JBC Anatrace, Inc.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M411241200 on November 29, 2004

J. Biol. Chem., Vol. 280, Issue 7, 5960-5971, February 18, 2005
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
280/7/5960    most recent
M411241200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Jyrkkärinne, J.
Right arrow Articles by Honkakoski, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Jyrkkärinne, J.
Right arrow Articles by Honkakoski, P.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Amino Acids Important for Ligand Specificity of the Human Constitutive Androstane Receptor*

Johanna Jyrkkärinne{ddagger}§, Björn Windshügel§, Janne Mäkinen{ddagger}, Markku Ylisirniö||, Mikael Peräkylä||, Antti Poso**, Wolfgang Sippl¶, and Paavo Honkakoski{ddagger}{ddagger}{ddagger}

From the Departments of {ddagger}Pharmaceutics, ||Chemistry, and **Pharmaceutical Chemistry, University of Kuopio, P. O. Box 1627, FIN-70211 Kuopio, Finland and the Department of Pharmaceutical Chemistry, Martin-Luther-University, Halle-Wittenberg, Wolfgang-Langenbeck-Strasse 4, D-06120 Halle (Saale), Germany

Received for publication, October 1, 2004 , and in revised form, November 12, 2004.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
The human constitutive androstane receptor (CAR, NR1I3) is an important ligand-activated regulator of oxidative and conjugative enzymes and transport proteins. Because of the lack of a crystal structure of the ligand-binding domain (LBD), wide species differences in ligand specificity and the scarcity of well characterized ligands, the factors that determine CAR ligand specificity are not clear. To address this issue, we developed highly defined homology models of human CAR LBD to identify residues lining the ligand-binding pocket and to perform molecular dynamics simulations with known human CAR modulators. The roles of 22 LBD residues for basal activity, ligand selectivity, and interactions with co-regulators were studied using site-directed mutagenesis, mammalian co-transfection, and yeast two-hybrid assays. These studies identified several amino acids within helices 3 (Asn165), 5 (Val199), 11 (Tyr326, Ile330, and Gln331), and 12 (Leu343 and Ile346) that contribute to the high basal activity of human CAR. Unique residues within helices 3 (Ile164 and Asn165), 5 (Cys202 and His203), and 7 (Phe234 and Phe238) were found control the selectivity for CAR activators and inhibitors. A single residue in helix 7 (Phe243) appears to explain the human/mouse species difference in response of CAR to 17{alpha}-ethynyl-3,17{beta}-estradiol.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Nuclear receptors (NRs)1 are ligand-inducible transcription factors that govern many physiological processes. Receptors for steroid hormones, thyroid hormone and retinoic acid (RAR) are critical for cellular differentiation and development, whereas many "adopted orphan" receptors are metabolic sensors regulating cholesterol, fatty acid, and glucose turnover (1–3). The most important structural component found in NRs is the ligand-binding domain (LBD). The NRs display great selectivity for ligands because of distinct sizes, contours, and lipophilicities of their ligand-binding pockets (LBP) that are caused by variations of the lining residues. Ligand binding induces conformational changes in the LBD. These changes allow the NR to recruit co-activators needed for histone acetylation, or co-repressors required for histone deacetylation at the NR target gene (4, 5). In most NRs studied, the position of helix 12 determines which type of co-regulator is able to bind to the LBD. After binding of an agonist, helix 12 is stabilized in an orientation where it forms a hydrophobic co-activator surface together with helices 3 and 4, and the conserved "charge clamp" residues glutamate (helix 12) and lysine (helix 3) are essential for co-activator binding and ligand-induced NR activation. This co-activator surface overlaps with the binding site for co-repressors that are composed of helices 3 and 4. Antagonist binding stabilizes another orientation of helix 12, which prevents co-activator recruitment and promotes co-repressor binding (6).

The human constitutive androstane receptor (CAR, NR1I3) belongs to the nuclear receptor subfamily 1I together with pregnane X (PXR, NR1I2) and vitamin D receptors (VDR, NR1I1) (7, 8). CAR and PXR share some ligands including a variety of xenobiotics, steroid hormones, and prescription drugs. These two receptors are important activators of overlapping sets of genes that code for cytochrome P-450, conjugative enzymes, and transport proteins (9–15), and their induction can lead to adverse drug effects or harmful drug-drug interactions (16). CAR also protects against the toxicity of endogenous cholestatic bile acids and bilirubin by enhancing their metabolism (17, 18). Important hepatic enzymes that are involved in fatty acid oxidation and energy metabolism such as squalene epoxidase and phosphoenolpyruvate carboxykinase 1 are repressed by CAR (9, 19).

Unlike most NRs, human and mouse CAR have high constitutive activity in the absence of any added ligand (20, 21). This seems to be because of the constitutive interaction of CAR with several co-activators including SRC-1, GRIP-1, PGC-1, and TIF2 (21–26). The prototypic cytochrome P-450 inducer phenobarbital stimulates translocation of CAR, but has not been shown to bind CAR LBD unlike many other activators (12, 13). The differences in ligand specificity between human and mouse CAR are significant. For example, mouse CAR is inhibited by a limited set of steroids related to 3{alpha}-androstenol (ANDR) (21, 24), whereas TCPOBOP (22), chlorpromazine (27), and 17{alpha}-ethynyl-3,17{beta}-estradiol (EE2) (28) are potent activators of mouse CAR. However, they do not activate human CAR at all (13) and EE2 even inhibits it (24). Thus, of the few known modulators of human CAR, EE2 and ANDR are repressors (24, 29, 30), whereas 6-(4-chlorophenyl)imidazo-(2,1-b)(1,3)thiazole-5-carb-aldehyde O-(3,4-dichlorobenzyl)oxime and tri-(p-methylphenyl)phosphate (TMPP) are activators (31, 32). The effect of clotrimazole on human CAR is not clear: the effects range from repressing (33) to no effect (34) to activating (30). A likely explanation for this discrepancy is the use of different cell lines in co-transfection assays. It is known that the responses of selective modulators of estrogen and progesterone receptors vary because of differential content and selection of NR co-regulators in cell lines (35, 36). In our studies with HEK293 cells, clotrimazole has consistently activated human CAR (24, 30, 32).

The crystal structure of human CAR LBD has not been reported, and partly for that reason the molecular determinants for its ligand specificity remain obscure. We wished to study the importance of the amino acid residues forming the LBP of human CAR. To select the critical residues for mutation analysis, homology models of human CAR LBD were created. Twenty-two LBP amino acids were identified and alanine scanning mutagenesis was carried out. The activities and ligand specificities of these mutants were tested in mammalian activity assays. The yeast two-hybrid assay was applied to study the interactions between the mutated LBDs and the steroid receptor coactivator-1 (SRC-1) and the nuclear receptor co-repressor (NCoR). Then, to gain insight to the molecular mechanisms of ligand binding, improved CAR models were constructed, and the ligands were docked and run through molecular dynamics (MD) simulations. Various ligand derivatives were used to complement the mutation analysis. These studies identified, for the first time, several amino acids that (i) contribute to the high basal activity of human CAR, (ii) control the ligand selectivity, and (iii) explain some of the species differences in CAR ligand specificity.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Chemicals—All steroids were at least 99% pure and from Steraloids Inc. (Newport, RI). TMPP and triphenyl phosphate (TPP) were synthesized and purified to more than 99% purity (32). All other chemicals were of highest quality and bought from Sigma or Biomol (Plymouth, PA). Oligonucleotides were purchased from Sigma.

