| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Institut National de la Santé et de la Recherche Médicale (INSERM) U-439, Pathologie Moléculaire des Récepteurs Nucléaires et Service d’Hormonologie, Centre Hospitalier Universitaire (CHU) de Montpellier (N.P., S.L., B.T., J.-M.A.L., C.S.); Centre de Biochimie Structurale, INSERM U-554 and Centre National de la Recherche Scientifique U-5048, Montpellier (N.P., W.B.); and Unité d’Endocriologic Pédiatrique, CHU (C.S.), 34295 Montpellier, France
Address all correspondence and requests for reprints to: Prof. Charles Sultan, INSERM, U-439, 70 rue de Navacelles, 34090 Montpellier, France. E-mail: chsultan{at}montp.inserm.fr.
Abstract
To decipher the clues of genotype-phenotype mapping in androgen insensitivity syndromes (AIS), we integrated clinical, molecular, and structural data in an investigation into the characteristics of androgen receptor (AR) ligand binding and activation. We looked for residues substituted in AIS that are conserved among the different AR species but that diverge in the other steroid receptors, thus suggesting a role in androgen binding specificity. Of the residues fitting these characteristics, we focused on the glycine at position 743, for which natural substitutions (glutamic acid and valine) have been associated with different androgen resistance phenotypes. The consequences of both substitutions were evaluated along with the minimal glycine to alanine mutation. The gradual impairment of binding and trans-activation efficiencies in AR mutants ranging from alanine to valine and subsequently glutamic acid were highlighted by in vitro experiments. Structural analyses showed the consequences of these substitutions at the atomic level and suggested a difference in local organization among the nuclear receptor superfamily. Overall, this integrative approach provides a useful tool for further investigations into the AR structure-function relationship in AIS.
THE ANDROGEN RECEPTOR (AR) is a transcription factor involving high affinity androgen binding to initiate various molecular events, leading to the specific gene activation required for male sex development. The multiplicity of AR gene mutations identified in patients presenting androgen resistance (1) attests to the crucial role of this nuclear receptor. Androgen insensitivity syndrome (AIS) encompasses a wide spectrum of male pseudohermaphroditisms ranging from complete AIS (CAIS) in subjects with female phenotype to partial AIS (PAIS) in men with infertility and/or stigmata of undervirilization (2).
AR is a member of the nuclear receptor superfamily that includes receptors for steroids and thyroid hormones, vitamin D3 and retinoic acids, and numerous orphan receptors for which no ligand has yet been identified (3). Hence, AR can be divided into separable domains with specific functions, such as ligand binding, dimerization, DNA binding, and trans-activation. Hormone binding induces a transconformation of the receptor and allows its translocation into the nucleus, where it initiates transcription through specific interaction with the transcription machinery (for review, see Ref.4).
The crystal structures of unliganded and liganded ligand-binding domains (LBD) of various nuclear receptors have been solved (for review, see Refs.5 and6) and have revealed a similar fold, with the major difference between the apo and holo state being the position of helix H12 that encompasses residues of the activating domain of the transcription activation function 2. In a previous work we developed a model of the AR LBD (7) based on homology with the liganded progesterone receptor LBD crystal structure (8). More recently, two crystal structures of the AR LBD were published (9, 10), indicating that our homology model was correct concerning the anchoring of the 3-keto group of the androgen ligand by R752 and Q711 and the role of N705 as the anchoring site of the 17ß-hydroxyl group.
The number of natural mutations occurring in patients with complete or partial AIS provides a powerful model to study the structure-function relationship of AR. Using amino acid alignments, crystal structures, and three-dimensional models, we identified divergent residues potentially involved in the specificity of activation of nuclear receptors. Among them, glycine 743 seems to be of particular interest. Two different natural substitutions at this position have been associated with different phenotypes: G743V has been previously reported in PAIS (11, 12), and we describe here a G743E substitution identified in a patient with CAIS. An integrative approach was thus developed to evaluate the consequences of the replacement of G743 for the structure and function of the AR. For this purpose we introduced the valine or glutamic acid substitutions in the AR expression vector along with an alanine, which represents the minimal substitution. Biochemical analysis of these mutant receptors revealed that the substitution of glycine at position 743 gradually impairs AR function when the amino acid ranges from alanine to valine and subsequently glutamic acid. Structural analysis contributed to a better understanding of the molecular basis of these functional defects.
