1042
Biochemical Society Transactions (2003) Volume 31, part 5
The androgen receptor transactivation domain: the interplay between protein conformation and protein–protein interactions J. Reid1 , R. Betney, K. Watt and I.J. McEwan2 Department of Molecular and Cell Biology, Institute of Medical Sciences, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, U.K.
Abstract The AR (androgen receptor) belongs to the nuclear receptor superfamily and directly regulates patterns of gene expression in response to the steroids testosterone and dihydrotestosterone. Sequences within the large N-terminal domain of the receptor have been shown to be important for transactivation and protein– protein interactions; however, little is known about the structure and folding of this region. Folding of the AR transactivation domain was observed in the presence of the helix-stabilizing solvent trifluorethanol and the natural osmolyte TMAO (trimethylamine N-oxide). TMAO resulted in the movement of two tryptophan residues to a less solvent-exposed environment and the formation of a protease-resistant conformation. Critically, binding to a target protein, the RAP74 subunit of the general transcription factor TFIIF, resulted in a similar resistance to protease digestion, consistent with induced folding of the receptor transactivation domain. Our current hypothesis is that the folding of the transactivation domain in response to specific protein–protein interactions creates a platform for subsequent interactions, resulting in the formation of a competent transcriptional activation complex.
Introduction “. . . the very point which appears to complicate a case is, when duly considered and scientifically handled, the one which is most likely to elucidate it.” Sherlock Holmes to Dr Watson in The Hound of the Baskervilles by Sir Arthur Conan Doyle The androgenic steroid hormones testosterone and the 5αreduced metabolite dihydrotestosterone alter patterns of gene expression in target tissues by activating an intracellular receptor protein termed the AR (androgen receptor) [1]. The AR belongs to the class I subfamily of steroid receptors that also includes GR (glucocorticoid receptor), PR (progesterone receptor) and MR (mineralocorticoid receptor): these receptors all bind related DNA-response elements. The consensus DNA-binding sequence for the AR consists of two palindromic half sites separated by a three-nucleotide spacer, 5 -AGWACWNNNWGTTCY-3 (where W = A or T and Y = T or C). The AR has a well-defined domain organization (Figure 1A), consisting of a C-terminal LBD
Key words: fluorescence spectroscopy, mutagenesis, secondary-structure prediction, steroid receptor, transactivation. Abbreviations used: AR, androgen receptor; GR, glucocorticoid receptor; PR, progesterone receptor; MR, mineralocorticoid receptor; LBD, ligand-binding domain; DBD, DNA-binding domain; NTD, N-terminal domain; TAD, transactivation domain; TAU, transcriptional activation unit; TFE, trifluorethanol; TMAO, trimethylamine N-oxide; TBP, TATA-binding protein. 1 Present address: Cancer Research UK, London Research Institute, Claire Hall Laboratories, Blanche Lane, South Mimms, Hertfordshire, EN6 3LD, U.K. To whom correspondence should be addressed (e-mail
[email protected]).
2
C 2003
Biochemical Society
(ligand-binding domain), a central ‘Zn-finger’ DBD (DNAbinding domain) and a structurally distinct NTD (N-terminal domain) important for transcription activation ([1] and references therein).
Hormone binding The general mechanism of AR action and the role of different macromolecular interactions is summarized in Figure 1(B). Upon binding to the hormone testosterone the LBD undergoes a rearrangement of a three layer αhelical sandwich conformation, resulting in a realignment of the most C-terminal helix (helix 12) with the core of the domain and a more compact structure [2]. In the inactive state the receptor is complexed with a number of molecular chaperones such as Hsps 90 and 70, p60 and p23 polypeptides and the immunophilin FKBP (FKB506-binding protein) (see [3,4]). The dissociation of these chaperone molecules is thought to occur concomitant with hormone binding and the LBD structural rearrangement and result in the translocation of the activated receptor into the nucleus (Figure 1B). However, recent evidence suggests that the molecular chaperones may accompany the receptor into the nucleus [4]. The binding site of Hsp90 within the GR has been mapped to seven amino acids within helix 1 of the LBD, corresponding to amino acids 670–676 of the human AR [3]. Genetic analysis in the budding yeast Saccharomyces cerevisiae has emphasized the role of molecular chaperones in maintaining the class I receptors, GR and AR, in a
Intermolecular Associations in 2D and 3D
Figure 1 The molecular mechanism of AR action (A) Schematic representation of the human AR, showing the domain organization and the TAD (transactivation domain), amino acids 142–485. LBD, ligand-binding domain; DBD, DNA-binding domain; TAU, transcriptional activation unit. (B) General mechanism of androgen action. The lipophilic steroid testosterone enters the target cells and binds to the receptor complexed with molecular chaperones (i.e. Hsp90) and associated proteins. The activated receptor then enters the nucleus and binds as a homodimer to DNA-response elements and through protein–protein interactions regulates gene expression. NTD, N-terminal domain; TBP, TATA-binding protein.
