Incongruent patterns of local and global genome size evolution in cotton

Abstract

Genome sizes in plants vary over several orders of magnitude, reflecting a combination of differentially acting local and global forces such as biases in indel accumulation and transposable element proliferation or removal. To gain insight into the relative role of these and other forces, approximately 105 kb of contiguous sequence surrounding the cellulose synthase gene CesA1 was compared for the two coresident genomes (AT and DT) of the allopolyploid cotton species, Gossypium hirsutum. These two genomes differ approximately twofold in size, having diverged from a common ancestor approximately 5-10 million years ago (Mya) and been reunited in the same nucleus at the time of polyploid formation, approximately 1-2 Mya. Gene content, order, and spacing are largely conserved between the two genomes, although a few transposable elements and a single cpDNA fragment distinguish the two homoeologs. Sequence conservation is high in both intergenic and genic regions, with 14 conserved genes detected in both genomes yielding a density of 1 gene every 7.5 kb. In contrast to the twofold overall difference in DNA content, no disparity in size was observed for this 105-kb region, and 555 indels were detected that distinguish the two homoeologous BACs, approximately equally distributed between AT and DT in number and aggregate size. The data demonstrate that genome size evolution at this phylogenetic scale is not primarily caused by mechanisms that operate uniformly across different genomic regions and components; instead, the twofold overall difference in DNA content must reflect locally operating forces between gene islands or in largely gene-free regions.

Figure 1

The evolutionary history of diploid and tetraploid Gossypium, as inferred by numerous chloroplast and nuclear data sets (Seelanan et al. 1997; Small et al. 1998; Cronn et al. 2002). Genome groups designate closely related species, as determined by interspecific meiotic pairing and chromosome size (Endrizzi et al. 1985). All diploid species have the same base chromosome number (n = 13); however, each genome group varies in genome size (1C content indicated in circles). Polyploid species are thought to have originated 1–2 Mya, following divergence of their diploid progenitors 5–10 Mya.

Figure 2

Pairwise alignment of CesA1 homoeologous BACs, A_T and D_T, to scale. A_T and D_T are genes shown as block diagrams: numbered boxes are predicted corresponding to the list presented in Table 1; rTE1, rTE2, and rTE3 represent the three largely intact retrotransposons identified (rTE1 encompasses two predicted copia elements); the POGO and MuDR-like TEs are indicated individually, as is the ycf2 fragment of plastidial origin. The bottom panel indicates a continuous window of sequence identity between the two BACs, scaled from 50% to 100%.

Figure 3

Nested insertions of retroelements in the A_T BAC of Gossypium hirsutum. The outer copia is shown in gray and the inner copia in black. Four LTRs, corresponding to the two copia insertions, are shown as triangles. The three coding domains of the copias, reverse transcriptase (RT), integrase (INT), and protease (PRO), are designated by the labeled boxes within the LTRs. Surrounding the copia nest is a single POGO element that is shared by A_T and D_T, and which was split in two when the copias inserted.

Figure 4

The spectrum of small indels inferred from sequence alignment of the A_T and D_T CesA1 BACs. For A_T (solid bars), “differences” are gapped positions relative to D_T, whereas for D_T (open bars), differences reflect gaps relative to A_T. These indels are not although the phylogenetically polarized, spectrum of indels is equivalent in the two genomes.