The participation of AtXPB1, the XPB/RAD25 homologue gene from Arabidopsis thaliana, in DNA repair and plant development

Authors


For correspondence (fax +55 11 3818 7354; e-mail: cfmmenck@usp.br).

Summary

Nucleotide excision repair in Arabidopsis thaliana differs from other eukaryotes as it contains two paralogous copies of the corresponding XPB/RAD25 gene. In this work, the functional characterization of one copy, AtXPB1, is presented. The plant gene was able to partially complement the UV sensitivity of a yeast rad25 mutant strain, thus confirming its involvement in nucleotide excision repair. The biological role of AtXPB1 protein in A. thaliana was further ascertained by obtaining a homozygous mutant plant containing the AtXPB1 genomic sequence interrupted by a T-DNA insertion. The 3′ end of the mutant gene is disrupted, generating the expression of a truncated mRNA molecule. Despite the normal morphology, the mutant plants presented developmental delay, lower seed viability and a loss of germination synchrony. These plants also manifested increased sensitivity to continuous exposure to the alkylating agent MMS, thus suggesting inefficient DNA damage removal. These results indicate that, although the duplication seems to be recent, the features described for the mutant plant imply some functional or timing expression divergence between the paralogous AtXPB genes. The AtXPB1 protein function in nucleotide excision repair is probably required for the removal of lesions during seed storage, germination and early plant development.

Introduction

The genetic material of living cells is constantly suffering spontaneous or induced damage. These damages are critical for the normal functioning of the DNA molecule, inhibiting essential cellular processes, such as DNA transcription and replication, and eventually yielding mutagenesis or leading up to death of the cell. During evolution, organisms have acquired efficient mechanisms that recognize and repair these DNA lesions. The studies of DNA repair mechanisms have focused largely on bacteria, yeast and human beings, but relatively little is known about DNA repair systems in plants (Britt, 1999). These organisms are continuously exposed to environmental damaging agents due to their sessile lifestyle. Sunlight, for example, has a component of ultraviolet (UV) radiation, which, apart from causing alterations in physiological processes, growth and development (Jansen et al., 1998;Veleminsky and Angelis, 1990; Vonarx et al., 1998), also induces DNA lesions such as cyclobutane pyrimidine dimer (CPD) and pyrimidine (6–4) pyrimidinone dimer [(6–4) photoproduct]. The impact of these DNA damages in plants may be deleterious to them, thus adversely affecting agronomically important crops. This is well illustrated by the recent findings of DNA rearrangement induction in the model Arabidopsis and tobacco plants by UVB irradiation, at doses that may be ecologically relevant (Ries et al., 2000).

The removal of UV induced DNA lesions in plants seems to be a co-ordinated action of the two main mechanisms, light and dark repair (Quaite et al., 1994). It was demonstrated that in low frequencies of damage, the light dependent repair pathway is mainly recruited. This pathway, known as photoreactivation, reverts the damaged DNA to a normal configuration through the action of an enzyme called photolyase (Yasui and Eker, 1998). Studies of photoreactivation in Arabidopsis revealed the existence of two active photolyases, specific for CPDs or (6–4) photoproducts, respectively (Jiang et al., 1997). The light independent pathway, involved in UV lesion removal, is recruited only in the presence of high doses of lesions. This pathway, denominated nucleotide excision repair (NER), is more general and flexible than photoreactivation because it is able to remove a large spectrum of structurally unrelated lesions, but with the common characteristic of causing considerable distortion in the double helix. The basic principle of NER action consists of the removal of a DNA fragment containing the lesion through double incision of the damaged strand. The resulting gap is filled using the intact strand as a template (de Laat et al., 1999).

Studies of NER action in plants are very incipient when compared with those of yeast and human cells. The slow removal of CPD in dark conditions was reported in several different plants, including Arabidopsis (Pang and Hays, 1991). The identification of those UV-sensitive mutants related to this repair in Arabidopsis, through classical genetics, confirms that NER may be an important pathway helping plant cells to cope with DNA damage (Jenkins et al., 1995; Jiang et al., 1997). More recently, based on the high conservation of DNA repair pathways in eukaryotes, the search for DNA sequences similar to those in yeast and mammalian cells has led to the cloning of many plant genes, which code for proteins involved in NER. A DNA ligase I homologue, capable of complementing the corresponding mutant Saccharomyces cerevisiae cdc9, was isolated from Arabidopsis (Taylor et al., 1998). In Arabidopsis, the human homologue of XPF NER endonuclease was cloned (Gallego et al., 2000; Liu et al., 2000) and its associated protein, the human homologue ERCC1, was identified in Lilium longiflorum (Xu et al., 1998). More recently, a mutant plant at the AtRAD2/XPG gene was described as sensitive to UV-light, H2O2 and ionizing radiation (Liu et al., 2001). In carrots, as is found in human cells, two homologues for hHR23, involved in lesion recognition, were identified (Sturm and Lienhard, 1998). The recent publication of the complete DNA sequence of Arabidopsis genome confirms these previous observations, and most of the NER involved genes are found in this plant, indicating that this repair pathway is conserved, besides keeping most of the same components, and, probably, functions (The Arabidopsis Genome Initiative, 2000).

