Roles of the AtErcc1 protein in recombination

Authors


(fax 33 4 73 40 77 77; e-mail chwhite@univ-pclermont.fr).

Summary

Atercc1, the recently characterized Arabidopsis homologue of the Ercc1 (Rad10) protein, is a key component of nucleotide excision repair as part of a structure-specific endonuclease which cleaves 5′ to UV photoproducts in DNA. This endonuclease also acts in removing overhanging non-homologous DNA ‘tails’ in synapsed recombination intermediates. We have previously demonstrated this recombination function of the Arabidopsis thaliana Xpf homologue, AtRad1p, and show here that recombination between plasmid DNA substrates containing non-homologous tails is specifically reduced 12-fold in atercc1 mutant plants compared with the wild type. Furthermore, using chromosomal tandem-repeat recombination substrates, we show that AtErcc1p is required for bleomycin induction of mitotic recombination in the chromosomal context. This work thus confirms both the specific and general recombination roles of the Atercc1 protein in recombination in Arabidopsis.

Introduction

The sessile nature and the need for sunlight for photosynthesis of plants means that they are particularly exposed to UV radiation. The two major DNA lesions produced by UV are cyclobutane (CPD) and pyrimidine (6-4) pyrimidone (6-4PP) dimers, which can block DNA replication, transcription and induce mutations (Friedberg et al., 1995). As is the case for other organisms, plant cells dispose of pathways for repair DNA damage (reviews by Britt, 1999; Hays, 2002; Tuteja et al., 2001).

The primary repair pathway for UV-induced CPDs is photoreactivation, in which the photoreactivating enzyme uses the energy of visible light to directly repair CPDs (Sancar, 1994). Arabidopsis plants have both CPD (Uvr2) and 6-4PP (Uvr3) photoreactivation activities (reviewed by Britt, 1999). In the absence of visible light, nucleotide excision repair (NER) removes a wide range of distorting lesions including CPDs and 6-4PPs. The proteins and processes involved in this process are highly conserved and have been the subject of recent reviews (Christmann et al., 2003; Friedberg et al., 1995; Hays, 2002; Hoeijmakers, 2001; Prakash and Prakash, 2000; Sinha and Hader, 2002; Wood, 1996). The NER pathway involves recognition of the damage, incision and removal of the damaged DNA strand followed by filling in of the resultant single-stranded gap by DNA synthesis and ligation. Xpc/hHR23B (Rad4/Rad23 in yeast) acts as the damage detector and initiates the NER. Xpa (Rad14), Rpa (Rfa), and TFIIH form an open complex where Xpb (Rad25) and Xpd (Rad3) bind and help in unwinding DNA. The single-strand incision 3′ of the damaged site is performed by Xpg (Rad2) and the 5′ incision by the Xpf/Ercc1 complex (Rad1/10).

Nucleotide excision repair also occurs in plants (reviews by Britt, 1999; Hays, 2002; Tuteja et al., 2001) and a number of mutants have been identified in this process. The Arabidopsisuvh3 mutant is the result of mutation of the homologue of the RAD2/XPG gene and shows hypersensitivity to both UV-B and UV-C (Liu et al., 2001). Atrad1, the Arabidopsis homologue of Xpf (yeast Rad1), has been characterized and atrad1 mutant plants and cells are UV-hypersensitive and have been shown to be defective in the removal of UV photoproducts in the dark (Fidantsef et al., 2000; Gallego et al., 2000; Li et al., 2002). Ercc1 (Rad10) homologues have been described in Lillium longiflorium and Arabidopsis thaliana. The lily Ercc1 protein partially complements the repair defect of Ercc1-deficient Chinese hamster ovary cells (Xu et al., 1998) and the Arabidopsis EMS-induced uvr7 mutant (atercc1) is hypersensitive to UV, gamma-rays and MitomycinC and has a slight sensitivity to methyl methane sulphonate (Hefner et al., 2003). Following gamma irradiation, the cells of the uvr7 mutant seedlings show a post-irradiation G2 phase arrest phenotype (Hefner et al., 2003).