Materials and Methods
Amino acid substitutions and patients
The G743V substitution has been reported previously (12). Briefly, the patient was referred in the neonatal period for evaluation of inguinal bilateral masses. External genitalia were female, but slight clitoral hypertrophy led us to classify the patient as grade V of Quigley’s classification (2). Plasma testosterone level rose to 30 nM after the human chorionic gonadotropin stimulation test (1500 IU, seven doses). The blood karyotype was actually 46,XY. Sequencing of genomic DNA identified a G to T transversion leading to the substitution of the glycine at position 743 by a valine. This mutation was also reported by Nakao et al. (11) in a patient with less severe PAIS.
The G743E AR mutant has not been described to date. This patient was referred for primary amenorrhea, normal breast development, and complete absence of pubic and axillary hair by the age of 15 yr. Plasma testosterone levels were within the normal range of male levels (24 nM), and the karyotype was 46,XY. Sequencing confirmed the diagnosis of CAIS (grade VI) by identifying a G to A mutation in exon 5 of the AR gene corresponding to the substitution of glycine by a glutamic acid. Informed consent was obtained from patients in accordance with institutional guidelines.
The G743A substitution has never been identified in AIS; however, it has been manufactured because it represents the most conservative substitution, which may have the weakest effect on AR function.
Recombinant plasmids
The three point mutations were introduced in the wild-type (wt) AR expression vector pCMV5hAR with the Altered Site Directed Mutagenesis System (Promega Corp., Madison, WI), as previously described (13). The sequences of the oligonucleotides used for generating the various mutations are summarized as follows; the first underlined nucleotide was introduced in the vector as a positive selection restriction site and the second as the amino acid substitution: G743A, 5'-CAGTATTCCTGGATGGCGCTCATGGTG-3'; G743V, 5'-CAGTATTCCTGGATGGTGCTCATGGTG-3'; and G743E, 5'-CAGTATTCCTGGATGGAGCTCATGGTG-3'.
Three reporter genes were used in this study. The p(TAT)-tk-Luc (rat tyrosine aminotransferase enhancer and thymidine kinase promoter-luciferase) was provided by Dr. M. Pons and constructed as previously described (14). The mouse vas deferens protein (MVDP)-Luc reporter gene was constructed as previously described from the promoterless basic plasmid pGL3 (Promega Corp., Lyon, France) expressing luciferase under control of the promoter of the 0.8-kb fragment of MVDP (13, 15). The reporter pARE-tk (androgen response element and thymidine kinase promoter)-Luc was constructed by replacing the MMTV promoter (HindIII-HindIII fragment) of pFC31-Luc (gift from Dr. H. Richard-Foy, Toulouse, France) (16) with the HindIII-XhoI fragment of pARE-tk-CAT (gift from Dr. G. Veyssière, Aubières, France). Both vector and insert were end-filled with Klenow fragment.
Androgen binding, thermostability, and dissociation rate kinetics
COS-7 cells were cultured in DMEM (Life Technologies, Inc., Cergy, France) with 10% fetal calf serum, penicillin (100 U/ml), and streptomycin (100 mg/ml) in a humidified 5% CO2 incubator.
Unlabeled 17
-methyltrienolone (R1881) and [3H] R1881 (87 Ci/mmol) were purchased from NEN Life Science Products (Paris, France). To determine the apparent equilibrium binding affinity of the wt and mutant ARs, 1.5 x 105 COS-7 cells were transiently transfected in 12-well dishes with 25 ng wt pCMV5hAR or mutant AR cDNA/well, using the calcium phosphate coprecipitation method. Cells were incubated 48 h after transfection for 2 h at 30 or 37 C with increasing concentrations of [3H]R1881 from 0.05–4 nM, with or without a 1000-fold molar excess of unlabeled R1881. Cells were placed on ice for 10 min, washed twice with cold PBS, and harvested in 300 µl lysis buffer [25 mM Tris phosphate (pH 7.8), 2 mM dithiothreitol, 2 mM EDTA, 1% Triton X-100, and 10% glycerol]. Radioactivity was determined by scintillation counting, and specific binding was determined as the difference between total and nonspecific counts. The data were analyzed by the Scatchard method.
The thermostability of the wt and mutant receptors was examined by incubating COS-7 cells, transfected as described above, at 37 and 41 C with 4 nM [3H]R1881 for 2 h, alone (total binding) or with a 1000-fold excess of unlabeled R1881 (nonspecific binding). Maximal R1881-binding activity was calculated for the two temperatures. The thermolability of AR-R1881 complexes refers to the percentage decrease in binding at 41 C compared with the binding at 37 C.
