The Gransden wild-type strain of P.patens (13) was used in this study. Protonemal tissue was propagated on PpNO₃ medium (14), supplemented with 2.7 mM NH₄-tartrate (standard medium). Cultures were grown in 9 cm Petri dishes on medium solidified with 0.7% Agar (Biomar) and overlaid with a cellophane disk (Cannings, Bristol). Cultures were illuminated with a light regime of 16 h light/8 h darkness and a quantum irradiance of 80 µE m⁻² s⁻¹ (standard conditions). Sporogenesis was performed in Magenta box in which the tested strains were grown aside the self sterile but cross fertile strain nicB5ylo6 (13) on minimal PpNO₃ medium. Crosses were grown in standard conditions for 6–8 weeks (i.e. until about 50 gametophores reached full development), then irrigated with sterile water and transferred for 2 weeks in growth chambers set at 15°C with 10 h of light per day and a quantum irradiance of 15 µE m⁻² s⁻¹. The development of archegonia, antheridia and spore capsules was followed visually during the next month. Antheridia and archegonia were manually dissected for further microscopic observation.

We used standard methods for all molecular cloning (15). Using genomic P.patens DNA as starting template, we amplified a 2340 bp PCR fragment covering an internal fragment of the MSH2 genomic region (Gene Bank: DQ117988) and inserted it into the TA-cloning vector pCR^®II (Invitrogen, Groningen, The Netherlands) to produce PpMSH2 (oligonucleotide sequences can be provided on request). To obtain PpMSH2-KO, a 465 bp 5′ targeting fragment (coordinate 2557–3022 in DQ117988) and a 598 bp 3′ targeting fragment (coordinate 4614–5212 in DQ117988) were recovered by restriction from PpMSH2 and inserted upstream (5′) and downstream (3′) of the loxP sites flanking the resistance cassette of pBSK35S-NPTII-lox (supplied by Dr Laloue, INRA Versailles). Vector pBSK35S-NPTII-lox carries a neomycin phosphotransferase resistance gene driven by the 35S CaMV promoter and flanked by two LoxP sites in direct orientation. The antibiotic resistance cassette is in the same orientation with regard to the transcription of PpMSH2.

To obtain PpAPT-KO, a 3995 bp Xba1 and EcoRI fragment covering the APT genomic sequence (Gene Bank: DQ117987), was inserted in pBlueScript KS (Stratagene), forming PpAPT1. The ATG codon is 1217 bp 3′ of the XbaI site and the stop codon is 881 bp 5′ of the EcoRI site in this PpAPT fragment. An internal 1102 bp BamHI fragment was deleted by digestion of the PpAPT1 vector, to produce PpAPTΔ. We generated homeologous variants of this targeting cassette by mutating the PpAPTΔ sequence by mutagenic PCR using the Cadwell and Joyce (16) protocol. Successive APT variants were generated until we reached 3% divergence within the PpAPT original fragment. A BamHI fragment containing the hygromycin phosphotransferase gene (HptII) driven by the 35S CaMV promoter was recovered from plasmid pGL2 (17) and ligated into the BamHI site of PpAPTΔ and of the variants to produce the homologous and homeologous PpAPT-KO vectors (Figure 4).

Protoplasts were isolated from 6-day-old protonema by incubation for 40 min in 1% Driselase (Fluka 44585) dissolved in 0.47 M mannitol. The suspension was filtered successively through 80 and 40 µm stainless steel sieves. Protoplasts were sedimented by low-speed centrifugation (600 g for 5 min at 20°C) and washed twice with 0.47 M mannitol. Protoplasts were then resuspended at 1.2 × 10⁶ protoplasts/ml in MMM solution (0.47 M mannitol, 15 mM MgCl₂ and 0.1% MES, pH 5.6) for transformation. For growth assay and phenotypic analyses, protoplasts were resuspended at a density of ca. 1–3 × 10⁴ protoplasts/ml in mannitol 0.47 M containing 0.5% agar. Two ml of suspension were spread on cellophane disks covering standard medium supplemented with 0.5% glucose and 0.36 M mannitol (PPM). Regeneration was conducted in standard growth conditions. After 1 week of regeneration, protoplasts were transferred to standard medium without mannitol.