Ercc1 (Rad10) and Xpf (Rad1) act in a complex to make the DNA incision 5′ to the lesion site (Bardwell et al., 1993, 1994; Davies et al., 1995; Tomkinson et al., 1993). Mutations of the human XPF gene result in xeroderma pigmentosum (XP), characterized by UV hypersensitivity, pigmentation abnormalities and a 1000-fold increased risk of skin cancer (Sijbers et al., 1996). Ercc1-deficient mice are severely runted and die before weaning, of liver failure (Hsia et al., 2003; McWhir et al., 1993; Nunez et al., 2000).

In addition to their role in NER, Ercc1 and Xpf are involved in homologous recombination, double-strand break (DSB) repair and the repair of inter-strand cross-links. ercc1 mutant CHO cells are hypersensitive to inter-strand cross-linking agents, although mouse ercc1 cells are much less so (Melton et al., 1998). In Saccharomyces cerevisiae, mutations in RAD1 and in RAD10 reduce intra-chromosomal recombination between directly repeated sequences and also decrease the efficiency of homologous integration of linear DNA fragments and circular plasmids (Fidantsef et al., 2000; Klein, 1988; Prado and Aguilera, 1995; Schiestl and Prakash, 1988, 1990).

In mammalian cells, Ercc1 is involved in removing non-homologous tails from ends-in gene targeting constructs in Chinese hamster cells (Adair et al., 2000; Sargent et al., 2000) and is required for ends-out targeted gene replacement in mouse embryonic stem cells in the absence of non-homology of the ends of the targeting construct with respect to the targeted genomic locus (Niedernhofer et al., 2001). ercc1 mutant CHO cells show elevated levels of recombination-dependent deletion formation of a chromosomal tandem duplication (Sargent et al., 1997, 2000). Furthermore, a recent report has shown that the function of the Errc1/Xpf complex is modulated by a direct interaction with hRad52 and has proposed that hRad52 could target and activate Ercc1/Xpf in recombination reactions (Motycka et al., 2004).

In plants, the study of the mechanisms of recombination lags far behind that of mammals and yeast. With the recent availability of tagged mutant collections and the genome sequence of Arabidopsis, this is rapidly changing (reviews by Bhatt et al., 2001; Britt, 1999; Gorbunova and Levy, 1999; Hanin and Paszkowski, 2003; Tuteja et al., 2001). We have previously shown that Atrad1 (Arabidopsis Xpf/Rad1 homologue) is involved in homologous recombination (Dubest et al., 2002) and here we present the characterization of the recombination role of the Atercc1 protein, testing specifically its involvement in removing non-homologous tails to permit productive homologous recombination between linear DNA molecules in A. thaliana. We confirm that this recombination role depends upon both the Atrad1 and the Atercc1 proteins, thus strongly supporting that this recombination role is performed by the Atrad1/Atercc1 endonuclease. We also show that AtErcc1p is involved in chromosomal recombination between tandem repeats in both direct and inverted orientations and is needed for the bleomycin induction of recombination in this plant. Thus, our data support a general recombination role for AtRad1/AtErcc1 in plants in addition to their NER role.

Results

Isolation of Arabidopsis atercc1 mutants

Two Arabidopsis atercc1 mutant lines, Salk_033397 and Salk_077000, with T-DNA insertions in the AtERCC1 coding sequence were identified in the Salk Institute SIGnAL T-DNA insertion Arabidopsis mutant collection (Alonso et al., 2003). We have designated the mutant atercc1 alleles present in these lines as atercc1-1 (Salk_033397) and atercc1-2 (Salk_077000).