DNA Plasmids—The cytomegalovirus promoter-driven GAL4-human CAR LBD (residues 108–348) fusion protein plasmid has been described (30). The GAL4-human CAR{Delta}8 (residues 108–340) plasmid lacking eight residues of helix 12 was generated by standard PCR methods. Alanine scanning mutants of the human CAR LBD were done according to instructions in the QuikChange kit (Stratagene, La Jolla, CA). The GAL4-dependent reporter UAS4-tk-luciferase plasmid was a gift from R. M. Evans, and pCMV{beta} was purchased from Clontech, Inc. All plasmids were purified with Qiagen columns (Hilden, Germany) and LBD regions were verified by dideoxy sequencing.

Co-transfection Assays—HEK293 cells (ATCC CRL-1573) were cultured and transfected with GAL4-LBD expression vector, GAL4-responsive reporter, and {beta}-galactosidase control plasmid as described (24, 28). Every transfection experiment also contained expression vectors for GAL4 lacking any LBD and wild-type human CAR LBD for standardization. After transfection, cells were cultured for 24 h in the presence of vehicle (0.1%, v/v) or test chemicals. The optimum concentrations were chosen from preliminary dose-response tests. The cells were washed, lysed, and assayed for {beta}-galactosidase and luciferase activities (24, 28) with the Wallac Victor2 luminescence and absorbance reader for 96-well plates (PerkinElmer Life Sciences). The normalized luciferase activities are from three to six independent transfections and expressed as mean ± S.E.

Yeast Two-hybrid Assays—The human CAR and all mutant LBDs were inserted between the EcoRI and BamHI sites in pGBKT7 plasmid (Clontech). pGADT7 plasmids containing human NCoR and human SRC-1 NR interaction domains (NRID) have been reported (24, 28). The NRID peptides from thyroid hormone receptor-associated protein (residues 501–738 (37)) and peroxisome proliferator-activated receptor {gamma} coactivator-1 (PGC-1; residues 78–371 (38)) were amplified from human RNA and inserted into pGADT7 accordingly. Yeast colonies were selected, amplified, and treated with vehicle or CAR modulators for 3.5 h, and {beta}-galactosidase activities and cell densities were measured as described previously (24, 28). Each experiment included recombinant yeast expressing GAL4 alone or wild-type human CAR for standardization. The normalized {beta}-galactosidase activities (mean ± S.E.) are from two separate experiments, each with three independent colonies.

Computational Methods—The initial homology model was built on the VDR LBD template (39) using Quanta98 software (Molecular Simulations Inc., San Diego, CA) and CLUSTALW alignment (40). The coordinates of CAR side chains were taken from VDR when identical or from Ponder's rotamer library (41). Minimization in vacuo was done using the Adopted-Basis Newton Raphson method of the CHARMM program (42). The ligands were docked manually, and MD simulations were done using the AMBER7.0 simulation package (University of California, San Francisco, CA) and the Cornell et al. (43) force field. The parameters of the ligands were generated with the Antechamber suite of AMBER7.0 in conjunction with the general amber force field. The atomic point charges of ligands were calculated with the two-stage RESP (44) fit at the HF/6–31G* level using ligand geometries optimized with the semi-empirical PM3 method using the Gaussian98 program (Gaussian Inc., Pittsburgh, PA).

The final homology models of the human CAR LBD with NRID peptides were built using INSIGHT II (Accelrys Inc., San Diego, CA) based on CLUSTAL W alignments (40) (see Fig. 2). PXR (45) and VDR (39) were used as template structures for human CAR and CAR/SRC-1 complex, whereas the structure of peroxisome proliferator-activated receptor {alpha} complexed with silencing mediator of retinoid and thyroid receptors (46) was included to model the human CAR/NCoR complex. Non-conserved side chains were assigned by SCWRL 2.95 (47). This program has given excellent agreement within the LBP residues when applied to crystal structures of estrogen receptor, VDR, and PXR. The protonation state was adjusted to mimic pH 7.4. MD simulations of the models were performed in a solvent box filled with SPC water (48). Na+ and Cl- ions were added to ensure the overall neutrality of the systems. The models were minimized using 2000 steps of steepest descent within the GROMOS96 force field (GROMACS version 3.14).



View larger version (54K):
[in this window]
[in a new window]
  		FIG. 2.
Sequence alignment of human CAR. Top, the first and the last residue of each helix and the {beta}-sheet in the human CAR model are shown in the graph, and the positions of mutated LBP residues are indicated by an X. Bottom, the alignment of human CAR, VDR, and PXR sequences. Human CAR residues 106–128, 158–223, 249–307, and 334–337 were modeled based on human PXR residues 143–165, 240–305, 330–388, and 415–418, respectively. The template for human CAR residues 129–157, 224–248, 308–333, and 340–348 was taken from hVDR residues 147–164 plus 216–226, 295–319, 397–404 and 415–423, respectively. The mutated CAR residues are shown in bold and hydrophobic residues are underlined. The VDR and PXR residues lining the LBPs are shown in bold and italics. Residues in lowercase are missing in the template VDR and PXR structures.

 
MD simulations with periodic boundary conditions were performed at 310 K using the program package GROMACS. The Particle-Mesh-Ewald method was applied for long-range electrostatic interactions, a cutoff of 0.9 nm was used for van der Waals interactions, and constant pressure was maintained at 105 pascal. The CAR model without any NRID peptide was equilibrated with decreasing constraints (from 1000 to 100 KJ/mol) on backbone atoms for 400 ps, followed by a free MD simulation of 2.25 ns. For the models with NRID peptides, an equilibration period of 250 ps with constraints of 1000 KJ/mol on the backbone atoms (excluding residues 24–52 because the VDR template was engineered in this part) was followed by a free MD simulation carried out for 2.25 ns. Bonds between heavy atoms and their corresponding hydrogen atoms were constrained to their equilibrium length using the LINCS algorithm. Because the interaction between the ligand-free CAR and NCoR is absent or very weak (24, 30), ANDR and EE2 were first docked into the LBP prior MD simulations (docking procedure described below). Trajectories of the free MD simulations were analyzed using NMRCLUST (49). Resulting clusters were examined and the representative frames were minimized with GROMACS.

