Expression and functional analyses of the Dxpa gene, the Drosophila homolog of the human excision repair gene XPA.

Xeroderma pigmentosum (XP) is a human hereditary disease characterized by a defect in DNA repair after exposure to ultraviolet light. Among the seven groups of XP, group A (XP-A) patients show the most severe deficiency in excision repair and a wide variety of cutaneous and neurological disorders. We have cloned homologs of the human XPA gene from chicken, Xenopus, and Drosophila, and sequence analysis revealed that these genes are highly conserved throughout evolution. Here, we report characterization of the Drosophila homolog of the human XPA gene (Dxpa). The Dxpa gene product shows DNA repair activities in an in vitro repair system, and Dxpa cDNA has been shown to complement a mutant allele of human XP-A cells by transfection. Polytene chromosome in situ hybridization mapped Dxpa to 3F6-8 on the X chromosome, where no mutant defective in excision repair was reported. Northern blot analysis showed that the gene is continuously expressed in all stages of fly development. Interestingly, the Dxpa protein is strongly expressed in the central nervous system and muscles as revealed by immunohistochemical analysis using anti-Dxpa antibodies, consistent with the results obtained in transgenic flies expressing a Dxpa-β-galactosidase fusion gene driven by the Dxpa promoter.

Xeroderma pigmentosum (XP) is a human hereditary disease characterized by a defect in DNA repair after exposure to ultraviolet light. Among the seven groups of XP, group A (XP-A) patients show the most severe deficiency in excision repair and a wide variety of cutaneous and neurological disorders. We have cloned homologs of the human XPA gene from chicken, Xenopus, and Drosophila, and sequence analysis revealed that these genes are highly conserved throughout evolution. Here, we report characterization of the Drosophila homolog of the human XPA gene (Dxpa). The Dxpa gene product shows DNA repair activities in an in vitro repair system, and Dxpa cDNA has been shown to complement a mutant allele of human XP-A cells by transfection. Polytene chromosome in situ hybridization mapped Dxpa to 3F6 -8 on the X chromosome, where no mutant defective in excision repair was reported. Northern blot analysis showed that the gene is continuously expressed in all stages of fly development. Interestingly, the Dxpa protein is strongly expressed in the central nervous system and muscles as revealed by immunohistochemical analysis using anti-Dxpa antibodies, consistent with the results obtained in transgenic flies expressing a Dxpa-␤galactosidase fusion gene driven by the Dxpa promoter.
Xeroderma pigmentosum (XP) 1 is a human autosomal recessive disease characterized by hypersensitivity to UV light and a high incidence of skin cancer because of a defect in nucleotide excision repair (NER) (Cleaver and Kraemer, 1989). Complementation analyses by cell fusion between cells from XP pa-tients have revealed seven complementation groups (A through G) and a variant form, representing different components that are thought to work together in NER. Five of the relevant genes (complementing XP-A, -B, -C, -D, and -G mutations) have been already identified (Tanaka et al., 1990;Weeda et al., 1990;Legerski and Peterson, 1992;Masutani et al., 1994;Flejter et al., 1992;O'Donovan and Wood, 1993;Scherly et al., 1993). Some of these have been shown to correspond to excision repair cross-complementing rodent repair deficiency (ERCC) genes, which are human genes that correct NER defects of a set of UV-sensitive rodent cell lines comprising 11 different complementation groups; the XPB gene is equivalent to the ERCC3 gene (Weeda et al., 1990), XPD to ERCC2 (Flejter et al., 1992), and XPG to ERCC5 (O'Donovan and . Homologs of the XP and ERCC genes have been identified in many organisms, most notably in Saccharomyces cerevisiae. A number of yeast radiation-sensitive mutants (rad) have been isolated, and some of the RAD genes are known to be the homologs; XPA, XPB, XPC, XPD, XPG, and ERCC1 share extensive sequence homology at the protein level with RAD14 (Bankmann et al., 1992), RAD25/SSL2 (Park et al., 1992;Gulyas and Donahue, 1992), RAD4 (Legerski and Peterson, 1992;Masutani et al., 1994), RAD3 (Weber et al., 1990), RAD2 (Scherly et al., 1993), and RAD10 (van Duin et al., 1986), respectively. Therefore, the basic features of the NER mechanism are likely to be conserved among eukaryotes. Indeed, expression of the XPD gene in S. cerevisiae was reported to complement the lethality of a mutation in the RAD3 gene (Sung et al., 1993).
