BRD4 is an atypical kinase that phosphorylates Serine2 of the RNA Polymerase II carboxy-terminal domain

Abstract

The bromodomain protein, BRD4, has been identified recently as a therapeutic target in acute myeloid leukemia, multiple myeloma, Burkitt’s lymphoma, NUT midline carcinoma, colon cancer, and inflammatory disease; its loss is a prognostic signature for metastatic breast cancer. BRD4 also contributes to regulation of both cell cycle and transcription of oncogenes, HIV, and human papilloma virus (HPV). Despite its role in a broad range of biological processes, the precise molecular mechanism of BRD4 function remains unknown. We report that BRD4 is an atypical kinase that binds to the carboxyl-terminal domain (CTD) of RNA polymerase II and directly phosphorylates its serine 2 (Ser2) sites both in vitro and in vivo under conditions where other CTD kinases are inactive. Phosphorylation of the CTD Ser2 is inhibited in vivo by a BRD4 inhibitor that blocks its binding to chromatin. Our finding that BRD4 is an RNA polymerase II CTD Ser2 kinase implicates it as a regulator of eukaryotic transcription.

BRD4 is a BET family protein that was identified originally as a ubiquitously expressed chromatin adapter that maintains epigenetic memory and regulates cell cycle progression (1). More recently, it has been characterized as a key determinant in acute myeloid leukemia, multiple myeloma, Burkitt’s lymphoma, NUT midline carcinoma, colon cancer, and inflammatory disease (2–7). It suppresses tumor metastasis in mice, and its expression is a prognostic signature of breast cancer survival (8). BRD4 has been proposed to be a structural scaffold that regulates transcription indirectly by recruiting the elongation factor, PTEFb, to the transcription preinitiation complex (9).

Productive transcription depends on the phosphorylation of the carboxyl-terminal domain (CTD) of RNA polymerase II (Pol II). The Pol II CTD consists of consecutive repeats of the heptad Y₁S₂P₃T₄S₅P₆S₇, which vary in number between 26 in yeast and 52 in mammals (10). Phosphorylation of the CTD residues serine 5 (Ser5) and serine 2 (Ser2) is necessary for the recruitment of RNA capping and splicing factors, respectively (11). The order, pattern, and temporal separation of these phosphorylation events ensure an orderly transition from initiation to productive transcription elongation (11–13). CTD Ser5 residues are phosphorylated primarily by the CDK7 kinase component of TFIIH (14), whereas the subsequent Ser2 phosphorylation, currently attributed primarily to the CDK9 subunit of PTEFb and/or CDK12/13, releases Pol II from an early elongation block (15, 16). PTEFb nuclear localization and activation depends on BRD4, which is also known to interact with a variety of other factors, including the mediator complex (1, 17). Despite the preponderance of evidence indicating a vital role for BRD4 in transcription, its precise molecular function is unknown.

Here, we show that BRD4 is an atypical protein kinase that exhibits both auto- and transphosphorylation. Furthermore, BRD4 directly and specifically phosphorylates the Pol II CTD at the Ser2 position, and it is distinct from all other known CTD kinases. Our data, indicate a direct and active role for BRD4 in regulating transcription.

Results

BRD4 Is an Atypical Kinase with Intrinsic Kinase Activity.

BRD4 exhibits sequence homology with human BRD2 that is reported to be an atypical protein kinase (PK) (18, 19). Although typical eukaryotic protein kinases consist of short consensus motifs tightly clustered in a known pattern, atypical PKs (only 40 of which are known) have weakly similar motifs dispersed across the protein (Fig. S1) (19, 20). BRD4 contains eight regions with homology to kinase subdomain motifs scattered across its N-terminal region (Fig. 1). As predicted by this analysis, we found that recombinant BRD4, purified from insect SF9 cells, autophosphorylated (Fig. 1A). MS analysis validated the purity of this rBRD4 (SI Text). Importantly, the intrinsic kinase activity of BRD4 was directly shown in an in-gel kinase assay, where purified BRD4 was run on a denaturing gel, refolded, and subjected to a kinase assay in situ (Fig. 1B). Furthermore, rBRD4 purified from Escherichia coli also autophosphorylated (Fig. S2).

