Multiple paleopolyploidizations during the evolution of the Compositae reveal parallel patterns of duplicate gene retention after millions of years

Abstract

Of the approximately 250,000 species of flowering plants, nearly one in ten are members of the Compositae (Asteraceae), a diverse family found in almost every habitat on all continents except Antarctica. With an origin in the mid Eocene, the Compositae is also a relatively young family with remarkable diversifications during the last 40 My. Previous cytologic and systematic investigations suggested that paleopolyploidy may have occurred in at least one Compositae lineage, but a recent analysis of genomic data was equivocal. We tested for evidence of paleopolyploidy in the evolutionary history of the family using recently available expressed sequence tag (EST) data from the Compositae Genome Project. Combined with data available on GenBank, we analyzed nearly 1 million ESTs from 18 species representing seven genera and four tribes. Our analyses revealed at least three ancient whole-genome duplications in the Compositae-a paleopolyploidization shared by all analyzed taxa and placed near the origin of the family just prior to the rapid radiation of its tribes and independent genome duplications near the base of the tribes Mutisieae and Heliantheae. These results are consistent with previous research implicating paleopolyploidy in the evolution and diversification of the Heliantheae. Further, we observed parallel retention of duplicate genes from the basal Compositae genome duplication across all tribes, despite divergence times of 33-38 My among these lineages. This pattern of retention was also repeated for the paleologs from the Heliantheae duplication. Intriguingly, the categories of genes retained in duplicate were substantially different from those in Arabidopsis. In particular, we found that genes annotated to structural components or cellular organization Gene Ontology categories were significantly enriched among paleologs, whereas genes associated with transcription and other regulatory functions were significantly underrepresented. Our results suggest that paleopolyploidy can yield strikingly consistent signatures of gene retention in plant genomes despite extensive lineage radiations and recurrent genome duplications but that these patterns vary substantially among higher taxonomic categories.

FIG. 1.—

Histograms of Compositae gene duplication ages with mixture model and SiZer analyses. (A) Gerbera hybrida, Mutisieae. (B) Carthamus tinctorius, Cardueae. (C) Cichorium intybus, Cichorioideae. (D) Helianthus argophyllus, Heliantheae. Plots of normal distributions were fitted from mixture model analyses; green lines represent the basal Compositae paleopolyploidization, whereas the blue and orange lines represent the independent Mutisieae and Heliantheae paleopolyploidizations, respectively. SiZer maps below histograms identify significant features at corresponding K_s values with blue areas indicating significantly increasing slopes, red indicating significantly decreasing slopes, purple representing no significant slope change, and gray indicating not enough data for the test.

FIG. 2.—

Phylogeny of Compositae tribes and outgroups displaying observed paleopolyploid events and the rapid radiation of tribes. Branch lengths are mean rate corrected K_s values from 36 nuclear orthologs (table S1, Supplementary Material online). Black dots indicate inferred paleopolyploidizations in the evolution of the Compositae. Topology is a consensus phylogeny of the 36 nuclear orthologs (supplementary fig. S1, Supplementary Material online).

FIG. 3.—

GO annotations of Compositae whole transcriptome and paleologs. The leftmost column displays the pooled Compositae transcriptome of 18 species, whereas the remaining columns represent paleologs retained in each tribe from the basal Compositae genome duplication and the basal Heliantheae genome duplication. Colors represent percent of transcriptome a particular GO category composes. Superscripts indicate significantly different groups as determined by chi-square tests (P < 0.05). GO categories that are significantly enriched or reduced among paleologs relative to nonpaleologs in at least three comparisons are indicated with ± signs.