The Peptide Near the C Terminus Regulates Receptor CAR Nuclear Translocation Induced by Xenochemicals in Mouse Liver

Abstract

In response to phenobarbital (PB) and other PB-type inducers, the nuclear receptor CAR translocates to the mouse liver nucleus (T. Kawamoto et al., Mol. Cell. Biol. 19:6318–6322, 1999). To define the translocation mechanism, fluorescent protein-tagged human CAR (hCAR) was expressed in the mouse livers using the in situ DNA injection and gene delivery systems. As in the wild-type hCAR, the truncated receptor lacking the C-terminal 10 residues (i.e., AF2 domain) translocated to the nucleus, indicating that the PB-inducible translocation is AF2 independent. Deletion of the 30 C-terminal residues abolished the receptor translocation, and subsequent site-directed mutagenesis delineated the PB-inducible translocation activity of the receptor to the peptide L³¹³GLL³¹⁶AEL³¹⁹. Ala mutations of Leu313, Leu316, or Leu319 abrogated the translocation of CAR in the livers, while those of Leu312 or Leu315 did not affect the nuclear translocation. The leucine-rich peptide dictates the nuclear translocation of hCAR in response to various PB-type inducers and appears to be conserved in the mouse and rat receptors.

Organisms are capable of inducing numerous xenochemical- and steroid-metabolizing enzymes such as cytochromes P450 (CYPs) in liver microsomes, increasing metabolic capability against xenochemical toxicity and carcinogenicity. Phenobarbital (PB), the prototype of structurally diverse xenochemicals, induces pleiotropically hepatic genes in mammalian species from mouse to human, with the CYP2B genes being most dramatically upregulated. A conserved 51-bp PB-responsive enhancer module (PBREM) has recently been located in the mouse, rat, and human CYP2B genes (9, 21). PB responsiveness of the region, including the 51-bp sequence, has been observed independently by various laboratories (16, 19, 20). Moreover, the nuclear receptor CAR (constitutive active receptor or constitutive androstane receptor) has been found to regulate the induction of the CYP2B genes, as well as the transactivation of PBREM (10, 21). CAR is retained in the cytoplasm of control hepatocytes and translocates to the nucleus following PB treatment (13). Nuclear translocation appears to be a general process through which CAR regulates the CYP2B induction, since various PB-type inducers (e.g., chloropromazine, chlorinated biphenyls, and methoxychlor) are also capable of translocating CAR into the nucleus in livers. The molecular and cellular mechanisms underlying the xenochemical-responsive nuclear translocation regulated by PB-type inducers remain a question of major interest.

The fluorescent protein-tagged hCAR, expressed in transformed cell lines such as HepG2 and HEK293, translocates spontaneously to the nucleus without exposing the cells to PB-type inducers (13). These in vitro systems provided a practical tool for the initial identification of signal sequences of CAR that regulate its nuclear translocation. To understand the molecular basis underlying CAR for nuclear translocation, we constructed various fluorescent protein-tagged CAR expression vectors and investigated the nucleocytoplasmic localization of these truncated or mutated CAR proteins expressed in HEK293 cells. Subsequently, these regulatory sequences were tested for their functions in livers using in situ injection of various CAR expression vectors. We herein describe experimental considerations that lead us to propose that the leucine-rich sequence near the C terminus regulates the xenochemical-induced nuclear translocation of CAR in mouse livers in vivo.

MATERIALS AND METHODS

Plasmids.