For dissociation rate kinetics, transiently transfected COS-7 cells were incubated at 30 C with 4 nM [3H]R1881 for 2 h, followed by the addition of a 1000-fold molar excess of unlabeled R1881. Nonspecific binding was determined from cells that were treated with 4 nM [3H]R1881 in the presence of a 1000-fold excess of unlabeled R1881. Cells were washed and lysed after increasing times (30, 60, 90, and 120 min), and the percentage of remaining specific androgen binding was plotted semilogarithmically against time.
All experiments were performed at least twice in triplicate.
Immunoblotting
Immunoblotting was performed in the COS-7 cell line as previously described (17). Briefly, 106 cells were plated on 10-cm dishes and transfected for each AR variant with 8 µg pCMV5hAR plasmid. After 48 h, the cells were washed twice with PBS and lysed in the presence of a protease inhibitor cocktail (Sigma-Aldrich, Saint-Quentin Fallavier, France). Cell lysates were submitted to electrophoresis, and gels were blotted onto Hybond membrane (Amersham Pharmacia Biotech, Paris, France). For AR detection, Western blotting was performed using the rabbit polyclonal antibody N-20 directed against a peptide corresponding to amino acids 2–21 mapping at the N terminus of the human AR (Santa Cruz Biotechnology, Inc., Tebu, France). For detection of ß-actin, membranes were incubated with the polyclonal antibody developed in rabbits using the C-terminal actin fragment (Sigma-Aldrich). Blots were developed using a chemiluminescent detection system (Pierce Chemical Co., Interchim, Montluçon, France).
Trans-activation capacity assays
CV-1 cells were grown in DMEM containing 10% fetal calf serum, penicillin (100 U/ml), and streptomycin (100 mg/ml). Transient transfections were performed in 12-well tissue culture plates using the calcium phosphate coprecipitation method with either 100 ng wt or mutant AR expression vectors; 1 µg MVDP-Luc, TAT-tk-Luc, or ARE-tk-Luc as reporter genes; and 250 ng pCMV5-ß-galactosidase. After 12 h, the precipitate was removed and substituted by DMEM with various concentrations of R1881 (10-12–10-6 M). R1881 was used instead of physiological androgens to minimize metabolization during incubation. In some experiments androgen antagonists were used: cyproterone acetate (CPA) and RU23908 (nilutamide) at 10-8, 10-7, and 10-6 M, with or without R1881 [CPA was obtained from Sigma and nilutamide was a gift from Roussel-UCLAF (Romainville, France)].
Thirty hours later, CV-1 cells were lysed with 300 µl lysis buffer. The induction of luciferase activity is indicated in arbitrary units, corrected by ß-galactosidase activity. Three independent assays were performed in duplicate.
Experiments involving cell manipulations were performed in accordance with institutional guidelines and with the agreement of INSERM.
Results and Discussion
Androgen binding, thermostability, and dissociation rate of wt and mutant ARs
The apparent equilibrium binding constants (Kd) for [3H]R1881, determined at 30 C, were in a similar range for wt AR, G743A, and G743V and were slightly increased for G743E (Table 1
). At 37 C, however, we were unable to determine an apparent equilibrium binding constant for G743E, presumably because of the instability of the ligand-AR complex at this temperature. This effect of temperature on the specific androgen-binding activity was even more drastic at 41 C for both G743V and G743E mutants. When complexed to R1881 in transfected COS-7 cells, G743E and G743V exhibited increased thermolability with, respectively, 75% and 53% decreases in binding at 41 C compared with binding at 37 C (Table 1).
|
).
|
), indicating that alteration of their ligand-binding activities was not attributable to abnormal expression or protein instability.
|
In the presence of androgen.
To quantify the functional consequences of these mutations, we tested the abilities of the various ARs to promote a transcriptional activation of several reporter genes. With TAT-tk-Luc the trans-activation efficiencies of wt AR and G743A were similar (Fig. 3A). It was difficult to discriminate the G743V and G743E substitutions based on their transcriptional properties, although the clinical presentations were clearly different. This could be due to a lack of specificity of the reporter gene, as TAT-tk-Luc has been shown to also be activated by glucocorticoids, progestagens, and mineralocorticoids (18). However, similar patterns of transcriptional activity were observed when the reporter gene was restricted to a minimal ARE (ARE-tk-Luc) devoid of any potentially specific adjacent sequences (Fig. 3B). We also used another luciferase reporter gene driven by the proximal fragment of the MVDP gene previously described as an androgen target gene (15). The R1881 doses that were sufficient to obtain 50% maximal fold induction shifted from 3 x 10-11 for wt AR to 10-10, 5 x 10-10, and 4 x 10-9 M for G743A, G743V, and G743E, respectively (Fig. 3C). Although the relative activities of the different receptors were overall similar to those observed with the two other promoters, the MVDP-Luc reporter gene allowed more accurate discrimination between the mutants at a low concentration of ligand.