The homologue for human XPB NER helicase was also identified in Arabidopsis thaliana by sequence similarity (Ribeiro et al., 1998). The XPB protein is a component of the transcription factor TFII-H, and apart from its action in NER, it shows an essential activity in RNA transcription (Coin et al., 1999). In human beings, a defect in this protein can result in clinical heterogeneity, associated with at least three human syndromes: xeroderma pigmentosum (XP), Cockayne's syndrome (CS) and trichothyodistrophy (TTD), which share the common feature of sensitivity to sunlight (de Boer and Hoeijmakers, 1999). The identification of the plant gene revealed that the putative protein is approximately 50% identical and 70% similar to the corresponding yeast and human homologues. However, compilation of data available from the Arabidopsis Sequence Genome Project (The Arabidopsis Genome Initiative, 2000) disclosed the presence of two gene copies for XPB in Arabidopsis. In fact, an analysis of the Arabidopsis genome indicated that more than 50% of its DNA sequence is duplicated, possibly due to a polyploidy event, several million years ago (The Arabidopsis Genome Initiative, 2000). The duplicated genes are very similar, but in some cases evolution may lead to divergent functions. In the case of AtXPB genes, named AtXPB1 and AtXPB2, they are arranged as a tandem head to tail duplication, and the putative proteins keep high identity (95%), with most of the divergent aminoacid restricted to their carboxyl termini. This high similarity and genomic positions indicate that AtXPB duplication is, in fact, very recent and not due to chromosome polyploidization.

In this work, one of the XPB genes from A. thaliana, AtXPB1 (Ribeiro et al., 1998) was investigated in order to assign its role in the plant, and to search for possible divergent functional features with its duplicated copy. The role of AtXPB1 in NER and RNA polymerase II transcription was analyzed through complementation assays in yeast strains carrying mutations in the corresponding homologous gene (Rad25). Moreover, a mutant plant containing the altered AtXPB1 gene was obtained by screening a plant library with random T-DNA insertions (Bouchez et al., 1993). Homozygous plants of this mutation were obtained, indicating that the gene product is not essential, possibly due to gene duplication. However, phenotypic analysis of these plants revealed interesting features, including delayed development and sensitivity to genotoxic agents. These characteristics suggest inefficient DNA lesion removal and potentially divergent functions and/or mRNA expression for AtXPB1 and AtXPB2.

Results

AtXPB1 partially corrects the UV sensitivity of a yeast rad25 mutant

The Saccharomyces cerevisiae strain KG119 bears a mutant allele of Rad25 (ssl2-xp) (Gulyas and Donahue, 1992) that mimics the mutation found in a defective human allele (cells XP11BE-FS740, Weeda et al., 1990a). This mutation corresponds to a deletion of the carboxyl-terminal region of the protein, and confers hypersensitivity to UV-light (Gulyas and Donahue, 1992). This mutant is completely deficient for all NER functions (Sweder and Hanawalt, 1994). The null mutation is not viable due to the essential role of RAD25 protein in RNA polymerase II transcription. Thus, KG119 is also called the ssl2-xp/rad25 strain.

AtXPB1 cDNA was subcloned into yeast expression vectors pYES (Elledge et al., 1991) and pYEP52 (Hill et al., 1986) and introduced into a wild type (KG99) and ssl2-xp/rad25 (KG119) yeast strains, respectively. These vectors confer auxotrophy for uracil (pYES) or leucine (pYEP52), necessary for vector selection in each yeast strain. As controls, the same transformations were conducted using empty vectors. The transformed strains were tested for UV sensitivity by measuring the ability of the irradiated cells to form colonies. They were exposed to increasing doses of UV light and kept in dark conditions, in order to allow for NER action. As shown in Figure 1, the yeast mutant is more sensitive to UV light when compared with the wild type strain. The expression of the plant protein resulted in a clear increase of mutant yeast survival, with no observed effect in the wild type strain. This complementation of UV sensitivity, although partial, suggests the involvement of AtXPB1 in the NER pathway.

Figure 1.

Partial complementation of ssl2-xp/rad25 yeast sensitivity to UV light by AtXPB1.

