|Home | About | Journals | Submit | Contact Us | Français|
Respiratory syncytial virus (RSV) is a negative-sense single-stranded RNA virus responsible for lower respiratory tract infections (LRTIs) in humans. In experimental models of RSV LRTI, the actions of the nuclear factor κB (NF-κB) transcription factor mediate inflammation and pathology. We have shown that RSV replication induces a mitogen-and-stress-related kinase 1 (MSK-1) pathway that activates NF-κB RelA transcriptional activity by a process involving serine phosphorylation at serine (Ser) residue 276. In this study, we examined the mechanism by which phospho-Ser276 RelA mediates expression of the NF-κB-dependent gene network. RelA-deficient mouse embryonic fibroblasts (MEFs) complemented with the RelA Ser276Ala mutant are deficient in CXCL2/Groβ, KC, and interleukin-6 (IL-6) expression, but NFKBIA/IκBα is preserved. We show that RSV-induced RelA Ser276 phosphorylation is required for acetylation at Lys310, an event required for transcriptional activity and stable association of RelA with the activated positive transcriptional elongation factor (PTEF-b) complex proteins, bromodomain 4 (Brd4), and cyclin-dependent kinase 9 (CDK9). In contrast to gene loading pattern of PTEF-b proteins produced by tumor necrosis factor (TNF) stimulation, RSV induces their initial clearance followed by partial reaccumulation coincident with RelA recruitment. The RSV-induced binding patterns of the CDK9 substrate, phospho-Ser2 RNA polymerase (Pol) II, follows a similar pattern of clearance and downstream gene reaccumulation. The functional role of CDK9 was examined using CDK9 small interfering RNA (siRNA) and CDK inhibitors, where RSV-induced NF-κB-dependent gene expression was significantly inhibited. Finally, although RSV induces a transition from short transcripts to fully spliced mRNA in wild-type RelA (RelA WT)-expressing cells, this transition is not seen in cells expressing RelA Ser276Ala. We conclude that RelA Ser276 phosphorylation mediates RelA acetylation, Brd4/CDK9 association, and activation of downstream inflammatory genes by transcriptional elongation in RSV infection.
Respiratory syncytial virus (RSV) is a negative-sense, single-stranded RNA virus that is responsible for respiratory tract infections and recurrent otitis media in humans (20). RSV-induced lower respiratory tract infection (LRTI), a complication of RSV infection of immunologically naïve children, represents the more clinically significant of these diseases. In the United States alone, LRTI accounts for 120,000 hospitalizations annually and is associated with postinfectious sequelae of recurrent episodic wheezing (36, 43, 44). Despite the fact that nearly 100% of U.S. children are infected by RSV before the age of 3, there is no efficacious vaccine or treatment (51). Because most patients with RSV-induced LRTI present at the time when viral titers are falling (57), the host signaling response to RSV infection is thought to play a significant role in disease pathogenesis. Our work and that of others have indicated that this paramyxovirus activates innate inflammatory signaling pathways in the airway epithelium that contribute to disease pathogenesis.
Upon inoculation, RSV replicates in the nasal mucosa, spreading from cell to cell into the lower respiratory tract through intraepithelial cellular bridges or via free virus in respiratory secretions binding to ciliated epithelial cells (19, 54). RSV replicates principally in epithelial cells in the mucosa, where it produces bronchiolar epithelial inflammation, epithelial necrosis, peribronchial monocytic infiltration, and submucosal edema (1, 52). The role of the innate immune response has been studied in rodent models of acute disease (16–18). These studies show that members of a cytokine network including interleukin-1 (IL-1), IL-6, tumor necrosis factor (TNF), KC, granulocyte colony-stimulating factor (G-CSF), and macrophage inflammatory protein 1α (MIP-1α) are rapidly secreted into the airway (9, 18); this event is followed by recruitment of neutrophils and macrophage/monocytes into the airway lumen and peribronchiolar space (9). Subsequently, clinical manifestations (weight loss, dyspnea, and increased airway resistance) are seen (9).
Several studies have shown a role of the NF-κB transcription factor in mediating disease pathogenesis. NF-κB is a family of inducible cytoplasmic transcription factors that plays a central role in controlling expression of inflammatory gene networks through activation and nuclear translocation of the potent RelA transcriptional activator (4, 6, 49). In RSV infection, NF-κB is activated in airway epithelium early in the course of infection (17). Inhibition of NF-κB activation reduces cytokine production and clinical disease, without decreasing viral replication (16). Together these findings suggest that activation of the host inflammatory response via NF-κB is a central step in the immunopathogenesis of LRTI.
As a result, the mechanisms by which RSV activates RelA have been intensively studied. In resting epithelial cells, RelA is retained in an inactivated state by its association with cytoplasmic ankyrin domain-containing inhibitors, predominately IκBα and the 100-kDa NF-κB2 precursor (24, 33). RSV induces cytoplasmic RelA release from IκBα by a mechanism mediated by activated IκB kinase (IKK) downstream of the RIG-I DEXD/H box-containing RNA helicase-MAVS complex (35), known as the canonical pathway. In addition, RSV activates a separate kinase composed of the NF-κB-inducing kinase (NIK)·IKKα complex, also associated with RIG-I-MAVS, that releases RelA from the NF-κB2/p100 complex (12, 33), known as the “cross talk” arm of the noncanonical pathway (6).
NF-κB release from cytoplasmic inhibition is necessary, but not sufficient, for target gene induction. Recent studies have shown that a separate, reactive oxygen species (ROS)-initiated and mitogen stress-related kinase 1 (MSK-1) pathway is required to induce RelA's transcriptional activity (22). This pathway converges on NF-κB phosphorylation at serine (Ser) residue 276. RelA Ser276 phosphorylation is coincident with its release from IκBα in the canonical pathway (22, 23, 39). Moreover, the fact that selective inhibition of Ser276 phosphorylation (without affecting Ser536 phosphorylation) blocks RSV-induced IL-8 expression (22) suggests that RelA Ser276 phosphorylation is a critical step in RSV-induced gene activation. The mechanism by which RSV-induced RelA Ser276 phosphorylation induces transcriptional activation is not completely understood. In other contexts, activated NF-κB mediates promoter-specific recruitment of p300/CBP, Src 1, and nuclear hormone receptor corepressor (N-CoR) coactivators to produce chromatin remodeling (45), transcription factor acetylation (10, 21), and stable enhanceosome formation (23) on inflammatory gene promoters. Whether these mechanisms are important in virus-induced NF-κB activation is not yet known.
