Mathematical modeling of bacteria-virus interactions in Lake Michigan incorporating phosphorus content

Abstract Viruses are ubiquitous in aquatic ecosystems and play a critical role in nutrient cycling because viral lysis of cells releases phosphorus, an essential and often limiting nutrient. Previous models have examined dynamics of bacterial and viral communities, but with limited analysis and without explicit consideration of nutrients. A recent model ( Fuhrman et al., 2011 ; Math. Comp. Model. 53, 716-730) incorporated internal nutrient content of bacteria and viruses. In the present study, we modified and tested the model with data from natural planktonic communities from Lake Michigan. Replicate 20 L water samples (135 μm screened), were either untreated or enriched with 8 μM phosphate, and bacterial and viral abundance, chl a fluorescence, and phosphorus (total and dissolved inorganic) were monitored for two weeks. Fuhrman et al.'s model (modified to include phytoplankton) was applied to the data, fitting burst size, lytic latent period and virus decay rate. For enriched samples, model fits were good and parameters were consistent with measurements in other freshwater ecosystems. However, for unenriched samples, where nutrient concentrations approached detection limits, model fits were relatively poorer. The model predicted similar viral decay rates but higher burst sizes and longer latent period in phosphorus-limited versus enriched conditions, underlining the potential importance of nutrients in host–virus interactions. The model is likely to be most useful in meso- to eutrophic systems; requirements for future model development and parameter estimation for application to oligotrophic lakes are discussed.