The tested yeast strains are of the wild type (KG99, circles) and the ssl2-xp/rad25 (KG119, squares) bearing expression vectors carrying the plant gene AtXPB1 (closed symbols) or empty (open symbols). Error bars represent SD of the average of four independent experiments. The results for the UV irradiated KG119 strains carrying the AtXPB1 gene were statistically different from the strain without the gene (P < 0.05, as indicated by the Tukey test, Sokal and Rohlf, 1981).

AtXPB1 is unable to correct a transcription defect of the rad25 mutant protein

Apart from its function in NER, RAD25 and XPB are essential for RNA polymerase II transcription in yeast and human cells, respectively. Studies conducted with the human homologue protein (XPB) suggested its involvement during promoter unwinding (Coin et al., 1999). The yeast complementation system was also employed to examine the AtXPB1 function in this process. Its participation in transcription was ascertained through testing the complementation of a thermo-sensitive conditional mutant rad25 yeast strain. The EPY82–24 strain (Park et al., 1992) is unable to grow in temperatures above 37°C due to a point mutation in the RAD25 protein that inhibits its transcriptional activity at that temperature, although it can be kept at 28°C, with normal growth. The yeast strain was transformed with pYEP/AtXPB1 and tested for growth. After 48 h of incubation at 28°C growth was normal, while no growth was observed at 37°C (results not shown). This negative result is probably due to AtXPB1 inefficiency when acting in the transcription of yeast cells, whereas expression of the plant gene in the transformed strains was confirmed by RT–PCR (data not shown).

Isolation of the Arabidopsis plant line with a T-DNA insertion in AtXPB1 gene

A collection of approximately 30,000 Arabidopsis plant lines containing random insertions of Agrobacterium tumefaciens T-DNA was screened by PCR using a combination of pairs of primers specific for the AtXPB1 gene and T-DNA border sequences, respectively (Bouchez et al., 1993). The presence of AtXPB1/T-DNA insertions inside 39 hyperpools, each containing genomic DNA from 768 mutant Arabidopsis lines, was searched for by separate PCRs using eight possible primer combinations: one of the four AtXPB1 primers with either the tag3 (left border) or the tag5 (right border) T-DNA primer (Figure 2a). Hybridization of PCR products to both an AtXPB1 and a T-DNA probe specifically identified one positive DNA band of approximately 1.0 kbp, derived from primer combination XPB2/tag5. The PCR screening employed the hierarchical pooling strategy up to the identification of a single Arabidopsis line. The occurrence of a T-DNA insertion within the AtXPB1 gene was confirmed by sequence analysis of the PCR product.

Figure 2.

The atxpb1–/– plant has a homozygous genotype.

(a) Schematic representation of T-DNA insertion into the AtXPB1 sequence. The structure of the AtXPB1 and the T-DNA insertion were deduced from PCR analysis. The position of the main protein domains are represented: NLS-nuclear localization signal; DNA bind. DNA binding; black boxes: helicase domains; LB-T-DNA left border and RB-right border. The arrows indicate the position of primers.

(b) Determination of the presence or absence of a mutant allele through the electrophoretic separation of PCR products. DNA from wild type (lanes 1–3) and atxpb1–/– (lanes 4–6) plants were PCR-amplified using the primer combinations 12T/XPB2 (lanes 1, 4); 11T/13T (lanes 2, 5) and XPB2/tag5 (lanes 3, 6). L represents a DNA molecular weight marker.

Molecular analysis of the mutant atxpb1 plant line

The XPB2/tag5 1.0 kbp PCR amplified fragment was sequenced. The T-DNA is located in the 15th exon, more precisely at position 2023 of the cDNA. This corresponds to a T-DNA insertion at the last codon of the corresponding VI helicase domain of the putative protein. The T-DNA insertion is positioned from the left to the right border (Figure 2a).

The next step was the identification of wild type and/or mutant alleles for AtXPB1 in the selected T-DNA line by PCR assay using specific primers for AtXPB1 sequences surrounding the insertion. The results are shown in Figure 2(b) and while DNA bands from the gene are observed for the wild type plant DNA, only one band from the pair of primers that identified the insertion was seen for the mutated plant. It should be noted that the PCR conditions would not amplify long DNA fragments, as those expected to include the plant line containing the T-DNA insertion. These results demonstrate that mutated plants are homozygous for the mutated allele; therefore it was then named atxpb1–/–. The sister plant, which does not contain the mutated allele, is called AtXPB1+/+. The presence of only one T-DNA insertion in the homozygous mutant plant at AtXPB1 locus and the absence of a T-DNA insertion in the wild type plant were confirmed by Southern blot analysis (data not shown). These two identified plants, wild type and mutant, and their progeny were used in all subsequent experiments of molecular and phenotypic characterization.

