2t null are male sterile at 25Њ. The genetic interac- et al. 1994). As for SSL2/RAD25 of yeast, extreme alleles tion between hay nc2 and B2t null suggested the haync2 allele of haywire are lethal (Regan and Fuller 1990), whereas encoded a poison product because flies transheterozy- viable mutant alleles are UV sensitive (Mounkes et al. gous for B2t mutations and a deficiency of haywire were 1992). Only a few cases of XP-B are known to date, male fertile (Regan and Fuller 1988). Revertant alleles suggesting that ERCC3 is also essential in humans, and of hay nc2 abolish the genetic interaction between hay nc2 that only patients carrying special weak mutant alleles of and B2t null, but still result in novel abnormal phenotypes ERCC3 survive to birth (Weeda et al. 1990; Vermeulen et with respect to male fertility (Regan and Fuller 1990), al. 1994). A viable point mutation in ERCC3 now defines suggesting that a poison product encoded by the hay nc2 a new trichothidystrophy complementation group which, allele is responsible for the failure to complement tu- however, does not cause XP-B (Weeda et al. 1997). bulin mutations. The fly, yeast, and human ERCC3 homologs show hall- In this article, we report the molecular characteriza- mark conserved domains characteristic of the superfam- tion of mutations that identify regions of haywire that are important either for viability or DNA repair. In addition, we show data indicating that the hay nc2 allele causes Corresponding author: Margaret T. Fuller, Department of Develop- production of a 68-kD polypeptide that accumulates mental Biology, Stanford University School of Medicine, Stanford, CA 94305-5427. E-mail: [email protected] to much higher levels than wild-type haywire protein. 1 Present address: California Pacific Medical Center Research Institute, Mutant haywire alleles that revert the genetic interaction Stern Building, 2330 Clay St., San Francisco, CA 94115. between hay nc2 and B2t null abrogate production of the

Genetics 152: 291–297 ( May 1999) 292 L. C. Mounkes and M. T. Fuller stable, truncated protein, suggesting that the same de- haywire antibody production and purification: A 2.5-kb fect causes accumulation of the 68-kD polypeptide and HindIII-NotI fragment from the haywire cDNA plasmid was treated with Klenow (Mounkes et al. 1992), inserted into the poison product activity. pQE30 (Qiagen) digested with SmaI, and phosphatase treated to make a construct encoding full-length haywire protein tagged with six histidine residues at the N terminus of haywire. MATERIALS AND METHODS Induction of fusion protein expression with 2 mm IPTG Fly strains and culture: Fly stocks and crosses were main- (Sigma, St. Louis) and harvesting of fusion protein were per- tained using standard fly media at 25Њ. Balancer chromosomes, formed as described (Mounkes and Fuller 1998) with the visible markers, Df(1)w, B2t null, and Df(3L)E(z)7e11 (67E1-4;67F1- following differences. haywire fusion protein was not soluble in 8 m urea and was solubilized in 6 m guanidine/TN1. Super- 3) are described in Lindsley and Zimm (1992). The following 2ϩ haywire alleles were used: hay nc2, a recessive male sterile muta- natants with soluble protein were applied to a Ni –NTA col- tion that fails to complement mutations in B2t (Regan and umn (Qiagen) equilibrated in 6 m guanidine/TN1. The col- Fuller 1988), and six alleles recovered as revertants of the umn was washed with 6 m guanidine/TN1, and fusion protein failure of hay nc2 to complement B2t null (Regan and Fuller was eluted in TN1/8 m urea and 60 mm imidazole. Antiserum 1990). Fertility of mutant combinations of alleles was deter- to purified protein was raised in rabbits (Josman Laboratories, mined by testis dissections, as described (Mounkes et al. 1992). Napa, CA). Antibodies were affinity purified using the pure, Sequencing of haywire alleles: The coding region of haywire original antigen coupled to Affi-gel 10 (Bio-Rad, Richmond, was divided into halves by two sets of PCR primers, each ampli- CA) essentially as described (Mounkes and Fuller 1998). kb fragments from haywire. Primers used to amplify Western analysis: Protein samples were made by grinding-1.5ف fying the 5Ј half of haywire were 5Ј-GGAAGCTTGCCATACGTGCT 20 adult flies in an