As frequently observed in plants from the Salk mutant collection, the kanamycin resistance marker present in Salk_033397 is not expressed; homozygous plants were thus identified by a PCR genotyping assay. Primers o397/o399, which span the insertion site were used to amplify the wild-type AtERCC1 allele and these primers with o405 (in the T-DNA LB sequence) the atercc1-1 allele (Figure 1a,b). The T-DNA insertion is surrounded by two incomplete left borders (LB) in diverging orientation. The insertion point of the T-DNA downstream end (LB1) was mapped by the sequencing of the PCR product generated by primers o406/o432 and shown to lie in exon 3 of the AtERCC1 gene (Figure 1a). Junction LB1 has 16 bp of unknown filler DNA between the AtERCC1 sequence and the beginning of the pROK2 vector LB sequence (Figure 1d). The T-DNA sequence starts 17 nt after the nicking site. The insertion point of the T-DNA upstream end (LB2) was mapped by the sequencing of the PCR product generated by primers o397/o406. LB2 has an insertion of 32 bp of unknown filler DNA, with no loss of T-DNA sequence (Figure 1d). Twenty-five base pairs of the AtERCC1 exon3 sequence were deleted at the insertion point. This 25 bp is shown as 13 and 12 bp blocks in Figure 1(d) in order to facilitate comparison. RT-PCR analysis, using primers in the AtERCC1 coding region upstream of the insertion (o435/o436), was used to check for AtERCC1 transcription in mutant and control plants. The expected 167 bp fragment was amplified by o435/o436 in the wild-type sample but no transcript was detected from the atercc1-1 locus. In addition, no RT-PCR product was amplified with primers o405/o432 on RNA from mutant plants, indicating that no hybrid transcript was transcribed from the mutant locus. As expected, the o405/o432 pair does amplify the 539 bp genomic fragment on DNA from mutant plants (data not shown). Control RT-PCR amplifications of the APT transcript (Moffatt et al., 1994) showed equivalent quantities of mRNA in the wild-type and mutant samples (Figure 1c). The T-DNA insertion thus inactivates the AtERCC1 gene in atercc1-1. Should a transcript be produced from the atercc1-1 locus, notwithstanding our inability to detect one, it would encode a putative protein consisting of the first 149 (of 410) amino acids of the native Atercc1 protein fused to 43 amino acids encoded by the inserted T-DNA sequence. This putative polypeptide would conserve the predicted XPA-binding region but lacks the XPF-binding regions of the Atercc1 protein (de Laat et al., 1998; Li et al., 1995).

Figure 1.

Molecular analysis of the atercc1-1 allele.
(a) Genomic organization of the AtERCC1 locus. Exons (grey boxes) and the position of the T-DNA insertion in exon 3 of the gene is shown. Primers used for genotyping and RT-PCR analysis are indicated as arrowheads.
(b) PCR genotyping of atercc1-1 and wild type. Primers used for genotyping are indicated in (a).
(c) RT-PCR. RNA isolated from 3-week-old wild-type and atercc1-1 plants. The APT1 transcript was amplified as a control of RNA quantity and quality.
(d) Sequence of the T-DNA insertion junctions aligned to the AtERCC1 and T-DNA (pROK2) sequences. Filler DNA is in bold lowercase.

The atercc1-1 plants do not show any obvious defects in growth and development under normal conditions and are fertile, in agreement with the observations of Hefner et al. (2003) and in contrast to the situation in Ercc1-deficient mice, which are severely runted and die before weaning (McWhir et al., 1993; Nunez et al., 2000).

The UV hypersensitivity of the progeny of self-fertilized heterozygous atercc1/AtERCC1 plants was quantified by examining the effect of UV irradiation on root growth. The seeds of the atercc1-1 mutant and wild-type lines, germinated and grown for 10 days on vertical agar plates, were subjected to 300 J m−2 of UV-C. Following irradiation, the plates were wrapped in aluminium foil to avoid photoreactivation and the growth of the main root of each seedling was measured daily for 3 days. After UV treatment, the atercc1-1 plants showed a complete root growth arrest while the wild-type roots continued to grow normally (data not shown). Of the 113 plants screened, 29 (25.7%) were UV hypersensitive, which corresponds to the 3:1 segregation expected for a single Mendelian locus (chi-squared, 1 d.f. = 0.03). PCR genotyping was used to verify that all the UV-hypersensitive seedlings were homozygous atercc1-1 mutants.

