NBS1 is involved in DNA repair and plays a synergistic role with ATM in mediating meiotic homologous recombination in plants


(fax +44 (0) 113 343 3144; e-mail c.e.west@leeds.ac.uk).


The ability of plants to repair DNA double-strand breaks (DSBs) is essential for growth and fertility. The Arabidopsis DSB repair proteins AtRAD50 and AtMRE11 form part of an evolutionarily conserved complex that, in Saccharomyces cerevisiae and mammals, includes a third component termed XRS2 and NBS1, respectively. The MRN complex (MRX in yeast) has a direct role in DSB repair and is also required for DNA damage signaling and checkpoint activation in a pathway mediated by the protein kinase ATM. This study characterizes Arabidopsis and maize NBS1 orthologues that share conserved protein motifs with human NBS1. Both plant NBS1 proteins interact with the corresponding MRE11 orthologues, and deletion analysis of AtNBS1 defines a region towards the C-terminus (amino acids 465–500) that is required for interaction with AtMRE11. Arabidopsis lines homozygous for a T-DNA insertional mutation in AtNBS1 display hypersensitivity to the DNA cross-linking reagent mitomycin C, and this phenotype can be rescued by complementation with the wild-type gene, consistent with a function for AtNBS1 in plant DSB repair. Analysis of atnbs1-1 atatm double mutants revealed a role for AtNBS1 in meiotic recombination. While atatm mutants produce reduced seed numbers, plants deficient in both AtATM and AtNBS1 are completely infertile. Cytological analysis of these double mutants revealed incomplete chromosome pairing and synapsis in meiotic prophase, and extensive chromosome fragmentation in metaphase I and subsequent stages. These results suggest a novel role for AtNBS1 that is independent of AtATM-mediated signaling and functions in the very early stages of meiosis.


As sessile, photosynthetic organisms, plants are necessarily exposed to high levels of environmental stresses. Effective mechanisms have evolved to cope with DNA damage, including cell-cycle delay or arrest and activation of DNA repair pathways (Zhou and Elledge, 2000). These pathways are of additional importance in plants as somatic cells give rise to germ cells at a relatively late stage in development. Of the different forms of DNA damage, double-strand breaks (DSBs) are one of the most cytotoxic, and result in chromosome fragmentation and loss of genetic information if left unrepaired. DSBs also play an important role during meiosis, where induced breaks facilitate the alignment and crossing over of homologous chromosomes in most organisms. In plants, meiotic DSBs are repaired by homologous recombination (HR), whereas the majority of DSBs occurring in somatic cells are repaired by non-homologous end joining (NHEJ; Puchta, 2005). In eukaryotes, the response to DSBs involves a cascade of signaling events that lead to recruitment of repair proteins to the damage site and arrest of the cell cycle while the damage is repaired. The phosphoinositide 3-kinase-like protein kinases (PIKK) ATM (ataxia telangiectasia mutated) and ATR (ATM- and RAD3-like) play a central role in coordinating these events, regulating the activities and localization of a number of proteins involved in DNA repair and cell-cycle control (Falck et al., 2005).

The evolutionarily conserved MRE11 and RAD50 proteins form a complex that is involved in the repair of DNA DSBs (D’Amours and Jackson, 2002; Daoudal-Cotterell et al., 2002). The nuclease and helicase activities of MRE11 may be important in processing broken ends before they are rejoined, and RAD50 has long coiled-coil domains that interact with one another, bridging broken ends and holding them together (Hopfner et al., 2002). In mammalian cells, a third component of this complex, NBS1, is important in controlling the activity and localization of the complex and also functions in DNA damage signaling (Tauchi et al., 2002). NBS1 was characterized as the genetic basis of human Nijmegen breakage syndrome, a rare disease associated with chromosomal instability. Null mutations of Nbs1 are lethal in mammals, but humans homozygous for a hypomorphic Nbs1 allele display hypersensitivity to irradiation resulting from an S-phase checkpoint defect (Maser et al., 2001). NBS1 links the detection of DNA damage to signaling events by recruiting ATM to the DSB (Falck et al., 2005; Lee and Paull, 2005). ATM activation leads to phosphorylation of numerous nuclear proteins including ATM itself, NBS1, CHK1, CHK2 and the histone 2A isoform H2AX in a cascade that results in cell-cycle arrest and activation of DNA repair pathways (Lavin et al., 2005).

Null mutations in any component of the MRE11/RAD50/NBS1 (MRN) complex result in lethality in mammals (Stracker et al., 2004), whereas Arabidopsis mre11 and rad50 mutants are viable but show defects in DNA repair resulting in hypersensitivity to the alkylating agent methyl methanosulfonate (MMS; Gallego et al., 2001; Bundock and Hooykaas, 2002). The MRN complex also functions in meiosis, where it has been implicated in resecting exposed DNA ends to produce 3′ single-stranded DNA tails. These form a substrate for RAD51-mediated homology search and strand invasion, facilitating the alignment and crossing over of homologous chromosomes (Aylon and Kupiec, 2004). The essential functions of AtMRE11 and AtRAD50 in meiotic HR means that Arabidopsis mutants deficient for these genes fail to form viable gametes (Gallego et al., 2001; Bundock and Hooykaas, 2002; Bleuyard et al., 2004; Puizina et al., 2004).

Although significant progress has been made in the study of AtRAD50, AtMRE11 and core components of the Arabidopsis NHEJ pathway, an Arabidopsis NBS1 orthologue has only recently been reported (Akutsu et al., 2007). Here we provide novel information on the function of NBS1 in plant cells through the analysis of an atnbs1-1 T-DNA insertional mutant. Phenotypic analysis of this line is consistent with a role for this protein in DNA repair in vegetative tissues. Furthermore, we also demonstrate a meiotic role for AtNBS1, which is essential for the residual fertility observed in atatm mutants.