XP patients manifest a wide variety of symptoms, the most characteristic of which is hypersensitivity to UV light and certain chemical mutagens. XP patients have a 2000-fold increased frequency of skin cancer upon UV exposure compared with the general population (Cleaver and Kraemer, 1989). Among the seven complementation groups, XP-A is the most severe form, and these patients exhibit a wide variety of neurologic abnormalities including microcephaly, progressive mental deterioration, ataxia, abnormal reflexes, and sensory deafness (Cleaver and Kraemer, 1989). The human XPA gene encodes a protein of 273 amino acids with a zinc-finger motif (Tanaka et al., 1990). Replacement of each of the 4 cysteine residues of the zinc-finger structure by serine or glycine resulted in loss of repair activity, indicating the functional importance of the motif (Miyamoto et al., 1992). The XPA protein binds to DNA with a preference for UV-irradiated over unirradiated DNA, suggesting that XPA functions as a key component in recognition of DNA damage during repair (Robins et al., 1991;Jones and Wood, 1993;Asahina et al., 1994).
Recently, the haywire gene of Drosophila was found to encode a protein with 66% identity to that encoded by ERCC3, the human gene associated with XP-B (Mounkes et al., 1992;Koken et al., 1992). Flies compromised for haywire function display phenotypes including UV sensitivity, central nervous system (CNS) defects, ataxia, and lethality, suggesting that a Drosophila mutant defective in the function of a human XP homolog could be useful as an animal model of XP. Sophisticated genetic tools available in Drosophila should help in investigating the functions of XP genes.
Previously, we cloned a Drosophila homolog of the human XPA gene (Dxpa) (Shimamoto et al., 1991). The Dxpa cDNA encodes a protein of 296 amino acids with 45% identity to that encoded by the human XPA gene. In this report, we show that the Dxpa protein is involved in NER both in vitro and in vivo. We also reveal that the Dxpa gene is expressed strongly in the CNS of Drosophila and determine the localization of the Dxpa gene to 3F6 -8 on the X chromosome, where no mutant defective in NER has yet been reported.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture Conditions-The cell lines Kc and C10 were derived from embryos and neurons, respectively, of wild-type Drosophila melanogaster. The cell line mei-9 was derived from a Drosophila mutant defective in both meiosis (Baker and Carpenter, 1972) and NER (Boyd et al., 1976). Kc and C10, and mei-9 cell lines were kind gifts from Drs. Y. Miyake and R. Ueda (Mitsubishi Kagaku Institute of Life Science), respectively, and were maintained in M3 (BF) medium (Cross and Sang, 1978) supplemented with heat-inactivated 10% fetal calf serum. For culture of C10 cells, the medium was also supplemented with 10 g/ml human insulin. WI38VA13 and XP12ROSV were derived from normal human fibroblasts and those from an XP-A patient, respectively, and were cultured in ␣-modified Eagle's medium supplemented with 10% fetal calf serum.
Production and Affinity Purification of Anti-Dxpa Protein Antibodies-The Dxpa coding sequence was amplified from the pUC Dxpa cloning vector by polymerase chain reaction using a 5Ј-forward primer containing a BamHI site (5Ј-GGTCGGATCCAGATGTCTGCGGAGGT-CTCTACCAACGAA-3Ј) and a 3Ј-reverse primer containing a BglII site (5Ј-TGACAGATCTCACTACATCTTCTCGTAGGTCTCGCTGTACG-GA-3Ј). This facilitated its cloning into the bacterial expression vector pET-3c via the BamHI site. The recombinant protein starts with 13 amino acids (MASMTGGQQMGRI) derived from the vector, followed by the entire sequence of the Dxpa protein. The recombinant Dxpa protein was purified as described previously (Miura et al., 1991;Asahina et al. 1994). Rabbits were immunized with the protein, and the anti-Dxpa antibodies were then purified by affinity chromatography using CNBractivated Sepharose 4B (Pharmacia Biotech Inc.) coupled with the recombinant Dxpa protein.
