We are using experimental evolution of RNA viruses to understand the basis of viral population structure. In earlier work we demonstrated that different viruses have different mutation frequencies in the same host (Schneider and Roossinck, 2000), and that the same virus has different mutation frequencies in different hosts (Schneider and Roossinck, 2001). We analyzed artificial populations of CMV in tobacco and found that significant bottlenecks occurred during systemic infection (Li and Roossinck, 2004), which could serve to limit mutation frequencies in a host specific manner. We also demonstrated that significant bottlenecks occur during horizontal transmission of CMV by two aphid species (Ali et al., 2006).

Genetic bottlenecks in different hosts: We are using our artificial CMV populations to compare bottlenecks during systemic infection of different hosts. For these studies, young leaves of test plants are inoculated with CMV. The first leaf to show systemic symptoms is designated as the primary systemic leaf, and the smallest leaf visible at that time is designated as the secondary systemic leaf. Preliminary results indicate that bottlenecks are different for different hosts, and during primary versus secondary systemic infection. This is likely due to different plasmodesmatal structures in different plants.

Replicase fidelity: We use the CMV satRNA as a convenient reporter for replicase fidelity studies. The satRNA uses the viral replicase, but is dispensable. We constructed a deletion mutant of the satRNA that is lacking the (+) strand promoter. Inoculation of this reporter onto CMV-infected plants generates only (-) strand progeny, which we can recover and analyze. With this system we demonstrated that the indel rate of the CMV replicase is significantly different in two experimental hosts, tobacco and pepper, and that the rate varies with the secondary structure of the satRNA (Pita et al., 2007). Currently we are using this system to analyze the substitution rate of the CMV replicase.

Point mutations in CMV RNA 3 created restriction enzyme markers that could be easily tracked. When a bottleneck a occurs the number of mutants recovered is reduced in a stochastic manner.

Viral determinants of genetic variation: We are using two strains of CMV, LS and Fny, to map the viral determinants of genetic variation. These two viruses differ in their mutation frequencies in some hosts, but are very similar in others. We have constructed reassorted viruses and are analyzing their mutation frequencies in several different hosts. This study will likely provide insights into interactions in the two replicase proteins that are involved in generating variation. Once determinants are mapped to a specific viral RNA we will use recombinant viruses, and ultimately site-directed mutants, to fine-map viral determinants of genetic variation.

In nature viruses are often found in mixed infections. We have used two related viruses, CMV and Cowpea chlorotic mottle virus, in individual and mixed infections to see how viruses might affect each other. In most cases the viruses show a trend toward lower mutation frequencies in mixed infections than in individual infections. This is not what would be expected if the viruses were competing for essential replication materials such as nucleotide pools. We are exploring the basis for these unexpected results.

Recombination rates in virus replication: Recombination has been studied extensively in the bromoviruses, but not in CMV. We are establishing a system to measure recombination rates in CMV by introducing point mutations that create unique restriction enzyme sites in different regions of RNA 3. By infecting plants with mixtures of RNAs 3 with different unique sites we will monitor recombination rates by determining how often we recover progeny that have either lost or gained restriction sites. This will give us a more complete picture of the myriad factors that affect evolution in our model virus system.