In this proposal we seek to perform a rigorous experimental and computational analysis of the dynamics of gene expression in a lentiviral model of HIV-1. The lentiviral system is derived from HIV-1 but leaves the Tat-mediated positive feedback loop controlling transcriptional activity intact. This feedback loop motif is a common gene regulatory architecture across all organisms and cellular subsystems thus measurements and conclusions derived from this study will apply broadly. Based on physical chemical principles and simple models of the system, we hypothesize (and show initial experimental results) that gene expression from this system is highly stochastic and that the Tat-mediated feedback is activated sporadically such that after the virus infects its target cell and integrates into the host genome, there may be significant periods of time elapsed before viral expression reaches the point where reproduction can occur and then propagate. This time could be enough to allow an activated T-cell to transition to its memory state thereby trapping the lentivirus in an inactive form until such time as that memory cell is reactivated. Thus, it is possible that this hypothesized noise is large enough to contribute to the formation of the latent pool of virus that makes HIV-1 so hard to treat. However, there is little experimental evidence of this possibility and further it has never been rigorously shown that mammalian gene expression is significantly stochastic. We therefore set out a program for quantitatively measuring the integration-point dependent kinetics of the key steps in this lentiviral, HIV-1 based, autoactivated gene expression system. We use specially designed viral constructs with fluorescent markers of expression and quantitative microscopy techniques to tease out all the parameters for a model of the eukaryotic gene expression model. We use the models to estimate the role of each of the steps in the process from integration point, to transcription initiation, to elongation control by tar, to Tat production/ degradation, to Tat/tar interaction in generating noise in the expression process and allowing delays until viral production. We also use the comparison of models to data to determine which parameters are not constrained by optimizations and thus direct measurements. The resultant experimentally-validated model of mammalian viral gene expression will be a resource for others studying the kinetics and role of noise in eukaryotic cellular processes.
Showing the most recent 10 out of 18 publications
