Role of Mediator in Transcriptional Activation by the Aryl Hydrocarbon Receptor

The aryl hydrocarbon receptor (AHR) binds many aromatic hydrocarbon compounds and mediates their carcinogenesis. We demonstrate that the endogenous AHR physically associates with the endogenous TRAP/DRIP/ARC/Mediator complex in a ligand-dependent manner. The Med220 subunit, which is known to interact with several nuclear hormone receptors through its LxxLL motifs, potentiates AHR-dependent reporter gene activity in a LxxLL-independent manner. Depletion of Med220 substantially reduces endogenous AHR-mediated CYP1A1 transcription. Both Med220 and CDK8 (another subunit of TRAP/DRIP/ARC/Mediator) are recruited to the CYP1A1 enhancer in a TCDD (2,3,7,8-Tetrachlorodibenzo-p-dioxin) -dependent fashion in vivo and Med220’s LxxLL motifs are not required. Med220 rapidly and persistently associates with the enhancer but not the promoter of the CYP1A1 gene after TCDD treatment with similar kinetics as AHR and the coactivators p300 and p/CIP. Our findings demonstrate a novel role for Med220 in AHR-regulated transcription, which differs mechanistically from its role in transcriptional regulation by other previously studied transcription factors. as early as 15 minutes after TCDD a maximum after about 30 minutes, and thereafter. Pol II occupancy over the promoter mirrored the recruitment of these coactivators. Our observations are consistent with previous nuclear run-on assays indicating that a maximal transcription rate is achieved within 30-40 minutes after TCDD treatment of Hepa-1 cells and is maintained for several hours (54, 55) and a previous restriction enzyme accessibility/micrococcus nuclease sensitivity study (8, 9) indicating a persistent open chromatin configuration and sustained maximum transcription rate after TCDD treatment of Hepa-1 cells. Our results indicate that the dynamics of cofactors association with the CYP1A1 enhancer is different from that of ER and TR. Med220 and a p160 coactivator associate and dissociate in unison on estrogen responsive promoters/enhancers in a cyclic manner similar to the estrogen receptor, whereas the recruitment of p300 is transient. In the case of TR, the p160 coactivator and p300 occupancies are rapid but not sustained, returning to the basal level after 2 hours of ligand treatment: Med220 recruitment follows a much slower kinetics. The unique kinetic profile of p300 in the case of AHR may be a reflection of different roles of p300 in AHR and NR mediated transcription. Our study reveals an integral role for Med220 in AHR driven transcription. Given the multimeric composition of the Mediator complex and potential tissue-specific variation and modification of its subunits, this complex may play a significant role in determining tissue and developmental variations in response to ligands of AHR.


Introduction:
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) and certain polycyclic aromatic hydrocarbons (PAHs) are widespread environmental pollutants that are designated human carcinogens and have been shown to be potent toxicants and carcinogens in experimental animals (1). The Aryl Hydrocarbon Receptor (AHR), a ligand activated transcription factor, is the intracellular protein that mediates most of the toxic effects of these compounds. Additionally, AHR may also play an important role in normal developmental and physiological setting possibly via an endogenous ligand(s). The evidence in support of this is largely derived from AHR knock-out mice. These mice are viable, but exhibit many physiological changes and developmental abnormalities, including reduced liver weight, a depressed immune system and reproductive defects such as small litter size (2).
Unliganded AHR forms a complex with the chaperone protein hsp90, a co-chaperone protein called p23 (3) and the Hepatitis B Virus X-associated protein 2 (XAP2) (4). This complex resides in the cytoplasm. Upon ligand binding, the AHR undergoes a conformation change and dissociates from the complex. The resulted exposure of its nuclear localization sequence triggers the nuclear translocation of AHR (5). Inside the nucleus, AHR dimerizes with a related nuclear protein called the Aryl Hydrocarbon Receptor Nuclear Translocator (ARNT). Biochemical). 400 µg of nuclear extract was pre-cleared with 35 µl of protein A Sepharose at 4 °C for one hour, and immuno-precipitations were performed overnight at 4 °C with AHR antibody, ARNT antibody or pre-immune IgG. Immuno-complexes were collected using 35 µl protein A sepharose containing 1mg/ml BSA. The sepharose beads were washed 4 times with 20 mM HEPES pH 7.9, 12.5% glycerol, 250mM KCl, 1.5mM MgCl 2 , 0.2 mM EDTA and 0.1% NP40. Immuno-complexes were eluted with SDS-sample buffer and separated in 6% SDS-PAGE followed by western blot analysis, probing with Med220, Med130, AHR or ARNT antibodies.