Despite the understanding that NF-κB is a common integrator of diverse inflammatory signals, the pathways and regulatory kinases controlling its actions are stimulus specific. Specifically, RSV induces NF-κB, at least in part, via the activation of the noncanonical/cross talk pathway, whereas cytokine stimulation selectively activates the canonical pathway (33). Second, RSV-induced ROS formation is mediated by cytoplasmic oxidases and NOX2 (8, 13), whereas TNF induces mitochondrial ROS via a TRAF-dependent mechanism (30). Third, RelA Ser276 phosphorylation induced by RSV is mediated by nuclear actions of MSK-1, whereas TNF induces Ser276 phosphorylation by activation of a cytoplasmic IKK-associated catalytic subunit of protein kinase A (23, 56). These observations indicate that the structure and regulation of intracellular signaling pathways modifying the NF-κB activation are fundamentally distinct in response to viral infections.
In this study, we examined mechanism by which RelA Ser276 phosphorylation mediates RSV-inducible gene expression. Analysis of RSV-inducible gene expression in cells selectively expressing the RelA Ser276Ala mutant shows reduced expression of a subnetwork of inflammatory genes, including those for CXCL2/Groβ, IL-6, and KC. Stimulus-inducible RelA Ser276 phosphorylation is required for RelA Lys310 acetylation and stable association with the activated PTEF-b complex. RelA acetylation is required for transcription, because a RelA Lys310Arg mutant does not activate targets in RelA−/− mouse embryonic fibroblasts (MEFs) and, separately, functions as a dominant negative inhibitor of endogenous RelA in epithelial cells. In contrast to the pattern demonstrated by TNF, RSV-inducible recruitment of bromodomain 4 (Brd4) and cyclin-dependent kinase 9 (CDK9) is associated with their rapid clearance and later reassociation to NF-κB-dependent promoters. Small interfering RNA (siRNA)-mediated knockdown and the use of selective pharmacological inhibitors of CDK9 showed that CDK9 is functionally coupled to Ser2 phosphorylation of the RNA polymerase (Pol) II carboxyl-terminal domain (CTD) and RSV-inducible transcriptional elongation, whereas inhibition of the transcriptional initiation factors CDK7 and cyclin H did not have a detectable effect. These data indicate that RSV-induced RelA Ser276 phosphorylation mediates acetylation and inducible complex formation with PTEF-b that induces transcriptional elongation of an inflammatory gene network in infected epithelial cells.
Human A549 pulmonary epithelial cells (American Type Culture Collection, Manassas, VA) were grown in F12K medium (Invitrogen, Carlsbad, CA) with 10% fetal bovine serum (FBS), penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C in a 5% CO2 incubator (34). MEFs were cultured in Eagle's minimum essential medium (Gibco) with 0.1 mM nonessential amino acids, 1.0 mM sodium pyruvate, and 10% FBS (34).
A 300-bp duplex RNA was synthesized using as a template a fragment of human β-tubulin amplified from cDNA using flanking primers containing T7 RNA polymerase binding sites (29). The sense primer sequence was 5′-TAATACGACTCACTATAGGCCAC ACTCAGGCTGGTCAGT-3′, and the antisense primer sequence was 5′-TAATACGACTCACTATAGGTA CCACATCCAGGACAGAATC-3′ (T7 polymerase sites are underlined). The plasmid was synthesized in vitro using T7 bacterial polymerase (Megascript; Ambion, Inc.) and purified according to the manufacturers recommendation. Double-stranded RNA (dsRNA) was transfected into 2 × 106 freshly isolated cells in suspension (34) (Amaxa).
The human RSV A2 strain was grown in Hep-2 cells and prepared as described previously (50). The viral titer of purified RSV pools was varied from 8 to 9 log PFU/ml, as determined by a methylcellulose plaque assay. Viral pools were aliquoted, quick-frozen on dry ice-ethanol, and stored at −70°C until they were used.
Full-length human RelA and site mutants were expressed as FLAG-enhanced green fluorescent protein (EGFP) fusion proteins using the pCX4-Pur expression vector (2, 27). Retrovirus stocks produced in Bosc23 cells were cotransfected with pCX4-pur-EGFP-RelA and amphotrophic packaging plasmid pCL-10A1 (37). RelA−/− MEFs and A549 cells were infected in the presence of 8 μg/ml Polybrene and selected for puromycin (2 μg/ml) resistance.
Proteins from whole-cell extractions were fractionated by 10% SDS-polyacrylamide gel electrophoresis and electrotransferred to a polyvinylidene difluoride membrane. Membranes were blocked in 5% nonfat dry milk in TBS-T (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 0.1% Tween) and probed with the indicated primary antibody (Ab). Membranes were washed and incubated with IRDye 700-conjugated anti-mouse Ab or IRDye 800-conjugated anti-rabbit Ab (Rockland Inc., Gilbertsville, PA). TBS-T-washed membranes were imaged with an Odyssey infrared scanner (LiCor BioSciences, Lincoln, NE). Anti-RelA rabbit polyclonal Ab was from Santa Cruz Biotechnology (Santa Cruz, CA), anti-FLAG M2 was from Sigma (Sigma-Aldrich, St. Louis, MO), and anti-acetyl-RelA (anti-Ac-RelA), anti-CDK7 and -9, anti-cyclin H, and anti-α-tubulin were from Abcam (Abcam Inc., Cambridge, MA).
The indicated concentrations of whole-cell lysates from FLAG-EGFP-RelA-expressing A549 cells were incubated with anti-FLAG M2 affinity gel for 16 h at 4°C (55). The beads were washed with chilled lysis buffer (50 mM Tris-HCl [pH 7.5], 1 mM EGTA, 1 mM EDTA, 1% Triton X-100, 0.27 M sucrose, 0.1% [vol/vol] β-mercaptoethanol, 1 mM sodium orthovanadate, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, complete protease inhibitor cocktail [Sigma Aldrich] and 1 mM phenylmethylsulfonyl fluoride) four times by centrifugation (5,000 × g, 2 min, 4°C). The FLAG-EGFP-RelA was eluted by addition of 150 μg of 3× FLAG peptide and incubated with mixing for 1 h at 4°C. The mixture containing FLAG-EGFP-RelA protein complex and gel was filtered through empty spin column and flowthrough was collected.
For gene expression analyses, 1 μg of RNA was reverse transcribed using SuperScript III in a 20-μl reaction mixture (34). One microliter of cDNA product was diluted 1:2, and 2 μl of diluted product was amplified in a 20-μl reaction mixture containing 10 μl of SYBR green Supermix (Bio-Rad) and 0.4 μM (each) forward and reverse gene-specific primers (Table 1). The reaction mixtures were aliquoted into a Bio-Rad 96-well clear PCR plate, and the plate was sealed with Bio-Rad Microseal B film before being put into the PCR machine. The plates were denatured for 90 s at 95°C and then subjected to 40 cycles of 15 s at 94°C, 60 s at 60°C, and 1 min at 72°C in an iCycler (Bio-Rad). PCR products were subjected to melting curve analysis to ensure that a single amplification product was produced.
