The nuclear receptor ligand-binding domain - CMBI

384

The nuclear receptor ligand-binding domain: structure and function Dino Moras* and Hinrich Gronemeyert In the past few years our understanding of nuclear receptor action has dramatically improved as a result of the elucidation of the crystal structures of the empty (apo) ligand-binding domains of the nuclear receptor and of complexes formed by the nuclear receptor's ligand-binding domain bound to agonists and antagonists. Furthermore, the concomitant identification and functional analysis of co-regulators (transcriptional intermediary factors [TIFs], comprising co-activators and co-repressors) previously predicted from squelching studies, have deepened this understanding. Recent data have provided the structural basis for the specific recognition of ligands and the molecular mechanisms of agonism and antagonism, enabling us to gain a comprehensive view of the early steps of nuclear receptor action.

Addresses

Institut de G6n~tique et de Biologie Mol6culaire et Cellulaire (IGBMC), CNRS/INSERM/ULP, BP 163, 6?404 IIIkirch Cedex, Cit~ Universitaire de Strasbourg, France *e-mail: [email protected] re-mail: [email protected] Current Opinion in Cell Biology 1998, 10:384-391

http://biomednet.com/elecref/0955067401000384

development, control maintenance of homeostasis, and induce or inhibit cellular proliferation, differentiation and death. Some members of this superfamily act as transcriptional suppressors, whereas for a large group of so-called 'orphan' NRs no ligands have yet been found. Evidence has been presented that ligand binding is an ability acquired during evolution and that the ancestral NR was an orphan [1"]. T h e genetic activities of NRs result from both direct modulation of the activity of cognate gene programmes and mutual interference with the activity of other signalling pathways. About 70 NRs have been identified to date and with the exception of some unusual NRs, which appear to contain only regions homologous to the DNA- or ligand-binding domains (DBD and LBD, respectively), all members display an identical structural organization with an aminoterminal region A/B, followed by a DBD consisting of two zinc fingers (region C), a linker region D, and the LBD. Some but not all NRs contain a carboxy-terminal region F of unknown function. Two autonomous transactivation functions, a constitutively active activation function (AF)-I originating in region A/B and a ligand-dependent AF-2 arising in the LBD, are responsible for the transcriptional activity of NRs [2].

© Current Biology Ltd tSSN 0955-0674 Abbreviations AP1 activatingprotein 1 AD transactivationdomain

AF CBP DBD

ERc~ H LBD

LBP NID NR N-CoR

p/CAF PPAR3' RAR RXR SMRT

activationfunction cAMP response element binding protein DNA-bindingdomain estrogenreceptor ct helix ligand-bindingdomain ligand-bindingpocket NR interaction domain nuclearreceptor nuclearreceptor co-repressor CBP associated factor peroxisome proliferator activated receptor y retinoicacid receptor retinoidX receptor

SRC TIF TR

silencingmediator for RXR and TR steroid receptor co-activator transcriptionalintermediary factor thyroid hormone receptor

VDR

vitaminD receptor

Introduction Nuclear receptors (NRs), such as the receptors for steroids and thyroid hormones, retinoids and vitamin D3, are ligand-inducible transcription factors present in vertebrates, arthropods, and nematodes. T h e y regulate complex physiological events that trigger key steps during

NRs bind as homodimers (e.g. steroid receptors, retinoid X receptor [RXR]) and/or heterodimers (e.g. the retinoic acid receptor [RAR], thyroid hormone receptor [TR] and vitamin D receptor [VDR]) along with the promiscuous heterodimerization partner, RXR, to cognate response elements of target genes ([3] and references cited therein). NR activities are not confined to cognate target genes since some NRs can also 'crosstalk' in a ligand-dependent fashion with other signalling pathways, leading to mutual interference (which can be positive or negative) with the transactivation potentials of the involved factors (e.g. activating protein 1 [AP1], nuclear factor ~(B) [4]. Furthermore, other signalling pathways (e.g. those involving mitogen-activated protein kinase [5], and cyclin-dependent kinase 7 [6]) can target NRs directly and modify the activity of their AFs. Recent years have seen a dramatic enhancement of our understanding of NR action at the molecular level. This is due to progress in various areas, the first of which is the identification and characterization of several novel classes of transcriptional mediators (transcriptional intermediary factors [TIFs]/co-regulators, comprising co-activators and co-repressors) for the AF-2 domain present in NR LBDs (for an illustration of the TIF2/steroid receptor co-activator (SRC)-I co-activator family see Figure 1). Furthermore,

The nuclear receptor ligand-binding domain: structure and function Moras and Gronemeyer

the first steps have been taken towards the decryption of a plethora of interactions reported to occur between NRs, mediators, chromatin and the basal transcription machinery (for reviews and further references see [7-9]; see Figure 2 for the current model of N R action). T h e second area in which progress has been made is the genetic analysis of NR function (for reviews see [10,11]); the third

385

is the identification of novel candidate ligands for known and novel 'orphan' receptors [12,13"]. T h e last area of progress is the determination of the three-dimensional structures of the empty (apo-), agonist-bound (holo-), and antagonist-bound LBDs of several NRs [14-16,17"'] together with the prediction of a common fold for all NR LBDs [18].

