ABSTRACT
The nuclear receptor CAR (NR1I3) regulates hepatic drug and energy metabolism as well as cell fate. Its activation can be a critical factor in drug-induced toxicity and the development of diseases, including diabetes and tumors. CAR inactivates its constitutive activity by phosphorylation at threonine 38. Utilizing receptor for protein kinase 1 (RACK1) as the regulatory subunit, protein phosphatase 2A (PP2A) dephosphorylates threonine 38 to activate CAR. Here we demonstrate that CAR undergoes homodimer-monomer conversion to regulate this dephosphorylation. By coexpression of two differently tagged CAR proteins in Huh-7 cells, mouse primary hepatocytes, and mouse livers, coimmunoprecipitation and two-dimensional gel electrophoresis revealed that CAR can form a homodimer in a configuration in which the PP2A/RACK1 binding site is buried within its dimer interface. Epidermal growth factor (EGF) was found to stimulate CAR homodimerization, thus constraining CAR in its inactive form. The agonistic ligand CITCO binds directly to the CAR homodimer and dissociates phosphorylated CAR into its monomers, exposing the PP2A/RACK1 binding site for dephosphorylation. Phenobarbital, which is not a CAR ligand, binds the EGF receptor, reversing the EGF signal to monomerize CAR for its indirect activation. Thus, the homodimer-monomer conversion is the underlying molecular mechanism that regulates CAR activation, by placing phosphorylated threonine 38 as the common target for both direct and indirect activation of CAR.
INTRODUCTION
A group of nuclear receptor superfamily members exhibit high constitutive activity. For these nuclear receptors, the molecular mechanism by which they control their constitutive activities to acquire responsiveness to activation is not understood at this time. Constitutively active/androstane receptor (CAR) is a member of this group of nuclear receptors. CAR regulates hepatic drug metabolism, disposition, and energy metabolism as well as cell signaling, through which it plays central roles in the development of drug-induced toxicity and diseases, such as type 2 diabetes and hepatocellular carcinoma (1 – 4). Understandably, determining the CAR activation mechanism is necessary to elucidate CAR-mediated disease processes and to predict and prevent them.
Constitutively active CAR is inactivated by phosphorylation at Thr38 within its DNA binding domain (DBD) and is retained in the cytoplasm of hepatocytes (5). Protein phosphatase 2Ac (PP2Ac) utilizes receptor for activated C kinase 1 (RACK1) as a regulatory subunit to dephosphorylate Thr38 (6). Then, nonphosphorylated CAR is activated and translocates into the nucleus. Epidermal growth factor (EGF) represses CAR activation by stimulating extracellular signal-regulated kinase 1/2 (ERK1/2) to directly bind phosphorylated CAR, preventing dephosphorylation of Thr38 and inactivating the receptor (7, 8). Phenobarbital (PB), the classic CAR-activating drug, antagonizes EGF by directly binding to the epidermal growth factor receptor (EGFR), reversing its signal to induce dephosphorylation of Thr38 for indirect activation of CAR (6). However, further insight into this antagonistic activation mechanism requires an understanding at the molecular level of an additional mechanism by which CAR converts the EGF-EGFR signal for its activation. An X-ray crystal structure of the mouse CAR ligand binding domain (LBD) revealed a 2-fold axis homodimer structure (9). The PP2A/RACK1 binding site of CAR was previously delineated to be on a loop near the N terminus of the LBD (6). This configuration of the CAR homodimer is found to bury the PP2A/RACK1 binding site within its interface. Therefore, we hypothesized that this homodimer is the underlying molecular mechanism through which CAR regulates PP2A/RACK1 binding in response to the EGF signal.
In the present study, coimmunoprecipitation, chemical cross-linking, two-dimensional electrophoresis, and fluorescence resonance energy transfer (FRET) analyses were employed to support the reported homodimer configuration of CAR and to demonstrate that the phosphorylation state of CAR at Thr38 regulates the existence of the homodimer to regulate RACK1 binding in response to the EGF-ERK1/2 signal. Subsequently, it was found that CAR activators do in fact dissociate the CAR homodimer either directly or indirectly in mouse liver. These results shed light on how a phosphorylation-based system provides the basis for an intermolecular switch as the general mechanism that enables CAR to confer responsiveness by controlling constitutive activity. Moreover, since Thr38 is conserved as a phosphorylation motif in a majority of nuclear receptors (10), phosphorylation and homodimerization may be common mechanisms for many constitutively activated nuclear receptors to regulate their activities.