Transient transfections and reporter gene assays:
Transfections of expression plasmids were performed using the Superfect reagent (Qiagen). For Hepa-1 and HepG2 cells, a dual-luciferase system (Promega) was used. DNA mixtures of the indicated amount of Med220 wild-type or mutant constructs were cotransfected into 12 well plates with 40ng of Renilla luciferase reporter (pRL-TK) driven by the HSV thymidine kinase promoter, and 200ng firefly luciferase reporter (pGL-CYP1A1) driven by 4.2 kb of the 5' upstream regulatory region of the rat CYP1A1 gene, including its promoter and multiple XREs. The final DNA concentrations were adjusted using empty expression vector to ensure that equal amounts of DNA were used in each plate. After 24 hours, transfected cells were treated with either TCDD or vehicle (DMSO) for 16 hours. Cells were then harvested and lysed in Passive lysis buffer (Promega).
Luciferase activities were measured using the luciferase assay reagent (Promega). The Renilla luciferase activity was used to normalize the firefly luciferase activity of each sample. Each transfection was carried out in triplicate.

RNA interference in HepG2 cells:
A Med220 siRNA oligo (siMed220), 5'-GGCUCUCAAAGUAACAUCU(TT)-3' , a scrambled RNA oligo(SC) with similar GC content 5'-GAGUAAGUCCAUCAC CUAU(TT)-3' and their respective complementary oligos synthesized by Dharmacon were deprotected and annealed per the manufacturer's protocol. For ectopic reporter assays, siMed220 or SC were cotransfected into 12 well plates with a mixture of 200 ng pGL-CYP1A1 firefly luciferase reporter and 40 ng pRL-TK renilla luciferase reporter, or a mixture containing 50 ng of GAL4-VP16 expression construct, 200 ng GAL4 firefly luciferase reporter and 40ng pRL-TK renilla luciferase reporter. Transfections were performed at 90% cell confluency with Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Transfection mixtures were replaced with fresh medium 5 hours later. 36 hours post transfection, cells transfected with the pGL-CYP1A1 reporter were treated with DMSO or TCDD for an additional 16 hours, whereas the cells transfected with the GAL4 reporter were left undisturbed. Cell lysates were harvested and reporter activities were determined as described above. For the effect of Med220 knock-down on endogenous CYP1A1 induction, siMed220 or SC oligos were transfected into 6 well plates at 30% confluency using oligofectamine (Invitrogen). 24 hours later, cells were transfected again with the same procedures for an additional 24 hours, cells were treated with either DMSO or TCDD for 6 hours. Total RNA was then isolated using Trizol (Invitrogen) and assayed for CYP1A1, AHR, and β-actin mRNA levels by real time PCR, as descried below. For detection of Med220 protein levels, cells were lysed in 50mM Tris-HCl pH 7.9, 300mM NaCl, 1mM EDTA, 0.5% NP40, 10% Glyerol, 1mM DTT and 1x complete protease inhibitor (Roche Biochemical) and 40 µg of whole cell extract was then separated on 6% SDS-PAGE followed by western blotting with either the Med220 antibody or AHR antibody.

Reverse Transcription and Real-time quantitative PCR:
Reverse transcriptions were carried out with Thermoscript RT kit (Invitrogen) using 500 ng of Total RNA in 20 µl volume. 10% of the reaction product was used together with SYBER-GREEN PCR master mix (Applied Biosystem) to assemble a 100 µl reaction that was divided into three 30 µl replicas and run on I-cycler PCR machine (Bio-Rad). Real

AHR/ARNT interacts with the Mediator Complex in vivo.
In order to explore the potential involvement of Mediator in AHR/ARNT regulated transcription, we set out to test whether the AHR/ARNT dimer can interact with Mediator.
