CHAPTER
Dimerization of Nuclear Receptors
2
Pierre Germain*,{ and William Bourguet*,{ *
Inserm U1054, Centre de Biochimie Structurale, Montpellier, France CNRS UMR5048, Universite´s Montpellier 1 & 2, Montpellier, France
{
CHAPTER OUTLINE Introduction .............................................................................................................. 22 2.1 Methods ............................................................................................................ 24 2.1.1 Studies of NR–NR Interactions through Protein Crystallization ............. 24 2.1.1.1 Required Materials .....................................................................26 2.1.1.2 Protocol......................................................................................26 2.1.2 Expression and Purification of NR–NR Complexes ............................... 27 2.1.2.1 Required Materials .....................................................................28 2.1.2.2 Protocol......................................................................................29 2.1.3 Monitoring NR–NR Interactions by Noncovalent Electrospray Ionization Mass Spectrometry ........................................................................... 29 2.1.3.1 Required Materials .....................................................................30 2.1.3.2 Protocol......................................................................................30 2.1.4 Monitoring NR–NR Interactions by Electrophoretic Mobility Shift Assays .......................................................................................... 31 2.1.4.1 Required Materials .....................................................................31 2.1.4.2 Protocol......................................................................................31 2.1.5 Two-hybrid Assays to Define NR–NR Interactions in Living Cells........... 32 2.1.5.1 Required Materials .....................................................................34 2.1.5.2 Protocol......................................................................................34 2.1.6 Fluorescence Cross-Correlation Spectroscopy to Measure the Concentrations and Interactions of NRs in Living Cells ........................ 35 2.1.6.1 Required Materials .....................................................................36 2.1.6.2 Protocol......................................................................................36 Acknowledgments ..................................................................................................... 39 References ............................................................................................................... 39
Methods in Cell Biology, Volume 117 Copyright © 2013 Elsevier Inc. All rights reserved.
ISSN 0091-679X http://dx.doi.org/10.1016/B978-0-12-408143-7.00002-5
21
22
CHAPTER 2 Dimerization of Nuclear Receptors
Abstract Multicellular organisms require specific intercellular communication to properly organize the complex body plan during embryogenesis and maintain its properties and functions during the entire life. While growth factors, neurotransmitters, and peptide hormones bind to membrane receptors, thereby inducing the activity of intracellular kinase cascades or the JAK–STAT signaling pathways, other small signaling compounds such as steroid hormones, certain vitamins, and metabolic intermediates enter, or are generated, within the target cells and bind to members of a large family of nuclear receptors (NRs). NRs are ligand-inducible transcription factors that control a plethora of biological phenomena, thus orchestrating complex events like development, organ homeostasis, immune function, and reproduction. NR–NR interactions are of major importance in these regulatory processes, as NRs regulate their target genes by binding to cognate DNA response elements essentially as homo- or heterodimers. A number of structural and functional studies have provided significant insights as to how combinatorial NRs rely on protein–protein contacts that discriminate geometric features of their DNA response elements, thereby allowing both binding site diversity and physiological specificity. Here, we will review our current understanding of NR–NR interactions and provide protocols for a number of experimental approaches that are useful for their study.
INTRODUCTION Nuclear receptors (NRs) are members of a large superfamily of evolutionarily related transcription factors that regulate genetic programs involved in a broad spectrum of physiological phenomena. The human NR family is classified by sequence into six evolutionary groups of unequal size. The reader is referred to several databases providing comprehensive annotation of the literature on the pharmacology of the 48 human NR gene products, together with relevant information on their structure and function (IUPHAR-DB (http://www.iuphar-db.org/), Nuclear Receptor Signaling Atlas (NURSA, http://www.nursa.org/), and NureXbase (http://nurexbase.prabi. fr)). Among NRs, ligands have been identified for only 24 family members. These receptors are ligand-dependent transcriptional factors that respond directly to a large variety of hormonal and metabolic substances that are hydrophobic, lipid-soluble, and of small size (e.g., retinoic acid or estradiol). The other class of NRs is the group of so-called orphan receptors, for which regulatory ligands are still unknown or may not exist (true orphans) or for which candidates have only recently been identified (adopted orphans). In both groups, NRs can exist as different receptor subtypes that originate from different genes such as the retinoic acid receptors a, b, and g (RARa (NR1B1), RARb (NR1B2), and RARg (NR1B3)), which can exist themselves as several isoforms originating from alternative splicing and differential promoter usage (e.g., RARb1–RARb4) (Germain, Staels, Dacquet, Spedding, & Laudet, 2006; Gronemeyer, Gustafsson, & Laudet, 2004).