Fig. 1.

BRD4 is an atypical kinase that binds to Pol II CTD. (A) BRD4 autophosphorylation in an in vitro kinase reaction (1 μg rBRD4, γ[³²P]ATP). CE, control extract. (B) BRD4 has intrinsic kinase activity. After denaturing gel electrophoresis and renaturation, BRD4 was subjected to an in-gel kinase reaction with γ[³²P]ATP. (C) BRD4 is an atypical kinase. (Upper) Map of putative kinase subdomains (III–IX), ATP binding site (★), and mutants. (Lower) Kinase assays of 1 μg BRD4 and equimolar amounts of BRD4 mutants. (D) BRD4 interacts with Pol II in vivo. BRD4 was immunoprecipitated from HeLa nuclear extract using α-BRD4 and immunoblotted with α-Pol II. (E) BRD4 interacts directly with Pol II CTD. BRD4 (0.25, 0.5, and 1 μg) was pulled down with 0.5 μg GST-CTD immobilized on GST-Sepharose beads. (F) CTD binds BRD4 kinase subdomains. (Upper) α-GST immunoblot of 0.2 μg GST-CTD (input) recovered by pull-down with 1 μg WT Flag-BRD4 or equimolar amounts of Flag-BRD4 mutants. (Lower) Quantitation of recovered GST-CTD.

BRD4 kinase domains were mapped by deletion of the putative kinase subdomains in the regions of bromodomain-1 (ΔB1; amino acids 1–146), ΔB2 (amino acids 350–599), or both (Fig. 1C). Whereas the kinase activities of the ΔB1 and ΔB2 mutants were modestly reduced, the activity of the ΔB1B2 mutant, where seven of the kinase subdomains were deleted, was significantly reduced. A mutant deleted of all eight known subdomains (ΔN; amino acids 1–699) had no residual activity. Deletion of the C-terminal region (ΔC; amino acids 699–1,400) did not abrogate activity. Thus, BRD4 is an atypical kinase with kinase activity that maps to its N terminus, spanning the region encompassing the predicted kinase subdomains. [We were unable to generate BRD4 kinase-dead point mutations, consistent with attempts reported for other atypical kinases (21, 22), possibly because of redundant kinase subdomains.]

BRD4 Phosphorylates RNA Pol II CTD Ser2 Sites.

Because of the known role of BRD4 in transcription, we next asked whether BRD4 interacted with and phosphorylated RNA Pol II. Indeed, BRD4 coeluted with Pol II in fast protein liquid chromatography (FPLC) fractions of C8166 whole-cell extracts (Fig. S3A) and efficiently coimmunoprecipitated Pol II from HeLa nuclear extracts, suggesting an interaction between BRD4 and Pol II (Fig. 1D). Significantly, in in vitro pull-down assays, a purified GST-tagged CTD segment of the RPB1 subunit of Pol II retained rBRD4 on anti-GST beads (Fig. 1E), showing a direct physical interaction between rBRD4 and the Pol II CTD. Binding of the Pol II CTD mapped to the BRD4 bromodomain region (Fig. 1F).

