A role for cofactor-cofactor and cofactor-histone interactions in targeting p300, SWI/SNF and Mediator for transcription

Abstract

Transcriptional activation from chromatin by nuclear receptors (NRs) requires multiple cofactors including CBP/p300, SWI/SNF and Mediator. How NRs recruit these multiple cofactors is not clear. Here we show that activation by androgen receptor and thyroid hormone receptor is associated with the promoter targeting of SRC family members, p300, SWI/SNF and the Mediator complex. We show that recruitment of SWI/SNF leads to chromatin remodeling with altered DNA topology, and that both SWI/SNF and p300 histone acetylase activity are required for hormone-dependent activation. Importantly, we show that both the SWI/SNF and Mediator complexes can be targeted to chromatin by p300, which itself is recruited through interaction with SRC coactivators. Furthermore, histone acetylation by CBP/p300 facilitates the recruitment of SWI/SNF and Mediator. Thus, our data indicate that multiple cofactors required for activation are not all recruited through their direct interactions with NRs and underscore a role of cofactor-cofactor interaction and histone modification in coordinating the recruitment of multiple cofactors.

Fig. 1. Hormone-dependent activation by AR and TR is associated with chromatin remodeling with changes in DNA topology, histone acetylation and specific targeting of SRC-3, p300, BRG1 and TRAP220. (A) R1881-stimulated activation by AR from a MMTV-LTR reporter assembled into chromatin in Xenopus oocytes. The Ctrl represents the primer extension product from the oocyte storage histone H4 RNA. The expression of AR was detected by western blot using a FLAG-tag-specific antibody (M2; Sigma). (B) The same groups of oocytes as in (A) were analyzed for chromatin remodeling by a DNA topology assay. NC, nicked DNA. Faster migration in lane 4 in comparison with the control DNA indicates the loss of negative superhelical turns. (C) ChIP assays reveal histone acetylation and recruitment of different cofactors and RNA Pol II during R1881-stimulated activation by AR. The injection and treatment of oocytes were as in (A). Input DNA used for PCR was 5% of the total input DNA. Negative control (–) was performed without addition of antibody. (D) T3-dependent activation by TR is also associated with histone acetylation and promoter targeting of different cofactors and RNA Pol II. ChIP assays were performed with oocytes injected with 4×UAS.TRβA reporter and mRNA encoding Gal4–TR and treated with 50 nM T3 as described.

Fig. 2. Targeting BRG1 to chromatin is sufficient for chromatin remodeling with changes in DNA topology, and a functional BRG1 is required for AR activation. (A) Histone acetylation is not the reason for changes in DNA topology. (B) Targeting of a wild type BRG1 but not an ATPase-deficient mutant to chromatin resulted in chromatin remodeling with change in topology. Groups of oocytes were injected with mRNA encoding Gal4–BRG1 or Gal4–BRG1m (K785R) and 4×UAS.TRβA reporter as indicated. After overnight incubation, the oocytes were used for analysis of DNA topology by supercoiling assay (left panel) and protein expression by western blot (bottom panel). The right panel shows that Gal4–BRG1 did not activate transcription in comparison with Gal4–TR in the presence of T3. (C) Expression of wild-type BRG1 but not the mutant stimulated AR activation in Xenopus oocytes. The expression of AR and wild-type or mutant BRG1 was detected by western analysis using a FLAG-specific antibody (M2, Sigma). (D) R1881-dependent activation by AR was impaired in SW13 cells and could be restored by co-transfection of BRG1. SW13 cells were transfected with a MMTV-LTR-driven luciferase reporter and AR or BRG1 expression construct as indicated. All experiments were carried out in triplicate and repeated twice; error bars represent the SEM.

Fig. 3. Inhibition of CBP/p300 HAT activity or blocking SRC–CBP/p300 interaction abrogates hormone-dependent activation. (A) Endogenous CBP/p300 HAT activity is required for AR activation. The requirement of CBP/p300 HAT activity for AR activation was tested using a CBP/p300 selective HAT inhibitor, Lys-CoA. The oocytes were first injected with AR mRNA and then injected with ssDNA of the MMTV-LTR reporter (50 ng/µl, 18.4 nl/oocyte) containing 11.1 µM (+), 33.3 µM (++) or 100 µM (+++) of Lys-CoA. After overnight incubation with or without 50 nM R1881, the levels of transcription from the MMTV-LTR reporter were detected by primer extension assay and the level of AR protein is shown at the bottom panel by western blot. (B) ChIP assay revealed that Lys-CoA blocked hormone-dependent histone acetylation induced by AR. The groups of oocytes were injected with AR mRNA and ssDNA of the MMTV-LTR reporter with or without 100 µM of Lys-CoA as indicated. (C) Transcriptional activation by liganded TR also requires CBP/p300 HAT activity. (D) Expression of a CBP fragment (amino acids 2057–2170) containing the SRC interaction domain (SID) is sufficient to block AR activation.