Introduction
FIGURE 2.1 Structural and functional organization on nuclear receptors. The structural organization of the DBD and LBD is illustrated by the crystal structures of the RXRa–RARa DBD heterodimer bound to a DR1 response element (Protein Data Bank code 1dsz) and the 9-cis-retinoic acid-bound RXRa–RARa LBD heterodimer (Protein Data Bank code 1xdk). Helices are represented as ribbons and labeled from H1 to H12 (LBD) or H1 and H2 (DBD). 9-cis-retinoic acid in RARa and RXRa LBDs is represented by yellow and red spheres. The gray spheres in the DBD indicate Zn2þ ions. Residues mediating the weak DBD intersubunit contacts are displayed in yellow. The cysteine residues coordinating the Zn2þ ions are shown in violet.
All NR proteins exhibit a characteristic modular structure that consists of five to six domains of homology (designated A–F, from the N-terminal to the C-terminal) based on regions of conserved sequence and function (Fig. 2.1). The DNA-binding domain (DBD; region C) is the most highly conserved domain and encodes two zinc finger modules. The ligand-binding domain (LBD; region E) is less conserved and mediates ligand binding, dimerization, and a ligand-dependent transactivation function, termed AF-2. The N-terminal A–B region contains a cell- and promoter-specific transactivation function termed AF-1. The region D is considered as a hinge domain. The F region is not present in all receptors and its function is poorly understood. Except for the dosage-sensitive sex reversal, adrenal hypoplasia congenital critical region on the X chromosome gene 1 (DAX-1 (NR0B1)), and the short heterodimer partner (SHP (NR0B2)), which both lack an essential DBD, NRs can bind their cognate sequence-specific promoter elements on target genes either as monomers, homodimers, or heterodimers with retinoid X receptors (RXRa (NR2B1), RXRb (NR2B2), and RXRg (NR2B3)). They all recognize the same hexameric DNA core motif, 50 -PuGGTCA (Pu ¼ A or G), but mutation, extension, duplication, and distinct relative orientations of repeats of this motif have generated response elements that are selective for a given class of receptors. For example, the RXR– RAR heterodimers bind to retinoic acid response elements that mainly correspond to direct repeats of the motif 50 -AGGTCA separated by five, two, or one nucleotides (referred to as DR5, DR2, and DR1; Fig. 2.1), whereas steroid receptors such as the
23
24
CHAPTER 2 Dimerization of Nuclear Receptors
androgen receptor (AR, (NR3C4)) bind essentially as homodimers to response elements containing palindromic repeats of the hexamerix half-site sequence 50 -AGAACA-30 (Khorasanizadeh & Rastinejad, 2001). Dimerization is a general mechanism to increase binding site affinity, specificity, and diversity due to (i) cooperative DNA binding, (ii) the lower frequency of two hexamer-binding motifs separated by a defined spacer compared to that of single hexamers, and (iii) heterodimers that may have recognition sites distinct from those of homodimers. In this regard, RXRs play a central role in various signal transduction pathways since they can both homodimerize and act as promiscuous heterodimerization partner for almost fifteen NRs. Crystal structures of DBD and LBD homo- and heterodimers have defined the surfaces involved in dimerization. It is important to point out that the response element repertoire for receptor homo- and heterodimers is dictated by the DBD while LBDs stabilize the dimers, but do not contribute to response element selection. Two types of dimerization functions mediate homoand heterodimerization. One involves several surface residues in the DBD that establish weak response element-specific interfaces with corresponding surfaces in the partner DBD (Fig. 2.1). The second is a strong dimerization function in the LBDs of both partners that differs between homo- and heterodimers and to some extent between the partners of RXR. We present here different procedures, which comprise both cellular and biochemical approaches, to investigate the interactions that occur between the NRs, that is, the formation of homodimers or heterodimers.