Remarkably, rBRD4 phosphorylated the Pol II CTD (Fig. 2A). Phosphorylation depended on the kinase domains of BRD4, and its extent was comparable to that of CDK9/PTEFb (Fig. 2A and Fig. S3B). Indeed, apparent K_m values of BRD4 for the Pol II CTD and ATP (85 nM and 10 μM, respectively) (Fig. S3C) are comparable with those values reported for PTEFb (55 nM and 36 μM, respectively) (23, 24). BRD4 kinase activity is not promiscuous; it did not phosphorylate GST (Fig. 2A), TBP, TFIIB, or TFIIH (Fig. S3D). BRD4 can use either ATP or GTP but not other nucleotides as phosphate donors (Fig. S3E). The Pol II CTD consists of consecutive repeats of the heptad Y₁S₂P₃T₄S₅P₆S₇. To identify the sites of BRD4 phosphorylation on the CTD, we used as substrates a series of GST–CTD fusions with 25- or 16-heptad repeats bearing alanine substitutions in every heptad at Ser2, Ser5, Ser7, or Thr4. BRD4 phosphorylated CTD proteins containing the consensus sequence and Ala substitutions at Ser5 and Ser7, but it did not phosphorylate Ser2 substitutions either alone or combined with Ser5 (Fig. 2B). Thr4 substitutions were phosphorylated but less efficiently, suggesting that Thr4 may contribute to the specificity of the kinase. Thus, BRD4 is an atypical serine kinase whose phosphorylation of the CTD requires Ser2.

Fig. 2.

BRD4 phosphorylates the Pol II CTD Ser2. (A) BRD4 phosphorylates CTD. In vitro kinase assay with 0.6 μg BRD4 and 0.25 μg GST-human CTD or GST peptide substrates. Equimolar amounts of PTEFb or Sf9 control extract (CE) were positive and negative kinase controls respectively. BRD4 autophosphorylation is not shown. (B) BRD4 requires Ser2 sites to phosphorylate CTD. GST-CTD (0.2 μg) with 25- or 16-heptad repeats was phosphorylated with 0.5 μg BRD4. The CTD proteins had unmodified sequences (WT) or alanine substitutions at every Thr4, Ser2, Ser5, Ser7, or both Ser2 and Ser5 sites. (C) BRD4 phosphorylates CTD Ser2. Immunoblots with α-Ser2P antibodies (3E10 and H5) after in vitro kinase assays of 0.1 μg GST-CTD with equimolar amounts of BRD4 (0.2 μg), PTEFb, or CE (−). Parallel kinase reactions with γ[³²P]ATP show total incorporation.

To distinguish whether Ser2 is the actual phosphorylation site or a nontarget requirement, BRD4-phosphorylated CTD was analyzed by immunoblotting with two anti-CTD Ser2P antibodies: the highly specific 3E10 antibody (25) and the conventional H5 antibody. The 3E10 antibody detected CTD phosphorylated by BRD4, showing that BRD4 phosphorylates CTD Ser2 directly (Fig. 2C). Surprisingly, the H5 antibody did not detect CTD Ser2 phosphorylation by BRD4, although it detected CTD phosphorylated by PTEFb in the same immunoblot (Fig. 2C). BRD4 did not phosphorylate either CTD Ser5 or Ser7 as assessed by specific Ser5P and Ser7P antibodies (Fig. S4 B and C). Total CTD phosphorylation in parallel kinase assays with radiolabeled ATP showed that BRD4 phosphorylated the CTD as efficiently as PTEFb (Fig. 2C). These results show that BRD4 phosphorylates the Pol II CTD exclusively at Ser2. The failure of the H5 antibody to detect BRD4-phosphorylated CTD Ser2 may explain why it has not been observed earlier. In agreement with recent reports that the H5 antibody has greater affinity for CTD phosphorylated at both Ser2 and Ser5 rather than Ser2 alone (10, 16, 25, 26), we observed that PTEFb phosphorylates both Ser5 and Ser2 (SI Text and Fig. S4A). Thus, BRD4 is a CTD Ser2 kinase that is distinct from PTEFb in its substrate specificity.

BRD4 Is a Distinct Pol II CTD Kinase.