Figure 1

NLS bHLH PAS domain

NID

CID/AD1

AD2

(LLxxLxxxL) r

-I

I-c mGRIP1/NCoA-2 hTIF2

,

NID

'

AD1 '

t

AD2

dn I

coactivating

@ N.

.

.

mSRC-1c _ _ ,mSRC.ld~ 4 l I el -

. t _ _ , NID

i

'

AD1

!

,hSRC-l,e| i

= I AD2

hSRC-la -C

mSRC-1 a

i

dn

I-IAT 'p/CAIF.. ,-- h A C T R

l

,,,-

'

NID

,

30-420/0

ioll--i i-c

p/CIPJ[)23(Q RAC3/AIB1 • (Q)28/30 I

57-620/0

''

A38

hAIB 1/TRAM- 1/RAC3/ACTR mp/ClP

I

p/CAF , CID ,' HAT

38-42O/o

' AD2

V

34-440/0

Current Opinion in Celt Biology

The TIF2/SRC-1 family of nuclear receptor co-activators for the AF-2 function of nuclear receptors (TIF, transcriptional intermediary factor; SRC, steroid receptor co-activator). Schematic alignment, drawn to scale, of the peptide sequences of the three known human orthologues (prefix 'h') TIF2 [43], SRC-la [44] and AIB1 [45°°]/RAC3 [46]/ACTR [47]/TRAM-1 [48]. The names of the respective murine hornologs are indicated on the right (prefix 'm'). Isoforms (splicing variants) are represented by square brackets above the sequences, conserved motifs by vertical bars (NR boxes I,I1,111and iV) or shaded boxes, and alignment gaps by dotted lines. Polyglutamine insertions of different lengths are indicated for the bottom sequence; Q, glutamine-rich region. Note that the few significant peptide sequence differences between mp/CIP [49] and hAIB1, including the distinct carboxyl terminus, can entirely be attributed to reading frame shifts; therefore mp/CIP represents the murine homolog of AIBI/RAC3/ACTR/TRAM-1. Likewise mNCoA-2 corresponds to mGRIP1. General features based on sequence homology (NLS, nuclear localization signal; bHLH, basic-helix-loop-helix motif; PAS, Per-Amt-Sim homology region) or functional evidence (NID, NR interaction domain; CID, CBPtp300 interaction domain; AD1, AD2, activation domains) are denoted at the top of the figure. Below the sequences, square brackets refer to functional regions determined by mutational analysis (dn, dominant-negative; HAT, histone acetyltransferase activity; p/CAF, p/CAF interacting regions). Percentage amino acid identities between the three orthologues are denoted for each region. Sequence alignment was performed using the CLUSTALW program [50]. GenBankJEMBL accession numbers: hTIF2, X97674; mGRIP1, U39060; mNCoA-2, AF000582; hSRC-la, U90661; mSRC-la, U56920; hHin2 (part of hSRC-le), U19179; mSRC-1 (mSRC-le), U64828; hAIB1, AF012108 (identical: TRAM1, AF016031 ; hACTR, AF036892, but note an insertion of nine residues in the PAS domain and putative splice at the position of the TIF2 splice); hRAC3, AF010227; mp/CIP, AF000581.

386

Nucleus and gene expression

Figure 2

a)

(b)

/~ytoplasm

f

Cytoplasm (

N/~ucleus ~ ~

~

Histone ~ Deacetylation

f

N

u

c

~

Acetylation

ligand (ho~one)

aganW~-

~r"

/ I

Ifntrt~:m e

k I

( ~-n~"'l,I

~

F

I

~ /

~

Histones

RNA Pol II

,

\ Current Opinion in Cell Biology

Proposed models for NR function. An NR homodimer or heterodimer targeted to its response element (RE) is depicted as two sets of three spheres representing the three functional domains AFol (hidden), DNA-binding domain (DBD) and ligand-binding domain (apo- or holo-LBD). (a) Silencing (transrepression) involves the recruitment of co-repressors (N-CoR, SMRT) and, indirectly, histone deacetylases such as RPD3; transcription is blocked, presumably due to chromatin compaction. (b) Transactivation is triggered by binding of the respective ligand (black triangles; originating from intracellular or extracellular sources), which enables the association of co-activators (e.g. members of the TIF2/SRC-1 family and CBP/p300) that possess histone acetyltransferase activity or recruit the histone acetyltransferase p/CAR Together these factors form a co-activator complex which 'decondenses' chromatin, thus permitting the establishment of a transcription initiation complex (basal factors, TBP [TATA box binding protein], RNA polymerase II).

The mouse

trap model

To date the crystal structures of five NR LBDs have been described; the dimeric apo-RXR(~ [14], monomeric holo-RARy [15], monomeric holo-TR~ [16], dimeric holo (estrogen)- and antagonist (raloxifen)-bound estrogen receptor o~ (ERct) [17°°], and apo- and holo (thiazolidinedione)-peroxisome proliferator activated receptor y (PPARy) [19]. All these N R LBDs display a common fold, as originally predicted [14,15,18], with 12 o~ helices (numbered H1-H12) and one 13 turn arranged as an antiparallel ~helical 'sandwich' in a three layer structure (Figure 3). Note that some variability exists; for example, no helix H2 was found in RARy [15], while an additional short helix H2' is present in PPARy [19]. Together the various apo- and holo-LBD structures suggest a common mousetrap-like mechanism by which AF-2 becomes competent to activate transcription: upon ligand binding, H l l is repositioned in the continuity of HI0, and the concomitant swinging of H12 unleashes the f2-1oop (between H2 and H3) which flips over underneath H6, carrying along the amino-terminal part of H3. Figure 3 illustrates these structural alterations upon ligand binding (referred to as LBD transconformation) which includes bending of H3. In its final position, H12 seals as a 'lid' the ligand-binding cavity and further stabilizes ligand binding (in some but not all NRs) by contributing to additional ligand-protein interactions. Note that in PPARy without bound ligand H12 occupies a position which is much closer to its position in the holo-LBD [19]. Recently, the identification of a constitutively active RXR mutant has