The stereochemical quality of the models was evaluated using PROCHECK (50) and PROFILES-3D (51). The size of the LBP was determined using the program SURFNET (52). Ligands were built within SYBYL 6.9 considering their protonation state at pH 7.4 and minimized using the TRIPOS force field and default settings (Tripos Inc., St. Louis, MO). Docking procedures were performed with GOLD 2.1 (Cambridge Crystallographic Data Centre, Cambridge, United Kingdom). Goldscore was chosen as fitness function. For each ligand, docking runs were performed with a maximum allowed number of 30 poses. These were grouped into clusters (Table I). For dockings of clotrimazole and TMPP in CAR-SRC1 as well as ANDR and EE2 in CAR-NCoR models, MD simulations were performed with the settings described above. Missing parameters for the ligands were obtained manually using the GROMOS96 topology. Figures were prepared using VMD (53).


View this table:
[in this window]
[in a new window]
  		TABLE I
Docking of ligands into final homology models

 

   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
Modulation of Human CAR Activity—Fig. 1A shows that TMPP and clotrimazole enhanced the basal activity of GAL4-human CAR LBD by 2.3- and 1.8-fold, whereas EE2 and ANDR inhibited it by 40 and 30%, respectively. These results were expected from data on full-length human CAR (28–30, 32, 33). None of the chemicals changed the activity of GAL4 only. In competition assays (Fig. 1, B and C), the presence of TMPP and clotrimazole shifted the inhibition curves of EE2 and ANDR to the right, with 5–10-fold increases in apparent IC50 values. Therefore, the inhibitors and activators compete for the same or overlapping binding site within the human CAR LBD. Because the ligand response is a net result from interactions of human CAR with associated co-regulators, yeast two-hybrid assays with NRID peptides from SRC-1 and NCoR were set up. There was a very strong basal interaction between human CAR and the SRC-1 NRID (Fig. 1D) that was not much affected by TMPP, increased 35% by clotrimazole, and decreased 10–20% by the inhibitors ANDR and EE2. The very weak basal interaction between human CAR and the NCoR NRID was not affected by TMPP, modestly increased (3-fold) by clotrimazole, and strongly enhanced (over 12-fold) by inhibitors ANDR and EE2. Our data support the results that human CAR inhibitors act by promoting NCoR interaction and activators slightly increase SRC-1 interaction (24, 30, 31).



View larger version (18K):
[in this window]
[in a new window]
  		FIG. 1.
Modulation of human CAR activity. A, reporter activities of empty GAL4 and GAL4-human CAR LBD (hCAR) in HEK293 cells after exposure to the indicated modulators TMPP (10 µM), clotrimazole (2 µM), EE2 (10 µM), and ANDR (10 µM). B and C, dose-response curves of inhibitors EE2 (B) and ANDR (C) in the absence or presence of activators TMPP (10 µM) and clotrimazole (2 µM). D, the modulator-elicited responses in recruitment of SRC-1 (left panel) and NCoR peptides (right panel) with empty GAL4 and GAL4-human CAR LBD (hCAR) in yeast cells. The data are expressed as mean ± S.E. of at least three independent experiments.

 
Homology Models of Human CAR—The initial homology model was built on the VDR template to identify residues that form the human CAR LBP. Fig. 2 shows 22 residues that were selected for subsequent mutation analysis. Similarly to PXR, most of the residues lining the pocket (15/22) are hydrophobic, and polar residues (Asn165, Cys202, His203, Thr209, Tyr224, Tyr326, and Gln329) were rather evenly distributed. The four modulators were then docked to this model and MD simulations were carried out. However, the number of possible conformations obtained was too high for any meaningful interpretation of the functional data. This indicated that more careful selection and validation of the protein model is of utmost importance. After these initial models were completed, the human PXR crystal coordinates (45) became available, and improved knowledge of co-regulator binding of other NRs enabled us to create more advanced models with and without NCoR or SRC-1 NRID peptides (see Fig. 3, A and B). To account for the structural differences between PXR and CAR (8), the VDR structure (39) was used as scaffold for the modeling of the region connecting helices 1 and 3, and helices 6 and 7. Helices 10–12 were also built upon the VDR template. In the final model, the LBP is formed by helices 3, 5–7, and 10/11 and the {beta}-sheet. The 22 residues used in mutation analysis are lining the LBP as well (Fig. 3, C and D). The evaluation of the models indicated that they were of high stereochemical quality. In the Ramachandran plots, 86.4–92.2% of the {varphi}/{Psi} torsion angles were within the favorable region for all CAR models, the Profiles-3D scores were close (89.1–99.1%) to values expected for a high quality model of corresponding size, and the protein folds remained stable during MD simulations. Finally, only one or two preferred conformations of ligands (Table I) with excellent scoring values emerged.



View larger version (44K):
[in this window]
[in a new window]
  		FIG. 3.
Homology models of human CAR. A, superposition of LBDs in different CAR models. COOH-terminal helix 12 remains in its active conformation without (gray) or with SRC-1 (orange). In the presence of the SRC-1 peptide, helix 12 is pushed closer to the LBD. Helix 12 is displaced from its original position by the NCoR peptide (cyan). Here, only the COOH-terminal part of CAR is shown for clarity. NRID peptides were SLTERHKILHRLLQE (SRC-1) and NLGLEDIIRKALMGSFDDK (NCoR) with core sequences underlined. B, schematic representation of contacts between helix 12 and the rest of the LBD. The lines indicate interactions in the presence (solid) or absence (dotted) of SRC-1 peptide. C and D, detailed views of residues lining the LBP. Parts of the contributing structural elements are shown as tubes. LBP residues that have been mutated are shown in black, other residues are kept gray. C, side view of the LBP, lined by mainly hydrophobic amino acids. The hydrogen bond between Asn165 and Tyr326 is indicated. D, rotation about 90 degrees, view from the {beta}-sheet into the LBP.

 
These CAR LBD models (Fig. 3A) suggested that even without ligand, helix 12 adopts the active conformation because of hydrophobic interactions of Leu343 with LBD residues Tyr326 and Ile330, and of Ile346 with residue Val199. Of importance is the hydrogen bonding between Asn165 and Tyr326 that will stabilize Tyr326 in a position that allows interactions with Leu343 in helix 12 (Fig. 3B). The central role of residues Asn165, Val199, Tyr326, Ile330, and Leu343 for basal activity was confirmed by dramatic decreases in activity and SRC-1 interaction by their mutation to alanine (see Fig. 4, A and C). The presence of SRC-1 seemed to further stabilize helix 12 in the active conformation and push it closer to the LBD in such a way that Tyr326 could now contact with Ile346 (Fig. 3, A and B). Thus, helix 12 adapts to SRC-1 binding by improving the fit between the NRID peptide and the LBD. Another example of this adaptive fit was that Gln331 (helix 11) made a hydrogen bond to Ser348 (helix 12) only in the presence of SRC-1 (Fig. 3B), suggesting its involvement in proper aligning of helix 12. Subsequent mutagenesis of Q331A decreased the basal activity of human CAR by about 70% (data not shown). More detailed MD studies on the structural basis of constitutive activity are as described in Ref. 54. In the presence of NCoR, helix 12 could not acquire the active conformation (Fig. 3A).