BRD4 can be distinguished from all known mammalian CTD kinases by its pH optimum, resistance to specific kinase inhibitors, or substrate specificity. BRD4 remains active at pH 9.0, whereas PTEFb is almost inactive above pH 7.5 (Fig. S4D). BRD4 phosphorylation of the CTD was not affected by either flavopiridol or roscovitine, which inhibits all known CDK Pol II CTD kinases (CDK1, CDK2, CDK7, CDK8, CDK9, and CDK12/13) (Fig. 3 A and B and Fig. S5). Inhibitors of DNA-PK and MAP (Erk1/2) CTD kinases as well as common broad spectrum kinase inhibitors had no effect on BRD4 (Fig. 3A and Fig. S5A). Conversely, BRD4 kinase activity was sensitive to 5,6-dichloro-1-β-D-ribofuranosylbenzimidazole (DRB), a known CTD kinase inhibitor, and apigenin (Fig. 3A and Fig. S5A). BRD4 and CK2 CTD kinase activities were distinguished by their differential substrate specificities: whereas CK2 phosphorylates the CTD exclusively on three nonconsensus repeats at its C-terminal end (27), BRD4 phosphorylated all consensus CTD heptad repeats (Fig. S6). Thus, BRD4 is a distinct CTD kinase.

Fig. 3.

BRD4 kinase activity is distinct from other Pol II CTD kinases. (A) Effect of kinase inhibitors on BRD4. In vitro kinase assays with 0.25 μg GST-CTD, 0.6 μg BRD4, and inhibitors (1× and 5× IC₅₀) flavopiridol (50 and 250 nM), roscovitine (20 and 100 μM), NU 7441 (DNA-PK inhibitor; 14 and 70 nM), FR 180204 (MEK1/2 inhibitor; 0.4 and 2 μM), and apigenin (50 and 250 μM). (B) BRD4 kinase activity is distinct from PTEFb. GST-CTD (0.2 μg) was phosphorylated by recombinant murine BRD4 (0.2 μg), human BRD4 immunoprecipitated from HeLa nuclear extract, or PTEFb (0.1 μg) with or without 1 μM flavopiridol and 100 μM roscovitine. (C) BRD4 phosphorylates CTD Ser2 during in vitro transcription. In vivo transcription reactions incubated with flavopiridol (1 μM), roscovitine (100 μM), or apigenin (50 μM) were immunoblotted for Ser2P. (D) BRD4 phosphorylation of CTD is affected by prior phosphorylation. (Left) Experimental design. (Right) Autoradiograph of BRD4 phosphorylation of GST-CTD (0.2 μg) unphosphorylated, or prephosphorylated by PTEFb (0.1 μg), TFIIH (0.2 μg), or both.

BRD4 Phosphorylates the Pol II CTD in in Vitro Transcription Reactions.

The finding that BRD4 phosphorylates Pol II CTD suggested that it functions directly in transcription. Indeed, in in vitro transcription reactions performed in the presence of flavopiridol and roscovitine, which inhibit TFIIH and PTEFb, ∼30% of Pol II is still phosphorylated de novo (Fig. S7). Furthermore, the extent of Pol II CTD Ser2 phosphorylation was not affected by the presence of either flavopiridol alone or combined with roscovitine (Fig. 3C). In contrast, apigenin, which maximally inhibits BRD4 kinase activity by 90% (Fig. S5D), reduced the Ser2P level by 90%. These results show that BRD4 phosphorylates RNA Pol II CTD Ser2 in an in vitro transcription reaction.

Existing models for CTD phosphorylation during transcription initiation/elongation posit an initial Ser5 phosphorylation by CDK7/TFIIH followed by phosphorylation of Ser2. We, therefore, asked whether prior phosphorylation of the CTD by either PTEFb or TFIIH would affect subsequent CTD phosphorylation by BRD4. Prephosphorylation of the CTD by TFIIH dramatically enhanced subsequent BRD4 phosphorylation of the CTD (Fig. 3D). In contrast, prephosphorylation of the CTD by PTEFb did not significantly change the magnitude of BRD4-mediated phosphorylation, suggesting that they do not compete for the same Ser2 phosphorylation sites. Thus, although BRD4 does not require Ser5 to phosphorylate Ser2, CTD phosphorylated at Ser5 is a better BRD4 substrate than unphosphorylated CTD. Prephosphorylation of the CTD by both TFIIH and PTEFb markedly reduced subsequent BRD4 phosphorylation, which is consistent with the interpretation that Ser5 phosphorylation of CTD by TFIIH leads to more efficient phosphorylation of Ser2 by PTEFb, leaving only a small fraction of unphosphorylated Ser2 sites open for BRD4 (Fig. 3D). These findings raise the possibility that BRD4 functions after TFIIH to phosphorylate the CTD during transcription initiation.