Moss protoplasts (1.2 × 10⁶) were transformed as described previously (18) with PpMSH2-KO digested with AgeI and AatII, which generate DSBs within the genomic MSH2 sequence. Primary transformants were selected in the presence of 50 mg/l paramomycin (Duchefa P0141, Haarlem, The Netherlands). Small pieces of the protonema tissue (∼2 mm in diameter) of primary P.patens transformants were then transferred onto standard medium without selection and cultured for 15 days. A second round of selection allowed the isolation of stable transformants, by transferring small pieces of protonema tissue from the previous colonies onto standard medium containing 50 mg/l G418 (Duchefa G0175, Haarlem, The Netherlands) (paramomycin and G418 belong to the kanamycin family of antibiotics).

Moss DNA was extracted from 200 mg of fresh protonema tissue ground in 220 µl of extraction buffer (220 mM Tris–HCl, pH 7.5, 250 mM NaCl, 25 mM EDTA and 0.5% SDS) with an electric pestle in the presence of sand. After centrifugation, DNA was precipitated with isopropanol at room temperature. The pellet was dried and resuspended with 100 µl sterile water. Stable disruptants of the PpMSH2 gene were screened by PCR for the incorporation of the resistance cassette into the MSH2 locus. We used primers specific for the resistance cassette (Uc: TATTTTTGGAGTAGACAAGCGTGTCGT and Lc: CCGCTTCCTCGTGCTTTACGGTAT) and primers specific for the PpMSH2 sequence outside the transformation construct (PpMSH2#4: CAATGTCACTGAGAGCAAAACG and PpMSH2#5: CACACTGATCTCTGCTCGGTCG).

Moss RNA was isolated using the RNAeasy kit (Qiagen) starting with 100 mg of fresh protonema. We obtained doubled-stranded cDNAs corresponding to mRNAs expressed in 6-day-old protonema cultures using the SMART-PCR cDNA Synthesis Kit (Clontech). The primers PpMSH2#4 and PpMSH2#2 (CACCAACACGCGCAAAGATG) were used for PCR on these cDNAs.

Direct repeats of LoxP recombination sites from the Cre/lox system (19) were located on either side of the 35S-nptII resistance cassette in PpMSH2-KO. Transient expression of a constitutive Cre expression cassette (20) in protoplasts allowed excision of the resistance gene from the PpMSH2 recombinant locus (D.G. Schaefer, A. Finka and J.P. Zrÿd, manuscript in preparation). Protoplasts were regenerated at a low density on non-selective medium for 2 weeks and fragments of single protoplast-derived colonies were then replicated on selective (i.e. G418 50 mg/l) and non-selective medium. Loss of antibiotic resistance was determined after an additional week of growth. Replicates growing on non-selective medium were used for subsequent amplification and molecular analyses. Clone Ppmsh2Δ was selected this way and used for further analyses. We analyzed by PCR the recombinant locus with the primers PpMSH2#4 and PpMSH2#2.

Binocular observations were made with a Nikon SMZ1000, and microscopic observations with a Leitz (type 090-122.012) inverted microscope.

Mutations in the APT gene confer resistance to 2-FA (2-Fluoroadenine, from Dr Laloue, INRA Versailles, France), a toxic compound for cells. The number of 2-FA resistant colonies reflects the frequency of spontaneous mutations. Protoplasts of wild-type and Ppmsh2 mutants were regenerated for 6 days regeneration on PPM (∼10⁵/Petri dish) and then transferred on standard solid medium supplemented with 10 µM 2-FA. After 2 weeks, the number of resistant colonies was counted. We repeated this experiment three times. The APT genomic sequence from 2-FA resistant clones was subsequently PCR amplified and sequenced to identify the mutation responsible for the resistance (oligonucleotide sequences can be provided on request).

Cisplatin (cis-diamminedichloroplatinum(II)) (Calbiochem, Darnstadt, Germany), MNU (N-methyl-N-nitrosourea) and O⁶-bzGua (O⁶-benzylguanine) (Sigma-Aldrich, Steinheim, Germany) were dissolved in dimethyl sulfoxide at 100 mM each for storage at −20°C and at 25 mM for O⁶-bzGua immediately before use.

Wild-type and Ppmsh2 mutant protoplasts in liquid PPM were treated for 20 h with 0–5 mM cisplatin or MNU (2 × 10⁵ protoplasts at each concentration). From 1 h before exposure to MNU and throughout the whole procedure, 25 µM O⁶-bzGua was present in the medium to inhibit the removal of methyl groups from the O⁶ position of guanine by endogenous methyl-transferase activity. After 20 h in the dark, protoplasts were washed and spread on solid medium at low concentrations.