The transcriptional characterization of the mutated gene was performed by AtXPB1 mRNA detection. Previous RT–PCR analysis, performed on the wild type A. thaliana, did not detect any expression differences for AtXPB1 and AtXPB2 in adult plants, thus suggesting that the expression of both genes occurs in all plant tissue at this stage (P. G. Morgante and M. A. Van Sluys, pers. comm.). The analysis obtained by Northern blot revealed an alteration of the atxpb1 transcript size and an increase in the level of transcription of atxpb1 mRNA in mutant plantlets (Figure 3). The size of the mRNA product presents a significant reduction of approximately 0.7 kb. The mutant mRNA size corresponds to the position of T-DNA insertion in the cDNA. Thus, the T-DNA insertion seems to disturb the processing of AtXPB1 mRNA, resulting in an altered product; this also affects mRNA expression or its turnover. As the probe employed does not discriminate the two homologues, the faint band of 2.8 kb observed in the mutant plant most probably corresponds to the expression of AtXPB2 gene. RT–PCR experiments on total RNA confirmed the presence of AtXPB2 product. Specific primers for both genomic sequences were used and resulted in the amplification of a fragment of approximately 400 bp, which covers a corresponding internal region for AtXPB1 and AtXPB2. Digestion with PvuII confirms the origin of the amplified fragments, as only the AtXPB1 sequence contains an internal restriction site for this enzyme. As shown in Figure 3(c), the mRNAs for AtXPB1 and AtXPB2 are detected in these plants, supporting the hypothesis that the observed 2.8 kb band in the Northern blot corresponds to AtXPB2 mRNA.

Figure 3.

AtXPB1 mRNA expression in plantlets.

(a) Northern blot analysis of total RNA extracted from seedlings of AtXPB1+/+ and atxpb1–/–. The 32P-labelled (AtXPB1) probe was the complete cDNA of the AtXPB1 gene. The blot was rehybridized with the 16S rDNA gene as a probe for normalizing the radioactive signals.

(b) The radioactive bands were quantified using a GS-700 Imaging Densitometer (BioRad Laboratories, California, USA) (expression relative to 16S RNA signals).

(c) Confirmation of the presence of AtXPB1 and AtXPB2 mRNAs through the electrophoresis separation of RT–PCR products. RNA from wild type (lanes 1, 2) and atxpb1–/– (lanes 3, 4) plants were amplified (RT–PCR) using the primer combinations 3 A/3 Aa, specific for AtXPB2 (lanes 1, 3) and 3 A/3 Ab, specific for AtXPB1 (lanes 2, 4). L represents a DNA molecular weight marker.

Phenotypic characterization of mutant atxpb1–/– plants

The growth and development of the mutant plant line was accompanied through different stages and compared with wild type plants. atxpb1–/– plants, when grown in soil, showed a slight delay in their development from germination to flowering. However, no morphological anomaly was observed. A careful comparison of seed germination and plant development in vitro confirmed this delay. These observations were quantified, and the data are presented in Figure 4. Three main conclusions can be drawn as a result of the mutation of the AtXPB1 gene: (i) absence of germination synchrony; (ii) lower seed germination rate; and (iii) delay in organ development. After 12 days, most of the atxpb1–/– plantlets proceeded with normal growth and development. These results indicate participation of the AtXPB1 protein at early stages in plant development.

Figure 4.

Developmental patterns of wild type (white) and mutant atxpb1–/– (black) plants.

Bars indicate the percentage of plantlets presenting (a) roots (b) cotyledons and (c) first leaf pair for 11 days after germination.

An interesting finding was observed during the manipulation of plant seeds for germination. Surface sterilization of seeds with a hypochlorous acid (HOCl) solution induced lower germination efficiency in seeds from atxpb1–/– plants. This was rechecked and the number of seeds that germinated was defined after increasing the time of HOCl treatment; results are shown in Figure 5. Seeds from wild type plants also displayed certain sensitivity to this solution, starting after 30 min of incubation. However, the seeds from mutated plants were much more sensitive than the wild type to HOCl, as seed germination is almost abolished after the same period and shorter periods of incubation lead to a decrease in seed viability.

Figure 5.

Decreased seed germination viability in an atxpb1–/– plant after treatment with hypochlorite.

(a) Seeds were incubated for the indicated periods of time in a HOCl solution: AtXPB1+/+ (white bars) and atxpb1–/– (black bars).

(b) General aspect of wild type (upper region of the Petri dish) and mutant atxpb1 (lower region) 7-day-old plantlets with seeds previously treated for 30 mins in a HOCl solution.

