Cheating leads to the evolution of multipartite viruses

Abstract

In multipartite viruses, the genome is split into multiple segments, each of which is transmitted via a separate capsid. The existence of multipartite viruses poses a problem, because replication is only possible when all segments are present within the same host. Given this clear cost, why is multipartitism so common in viruses? Most previous hypotheses try to explain how multipartitism could provide an advantage. In so doing, they require scenarios that are unrealistic and that cannot explain viruses with more than 2 multipartite segments. We show theoretically that selection for cheats, which avoid producing a shared gene product, but still benefit from gene products produced by other genomes, can drive the evolution of both multipartite and segmented viruses. We find that multipartitism can evolve via cheating under realistic conditions and does not require unreasonably high coinfection rates or any group-level benefit. Furthermore, the cheating hypothesis is consistent with empirical patterns of cheating and multipartitism across viruses. More broadly, our results show how evolutionary conflict can drive new patterns of genome organisation in viruses and elsewhere.

Fig 1. The evolution of multipartitism via cheating.

The ancestral monopartite population [1] consists only of cooperative viruses that each encode a full viral genome. This population is invaded first by 1 type of cheat [2], and then by a second type of cheat [3]. Each cheat has an advantage when coinfecting cells with the cooperator, and when each different type of cheat infects the same host cells, they are able to complement one another in coinfection. Consequently, provided coinfection is frequent enough, the cheats are able to drive the cooperator extinct, resulting in a multipartite population [4]. This mechanism can occur even when the final multipartite population [4] has a lower level of population productivity than the ancestral monopartite population [1]. Figure was created using BioRender.com.

Fig 2. The costs and benefits of cheating.

(a) We assume that the expected number of viral genomes produced in an infected cell depends on whether cells are infected by cooperators (that encode both genes), cheats (that each encode one gene but not the other), or both. (b) The number of progeny viral genomes of each type can be captured in a payoff matrix. Here, each entry reflects the number of viral progeny that each strategy on the left (rows) receives when it coinfects a cell alongside the strategy listed at the top (columns). (c) We analyse the dynamics of the payoff matrix using replicator dynamics, yielding simple equations for the change in relative frequency of cheats and cooperators (Eq 1 in main text). Figure was created using BioRender.com.

Fig 3. Cheats can drive the evolution of multipartitism under realistic conditions.

We used existing experimental data to derive estimates for the parameters in our analytical model (Methods; Table B in S1 Text). We then used these parameters to determine whether our model would predict the evolution of multipartitism. We plot the fraction of cells infected by multiple viral genomes (β) on the x-axis, and the minimum productivity of cells coinfected by cheats, relative to cells infected by cooperators (e/d) on the y-axis. In the shaded regions, our model predicts that multipartitism evolves; in the unshaded regions, the population remains monopartite. The top 4 panels provide examples of species with defective interfering genomes: poliovirus, vesicular stomatitis virus, rabies virus, and Bunyamweravirus. In such species, cheating can favour the evolution of multipartitism when as few as half of all cells are coinfected, even when there is no benefit to being multipartite (e/d ≤ 1; highlighted by the dashed red line). The bottom 2 panels provide examples of cheats derived from point mutations or small deletions: Phage Φ6 and Phage MS. In such species, our model predicts that the evolution of multipartitism requires both higher rates of coinfection to evolve and some group benefit to multipartitism. This figure can be generated using the data and code at https://doi.org/10.17605/OSF.IO/PBE4N.

Fig 4. Cheating drives the evolution of multipartite viruses with more than 2 genome segments.

This figure plots the cumulative fraction of simulations that led to different numbers of multipartite genome segments, for a viral genome containing 8 genes. Splits to higher numbers of genome segments (indicated by darker orange) were more likely when a larger average number of viruses infected each host cell (a higher multiplicity of infection or MOI; λ in our model). Each vertical bar represents 500 simulation runs over 10,000 generations. This figure can be generated using the data and code at https://doi.org/10.17605/OSF.IO/PBE4N.

Fig 5. Multipartite viruses were more resistant to exploitation by full cheats.

Here, we plot the relative abundance of “full cheats” that encode no genes whatsoever, across simulation runs that resulted in multipartite viral genomes with different numbers of genome segments. We found that when multipartitism evolved, the viral population was subsequently less exploited by full cheats. This effect was stronger for multipartite viral populations with higher numbers of genome segments and for multipartite viruses with more uneven distribution of genes across their genome segments (lighter orange shading). Each point represents an individual simulation run for a viral genome containing 8 genes. This figure can be generated using the data and code at https://doi.org/10.17605/OSF.IO/PBE4N.

Fig 6. Cheating can favour the evolution of segmented viral genomes.

We extended the simulation to allow multiple viral genomes to be packaged inside the same virion. Genome fragmentation evolved at lower multiplicities of infection when multiple genomes were packaged inside the same virion (reflecting the evolution of segmented viruses), than when genomes were packaged inside separate virions (reflecting the evolution of multipartite viruses). Each point represents the fraction of 100 simulation runs in which the full-length cooperator was driven extinct and replaced with complementing sets of viral cheats (genome fragmentation). This figure can be generated using the data and code at https://doi.org/10.17605/OSF.IO/PBE4N.

Fig 7. Cheating is associated with multipartitism across the virosphere.

The fraction of genera known to contain defective interfering genomes is plotted against the fraction of genera known to be multipartite for 4 viral Realms, where each Realm represents a likely independent origin of viruses (Methods). This figure can be generated using the data and code at https://doi.org/10.17605/OSF.IO/PBE4N.