Table 1.
Sequences of qRT-PCR primers for human IRF3 dependent genes.
Figure 1.
Time series of mRNA levels of TNFAIP3, IκBα, RIG-I and IFNβ following stimulation by 4 µg dsRNA in hAECs cells for 0, 0.5, 1, 2, 4, and 6 hr. Gene expression estimated using Q-RT-PCR was as described previously [10]. Experimental measurements, black circles with empirical 95% confidence intervals based on triplicate measurements; means of 100 simulated single-cell trajectories, blue lines; 95% confidence bands based on simulations, red lines. Two types of simulations presented (A) under extrinsic noise, (B) under extrinsic and intrinsic noise. Horizontal axis - time (hr); vertical axis - number of molecules. Absolute values of experimental measurements scaled to simulation data (see the text for details).
Figure 2.
Time series of phosphorylated proteins, following stimulation by 4 µg dsRNA in hAECs, obtained using the Selected Reaction Monitoring (SRM) assay: In the cytoplasm, RIG-I, MAVS, IKK1, IKK2, RelA, IκBα, and IRF3; and in the nucleus, RelA, IκBα, and IRF3. Experimental measurements, black circles with empirical 95% confidence intervals based on triplicate measurements; means of 100 simulated single-cell trajectories, blue lines; 95% confidence bands based on simulations, red lines. Two types of simulations presented (A) under extrinsic noise, (B) under extrinsic and intrinsic noise. Horizontal axis: time (hr); vertical axis: number of molecules. Absolute values of experimental measurements scaled to simulation data (see the text for details).
Figure 3.
Gene knockdowns using siRNA specific to target genes in hAECs. The experiments were carried out using target gene-specific siRNA and the control nonspecific siRNA, which were reverse-transfected into hAECs at the concentration of 100 nM. The data presented are the corresponding mRNA levels at 6 hr after electroporation (fold change, in logarithmic scale). (A), knockdown of RelA, IRF3, RIG-I and IKKγ. (B), knockdown of TNFAIP3/A20, NFKBIA/IκBα, ISG56 and IFNβ. The mRNA levels of the indicated genes (at top) were determined by RT-PCR. Results were represented as normalized fold change of expression compared to control non dsRNA-induced cells transfected with scrambled (non-target control) siRNA. Gray bars, experiment (dark, no dsRNA stimulation; light, 4 µg dsRNA); red lines, model simulation. 95% confidence intervals of experimental data (based on 3 replicates) are distorted by the logarithmic scale. In linear scale, the relative error rate is approximately 10%. Please notice that in IFNβ and ISG56 charts, the model values at time 0 are equal to 0, which is impossible to depict on the log scale. (C), Effect of RIG-I knockdown on NF-κB and IRF3 dependent gene expression. hAECs were transfected with siRNA to RIG-I or control siRNA (Con). Left panel, effect of RIG-I siRNA on RIG-I expression. Note that dsRNA induced RIG-I expression is largely inhibited in hAECs transfected with RIG-I siRNA. Middle panel, effect of RIG-I knockdown on NF-κB-dependent TNFAIP3/A20 gene expression. dsRNA induces TNFAIP3/A20 expression in RIG-I knockdown cells. Right panel, effect of RIG-I knockdown on IRF3-dependent gene expression. RIG-I knockdown significantly blunts IRF3-dependent ISG56 expression. We conclude from these data that RIG-I is primarily coupled to IRF3 signaling in hAECs. (D), Effect on RIG-I expression in murine MEF cells with both IRF3 and IRF7 genes knocked down. RIG-I is down-regulated, which suggests IRF7 may play a key role in RIG-I up-regulation. Compare with RIG-I results in Panel A where the IRF3 knock-down does not seem to down-regulate RIG-I expression.
Figure 4.
