Evolutionary dynamics of the LTR retrotransposons roo and rooA inferred from twelve complete Drosophila genomes

Abstract

Background: Roo is the most abundant retrotransposon in the fruit fly Drosophila melanogaster. Its evolutionary origins and dynamics are thus of special interest for understanding the evolutionary history of Drosophila genome organization. We here study the phylogenetic distribution and evolution of roo, and its highly diverged relative rooA in 12 completely sequenced genomes of the genus Drosophila.

Results: We identify a total of 164 roo copies, 57 of which were previously unidentified copies that occur in 9 of the 12 genomes. Additionally we find 66 rooA copies in four genomes and remnants of this element in two additional genomes. We further increased the number of elements by searching for individual roo/rooA sequence domains. Most of our roo and rooA elements have been recently inserted. Most elements within a genome are highly similar. A comparison of the phylogenetic tree of our roo and rooA elements shows that the split between roo and rooA took place early in Drosophila evolution. Furthermore there is one incongruency between the species tree and the phylogenetic tree of the roo element. This incongruency regards the placement of elements from D. mojavensis, which are more closely related to D. melanogaster than elements from D. willistoni.

Conclusion: Within genomes, the evolutionary dynamics of roo and rooA range from recent transpositional activity to slow decay and extinction. Among genomes, the balance of phylogenetic evidence, sequence divergence distribution, and the occurrence of solo-LTR elements suggests an origin of roo/rooA within the Drosophila clade. We discuss the possibility of a horizontal gene transfer of roo within this clade.

Figure 1

Roo element structure and phylogenetic tree of the 12 sequenced Drosophila species. a) The canonical roo element is 9092 basepairs (bp) long, has one 7083 bps long ORF (light gray box), and is flanked by 428/429 bps long LTR sequences (dark gray boxes). The ORF encodes the gag, pol and env genes. b) The phylogenetic tree of the Drosophila species is taken from [43]. The number of identified roo and rooA elements is listed for each species. Roo and rooA elements are highlighted with red and blue bars, respectively.

Figure 2

Insertion time distribution. The histograms show the estimated insertion time distribution of roo and rooA (in red and blue bars, respectively), based on the intra-element LTR divergence. Notice the different scales on the y-axis. The vertical red and blue lines indicate the median insertion time in each genome for roo and rooA, respectively. Two roo elements of D. mojavensis had to be excluded because their LTR sequences were falsely identified. The histogram in the bottom right corner shows the insertion times of all roo and rooA elements. myr: million years.

Figure 3

Phylogenetic tree of the 230 identified roo and rooA elements. The tree is based on the protein coding sequence of roo and rooA. It was constructed using PhyML with an approximate likelihood ratio test to estimate the statistical support of the tree topology [27], as shown by the numbers at the branches. All branches have a very high support. There is a clear division between roo (red background) and rooA (blue background) elements on the tree. Specifically, all roo elements from one species form a clade and the same holds for rooA. The black triangles at some leaves indicate the divergence of the elements associated with this leaf, with long triangles indicating great divergence. The number in brackets behind each species name indicates how many elements are present in the respective species.

Figure 4

Three hypotheses for the evolutionary dynamics of the roo and rooA element in the 12 Drosophila genomes. Red and blue branches indicate species harboring roo and rooA, respectively. Black branches indicate the loss of both elements. a) Roo/rooA may have originated in the common ancestor of all 12 Drosophila genomes with the split between the two elements shortly thereafter. Both elements were then vertically transmitted, as shown by the red and blue branches. Roo was lost in 3 species and rooA in six or seven (depending on the correctness of the rooA element in D. mojavensis (see text for details)). b) Roo/rooA may have arisen in the common ancestor of the Sophophora group and split shortly thereafter into separate lineages. Both elements were vertically transmitted to all descendant genomes, with a loss of roo in one genome and of rooA in four genomes. In this scenario, a horizontal transfer of roo to the genome of D. mojavensis took place after the split from D. willistoni (red arrow), but the exact time point of transmission cannot be determined. c) Roo/rooA might have originated in the common ancestor of all 12 genomes. With the split of the Sophophora group the element evolved into the roo element in the Sophophora group, and into the rooA element in the other three genomes. While roo was vertically transferred to all genomes, rooA is today only present in the D. mojavensis member of this clade. A horizontal transfer of roo to the genome of D. mojavensis took place after the split from D. willistoni. Later, another horizontal transfer, this time of rooA, took place from D. mojavensis to the common ancestor of the melanogaster subgroup where it then spread in all five genomes.