Floral gene resources from basal angiosperms for comparative genomics research

Abstract

Background: The Floral Genome Project was initiated to bridge the genomic gap between the most broadly studied plant model systems. Arabidopsis and rice, although now completely sequenced and under intensive comparative genomic investigation, are separated by at least 125 million years of evolutionary time, and cannot in isolation provide a comprehensive perspective on structural and functional aspects of flowering plant genome dynamics. Here we discuss new genomic resources available to the scientific community, comprising cDNA libraries and Expressed Sequence Tag (EST) sequences for a suite of phylogenetically basal angiosperms specifically selected to bridge the evolutionary gaps between model plants and provide insights into gene content and genome structure in the earliest flowering plants.

Results: Random sequencing of cDNAs from representatives of phylogenetically important eudicot, non-grass monocot, and gymnosperm lineages has so far (as of 12/1/04) generated 70,514 ESTs and 48,170 assembled unigenes. Efficient sorting of EST sequences into putative gene families based on whole Arabidopsis/rice proteome comparison has permitted ready identification of cDNA clones for finished sequencing. Preliminarily, (i) proportions of functional categories among sequenced floral genes seem representative of the entire Arabidopsis transcriptome, (ii) many known floral gene homologues have been captured, and (iii) phylogenetic analyses of ESTs are providing new insights into the process of gene family evolution in relation to the origin and diversification of the angiosperms.

Conclusion: Initial comparisons illustrate the utility of the EST data sets toward discovery of the basic floral transcriptome. These first findings also afford the opportunity to address a number of conspicuous evolutionary genomic questions, including reproductive organ transcriptome overlap between angiosperms and gymnosperms, genome-wide duplication history, lineage-specific gene duplication and functional divergence, and analyses of adaptive molecular evolution. Since not all genes in the floral transcriptome will be associated with flowering, these EST resources will also be of interest to plant scientists working on other functions, such as photosynthesis, signal transduction, and metabolic pathways.

Figure 1

Well- supported evolutionary relationships among FGP species and other genomic models are shown in this phylogenic tree of seed plants. Red taxon names indicate those species for which we aim to sequence 10,000 ESTs, green taxon names indicate species for which we are sequencing 2000 ESTs, and blue taxon names indicate species for which large EST sets are already available in public databases [24, 25, 28, 29, 61] or will soon become available [70].

Figure 2

The relative frequencies of ESTs assigned to GO Biological process classes are quite similar across our study taxa. Class frequencies are shown for ten EST sets, the inferred Arabidopsis proteome, and Arabidopsis genes with moderate-to-high expression in young inflorescences (stage 3).

Figure 3

TribeMCL gene clusters (Tribes) with floral development genes vary in size and tend to include similar numbers of rice and Arabidopsis genes (left). These gene families are well represented in our EST sets (right and Tables 2 and 3). *The unigene counts (right) for the CLAVATA gene family have been halved.

Figure 4

A phylogeny for the DUF642-domain gene family indicates that two Arabidopsis genes with differential expression in petals [40] are not the products of a recent duplication event. Genes with the plant-specific but functionally uncharacterized domain family DUF642 form three clades (A, B, and C). In addition to the Arabidopsis genes with petal-specific expression patterns, Clade A includes asterid, Vitis, monocot and magnoliid genes, indicating that the clade predates the divergence of these lineages (Fig. 1). The genomes of basal-most angiosperm lineages (Amborella and Nymphaeales) and gymnosperms may also contain unsampled Clade A genes. Clade B genes were sampled from Nuphar (Nymphaeales) and Clade C genes were identified from Amborella and Pinus. Bootstrap support values (>50%) are shown above each branch. Abbreviated taxon names associated with some gene sequences: Os, Oryza sativa (rice); At, Arabidopsis thaliana; Gossyp, Gossypium spp. (cotton); G.max, Glycine max (soybean); L.japo, Lotus japonicus; M.trun, Medicago truncatula; V.vini, Vitis vinifera (grape); L.sati Lactuca sativa (lettuce); S.tube, Solanum tuberosum (potato); S.escu, Solanum esculentum (tomato); M.crys, Mesembryanthemum crystallinum (ice plant); A.cepa, Allium cepa (onion); T.aest, Triticum aestivum (wheat); H.vulg, Hordeum vulgare (barley); S.bico, Sorghum bicolor; S.offi Saccharum officinarum (sugarcane); Z.mays, Zea mays (maize); L.tuli, Liriodendron tulipifera (tulip poplar); P.amer, Persea americana (avocado); N.adve Nuphar advena; A.tric, Amborella trichopoda.