Do Viruses Exchange Genes across Superkingdoms of Life?

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

Remarkably, the 29 FSFs shared exclusively between bacterioviral and bacterial proteomes included several viral hallmark proteins involved in phage (virus) assembly such as the gp9 and gp10 proteins, head-binding, head-to-tail joining, head decoration, and tail proteins, along with the major coat proteins of ssDNA harboring bacterioviruses (Inoviridae) and the dimerization domain of bacteriophage T4 recombination endonuclease VII (Table 1 ). In addition, the coiled-coiled domain of bacterial neurotoxin involved in host virulence was also detected. Interestingly, the majority of FSFs in group B had f -values close to 0 indicating their rare presence in bacterial proteomes (Table 1) . Collectively, therefore, the enrichment of the B Venn group in viral hallmark functions with negligible presence in bacterial proteomes suggests that these genes were likely acquired by bacterial cells from viruses via virus-to-cell HGT, a phenomenon that has been assumed to be relatively less frequent than cell-to-virus HGT (Moreira and Lopez-Garcia, 2009) , though now increasingly being revisited (Forterre, 2016) . Similarly, viral hallmark proteins such as the viral capsid and coat-related proteins (e.g., the "jelly-roll" and "double jelly-roll" folds) (Abrescia et al., 2012) , viral glycoproteins and matrix proteins, the integrase proteins of retroviruses and HIV, and toxins were part of the 37 FSFs shared exclusively between eukaryoviruses and Eukarya (the E Venn group) with low f -values in eukaryal proteomes ( Table 1) . These viral hallmark proteins shared exclusively between eukaryoviruses and eukaryotes could therefore also represent episodes of virus-to-cell gene transfer. In turn, other E FSFs such as the CCCH zinc finger domains (involved in regulation and DNA binding), CAD and PB1 domains (cell cycle and apoptosis), CRAL/TRIO domains (likely functional components of the visual cycle), TRAF-domain like (involved in stress response, immunity, apoptosis, among other roles), and others were near ubiquitous in eukaryotic proteomes (i.e., f -value close to 1.0, FSFs shared exclusively between the proteomes of host superkingdoms, Archaea (A), Bacteria (B), and Eukarya (E), and the proteomes of their viruses, archaeoviruses (AV), bacterioviruses (BV), and eukaryoviruses (EV). FSFs are identified both by SCOP numeric IDs and alpha-numeric concise classification strings (ccs). FSF distribution (f-values, number of proteomes in a superkingdom or virus group encoding an FSF/total number of proteomes in that superkingdom or virus group) are also listed. FSF b.57.1 was also detected in eukaryoviruses in addition to Bacteria and bacterioviruses and FSFs b.121.2 and b.121.5 were also detected in bacterioviruses in addition Eukarya and eukaryoviruses possibly indicating genetic crosstalk or ancient ancestry (read text). These FSFs are highlighted in bold. Table 1 ). These "cell-like" proteins detected in eukaryoviruses could therefore suggest recent gene capture by viruses from cells (i.e., cell-to-virus HGT) as likely part of viral mimicry of cellular proteins to interfere with the antiviral response (Elde and Malik, 2009 ). In summary, a large number of FSFs shared exclusively between viruses and their host genomes had rare presence in hosts and were involved in virus-hallmark functions suggesting these genes likely originated in viral lineages and were later transferred to their host cells.