Eppendorf tube in 0.15 ml of solution GTGTTAC-3Ј, and 5Ј-CCAAGCTTCCAGCACCATGATGC A (0.1 m Tris, pH 9.0, 0.1 m EDTA, 1% SDS, 0.5% DEPC) CCC-3Ј. Primers used to amplify the 3Ј half of haywire were plus 1 mm PMSF. Samples were spun in the microfuge for 30 5Ј-GGGCTGTGGAATTCTTGTGACCACATAC-3Ј and 5Ј-GGA sec to pellet large body parts. Supernatants were transferred AGCTTGGGTTACTGCAGTGGTAAAG-3Ј. For the viable al- to a new tube, 10 ␮l each were brought to 1ϫ sample buffer leles hay nc2 and hay nc2rv8, templates for the PCR reactions were (Laemmli 1970), and proteins were separated by 10% SDS- genomic DNA made from flies heterozygous for the viable PAGE. Gels were blotted to nitrocellulose (Schleicher & haywire allele and a deficiency for haywire, Df(3L)E(z)7e11. For Schuell, Keene, NH) as described (Towbin et al. 1979). haywire lethal haywire alleles, template DNA for the PCR reactions was was visualized by enhanced chemiluminescence (ECL) stain- made from flies heterozygous for the lethal revertant allele ing as described by the manufacturer (Amersham). Affinity- and the parent chromosome, hay nc2. PCR products were puri- purified primary antibody was used at 1:1500. Horseradish fied and sequenced as described (Sanger et al. 1977; Biggin peroxidase-conjugated donkey anti-rabbit secondary antibody et al. 1983; Kretz et al. 1989). One strand of the revertant (Amersham) was used at 1:4000. alleles was sequenced, except in areas where there were Germline transformations: PCR mutagenesis of a 380-nucle- changes. Changes were verified by sequencing the comple- otide BglII-SacI fragment of the 3Ј end of the haywire cDNA mentary strand. was performed to introduce the same C-to-T transition found Northern analysis: Total RNA was isolated by RNAzol extrac- in the hay nc2 allele. The mutagenic oligonucleotide had the tion according to Cinna Scientific. RNA (30 ␮g) was run on sequence 5Ј-CCAGATCTCTTCGCATGGCGGCTCTTGTCGT 0.1 m sodium phosphate gels, blotted onto a Hybond mem- CAGGAGG-3Ј. The mutagenized BglII-SacI fragment was sub- brane (Amersham, Arlington Heights, IL), and hybridized as cloned into a 4.5-kb BamHI fragment of the 3Ј half of a geno- described (Sambrook et al. 1989). Blots were probed with mic haywire clone (Mounkes et al. 1992). The resulting BamHI a haywire riboprobe made from a HindIII-linearized cDNA fragment was then used to replace the corresponding wild-type construct (Mounkes et al. 1992) and a random-primer-labeled fragment by cloning into a 5.2-kb BamHI fragment obtained by 1.5-kb EcoRI-SalI fragment from the plasmid containing rp49 partial digestion of the haywire genomic clone. The entire 7.8- (O’Connell and Rosbash 1984). kb haywire genomic fragment with the C-to-T change was then

TABLE 1 Alterations in lethal and viable haywire alleles

Nucleic acid Amino acid Allele Typea change change Residue bhaync2rv1 L CAG TAG Gln STOP 657 → → bhay nc2rv2 L TGC TAC Cys Tyr 377 → → bhay nc2rv3 L CGA TGA Arg STOP 380 → → bhay nc2rv7 L TGG TGA Trp STOP 441 → → bhay nc2rv8 V, uv GAG AAG Glu Lys 278 → → chay nc2rv12 L Rearranged Deletes GAGKS 265–269 bhay nc2 V, uv CGT TGT Arg Cys 652 → → a haywire alleles are either lethal (L) or viable (V) and UV sensitive (uv). b Single nucleotide changes in hay nc2rv1,2,3,7,8 and haync2 are consistent with the mode of induction of these alleles by EMS mutagenesis (Regan and Fuller 1990), a potent inducer of point mutations (Ashburner 1989) that favors GC-to-AT transitions (Williams and Shaw 1987). c The hay nc2rv12 mutation was induced by diepoxybutane (L. C. Mounkes and K. Schuske, unpublished results), a known inducer of deﬁciencies (e.g., Olsen and Green 1982). haywire Alleles 293

Figure 2.—The hay nc2 mutation results in accumulation of a 68-kD polypeptide (nc2) not seen in the wild type. Western analysis of whole-fly homogenates from haywire alleles. WT in first lanes (A and B) indicates homogenates from the wild- type background strain, red e, on which the hay nc2 mutation and subsequent revertant alleles were made. Positions of molecular weight markers are indicated on the left in kilodaltons. (A) Mutant haywire alleles over a wild-type balancer. (B) Compari- son of mutant haywire alleles in combination with a deficiency (Df ) or wild-type balancer chromosome (ϩ).