Hepa-1 cells were treated with TCDD or vehicle (DMSO) for 60 minutes. Nuclear extracts from these cells were prepared and immuno-precipitations were performed using an AHR antibody. Two of the Mediator subunits Med220 and Med130, were coimmunoprecipitated with AHR ( Figure 1A) from cells treated with TCDD, but not from cells not treated with this AHR ligand. Whereas Med220 may be a loosely attached subunit, since nullification of Med220 does not seem to compromise the overall integrity of the Mediator complex (19), Med130 is a an integral part of the Mediator: absence of Med130 destabilizes the associations of Med100 and Med95 with the Mediator complex (22). Hence this co-IP experiment indicates that AHR/ARNT interacts with the Mediator complex. We also immunoprecipitated ARNT from a nuclear extract of Hepa-1 cells untreated with TCDD. We were unable to detect co-immunoprecipitation between Med220, Med130 and ARNT ( Figure 1B). These last results indicate ARNT, at least when not associated with AHR, does not interact with the Mediator complex.
Med220 and also Med220 with mutated LxxLL motifs enhance AHR/ARNT mediated reporter gene activity in Hepa-1 and HepG2 cells. peroxisome proliferator-activated receptor-γ (PPARγ) and estrogen receptor (ER) (23,14,19,16). Mutations of two LxxLL motifs in Med220 abolish its interaction and transcriptional coactivation activity with NRs (24,25). Another important class of NR coactivators is the p160 family including SRC-1, NCoA2 and p/CIP, which can also interact with NRs via their LxxLL motifs (26). The p160 family of coactivators have also been implicated in AHR/ARNT mediated transcription (27,28). SRC-1 can interact with both AHR and ARNT (27), and the interaction between AHR and SRC-1 depends on the latter's LxxLL motifs (28). We therefore extended our study using Med220 derivatives with mutated LxxLL motifs. We found that Med220 with either one or both of its LxxLL motifs mutated enhanced AHRC regulated reporter activity with equal potency as wild-type Med220 in Hepa-1 and HepG2 cells ( Figure 2B and 2C).
We used RNA interference (RNAi) to deplete endogenous Med220 in order to examine the contribution of Med220 to AHR regulated transcription. We were precluded from using Med220 wild type and Med220 knockout mouse embryo fibroblasts (MEF)[see later] in this study because CYP1A1 expression is known to be repressed in MEF (29) and also because the receptor appeared to be more abundant in Med220-/-MEF than in Med220+/+ MEF (data not shown). HepG2 cells were transfected with a small interference RNA oligo against human Med220 (siMed220) or a corresponding scrambled control oligo (SC) for 48 hours, followed by DMSO or TCDD treatment for 6 hours. Total RNA was then prepared and analyzed by quantitative real time RT-PCR to determine Med220 mRNA levels. Whole cell extracts of the transfected HepG2 cells were also analyzed for Med220 protein levels. As shown in Figure 3A and 3B, the siMed220 oligo drastically reduced Med220 mRNA and protein levels but did not affect AHR protein levels. For Med220's effect on AHR/ARNT mediated reporter gene transcription, either siMed220 or SC oligos was cotransfected with the CYP1A1-luciferase reporter plasmid into HepG2 cells. 36 hours after transfection, cells were treated with DMSO or TCDD for an additional 16 hours and reporter activities were then determined. As showed in Figure 3C, siMed220 treatment significantly reduced TCDD induced reporter activity, but SC had no effect. When the same oligos were cotransfected into HepG2 cells along with GAL4-VP16 expression and GAL4 based reporter plasmids, no significant difference in reporter activities were observed between siMed220 treated and SC treated cells ( Figure 3D), in accordance with previous studies showing that VP16 targets the TRAP80 subunit in the Mediator complex and that GAL4-VP16 mediated reporter activities are unaffected by the absence of Med220 (30). To study the effects of siRNA on the endogenous CYP1A1 gene, HepG2 cells were transfected with the siMed220 oligo or the control oligo SC for 48 hours, followed by DMSO or TCDD treatment for 6 hours. Total RNA was then prepared and analyzed by quantitative real time RT-PCR. TCDD triggered CYP1A1 transcription was significantly compromised in siMed220 treated cells ( Figure 3E).