Quantification of relative changes in gene expression was done using the ΔΔCT method. In brief, the ΔCT value was calculated (normalized to glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) for each sample by using the equation ΔCT = CT (target gene) − CT (GAPDH). Next, the ΔΔCT was calculated by using the equation ΔΔCT = ΔCT (experimental sample) − ΔCT (control sample). Finally, the fold differences between the experimental and control samples were calculated using the formula 2−ΔΔCT. Quantification of absolute concentrations of the short and spliced RNA transcripts was performed by estimating transcript number relative to serial dilutions of cDNA standards in RT-PCR.
Control, CDK7, CDK9, and cyclin H siRNAs (Dharmacon, ThermoFisher Scientific, Lafayette, CO) were reverse transfected at a 100 nM concentration into A549 cells by use of TransIT-siQUEST transfection reagent (Mirus Bio Corp), and 48 h later, cells were infected or not with RSV (multiplicity of infection [MOI], 1.0) for another 24 h. Cells were washed with phosphate-buffered saline (PBS) twice and lysed in TRI reagent (Sigma-Aldrich).
Two-step chromatin immunoprecipitation (XChIP) was performed as described previously (38). A549 cells (4 ×106 to 6 ×106 per 100-mm dish) were washed twice with PBS. Protein-protein cross-linking was first performed with disuccinimidyl glutarate (Pierce), followed by protein-DNA cross-linking with formaldehyde. Equal amounts of sheared chromatin were immunoprecipitated overnight at 4°C with 4 μg of the indicated Ab in ChIP dilution buffer (38). Immunoprecipitates were collected with 40 μl protein A magnetic beads (Dynal Inc.), washed, and eluted in 250 μl elution buffer for 15 min at room temperature. Samples were de-cross-linked in 0.2 M NaCl at 65°C for 2 h. The precipitated DNA was phenol-chloroform extracted, precipitated with 100% ethanol, and dried.
Gene enrichment in XChIP was determined by quantitative real-time genomic PCR (Q-gPCR) as previously described (38, 48) using region-specific PCR primers (Table 2). Standard curves were generated using a dilution series of genomic DNA (from 40 ng to 25 μg) for each primer pair. The fold change of DNA in each immunoprecipitate was determined by normalizing the absolute amount to the input DNA reference and calculating the fold change relative to that amount in unstimulated cells.
EGFP-RelA-expressing MEFs were plated on cover glasses pretreated with rat tail collagen (Roche Applied Sciences). After the indicated stimulation, the cells were fixed with 4% paraformaldehyde in PBS and counterstained with 4′,6-diamidino-2-phenylindole (DAPI). The cells were visualized with a Zeiss fluorescence LSM510 confocal microscope at a magnification of ×63 (34, 35).
All images were analyzed using CellProfiler (28), a platform for automated image analysis. A custom-built pipeline was used to segment and quantify these images. In each image field, cells that were overlapping other cells, that were dividing, or whose outlines were only partially present within the image view were excluded from analysis. Images were corrected for uneven illumination using the Correction_Illumination module for both DAPI (blue) and EGFP (green) channels using the “regular” method and “Gaussian filter” as smoothing methods, respectively. The illumination functions were then applied to each input image to correct for the variation in illumination pattern. Cell identification was initiated by using nuclei as “seed,” and then the entire cell was identified using the “propagation” method of segmentation using fluorescence intensities. Upon completion of the automated cellular segmentation, the images were manually reviewed and the intensity for each channel was quantified. Where indicated, the nuclear EGFP-RelA fluorescence signal was divided by the total cellular EGFP-RelA signal to obtain the fraction of the NF-κB/RelA population that translocates into the nucleus.
One-way analysis of variance (ANOVA) was performed when looking for time differences, followed by Tukey's post hoc test to determine significance. Mann-Whitney tests were used for nonparametric data. A P value of <0.05 was considered significant.
Human type II-like A549 airway epithelial cells (AECs) productively replicate RSV and are a well-established model for study of the lower airway epithelial cell response to RSV infection (14, 15, 25). Here, RSV replication can be detected within 6 to 9 h after viral adsorption at an MOI of 1.0, with infective virions detectable in the cell culture supernatant within 15 h, at times coincident with activation of the canonical and cross talk NF-κB pathways (12, 24). To establish the time course of RelA Ser276 phosphorylation, whole-cell lysates of RSV-infected A549 cells were subjected to Western immunoblotting for phospho-Ser276 RelA. A 2-fold increase in phospho-Ser276 RelA was observed at 15 h after infection, and phosphorylation was strongly induced by 24 h (Fig. 1A).
Previous work from our group (47) has shown that cytokine activation of the NF-κB pathway results in expression of distinct temporal waves of its downstream gene network, with a group of inflammatory cytokines rapidly responding to stimulation (“early” genes), whereas others are delayed in their response (“late” genes). To determine if the NF-κB network showed distinct profiles in response to RSV, a time course of mRNA expression was analyzed for 6 genes by Q-RT-PCR. Here, RSV strongly induced the parallel expression of CXLC2/Groβ, IL-8, TNFAIP3/A20, IL-6, CCL5/RANTES, and NFKBIA/IκBα, with a increase in mRNA expression first detectable at 6 h and increasing monotonically until 24 h (Fig. 1B). All genes measured showed qualitatively similar profiles and did not demonstrate the cytokine-induced waves of expression characteristic of tonic cytokine stimulation. We also noted that the mRNA expression changes paralleled the profile of phospho-Ser276 RelA formation (Fig. 1A). Viral replication, determined by measurement of RSV transcription, also followed an indistinguishable profile (Fig. 1B). Together these data indicate that the NF-κB gene network is synchronously activated in a time course paralleling that of RSV replication.
To determine whether phospho-Ser276 RelA was a posttranslational modification required for the activation of the NF-κB-dependent gene network, RelA−/− MEFs stably transfected with either FLAG-EGFP-tagged wild-type RelA (RelA WT) or an FLAG-EGFP-tagged nonphosphorylatable RelA with a Ser276-to-Ala mutation (RelA Ser276Ala) were stimulated in the absence or presence of RSV. Total RNA was extracted, and changes in murine CXCL2/Groβ, KC, IL-6, and NFKBIA/IκBα gene expression were determined by Q-RT-PCR. In RelA WT-expressing MEFs, RSV induced a potent 20-fold increase mCXCL2/Groβ mRNA expression after 6 h, which increased to a 120-fold induction at 24 h (Fig. 1C). In contrast, in RelA Ser276Ala-expressing cells, the expression of mCXCL2/Groβ was significantly blunted, with less than half of the mRNA expression seen in RelA WT-expressing cells at any time point after RSV infection. Similar results were observed for the KC and IL-6 genes; however, NFKBIA/IκBα was inducible to similar levels in both cell types, indicating that its expression was independent of the requirement for RelA Ser276 phosphorylation (Fig. 1C). Additionally, similar levels of RSV replication were observed in both cell types (Fig. 1D). These data suggest that the phospho-Ser276 RelA modification is required for the activation of a subset of inflammatory genes in response to RSV but does not affect the RelA-dependent activation of its autoregulatory inhibitory loop.