Protoplasts in liquid mannitol medium were exposed to ionizing radiation (1 Gy/s) using a ¹³⁷Cs irradiator (IBL-637 CIS-Bio-International, Institut Curie, Orsay). After 20 h in the dark, the protoplasts were spread on solid medium at low concentrations.

Protoplasts spread on solid medium (5 × 10⁴ protoplasts at each concentration in four Petri dishes) were exposed to UV-B light (60 J/m²/s) from a 312 nm TFX lamp. We calculated the flux with a UV-Elektronik GmbH dosimeter. The protoplasts were then left in the dark for 20 h.

After 6 days regeneration, protoplasts were transferred onto standard solid medium. After 1 week, the number of survivors was counted. We repeated these experiments two to three times. Statistical analyses have been made with the χ² test.

Targeting APT fragments were produced by digestion of PpAPT-KO vector with NotI and EcoRI enzymes. We selected stable transformants after transformation of wild-type and Ppmsh2 mutant protoplasts with homologous and homeologous APT disruption fragments. Small pieces of protonema tissue (∼2 mm in diameter) from stable transformants were then transferred onto standard medium containing 10 µM of 2-Fluoroadenine to detect APT gene targeting events. Experiments were repeated three times.

Moss protoplasts were transformed with vector PpMSH2-KO and four independent clones were selected for further investigation (Ppmsh2#1 to Ppmsh2#4). In each clone, disruption of the chromosomal PpMSH2 locus was controlled by PCR and absence of the PpMSH2 transcript was established by RT–PCR (see Ppmsh2#1 for example Figure 1B–D). In order to obtain a simple deletion mutant, the resistance cassette inserted at the PpMSH2 locus of the Ppmsh2#1 clone was eliminated by the CRE recombinase to produce clone Ppmsh2Δ (see Materials and Methods and Figure 1A). In this clone, we confirmed deletion of exon 7 of the PpMSH2 gene by PCR and checked modification of the transcript by RT–PCR (Figure 1C and D). The four independent PpMSH2 knockouts (Ppmsh2#1 to Ppmsh2#4) and the deletion mutant (Ppmsh2Δ) behaved identically under all circumstances tested except in aged cultures. Therefore, Ppmsh2Δ mutant was used for further characterization except for late developmental phenotypes (see below). To avoid the genetic drift resulting to the expected mutator phenotype associated with msh2 mutations, all the experiments were performed using the primary stock of protonema suspension from Ppmsh2Δ stored at 4°C as starting material.

The haploid gametophytic development of P.patens from regenerating protoplasts starts with a filamentous branched structure called the protonema that grows by apical cell divisions and lateral division of subapical cells. There are two cell types in protonema filaments: chloronema and caulonema. Regenerating protoplasts first develop into photosynthetically active chloronemal filaments, formed by cells filled with chloroplasts and separated by walls perpendicular to the axis of the filaments. After ∼5 days, chloronemal apical cells occasionally differentiate into caulonemal cells characterized by oblique cell walls and smaller and fewer chloroplasts. After 8–10 days, primitive meristems called buds differentiate from caulonemal side branch initials and further differentiate into leafy shoots, also known as gametophores. Each leafy shoot resembles a small plant with an unbranched stem, filamentous basal and aerial rhizoids and leaves displaying clear phyllotaxis along the stem axis (21). We found that the msh2 mutants had several pronounced phenotypes affecting the development of both protonema and leafy shoots.

The strong developmental defects observed in Ppmsh2 mutants could be related to accumulation of spontaneous mutations. This phenotype has already been observed in msh2 mutants of other organisms (2). To further quantify this mutator phenotype, we investigated the effect of the loss of the PpMSH2 gene on mutation frequency of the adenine phosphoribosyl transferase (PpAPT) reporter gene. Spontaneous mutations leading to loss of APT activity confers resistance to adenine analogues, such as 2-fluoroadenine (2-FA) and can be selected on standard medium supplemented with 10 µM 2-FA. We tested a total of 10⁷ Ppmsh2Δ protoplast-derived colonies and 3 × 10⁶ wild-type colonies for resistance to 2-FA.