provided insight into the initial events occurring upon ligand binding [20"]. T h e transconformation of H12, together with additional structural changes (such as bending of helix H3; compare Figures 3a, b), brings it into a distinct receptor environment, thus creating the surface(s) which allow binding by bona fide co-activators, such as the members of the SRC-1/TIF2 family. Structure-function analyses [21-23] have revealed that co-activators are composed of distinct domains responsible for N R binding (NR interaction domain [NID]) and transactivation (transactivation domain [AD]I and AD2). N I D s can contain one [24"] or several redundant [25"] L x x L L N R boxes (where x is any amino acid and L is leucine in single-letter code for amino acids). Interestingly, T I F 2 AD1 is indistinguishable from the domain that interacts with cAMP response element binding protein (CBP), denoted CBP interaction domain (CID) in Figure 2. This suggests that AD1 functions via CBP recruitment, while AD2 does not interact with CBP [21]. It appears therefore that N R AF-2s function via the establishment of a multiprotein co-activator complex comprising the receptor, member(s) of the SRC-1/TIF2 co-activator family, CBP and, possibly, the CBP-binding factor p/CAF (Figure 2). Several of these factors have been shown to possess acetyltransferase activity, and it is believed that chromatin disruption is a critical, albeit apparently insufficient, step in NR-dependent gene activation [26]. It is therefore possible that the ligand-dependent AF-2 of NRs serves to assemble a complex which can open up chromatin, thus allowing


387

Figure 3

(a)

/N

(b)

N,

(c)

N,

_

f~ loop

Current Opinion in Cell Biology

Schematic drawing of three conformational states of nuclear receptor ligand-binding domains (LBDs). (a) The apo-LBD from the crystal structure of the apo-retinoic acid receptor (RAR)~ LBD. (b) The holo-(agonist-bound) LBD according to the structure of the all.trans retinoic acid-RARy LBD complex. Agonist is shown in 'stick' form in the molecule's center. (c) The antagonist-bound LBD according to the observed conformation of the raloxifen-bound estrogen receptor c( LBD. Raloxifen is shown as a bent cylinder in the center of the LBD. Note the different position of helix H12 in each situation. Numbered rods represent c~ helices. The IB turn is depicted as broad arrows, and the amino and carboxyl termini are labeled N and C, respectively.

assembly of the transcriptional machinery. Assembly may be governed by the modification of general transcription factors [27 °] and/or by ligand-independent interactions between either NRs or components of the co-activator complex with general transcription factors. Note that the physical existence of a co-activator complex has yet to be demonstrated. In the absence of ligand some NRs, such as T R and RAR, recruit a complex of opposite functionality, composed of co-repressors (NR co-repressor, [N-CoR] [28] or silencing mediator for RXR and T R (SMRT) [29]), Sin3 and Rpd3, the last displaying histone deacetylase activity. It is believed that this complex is responsible for the gene silencing ability of some NRs (Figure 2a; for details and further references, see the review by J Torchia, C Glass and MG Rosenfeld in this issue, pp 373-383). For a given receptor, the equilibrium between the apo and holo (or apo and antagonist) conformational states of an NR LBD can be altered through intramolecular interactions of H12, such as a salt bridge (holo LBD of RARy [15]) or hydrophobic contacts (as suggested for apo-ER [30°]). This implies that the apo conformation is not necessarily the default state, so that some NRs may be constitutive activators or repressors without possessing

a cognate ligand. Moreover, an increase in co-activator concentration can generate a transcriptionally competent RAR under certain conditions [21] and the apo-ER conformation may be destabilized by phosphorylation [30e,31]. Thus, overexpression of co-activators or receptor modification may generate ligand-independent receptors. Such scenarios could have significant implications for endocrine cancer therapies, because antagonist action may be impaired in cancer cells that overexpress co-activators.