conformation capable of binding ligand and not simply acting to block entry into the nucleus ([5] and references therein).
DNA binding The regulation of gene expression requires the specific tethering of the AR homodimer to the promoter and/or enhancer sequences of target genes (Figure 1B). Recognition and binding of DNA is achieved by the DBD (amino acids 550–624) [6]. NMR spectroscopy and X-ray crystallography studies have revealed a common structural fold for the DBD of both steroid and non-steroid receptors (reviewed in [7]). The main feature of this fold consists of two α-helices positioned perpendicular to each other. The P-box residues, which are involved in part in DNA sequence recognition,
are located at the N-terminus of the first helix. This helix is positioned within the major groove of the DNA double helix, forming a recognition helix. The P-box residues found within the AR are Gly-575, Ser-576 and Val-579, which are identical to those found within the GR, MR and PR. Clearly, for the class I receptors such as AR and GR determinants other than the P-box residues must also play a role in determining DNA recognition and binding specificity in cells where both receptors are present (for recent review see [6]). Claessens and co-workers [6] have provided evidence to suggest that on AR-specific response elements, found in certain androgenregulated genes, the AR recognizes the sequence of DNA as a direct repeat sequence and adopts a different conformation binding in a ‘head-to-tail’ configuration instead of ‘head-tohead’ ([6] and references therein). DNA-binding site selection C 2003
Biochemical Society
1043
1044
Biochemical Society Transactions (2003) Volume 31, part 5
studies by Rennie and co-workers [8] suggest that the nucleotides within the spacer and/or the flanking sequences may also play a role in AR-binding specificity.
Transcription activation Once targeted to specific genes the AR makes direct and/or indirect interactions with the transcription machinery in order to regulate gene expression (Figure 1B). Regions of the receptor protein important for transactivation activity have been mapped to the N-terminal domain of the receptor [9– 11]. Deletion analysis and point mutations have highlighted the modular nature of the receptor TAD (transactivation domain) and have delineated the sequences between amino acids 142 and 485 as being important for transcriptional activation: this includes TAU-1 (transcriptional activation unit 1; amino acids 101–370) and TAU-5 (amino acids 360–485) [10,11].
Protein–protein interactions In the context of the full-length AR, interactions between the N- and C-terminal domains are important for receptordependent gene transcription. The residues important for this interaction map out with the TAD: and include amino acids N-terminal of helix 12 in the AR-LBD and the first 30 residues of the AR-NTD [12,13]. A FXXLF sequence between amino acids 23 and 27 appears to play a key role in the N/C-terminal interactions. We have investigated the structure–function relationship of a region of the AR N-terminal domain that retains 70% of the activity of the full-length N-terminus [14]. This region, amino acids 142–485 (numbering for human AR with polyglutamine and poly-glycine repeats of 20 and 16 respectively) is defined as the AR-TAD (or AF-1) and comprises the previously identified TAU-1 and TAU-5 activities. To understand the mechanism of transactivation we initially identified targets for the AR-TAD within the transcriptional machinery. These studies revealed a selective interaction with the large subunit of the general transcription factor TFIIF. This interaction is capable of reversing AR-dependent squelching in both yeast [15] and mammalian cell nuclear extracts (A. Ball and I.J. McEwan, unpublished work). Subsequently, work in other laboratories have described other targets for the ARNTD, including the p160 family of co-activators