Identification of multiple independent horizontal gene transfers into poxviruses using a comparative genomics approach

Abstract

Background: Poxviruses are important pathogens of humans, livestock and wild animals. These large dsDNA viruses have a set of core orthologs whose gene order is extremely well conserved throughout poxvirus genera. They also contain many genes with sequence and functional similarity to host genes which were probably acquired by horizontal gene transfer.Although phylogenetic trees can indicate the occurrence of horizontal gene transfer and even uncover multiple events, their use may be hampered by uncertainties in both the topology and the rooting of the tree. We propose to use synteny conservation around the horizontally transferred gene (HTgene) to distinguish between single and multiple events.

Results: Here we devise a method that incorporates comparative genomic information into the investigation of horizontal gene transfer, and we apply this method to poxvirus genomes. We examined the synteny conservation around twenty four pox genes that we identified, or which were reported in the literature, as candidate HTgenes. We found support for multiple independent transfers into poxviruses for five HTgenes. Three of these genes are known to be important for the survival of the virus in or out of the host cell and one of them increases susceptibility to some antiviral drugs.

Conclusion: In related genomes conserved synteny information can provide convincing evidence for multiple independent horizontal gene transfer events even in the absence of a robust phylogenetic tree for the HTgene.

Figure 1

A schematic representation of the ML species tree topology. Branch lengths are not to scale. Numbers above branches are percent bootstrap values, numbers below the branches indicate the percentage of gene trees that supported the branch. Taxonomic groups are labelled and shaded for clarity. The term "clade II" poxviruses is used for simplicity and refers to the group of yatapox, deerpox, capripox and suipox, as per convention [11].

Figure 2

Model gene order comparisons for inferring the number of horizontal gene transfer events. The relative order of genes around the gene of interest is illustrated. Orthologous genes are shaded in the same color and lined up vertically. A white box indicates that the gene is present in the genome, but not at the expected (i.e., equivalent) location. Horizontal lines indicate genomic segments. Thick horizontal lines indicate that the connected genes are neighboring genes of the specified type (where gene of interest is present) or have fewer than three intervening genes of the specified type (where gene of interest is absent). Thin horizontal lines signify three to six intervening genes. Vertical lines separate discontiguous genomic regions. (A) The gene of interest is in an equivalent location in all genomes where it is found. This supports a single origin (i.e., horizontal transfer) of this gene. (B) The gene of interest is found in two different genomic locations and the relative arrangement of other genes in each of these locations is conserved (i.e., there is no support for a local rearrangement). This supports two independent transfers of this gene. (C) The gene of interest is found in two different genomic locations but these regions are not conserved between genomes, so it is not possible to infer if it was a single transfer followed by genome rearrangement, or two independent transfer events.

Figure 3

Synteny conservation around IL-10 in poxvirus genomes. As for Fig. 2. The gene of interest, IL-10, is colored red with a bold outline. A horizontal line that crosses a box, indicates that this gene is not part of the contiguous segment. (A) Neighborhood with respect to "core families". The Serine/Threonine Kinase gene has experienced a local inversion in the parapox genomes and is shown in two columns for clarity; (B) Neighborhood with respect to "ortholog families".