Volume 15, Issue 6, Pages 937-949 (September 2004) Selective Phosphorylations of the SRC-3/AIB1 Coactivator Integrate Genomic Reponses to Multiple Cellular Signaling Pathways  Ray-Chang Wu, Jun Qin, Ping Yi, Jiemin Wong, Sophia Y. Tsai, Ming-Jer Tsai, Bert W. O'Malley  Molecular Cell  Volume 15, Issue 6, Pages 937-949 (September 2004) DOI: 10.1016/j.molcel.2004.08.019

Figure 1 SRC-3 Phosphorylation Sites and Generation of Phosphorylation State-Specific Antibodies (A) A schematic representation of the SRC-3 protein and the known functional domains it contains. Shown below are the four identified phosphopeptides, and the putative phosphoamino acids are indicated by asterisks. (B) Mutations to alanine abolished or reduced SRC-3 reactivity with phospho-specific antibodies. HEK293 cells grown in complete medium were transfected with the indicated SRC-3 constructs. Following immunoprecipitation of the Flag-SRC-3 by anti-Flag antibodies, the samples were separated by SDS-PAGE (8%) and probed with the indicated phosphorylation state-specific antibodies. The membrane was stripped and reprobed with anti-Flag antibodies to control the protein loading. Experiments shown in this study were repeated an average of five times, but never less than three times. (C) Dephosphorylation of SRC-3 abolished the reactivity of antibodies to SRC-3. Purified SRC-3 was dephosphorylated by lambda phosphatase where indicated by (+) and separated along with untreated SRC-3 on a SDS-PAGE (8%). Immunoblotting was performed using the indicated phosphorylation state-specific antibodies or an anti-SRC-3 antibody that recognized both phosphorylated and nonphosphorylated SRC-3 (Wu et al., 2002). Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)

Figure 2 SRC-3 Phosphorylation Was Induced In Vivo (A) Estradiol induced SRC-3 phosphorylation at the six identified sites. HEK293 cells grown in phenol red-free DMEM (supplemented with 5% charcoal-dextran stripped FBS) were cotransfected with expression plasmids for ER and wild-type SRC-3. Forty-eight hours posttransfection, estradiol (10−8 M) was added to the cells for 1 hr where indicated by (+). Flag-SRC-3 was immunoprecipitated by anti-Flag antibodies and separated on an SDS-PAGE (8%). Immunoblotting was performed using the indicated phosphorylation state-specific antibodies. All membranes used for experiments were stripped and reprobed with anti-Flag antibodies to control protein loading (total f-SRC-3); a representative blot is shown. Expression of the ER was confirmed by an ER antibody and is also shown. (B) An experiment as in (A) was performed except that AR was cotransfected with wild-type SRC-3 and ligand (R1881, 10−8 M) for AR was used as indicated by (+). (C) Estradiol induced phosphorylation of endogenous SRC-3 in MCF7 cells. MCF7 cells grown in phenol red-free DMEM (supplemented with 5% charcoal-dextran stripped FBS) were treated with estradiol (10−8 M) for 1 hr where indicated by (+). SRC-3 was first immunoprecipitated by a polyclonal antibody and immunoblotted by the indicated phosphorylation state-specific antibodies. The membrane was stripped and reprobed with SRC-3 antibodies in all experiments to control the protein loading and a representative blot is shown. (D) HEK293 cells were transfected with wild-type Flag-SRC-3 and treated with TNF-α (10 ng/ml) for the indicated times. Afterward, Flag-SRC-3 was immunoprecipitated with anti-Flag antibodies, and phosphorylation of the precipitated Flag-SRC-3 was detected by the phosphorylation state-specific antibodies. Membranes were stripped and reprobed with anti-Flag antibodies to control protein loading. Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)