Polymerase chain amplification was employed using pT7HisMyc-hCAR plasmid, Pfu DNA polymerase, and specific sets of primers to generate deletion mutants of CAR with newly created XhoI and EcoRI sites at the 5′ and 3′ ends, respectively. PCR amplification was done with primers that amplified the nucleotides encoding amino acids 1 to 348 (hCAR_wt), 1 to 338 (hCAR_1–338), 1 to 328 (hCAR_1–328), 1 to 318 (hCAR_1–318), and 1 to 308 (hCAR_1–308). Amplified fragments were digested with these two enzymes and inserted into pEGFP-C1, pECFP-C1, or pEYFP-C1 vectors (Clontech, Palo Alto, Calif.) to give an in-frame N-terminal fusion with one of the green (GFP), cyan (CFP), or yellowish (YFP) fluorescent proteins. A human glucocorticoid receptor (hGR)cDNA (provided by John Cidlowski) was also cloned into pEYFP-C1 or pECFP-C1 vector. Site-directed mutagenesis of hCAR and mouse CAR (mCAR) in pEYFP-C1 was conducted according to the instruction manual for the QuickChange site-directed mutagenesis system (Stratagene, La Jolla, Calif.). Nucleotides encoding leucines 312, 313, 315, 316, and 319 in hCAR and leucines 326 and 329 in mCAR, as well as Gly314 and Glu318 in hCAR, were mutated to encode alanines. All mutations and deletions were confirmed by sequencing, and plasmids bearing the CAR cDNAs with mutations or deletions were prepared using Qiagen Plasmid Maxi Kit (Qiagen, Valencia, Calif.). PBREM-tk-luciferase reporter gene was constructed in the pGL3 plasmid (21). A mouse SRC-1 cDNA, corresponding to amino acid residues from 633 to 1405, was amplified by using Pfu DNA polymerase and a specific set of primers having EcoRI and HindIII restriction sites at the 5′ and 3′ ends, respectively. Amplified cDNA was cloned into pcDNA3.1 Myc/His plasmid (Invitrogen) with the Kozak sequence at the 5′ end. The SRC-1 cDNA was verified by sequencing.

DNA transfections and visualization of fluorescent protein-tagged CAR in HEK293 cells.

Cells were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum and 100 U of penicillin per ml and 100 μg of streptomycin per ml at 37°C under a 5% CO₂ atmosphere. For analysis of nuclear translocation, cells in an 8-well Lab-Tek Chamber Slide (Nalge Nunc, Naperville, Ill.) were switched to DMEM without phenol red, transfected with various plasmids using the calcium phosphate method (CellPhect Kit; Pharmacia), and cultured for an additional 24 h. Fluorescence-positive cells were identified with Axiovert 35 inverted fluorescence microscope (Zeiss) equipped with filters from Omega optics (Brattleboro, Vt.). YFP and GFP fluorescences were detected by using an XF104 filter, while CFP fluorescence was visualized by using an XF113 filter. Fluorescent images of liver cells were captured with Spot II cooled charge-coupled device camera (Diagnostic Instruments, Sterling Heights, Mich.) and processed using the accompanying software package. Nucleocytoplasmic localization of fluorescent protein-tagged CAR or GR was determined by counting cells. GFP fluorescence-positive cells were classified into three different categories: N<C for predominantly cytoplasmic fluorescence, N=C for equal fluorescence distribution in both the cytoplasmic and the nuclear regions, and N>C for preferentially nuclear fluorescence.

For analysis of SRC-1 coactivation HEK293 cells were transfected with pGL3-tk-mPBREM and pRL-CMV plasmids using calcium phosphate coprecipitation. The pcDNA3.1 Myc/His (Invitrogen) plasmid containing cDNA for wild-type hCAR and its mutant forms was cotransfected alone or with pcDNA3.1–mSRC-1 vector, and the luciferase activity was assayed 24 h after transfection. The promoter activities were determined from three independent transfections and normalized against Renila luciferase activities.

Expression of fluorescent protein-tagged CAR in livers.

A portion of the liver was exposed through a ventral midline incision, and 80 μg of a given plasmid DNA in 300 μl of minimal essential medium was injected into three different spots, as previously described (21), or the expression plasmids (10 μg) were injected through the tail vein using TransIT In Vivo Gene Delivery System (Mirus, Madison, Wis.) according to the manufacturer's protocol. Then, two doses of xenochemicals were administered intraperitoneally at 2 and 5 h after the injection: chloropromazine (CPZ; 50 and 25 mg/kg, respectively), PB (100 mg/kg), or 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene (TCPOBOP; 0.33 mg/kg). The mice were sacrificed 3 h after the last treatment with xenochemical, and a block of the liver was frozen for preparing the sections (8-μm thickness) using Tissue-Tek OTC compound (Sakyra Finetek, Torrance, Calif.). One of every other four consecutive sections was placed on a cover glass (three sections per glass). Liver sections on cover glasses were fixed with 4% paraformaldehyde, washed once with phosphate-buffered saline, dehydrated in methanol, and stained for nuclei using 0.5 μg of Hoechst S-33258 per ml in 80% glycerol.