(Figure 1). This missense mutation was present in all the revertant alleles of hay nc2, as expected. As hay nc2 homozygotes are viable, arginine 652 is not absolutely required for haywire function in basal transcription, but it may be important in DNA repair since the hay nc2 allele is UV sensitive (Mounkes et al. 1992). Figure 1.—Alterations in lethal and viable haywire alleles. The hay nc2 mutation resulted in accumulation of a Amino acid changes are indicated above the appropriate resi- 11 polypeptide that migrated on SDS polyacrylamide gels dues. X indicates sequences deleted in the e wild-type strain; as 68 kD, cross-reacted with anti-haywire antisera, and $ indicates a mutation to a stop codon. The conserved helicase domains are defined by dashed lines underneath the se- was present at much higher levels in the mutant than in quences. The asterisk indicates the conserved lysine within the wild type (Figure 2). The 68-kD apparent molecular the P loop of the nucleotide-binding domain. WT, wild type; weight polypeptide was 26 kD less than expected for rv2, hay nc2rv2; rv3, hay nc2rv3; rv8, hay nc2rv8; rv12, hay nc2rv12; rv7, nc2rv7 nc2rv1 the full-length protein, suggesting the possibility that hay ; rv1, hay . Alleles in boldface type are homozygous nc2 viable. All other mutant haywire alleles are lethal. the hay allele results in accumulation of a shortened form of the haywire protein. Although truncation at the site of the hay nc2 mutation (Arg652-to-Cys) would pro- excised with HindIII and cloned into the pCaSpeR transforma- duce a predicted 68-kD N-terminal peptide, the region tion vector (Pirotta 1988) previously linearized with HindIII of haywire protein present in the 68-kD polypeptide is and treated with calf intestinal phosphatase (CIP). All cloning not known. It is also formally possible that the 68-kD junctions within haywire and the mutated site were sequenced to verify sequence integrity at the cloning junctions and intro- polypeptide does not correspond to a fragment of the duction of the mutation. Restriction enzymes, CIP, and Taq haywire protein, but instead is an unrelated polypeptide DNA polymerase were purchased from Boehringer Mannheim that is overproduced specifically in the hay nc2 mutant (Indianapolis). DNA (200 ␮g/ml) twice purified by cesium background and that cross-reacts with the anti-haywire chloride banding was injected by standard methods (Rubin antisera. The 68-kD polypeptide did not accumulate in and Spradling 1982) into Df(1)w embryos with p[ryϩ, ⌬2–3] (Robertson et al. 1988) at 200 ␮g/ml as a source of trans- either the wild-type background strain (red e, Figure 2) posase. One transformant linked to the X chromosome was or in flies carrying revertant alleles of hay nc2 (Figure isolated. 2), indicating that accumulation of the 68-kD product results from the hay nc2 allele. The affinity-purified, anti- haywire antibodies recognized many bands smaller than RESULTS the expected wild-type size in Western blots. Although Molecular lesion in the hay nc2 allele responsible for some of these smaller peptides could be proteins with the dominant genetic interaction with B2t: The original cross-reacting epitopes, some could also be degradation hay nc2 mutation caused replacement of arginine 652 with products of haywire. a cysteine (Table 1) in conserved helicase domain VI To demonstrate that the Arg652-to-Cys mutation was 294 L. C. Mounkes and M. T. Fuller

TABLE 2 Transgenic animals made by introduction of the Arg652-to-Cys change of hay nc2 into a genomic rescue fragment show a genetic interaction with the B2t null allele