View larger version (19K):
[in this window]
[in a new window]
  		FIG. 4.
The effects of LBP mutations on human CAR activity. A and B, the reporter activities were measured in HEK293 cells as described in the legend to Fig. 1 for activators (A) or inhibitors (B), and normalized to empty GAL4 values (set at 1.0). C and D, the same constructs were then assayed in yeast for interactions with NRID peptides SRC-1 (C) or NCoR (D). The SRC-1 and NCoR results were normalized to wild-type CAR treated with vehicle (=100, C) or with 10 µM ANDR (=100, D), respectively. The data are expressed as mean ± S.E. of three to six independent experiments.

 
Significantly, the size of the LBP in the final model (480 Å3) was smaller than that of crystallized PXR (1294 Å3) (45) and of previous CAR models (1150–1170 Å3) that were built on PXR but not validated by MD simulations (26, 55, 56). The small size of LBP is because of protruding, bulky aromatic (e.g. Phe129, Phe161, Phe217, Tyr224, Phe234, Phe238, and Tyr326) and hydrophobic (e.g. Ile164, Met168, Val199, and Ile330) residues (Fig. 3, C and D). The smaller size of the CAR pocket relative to PXR is also consistent with the fact that fewer and smaller chemicals may act as ligands for CAR than for PXR (12, 13).

Basal Activities of Human CAR Mutants—The activities of the 22 mutants were measured and compared with those of the wild-type human CAR (Fig. 4, A and B). Interestingly, the basal activities (white columns) of 16 mutants were decreased to 10% or less of the wild-type activity. Our interpretation is that mutation of aromatic (Phe161, Phe234, Phe238, and Tyr326) or hydrophobic (Cys202, Ile164, Met168, Ile330, Ile333, and Met339) residues that protrude into the LBP to alanine will create more space and reorganize the surrounding LBP residues. This re-organization would decrease the basal activity in most cases via the central residue Tyr326, the position of which is crucial for the stabilization of helix 12 in active conformation. Second, at least three other residues (Val199, Asn165, and Leu343) are also involved in stabilization of helix 12 (see above); accordingly, their mutation also decreased the basal activity. Third, three mutants (F129A, F217A, and Y224A) also had low basal activity, but they could not be activated in mammalian cells (Fig. 4A) or could not elicit a SRC-1 or NCoR response in yeast (Fig. 4, C and D) like the other 19 mutants did. Residues Phe217 and Tyr224 form a wall in the LBP (Fig. 3C), and their change to alanine is likely to disrupt protein folding locally. Finally, the remaining six mutants (H203A, L206A, T209A, L242A, F243A, and Q329A) retained 30–80% of the wild-type activity. These residues were located on the top and back of LBP (Fig. 3, C and D), and apart from His203, their side chains were not clearly projected to the center of LBP or the central residue Tyr326. Therefore, their larger distance from Tyr326 was consistent with the finding that their mutations could not profoundly attenuate the basal activity via influence on position of Tyr326.

The majority of mutants (13/16) with most dramatic losses in basal activity also had corresponding losses in SRC-1 interaction (Fig. 4C, white columns). The remaining three low activity mutants (N165A, C202A, and I333A) showed only modest reduction in SRC-1 interaction (20–40%). Of these three exceptions, N165A interacted strongly with NCoR even without any ligand (Fig. 4D). This increase in co-repressor binding explains well the loss of net activity of this mutant in mammalian cells. The reason for low basal activities of C202A and I333A was not directly apparent from the NCoR and SRC-1 recruitment. Therefore, we tested the possibility that these two mutants differed in their association with co-activators other than SRC-1. The recruitment of NRID peptides from thyroid hormone receptor-associated protein or PGC-1 was significantly lower by C202A and I333A mutants than by the wild-type CAR (Fig. 5A), providing a basis for their low net activity in mammalian cells that are known to express a variety of co-activators. In the remaining six mutants retaining about one-third or more of wild-type activity (H203A, L206A, T209, L242A, F243A, and Q329A), the correspondence between SRC-1 association and co-transfection assay was good (T209A, F243A; Fig. 4, A and C) or compensated for by progressively larger decreases in PGC-1 interaction (H203A, L206A, L242A; Fig. 5A), thus leaving only one mutant (Q329A) out of 22 without a clear match between co-transfection and yeast two-hybrid results.



View larger version (42K):
[in this window]
[in a new window]
  		FIG. 5.
Differential association of coactivators with human CAR mutants. A, the basal interaction of CAR and selected mutants with NRID peptides from coactivators SRC-1 (black), thyroid hormone receptor-associated protein (TRAP) (gray), and PGC-1 (white). The assays were conducted as described in the legend to Fig. 4C, and normalized to wild-type CAR (=100). B, the EE2- and ANDR-elicited interactions of CAR and selected mutants with NRID peptide from coactivator SRC-1. The assays were done as described in the legend to Fig. 4C, and normalized to vehicle-treated wild-type CAR (=100). The data are expressed as mean ± S.E. of three independent experiments.

 
Docking and MD Simulation of Ligand Binding—To find determinants of ligand binding, the four modulators were docked into appropriate human CAR/NRID peptide models, and MD simulations were carried out. Docking of TMPP into CAR-SRC-1 yielded 18 favored poses that could be grouped in two clusters that differed only slightly from each other. For clotrimazole, 19 favored poses showing also just minor differences between each other were obtained (Table I). The two different binding orientations were analyzed visually and compared considering their scoring values. The top-ranking pose for TMPP and clotrimazole was then further investigated by MD simulation. A representative frame of the MD simulation, calculated using NMRCLUST (49), of the CAR/SRC-1 model with bound activators is shown in Fig. 6, A and B, respectively. Clotrimazole was bound deep in the LBP without any contact to helix 12. The phenyl rings of clotrimazole made contacts with residues Phe161, Phe217, Tyr224, Phe234, and Tyr326, whereas no LBP residue formed a clear hydrogen bond with the imidazole ring. During MD simulation, the distance between Tyr326 and helix 12 was decreased. Phe161 reoriented to the interface between helix 12 and LBD, interacting with the hydrophobic Met339, and Phe243 now pointed away from the LBP. All these movements enlarged the LBP to a volume of 750 Å3. TMPP shows a different binding mode (Fig. 6A). It contacted residues Ile164, Asn165, Met168, Val199, Cys202, His203, Phe217, and Tyr224. In addition, one of the methylphenyl groups interacted directly with helix 12 residues Leu343 and Ile346. Two other methylphenyl groups contacted Phe234 and Tyr326 and also interacted with Phe161, pushing it deeper into the LBP. TMPP was also held in place by strong hydrogen bonds between the phosphate group and Asn165 and Tyr326 that remained stable during the entire MD simulation. Upon TMPP binding, helix 12 moved closer to LBD, and the LBP size increased to about 560 Å3. In summary, the binding of activators caused an expansion of the LBP volume, and significantly, helix 12 was packed closer to the LBD. This shift of helix 12 was expected to improve subsequent binding of the SRC-1 peptide (compare with Fig. 3A).



