A conserved protein motif is required for full modulatory activity of negative elongation factor subunits NELF-A and NELF-B in modifying glucocorticoid receptor-regulated gene induction properties

Abstract

NELF-B is a BRCA1-interacting protein and subunit (with NELF-A, -C/D, and -E) of the human negative elongation factor (NELF) complex, which participates in RNA polymerase II pausing shortly after transcription initiation, especially for synchronized gene expression. We now report new activities of NELF-B and other NELF complex subunits, which are to attenuate glucocorticoid receptor (GR)-mediated gene induction, reduce the partial agonist activity of an antagonist, and increase the EC50 of an agonist during nonsynchronized expression of exogenous and endogenous reporters. Stable knockdown of endogenous NELF-B has the opposite effects on an exogenous gene. The GR ligand-binding domain suffices for these biological responses. ChIP assays reveal that NELF-B diminishes GR recruitment to promoter regions of two endogenous genes. Using a new competition assay, NELF-A and NELF-B are each shown to act independently as competitive decelerators at two steps after the site of GR action and before or at the site of reporter gene activity. A common motif in each NELF was identified that is required for full activity of both NELF-A and NELF-B. These studies allow us to position the actions of two new modulators of GR-regulated transactivation, NELF-A and NELF-B, relative to other factors in the overall gene induction sequence.

Keywords: Cofactor Action; Competitive Decelerator; EC50; Gene Induction; Gene Transcription; Glucocorticoid Receptor; Mathematical Modeling; NELF-A and NELF-B; Steroid Hormone; Transcription Factors.

FIGURE 1.

Characteristics of NELF-B modulation of GR induction properties of exogenous reporter. A and B, effect of exogenous NELF-B on A_max, -fold induction, PAA, and EC₅₀ of transfected reporter in U2OS cells. Triplicate wells of cells were transiently transfected with GREtkLUC reporter (100 ng in A, 12 ng in B) and the indicated amounts of chimeric NELF-B plasmid (A) or full-length NELF-B plasmid (B), without or with GR plasmid (0.5 ng in A, 0.2 ng in B), induced by steroid, and analyzed as described under “Experimental Procedures.” *, p < 0.05, **, p < 0.01 versus no NELF-B. y axis numbers are the quantitative values for the categories of the x axis (average ± S.E., n = 3 and 4 independent experiments in A and B, respectively). DM, Dex-Mes. C–E, modulation of GR activity with exogenous reporter upon reducing endogenous NELF-B. C, levels of NELF-B and control (LAD1) mRNAs in T47D cells stably transfected with shEGFP or shNELF-B RNA. Duplicate mRNA samples were prepared as described under “Experimental Procedures” and separated on agarose gels. The quantitative abundance was determined by qRT-PCR of the same original samples. D, representative dose-response curves for Dex induction of GREtkLUC in T47D cells stably transfected with shEGFP or shNELF-B RNA after transient transfection without or with GR plasmid. E, -fold changes in GR induction parameters from 5 independent experiments (± S.E.) in T47D cells stably transfected with shEGFP or shNELF-B RNA after transient transfection without or with GR plasmid. Thick horizontal dashed line represents no difference between shEGFP and shNELF-B cells. *, p < 0.02 versus no GR. F, graphic representation of domains of full-length rat GR and GR C terminus fused to GAL4 DBD. Cross hatching = GR DBD; shading = GAL4 DBD; striped = GR LBD. G, intracellular levels of NELF-B affect the EC₅₀ of gene induction by GAL/GR525C. Representative dose-response curves for Dex induction of GREtkLUC in T47D cells stably transfected with shEGFP or shNELF-B RNA after transient transfection without or with GAL/GR525C plasmid are shown. H, -fold changes in GR induction parameters from 7 independent experiments (± S.E.) in T47D cells stably transfected with shEGFP or shNELF-B RNA after transient transfection without or with GAL/GR525C plasmid. Thick horizontal dashed line represents no difference between shEGFP and shNELF-B cells. *, p < 0.05, ***, p < 0.0005 versus -fold difference = 1. I, binding of NELF-B to overexpressed GAL/GR525C. Cytosolic extracts of Cos-7 cells that had been transiently transfected with GAL/GR525C plus FLAG/NELF-B were treated with sodium sulfate, with or without EtOH (E) or Dex (D). Receptors were then analyzed for co-immunoprecipitation (IP) with FLAG/NELF-B using anti-FLAG antibody as described under “Experimental Procedures.” W, Western blot. J, NELF-B binds only to activated endogenously expressed GRs. A cytosolic solution of U2OS.rGR cells with endogenous GR and transiently transfected FLAG-tagged NELF-B was split in half, treated with sodium molybdate (to block activation) or sodium sulfate, and then incubated with EtOH (E), Dex (D), or RU486 (R) before being immunoprecipitated (IP) with anti-FLAG antibody as described under “Experimental Procedures.” W, Western blot.