Fertilea Sterilea ϩ B2t null 1b 035 haync2 ϩ 2c B2t null/ϩ 35 0 3c haync2/ϩ 27 0 ϩ B2t null 4 32 0 Df(hay) ϩ ϩ B2t null 5dcP[A3, wϩ]/Ͼ; 151 Df(hay) ϩ ϩϩ 6 P[A3, wϩ]/Ͼ; 42 0 Df(hay) ϩ ϩϩ 7 P[A3, wϩ]/Ͼ; 23 8 ϩ B2t null ϩ Figure 3.—The hay nc2 missense mutation is responsible for 8 P[A3, wϩ]/Ͼ; 33 0 ϩ production of the 68-kD polypeptide (nc2 trunc). Western analysis of hay nc2 transgenic animal made by site-directed muta- Lines 1–3, Sibling progeny from the cross hay nc2/TM3 ϫ genesis of the wild-type gene. Molecular weight markers (kilo- B2t null/TM3. daltons) are indicated on the right. WT, wild-type haywire Line 4, Df(hay) indicates Df(3L)E(z)7e11, a deletion that re- protein; nc2 trunc, truncated product accumulating in hay nc2 moves haywire. Test class flies were derived from crossing flies. The lanes are as follows: ϩ, wild-type background strain Df(1)w; B2t null/TM3 and Df(1)w; E(z)7e11/TM3 flies. A deletion (red e); nc2/ϩ, hay nc2 over a wild-type balancer (TM3); of hay complements B2t null, while the original hay nc2 allele fails ϩ P[A3,w ], transgenic animal with mutated copy of haywire car- to complement B2t null. rying the hay nc2 C T transition introduced into an otherwise ϩ ϩ → Lines 5–7, sibling progeny from the cross P[A3, w ]/P[A3,w ]; wild-type haywire genomic fragment; w; ScO/CyO, genetic back- Df(3L)E(z)7e11/TM3 virgin females ϫ Df(1)w; riB2t nulle/TM6B ϩ ground in which the transgene was propagated; P[A, ry ], males. The hay nc2 transgene caused failure to complement wild-type copy of haywire introduced into flies. Note: Both B2t null in the presence of one endogenous wild-type copy of ϩ ϩ P[A3, w ] and P[A, ry ] have two wild-type endogenous copies haywire partially alleviated male sterility (line 7), suggesting of haywire. that the poison product encoded by the hay nc2 allele on the transgene acts as an antimorph and competes with wild-type haywire protein. responsible for the production of the 68-kD polypeptide Line 8, The transgene insert did not cause male sterility on nc2 its own. in hay animals, we constructed a transgenic animal a Fertility was scored by visual inspection of dissected testes. with the hay nc2 C-to-T transition introduced into an oth- b The original genetic interaction between hay nc2 and B2t null. erwise wild-type haywire genomic fragment. The mutated c Control class siblings. d ϩ haywire gene was introduced into a background wild P[A3, w ] indicates an insertion on the X chromosome of the mutant copy of haywire, introducing the same Arg652-to- type for haywire, and the resulting transformants were Cys change found in the haync2 allele. tested for production of the truncated hay nc2 product.

Several transformant lines established from a single Go fly all made and accumulated the 68-kD polypeptide although these flies were not as sterile as hay nc2 ϩ/ϩ seen in the original hay nc2 flies (Figure 3). B2t null. Thus, the genetic interaction appeared to be Transformants carrying the Arg652-to-Cys change sensitive to the ratio between mutant and wild-type hay- also showed the genetic interaction with B2t null charac- wire (Table 2, compare lines 5 and 7). Similar behavior teristic of the original hay nc2 mutation. One copy of the was observed in the failure of a missense mutation in altered gene carried on the insert plus one wild-type ␣-tubulin to complement B2t null (Hays et al. 1989), and it endogenous copy of haywire resulted in male sterility in is consistent with hay nc2 encoding an antimorphic poison flies heterozygous for the B2t null allele (Table 2, line 5). that can compete with the wild-type protein. The (The flies shown in Table 2, line 5, have one of the two Arg652-to-Cys altered transgene caused both accumula- endogenous copies of haywire deleted.) One copy of the tion of the 68-kD polypeptide and failure to comple- altered transgene also caused reduced male fertility in ment B2t mutations, demonstrating that the single mu- B2t null/ϩ heterozygotes, even in the presence of two wild- tation was responsible for both phenotypes associated type endogenous copies of haywire (Table 2, line 7), with hay nc2. haywire Alleles 295

A missense mutation (hay nc2rv2) and small rearrangement (hay nc2rv12) indicate that the nucleotide-binding domain of haywire is essential for viability. The hay nc2rv2 mutation replaced cysteine 377 with a tyrosine residue (Table 1). This cysteine, which lies eight amino acids downstream of the P loop (Walker et al. 1982) in the nucleotide- binding domain (Figure 1), is conserved among the Drosophila, human, mouse, and yeast homologs. The hay nc2rv12 lesion was a small chromosomal rearrangement located at the ‘GAGKS’ P loop of the nucleotide-binding site (brackets in Figure 1). All these revertants of the genetic interaction of hay nc2 also reverted the defect that leads to accumulation of the 68-kD polypeptide (Fig- ure 2). A viable revertant of hay nc2: The hay nc2rv8 viable mutation changed glutamic acid 278 to lysine, introducing Figure 4.