Gel shift assays.

The TNT-Coupled Reticulocyte Lysate System (Promega) was used to prepare in vitro-translated CAR and RXRα as described in our previous study (21). These in vitro-translated receptors were incubated with 10 μl of HEPES buffer (pH 7.6) containing 0.5 mM dithiothreitol, 15% glycerol, 2 μg of poly(dI-dC), 0.05% NP-40, 50 mM NaCl, and approximately 30,000 cpm of ³²P-end-labeled oligonucleotide.

RESULTS AND DISCUSSION

Nuclear localization of hCAR in HEK293 cells.

The PBREM-tk-CAT reporter gene is activated by hCAR in cotransfected cells such as HepG2 and HEK293 in the absence of inducers such as PB and TCPOBOP (10). The activation appears to occur because hCAR is inherently active and always in the nucleus of the transfected cells (13). YFP-tagged hCAR was expressed in the nucleus of the transfected HEK293 cells, whereas CFP-tagged hGR was expressed in the cytoplasm and dexamethasone treatment translocated the GR into the nucleus (Fig. 1A).

FIG. 1.

Nucleocytoplasmic distribution of the wild-type and truncated CARs in HEK293 cells. (A) Both YFP-CAR and CFP-GR expression plasmids were cotransfected into HEK293 cells. After being cultured for 24 h, the transfected cells were treated with dexamethasone (+DEX, 0.1 μM) or dimethyl sulfoxide (−DEX) for 3 h and visualized under a fluorescence microscope. (B) Various truncated hCARs were constructed, expressed in HEK293 cells, and visualized under a fluorescent microscope as described in Materials and Methods. (C) Deletion mutants are shown underneath the wild-type (hCAR_wt) with numbers indicating deletions from the C-terminal end. The cells expressing hCAR were visualized and evaluated for the nucleocytoplasmic distribution of the receptor: N<C for cytoplasmic dominant fluorescence; N=C for equal fluorescence distribution in both cytoplasmic and nuclear regions; and N>C for nuclear dominant fluorescence.

Residues were successively deleted from the C terminus of hCAR, and the C-terminal truncated receptors were constructed into pEGFP-C1 vector and transfected into HEK293 cells (Fig. 1B and C). Even when the first 10 residues were removed, the receptor (hCAR_1–338) still accumulated in the nucleus, suggesting that hCAR_1–338 retained the translocation capability. However, deletion of the additional 10 residues decreased dramatically the nuclear translocation capability of the receptor. In fact, none of the cells displayed nuclear dominant localization of the hCAR_1–328. The lack of the nuclear accumulation was not due to a stability of the deleted CAR protein since Western blot analysis of the cell extracts showed the expression of both hCAR_wt and hCAR_1–328 with the expected sizes at similar levels in the transfected HEK293 cells (data not shown). In further deletion, hCAR_1–318 behaved similarly to the hCAR_1–328 with respect to the translocation capability. These results suggested that a regulatory information for the nuclear translocation might lie within the 20-amino-acid residue segment between positions 319 and 338 of the receptor. We then examined whether these residues regulated the nuclear translocation of CAR in mouse liver in vivo following treatment with PB or other PB-type inducers.

PB-responsive nuclear translocation in livers.