We observed no resistant colonies in the wild type and a total of 398 2-FA resistant colonies in the Ppmsh2Δ mutant. APT mutation frequencies were thus lower than 3 × 10⁻⁷ in wild type and 4 × 10⁻⁵ ± 1.2 × 10⁻⁵ in Ppmsh2Δ. To confirm that 2-FA was effectively resulting from mutation in the APT gene, we partially sequenced the APT gene of randomly chosen 2-FA resistant clones. Point mutations in the APT sequence were effectively detected in more than 75% of the clones tested (Table 1). Mutations resulting from transitions and transversions were recovered, but we did not identify mutations resulting from insertion or deletion. These modifications were matching with highly conserved domain or amino acid of the APT protein and some of them were identified several times in independent 2-FA resistant clones. These results demonstrate that loss of MSH2 gene leads to the accumulation of point mutations in the moss genome and that the spontaneous mutation rate is increased at least by two orders of magnitude (130 times) in the Ppmsh2Δ mutant compared with the wild type. This last value is probably underestimated since the screen used here has selected only mutations that confer 2-FA resistance. Additionally, the type of mutations identified suggests that they result from defaults of post DNA replication proofreading processes.

MMR is known to recognize different types of DNA damages depending on the lesion and on the organism. MMR is responsible for the hypersensitivity to agents, such as radiation or some genotoxic chemicals that create DNA lesions (22). Loss of MMR may prevent cell death normally resulting from exposure to such genotoxic agents, thus conferring tolerance to these stresses. Therefore, we treated Ppmsh2Δ mutant protoplasts with UV-B and γ-radiation and chemicals such as cisplatin and MNU to determine whether P.patens cells with defective MMR would be affected or not. Survival was calculated as the ratio of protoplasts surviving after 15 days regeneration following treatment to the number of protoplasts undergoing normal regeneration without treatment.

In E.coli (29), Saccharomyces cerevisiae (5), Trypanosoma brucei (30) and mouse embryogenic stem cells (7), depletion of MutS or MSH2 leads to an increased frequency of recombination between homeologous DNA sequences. We examined the involvement of PpMSH2 in HR by measuring GT efficiency with homologous or homeologous DNA sequences at the APT locus. Wild-type and Ppmsh2Δ cells were transformed with APT constructs presenting 0, 1, 2 and 3% divergence to the PpAPT genomic sequence, respectively (Figure 4). Hygromycin resistant transgenic clones were initially selected and tested for 2-FA resistance. GT frequency was calculated as the ratio of 2-FA resistant clones to the total number of Hygromycin resistant transformants.

We observed a significant decrease (1.8-fold) in GT frequency for mutant cells compared to wild type (P = 0.001) when the recombination substrate was fully homologous to its genomic target (Table 2). In wild type, GT frequency decreased as the percentage of divergence increased, whereas GT frequency remained stable for the Ppmsh2Δ mutant whatever the divergence level (Table 2). In wild type GT frequency significantly decreased 22-fold with 3% mismatches compared to 0% mismatches (χ² test, P < 0.00001), whereas in Ppmsh2Δ mutant, the GT was as efficient with homeologous fragments as with homologous fragment (P = 0.40). Thus, GT efficiency with a homeologous fragments containing 3% divergence was increased 15-fold in Ppmsh2Δ mutant compared to the wild type. These data demonstrate that loss of function of MSH2 allows GT in P.patens with up to 3% divergent homeologous sequences, hence that the moss MSH2 protein has an antirecombination activity that prevents recombination between slightly divergent sequences.

Ppmsh2 mutant colonies develop strong developmental defects, including sterility, after a few weeks. This sterility may be due to the msh2 associated mutator phenotype increasing the rate of mutations that cause developmental problems in complex structures such as leafy shoots, antheridia and archegonia. Similar phenotypes have been observed in Arabidopsis thaliana msh2 mutant lines (10). After five generations, plants showed abnormalities in morphology and development, fertility, germination efficiency, seed/silique development and seed set. In human and mice, mutations in MSH2 result in a greatly increased likelihood of developing certain types of tumors (31). As in other eukaryotes, during growth and development, MSH2 in moss seems to be essential for the maintenance of genome integrity. Accordingly, we scored a 130-fold increase in the spontaneous mutation rate of the APT gene in Ppmsh2 mutants compared with wild type. This increased mutation rate is comparable with those observed in E.coli mutS mutant (20-fold increase) (32), or S.pombe (15-fold increase) (33) and S.cerevisiae (200-fold increase) (8) and other higher eukaryotes msh2 mutants (from 30- to several hundred-fold increase (7,10,34,35). This mutation rate can explain the phenotypes observed in the Ppmsh2 mutants. Assuming that the amount of mutations increases with the number of rounds of DNA replication, abnormalities should increase with the number of generations of vegetative propagation and this is what we observed after multiple successive subcultures of the colonies. Nevertheless, the phenotypes we observed in the mutant background are particularly strong, and two explanations can be proposed to account for this observation. The first one is linked to the fact that the major part of the moss life cycle is haploid. Thus, the deleterious effects of mutations in development are greater in moss than for diploid mice or A.thaliana, where the allelic locus would mask the phenotype resulting from randomly occurring mutations, at least in the first generation. Moreover, we cannot discard the hypothesis that, as in budding yeast, efficient GT in P.patens reflects the predominance of the HR pathway in the repair of DNA DSB. In this case, alteration of the MMR system, which participates in this DNA repair pathway, would lead to more drastic effects than in other higher eucaryotes, where HR is not the major pathway for DNA DSB repair. This is reminiscent of the strong reduction in spore viability observed in the msh2 yeast mutant (36) an organism where HR is also predominant.