Sequence analysis of plant NBS1 orthologues

A putative AtNBS1 orthologue was identified by blast search of the Arabidopsis genomic sequence database as the initial genome annotation failed to identify the full-length AtNBS1 open reading frame, with only the last two exons predicted to be protein-coding. AtNBS1 is predicted to encode a 60 kDa, 542 amino acid polypeptide with a pI of 5.1 (GenBank accession no. DQ167217, AGI code At3g02680), and maize NBS1 (ZmNBS1) encodes a 558 amino acid protein (GenBank accession no. DQ841175). The AtNBS1 cDNA isolated in the current study is 27 bp longer than that reported previously (Akutsu et al., 2007), resulting from differences in splicing at the junction of intron 9 and exon 10. Although the shorter version was never detected in our studies, it may represent the possibility of splicing isoforms of NBS1 in Arabidopsis. The nine amino acids missing in accession ABK59968 (Akutsu et al., 2007) are in the central region of the AtNBS1 protein sequence (Figure 1) and are conserved in the rice and maize orthologues.

Figure 1.

 Alignment of Arabidopsis NBS1, maize NBS1, rice putative NBS1 and human NBS1 protein sequences.
Protein sequences were aligned using Clustal W, the alignment was hand-finished using JalView and annotated using Boxshade. Black boxes are identical residues, gray boxes are conservative substitutions aligned in two or more sequences. The GenBank accession nos are 4505339 (human NBS1), DQ167217 (Arabidopsis NBS1), DQ841175 (maize NBS1) and 18087867 (rice NBS1-like). FHA and BRCT domains were identified using RPS-BLAST (Marchler-Bauer and Bryant, 2004).

Analysis of both predicted Arabidopsis and maize NBS1 protein sequences (Figure 1) reveals a significant match (e < 10−8) to the consensus forkhead-associated (FHA) domain common to mammalian NBS1 and S. cerevisiae XRS2 (D’Amours and Jackson, 2001; Zhao et al., 2002; Marchler-Bauer and Bryant, 2004). Mammalian NBS1 proteins have a domain with a weak match to a consensus BRCT (BRAC1 C-terminal) domain adjacent to the FHA domain; AtNBS1 and ZmNBS1 also show only a weak match with the BRCT domain consensus, and a similarly distant match to the mammalian NBS1 sequences (e = 0.11, residues 127–194). The NBS1 orthologues from different species display the greatest sequence divergence of the MRN complex members. Arabidopsis and human NBS1 share 45% similarity over the N-terminal 246 amino acids, which include the FHA and BRCT domains. Yeast XRS2 displays no significant sequence similarity with Arabidopsis or human NBS1 proteins in simple gapped sequence alignments.

Sequence analysis also identified a Ser-Gln (SQ) motif that is conserved in all putative plant NBS1 homologues represented in EST and genome sequence databases (Figure 1, and data not shown). This sequence motif has been identified in mammals as a consensus ATM phosphorylation site, and is found in a number of proteins involved in DNA repair and intracellular DNA damage signaling (Lavin et al., 2005). Plant and mammalian NBS1 orthologues contain conserved regions that are involved in interaction with MRE11 (detailed below) and ATM, including a conserved phenylalanine that is important in the interaction between Nbs1 and HEAT repeats of mammalian ATM (You et al., 2005).

Interaction between plant NBS1 and MRE11 orthologues

Yeast two-hybrid analysis was used to investigate interactions between AtNBS1, AtMRE11 and AtRAD50 (Figure 2). Strong lacZ expression was found in yeast expressing AtRAD50–GAL4-DB and AtMRE11–GAL4-AD fusion proteins (Figure 2a), confirming the previous report of interaction between these two proteins in Arabidopsis cell extracts (Daoudal-Cotterell et al., 2002). Specific interaction was also seen between AtMRE11 and AtNBS1(217-542), but no interaction was observed between AtRAD50 and AtNBS1(217-542).

Figure 2.

 Yeast two-hybrid analysis of NBS1 interaction with MRE11.
(a) AtNBS1 (residues 217–542), AtMRE11, AtRAD50 or control proteins were expressed as GAL4 DNA binding domain (GAL4-DB) or GAL4 activation domain (GAL4-AD) fusion proteins. Interaction between DB/AD fusions results in β-galactosidase expression.
(b) Interaction of AtNBS1 deletions with AtMRE11 or controls.
(c) Schematic of maize MRE11 genes. Exons are shown as boxed regions. Gray shaded boxes are untranslated regions. White boxes encode the DNA binding domain of ZmMRE11. The cross-hatched region is unique to ZmMRE11B.
(d) Interaction between ZmNBS1 and ZmMRE11 resulting in histidine prototrophy.
(e) Radiolabelled AtNBS1 full-length protein or polypeptide regions were applied to AtMRE11-coated beads (+) or control beads incubated with E. coli cell extract (−). Interaction with AtMRE11 was detected by SDS–PAGE and autoradiography.
(f) Schematic comparing regions of AtNBS1 analyzed in the in vitro interaction studies and the constructs used in the yeast two-hybrid analysis. The minimal region of AtNBS1 required for interaction with AtMRE11 is indicated.