Fig. 4. The SWI/SNF complex is unlikely to be recruited directly by liganded TR but can be targeted to chromatin through a SRC family coactivator. (A) The interaction of SRC family coactivators, Mediator and SWI/SNF complex with TR was analyzed by GST–TR pulldown assay in vitro. Note that SRC-1, SRC-3 and TRAP220 were bound to GST–TR in a T3-dependent manner, whereas no interaction was observed between BRG1 and GST–TR. (B) Tethering SRC-3 to chromatin is sufficient for transcriptional activation. The oocytes were injected with or without mRNA encoding a Gal4–SRC-3 fusion protein (200 ng/µl, 18.4 nl/oocyte) and ssDNA of the 4×UAS.TRβA reporter. After overnight incubation, transcription was analyzed by primer extension in (B), DNA topology assay in (C) and ChIP assays in (D).

Fig. 5. The ability of SRC-3 to induce transcriptional activation, chromatin remodeling and recruitment of SWI/SNF and Mediator is correlated with its interaction with CBP/p300. (A) Diagram of the different functional domains of SRC-3 fused to Gal4 DBD. bHLH/PAS, basic helix–loop–helix and PAS dimerization domain; NID, nuclear receptor interaction domain; CID, CBP/p300 interaction domain; HAT, histone acetyltransferase domain. (B) Only the fragment containing the CID is sufficient for transcriptional activation when tethering to chromatin. The concentrations of mRNA used: +, 100 ng/µl and ++, 300 ng/µl. The transcription was assayed by primer extension and the protein expression was revealed by western analysis using a Gal4 DBD-specific antibody. (C) Topology assay showed that only Gal3-SRC-3C is capable of inducing DNA topological change. The concentration of mRNA used for each fusion protein is 300 ng/µl. (D) ChIP assays show that p300, BRG1 and TRAP220 can be recruited by Gal4–SRC-3C.

Fig. 6. SWI/SNF and Mediator complexes can be recruited by p300 and this recruitment is facilitated by histone acetylation. (A) Tethering of p300 to chromatin is sufficient for transcriptional activation and this activation can be inhibited by Lys-CoA. Oocytes were injected with Gal4–p300 mRNA (200 ng/µl) and ssDNA of the 4×UAS.TRβA reporter with or without addition of Lys-CoA and assayed for transcription after overnight incubation. (B) Injection was as in (A) but assayed for chromatin remodeling by DNA topology assay. Note that Lys-CoA partially inhibited loss of negative superhelical density. (C) ChIP assays revealed that SWI/SNF and Mediator can be recruited by Gal4–p300, and co-injection of Lys-CoA abolished histone acetylation and partially inhibited the recruitment of BRG1 and TRAP220. The results for BRG1 and TRAP220 were also analyzed by quantitative real-time PCR and the average values of three independent experiments are shown. (D) Lys-CoA has no effect on chromatin remodeling instigated by Gal4–BRG1.

Fig. 7. Chromatin remodeling by liganded receptors is also partially dependent on histone acetylation, and blocking the recruitment of SRC coactivators inhibited chromatin remodeling. (A) Injection was similar to experiments in Figure 1A except that Lys-CoA (100 µM) was included as indicated. Note that Lys-CoA partially inhibited chromatin remodeling induced by liganded AR. (B) Expression of the RID of SRC-3 (amino acids 582–843) inhibited T3-dependent activation (left panel) and chromatin remodeling (right panel). (C) A sequential model showing that recruitment of SRC/p300, SWI/SNF and Mediator is achieved through independent interaction with liganded NRs. (D) A new combinatorial model showing that, in addition to NR–cofactor interaction, cofactor–cofactor and cofactor–histone interaction also contribute to the recruitment of multiple cofactors required for NR activation. CBP/p300 can mediate the recruitment of SWI/SNF and Mediator. Furthermore, histone acetylation by CBP/p300 facilitates subsequent recruitment of SWI/SNF and Mediator.