2.1 METHODS 2.1.1 Studies of NR–NR interactions through protein crystallization One of the very effective structural methods to study protein–protein interactions is based on the preparation of crystals of the complexes followed by their analysis using X-ray beams. The determination of the high-resolution three-dimensional (3D) structures that derive from this technique provides atomic-level information on the protein–protein interface. The vast majority of NR–NR interactions are mediated by the dimerization surface located in the LBD of receptors. In this regard, several structures of homo- and heterodimers of NRs have been reported, thereby identifying the structural organization of NR dimers. In particular, crystal structures of several NRs in their homodimeric form (Bourguet, Ruff, Chambon, Gronemeyer, & Moras, 1995; Brzozowski et al., 1997; Greschik et al., 2002; Nolte et al., 1998) or containing RXR LBD in complex with various partner LBDs, including RAR (Bourguet, Vivat, et al., 2000; Pogenberg et al., 2005; Sato et al., 2010), PPAR (Gampe et al., 2000), TR (Putcha, Wright, Brunzelle, & Fernandez, 2012), LXR (Svensson et al., 2003), or CAR (Suino et al., 2004; Xu et al., 2004), have been reported. All these structures demonstrated a topologically conserved dimerization surface with identical structural elements generating the interface in homo- and heterodimers. However, amino acid variations at the surface of the various NRs determine their specific dimerization characteristics. An extensive analysis of the dimerization interfaces of NR LBDs,
2.1 Methods
FIGURE 2.2 Structural analysis of NR dimers. (A) The dimeric arrangement of NRs is illustrated by the structure of RARa–RXRa LBD heterodimer (Protein Data Bank code 1dkf) viewed along (left) or perpendicular (right) to the dimer axis. The secondary structural elements involved in the dimer interface are displayed in orange and labeled. The ligands in both subunits are represented as yellow sticks. (B) Some important intersubunit interactions of the RARa–RXRa LBD heterodimer are shown. For clarity, not all the contacts are displayed. (C) The dimeric arrangement of the GR LBD homodimer as seen in the crystal (Protein Data Bank code 1m2z). The secondary structural elements involved in the canonical dimerization surface used by other NRs are labeled together with the C-terminal b-strand that masks a portion of this surface. The structural elements that are involved in the GR–GR interaction are indicated.
based on crystal structure and sequence alignment, has been reported (Bourguet, Vivat, et al., 2000). The reader is referred to this publication for details and an interpretation of the structural basis that accounts for the homo- and heterodimerization pattern of distinct members of this family. Briefly, the structures show that the dimeric arrangements are closely related, with residues from helices H7, H9, and H10 and loops L8–9 and L9–10 of each protomer forming an interface comprising a network of complementary hydrophobic and charged residues and further stabilized by neutralized basic and acidic surfaces (Fig. 2.2A and 2.2B). Interestingly, the recently reported crystal structures of the entire PPARg–RXRa heterodimer and HNF-4a homodimer bound
25
26
CHAPTER 2 Dimerization of Nuclear Receptors
to their DNA response elements fully validated the studies on the isolated LBDs (Chandra et al., 2008, 2013). Indeed, both the general dimeric organization and the details of the inter-LBD interactions observed in the LBD dimers are strictly conserved in the context of the full-length receptors. A notable exception to this conserved dimeric arrangement is the case of 3-keto-steroid receptors (namely, the androgene AR, progesterone (PR (NR3C3)), glucocorticoid (GR (NR3C1)), and mineralocorticoid (MR (NR3C2)) receptors) whose dimerization surface is partly masked by a C-terminal extension folding as a b-strand and forming a b-sheet interaction with residues located between helices H8 and H9 (Fig. 2.2C). The 3D structure of the GR LBD homodimer (Bledsoe et al., 2002) suggested an alternative mode of dimerization involving residues from a b-turn located between H5 and H6 and the extended loop between H1 and H3, as well as the last residue of H5 (Fig. 2.2C). However, formation of ˚ 2 of solvent-accessible surface as compared with the GR homodimer buries only 600 A 2 ˚ the 1000–1700 A of buried surfaces observed in the other NRs so that the biological relevance of this mode of dimerization remains to be confirmed. Considering that the methods used for the crystallization of protein complexes are diverse and that crystallization conditions are unique to each protein (or protein complex), we provide here a very general procedure that may apply to all kinds of NR LBDs in complex or in isolation. The following protocol is an example given for the crystallization of the RARa–RXRa LBD heterodimers bound to BMS614, a RARa-selective antagonist.