A series of AtNBS1 deletion constructs were tested to further characterize the region of AtNBS1 mediating AtMRE11 interaction (Figure 2b). Removal of the C-terminal 98 amino acids of AtNBS1 (AtNBS1(217-444)) abolished interaction with AtMRE11 (Figure 2b), although a C-terminal 53 amino acid fragment of AtNBS1 (AtNBS1(490-542)) was insufficient for interaction with AtMRE11. A larger C-terminal fragment representing the C-terminal third of AtNBS1 (AtNBS1(362-542)) did interact with AtMRE11, and was further characterized by in vitro interaction studies (see below).

Yeast two-hybrid analysis indicated similar interactions between ZmNBS1 and ZmMRE11 orthologues. Maize has two putative Mre11 homologues, termed ZmMRE11A and ZmMRE11B (Figure 2c; C.A and C.F.W., unpublished data) [accession numbers EF584428 (ZmMRE11A) and EF599099 (ZmMRE11B)]. A splice variant of ZmMRE11A has also been detected that would encode a truncated protein, and this had previously been submitted to the sequence database (P.B. Mahajan, Pioneer Hi-Bred, accession number AX287091). For ZmMRE11A, interaction was detected with ZmNBS1 but not a ZmNBS1(1-433) C-terminal deletion construct, consistent with removal of the MRE11 interaction domain (Figure 2d). No interaction was observed between ZmNBS1 and either ZmMRE11B or the truncated MRE11A splice variant (Figure 2d). This indicates a phylogenetically conserved interaction between NBS1 and MRE11 in plants. Both ZmMRE11B and the truncated MRE11A splicing variant in maize lack the C-terminal domain of ZmMRE11A, which may therefore be involved in the interaction with ZmNBS1. The functions of the maize MRE11 isoforms that do not interact with ZmNBS1 are the subject of ongoing studies.

The region of AtNBS1 that mediates AtMRE11 interaction was further defined using in vitro binding assays. AtNBS1 peptide domains were produced by in vitro transcription/translation, incubated with paramagnetic beads coated with AtMRE11 or uncoated controls, and analyzed by SDS–PAGE and autoradiography (Figure 2e). N-terminal deletions of AtNBS1 (362–542, 420–542 and 465–542) and C-terminal deletions up to residues 525 and 500 retained the ability to bind AtMRE11 in vitro (Figure 2e), whereas deletion of the N-terminal 473 amino acids abolished this interaction (Figure 2e). These data indicate that residues 465–525 of AtNBS1, including the conserved FKR/KFR/KK motif, is the minimal region required for interaction with AtMRE11 (Figure 2f).

Identification of an insertional atnbs1-1 mutant line

An atnbs1-1 T-DNA insertion line was identified in the GABI-KAT collection (Rosso et al., 2003), and both sulphadiazine resistance and Southern blot analysis were consistent with a single T-DNA insertion in this line (Figure 3a,b, and data not shown). The 5.8 kb T-DNA was inserted such that the left border was located at boundary of intron 4 and exon 5, with four bases of microhomology between the T-DNA and the AtNBS1 genomic sequence (Figure 3c). The right border insertion site caused a 17 bp deletion of exon 5 and included a 27 bp insertion most of which was probably derived from the AtNBS1 sequence, with a 16 bp fragment copied from the region adjacent to the left border and a 10 bp fragment that may be copied from the first intron. No full-length AtNBS1 transcript could be detected in the homozygous mutant background, although a 5′ transcript was expressed and is predicted to encode an N-terminal fragment of AtNBS1 including the FHA domain and half the BRCT domain (data not shown). The pAC161 T-DNA used to make this mutant carries the CaMV 35S promoter at the right border for activation tagging studies. Chimeric AtNBS1 transcript originating from this promoter could be detected at approximately 2000-fold higher levels than endogenous AtNBS1 (Figure 3d). The first ATG of this chimeric mRNA is out of frame, but is in a weak Kozak context raising the possibility that translation could initiate from the second ATG (Figure 3c). The second ATG is in-frame, and encodes a protein starting within the AtNBS1 BRCT domain. Translation of this downstream AtNBS1 fragment occurred both in vitro (Figure 3e) and in vivo; a fusion of the chimeric 3′ fragment to a GFP reporter shows that high levels of expression are initiated at the second AUG in tobacco cells (Figure 3f,g). The N-terminal truncated AtNBS1 protein lacks both the FHA and BRCT domains but retains the AtMRE11 and putative AtATM interaction motifs. The atnbs1-1 mutant closely resembles mammalian NBS1 hypomorphic mutants in which internal translation initiation of the NBS1 mRNA results in an N-terminal truncated NBS1 fragment that is sufficient to rescue the lethality observed in true NBS1 null lines (Maser et al., 2001).

Figure 3.

 Position of T-DNA insertion in atnbs1-1.
(a) Schematic showing the arrangement of the T-DNA in the AtNBS1 gene, with introns shown as a line, exons as boxes and the 3′ untranslated region in gray. LB, T-DNA left border; RB, T-DNA right border.
(b) Position corresponding to the T-DNA insertion in the predicted AtNBS1 protein sequence.
(c) Sequence analysis of the left and right border regions, with the four-base microhomology underlined, the vector sequence shown in bold, the exon sequence shown in capitals, and the insertions in italics. The AtNBS13′ transcript obtained by RT-PCR using atnbs1-1 RNA is shown translated in three frames with potential start codons boxed.
(d) Real-time PCR quantification of AtNBS1 3′ transcript levels in atnbs1-1 mutant and Col-0 plants normalized to actin levels.
(e) In vitro translation of AtNBS1 3′ transcript and full-length AtNBS1.
(f) Transient expression of the AtNBS1 3′ transcript–GFP fusion in tobacco cells analyzed by confocal microscopy. Scale bar = 20 μm.
(g) Schematic of the T-DNA–AtNBS1–GFP fusion construct.

