1054
Biochemical Society Transactions (2006) Volume 34, part 6
The human androgen receptor AF1 transactivation domain: interactions with transcription factor IIF and molten-globule-like structural characteristics D.N. Lavery and I.J. McEwan1 *School of Medical Sciences, Institute of Medical Sciences, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, Scotland, U.K.
Abstract The AR (androgen receptor) is a ligand-activated transcription factor and member of the steroid receptor superfamily. The AR responds to the ligands testosterone and dihydrotestosterone and activates multiple downstream genes required in development and reproduction. During the events of transactivation, the AR makes specific protein–protein interactions with several basal transcription factors such as TBP (TATAbox-binding protein) and TFIIF (transcription factor IIF). These interactions occur predominantly within a defined region termed AF1 (activation function-1) located within the highly disordered N-terminal domain of the receptor. Our focus is on the structural aspects of AF1 and how this flexible and disordered domain generates functional interactions with regulators of transcription. Our working hypothesis is that AR-AF1 domain exhibits induced folding when contacted by transcription regulators (such as TFIIF) into a more compact and ‘active’ conformation, enabling further co-regulator recruitment and ultimately transcription. Structural flexibility and intrinsic disorder of AR-AF1 were studied using predictive algorithms and fluorescence spectroscopy under different experimental conditions and the results revealed this domain retains characteristics indicative of molten-globule or pre-molten-globule-like structures. We hypothesize that this partially folded intermediate state is important for, and enables the AF1 domain to make, multiple protein–protein interactions. The structural aspects of AR-AF1 and interactions with TFIIF are discussed.
Introduction The AR (androgen receptor) is a member of the SHR (steroid hormone receptor) family, which also includes the oestrogen, progesterone, mineralocorticoid and glucocorticoid receptors [1]. This family of receptors has a modular and well-defined domain organization (Figure 1A). The central domain and CTD (C-terminal domain) are involved in DNA recognition and ligand binding respectively. The structures of these domains are well characterized and have been solved by crystallography for most of the SHRs. Interestingly, members of this family have an apparently disordered and structurally flexible N-terminal domain, containing residues important for creating protein–protein interactions during the events of transcription, termed AF1 (activation function-1) (for a review, see [2]). In response to stimuli, the AR in complex with heat-shock/ chaperone proteins undergoes a conformational change and translocates into the nucleus. Subsequently, nuclear AR forms homodimers and binds to specific response elements upstream of regulated genes. At the promoters of these genes, the AR is able to recruit members of the basal transcription Key words: basal transcription factor, molten globule, protein–protein interaction, steroid hormone receptor, transactivation domain, transcription. Abbreviations used: AF, activation function; ANS, 1-anilino-8-naphthalenesulfonate; AR, androgen receptor; CTD, C-terminal domain; RAP, RNA polymerase II-associated protein; SHR, steroid hormone receptor; SPR, surface plasmon resonance; TBP, TATA-box-binding protein; TFIIF, transcription factor IIF; TMAO, trimethylamine-N-oxide. 1 To whom correspondence should be addressed (email
[email protected]).
C 2006
Biochemical Society
machinery [such as TBP (TATA-box-binding protein) and TFIIF (transcription factor IIF)] in addition to a variety of co-regulators {such as p160 family members and CBP [CREB (cAMP-response-element-binding protein)-binding protein]}. It is thought that the flexible nature of the AF1 domain allows multiple contacts to be made with numerous basal transcription factors and co-regulators, which ultimately results in the recruitment of RNA polymerase II and the initiation of transcription (for a review, see [2]).