View larger version (40K):
[in this window]
[in a new window]
  		FIG. 6.
The conformations of human CAR modulators within the LBP. A, activators clotrimazole (gray) and TMPP (black) were docked and simulated in the LBP of CAR-SRC-1 model. Clotrimazole is located deep within the LBP without any hydrogen bonds or direct contacts to helix 12. TMPP interacts with helix 12 and is bound by hydrogen bonds to Asn165 and Tyr326. In both cases, ligand binding decreases the distance between Tyr326 and helix 12 as compared with ligand-free CAR. B, inhibitors ANDR (gray) and EE2 (black) docked in the LBP of CAR-NCoR model. A stable hydrogen bond is formed between the 3-hydroxyl group of EE2 and Asn165, whereas the 3{alpha}-hydroxyl of ANDR was not involved in hydrogen bonding. C, comparison of CAR with docked EE2 (light gray) and a representative frame from the subsequent MD simulation (black). After MD, the position of EE2 showed only a minor change. Phe161 was reoriented toward the interface between LBD and helix 12. The distance between the LBD and helix 12 was increased (black lines, see also D). Parts of the protein backbone were removed for clarity. D, distances between C{alpha} atoms of Tyr326 and Leu343 were plotted for each frame of the MD. In empty CAR (gray), this distance is about 10.7 Å. Higher values are found at the beginning of the MD because of equilibration of the model. In complex with docked EE2 (black), the distance is increased by about 2 Å with much larger fluctuations.

 
In the CAR/NCoR model, only one energetically favorable solution was obtained when ANDR or EE2 were docked. In both cases, the ligand was interacting with the same set of side chains: Phe161, Asn165, and Met168 from helix 3, Val199, Cys202, His203, and Leu206 from helix 5, and Phe217, Tyr224, Phe234, and Tyr326 (Fig. 6B). However, the plane and orientation of the steroids were quite opposite: the A-ring of ANDR was located in the same region as the D-ring of EE2, and their C18 methyl groups were facing helix 5 and helix 3, respectively. ANDR did not form any clear hydrogen bonds, whereas the 3-hydroxyl of EE2 made a stable bond with Asn165. The conformation of EE2 suggested that EE2 might inhibit CAR by disturbing the hydrophobic interactions between helix 12 and the LBD with the protruding 3-hydroxyl group that pointed toward Leu343 in the CAR/SRC-1 model. MD simulation of the CAR-EE2 complex revealed that binding of EE2 to CAR LBD increased the distance between Tyr326 and helix 12, whereas EE2 itself rotated by 30 degrees along its long axis and forced reorientation of Phe161 (Fig. 6, C and D). Similar weakening of interactions between helices 11 and 12 because of reorientation of Tyr326 was seen in the CAR/SRC-1 model upon EE2 addition (data not shown). For ANDR inhibition, a similar mechanism could not be suggested. Nevertheless, Tyr326 was located much deeper in LBP, increasing its distance from helix 12 (Fig. 6B).

Ligand Specificities of Human CAR Mutants—Next, the responses of 22 mutants to activators were measured in co-transfection and yeast two-hybrid assays (Fig. 4, A and C). We first considered mutants with low basal activity: ligand responsiveness was totally lost by only three mutants (F129A, F217A, and Y224A). Four mutants (V199A, I330A, M340A, and L343A) did not respond to activators but inhibitors provoked their association with NCoR (Fig. 4D). Three of these residues (Val199, Ile330, and Leu343) were found to keep helix 12 in the active position (see above), which explains the lack of SRC-1 recruitment by these mutants even after addition of an activator. The remaining 15 mutants had either low-to-moderate basal activity that could be increased by TMPP and/or clotrimazole to variable degrees (Fig. 4A). Of these, eight mutants (M168A, C202A, L206A, T209A, L242A, F243A, Q329A, and I333A) were activated by TMPP to the same or higher degree than by clotrimazole, thus resembling the preference of wild-type human CAR. These residues were located on the top and back of the LBP (Fig. 3, A and B), away from the residues in contact with the activators (Fig. 6A), which probably explains why these mutations had only weak effects on activator preference. Four mutants were activated by TMPP only (H203A, F234A, F238A, and Y326A). These aromatic residues form hydrophobic and/or stacking interactions with clotrimazole in MD simulations (Fig. 6A) and their replacement with alanine explains the loss of activation. Finally, three helix 3 mutants (F161A, I164A, and N165A) preferred clotrimazole over TMPP (Fig. 4, A and C). A likely explanation is that residues Ile164 and Asn165 interact with the TMPP phosphate and methylphenyl groups (but not with clotrimazole) and these interactions would be lost upon mutation (Fig. 6A). Because Phe161 interacts with both TMPP and clotrimazole, it is easy to see why the F161A mutation decreased activation and SRC-1 recruitment by both activators (Fig. 3, A and C).

Responses to human CAR inhibitors were investigated next (Fig. 4, B and D). Among the 16 low-activity mutants, any further inhibition was difficult to observe. However, two helix 3 mutants displayed significant activation by EE2 only (N165A) or by both inhibitors (M168A). These results were supported by yeast two-hybrid assays. First, only ANDR but not EE2 further enhanced NCoR recruitment by N165A, whereas the NCoR responses of M168A were greatly decreased (EE2) or abolished (ANDR) (Fig. 4D). Second, both EE2 and ANDR increased interactions of these mutants with SRC-1 (Fig. 5B). In NCoR assays, three additional mutants (V199A, C202A, and I330A) were responsive to EE2 but not to ANDR. Among the five mutants with moderate basal activity, none were inhibited by EE2 or ANDR anymore; in fact, there was about a 2-fold activation by either steroid (L242A and F243A) or both (H203A). Consistent with the loss of inhibition, the responses to EE2 and ANDR of these mutants in NCoR assays were attenuated by 70–90% (Fig. 4D). SRC-1 assays with EE2 and ANDR indicated a clear-cut activation of F243A by EE2 only (Fig. 5B). Taken together, three amino acids (Asn165, Met168, and Phe243) appear to regulate the response of EE2 consistently in both types of assays, whereas additional residues (Val199, Cys202, and Ile330) may contribute to recognition of ANDR. Of these residues, at least Asn165, Met168, Val199, and Cys202 are located close to the steroids, and residue Asn165 formed a hydrogen bond with EE2 in MD simulations (Fig. 6B). Because ANDR was not fixed by any hydrogen bonds, mutation of closely located residues Val199, Cys202, and Ile330 to alanine may create sufficient space to completely reorient ANDR but not EE2, explaining the selectivity of these residues.