2.1.1.1 Required materials – – – – – – –
Crystallization robot (X8 nanosystem, Cartesian; Freedom EVO, Tecan) Crystallization screens (Molecular Dimension, Hampton Research, etc.) 96-Well crystallization plates (Greiner Bio-One) 24-Well crystallization plates (Greiner Bio-One, Molecular Dimension) Siliconized glass coverslips (Greiner Bio-One) Stereo microscope (Leica) X-ray generator (Rigaku HF 007) with a MAResearch image plate
2.1.1.2 Protocol Prior to crystallization trials, the purified heterodimer (see the succeeding text for the description of a general purification method of NR heterodimers) is mixed with threefold molar excess of the ligand, concentrated to 5–10 mg/mL, and centrifuged for 30 min at 13,000 rpm to get rid of any aggregated material. An initial search for crystallization conditions is undertaken by mixing the purified protein with sparse matrix screening conditions using crystallization robots and dedicated 96-well crystallization plates. This setup allows for the rapid screening of hundreds of crystallization conditions commercially available in a 96-well format with a reduced amount of protein. The robot automatically performs each crystallization trial by mixing 0.1 mL of protein with 0.1 mL of precipitant condition in a well suspended over a reservoir containing 100 mL of the corresponding condition. Once the 96 conditions have been dispensed, the plates are sealed with a transparent
2.1 Methods
plastic film and incubated at a fixed temperature, usually around 20 C. The drops are observed on a daily basis with a stereo microscope to detect any crystal growth. Numerous crystallization kits in a 96-well format are provided by several companies. They correspond to a collection of diverse kinds of buffers, precipitants, and additives designed to cover a wide range of conditions. The number of screens to be used depends on protein availability. The crystals obtained at this step are generally small (
Dimerization of Nuclear Receptors
2
Pierre Germain*,{ and William Bourguet*,{ *
Inserm U1054, Centre de Biochimie Structurale, Montpellier, France CNRS UMR5048, Universite´s Montpellier 1 & 2, Montpellier, France
{
CHAPTER OUTLINE Introduction .............................................................................................................. 22 2.1 Methods ............................................................................................................ 24 2.1.1 Studies of NR–NR Interactions through Protein Crystallization ............. 24 2.1.1.1 Required Materials .....................................................................26 2.1.1.2 Protocol......................................................................................26 2.1.2 Expression and Purification of NR–NR Complexes ............................... 27 2.1.2.1 Required Materials .....................................................................28 2.1.2.2 Protocol......................................................................................29 2.1.3 Monitoring NR–NR Interactions by Noncovalent Electrospray Ionization Mass Spectrometry ........................................................................... 29 2.1.3.1 Required Materials .....................................................................30 2.1.3.2 Protocol......................................................................................30 2.1.4 Monitoring NR–NR Interactions by Electrophoretic Mobility Shift Assays .......................................................................................... 31 2.1.4.1 Required Materials .....................................................................31 2.1.4.2 Protocol......................................................................................31 2.1.5 Two-hybrid Assays to Define NR–NR Interactions in Living Cells........... 32 2.1.5.1 Required Materials .....................................................................34 2.1.5.2 Protocol......................................................................................34 2.1.6 Fluorescence Cross-Correlation Spectroscopy to Measure the Concentrations and Interactions of NRs in Living Cells ........................ 35 2.1.6.1 Required Materials .....................................................................36 2.1.6.2 Protocol......................................................................................36 Acknowledgments ..................................................................................................... 39 References ............................................................................................................... 39
Methods in Cell Biology, Volume 117 Copyright © 2013 Elsevier Inc. All rights reserved.