RESULTS
CAR forms a homodimer in response to EGF.EGF is known to prevent PP2A/RACK1 from binding to phosphorylated CAR and dephosphorylating Thr38 (6). Coimmunoprecipitations were performed to examine whether or not CAR forms its homodimer in the cytoplasm and if EGF regulates this homodimerization. A CAR T38D mutant was selected for experiments because it is the target of dephosphorylation and is also sequestered in the cytoplasm. FLAG-tagged and green fluorescent protein (GFP)-tagged CAR T38D constructs were ectopically coexpressed in Huh-7 cells, whole extracts of which were subjected to coimmunoprecipitation assays to examine the interaction between FLAG- and GFP-tagged CAR T38D mutants and to determine whether or not EGF regulates this interaction. An anti-GFP antibody coprecipitated FLAG-tagged CAR, and EGF treatment increased this coprecipitation (Fig. 1A). Consistent with this coimmunoprecipitation, an anti-FLAG antibody also coprecipitated GFP-CAR T38D, and EGF increased the coprecipitation (see Fig. S1A in the supplemental material). Thus, phosphorylated CAR enabled formation of the homodimer in response to EGF treatment. Although a CAR T38A mutant was also found to form its homodimer in coimmunoprecipitation assays, this homodimer was only poorly detected compared to that of the CAR T38D mutant (Fig. S1A). The primary reason for this poor detection may result from its exclusive localization (Fig. S1B). Wild-type (WT) CAR, which is expressed in both the cytoplasm and the nucleus, formed its homodimer only in the cytoplasm (Fig. S1C). Consistent with this nuclear localization of CAR T38A, when CAR T38A and CAR T38D were coexpressed, they formed a heterodimer, although the interaction levels were greatly reduced compared to those of the CAR T38D homodimer (Fig. S1D).
EGF regulates dimerization of CAR. (A) Coimmunoprecipitation (IP) assays were performed using Huh-7 cells transfected with FLAG- and GFP-tagged CAR T38D and treated with EGF (10 ng/ml) for 30 min. Cell extracts were immunoprecipitated with an anti-GFP antibody and immunostained with an anti-FLAG antibody. (B) Huh-7 cells were transfected with expression plasmids for CFP- and YFP-tagged CAR T38D or with CFP and YFP plasmids and treated with EGF (10 ng/ml) for 12 h. FRET efficiency was calculated for 10 cells of each sample. *, P < 0.05 (Tukey-Kramer test). (C) Cell lysates were prepared from EGF-treated Huh-7 cells overexpressing GFP-tagged CAR T38D. For blue-native PAGE/SDS-PAGE two-dimensional electrophoresis, GFP-tagged CAR T38D (65 kDa) was detected by Western blotting with anti-GFP antibody. Molecular masses (in kilodaltons) for blue-native PAGE are indicated above the gels.
FRET analysis with CFP-CAR T38D and YFP-CAR T38D in Huh-7 cells further confirmed our finding (Fig. 1B). Thus, EGF strengthened the interaction between CAR molecules in Huh-7 cells. Subsequently, the nature of this interaction was examined by two-dimensional gel electrophoresis (blue-native PAGE/SDS-PAGE). CAR T38D appeared at a size consistent with a monomer in phosphate-buffered saline (PBS)-treated Huh-7 cells and shifted to that of a homodimer after EGF treatment (Fig. 1C). The results obtained by electrophoresis suggest that the interactions observed by coimmunoprecipitation and FRET analyses are reflective of CAR homodimerization.
Next, Huh-7 cells were pretreated with the EGFR tyrosine kinase inhibitor erlotinib to examine EGFR involvement in CAR dimer formation. Erlotinib treatment antagonized the EGF signal, as indicated by dephosphorylation of EGFR and ERK1/2, and dissociated the CAR homodimer (Fig. 2A). In mouse primary hepatocytes, 0.1 μM erlotinib treatment inhibited phosphorylation of CAR at Thr38 (Fig. 2B) and dose dependently increased the Cyp2b10 mRNA level (Fig. 2C). This mRNA increase was significantly lower than that induced by PB, maybe because PB also regulates the multiple steps of CAR-mediated transcription in the nucleus (11), while erlotinib can regulate only nuclear CAR translocation.
Inhibition of EGFR attenuates interaction of CAR and activates CAR. (A) Huh-7 cells transfected with FLAG- and GFP-tagged CAR T38D were treated with DMSO (0.1%) or erlotinib (1 μM) for 1 h before 10 ng/ml EGF treatment. Cell extracts were immunoprecipitated with an anti-GFP antibody and immunostained with an anti-FLAG antibody. (B) Primary mouse hepatocytes were treated with DMSO (0.1%), erlotinib (0.01 to 10 μM), or PB (2 mM) in medium containing 10% FBS for 1.5 h. Cell lysates were subjected to Western blot analysis. (C) Primary mouse hepatocytes were treated with vehicle (0.1% DMSO), erlotinib (0.01 to 10 μM), or PB (2 mM) in medium containing 10% FBS for 18 h. Total RNA was subjected to quantitative RT-PCR analysis of Cyp2b10. Data were normalized by use of Gapdh. Values are means with standard deviations (SD) (n = 3). *, P < 0.05 versus vehicle group (Dunnett's test).
