The long non-coding RNA LINDA restrains cellular collapse following DNA damage in Arabidopsis thaliana

The genomic integrity of every organism is endangered by various intrinsic and extrinsic stresses. To maintain the genomic integrity, a sophisticated DNA damage response (DDR) network is activated rapidly after DNA damage. Notably, the fundamental DDR mechanisms are conserved in eukaryotes. However, knowledge about many regulatory aspects of the plant DDR is still limited. Important, yet little understood, regulatory factors of the DDR are the long non-coding RNAs (lncRNAs). In humans, 13 lncRNAs functioning in DDR have been characterized to date, whereas no such lncRNAs have been characterized in plants yet. By meta-analysis, we identified the long intergenic non-coding RNA induced by DNA damage (LINDA) that responds strongly to various DNA double-strand break-inducing treatments, but not to replication stress induced by mitomycin C. After DNA damage, LINDA is rapidly induced in an ATM- and SOG1-dependent manner. Intriguingly, the transcriptional response of LINDA to DNA damage is similar to that of its flanking hypothetical protein-encoding gene. Phylogenetic analysis of putative Brassicales and Malvales LINDA homologs indicates that LINDA lncRNAs originate from duplication of a flanking small protein-encoding gene followed by pseudogenization. We demonstrate that LINDA is not only needed for the regulation of this flanking gene, but also for fine-tuning of the DDR after the occurrence of DNA double-strand breaks. Moreover, Δlinda mutant root stem cells are unable to recover from DNA damage, most likely due to hyper-induced cell death. We performed RNA-Seq experiments on wildtype (Col-0) and Δlinda mutants after induction of DNA damage by treatment of 14-day-old seedlings with either 80 µg/ml Zeocin or 2 kJ UV-C. Mock treated seedlings or seedlings grown under normal light conditions were used as controls for Zeocin or UV-C treatment, respectively. Samples were harvested 3 hours after induction of DNA damage. We harvested four samples per genotype per treatment.