TNF-induced Ser276 phosphorylation controls its association with the p300/CBP histone and transcription factor acetyltransferase (11, 23, 41); we therefore examined whether RSV induces RelA acetylation. To assess acetylation of endogenous RelA, the protein was enriched from a time course of RSV infection by immunoprecipitation prior to Western immunoblotting by anti-Ac-Lys310 RelA Ab. Although Ac-Lys310 RelA could be detected in uninfected cells, RSV induced a 2-fold accumulation of Ac-Lys310 RelA at 15 and 24 h after infection (Fig. 2 A). To confirm that the EGFP-RelA protein was modified in a similar manner, immunoprecipitation and Western blotting were performed on EGFP-RelA cells from a time course of infection with RSV. An indistinguishable pattern of acetylation was observed (Fig. 2B). We note that the magnitude of Ac-Lys310 RelA formation induced by RSV was similar to that produced by activation of the TNF receptor pathway (Fig. 2B, right panel).
We noted the parallel induction of RelA phosphorylation and acetylation in our time course experiments. To establish whether there was an interrelationship between the two posttranslational modifications, we measured RelA acetylation in RelA WT- or RelA Ser276Ala-expressing MEFs. Consistent with the observations on A549 cells, RelA Lys310 acetylation could be detected after 24 h of RSV infection; however, the inducible RelA Lys310 acetylation was abolished in the RelA Ser276Ala-expressing cells (Fig. 2C), indicating that Ser276 phosphorylation was required for RelA Lys310 acetylation in response to RSV infection.
To examine the requirement of RelA acetylation for gene expression, the EGFP-RelA Lys310Arg mutation was stably expressed in RelA−/− MEFs, where similar levels of expression of mutant and wild-type RelA proteins were observed by Western immunoblotting (Fig. 3A). Since acetylation may influence nuclear translocation, we first examined the transfectants by fluorescence microscopy. In order to synchronize the cellular response, cells were transfected in the absence or presence of dsRNA prior to confocal imaging. In RelA WT-expressing cells, the EGFP fluorescence was cytoplasmic in unstimulated cells. Upon exposure to dsRNA, a strong nuclear translocation was observed (Fig. 3B). A similar finding was observed for RelA Lys310Arg, with the protein in a cytoplasmic distribution in unstimulated cells and the majority demonstrating nuclear translocation upon stimulation (Fig. 3B, lower panels). The images were subjected to quantitative image analysis, where the fraction of nuclear translocation was plotted as a histogram of the number of cells observed by treatment and RelA mutation (Fig. 3C). In cells expressing EGFP-RelA WT, the median nuclear/cytoplasmic ratio of control cells is 17% (n = 52 cells observed), versus 28% for those stimulated with dsRNA (n = 56) (P < 0.0001). Similarly, for cells expressing EGFP-RelA Lys310Arg, the median nuclear/cytoplasmic ratio in control cells is 17% (n = 40), versus 25% for dsRNA-stimulated cells (n = 33) (P < 0.0001). There was no significant difference between nuclear/cytoplasmic ratios of dsRNA-stimulated EGFP-RelA WT- and EGFP-RelA Lys310Arg-expressing transfectants. We therefore concluded that EGFP-RelA Lys310Arg mutant translocation was inducible in a manner indistinguishable from that for EGFP-RelA WT.
We next asked whether RelA Lys310Arg could stably interact with its endogenous target genes. For this purpose, we performed a quantitative two-step chromatin immunoprecipitation (XChIP) assay with EGFP-RelA Lys310Arg-expressing cells using anti-RelA Ab, followed by quantification with primers spanning the mCXCL2/Groβ RelA binding sites using Q-gPCR (48). Basal EGFP-RelA Lys310Arg binding could be seen by comparing mGroβ enrichment with anti-RelA Ab with that produced by IgG (Fig. 3D). In response to RSV, a 7-fold-increased binding of EGFP-RelA Lys310Arg to the mCXCL2/Groβ promoter over that in uninfected cells was observed, indicating that RelA Lys310Arg inducibly binds to NF-κB targets within native chromatin.
To assess the requirement of RelA Lys310 acetylation for RSV-induced NF-κB-dependent gene expression, changes in NF-κB-dependent gene expression were determined by Q-RT-PCR in RelA WT- and RelA Lys310Arg-expressing MEFs. A significant reduction in RSV-induced CXCL2/Groβ KC, IL-6, and NFKBIA/IκBα gene expression was observed (Fig. 3E). No difference in RSV replication was seen (Fig. 3F), indicating that the mutation does not affect the innate antiviral response of A549 cells.
To further support the conclusion that RelA Lys310Arg is deficient in transcriptional activation, we measured its actions as a dominant negative inhibitor. In this experiment, A549 cells endogenously expressing RelA WT were stably transfected with EGFP-RelA WT or EGFP-RelA Lys310Arg retroviruses. The expression levels of EGFP-RelA were ~4-fold greater than those of endogenous RelA in both cell types in Western blotting (Fig. 3G). Strikingly, the expression of NF-κB-dependent genes in response to RSV was dramatically inhibited in cells expressing EGFP-RelA Lys310Arg versus those expressing EGFP-RelA WT (Fig. 3H). For example, the RSV-induced 25-fold induction of IL-6 expression in EGFP-RelA WT-expressing cells was reduced to 4-fold in those expressing EGFP RelA Ser276Ala and to 10-fold in those expressing EGFP-RelA Lys310Arg (Fig. 3H, bottom panel). We interpret these data to indicate that the RelA Lys310Arg mutant is transcriptionally defective, functioning as a dominant negative inhibitor.