Homodimer interface.CAR formed its homodimer in Huh-7 cells, but what configuration did it use? An X-ray crystal structure of the CAR homodimer was previously reported (PDB ID 1XNX ) (9), in which three loops (loop 1 [142-FIHH-145], loop 2 [210-FCLQTQNFL-218], and loop 3 [301-RRPRDRF-307], between helices H3 and H4, H6 and H7, and H10 and H11, respectively) form the dimer interface (Fig. 3A). Consistent with the notion that this interface is strengthened by both a salt bridge and polar interactions, NaCl at high concentrations greatly decreased CAR T38D dimerization in coimmunoprecipitation assays (Fig. 3B). To further test this possible interface, coimmunoprecipitation assays were performed in the presence of a peptide of each loop within the interface. Both loop 1 and loop 3 peptides inhibited CAR to form a homodimer in a dose-dependent manner (Fig. 3C). The glucocorticoid receptor (GR) is reported to also form a homodimer, with a configuration similar to that of the CAR homodimer (12). A peptide of loop 1 within the GR homodimer was unable to inhibit CAR dimerization and was used as a negative control (Fig. 3C). In addition to peptide competition assays, a mutation of Asp305 to arginine (D305R in the sequence 301-RRPRRRF-307) was introduced into loop 3 so that these loops of two CAR monomers repel each other. This mutation also prevented dimerization of CAR T38D (Fig. 3D). Thus, these results are consistent with the concept that CAR utilizes these loops to form its homodimer in the configuration suggested by the CAR crystal structure. Thus, all three loops are involved in the formation of the homodimer interface.
Characterization of the CAR dimer interface. (A) CAR dimer interface showing the key residues and the pseudostrands. The interaction motifs (F142 to H145, F210 to L218, and R301 to F307) of both monomers are labeled and colored red. (B) Cell lysates from Huh-7 cells overexpressing FLAG- and GFP-tagged CAR T38D and treated with EGF were subjected to coimmunoprecipitation assay with a buffer containing 100 mM or 500 mM NaCl. (C) Cell lysates from panel B were incubated with small peptides (1 to 1,000 μM) for 16 h and then subjected to coimmunoprecipitation assay. Pep1 and Pep3, 138PAHLFIHHQP147 and 299QQRRPRDRFL308 of CAR, respectively; NS, one interacting loop of GR (residues 541 to 557; PEVLYAGYDSSVPDSTW). (D) Huh-7 cells overexpressing FLAG- and GFP-tagged CAR T38D or CAR T38D/D305R were treated with EGF for 30 min. Cell lysates were subjected to coimmunoprecipitation assay. (E) Primary mouse hepatocytes from Car −/− mice were transfected with 5×(NR1)-TK-pGL3, phRL-TK, and an expression plasmid for wild-type CAR or CAR D305R and treated with PB (2 mM) for 24 h. Reporter activities were measured, and the reported values are means with SD (n = 4). *, P < 0.05 (Tukey-Kramer test).
In reporter assays using the CAR binding motif expressed from the CYP2B6 promoter, we determined the influence of the D305R mutation on CAR activation (Fig. 3E). Basal levels of reporter activity of this mutant were higher than those of the WT, and PB did not induce reporter activity of the D305R mutant (Fig. 3E), suggesting that this mutation inhibits dimerization of CAR in its X-ray crystal configuration and increases its nuclear translocation.
The interaction energy between the CAR monomers upon formation in this form of homodimer was calculated and compared to that between CAR and retinoid X receptor α (RXRα) upon formation of their heterodimer. CAR utilizes a surface composed of helices H8, H10, and H11 (these helices are numbered H7, H9, and H10 in references 13 and 14), which resides opposite its location in the CAR homodimer, to form a heterodimer (Fig. S2). Molecular dynamics (MD) simulation followed by energy calculations gave results for the CAR-RXRα heterodimer and the CAR homodimer of −5,345 ± 124 kcal/mol and −5,163 ± 117 kcal/mol, respectively. Homodimerization of CAR contributes −210 ± 26 kcal/mol to the total dimer energy, and this dimer interaction energy is less than a third of the CAR-RXRα heterodimer interaction energy of −724 ± 58 kcal/mol. Thus, CAR can be expected to form a stable homodimer in the cytoplasm, where RXRα is not present, in the configuration observed in the CAR crystal structure. On the other hand, this energy-based consideration suggests that CAR is unable to form its homodimer in the presence of RXRα but instead forms a heterodimer with RXRα in the nucleus.
RACK1 binds to CAR monomers.The X-ray crystal structure of the CAR homodimer suggests that the PP2A/RACK1 binding site is buried within its interface structure (Fig. 3A). As suggested above, EGF treatment, which stimulated CAR homodimerization, decreased coimmunoprecipitation of CAR by an anti-RACK1 antibody (Fig. 4A). On the other hand, EGF was unable to decrease coimmunoprecipitation of the CAR D305R mutant that disrupted homodimer formation (Fig. 4A). The homodimer was produced by chemically cross-linking in vitro-translated CARs with dimethyl pimelimidate (DMP) (Fig. 4B). Subsequent glutathione S-transferase (GST) pulldown assays showed that GST-RACK1 bound the CAR monomers before cross-linking but not the homodimer after cross-linking (Fig. 4B). These results indicated that RACK1 binds only to monomers. It was previously shown that inactivation of ERK1/2 by the MEK1/2 inhibitor U0126 results in the dephosphorylation of CAR at Thr38 (7). We examined if U0126 dissociates the CAR homodimer. After pretreatment with U0126 or its inactive derivative U0124, coimmunoprecipitation assays were performed and showed that U0126, but not U0124, increased the amount of CAR monomers and RACK1 binding to these monomers (Fig. 5). Thus, these observations indicate that EGF stimulates CAR homodimerization through ERK1/2 and prevents RACK1 binding to phosphorylated CAR.