Phenotypic characterization of atnbs1-1

Homozygous atnbs1-1 mutants displayed normal growth under standard conditions and no defects in fertility, in contrast to atmre11 and atrad50 mutants, which are sterile (Gallego et al., 2001; Bundock and Hooykaas, 2002). No differences were observed between atnbs1-1 mutant and wild-type plants in the response to UV-C (up to 3 kJ m−2) or to 300 Gy X-rays, which is lethal to the NHEJ mutants atku80 and atku70 (data not shown). However, when grown in the presence of MMS, atnbs1-1 plants were significantly smaller, with reduced root growth (Figure 4a) relative to wild-type controls, indicating a role for AtNBS1 in DNA repair. This was further investigated using the bi-functional alkylating agent mitomycin C (MMC), which reacts with DNA to produce monoadducts and DNA crosslinks. Homozygous atnbs1-1 mutants displayed clear MMC hypersensitivity resulting in significantly reduced mass (Figure 4b; P < 0.05) compared with Col-0 controls. In addition, significantly more atnbs1-1 mutants failed to produce true leaves after 3 weeks’ growth in the presence of MMC compared with wild-type (Figure 4c). Complementation with a genomic clone encompassing the AtNBS1 coding region and 1.6 kb of promoter resulted in significantly reduced hypersensitivity to MMC (P < 0.05; Figure 4d) relative to atnbs1-1, and restored growth to near wild-type levels (Figure 4d, e).

Figure 4.

 Phenotypic analysis of atnbs1-1 mutants exposed to MMS and MMC.
(a) Root length of atnbs1-1 mutant (squares) and Col-0 (circles) seedlings exposed to MMS.
(b) Mass (fresh weight) of atnbs1-1 mutant (squares) and Col-0 (circles) plants grown in the presence of various concentrations of MMC. Error bars represent the standard error for six replicates of ten plants for both (a) and (b).
(c) Plants were scored for the presence of true leaves >1 mm in size. Black bars, atnbs1-1; white bars, Col-0.
(d) Fresh weight of plants grown in the presence of 8 or 10 μg ml−1 MMC for 3 weeks as indicated. White bars, Col-0; black bars, atnbs1-1; gray bars, complemented atnbs1-1 AtNBS1. Error bars represent the standard error for six replicates of ten plants for both (c) and (d).
(e) Appearance of plants exposed to 8 μg ml−1 MMC for 3 weeks (bar = 1 cm).
All experiments were performed independently apart from (e) which shows representative groups of ten plants from (d).

Characterization of an atnbs1-1 atatm-3 double mutant

The wild-type fertility of atnbs1-1 mutants differed from the severe meiotic defects observed in atrad50 and atmre11 lines. This was confirmed by cytological analysis of atnbs1-1 pollen mother cells, which indicated normal progression through meiosis (data not shown). To investigate whether AtATM compensates for AtNBS1 deficiency, atnbs1-1 plants were crossed into an atatm-3 mutant line. The atatm-3 mutant (SALK 089805) contains a T-DNA insertion in exon 20 (Figure 5a). PCR and Southern analysis of homozygous atatm-3 mutant plants confirmed the presence of a single T-DNA insertion that knocked out AtATM expression (Figure 5b,c, and data not shown). The left border of the T-DNA insertion was inserted into exon 20 at a two-base microhomology. The right border included 693 bp of pROK2 vector sequence beyond the right border and joined intron 19 of AtATM, resulting in a 20 base deletion and insertion (Figure 5c). Null atatm-3 mutants displayed similar phenotypes comparable to those previously documented for atatm-1 and atatm-2 (Garcia et al., 2003), including hypersensitivity to X-rays (Figure 5d) and failure to induce AtRAD51 transcript 1 h after X-irradiation (10 Gy; Figure 5e). The two mutant lines were crossed to generate an atnbs1-1 atatm-3 double mutant, which was viable and showed normal vegetative growth. While atatm mutants show meiotic defects, including reduced fertility with an approximately 90% reduction in seed numbers (Garcia et al., 2003), the atnbs1-1 atatm-3 double mutants displayed complete sterility, producing very small siliques devoid of seeds (Figure 5f).

Figure 5.

 Characterization of the atatm-3 mutant phenotype and analysis of atnbs1-1 atatm-3 double mutants.
(a) Schematic of the AtATM genomic region with exons shown as boxed regions. The positions of the atatm-1 and atatm-2 alleles (Garcia et al., 2003) and the atatm-3 allele (this study) are shown.
(b) PCR analysis of homozygous atatm-3 mutants showing the absence of wild-type AtATM allele (left panel) and the absence of AtATM transcript (right panel).
(c) Sequence analysis of the T-DNA border regions in atatm-3. Bold font indicates T-DNA or pROK2 vector sequence, AtATM exon 20 is in capitals, and a 20 bp insertion is shown in italics.
(d) Col-0 and atatm-3 mutants 3 weeks after treatment with 100 Gy X-rays, and untreated controls.
(e) Real-time RT-PCR analysis of AtRAD51 transcript levels (normalized to Actin2) in Col-0 and atatm-3 plants 1 h after a 10 Gy X-ray dose, or untreated controls. Standard errors are shown.
(f) Mature siliques from mutant and wild-type plants as indicated. From left to right: atatm-3 AtNBS1+/−, atatm-3, atnbs1-1 atatm-3 double mutants and Col-0.
(g) Mean numbers of seeds per silique in atatm-3, AtNBS1+/−atatm-3 and atnbs1-1−/−atatm-3 double mutants. Standard errors are shown.

