HEXIM1 forms a transcriptionally abortive complex with glucocorticoid receptor without involving 7SK RNA and positive transcription elongation factor b

Abstract

The HEXIM1 protein has been shown to form a protein–RNA complex composed of 7SK small nuclear RNA and positive transcription elongation factor b (P-TEFb), which is composed of cyclin-dependent kinase 9 (CDK9) and cyclin T1, and to inhibit the kinase activity of CDK9, thereby suppressing RNA polymerase II-dependent transcriptional elongation. Here, we biochemically demonstrate that HEXIM1 forms a distinct complex with glucocorticoid receptor (GR) without RNA, CDK9, or cyclin T1. HEXIM1, through its arginine-rich nuclear localization signal, directly associates with the ligand-binding domain of GR. Introduction of HEXIM1 short interfering RNA and adenovirus-mediated exogenous expression of HEXIM1 positively and negatively modulated glucocorticoid-responsive gene activation, respectively. In the nucleus, HEXIM1 was shown to localize in a distinct compartment from that of the p160 coactivator transcriptional intermediary factor 2. Overexpression of HEXIM1 decreased ligand-dependent association between GR and transcriptional intermediary factor 2. Antisense-mediated disruption of 7SK blunted the negative effect of HEXIM1 on arylhydrocarbon receptor-dependent transcription but not on GR-mediated one, indicating that a class of transcription factors are direct targets of HEXIM1. These results indicate that HEXIM1 has dual roles in transcriptional regulation: inhibition of transcriptional elongation dependent on 7SK RNA and positive transcription elongation factor b and interference with the sequence-specific transcription factor GR via a direct protein–protein interaction. Moreover, the fact that the central nuclear localization signal of HEXIM1 is essential for both of these actions may argue the crosstalk of these functions.

Transcription is a complex multistep process that relies on highly coordinated actions of a number of cis- and transacting elements (1–3). RNA polymerase II (RNAP II), as the central player in transcription of class II genes, carries out a series of events that include promoter binding, transcription initiation, promoter escape, transcription elongation, and transcription termination. During transcription by RNAP II, phosphorylation of the C-terminal domain of the largest subunit of RNAP II by positive transcription elongation factor b (P-TEFb), which is composed of cyclin-dependent kinase 9 (CDK9) and cyclin T1, is crucial for the transition from the abortive to the productive phase of transcriptional elongation, leading to the generation of full-length RNA transcripts. P-TEFb has been shown to lose its ability to phosphorylate the C-terminal domain when associated with 7SK RNA, a 330-nt small nuclear RNA (4–6).

HEXIM1 was first identified as a protein whose expression is induced in vascular smooth muscle cells (VSMC) in response to hexamethylene bisacetamide (HMBA) treatment (7, 8). The HEXIM1 protein consists of 359 amino acids and is tentatively divided into three regions: an N-terminal proline-rich region (amino acids 1–149), a lysine-arginine rich central nuclear localization signal (NLS)-like region (amino acids 150–177), and a C-terminal acidic region enriched in aspartic and glutamic acid residues (amino acids 178–359) (8). Recently, HEXIM1 was shown to bind 7SK via the NLS, which is considered to be an arginine-rich RNA-binding motif, to form a complex with P-TEFb and to potently and specifically inhibit the kinase and transcriptional activities of P-TEFb (9–12). Roughly half of nuclear P-TEFb in HeLa cells is considered to be sequestered in an inactive state with 7SK and HEXIM1. However, this population of P-TEFb can be rapidly dissociated from 7SK and HEXIM1 upon various treatment of cells, some of which in turn cause the stimulation of C-terminal domain phosphorylation and global increase in RNA and protein synthesis (9, 10, 13, 14). It is therefore of particular importance to address how association of HEXIM1 with 7SK and P-TEFb is controlled. On the contrary, HEXIM1 appears to be in large excess over P-TEFb (11). Moreover, it has recently been reported that HEXIM1 might interact with cellular factors other than 7SK/P-TEFb and perform distinct functions (15). Interestingly, a hypothetical protein gene located in a locus adjacent to that of HEXIM1 has a great sequence similarity to HEXIM1. This protein, named HEXIM2, was recently shown to be expressed in testes and to have an elongation-suppressive function as well (16, 17). For further understanding of the role of HEXIMs in transcriptional regulation and gene expression, it is therefore essential to identify partner factors of HEXIMs.