In summary, based on carefully constructed molecular models, we were able to identify, for the first time, residues critical for ligand binding and to explain the effects of mutations with all types of human CAR modulators. This was not possible with simpler, single template-based models or without MD simulations. To validate our models further, we docked TPP, a derivative of TMPP lacking the three methyl groups, into CAR models and assayed the activities of selected human CAR mutants. The applied docking procedure led to two different binding orientations (Table I). One binding mode resembled the one obtained for TMPP that displayed hydrogen bonds with Tyr326 and Asn165. In the second cluster, the phosphate group of TPP was reoriented and made a hydrogen bond to His203 instead (Fig. 7A). Docking of TPP into the CAR-SRC-1 model resulted in a similar pose showing the same hydrogen bond to His203. This suggested that TMPP and TPP interact differentially with residues Asn165 and Tyr326. The N165A mutation was expected to eliminate hydrogen bonding to Tyr326 and increase its flexibility, thereby decreasing TMPP-induced CAR activity, but not affect TPP-induced activity very much. This prediction was borne out exactly in activation assays (Fig. 7B).



View larger version (29K):
[in this window]
[in a new window]
  		FIG. 7.
The predicted interactions of modulator derivatives within the human CAR LBP. A, docked conformations of TPP (red) and TMPP (blue) in CAR LBP. Unlike TMPP, TPP is not bound to Asn165 or Tyr326 but it formed a hydrogen bond with His203. Helix 12 is shown in orange ribbon. B, the TMPP- and TPP-elicited activities of selected mutants were measured as described in the legend to Fig. 1. C, docked conformations of E2 (green), EE2 (red), and mestranol (blue) in CAR LBP. Residues important for ligand recognition are depicted with capped sticks. D, the estrogen-elicited activities of selected mutants were measured as described in the legend to Fig. 1.

 
In another test, derivatives of EE2 were studied (Fig. 7C). Docking of mestranol (3-hydroxyl is methylated) into CAR resulted in just one favorable conformation, whereas for E2 (lacking the 17{alpha}-ethynyl group) two different binding orientations were observed (Table I). About 80% of all poses could be grouped in a major cluster in which the 17{beta}-hydroxyl group of E2 pointed toward the helix 12/LBD interface. The minor cluster revealed a reversed orientation of the steroid showing much lower scoring values. For EE2, 23 poses in two different clusters were obtained. In the top-ranking pose, the 3-hydroxyl group of the steroid was pointing toward the helix 12/LBD interface (i.e. a reversed binding mode compared with E2 and mestranol). This conformation of EE2 is similar to the one obtained when docking the ligand into the CAR/NCoR model. The second cluster contains poses with lower scores and a conformation similar to that obtained for mestranol. Keeping in mind the top-ranking binding orientations of E2 and EE2 (Fig. 7C) and the above suggested mechanism of EE2 inhibition, these calculations imply that E2 and EE2 would inhibit human CAR but mestranol would not. The experimental results confirmed this prediction (Fig. 7D). Because both Asn165 and Phe243 residues were thought important for EE2 recognition (see above), we also deduced that all of these estrogens would be activators of the N165A mutant, regardless of their orientation. This would be the case because of lack of hydrogen bonding between Ala165 and Tyr326. The effect of F243A would, on the other hand, depend on how well the estrogen was fixed into the LBP by hydrophobic interactions. In co-transfections, N165A was activated by 2–6-fold by estrogens, whereas the activity of F243A was enhanced 2–6-fold by EE2 and mestranol but not by E2 (Fig. 7D). We thought that the inactivity of E2 was because of its complete reorientation within the LBP of the F243A mutant but mestranol was still able to interact with and stabilize Tyr326 via its 17{alpha}-ethynyl group.

Previously we have found, based on mouse/human CAR chimeras, that replacement of human CAR amino acids 190–253 with corresponding mouse residues converted EE2 from an inhibitor to an activator (24). In this region, LBP residues that differed between the human and mouse CAR were C202L, L242I, and F243L. We made these human-to-mouse mutations and analyzed the effect of EE2 on CAR activity. We found that although inhibition by EE2 was reduced by C202L and L242I, only F243L mutation was clearly activated by EE2 (Fig. 8). In all, the critical role of Phe243 in EE2 recognition was supported by docking into homology models, mutagenesis studies, and naturally occurring species variation.



View larger version (11K):
[in this window]
[in a new window]
  		FIG. 8.
The species difference in CAR response to EE2. The reporter activities of the indicated CAR constructs were transfected into HEK293 cells and treated with vehicle or 10 µM EE2 and assayed as described in the legend to Fig. 1. The data are expressed as mean ± S.E. of three independent experiments.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Factors Contributing to Basal Activity of CAR—Our homology models of CAR indicated that residues Ile346 and Leu343 in helix 12 made hydrophobic contacts with Val199 (helix 5) and Tyr326 and Ile330 (helix 11), respectively. Furthermore, the central residue Tyr326 was stabilized by Asn165. All these interactions contributed to helix 12 acquiring the active position, because mutation of each residue of alanine caused marked decreases in the basal activity. In addition, novel contacts between helices 11 and 12 were formed upon SRC-1 binding (Gln331-Ser348, Tyr326-Ile346). Other NRs with significant basal activity such as Nurr1 (57), ERR{gamma} (58), ROR{beta} (59), and LRH-1 (60) also have similar contacts in their crystal structures (see Table II). However, the interaction in each case may involve not only hydrophobic contacts but also hydrogen bonding, aromatic stacking, and even salt bridges. Even though residues corresponding to the pair Ile330-Leu343 are often seen, the interacting amino acid pairs do not uniformly match the pairs in human CAR. Nevertheless, these findings indicate that the presence of extensive contacts between helices 11 and 12 may be fundamental to basal activity of NRs in general. The possibility of such interactions has been pointed out in other homology models of CAR although no MD simulations had been conducted (26, 61). We were not able to detect any hydrogen bonds between Cys347 and Tyr326, as described for mouse and human CAR (26, 61), but we agree on the importance of a hydrophobic contact between Ile330 and Leu343 in human CAR (26).


View this table:
[in this window]
[in a new window]
  		TABLE II
The interactions between helices 11 and 12 in constitutively active NRs

 
A charge clamp between Lys205 in helix 4 and the negatively charged carboxyl terminus (Ser358) was suggested important for basal activity (55). This notion was based on a mouse CAR model built on PXR and loss of activity because of extension of helix 12 by one helical turn (55). We could not detect such an interaction in any of our models. The residue corresponding to Lys205 (Lys195 in human CAR) is highly conserved among NRs and it has well documented interactions with NRID peptides in other NR crystal structures (e.g. Refs. 45 and 62). In a crystal structure of an agonist-bound NR that lacks the SRC-1 peptide, the corresponding lysine residue does not contact helix 12 at all (39). In addition, extension of helix 12 by three residues in mouse CAR did not influence the basal activity (61). Mutation of Lys205 is therefore likely to decrease basal activity of mouse CAR by loss of SRC-1 binding rather than by destabilization of helix 12. On another note, residue Asn165 (helix 3) exerts its effects through Tyr326, whereas residue Val199 (helix 5) appears to contribute to the stability of helix 12 with our CAR and CAR/SRC-1 models. Although the role of Asn165 may be of unique importance to CAR, hydrophobic contacts between residues matching Val199 and helix 12 are present at least in VDR (Ile268-Phe422) and PXR (Phe281-Phe429) structures (39, 45).