ISSN 0091-679X http://dx.doi.org/10.1016/B978-0-12-408143-7.00002-5
21
22
CHAPTER 2 Dimerization of Nuclear Receptors
Abstract Multicellular organisms require specific intercellular communication to properly organize the complex body plan during embryogenesis and maintain its properties and functions during the entire life. While growth factors, neurotransmitters, and peptide hormones bind to membrane receptors, thereby inducing the activity of intracellular kinase cascades or the JAK–STAT signaling pathways, other small signaling compounds such as steroid hormones, certain vitamins, and metabolic intermediates enter, or are generated, within the target cells and bind to members of a large family of nuclear receptors (NRs). NRs are ligand-inducible transcription factors that control a plethora of biological phenomena, thus orchestrating complex events like development, organ homeostasis, immune function, and reproduction. NR–NR interactions are of major importance in these regulatory processes, as NRs regulate their target genes by binding to cognate DNA response elements essentially as homo- or heterodimers. A number of structural and functional studies have provided significant insights as to how combinatorial NRs rely on protein–protein contacts that discriminate geometric features of their DNA response elements, thereby allowing both binding site diversity and physiological specificity. Here, we will review our current understanding of NR–NR interactions and provide protocols for a number of experimental approaches that are useful for their study.
INTRODUCTION Nuclear receptors (NRs) are members of a large superfamily of evolutionarily related transcription factors that regulate genetic programs involved in a broad spectrum of physiological phenomena. The human NR family is classified by sequence into six evolutionary groups of unequal size. The reader is referred to several databases providing comprehensive annotation of the literature on the pharmacology of the 48 human NR gene products, together with relevant information on their structure and function (IUPHAR-DB (http://www.iuphar-db.org/), Nuclear Receptor Signaling Atlas (NURSA, http://www.nursa.org/), and NureXbase (http://nurexbase.prabi. fr)). Among NRs, ligands have been identified for only 24 family members. These receptors are ligand-dependent transcriptional factors that respond directly to a large variety of hormonal and metabolic substances that are hydrophobic, lipid-soluble, and of small size (e.g., retinoic acid or estradiol). The other class of NRs is the group of so-called orphan receptors, for which regulatory ligands are still unknown or may not exist (true orphans) or for which candidates have only recently been identified (adopted orphans). In both groups, NRs can exist as different receptor subtypes that originate from different genes such as the retinoic acid receptors a, b, and g (RARa (NR1B1), RARb (NR1B2), and RARg (NR1B3)), which can exist themselves as several isoforms originating from alternative splicing and differential promoter usage (e.g., RARb1–RARb4) (Germain, Staels, Dacquet, Spedding, & Laudet, 2006; Gronemeyer, Gustafsson, & Laudet, 2004).
Introduction
FIGURE 2.1 Structural and functional organization on nuclear receptors. The structural organization of the DBD and LBD is illustrated by the crystal structures of the RXRa–RARa DBD heterodimer bound to a DR1 response element (Protein Data Bank code 1dsz) and the 9-cis-retinoic acid-bound RXRa–RARa LBD heterodimer (Protein Data Bank code 1xdk). Helices are represented as ribbons and labeled from H1 to H12 (LBD) or H1 and H2 (DBD). 9-cis-retinoic acid in RARa and RXRa LBDs is represented by yellow and red spheres. The gray spheres in the DBD indicate Zn2þ ions. Residues mediating the weak DBD intersubunit contacts are displayed in yellow. The cysteine residues coordinating the Zn2þ ions are shown in violet.