Our previous immunochemical analyses of mCAR in mouse livers and primary hepatocytes have shown that mCAR is cytoplasmic and translocates to the nucleus in response to PB and PB-type inducers such as TCPOBOP and CPZ (13). To examine whether hCAR was also retained in the cytoplasm and translocated into the nucleus after treatment with PB-type inducers, the CFP-tagged hCAR and YFP-tagged hGR were coexpressed in mouse liver in vivo. After injection with the corresponding expression plasmids, frozen sections of the livers were used to visualize these receptors under a fluorescent microscope. The wild-type hCAR_wt was expressed in the cytoplasm of liver cell, while hGR was primarily localized in the nucleus of the same cell (Fig. 2A), which is opposite from what happened in the transfected HEK293 cells (Fig. 1A). The cytoplasmic and nuclear localization of hCAR and hGR, respectively, appeared to be a true reflection of the nucleocytoplasmic localization supposed to be observed with these receptors in the unexposed livers. To examine whether the cytoplasmic hCAR_wt could translocate to the nucleus following treatment with PB-type inducers, the mice bearing livers injected with the expression plasmid were treated with CPZ (Fig. 2B). In fact, none of the 70 cells examined exhibited the exclusive nuclear localization of CFP-hCAR_wt in control livers (Fig. 2C). In sharp contrast to the unexposed liver cells, the majority of the 63 cells examined showed the nuclear localization of the CFP fusion protein in the CPZ-exposed livers. These results thus indicate that the expressed hCAR_wt is capable of translocating to the nucleus following CPZ treatment in mouse livers.

FIG. 2.

Nucleocytoplasmic distribution of hCAR and its deletions in livers in vivo. (A) YFP-hGR and CFP-hCAR_wt were coexpressed in the livers of nontreated mice using the in situ DNA injection and then visualized in the same cell under a fluorescence microscope. Panels a and b represent YFP-hGR and CFP-hCAR_wt, respectively, while panel c shows the stained nuclei. (B) CFP-hCAR_wt and YFP-hCAR_1–338 were coexpressed in the livers of the nontreated and CPZ-treated mice using the in situ DNA injection and then visualized in the same cell under a fluorescence microscope. Panels a and b represent YFP-hCAR_1–338 and CFP-hCAR_wt, respectively, while panel c was stained for nuclei. The three images are imposed in each panel d. (C) The expression plasmids for various hCARs as GFP-tagged protein were directly injected into mouse livers, and the nucleocytoplasmic distribution of the expressed CARs was analyzed in the nontreated (−) and CPZ-treated (+) livers as described in Materials and Methods. The numbers indicate the cell populations as follows: N, nuclear distribution; N/C, distribution in cytoplasm and nucleus; and C, cytoplasmic distribution.

After establishing the in vivo liver system in which the nuclear translocation of hCAR occurred in response to a PB-type inducer, we attempted to define the role of the C-terminal region of hCAR in nuclear translocation. For this, various C-terminal truncated CARs were coexpressed with the wild-type receptor in the same cell. The YFP-tagged hCAR_1–338, lacking C-terminal 10 residues (i.e., AF2 domain), was sequestered in the cytoplasm of unexposed livers and translocated to the nucleus following CPZ treatment, which also occurred with the wild-type in same liver cells (Fig. 2B and C). The expressions of CFP-tagged hCAR_wt and YFP-tagged hCAR_1–338 overlapped exactly in the unexposed and CPZ-induced liver cells. In fact, the hCAR_1–338 responded not only to CPZ but also to PB and TCPOBOP, indicating that the AF2 domain was nonessential for the nuclear translocation induced by PB-type inducers.

The nuclear translocation of hCAR, however, began to be affected by an additional deletion of the C-terminal residues. When the 20 residues were removed, none of the 32 cells showed the dominant nuclear localization in the CPZ-induced livers (Fig. 2C). However, the hCAR_1–328 retained some degree of the nuclear translocation capability in the liver since the receptor was localized equally in the cytoplasm and nucleus of half of these cells. Given the differences in the translocation capability of the truncated receptor in the livers compared with the complete loss in the HEK293 cells, we removed the 30 residues up to position 318 of the hCAR. The YFP-tagged hCAR_1–318 was always localized in the cytoplasm even after treatment with CPZ, indicating that the hCAR_1–318 lost the nuclear translocation capability in mouse liver (Fig. 3). These results indicated that, first of all, the AF2 domain plays no role in the nuclear translocation in the liver and that the key residue responsible for the hepatic translocation may reside between positions 319 and 328 of the receptor.

