Transposable elements drive intron gain in diverse eukaryotes

Abstract

There is massive variation in intron numbers across eukaryotic genomes, yet the major drivers of intron content during evolution remain elusive. Rapid intron loss and gain in some lineages contrast with long-term evolutionary stasis in others. Episodic intron gain could be explained by recently discovered specialized transposons called Introners, but so far Introners are only known from a handful of species. Here, we performed a systematic search across 3,325 eukaryotic genomes and identified 27,563 Introner-derived introns in 175 genomes (5.2%). Species with Introners span remarkable phylogenetic diversity, from animals to basal protists, representing lineages whose last common ancestor dates to over 1.7 billion years ago. Aquatic organisms were 6.5 times more likely to contain Introners than terrestrial organisms. Introners exhibit mechanistic diversity but most are consistent with DNA transposition, indicating that Introners have evolved convergently hundreds of times from nonautonomous transposable elements. Transposable elements and aquatic taxa are associated with high rates of horizontal gene transfer, suggesting that this combination of factors may explain the punctuated and biased diversity of species containing Introners. More generally, our data suggest that Introners may explain the episodic nature of intron gain across the eukaryotic tree of life. These results illuminate the major source of ongoing intron creation in eukaryotic genomes.

Keywords: comparative genomics; evolution; genome structure; intron; splicing.

Fig. 1.

Diversity and characteristics of Introners across eukaryotes. Results are shown from 130 genomes representing 32 lineages with putatively independent acquisitions of Introners (different colors). Leaf tip colors indicate the total number of predicted Introners for each genome. Proportion in genes is shown by the red mark, which consistently exceeds the expected values as determined by randomization within each genome (black box plots; center line denotes median; box limits denote upper and lower quartiles; whiskers denote 1.5x interquartile range). Heat maps represent predicted nucleosome occupancy for Introner insertion sites and surrounding genomic regions for genomes in which accurate nucleosome occupancy prediction was possible (see Methods). Introner insertion sites consistently show reduced histone occupancy (dark) relative to surrounding regions. For genomes within the same genus, multiple genomes are shown only if the genomes have different complements of Introner families (see Methods).

Fig. 2.

Examples of diverse Introner intron creation mechanisms. Splice sites are shown in bold, Introner boundaries are denoted by a vertical bar, and introns and exons are represented by lines and boxes, respectively. (A) Introners in Alternaria alternata do not exhibit specific sequence features associated with known DNA transposition mechanisms and appear to replicate via direct insertion. (B) Introners in Symbiodinium microadriaticum show clear evidence of 4 bp TSDs (shown in green) and TIRs (underlined), consistent with many known DNA transposons, carry their 3′ splice site, and co-opt their 5′ splice site from their TSDs upon insertion. (C) Introners in Chrysochromulina sp. show evidence of 4 bp TSDs but no evidence of TIRs, carry their 5′ splice site, and co-opt their 3′ splice site. (D) Introners in Acanthoeca sp. show clear evidence of TIRs but no TSDs. (E) Introners in Florenciella sp. do not carry either splice site and instead co-opt both from their insertion site. (F) Introners in Aureococcus sp. carry both splice sites but add an extra 12 bp into the transcript upon insertion (4 bp from the Introner + 8 bp from the TSD), (G) resulting in the addition of four amino acids to the respective protein when compared to an ortholog from a different isolate which lacks an Introner at that position.

Fig. 3.

Relative to other introns, Introners are more efficiently spliced but less frequently found in highly expressed genes. (A) Explanation of PSI. Species with greater PSI coefficients (above 0, red) have Introners spliced in more frequently than other introns in the same genome. (B) Introners are more efficiently spliced than other introns in most species, as indicated by PSI coefficients less or equal to 0 for 31/36 species. (C) Introners are overrepresented in lowly expressed genes for most species. Relative log-normalized gene expression values for Introner-containing genes relative to other Intron-containing genes are shown.