We noticed that siMed220 decreases both the reporter gene activity and the endogenous CYP1A1 mRNA levels in the absence of TCDD. This is probably due to transcription of CYP1A1 in the absence of TCDD also depending on AHR. Several lines of evidence support this last claim: constitutive levels of AHR-responsive genes CYP1A2 and UGT1 * 06 decrease dramatically in AHR null mice as compared with wild type mice (2) and a substantial decrease of CYP1B1 basal transcription is also observed in primary AHR-/-MEF when compared with AHR+/+ MEF (31). To provide additional support for this notion, we used real time QPCR to compare CYP1A1 mRNA levels in Hepa-1 cells with those in c35, a mutant derivative of Hepa-1 which contains a mutant, substantially inactivated form of AHR (32). We found indeed that in the absence of TCDD, c35 had much less CYP1A1 mRNA than Hepa-1 ( Figure 3F). The measurable level of CYP1A1 transcription in the absence of the exogenous ligand is probably due to the effect of an endogenous ligand in Hepa-1 cells (33) or to the presence of a small amount of AHR inside the nucleus at any given time resulted from trafficking of the receptor between the cytoplasm and nucleus (34).

Mediator associates with the murine CYP1A1 enhancer in vivo in a TCDD dependent fashion.
We carried out Chromatin Immuonprecipitation (ChIP) experiments to investigate whether the Mediator complex is recruited to the endogenous the CYP1A1 gene upon its transcription activation. Hepa-1 cells were treated with DMSO or TCDD for 60 minutes, followed by formaldehyde crosslinking. Cells were lysed in lysis buffer and sonicated so that the size of the majority of chromatin fragments were reduced below 600 bps.
Antibodies against Med220 and CDK8 were then used in order to pull down their associated chromatin fragments. Reversal of crosslinking and deproteinization were then carried out. DNA fragment were purified and amplified by PCR with primers flanking the mCYP1A1 enhancer region ( Figure 4A). We thereby detected TCDD dependent enhancer association for both Med220 and CDK8 ( Figure 4B), indicating that the Mediator was demonstrated TCDD inducible binding of Med220 and AHR to the CYP1A1 enhancer in these cells ( Figure 4C). However, no Med220 association with the enhancer was observed in Med220-/-MEF, although TCDD-inducible binding of AHR to the enhancer still occurred ( Figure 4C). Moreover, TCDD inducible binding of CDK8 to the enhancer could only be detected in Med220+/+ MEF ( Figure 4C). This strongly suggests that Med220 is critical for the recruitment of the Mediator complex by AHR/ARNT. In agreement with the fact that CYP1A1 expression is suppressed in MEFs, ChIP assays in these cells revealed that although receptor binding and Med220 recruitment to the enhancer were readily observed 60 minutes post TCDD treatment, no increase of Pol II occupancy over the promoter could be detected even after 2 hours of TCDD treatment (data now shown), suggesting that AHR is capable of recruiting Mediator even in the absence of a stabilized preinitiation complex over the promoter.
To study the role of the LxxLL motifs of Med220 in the process of recruitment, we performed ChIP experiments in Med220 null MEF stably infected with a mutant Med220 in which both of LxxLL motifs are mutated. As shown in Figure 4D, Med220 with mutated LxxLL motifs can still be recruited to the enhancer in response to TCDD.
Collectively, therefore these results show that Med220 is recruited to the murine CYP1A1 enhancer after TCDD treatment; that the LxxLL motifs in Med220 are not required for recruitment; and that the large complex of the Mediator containing CDK8 is employed.
Kinetics of assembly of AHR, Med220 and other coactivators at the CYP1A1 enhancer after TCDD treatment.
Different transcription activators recruit different coactivator complexes in distinct kinetic manners to their target promoters, as is evident from a few well-studied nuclear receptor systems. The estrogen receptor (ER) recruits Med220 and the p/CIP member of the p160 family of coactivators simultaneously to the estrogen responsive regulatory sequences and these proteins appear to associate and disassociate from the regulatory sequences of the pS2 and the cathepsin D genes in a cyclic fashion, the timing coinciding with receptor DNA binding and Pol II occupancy over the regulatory region (36,37). In the case of thyroid hormone receptor (TR), the receptor mediated recruitment of p160 family of coactivators precedes Med220 recruitment (38). We carried out ChIP kinetic studies on the murine CYP1A1 regulatory regions using antibodies against AHR, Med220 and various other cofactors in Hepa-1 cells. The use of sonication conditions that reduced the sizes of the majority of chromatin fragments to below 600 bps enabled us to distinguish binding to the promoter from binding to the enhancer element, located 1 kb from the promoter. We AHR nor Med220 was detected at the promoter ( Figure 5C) suggesting that the Med220 is probably located more proximal to the enhancer than to the promoter in the multimeric protein complex that is associated with the CYP1A1 gene during transcriptional activation.