In the present study, we demonstrate that HEXIM1 forms a distinct complex with the glucocorticoid receptor (GR) in an RNA-independent fashion. GR is a classical member of the nuclear receptor superfamily and has a modular structure consisting of the N-terminal transactivating domain [also termed activation function-1 (AF1)], the central DNA-binding domain (DBD), and the C-terminal ligand-binding domain (LBD), and AF2 at the C terminus (18, 19). It is believed that, upon ligand binding, GR translocates into the nucleus and binds to a target DNA sequence, thereby activating transcription via complex interplay with coactivators, mediators, and target DNA (18, 19). HEXIM1, through its arginine-rich NLS, directly associates with the GR LBD and appears to interfere with productive communication of GR with the p160 coactivator transcriptional intermediary factor 2 (TIF2). These results indicate that HEXIM1 has dual roles in transcriptional regulation by means of a single arginine-rich domain: inhibition of transcriptional elongation dependent on 7SK RNA and P-TEFb and interference with GR by a direct protein–protein interaction.

Materials and Methods

Recombinant DNA, Antibodies, and Cells. Expression plasmids for FLAG-tagged and GST-fused wild-type and mutant HEXIM1 were described in ref. 8. pGEX/hHEXIM1/Δ150–177+SV, pGEX/hHEXIM1/150–180, pCMX/6His-GR, and its mutants were generated by cloning appropriate PCR fragments into pGEX6P-3 (Amersham Biosciences) or pCMX (20). Expression plasmids for double-stranded hairpin RNAs were constructed from pSilencer3.1-H1 neo (Ambion, Austin, TX). The target sequence used was 5′-gaagaagcggcattggaaa-3′ for HEXIM1. Recombinant adenovirus encoding FLAG- and 6His-HEXIM1 (AdCALNL/FHhHEXIM1) was generated by using an Adenovirus Cre/loxP-regulated Expression Vector set (TaKaRa). The following antibodies were used in this study: GR, TIF2 (catalog nos. 611227 and 610985, respectively, BD Transduction Laboratories), CDK9, cyclinT1, CBP (catalog nos. sc-484G, sc-8127, and sc-369, respectively, Santa Cruz Biotechnology), FLAG, α-actinin (catalog nos. F3165 and A5044, respectively, Sigma), arylhydrocarbon receptor (AhR; catalog no. SA-210, Biomol, Plymouth Meeting, PA), and HEXIM1 (8). HeLa, COS7, and HepG2 cells were obtained from RIKEN Cell Bank (Tsukuba, Japan) and cultured in DMEM (Sigma) supplemented with steroid-stripped 10% FBS (21). Human VSMC were obtained from Kurabo (Osaka, Japan) and maintained as described in ref. 8.

Transfection and Luciferase Assay. Cells (1.5 × 10⁵) were transfected in six-well plates with 500 ng of glucocorticoid-responsive luciferase reporter plasmid (GRE-Luc) or xenobiotic-responsive luciferase reporter plasmid using TransIt-LT1 (Panvera, Madison, WI) (22). After treatment with dexamethasone (DEX) (Sigma) or 3-methylchoranthrene (Sigma) for 24 h, cellular luciferase activity was measured by using a luciferase assay system (Promega). For cotransfection assay with plasmids and antisense 7SK deoxyoligonucleotide, Lipofectamine reagent and Plus reagent (Invitrogen) were used. The deoxyoligonucleotide sequences used were 5′-ccttgagagcttgtttggagg-3′ for antisense 7SK and 5′-cgtcgatgtgatgctgtgtga-3′ for scrambled deoxyoligonucleotide (14).

Affinity Purification and Mass Spectrometry. Bacterially expressed GST-HEXIM1 proteins were immobilized onto glutathione Sepharose 4B resin (Amersham Biosciences). HeLa cell nuclear extracts prepared as described in ref. 23 were incubated with the affinity beads for 1.5 h at 4°C and washed with binding buffer (10 mM Hepes, pH 7.9/10% glycerol/50 mM KCl/6 mM MgCl₂/0.1 mM EDTA/0.5 mM DTT/0.5 mM PMSF/0.1% Nonidet P-40). The bound proteins were eluted with elution buffer (10 mM Hepes, pH 7.9/10% glycerol/1 M NaCl/0.1 mM EDTA/0.5 mM DTT/0.5 mM PMSF/0.1% Nonidet P-40) and separated by 5–20% SDS/PAGE, and the amino acid sequence was determined by mass spectrometry as described in ref. 24.