Earlier we discovered that RelA Ser276 phosphorylation was required for its TNF-induced association with CDK9/CcnT1, kinase components of the positive transcription elongation factor (PTEF-b) complex (39). PTEF-b is distributed in two states: as an inactive complex with 7SK snRNA or in an activated state associated with the bromodomain 4 (Brd4) histone reader (26). To determine whether RSV induces RelA-Brd4 association and whether this association is phospho-Ser276 RelA dependent, coimmunoprecipitation assays were performed with RelA−/− MEFs expressing RelA WT and RelA Ser276Ala in the absence or presence of RSV infection. Whole-cell lysates were enriched for RelA with anti-FLAG affinity beads and subjected to Western blotting using anti-Brd4 Ab. Although Brd4 could be only weakly identified in the RelA complexes in the absence of infection, in response to RSV, Brd4 was significantly enriched (Fig. 4A, top panel). In contrast, Brd4 recruitment to the RelA complex was completely inhibited by the RelA Ser276Ala mutation.
To confirm that RelA Ser276 phosphorylation was required for association with the PTEF-b complex, we also measured binding of CDK9, the kinase responsible for serine phosphorylation of the heptad repeats of the RNA Pol II carboxyl-terminal domain (CTD). As with its binding to Brd4, RSV induced RelA association with CDK9, and this association was significantly reduced on the RelA Ser276Ala mutant (Fig. 4A). We conclude that Ser276 phosphorylation is a key posttranslational event regulating RelA association with the activated PTEF-b complex.
To extend these studies, we examined the recruitment of Brd4 to endogenous target genes by XChIP. In TNF-stimulated A549 cells, RelA was rapidly recruited to the CXCL2/Groβ promoter within 15 min (Fig. 4B, top). Similarly, Brd4 was rapidly recruited 10-fold to the CXCL2/Groβ promoter within 15 min, declining thereafter, despite the continued presence of RelA binding (Fig. 4B). In contrast, a distinct binding profile was observed in response to RSV. Here, RelA is recruited to the CXCL2/Groβ promoter over 15 to 24 h, in a pattern consistent with its activating Ser276 phosphorylation and expression of CXCL2/Groβ mRNA. In contrast with the pattern produced by TNF, the RSV-induced Brd4 binding was significantly reduced after 15 h of infection, to less than 20% of the levels seen in control samples (Fig. 4C). Brd4 subsequently transiently increased at the 24- and 30-h time points, but it did not fully reaccumulate to baseline levels.
Because Brd4 binds preferentially to Ac-Lys5 histone H4 (26), we asked whether RSV affected the dynamics of Ac-Lys5 H4 association with NF-κB-dependent genes. Interestingly, although RSV strongly induced the association of Ac-Lys5 H4 on CXCL2/Groβ, Brd4 binding was inversely related, with a peak of Ac-Lys5 H4 occurring at a time when Brd4 was being cleared from the promoter (Fig. 4C). A similar out-of-phase binding pattern of Brd4 and Ac-Lys5 H4 was also observed on the IL-8 gene promoter by XChIP (Fig. 4D).
Promoter recruitment of the P-TEFb complex induces phospho-Ser2 formation on the CTD of RNA Pol II via CDK9, resulting in clearance of the CDK9-RNA Pol II complex from the proximal promoter and its movement toward the 3′ end of the gene. To measure this dynamic, we assayed the pattern of CDK9-RNA Pol II complex binding by gene region-specific XChIP using PCR primers designed to amplify the CXCL2/Groβ 5′ untranslated region (5′UTR), intron 3 (I3), or 3′UTR (Fig. 5A). XChIP assays were performed with anti-CDK9 Ab and anti-phospho-Ser2 RNA Pol II CTD Abs on chromatin prepared from a time course of TNF stimulation. Here CDK9 binding to the CXCL2/Groβ 5′UTR could be detected in the absence of stimulation (compare unstimulated IgG and anti-CDK9 in Fig. 5B, top left). TNF induced a rapid 3-fold recruitment of CDK9 binding to the 5′UTR within 15 min, an interaction that persisted for up to 1 h after stimulation. CDK9 binding to intron 3 and the 3′UTR was delayed relative to that seen on the 5′UTR, peaking 10-fold at 0.5 h (for I3 primers) and 6-fold at 1 h (for 3′UTR primers) (Fig. 5B, top). These data suggest that TNF induces CDK9 loading on the proximal promoter, followed by progression onto the downstream regions of the gene.
A time-dependent regional analysis of phospho-Ser2 RNA Pol II binding was also performed. Basal phospho-Ser2 RNA Pol II binding to the 5′UTR rapidly increased 4.8-fold within 15 min of TNF stimulation and plateaued at 3-fold after 1 h of stimulation (Fig. 5B, lower panels). Phospho-Ser2 RNA Pol II loading to intron 3 and the 3′UTR was also increased and temporally delayed from that pattern seen on the 5′UTR. A 10-fold increase of phospho-Ser2 RNA Pol II binding was observed at 1 h after stimulation on intron 3, and a 17-fold increase was observed after 1 h on the 3′UTR (Fig. 5B, lower panels). Together these data indicated that TNF induces CDK9 and phospho-Ser2 RNA Pol II loading on the CXLC2/Groβ proximal promoter and its downstream clearance.
A similar regional analysis of CDK9 and phospho-Pol II was performed with chromatin isolated from RSV-infected AECs. Strikingly in contrast to that behavior seen in TNF stimulation, RSV infection resulted in the clearance of CDK9 binding on the 5′UTR to less than 50% of basal CDK9 binding 24 h after infection, followed by a rebound at 30 h to levels greater than those seen in uninfected cells. This initial clearance of CDK9 was also observed for its association with the intron 3 and 3′UTR regions, where a pattern that was qualitatively similar, but reduced in magnitude, was seen (Fig. 5C, top panels). We also note that the profiles of Brd4 binding are quite similar.
We also applied regional analysis to quantify RSV-induced temporal changes in phospho-Ser2 RNA Pol II binding. Here, RSV also produced an initial clearance of phospho-Ser2 RNA Pol II at 15 and 24 h after RSV infection, falling to 10% of the initial levels at 24 h. Although phospho-Ser2 RNA Pol II increased at 30 h, these levels did not yet approach those seen in uninfected cells (Fig. 5C, bottom panels). The intron 3 and 3′UTR data are qualitatively similar to those patterns seen on the 5′UTR. Together these data indicate that RSV induces promoter clearance of CKD9 and phospho-Ser2 RNA Pol II on NF-κB-dependent target genes.
