HIV promoter integration site primarily modulates transcriptional burst size rather than frequency

Abstract

Mammalian gene expression patterns, and their variability across populations of cells, are regulated by factors specific to each gene in concert with its surrounding cellular and genomic environment. Lentiviruses such as HIV integrate their genomes into semi-random genomic locations in the cells they infect, and the resulting viral gene expression provides a natural system to dissect the contributions of genomic environment to transcriptional regulation. Previously, we showed that expression heterogeneity and its modulation by specific host factors at HIV integration sites are key determinants of infected-cell fate and a possible source of latent infections. Here, we assess the integration context dependence of expression heterogeneity from diverse single integrations of a HIV-promoter/GFP-reporter cassette in Jurkat T-cells. Systematically fitting a stochastic model of gene expression to our data reveals an underlying transcriptional dynamic, by which multiple transcripts are produced during short, infrequent bursts, that quantitatively accounts for the wide, highly skewed protein expression distributions observed in each of our clonal cell populations. Interestingly, we find that the size of transcriptional bursts is the primary systematic covariate over integration sites, varying from a few to tens of transcripts across integration sites, and correlating well with mean expression. In contrast, burst frequencies are scattered about a typical value of several per cell-division time and demonstrate little correlation with the clonal means. This pattern of modulation generates consistently noisy distributions over the sampled integration positions, with large expression variability relative to the mean maintained even for the most productive integrations, and could contribute to specifying heterogeneous, integration-site-dependent viral production patterns in HIV-infected cells. Genomic environment thus emerges as a significant control parameter for gene expression variation that may contribute to structuring mammalian genomes, as well as be exploited for survival by integrating viruses.

Figure 1. HIV LTR expression distributions are wide and highly skewed.

A) Experimental system: Sample clonal histograms, spanning a range of ‘dim’ and ‘bright’ integrations, with autofluorescence included for comparison, represent fluorescence measurements on approximately 4000 cells. Smooth curves through each are the result of an optimized low-pass Fourier filter, and are used for model fitting. All measurements are given in cytometer-based relative fluorescent units (RFU). B) Trend in distribution-shape variation: Log-log linear regression coefficients quantify trends in distribution-shape variation as a power-law relationship between distribution variance (σ²) and mean (μ): ; (R² = 0.89, coefficients ±95% confidence). Dashed lines demonstrate Poisson-like scaling (‘Poisson’, α = 1) and over-all distribution scaling (‘Scaling’, α = 2). C) Characteristic distribution shape: Smoothed, autoflorescence-deconvolved histograms were shifted by a constant fluorescence to a fixed mean (specified by the median over the set of integration clones, μ₀), and fluorescence values were scaled about that mean according to the variance regression in 1B (inset), by a factor with α = 1.7 (thin curves). The grey curve averages the transformed distributions and represents a ‘typical’ HIV-LTR expression profile that is wide (coefficient of variation = σ₀/μ₀∼60%) and highly skewed.

Figure 2. Transcript production in bursts qualitatively explains HIV LTR distribution shapes and variation with integration site.

A) Model schematic. φ_a/r = active/repressed gene state, T = Transcript, P = Protein, X = Degraded, κ = Probability/unit time. Bold parameters are considered to be integration-site dependent, while others were measured separately and fixed for all clones. All rates are measured relative to the transcript degradation rate, κ_t⁻. B) Model regimes depend on the ratio of gene-state to transcriptional dynamics: ‘Fast’ (), ‘Slow’ (), ‘Intermediate’ (), ‘Bursting’ (). Transcript production rates (κ_t⁺) for sample distributions are set to reproduce the same mean number of transcript copies (left) and protein copies (right) at steady state as predicted for the ‘typical’ experimental distribution in Fig. 1C; autoflourescence is not included, and distributions are binned on a linear RFU scale for comparison. C) Distribution shape-variation in the bursting regime: burst-frequency variation (κ_a) leads to an approximate Poisson-like shape-variation, burst-size (b) yields an approximate distribution-scaling shape variation, and the combination with (combined) gives a shape-variation most closely resembling the experimental data (compare Fig. 1B). Insets are sample, log-binned distributions with varying burst size or frequency. The fixed parameter in each sample is set to the value that approximately reproduces the ‘typical’ distribution shape in Fig. 1C.

Figure 3. Transcriptional bursts are short, but only an upper bound on their duration can be resolved.

A) Sample fits for several fixed values of burst duration, τ (measured relative to the transcript degradation time): experimental distribution (solid curve with 95% confidence region in grey); optimal fit at small τ (long dash); fits for larger (short dash), demonstrating increased deviations (τ increasing along arrow in for inset). B) Relative fit deviation, Dev_r decreases for shorter active duration τ, for each clone (solid lines), and are optimal (Dev_r = Dev_r^Opt) when burst durations are shortest (i.e. in the bursting regime). Dev_r−Dev_r^Opt = 1 (dashed line) is considered a cut-off, beyond which fit quality is significantly worse than the optimum, specifying a distinguishability cut-off upper-bound on τ, marked by the intersection of dashed line and the solid lines and referred to as for each clone. C) Resolution of bursting dynamics: Calculated upper-bound active-duration () and optimal transcriptional burst size (b^Opt) for each clone. Predicted large transcriptional bursts (b^Opt≫1) identify clones for which the inferred transcriptional dynamics differ significantly from continuous transcription at a single fixed rate, and small indicates good resolution of short bursts from less ‘noisy’ dynamics.

Figure 4. Modulation of transcriptional bursts by integration site.

Best-fit transcriptional burst size (b) and burst frequency (κ_a) that minimize the relative fit deviation (Dev_r) were calculated at (which specifies a short active duration that was nearly optimal for all clones). Error bars represent the maximum 95% confidence interval for simultaneous parameter variations that increase Dev_r by 1. Log-log regression coefficients represent power-law scaling of fit parameters with distribution mean (μ, measured in cytometer RFU), of the form (x = b or κ_a), and are given with 95% confidence intervals. A) b: α = 0.76±0.14; β = 0.13±0.2; R² = 0.66. B) κ_a: α = 0.2±0.15; β = −0.5±0.2; R² = 0.2.

Figure 5. Active-fraction variation distinguishes modes of transcriptional regulation.

The best-fit restriction of τ below τ^Max in Fig. 3 specifies an upper bound on the predicted fraction of cells with the LTR in the active state (), which is marked by bars for each clone. Each Mode of integration-site modulation of transcriptional dynamics leads to a different expected variation of that distinguishes them. For Mode 1, active-state stability varies over integration clones, with the active-state transcription rate fixed ( was used for this example), while for Mode 2 the active-state transcription rate varies over integrations, with the active duration fixed ( was used for this example).

Figure 6. Basal promoter fluctuations as determinants of infected-cell fate.

Possible decomposition of the ‘space’ of basal burst-parameters inferred by the current analysis into ranges of parameter combinations that, in the presence of positive feedback from Tat, may lead to active viral replication vs. latent fates. Region I: Basal transcription pattern never leads to Tat-transactivation. Region II: Fast transactivation always leads to a stable highly expressing state. Region III: Bimodal expression patterns, where large fluctuations in basal transcriptional bursting can infrequently drive transitions from basal to transactivated states. Inset histograms demonstrate representative expression patterns for single-integration clones of a similar vector that includes Tat , , and region boundaries are hypothetical.