Comparative genomics of Brassica oleracea and Arabidopsis thaliana reveal gene loss, fragmentation, and dispersal after polyploidy

Abstract

We sequenced 2.2 Mb representing triplicated genome segments of Brassica oleracea, which are each paralogous with one another and homologous with a segmentally duplicated region of the Arabidopsis thaliana genome. Sequence annotation identified 177 conserved collinear genes in the B. oleracea genome segments. Analysis of synonymous base substitution rates indicated that the triplicated Brassica genome segments diverged from a common ancestor soon after divergence of the Arabidopsis and Brassica lineages. This conclusion was corroborated by phylogenetic analysis of protein families. Using A. thaliana as an outgroup, 35% of the genes inferred to be present when genome triplication occurred in the Brassica lineage have been lost, most likely via a deletion mechanism, in an interspersed pattern. Genes encoding proteins involved in signal transduction or transcription were not found to be significantly more extensively retained than those encoding proteins classified with other functions, but putative proteins predicted in the A. thaliana genome were underrepresented in B. oleracea. We identified one example of gene loss from the Arabidopsis lineage. We found evidence for the frequent insertion of gene fragments of nuclear genomic origin and identified four apparently intact genes in noncollinear positions in the B. oleracea and A. thaliana genomes.

Figure 1.

Phylogenetic Relationships of Conserved Genes. Left, overview of the positions of the members of the analyzed gene family across the sequence contigs. Right, phylogenies illustrated as maximum likelihood cladograms. Bootstrap support, calculated from 1000 replicates, is shown at each branch as a percentage.

Figure 2.

Alignment of Conserved Genes. Vertical lines denote sequence contigs, and horizontal lines join members of homologous/paralogous gene families. The triangle by each gene model name indicates the coding stand of the gene and is color-coded to indicate the Munich Information Center for Protein Sequences FunCat functional classification of the predicted protein, as shown in the key at bottom.

Figure 3.

Phylogenic Analysis of an Inverted Duplicate Array Present in Both A. thaliana Daughter Segments of the α Genome Duplication. Left, overview of the positions of the members of the analyzed gene family across the sequence contigs. Right, phylogeny illustrated as a maximum likelihood cladogram. Bootstrap support, calculated from 1000 replicates, is shown at each branch as a percentage.

Figure 4.

Phylogenic Analysis of a Parallel Triplicate Array Present in Both A. thaliana Daughter Segments of the α Genome Duplication. Left, overview of the positions of the members of the analyzed gene family across the sequence contigs. Right, phylogeny illustrated as a maximum likelihood cladogram. Bootstrap support, calculated from 1000 replicates, is shown at each branch as a percentage. The subgroup represents a cluster of sequences (At5g47350, B.m00415, At5g47340, and At5g47330) for which a robust phylogeny could not be constructed.