Monomer-dimer conversion regulates RACK1 binding to CAR. (A) Huh-7 cells overexpressing FLAG-tagged CAR T38D or CAR T38D/D305R were treated with EGF. Cell lysates were subjected to coimmunoprecipitation assay using a RACK1 antibody. (B) In vitro-translated FLAG-CAR T38D was cross-linked by use of 1 mg/ml DMP and then incubated with recombinant GST or GST-RACK1 protein and glutathione-Sepharose 4B. Proteins were extracted from glutathione-Sepharose 4B with SDS-PAGE sample buffer. Immunoblotting was performed using an anti-FLAG antibody. Open and shaded arrowheads indicate expected FLAG-CAR monomer and dimer sizes, respectively.
Phosphorylation of ERK1/2 regulates the dimer-monomer proportions of CAR. (A) Huh-7 cells transfected with FLAG- and GFP-tagged CAR T38D were pretreated with 10 μM U0126 or U0124 for 1 h before treatment with 10 ng/ml EGF. Cell extracts were immunoprecipitated with an anti-GFP antibody. (B) Huh-7 cells transfected with FLAG-tagged CAR T38D were pretreated with 10 μM U0126 or U0124 for 1 h before treatment with 10 ng/ml EGF. Cell extracts were immunoprecipitated with an anti-RACK1 antibody.
Ligand-induced homodimer dissociation.The phosphatase inhibitor okadaic acid is known to repress ligand-induced nuclear translocation of CAR (15). Moreover, the GFP-CAR T38D mutant remains sequestered in the cytoplasm even after treatment with the human CAR (hCAR) ligand CITCO (5). These findings suggested that CAR must be dephosphorylated either before ligand binding or by ligand binding. Actually, both CITCO and mouse CAR ligand 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP) treatments increased dephosphorylation of CAR in human and mouse primary hepatocytes (16) (Fig. S3). To this end, coimmunoprecipitation assays were employed to examine how CITCO binding affected CAR dimerization and/or RACK1 binding. CITCO effectively decreased coimmunoprecipitation between FLAG- and GFP-tagged CAR T38D constructs in the presence of EGF (Fig. 6A). Conversely, CITCO increased coimmunoprecipitation of CAR T38D with RACK1 (Fig. 6B). Moreover, CITCO acted as an antagonist of EGF to decrease the ERK1/2 interaction with CAR T38D (Fig. 6C). However, CITCO did not inhibit either phosphorylation of EGFR or that of ERK1/2 (Fig. 6D). Thus, CITCO induced CAR monomerization and RACK1 binding directly, not indirectly through the EGF-EGFR-ERK1/2 signal.
CAR ligand induces monomerization of CAR. (A) Huh-7 cells transfected with FLAG- and GFP-tagged CAR T38D were pretreated with 1 μM CITCO for 1 h before treatment with 10 ng/ml EGF. Cell extracts were immunoprecipitated with an anti-GFP antibody. (B and C) Huh-7 cells transfected with FLAG-tagged CAR T38D were pretreated with 1 μM CITCO for 1 h before treatment with 10 ng/ml EGF. Cell extracts were immunoprecipitated with an anti-RACK1 antibody (B) or anti-ERK1/2 antibody (C) and immunostained with an anti-FLAG antibody. (D) Huh-7 cells were treated with CITCO (1 μM) or erlotinib (1 μM) for 2 h and then subjected to Western blot analysis.
CAR homodimer in mouse liver.FLAG-tagged human CAR-expressing transgenic mice were crossed with Car −/− mice to generate a mouse line (named FC/KO) that expresses human but not mouse CAR (Fig. S4A). CITCO treatment induced nuclear translocation of CAR and increased the Cyp2b10 mRNA level in the livers of FC/KO mice (Fig. S4B and C). Blue-native PAGE/SDS-PAGE of liver extracts and Western blot analysis with an anti-FLAG antibody detected a FLAG-hCAR complex of around 300 kDa, and this CAR complex dissociated after treatment with PB or CITCO (Fig. S5). Given these facts, FC/KO mice were used to examine CAR homodimerization by expression of GFP-CAR T38D administered through tail vein injection with the expression plasmid. Whole-liver lysates were then prepared and subjected to coimmunoprecipitation assays and two-dimensional blue-native PAGE/SDS-PAGE. CITCO treatment decreased interactions between FLAG-CAR and GFP-CAR T38D (Fig. 7A). Two-dimensional blue-native PAGE/SDS-PAGE was employed to examine the nature of these protein-protein interactions by Western blotting. A spot was detected at a position consistent with the CAR homodimer for the livers before treatment, and it was converted to a CAR monomer position after CITCO treatment (Fig. 7B). Similarly, PB treatment converted the CAR dimer to monomers (Fig. 7C and D). These results suggest that the CAR homodimer dissociates to form monomers in the mouse liver after treatments with CAR activators, as observed in Huh-7 cells.
CAR activators dissociate dimers in the mouse liver. FC/KO mice were injected with the expression plasmid for GFP-tagged CAR T38D. (A and C) Seven hours after injection, mice were treated with corn oil (CO; 10% DMSO) or CITCO (10 mg/kg) (A) or with PBS or PB (100 mg/kg) (C) for 1 h. Liver lysates were subjected to coimmunoprecipitation assay. (B and D) Four hours after injection, mice were treated with corn oil (CO; 10% DMSO) or CITCO (10 mg/kg) (B) or with PBS or PB (100 mg/kg) (D) for 4 h. Liver lysates were subjected to blue-native PAGE/SDS-PAGE two-dimensional electrophoresis. GFP-tagged CAR T38D (65 kDa) was detected by Western blotting with anti-GFP antibody.
