To analyze plant mechanisms for resistance to UV radiation, mutants of Arabidopsis that are hypersensitive to UV radiation (designated uvh and uvr) have been isolated. UVR2 and UVR3 products were previously identified as photolyases that remove UV-induced pyrimidine dimers in the presence of visible light. Plants also remove dimers in the absence of light by an as yet unidentified dark repair mechanism and uvh1 mutants are defective in this mechanism. The UVH1 locus was mapped to chromosome 5 and the position of the UVH1 gene was further delineated by Agrobacterium-mediated transformation of the uvh1-1 mutant with cosmids from this location. Cosmid NC23 complemented the UV hypersensitive phenotype and restored dimer removal in the uvh1-1 mutant. The cosmid encodes a protein similar to the S. cerevisiae RAD1 and human XPF products, components of an endonuclease that excises dimers by nucleotide excision repair (NER). The uvh1-1 mutation creates a G to A transition in intron 5 of this gene, resulting in a new 3′ splice site and introducing an in-frame termination codon. These results provide evidence that the Arabidopsis UVH1/AtRAD1 product is a subunit of a repair endonuclease. The previous discovery in Lilium longiflorum of a homolog of human ERCC1 protein that comprises the second subunit of the repair endonuclease provides additional evidence for the existence of the repair endonuclease in plants. The UVH1 gene is strongly expressed in flower tissue and also in other tissues, suggesting that the repair endonuclease is widely utilized for repair of DNA damage in plant tissues.
Plants are exposed to UV-B radiation in sunlight (wavelengths 280–320 nm) that reaches the Earth's surface. Although plants protect themselves by synthesis of UV-absorbing molecules ( Li et al. 1993 ; Stapleton & Walbot 1994), UV radiation penetrates plant tissues and damages DNA and other cellular targets such as Photosystem II and plasma membrane ATPase ( Stapleton 1992). Two major DNA lesions, cyclobutane pyrimidine dimers (CPDs) and pyrimidine (6-4) pyrimidinone dimers [(6-4) photoproduct] which make up approximately 70–80% and 20–30%, respectively, of the total UV photoproducts ( Mitchell & Nairn 1989) are produced. If not removed, these lesions block both DNA replication ( Painter 1985) and transcription ( Protic-Sabljic & Kraemer 1985).
One mechanism for removing UV-induced DNA damage in both plants and other organisms is an enzymatic reversal of dimers by photoreactivation. Photolyases bind specifically to CPDs and (6-4) photoproducts and catalyze their reversal by absorbing light in the 300–600 nm wavelength range ( Sancar 1994). Based on sequence and function, photolyases fall into several classes, some members of which are specific for CPDs and colleagues for (6-4) photoproducts ( Yasui & Eker 1998). Arabidopsis uvr2 mutants are defective in the photoreactivation of CPDs and the uvr3 mutant in the photoreactivation of (6-4) photoproducts ( Jiang et al. 1997a ; Landry et al. 1997 ). Both CPD and (6-4) photolyase genes have been cloned from Arabidopsis using degenerate PCR primers, and uvr2 and uvr3 mutants carry mutations in these genes ( Ahmad et al. 1997 ; Landry et al. 1997 ; Nakajima et al. 1998 ).
CPDs and (6-4) photoproducts are also repaired in plants ( Howland 1975; Quaite et al. 1994 ; Taylor et al. 1996 ) and other organisms in the absence of photoreactivating light by a dark repair mechanism. For example, using damage specific antibodies, the removal of (6-4) photoproducts is readily detected in Arabidopsis seedlings grown in the dark after UV irradiation ( Britt et al. 1993 ). In many organisms, both CPDS and (6-4) photoproducts are excised and repaired in the dark by nucleotide excision repair (NER), as illustrated in Fig. 1 (Sancar 1996). NER also repairs a variety of other types of DNA damage, including certain kinds of chemically damaged bases and DNA cross-links (see Huang et al. 1994 ). However, there is presently no evidence that UV damage is removed by a similar mechanism in plants.