|
CPA displayed, as expected, a partial agonist property with wt AR but produced a lower induction with G743A and no induction even at 10-6 M with G743V and G743E (Fig. 4A). In contrast, no transcriptional activity was seen for the nonsteroidal antiandrogen nilutamide whatever the concentration or expression vector (wt AR and mutants; Fig. 4B).
|
). At 10-7 M CPA, the antagonist activity was complete for G743A and G743V, whereas it was partial for wt AR. For G743E mutant, a strong antagonistic activity was observed at a high CPA concentration (10-6 M). Competitive binding experiments correlated with these results by showing that at this concentration (10-6 M) CPA is able to displace R1881 when bound to wt AR as well to the various mutants (data not shown). These data suggest that although CPA can still bind to the various AR mutants, it is not able to stabilize their agonistic conformation. Therefore, the loss of partial CPA agonist activity corresponds to an apparent enhancement of its antagonistic properties.
With the nonsteroidal nilutamide, a significant loss in antagonist properties was observed for the three AR mutants (Fig. 4D). This was particularly evident for the G743V and G743E variants at a high concentration (10-6 M). In correlation with these data, nilutamide, even at 10-6 M, was not able to correctly displace R1881 bound to either G743V or G743E (data not shown). The different behavior of the nonsteroidal nilutamide compared with that of the steroidal CPA probably reflects their different binding modes to AR-LBD, as would be expected from their chemical structures.
Structural analysis of AR mutants
Analysis of AR mutants with different substitutions at the same position is an important step in understanding the structure-function relationship of AR. This type of molecular investigation of the AR-LBD mutations associated with AIS (24, 25, 26, 27, 28) has given valuable insights into the subtleness of ligand-binding activity. The recent publications of the AR-LBD crystal structure (9, 10) led us to study natural and engineered AR mutants described in AIS by integrating structural data with clinical and functional information. Structural analysis may provide a detailed molecular basis of AR functional defects (29).
Structural data have shown that the ligand-binding domains of nuclear receptors share a very similar overall fold (5, 30). Whether currently identified or not, the specificity of each is restricted to limited regions, which are constituted by a rather low number of amino acids. Crystallographic data, for example, have confirmed the importance of a single residue in the discrimination between 3-keto and 3-hydroxyl substituents in steroids (31), which was previously suggested by an in vitro study (32). To identify the amino acids that might be implicated in the specificity of nuclear receptor activation, we looked for divergent residues in the superfamily. As shown in the partial amino acid alignment (Fig. 5), residues at positions homologous to 743 in AR are not conserved among the nuclear receptors. Such diversity could reflect the weak conservation of the function of this residue and, thus, its low importance. Amino acid substitution would thus result in minor consequences. On the other hand, this diversity could also be related to the specificity of each nuclear receptor. Indeed, a glycine was always observed in AR whatever the species, suggesting a specific role for this residue in AR function.
|
The structural consequences of G743 mutations were evaluated on the basis of the AR-LBD structures (9, 10). G743 belongs to helix H5 in AR-LBD (Fig. 5). Helix H5 is an important structural element, as several residues contacting the ligand are located in it (e.g. M742, M745, V746, and R752; Fig. 6). Therefore, any modification destabilizing helix H5 should perturb the ligand-binding characteristics of the receptor.
|
). The weak effect of the G743A mutation on AR function supports this view.
The second way in which mutation of G743 can destabilize helix H5 is by perturbing the local environment due to a lack of complementarity between the physico-chemical properties of the new side-chain and the surrounding residues. This is buttressed by the dramatic effect of the introduction of a glutamic acid, which, in addition to a high average volume (155 Å3), brings a negative charge in a mostly hydrophobic environment. It is noteworthy that many NRs [e.g. estrogen receptor (ER), retinoic X receptor, thyroid hormone receptor, vitamin D receptor, peroxisomal proliferator-activated receptor, and retinoic acid receptor] contain an acidic residue (E or D) at the position occupied by G743 in AR (Fig. 5). As an example, in ER, for which several crystallographic structures are available (33, 34), E385 (homologous to AR-G743) is hydrogen-bonded to S456 (H8). Other residues, such as N455 H8 and S518 H10, do not interact directly with E385, but generate a hydrogen bond network involving several water molecules and the carboxylate moiety of E385. Most importantly, although R515 H10 is not engaged in a classical salt bridge with E385, it places its guanidium group at a distance (4.5 Å) that permits the stabilization of the negatively charged E385. Interestingly, multiple sequence alignment reveals that all nuclear receptors having an acidic residue at the position equivalent to G743 also harbor an arginine in helix 10 (Fig. 5).