Kinetics of AR-AF1–TFIIF interactions The general transcription factor TFIIF is a heterotetramer (α 2 β 2 ) of RAP30 (RNA polymerase II-associated protein 30) and RAP74 subunits. During transcription, TFIIF plays multiple roles in pre-initiation complex assembly, stability and elongation efficiency (for a review, see [3]). We originally identified RAP74 in a screen for interacting partners of ARAF1 [4]. In further studies, the binding site for AR-AF1 was mapped to the N- and C-terminal regions of RAP74 and using site-directed mutagenesis amino acids in AF1 that selectively disrupt binding of RAP74 were identified [5,6]. We have recently calculated the kinetic rate constants for the interaction of AR-AF1 with the full-length and subdomains of the large subunit TFIIF using SPR (surface plasmon resonance). Different kinetic profiles were observed and collectively we observe K d values in the nanomolar
Molecular Basis of Transcription
Figure 1 AR overview: domain structure and structural characteristics (A) Schematic representation of the human AR. The AR protein is organized into discrete functional domains, notably a central DNA-binding domain (DBD) and a C-terminal ligand-binding domain (LBD). Both domains are highly ordered and structured. The large N-terminal domain (NTD) contains the AF1 domain (amino acids 142–485) and is proposed to be highly flexible and disordered. (B) Secondary structure prediction (NNpredict) of the AR-AF1 domain, highlighting the limited amounts of structure; H, helix; E, strand.
range for these interactions (D.N. Lavery and I.J. McEwan, unpublished work). To further analyse the interaction of AR-AF1 with RAP74, mutations were introduced into the highly conserved CTD of RAP74. Crystal structures demonstrate that RAP74-CTD is composed of three helices followed by an antiparallel β-sheet forming a tight globular domain [7]. By substituting conserved hydrophobic residues for proline, the individual helices that support the structure of RAP74CTD were targeted for disruption. In subsequent protein– protein interaction assays, it was observed that helix 3 played an important role in AR-AF1 interactions (D.N. Lavery and I.J. McEwan, unpublished work). Additionally, in a cell-free transcription system, unlike wild-type, the mutant RAP74 when reconstituted with RAP30 could not support transcription in TFIIF-depleted cell extracts (M.A. Choudhry, D.N. Lavery and I.J. McEwan, unpublished work).
Intrinsic disorder and molten globule characteristics of AR-AF1 It has become increasingly clear that numerous proteins contain intrinsically disordered regions with limited sec-
ondary structure (for a review, see [8]). Furthermore, many proteins show induced folding when binding target proteins. Induced folding has several advantages over interactions between folded proteins, including the possibility of multiple conformers and therefore multiple binding targets, loose assembly of multiprotein complexes and specificity without the need for high-affinity binding (for a review, see [8]). Our recent studies on the structural aspects of AR-AF1 suggest that it exists as a protein domain with limited secondary structure and high proportions of intrinsic disorder (Figure 1). Using CD and Fourier-transform infrared spectroscopy, it was found that the AF1 domain exists in an unstructured state in aqueous solution [9,10]. Interestingly, TMAO (trimethylamine-N-oxide), a natural organic osmolyte, induced folding of this domain and increased the levels of α-helical structure at the expense of β-strand and turn [9,10]. Significantly, incubation with the RAP74CTD induces a similar change in AR-AF1 structure [9,10]. Using fluorescence spectroscopy, the λmax for tryptophan emission was found to be intermediate between that expected for a random coil (fully solvent-exposed) and globular structured protein (buried) and the tertiary structure of ARAF1 could be both stabilized or destabilized by addition C 2006
Biochemical Society
1055
1056
Biochemical Society Transactions (2006) Volume 34, part 6
Figure 2 Flexibility of the AR-AF1 domain (A) Bacterially expressed and Ni2+ -agarose affinity-purified AR-AF1. (B) Emission spectra of AR-AF1 (25 µg/ml) in both structure-stabilizing (TMAO) and -destabilizing (urea) environments. Note the emission peaks of tryptophan (W) at 345 nm and that of tyrosine (Y) at 309 nm. In a destabilizing environment, we observed an increase in both W and Y emission and a spectrum red shift (W at 351 nm). Incubation with TMAO stabilizes AF1 structure, as evident from the loss of Y emission and spectrum blue shift (W at 335 nm).
of TMAO and urea respectively (Figure 2). Together, these observations suggest that AR-AF1 exists in an unfolded state that is not random coil, but may resemble a molten-globulelike conformation. Using the hydrophobic probe ANS (1-anilino-8-naphthalenesulfonate), the molten-globule-like characteristics of AR-AF1 were analysed further. This probe has been used previously to dissect out protein folding states due to the fact that ANS will interact with only moltenglobule intermediates (i.e. not random coil or fully folded forms) [11,12]. The initial findings are consistent with ARAF1 having structural features characteristic of pre-molten or molten-globule states (D.N. Lavery and I.J. McEwan, unpublished work). Furthermore, when folded, AR-AF1 can interact with multiple binding targets, highlighting the structural and flexible nature of this domain (D.N. Lavery and I.J. McEwan, unpublished works).