Immunoprecipitation of in Vitro Translated Products-pGEM Dxpa was generated by inserting the Dxpa cDNA HindIII-EcoRI fragment containing the entire coding sequence into pGEM-7Zf(Ϫ). The plasmid pET-XPAH, which contains the full-length human XPA cDNA under the control of the T7 promoter, was kindly provided by Dr. H. Asahina (Kobe University) (Miura et al., 1991). These plasmids were cut with EcoRI, and in vitro transcription-translation reactions were carried out using SP6 RNA polymerase or T7 RNA polymerase and then rabbit reticulocyte lysate (Amersham Corp.) in the presence of [ 35 S]methionine. Ten-l aliquots of each lysate were incubated with 20 g/ml affinity-purified rabbit anti-Dxpa antibodies or control rabbit IgG in 400 l of lysis buffer (50 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 0.1% Nonidet P-40, and 0.25% gelatin) for 30 min at 4°C, followed by incubation with 40 l of protein A-Sepharose 6MB (Pharmacia Biochem Inc.) for 1 h at 4°C. Immune complexes were collected by centrifugation at 10,000 ϫ g for 15 s. The pellets were washed five times with lysis buffer and suspended in sample buffer (20 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.01% bromphenol blue, and 0.1 M dithiothreitol). After heating at 100°C for 2 min, samples were loaded onto 12.5% SDS-polyacrylamide gels, electrophoresed, and autoradiographed.
Immunoblotting-Cells (1 ϫ 10 6 ) were lysed in 50 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, and 2 mM EDTA containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride and 2 g/ml aprotinin). After centrifugation at 10,000 ϫ g for 10 min at 4°C, the supernatants were collected, and 20 g of each sample was mixed with 15 l of sample buffer, boiled for 3 min, and separated as described above. The blots were probed by enhanced chemiluminescence (ECL, Amersham Corp.) using a 1:2000 dilution of rabbit anti-Dxpa serum as a primary antibody and a 1:20,000 dilution of horseradish peroxidase-conjugated anti-rabbit IgG as a second antibody (Promega).
Preparation of Whole-cell Extract-The Kc cell line was grown at 25°C in 150-mm tissue culture plates (Falcon) and collected by gentle pipetting with phosphate-buffered saline (PBS). The harvested cells were washed with PBS, and whole-cell extract was prepared as described previously (Manley et al., 1980;Wood et al., 1988). Protein concentration was determined by the method of Bradford (1976) with bovine serum albumin as a standard. The extract contained 15 mg of protein/ml.
Cell-free DNA Repair Assay-The assay was carried out as described previously . The standard reaction mixture (20 l) contained 40 mM creatine phosphate/Tris, pH 7.7, 1 mM dithiothreitol, 10 mM MgCl 2 , 2 mM ATP, 50 M dATP, 50 M dGTP, 50 M dTTP, 10 M (1 Ci) [␣-32 P]dCTP, 0.5 g of phosphocreatine kinase (Type I, Sigma), 6.4 g of bovine serum albumin, whole-cell extract (80 g of protein), 0.3 g of unirradiated pUC19 replicative form I DNA, and UV-irradiated (200 J/m 2 ) or unirradiated SV40 minichromosomes (0.3 g of viral DNA). After incubation at 30°C for 3 h, the reaction was terminated by the addition of 1 l of 0.5 M EDTA and cooling to 0°C. Then, 1 l of 1 mg/ml bovine pancreatic ribonuclease A was added, and the mixture was incubated at 37°C for 30 min. The products were further treated with a solution of 200 g/ml proteinase K containing 0.5% SDS. After incubation at 37°C for 1 h, the products were treated with phenol/chloroform/isoamyl alcohol (25:24:1) and then precipitated with ethanol. The precipitates were solubilized, digested with EcoRI, and loaded onto 1% agarose gels. After electrophoresis at 20 -30 V for 12-18 h, the gels were dried and autoradiographed at Ϫ80°C using Fuji New RX x-ray film.
Transfection Experiments and UV Survival-For transfection, the Dxpa cDNA fragment in which the HindIII site was replaced with an EcoRI site by linker ligation was inserted into the EcoRI site of pCAGGSneo, constructed from the mammalian expression vector pCAGGS (Niwa et al., 1991) by introducing the neomycin resistance gene MC1neo into the SalI site. Transfection of this vector into the XP-A cell line XP12ROSV was performed by the method of Chen and Okayama (1987). Transfected cells were selected in the presence of the neomycin analog G418 (250 g/ml). Stable transfectants were cloned by limiting dilution, and expression of the Dxpa protein was checked by immunoblotting.