The T-DNA insertion in the second atercc1 mutant line, atercc1-2 (Salk_077000), was analysed by PCR using the primers o406/o397 and UV hypersensitivity verified as described above. All atercc1-2 plants present the expected hypersensitivity to UV irradiation (data not shown). The insertion point of the T-DNA left end, mapped by the sequencing of the PCR products, is in AtERCC1 gene exon 4.

The UV hypersensitivity of these lines is considered to be a result of the mutation of the AtERCC1 locus. This was confirmed by crossing atercc1-1 and atercc1-2 plants and examining the UV sensitivity of the progeny. All progeny (atercc1-1/atercc1-2) were UV hypersensitive, confirming that the UV hypersensitivity is a consequence of the defective atercc1 genes. Should another unidentified mutation have been responsible for the UV hypersensitivity in one or both lines, the progeny should all have been UV resistant. As these two mutant alleles result from independent insertional mutagenesis events, the probability of both lines carrying a mutation of the same, unidentified, second locus is extremely low.

AtErcc1p function in recombination in the presence of non-homologous overhangs

We have recently shown that Atrad1p (Xpf homologue) is implicated in DSB repair by homologous recombination (single-strand annealing) in A. thaliana and specifically in removing non-homologous overhangs to permit productive homologous recombination between linear-transforming DNA molecules (Dubest et al., 2002). In this work, we test whether the Arabidopsis Atercc1 protein is also needed to carry out this function in recombination. In order to monitor inter-molecular recombination, linear recombination substrates were prepared from pCW344 and pCW345 plasmids by restriction with enzymes that leave (with tails) or not (no tails) non-homologous overhangs flanking the region of DNA sequence homology (Figure 2). Plasmid pCW344 contains the C-terminal region of the β-glucuronidase (GUS) gene, whereas pCW345 contains the N-terminal region of GUS. The common overlapping homologous region is 590 bp long. The linearized plasmids were used to co-transform leaves from the wild-type and atercc1-1 plants and the number of recombinant spots per leaf counted after histochemical staining.

Figure 2.

Recombination assay.
The leaves of the atercc1-1 and wild-type plants were transformed with two plasmids containing two incomplete copies of the GUS gene. These substrates were digested by restriction enzymes that leave (with tails) or not (no tails) non-homologous DNA overhangs.
(a) After transformation one strand on each side is resected in the 5′ to 3′ direction leaving 3′ tails.
(b) When complementary sequences are exposed, they can anneal, forming the synapsed intermediate. In the ‘with tails’ experiments, this structure has overhanging non-homologous tails, which must be removed (c). We test whether the single-stranded tail is removed by AtRad1/AtErcc1 endonuclease.
(d) The gap is filled in, and any remaining nicks are ligated, finally resulting in a deletion product.
(e) and (d) In the ‘no tails’ constructs, the gap is repaired by a simple annealing, filling in and ligation.

As shown in Table 1(a) and Figure 3, the frequencies of recombination events with the ‘no tail’ constructs do not differ significantly between the wild-type and the atercc1-1 plants. In the ‘with tails’ transformation, the number of recombinant blue spots in atercc1-1 leaves is significantly (12-fold) reduced with respect to the wild-type leaves. Transformation efficiency was controlled in parallel by transforming with circular pGUS23 plasmid DNA, which contains the entire GUS transcription cassette. This result thus confirms that the Atercc1 protein is implicated in DSB repair by homologous recombination in the presence of non-homologous DNA overhangs.

