Long-read sequencing characterizes mitochondrial and plastid genome variants in Arabidopsismsh1mutants

Abstract

The abundant repeats in plant mitochondrial genomes can cause rapid genome rearrangements and are also a major obstacle in short-read sequencing studies. Nuclear-encoded proteins such as MSH1 are known to suppress the generation of repeat-associated mitochondrial genome variants, but our understanding of these mechanisms has been constrained by the limitations of short-read technologies. Here, we used highly accurate long-read sequencing (PacBio HiFi) to characterize mitochondrial and plastid genome variants in Arabidopsis thaliana msh1 mutant individuals. The HiFi reads successfully quantified parental and crossover recombination products of both large and small repeats, providing a global view of recombination dynamics. We found that recombination breakpoints were distributed relatively evenly across the length of repeated sequences and observed extensive non-crossover recombination (gene conversion) between pairs of imperfect repeats in the mitochondrial genome of msh1 mutants. Comparison of long-read assemblies of seven other Arabidopsis thaliana accessions showed that mitochondrial genome structural divergence between accessions paralleled the repeat-mediated dynamics observed in msh1 mutants. The Arabidopsis plastid genome generally lacks small repeats and exhibited a very different pattern of variant accumulation in msh1 mutants compared with the mitochondrial genome. Our data illustrate the power of HiFi technology in studying repeat-mediated recombination in plant mitochondrial genomes and improved the sequence resolution for recombinational processes suppressed by MSH1.