Genetic analysis of NER-deficient yeast and mammalian cell mutants and biochemical studies have revealed that multiple proteins are required for NER, that these proteins are highly conserved structurally and functionally, and that a number of them are transcription factors ( Sancar 1996; Thompson 1998). Certain yeast rad mutants and rodent cell mutants, and the seven complementation groups of the human disease xeroderma pigmentosum (XPA-XPG) are deficient in these factors. The availability of plant mutants provides an opportunity to analyze the role of similar proteins in plants.
To position the UVH1 locus on chromosome 5, RFLP and PCR-based markers were used to analyze 776 recombinant chromosomes derived from a cross between the uvh1-1 mutant and the Landsberg strain. The resulting linkage data are given in Fig. 2(a). The UVH1 locus was found to be tightly linked to marker CR3. Using this marker as a starting point, a BAC contig covering 300 kb of the corresponding genomic region was constructed. New markers 7K13L, XPB2 and B4 within this contig were then developed and used to localize the UVH1 locus to within a 60 kb region. Two tandem copies of a DNA repair gene designated AtXPB1 ( Ribeiro et al. 1998 ; GenBank accession number U29168) and AtXPB2 (Z. Liu and D. Mount, unpublished results) were found at this location. The products of these genes are homologous to human XPB protein which is involved in NER as a component of TFIIH ( Fig. 1). However, neither of these genes was UVH1 as judged by complementation tests and DNA sequencing (data not shown).
In order to perform more detailed mapping and complementation studies in this region, a cosmid contig spanning this region was constructed. Cosmid clones were classified into nine groups ( Fig. 2a) and a representative cosmid from each group was introduced into Agrobacterium for transformation of the uvh1-1 mutant. Cosmid NC23 complemented the UV-sensitive phenotype of the uvh1-1 mutant ( Fig. 2b) and the transformed uvh1-1 plants were capable of NER ( Fig. 2d). These results reveal that the cosmid clone NC23 carries the UVH1 gene. The UVH1 locus was then located within a 10–15 kb region of cosmid NC23 that included marker 7K13L on the left end of BAC T7K13. Database similarity searches using the 7K13L sequence as a query revealed that this same region matches a newly released sequence of PAC clone MEE6 (GenBank accession number AB010072) and that this region encodes a protein (AtRAD1) that is strikingly similar to S. cerevisiae Rad1 and human XPF DNA repair proteins.
Cloning of the UVH1/AtRAD1 gene
In order to obtain the sequence of the encoded protein, a cDNA copy of the gene was identified. A 10 kb genomic fragment isolated from cosmid clone NC23 was used to probe a cDNA library made from Arabidopsis seedlings. From 1 × 106 phage clones, nine cDNA clones of different sizes were identified. The longest cDNA clone contained an open reading frame of length 2868 bp, 94 bp of untranslated 5′ leader sequence and 210 bp of 3′ sequence. An additional 500 bp of untranslated 5′ leader was amplified from UVH1 mRNA by RT-PCR indicating the presence of a long 5′ leader sequence. However, the exact transcriptional start site has not been established. In the promoter region, putative TATA and CCAAT boxes were identified 742 bp and 721 bp upstream from the ATG codon, respectively. A repeat of nine GAA sequences 656 bp upstream from the ATG codon was also found. The structure of the UVH1/AtRAD1 gene is shown in Fig. 3(a). The gene contains eight introns and nine exons. The predicted UVH1/AtRAD1 protein has a length of 956 amino acids, a molecular mass of 107.6 kDa and a pI of 8.10. The UVH1/AtRAD1 gene is present as a single copy in the Arabidopsis genome, as revealed by Southern blot analysis (data not shown).