The sensitivity of cells to agents that create lesions on DNA has been attributed to attempts of MMR to correct the resulting damage (22). Depending on the type of DNA damage, loss of MMR may result in increased mutagenesis, loss of cell-cycle control and resistance to apoptosis (37). The resistance of msh2-deficient cells to DNA damage is not so clear because it depends on the organism and on the specificity of the lesions. In C.elegans, MSH2 confers sensitivity to ionizing radiation (9). In mouse cells, conflicting results have been reported, although the effects are generally minor. Msh2-nullizygous cells were found to exhibit slight higher levels of survival after exposure to ionizing radiation (38,39). However, it has recently been shown that both mouse embryogenic fibroblasts (MEFs) and mouse colorectal carcinoma cells defective in MSH2 were slightly more sensitive (1.5-fold) to X-ray irradiation than their MSH2-non-defective counterparts (40). In contrast, in moss, as in drosophila and human cells, the loss of MSH2 does not affect the sensitivity to γ-radiation, suggesting that MSH2 is not involved in the detection of lesions due to γ-radiation (41,42).

In E.coli, T.brucei, mouse and human cells, MSH2 confers sensitivity to MNU (27,38,43,44). In contrast, in the yeast S.cerevisiae, the MMR system did not affect sensitivity to methylating agents (45). In our experiments, Ppmsh2 mutants showed only a weak resistance to MNU at high concentration. This suggests that only a fraction of the MNU cytotoxicity in moss cells may be dependent on MSH2 function. Other mechanisms must be involved to repair DNA lesion in conjunction or not with MMR. HR can repair DNA lesions, like it has been shown in yeast for O⁶-meGua damages (45). Absence of MMR in S.cerevisiae did not affect sensitivity to methylating agents, while HR inactivation sensitized cells to this agent. In HR-deficient cells, defects in the MSH2 gene rescue the original sensitivity (45). These authors conclude that methylation damage processing and repair by HR masks the sensitivity of MMR-proficient S.cerevisiae cells toward methylating agents (45). In our experiments, Ppmsh2 mutants showed a weak resistance to MNU only at high concentration, suggesting that only a fraction of the MNU cytotoxicity in moss cells may be dependent on MSH2 function. Like in yeast, moss MMR could be overwhelmed by other mechanisms such as HR, thus preventing the cytotoxicity due to methylating agents to be detectable in wild type.

Peters et al. (37) have shown that in MEF (mouse embryogenic fibroblasts), MSH2 may sense damage due to photoproducts or may contribute to the processing of UV-B-induced DNA damage, thereby initiating apoptosis and influencing the activation of p53 (a tumor suppressor which functions as a transcription factor involved in cell-cycle checkpoints and apoptosis). They showed that Msh2-null MEFs were partially resistant to the cytotoxic effects of UV-B, and that this resistance led to a reduction in apoptosis. In contrast, we have shown that the Ppmsh2 mutant is more sensitive to UV-B radiation than the wild type. Therefore, MSH2 in moss may play a more important role in the recognition of UV-B lesions than that reported in MEF.

Experiments in E.coli have shown that loss of MMR activity, due to mutations in mutS, was correlated with cisplatin resistance (46,47). There are conflicting results in the response of cisplatin treatment in higher eukaryotes. It was thought that mouse MSH2 conferred sensitivity to cisplatin (48). However, Claij and te Riele (49) have recently shown that msh2-deficient mouse cells do not respond specifically to cisplatin lesions compared with wild type (49). In contrast, data from Trypanosoma cruzei have demonstrated a role of MSH2 in sensitivity to cisplatin (50), like in prokaryotes and in yeast (46,51). Our results in moss show that Ppmsh2 mutants are tolerant to cisplatin.

