The role of genomic data in the discovery, annotation and evolutionary interpretation of the interferon-lambda family

Abstract

Background: Type-I interferons, type-II interferons, and the IL-10 family are helical cytokines with similar three-dimensional folds. However, their homologous relationship is difficult to detect on the basis of sequence alone. We have previously described the discovery of the human type-III interferons (IFN lambda-1, -2, -3 or IL-29, IL-28A, IL-28B), which required a combination of manual and computational techniques applied to predicted protein sequences.

Principal findings: Here we describe how the use of gene structure analysis and comparative genomics enabled a more extensive understanding of these genes early in the discovery process. More recently, additional mammalian genome sequences have shown that there are between one and potentially nine copies of interferon lambda genes in each genome, and that several species have single exon versions of the interferon lambda gene.

Significance: The variable number of single exon type-I interferons in mammals, along with recently identified genes in zebrafish homologous to interferons allows a story of interferon evolution to be proposed. This model suggests that the gene duplications and single exon retrotransposons of mammalian type-III interferons are positively selected for within a genome. These characteristics are also shared with the fish interferons and could be responsible for the generation of the IL10 family and also the single exon type-I interferons.

Figure 1. Gene structure of human IFNL1, IFNL2 and IFNL3.

Each numbered box represents an exon (not drawn to scale). The shorter extensions at the terminal exons represent noncoding DNA. The letters inside the boxes are protein structural features: (S)ecretion peptide, helix (A), helix (B), helix (C) and helix (D).

Figure 2. Phylogenetic tree using the actual or predicted DNA sequences of spliced interferon lambda genes or pseudogenes in human, mouse, dog and guinea pig.

Human and mouse gene loci are labeled as in Lasfar et. al. . The labels for dog and guinea pig are not intended to be suggestions for permanent names. The “dog1” gene is the single exon gene from dog on chromosome 24, while “dog2” and “dog3” are multi exon genes located near each other on chromosome 1 (see Supplementary Figure S1 for gene sequences). The sequence for the more distant mouse pseudogene, named interferon lambda 4 in Lasfar et. al., was excluded because we were unable to find the locus. The optimal tree is drawn to scale, with bootstrap values shown at branch points and branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree (sum of branch length = 1.0194). All positions containing gaps and missing data were eliminated from the dataset, and the full codons were used. There were a total of 314 nucleotide positions in the final dataset, compared to the longest sequence of 622 nucleotides (the pseudogenes shortened the total number of columns which were available for alignment).