Indirect Immunofluorescence. Cells were plated on coverslips, fixed in 3.7% paraformaldehyde for 20 min at room temperature, permeabilized with 0.1% Triton X-100 in PBS for 15 min, and incubated with blocking buffer (3% BSA/0.1% Triton X-100 in PBS) for 1 h. Cells were incubated for 1 h with primary antibodies and stained with Alexa Fluor 594- or 488-conjugated secondary antibodies (Molecular Probes). Cellular fluorescence was analyzed with Olympus laser scanning confocal microscopy. For digital image analysis, Olympus fluoview software was used as described in ref. 25.

GST Pull-Down Assays and Immunoprecipitations. 6His-GR and its mutants were in vitro-translated with [³⁵S]methionine by using TNT Coupled Reticulocyte Lysate systems (Promega) and incubated for 1.5 h with GST-HEXIM1 immobilized beads. Bound proteins were eluted with sample buffer, separated by SDS/PAGE, and detected by fluorography. For immunoprecipitation, HeLa cell whole-cell extracts (500 μg) were prepared in radio-immunoprecipitation assay buffer (10 mM Tris, pH 7.9/150 mM NaCl/0.1% SDS/1% Triton X-100/1% deoxycholate/1 mM DTT/1 mM PMSF), incubated with 5 μg of anti-HEXIM1 or anti-CDK9 antibody for 1.5 h with or without RNase A, and incubated with protein A-Sepharose (Amersham Biosciences) for 1.5 h. HeLa cell nuclear extracts (200 μg) were prepared as described in ref. 26, incubated with 1 μg of anti-TIF2 antibody for 6 h in immunoprecipitation buffer (20 mM Hepes, pH 7.9/50 mM NaCl/1 mM EDTA/1 mM EGTA/0.1% Nonidet P-40/1 mM DTT/1 mM PMSF/10 μM DEX), and incubated with Protein G-Sepharose (Amersham Biosciences) for 1.5 h. Bound proteins were resolved by SDS/PAGE followed by Western blotting.

RT-PCR and DNA Microarray. Reverse transcription was performed by using SuperScript II (Invitrogen). Specific primers for PCR are described elsewhere (8, 27–29). Total RNA of HepG2 cells was prepared as recommended by Affymetrix. Samples were run by using a GeneChip Human Genome Focus Array (Affymetrix, Santa Clara, CA) comprising 8,746 probe sets representing ≈8,400 human genes. Genes whose expression was significantly detected at any time point were considered to be valid for further analysis, and 5,288 genes were analyzed by using genechip software (Affymetrix) as described in ref. 30.

Results

Complex Formation of HEXIM1 with GR Independent of RNA and P-TEFb. HEXIM1-binding proteins were affinity-purified by using GST-HEXIM1 and identified with mass spectrometry. In the absence of treatment with RNase A, HEXIM1 bound numerous proteins participating in RNA metabolism, probably because of its RNA-binding property (data not shown). In clear contrast, in the presence of an excess of RNase A, we detected a single HEXIM1-binding protein with a molecular weight of 94 kDa (Fig. 1A). Mass spectrometric analysis revealed that the protein is GR. Western blot analysis showed that eluates from HEXIM1 affinity beads contained CDK9 and cyclin T1, both of which were diminished in the presence of RNase A (Fig. 1B). In contrast, RNase A did not affect the interaction between HEXIM1 and GR (Fig. 1B). Immunoprecipitation of HeLa whole-cell extracts with anti-HEXIM1 antibody demonstrated that HEXIM1 forms distinct complexes with P-TEFb and GR in RNA-dependent and -independent manners, respectively (Fig. 1C). Because samples immunoprecipitated from the extracts with anti-CDK9 antibody contained HEXIM1 but not GR (Fig. 1D), we concluded that HEXIM1 forms at least two protein complexes, one involving P-TEFb and 7SK and the other containing GR. We also concluded that the interaction between HEXIM1 and GR does not require RNA.

Fig. 1.