FIG. 3.

Nuclear translocation in response to various PB-type iducers. The expression vectors CFP-hCAR_wt and YFP-hCAR_1–318 were simultaneously injected into the livers of mice, and subsequently the mice were treated with PB, CPZ, or TCPOBOP as described in Materials and Methods. For each group, panels a and b display YFP-hCAR_1–318 and CFP-hCAR_wt, respectively; panel c was stained for nuclei; and panel d shows an imposition of all three images.

To search for residue(s) that may regulate the nuclear translocation in the liver, various amino acid residues between positions 319 and 338 were mutated: S321A, I322A, Y326A, Y328A, I330A, I333A, and L336A. These mutations, however, did not affect the nuclear translocation of hCAR (data not shown). Amino acid sequence analysis then revealed a small region of CAR that contains a cluster of Leu residues starting with Leu319 toward the N terminus: L³¹²LGLLAEL³¹⁹ (Fig. 4). Each of these Leu residues was mutated to Ala, and the mutated YFP-tagged hCARs were coexpressed with the CFP-tagged wild-type receptor in the mouse livers using gene delivery through tail vein injection. Since this delivery system provided us with high transfection efficiency, the expression of a given fluorescent protein-tagged CAR could be observed in multiple cells on same field. The hCAR[L312A] and hCAR[L315A] mutants translocated to the nucleus after CPZ treatment as the wild-type CAR did, whereas the mutants hCAR[L313A], hCAR[L316A], and hCAR[L319A] abrogated the nuclear translocation capability and remained in the cytoplasm even after CPZ treatment (Fig. 5). Thus, the L³¹³XXL³¹⁶XXL³¹⁹ appeared to be a motif that regulates the CPZ-inducible nuclear translocation of hCAR in the mouse livers. Retrospectively, the lack of Leu319 was a major reason that hCAR_1–318 was unable to translocate to the nucleus in the CPZ-treated livers. Although a role for residues 329 to 338 could not be completely ruled out, any such role might be secondary to the role played by these Leu residues in the nuclear translocation. The L³¹²LGLLAEL³¹⁹ sequence is conserved in the mouse and rat CARs as L³²²MGLLADL³²⁹ and L³²²MGLLAEL³²⁹, respectively (Leu residues in the motif are underlined). As expected, the mutation of Leu326 or Leu329 to Ala abolished the nuclear translocation of mCAR in the mouse livers (Fig. 6). The C-terminal leucine-rich peptide conserved as L/(M) XXLXXL appeared to be a general response signal that dictates the receptor CAR to translocate to the nucleus following CPZ treatment. We have designated this leucine-rich peptide as the xenochemical response signal (XRS).

FIG. 4.

Alignment of a carboxy-terminal sequence of hCAR with the corresponding region of hGR. The hGR sequence and the localization of the predicted secondary structures are depicted from the Wurtz's multiple alignments (22). The hCAR sequence was then aligned underneath of the sequence of hGR. Residues within the putative LXXLXXL sequence designated the XRS are boxed. The presumed α-helices 10, 11, and 12 are indicated by arrows. The numbers indicate the positions of the residues of each receptor. The sizes of DBD and LBD are arbitrary.

FIG. 5.

Nucleocytoplasmic localization of the hCAR mutants. (A) Various mutated hCARs were coexpressed with the wild-type hCAR using the gene delivery system through tail vein injection of their fluorescent protein-tagged expression vectors in the nontreated (−) and CPZ-treated (+) mouse livers as described in Materials and Methods. All three images of YFP-mutated hCAR, CFP-hCAR_wt, and nuclei stained with Hoechst S-33258 are imposed in these pictures. (B) Semiquantification of the intracellular distribution of various mutated hCAR in liver cells. The numbers indicate the cell populations as follows: N, nuclear distribution; N/C, distribution in cytoplasm and nucleus; and C, cytoplasmic distribution.