All NR proteins exhibit a characteristic modular structure that consists of five to six domains of homology (designated A–F, from the N-terminal to the C-terminal) based on regions of conserved sequence and function (Fig. 2.1). The DNA-binding domain (DBD; region C) is the most highly conserved domain and encodes two zinc finger modules. The ligand-binding domain (LBD; region E) is less conserved and mediates ligand binding, dimerization, and a ligand-dependent transactivation function, termed AF-2. The N-terminal A–B region contains a cell- and promoter-specific transactivation function termed AF-1. The region D is considered as a hinge domain. The F region is not present in all receptors and its function is poorly understood. Except for the dosage-sensitive sex reversal, adrenal hypoplasia congenital critical region on the X chromosome gene 1 (DAX-1 (NR0B1)), and the short heterodimer partner (SHP (NR0B2)), which both lack an essential DBD, NRs can bind their cognate sequence-specific promoter elements on target genes either as monomers, homodimers, or heterodimers with retinoid X receptors (RXRa (NR2B1), RXRb (NR2B2), and RXRg (NR2B3)). They all recognize the same hexameric DNA core motif, 50 -PuGGTCA (Pu ¼ A or G), but mutation, extension, duplication, and distinct relative orientations of repeats of this motif have generated response elements that are selective for a given class of receptors. For example, the RXR– RAR heterodimers bind to retinoic acid response elements that mainly correspond to direct repeats of the motif 50 -AGGTCA separated by five, two, or one nucleotides (referred to as DR5, DR2, and DR1; Fig. 2.1), whereas steroid receptors such as the
23
24
CHAPTER 2 Dimerization of Nuclear Receptors
androgen receptor (AR, (NR3C4)) bind essentially as homodimers to response elements containing palindromic repeats of the hexamerix half-site sequence 50 -AGAACA-30 (Khorasanizadeh & Rastinejad, 2001). Dimerization is a general mechanism to increase binding site affinity, specificity, and diversity due to (i) cooperative DNA binding, (ii) the lower frequency of two hexamer-binding motifs separated by a defined spacer compared to that of single hexamers, and (iii) heterodimers that may have recognition sites distinct from those of homodimers. In this regard, RXRs play a central role in various signal transduction pathways since they can both homodimerize and act as promiscuous heterodimerization partner for almost fifteen NRs. Crystal structures of DBD and LBD homo- and heterodimers have defined the surfaces involved in dimerization. It is important to point out that the response element repertoire for receptor homo- and heterodimers is dictated by the DBD while LBDs stabilize the dimers, but do not contribute to response element selection. Two types of dimerization functions mediate homoand heterodimerization. One involves several surface residues in the DBD that establish weak response element-specific interfaces with corresponding surfaces in the partner DBD (Fig. 2.1). The second is a strong dimerization function in the LBDs of both partners that differs between homo- and heterodimers and to some extent between the partners of RXR. We present here different procedures, which comprise both cellular and biochemical approaches, to investigate the interactions that occur between the NRs, that is, the formation of homodimers or heterodimers.
2.1 METHODS 2.1.1 Studies of NR–NR interactions through protein crystallization One of the very effective structural methods to study protein–protein interactions is based on the preparation of crystals of the complexes followed by their analysis using X-ray beams. The determination of the high-resolution three-dimensional (3D) structures that derive from this technique provides atomic-level information on the protein–protein interface. The vast majority of NR–NR interactions are mediated by the dimerization surface located in the LBD of receptors. In this regard, several structures of homo- and heterodimers of NRs have been reported, thereby identifying the structural organization of NR dimers. In particular, crystal structures of several NRs in their homodimeric form (Bourguet, Ruff, Chambon, Gronemeyer, & Moras, 1995; Brzozowski et al., 1997; Greschik et al., 2002; Nolte et al., 1998) or containing RXR LBD in complex with various partner LBDs, including RAR (Bourguet, Vivat, et al., 2000; Pogenberg et al., 2005; Sato et al., 2010), PPAR (Gampe et al., 2000), TR (Putcha, Wright, Brunzelle, & Fernandez, 2012), LXR (Svensson et al., 2003), or CAR (Suino et al., 2004; Xu et al., 2004), have been reported. All these structures demonstrated a topologically conserved dimerization surface with identical structural elements generating the interface in homo- and heterodimers. However, amino acid variations at the surface of the various NRs determine their specific dimerization characteristics. An extensive analysis of the dimerization interfaces of NR LBDs,
2.1 Methods
FIGURE 2.2 Structural analysis of NR dimers. (A) The dimeric arrangement of NRs is illustrated by the structure of RARa–RXRa LBD heterodimer (Protein Data Bank code 1dkf) viewed along (left) or perpendicular (right) to the dimer axis. The secondary structural elements involved in the dimer interface are displayed in orange and labeled. The ligands in both subunits are represented as yellow sticks. (B) Some important intersubunit interactions of the RARa–RXRa LBD heterodimer are shown. For clarity, not all the contacts are displayed. (C) The dimeric arrangement of the GR LBD homodimer as seen in the crystal (Protein Data Bank code 1m2z). The secondary structural elements involved in the canonical dimerization surface used by other NRs are labeled together with the C-terminal b-strand that masks a portion of this surface. The structural elements that are involved in the GR–GR interaction are indicated.