Role of H12 positioning H12 is a crucial helical component of the NR LBD, because its ligand-induced repositioning in the holo NR provides the surface(s) for co-activator interaction and thereby generates the transcriptional activity of the AF-2 domain. T h e critical importance of a carboxy-terminal LBD motif that displays features conserved in all AF-2containing NRs and that is generally referred to as the core of the AF-2 AD had been demonstrated by mutational analyses before the LBD structure was solved and was shown to correspond to the H12 amphipathic ct helix [2]. T h e first structures of apo- and holo-LBDs to be described highlighted the importance of the ligand-induced conformational changes and immediately suggested that the interactions between H12, or residues in its proximity, and the ligand was critical for the control of ag-

388


onist/antagonist properties of nuclear receptors [14,15,18]. T h e crystal structure of the ERc~ LBD-raloxifen complex confirmed this hypothesis and showed an alternative stable 'antagonist position' for H12, nested between H4 and H3 [17°']. Steric hindrance upon binding of a ligand with bulky substitutions projecting towards the 'lid' of the LBP prevents the proper positioning of H12 in its 'agonistic' site. Although the necessity of a unique binding site for all antagonists is not obvious in view of the existence of distinct types of antagonists which differently modulate co-regulator interaction, this flip-flop mechanism between two positions is appealing. To summarize, upon ligand binding H I 2 would dynamically adopt one of (at least) two stable positions, as dictated by the binding characteristics of the ligand; in the absence of ligand H12 would adopt the apo form. Going from the apo to the agonist form necessitates conformational changes which are most likely to be responsible for disrupting the interactions with the co-repressors. Whether transconformation of H12 is directly involved in ligand-induced co-repressor dissociation is not proven, but is suggested by the observation that deletion of H12 results in the constitutive binding of the SMRT co-repressor [29]. While the position of H12 in its active (holo) agonist conformation creates the correct surface which allows the binding of co-activators, its positioning in any other place changes the shape and the charge distribution of the cognate recognition surface(s) and no co-activator interaction can occur. Consequently, ligands inducing non-holo H12 positions act as antagonists [17"']. Topological considerations predict that the cognate surface(s) for co-activator interaction encompass(es) H12 and part of H4 and H3. This hypothesis has recently been confirmed by co-crystallization of a 13 residue peptide of the T I F 2 / G R I P - I N R box 2 [21,23] with the LBD of TRI3 to about 3.7~, resolution (B Darimont et al. personal communication). In solution, this N R box peptide has no detectable secondary structure, but forms an amphipathic (x helix in complex with TR[~. T h e hydrophobic site of this helix, formed by the conserved, hydrophobic residues of the NR-box motif, interacts with hydrophobic residues from three helices (H3, H4 and HI2) which form a shallow hydrophobic cleft as in RARy and ERot. This cleft is where H12 binds in its antagonist position in the raloxifen-ER~ complex. Based on mutant analyses it has been proposed that helix H1 (encompassing the so-called 'CoR [co-repressor] box') is directly involved in co-repressor binding [28]. From the present crystal structures, however, it is evident that the mutated residues cannot make direct interactions with the protein partner, since their sidechains point towards the core of the LBD. T h e most probable effect of the so-called A H T mutation (a mutation of three residues within the region of H1 postulated to correspond to the N-CoR binding site [28]) is rather to dissociate H1 from the core of the LBD, thus disrupting the surface of the

LBD at that location. T h e observation that TR[3 LBD mutants with gross deletions still interact efficiently with N-CoR [28] would imply that co-repressor binding does not require LBD structural integrity and deserves further attention.

The dimer interface T h e majority of studies addressing NR--co-regulator interactions have considered only N R monomers. Our own data indicate that heterodimers can exhibit a very different and, in view of the monomer results, sometimes surprising plethora of co-regulator interactions (C Zechel and H Gronemeyer, unpublished data). In view of the fact that in most cases N R homodimers and heterodimers are the biologically active species, the roles of HI, H12 and ligand-dependent intradimeric allosteric effects have to be taken into account to reveal a comprehensive molecular view of N R LBD function. Three crystal structures [14,17"', 19] report atomic descriptions of NR homodimers. As predicted from mutagenesis studies, the homodimeric contacts involve helix H10. What was unexpected was the high degree of similarity between the dimer interfaces of the three structures, first described in the crystal structure of apo-RXR~ and now observed in ERot and PPARy. T h e key player which locks the dimer interface is indeed H10, which can self-associate through hydrophobic contacts, but helices H9, H7 and the loops connecting H7 and H8 also contribute to dimer stability. T h e interesting feature of holo-homodimers is the proximity of H12 of one monomer to H7 in the other one, which could provide the basis for an allosteric crosstalk between subunits. If heterodimers were similarly associated, which remains to be confirmed, this close proximity of H12 to the other subunit could play a significant role in the function of the NR heterodimer. One such role might be RXR subordination--the inability of RXR to respond to cognate agonists when RXR forms a heterodimer with non-liganded RAR, T R or vitamin D receptor ([20"] and references therein). A peculiar aspect of the oligomeric state of NRs is the observation that apo-RXR, in contrast to other apo-NRs, can form tetramers which dissociate upon ligand binding ([32] and references therein). T h e functional meaning and the structure of these entities is not yet established but an attractive view is that these tetramers represent a storage form of this promiscuous heterodimerization partner.