FIGURE 2.

NELF-B modulates GR induction parameters for endogenous genes. A, NELF-B increases the EC₅₀ for Dex induction of IGFBP1 and IP6K3. Representative dose-response curves for Dex induction of IGFBP1 or IP6K3 mRNA in U2OS cells after transfection with GR and NELF-B or control plasmid (hSA/pSG5) were determined by qRT-PCR as described under “Experimental Procedures.” B–E, summary of effect of NELF-B versus control plasmid transfection on -fold induction by saturating concentrations of Dex, PAA with 1 μm Dex-Mes (DM), and EC₅₀ (when determinable) for IGFBP1 (B; n = 7), IP6K3 (C; n = 5), GILZ (D; n = 6), and LAD1 (E; n = 6). Error bars represent S.E. *, p < 0.05, **, p < 0.005, ***, p < 0.0005 versus no NELF-B.

FIGURE 3.

Binding of GR and NELF-B to elements of IP6K3 and IGFBP1 genes and GREtkLUC reporter. A, the intron 1 region of IP6K3 gene contains a GR-inducible enhancer. A summary of effects of exogenous NELF-B on A_max, -fold induction, PAA, and EC₅₀ of transfected reporter (IP6K3 intron 1 fused upstream of tkLUC plasmid) in CV-1 cells from 5 independent experiments (± S.E., *, p < 0.05) is shown. B, GR inducibility of IP6K3 intron 1 fragment is due to a GRE. The average A_max (n = 4, ± S.E., *, p < 0.05 versus no NELF-B) with 1 μm Dex for the empty tkLUC reporter, the IP6K3 intron 1/tkLUC reporter with a mutation in the putative GRE, and the wild type IP6K3 intron 1/tkLUC reporter was determined in transiently transfected CV-1 cells. C–F, recruitment control (C) of GR (D) and of FLAG/NELF-B (E) to regions of endogenous IP6K3 gene in U2OS cells. ChIP assays were performed as described under “Experimental Procedures.” The recruitment of FLAG/NELF-B above background (F) was calculated as the signal with anti-FLAG antibody ± added FLAG/NELF-B. The average values from 5 (GR) and 4 (NELF-B) experiments are plotted (± S.E., *, p < 0.05 versus no NELF-B in D, versus 0 in F). G and H, recruitment of GR (G) and FLAG/NELF-B (H) to regions of exogenous GREtkLUC reporter in CV-1 cells. The average recruitment of GR (above IgG controls) and FLAG/NELF-B from 5 and 3 independent ChIP assays, respectively, is plotted (± S.E., **, p < 0.008 versus no NELF-B). I, biological activity of different IGFBP1 GRE sequences. -Fold induction of luciferase activity by 1 μm Dex in U2OS cells transiently transfected by GR plasmid and no reporter (None), tkLUC (Promoter), tkLUC with upstream fusions of either IGFBP1 intron 1 sequence (data not shown) (Intron), or the same IGFBP1 intron 1 sequence with the mutant palindromic GRE (wt = AGAACATAATGTGAG; mutant = AGAAtATAAacTGAG). *, p < 0.05, **, p < 0.005 versus intron (n = 4; n = 2 for None) is shown. J and K, ChIP assays of endogenous IGFBP1 gene in U2OS cells. Experiments were conducted as for C and D with primers directed toward the two previously reported GREs (IG2 and IG1, which are just 5′ of the TSS) (47), the putative pause site at +50 bp, and the intron 1 GRE. *, p < 0.05 versus no NELF-B.

FIGURE 4.

Levels of NELF subunits in different mouse tissues. cDNA was prepared from the indicated mouse tissues. The amount of each NELF subunit in each tissue was then determined by qRT-PCR using the primers in Table 3. Similar results are seen on the BioGPS website. Adrenal G, adrenal gland; error bars represent S.D. of triplicates.

FIGURE 5.

