Do Viruses Exchange Genes across Superkingdoms of Life?

Malik, Shahana S.; Azem-e-Zahra, Syeda; Kim, Kyung Mo; Caetano-Anollés, Gustavo; Nasir, Arshan

The comparative genomic approach presented here is useful to postulate data-driven hypotheses regarding viral evolution, especially because large-scale sequence-based phylogenetic analysis on viral genes and genomes is sometimes prohibitive due to high nucleotide and amino acid sequence variability within and between viral genome groups. The comparative genomic approach however does suffer from some limitations. First, in the absence of phylogenetic reconstruction, structural similarities are considered homologies. Protein domain sharing could be a result of convergent evolution, HGT and vertical evolution. However, protein domains grouped into FSFs are believed to have evolved from a common evolutionary ancestor and thus cannot (by SCOP definitions) be subject to convergent evolution. Specifically, the interlocking of amino acid side-chains in the buried cores of protein domain structures represents a distinctive "fingerprint, " which is recognizable among member domains of any particular superfamily. Amino acid substitutions that occur over evolutionary timespans do not distort the 3D fingerprint characteristic of each superfamily without risking loss of the protein fold, and ultimately its biochemical function (e.g., bacterial MreB and FtsZ proteins that are prokaryotic homologs of eukaryotic actin and tubulin, respectively). That is the reason why despite low sequence identities, member protein domains of SCOP FSFs share recognizable structural and biochemical similarities, which are taken as evidence for common origin. Empirically, the odds of originating the same fingerprint (a product of multiple interactions occurring between many amino acid side chains) independently are considered to be extremely low (e.g., between 3 and 5% in Gough, 2005) . In other words, each known fold or FSF is a unique discovery in evolution. Given the small number of expected folds that exist in nature (∼1,500), convergence becomes an unlikely scenario.