The mitochondrial genome of soybean reveals complex genome structures and gene evolution at intercellular and phylogenetic levels

Abstract

Determining mitochondrial genomes is important for elucidating vital activities of seed plants. Mitochondrial genomes are specific to each plant species because of their variable size, complex structures and patterns of gene losses and gains during evolution. This complexity has made research on the soybean mitochondrial genome difficult compared with its nuclear and chloroplast genomes. The present study helps to solve a 30-year mystery regarding the most complex mitochondrial genome structure, showing that pairwise rearrangements among the many large repeats may produce an enriched molecular pool of 760 circles in seed plants. The soybean mitochondrial genome harbors 58 genes of known function in addition to 52 predicted open reading frames of unknown function. The genome contains sequences of multiple identifiable origins, including 6.8 kb and 7.1 kb DNA fragments that have been transferred from the nuclear and chloroplast genomes, respectively, and some horizontal DNA transfers. The soybean mitochondrial genome has lost 16 genes, including nine protein-coding genes and seven tRNA genes; however, it has acquired five chloroplast-derived genes during evolution. Four tRNA genes, common among the three genomes, are derived from the chloroplast. Sizeable DNA transfers to the nucleus, with pericentromeric regions as hotspots, are observed, including DNA transfers of 125.0 kb and 151.6 kb identified unambiguously from the soybean mitochondrial and chloroplast genomes, respectively. The soybean nuclear genome has acquired five genes from its mitochondrial genome. These results provide biological insights into the mitochondrial genome of seed plants, and are especially helpful for deciphering vital activities in soybean.

Figure 1. The circular map of the mitochondrial genome of G. max.

Features on the clockwise- and counter-clockwise-transcribed strands are drawn on the inside and outside of the circle, respectively. The figure was drawn using OGDraw v1.2 .

Figure 2. Relationships revealed by multiple alignments among five R1 large repeats (marked as R1a–e).

R1b, R1d, R1e are in the some orientation; R1a and R1c are in the reverse orientation. The red segments S6* in R1b and R1c are homologous, but different from S6 in the other three R1 large repeats.

Figure 3. Reversible reorganization of the soybean mtDNA may produce multiple subgenomic circles mediated by large repeats.

(A) Arrows of the same color denote homologous large repeats and their sequence orientation. (B) shows subgenomic circles of various sizes produced by rearrangements of the five pairs of large repeats. (C) The three small circles may be produced by three pairs of large repeats (R1b-R1d, R1d-R1e and R1b-R1e).

Figure 4. Isometric genome structures formed by rearrangements within eight pairs of inverted large repeats.

(A) is structurally the same as Figure 3A. (B–I) shows the eight isometric master genome structures of (A). The red arrows in the circles denote the inverted regions mediated by the repeats. Combinations of inverted repeats for each isometric master circle are marked in the circles.

Figure 5. Chromosome coverage of hits obtained by searching the mitochondrial genome against the nuclear assembly.

Rectangles show the lengths of matches covering the chromosomes. The lines show percent coverage by the matches on the soybean chromosomes.

Figure 6. Characteristics of nuclear-mitochondrial sequences in soybean.

Results are based on a BLAST e-value cutoff of 1e^–12. (A) Distributions of percent identities between shared nuclear-mitochondrial matches. The number of matches is shown by brown boxes and is plotted on the left ordinate. The red and blue square lines show the coverage of matches on nuclear and mitochondrial genomes, respectively, and are plotted on the right ordinate. (B) Distributions of lengths between shared nuclear-mitochondrial matches; the notation method is the same as for (A).

Figure 7. Phylogeny of four Faboideae mitochondrial genomes.

Numbers above each node represent bootstrap values from 1000 replicates. Branch lengths are in units of synonymous substitutions per synonymous site.

Figure 8. DNA transfers among the nuclear, chloroplast and mitochondrial genomes in soybean.

Figure 9. Phylogenetic tree of representative mitochondrial genomes in higher plants.

Phylogenetic trees were constructed with 28 representative plant mitochondrial genomes using the coding sequences of 22 genes under the GTR+G+I model . Mitochondrial-like tRNA genes and protein-coding genes eliminated during evolution are shown in (A) by arrowed black boxes in the evolutionary clades. Genes exceptionally maintained in the branches are shown by arrowed red boxes. (B) shows the chloroplast-derived tRNA genes transferred to the mitochondrial genome as boxes arrowed to a clade, and the genes lost in a branch are shown with an arrow to a box. trnC (GCA) represent tRNA genes of unknown origin, here attributed to chloroplast genes.