Search

The genomes of cellular organisms display CpG and TpA dinucleotide composition biases. Such biases have been poorly investigated in dsDNA viruses. Here, we show that in dsDNA virus, bacterial, and eukaryotic genomes, the representation of TpA and CpG dinucleotides is strongly dependent on genomic G + C content. Thus, the classical observed/expected ratios do not fully capture dinucleotide biases across genomes. Because a larger portion of the variance in TpA frequency was explained by G + C content, we explored which additional factors drive the distribution of CpG dinucleotides. Using the residuals of the linear regressions as a measure of dinucleotide abundance and ancestral state reconstruction across eukaryotic and prokaryotic virus trees, we identified an important role for phylogeny in driving CpG representation. Nonetheless, phylogenetic ANOVA analyses showed that few host associations also account for significant variations. Among eukaryotic viruses, most significant differences were observed between arthropod-infecting viruses and viruses that infect vertebrates or unicellular organisms. However, an effect of viral DNA methylation status (either driven by the host or by viral-encoded methyltransferases) is also likely. Among prokaryotic viruses, cyanobacteria-infecting phages resulted to be significantly CpG-depleted, whereas phages that infect bacteria in the genera Burkolderia and Staphylococcus were CpG-rich. Comparison with bacterial genomes indicated that this effect is largely driven by the general tendency for phages to resemble the host's genomic CpG content. Notably, such tendency is stronger for temperate than for lytic phages. Our data shed light into the processes that shape virus genome composition and inform manipulation strategies for biotechnological applications., (© 2024 The Authors. Molecular Ecology published by John Wiley & Sons Ltd.)

Coronaviruses (CoVs) have complex genomes that encode a fixed array of structural and nonstructural components, as well as a variety of accessory proteins that differ even among closely related viruses. Accessory proteins often play a role in the suppression of immune responses and may represent virulence factors. Despite their relevance for CoV phenotypic variability, information on accessory proteins is fragmentary. We applied a systematic approach based on homology detection to create a comprehensive catalogue of accessory proteins encoded by CoVs. Our analyses grouped accessory proteins into 379 orthogroups and 12 super-groups. No orthogroup was shared by the four CoV genera and very few were present in all or most viruses in the same genus, reflecting the dynamic evolution of CoV genomes. We observed differences in the distribution of accessory proteins in CoV genera. Alphacoronaviruses harboured the largest diversity of accessory open reading frames (ORFs), deltacoronaviruses the smallest. However, the average number of accessory proteins per genome was highest in betacoronaviruses. Analysis of the evolutionary history of some orthogroups indicated that the different CoV genera adopted similar evolutionary strategies. Thus, alphacoronaviruses and betacoronaviruses acquired phosphodiesterases and spike-like accessory proteins independently, whereas horizontal gene transfer from reoviruses endowed betacoronaviruses and deltacoronaviruses with fusion-associated small transmembrane (FAST) proteins. Finally, analysis of accessory ORFs in annotated CoV genomes indicated ambiguity in their naming. This complicates cross-communication among researchers and hinders automated searches of large data sets (e.g., PubMed, GenBank). We suggest that orthogroup membership is used together with a naming system to provide information on protein function., (© 2022 The Authors. Molecular Ecology published by John Wiley & Sons Ltd.)

Centromeres have central functions in chromosome segregation, but centromeric DNA and centromere-binding proteins evolve rapidly in most eukaryotes. The selective pressure(s) underlying the fast evolution of centromere-binding proteins are presently unknown. An attractive possibility is that selfish centromeres promote their preferential inclusion in the oocyte and centromeric proteins evolve to suppress meiotic drive (centromere drive hypothesis). We analysed the selective patterns of mammalian genes that encode kinetochore proteins and microtubule (MT)-destabilizing factors. We show that several of these proteins evolve at the same rate or faster than proteins with a role in centromere specification. Elements of the kinetochore that bind MTs or that bridge the interaction between MTs and the centromere represented the major targets of positive selection. These data are in line with the possibility that the genetic conflict fuelled by meiotic drive extends beyond genes involved in centromere specification. However, we cannot exclude that different selective pressures underlie the rapid evolution of MT-destabilizing factors and kinetochore components. Whatever the nature of such pressures, they must have been constant during the evolution of eutherian mammals, as we found a surprisingly good correlation in dN/dS (ratio of the rate of nonsynonymous and synonymous substitutions) across orders/clades. Finally, when phylogenetic relationships were accounted for, we found little evidence that the evolutionary rates of these genes change with testes size, a proxy for sperm competition. Our data indicate that, in analogy to centromeric proteins, kinetochore components are fast evolving in mammals. This observation may imply that centromere drive plays out at multiple levels or that these proteins adapt to lineage-specific centromeric features., (© 2021 John Wiley & Sons Ltd.)