Table 1.  Recombination in atercc1 and wild-type plants
DNAWild typeatercc1Fold reductionChi-squared
nNXnNXWild typeatercc1
  1. Plasmid-based (a) and chromosomal (b,c) recombination assays of the wild-type and atercc1-1 plants. The number of leaves (a) or plants (b,c) tested (n), total GUS+ recombinant spots (N) and the mean number of spots per leaf or plant (X) are given. Chi-squared (1 d.f.) values for the null hypothesis that all leaves or plants from a given experiment (WT and atercc1-1) are from the same population with respect to the number of blue spots. Non-parametric statistical analysis was carried out as described in materials and methods. (a) In the ‘with tails’ transformation, the number of blue spots in the atercc1 mutant is significantly (12.3-fold) reduced. No significant differences are observed in the transformation efficiency (pGUS23) or the ‘no tails’ control transformations. (b) Spontaneous intrachromosomal recombination with the inverted repeat substrate (IR) is significantly reduced in atercc1-1 plants compared with the wild type. (c) Bleomycin-induced intrachromosomal recombination in inverted (IR) and direct (DR) repeat substrates is significantly reduced in atercc1-1 mutant plants compared with the wild type.

(a)
pGUS1015515.511165151.00.80.4
No tails30411.428341.21.10.10.1
With tails301123.73090.312.33030
(b)
7 (IR)1123242.9121690.64.88594.2
31 (IR)1123242.9631001.61.827.275.3
(c)
7 (IR)100123612.4961661.7746.296
31 (IR)100123612.4812543.1417.677
17 (DR)100117011.71001871.9692.2100
Figure 3.

Recombination in the atercc1-1 mutant and wild-type leaves.
Effect of the non-homologous tails on the recombination rate in the wild-type (empty bars) and in atercc1-1 plants (filled bars). No significant difference between the mutant and wild-type plants is seen in the absence of non-homologous tails (no tails). The presence of non-homologous overhangs (with tails) specifically reduces the frequency of the recombinant spots in the atercc1-1 line.

AtErcc1p plays a role in chromosomal recombination

In order to determine whether AtErcc1 acts in recombination in chromosomal DNA, we tested chromosomal homologous recombination in planta. We used a specific assay (Gherbi et al., 2001; Swoboda et al., 1994) in which homozygous atercc1-1 plants were crossed with plants homozygous for a tandem repeat of two non-functional overlapping fragments of the GUS marker gene. A functional GUS gene can be obtained by homologous recombination between the two overlapping regions (618 bp of overlapping homology) and is visualized by blue sectors after histochemical staining of the plants.

We selected atercc1-1 and AtERCC1 plants (in F2 progeny) by PCR using the primers o397/o399 (wild-type AtERCC1 locus) and o399/o406 (atercc1 locus), both homozygous for the GUS recombination substrates in direct and in inverted orientations. The number of recombinant spots was determined both for untreated (Figure 4a) and for bleomycin-treated (Figure 4b–d) plants. Bleomycin is a gamma-ray mimetic agent, which induces DNA DSB (Favaudon, 1982). Spontaneous recombination levels in the untreated plants are significantly reduced in the atercc1-1 compared with the wild-type plants. Bleomycin treatment stimulated recombination levels in both the wild-type and the atercc1-1 plants and the higher recombination levels induced by this treatment permit a much clearer visualization of the reduced recombinational in the ercc1-1 mutant (Figure 4). A total of 100 AtERCC1 plants, 96 atercc1-1 (line 7) and 81 atercc1-1 (line 31) were tested for the inverted repeat substrate. A hundred AtERCC1 and 100 atercc1-1 (line 17) plants were tested for the direct repeat substrate. The bleomycin concentration chosen for this experiment had no visible effect on the growth of either wild-type or mutant plants (Figure 4e).

Figure 4.