Pol II CTD Ser2 Phosphorylation Depends on BRD4 in Vivo.

We next examined the role of BRD4 in phosphorylating the CTD Ser2 in vivo. Overexpression of BRD4 in HeLa cells resulted in a more than fourfold increase in Ser2P levels (Fig. 4A). Since BRD4 is known to recruit CDK9/PTEFb into the nucleus (28), we tested a BRD4 mutant, BRD4 FEE-AAA, that is incapable of binding PTEFb (29). Overexpression of BRD4 FEE-AAA also significantly increased Ser2P levels as detected by the 3E10 antibody (Fig. 4A). In contrast, the H5 antibody detected no PTEFb-mediated increase in Ser2,5 phosphorylation. These findings show that BRD4 overexpression is capable of independently increasing CTD Ser2P levels in vivo.

Fig. 4.

BRD4 phosphorylates CTD Ser2 in vivo. (A) BRD4 overexpression increases Ser2P levels independent of PTEFb recruitment; 5 μg human Brd4, a BRD4 mutant incapable of binding PTEFb (FEE-AAA) or vector alone, were transfected into HeLa cells. Whole-cell extracts of these cells were immunoblotted for CTD Ser2P levels (Upper Right). Quantitation of averages ± SEM from two independent experiments (Upper Left). PTEFb components CDK9 and Cyclin T1 do not coimmunoprecipitate with overexpressed Flag-BRD4 mutant FEE-AAA (Lower). (B) Treatment with JQ1 derivative CPI203 inhibits BRD4 phosphorylation of CTD Ser2 in vivo. HeLa cells were transfected as above and treated with increasing concentrations of CPI203. Whole-cell extracts of these cells were immunoblotted for CTD Ser2P levels. Results were normalized to total Pol II levels (Lower). (C) In vivo CTD Ser2P levels are independent of CDK9/PTEFb in stem cells. CTD Ser2P, CDK9, and BRD4 levels in cell extracts from normal trabecular bone cells (NTB), human bone marrow stromal stem cells (BMSc), mature HeLa cells, MEFs or from MEFs transfected with control siRNA, BRD4, or CDK9-specific siRNA (Upper). Quantitation of averages ± SEM from two independent experiments (Lower).

BRD4 recruitment to acetylated chromatin is specifically blocked by the small molecule thienodiazepine analog (S)-tert-butyl 2-(4-(4-chlorophenyl)-2,3,9-trimethyl-6H-thieno[3,2-f][1,2,4]triazolo[4,3-a][1,4]diazepin-6-yl)acetate (JQ1), which competitively occupies the acetyl-binding pockets of BET-bromodomains. JQ1 treatment results in the growth inhibition of acute myelogenous leukemia (AML), lymphoma, and multiple myeloma cells in culture and animals models (2–4). We examined the effect of pharmacologic blocking of the BRD4 bromodomains in vitro and in vivo on BRD4 kinase activity using a highly related derivative of JQ1: (S)-2-(4-(4-chlorophenyl)-2,3,9-trimethyl-6H-thieno[3,2-f][1,2,4]triazolo[4,3-a][1,4]diazepin-6-yl)acetamide (CPI203). Like JQ1, CPI203 inhibits BRD4 in vitro and in cells (Fig. S8A). CPI203 did not affect BRD4 kinase activity in vitro (Fig. S8B). In contrast, in vivo, specific Ser2 phosphorylation by either endogenous BRD4 or exogenous BRD4 FEE-AAA was markedly decreased by treatment with the inhibitor (Fig. 4B). Thus, BRD4-mediated Pol II CTD Ser2 phosphorylation in vivo is suppressed by specific inhibition of BRD4 recruitment to chromatin.