For UV survival experiments, cells were plated at densities varying from 8 ϫ 10 2 to 1.6 ϫ 10 6 cells/100-mm Petri dish, depending on the cell line and UV dose used. Cells were rinsed with PBS and exposed to UV light approximately 1 day after plating. A series of dishes was irradiated for each cell line, receiving a single UV dose (three dishes/UV dose). Colonies were fixed and stained with 0.1% crystal violet approximately 10 days after UV irradiation, and percent colony-forming abilities were determined by comparing the colony counts of the irradiated plates with those of unirradiated control plates.
Chromosomal in Situ Hybridization-Salivary gland polytene chromosomes from late third instar larvae of Canton S strain flies were prepared as described by Zucker et al. (1985). The probe was prepared by labeling Dxpa cDNA with digoxygenin-dUTP, and hybridization was detected using an anti-digoxygenin monoclonal antibody coupled to alkaline phosphatase (Boehringer Mannheim).
Cloning of Genomic Dxpa-A EMBL3 Drosophila genomic library (a gift from Dr. Y. Nishida) was screened with the Dxpa cDNA HindIII-EcoRI fragment as a probe according to standard methods (Sambrook et al., 1989). Several clones were picked up and checked by restriction enzyme digestion followed by electrophoresis. A fragment consisting of 1.3 kb of 5Ј-flanking region, 1.2 kb of transcribed region, and 1.5 kb of 3Ј-flanking region was sequenced by the chain termination method (Sanger et al., 1977) using a Sequenase kit (United States Biochemical Corp.).
Chimeric Gene Construction and Germ Line Transformation-The SalI-HaeII fragment of genomic Dxpa consists of 1.3 kb of Dxpa gene promoter, the first exon, a 65-bp intron, and 33 bp of the second exon (see Fig. 6B). This fragment was introduced into pMC1871, a vector for constructing a fusion protein with Escherichia coli ␤-galactosidase. The correct reading frame was confirmed by DNA sequencing of the fusion junction. A 4.4-kb SalI-SalI fragment containing the fusion gene construct was then subcloned into the P-element transformation vector Carnegie 20 (Rubin and Spradling, 1983) at the SalI site (see Fig. 6C).
Germ line transformation was carried out essentially by the method of Rubin and Spradling (1982). The recipient Drosophila strain was rosy 506 /Ki p p P[D2-3, ry ϩ ], and the DNA concentration for injection was 1 mg/ml of the Carnegie 20 construct. The embryos for injection were collected, dechorionated, and aligned to the edge of a coverslip. Following injection, embryos were placed in a moist chamber and incubated at 20°C until hatching. Surviving larvae were transferred to standard food vials and allowed to develop into adults (G 0 ), which were then backcrossed to rosy 506 flies. G 1 progeny with the wild-type eye color were crossed again to rosy 506 flies, and the resulting wild-type G 2 progeny were mated to each other to establish homozygous lines. Southern blotting of genomic DNA was performed to confirm the presence of the inserted DNA in transformed lines.
X-Gal Staining and Immunohistochemistry-The spatial distribution of ␤-galactosidase expression was determined by X-gal staining as described previously (Fortini and Rubin, 1990). Flies were frozen in Tissue Tek O.C.T. compound (Miles Inc.), cut into sections 10 m thick on a cryostat, and thaw-mounted onto glass slides pretreated with VECTABOND TM (Vector Laboratories, Inc.). After air-drying for 30 min, sections were fixed in 1% glutaraldehyde in PBS for 15 min at room temperature, followed by three washes in PBS for 10 min each time. The sections were submerged in X-gal staining buffer (0.1 M sodium phosphate buffer, pH 7.2, 20 mM potassium ferrocyanide, 20 mM potassium ferricyanide, 0.02% X-gal, and 1 mM MgCl 2 ) at 37°C for 16 h and then washed three times in PBS for 10 min each time. They were finally mounted with 80% glycerol in PBS and examined under a light microscope.