Identification of the UVH1/AtRAD1 gene product
The sequence of the predicted UVH1 product is 37% and 29% identical in pair-wise alignments to the human XPF and S. cerevisiae Rad1 proteins, respectively, sufficient to conclude that the function of these proteins is almost certainly identical. A multiple sequence alignment of the UVH1 protein and the homologous human, S. cerevisiae, D. melanogaster and S. pombe proteins is shown in Fig. 4. Three conserved regions are present in UVH1/AtRAD1 protein: (1) a leucine-rich region at residues 1–53 possibly involved in protein–protein interactions; (2) two highly conserved regions at residues 87–436 and 629–866 that are essential for function ( Sijbers et al. 1996 ; Matsumura et al. 1998 ); and (3) a C-terminal domain from residues 867–948 that interacts with the ERCC1 component of the endonuclease ( de Laat et al. 1998 ). The poorly conserved region between residues 437 and 610 is known to harbor putative nuclear targeting sequences ( Dingwall & Laskey 1991). Sixty per cent of point mutations that inactivate XPF function are located in the corresponding region of the gene ( Fig. 4).
The uvh1-1 mutation alters pre-mRNA splicing and causes premature translational termination of UVH1 mRNA
The genomic region of 7K13L was examined for possible sequence polymorphisms between the uvh1-1 mutant and wild-type alleles. As illustrated in Fig. 2(c), digestion of 7K13L DNA amplified from the uvh1-1 mutant and the C10 parental plant by Dde I reveals a polymorphism between mutant and wild-type plants. Sequence analysis showed that a G to A transition mutation identified within the predicted Arabidopsis XPF gene in uvh1-1 mutant creates a new restriction site for Dde I. Comparison of the UVH1 genomic and cDNA sequences revealed that the G to A mutation in the uvh1-1 mutant is located at the 3′ end of intron 5, 6 bp before the 3′ splicing donor site AGGAT. Since this mutation does not change the amino acid sequence of the UVH1/AtRAD1 protein, the uvh1-1 mutant was further examined for a possible splicing defect.
Based on an analysis of consensus splice sites in Arabidopsis, the uvh1-1 mutation appears to create a new 3′ splice site ( Brown et al. 1996 ) that could be used instead of the wild-type site. To investigate this possibility further, primer pairs P1 and P2, which flank the uvh1-1 mutation, were used to amplify a 750 bp fragment from both wild-type and uvh1-1 mRNA by RT-PCR ( Fig. 3a). As shown in Fig. 3(b), the uvh1-1 fragment migrated more slowly than the wild-type fragment on 2% Metaphor agarose. Sequencing of these two bands revealed that the uvh1 fragment carries an additional 6 bp, AATTAG, that introduces an in-frame stop codon TAG ( Fig. 3c,d). Six additional uvh1-1/uvh1-1 homozygous lines that had been derived from a back-cross of the uvh1-1 mutant to the parental C10 line were also found to have the same sequence change and splicing defect. Thus, the uvh1-1 mutation creates a new 3′ splice site and this site is preferred over the wild-type splice site by the plant spliceosome.
Expression of the UVH1/AtRAD1 gene
To investigate UVH1 expression in specific plant tissues, primer pairs were designed for amplifying specific mRNAs by RT-PCR, including UVH1 mRNA, UVR2 and UVR3 photolyase mRNAs for comparison with UVH1 mRNA, and UBQ3 mRNA as a constitutive control. Under the conditions used (see Experimental procedures), the amount of RT-PCR product was proportional to the input amount of total RNA over a range of 1–50 ng ( Fig. 5a). UVH1 mRNA was detected in all tissues ( Fig. 5b) with the greatest amount in flower bud tissue and the least amount in leaf tissue. UVR2 mRNA was also most abundant in flower bud tissue but barely detectable in other tissues. A similar result for UVR2 transcription by Northern blot analysis was reported previously ( Ahmad et al. 1997 ). UVR3 mRNA was only detectable in flower bud and meristem tissues. Although transcripts of all three DNA repair genes were detectable in flower bud tissue, UVH1 gene levels were reproducibly the greatest and UVR3 levels were the least. UVH1 transcripts were the only one of the three repair gene transcripts detected in root and stem tissues. UBQ3 mRNA was equally well detected in all tissues, as shown in Fig. 5(b). Similar results for each gene were obtained when at least two other primer sets were used (data not shown).