To further investigate the role BRD4 plays in vivo, we quantified CTD Ser2P levels in adult stem cells that have minimal CDK9 expression (30). Although BRD4 levels were comparable with their mature cell counterparts, both human bone marrow stromal stem cells and murine embryonic fibroblasts (MEFs) have markedly reduced levels of CDK9. Remarkably, the total levels of Ser2 phosphorylation, as measured by the 3E10 antibody, were as great in the immature cells as in the mature cell controls (Fig. 4C). Furthermore, depletion of BRD4 by siRNA knockdown in MEFs dramatically reduced the level of Ser2P, which was detected by the 3E10 antibody, and suppressed global RNA synthesis (Fig. 4C and Fig. S9). In marked contrast, depletion of CDK9 only minimally affected the 3E10-detected Ser2P levels. As expected, because BRD4 recruits PTEFb, both BRD4 and CDK9 depletions reduced H5-detected Ser2,5P levels. [CTD phosphorylation by the only other known CTD Ser2 kinase, CDK12/13, is not recognized by the 3E10 antibody (16).] Taken together, these results show that BRD4 mediates CTD Ser2 phosphorylation in vivo.

Discussion

The present studies show that the bromodomain protein BRD4 is an atypical kinase that binds to and phosphorylates the Pol II CTD Ser2 in vitro and in vivo. These findings suggest a previously unappreciated role for BRD4 in transcription. Although PTEFb, which is recruited by BRD4, was thought to be the predominant CTD Ser2 kinase thus far, we now find that BRD4 is equally capable of phosphorylating the CTD Ser2. As we show, CTD Ser2 is phosphorylated both under conditions where PTEFb is not recruited to promoters and in stem cell lines deficient in PTEFb (Fig. 4). A recent report that BRD4 activates transcription of a subset of genes independent of PTEFb supports our conclusions (31). Our findings may provide a mechanistic basis for several functional studies that showed that loss of BRD4 causes transcription termination and embryonic lethality (28, 32, 33). Our future studies will be directed at testing our working model of transcription initiation, in which the initial phosphorylation of the Pol II CTD Ser2 is mediated by BRD4 during the transition from transcription initiation to elongation and only subsequently by PTEFb during elongation.

Materials and Methods

Cell Culture.

Bone marrow stromal stem cells, processed from marrow aspirates of normal human adult volunteers (20–35 y), were purchased from Poietic Technologies. About 1 × 10⁶ STRO-1–positive cells per sample were sorted using a FACStar^PLUS (Becton Dickinson) and then seeded for clone formation into six-well plates containing the α-modification of Eagle's medium supplemented with 15% (vol/vol) FCS. Colonies containing proliferative cells were pooled and grown in a 175-cm² flask. Whole-cell extracts were prepared from the cells after 4 d using the tissue extraction reagent (Invitrogen).

Reagents.

Antibodies used for either immunoprecipitation or immunoblotting were anti-RNA Pol II: mAb 8WG16 (Santa Cruz), anti-BRD4: H-250 (Santa Cruz), anti-Pol II phosphoSer2: mAb 3E10 (Millipore) or mAb H5 (Covance), anti-Pol II phosphoSer5: mAb 3E8 (Millipore), anti-Pol II phosphoSer7: mAb 4E18 (Millipore), anti-CDK9: mAb D-7 (Santa Cruz), anticyclin T1 (Santa Cruz), or antitubulin (Abcam). Kinase inhibitors NU 7441 and FR 180204 from Tocris and flavopiridol, roscovitine, apigenin, and genestein (Sigma) were dissolved in DMSO to 10 mM. DRB (Sigma) was dissolved in ethanol to 10 mM. Staurosporin (Ascent Scientific) was dissolved in DMSO to 1 mM. Stock solutions of all kinase inhibitors were stored at −80 °C. The JQ1 derivative CPI203 was prepared by liberation of the protected carboxylic acid in JQ1 followed by primary amide bond formation (US Patent no. US005712274A).

Plasmid Constructs.

