Noise-induced bistability in the quasineutral coexistence of viral RNA under different replication modes

Abstract

Evolutionary and dynamical investigations on real viral populations indicate that RNA replication can range between two extremes given by so-called stamping machine replication (SMR) and geometric replication (GR). The impact of asymmetries in replication for single-stranded, (+) sense RNA viruses has been up to now studied with deterministic models. However, viral replication should be better described by including stochasticity, since the cell infection process is typically initiated with a very small number of RNA macromolecules, and thus largely influenced by intrinsic noise. Under appropriate conditions, deterministic theoretical descriptions of viral RNA replication predict a quasineutral coexistence scenario, with a line of fixed points involving different strands' equilibrium ratios depending on the initial conditions. Recent research on the quasineutral coexistence in two competing populations reveals that stochastic fluctuations fundamentally alters the mean-field scenario, and one of the two species outcompetes the other one. In this manuscript we study this phenomenon for RNA viral replication modes by means of stochastic simulations and a diffusion approximation. Our results reveal that noise has a strong impact on the amplification of viral RNA, also causing the emergence of noise-induced bistability. We provide analytical criteria for the dominance of (+) sense strands depending on the initial populations on the line of equilibria, which are in agreement with direct stochastic simulation results. The biological implications of this noise-driven mechanism are discussed within the framework of the evolutionary dynamics of RNA viruses with different modes of replication.

Stability of a Stochastically Perturbed Model of Intracellular Single-Stranded Rna Virus Replication

“…That is, if here exists q > 0 such that (17) − q + 2p 11 γ 1 < 0, −1 + 2p 22 γ 2 < 0 hold, then Lv(y(t)) is negative definite, whereas v(y) is positive definite, and, therefore, the zero solution of equation ( 13) is asymptotically mean square stable. Substituting (15) into (17), we obtain…”

“…Consequently, the outcome of infection is highly likely to be affected by the variability in the initial molecular interactions between the virus and host cell resources. The impact of this inherent stochasticity in ssRNA virus replication has not been properly addressed yet from a theoretical perspective, with the recent paper by Sardanyés et al [17] being a first attempt in this direction. This study confirmed the important role of noise in the amplification of ssRNAs and found that it may even induced bistability in the system, a theoretical prediction that has not yet been observed in real viral systems.…”