The ligand-binding pocket In all crystal structures presently available the ligand is embedded within the protein with no clear accessible entry or exit site. PPARy seems the only exception to that rule, since a potential access cleft to the ligand-binding pocket (LBP) was observed between H3 and the 13 turn which could be of sufficient size to allow entry of small ligands without major adaptation [19]. For all other receptors of known structure, significant conformational

The nuclear receptor ligand-binding domain: structure and function Moras and Gronemeyer 389

changes are necessary to generate potential entry sites. T h e mouse trap model provides an easy solution to the problem: the mobility of H12 opens a channel by removing the 'lid' from the LBP. A distinct entry site close to the 13 turn has been proposed by Wagner et al, [16]. While this hypothesis may be true for some NRs, in most receptors the necessary conformational changes at that location would require the disruption of multiple hydrogen bonds. In addition, the electrostatic considerations discussed by Renaud eta/. [15] would make this type of entry energetically very costly for most NRs. T h e LBPs are predominantly lined with hydrophobic residues. A few polar residues at the deep end of the pocket near the 13 turn act as anchoring points for the ligand or play an essential role in the correct positioning and enforce the selectivity of the pocket. Most NRs contain a conserved arginine within helix H5 which points into this part of the cavity. T h e anchoring polar residues are conserved within a given NR subfamily (e.g. retinoid or steroid receptors) and are indicative of the polar group characteristics of each family of ligands (i.e. carboxylate for retinoids and ketone for steroids). Their conservation strongly suggests that within each subfamily the ligands are positioned in the same way: this is supported by LBD swapping data which indicate that testosterone is identically oriented in its receptor LBP [33] as estradiol in the ER LBP [17"].

Ligand selectivity As shown in the cases of RARy [15] and TR13 [16], the shape of the LBP matches that of the ligand. This structural information is apparently built into the protein sequence as is demonstrated by the homology modeling of vitamin D receptor (VDR) based on the crystal structure of holo-RARy. This model exhibits a LBP which can perfectly accommodate vitamin D3 and its analogs and predicts critical LBP-ligand contacts [34]. T h e accordance of shape and volume maximize the number of hydrophobic contacts, thus contributing to the stability of the complex and the selectivity of the pocket for the cognate ligand. RAR possesses an interesting LBD, since it can bind equally well two chemically different ligands: all-trans retinoic acid and its 9-cis isomer. Crystallographic analysis [35 °] of the two ligands in the RARy LBP showed that both adapt conformationally to the LBP, which acts as matrix. Moreover, the conformation of a RARy-selective agonist was also shown to closely match that of the natural ligands in their bound state [35"]. T h e adaptation of ligands to the protein leads to an optimal number of interactions for binding and selectivity, and justifies modelling approaches for ligand design. For steroid receptors, the LBP volume is significantly larger than the volume of the corresponding ligands and the rigidity of the ligands does not allow adaptability. Therefore, selectivity cannot be driven by multiple hydrophobic contacts which would anyway not suffice to

discriminate between small structurally similar ligands. In this case specific key interactions are more important. Note that very large LBP volumes allow for the binding of multiple ligands with different stereochemistries, such as in the case of PPARy [19], often at the expense of lower binding affinities. As we observed in a structure-based sequence alignment [15] that only three residues diverged in the LBPs of RARc~, 13 and y, we predicted that these divergent residues were critically involved in differentiating between isotype-selective retinoids. Indeed, swapping of these residues confirmed our hypothesis (M Gehin et al., unpublished data). Moreover, swapping of these residues not only conferred isotype-selective binding but also conferred the agonistic/antagonistic ligand response onto any other RAR isotype.

Conclusions Despite, or because of, the rapid accumulation of knowledge about NR action, numerous questions remain to be answered. There are four major questions. One concerns the structure and function of the so-called orphan receptors, some of which may not necessarily be orphans (see above) but rather represent descendents of primordial NR activators/repressors that have not acquired ligand-binding capability. T h e second question concerns the biochemical definition of the stoichiometries of complexes formed by co-activators or co-repressors and NR homodimers or heterodimers. It includes defining the contribution of the so-far neglected AF-1, and gaining an understanding of the structural basis and NR specificity of the recognition principles that establish these complexes. It is surprising that, at present, only the androgen receptor can apparently afford its own co-activator, ARAT0 [36], while all other bona fide or putative co-activators are promiscuous. T h e first reports of NR-specific differences in co-regulator interaction are now emerging [23,37]. Further three-dimensional structures of NR-co-regulator complexes are required to determine the details of the interaction surfaces, which may represent a novel type of drug target. The third area for focused research is the elucidation of intramolecular and intermolecular allosteric effects, including the structural roles of NR regions D and F [38"]. Examples of intramolecular effects are the following: the constitutively active AF-I is repressed by apo-LBD; agonists and some, but not all, antagonists relieve this repression; some of these antagonists can thus act as cell- and promoter-specific AF-1 agonists (see [2] and references therein). Intermolecular effects are exemplified by RXR subordination [20°] and can be caused by DNA binding [39]. T h e last question involves determining the mechanisms and structural basis of the multiply faceted signal transduction crosstalks, such as ER signalling through API sites [40,41"] and mutual transrepression between several NRs and AP1 or nuclear factor K:B [4], as well as receptor [5,6,42] and co-regulator modifications which result in modulation of NR activity.

390


After having accomplished a major leap towards a molecular understanding of the structure and function of the NR LBD we should soon have the answers to most of the above questions. In particular we look forward to seeing the structure of complexes between NRs and their cognate co-regulatory proteins.

Note added in proof An interesting cold-inducible NR co-activator has been identified very recently. This factor harbours a single LxxLL NR box motif and interacts promiscuously with NRs. But in contrast to other T I F s it exhibits a tissue-selective expression [51]. The allosteric effects that DNA binding can impose on transactvators have been comprehensively discussed in a recent review [52].