Figure 3 Mutation of SRC-3 Phosphorylation Sites to Nonphosphorylatable Alanine but Not Glutamic Acid Reduced SRC-3 Activity with ER, AR, and NF-κB (A) (Left panel) Transfection of HeLa cells was carried out as previously described (Lee et al., 2002). In brief, HeLa cells were cotransfected with an ER response element/luciferase reporter gene (150 ng/well), expression plasmids for ER (5 ng/well), and the indicated SRC-3 (150 ng/well). Estradiol (10−8 M) was added to the transfected cells 18 hr posttransfection and incubated for an additional 24 hr. Luciferase activity was determined and normalized to the amount of total input protein. A representative analysis of the levels of the transfected Flag-SRC-3 from the estradiol-treated samples is shown in the inset. (Right panel) Experiments were performed as described for the left panel except that the indicated glutamic acid substituted SRC-3 mutants were used. (B) Transfection of HeLa cells and analysis of luciferase activity were performed as in (A), except that AR and an MMTV-luciferase reporter were used. (C) (Left panel) HeLa cells were cotransfected with an NF-κB response element/luciferase reporter gene (100 ng/well) and expression plasmids for the indicated SRC-3 (150 ng/well). Forty-eight hours posttransfection, TNF-α (10 ng/ml) was added for 4 hr where indicated. Luciferase assays were determined and presented as in Figure 3A. (Right panel) Experiments were performed as in the left panel except that the indicated glutamic acid substituted SRC-3 mutants were used. All results were expressed as relative activity to the activity of sample from no SRC-3 and no estradiol or TNF-α added (set at 1). All error bars indicate SEM of five independent experiments performed in duplicate. Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)

Figure 4 Mutation of SRC-3 Phosphorylation Sites to Alanine Affects Protein-Protein Interactions (A) HEK293 cells grown in phenol red-free DMEM (supplemented with 5% charcoal-dextran stripped FBS) were cotransfected with ER and the indicated SRC-3 constructs; estradiol (10−8 M) was added where indicated by (+). A coimmunoprecipitation (Co-IP) assay was performed using the ER antibodies (top left panel), and coprecipitated SRC-3 was detected by the anti-Flag antibodies (middle left panel). Control expression levels of the SRC-3 constructs are shown (bottom left panel). Interactions between ER and SRC-3 substitutions with glutamic acid were determined by a similar Co-IP assay and were not affected (right panels). (B) HEK293 cells were transfected with the indicated SRC-3 expression plasmids and treated with TNF-α for 1 hr. A Co-IP was performed with anti-Flag antibodies; the coprecipitated NF-κB was detected by NF-κB (p65) antibody. (C) Interactions between CBP and phosphorylation-defective SRC-3 mutants were impaired. HEK293 cells grown in complete medium were cotransfected with HA-CBP and the indicated SRC-3 expression plasmids. A Co-IP assay was performed using anti-HA antibodies (top left panel), the coprecipitated Flag-SRC-3 was detected by anti-Flag antibodies (middle left panel). Control expression levels of Flag-SRC-3 are shown (bottom left panel). Interactions between CBP and SRC-3 molecules substituted with glutamic acid were determined by a similar Co-IP assay and were not affected (right panels). (D) Interactions between CARM1 and phosphorylation-defective SRC-3 mutants were not affected. HEK293 cells grown in complete medium were cotransfected with HA-CARM1 and indicated SRC-3 expression plasmids. Co-IP assays were performed using anti-HA antibodies except for the last lane (anti-Flag) (top left panel); coprecipitated Flag-SRC-3 was detected by anti-Flag antibodies (middle left panel). Control expression levels of Flag-SRC-3 are shown (bottom left panel). Interactions between CARM1 and SRC-3 substituted with glutamic acid were determined by a similar Co-IP assay and were not affected (right panels). Shown are representative results from at least ten experiments with different stimuli and receptors in which every specific Co-IP was carried out two to three times for each interaction point. Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)

Figure 5 Identification of Candidate Kinases (A) Purified GST fusion proteins containing the bHLH/PAS domain, S/T rich region, or CBP-interaction domain (CID) of SRC-3 were used in the in vitro phosphorylation assays with the indicated kinases. Following the reaction, phosphorylation of the recombinant proteins was detected by immunoblotting using the indicated phosphorylation state-specific antibodies. (B) Activation of kinases in MCF7 cells. Estradiol (10−8 M) was added to MCF7 cells grown in phenol red-free medium (supplemented with 5% charcoal-dextran stripped FBS) for 1 hr. Activation of MAPK kinases was determined by immunoblotting with the phosphorylation state-specific antibodies for each of the kinases shown, and activation of IKK was determined by a kinase assay using SRC-3 as substrate. (C) Pretreatment with kinase inhibitors to p38 MAPK (SB203580) and ERK1/2 (PD98059) reduced interactions between ER and SRC-3. HEK293 cells transfected with Flag-SRC-3 were pretreated with kinase inhibitors SB203580 or PD98059 (2 hr) prior to the addition of estradiol (1 hr). A GST pull-down assay was performed by incubation of GST-ER with the indicated lysate. The coprecipitated Flag-SRC-3 was detected by anti-Flag. Total input GST-ER and Flag-SRC-3 are shown. (D) Phosphorylation of the SRC-3 CID enhances interaction with CBP. The GST-CID was phosphorylated with the indicated kinases, and the phosphorylation of target amino acids was confirmed by immunoblotting with phosphorylation state-specific antibodies. Subsequently, a GST pull-down assay was performed by incubating the repurified phosphorylated or nonphosphorylated CID with cell lysate. Coprecipitated amounts of CBP were determined by immunoblotting with CBP antibodies. The input amount of the GST-CID is shown in the bottom panel. Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)