The goal of this work was to identify the function of the Arabidopsis UVH1 DNA repair gene. Since uvh1 mutants are defective in the removal of UV damage in the dark, the UVH1 product might be expected to be involved in NER. The gene has been cloned and identified by positional mapping techniques. The location on chromosome 5 includes a putative DNA repair gene (AtRAD1) that encodes a protein with significant sequence similarity to human XPF and S. cerevisiae Rad1 proteins. AtRAD1 was identified as UVH1 based upon complementation of the UV-hypersensitive phenotype of the uvh1-1 mutant with a cosmid that carries the intact AtRAD1 repair gene. Removal of UV damage was restored in the complemented uvh1-1 mutant plants. The uvh1-1 mutation leads to erroneous mRNA splicing and truncation of the protein. We conclude that the UVH1/AtRAD1 gene is a homolog of the S. cerevisiae Rad1 and human XPF genes. The AtRAD1 gene has been independently identified and characterized ( Gallego et al. 2000 ). The gene product is predicted to be a component of a repair endonuclease that is utilized by plants to excise UV damage (CPD and (6-4) photoproducts). All organisms that carry Rad1/XPF genes depend upon them for resistance to UV radiation through removal of UV damage by NER ( Boyd et al. 1976 ; Carr et al. 1994 ; Cleaver & Kraemer 1995; Jiang et al. 1997b ; Reynolds & Friedberg 1981). We conclude that plants utilize an NER pathway for the repair of UV damage similar to that of yeasts, humans and other organisms.
In S. cerevisiae, the repair endonuclease is a complex of Rad1 and Rad10 proteins that acts in NER ( Sung et al. 1993 ) and genetic recombination ( Bardwell et al. 1994 ). The human XPF and ERCC1 proteins form a similar complex ( Park et al. 1995 ). As illustrated in Fig. 1, this complex makes an incision on the 5′ side of the damaged site and another endonuclease (human XPG and S. cerevisiae Rad2 proteins) makes a similar incision on the 3′ side of the damage. A repair endonuclease with specificity for UV-induced pyrimidine dimers has been identified in extracts of Daucus carota ( McClennan & Eastwood 1986). This activity could correspond to one of the plant repair endonucleases.
Additional plant homologs to known NER genes have been identified by sequence similarity searches ( Britt 1999). An ERCC1 homolog of the plant Lilium longiflorum and Arabidopsis homologs of the S. cerevisiae RAD23 gene ( Schultz & Quatrano 1997) and S. cerevisiae RAD25/human XPB gene family ( Ribeiro et al. 1998 ) have been identified. Two tandem copies of the RAD25/XPB homolog (Z. Liu and D. Mount, unpublished observations) were mapped approximately 60 kb from the UVH1 locus indicating a possible cluster of NER genes. Additional Arabidopsis mutants that are defective in excision of (6-4) photoproducts have been mapped to other locations ( Jiang et al. 1997b ).
A review of the known biological functions of the Rad1/XPF gene family suggests that the UVH1 gene may have additional roles in plant growth and metabolism. Transgenic mice with a knockout of the ERCC1 component of the repair endonuclease show abnormal development and premature senescence. These effects could be due to an inhibition of cell division by unrepaired DNA damage ( McWhir et al. 1993 ). Similarly, some Arabidopsis uvh and uvr mutants defective in dark repair frequently exhibit an early senescence phenotype ( Jenkins et al. 1995 ) and growth of new tissues is delayed when uvh1-1 mutant seeds are germinated (Z. Liu and D. Mount, unpublished observations). In yeasts and Drosophila, Rad1/XPF genes play a role in homologous genetic recombination ( Carpenter 1982; Gutz & Schmidt 1985; Ivanov & Haber 1995; Schiestl & Prakash 1990). XPF/ERCC1 and Rad1/Rad10 complexes function as junction-specific endonucleases that remove 3′ single strands from duplex DNA molecules ( Bardwell et al. 1994 ). Similarly, the UVH1 gene and the plant ERCC1-homolog could play a role in homologous recombination in plants.