The fact that Ppmsh2 mutants have opposite responses to lesions due to UV-B (sensitivity) and cisplatin (resistance) suggests that moss may have different pathways for managing lesions involving PpMSH2 proteins. PpMSH2 may signal DNA lesions and induce cell-cycle arrest followed by either DNA repair or apoptosis to prevent accumulation of a large number of mutations. In the absence of PpMSH2, we suggest that, depending on the type of DNA lesions, different responses may be engaged. When the lesion is too severe (exposure to UV-B), the cells are unable to repair them and die. For other lesions, they can be bypassed by a translesion synthesis polymerase (52,53) or be repaired by other mechanisms such as base excision repair, nucleotide excision repair or by HR (45,54,55). In these cases, the cells can survive almost normally, as observed with moss msh2 mutant cells exposed to cisplatin. We suggest that in wild-type cells, the DNA lesions caused by cisplatin trigger cell death and/or prolonged cell-cycle arrest through PpMSH2. We have evidence that the cell cycle is perturbed in the absence of PpMSH2: during the regeneration of protoplasts, Ppmsh2 mutants showed a slightly higher rate of cell division compared with the wild type. This suggests that Ppmsh2 mutants may have an accelerated cell cycle, which may be explained if there was shorter arrest at cell-cycle checkpoints. Indeed, in mice it has been shown that after DNA damage, the absence of MSH2 results in premature release from the early G2/M arrest and an inability to activate properly the checkpoint kinase proteins (40). It was suggested that an active MSH2 is required for a correct response to DNA damage in the G2 phase of the cell cycle, possibly connecting DSB repair to checkpoint signaling. Moreover, in human, MMR is required for S-phase and G2-M checkpoint regulation (56,57). Consistent with these findings, we suggest that PpMSH2 is required for efficient cell-cycle arrest induced by DNA damage and may be a key component acting on DNA damage-activated checkpoints.

We examined the influence of MMR on HR in P.patens by assaying the frequency of recombination between a foreign DNA fragment released in vivo and a genomic target sequence. We found that the frequency of GT via HR between identical DNA was decreased 1.8-fold in the Ppmsh2-defective cells compared with the wild type. In mice, msh2 mutants show a slight decrease of GT efficiency with homologous fragment compared with wild-type mice, and this decrease is comparable with our results (1.5-fold) (7). In yeast, a recent study has shown that the frequency of GT was reduced 3- to 4-fold in the msh2 strain (58). Together with our results, these suggest that MSH2 in P.patens is probably involved in mechanisms that regulate GT of perfectly matching sequences in the genome. Further analyses are needed to determine the mechanisms involved.

We tested the influence of MMR on homeologous recombination in P.patens by assaying the frequency of recombination between a mutated DNA fragment and a genomic target sequence. Our results show that DNA sequence identity is important to determine the efficiency of gene replacement in the wild type. We have also shown that lack of PpMSH2 allows homeologous sequences with up to 3% divergence to be as efficiently targeted as perfectly identical sequences to a genomic locus. This shows that PpMSH2 plays a major role to prevent recombination between homeologous sequences. This is consistent with the similar role described for MMR in E.coli (29), in Trypanosoma (30), in yeast (5) and in mouse cells (7,59,60), suggesting that this role of MSH2 is strongly preserved during evolution. However, this is the first time that MSH2 involvement in HR and homeologous recombination is shown in plants.

In conclusion, we have found that in moss, MSH2 is involved in mismatch recognition and repair that most likely results from DNA replication errors, and in the response to certain genotoxic agents. Moreover, MSH2 plays a role in HR and in the recognition of mismatches that occur during heteroduplex formation between two not perfectly complementary DNA strands. These mechanisms are essential for maintaining genome integrity in all organisms and to establish interspecific barrier. This is particularly true for plants as they are unable to move and must cope with exposure to environmental mutagens. Thus, an efficient mechanism for genomic stability is needed in plants, and even more in mosses, where the major developmental stage of the life cycle is the haploid gametophyte. This is emphasized by the dramatic deleterious effect on development observed in an msh2 P.patens mutant compared with its A.thaliana counterpart. Our results show for the first time, the crucial role of MMR in a pluricellular organism where HR seems to be one of the major pathways to repair DNA lesions.