DISCUSSION
The identification of CAR as a PB-activated nuclear receptor in 1999 was a watershed moment in the quest to answer the question, first broached in 1962, as to how PB induces hepatic drug metabolism (17, 18). However, CAR's high constitutive activity has impeded our efforts to determine the molecular mechanism of CAR activation. In addition, the unexpected fact that PB is not a CAR ligand and indirectly activates CAR further complicated the matter. Superseding these complications was the finding that CAR is inactivated by phosphorylation of Thr38, a residue residing between the two zinc fingers of the DBD (5). This finding provided us with critical information to begin to untangle the complexities of CAR activation. First, PB was found to antagonize the EGF signal to elicit dephosphorylation at Thr38 by PP2A/RACK1 for CAR activation (6). Our introduction of the concept that CAR can form a homodimer now greatly broadens our understanding of mechanisms by which the receptor is activated. In response to EGF, CAR forms a homodimer in the cytoplasm to repress dephosphorylation by burying the PP2A/RACK1 binding site within its interface. CITCO directly binds this homodimer to dissociate it into monomers for dephosphorylation and activation. PB reverses the EGF-EGFR signal to dissociate the homodimer for indirect activation. Thus, the homodimer-monomer conversion of phosphorylated CAR is the underlying molecular mechanism by which CAR regulates its activation.
Figure 8 depicts a simplified model of the homodimer-monomer conversion of the CAR activation mechanism based on our present findings. In our present experiments with FC/KO mice, endogenous human CAR appeared as a large complex with an apparent molecular mass of about 300 kDa. This large complex dissociated as the transiently expressed CAR homodimer did after CITCO treatment. Although this should be confirmed in future investigations, the CAR homodimer may constitute a core of the large complex, implying that CAR monomerization results in dissociation of the large complex. The chaperone HSP90 and the cochaperone cytoplasmic CAR retention protein (CCRP; also known as DNAJ7) are known to interact with CAR in the cytoplasm (15, 19). Although these proteins may be included in the large complex, our investigations are not sufficient to integrate these factors into the present homodimer model. Although HSP90 appeared to be able to interact with a surface that forms a dimer interface, HSP90 was not critically involved in CAR homodimerization (our unpublished observation). Studies with CCRP KO mice found that while CCRP regulates PB-induced nuclear accumulation of CAR, CCRP also regulates CAR-mediated transcription in the nucleus (11). Active ERK1/2 binds to the XRS peptide near the C terminus of the CAR LBD to keep CAR phosphorylated (7). This binding enables CAR to form its homodimer (Fig. 5A). It is anticipated that continuous investigations in this line will allow us to understand the cytoplasmic form of CAR and its regulatory mechanism at the molecular level in the future.
Depicted model of CAR activation. EGF elicits ERK1/2 binding to a short peptide near the C terminus of CAR, called XRS (7). This binding enables CAR to form its homodimer and prevents CAR from binding to PP2A/RACK1, thereby retaining CAR phosphorylation. The ligand binds the CAR homodimer, dissociating it from ERK1/2 to convert it into monomers. PP2A and RACK1 then bind the monomers to dephosphorylate them.
Whereas the indirect activation mechanism by which PB activates CAR via the EGF receptor is clear (6), the direct activation mechanism induced by CAR ligands has been elusive. CITCO, a human CAR ligand, is unable to activate the CAR T38D mutant (5), and CITCO treatment results in the dephosphorylation of CAR in human primary hepatocytes (16). Ligand activation of CAR is also under cell signal regulation, since EGF suppresses activation in mouse primary hepatocytes (8). However, these observations were unable to explain whether CITCO induced dephosphorylation or activated an already dephosphorylated CAR or how EGF regulated ligand activation. We have now found that CITCO dissociates the homodimer, enabling RACK1 to bind to CAR for dephosphorylation. Since CITCO did not antagonize EGF activation of the EGF receptor as observed with PB, CITCO appears to indirectly antagonize the EGF signal to dissociate the homodimer by directly binding to and displacing CAR from ERK1/2. Therefore, phosphorylation-mediated homodimerization has integrated both direct and indirect activation into a single activation mechanism.