originally identified as binding to the LBD of nuclear receptors in a ligand-dependent manner [16–18]. Using deletion constructs of RAP74 (large subunit of transcription factor TFIIF) and SRC (steroid receptor coactivator)-1a we have mapped the AR-binding sites to the N- and C-termini of RAP74 and the C-terminus of SRC-1. A competition binding assay with holo-TFIIF, reconstituted from RAP30 and RAP74 subunits, suggested that the Cterminal site contained the main interacting surface for the AR-TAD [14]. Introducing point mutations into the ARTAD has helped define distinct binding sites for TFIIF and SRC-1a (Table 1). TFIIF appears to bind to the N-terminal sequences of the AR-TAD, which include the TAU-1 activity, C 2003
Biochemical Society
Table 1 Summary of the functional activity of wild-type and mutant AR-TADs The transactivation activity of the AR-TAD-Lex fusion proteins was measured using a β-galactosidase reporter gene assay in the budding yeast S. cereivisae [14,18a]. The AR-TAD (amino acids 142–485) retains about 70% of the activity of the full-length AR-NTD [14] and has been set at 100% to permit comparison with mutant AR-TAD polypeptides: ND, not determined. Under ‘Binding’ are shown the results from protein–protein interaction studies with immobilized AR polypeptides and radiolabelled RAP74 (TFIIF) and SRC-1a-CTD (amino acids 977– 1441). −, +/− and + represent reduced binding, of 63, 40 and 27% respectively [14,18a] ‘++’ represents wild-type binding (100%). Binding Protein
Mutations
Activity (%)
TFIF
SRC-1a
AR-TAD
–
100
++
++
M1 M2 M3
I229A/L234A V240A/V242A M244A/L246A/V248A
70 58 36
++ ++ +/−
++ ++ ++
M4 M5 M6
I251A/L254A I181A/L182A S159A/S162A
70 30 ND
++ ++ −
++ ++ ++
M7
S340A/S343A
ND
+
++
whereas SRC-1a is likely to bind to the C-terminus, which includes TAU-5 [14,18a].
Structural analysis of the AR-TAD Although clearly important for AR-dependent gene regulation little is known about the structural basis of receptordependent transactivation. Secondary structure prediction analysis (Figure 2) suggested four regions of α-helix and a total helical content of 20%, which is in close agreement with the value determined from CD spectroscopy analysis (13– 15%). The α-helix content increased in the presence of increasing amounts of TFE (trifluorethanol) a solute thought to stabilize secondary structure [19]. Limited proteolysis with trypsin further supported the view that the AR-TAD contains limited but defined elements of secondary structure, which could be disrupted by introducing helix-breaking mutations [19]. Similarly, analysis of the intrinsic fluorescence from tyrosine and tryptophan residues emphasized the lack of stable structure and the ability of the AR-TAD to adopt a more folded conformation in the presence of the natural osmolyte TMAO (trimethylamine N-oxide). Taken together, the biophysical and biochemical analysis of the AR-TAD conformation revealed a structurally flexible polypeptide that could fold into a more compact conformation in the presence of structure stabilizing solutes (TFE, TMAO); this folded structure may involve a coil-to-helix transition. Similar structural properties have been described for the AF-1 domains of the GR [20,21], oestrogen receptor [22] and the non-steroid nuclear receptor peroxisome proliferator-activated receptor γ [23]. Furthermore, it appears that induced folding is a
Intermolecular Associations in 2D and 3D
Figure 2 Secondary-structure predictions for the AR-TAD The secondary-structure content for the AR-TAD and the mutant polypeptides M3 and M6 was estimated using the consensus secondary-structure prediction option, Network Protein Sequence Analysis (or NPS@), at http://pbil.ibcp.fr [26].
common property of transactivation domains and may have an advantage in terms of determining binding specificity, while maintaining the ability to participate in multiple protein–protein interactions (see [24] for review).