Immunohistochemical staining was performed using a Vectastain Elite ABC kit (Vector Laboratories, Inc.). Sections 10 m thick were fixed with 2% paraformaldehyde in PBS and washed in PBS. After preincubation with PBSG buffer (PBS, 1% normal goat serum, 0.01% Triton X-100, and 0.2% bovine serum albumin) at room temperature for 30 min, they were incubated with affinity-purified rabbit anti-Dxpa antibodies or control rabbit IgG in PBSG buffer at room temperature for 1 h and washed with washing buffer (PBS and 0.01% Triton X-100). The sections were incubated with biotinylated goat anti-rabbit IgG for 30 min at room temperature and washed, followed by incubation with ABC complex for 30 min at room temperature. After washing, the sections were submerged in peroxidase substrate solution (0.1 M Tris-HCl, pH 7.2, 0.5 mg/ml diaminobenzidine, and 0.02% H 2 O 2 ) for 2-5 min, washed again, and mounted. Immunohistochemistry of whole embryos was performed as described above except for the procedures of fixation and permeabilization; dechorionated embryos were fixed in 0.1 M PIPES, 2 mM EGTA, 1 mM MgSO 4 , pH 6.9, 3.7% formaldehyde solution saturated with n-heptane, and the concentrations of Triton X-100 were raised to 0.1% in PBSG buffer and 0.3% in washing buffer.

Identification of Dxpa Protein
Expression-To characterize the Dxpa protein expression pattern, rabbit polyclonal antibodies were raised against a fusion protein comprising 13 amino acids derived from the bacterial expression vector pET-3c and the full-length Dxpa protein (see "Experimental Procedures" ;  Fig. 1A). The affinity-purified anti-Dxpa antibodies specifically immunoprecipitated two proteins with molecular masses of 43 and 40 kDa in in vitro transcription-translation products of the cloned Dxpa cDNA, while control rabbit IgG did not (Fig. 1B,  lanes 1 and 2). These two proteins derived from Dxpa cDNA are thought to correspond to human XPA proteins that migrated with apparent molecular masses of 40 and 38 kDa (Miura et al., 1991). The anti-Dxpa antibodies also cross-reacted with in vitro translated products of the cloned human XPA cDNA (Fig. 1B,  lane 3), indicating that both Dxpa and human XPA proteins have common epitopes recognized by anti-Dxpa antibodies. To examine the expression of the Dxpa proteins in vivo before and after UV irradiation, immunoblotting analysis was carried out using total lysates from the Drosophila cell lines Kc, C10, and mei-9. Kc and C10 are repair-proficient wild-type cell lines, while the mei-9 cell line is derived from a mutant defective in NER. The antibodies recognized two proteins of the same sizes as those of the in vitro translated Dxpa proteins in all three cell lines (Fig. 1C). The amounts of Dxpa proteins were not increased by UV irradiation. This result coincides with the observed lack of UV inducibility of the human XPA protein (Miura et al., 1991).

Involvement of Dxpa Protein in DNA Repair in Drosophila-
The in vitro DNA repair system was originally developed on the basis of human cell extracts Wood et al., 1988). In the cell-free system, repair synthesis is detected in SV40 minichromosomes that have been irradiated with UV light. Since the SV40 minichromosome forms a eukaryotic chromatin structure, this in vitro repair system is thought to reflect in vivo repair conditions faithfully. As an internal control, unirradiated pUC19 plasmids were mixed with the irradiated DNA. A mixture of the two closed circular molecules, SV40 minichromosomes and pUC19, was incubated with whole-cell extract in buffer containing deoxyribonucleotides, [␣-32 P]dCTP, ATP, and an ATP-regenerating system. During the incubation, some of the pyrimidine dimer photoproducts are removed from irradiated DNA by NER. DNA repair synthesis can be visualized after recovery of the DNAs from the reaction mixture by linearization of the DNAs with a restriction enzyme, gel electrophoresis, and autoradiography.
To investigate whether this system would also work for Drosophila, a whole-cell extract was prepared from growing Kc cells by   1, 4, and 7), 4 h (lanes 2, 5, and 8), and 8 h (lanes 3, 6, and 9), and immunoblotting using affinity-purified rabbit anti-Dxpa antibodies was carried out as described under "Experimental Procedures." (1988)). Fig. 2A shows that Drosophila Kc whole-cell extracts supported specific incorporation of deoxyribonucleotides into UV-irradiated SV40 minichromosomes. Little nonspecific incorporation was detected in unirradiated SV40 minichromosomes and pUC19, indicating that this system could detect Drosophila NER in vitro ( Fig. 2A, lanes 1 and 2). When anti-Dxpa antibodies were added to the reaction mixtures, DNA repair synthesis with UV-irradiated SV40 minichromosomes was inhibited in a dose-dependent manner ( Fig. 2A, lanes  5-12), whereas control rabbit IgG had no effect (lanes 3 and 4). Addition of the recombinant Dxpa protein to the reactions after mixing the anti-Dxpa antibodies with extracts restored the DNA repair synthesis inhibited by the antibodies in a dose-dependent manner (Fig. 2B). These findings are consistent with results obtained in experiments using the human XPA protein , verifying that the Dxpa protein is involved in DNA repair synthesis in Drosophila.