Sensitivity of atxpb1–/– mutant plants to different genotoxic agents

The xpb/rad25 human and yeast mutant cells are very sensitive to UV induced damage (Weeda et al., 1990a;1990b; Gulyas and Donahue, 1992; Park et al., 1992). Thus, the atxpb1–/– mutated plants were tested for their UV sensitivity. Three-day-old plantlets were exposed to different UV (254 nm) doses (up to 2000 Jm−2.s−1) and were kept under light or dark conditions, in order to distinguish the ability of these plantlets to remove UV-induced lesions by photoreactivation and NER, respectively. Surprisingly, although higher UV doses induced a decreased leaf growth and an increased yellowish pigmentation, no phenotypic differences were observed among the wild type and mutated plants at any UV analyzed dose (data not shown).

Other than UV light, xpb mutant mammalian cells show high sensitivity to the alkylating methyl methane sulfonate (MMS) agent (Weeda et al., 1990b). Therefore, the sensitivity of atxpb1–/– plants to this chemical was also tested. Six-day-old plantlets were transferred to a nutrient medium containing increasing concentrations of MMS. As is illustrated in Figure 6(a), after 3 weeks, wild type plants suffer from MMS treatment at high doses, as their growth is inhibited at 100 and 120 p.p.m. The increased sensitivity of mutant plants to MMS was confirmed by a second strategy, as 3-day-old plantlets were grown with 80 p.p.m. of MMS in liquid medium for 1 week. This treatment resulted in a higher death rate among atxpb1–/– plantlets (75%), when compared with wild type plantlets (33%) (Figure 6b). These results confirm a more pronounced sensitivity of the mutant plant line to this genotoxic agent.

Figure 6.

Increased sensitivity of atxpb1–/– plants to MMS.

(a) Six-day-old plantlets were transferred to a medium containing indicated MMS doses, where they were allowed to grow for 3 weeks.

(b) Sensitivity difference between wild type and atxpb1-/–plantlets after 1 week of treatment with 80 p.p.m. of MMS.

Discussion

The aim of this work is to functionally characterize the Arabidopsis AtXPB1 gene. AtXPB1 codes for a protein with a high aminoacid sequence similarity with the human XPB and S. cerevisiae Rad25 helicases. Arabidopsis differs from other eukaryote organisms as it presents two paralogous genes: AtXPB1 (Ribeiro et al., 1998) and AtXPB2 (The Arabidopsis Genome Initiative, 2000). From the predicted sequence, both genes encode very similar proteins (95% identical) and differences are mostly located at the C-terminal, not affecting the helicases motifs. However, differences at this region strongly affect DNA repair in the human gene (Weeda et al., 1990b). Preliminary studies performed with 10-day-old plantlets and roots, leaves and flower tissues from adult plants did not reveal any specific expression of these two genes. However, these results do not eliminate the possibility of differential tissue expression during the first stage of plant development. These features raise some intriguing questions concerning the functions of the two homologues and how they are co-ordinated by cell DNA metabolism.

The role of AtXPB1 in NER and RNA polymerase II transcription was initially ascertained through complementation assays in mutant rad25 yeast strains. The expression of the plant protein resulted in a partial survival increase after UV exposure (Figure 1). This functional complementation implies that AtXPB1 has the potential to repair UV-induced damage through the NER pathway. However, the plant protein was unable to complement transcription deficiency in the thermo-sensitive conditional yeast strains. The heterologous expression of the plant gene may produce a not completely functional protein in yeast. In fact, it is reasonable to suppose that the interaction of AtXPB1 with TFII-H yeast proteins is not very stable. The Arabidopsis genome data revealed important differences between the TFII-H components of this plant and those from other eukaryotic organisms (The Arabidopsis Genome Initiative, 2000). For example, in plants the cdk-activating kinase (CAK) subcomplex, normally required for RNA Pol II transcription initiation (de Laat et al., 1999), differs considerably from that found in human cells. This condition may explain the partial AtXPB1 action on double helix unwinding during NER and its ineffective complementation of transcription. In agreement with these results, previous experiments trying to obtain heterologous complementation of the AtXPB1 in rodent cells deficient at the XPB homologue were unsuccessful (Ribeiro et al., 1998). Therefore, the biological role of AtXPB1 protein transcription, as well as in NER, has to be addressed more precisely in Arabidopsis.