Because the distinctly longer time scales for NF-κB recruitment produced by RSV infection versus TNF stimulation may allow us to detect CDK9/Pol II clearance more easily, we performed similar experiments using dsRNA administration, a treatment that more rapidly activates NF-κB. Within 4 h of dsRNA stimulation, the level of RelA binding to the CXCL2/Groβ proximal promoter was increased 4.8-fold, a pattern that more closely mimics the rapid TNF response (Fig. 5D). Here, CDK9 binding to CXCL2/Groβ was also reduced to 40% of that seen in untreated cells and remained reduced for 4 h (Fig. 5E). CDK9 binding was also reduced at 1 h on the intron 3 and the 3′UTR and subsequently returned to baseline after 4 h of stimulation (Fig. 5E, top panels). Similar qualitative behavior was seen for the patterns of phospho-Ser2 RNA Pol II, with rapid clearance from the 5′UTR and gradual reaccumulation on the intron 3 and 3′UTR regions of the gene (Fig. 5E, bottom panels). Together these data indicate that dsRNA stimulation and RSV replication induce promoter clearance of CDK9 and phospho-Ser2 Pol II binding in a manner qualitatively different from those induced by TNF stimulation.
Together our data indicate that the role of the Ser276 phosphorylation-Lys310 acetylation pathway in the activation of NF-κB-dependent genes was due, at least in part, to the process of transcriptional elongation mediated by Brd4-CDK9 recruitment. To further explore the mechanism of transcriptional activation, we also compared the effects of TNF and RSV on recruitment of the transcriptional initiation complex proteins CDK7 and cyclin H by XChIP. Here, TNF stimulation resulted in a rapid 5-fold recruitment of both CDK7 and cyclin H to the proximal CXLC2/Groβ promoter 30 min after stimulation (Fig. 5F). In contrast, RSV induced both subunits <2-fold 24 h after infection (Fig. 5G). Together these data indicate that TNF induces components of both transcriptional elongation and initiation complexes to bind NF-κB-dependent promoters.
Based on this anomalous behavior of PTEF-b binding in RSV infection, we sought to examine the functional role of CDK9 in NF-κB-dependent gene expression. Two approaches were therefore used: we modulated its abundance using siRNA transfection and inhibited kinase activity using pharmacologic CDK9 inhibitors. The levels of CDK9 were inhibited by siRNA transfection in AECs (Fig. 6A). Here, the transfection of CDK9 siRNA reduced the basal CDK9 mRNA to ~25% of that with nonspecific siRNA (Fig. 6A). Interestingly, CDK9 mRNA expression was induced by RSV infection. In both basal and RSV infection conditions, CDK9 expression was significantly reduced by CDK9 siRNA. The expression of NF-κB-dependent genes was then assessed by Q-RT-PCR. CXCL2/Groβ, IL-8, IL-6, and CCL5/RANTES were significantly inhibited in CDK9-silenced cells, whereas NFKBIA/IκBα and TNFAIP3/A20 were not affected (Fig. 6B). Moreover, CDK9 inhibition had no effect on RSV transcription (Fig. 6C).
To evaluate the relative functional role of transcriptional initiation in RSV-induced gene expression, CDK7 and cyclin H were silenced in A549 cells and the effect on TNF- and RSV-inducible gene expression determined by Q-RT-PCR. Upon transfection of CDK7 siRNA, CDK7 mRNA expression was reduced to ~16% of that seen in control treated cells (Fig. 6D). Similarly, cyclin H siRNA reduced cyclin H mRNA to ~23% of that seen in control cells. The RSV-inducible expression of IL-6 and TNFAIP3/A20 was slightly enhanced by CDK7 or cyclin H knockdown (Fig. 6D), without an effect on RSV transcription. These data indicate that transcriptional initiation does not play a major role in expression of NF-κB-dependent genes in response to RSV.
Despite multiple attempts at optimization, including timing, repeated dosing, and variation of transfection methods, we were never successful in completely inhibiting CDK9 expression using siRNA, a finding consistent with its essential role in cell survival. To confirm its role in RSV-induced NF-κB activation, we also evaluated the effect of flavopiridol (FP), a synthetic flavenoid with potent CDK inhibitory activity (42). FP-treated AECs were infected with RSV for various times, and inducible changes in NF-κB-dependent mRNAs were detected by Q-RT-PCR (Fig. 7A). Compared to vehicle treatment alone, FP was a potent inhibitor of RSV-induced CXCL2/Groβ, IL-8, IL-6, and CCL5/RANTES, relatively sparing NFKBIA/IκBα in a pattern similar to that produced by CDK9 silencing (compare Fig. 7A and and6B).6B). Consistent with its potent inhibitory effects on gene expression, FP also produced a significant inhibition of NF-κB-dependent cytokine secretion of IL-8, MCP-1, IP-10, IL-6, and CCL5/RANTES into the medium as measured by multiplex enzyme-linked immunosorbent assay (ELISA) (Fig. 7B).
To confirm that CDK9 kinase activity mediates RSV-induced NF-κB-dependent gene expression, we also tested the unrelated CDK inhibitor CAN508. CAN508 has a 10-fold lower 50% inhibitory concentration (IC50) for CDK9 than for CDK1 to -7 and therefore is CDK9 selective (31, 32). Dose-out studies of CAN508 were performed to identify the lowest effective concentration of CAN508 (data not shown). The treatment of AECs with 100 to 200 μM CAN508 potently inhibited RSV-induced CXCL2/Groβ, IL-8, IL-6, CCL5/RANTES, and TNFAIP3/A20, with a lesser effect on NFKBIA/IκBα expression (Fig. 8A). As observed with FP, CAN508 at these doses has no effect on RSV transcription (Fig. 8B).
To establish that CDK9 mediates RSV-induced RNA Pol II Ser2 phosphorylation, we conducted XChIP on chromatin from control or CAN508-treated cells in the absence or presence of RSV infection. In the presence of CAN508, the promoter abundance of phospho-Ser2 Pol II was significantly reduced (Fig. 8C). Together these data indicate that CDK9 kinase activity is required for RSV-inducible RNA Pol II phosphorylation and NF-κB-dependent gene expression.
A rapid response of innate immune response genes is essential for host survival. Our work and that of others have demonstrated that a subset of highly inducible NF-κB-dependent genes are maintained in an open chromatin configuration bound by RNA Pol II (39). Here, hypophosphorylated RNA Pol II nonproductively cycles and produces abortive transcripts of ~30 nucleotides (nt) in length (46). Upon Ser2 phosphorylation, RNA Pol II enters transcriptional elongation mode, producing full-length, fully spliced mRNAs. To measure the transcriptional elongation, primers specific for the 5′UTR were designed to measure total RNA (which will detect abortive transcripts and spliced RNA) and spliced RNA (schematically diagrammed in Fig. 9A).