—Flies carrying the hay nc2 allele accumulate normal levels of haywire transcript. Northern analysis of flies hemi- a charge change and possibly affecting protein structure or heterozygous for mutant haywire alleles. Molecular weight or stability. hay nc2rv8 homozygotes are viable, indicating markers (in kilobases) are indicated to the left of the blot. A that glutamic acid 278 is not essential for basal transcrip- 2.5-kb haywire transcript (hay) was observed in all haywire al- tion. Glu278 could be important for the repair of UV leles. The blot was reprobed for rp49 (20) as a loading control. damage, since this allele also showed UV sensitivity Lanes: 1, red e;2,E(z)7e11/TM3;3,hay nc2/E(z)7e11;4,hay nc2rv8/ E(z)7e11;5,hay nc2/TM3; 6, hay nc2rv8/TM3; 7, hay nc2rv1/TM3; 8, (Mounkes et al. 1992). Alternatively, the UV sensitivity hay nc2rv2/TM3; 9, hay nc2rv3/TM3; 10, hay nc2rv7/TM3; 11, hay nc2rv12/ could be caused by the still-present hay nc2 lesion, while TM3. the hay nc2rv8 reversion merely alleviated the poison nature of the hay nc2 product. Again, the hay nc2rv8 missense mutation also reverted the defect leading to accumulation The high levels of the 68-kD possible fragment of hay- of the 68-kD polypeptide characteristic of hay nc2. wire accumulating in the hay nc2 mutant did not appear Altered amino acids encoded by wild-type haywire al- to result from increased transcription or stability of the leles: The sequences of three wild-type alleles from dif- haywire mRNA encoded by hay nc2. Northern blot analysis ferent genetic backgrounds revealed regions of haywire/ of all the alleles sequenced revealed relatively normal ERCC3 that are not essential (Figure 1). A wild-type levels of haywire transcript in all cases (Figure 4), with cDNA from a dp cn bn strain and two different wild-type the possible exception of the hay nc2rv3 and hay nc2rv7 alleles. genomic sequences (red e and e11) differed at several hay nc2rv3/TM3 and hay nc2rv7/TM3 flies appeared to have positions within the amino acid coding region. Alanine reduced levels of haywire mRNA compared to other al- residues at positions 238 and 244 in the dp cn bn and red e leles tested over the same balancer chromosome. As backgrounds were changed to threonine and proline, both of these alleles introduce stop codons early in the respectively, in wild-type haywire from the e11 chromo- haywire protein-coding region (below), they may cause some. Amino acid 663 was a methionine in the e11 strain destabilization of the mRNA encoding the mutant allele but a leucine in both the red e and dp cn bn strains. In (Talesa et al. 1995). In addition, accumulation of the addition, the e11 wild-type allele of haywire had a four- 68-kD polypeptide in hay nc2 did not appear to be caused amino-acid deficiency at positions 253–256 compared by translation of an abnormally processed message, as with haywire on both the red e and dp cn bn chromosomes. no alterations in haywire message size were detected in The four consecutive amino acids missing in the e11 flies carrying the hay nc2 allele. background are valine/valine/alanine/alanine. Most of Lethal alleles isolated as revertants of hay nc2: Several the wild-type polymorphisms occur in regions that are mutations that revert the failure of hay nc2 to complement poorly conserved between the Drosophila, yeast, and mutations in B2t introduce stop codons into the haywire mammalian homologs, and may identify residues of hay- coding region, supporting the hypothesis that hay nc2 fails wire not essential for function. to complement B2t because of the production of a poison product (Table 1). hay nc2rv1 introduces a stop codon DISCUSSION in helicase domain VI, hay nc2rv3 results in a stop codon in the conserved nucleotide-binding domain, and hay nc2rv7 The single nucleotide mutation in the hay nc2 allele has a stop codon six amino acids before the predicted that causes an arginine-to-cysteine change is responsible magnesium-binding domain (helicase domain II, Figure for both the production of a shortened polypeptide 1). These nonsense mutations