The study of the AtXPB1 function in Arabidopsis was performed by the identification and characterization of a plant line containing the AtXPB1 gene interrupted by a T-DNA insertion. The screened mutant presents the T-DNA insertion interrupting the 3′ region of AtXPB1 genomic sequence, at a position that interrupts the 15th exon, at the last codon of the ultimate helicase protein domain. This insertion results in a high expression level of an mRNA with a reduced size, possibly due to changing intron-processing signals. Consequently, the putative protein probably lacks part of its carboxyl-terminal region. The atxpb1–/– mutant is viable, as homozygous plants were obtained. This is not a trivial event, since XPB and RAD25 are essential genes for both mammalian and yeast cells. Mutant cells that are homozygous for the disrupted allele are not viable (de Boer and Hoeijmakers, 1999; Gulyas and Donahue, 1992). A smaller deficiency of the C-terminal in the human XPB protein results in a severe clinical syndrome with the manifestation of xeroderma pigmentosum associated with the Cockayne syndrome. The patient with this syndrome presents high photosensitivity, resulting in a high skin cancer development rate, and severe neurological and developmental problems (Weeda et al., 1990b). In contrast to the human patient, the mutant plant did not show a severe developmental problem, although delay in development is observed. Two hypotheses can be drawn: first the biological function of AtXPB1 is not required during plant development; and second that AtXPB2 is complementing AtXPB1 deficiency. As their predicted sequences are very similar the expression of a functional AtXPB2 is probably sufficient to maintain most of the transcriptional functions in the plant cell.

Interestingly, the data obtained with the mutant plant line suggest the requirement of AtXPB1 protein during the first stages of germination, since delay in germination was observed. If it is assumed that AtXPB2 is able to complement the AtXPB1 deficiency in later stages, it may be possible that during germination, differences in timing expression of both genes or different functional specialization of both proteins result in this phenotype. This developmental delay could also be a result of transcription deficiency during the first stages of plant development. However, a developmental delay was also observed in XPF Arabidopsis mutant (Liu et al., 2000). The latter protein is not directly involved in the transcription process, but both XPB and XPF human proteins participate in DNA repair. Thus, it is possible that the defective action of AtXPB1 in NER is responsible for the early developmental delay in Arabidopsis. The importance of DNA repair in embryos has already been observed in mature seeds, immediately after rehydration, as one of the early events after germination (Dandoy et al., 1987; Elder et al., 1987; Osborne et al., 1980). The deficiency of AtXPB1 in removing DNA lesions may also be responsible for the lack of plant growth and germination synchrony. In mature seeds, most embryonic cells stall at the G1 phase of the cell cycle (Vonarx et al., 1998), and the deficiency of atxpb1–/– seeds may have required longer periods for dealing with damaged DNA, thus disturbing early embryonic growth. Also, the observed decrease of seed viability may be a consequence of DNA repair deficiency of atxpb1–/– plants. This decreased germination efficiency was more pronounced after incubation with HOCl for longer periods of time (Figure 4). The higher sensitivity of mutated seeds to HOCl is a strong indication that AtXPB1 action in DNA repair is required during germination. These observations are in accordance with suggestions that DNA damage is induced during the storage and dehydration of seeds and thereby repair mechanisms are required during germination (Osborne et al., 1980). It is known that HOCl generates reactive oxidant species (Shen et al., 2000), which can damage the DNA. Spencer et al. (2000) have shown that human cells exposed to HOCl have significant increase in oxidative DNA damage. In human beings, XPB protein is required for removing DNA oxidative lesions from transcribed strands (Le Page et al., 2000). Interestingly, recent data indicate that the AtRAD2/XPG protein also acts to repair oxidative damage from DNA (Liu et al., 2001). It is reasonable to suppose that the generated DNA oxidative lesions were less efficiently removed in the atxpb1–/– mutant plant line, resulting in a decrease of seed viability.

Contrary to these results, the atxpb1–/– plants did not show any sensitivity to UV light. Differing from AtXPB1, the AtXPF and AtXPG mutant plants showed high UV sensitivity (Gallego et al., 2000; Liu et al., 2000; Liu et al., 2001). The most reasonable explanation for these results is the complementing action of AtXPB2 in the repair of UV-induced lesions. This complementation may be only partial for MMS induced lesions, whereas mutant atxpb1–/– plants were more sensitive to this chemical than wild type plants (Figure 6). This difference in sensitivity of atxpb1–/– plant to these two agents is intriguing because the removal of UV-induced lesions has always been related to NER. The divergent results for these genotoxic agents may be due to the different types of treatment. During UV irradiation, there is a brief exposure to the damaging agent. For MMS treatment, the plants were exposed for a long period with this chemical, with continuous lesion induction. Thus, in this latter condition the presence of AtXPB2 protein may be not sufficient for removing the MMS induced lesions.

