From Fossils to Function: Exploring the Evolutionary And Functional Diversity Of HERV Envelope Proteins Across Primate Lineages

Endogenous retroviruses (ERVs) are considered “fossils” as they represent the footprints of previous retroviral infections and are thereby vertically transmitted through the germline in a Mendelian fashion. Throughout evolution, ERVs were subjected to numerous amplification and transposition events resulting in multiple copies of proviruses in the DNA of all the cells. Overall, ERVs are present in abundance in almost all the vertebrates, particularly ~8% of the human genome consists of human ERVs (HERVs). ERVs genome is made up of four main protein coding genes i.e. gag, pro, pol and env flanked by the long terminal repeats (LTRs) at both 5’ and 3’ ends. Among the four genes, the env gene that encoding Envelope glycoprotein (Env), mediates entry and attachment of the virus with cellular receptors thereby promoted the endogenization activities. Overtime, env gene experienced various selection pressures that shaped its evolution and function. Therefore, the env gene of the ERVs is extensively investigated for its role in various diseases such as cancer, autoimmunity, neurodegenerative diseases and inflammation.

Hence, the overall aim of the present project was to characterize in detail env diversity across primate species, to characterize their acquisition and evolutionary dynamics within the host genome and gain insights about their past and residual coding potential, to subsequently investigate the tropism of ancestral and current HERV Env proteins.

In particular, the first part of the project was focused on the HERVK(HML2) group, being highly relevant for its recent diffusion in the human genome that led to a high degree of functional conservation of its members. In fact, even though ERVs are present in all the vertebrates, some are specific to primates such as HERV-K, consisting of 10 HML subtypes and including the most recently acquired elements. Particularly, HML2 is the youngest clade, having some human-specific integrations, and while it has been widely described in humans its presence and distribution in non-human primates remain poorly characterized. To investigate HML2 distribution in non-human primates, the present study focused on the characterization of HML2 integrations in Macaca fascicularis and Macaca mulatta which are the most evolutionarily distant species related to humans in the Catarrhini parvorder. We identified overall 208 HML2 proviruses for M. fascicularis (77) and M. mulatta (131). Among them, 46 proviruses are shared by the two species while the others are species specific. Only 12 proviruses were shared with humans, confirming that the major wave of HML2 diffusion in humans occurred after macaques’ divergence. Phylogenetic analysis confirmed structural variations between HML2 macaques’ species-specific proviruses, and the ones shared between macaques and humans. The HML2 loci were characterized in terms of structure, focusing on potential residual open reading frames (ORFs) for gag, pol, and env genes for the latter being reported to be expressed in human pathological conditions. The analysis identified highly conserved gag and pol genes, while the env genes had a very divergent nature. Of the 208 HML2 proviral sequences present in Macaca species, 81 sequences form a cluster having a MER11A, a characteristic HML8 LTR sequence, insertion in the env region indicating a recombination event that occurred between the HML2 env gene and the HML8 LTR. This recombination event, which was shown to be present only in a subset of macaques’ shared sequences and species- specific sequences, highlights a recent viral activity leading to the emergence of an env variant specific to the Old World Monkeys (OWMs).

The exhaustive analysis of HML2 in humans and Macaque genomes indicated the divergent nature of env gene due to recombinations, we further aimed to reconstruct representative Env prototype sequences and screen them across the primate genomes to study the diversity patterns of ERVs. Hence, we reconstructed 32 Env sequences representing the prototypes of these ancestral proteins in Class I, Class II, and Class III HERVs. These reconstructed Envs were then employed in diverse methods comprising similarity search, phylogenetic analysis, and examination of recombination events occurred within primates’ genomes that were applied to 43 primate species across the Catarrhini and Platyrrhini parvorders. Through a comprehensive pipeline we reconstitute a phylogenetic distribution of ERV based specifically on the env genes, showing that the ERVs have been prevalent and widely distributed across the primate lineage. We observed for the first time the presence of some HML group in the Platyrrhini parvorder, indicating initiation of spread of HML supergroup before the split between New World Monkeys (NWM) and Old World Monkeys (OWM) i.e. even before 40 mya. Importantly, we confirmed notable interclass and intra-class env recombination events showing the phenomenon of “env snatching” among primates’ ERVs. As a result, we demonstrate that tracing the diversity patterns of ERVs’ env provides relevant insights into the retroviral evolutionary history of ERVs in Catarrhini and Platyrrhini parvorders. Overall, our findings reveal that env recombination contributes to the diversification of ERVs, thereby broadening our comprehension of retroviral and primate evolution.
One aspect of the study was to understand the evolutionary dynamics of HERVs with respect to the diversity of env gene and how such divergence leads to gain or loss of function. The other aspects were to study the interaction of HERV Env with the human cellular proteins. HERVs Env protein are known to be expressed in various physiological and pathological conditions such as cancer, autoimmune diseases, neurodegenerative diseases, etc. and hence identifying their interactions with the cellular proteins will help better understand the role of HERVs in diseased and normal conditions. To serve this purpose, we applied deep learning to our study and build an unsupervised learning autoencoder model to detect the protein-protein interactions for HERV Env proteins. With this model, we detected several human cellular receptors that interact with the HERVs’ Env proteins. With the help of the model, we predicted interactions for Syncytin-1, HERV-W, HERV-T, HML2-Rec and Np9. A significant pattern emerged from the analysis concerning the role of Syncytin-1, HERV-W, HERV-T, HML2-Rec and Np9 in sensory and signaling pathways. Although each HERV seems to make distinct contributions to specific pathways, they all share a focus on sensory perception and cellular communication highlighting their role in pathological diseases such as cancer, neurodegenerative diseases and autoimmune diseases. This indicates that HERVs may together play a role in the evolution of complex human biology. Additionally, the shared pathways among HERVs may reflect a common evolutionary strategy employed by retroviral elements to affect host biology. Therefore, experimentally exploring the specific functions of HERVs in these pathways could offer important insights into their impact on human health and disease.