Thr38 of CAR is conserved as a phosphorylation motif in a majority (41 of 46) of human nuclear receptors (10). Thr38 directs its side chain toward the surface of the LBD in our model structure of CAR with both the DBD and the LBD (5). Phosphorylation of the corresponding Ser78 of hepatocyte nuclear factor 4α (HNF4α) is associated with the interaction between the DBD and the LBD in a crystal structure (20). The CAR homodimer that helped our investigations utilizes a surface, constrained with three loops of the LBD (9), which resides on a side opposite the surface of CAR that forms a heterodimer with retinoid X receptor α (RXRα) (13, 14). This novel homodimer configuration enables CAR to utilize phosphorylation to regulate its activation by burying the PP2A/RACK1 binding site within the interface. As mentioned previously, the majority of human nuclear receptors conserve this phosphorylation motif within their DBDs (10). In addition to CAR, endogenous estrogen receptor alpha (ERα) and farnesoid X receptor (FXR) are found to be phosphorylated at this conserved motif in mouse tissues and cells in vivo (10, 21). Like CAR T38D, phosphomimetic mutants of the vitamin D receptor (VDR S51D) and RAR-related orphan receptor α (RORα S100D) also formed homodimers as observed with CAR (data not shown). Thus, numerous nuclear receptors should be able to integrate this conserved phosphorylation into their functional regulation through homodimerization. In addition, homodimerization can also be utilized independently of the conserved phosphorylation. GR, which does not carry this motif, is known to form its homodimer in the cytoplasm (22, 23), with the same configuration as that observed for the CAR homodimer (12). This GR homodimerization may retain GR in the cytoplasm, since a plant-derived GR antagonist, compound A, inhibits GR from forming this homodimer, instead translocating it into the nucleus (23). Therefore, the ability of nuclear receptors to form their homodimers in this novel configuration should greatly diversify their functions and regulations.
Retrospectively, homodimerization of nuclear receptors has generally been investigated with respect to DNA binding and/or gene activation capabilities. For example, whereas HNF4α and ERα utilize the same surface for homodimerization that is also used for heterodimerization with RXRα (20, 24, 25), it remains to be investigated whether these also form homodimers in the cytoplasm. Recent investigations of nuclear receptors have increasingly emphasized biological roles outside the nucleus (26, 27). For example, GR has been suggested to regulate inflammation signaling in the cytoplasm (28), and CAR is known to repress c-Jun N-terminal kinase 1 (JNK1) signaling via direct interaction with GADD45β (29). The cell membrane-bound progesterone receptor (PR) appears to regulate an ion channel (26, 30, 31). The homodimers observed with CAR should provide us with an excellent basis to gain additional insight into the structural biology of nongenomic functions of nuclear receptors either in the cytoplasm or on membranes.
In conclusion, we now comprehend CAR activation as a unified mechanism by which drugs and xenobiotics activate CAR either directly or indirectly. The underlying principle of this mechanism is the phosphorylation-mediated homodimerization of CAR. EGF stimulates homodimerization, retaining CAR in its phosphorylated (i.e., inactivated) form, while CAR activators antagonize EGF to dissociate the homodimer for activation. CAR activation can be either a benefit or a risk factor to human health. Knowing the activation mechanism should help us to understand the relative benefits and risks. Moreover, phosphorylation-mediated homodimerization may provide us with a collective structural basis by which to integrate numerous nuclear receptors into a unified regulatory mechanism far beyond that of CAR.
MATERIALS AND METHODS
Reagents and materials.EGF, U0124, and okadaic acid were purchased from Calbiochem (San Diego, CA). Erlotinib was purchased from Selleckchem (Houston, TX). U0126 was purchased from Promega (Madison, WI). Phenobarbital, CITCO, TCPOBOP, anti-FLAG M2 affinity gel (A2220), and horseradish peroxidase (HRP)-conjugated anti-FLAG M2 (S8592) were purchased from Sigma-Aldrich (St. Louis, MO). Protein L- agarose (sc-2336), anti-TATA-binding protein (TBP) antibody (sc-273), and HRP-conjugated antibodies against rabbit IgG (sc-2004) and mouse IgM (sc-2064) were purchased from Santa Cruz Biotechnology (Dallas, TX). Antibodies against RACK1 (61078) and HSP90 (610419) and mouse normal IgM (550963) were obtained from BD Biosciences (San Jose, CA). Antibody against CAR (PP-N4111-00) was purchased from Perseus Proteomics (Tokyo, Japan). Antibody against green fluorescent protein (GFP) (HRP conjugated; ab6663) was purchased from Abcam (Cambridge, MA). Antibodies against EGFR (2232S), phospho-EGFR (Tyr1173) (4407S), p44/42 mitogen-activated protein kinase (MAPK) (ERK1/2) (4695S), and phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) (4370S) were obtained from Cell Signaling Technology (Danvers, MA). Antibody against the phosphorylated Thr38 residue of the CAR peptide was produced in our previous work (5). The enhanced chemiluminescence reagent WesternBright was purchased from GE Healthcare (Piscataway, NJ).
Plasmids.Human CAR (hCAR) cDNA was previously cloned into pEYFP-c1 (Clontech Laboratories, Palo Alto, CA) (YFP-CAR) (32), pECFP-c1 (Clontech Laboratories) (CFP-CAR) (32), pEGFP-c1 (Clontech Laboratories) (GFP-CAR) (32), pCDNA3.1/V5-His TOPO (Invitrogen, Carlsbad, CA) (CAR-V5-His) (19), and pCR3 (Invitrogen) (33). A DNA fragment of the FLAG tag was placed at the 5′ end of CAR in pCR3 (FLAG-CAR) (7). The 5×(NR1)-TK-pGL3 plasmid was constructed previously (34). The phRL-TK control vector was from Promega. Using PrimeSTAR Max DNA polymerase (Clontech Laboratories) and appropriately mutated primers, we constructed and confirmed all of the mutants used by nucleotide sequencing.