Murine Flag-BRD4 WT and ΔB1, ΔB2, ΔB1B2, and ΔC mutants are as described before (34). The BRD4 ΔN (amino acids 699–1,400) was generated as a dropout mutant in the pAcHLT-C baculovirus transfer vector from the Flag- BRD4 WT. The plasmid bearing GST-tagged human BRD4 in the pGEX-6P-1 bacterial expression vector was obtained from Addgene (plasmid 14447). The human Flag-tagged BRD4 WT and FAA-AAA mutant in the pCDNA mammalian vector were a gift from E. Verdin (Gladstone Institute of Virology, University of California, San Francisco, CA). The human GST-CTD construct in pGEX vector was a gift from J. N. Brady (National Cancer Institute, National Institutes of Health, Bethesda, MD).

Protein Purification.

Flag-tagged BRD4 and associated mutant proteins were purified as described earlier (34). The Flag peptide was eliminated on a micron column (Millipore), and proteins were recovered in 20 mM Hepes, pH 7.9, 100 mM KCl, 0.2 mM EDTA, and 20% vol/vol glycerol. A control extract (CE) from uninfected Sf9 cells was prepared similarly. In E. coli cells (BL21 Rosetta-gami strain; Novagen), GST-CTD, GST-BRD4, and GST protein were induced using 0.8 mM isopropyl_β-D-1-thiogalactopyranoside (IPTG) overnight at 18 °C and purified using GST Sepharose beads (Amersham). The proteins were concentrated on a microcon column (Millipore) and recovered in 20 mM Hepes, pH 7.9, 100 mM KCl, 0.2 mM EDTA, and 20% vol/vol glycerol. The GST-CTD 25- and 16-heptad fusion proteins were made from plasmid constructs as described earlier (35). Purified PTEFb was procured from Millipore, whereas purified TFIIH, TFIIB, and TBP were purchased from ProteinOne. Purified CK2 enzyme was purchased from New England Biolabs.

FPLC Fractionation.

FPLC fractions of whole-cell extracts from C8166 T cells were obtained as previously described (13). The proteins in the fractions were separated on SDS/PAGE gels and analyzed by immunoblotting with antibodies specific for RNA Pol II (8WG16) and BRD4.

In Vitro Kinase Assays.

In vitro kinase assays with BRD4, PTEFb, and TFIIH were performed in 20 μL 50 mM Tris, pH 7.5, 5 mM DTT, 5 mM MnCl₂, and 5 mM MgCl₂ with 10 μCi γ³²P ATP (6,000 Ci/mM) and/or 40 μM ATP where indicated. The kinase reactions were incubated for 1 h at 30 °C, and then, the proteins were resolved by SDS/PAGE; the extent of phosphorylation was quantitated by a phosphorimager. For kinetic measurements, the quantitated values were plotted on a Lineweaver–Burke plot to calculate V_max and K_m values. When CTD phosphorylation was determined by immunoblotting with specific antibodies as indicated in Figs. 2–4, kinase assays were performed with unlabeled (cold) ATP. When kinase inhibitors were used, appropriate dilutions of the inhibitor were added at the start of the kinase reaction. Mock-treated kinase reactions were treated with equivalent volumes of DMSO.

In-Gel Kinase Assay.

The in-gel kinase assay was done as described earlier (36) with minor modifications. Briefly, three sets of 4 μg purified BRD4 and control BSA were run on an 8% SDS/PAGE gel, which was then cut into strips with one set of proteins each. One set was denatured with 6 M guanidine hydrochloride for 1.5 h and renatured for 18 h; then, the gel strip was soaked in kinase buffer supplemented with 20 Ci/mL γ³²P ATP for 1.5 h. The gel was washed stringently for 1 h and dried, and phosphorylated proteins were detected by a phosphorimager. The autoradiograph was aligned with the other two gel strips with BRD4 and BSA that were Coomassie-stained or Western-blotted with anti-BRD4 antibody.

CPI203 and JQ1 Comparative Assays.