Chromosomal tandem repeat recombination in planta.
(a–d) Frequency distribution histograms showing the proportions of plants (atercc1-1 lines 7 and 31, and AtERCC1) with a given number of blue spots in the inverted repeat [(a) without bleomycin induction, (b,c) with bleomycin induction] and in direct repeat (atercc1-1 line17 and AtERCC1) [(d) with bleomycin induction]. The schematic representation of recombination substrate is also presented. The recombination substrate consists of two overlapping fragments of the GUS gene separated by hygromycin resistance gene (hpt) either in direct or in inverted orientation.
(e) Bleomycin treatment has no visible effect on growth in the atercc1-1 (upper) and in wild-type plants (lower).

The results are summarized in Table 1(b,c) and are presented as frequency distribution histograms in Figure 4. Non-parametric statistical analysis (see Gherbi et al., 2001) confirmed the reduced recombination levels in the mutant versus wild-type plants in each case (Table 1). The atercc1-1 mutant is thus mitotic hypo-rec for these chromosomal tandem repeat recombination tests, thereby implicating this protein in recombination in the chromosomal context in Arabidopsis.

Discussion

In this study, we describe the identification and characterization of two new T-DNA insertion mutant alleles of the ArabidopsisERCC1 homologue, AtERCC1. The mutant alleles are recessive and the atercc1 plants are viable, develop normally and are fertile, in agreement with the data of Hefner et al. (2003). The atercc1 mutants presented here are UV hypersensitive. We have also confirmed that the atercc1-1 mutant plants are hypersensitive to UV-C irradiation and the radiomimetic agent bleomycin, in agreement with the gamma-ray sensitivity demonstrated by Hefner et al. (2003) for uvr7 (data not shown).

Together with the Xpf protein, the Ercc1 protein forms a structure-specific DNA endonuclease responsible for making a single-strand incision 5′ to the damaged site in NER. This endonuclease also acts in recombination and we have previously shown that the Arabidopsis Xpf/Rad1 homologue (AtRad1) acts in removing overhanging DNA tails in recombination in this plant (Dubest et al., 2002). Here, we test the involvement of the other component of this endonuclease, the AtErcc1 protein (Ercc1/Rad10 homologue) in this process. In the ‘with tails’ transformation, the number of recombinant spots in atercc1-1 leaves is significantly reduced (12-fold) compared with the wild type, while no difference was observed between the wild-type and mutant plants with ‘no tails’ transformation. This is comparable with the 10-fold decrease of recombinants in the presence of non-homologous tails that we have previously described in atrad1 mutant plants. Taken together, these data confirm that Atercc1 plays a role in recombination equivalent to that of AtRad1 and provides strong support for the conclusion that AtErcc1 and AtRad1 form a heterodimeric endonuclease, which functions in removing non-homologous overhangs in recombination in Arabidopsis.

The test we have used to assay the specific DNA ‘tail-removing’ activity of these proteins is plasmid transformation based and although it permits the demonstration of this activity, it is difficult to draw concrete conclusions concerning the importance of this in the chromosomal DNA context. We therefore tested chromosomal recombination in planta using a specific tandem-repeat assay originally developed by the group of B. Hohn (Swoboda et al., 1994). The low number of spontaneous somatic recombination events in the wild-type control plants makes it difficult to draw definite conclusions concerning the relative recombination frequency in the atercc1-1 mutant plants, which nevertheless appear to have a reduced level of recombination. However, in the case of bleomycin exposure, the atercc1-1 mutant plants show a clear absence of induced recombination compared with the wild-type controls.