RIG-I synthesis is IRF7 dependent. (A) IRF3 siRNA knockdown of A549 cells. A549 cells were transfected with scrambled control (Con) or IRF3-specific siRNAs and stimulated in the absence or presence of poly(I:C). mRNA measured by Q-RT-PCR and the results are represented as the normalized fold change expression compared to control cells transfected with scrambled (non-target control) siRNA. Shown are normalized mRNA expression for IRF3, RIG-I and ISG56 mRNA on a linear scale. Note that poly(I:C) induced ISG56 expression is reduced by 86%, whereas RIG-I expression is reduced by <10%. (B) Effect of IRF3/IRF7 deficiency on RIG-I expression. WT or IRF3/IRF7-/- MEFs were transfected with poly(I:C) and expression of RIG-I determined by Q-RT-PCR. Shown is a time course of wild type or IRF3/7-/- double knockout cells in response to dsRNA. Note that RIG-I expression is completely blocked in IRF3/7-/- cells. Together these data indicate that RIG-I expression is largely independent of IRF3, but requires IRF7.
Figure 5.
Observed and simulated dsRNA dose-dependence curves in hAECs.
Experimental measurements, black circles with empirical 95% confidence intervals based on triplicate measurements; means of 100 simulated single-cell trajectories, blue lines; 95% confidence bands based on simulations, red lines. Two types of simulations presented (A) under extrinsic noise, (B) under extrinsic and intrinsic noise. Horizontal axis: dsRNA dose (µg); vertical axis: number of molecules.
Figure 6.
Single cell nucleus/cytoplasm ratios under 5 µg of dsRNA stimulation (RelA, green channel fluorescence).
EGFP-RelA stable hAECs were electroporated with different dosages of synthetic dsRNA analog Poly IC and dynamic live cell imaging was performed. Time presented in hr. Green trend lines are third-order polynomials, fitted using least-squares minimization. Upper row: Raw time series. Middle row: Detrended time series. Bottom row: Fourier periodograms. Columns 1–4: Observed single cells. Column 5: Two simulated cells.
Figure 7.
Single cell nucleus/cytoplasm ratios under 50 µg of dsRNA stimulation (RelA, green channel).
EGFP RelA stable hAECs were electroporated with different dosages of synthetic dsRNA analog Poly IC and dynamic live cell imaging was performed. Time is in hr. Green trend lines are third-order polynomials, fitted using least-squares minimization. Upper row: Raw time series. Middle row: Detrended time series. Bottom row: Fourier periodograms. Columns 1–5: Observed single cells.
Figure 8.
Single cell nucleus/cytoplasm ratios under 5 µg of dsRNA stimulation (IRF3, red channel fluorescence).
Strawberry -IRF3 hAECs were electroporated with different dosages of synthetic dsRNA analog Poly IC and dynamic live cell imaging was performed. Time presented in hr. Green trend lines are third-order polynomials, fitted using least-squares minimization. Upper row: Raw time series. Middle row: Detrended time series. Bottom row: Fourier periodograms. Columns 1–4: Observed single cells. Column 5: Two simulated cells.
Figure 9.
Single cell nucleus/cytoplasm ratios under 50 µg of dsRNA stimulation (IRF3, red channel).
Strawberry -IRF3 hAECs were electroporated with different dosages of synthetic dsRNA analog Poly IC and dynamic live cell imaging was performed. Time is in hr. Green trend lines are third-order polynomials, fitted using least-squares minimization. Upper row: Raw time series. Middle row: Detrended time series. Bottom row: Fourier periodograms. Columns 1–5: Observed single cells.
Figure 10.
Simplified schematic of the IRF3-NF-κB model. Only dsRNA, proteins (in the cytoplasm) and genes (in the nucleus) are shown. Solid green lines on the top denote direct chemical binding. Green dotted lines denote activation. Vertical thick colored arrows denote translocation of activated transcription factors into the nucleus. Red dotted lines denote inhibition. Horizontal solid black arrows in the nucleus denote gene transcription, with plus or minus signs denoting activation or repression, respectively. Transcripts and inactive forms of the proteins are omitted for simplicity.
Table 2.
Base constant rates used (BCR) to fit the model to experiments.
Table 3.
Modeled effects of theTNFAIP3 knockdown at 6