Several observations provide evidence that NER is a significant mechanism of DNA repair in plants. First, uvh1-1 mutant plants were hypersensitive to UV radiation even when light was available for photoreactivation, thus indicating that they also depend on NER for repair of UV damage ( Harlow et al. 1994 ). Second, the UVH1 and UVR2 photoreactivating genes were both expressed most strongly in flower bud tissue, possibly in order to remove as much damage as possible and avoid germline mutations. Furthermore, the ERCC1 homolog of Lilium oblongiflorum was most highly expressed in pollen generative cells ( Xu et al. 1998 ). Third, UVH1 was the only DNA repair gene found ( Fig. 5b) to be expressed in root tissue. Since UV damage is not expected in this tissue, the UVH1 product might play a role in the repair of other types of DNA damage. Finally, Arabidopsis uvh1 mutants have increased sensitivity to ionizing radiation (IR) damage ( Harlow et al. 1994 ; Jenkins et al. 1995 ; Jiang et al. 1997b ), suggesting that the UVH1 product may by involved in the repair of DNA single-strand or double-strand breaks or of oxygen-damaged bases that are produced by ionizing radiation ( Ward 1998).
The uvh1 mutation changes the splicing pattern of intron 5 by creating a new 3′ splice site 6 bp upstream of the site used in wild-type plants. This new site is preferentially used by the spliceosome during pre-mRNA splicing and the resulting mature mRNA contains an additional 6 bp that introduces an in-frame stop codon. To the best of our knowledge, this type of splicing defect in Arabidopsis mutants has not been described previously. Most splicing mutants have a sequence change in the 3′ splice site which leads to the use of an alternative, cryptic site usually downstream in the sequence but occasionally upstream in the intron ( Brown 1996). The scanning model for 3′ splice selection in eukaryotes proposes that spliceosomes scan the primary RNA transcript starting from a branchpoint recognition site in the intron and usually select the first AG: dinucleotide as the 3′ splice site. If more than one AG: lies in close proximity, local scanning in this region selects the best 3′ splice site on the basis of sequence context ( Smith et al. 1989 ; Smith et al. 1993 ). Plant spliceosomes use a similar mechanism to select the 3′ splice site ( Goodall & Filipowicz 1989; Liu & Filipowicz 1996). A branchpoint sequence, CTCAT, similar to consensus sequences in Arabidopsis ( Brown et al. 1996 ), is found at position −33 in intron 5. In agreement with the scanning model, the new 3′ splice site created in the uvh1 mutant is physically 6 bp closer to the branchpoint site than the wild-type splice site.
Strains, growth conditions and UV irradiation treatment
Plant growth conditions, generation of C10 and uvh1 lines, and treatment of plants with UV radiation have been described previously ( Harlow et al. 1994 ; Jenkins et al. 1995 ). Plants were grown at ambient room temperature and humidity in soil (Sunshine All purpose potting Mix Plus; Fisons, Vancouver, BC, Canada) with continuous illumination from two 40-W cool white fluorescent bulbs set 35 cm above the plants (2000-lux intensity) and were watered as needed with diluted commercial fertilizer.
Genetic mapping of UVH1 gene
From a cross between the uvh1 mutant (in the C10 Columbia ecotype background) and Landsberg erecta (Ler), UV-sensitive (uvh1/uvh1) F2 plants were isolated and analyzed with PCR-based markers PhyC, LEY3, DFR, CATT0191 and RFLP markers M247 and CR3. Plant genomic DNA preparation and Southern blot analysis were performed as described previously ( Harlow et al. 1994 ). A filter of the TAMU BAC library obtained from the ABRC was used to identify clones that carry marker CR3. The generation of BAC ends was as described by Woo et al. (1994) . BAC ends longer than 1.5 kb were sequenced and used for development of new CAPS markers. Marker 7K13L (5′ ATGAGAAGGTGACTATGAAAGCAAT- GAC 3′ and 5′ CTCTTATGGCTGCTGCGTCTTCTATTCG 3′) detects a number of polymorphisms between C10 and Ler using restriction enzymes AvaI, DdeI, HinfI, NdeII or TagI. Marker XPB2 is detected by using primers (5′ GTGTCCACGGTTTTACT- GGT 3′ and 5′ ATCAATATGGGAAACGGTAATGGTCTCG 3′) and TagI digestion. RFLP probe B4 (a 12 kb fragment) was isolated by BamHI digestion of the T7L5 clone and a polymorphism between C10 and Ler is detected using Bsu36I.