are all lethal, supporting associated with the hay nc2 allele and the genetic interac- an essential role in flies for the haywire gene product, tion observed between hay nc2 and B2t mutations. Several probably in basal transcription. molecular mechanisms of reverting the genetic interac- 296 L. C. Mounkes and M. T. Fuller tion between hay nc2 and B2t null are suggested by sequence mulates in hay nc2 flies is the product of some other gene analysis of the haywire alleles obtained as revertants of that is upregulated in this particular mutant back- hay nc2. Destruction of the P loop of the nucleotide-bind- ground. If so, however, the product of this hypothetical ing domain reverted the genetic interaction between upregulated gene would have to cross-react with the hay nc2 and B2t null (hay nc2rv12, Table 1) possibly by destroy- affinity-purified, anti-haywire antiserum. In addition, ing the ATP-binding or helicase function of the hay nc2 upregulation of the other hypothetical gene appears to primary product. Two other revertant alleles, hay nc2rv2 depend strictly on the poison product effect of the hay nc2 and hay nc2rv8, caused missense mutations that could alter allele, as the polypeptide was not detected in any of the protein conformation or stability. hay nc2rv2, which caused six revertants of hay nc2. a change from a cysteine to tyrosine, altered a potential The hay nc2 allele genetically interacts with mutations disulfide bond partner. hay nc2rv8, which caused a change in the B2t locus (Regan and Fuller 1988), which en- from a glutamic acid to lysine, caused a change from a codes a testis-specific ␤-tubulin isoform (Kemphues et negatively charged residue to a positively charged resi- al. 1979). Levels of ␤2-tubulin are critical for spermato- due. Finally, nonsense mutations hay nc2rv1, hay nc2rv3, and genesis in flies. Flies carrying a deficiency for B2t are nc2rv7 nc2 hay reverted the genetic interaction between hay already close to a threshhold requirement of ␤2-tubulin, and Bt2 null. The finding that at least four out of six since B2t null heterozygotes are fertile at 25Њ but sterile revertants of hay nc2 (the three stop codons and the at 18Њ (Fuller et al. 1989). Further decreasing the level P-loop mutation) are mutations that should knock out of transcription by altering the function of a basal tran- nc2 haywire function supports the idea that the original hay scription factor might bring the level of expressed ␤2- allele encodes a poison product. tubulin below the critical threshold required for the The shortened polypeptide that accumulates in hay nc2 many microtubule-mediated tasks necessary for sperma- flies is likely to be either responsible for or a direct togenesis. It is feasible that either full-length haywire result of the poison nature of the hay nc2 allele, as the protein with the Arg652-to-Cys mutation or accumula- 68-kD polypeptide no longer accumulated in six out of tion of a 68-kD fragment of the haywire protein in hay nc2 six mutants that reverted the dominant enhancer effect flies could cause such a deleterious effect on transcrip- of hay nc2 (Figure 2). The 68-kD polypeptide that accumu- tion, causing the hay nc2 mutation to act as a dominant lates in hay nc2 flies is not likely to be caused by a simple enhancer of B2t mutations. truncation of the haywire protein near the site of the nc2 We thank Barbara Robertson and Kim Schuske, who conducted nc2rv1 mutation. The stop codon mutation in hay (Table 1) genetic screens to identify new revertant alleles of hay nc2, and Cricket suggests that a simple truncation event near amino acid Wood for technical assistance. We thank Stig Hansen and Robert 652 results in neither production of a poison product Tjian for help and advice in producing affinity-purified antibodies to nor accumulation of the 68-kD polypeptide (Figure 2). the haywire protein. We acknowledge the protein and nucleic acid nc2rv1 (PAN) facility, Stanford University, for synthesizing oligonucleotides Furthermore, the hay allele is lethal, perhaps be- used in this work. This work was supported by National Institutes of cause of truncation of the protein within the last con- Health grants HD-18127 and HD-29194 to M.T.F. served