Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Dec 15;5:17661.
doi: 10.1038/srep17661.

Distinct Promoter Activation Mechanisms Modulate Noise-Driven HIV Gene Expression

Affiliations
Free PMC article

Distinct Promoter Activation Mechanisms Modulate Noise-Driven HIV Gene Expression

Arvind K Chavali et al. Sci Rep. .
Free PMC article

Abstract

Latent human immunodeficiency virus (HIV) infections occur when the virus occupies a transcriptionally silent but reversible state, presenting a major obstacle to cure. There is experimental evidence that random fluctuations in gene expression, when coupled to the strong positive feedback encoded by the HIV genetic circuit, act as a 'molecular switch' controlling cell fate, i.e., viral replication versus latency. Here, we implemented a stochastic computational modeling approach to explore how different promoter activation mechanisms in the presence of positive feedback would affect noise-driven activation from latency. We modeled the HIV promoter as existing in one, two, or three states that are representative of increasingly complex mechanisms of promoter repression underlying latency. We demonstrate that two-state and three-state models are associated with greater variability in noisy activation behaviors, and we find that Fano factor (defined as variance over mean) proves to be a useful noise metric to compare variability across model structures and parameter values. Finally, we show how three-state promoter models can be used to qualitatively describe complex reactivation phenotypes in response to therapeutic perturbations that we observe experimentally. Ultimately, our analysis suggests that multi-state models more accurately reflect observed heterogeneous reactivation and may be better suited to evaluate how noise affects viral clearance.