Acknowledgements We apologize to all our colleagues whose original contributions could not be cited due to the space limitations; these publications are either cited in the reviews mentioned or can be obtained from the authors. We thank Matthias Heine and Raymond Ripp for compilation of Figures 1 and 3, respectively. Studies in our laboratories are supported by funds from Institut National de la Sant6 et de la Recherche M6dicale, Centre National de la Recherche Scientifique, Centre Hospitalier Universitaire, Fondation pour la Recherche Mddicale, the Ministdre de la Recherche et de la Technologie, the Colldge de France, the EC BIOMED Progamme, Schering AG, RousseI-Uclaf and Bristol-Myers Squibb. ~ ......

References and recommended reading

12. 13. ••

Kliewer SA, Moore .IT, Wade L, Staudinger JL, Watson MA, JonesSA, McKee DD, Oliver BB, Willson TM, Zetterstr~m RH, Perlmann T, Lehmann JM: An orphan nuclear receptor activated by pregnanes defines a novel steroid signaling pathway. Cell 1998, 92:73-82. Cloning and analysis of a novel nuclear receptor, PXR, which activates CYP3A gene expression in response to the binding of various natural and synthetic agonists and antagonists and is likely to be involved in the regulation of steroid metabolism. 14.

Bourguet W, Ruff M, Chambon P, Gronemeyer H, Moras D: Crystal structure of the ligand-bindJng domain of the human nuclear receptor RXRc~. Nature 1995, 375:377-382.

15.

RenaudJP, Rochel N, Ruff M, Vivat V, Chambon P, Gronemeyer H, Moras D: Crystal structure of the RARy ligand-binding domain bound to all-trans retinoic acid: Nature 1995, 378:681-689.

16.

Wagner RL, Apriletti JW, McGrath ME, West BL, Baxter JD, Fletterick RJ: A structural role for hormone in the thyroid hormone receptor. Nature 1995, 378:690-697.

t 7. o•

Brzozowski AM, Pike ACW, Dauter Z, Hubbard RE, Bonn T, EngstrsmO, Ohman L, Greene GL, Gustafsson J-A, Carlquist M: Molecular basis of agonism and antagonism in the •estrogen receptor. Nature 1997, 389:753-758. First crystal structure of a steroid receptor ligand-binding domain in the presence of estradiol or the AF-2 (but not AF-1) antagonist raloxifen. This study reveals that helix 12 can adopt a holo (agonist) and a distinct antagonist conformation (compare Figure 2). 18.

Wurtz J-M, Bourguet W, Renaud J-P, Vivat V, Chambon P, Moras D, Gronemeyer H: A canonical structure for the Ugand-binding domain of nuclear receptors Nat Struct Bio/1996, 3:87-94. [Published erratum appears in Nat Struct Bio/1996, 3:206].

19.

MilburnMV, Charifson P, Lambert M, Cobb J, Wisely GB: Ligand binding and activation of PPARs-crystal structures of PPARy. [Abstract]. EMBO Workshop on 'Structure and Function of Nuclear Receptors', May 2-5, 1997, Erice, Italy.

Papers of particular interest, published within the annual period of review, have been highlighted as: • of special interest 0• of outstanding interest 1. •

Escriva H, Sail R, Hanni C, Langlois MC, Saumitou-Laprade P, Stehelin D, Capron A, Pierce R, Laudet V: Ligand binding was acquired during evolution of nuclear receptors. Proc Natl Acad Sci USA 1997, 94:6803-6808. This study shows that nuclear receptors are specific to metazoans and that their diversity was generated by two waves of gene duplications during evolution. The authors conclude from their study that the primordial NRs acted as 'orphan' activators or repressors which acquired ligand-binding ability during evolution.

Mangelsdorf DJ, Evans RM: The RXR heterodimers and orphan receptors. Ceil 1995, 83:841-850.

20. •

Vivat V, Zechel C, Wurtz J-M, Bourguet W, Kagechika H, Umemiya H, Shudo K, Moras D, Gronemeyer H, Chambon, P: A mutation mimics the ligand-induced transconformation of the retinoid X receptor. EMBO J 1997, 16:5697-5709. The identification of a constitutive RXR mutant suggests that the initial event occurring in the RXR LBD upon ligand binding is the disruption of a network of van der Waals interactions. The mutant is used to support the concept of RXR subordination, i.e. the inability of RXR to respond to cognate agonists in heterodimers with RAR, TR and VDR. 21.

Voegel JJ, Heine MJS, Tini M, Vivat V, Chambon P, Gronemeyer H: The coactivator TIF2 contains three nuclear receptor-binding motifs and mediates transactivation through CBP bindingdependent and -independent pathways. EMBO J 1998, 17:507519.

22.

KalkhovenE, Valentine JE, Heery DM, Parker MG: Is•forms of steroid receptor coactivator 1 differ in their ability to potentiate transcription by the •estrogen receptor. EMBO J 1998, 17:232243.

23.

Ding XF, Anderson CM, Ma H, Hong H, Uht RM, Kushner PJ, Stallcup MR: Nuclear receptor binding sites of coactivators GRIPI and SRC-I: multiple motifs with different binding specificities. Mol Endocrino11998, in press.

2.

Gronemeyer H, Laudet V: Nuclear Receptors. Protein Profile 1995, 2:1173-1308.