FIG. 6.

Nucleocytoplasmic localization of the mCAR mutants. Expression vectors of the wild-type mCAR and of either mCAR mutants were simultaneously injected in mice through the tail vein, and the mice were subsequently treated with CPZ (+CPZ) or dimethyl sulfoxide (−CPZ) as described in Materials and Methods. For each group, panels a and b display the mutated YFP-mCAR and wild-type CFP-mCAR, respectively; panel c was stained for nuclei; and panel d shows an imposition of all three images.

The XRS region overlaps with the region that could be involved in various protein-protein interactions of nuclear receptors. The recent crystal structures of RXR-PPAR and RXR-RAR heterodimers revealed that N-terminal region of helix 10 is located on the dimerization interface of these nuclear receptors (3, 7). Site-directed mutagenesis of residue in the same N-terminal region resulted in the inhibition of receptor heterodimerization (1). Therefore, we examined what degree the Leu mutations of XRS could affect the heterodimerization of hCAR with hRXRα. For this, gel shift assays were performed using ³²P-labeled NR1 oligonucleotide as a probe (Fig. 7A). Although mutations of each of these Leu residues slightly decreased the band intensities representing the CAR-RXRα heterodimer, there was no significant difference in the ability of various mutated hCARs to form RXRα heterodimer, regardless of their specific effects on the nuclear translocation of hCAR in mouse livers. We also examined whether the mutations altered the coactivation of hCAR by SRC-1 in HEK293 cells (Fig. 7B). However, the coexpression of SRC-1 resulted in an increase of ca. 70% of all hCAR-mediated transactivations of PBREM in the cotransfected cells. Our present experiments do not suggest that the region of XRS is not involved in various heterodimerizations, but they clearly show that both the RXRα heterodimerization and the SRC-1 coactivation are not affected by the Leu mutations that inhibited the nuclear translocation of hCAR. These results suggest that the Leu residues regulate the translocation but not activation of the receptor. Whether the XRS region constitutes and acts as a sequence motif remains an interesting question for future investigation.

FIG. 7.

Effect of Leu mutations on the heterodimerization with RXRα and the coactivation by SRC-1. (A) Various mutated hCARs were prepared by in vitro translation and mixed with the similarly prepared RXRα for gel shift assays using ³²P-labeled NR1 oligonucleotide as a probe. (B) pcDNA3.1 plasmids bearing various hCARs were cotransfected with a PBREM-tk-luciferase reporter plasmid (in pGL3) into HEK293 cells. After being cultured with or without the additional transfection of mSRC-1 for 24 h, the cells were lysed for luciferase activity. The activities were normalized against those of pRL-CMV, and coactivation was calculated as the percentage of the increase by coexpression of SRC-1.

XRS resides on α-helix 10 near the C terminus of the CAR molecule, as predicted from the alignment of multiple nuclear receptor sequences (25). The C-terminal α-helix 12 bears the AF2 domain; ligand-dependent functions are generally associated with the AF2 domain of nuclear receptors (5, 8). Recent X-ray crystal structures have revealed that α-helix 12 is displaced upon binding of the ligand, allowing its association with a coactivator (15, 22). In addition, AF2 has also regulated the ligand-dependent nuclear translocation of vitamin D receptor, in which deletion of the AF2 domain (AF-2del-VDR-GFP) abolished the translocation to the nucleus in response to the vitamin (18). In the case of CAR, however, our present studies have clearly shown that the receptor does not require the AF2 for the PB-induced nuclear translocation in liver in vivo. It is known that the removal of residues within the AF2 domain abrogated the constitutive transactivation activity of hCAR (4). Thus, the nuclear translocation and transactivation appear to be regulated by the different mechanism, and only the latter function is linked to the AF2 domain observed. Instead, the nuclear translocation function resides in the C-terminal LXXLXXL sequence, XRS.