Given the structure of the tandem-repeat recombination substrates, a physical exchange of flanking DNA sequences is necessary to produce a functional recombined GUS gene detected by the blue coloration. In the case of the direct repeat, this could occur by Holliday junction resolution (crossover) or single-strand annealing mechanisms. However, for the inverted repeat, only a crossover would work. As we have previously discussed for these substrates (Gherbi et al., 2001), the equivalence of the magnitude of the effects observed in the direct and inverted repeat cases implies that the events occur intra-chromosomally (within a repetition) rather than between chromatids or chromosomes. The latter two cases would lead to the recombined GUS gene residing on an acentric or dicentric chromosome and would thus not be expected to be stable. We thus conclude that the AtErcc1 protein (and presumably AtRad1) plays an important role in planta, not only in SSA recombination as measured by the plasmid assay, but also in gene-conversion/crossing over in chromosomal DNA. This conclusion is of particular interest in the light of the recent report of a direct interaction between the Xpf/Ercc1 endonuclease and the hRad52 protein in human cell-free extracts (Motycka et al., 2004) and that this complex is needed for ends-out gene targeting in mouse cells in the absence of non-homologous overhangs (Niedernhofer et al., 2001), implying a more general role for this endonuclease in recombination than the removal of non-homologous DNA overhangs from recombination intermediates.

Experimental procedures

Strains and growth conditions

Arabidopsis thaliana were Columbia ecotype. The atercc1 T-DNA insertion lines (Salk_033397 and Salk_077000) were identified by searching the T-DNA express database established by the Salk Institute Genomic Analysis Laboratory (Alonso et al., 2003). The T-DNA Express database is accessible from the SIGnAL website at http://signal.salk.edu. After surface sterilization with sodium hypochlorite (7%, w/v), Atrad10 and wild seeds (ecotype col0) were sown on ATG germination medium [Murashige and Skoog basal medium with vitamins (Sigma, St Louis, MO, USA), supplemented with 30 g l−1 of sucrose, pH 5.8]. Plates were incubated in a growth chamber at 22 °C (16 h light/8 h dark).

PCR and sequencing of T-DNA insertion sites

Plants heterozygous and homozygous for the atercc1 mutation were identified by a PCR genotyping assay. The following primer combinations were used to amplify the wild-type AtERCC1 locus, o397 (5′-ACTAAATATCCACTTA TTTTTGTTGGG-3′) and o399 (5′-GTAGAACCAAGTGTAACTACATCGC-3′). The mutant locus was identified by PCR with o399 and o405 (5′-TGGTTCACGTAGTGGGCCATCG-3′). Mutant identification was confirmed by testing UV-C sensitivity.

The following primer combinations were used to amplify DNA flanking the T-DNA in line Salk_033397: at the LB1 left border, o432 (5′-TTTTATCGTTTC CAAATAACGGGC-3′) and o406 (5′-GCGTGGACCGCTTGCTGCAACT-3′), and at the LB2 left border, o397and o406. The Salk_077000 line was analysed by PCR using the primers o406/o397. The PCR products were then purified on a QIAquick column (Qiagen, Courtaboeuf, France) and sequenced commercially (MWG Biotech, Les Ulis, France).

RT-PCR

Tissue sample of 80–100 mg was used for RNA isolation from 3-week-old plants according to directions provided in the RNeasy Plant Mini Kit (Qiagen). After treatment with RNase-free, DNaseI (Roche, Meylan, France), analysis of gene expression was performed using the Titan One Tube RT-PCR kit as directed by the manufacturer (Boehringer-Mannheim GmbH, Mannheim, Germany). The primer pairs used for the analysis of AtERCC1 expression were: (i) o435 (5′ GGCGCTT CTCAGGTGCCCC 3′) and o436 (5′ GATATGTTTAAGAAGCGGGTTCC 3′) for amplification of the 5′AtERCC1; (ii) o405 and o432. The APT1 transcript (adenine phosphorybosyl transferase) (Moffatt et al., 1994) was used as a control for mRNA quantity and quality and was amplified with primers apt1 (5′-TCCCAGAATCGCTAAGATTGC-3′) and apt2 (5′-CCTTTCCCTTAA-GCTCTG-3′).