Screening cosmid libraries and plant transformation
To isolate cosmid clones, inserts ranging from 85 to 100 kb were isolated from BAC clones T7L5 and T1O1, respectively, using NotI digestion, and were labeled for use as probes of cosmid libraries ( Olszewski et al. 1988 ). Cosmid DNA was purified and introduced into Agrobacterium strain AGL1 by electroporation. The transformed Agrobacterium cells were selected on LB medium containing 30 carbenicillin mg l−1 and 60 mg kanamycin l−1. Cosmid integrity was verified by restriction digestion of cosmid DNA isolated from Agrobacterium cells. Transformation of uvh1 plants was performed according to the vacuum infiltration method ( Bechtold et al. 1993 ). A previous study reported decreased T-DNA transformation in uvh1-1 mutant plants, a process that might involve the repair of strand breaks ( Sonti et al. 1995 ). However, this result has not been reproduced in two other laboratories ( Nam et al. 1998 ; Preuss et al. 1999 ) and, in the present experiments, uvh1 plants were quite readily transformed. Kanamycin- or hygromycin-resistant plants were selected and transferred to soil. UV resistant transformants were initially screened by irradiation of cosmid-transformed plants with 300 J m−2 UV-C radiation. The second generation of selfed progeny of the transformants were additionally analyzed to confirm the UV-resistant phenotype.
Analysis of the UVH1 gene structure and the uvh1-1 mutation
UVH1 DNA was amplified from the uvh1-1 mutant using primer pair (5′ TCTTCCATCTCCACACTCAACAACATCC 3′ and 5′ ATG- AGAAGGTGACTATGAAAGCAATGAC 3′). The resulting 6.7 kb product was sequenced in both directions to identify the uvh1 mutation. Primer pair P1 (5′ CTGGTGAAGAACATTTGGTAG- ACTGCTC 3′) and P2 (5′ CTCTTATGGCTGCTGCGTCTTCTATTCG 3′) was used for amplification of a cDNA fragment flanking the uvh1 mutation by RT-PCR. The amplified mutant and wild-type fragments were separated on 2% MetaPhor agarose gel in 0.5× TBE (FMC BioProducts, Rockland, ME, USA). The resulting bands were isolated and sequenced.
RT-PCR and PCR amplification
Total RNA was isolated from 4-week-old leaf, root, stem and unbolted flower bud tissues, and 2-week-old apical meristem tissue, respectively, and stored at −80°C. About 50–100 mg of tissue sample was used for RNA isolation according to directions provided in the Rneasy Plant Mini Kit (Qiagen, Valencia, CA, USA). RNA quality was examined by gel electrophoresis and RNA concentration was determined by spectrophotometry. Analysis of gene expression was performed using the Titan One Tube RT-PCR kit as directed (Boehringer Mannheim GmbH, Germany). Fifty ng total RNA and 27 cycles were used, including 50°C for 30 min for the reverse transcription step, followed by 10 cycles of 94°C for 1 min, 65°C for 1 min and 68°C for 2 min, and then followed by 17 cycles of the same series of steps plus a 5 sec increment at each elongation step starting at 2 min 5 sec. Under these conditions, the amount of PCR product using the UBQ3 primer pair was proportional to the input amount of RNA, as shown in Fig. 5(a). Unless indicated otherwise, one-third of the 25 UL reaction mixture was loaded onto the gel and the bands resolved by electrophoresis on 3% agarose. The identity of each RT-PCR product was confirmed by Southern blot analysis and sequencing. To rule out the possibility of DNA contamination, RNA was amplified without using reverse transcriptase and no products were detected. For the gel presentation in Fig. 5, the gel picture was inverted using the image processing features of CorelDraw (Corel Corporation, Ontario, Canada).