The BRD4 α-screen assay was done as described previously (37). The MYC cellular assay was also done as previously described (4) with minor changes summarized in SI Text. For the IL-6 release assay, THP-1 monocytic leukemia cells were pretreated with the inhibitors for 2 h before the addition of LPS. IL-6 release was detected by ELISA 16 h after the addition of LPS. IC₅₀ values were calculated using a 10-point serial dilution of the BET inhibitor in these assays.

Immunoprecipitation and Coimmunoprecipitation.

To coimmunoprecipitate Pol II with BRD4 from HeLa nuclear extracts, magnetic beads (Protein A; Dynabeads) were coated with 5 μg anti-BRD4 antibody and incubated with 25 μg HeLa nuclear extract (Promega) for 3 h at 4 °C. The beads were then washed three times with 50 mM Tris, pH 8.0, 200 mM NaCl, and 0.2% Nonidet P-40. Bound proteins were separated on SDS/PAGE gels and immunoblotted with antibodies against Pol II and BRD4. To test for BRD4 binding to Pol II CTD, 0.5 μg purified GST- CTD were immobilized on GST agarose beads and incubated with 0.25, 0.5, and 1 μg purified BRD4 for 3 h at 4 °C. Beads were washed two times with 50 mM Tris, pH 8.0, 150 mM NaCl, and 0.2% Nonidet P-40, and bound proteins were immunoblotted. CTD binding to BRD4 mutants was detected by immobilizing equimolar amounts of BRD4 proteins on anti-Flag M2 agarose beads and incubating with 200 ng GST-CTD for 2 h at 4 °C. The beads were washed two times with 50 mM Tris, pH 8.0, 150 mM NaCl, and 0.2% Nonidet P-40, and bound GST-CTD was immunoblotted with anti-GST antibody. All immunoblot analyses were performed using the Odyssey infrared scanner and secondary antibodies from Li-Cor.

In Vitro Transcription.

In vitro transcription reactions were done by preincubating 100 μg HeLa nuclear extract (Promega) with or without specific kinase inhibitors as indicated in Figs. 2–4 for 15 min at 23 °C; 2 μg DNA, MHC class I promoter construct and −313CAT described earlier (13) with 0.8 mM rNTPs, 6 mM MgCl₂, and 250 μM trichostatin A were then added, and reactions were incubated at 23 °C for 30 min. The reactions were run on an SDS/PAGE gel to analyze the phosphorylation status of RNA Pol II through immunoblotting. Alternatively, the reactions were spiked with 10 μCi γ³²P ATP (6,000 Ci/mM) when the phosphorylation status of RNA Pol II was analyzed through a phosphorimager.

Transient Transfections.

Transient transfections of Flag-hBRD4 and Flag-hBRD4 FEE-AAA mutant constructs were done in HeLa cells using Lipofectamine (Invitrogen). Where indicated, cells were treated with CPI203 for 48 h. Whole-cell extracts were made from the cells and analyzed through Flag immunoprecipitation and immunoblotting for PTEFb binding and RNA Pol II phosphorylation. BRD4 and CDK9 knockdown was done by lipofectamine-mediated transfection of 80 nM siRNA (ON-TARGETplus smartpool human BRD4 and CDK9 siRNA or siGENOME nontargeting siRNA pool; Dharmacon) into MEFs. Time required for maximal knockdown was empirically determined. Whole-cell extracts were made from the cells 72 h after siRNA transfection and used for immunoblotting.

Immunodetection of Nascent RNA.

Seventy-two hours after siRNA transfection, MEFs were pulsed with 2 mM 5-fluorouridine (Sigma) for 12 min. Treated cells were fixed with 2% paraformaldehyde in PBS for 5 min and permeabilized with 0.5% Triton X-100 for 5 min. The incorporation of 5-fluorouridine into nascent RNA was detected with an antibody against halogenated UTP (1:400, anti-BrdU clone BU-33; Sigma) and a Texas Red-conjugated secondary antibody (Alexa 568). The cells were costained with 0.1 μM DAPI, and images recorded through a confocal microscope.