Animal experiments.All mice were maintained in the temperature- and light-controlled National Institute of Environmental Health Sciences (NIEHS) animal facility and had free access to water and diet. All animal procedures were approved by the Animal Ethics Committee, NIEHS, National Institutes of Health. A colony of FLAG-hCAR transgenic B6 mice was established at Xenogen Biosciences. The human CAR gene was amplified from FLAG-CAR-pCR3 and cloned into the pLIVE vector (Mirus, Madison, WI). The vector was linearized with PacI and NdeI and electroporated into one-cell-stage embryos derived from C57BL/6 mice. The positive embryonic stem (ES) cell clones were injected into 486 one-cell-stage embryos of C57BL/6 mice. The obtained FLAG-hCAR transgenic mice were crossed with Car −/− (C57BL/6) mice (35). We confirmed the transgenic FLAG-hCAR cassette and the knockout of mCAR genes by genotyping using specific primer sets for mouse and human CAR.
The GFP-CAR T38D expression plasmid was injected via the tail vein into FC/KO mice by using the TransIT in vivo gene delivery system (Mirus) according to the manufacturer's instructions. The mice were intraperitoneally administered corn oil (10% dimethyl sulfoxide [DMSO]; 10 ml/kg of body weight) or CITCO (10 mg/kg) 4 and 7 h after the injection of plasmid and sacrificed 4 or 1 h after the drug administration. Liver sections were collected and subjected to several analyses.
Cell culture.Primary hepatocytes were isolated from 8- to 10-week-old C3H/HeNCrIBR males by a two-step collagenase perfusion method as described previously (36). Hepatocytes were seeded into collagen-coated plates (BD Biosciences) and cultured in Williams' medium E (Invitrogen) containing 10% fetal bovine serum (FBS), 1 mM sodium pyruvate, and 100 U/ml penicillin-streptomycin. Three hours after seeding, hepatocytes were treated with DMSO (0.1%), erlotinib (0.01 to 10 μM), or PB (2 mM) in Williams' medium E containing 10% FBS, 1 mM sodium pyruvate, and 100 U/ml penicillin-streptomycin. Huh-7 cells were cultured in minimum essential medium (Invitrogen) supplemented with 10% FBS, 2 mM l-glutamine, and 100 U/ml penicillin-streptomycin. Twenty-four hours after seeding, the culture medium was changed to prewarmed minimum essential medium without FBS, and plasmids were transfected by use of Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Twenty-four hours after transfection, cells were treated with 10 ng/ml EGF.
Fractionation.Huh-7 cells were collected and homogenized in 10 mM HEPES buffer (pH 7.6) containing 10 mM KCl, 1.5 mM MgCl2, 0.3% NP-40, and Complete protease inhibitor cocktail tablets (Roche Diagnostics, Indianapolis, IN). Homogenates were centrifuged at 2,000 × g for 5 min to obtain cytosolic extracts. The resulting pellets were resuspended in 10 mM HEPES buffer (pH 7.6) containing 400 mM NaCl, 100 mM KCl, 3 mM MgCl2, 0.1 mM EDTA, and Complete protease inhibitor cocktail tablets and ultracentrifuged at 20,000 × g for 30 min to obtain nuclear extracts.
Livers from the mice were homogenized in 10 mM HEPES (pH 7.6) containing 25 mM KCl, 2 M sucrose, 10% glycerol, and 1 mM EDTA. The homogenate was ultracentrifuged at 25,000 rpm for 45 min with an SW-28 rotor (Beckman-Coulter, Miami, FL). The precipitate was homogenized in 100 ml of lysis buffer (10 mM HEPES [pH 7.6], 100 mM NaCl, 10% glycerol, 3 mM MgCl2, 0.1 mM EDTA, and Complete protease inhibitor cocktail) by use of a Dounce homogenizer. Sodium chloride solution was added to the homogenate to bring the final concentration to 0.4 M, and the homogenate was incubated for 30 min and centrifuged at 20,000 × g for 30 min. The supernatant was subjected to Western blotting.
Coimmunoprecipitation.Coimmunoprecipitation was performed as described previously (6), with minor modifications. Cell lysate was incubated with anti-FLAG M2 affinity gel or anti-GFP–Sepharose at 4°C for 16 h or incubated with the indicated antibody for 12 h and then added, with protein L-Sepharose or protein G-coupled Dynabeads, to the reaction mixture, and the mixture was further incubated at 4°C for 6 h. For peptide competition, 1 to 1,000 μM peptide mixtures corresponding to sequences in loop 1 (residues 138 to 147; PAHLFIHHQP) and loop 3 (residues 299 to 308; QQRRPRDRFL) of CAR or the loop of GR (residues 541 to 557; PEVLYAGYDSSVPDSTW) were incubated with anti-FLAG M2 affinity gel for 16 h. The immune complexes were eluted with SDS-PAGE sample buffer and subjected to immunoblotting with the indicated antibody.
Determination of mRNA and protein levels.Quantitative reverse transcription-PCR (RT-PCR) was performed by using an ABI Prism 7700 sequence detector (Applied Biosystems, Foster, CA), TaqMan universal PCR mix, and primers (Applied Biosystems) for mouse Cyp2b10 (Mm00456591_m1) or Gapdh (4352934E).
Cultured cells were lysed in buffer containing 8 M urea, 0.1% Tween 20, and 0.1% SDS and then sonicated. Protein concentrations were determined with protein assay dye reagent (Bio-Rad Laboratories, Hercules, CA). Western blot analysis was carried out as described previously (7).