based on crystal structure and sequence alignment, has been reported (Bourguet, Vivat, et al., 2000). The reader is referred to this publication for details and an interpretation of the structural basis that accounts for the homo- and heterodimerization pattern of distinct members of this family. Briefly, the structures show that the dimeric arrangements are closely related, with residues from helices H7, H9, and H10 and loops L8–9 and L9–10 of each protomer forming an interface comprising a network of complementary hydrophobic and charged residues and further stabilized by neutralized basic and acidic surfaces (Fig. 2.2A and 2.2B). Interestingly, the recently reported crystal structures of the entire PPARg–RXRa heterodimer and HNF-4a homodimer bound
25
26
CHAPTER 2 Dimerization of Nuclear Receptors
to their DNA response elements fully validated the studies on the isolated LBDs (Chandra et al., 2008, 2013). Indeed, both the general dimeric organization and the details of the inter-LBD interactions observed in the LBD dimers are strictly conserved in the context of the full-length receptors. A notable exception to this conserved dimeric arrangement is the case of 3-keto-steroid receptors (namely, the androgene AR, progesterone (PR (NR3C3)), glucocorticoid (GR (NR3C1)), and mineralocorticoid (MR (NR3C2)) receptors) whose dimerization surface is partly masked by a C-terminal extension folding as a b-strand and forming a b-sheet interaction with residues located between helices H8 and H9 (Fig. 2.2C). The 3D structure of the GR LBD homodimer (Bledsoe et al., 2002) suggested an alternative mode of dimerization involving residues from a b-turn located between H5 and H6 and the extended loop between H1 and H3, as well as the last residue of H5 (Fig. 2.2C). However, formation of ˚ 2 of solvent-accessible surface as compared with the GR homodimer buries only 600 A 2 ˚ the 1000–1700 A of buried surfaces observed in the other NRs so that the biological relevance of this mode of dimerization remains to be confirmed. Considering that the methods used for the crystallization of protein complexes are diverse and that crystallization conditions are unique to each protein (or protein complex), we provide here a very general procedure that may apply to all kinds of NR LBDs in complex or in isolation. The following protocol is an example given for the crystallization of the RARa–RXRa LBD heterodimers bound to BMS614, a RARa-selective antagonist.
2.1.1.1 Required materials – – – – – – –
Crystallization robot (X8 nanosystem, Cartesian; Freedom EVO, Tecan) Crystallization screens (Molecular Dimension, Hampton Research, etc.) 96-Well crystallization plates (Greiner Bio-One) 24-Well crystallization plates (Greiner Bio-One, Molecular Dimension) Siliconized glass coverslips (Greiner Bio-One) Stereo microscope (Leica) X-ray generator (Rigaku HF 007) with a MAResearch image plate
2.1.1.2 Protocol Prior to crystallization trials, the purified heterodimer (see the succeeding text for the description of a general purification method of NR heterodimers) is mixed with threefold molar excess of the ligand, concentrated to 5–10 mg/mL, and centrifuged for 30 min at 13,000 rpm to get rid of any aggregated material. An initial search for crystallization conditions is undertaken by mixing the purified protein with sparse matrix screening conditions using crystallization robots and dedicated 96-well crystallization plates. This setup allows for the rapid screening of hundreds of crystallization conditions commercially available in a 96-well format with a reduced amount of protein. The robot automatically performs each crystallization trial by mixing 0.1 mL of protein with 0.1 mL of precipitant condition in a well suspended over a reservoir containing 100 mL of the corresponding condition. Once the 96 conditions have been dispensed, the plates are sealed with a transparent
2.1 Methods
plastic film and incubated at a fixed temperature, usually around 20 C. The drops are observed on a daily basis with a stereo microscope to detect any crystal growth. Numerous crystallization kits in a 96-well format are provided by several companies. They correspond to a collection of diverse kinds of buffers, precipitants, and additives designed to cover a wide range of conditions. The number of screens to be used depends on protein availability. The crystals obtained at this step are generally small (