helicase domain, whereas the haync2 allele is viable (Table 1). The hay nc2 allele encodes a product that must contain LITERATURE CITED nc2 at least some wild-type function, as hay /Df(hay) flies Ashburner, M., 1989 Drosophila: A Laboratory Handbook. Cold Spring were fully viable. As the mutation responsible for caus- Harbor Laboratory Press, Cold Spring Harbor, NY. ing accumulation of the truncated product is not a stop Biggin, M. D., T. J. Gibbon and G. F. Hong, 1983 Buffer gradient gels and 35S label as an aid to rapid DNA sequence determination. codon, full-length haywire protein could be initially ex- Proc. Natl. Acad. Sci. USA 80: 3963–3965. pressed from the hay nc2 allele. It is possible that this Drapkin, R., J. T. Reardon, A. Ansari, J.-C. Huang, L. Zawel et full-length but mutant haywire protein could provide al., 1994 Dual role of TFIIH in DNA excision repair and in transcription by RNA polymerase II. Nature 368: 769–772. sufficient haywire function for viability when initially ex- Feaver, W. J., J. Q. Svejstrup, L. Bardwell, A. J. Bardwell, S. pressed. However, defects in functioning or processing Buratowski et al., 1993 Dual roles of a multiprotein complex of the defective protein because of the Arg652-to-Cys from S. cereviseae in transcription and DNA repair. Cell 75: 1379– 1387. mutation could alter or disrupt normally occurring pro- Fuller, M. T., C. L. Regan, L. L. Green, B. Robertson, R. Deuring cessing of the full-length haywire protein, resulting in et al., 1989 Interacting genes identify interacting proteins in- accumulation of a shortened form of the protein in volved in microtubule function in Drosophila. Cell Motil. Cytoskel- nc2 eton 14: 128–135. hay . The observation that the 68-kD shortened poly- Gorbalenya, A. E., E. V. Koonin, A. P. Donchenko and V. M. Blinov, peptide accumulated to much higher levels than the 1989 Two related superfamilies of putative helicases involved wild-type haywire product also suggests the possibility in replication, recombination, repair and expression of DNA and RNA genomes. Nucleic Acids Res. 17: 4713–4730. of turnover of haywire/ERCC3 as a normal part of its Guzder, S. N., P. Sung, V. Bailly, L. Prakash and S. Prakash, function. We would like to raise the possibility that the 1994 RAD25 is a DNA helicase required for DNA repair and haywire protein might normally be degraded as part of RNA polymerase II transcription. Nature 369: 578–581. Hays, T., R. Deuring, B. Robertson, M. Prout and M. T. Fuller, the functional cycle of TFIIH in transcription initiation. 1989 Interacting proteins identified by genetic interactions: a It is also possible that the 68-kD polypeptide that accu- missense mutation in ␣-tubulin fails to complement alleles of haywire Alleles 297

the testis-specific ␤-tubulin gene of Drosophila. Mol. Cell. Biol. 9: Regan, C. L., and M. T. Fuller, 1990 Interacting genes that affect 875–884. microtubule function in Drosophila melanogaster: two classes of Kemphues, K. J., R. A. Raff, T. C. Kaufman and E. C. Raff, 1979 mutation revert the failure to complement between haync2 and Mutation in a structual gene for a ␤-tubulin specific to testis in mutations in tubulin genes. Genetics 125: 77–90. Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 76: 3991–3995. Robertson, H. M., C. R. Preston, R. W. Phillis, D. M. Johnson- Kretz, K. A., G. S. Carson and J. S. O’brien, 1989 Direct sequencing Schlitz, W. K. Benz et al., 1988 A stable genomic source of from low-melt agarose with Sequenase. Nucleic Acids Res. 17: P element transposase in Drosophila melanogaster. Genetics 118: 5864. 