Nuclear localization of a given protein is determined by a balance of nuclear import and export (11, 17), which can be actively regulated by peptide sequences such as the nuclear localization signal (NLS) and the nuclear export signal (NES) built into the protein. XRS does not resemble its sequence with so-called monopartite or bipartite NLSs consisting of a single or repeated cluster of basic amino acids. The NLS associates with specific factors such as importins that carry NLS-bearing proteins into nucleus. Based on its sequence, XRS is not a typical NLS, although the possibility of XRS being a novel NLS still remains an interesting question for future research. The sequence of XRS is somewhat similar to those of the NESs, occurring in the Ah receptor, heat-stable inhibitor of cyclic AMP-dependent protein kinase and human immunodeficiency virus type 1 Rev (6, 12, 24). The XRS acting as an NES may be an alternative possibility: in the liver treated with PB-type inducers, the XRS activity as NES is masked by the heterodimerization with RXR, for example, resulting in the accumulation of CAR in the nucleus. This is a less likely possibility since the CAR without XRS is retained in cytoplasm and is not accumulated in the nucleus of both livers as well as HEK293 cells. Moreover, the Leu mutations of XRS did not abolish the heterodimerization of hCAR with RXRα. Thus, the XRS as an NLS is still an interesting possibility.

The molecular and cellular mechanism of how XRS regulates the nuclear translocation of CAR in response to PB-type inducers remains a major interest in future research. Is direct PB binding to CAR essential for the receptor's nuclear translocation? The most potent PB-type inducer TCPOBOP induces the Cyp2b10 gene at a concentration of 10 to 50 nM in mouse primary hepatocytes compared with other PB-type inducers that require micromolar to millimolar concentrations to induce the gene, making it the best candidate for demonstrating its binding. Using sensitive but indirect in vitro binding assays, it has recently been shown that TCPOBOP can bind to mCAR but not to hCAR (14, 23). PB did not exhibit a meaningful binding to either mCAR or hCAR. Under the present circumstances, in which the direct binding is not firmly established, it is difficult to speculate how PB-type inducers activate XRS and translocate the receptor into the nucleus. Our finding of the translocation of hCAR as well as mCAR in the mouse livers following treatment with TCPOBOP, however, is insightful. CAR may translocate to the nucleus in the absence of the direct binding of TCPOBOP to the receptor. The protein phosphatase inhibitor okadaic acid is known to suppress the PB-induced nuclear translocation of mCAR in mouse primary hepatocytes (13). PB-type inducers may elicit a phosphorylation/dephosphorylation pathway that regulates the nuclear translocation of CAR in mouse livers. Although the direct binding mechanism should be considered, the indirect regulation via a protein dephosphorylation offers an alternative direction for future research.

Steroid hormone receptors are sequestered in the cytoplasm; as a multimeric complex with the hsp90 chaperonic system, a detailed hypothesis of the hormone-dependent release of the receptors has been proposed (2). GR is retained in the cytoplasm of transfected HEK293 cells prior to dexamethasone treatment. However, hCAR could not be retained in the cytoplasm and accumulated in the nucleus of the cells. The ability of HEK293 cells to retain GR but not hCAR in the cytoplasm leads us to think that these two receptors are regulated distinctly with respect to their translocation. In fact, the CAR also appeared to exist in the complex with hsp90 in liver cytoplasm (K. Yoshinari et al., unpublished observation), suggesting the presence of factors specific to CAR that regulate the receptor translocation. Since the α-helix 10 contains multiple sequences responsible for intramolecular and intermolecular interactions of the nuclear receptors (15), XRS may be one of these sequences and may bind to the specific factors regulating the PB-induced nuclear translocation of CAR in liver in vivo. Identifications of the intramolecular interaction with XRS and of proteins that associate with XRS would help us to uncover the mechanism by which XRS regulates the receptor CAR for the nuclear translocation. To this end, the purification and characterization of CAR as a large complex from the liver cytosol of untreated mice is now under way in this laboratory. Nevertheless, further characterization of XRS should provide insight into the nuclear translocation of CAR and the induction of various genes following exposure to PB-type inducers.