Figure 6 In Vivo Kinase Activities and Phosphorylation of SRC-3 Are Important for SRC-3 Function (A) Knockdown of the kinases by siRNAs was demonstrated by immunoblotting with the indicated antibodies. For protein loading controls, levels of β-actin were determined after stripping and reprobing of the membranes. (B) Knockdown of the kinases affected phosphorylation of SRC-3. In a parallel experiment, cells transfected with the indicated siRNAs were subsequently transfected with Flag-SRC-3 and treated with estradiol (10−8 M) for 1 hr or left untreated. The phosphorylations of the precipitated SRC-3 at the indicated sites were determined by immunoblotting with the corresponding phosphorylation state-specific antibodies. (C) Knockdown of kinase affected SRC-3 activity in ER-mediated transactivation. Forty-eight hours posttransfection of the siRNA to knockdown the expression of the kinases, HeLa cells were further transfected with an ER response element/luciferase reporter gene and expression plasmids for ER and SRC-3. Eighteen hours posttransfection, estradiol (10−8 M) was added where indicated for an additional 24 hr. Luciferase assays were performed and presented as in Figure 3. (D) Knockdown of kinases did not affect SRC-1 activity in ER-mediated transactivation. Knockdown and transfection were performed as in (A) except that SRC-1 was used for cotransfection with ER. Luciferase assays were performed and presented as in Figure 3. (E) Expression of endogenous ER target genes was affected by the knockdown of the kinases in MCF7 cells. Total RNAs prepared from MCF7 cells receiving the indicated siRNA were used for RT-PCR. Reactions were performed with primers for GAPDH as internal controls and primers for pS2 in the same tube. PCR products were separated on a 2% agarose gel, photographed, and quantified by densitometer. In each group, levels of pS2 mRNA expression after induction by estradiol were expressed relative to that of the untreated cells, which was set at 1. All levels were normalized against GAPDH. A representative gel from two independent experiments is shown (inset). Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)

Figure 7 Physiological Functions of SRC-3 (A) IL-6 is a target gene of SRC-3 in response to TNF-α. Total RNAs from primary lymphocytes of age-matched wild-type or SRC-3 null mutant mice, and MEFs of the indicated genotypes, were analyzed by RNase protection assays after TNF-α treatment for 4 hr, indicated by (+). Induction of IL-6 expression by TNF-α is shown to be dependent on SRC-3 in both primary lymphocytes and MEFs. L32 and GAPDH were used as RNA loading controls. (B) Reconstituted SRC-3 expression and induction of IL-6 in SRC-3−/− MEFs. SRC-3−/− MEFs were transduced with retrovirus carrying the indicated SRC-3 constructs and treated with TNF-α for 4 hr as indicated by (+). Expression of SRC-3 was confirmed by Western blots 2 days after infection (data not shown). After TNF-α treatment for 4 hr, total RNA from infected cells was prepared and analyzed by RPA to detect the induction of IL-6. Numbers in parenthesis represent the normalized average density of IL-6 expression from two independent experiments. (C) The activity of each sample from (B) was then quantified and expressed as fold stimulation over that of the control nonstimulated SRC-3+/+ MEFs, which was set as 1. (D) In separate experiments, the oncogegnic potential of SRC-3 was determined using a colony focus formation assay. Focus formation assays were performed using SRC-3−/− MEFs stably expressing RasV12 or RasV12 and the indicated SRC-3 constructs. Results from a representative focus formation assay are shown, and the calculated number of foci from three experiments is summarized in the table. Also shown is the expression of RasV12 and SRC-3 detected by Western blot. Formation of foci was examined and quantified using microscopy. Molecular Cell 2004 15, 937-949DOI: (10.1016/j.molcel.2004.08.019)