Interspecies Functional Complementation of Dxpa Gene in Human XP-A Cells-When human XPA cDNA under the constitutive ␤-actin promoter was expressed in a repair-deficient human cell line, XP12ROSV, which is known not to express the XPA protein, transfectants restored the UV resistance completely (Miyamoto et al., 1992). To confirm functional conservation between Drosophila and human XPA proteins, we introduced a Dxpa expression plasmid into XP12ROSV cells. Of stable transfectants, three independent clones strongly expressing the Dxpa protein were selected by immunoblotting (Fig.  3A). The human XPA protein was not detected in the human repair-proficient cell line WI38VA13 with anti-Dxpa antibodies, although it was recognized by immunoprecipitation (Figs. 1B and  3A, lane 1). Expression of the Dxpa protein in the three transfectants resulted in the restoration of partial UV resistance (Fig.  3B). This result shows that Drosophila Dxpa partially comple-  3,5,7,9,and 11) or unirradiated (lanes 2,4,6,8,10,and 12) SV40 minichromosomes were incubated in standard reaction mixtures, which were preincubated in the presence of 1 g of control rabbit IgG (lanes 3 and 4) or 1 ng (lanes 5 and 6), 10 ng (lanes 7 and 8), 100 ng (lanes 9 and 10), and 1000 ng (lanes 11 and 12) of anti-Dxpa antibodies for 30 min on ice. Plasmids were linearized and resolved by 1% agarose gel electrophoresis as described under "Experimental Procedures." Upper panel, autoradiogram; lower panel, ethidium bromide staining of the gel. B, restoration by recombinant Dxpa protein of the repair synthesis inhibited by anti-Dxpa antibodies. One ng (lanes 9 and 10), 10 ng (lanes 11 and 12), 100 ng (lanes 13 and  14), or 1000 ng (lanes 5, 6, 15, and 16) of recombinant Dxpa protein was added to the standard reaction mixtures containing 1 g of control rabbit IgG (lanes 3-6) or 100 ng of anti-Dxpa antibodies (lanes 7-16), and the DNA repair reaction was performed in the mixtures in the presence of UV-irradiated (lanes 1, 3, 5, 7, 9, 11, 13, and 15) or unirradiated (lanes 2, 4, 6, 8, 10, 12, 14, and 16)  ments the defect of excision repair in human XP-A cells, indicating that Drosophila Dxpa and human XPA proteins are functionally conserved in evolution (see "Discussion").
Dxpa Is a Novel Excision Repair Gene in Drosophila-We determined the Dxpa gene locus by in situ hybridization on polytene chromosomes using Dxpa cDNA as a probe (Fig. 4). Dxpa was mapped at region 3F6 -8 on the X chromosome, where neither mutagen-sensitive nor DNA repair-deficient mutants have yet been localized. The mei-9 mutant defective in NER and meiosis was mapped previously within 4B on the X chromosome (Yamamoto et al., 1990). These results indicate that the Dxpa gene is a novel excision repair gene in Drosophila.
Expression of Dxpa mRNA during Fly Development-To investigate Dxpa mRNA expression during development, Northern hybridization analysis was carried out using the Dxpa cDNA as a probe. Fig. 5 shows a single band of 1.4 kb detected on Northern blots prepared with total RNA from embryos, third instar larvae, early pupae, middle pupae, and adults. The level of Dxpa mRNA was changed during development. We compared the relative amount of Dxpa mRNA with the internal control ribosomal protein 49 (rp49) mRNA by densitometric measurement (data not shown); the expression was the lowest at the embryonic stage, increased as development proceeded (ϳ2fold (at larval stage) or ϳ4-fold (at pupal stage) as compared with the embryonic stage), and finally reached the maximal level in adults (ϳ5-6-fold). The abundance of Dxpa mRNA appeared to be lower in the head than thorax and abdomen fractions.