Analysis of the bat viruses most closely related to SARS-CoV-2 indicated that the virus probably required limited adaptation to spread in humans. Nonetheless, since its introduction in human populations, SARS-CoV-2 must have been subject to the selective pressure imposed by the human immune system. We exploited the availability of a large number of high-quality SARS-CoV-2 genomes, as well as of validated epitope predictions, to show that B cell epitopes in the spike glycoprotein (S) and in the nucleocapsid protein (N) have higher diversity than nonepitope positions. Similar results were obtained for other human coronaviruses and for sarbecoviruses sampled in bats. Conversely, in the SARS-CoV-2 population, epitopes for CD4 + and CD8 + T cells were not more variable than nonepitope positions. A significant reduction in epitope variability was instead observed for some of the most immunogenic proteins (S, N, ORF8 and ORF3a). Analysis over longer evolutionary time frames indicated that this effect is not due to differential constraints. These data indicate that SARS-CoV-2 evolves to elude the host humoral immune response, whereas recognition by T cells is not actively avoided by the virus. However, we also found a trend of lower diversity of T cell epitopes for common cold coronaviruses, indicating that epitope conservation per se is not directly linked to disease severity. We suggest that conservation serves to maintain epitopes that elicit tolerizing T cell responses or induce T cells with regulatory activity., (© 2020 The Authors. Molecular Ecology published by John Wiley & Sons Ltd.)

The Old World (OW) arenavirus complex includes several species of rodent-borne viruses, some of which (i.e., Lassa virus, LASV and Lymphocytic choriomeningitis virus, LCMV) cause human diseases. Most LCMV and LASV infections are caused by rodent-to-human transmissions. Thus, viral evolution is largely determined by events that occur in the wildlife reservoirs. We used a set of human- and rodent-derived viral sequences to investigate the evolutionary history underlying OW arenavirus speciation, as well as the more recent selective events that accompanied LASV spread in West Africa. We show that the viral RNA polymerase (L protein) was a major positive selection target in OW arenaviruses and during LASV out-of-Nigeria migration. No evidence of selection was observed for the glycoprotein, whereas positive selection acted on the nucleoprotein (NP) during LCMV speciation. Positively selected sites in L and NP are surrounded by highly conserved residues, and the bulk of the viral genome evolves under purifying selection. Several positively selected sites are likely to modulate viral replication/transcription. In both L and NP, structural features (solvent exposed surface area) are important determinants of site-wise evolutionary rate variation. By incorporating several rodent-derived sequences, we also performed an analysis of OW arenavirus codon adaptation to the human host. Results do not support a previously hypothesized role of codon adaptation in disease severity for non-Nigerian strains. In conclusion, L and NP represent the major selection targets and possible determinants of disease presentation; these results suggest that field surveys and experimental studies should primarily focus on these proteins., (© 2017 John Wiley & Sons Ltd.)

The complement system is an innate immunity effector mechanism; its action is antagonized by a wide array of pathogens and complement evasion determines the virulence of several infections. We investigated the evolutionary history of the complement system and of bacterial-encoded complement-interacting proteins. Complement components targeted by several pathogens evolved under strong selective pressure in primates, with selection acting on residues at the contact interface with microbial/viral proteins. Positively selected sites in CFH and C4BPA account for the human specificity of gonococcal infection. Bacterial interactors, evolved adaptively as well, with selected sites located at interaction surfaces with primate complement proteins. These results epitomize the expectation under a genetic conflict scenario whereby the host's and the pathogen's genes evolve within binding avoidance-binding seeking dynamics. In silico mutagenesis and protein-protein docking analyses supported this by showing that positively selected sites, both in the host's and in the pathogen's interacting partner, modulate binding., (© 2016 John Wiley & Sons Ltd.)

Plasmodium falciparum, the causative agent of the deadliest form of malaria, is a member of the Laverania subgenus, which includes ape-infecting parasites. P. falciparum is thought to have originated in gorillas, although infection is now restricted to humans. Laverania parasites display remarkable host-specificity, which is partially mediated by the interaction between parasite ligands and host receptors. We analyse the evolution of BSG (basigin) and GYPA (glycophorin A) in primates/hominins, as well as of their Plasmodium-encoded ligands, PfRH5 and PfEBA175. We show that, in primates, positive selection targeted two sites in BSG (F27 and H102), both involved in PfRH5 binding. A population genetics-phylogenetics approach detected the strongest selection for the gorilla lineage: one of the positively selected sites (K191) is a major determinant of PfRH5 binding affinity. Analysis of RH5 genes indicated episodic selection on the P. falciparum branch; the positively selected W447 site is known to stabilize the interaction with human basigin. Conversely, we detect no selection in the receptor-binding region of EBA175 in the P. falciparum lineage. Its host receptor, GYPA, shows evidence of positive selection in all hominid lineages; selected codons include glycosylation sites that modulate PfEBA175 binding affinity. Data herein provide an evolutionary explanation for species-specific binding of the PfRH5-BSG ligand-receptor pair and support the hypothesis that positive selection at these genes drove the host shift leading to the emergence of P. falciparum as a human pathogen., (© 2015 John Wiley & Sons Ltd.)