Figures

Figure 1
Figure 1. Varying basal transcription rate and strength of positive feedback affects protein heterogeneity in a one-state promoter model.
(A) Conceptual schematic based on biological understanding of an ‘ideal’ promoter configuration for HIV reactivation as a fully active provirus with all transcriptional machinery available for sustained viral mRNA synthesis. (B) Schematic depiction of a one-state computational model with positive feedback. Reactions involving transcription, translation, degradation, and positive feedback are depicted. (C) Violin plots of the steady-state endpoint protein distributions for a range of basal transcription rates and positive feedback strengths. Violin plots capturing endpoint distributions for a system without feedback are shown on the far right (white shading; ‘No fb’). The stochastic model for every parameter set was run 1000 times for a period of 10 days. (D) Fano factors for the one-state model with positive feedback are shown for different basal transcription rates of 0.01 (black), 0.1 (red), 0.5 (blue), 1 (magenta), and 10 (green) across varying strengths of positive feedback. The x-axis is presented in log scale. (E) Mean onset times for the one-state model with positive feedback are shown for different basal transcription rates of 0.01 (black), 0.1 (red), 0.5 (blue), 1 (magenta), and 10 (green) across varying strengths of positive feedback. If a cell did not activate during the simulation, onset time was set to 10 days.
Figure 2
Figure 2. Conceptual schematic and transcriptional bursting behaviors in a two-state promoter model with positive feedback.
(A) Conceptual diagram of the underlying biological mechanisms that define inactive and active LTR states. (B) Schematic depiction of a two-state computational model with positive feedback. (C) Heat map of the deterministic steady-state Tat protein levels for the two-state model with positive feedback simulated over a range of burst sizes and burst frequencies (presented in log scale). Burst size is computed as the Tat-independent transcription rate divided by the promoter inactivation rate (αb/ki). Burst frequency is computed as the promoter activation rate divided by the transcript degradation rate (kbm). The color map indicates the deterministic steady state value of Tat (in log scale). The black dotted line at a burst size of 0.03 is equivalent to the Tat-mediated transcription rate used in the active one-state model with positive feedback. Three distinct regions are depicted on the color map (as I, II, and III), and these correspond to regions of unproductive, switching, or productive behavior, respectively. (D) Violin plots of the steady-state protein distributions for the corresponding parameter sets of burst size and burst frequency in (C). The green dotted line indicates the threshold for activation (Supplementary Figure S2). The black dotted line is the same as in (C). Numbers in red indicate cell activation percentages.
Figure 3
Figure 3. Assessing system heterogeneity under varying transcriptional burst sizes and frequencies for the two-state promoter model.
(A–E) A two-state model with positive feedback was simulated for a range of burst sizes and burst frequencies and the following metrics were calculated based on the final Tat protein values: (A) Mean protein counts; (B) cell activation; (C) mean first passage time (days); (D) coefficient of variation squared (presented in log scale); and (E) Fano factor. Color bars in each panel indicate the range of values for each metric. (F,G,I,J) Violin plots capturing endpoint protein distributions across paths charted in (E). (H,K) Plots of Fano factor across increasing burst size or frequency correspond to paths charted in (E).
Figure 4
Figure 4. Conceptual schematic and transcriptional bursting behaviors in a three-state promoter model with positive feedback.
(A) Conceptual diagram of possible biological mechanisms underlying three promoter states (two inactive and one active). (B) Schematic depiction of a three-state computational model with positive feedback. (C) Three-state models with positive feedback were fixed at different values of kON and kOFF and then simulated over a range of burst sizes and burst frequencies. The following metrics were calculated based on the final Tat protein values: mean protein counts (in log scale), cell activation, Fano factor (in log scale), and mean first passage time (days). Color bars in each panel indicate the range of values for each metric.
Figure 5
Figure 5. Distinctive activation and noise profiles at low and high burst frequency for two- and three-state promoter models with positive feedback.
(A–C) Two- and three-state models were fixed at a low value of burst frequency (0.3) and then simulations were run for a range of burst sizes. Final Tat protein values were used to calculate (A) cell activation, (B) Fano factor (in log scale), and (C) mean protein counts (in log scale). (D–F) Same as in (A–C) but burst frequency was set to a high value (1.5). The three-state model was simulated with kON = kOFF = 0.01 day−1 (‘3 prom_0.01/0.01’), kON = kOFF = 1 day−1 (‘3 prom_1/1’), and kON = kOFF = 10 day−1 (‘3 prom_10/10’). Two-state model is labeled as ‘2 prom’.
Figure 6
Figure 6. Comparing sample experimental and simulated latent HIV infections under basal and stimulated conditions.
(A,B) Conceptual diagrams illustrating the action of Aza and TNF on (A) three-state and (B) two-state models. (C,D) Flow cytometry histograms of GFP expression for two HIV latency cell line models that are (C) repressed or (D) permissive for activation by TNF. Histograms are shown under basal conditions and following stimulation with 10 ng/ml TNF and/or 5 μM Aza. (E) Three-state model simulation of the experimental data presented in (C). TNF is assumed to increase burst size and burst frequency and Aza is assumed to increase the transition between inactive states (kON). (F) Two-state model simulation of the experimental data presented in (D).

Similar articles

See all similar articles

Cited by 5 articles

References

    1. Losick R. & Desplan C. Stochasticity and cell fate. Science. 320, 65–68 (2008). - PMC - PubMed
    1. Raser J. M. & O’Shea E. K. Noise in gene expression: origins, consequences, and control. Science. 309, 2010–2013 (2005). - PMC - PubMed
    1. Miller-Jensen K., Dey S. S., Schaffer D. V. & Arkin A. P. Varying virulence: epigenetic control of expression noise and disease processes. Trends Biotechnol. 29, 517–525 (2011). - PubMed
    1. Raj A. & van Oudenaarden A. Nature, nurture, or chance: stochastic gene expression and its consequences. Cell. 135, 216–226 (2008). - PMC - PubMed
    1. Swain P. S., Elowitz M. B. & Siggia E. D. Intrinsic and extrinsic contributions to stochasticity in gene expression. Proc. Natl. Acad. Sci. USA 99, 12795–12800 (2002). - PMC - PubMed

Publication types

Substances

Feedback