The above data indicate that LBD residues that interact with helix 12 also influence SRC-1 and/or NCoR binding. This view supports reports on both naturally occurring or alanine scanning mutants of other NRs such as RAR{alpha} and thyroid hormone (63–66). For example, RAR{alpha} alanine scanning (66) indicated that mutation of amino acids in close proximity to helix 12 such as L266A (Val199 in CAR) and V395A (Ile330) or indirectly affecting helix 12 such as T233A (Asn165) exhibit no or inefficient release of co-repressors upon agonist binding. Even though the same residues are implied, the actual effect of corresponding mutations in CAR was opposite: N165A mutation increases co-repressor recruitment in CAR but the corresponding mutation T233A decreases it in RAR{alpha}, whereas the others (V199A, Q329A, and I330A) cannot interact with NCoR. This difference in the direction of effect is very likely because of constitutive interaction of CAR with co-activators and RAR{alpha} with co-repressors.

Ligand Specificity of Human CAR—To our knowledge, this is the first report that addresses the LBP residues affecting the ligand specificity of human CAR. Table III summarizes the selective effects of LBP residues on ligand specificity. Of importance are the roles of helix 3 residues in regulating the preference to TMPP, the requirement of several aromatic residues for clotrimazole recognition, and finally, the distinction between EE2 and ANDR by residues Cys202 and Phe243. A further support to our models came from the finding that the nature of residue 243 regulates species-specific recognition of EE2. The single mutation of Phe243 to the corresponding leucine in mouse CAR converted EE2 to an activator. As expected from ligand-dependent effects of mutations described in Table III, it should be emphasized that Phe243 probably is not the only regulator of human/mouse differences. Indeed, activation by the mouse CAR-selective agonist TCPOBOP seems to depend on residue Thr350 (24).


View this table:
[in this window]
[in a new window]
  		TABLE III
Summary of selective ligand responses affected by specific CAR LBP residues

 
We noticed intriguing differences when comparing the CAR LBP to that of VDR and PXR, the templates for our homology models. The residue corresponding to the key residue Tyr326 in CAR contacts the ligand in PXR (His407) and in VDR (His397) structures and their mutation appears to modulate ligand selectivity (45, 67) or eliminate ligand binding (68, 69), respectively. Residue Ser237 in VDR is important for both ligands 1{alpha},25-dihydroxyvitamin D and lithocholic acid (68, 70) but the corresponding Met168 in CAR shows selectivity to CAR inhibitors only. Similarly, helix 5 residues are important to the ligand selectivity of VDR (Ser274 and Ser278 (70)) but they have only a minor impact on CAR. Finally, residues Asn165 or Phe243 that are crucial to CAR are not implicated in VDR or PXR ligand selectivity at all. This indicates that even though most of the LBP residues may correspond spatially, their functions differ tremendously.

In conclusion, we have developed well defined homology models for human CAR. When coupled with extensive functional analysis, these models have helped us to suggest mechanisms that contribute to the high basal activity of human CAR, to identify residues that impart selectivity to the CAR ligand recognition and finally, have given the first explanations to the wide species differences in CAR ligand specificity.


   FOOTNOTES
 
* This work was supported by Academy of Finland Grants 44040 and 51610 (to P. H.) and 104622 (to M. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

§ Both authors contributed equally to this work. Back

{ddagger}{ddagger} To whom correspondence should be addressed: Dept. of Pharmaceutics, University of Kuopio, P. O. Box 1627, FIN-70211 Kuopio, Finland. Tel.: 358-17-162490; Fax: 358-17-162252; E-mail: paavo.honkakoski{at}uku.fi.

1 The abbreviations used are: NR, nuclear receptor; ANDR, 3{alpha}-androstenol; CAR, constitutive androstane receptor; EE2, 17{alpha}-ethynyl-3,17{beta}-estradiol; LBD, ligand-binding domain; LBP, ligand-binding pocket; MD, molecular dynamics; NCoR, nuclear receptor corepressor; NRID, NR interaction domain; PGC-1, PXR, pregnane X receptor; RAR, retinoic acid receptor; SRC-1, steroid receptor coactivator-1; TMPP, tri-(p-methylphenyl)phosphate; TPP, triphenyl phosphate; VDR, vitamin D receptor. Back