HEXIM1 forms a distinct complex with GR independent of RNA and P-TEFb. (A and B) Identification of GR as a HEXIM1-binding protein. Nuclear extracts (NE) from HeLa cells were incubated with GST or GST-HEXIM1 immobilized beads. Bound fractions were analyzed with SDS/PAGE followed by silver staining (A) and analyzed on Western blots (WB) (B). (C and D) HEXIM1-GR complex does not contain 7SK RNA, CDK9, or cyclin T1. Whole-cell extracts (WCE) (C) or nuclear extracts (D) from HeLa cells were immunoprecipitated with anti-HEXIM1 (C) or anti-CDK9 (D) antibodies in the absence or presence of RNase A. Immunocomplexes were analyzed on Western blots as indicated. IP, immunoprecipitation; FT, flow through; B, bound.

HEXIM1 Negatively Modulates Glucocorticoid-Mediated Gene Expression. Adenovirus-mediated overexpression of HEXIM1 in HepG2 cells decreased DEX-induced mRNA expression of endogenous glucocorticoid-regulated genes without significant alteration in GR protein levels in whole-cell extracts (Fig. 2A). DNA microarray analysis revealed that expression of 2.5% of the 5,288 annotated genes was repressed by the exogenous FLAG-HEXIM1, possibly because of the inhibition of P-TEFb. Concerning glucocorticoid-regulated genes, 135 genes (2.6% of total) were positively regulated by DEX, and the induction responses of 125 genes (92.6% of 135 genes) were canceled in the presence of exogenous FLAG-HEXIM1 (Fig. 2B; see also Table 1, which is published as supporting information on the PNAS web site). Short interfering RNA-mediated knockdown of HEXIM1 did not influence total amounts of GR (data not shown) but enhanced GR-dependent reporter gene expression, which was again canceled by exogenous expression of FLAG-HEXIM1 (Fig. 2C). These results indicate the negative regulatory role of HEXIM1 in GR-mediated transcription. In concert with this idea, treatment of VSMC with HMBA, which resulted in enhancement of HEXIM1 protein expression (8), decreased GR-dependent transcription (Fig. 2D).

Fig. 2.

Effects of HEXIM1 on glucocorticoid-mediated gene expression. (A) HepG2 cells were infected with a recombinant adenovirus carrying loxP-flanked FLAG-HEXIM1 cDNA (AdCALNL/FHhHEXIM1; multiplicity of infection = 30) alone (Cre -) or along with a recombinant adenovirus expressing Cre recombinase (Cre +). After a 2-h treatment with 1 μM DEX, mRNA and protein expression levels were analyzed with RT-PCR and Western blots (WB), respectively. ACE2, angiotensin I converting enzyme 2; ADH1A, alcohol dehydrogenase 1A; SGK2, serum/glucocorticoid-regulated kinase 2. (B) HepG2 cells were infected with these recombinant adenoviruses and cultured in the presence of 1 μM DEX for 1, 2, or 4 h, and total RNA was isolated and subjected to DNA microarray analysis. Genes whose mRNA expression was increased >2-fold by DEX alone were selected, and their relative mRNA expression levels compared with vehicle-treated cells (0 h) were plotted versus time after treatment with DEX. (C) HeLa cells were cotransfected with GRE-Luc, expression plasmids for FLAG-HEXIM1 (FLAG-HEXIM1 +) or empty vector (FLAG-HEXIM1 -), and short interfering RNA against HEXIM1 (siHEXIM1 +) or control vector (siHEXIM1 -), as indicated. After 24 h of treatment with 100 nM DEX, whole-cell extracts were prepared and subjected to luciferase assay and Western blots. (D) Human VSMC were transfected with GRE-Luc and cultured in the presence or absence of 5 mM HMBA for 12 h. After 24 h of treatment with 100 nM DEX, whole-cell extracts were prepared and subjected to luciferase assay and Western blots. Tx, treatment; RLU, relative light units.

