Intron gain and loss in segmentally duplicated genes in rice

Abstract

Background: Introns are under less selection pressure than exons, and consequently, intronic sequences have a higher rate of gain and loss than exons. In a number of plant species, a large portion of the genome has been segmentally duplicated, giving rise to a large set of duplicated genes. The recent completion of the rice genome in which segmental duplication has been documented has allowed us to investigate intron evolution within rice, a diploid monocotyledonous species.

Results: Analysis of segmental duplication in rice revealed that 159 Mb of the 371 Mb genome and 21,570 of the 43,719 non-transposable element-related genes were contained within a duplicated region. In these duplicated regions, 3,101 collinear paired genes were present. Using this set of segmentally duplicated genes, we investigated intron evolution from full-length cDNA-supported non-transposable element-related gene models of rice. Using gene pairs that have an ortholog in the dicotyledonous model species Arabidopsis thaliana, we identified more intron loss (49 introns within 35 gene pairs) than intron gain (5 introns within 5 gene pairs) following segmental duplication. We were unable to demonstrate preferential intron loss at the 3' end of genes as previously reported in mammalian genomes. However, we did find that the four nucleotides of exons that flank lost introns had less frequently used 4-mers.

Conclusion: We observed that intron evolution within rice following segmental duplication is largely dominated by intron loss. In two of the five cases of intron gain within segmentally duplicated genes, the gained sequences were similar to transposable elements.

Figure 1

Flow chart for the identification of intron gain and intron loss within segmentally duplicated rice genes. TE, transposable element.

Figure 2

Example of intron loss. Multiple alignment of the two duplicated rice genes (top; LOC__Os03g18690.1, LOC_Os07g49150.1) and their putative orthologous Arabidopsis gene (bottom; At4g29040.1) suggests that the third intron of LOC_ Os07g49150.1 was lost. Yellow inserts indicate conserved introns across the three genes while red indicates lost intron. The phase of the intron is inserted into the alignment. All conserved introns are phase 0 whereas the lost intron is phase 2. The two rice genes and putative Arabidopsis ortholog encode a 26S proteasome regulatory subunit 4.

Figure 3

Distribution of the sizes of the lost and gained introns. Intron lengths were binned into 100 bp bins and the number of lost and gained introns in each bin was determined and plotted against the frequency of 33,011 FLS introns within the rice genome.

Figure 4

Intron loss along the coding sequence. The positions of the lost introns were inferred from the retained intron of its corresponding duplicated gene. The whole coding sequence was divided into 10 bins. The positions of independently lost introns were placed into the corresponding bin and plotted against the frequency of all 33,011 FLS introns within the rice genome, which had been binned into the same 10 bins.

Figure 5

Extraction of the exonic 4-mers at the donor and acceptor splice sites of lost and retained introns. Duplicated rice gene 1 with a single exon and rice gene 2 and Arabidopsis orthologous gene with two exons and a single intron are shown in colored rectangles. Dashed lines indicate similar regions. Phylogeny analysis with Arabidopsis suggests an intron was lost in rice gene 1. The red ovals show the 4-mers extracted for SoR analysis.