The atnbs1-1 meiotic phenotype is recessive

The atnbs1-1 allele showed normal 1:2:1 segregation (wild-type: heterozygous: atnbs1-1 mutant) in the atatm-3 background, and the sterility phenotype showed a 1:3 distribution, correlating with homozygosity of the atnbs1-1 allele. This is consistent with the recessive nature of the mutation, rather than a dominant negative effect of overexpression of the 3′ end of AtNBS1. The possibility of a weak dominant effect was further investigated by comparison of seed numbers between atatm-3 plants that were heterozygous or wild-type for the atnbs1-1 allele (Figure 5g). Interestingly, plants homozygous for atatm-3 and heterozygous for the atnbs1-1 mutation produced, on average, a reduced number of seeds per silique compared with atatm-3 plants, although this effect was of marginal statistical significance (P = 0.053) due to high variability between siliques. This analysis indicates that there is no strong dominant effect due to the atnbs1-1 mutant allele, but does not rule out a slight reduction in seed numbers in atatm-3 plants heterozygous for the atnbs1-1 allele. Complementation of the atnbs1-1 DNA repair defects by re-introduction of a AtNBS1 genomic construct also demonstrates the recessive nature of the atnbs1-1 mutation and confirms the absence of any strong dominant effects from overexpression of the AtNBS1 3′ fragment (Figure 4d,e).

Cytological analysis of Col-0, atatm-3 and atnbs1-1 atatm-3 lines

Further investigation into the failures in gametogenesis underlying the sterility of the atnbs1-1 atatm-3 line was performed by analysis of meiotic events in pollen mother cells, and comparison with wild-type and atatm-3 mutants. Previous cytological analysis of atatm plants revealed variable chromosome fragmentation (column 2 in Figure 6), but otherwise normal progression through meiosis (Garcia et al., 2003). In contrast, atnbs1-1 atatm-3 plants display severely disrupted meiosis (column 3 in Figure 6). Cytological analysis revealed disruption of chromosome pairing in the double mutants (arrowed in Figure 6f), consistent with defects in HR. Those meiocytes that progressed through prophase displayed extensive chromosome fragmentation (Figure 6i) compared with atatm mutants (Figure 6h). This resulted in severe defects in metaphase II and tetrad formation in atnbs1-1 atatm-3 plants (Figure 6l,o).

Figure 6.

 Cytological analysis of meiosis in wild-type, atatm-3 mutant and the atnbs1-1 atatm-3 double mutant.
Meiotic spreads were prepared from pollen mother cells and stained with 4′,6-diamidino-2-phenylindole. Male meiosis in wild-type (a, d, g, j, m), atatm-3 (b, e, h, k, n) and atnbs1-1 atatm-3 (c, f, i, l, o).
(a–c) Leptotene stages are apparently normal in all cases.
(d–f) WT pachytene showing complete synapsis: (e) alignment and incomplete synapsis; (f) incomplete synapsis (arrowed).
(g–i) Metaphase I: (g) the wild-type cell shows five bivalents aligned on the metaphase plate; (h) a fragment (arrowed) is present in the atatm-3 mutant; (i) extensive fragmentation is present in the atnbs1-1 atatm-3 mutant.
(j–l) Metaphase II: (j) the wild-type has two sets of five chromosomes; (k) the atatm-3 mutant shows some fragmentation; (l) the atnbs1-1.atatm-3 shows extensive fragmentation.
(m) Normal tetrad formation in wild-type.
(n) The atatm-3 mutant with some fragmentation but possible normal chromosome complement (circled).
(o) The double mutant, demonstrating extreme fragmentation.


DNA repair mechanisms and DNA damage sensing and signaling pathways play essential roles in maintaining genome stability against a constant background of environmentally and endogenously induced DNA damage (Bray and West, 2005). DNA repair mutants that cause lethality in animals are often viable in plants (Stracker et al., 2004; Bray and West, 2005), which thereby provide a useful model system. The present study provides a functional characterization of plant NBS1 in DNA repair and identifies an important role for AtNBS1 in the early stages of meiosis.

The predicted protein sequences of maize and Arabidopsis NBS1 orthologues contain several conserved protein interaction motifs including FHA and BRCT domains. The FHA domain of yeast XRS2 is required for interaction with LIF1, the yeast orthologue of the mammalian and plant NHEJ protein XRCC4 (Palmbos et al., 2005), and in humans is required to localize MRE11 to sites of DNA damage (Desai-Mehta et al., 2001). The MRE11 interaction region of AtNBS1 contains an FKXFXK motif, and the two conserved lysine residues of this motif were recently shown to be important for the interaction between S. cerevisiae XRS2 and MRE11 (Shima et al., 2005). In mammals, NBS1 is involved in ATM-mediated DNA damage signaling, whereby NBS1 recruits ATM to the DSB site and activated ATM subsequently phosphorylates target proteins on Ser residues adjacent to Gln (SQ). Putative plant NBS1 protein sequences display absolute conservation of the SQ motif located at S275Q276 in AtNBS1 (Figure 1), raising the possibility of post-translational regulation of AtNBS1 activity. Arabidopsis PIKKs have been shown to be required for DNA damage signaling in vegetative tissues (Garcia et al., 2003; Culligan et al., 2004), with AtATM largely responsible for the transcriptional activation of genes in response to DSBs, and AtATR involved in DNA damage checkpoint activation in response to replication blocks. In γ-irradiated Arabidopsis cells, PIKK activation results in H2AX phosphorylation (Friesner et al., 2005), which functions to recruit DNA repair factors to the site of DNA damage and may amplify the DNA damage response (van Attikum et al., 2004).