Two-dimensional blue-native PAGE and SDS-PAGE.Huh-7 cells or liver samples were homogenized in buffer containing 20 mM HEPES (pH 7.6), 50 mM NaCl, 15% glycerol, 1% Triton X-100, and 1 mM EDTA and centrifuged at 20,000 × g for 10 min at 4°C. Supernatant was mixed with 10× BN sample buffer (100 mM bis-Tris-HCl [pH 7.0], 2.5% Coomassie brilliant blue [CBB] G-250, and 500 mM 6-aminocaproic acid) and subjected to electrophoresis on a 4 to 16% gradient native gel (Invitrogen) with cathode buffer (50 mM Tricine-HCl, 15 mM bis-Tris-HCl [pH 7.0], 0.02% CBB G-250) and anode buffer (50 mM bis-Tris-HCl [pH 7.0]). After electrophoresis, a 4 to 16% gradient native gel was vertically cut to make a ribbon strip which was vertically placed onto a 10% SDS-PAGE gel.
Reporter assay.Primary hepatocytes isolated from 8-week-old Car −/− mice (C3H/HeNCrIBR) (3) were seeded into a collagen-coated plate (BD Biosciences). Four hours after seeding, cells were transfected with 5×(NR1)-TK-pGL3, phRL-TK control vector (Promega), and CAR-V5-His or CAR D305R-V5-His by use of Lipofectamine 2000 (Life Technologies) and then treated with PB (2 mM). Twenty-four hours after transfection, cell lysate was subjected to a dual-luciferase assay (Promega). Firefly luciferase activity was normalized to renilla luciferase activity.
Cross-linking and GST pulldown assay.FLAG-CAR T38D was in vitro translated by use of a TNT Quick coupled transcription-translation system (Promega). In vitro-translated FLAG-CAR T38D was incubated with 1 mg/ml dimethyl pimelimidate (DMP) and 100 mM triethanolamine in PBS for 1 h. Cross-linked proteins were incubated with recombinant GST or GST-RACK1 (6) and glutathione-Sepharose 4B (GE Healthcare) at 4°C overnight. Proteins were extracted from the resin with SDS-PAGE sample buffer and separated in an SDS-PAGE gel. Immunoblotting was performed using anti-FLAG antibody.
FRET assay.For the detection of CAR dimers in living cells, a sensitized emission fluorescence resonance energy transfer (FRET) assay was used. Huh-7 cells were transfected with cyan fluorescent protein (CFP)- and yellow fluorescent protein (YFP)-expressing plasmids or CFP-CAR T38D- and YFP-CAR T38D-expressing plasmids and treated with EGF for 12 h. Cells were visualized using an LSM 710 microscope (Zeiss, Oberkochen, Germany) with laser excitation lines of 458 nm and 514 nm for CFP- and YFP-tagged receptors, respectively, and transmitted light. Images of CFP and YFP emission were recorded simultaneously with the transmitted light images. The bleedthrough values for CFP-CAR T38D and YFP-CAR T38D were calculated by using the values for Huh-7 cells transfected with CFP-CAR T38D and YFP-CAR T38D alone, respectively, and were subtracted from the FRET channel signal. The FRET values were normalized and the mean values determined with the FRET and Colocalization Analyzer of ImageJ software by following the user guide (37). Each data point was the mean for 10 cells, and the experiment was repeated independently three times.
MD simulations.Molecular dynamics (MD) simulations of hCAR homodimer and hCAR/hRXRα heterodimer systems were performed. The initial dimer structure of the hCAR homodimer was created after superimposing the hCAR coordinates from the hCAR/hRXRα heterodimer (PDB ID 1XV9 ) on the mouse CAR dimer found in the X-ray crystal structures of PDB ID 1XNX . The hCAR/hRXRα heterodimer structure was taken directly from the X-ray crystal structure of PDB ID 1XV9 . Each of the dimeric systems was solvated in a box of water after adding hydrogens and necessary counter ions. The homodimer system had 24,907 water molecules solvating it, and the heterodimer system had 26,384 water molecules. After 500-ps belly dynamics runs with fixed protein heavy atoms and an energy minimization, a low-temperature, constant-pressure MD run with constrained protein positions was performed for each system to ensure a reasonable starting density of around 1 g/ml. After another energy minimization, followed by a stepwise heating MD run at constant volume to bring the temperature up to 300 K within 1 ns, a constant-volume MD trajectory calculation was performed for 10 ns to equilibrate each system. Production simulation runs were extended for each system over a period of 50 ns at a temperature of 300 K. All MD trajectory calculations were performed with 2-fs steps, using the Pmemd module of AMBER.14 and the Amber.ff14SB force field representing proteins.
ACKNOWLEDGMENTS
We thank the NIEHS Protein Expression Core Facility for providing anti-GFP–Sepharose beads for immunoprecipitation.
This work was supported by National Institutes of Health intramural research program Z01ES1005-01.
FOOTNOTES
- Received 12 December 2016.
- Returned for modification 29 December 2016.
- Accepted 18 February 2017.
- Accepted manuscript posted online 6 March 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/MCB.00649-16 .
- Copyright © 2017 American Society for Microbiology.