Total RNA from MEFs stably expressing either RelA WT or RelA Ser276Ala in the absence or presence of RSV infection was isolated and analyzed by use of region-specific primers. In response to RSV infection, a 5.8-fold increase in total mCXCL2/Groβ transcripts was observed in RelA WT-expressing cells, which was significantly reduced (to less than 2-fold) in cells expressing the RelA Ser276Ala mutant (Fig. 9B, left panel). In contrast, RSV induced a 13-fold increase in spliced RNA transcripts in RelA WT-expressing cells (Fig. 9B, right panel), whereas the RSV-induced increase in spliced transcripts was completely abolished in RelA Ser276Ala-expressing cells (Fig. 9B, right panel). Estimation of the transcriptional elongation index was determined by calculating the ratio of spliced to total mGroβ transcript abundance. Although RSV infection increased the mGroβ RelA transcriptional elongation complex, this induction was lost in RelA Ser276Ala-expressing cells (Fig. 9B, inset).
To confirm the role of CDK9 in RSV-induced transcriptional elongation, quantification of RSV-induced total and spliced mGroβ transcripts in the presence or absence of CAN508 was performed. In the absence of RSV infection, CAN508 induced basal expression of total mRNA transcripts (Fig. 9C), consistent with its known inhibitory actions on negative elongation factors (5). In response to RSV, CAN508 dramatically reduced the expression of fully spliced transcripts (Fig. 9C). We interpret these data to indicate that RSV induces transcriptional elongation of NF-κB-dependent genes in a phospho-Ser276 RelA-CDK9 dependent mechanism.
The paramyxovirus RSV is a potent inducer of the innate immune response in AECs, being responsible for the activation of downstream inflammatory gene networks important in disease pathogenesis. Here, NF-κB is released from cytoplasmic inhibitors via both the canonical and cross talk pathways (12, 24, 33). Earlier work has shown that RSV replication induces an ROS and MSK-1 pathway whose action is required to induce RelA's transcriptional activity (22), converging on RelA phosphorylation at Ser276 (22). Because cytokine- and RSV-inducible NF-κB activation are mediated by distinct signaling pathways, we have examined the mechanism by which RelA Ser276 phosphorylation mediates RSV-inducible gene expression. Here, experiments involving expression of a RelA Ser276Ala mutant showed that RelA Ser276 phosphorylation is required for the synchronous activation of a subnetwork of inflammatory genes by RSV, including the CXCL2/Groβ, IL-6, IL-8, and CCL5/RANTES genes. Stimulus-inducible RelA Ser276 phosphorylation is required for RelA Lys310 acetylation and its stable association with activated PTEF-b complex, containing Brd4 and CDK9. Because of this association, we examined the functional role of transcriptional elongation in RSV-inducible inflammatory cytokine expression. siRNA-mediated knockdown and the use of selective pharmacological inhibitors of CDK9 showed that CDK9 actions are coupled to RNA Pol II Ser2 phosphorylation and inducible transcriptional elongation. Analysis of spliced and total RNA transcripts indicated that RelA Ser276 phosphorylation is required for RSV-induced transcriptional elongation. These data indicate that RelA Ser276 phosphorylation is a critical switch coupled to subsequent posttranslational modifications (Lys acetylation) and inducible complex formation with PTEF-b, producing CTD RNA Pol II phosphorylation, and that transcriptional elongation partially mediates the cellular response to viral replication.
Studies using cytokine stimulation showed that Ser276-phosphorylated RelA activates only a subset of early response genes, including those for IL-8, IL-6, and CXLC2/Groβ (39). Although our earlier work showed that phospho-Ser276 RelA is required for RSV-inducible CXLC2/Groβ expression (22), this study has extended the phospho-Ser276 RelA-controlled subnetwork to include the IL-6 and KC genes. Interestingly, NFKBIA/IκBα expression is not affected by the RelA Ser276Ala mutation, a finding that suggests that NF-κB activates the major negative feedback loop of the NF-κB pathway via a distinct transcriptional mechanism, independent of RelA Ser276 phosphorylation. Our Q-RT-PCR studies of CDK7 and cyclin H silencing indicate that the process of transcriptional initiation is not strongly involved in RSV-induced NF-κB-dependent gene expression. However, because we did not totally silence expression of these kinases, we cannot completely exclude the possibility that transcriptional initiation plays a role in RSV-induced NF-κB-dependent gene expression, but we interpret its contribution as being minor. Together, our data indicate that RelA Ser276 phosphorylation in response to RSV or TNF is a common convergence point that mediates transcriptional elongation of the inflammatory gene subnetwork in response to diverse stimuli.
In response to distinct stimuli, RelA is phosphorylated on multiple sites, including Ser205, -276, -281 -529, -536, and others (3, 22, 53). Of these, phosphorylations at Ser276 and -536 have been identified as major regulatory events in RSV-inducible gene expression and are controlled by distinct upstream kinases, with Ser276 being regulated by MSK-1 and Ser536 being modulated by IKKε or -β (22, 53). Because Ser276 phosphorylation is required for expression of proinflammatory genes related to pathogenesis, we have focused this study on its mechanisms of transactivation. Ser276 phosphorylation decreases intermolecular NH2- and COOH-terminal intermolecular interactions (56), an event that may be permissive for promoting protein-protein interactions. For example, phospho-Ser276 is required for its stable interaction with p300/CBP binding and stable enhanceosome formation (23). Previous studies from our group have shown that Ser276 RelA phosphorylation is a posttranslational modification necessary for TNF-induced complex formation with CDK9 and cyclin T1. Our finding that RSV induces RelA association with both Brd4 and CDK9 proteins indicates that RSV induces RelA association with activated PTEF-b. Coimmunoprecipitation experiments using the RelA Ser276Ala mutation further show that RSV-induced recruitment of the activated PTEF-b complex is phospho-Ser276 dependent. Further studies will need to be conducted to determine whether phosphorylation-inducible PTEF-b complex formation is due to conformational changes or stabilized by the Ac-Lys310 RelA modification.
In response to Ser276 phosphorylation, RelA is multiply acetylated at adjacent Lys residues 310, 314, and 315. Of these, Lys310 is the major site involved in transcriptional activation, with Lys314 and Lys315 playing minor roles, if any (7). Whether Ser276 phosphorylation-induced Lys310 acetylation is a general or stimulus-specific coupling has not yet been fully analyzed. Here we observe that RelA Lys310 acetylation is produced in response to RSV replication and that acetylation is phospho-Ser276 dependent, confirming that the linkage between phosphorylation and acetylation is also seen in response to single-stranded RNA (ssRNA) virus infection. Our experiments analyzing the transcriptional properties of the Lys310Arg site mutation indicate that inhibiting Lys310 acetylation also produces defective transcriptional responses to RSV and that this protein functions as a dominant negative inhibitor of endogenous RelA. We posit that part of this dominant negative behavior is due to disruption of acetylation-mediated protein-protein interactions with essential transcriptional coactivators and transcriptional elongation complexes.