3.

Gronemeyer H, Moras D: Nuclear receptors, How to finger DNA. Nature 1995, 375:190-191.

4.

Gsttlicher M, Heck S, Herrlich P: Transcriptional crosstalk, the second mode of steroid hormone action. J Mo/Med 1998, in press.

5.

Kato S, Endoh H, Masuhiro Y, Kitamoto T, Uchiyama S, Sasaki H, Masushige S, Gotoh Y, Nishida E, Kawashima H et a/. : Activation of the estrogen receptor through phosphorylation by mitogenactivated protein kinase. Science 1995, 270:1491-1494.

6.

Rochette-Egly C, Adam S, Rossignol M, Egly JM, Chambon P: Stimulation of RAR alpha activation function AF-1 through binding to the general transcription factor TFIIH and phosphorylation by CDK7. Ceil t 997, 90:97-10Z

7.

Glass CK, Rose DW, Rosenfeld MG: Nuclear receptor coactivators. Curt Opin Cell Bio/1997, 9:222-232.

Le Douarin B, Nielsen AL, Gamier JM, Ichinose H, Jeanmougin F, Losson R, Chambon P: A possible involvement of TIFla and TIFI~ in the epigenetic control of transcription by nuclear receptors. EMBO J 1996, 15:6701-6715. Contains the first description of a NR box (a LxxLL motif which is necessary and sufficient to mediate agonist-and AF-2 integrity-dependent interaction of TIF1(x with nuclear receptors) in the putative co-activator TIFI(x.

8.

Wolffe AP: Transcriptional control. Sinful repression. Nature 1997, 387:16-17.

25. •

9.

Pazin MJ, Kadonaga JT: What's up and down with histone deacetylation and transcription? Ce//1997, 89:325-328.

10.

Beat• M, Herrlich P, Sch9tz G: Steroid hormone receptors: many actors in search for a plot. Ce//1995, 83:851-857.

11.

KastnerP, Mark M, Chambon P: Nonsteroid nuclear receptors: What are genetic studies telling us about their role in real life. Cell 1995, 83:859-869.

24. •

Heery DM, Kalkhoven E, Hoare S, Parker MG: A signature motif in transcriptional co-activators mediates binding to nuclear receptors. Nature 1997, 387:733-736. Identification of the LxxLL NR boxes of RIPt 40 (nine copies), SRC-1 (four copies) and CBP (two copies) which interact with different efficiencies with the ER LBD in an agonist and AF-2 integrity-dependent manner. The authors show that mutation of the four NR boxes of SRC-1 abrogates its stimulation of •estrogen-induced ER transactivation. 26.

Wong J, Shi YB, Wolffe APi Determinants of chromatin disruption and transcriptional regulation instigated by the


thyroid hormone receptor: hormone-regulated chromatin disruption is not sufficient for transcriptional activation. EMBO J 1997, 16:3158-31 71. 27. •

39.

391

Starr DB, Matsui W, Thomas JR, Yamamoto KR: Intracellular receptors use a common mechanism to interpret signalling information at response elements. Genes Dev 1996, 10:12711283.

Imhof A, Yang X-J, Ogryzko W, Nakatani Y, Wolffe AP, Ge H: Acetylation of general transcription factors by histone acetyltransferases. Curt Biol 199?, 7:689-692. This paper reports that not only histones but also the general transcription factors TFIIE~ and TFilE can be substrates in vitro of the histone acetyltransferases pCAF, p300 and TAFII250.

Description of a novel type of transcription activation-independent liganddependent FIp recombinase assay which reveals the initial event occurring upon ligand binding the estrogen receptor. The results demonstrate that agonists and antagonists differently position helix 12 and implicate the F region in mechanisms of antagonist action.

28.

40.

29.

HOrleinAJ, N~i~r AM, Heinzel T, Torchia J, Gloss B, Kurokawa R, Ryan A, Kamei Y, S6derstrOm M, Glass CK, Rosenfeld MG: Ligand-independent repression by the thyroid hormone receptor mediated by a nuclear receptor co-repressor. Nature 1995, 377:397-404. Chen JD, Evans RM: A transcriptional co-repressor that interacts with nuclear hormone receptors. Nature 1995, 377:454-457.

30. •

White R, Sloberg M, Kalkhoven E, Parker MG: Ligandindependent activation of the oestrogen receptor by mutation of a conserved tyrosine. EMBO J 1997, 16:1427-1436. The authors describe that mutations of the human estrogen receptor Tyr573 to Asp, Glu or Ala generate estrogen receptor mutants that are transcriptionally active in the absence of a iigand. They hypothesize that the mutation may disrupt interactions which arrest helix H12 in its apo position and speculate that phosphorylation of this tyrosine may have the same effect, thereby circumventing ligand control of receptor activity. Similar results are reported in Weis eta/. 1996 [31].

Gaub MP, Bellard M, Scheuer I, Chambon P, Sassone-Corsi P: Activation of the ovalbumin gene by the estrogen receptor involves the fos-jun complex. Cell 1990, 63:1267-1276.

41.

Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J, Kushner PJ, Scanlan TS: Differential ligand activation of estrogen receptors ERcc and ER(3 at AP1 sites. Science 1997, 277:1508-1510. The two ER isotypes were found to transactivate differently in response to several ligands through AP1 sites, although an identical transactivation pattern was observed of a classical oestrogen response element-containing reporter. These results suggest that the two isoforms may have distinct signaling abilities in vivo. •

42.