The interplay between protein–protein interactions and protein conformation The fact that the AR-TAD adopted a more folded conformation in the presence of TFE or TMAO led us to investigate the role protein–protein interactions would have on conformation. Interactions with the C-terminal receptor binding domain of RAP74 lead to the protection of the ARTAD from limited proteolysis by trypsin, chymotrypsin and endoGlu-C. Strikingly, the pattern of protection was very similar to that seen with TFE and TMAO [19]. These results suggest that the binding of TFIIF leads to a more folded conformation of the AR-TAD that is resistant to protease attack. Induced folding of the oestrogen receptor AF-1 domain upon binding of another of the general transcription factors, the TBP (TATA-binding protein), has also been reported [22]. The folding of the AR-TAD in response to RAP74-binding suggests that this is likely to be a high-affinity interaction that is enthalpy-driven (see [24]). These studies have led us to propose the following model: protein–protein interactions result in induced folding of the AR-TAD, which in turn leads to the generation of surfaces for further interactions and the assembly of a transcriptional competent receptor complex. Preliminary evidence supporting such a model comes from glutathione S-transferase pull-down experiments. Pre-incubation of immobilized glutathione S-transferase-AR-TAD with TMAO or crucially TFIIF resulted in a significant enhancement of SRC-1a binding (R. Betney and I.J. McEwan,
unpublished work). The result was specific for TFIIF as incubation with a non-interacting protein had little affect on SRC-1 binding. Kumar et al. [25] observed similar results with the GR-AF-1 domain: with TMAO-induced protein folding resulting in an increase in the binding of the target proteins TBP, GRIP1 (glucocorticoid receptor interacting protein-1) and CBP (cAMP-response-element-binding protein-binding protein). On-going experiments aim to further study the binding affinities and thermodynamic profiles for different AR-TAD–target protein interactions. Such experiments are important in order to elucidate the significance of different protein–protein interactions for AR-TAD structure and function. The introduction of point mutations into the AR-TAD resulted in the identification of two mutant polypeptides selectively impaired for TFIIF binding (Table 1) [14,18a]. Interestingly, these mutant polypeptides have different structural properties, suggesting they may interfere with TFIIF binding by different mechanisms. Mutant ARTAD:M3 (M244A/L246A/V248A) is not predicted to have an altered secondary structure (Figure 2). Also the tyrosine and tryptophan fluorescence emission spectrum for the M3 polypeptide is essentially identical to that of the wild-type AR-TAD [18a]. In contrast, M6 (S159A/S161A) is predicted to alter an N-terminal β-strand to α-helix (Figure 2). The fluorescence emission spectrum for this polypeptide is consistent with alterations in conformation; with changes in the tyrosine and tryptophan emissions suggesting a more structured conformation [18a]. We conclude from the above studies that the binding of TFIIF is dependent upon the structural flexibility of the AR-TAD and the mutations in M6 disrupt binding by being structurally less flexible. While the mutations in M3 may directly constitute part of C 2003
Biochemical Society
1045
1046
Biochemical Society Transactions (2003) Volume 31, part 5
the binding surface for TFIIF and mutating the hydrophobic residues disrupts key intermolecular interactions. On-going studies are aimed at testing this hypothesis further.
Conclusions In conclusion, the current model for AR-dependent gene regulation involves multiple protein–protein interactions leading to the enhancement of mRNA synthesis of target genes. Our findings suggest that the N-terminal TAD of the AR is structurally flexible but adopts a more stable conformation upon specific protein–protein interactions. This in turn generates a surface that can serve as a binding site for other target proteins. It is also worth noting that specific DNA binding may also play a role in the assembly of multi-protein complexes involving the AR. Thus, the interplay between protein interactions and protein conformation provides a mechanism for establishing AR-dependent transcriptionally competent complexes at the promoter/enhancer regions of target genes. This work was supported in part by grants 1/C10407 from the Biotechnology and Biological Sciences Research Council (BBSRC) and 99-094 from the Association for International Cancer Research. We are grateful to Dr A.O. Brinkmann (Erasmus University, Rotterdam, The Netherlands) and Dr Z. Burton (Michigan State University, East Lansing, MI, U.S.A.), Professor B.W. O’Malley (Baylor College of Medicine, Houston, TX, U.S.A.) and Professor A.P.H. Wright (Sodertorns ¨ Hogskolan, ¨ Stockholm, Sweden) for the gift of plasmid DNA.
