An endogenous foamy-like viral element in the coelacanth genome

Abstract

Little is known about the origin and long-term evolutionary mode of retroviruses. Retroviruses can integrate into their hosts' genomes, providing a molecular fossil record for studying their deep history. Here we report the discovery of an endogenous foamy virus-like element, which we designate 'coelacanth endogenous foamy-like virus' (CoeEFV), within the genome of the coelacanth (Latimeria chalumnae). Phylogenetic analyses place CoeEFV basal to all known foamy viruses, strongly suggesting an ancient ocean origin of this major retroviral lineage, which had previously been known to infect only land mammals. The discovery of CoeEFV reveals the presence of foamy-like viruses in species outside the Mammalia. We show that foamy-like viruses have likely codiverged with their vertebrate hosts for more than 407 million years and underwent an evolutionary transition from water to land with their vertebrate hosts. These findings suggest an ancient marine origin of retroviruses and have important implications in understanding foamy virus biology.

Figure 1. Comparison of the genome structures between CoeEFV and typical exogenous foamy virus.

LTR, long-terminal repeat; PBS, primer-binding site.

Figure 2. Conserved domain alignment of the CoeEFV Env protein and foamy virus envelope protein domain (pfam03408).

Numbers refer to the position in the original CoeEFV env protein or conserved domain pfam03408. Identical amino acid residues are highlighted in red, and black and blue indicate gaps or different amino acid residues, respectively. The E-value was generated by Conserved Domain search.

Figure 3. Retrovirus phylogeny.

The phylogeny is the 50% majority-rule consensus tree reconstructed based on the conserved region of Pol protein of CoeEFV and various endogenous and exogenous retroviruses using MrBayes 3.1.2. Posterior probabilities are shown near the selected nodes. The foamy virus and lentivirus clades are highlighted in red and blue, respectively. The full tree and taxon labels are depicted in Fig. S2. WDSV, walleye dermal sarcoma virus; SnRV, snakehead retrovirus.

Figure 4. Phylogenetic congruence of foamy viruses (right) and their hosts (left).

Associations between foamy viruses and their hosts are indicated by connecting lines. The scale of the host phylogeny (left) indicates millions of years. The foamy virus phylogeny (right) is the 50% majority-rule consensus tree inferred from conserved region of foamy virus and Class III retrovirus Pol protein alignment with MrBayes 3.1.2. The Bayesian phylogeny is well supported with all nodes showing posterior probability of 1.00. Branch lengths are in expected amino acid changes per site. Coelacanth image courtesy of Robbie Cada.

Figure 5. A plot of the correlation between foamy virus divergence and their vertebrate hosts' divergence times.

The plot depicts host branch length (in millions of years) versus virus branch length (in expected amino acid substitutions per site) for every branch (both internal and external). The virus branch lengths are derived from the virus tree in Fig. 4.