Modulatory activity of NELF subunits. A, all NELF subunits reduce the EC₅₀ of gene induction by GR. Representative dose-response curves for Dex induction of GREtkLUC by each NELF subunit in transiently transfected U2OS cells are shown. B, the average -fold changes in GR induction parameters from 4 independent experiments (± S.E.) in U2OS cells treated as in panel B with vector alone (blank), GR and vector (GR), or GR and one NELF subunit. *, p < 0.05, **, p < 0.005, ***, p < 0.0005 versus GR alone. DM, Dex-Mes. C, overexpression of one NELF subunit does not affect the level of the other subunits. Western blots of NELF subunits in U2OS cytosols after transient transfection of each NELF subunit as described under “Experimental Procedures” are shown. FLAG/NELF-E was detected with anti-FLAG antibody. D, levels of GR, TIF2, NELF-B, and NELF-A protein are not affected during overexpression of various factors. U2OS cells were transiently transfected with different combinations of factors under the conditions of the competition assays of Fig. 6. Cytosolic extracts were then analyzed by Western blotting to show that only the level of the transfected protein changes, with β-actin being used as an internal control.

FIGURE 6.

Full-length NELF-B and NELF-A act at two steps in competition assays with GREtkLUC and with GR. U2OS cells were transiently transfected with the indicated varying amounts of GREtkLUC reporter, NELF-B, or NELF-A plasmid or GR plasmid as described under “Experimental Procedures.” The A_max and EC₅₀ were determined as described for Fig. 5D. Plots, after correcting for nonlinear expression of chimeric NELF-B as described under “Experimental Procedures,” of 1/EC₅₀ versus GREtkLUC for NELF-B (A), EC₅₀/100 × A_max versus NELF-B (B, D, and F) or versus NELF-A (H), and A_max/100 × EC₅₀ versus GR with NELF-B (C and E) and with NELF-A (G) were made and interpreted as described under “Results.” Similar results were seen in 3 or 4 additional independent experiments.

FIGURE 7.

NELF-B and NELF-A function additively as competitive decelerators at two steps with different cofactors. Competition assays were conducted as in Fig. 6 with GR, GREtkLUC, NELF-A, full-length wild type NELF-B, and/or TIF2. A, NELF-B reverses effects of TIF2 on GR induction parameters. The average -fold changes in GR induction parameters from 5 independent experiments (± S.E.) in U2OS cells transiently transfected with GR ± TIF2 plasmid and the indicated amounts (ng) of NELF-B plasmid are plotted in the same manner as Fig. 1A. *, p < 0.05, **, p < 0.005, ***, p < 0.0005 versus GR plus TIF2. B–E, graphical analysis of competition of NELF-B or NELF-A, with TIF2 for modulation of GR induction parameters. Experiments were performed with the indicated amounts of NELF-B (after correction for nonlinear expression) or NELF-A and TIF2 (after correction for nonlinear expression) and then analyzed as above in Fig. 6. The “a versus b plots” for A_max/EC₅₀ versus TIF2 graphs such as in panel C (see “Experimental Procedures”), which yield an unbiased determination of a common intersection point, gave average values for the x and y axis coordinates of the intersection point as −12.8 ± 6.4 and 0.73 ± 1.20 (S.D., n = 4), respectively. Western blots determined that endogenous TIF2 protein is equivalent to 4.1 ng of TIF2 plasmid, so −4.1 ng TIF2 plasmid is where there is no TIF2 in the cells, and the intersection point of the A_max/EC₅₀ versus TIF2 graphs is less than this. F and G, competition assay with NELF-A and NELF-B. A graph of EC₅₀/A_max versus corrected (cor'd) amount of NELF-B with indicated amounts of NELF-A (F) was constructed as for panel E. The graph of EC₅₀/A_max versus the combined amounts of NELF-A and NELF-B (G) was compiled as described under “Experimental Procedures.” Similar results were seen in 2 additional independent experiments.

FIGURE 8.

Conserved NELF motif is required for full activity of NELF-A and NELF-B. A, putative NELF motif. A motif identified by manual scanning of regions of NELF-B against the other NELF-proteins is aligned for each NELF. Numbers in parentheses indicate the amino acid residues shown. Shaded residues indicate ≥75% similarity ($ = charged amino acid). B, spatial distribution of NELF motif residues in an α-helical structure. C–F, comparison of plots of EC₅₀/A_max for wtNELF-A (C and E) or 4mtNELF-A (D and F) competition assays with GR (C and D) and with GREtkLUC (E and F) performed as in Fig. 6D. G and H, potency of NELF-B is reduced by mutations in NELF motif. Average potency of wtNELF-B and 4mtNELF-B (± S.E., n = 16) in reducing A_max/EC₅₀ in competition assays with GR was determined as described under “Experimental Procedures” and displayed as a vertical line, with shaded area indicating ± S.E.