In AR, the environment of G743 is more hydrophobic. S456 and S518 in ER are replaced by I815 and A870, respectively. In addition, the equivalent residue of ER-N455 is S814 in AR, which does not contribute to the definition of a hydrophilic environment for G743, because its side-chain orientation confers an interaction with the carbonyl group of L810 in H8 (Fig. 6). However, the main difference is the replacement of R515 H10 in ER by Q867 in AR (Fig. 5). In the G743E mutant, the lack of charge compensation is likely to alter the structural integrity of the region surrounding E743 (including H5) and, consequently, to account for the loss of function of the receptor.
With respect to the structural basis of the functional defect associated with the G743V mutant, as previously mentioned, AR-G743 is surrounded by mostly hydrophobic amino acids that allow its replacement by aliphatic residues within a limited range of size. These steric restrictions explain why an alanine (average volume, 91.5Å3) is less deleterious than a valine (141.7 Å3) to androgen binding and trans-activation efficiency. Indeed, the distance of the C
atoms of the valine residue to the closest amino acids (V866 and A870 in H10) would be 2.8 Å. Therefore, some structural adaptation, possibly affecting ligand binding, may be necessary to accommodate this change in side-chain size.
In conclusion, our integrative approach has provided a useful means to understand the phenotypes of some patients with androgen insensitivity. In the case of G743 substitutions, we observed good agreement among these three complementary investigations. This approach could also be applied to the study of mutations associated with prostate cancer, male infertility, and other diseases associated with AR defects.
Acknowledgments
We are grateful to Dr. T. R. Brown for providing the pCMV-AR plasmid, to Dr. H. Richard-Foy for pFC31-luciferase (MMTV-luciferase), to Dr. G. Veyssière for ARE-tk-CAT, and to Dr. M. Pons for TAT-tk-Luc. We thank Dr. G. Auzou for the luciferin preparation. We thank Fanja Rabanoelina for helpful technical assistance. We express our gratitude to Dr. J. C. Nicolas for helpful discussion.
Footnotes
This work was supported in part by INSERM and the Association pour la Recherche sur le Cancer (Grant 5205, 1999).
N.P. and S.L. contributed equally to this work.
Present address for J.-M.A.L.: Physiologie Comparée et Endocrinologie Moléculaire, Université Blaise Pascal, 63177 Aubières, France.
Abbreviations: AIS, Androgen insensitivity syndrome; AR, androgen receptor; CAIS, complete androgen insensitivity syndrome; CPA, cyproterone acetate; LBD, ligand-binding domain; MVDP, mouse vas deferens protein; PAIS, partial androgen insensitivity syndrome; pARE-tk, androgen response element and thymidine kinase promoter; p(TAT)-tk-Luc, rat tyrosine aminotransferase enhancer and thymidine kinase promoter-luciferase; wt, wild type.
Received March 28, 2002.
Accepted September 17, 2002.
References
Val) in the steroid-binding domain of the human androgen receptor leads to Reifenstein syndrome. J Clin Endocrinol Metab 77:103–107[Abstract]
: estrogen versus androgen discrimination. J Biol Chem 273:693–699This article has been cited by other articles:
 |
|
 |
|
  B. He, R. T. Gampe Jr., A. T. Hnat, J. L. Faggart, J. T. Minges, F. S. French, and E. M. Wilson Probing the Functional Link between Androgen Receptor Coactivator and Ligand-binding Sites in Prostate Cancer and Androgen Insensitivity J. Biol. Chem., March 10, 2006; 281(10): 6648 - 6663. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Kohler, S. Lumbroso, J. Leger, F. Audran, E. S. Grau, F. Kurtz, G. Pinto, M. Salerno, T. Semitcheva, P. Czernichow, et al. Androgen Insensitivity Syndrome: Somatic Mosaicism of the Androgen Receptor in Seven Families and Consequences for Sex Assignment and Genetic Counseling J. Clin. Endocrinol. Metab., January 1, 2005; 90(1): 106 - 111. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