Spatial and Temporal Expression of Dxpa Gene in Drosophila-Northern blot analysis suggested spatial and temporal variations in Dxpa gene expression during development. For further analysis of its expression, we generated transgenic flies expressing the E. coli lacZ marker gene under control of the Dxpa promoter. Genomic Dxpa DNA was cloned using Dxpa cDNA as a probe and sequenced by the dideoxy chain termination method (Sanger et al., 1977) (Fig. 6, A and B). A 1.5-kb SalI-HaeII fragment consisting of a 5Ј-flanking region, the first exon, the intron, and 33 bp of the second exon was fused in frame to the E. coli lacZ gene, and this fusion gene was inserted into the Carnegie 20 vector. By screening for rosy ϩ descendants among embryos injected with this vector and sib mating, three independent transformed lines were generated. Flies from these lines were assayed for ␤-galactosidase expression using the chromogenic substrate X-gal.
The expression pattern of the fusion protein in adult flies is shown in Fig. 7 (A-C). The fusion protein was strongly expressed in the nuclei of cells in the lamina and retina of the eyes and in the muscles of the whole body. The staining pattern was the same in all three transformants. The nuclear location was attributed to the presence of a putative nuclear localiza-tion signal in the first exon of the Dxpa gene (Shimamoto et al., 1991;Miyamoto et al., 1992). The signal is conserved evolutionarily, and the functioning of the signal of the human XPA protein has already been verified (Miyamoto et al., 1992). During embryogenesis, the fusion protein was expressed in all cells (data not shown), consistent with its universal function involved in NER. Fig. 7 (D-F) shows immunohistochemical staining of adult flies using affinity-purified anti-Dxpa antibodies. The staining pattern was almost the same as that in the transformants. Strong expression of the Dxpa protein was detected in the nuclei of cells in the CNS and the muscles, although all cells showed different levels of expression. No such staining was seen when control rabbit IgG was used (data not shown), indicating the specificity of the staining by the anti-Dxpa antibodies. Comparison of the staining of transformants using X-gal with that of wild-type flies by the anti-Dxpa antibodies showed that the latter was more sensitive than the former. Therefore, the X-gal staining of the transformants only reflected the strong expression of the Dxpa protein. Interestingly, the ventral nerve cord of embryos was strongly stained with the anti-Dxpa antibodies, but not with control rabbit IgG ( Fig. 7G; data not shown).

DISCUSSION
In this report, we have shown that the Dxpa gene product is involved in NER in vitro and in vivo. For the in vitro experiment, we employed a cell-free system for excision repair reactions on UV-irradiated SV40 minichromosomes. DNA repair synthesis was specifically suppressed when the anti-Dxpa antibodies were added to the reaction mixtures containing Kc cell extracts, and the inhibition was restored by adding the recombinant Dxpa protein. These results provided direct evidence for the involvement of the Dxpa protein in NER (Fig. 2). This is the first demonstration that the in vitro system can be used in the analysis of NER in Drosophila. This in vitro system might be an efficient assay to purify novel factors such as Mei-9 and Mus 201 gene products, which have not been cloned but are known to be involved in NER in Drosophila (Yamamoto et al., 1990;Todo and Ryo, 1992).
Furthermore, we demonstrated that the Dxpa gene could improve UV survival of the human XP-A cell line XP12ROSV (Fig. 3). The functions of the human and Drosophila XPA genes are thus evolutionarily conserved. However, UV resistance of transfectants conferred by Dxpa expression has not been restored completely (Fig. 3). This incomplete correction was not due to the quantity of the expressed Dxpa protein because the Dxpa gene was expressed strongly under control of the ␤-actin promoter in the transfectants at both mRNA and protein levels compared with the human XPA gene in WI38VA13 cells (data not shown). The big difference between primary structures of human XPA and Drosophila Dxpa is the presence or absence of the E-cluster region. There is no E-cluster region in Dxpa (Shimamoto et al., 1991), and the mutant human XPA protein shows very weak repair activities in human XP-A cells (Miyamoto et al., 1992). There is a possibility that the Dxpa protein having no E-cluster might show low repair activities in human XP-A cells. Previously, it was reported that functional cross-complementation was not observed in human XP-B cells transfected with ERCC3 Dm cDNA, which is a Drosophila homolog of the human XPB/ERCC3 gene . ERCC3 Dm is identical to the haywire gene and more highly conserved during evolution than the Dxpa gene (Mounkees et al., 1992). The cause of this discrepancy is unknown, but there are several possible explanations among the experimental procedures. First, Koken et al. (1992) cotransfected XP-B cells with an ERCC3 Dm expression vector together with a separate selection marker (neomycin resistance gene)-containing vector, while we used an expression vector in which a selection marker was already inserted. Second, they assayed the resulting G418-resistant transfectants without cloning, which might include those without ERCC3 Dm expression, while we picked up independent transformants and confirmed the expression of the FIG. 7. Expression of the Dxpa gene in embryos and adults. Dxpa gene expression was monitored by localizing ␤-galactosidase activity in transformants harboring a lacZ fusion gene under control of the Dxpa promoter as described under "Experimental Procedures" (A-C), and expression of the Dxpa protein was immunohistochemically detected in wild-type Drosophila (D-G). A-C, X-gal staining of adult head sections at lower (A) and higher (B) magnification and of the adult leg (C) of transformants. ␤-Galactosidase expression was detected in the nuclei of neurons in the brain and retina in the head and muscle cells in the leg. Re, retina; R1-R7, nuclear region of distal photoreceptor cells; R8, nuclear region of proximal photoreceptor cells; La, lamina; Me, medulla; Lo, lobula and lobula plate. D-G, immunohistochemical staining of sections of the adult head and thorax (D), the adult head at higher magnification (E), the adult leg (F), and embryos (G) using affinity-purified anti-Dxpa antibodies. The Dxpa protein was expressed strongly in the CNS and muscle cells in adults and in the ventral nerve cord (VNC) in embryos.