Another explanation is that UV and MMS induce different kinds of lesions, which are differentially managed by AtXPB1 and AtXPB2. In view of the data presented here, AtXPB1 would be implicated more in the repair of lesions induced by MMS. This is consistent with the ability of AtXPB1 to act in the DNA repair of oxidative lesions in embryo cells during germination. Finally, since the T-DNA disruption is at the 3′ end of the plant gene, there is the possibility that this truncated AtXPB1 protein is produced. This protein could still interact in the TFII-H complex, competing with the AtXPB2 protein. As the truncated protein may be partially or completely inactive, it would interfere with the normal functions of TFII-H, especially during germination and the early development of the mutated plants, yielding the phenotype shown. This would be even more noticeable if differential expression for AtXPB1 and AtXPB2 mRNA could result in low amounts of AtXPB2 in the plant cells during the first stages of Arabidopsis development. Until now, both mRNAs were found in different plant tissues investigated by RT–PCR, but small differences of expression, not detected by this methodology, could have profound phenotypic effects, especially during embryonic development.

The studies performed with AtXPB1 suggest its involvement in NER in A. thaliana. But both gene copies seem to be functional, whereas AtXPB2 was probably able to complement AtXPB1 deficiency for the transcription and removal of UV induced lesion. Thus, this duplication confers more flexibility and resistance to DNA damaging agents to this plant when compared with other organisms. However, this complementation was not fully efficient during the first stages of plant life, as atxpb1–/– seeds exhibited germination and early developmental problems. The atxpb1–/– plant line is also more sensitive to HOCl (seed) and MMS (plantlet) treatments, indicating DNA repair impairment. This could be the first indication of divergence of AtXPB1 and AtXPB2 proteins in their functional roles in plant cells. Not exclusively, quantitative differences in time or tissue expression of both mRNAs may explain the atxpb1–/– phenotype. Experiments to clarify these possibilities are being performed. They may help to elucidate whether two genes, products of a recent duplication, have already evolved to divergent functions.

Experimental procedures

Construction of plasmids for complementation analysis

The AtXPB1 cDNA was subcloned into the yeast expression vectors pYES and pYEP52, by normal cloning procedures (Sambrook et al., 1989). In these vectors the gene is under the control of the GAL10 promoter, and they can be selected in yeast by employing a medium lacking either uracil (pYES) or leucine (pYEP52).

Complementation assays in S. cerevisiae

S. cerevisiae media and classical genetic techniques are described elsewhere (Sambrook et al., 1989). The strains used in this study were KG99 (his7–2, leu2–3, leu-112, trp1–289, ura3–52, SSL2::TRP1, {p1533 SSL2 LEU2 CEN4}); KG119 (his7–2, leu2–3, leu-112, trp1–289, ura3–52, SSL2::TRP1, {YCp50 SSL2-XP URA3 CEN4}) (Gulyas and Donahue, 1992); conditional mutant EPY82–24 (MATa, leu2–3, leu2–112, trp1-Δ, ura3–52, rad25Δ, {pEP39–0 rad25ts24TRP1}) (Park et al., 1992). Yeast strains KG99 and KG119 were grown in a minimum medium supplemented with necessary amino acids, serially diluted and plated onto rich medium plates. For UV-irradiation, plates were immediately exposed at increasing times to a germicidal lamp (mainly 254 nm, rate 0.4 Jm−2s−1), then incubated at room temperature in the dark for 24 h, followed by at least 48 h at 30°C. Surviving colonies were counted and survival frequency was calculated as the ratio of irradiated to-non-irradiated colonies.

For the transcription complementation assay, 3 × 106 yeast and dilution series of 1/10, 1/102 and 1/103 were grown on a solid rich medium for 48 h at 28°C or 37°C (restrictive temperature).

Plant material and culture conditions

All experiments were performed using Arabidopsis thaliana ecotype Wassilewskija (WS). Seeds were sterilized in an HOCl solution (5% HOCl, 0.5% Tween 20) for 15 mins and grown on a standard culture medium (ABIS: KNO3 5 mm, MgSO4 2 mm, Ca(NO3)2 1 mm, Fe-EDTA 50 µm, sucrose 10 gl−1, agar 0.7 gl−1, micronutrients as described by Murashige and Skoog (1962), MES 1 mm, and K2HPO4 + KH2PO4 2.5 mm, pH = 6.0). For T-DNA selection kanamycin was added at a concentration of 50 µg ml-1. Plants were kept at 4°C for 48 h and then transferred to the greenhouse at 21°C with 10 h light (120 µEm−2 s−1)/14 h dark cycles. The conditions for plant growth in the soil were similar to those in an agar media.