The role of AtNBS1 in the response to DNA damage was further investigated by analyzing plants containing an activation tagging T-DNA inserted in the region of AtNBS1 encoding the BRCT domain. The T-DNA resulted in ectopic expression of a chimeric 3′AtNBS1 mRNA that produced an N-terminal truncated protein similar to an NBS1 fragment that allows growth and fertility in mammalian NBS1 hypomorphic mutants. The chimeric transcript showed very high levels of expression, although the first AUG codon was out of frame, raising the possibility that the mutant mRNA would not be translated. However, when the chimeric transcript was expressed as a GFP fusion in plant cells, intense GFP fluorescence was observed, consistent with high levels of translation (Figure 3f). In vitro data show that the AtNBS1 fragment retains the ability to bind AtMRE11, and sequence analysis indicated that the truncated AtNBS1 protein has a putative AtATM interaction domain but lacks the FHA and BRCT domains. The truncated AtNBS1 may retain some function that is responsible for the fertility of homozygous atnbs1-1 mutants. This would mirror the situation in mammals where an NBS1 fragment is required for growth in the otherwise non-viable Nbs1−/− mutants. Phenotypic analysis of atnbs1 mutants can therefore reveal the effects of the absence (or uncoupling) of the FHA and BRCT domains from the rest of AtNBS1 on plant growth and fertility.

Homozygous atnbs1-1 plants were not markedly hypersensitive to X-rays, in contrast to atmre11, atku70, atku80 and mammalian Nbs1ΔB/ΔB hypomorphic mutants (Bundock and Hooykaas, 2002; Riha et al., 2002; West et al., 2002). However, the hypersensitivity of atnbs1-1 to MMS was similar to the impaired growth observed with atmre11 and atrad50 mutants (Gallego et al., 2001; Bundock and Hooykaas, 2002). The difference in sensitivity to MMS compared with X-rays presumably reflects differences in the spectrum of DNA lesions and mechanisms of their repair. Mutant atnbs1-1 plants also displayed MMC hypersensitivity, and this growth inhibition could be rescued by complementation with the wild-type AtNBS1 gene (Figure 4e). MMC induces inter-strand DNA cross-links, and the repair of these lesions involves HR as indicated by a requirement for the RAD51 paralogues (Bleuyard et al., 2005). The defective responses to DNA damage observed in the hypomorphic atnbs1 lines demonstrate a role for AtNBS1 in DNA repair, and reflect the importance of the N-terminal domains of AtNBS1 for survival in conditions of genotoxic stress. A true atnbs1 null line in which AtMRE11 and AtATM complex formation is abolished is likely to display a more severe hypersensitivity to genotoxic agents including X-rays, as observed with atatm and atmre11 mutant lines.

The fertility of atnbs1-1 plants contrasts with the sterile phenotype of atmre11 and atrad50 mutants, and suggests that the FHA and BRCT domains of AtNBS1 are not required for meiotic HR (Gallego et al., 2001; Bundock and Hooykaas, 2002; Bleuyard et al., 2004; Puizina et al., 2004). This differs from the yeast XRS2 which requires the N-terminus for meiotic functions (Tsukamoto et al., 2005), including the MRX-dependent induction of meiotic DSBs (Shima et al., 2005). The N-terminal region of XRS2 also plays a minor role in the subsequent processing of meiotic DSBs (Shima et al., 2005). The role of mammalian NBS1 in meiosis is less clear as it is an essential gene, but meiosis is unaffected in most human and mouse hypomorphic NBS1 mutants (Kang et al., 2002; Williams et al., 2002), as is the case in hypomorphic RAD50ss mutants (Bender et al., 2002).

Interestingly, in Atm−/− null mice, the Nbs1ΔB/ΔB hypomorphic mutation results in lethality. This additive effect of the two mutations is consistent with functions of NBS1 that are independent of ATM signaling (Williams et al., 2002). To investigate this further in Arabidopsis, atnbs1-1 atatm-3 double mutants were isolated. In contrast to the mammalian studies, these plants displayed normal vegetative growth but were completely sterile. The sterile phenotype conferred by the atnbs1-1 allele was recessive, confirming that no strong dominant effect arises from the high expression levels of the 3′ fragment of AtNBS1. Complete failure of the atnbs1-1 atatm-3 mutant line to produce any seeds was indicative of more severe meiotic failure than that observed in atatm single mutants. This was confirmed by a cytological analysis of gametogenesis, which revealed extensive chromosome fragmentation and incomplete chromosome pairing in the double mutants (Figure 6). The increased severity of the double mutant phenotype compared to the atatm single mutant line could possibly result from an additive effect of accumulated DNA damage in the atnbs1 background, coupled with the defective checkpoint activation in atatm plants. However, cytological analysis of atnbs1 meiocytes revealed no evidence of DNA fragmentation, and fertility of the mutants was indistinguishable from wild-type plants. This indicates that the increased severity of the atnbs1 atatm double mutant is attributable to roles for AtNBS1 that are independent of AtATM-mediated signaling. This is similar to the observation of the synthetic lethality of the mammalian Atm−/−Nbs1ΔB/ΔB double mutant, although the non-viability of the mammalian mutants prevents analysis of possible meiotic defects (Williams et al., 2002). The mammalian study concluded that the ATM-independent role of NBS1 may be to mediate ATR-dependent signaling (Williams et al., 2002). In agreement with this, recent studies found that, in yeast, XRS2 is required for MEC1 activation in response to DSBs (Grenon et al., 2006), and, in mammals, NBS1 is involved in ATR activation in response to replication stress (Stiff et al., 2005). Our finding that the absence of both AtATM and AtNBS1 leads to very severe meiotic defects is consistent with a novel role of NBS1 in the control of meiotic events in plants independent of AtATM signaling. The meiotic defects may result from a failure to activate checkpoints that regulate passage through meiosis. However, the incomplete chromosome pairing observed in the atnbs1-1 atatm-3 background is indicative of defective recombination. This suggests that, in the absence of AtATM and AtNBS1, the MRN complex may not be activated/recruited, leading to reduced DSB processing and meiotic failure. While these studies have revealed a role for AtNBS1 that is independent of AtATM, the atnbs1-1 line provides no data on the role of AtNBS1 in mediating AtATM signaling, due the expression of the AtNBS1 C-terminal fragment. Further studies using null atnbs1 mutant lines are required to determine the function of AtNBS1 in the DNA damage response initiated by AtATM.

