Defective interfering (DI) virus-like particles arise during virus culture and may play a role in the transmission, pathogenesis and persistence of viruses in natural hosts. They are defective because they are not able to replicate without complementation by a co-infecting wild-type(fully infectious) virus. DI particles interfere with the replication of wild-type virus by acting as a parasite to the wild-type infection, competing for viral RNA-dependent RNA polymerase and biosynthetic resources of the cell. The propensity to generate and propagate DI particles is a common feature of many viruses, including bacteriophage, influenza virus, poliovirus, and vesicular stomatitis virus(VSV). We focus on VSV, a member of the rabies virus class, because it is minimally pathogenic and it has potential applications as a live-viral vaccine against AIDS and respiratory syncytial virus. VSV is also being developed as a self-perpetuating oncolytic (tumor-destroying) agent. It has long been observed that high multiplicities of infection (MOI) promote the emergence and outgrowth of DI particles during virus culture. However, the effects of MOI on DI particle production have not been elucidated.

We cultured VSV on baby hamster kidney(BHK 21) cells and studied how different fixed MOIs influence the dynamics of wild-type and DI particles of VSV. DI particles themselves can be difficult to measure with a direct biological assay as they require the presence of a helper virus to replicate. We thus used an indirect assay, yield reduction, to estimate the concentration of DI particles in a given sample. We further developed a physical assay to detect the different particle types. Virus and virus-like particles from culture were separated by density-gradient centrifugation and measures of particle concentrations, infectivity, and interference were determined for sample fractions by optical density measurements and by employing plaque and yield-reduction assays, respectively. As anticipated, we found that the emergence and accumulation of DI particles coincided with a loss of infectious virus particles in the population. Moreover, results from serial-passage culture of virus under controlled MOIs were used to parameterize mathematical models, providing an estimate for DI particle productivity (average yield of DI particles per co-infected cell) as well as the intrinsic yields of DI particle production from cells infected with only fully infectious virus. These results have implications for developing quantitative and predictive models of virus growth for clinical, therapeutic and technological applications.