Deficiency of both classical and alternative end-joining pathways leads to a synergistic defect in double-strand break repair but not to an increase in homology-dependent gene targeting in Arabidopsis.

In eukaryotes, double-strand breaks (DSBs) are either repaired by homologous recombination (HR) or non-homologous end-joining (NHEJ). In somatic plant cells, HR is very inefficient. Therefore, the vast majority of DSBs are repaired by two different pathways of NHEJ. The classical (cNHEJ) pathway depends on the heterodimer KU70/KU80, while polymerase theta (POLQ) is central to the alternative (aNHEJ) pathway. Surprisingly, Arabidopsis plants are viable, even when both pathways are impaired. However, they exhibit severe growth retardation and reduced fertility. Analysis of mitotic anaphases indicates that the double mutant is characterized by a dramatic increase in chromosome fragmentation due to defective DSB repair. In contrast to the single mutants, the double mutant was found to be highly sensitive to the DSB-inducing genotoxin bleomycin. Thus, both pathways can complement for each other efficiently in DSB repair. We speculated that in the absence of both NHEJ pathways, HR might be enhanced. This would be especially attractive for gene targeting (GT) in which predefined changes are introduced using a homologous template. Unexpectedly, the polq single mutant as well as the double mutant showed significantly lower GT frequencies in comparison to wildtype plants. Accordingly, we were able to show that elimination of both NHEJ pathways does not pose an attractive approach for Agrobacterium-mediated GT. However, our results clearly indicate that a loss of cNHEJ leads to an increase in GT frequency, which is especially drastic and attractive for practical applications, in which the in planta GT strategy is used.

The DNA-dependent protease AtWSS1A suppresses persistent double strand break formation during replication.

The protease WSS1A is an important factor in the repair of DNA-protein crosslinks in plants. Here we show that the loss of WSS1A leads to a reduction of 45S rDNA repeats and chromosomal fragmentation in Arabidopsis. Moreover, in the absence of any factor of the RTR (RECQ4A/TOP3α/RMI1/2) complex, which is involved in the dissolution of DNA replication intermediates, WSS1A becomes essential for viability. If WSS1A loss is combined with loss of the classical (c) or alternative (a) nonhomologous end joining (NHEJ) pathways of double-strand break (DSB) repair, the resulting mutants show proliferation defects and enhanced chromosome fragmentation, which is especially aggravated in the absence of aNHEJ. This indicates that WSS1A is involved either in the suppression of DSB formation or in DSB repair itself. To test the latter we induced DSB by CRISPR/Cas9 at different loci in wild-type and mutant cells and analyzed their repair by deep sequencing. However, no change in the quality of the repair events and only a slight increase in their quantity was found. Thus, by removing complex DNA-protein structures, WSS1A seems to be required for the repair of replication intermediates which would otherwise be resolved into persistent DSB leading to genome instability.

An Arabidopsis FANCJ helicase homologue is required for DNA crosslink repair and rDNA repeat stability.

Proteins of the Fanconi Anemia (FA) complementation group are required for crosslink (CL) repair in humans and their loss leads to severe pathological phenotypes. Here we characterize a homolog of the Fe-S cluster helicase FANCJ in the model plant Arabidopsis, AtFANCJB, and show that it is involved in interstrand CL repair. It acts at a presumably early step in concert with the nuclease FAN1 but independently of the nuclease AtMUS81, and is epistatic to both error-prone and error-free post-replicative repair in Arabidopsis. The simultaneous knock out of FANCJB and the Fe-S cluster helicase RTEL1 leads to induced cell death in root meristems, indicating an important role of the enzymes in replicative DNA repair. Surprisingly, we found that AtFANCJB is involved in safeguarding rDNA stability in plants. In the absence of AtRTEL1 and AtFANCJB, we detected a synergetic reduction to about one third of the original number of 45S rDNA copies. It is tempting to speculate that the detected rDNA instability might be due to deficiencies in G-quadruplex structure resolution and might thus contribute to pathological phenotypes of certain human genetic diseases.