Here we report the functional characterization of NBS1 in plants. The interaction of plant NBS1 and MRE11 orthologues is consistent with a role for the MRN complex in recognition and repair of DNA DSBs in planta. Phenotypic analysis suggests a role for AtNBS1 in HR-mediated DSB repair in both somatic and meiotic cells, and we provide evidence that, in addition to putative ATM-mediated signaling, AtNBS1 mediates ATM-independent signaling events that are important in controlling recombination events early in meiosis.

Experimental procedures

Plant material and growth conditions

Arabidopsis (Col-0) plants were raised in growth chambers under constant humidity (70%), with 16 h light and 8 h dark cycles at 20°C. Greenhouse-grown seedlings of maize inbred lines B73 and W22 were used for DNA and RNA extractions. Arabidopsis mutant line 570B09 was obtained from GABI-KAT (Rosso et al., 2003), and Col-0 and atatm-3 mutants (SALK_089805) were obtained from the Nottingham Arabidopsis Stock Centre. Double mutants were produced by crossing the atatm-3 mutants into the atnbs1-1 background and screening subsequent generations. Agrobacterium-mediated plant transformation and confocal analysis were performed as described previously (West et al., 2002; Foresti et al., 2006).

Nucleic acid purification, amplification and cloning

DNA procedures and bacterial manipulations were performed according to established protocols (Sambrook et al., 1989). RNA was isolated from above-ground tissues of flowering Arabidopsis using the SV total RNA isolation kit (Promega; http://www.promega.com/) according to the manufacturer’s instructions, and from maize tissues using Triazol reagent (Molecular Research Centre; http://www.mrcgene.com). AtNBS1 was cloned by RT-PCR using reverse transcriptase (Superscript II, Invitrogen; http://www.invitrogen.com/) for cDNA synthesis, followed by amplification using iPROOF DNA polymerase (Bio-Rad; http://www.bio-rad.com/). RACE-PCR was performed using the 3′/5′ RACE-PCR kit (Roche; http://www.roche-applied-sciene.com) according to the manufacturer’s instructions. PCR products were cloned using a TOPO-TA cloning kit and Escherichia coli TOP10 cells (Invitrogen), and plasmid DNA was prepared using spin columns (Qiagen; http://www.qiagen.com/) prior to fluorescent dye-terminator DNA sequencing (Perkin Elmer; http://www.perkinelmer.com). ZmNBS1 PCR amplification was performed using TaKaRa ExTaq HS (Takara Bio Inc.; http://www.clontech.com). The 35S-AtNBS1 3′ construct was amplified from atnbs1 genomic DNA and cloned into the binary vector pCB1300, and smrs-GFP (Davis and Vierstra, 1998) was inserted into a Csp45I site in exon 10 of AtNBS1 to create a translational fusion. To determine the subcellular localization of AtNBS1, the AtNBS1 cDNA was cloned into pCB1300 containing smrs-GFP under the control of the 35S promoter to create AtNBS1-smrs-GFP, and used for Agrobacterium-mediated plant transformation. Real-time PCR analyisis was performed on a iCycler thermocycler (Bio-Rad) using premix (Bio-Rad) and primers NBScDNA_3f (CTTCACTGATACCACCATCCGTTG), NBScDNA_3r (GTCCCAGCTATCGTATTCTTTGTATG) for the AtNBS1 3′ end, and qPCR_ACTf (CTCAGGTATCGCTGACCGTATGAG) and qPCR_ACTr (CTTGGAGATCCACATCTGCTGGAATG) for actin2 (At1g49240). AtRAD51 (At5G20850) was amplified using primers rad51RTf (GTTCTTGAGAAGTCTTCAAGAAGTTAG) and rad51rtr (GCTGAACCATCTACTTGCGCAACTAC) by real-time PCR. Complementation of the atnbs1-1 mutation was performed using a genomic clone of chromosome 3 region 574 667–579 579 amplified by PCR using iProof DNA polymerase (Bio-Rad) and primers nbs_gen_sacf (CATCATGAGCTCGCCTATATTTAGGTTGGCTCCATC) and nbs_gen_sacr (CATCATGAGCTCTTAACAGAGAAGTAATGGCCACATG). The PCR product was cloned into the SacI site of pCB1300 and used to transform atnbs1-1 plants according to standard protocols (Clough and Bent, 1998).