Discussion The AF1 domain of the AR makes specific protein– protein interactions with members of the basal transcription machinery, including TBP and TFIIF. We have previously characterized the interaction between the AF1 domain and the major subunit of TFIIF, RAP74 [4,5], and have observed differential interactions with the N-/C-termini of RAP74, with the CTD representing the major binding site [5]. The interaction with RAP74 is disrupted by selective point mutations in the AF1 domain [5,6]. More recently, using a unique dual-reporter gene cell-free transcription assay, these C 2006
Biochemical Society
AF1 mutants were found to impair the ability of the AF1 domain to squelch basal transcription in HeLa nuclear extracts [13]. Interestingly, holo-TFIIF can specifically reverse AR-AF1-dependent squelching of a promoter-proximal transcript, suggesting that this interaction has important functional roles during the early steps of transcription [13]. To investigate the interaction of AR-AF1 and TFIIF further, we used SPR to characterize the binding kinetics. Similar to our initial findings, we observe differential binding of AR-AF1 to the termini of RAP74. This differential binding may play a functional role in vivo as the AR must form homodimers before it can be fully active. The AF1 domain has been shown to be highly disordered, consistent with other transactivation domains. Using spectroscopic techniques, we have shown that the structure of AF1 domain can be induced by incubation with TMAO or interestingly RAP74-CTD. Using the probe ANS, we now have evidence that the AF1 domain behaves in a molten- or pre-molten-globule-like manner and that this may aid in multiple target protein binding. Dissecting out the dynamic interplay between AF1 domain structure and the recruitment of proteins that regulate transcription remains the focus of ongoing studies.
This work was supported by a PhD studentship funded by the AICR (Association of International Cancer Research), award 03-127, and a project grant, C18001, from the Biotechnology and Biological Sciences Research Council.
Molecular Basis of Transcription
References 1 Escriva, H., Bertrand, S. and Laudet, V. (2004) Essays Biochem. 40, 11–26 2 Lavery, D.N. and McEwan, I.J. (2005) Biochem. J. 391, 449–464 3 Dvir, A., Conaway, J.W. and Conaway, R.C. (2001) Curr. Opin. Genet. Dev. 11, 209–214 4 McEwan, I.J. and Gustafsson, J. (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 8485–8490 5 Reid, J., Murray, I., Watt, K., Betney, R. and McEwan, I.J. (2002) J. Biol. Chem. 277, 41247–41253 6 Betney, R. and McEwan, I.J. (2003) J. Mol. Endocrinol. 31, 427–439 7 Kamada, K., De Angelis, J., Roeder, R.G. and Burley, S.K. (2001) Proc. Natl. Acad. Sci. U.S.A. 98, 3115–3120 8 Wright, P.E. and Dyson, H.J. (1999) J. Mol. Biol. 293, 321–331
9 Reid, J., Kelly, S.M., Watt, K., Price, N.C. and McEwan, I.J. (2002) J. Biol. Chem. 277, 20079–20086 10 Kumar, R., Betney, R., Li, J., Thompson, E.B. and McEwan, I.J. (2004) Biochemistry 43, 3008–3013 11 Semisotnov, G.V., Rodionova, N.A., Razgulyaev, O.I., Uversky, V.N., Gripas’, A.F. and Gilmanshin, R.I. (1991) Biopolymers 31, 119–128 12 Bailey, R.W., Dunker, A.K., Brown, C.J., Garner, E.C. and Griswold, M.D. (2001) Biochemistry 40, 11828–11840 13 Choudhry, M.A., Ball, A. and McEwan, I.J. (2006) Mol. Endocrinol. 20, 2052–2061
Received 28 July 2006
C 2006
Biochemical Society
1057