UV sensitivity test

Ten-day-old seedlings, germinated as described above in plates placed vertically to allow roots to grow downward on the surface of the agar, were exposed to 300 J m−2 of UV-C light (GS Gene linker; Bio-Rad, Hercules, CA, USA). These were then wrapped in aluminium foil and replaced in the growth chamber in the dark in order to avoid photoreactivation. The growth of the main root of each seedling was measured daily, up to day 3.

Plasmid constructs

As the control for transformation efficiency, we used plasmid pGUS23 (Puchta and Hohn, 1991), which contains the cauliflower mosaic virus (CaMV) 35S promoter, the GUS gene and the nopaline synthetase (NOS) transcription terminator. We have previously described the construction of plasmids pCW344 and pCW345 (Dubest et al., 2002).

Biolistic transformation and recombination assay

Leaves from 3- to 4-week-old wild-type and atercc1-1 mutant plants were placed on solid germination medium in Petri dishes. Gold beads (Sigma; 0.6 μm diameter) were washed in water, then rinsed and resuspended in ethanol at 1 μμl−1. Fifty microlitres (50 μg) of plasmid DNA, 250 μl of 2.5 m CaCl2 and 100 μl of 1 m spermidine (Sigma) were added to 50 μl of beads. After 5 min on ice, the beads were pelletted with a brief centrifugation and 400 μl of supernatant removed, leaving the beads in a volume of 50 μl. Five microlitres (1/10th) of the resuspended, DNA-coated gold beads were used for each transformation with a helium particle gun – 6 bars, 28 kPa, 13 cm from the leaves (Finer et al., 1992). Leaves in water were then placed in a growth chamber at 22°C (16 h light/8 h dark) for 1 or 2 days and the GUS activity assay carried out as described by Jefferson et al., (1987). The number of blue spots per leaf was determined by visual observation using a dissecting microscope. Transformation efficiency was controlled by transforming wild and mutant plants with 25 μl (25 μg) of pGUS23. Plasmids pCW344 and pCW345 were linearized with Bsm1 or EcoRI, respectively, to prepare the ‘no-tail’ recombination substrate. ‘With-tail’ substrates were prepared by digesting pCW344 and pCW345 with ScaI or SpaI, respectively. In all cases, the overlapping homology region is 590 bp to which pCW344/ScaI adds a non-homologous tail of 816 bp and pCW345/SpaI a tail of 233 bp. In both ‘with-tails’ and ‘no-tails’ experiments, the transformation was carried out with 25 μl (25 μg) of each plasmid (pCW344 and pCW345). The statistical analysis was performed as described previously (Gherbi et al., 2001).

In plantarecombination assays

The recombination substrates (Swoboda et al., 1994) consist of two fragments of the GUS gene with 618 bp of overlap, either in direct (pGU.US) or inverted orientation (pU'G’.US). We have previously described the Arabidopsis lines used which carry single copies of these substrates integrated into chromosomal DNA (Gherbi et al., 2001). The two GUS lines were crossed to the homozygous atercc1-1Arabidopsis mutant. The hygromycin resistance markers and PCR with primers o397/o399 (wild-type AtERCC1 locus) and o399/o406 (atercc1-1 locus) were used to identify the F2 homozygous for the GUS recombination substrate and homozygous for either the atercc1 or the AtERCC1 alleles.

F3 seeds from this line were placed in ATG germination medium for 7 days and then transferred to liquid ATG (spontaneous recombination) or to ATG medium with 0.05 μg ml−1 of bleomycin (inducted recombination) for 7 days. The GUS histochemical staining and the statistical analysis were performed as described (Gherbi et al., 2001).

Acknowledgements

We thank the members of the recombination mechanisms group and BIOMOVE for their help and discussions. Thanks are also due to the group of Michel Bernard at INRA for kindly giving us access to their helium canon. This work was partially financed by the CNRS, Université Blaise Pascal and an European Community Research grant (QLG2-CT-2001-01397). S. Dubest is supported by a MENRT doctoral fellowship.

NASC accession numbers: N533397 (Salk_033397); N577000 (Salk_077000).

Ancillary