The 5′ untranslated UVH1 mRNA leader sequence was analyzed using the RT-PCR amplification. The primer pairs used for the analysis of gene expression are: (1) UVH1U1346 (5′ ATCGGT- CTGTTTGGTTATTTGC 3′) and UVH1L2214 (5′ Ccttgggagtctt- ttgtttcttc 3′) for UVH1 gene; (2) UVR2U191 (5′ CCGTCGTTT- TCAATCTGTTCG 3′) and UVR2L (5′ TCCCCGTCCATTTCTTCTTCG 3′) for UVR2/PHR1 gene (AF053365); (3) UVR3U68 (5′ ACC- TTCAAGTTATCGTCTCAATC 3′) and UVR3L638 (5′ AGGAAG- TGAGGAATAAGACATC 3′) for UVR3 gene (AB003687); and (4) UBQ3U1297 (5′CAATCTCTCCCAAAGCCTAAAG 3′) and UBQ3L2132 (5′ TCGACTCCTTTTGAATGTTGTAG 3′) for UBQ3 gene (L05363). The primer pairs used for defining the 5′ RNA leader sequence of the UVH1 gene are XFRTL2214 (5′ CCTTGG- GAGTCTTTTGTTTCTTC 3′), XFRTU451 (5′ CTCTCCCATCCTT- CACTCTCC 3′), XFRTU1088 (5′ TAAGTCCTTTGATGAGATTGT- GAC 3′) and XFRTU1631 (5′ AATGTGGCTACTGGCGTTGTTG 3′).
Assay of (6-4) photoproduct
Arabidopsis seeds were sterilized in 20% bleach (1.05% sodium hydrochlorite/0.1% triton X-100 for 20 min, washed five times with sterilized water, and germinated on vertically oriented agar medium containing half strength of B5 salt mixture. Five-day-old seedlings were irradiated with a UVB dose of 1.4 kJ m−2 (UV Stratalinker 2400). Irradiated seedlings were incubated in the dark and samples were harvested at 0 and 24 h after treatment. DNA was isolated from seedling extracts by the cetyltrimethyl ammonium bromide method ( Ausubel et al. 1992 ). DNA concentration was measured by spectrophotometry and confirmed by observations of band strength following gel electropheresis. The (6-4) photoproduct was quantified by a lesion-specific radiommunoassay ( Mitchell 1996).
Sequence alignments were performed using the local alignment program LALIGN with default scoring matrix and gap penalties, from ftp.virginia.edu/fasta. PRSS ( Pearson & Miller 1992) from the Fasta package was used to show that all alignments were significant (E < 0.05). SSEARCH also from the Fasta package and BLAST2 ( Altschul et al. 1997 ) were used for database similarity searches. The multiple sequence alignment shown in Fig. 4 was obtained using the Clustal W program with default parameters ( Higgins et al. 1996 ) and the editor GeneDoc ( http://http://www.cris.com/~ketchup/genedoc.shtml).
We are grateful to Frans Tax and Jennifer Hall for helpful comments on the manuscript. We also thank M. Walk for kindly providing marker CR3; G. Picard for marker CATT0191; N. Olszewski for the cosmid library; and the Arabidopsis Biological Resource Center (ABRC) for providing marker M247 (CD1-50), Landsberg erecta (CS20), cosmid (CD4-11) and cDNA (CD4-14, CD4-15, CD4-16) libraries, TAMU BAC filters (CD4-23F) and individual BAC clones. M.I. was supported by a CONACYT and CAIO fellowship from Mexico. This work was supported by grant MCB-9728125 from the National Science Foundation to D.W.M.
The paper AtRAD1, a plant homologue of human and yeast nucleotide excision repair endonucleases, is involved in dark repair of UV damages and recombination by Francesca Gallego, Oliver Fleck, Anatoliy Li, Joanna Wyrzykowska and Bruno Tinland, which is based on independent research, also reports the identification of Arabidopsis thaliana AtRAD1 and is published on pages 507–518 in this issue of The Plant Journal.