The Role of Tat Positive Feedback the Establishment of and Reactivation from HIV Latency

Proviral latency is the main persistence mechanism that precludes eradication of human immunodeficiency virus, HIV, from infected patients. Although latency is a viral phenotype, current theories posit that HIV latency is not 'hardwired' into viral circuitry but directly under cell-state control. Therefore, latency is thought to be a deterministic epiphenomenon with no role in the natural history of the virus. Here, we synthetically reengineer HIV regulatory circuits to define the role of viral gene circuitry and cellular state in regulating latency. The reengineered circuits demonstrate that latency is largely autonomous to cellular state in both minimal circuits and full-length replicating viruses. Strikingly, in primary cells, cell-driven silencing of viral transcription--the prevailing hypothesis for latency establishment--is overcome by tuning viral feedback strength. The reengineered minimal circuits also show that, through a combination of mathematical modeling and noise measurements, HIV transcription occurs through episodic bursts generating large stochastic fluctuations in HIV gene-expression. It is unclear if these stochastic fluctuations influence HIV Tat positive feedback, the decision-making circuit encoded by the virus. Surprisingly, upon stimulation of HIV transcription, Tat positive feedback immediately saturates, buffering against stochastic fluctuations once a fate decision has been made. This feedback saturation leads to three striking properties; (i) transcriptional kinetics and ( ii) steady-state output from the LTR are insensitive to variable Tat inputs, and (iii) the LTR converts unimodal, graded Tat inputs, into bimodal expression patterns. Thus, stochastic fluctuations in Tat levels will have the most profound affect early in viral infection because, upon commitment to active replication, HIV circuitry displays robustness to noise. Overall these results argue that HIV latency is an intrinsic, stochastic feature of the virus that most likely occurs early in the viral lifecycle. Since HIV circuitry can act autonomous to cell-state, latency seems to be `hardwired' into viral circuitry and not simply an epiphenomenon stemming from host-cell factors. Given the rapid mutation rate of HIV, selection for and conservation of the latency phenotype suggests it has a fitness role in the natural history of the virus.