Overexpression of AtMRE11

Full-length AtMRE11 cDNA was cloned into pET30EkLIC vector (Novagen; http://www.merckbiosciences.co.uk), generating both N- and C-terminal 6× His-tagged proteins, and E. coli BL21(DE3)pLysS was transformed with the constructs. Expression was induced with 1 mm IPTG (Promega) for 3 h. Bacteria were recovered by centrifugation at 5000 g, resuspended in Bugbuster (Novagen), and lysed by freezing and thawing. Nucleic acids were removed by benzonase treatment (25 U ml−1, Novagen) at 37°C for 15 min, and the extract cleared by centrifugation at 25 000 g for 30 min at 4°C. Lysate containing His-tagged AtMRE11 or from untransformed BL21(DE3) cells (control) was applied to 100 μl nickel-coated paramagnetic beads (Promega) and washed five times with 1 ml RS buffer (50 mm Tris-Cl pH 7.5, 50 mm NaCl, 100 mm imidazole, 5% v/v glycerol, 0.1% v/v Triton X-100), then resuspended in 100 μl RS buffer and used immediately for in vitro interaction studies.

In vitro expression and interaction studies

AtNBS1 partial cDNA fragments were PCR-amplified using primers that incorporated a T7 RNA polymerase promoter. PCR products were transcribed and translated in vitro (TnT kit, Promega) in the presence of 0.4 μCi μl−1 [35S]methionine (>1000 Ci mmol−1, MP Biomedicals Europe; http://www.mpbio.com). Then, 10 μl TnT mix was incubated with 10 μl of AtMRE11-bound or control paramagnetic beads in a final volume of 100 μl RS buffer for 10 min. Beads were washed twice with 1 ml RS buffer, once with 1 ml RS buffer + 500 mm NaCl, and once again with 1 ml RS buffer prior to elution in 50 μl tricine gel loading buffer (Bio-Rad) at 90°C for 10 min. Samples were analyzed by tricine gel SDS–PAGE and autoradiography.

Yeast two-hybrid analysis

Two-hybrid analysis was performed as described previously (Fields and Song, 1989; Durfee et al., 1993). Full-length AtMRE11 and AtRAD50 cDNAs were cloned into the plasmid pAS1 to create GAL4 DNA binding domain fusion proteins. AtNBS1 cDNA encoding amino acid residues 217–542 was cloned into the plasmid pGAD-C1 (James et al., 1996) to create a GAL4 activation domain fusion. AtNBS1(217–444) was constructed by BglII digestion of pGAD-C1-NBS1(217-542) and re-ligation, resulting in a C-terminal deletion. AtNBS1(490-542) was constructed by NdeI digestion of pGAD-C1-NBS1(217-542) and re-ligation, resulting in an N-terminal deletion. Expression of AtNBS1(217-542) as a GAL4-DB fusion resulted in lacZ expression in yeast, and this auto-activation was avoided by further N-terminal deletion to produce AtNBS1(362-542). The full-length coding regions of the ZmNBS1, ZmMRE11A and ZmMRE11B genes cloned into pCR/GW/TOPO were subcloned into yeast two-hybrid vectors pDEST32 and pDEST22 using LR recombinase (Invitrogen) according to the manufacturer’s instructions. S. cerevisiae strain Y190 (MATa, ura3-52, his3-200, ade2-101, lys2-801, trp1-901, leu2-3, 112, gal4Δ, gal80Δ, cyhr2, LYS2::GAL1UAS-HIS3TATA-HIS3, URA3::GAL1UAS-GAL1TATA-lacZ) was used for interaction studies, and plasmids pSE1111 and pSE1112 expressing the SNF1-GAL4 DNA binding domain and the SNF4-GAL4 activation domain, respectively, were used as controls (Durfee et al., 1993). S. cerevisiae were transformed as described previously (Soni et al., 1993). Individual yeast colonies carrying both plasmids were tested for HIS3 reporter gene activation by growing liquid cultures to OD600 = 0.5 and spotting 5 μl of tenfold serial dilutions of each transformed line on synthetic medium lacking Leu, Trp and His supplemented with 25 mm 3-amino-1,2,4-triazole (Sigma-Aldrich; http://www.sigmaaldrich.com/). Liquid β-galactosidase assays were performed according to published protocols (http://www.clontech.com/), and activities were normalized to cell densities.

Mutagen sensitivity tests

For MMS sensitivity assays, 5-day-old Arabidopsis seedlings were grown on MS plates kept vertical in a growth chamber and then transferred to MS plates containing 0–0.1% MMS. Plants were returned to the growth chamber and incubated vertically for 7 days, transferred to fresh MS + MMS plates, and incubated vertically for an additional 7 days. Roots were then measured by removing the plants and extending the roots. For irradiation assays, seeds were imbibed at 4°C for 2 days in 0.1% agarose, and irradiated with 100 Gy X-rays at 0.8 Gy min−1. Seeds were plated on MS agar and grown under standard conditions for 3 weeks. Mitomycin C sensitivity assays used seeds imbibed at 4°C for 2 days, then transferred to liquid MS medium (0.8 ml) in a 24-well plate and supplemented with mitomycin C at 0–15 μg ml−1 as indicated.


Pollen mother cells were prepared as described previously (Armstrong et al., 2001), with the modification that cytohelicase was omitted from the digestion mix, and the digestion time was extended to 75 min.


We acknowledge the financial support of the UK Biotechnology and Biological Sciences Research Council through research grants to C.M.B. and the award of a David Phillips Research Fellowship to C.E.W. Funding from the Royal Society to C.E.W. and from the Bi-national Agriculture Research and Development Fund to C.F.W. is also gratefully acknowledged.