Legume genome evolution viewed through the Medicago truncatula and Lotus japonicus genomes

Abstract

Genome sequencing of the model legumes, Medicago truncatula and Lotus japonicus, provides an opportunity for large-scale sequence-based comparison of two genomes in the same plant family. Here we report synteny comparisons between these species, including details about chromosome relationships, large-scale synteny blocks, microsynteny within blocks, and genome regions lacking clear correspondence. The Lotus and Medicago genomes share a minimum of 10 large-scale synteny blocks, each with substantial collinearity and frequently extending the length of whole chromosome arms. The proportion of genes syntenic and collinear within each synteny block is relatively homogeneous. Medicago-Lotus comparisons also indicate similar and largely homogeneous gene densities, although gene-containing regions in Mt occupy 20-30% more space than Lj counterparts, primarily because of larger numbers of Mt retrotransposons. Because the interpretation of genome comparisons is complicated by large-scale genome duplications, we describe synteny, synonymous substitutions and phylogenetic analyses to identify and date a probable whole-genome duplication event. There is no direct evidence for any recent large-scale genome duplication in either Medicago or Lotus but instead a duplication predating speciation. Phylogenetic comparisons place this duplication within the Rosid I clade, clearly after the split between legumes and Salicaceae (poplar).

Fig. 1.

Dot plots, synteny closeup, and chromosome correspondences. (A) In the dot plot, each dot represents the reciprocal best BLASTP match between gene pairs. Red dots show regions of synteny as identified by DiagHunter. Some Lj chromosome orientations have been flipped (“fLj5, fLj6, fLj2”) to visually correspond to Mt orientations. Both Mt and Lj have been scaled to occupy the same lengths. See supporting information for all dot plots and related results. (B) Closeup views of synteny. Lines in shades of blue or yellow indicate BLASTP E-values, with strongest correspondence blue (0.0) to weakest yellow (0.01). Only single strongest reciprocal hits are shown. Comparisons a and b (upper callout) show synteny in Lj2 × Mt5; comparisons c and d (lower callout) show synteny in Mt5 × Mt8 (see text). (C) Graphs showing percentages of individual pseudochromosomes with synteny between Mt and Lj. Coverage is calculated as the sum of block sizes (a block has dimensions of the end gene minus the start gene). Multiple synteny blocks over the same region are counted, so altogether coverage can exceed 100%.

Fig. 2.

Ks dating of duplication blocks and Mt/Lj synteny blocks. Age distributions of duplicated (Mt, Lj, and Arabidopsis) and collinear segments (Mt/Lj). The vertical axis indicates percent of Ks values, and the horizontal axis denotes Ks (one bin corresponds with a Ks value of 0.1). Ks values are averages of three adjacent homologs within a collinear segment, as described in Methods. The Arabidopsis distribution is taken from Simillion et al. (49).