Hu E, Kim JB, Darraf P, Spiegelmann BM: inhibition of adipogenesis through MAP kinase-mediated phosphorylation of PPARy. Science 1996, 274, 2100-2103.

43.

Weis KE, Ekena K, Thomas JA, Lazennec G, Katzenellenbogen BS: Constitutively active human estrogen receptors containing amino acid substitutions for tyrosine 537 in the receptor protein. Mol Endocrinol 1996, 10:1388-1398.

Voegel JJ, Heine MJS, Zechel C, Chambon P, Gronemeyer H: TIF2, a 160 kDa transcriptional mediator for the ligand-dependent activation function AF-2 of nuclear receptors. EMBO J 1996, 15:3667-3675.

44.

Oate SA, Tsai SY, Tsai MJ, O'Malley BW: Sequence and characterization of a coactivator for the steroid hormone receptor superfamily. Science 1995, 270:1354-1357.

32.

Chen ZP, lyer J, Bourguet W, Held P, Mioskowski C, Lebeau L, Noy N, Chambon P, Gronemeyer H: Ligand- and DNA-induced dissociation of RXR tetramers. J Mo/Biol 1998, 275:55-65.

45. •-

33.

Vivat V, Gofflo D, Garcia T, Wurtz .JM, Bourguet W, Philibert D, Gronemeyer H: Sequences in the ligand-binding domains of the human androgen and progesterone receptors which determine their distinct ligand identities. J Mo/Endocrinol 1997, 18:147-16.

34.

Wurtz JM, Guillot B, Moras D: 3D model of the ligand binding domain of the vitamin D nuclear receptor based on the crystal structure of holo RARy. In Vitamin D, Proceedings of the Tenth Workshop on Vitamin D. Edited by Norman AW, Bouillon R, Thomasset M. Riverside, California: University of California; 1997:165-172.

31.

Klaholz BP, Renaud .Ip, Mitschler A, Zusi C, Chambon P, Gronemeyer H, Moras D: Conformational adaptation of agonists to the human nuclear receptor RART. Nat Struct Biol 1998, in press. First report of the crystal structure of an RAR isotype-specific synthetic retinoid; the comparison of the crystal structures of the LBD complexes with three natural/synthetic ligands supports the concept that the ligand-binding pocket of some nuclear receptors acts as a matrix to which the ligand adapts its conformation.

Anzick SL, Kononen J, Walker RL, Azorsa DO, Tanner MM, Guan XY, Sauter G, Kallioniemi OP, Trent JM, Meltzer PS: AIB1, a steroid receptor coactivator amplified in breast and ovarian cancer. Science 1997, 277:965-958. Identification of a nuclear receptor co-activator which belongs to the TIF2/SRC-1 family (see Figure la). The AIB1 gene, located on chromosome 20q12, was found amplified in 10%, and showed high expression levels in 64%, of 105 unselected primary breast tumors. 46.

Li H, Gomes PJ, Don Chen J: RAC3, a steroid/nuclear receptorassociated coactivator that is related to SRC-1 and TIF2. Proc Nat/Acad Sci USA 1997, 94:8479-8484.

47.

Chen H, Lin PJ, Schiltz RL, Chakravarti D, Nash A, Nagy L, Privalsky ML, Nakatani Y, Evans RM: Nuclear receptor coactivator ACTR is a novel histone acetyltransferase and forms a multimeric activation complex with p/CAF and CBP/p300. Cell 1997, 90:569-560.

48.

Takeshita A, Cardona GR, Koibuchi N, Suen CS, Chin WW: TRAM-l, a novel 160-kDa thyroid hormone receptor activator molecule, exhibits distinct properties from steroid coactivator1. J Bio/ Chem 1997, 272:27629-27634.

49.

Torchia J, Rose DW, Inostroza J, Kamei Y, Westin S, Glass CK, Rosenfeld MG: The transcriptional co-activator p/ClP binds CBP and mediates nuclear-receptor function. Nature 1997, 387:677-664.

60.

Thompson JD, Higgins DG, Gibson TJ: ClustalW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 1994, 22:4673-4680. Puigserver P, Wu Z, Park CW, Graves R, Wright M, Speigelmann BM: A cold inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell 1998, 92:829839.

35. •

36.

37.

36. •

Yeh S, Chang C: Cloning and characterization of a specific coactivator, ARAT0, for the androgen receptor in human prostate cells. Proc Nat/Acac/Sci USA 1996, 93:5517-5521. Henttu PMA, Kalkhoven E, Parker MG: AF-2 activity and recruitment of steroid receptor coactivator 1 to the estrogen receptor depend on a lysine residue conserved in nuclear receptors. Mol Cell Biol 1997, 17:1832-1839. Nichols M, Rientjes JMJ, Stewart AF: Different positioning of the ligand-binding domain helix 12 and the F domain of the estrogen receptor accounts for functionaldifferences between agonists and antagonists. EMBO J 1998, 17:765-733.

51.

52.

Lefstin .IA, Yamamoto KR: AIIosteric effects of DNA on transcriptional regulators. Nature 1996, 392:685-886.