Screening of Arabidopsis T-DNA library

The PCR screen was performed on pooled genomic DNA from independently isolated T-DNA transformed lines (Bechtold et al., 1993; Bouchez et al., 1996). The oligonucleotide primers used to target AtXPB1 sequences were:

XPB1–5′-ATCGCTAATTGATAACT-3′;

XPB2–5′-GGTCAGCAATTCAAGAAACCTAAAGAC-3′;

XPB3–5′-GACATATTTGGTCATCTCTTATAGTAGAC-3′;

XPB4–5′-GGTCTGAGGAATCAGAGAAGATTATAG-3′

To target T-DNA sequences, a left border (LB) tag3 primer (5′-CTGATACCAGACGTTGCCCGCATAA-3′) and a right border (RB) tag5 one (5′-CTACAAATTGCCTTTTCTTATCGAC-3′) were used. They read in opposite directions towards the T-DNA surrounding sequences. Each PCR reaction contained, in 25 µl, 50 ng DNA, 50 µmol T-DNA primer, 100 µmol AtXPB1 primer, 1 unit Taq polymerase, 2.5 µl 10X reaction buffer (Invitrogen Corporation, California, USA), and 0.2 mm dNTP. The following reaction cycles were used: 94°C for 5 min; 55 cycles at 94°C for 45 sec, 60°C for 1 min and 72°C for 2.5 min; 72°C for 5 min. The DNA products were discriminated by electrophoresis, blotted to nylon membranes, and hybridized with 32P-labelled AtXPB1 and T-DNA probes under stringent conditions (Sambrook et al., 1989). The AtXPB1 probe was the full-length cDNA, and the T-DNA probe corresponded to two 800 bp HincII fragments from plasmid pGKB5 encompassing both extremities of the T-DNA region (Bouchez et al., 1993). The genetic analysis of homozygous plants of the mutant or wild type of AtXPB1 locus were distinguished by differential PCR. The oligonucleotide primers used were: 12T−5′-CCTGAAGCTAATGTG-3′; 11T−5′-GTACTCATCACTTAAG-3′; 13T−5′-TTCACATGGAGATACG-3′. The wild type AtXPB1 gene produced a 760-bp PCR fragment with 12T/2 primers; a 400-bp PCR fragment with 11T/13T primers, and no fragment with XPB2/tag5 primers (Figure 2). The mutant atxpb1 gene produced no fragment with 12T/2 and 11T/13T primers, and an 880-bp PCR fragment with XPB2/tag5 primers.

RNA extraction, DNA extraction, sequence analysis and gene expression

RNA extraction and hybridization was carried out as described (Sambrook et al., 1989). DNA preparation from pooled T 2 progeny Arabidopsis plants was done as previously described (Bouchez et al., 1993). PCR fragments were cycle sequenced with an Applied Biosystems 377 automated sequencer. For RNA expression, RT–PCR assays a 400-bp internal fragment of AtXPB1 and AtXPB2 were amplified with specific primers: AtXPB1–3 A-5′-GCCCCATGCACAACC-3′ and 3 Ab-5′-CATCCATGAGCAGTAAC-3′; AtXPB2–3 A and 3 Aa-5′-CTCGTCCATAAGCAGC-3′. Specificity was further confirmed by DNA digestion with PvuII restriction enzyme, which cleaves only AtXPB1 PCR product. Each RT–PCR reaction contained, in 50 µl, 1 µg RNA template, 0.2 mm for each primer, 25 µl 2X Reaction Mix, 1 µl RT/Taq Mix (Invitrogen Corporation). The following reaction cycles were used: 1 cycle at 50°C for 30 min and 94°C for 2 min; 40 cycles at 94°C for 15 sec, 55°C for 30 sec and 68°C for 2.5 min; 1 cycle of 72°C for 7 min. The DNA products were discriminated by electrophoresis in a 2% agarose gel.

UV and MMS sensitivity tests

Sterilized seeds from wild type and mutated plants were sown in a solid nutritive medium and cultured under normal plant conditions. Plates were placed vertically to allow the roots of the plantlets to grow downward on the surface of the agar. After 4 days of germination, plantlets were exposed to different UV irradiation (up to 2000 Jm−2 at rate 3 Jm−2 s−1) and kept under light and dark conditions. The plantlets' development was observed daily up to day 5. The phenotypic morphology of plantlets was also observed up to this day. For MMS treatment, seeds were allowed to germinate in a normal solid nutrient medium under normal conditions. After 6 days of germination, plantlets were transferred to a liquid nutritive medium containing different MMS doses (50, 80, 100 and 120 p.p.m.) during 3 weeks. Otherwise, 3-day-old plantlets were transferred to liquid medium containing MMS (80 p.p.m.) for 1 week.

Acknowledgements

This work was supported by FAPESP (São Paulo, Brazil, proc.# 98/11119–7) and PADCT/CNPq (Brasília, Brazil). The authors are grateful to Dr L.F.Agnez-Lima for help in statistical analysis.

Accession numbers: AtXPB1 –U29168/AtXPB2 –AF308595.

Ancillary