Dxpa protein in each clone by immunoblotting. This allowed us to examine UV survival with homogeneous cell populations expressing the Dxpa protein.
Although the calculated molecular mass of the Dxpa protein was 34 kDa, it migrated with masses of 43 and 40 kDa on SDS-polyacrylamide gel electrophoresis (Fig. 1, B and C). Since the recombinant Dxpa protein synthesized in E. coli was detected as a band of ϳ43 kDa (Fig. 1A) and could restore the DNA repair synthesis inhibited by the anti-Dxpa antibodies, the products seemed to be unmodified Dxpa proteins.
Northern hybridization indicated that the Dxpa gene is expressed continuously throughout fly development (Fig. 5). This is consistent with the important role of the gene product in NER. As compared with the internal control rp49, expression of the Dxpa gene changed during development and was stronger in the thorax and abdomen than in the head. For detailed examination of the expression of the Dxpa gene, we made transgenic flies expressing a marker protein, E. coli ␤-galactosidase, fused to a partial coding region of the Dxpa gene in frame under the control of the Dxpa promoter (Fig. 6). The fusion gene product was expressed strongly in the CNS in adults (Fig. 7, A and B). Immunohistochemical staining provided similar results and furthermore clarified the intense expression of the Dxpa protein in the CNS in embryos (Fig. 7, D, E, and G). Recently, haywire was identified as a Drosophila mutant defective in the homolog of the human XPB gene, and this mutant could be used as an animal model of human XP-B disease (Mounkes et al., 1992). In fact, haywire flies manifest neurological defects that are thought to mimic the CNS defects observed in XP-B patients. In a similar way, the Dxpa protein may play an important role in the CNS. Satoh et al. (1993) reported that the DNA repair defect of XP prevented removal of a class of oxygen free radical-induced base lesions. Their data suggested that accumulation of endogenous oxidative damage in cellular DNA of XP patients contributed to the neural degeneration occurring in serious cases of the syndrome. Since the O 2 consumption in the CNS is known to be particularly high compared with other organs, it is probable that the preferential accumulation of oxidative damage occurs in the CNS. The large amount of Dxpa protein in the CNS of Drosophila might correspond to the necessity of NER in the CNS and is suggestive of a mechanism of the variety of neurological symptoms exhibited by XP-A patients.
mei-9 was originally identified as a mutant defective in meiosis. This mutant also has a defect in NER, and it was possible that its locus was identical to that of the Dxpa gene. However, we verified that they are different genes on the basis of the following data. First, genome mapping demonstrated that their locations were different; we mapped the Dxpa gene to 3F6 -8 of the X chromosome (Fig. 4), while mei-9 had been mapped to 4B. Second, we identified the Dxpa protein in mei-9 whole-cell extracts. The size of the protein was the same as that of the wild-type cell lines Kc and C10, and the amount was almost equal to that of C10 (Fig. 1C). Thus, we concluded that the Dxpa gene is different from mei-9.
To date, no mutant defective in the Dxpa gene has yet been identified. Future isolation of such mutants will facilitate further characterization of the function of the Dxpa protein in vivo in addition to providing a potentially useful animal model of human XP-A.