461–470. Laemmli, U. K., 1970 Cleavage of structural proteins during the Rubin, G. M., and A. C. Spradling, 1982 Genetic transformation assembly of the head of bacteriophage T4. Nature 227: 680–685. of Drosophila with transposable element vectors. Science 218: 348– Linder, P., P. F. Lasko, M. Ashburner, P. Leroy, P. J. Nielsen et 353. al., 1989 Birth of the D-E-A-D box. Nature 337: 121–122. Sambrook, J., E. F. Fritsch and T. Maniatis, 1989 Molecular Clon- Lindsley, D. L., and G. G. Zimm, 1992 The Genome of Drosophila ing: A Laboratory Manual. Cold Spring Harbor Laboratory Press, melanogaster. Academic Press, San Diego. Cold Spring Harbor, NY. Ma, L., A. Westbroek, A. G. Jochemsen, G. Weeda, A. Bosch et al., Sanger, F., S. Nicklen and A. R. Coulson, 1977 DNA sequencing 1994a Mutational analysis of ERCC3, which is involved in DNA with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74: repair and transcription initiation: identification of domains es- 5463. sential for the DNA repair function. Mol. Cell. Biol. 14: 4126– Schaeffer, L., R. Roy, S. Humbert, V. Moncollin, W. Vermeulen 4134. et al., 1993 DNA repair helicase: a component of BTF2 (TFIIH) Ma, L., E. D. Siemssen, M. H. M. Noteborn and A. J. van der Eb, basic transcription factor. Science 260: 58–63. 1994b The xeroderma pigmentosum group B protein ERCC3 Sweder, K. S., and P. C. Hanawalt, 1994 The COOH terminus of produced in the baculovirus system exhibits DNA helicase activity. suppressor of stem loop (SSL2/RAD25) in yeast is essential for Nucleic Acids Res. 22: 4095–4102. overall genomic excision repair and transcription-coupled repair. Mounkes, L. C., and M. T. Fuller, 1998 The DUG gene of Drosophila J. Biol. Chem. 269: 1852–1857. melanogaster encodes a structural and functional homolog of the Talesa, V., E. Culetto, N. Schirru, H. Bernardi, Y. Fedon et al., S. cerevisiae SUG1 predicted ATPase associated with the 26S protea- 1995 Characterization of a null mutation in ace-1, the gene some. Gene 206: 165–174. encoding class A acetylcholinesterase in the nematode Caenorhab- Mounkes, L. C., R. S. Jones, B.-C. Liang, W. Gelbart and M. T. ditis elegans. FEBS Lett. 357: 265–268. Fuller, 1992 A Drosophila model for xeroderma pigmentosum Towbin, H., T. Staehelin and J. Gordon, 1979 Electrophoretic and Cockayne’s syndrome: haywire encodes the fly homolog of transfer of proteins from polyacrylamide gels to nitrocellulose ERCC3, a human excision repair gene. Cell 71: 925–937. sheets: procedure and some applications. Proc. Natl. Acad. Sci. O’Connell, P., and M. Rosbash, 1984 Sequence, structure, and USA 76: 4350–4354. codon preference of the Drosophila ribosomal protein 49 gene. Nu- Vermeulen, W., R. J. Scott, S. Rudger, H. J. Muller, J. Cole et al., cleic Acids Res. 12: 5495–5513. 1994 Clinical heterogeneity within xeroderma pigmentosum as- Olsen, O. A., and M. M. Green, 1982 The mutagenic effects of sociated with mutations in the DNA repair and transcription gene diepoxybutane in wild-type and mutagen-sensitive mutants of Dro- ERCC3. Am. J. Hum. Genet. 54: 191–200. sophila melanogaster. Mutat. Res. 92: 107–115. Walker, J. E., M. Saraste, M. J. Runswick and N. J. Gay, 1982 Dis- tantly related sequences in the a- and b-subunits of ATP synthase, Park, E., S. N. Guzder, M. H. Koken, D. I. Jaspers, G. Weeda et al., myosin, kinases and other ATP-requiring enzymes and a common 1992 RAD25 (SSL2), the yeast homolog of the human xero- nucleotide binding fold. EMBO J. 1: 945–951. derma pigmentosum group B DNA repair gene, is essential for Weeda, G., R. C. A. van Ham, W. Vermeulen, D. Bootsma, A. J. van viability. Proc. Natl. Acad. Sci. USA 89: 11416–11420. der Eb et al., 1990 A presumed DNA helicase encoded by Pirotta, V., 1988 Vectors for P-mediated transformation of Drosoph- ERCC-3 is involved in the human repair disorders xeroderma ila, pp. 437–456 in A Survey of Molecular Cloning Vectors and Their pigmentosum and Cockayne’s syndrome. Cell 62: 777–791. Uses, edited by R. L. Rodreguez and D. T. Denhardt. Butter- Weeda, G., E. Eveno, I. Donker, W. Vermeulen, O. Chevallier- worths, Boston. Lagente et al., 1997 A mutation in the XPB/ERCC3 DNA repair Qui, H., E. Park, L. Prakash and S. Prakash, 1993 The Saccharo- transcription gene, associated with trichothiodystrophy. Am. J. myces cerevisiae DNA repair gene RAD25 is required for transcrip- Hum. Genet 60: 320–329. tion by RNA polymerase II. Genes Dev. 7: 2161–2171. Williams, L. D., and B. R. Shaw, 1987 Protonated base pairs explain Regan, C. L., and M. T. Fuller, 1988 Interacting genes that affect the ambiguous pairing properties of O6-methylguanine. Proc. microtubule function: the nc2 allele of the haywire locus fails to Natl. Acad. Sci. USA 84: 1779–1783. complement mutations in the testis-specific ␤-tubulin gene of Drosophila. Genes Dev. 2: 82–92. Communicating editor: R. S. Hawley