Virulence evolution of a generalist plant virus in a heterogeneous host system

Abstract

Modelling virulence evolution of multihost parasites in heterogeneous host systems requires knowledge of the parasite biology over its various hosts. We modelled the evolution of virulence of a generalist plant virus, Cucumber mosaic virus (CMV) over two hosts, in which CMV genotypes differ for within-host multiplication and virulence. According to knowledge on CMV biology over different hosts, the model allows for inoculum flows between hosts and for host co-infection by competing virus genotypes, competition affecting transmission rates to new hosts. Parameters of within-host multiplication, within-host competition, virulence and transmission were determined experimentally for different CMV genotypes in each host. Emergence of highly virulent genotypes was predicted to occur as mixed infections, favoured by high vector densities. For most simulated conditions, evolution to high virulence in the more competent Host 1 was little dependent on inoculum flow from Host 2, while in Host 2, it depended on transmission from Host 1. Virulence evolution bifurcated in each host at low, but not at high, vector densities. There was no evidence of between-host trade-offs in CMV life-history traits, at odds with most theoretical assumptions. Predictions agreed with field observations and are relevant for designing control strategies for multihost plant viruses.

Keywords: Cucumber mosaic virus; multihost parasites; virulence evolution; virus emergence.

Figure 1

Predictions of co-infection models for one host (A) or for two hosts (B) for virulence evolution of CMV in Host 1 (tomato) and Host 2 (melon), as a function of the number of aphids per plant (i) and of the rate of transmission events (β_e). Graphs indicate areas in which there is no infection (S ) or where Y isolates (Y ), A isolates (A ), N isolates (N ) or A+N isolates (M ) are the most prevalent in the virus populations. Figure 1A for tomato was redrawn from the study by Escriu et al. (2003). For a series of i and β_e values, indicated by asterisks in the figure, the relative equilibrium frequency of the different virus genotypes in, and the average virulence of, the virus population is indicated.

Figure 2

Equilibrium density of susceptible noninfected plants, S, and of plants infected by CMV isolates Y, A, N or mixed infected by CMV isolates A+N (M) according to a co-infection model for two hosts, when within-host and between-host rates of transmission events (β_e) differ. Presented results are for within-host β_e = 0.1 and between-host β_e = 0.0001 (A) or when within-host β_e = 0.0001 and between-host β_e = 0.1 (B). Number of aphids per plant, i, varied from 1 to 10. Initial conditions were as follows: S₁ = 3.97, Y₁ = A₁ = N₁ = 0.01, M₁ = 0 and S₂ = 0.8, Y₂ = A₂ = N₂ = M₂ = 0 plants/m². Bars represent plant density for different i values for noninfected plants (S₁; S₂ ) or for plants infected by CMV isolates Y₁, Y₂ ( ), A₁, A₂ ( ), N₁, N₂ ( ) or mixed infected with A+N isolates (M₁, M₂ ).

Figure 3

Equilibrium density of susceptible noninfected plants, S, and of plants infected by CMV isolates Y, A, N or mixed infected by CMV isolates A+N (M) according to a co-infection model for two hosts when within-host and between-host rates of transmission events (β_e) differ. Presented results are for within-host β_e = 0.05 and transmission events from Host 1 to Host 2 β_e(H1–H2) = 0 and from Host 2 to Host 1 β_e(H2–H1) = 0.01 (A) or when within-host β_e = 0.05 and transmission events from Host 1 to Host 2 β_e(H1–H2) = 0.01 and from Host 2 to Host 1 β_e(H2–H1) = 0 (B). Number of aphids per plant, i, varied from 1 to 10. Initial conditions were as follows: S₁ = 3.97, Y₁ = A₁ = N₁ = 0.01, M₁ = 0 and S₂ = 0.8, Y₂ = A₂ = N₂ = M₂ = 0 plants/m². Bars represent plant density for different i values for noninfected plants (S₁; S₂ ) or for plants infected by CMV isolates Y₁, Y₂ ( ), A₁, A₂ ( ), N₁, N₂ ( ) or mixed infected with A+N isolates (M₁, M₂ ).

Figure 4

Dynamics of the populations of susceptible noninfected plants, S, and of plants infected by CMV isolates Y, A, N or mixed infected by CMV isolates A+N (M) according to a co-infection model for two hosts that rotate in time. Presented results are for a temporal overlap of hosts equivalent to half their life cycle (i.e. for 50 days) and for within-host β_e = 0.1 and between-host β_e = 0.0001 (A) or within-host β_e = 0.0001 and between-host β_e = 0.1 (B), for i = 0.5 and i = 2. Initial conditions were as follows: S₁ = 3.97, Y₁ = A₁ = N₁ = 0.01, M₁ = 0 and S₂ = 0.8, Y₂ = A₂ = N₂ = M₂ = 0 plants/m². Curves represent the variation in time of the density of noninfected plants (S ) or plants infected by Y isolates (Y ), by A isolates (A ), by N isolates (N ) or by A+N isolates (M ). The shadow indicates the overlapping of the life cycles of Host 1 and Host 2, and the arrow indicates the host that initiates the rotation.

Figure 5

Dynamics of the populations of susceptible noninfected plants, S, and of plants infected by CMV isolates Y, A, N or mixed infected by CMV isolates A+N (M) according to a co-infection model for two hosts that rotate in time. Presented results are for a temporal overlap of hosts equivalent to half their life cycle (i.e. for 50 days) and for within-host β_e = 0.1 and between-host β_e = 0.0001 (A) or within-host β_e = 0.0001 and between-host β_e = 0.1 (B), for i = 0.5 and i = 2. Initial conditions were as follows: S₁ = 4, Y₁ = A₁ = N₁ = M₁ = 0 and S₂ = 0.77, Y₂ = A₂ = N₂ = 0.01, M₂ = 0 plants/m². Curves represent the variation in time of the density of noninfected plants (S ) or plants infected by Y isolates (Y ), by A isolates (A ), by N isolates (N ) or by A+N isolates (M ). The shadow indicates the overlapping of the life cycles of Host 2 and Host 1, and the arrow indicates the host that initiates the rotation.