TCDD treatment stimulated association of Pol II with the promoter ( Figure 5C), but as expected, not with the enhancer ( Figure 5A). The kinetics of binding of Pol II to the promoter was very similar to the kinetics of AHR, Med220, p300 and p/CIP binding to the enhancer ( Figure 5C). A ChIP reimmunoprecipitation experiment using p300 and Med220 antibodies sequentially was performed ( Figure 5D). Briefly, cross-linked chromatins from TCDD treated Hepa-1 cells were first immuno-precipitated with the p300 antibody. The immuno-complex was eluted and subjected to immuno-precipitation with the Med220 antibody. Our results indicate that p300 and Med220 can co-occupy the same enhancer element.
These data suggest that though targeting similar pool of coavtivators, AHR recruits them to chromatin with a different kinetic profile from those nuclear receptors that have been studied (36,37,38).

Discussion:
Mammalian transcriptional activation is multi-faceted in the sense that inputs from different modulators influence final read-outs. Repression by nucleosomes needs to be overcome in order to allow promoter sequences to be accessible to the general transcription Interference between AHR/ARNT and NRs at this crucial step of transcription activation could contribute to endocrine disruption by TCDD, in addition to other proposed mechanisms, such as enhanced ligand metabolism (51), down regulation of nuclear receptor levels (52) and altered hormone synthesis (53).
To examine the contribution of Med220 to AHR regulated transcription, we monitored CYP1A1 induction after selective reduction of endogenous Med220 using RNAi.
Transient knock-down of Med220 using RNAi not only allowed us to study the effect in cell lines highly responsive to TCDD, but also potentially minimized possible perturbation or compensatory effects occurring after elimination of Med220 in long term experiments.
Our data show that RNAi reduction of Med220 in HepG2 cells selectively compromised AHR regulated CYP1A1 transcription. However, CYP1A1 induction was not completely abolished. There could be several mechanisms responsible for this remaining activity, including the effects of residual amounts of Med220, or the possibility of weak association between Mediator and AHR via additional subunits other than Med220. A parallel observation was made with TR, showing that it retains a small yet significant amount of transactivation in Med220-/-MEF when compared to Med220 wild type MEF (30).
Transcription of CYP1A1 in the uninduced state is silent. Positioned nucleosomes at the promoter block transcription. TCDD treatment induces chromatin remodeling over the promoter and rapid activation of transcription. The TAD of AHR is indispensable for this process. The remodeled state of chromatin at the promoter is stable after induction rather than being transient (9). Our kinetic ChIP studies on cofactor occupancy provide a means by which AHR mediated coactivator complex assembly can be examined during induction.
The kinetics of recruitment of Med220 to the enhancer could not be distinguished from that of p/CIP, p300 or even AHR. Association of all these proteins with the enhancer was detected as early as 15 minutes after TCDD treatment, reached a maximum after about 30 minutes, and stabilized thereafter. Pol II occupancy over the promoter mirrored the recruitment of these coactivators. Our observations are consistent with previous nuclear run-on assays indicating that a maximal transcription rate is achieved within 30-40 minutes after TCDD treatment of Hepa-1 cells and is maintained for several hours (54, 55) and a previous restriction enzyme accessibility/micrococcus nuclease sensitivity study (8,9) indicating a persistent open chromatin configuration and sustained maximum transcription rate after TCDD treatment of Hepa-1 cells. Our results indicate that the dynamics of cofactors association with the CYP1A1 enhancer is different from that of ER and TR.    the mRNA level of CYP1A1 normalized to that of β-actin was measured by real time PCR using primers specific for human Med220 and human β-actin. F. c35 and Hepa-1 were treated with DMSO or TCDD for 10 hours, total RNA were isolated using Trizol and subjected to reverse transcription. CYP1A1 mRNA level normalized to that of β-actin was measured by Real time PCR using primers specific for murine CYP1A1 and β-actin.  Similar experiments as in A were carried out except that shorter time points were used.
C. ChIP products from A were subjected to PCR reactions with primers flanking the TATA box of murine CYP1A1. D. ChIP was first carried out using the p300 antibody (first lane) , the immuno-complex was eluted by 10mM DTT. The elution was diluted and immuno-precipitated with the Med220 antibody (second lane). ChIP products were subjected to PCR reactions using primers flanking the enhancer of murine CYP1A1.