   ACKNOWLEDGMENTS
 
We thank CSC-Scientific Computing Ltd. (Espoo, Finland) for the SYBYL and GOLD software and hardware resources.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Mangelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Schütz, G., Umesono, K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., and Evans, R. M. (1995) Cell 83, 835-939[CrossRef][Medline] [Order article via Infotrieve]
  2. Chawla, A., Repa, J. J., Evans, R. M., and Mangelsdorf, D. J. (2001) Science 294, 1866-1870[Abstract/Free Full Text]
  3. Aranda, A., and Pascual, A. (2001) Physiol. Rev. 81, 1269-1304[Abstract/Free Full Text]
  4. Renaud, J. P., and Moras, D. (2000) Cell. Mol. Life Sci. 57, 1748-1769[CrossRef][Medline] [Order article via Infotrieve]
  5. Weatherman, R. V., Fletterick, R. J., and Scanlan, T. S. (1999) Annu. Rev. Biochem. 68, 559-581[CrossRef][Medline] [Order article via Infotrieve]
  6. Steinmetz, A. C., Renauld, J. P., and Moras, D. (2001) Annu. Rev. Biophys. Struct. 30, 329-359[CrossRef][Medline] [Order article via Infotrieve]
  7. Robinson-Rechavi, M., Carpentier, A., Duffraisse, M., and Laudet, V. (2001) Trends Genet. 17, 554-556[CrossRef][Medline] [Order article via Infotrieve]
  8. Moore, L. B., Maglich, J. M., McKee, D. D., Wisely, B., Willson, T. W., Kliewer, S. A., Lambert, M. H., and Moore, J. T. (2002) Mol. Endocrinol. 16, 977-986[Abstract/Free Full Text]
  9. Ueda, A., Hamadeh, H. K., Webb, H. K., Yamamoto, Y., Sueyoshi, T., Afshari, C. A., Lehmann, J. M., and Negishi, M. (2002) Mol. Pharmacol. 61, 1-6[Abstract/Free Full Text]
  10. Honkakoski, P., Zelko, I., Sueyoshi, T., and Negishi, M. (1998) Mol. Cell. Biol. 18, 5652-5658[Abstract/Free Full Text]
  11. Wei, P., Egan-Hafley, M., Liang, S., and Moore, D. D. (2000) Nature 407, 920-923[CrossRef][Medline] [Order article via Infotrieve]
  12. Willson, T. M., and Kliewer, S. A. (2002) Nature Rev. Drug Discov. 1, 259-266[CrossRef][Medline] [Order article via Infotrieve]
  13. Honkakoski, P., Sueyoshi, T., and Negishi, M. (2003) Ann. Med. 35, 172-182[CrossRef][Medline] [Order article via Infotrieve]
  14. Maglich, J. M., Stoltz, C. M., Goodwin, B., Hawkins-Brown, D., Moore, J. T., and Kliewer, S. A. (2002) Mol. Pharmacol. 62, 638-646[Abstract/Free Full Text]
  15. Rosenfeld, J. M., Vargas, R., Jr., Xie, W., and Evans, R. M. (2003) Mol. Endocrinol. 17, 1268-1282[Abstract/Free Full Text]
  16. Rodrigues, D., A., and Lin, J. H. (2001) Curr. Opin. Chem. Biol. 5, 396-401[CrossRef][Medline] [Order article via Infotrieve]
  17. Guo, G. L., Lambert, G., Negishi, M., Ward, J. M., Brewer, H. B., Jr., Kliewer, S. A., Gonzalez, F. J., and Sinal, C. J. (2003) J. Biol. Chem. 278, 45062-45071[Abstract/Free Full Text]
  18. Huang, W., Zhang, J., Chua, S. S., Qatani, M., Han, Y., Granata, R., and Moore, D. D. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 4156-4161[Abstract/Free Full Text]
  19. Yamamoto, Y., Kawamoto, T., and Negishi, M. (2003) Arch. Biochem. Biophys. 409, 207-211[CrossRef][Medline] [Order article via Infotrieve]
  20. Baes, M., Gulick, T., Choi, H.-S., Martinoli, M. G., Simha, D., and Moore, D. D. (1994) Mol. Cell. Biol. 14, 1544-1552[Abstract/Free Full Text]
  21. Forman, B. M., Tzameli, I., Choi, H. S., Chen, J., Simhal, D., Seol, W., Evans, R. M., and Moore, D. D. (1998) Nature 395, 612-615[CrossRef][Medline] [Order article via Infotrieve]
  22. Tzameli, I., Pissios, P., Schuetz, E. G., and Moore, D. D. (2000) Mol. Cell. Biol. 20, 2951-2958[Abstract/Free Full Text]
  23. Min, G., Kemper, J. K., and Kemper, B. (2002) J. Biol. Chem. 277, 26356-26363[Abstract/Free Full Text]
  24. Jyrkkärinne, J., Mäkinen, J., Gynther, J., Savolainen, H., Poso, A., and Honkakoski, P. (2003) J. Med. Chem. 46, 4687-4695[CrossRef][Medline] [Order article via Infotrieve]
  25. Shiraki, T., Sakai, N., Kanaya, E., and Jingami, H. (2003) J. Biol. Chem. 278, 11344-11350[Abstract/Free Full Text]
  26. Frank, C., Molnar, F., Matilainen, M., Lempiäinen, H., and Carlberg, C. (2004) J. Biol. Chem. 279, 22558-22566[Abstract/Free Full Text]
  27. Wei, P., Zhang, J., Dowhan, D. H., Han, Y., and Moore, D. D. (2002) Pharmacogenomics J. 2, 117-126[CrossRef][Medline] [Order article via Infotrieve]
  28. Mäkinen, J., Reinisalo, M., Niemi, K., Viitala, P., Jyrkkärinne, J., Chung, H., Pelkonen, O., and Honkakoski, P. (2003) Biochem. J. 376, 465-472[CrossRef][Medline] [Order article via Infotrieve]
  29. Kawamoto, T., Kakizaki, S., Yoshinari, K., and Negishi, M. (2000) Mol. Endocrinol. 14, 1897-1905[Abstract/Free Full Text]
  30. Mäkinen, J., Frank, C., Jyrkkärinne, J., Gynther, J., Carlberg, C., and Honkakoski, P. (2002) Mol. Pharmacol. 62, 366-378[Abstract/Free Full Text]
  31. Maglich, J. M., Parks, D. J., Moore, L. B., Collins, J. L., Goodwin, B., Billin, A. N., Stolzt, C. A., Kliewer, S. A., Lambert, M. H., Willson, M. T., and Moore, J. T. (2003) J. Biol. Chem. 278, 17277-17283[Abstract/Free Full Text]
  32. Honkakoski, P., Palvimo, J. J., Penttilä, L., Vepsäläinen, J., and Auriola, S. (2004) Biochem. Pharmacol. 67, 97-106[CrossRef][Medline] [Order article via Infotrieve]
  33. Moore, L. B., Parks, D. J., Jones, S. A., Bledsoe, R. K., Consler, T. G., Stimmel, J. B., Goodwin, B., Liddle, C., Blanchard, S. G., Willson, T. M., Collins, L. J., and Kliewer, S. A. (2000) J. Biol. Chem. 275, 15122-15127[Abstract/Free Full Text]
  34. Toell, A., Kröncke, K. D., Kleinert, H., and Carlberg, C. (2002) J. Cell. Biochem. 85, 72-82[CrossRef][Medline] [Order article via Infotrieve]
  35. Smith, C. L., Nawaz, Z., and O'Malley, B. W. (1997) Mol. Endocrinol. 11, 657-666[Abstract/Free Full Text]
  36. Lin, Z., Auboeuf, D., Wong, J., Chen, J. D., Tsai, S. Y., Tsai, M. J., and O'Malley, B. W. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 7940-7944[Abstract/Free Full Text]
  37. Yuan, C. X., Ito, M., Fondell, J. D., Fu, Z. Y., and Roeder, R. G. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 7939-7944[Abstract/Free Full Text]
  38. Puigserver, P., Wu, Z., Park, C., Graves, R., Wright, M., and Spiegelman, B. M. (1998) Cell 92, 829-839[CrossRef][Medline] [Order article via Infotrieve]
  39. Rochel, N., Wurtz, J. M., Mitschler, A., Klaholz, B., and Moras, D. (2000) Mol. Cell 5, 173-179[CrossRef][Medline] [Order article via Infotrieve]
  40. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) Nucleic Acids Res. 22, 4673-4680[Abstract/Free Full Text]
  41. Ponder, J. W., and Richards, F. M. (1987) J. Mol. Biol. 193, 775-791[CrossRef][Medline] [Order article via Infotrieve]
  42. Brooks, B. R., Bruccoleri, R. E., Olafson, B. D., States, D. J., Swaminathan, S., and Karplus, M. (1983) J. Comp. Chem. 4, 187-217[CrossRef]
  43. Cornell, W. D., Cieplak, P., Bayly, C. I., Gould, I. R., Merz, K. M., Ferguson, D. M., Spellmeyer, D. C., Fox, T., Caldwell, J. W., and Kollman, P. A. (1995) J. Am. Chem. Soc. 117, 5179-5197[CrossRef]
  44. Bayly, C. I., Cieplak, P., Cornell, W. D., and Kollman, P. A. (1993) J. Phys. Chem. 97, 10269-10280[CrossRef]
  45. Watkins, R. E., Davis-Searles, P., Lambert, M. H., and Redinbo, M. R. (2003) J. Mol. Biol. 331, 815-828