Although it is established that inducible NF-κB mediates promoter-specific recruitment of p300/CBP, Src 1, and N-CoR coactivators to produce chromatin remodeling (45) on target genes, it is not clear whether this is a general mechanism for all NF-κB-dependent genes. Our previous XChIP studies have shown that the IL-8 and CXCL2/Groβ genes are a subset of rapid and highly inducible genes that are maintained in an open chromatin configuration. In this NF-κB-dependent subnetwork, chromatin-remodeling activities initiated by activated RelA are not relevant for promoter activation. Instead, in these gene promoters, bound RNA Pol II is hypophosphorylated on the CTD of its large subunit and in a stalled configuration. Hypophosphorylated RNA Pol II nonproductively cycles and produces short, abortive transcripts ~30 nt in length. Upon CTD phosphorylation, RNA Pol II enters a processive transcriptional elongation mode, producing full-length, fully spliced mRNAs. Studies investigating the mechanisms controlling RNA Pol II CTD phosphorylation have shown that the essential role of the catalytic subunit of CDK9 (40). Our finding that phospho-Ser2 RNA Pol II is inhibited by CAN508 is consistent with this model. Moreover, our data showing that TNFAIP3/A20 expression does not strongly depend on CDK7 or cyclin H suggest that transcriptional elongation, rather than transcriptional initiation, plays a role in the rapid innate transcriptional response.
The XChIP results shown here indicate that the patterns of CDK9 association and clearance are distinct for cytokine versus dsRNA stimulation. In contrast to the pattern of CDK9 loading on the proximal promoter produced by TNF, both dsRNA and RSV induce proximal promoter clearance of activated PTEF-b, which is later followed by reassociation (Fig. 4 and and5).5). The initial clearance of the PTEF-b complex occurs prior to detectable changes in RelA binding, suggesting that this initial clearance is NF-κB independent. We speculate that other chromatin-modifying signals induced by RSV may mediate this event. For example, IKKα and p38/mitogen-activated protein kinase (MAPK), both of which are activated in response to RSV, are known chromatin modifiers through histone phosphorylation. We note that the later phases of PTEF-b reassociation with the proximal promoter and downstream regions of the gene are temporally coincident with RelA binding and independent of Ac-Lys5 histone H4 association. These findings suggest that CDK9 recruitment is RelA dependent, consistent with our earlier findings in response to TNF stimulation (39). The magnitude of PTEF-b reassociation is small, which may reflect the finding that RelA activation in RSV-infected cells is a stochastic process, where only 30% of the cells exhibit activated RelA at any one time (35). In this situation, XChIP assays measuring RelA-dependent promoter processes are averaged over a larger population of inactive cells.
Despite the distinct patterns of clearance/reassociation of PTEF-b produced by RSV infection, CDK9 activity was functionally important in downstream RSV-inducible gene expression, as demonstrated experimentally using siRNA-mediated silencing. Although CDK9 knockdown was incomplete, inhibition of RSV-inducible CXCL2/Groβ, IL-8, IL-6, and CCL5/RANTES expression was clearly evident (Fig. 6B). Importantly, this subnetwork of NF-κB-dependent genes is the same that is inhibited by (i) the RelA Ser276Ala mutation (Fig. 1C), (ii) ROS inhibition by antioxidant treatment (22), and (iii) MSK-1 knockdown (22). We interpret these results to suggest that this subnetwork is controlled by an ROS-MSK-1-Ser276 phosphorylation-PTEF-b pathway.
The effect of CDK9 silencing was further supported by inhibition of its kinase activity using relatively selective pharmacological inhibitors. We note that CAN508 is a compound with a 10-fold-lower IC50 for CDK9 than for CDK1 to -7 (31), and so it is CDK9 selective (32). As a measure of its effect in our system, we show that CAN508 inhibits the RSV-induced formation of phospho-Ser2 RNA Pol II on NF-κB-dependent targets (Fig. 8C). Because CAN508 has potent inhibitory actions on the RSV-inducible CXCL2/Groβ, IL-8, IL-6, and CCL5/RANTES subnetwork, we suggest that transcriptional elongation is an important component of inducible cytokine expression. An additional finding from this study is that FP and CAN508 are much more potent inhibitors of NF-κB-dependent gene expression than CDK9 siRNA. One explanation may be that nuclear CDK9 levels are in vast excess of that required to support inducible expression of NF-κB-dependent genes and that residual CDK9 activity in the siRNA experiment partially supports gene expression. An alternative explanation is that other, as-yet-unknown FP- or CAN508-sensitive cyclin-dependent kinases are participating in this process. Along these lines, we interpret the relative resistance of TNFAIP3 and NFKBIA/IκBα expression to CDK9 silencing (but not to pharmacological inhibition) to indicate that these genes may have a lower requirement for CDK9 activity in their inducible expression. More experiments will be required to understand the biochemistry of this dose-response phenomenon.
Our transcript-processing analysis indicates that a major regulatory action of RelA Ser276 phosphorylation is in converting the expression of RNA from short unprocessed RNA into fully spliced transcripts. Here, RSV strongly induced expression of total and fully spliced CXCL2/Groβ mRNA in RelA WT-expressing cells, consistent with the presence of phospho-Ser2 RNA Pol II loading on intron 3 and the 3′UTR of the CXCL2/Groβ gene (Fig. 5). This effect is more dramatically seen with dsRNA stimulation, where the proportion of cells and timing of stimulation are synchronized. In contrast, the production of fully spliced CXCL2/Groβ mRNAs in RSV-infected cells expressing the RelA Ser276Ala mutant was significantly reduced. Together these results suggest that transcriptional elongation is a major regulated process in RSV-induced airway inflammation.
In summary, we propose that RelA Ser276 phosphorylation is a key regulatory step in RSV-induced activation of downstream inflammatory genes (schematically diagrammed in Fig. 10). This phosphorylation is coupled to Lys310 acetylation and association with the activated PTEF-b complex. Downstream genes are induced to clear prebound PTEF-b complex, transiently associating with Ac-Lys H4 marking. This process is followed by RelA association and partial reaccumulation of the PTEF-b complex. The production of phospho-Ser2 RNA Pol II switches the polymerase into transcriptional elongation mode, resulting in the high-level expression of fully spliced mRNA transcripts. Selective CKD9 inhibition may be a rational target for modulating the inflammatory pathology associated with RSV-induced LRTI.
This work was supported, in part, by NIH grants AI062885 (A.R.B.) and GM086885 (A.R.B.) and NHLBI contract NIH-NLBI-HHSN268201000037C (A.R.B.). M.K.K. is a trainee of the Keck Center Computational Cancer Biology Training Program.
We thank Ruwen Cui for assistance with production of stable transfectants.
Published ahead of print on 7 September 2011.