Virus world hypothesis

Model

The virus-first model envisions the pre-cellular stage of life as a diverse pool of competing, virus-like replicators dwelling within networks of inorganic molecules (the so-called "primordial soup"), exchanging genetic material freely and lacking both full-fledged ribosomes and independent metabolism.^[1][8] RNA would have arisen first, followed by retroid elements and then DNA viruses,^[10] an order that links the hypothesis to the RNA world hypothesis, as contemporary viruses are argued to reflect evolution spanning from an early RNA world to the modern DNA/protein world.^[1][11] Viroids and ribozymes, as the simplest known self-replicating RNA elements, could be considered the closest extant analogues of these early replicators.^[11]

Within this framework, Patrick Forterre proposed that the transition from an RNA world to the modern DNA-based cellular world was itself driven by viruses: three independent DNA viruses may have each displaced the ancestral RNA genome in three separate primordial cell lineages, giving rise to the Bacteria, Archaea, and Eukarya.^[12][13] This accounts for the observed lack of homology between the core DNA polymerases used by the three domains.^[13] Forterre introduced the virocell concept to formalise the distinction between the metabolically inert extracellular virion and the metabolically active infected cell, arguing that the latter constitutes the true living form of the virus.^[14] Gill and Forterre extended the model by proposing that ancestral extracellular vesicles may have been evolutionary intermediates between simple lipid compartments and the first virions, with early ribovirocells (cells simultaneously capable of dividing and producing virions) representing a transitional stage before the existence of the last universal common ancestor (LUCA).^[15] Koonin further proposed that viral capsids may have functioned as a primitive compartment that helped pioneer the plasma membrane architectures eventually adopted by cells.^[16]

Phylogenetics

The "virus hallmark genes" are a distinct set of conserved genes shared across a wide range of virus groups and notably absent from cellular life.^[1] Some of the smallest virus genomes (such as parvoviruses or the single-stranded RNA tombusviruses) consist mostly of these hallmark genes, whereas in the largest viruses these genes comprise a minority.^[17] Since the hallmark genes possess only distant homologs in cellular life (which nonetheless appear also to be of viral origin), while also forming a network that connects almost all viruses, their existence therefore seems to suggest descent from a primordial genetic pool predating LUCA, thus constituting an uninterrupted gene flow from the pre-cellular stage,^[18] although Koonin specifies that double-stranded RNA and negative-strand RNA viruses are still likely to have independently evolved later.

The conventional assumption that when viral and host genes cluster phylogenetically, the parasitic virus must therefore have derived its genes from the host, has been particularly challenged by Luis P. Villarreal. Metagenomic surveys of marine environments show that the majority of photosynthesis genes detected are of viral origin,^[19] display a virus-like codon usage bias, and undergo selection independently of their hosts, suggesting they evolved within the virosphere rather than being captured from cells.^[20] Furthermore, viruses collectively constitute the largest reservoir of genetic diversity in the known biosphere, a fact which, with the addition of the hallmark genes having no clear cellular homologs, is argued by Villarreal to be difficult to reconcile with a model in which the virosphere is derived from cellular life.^[21]

Viroid-first

Unlike the main virus-first scenario, the "viroid-first" or "viroid early" version holds that viroids are not only direct descendants of pre-cellular replicators but also holdovers from the pre-protein RNA world, as evidenced by their use of ribozymes for a stage of replication. Theodor Otto Diener, the discoverer of viroids, proposed in 1989 that their small size, circular structure, ribozyme activity, and complete lack of protein-coding capacity make them more plausible relics of the RNA world than any other known entity, and described them as "living fossils" of pre-cellular evolution.^[22][23] However, a 2022 paper argued that since viroid-like agents have been found only in multicellular eukaryotes, with no representatives in bacteria or archaea, their origin likely postdates rather than predates cellular life, and their secondary structures are attuned to protein-based polymerases rather than the ribozyme polymerases of a pre-cellular setting. They proposed instead that viroids most plausibly arose by "escape" from retrozymes. The discovery of new viroid-like cccRNAs through metatranscriptome mining later caused the authors to reconsider Diener's theory.^[24][25]

Chimeric scenario

More recent work has treated the classical hypotheses as complementary. Krupovic, Dolja, and Koonin proposed a chimeric scenario in which viral replication modules descended from primordial pre-cellular replicators, consistent with the virus-first view, while genes encoding major capsid proteins were independently recruited from cellular hosts at multiple points during the diversification of the virosphere, consistent with the escape hypothesis.^[2]

In the main proposed chimeric scenario, the replication and morphogenetic modules of viruses are argued to have two different evolutionary origins:^[26]

• Replication modules may descend directly from the primordial genetic pool; no evidence exists that proteins such as RNA-dependent RNA polymerase, reverse transcriptase, or rolling-circle replication endonuclease were ever encoded by pure cellular genes, thus pointing to an early origin for viral replication, although the long course of their subsequent evolution did involve displacements by replicative genes from their cellular hosts.
• Morphogenetic modules by contrast had to have come later, as the most common capsid fold (single jelly-roll), exists in bacteria, archaea, and eukaryotes as TNF superfamily members. The cellular versions of these folds are far more widespread and varied than their viral counterparts, meaning capsid proteins were likely cellular at first.^[27]

New viral lineages could then continue to emerge from cellular sources throughout evolutionary history, while also possessing primordial replication genes, explaining the polyphyletic character of the modern viruses.^[28]

Giant viruses

Study of giant viruses has provided mixed evidence for the positioning of viruses in the tree of life.^[32] A study analyzing mimiviruses found that the majority of its core replication proteins group phylogenetically with eukaryotic cellular homologs, and that the replication genes show no evidence of recent horizontal gene transfer and are under purifying selection. The authors argued that because the eukaryotic resemblance cannot be explained by recent gene acquisition from a host, it reflects genuine shared ancestry, and concluded that mimiviruses most plausibly descended from a complex cellular ancestor by reductive evolution, contradicting the idea that it represents a relic of a pre-cellular viral lineage.^[33] Because viruses leave no fossil record, their evolutionary trajectories must be inferred entirely from extant molecular data, and no analysis to date has definitively resolved the question of viral origins.^[34]

Symbiogenic model

In a more complex iteration of the classical regressive hypothesis, the co-evolutionnary "symbiogenic" model proposed in 2012 hypothesizes that ancient cells coexisted with ancient virocells and that modern-day viruses evolved by genomic reduction.^[35] Their phylogenomic timeline places a communal cellular world first. Viruses would then emerge from a primordial cellular stem line by losing genes and complexity over time, with a key point being that the regression occurred before the advent of parasitism, and that the obligate parasitism of these cell-derived viruses developed much later.^[36]