7SK-Binding Domain of HEXIM1 Directly Interacts with GR LBD in a Ligand-Independent Manner. Bacterially expressed GST-HEXIM1 and its various mutants were tested for the interaction with [³⁵S]methionine-labeled in vitro-translated full-length GR. Deletion of either N- or C-terminal part of HEXIM1 did not affect binding to GR. However, deletion of the central NLS or its replacement with the simian virus 40 large T-antigen NLS abolished GR binding, and the HEXIM1 NLS alone could strongly bind GR (Fig. 3A), indicating that a mere cluster of basic amino acids acting as an NLS is not sufficient and that the native HEXIM1 NLS region is required for binding GR. Next, to identify a HEXIM1-binding domain of GR, bacterially expressed GST-HEXIM1 was immobilized, and in vitro-translated 6His-GR or its mutants were applied in the absence of ligand. Either full-length GR, AF1-deleted GR, or the LBD alone could still bind HEXIM1. However, neither AF1 nor DBD was trapped by HEXIM1. Moreover, deletion of the LBD from GR abolished HEXIM1 binding ability, indicating critical requirement of the LBD (Fig. 3B). In excellent agreement with this finding, in vitro binding of 7SK to HEXIM1 was specifically inhibited by the GR LBD (Fig. 3C). Because we obtained similar sets of results when DEX was included in the reactions (data not shown), it is suggested that classical agonists are not critical for the GR–HEXIM1 interaction.

Fig. 3.

HEXIM1 directly interacts with GR. (A and B) GST or GST-fused HEXIM1 mutants were immobilized on glutathione Sepharose beads and incubated with in vitro-translated ³⁵S-labeled GR or its mutants, and bound GR was analyzed with SDS/PAGE followed by fluorography. SV40, simian virus 40. (C) Four picomoles of GST or GST-fused HEXIM1 mutants were immobilized on glutathione Sepharose beads and incubated with ³²P-labeled in vitro-transcribed 7SK (lane 1) in the absence or presence of 12 pmol (lanes 5 and 7) or 60 pmol (lanes 6 and 8) of bacterially expressed 6His-GR LBD or DBD as indicated. Bound RNA was analyzed with denaturing PAGE followed by autoradiography. The radioactivity of each band quantified by using Fuji Film BAS2000 image analyzer is shown.

HEXIM1 Interferes with the Interaction Between GR and TIF2. Immunofluorescent analysis revealed that endogenous HEXIM1 constitutively localizes to discrete spots in the nucleus and that ligand-bound GR partially overlaps with HEXIM1 or TIF2 in HeLa cells (Fig. 4A). Because HEXIM1 barely colocalized with TIF2 (Fig. 4A), we hypothesized that HEXIM1 may compete with TIF2 for binding to GR in a DEX-treated cell nucleus. When HEXIM1 was overexpressed, exogenous expression of TIF2 did not efficiently restore transactivational function of GR (Fig. 4B). Moreover, immunoprecipitation of HeLa cell extracts with anti-TIF2 antibody recovered GR but not HEXIM1, and overexpression of HEXIM1 reduced complex formation between GR and TIF2 without significant alteration in the levels of TIF2 (Fig. 4C). These results suggest that increase in cellular HEXIM1 down-modulates GR association with TIF2.

Fig. 4.

HEXIM1 represses the functional interaction between GR and TIF2. (A) HeLa cells were treated with 100 nM DEX or vehicle [0.1% ethanol, Tx(-)] for 1 h and subjected to indirect immunofluorescence. Confocal laser microscopic images of GR, HEXIM1, and TIF2 are shown. (B) COS7 cells were cotransfected with GRE-Luc and expression plasmids for GR, HEXIM1, and TIF2, as indicated. After 24 h of treatment with 100 nM DEX, whole-cell extracts were prepared and subjected to luciferase assay. (C) HeLa cells were infected with AdCALNL/FHhHEXIM1 (multiplicity of infection = 30) alone (Cre -) or along with recombinant adenovirus expressing Cre recombinase (Cre +). After 1 h treatment with 1 μM DEX, nuclear extracts (NE) were prepared and immunoprecipitated (IP) with anti-TIF2 antibody. Immunocomplexes were analyzed on Western blots as indicated.