References 1 Gelman, E.P. (2002) J. Clin. Oncol. 20, 3001–3015 2 Matia, P.D., Donner, P., Coelho, R., Thomaz, M., Peixoto, C., Macedo, S., Otto, N., Joschko, S., Scholz, P., Wegg, A. et al. (2000) J. Biol. Chem. 275, 26164–26171 3 Davies, T.H., Ning, Y.-M. and Sanchez, E.R. (2002) J. Biol. Chem. 277, 4597–4600
C 2003
Biochemical Society
4 Kaul, S., Murphy, P.J.M., Chen, J., Brown, L., Pratt, W.B. and Stoney Simons, S. (2002) J. Biol. Chem. 277, 36223–36232 5 McEwan, I.J. (2001) Trends Genet. 17, 239–243 6 Verrijdt, G., Haelens, A. and Claessens, F. (2003) Mol. Gen. Metab. 78, 175–185 ˚ 7 Zilliacus, J., Wright, A.P.H., Carlstedt-Duke, J. and Gustafsson, J.-A. (1995) Mol. Endocrinol. 19, 389–400 8 Nelson, C.C., Hendy, S.C., Shukin, R.J., Cheng, H., Bruchovsky, N., Koop, B.F. and Rennie, P.S. (1999) Mol. Endocrinol. 13, 2090–2107 9 Simental, J.A., Sar, M., Lane, M.V., French, F.S. and Wilson, E.M. (1991) J. Biol. Chem. 266, 510–518 10 Jenster, G., van der Korput, H. A., Trapman, J. and Brinkmann, A.O. (1995) J. Biol. Chem. 270, 7341–7346 11 Chamberlain, N.L., Whitacre, D.C. and Miesfeld, R.L. (1996) J. Biol. Chem. 271, 26772–26778 12 He, B. and Wislon, E.M. (2002) Mol. Gen. Metab. 75, 293–298 13 Steketee, K., Berrevoets, C.A., Dubbink, H.J., Doesburg, P., Hersmus, R., Brinkmann, A.O. and Trapman, J. (2002) Eur. J. Biochem. 269, 5780–5791 14 Reid, J., Murray, I., Watt, K., Betney, R. and McEwan, I.J. (2002) J. Biol. Chem. 277, 41247–41253 ˚ (1997) Proc. Natl. Acad. Sci. U.S.A. 15 McEwan, I.J. and Gustafsson, J.-A. 94, 8485–8490 16 Alen, P., Claessens, F., Verhoeven, G., Rombauts, W. and Peeters, B. (1999) Mol. Cell. Biol. 19, 6085–6097 17 Ma, H., Hong, H., Huang, S.M., Irvine, R.A., Webb, P., Kushner, P.J., Coetzee, G.A. and Stallcup, M.R. (1999) Mol. Cell. Biol. 19, 6164–6173 18 Bevan, C.L., Hoare, S., Claessens, F., Heery, D.M. and Parker, M.G. (1999) Mol. Cell. Biol. 19, 8383–8392 18a Betney, R. and McEwan, I.J. (2003) J. Mol. Endocrinol., in the press 19 Reid, J., Kelly, S.M., Watt, K., Price, N.C. and McEwan, I.J. (2002) J. Biol. Chem. 277, 20079–20086 20 Dahlman-Wright, K., Baumann, H., McEwan, I.J., Almlof, ¨ T., Wright, A.P., ˚ and Hard, Gustafsson, J.-A. ¨ T. (1995) Proc. Natl. Acad. Sci. U.S.A. 92, 1699–1703 21 Baskakov, I.V., Kumar, R., Srinivasan, G., Ji, Y.S., Bolen, D.W. and Thompson, E.B. (1999) J. Biol. Chem. 274, 10693–10696 ˚ and Hard, 22 Warnmark, ¨ A., Wikstrom, ¨ A., Wright, A.P.H., Gustafsson, J.-A. ¨ T. (2001) J. Biol. Chem. 276, 45939–45944 23 Hi, R., Osada, S., Yumoto, N. and Osumi, T. (1999) J. Biol. Chem. 274, 35152–35158 24 Wright, P.E. and Dyson, H.J. (1999) J. Mol. Biol. 293, 321–331 25 Kumar, R., Lee, J.C. and Thompson, E.B. (2001) J. Biol. Chem. 276, 18146–18152 26 Combet, C., Blanchet, C., Geourjon, C. and Deleage, G. (2000) Trends Biochem. 25, 147–150
Received 20 June 2003
