The H3K4me1 histone mark recruits DNA repair to functionally constrained genomic regions in plants

Abstract

Mutation is the ultimate source of genetic variation in natural populations and crops. To study mechanisms determining mutation rate variation within plant genomes, we analyzed 43,483 de novo germline single base substitutions in 1,504 fast neutron mutation lines of the model rice cultivar Kitaake (Oryza sativa ssp japonica) (from Li et al. 2017). We find that, like previously observed for de novo germline mutations in Arabidopsis. thaliana, mutation rates are significantly lower in genomic regions marked by H3K4me1, a histone modification found in the gene bodies of actively expressed genes in plants. We also observed conservation in rice for PDS5C, a cohesion cofactor involved in the homology-directed repair pathway that in A. thaliana binds to H3K4me1 via its Tudor domain. By examining existing ChIP-seq data for PDS5C in A. thaliana, we find that it localizes to genome regions marked by H3K4me1: regions of low mutation rates, coding regions, essential genes, constitutively expressed genes, and genes under stronger purifying selection, mirroring mutation biases observed in rice as well. We searched the A. thaliana proteome for genes containing similar Tudor domains and found that they are significantly enriched for DNA repair functions (p<1×10-11), including the mismatch repair MSH6 gene (in both rice and A. thaliana ), suggesting the potential for multiple DNA repair pathways to specifically target gene bodies and essential genes through H3K4me1 reading. These findings inspire further research to characterize mechanisms that localize DNA repair via histone interactions, leading to hypomutation in functionally constrained regions and potentially tuning the evolutionary trajectories of plant genomes.

Report of mutation biases mirroring selection in Arabidopsis thaliana unlikely to be entirely due to variant calling errors

“…Distribution of de novo mutations in 1,504 fast-neutron-treated rice lines, with >99% validation rate of SNVs [43]; for further details see [44]. …”

“…2b). MSH2 functions as a dimer with MSH6, which is likely targeted to H3K4me1 via its Tudor domain [10,44]. Additional recent studies that support the core findings of [1] include reports of (i) reduced germline and somatic mutations rates in transcribed genes, in genome regions marked by epigenomic features characteristic for gene activation, and in gene bodies, observed in several algae, plants as well as humans and nematodes [17,21,36,42,[47][48][49][50][51][52]; (ii) relationships between histone modifications -especially those linked to gene activation -, mutation rates and DNA repair across diverse species [51,[53][54][55][56][57][58][59][60]; and 8 (iii) H3K4me1 affecting rates of CRISPR-mediated mutagenesis in plants [61].…”

“…Proteins with Tudor domains, as they are found in some plant DNA repair proteins, specifically bind to regions with lower mutation rates (Fig. 4d) [40,44]. A potential mechanistic basis (Fig.…”

“…2b). MSH2 functions as a dimer with MSH6, which is likely targeted to H3K4me1 via its Tudor domain [10,44].…”

“…As in A. thaliana, removal of C:G>T:A single-nucleotide mutations does not greatly change the observed patterns. More details can be found in [44]. [43]; for further details see [44].…”