Transcription Factor Selectivity of HEXIM1. We next tested the effects of HEXIM1 on AhR, because AhR has been shown to rely on P-TEFb and coactivators, including TIF2, for transcriptional regulation (31, 32). AhR does not directly bind HEXIM1 (data not shown). In HepG2 cell nuclear extracts, adenovirus-mediated overexpression of HEXIM1 did not influence protein levels of either GR or AhR (see Fig. 6, which is published as supporting information on the PNAS web site), and respective ligands increased nuclear fractions of GR and AhR. Transient introduction of antisense deoxyoligonucleotide of 7SK (AS7SK) for disruption of endogenous 7SK (14) resulted in a ≈2.5-fold activation of reporter genes driven by either AhR or GR, indicating the liberation of P-TEFb from 7SK and HEXIM1 to enhance P-TEFb kinase activity. Transactivational activity of AhR and GR was suppressed by HEXIM1 in a dose-dependent manner in the absence of AS7SK. As expected, AS7SK-mediated disruption of 7SK resulted in extinction of HEXIM1 inhibition of AhR. In clear contrast, introduction of AS7SK did not affect the inhibitory effect of HEXIM1 on GR-mediated transcription (Fig. 5), strongly supporting the notion that HEXIM1 suppresses GR-mediated transcription not only through the inhibition of P-TEFb but also by a distinct mechanism that does not involve 7SK. Together, we propose that HEXIM1, by directly binding with GR, may play a distinct role in transcriptional regulation.

Fig. 5.

P-TEFb-dependent and -independent suppression of transcription by HEXIM1. HepG2 cells were cotransfected with the xenobiotic responsive reporter plasmid XRE-Luc or GRE-Luc and indicated amounts of HEXIM1 expression plasmid, along with scrambled or 7SK antisense deoxyoligonucleotides, and treated with 1 μM 3-methylchoranthrene (3MC) or 100 nM DEX for 24 h. Whole-cell extracts were prepared and subjected to luciferase assay. Tx, treatment; RLU, relative light units.

Discussion

At this moment, the precise physiological function of HEXIM1 remains unknown. However, genetic disruption of the CLP-1 gene, which is a mouse homologue of HEXIM1, is reported to result in embryonic lethality with marked cardiac hypertrophy (33). Moreover, it is separately shown that hyperactivation of P-TEFb in cardiac myocytes produces cardiac hypertrophy (34). It is therefore likely that HEXIM1 plays an important role in cardiac development and disease pathogenesis. On the other hand, it has been shown that intracellular levels of HEXIM1 are variable and inducible (7, 8). We showed that treatment of VSMC with HMBA resulted in an increase of HEXIM1 expression and a concomitant decrease of GR-activated transcription (8) (see also Fig. 2D). Although the role of glucocorticoids in VSMC remains to be clarified, it is likely that cellular HEXIM1 levels may determine the sensitivity to the hormone in peripheral tissues.

It should be noted that binding of HEXIM1 to 7SK or GR is mediated by the common central arginine-rich motif. The use of the same domain or motif in recognition of nucleic acid and polypeptide molecules is often observed, and two strategies are known to achieve this type of interaction: macromolecular mimicry and induced fit mechanisms (35). The former is observed, for example, in tRNA and polypeptide release factor binding to the same domain of ribosome (36) and in TAFII230 and TATA box element DNA binding to TATA-binding protein (37). The latter is observed in interactions between RNA and arginine-rich motifs (38). The N-terminal part of the HEXIM1 NLS is almost perfectly aligned with the TAR RNA-binding motif of the HIV-1 Tat protein (10). It is known that arginine-rich motifs from different proteins adopt different conformations dependent on the RNA sites recognized and in some cases fold only in the presence of RNA (38). The formation of the Tat/TAR/P-TEFb complex in HIV-infected cells may, therefore, compete and preclude the formation of the inactive HEXIM1/7SK/P-TEFb complex with resultant activation of the HIV-1 transcription. Because GR also targets the NLS of HEXIM1, it is extremely intriguing to speculate that HEXIM1, by differential formation of distinct protein–protein and protein–RNA complexes through its NLS, modulates cellular gene expression and host defense systems. In this line, it is of particular importance to study how HEXIM1 distinguishes GR and 7SK to form distinct modules. Along this line, continued study is necessary to compare structures of TAR/Tat, HEXIM1/7SK, and HEXIM1/GR and to understand differential roles of HEXIMs in transcriptional regulation in more detail.

In conclusion, we showed that HEXIM1, in addition to the HEXIM1/7SK/P-TEFb complex, specifically binds GR to form a transcriptionally inactive complex. Together with the results of AhR, these data suggest that this mode of complex formation may be restricted to a certain class of transcription factors including GR. HEXIM1 might serve as a molecular device to give distinct biological cues: suppression of transcriptional elongation and repression of GR-mediated transcription.