Molecular Recognition of Agonist Ligands by RXRs (original) (raw)

Journal Article

,

1Laboratoire de Biologie et Génomique Structurales, Institut de Génétique et Biologie Moléculaire et Cellulaire, Centre National de la Recherche Scientifique/Institut National de la Santé et de la Recherche Médicale/Université Louis Pasteur, Parc d’Innovation BP163, 67404 Illkirch cedex, France

Search for other works by this author on:

,

1Laboratoire de Biologie et Génomique Structurales, Institut de Génétique et Biologie Moléculaire et Cellulaire, Centre National de la Recherche Scientifique/Institut National de la Santé et de la Recherche Médicale/Université Louis Pasteur, Parc d’Innovation BP163, 67404 Illkirch cedex, France

Search for other works by this author on:

1Laboratoire de Biologie et Génomique Structurales, Institut de Génétique et Biologie Moléculaire et Cellulaire, Centre National de la Recherche Scientifique/Institut National de la Santé et de la Recherche Médicale/Université Louis Pasteur, Parc d’Innovation BP163, 67404 Illkirch cedex, France

*Address all correspondence and requests for reprints to: Dino Moras, Laboratoire de Biologie et Genomique Structurales, Institut de Genetique et Biologie Moleculaire et Cellulaire, Centre National de la Recherche Scientifique/Institut National de la Santé et de la Recherche Médicale/Université Louis Pasteur, Parc d’Innovation BP163, 1 rue Laurent Fries, 67404 Illkirch cedex, France.

Search for other works by this author on:

Accepted:

21 December 2001

• PDF
• Split View
• • Article contents
  • Figures & tables
  • Video
  • Audio
  • Supplementary Data

Navbar Search Filter Mobile Enter search term Search

Abstract

The nuclear receptor RXR is an obligate partner in many signal transduction pathways. We report the high-resolution structures of two complexes of the human RXRα ligand-binding domain specifically bound to two different and chemically unrelated agonist compounds: docosa hexaenoic acid, a natural derivative of eicosanoic acid, present in mammalian cells and recently identified as a potential endogenous RXR ligand in the mouse brain, and the synthetic ligand BMS 649. In both structures the RXR-ligand-binding domain forms homodimers and exhibits the active conformation previously observed with 9-**cis-RA. Analysis of the differences in ligand-protein contacts (predominantly van der Waals forces) and binding cavity geometries and volumes for the several agonist-bound RXR structures clarifies the structural features important for ligand recognition. The L-shaped ligand-binding pocket adapts to the diverse ligands, especially at the level of residue N306, which might thus constitute a new target for drug-design. Despite its highest affinity 9-cis-**RA displays the lowest number of ligand-protein contacts. These structural results support the idea that docosa hexaenoic acid and related fatty acids could be natural agonists of RXRs and question the real nature of the endogenous ligand(s) in mammalian cells.

NUCLEAR RECEPTORS (NRs) are important transcriptional regulators involved in widely diverse physiological functions such as embryonic development, cell growth, differentiation and apoptosis, and homeostasis (1). Many of these receptors are regulated by the binding of small hydrophobic ligands such as steroids, retinoids, or thyronines. Some receptors called “orphan receptors” have no (yet) known ligands. In addition, these proteins are important therapeutic targets because a large number of them are implicated in diseases such as cancer, diabetes, metabolic diseases, or hormone resistance syndromes. All NRs share a common functional and structural organization with distinct modules. The E/F region of about 250 amino acids contains the moderately conserved ligand-binding domain (LBD) that carries the ligand-dependent activation function AF2 and is also responsible for the dimerization of the receptor and its interaction with transcriptional coactivators or corepressors (2). Ligand binding triggers major structural changes that affect the position of the C-terminal helix H12 carrying the AF2 activation function (3, 4). Upon agonist binding, the LBD adopts a unique active conformation, the activated receptor being competent for transcriptional coactivator binding.

Retinoids are vitamin A-derived metabolites that play important roles in development, cell growth, differentiation and death, and homeostasis (5). RA pleiotropic effects are transduced by two distinct classes of specific NRs: RARs and RXRs (6). Two natural isomers of RA are high-affinity ligands for these receptors: RXRs bind exclusively the 9-_cis-_isomer (7, 8) whereas RARs interact with both 9-cis and all-_trans-_isomers (9, 10). The crystal structures of RXR LBD bound to 9-_cis-_RA (11, 12) illustrated and explained the specificity of this ligand for RXR. Among NRs RXR occupies a central position as an ubiquitous heterodimerization partner with several other NRs such as VDR, PPAR, TR, RAR, and many orphan receptors (13). Although RXRs can be active as homodimers, the RXR heterodimers are the physiologically relevant molecular species (5, 6). Thus, RXR is a key receptor being an obligate partner in many signaling pathways (14).

Three types of RXR heterodimers have been described (13). In some heterodimers (e.g. RXR/TR or RXR/VDR) RXR might be a completely silent partner. In others (e.g. RXR/RAR), RXR seems to be a conditionally silent partner. In the last category (e.g. many RXR/orphan receptors) RXR appears to be a fully active and competent partner. Another property of RXR makes it a special member of the NR superfamily: its propensity to form autorepressed homotetramers in the absence of ligand (15–17).

The sequence homology across the NR superfamily indicates that its members evolved through divergent evolution from a common orphan ancestor. Phylogenetic analysis of NR LBD sequences led to the partition of the superfamily into six subfamilies with several subgroups (18). Interestingly, RARs and RXRs do not belong to the same subfamily. In fact, they appear to be phylogenetically quite distant (32% identity and 51% similarity). RXRs belong to the subfamily that also includes TRs and are also close to ERs that belong to another subfamily (31% identity and 48% similarity). In the RXR subfamily, ultraspiracle protein (USP) is the most closely related to RXR (51% identity and 73% similarity) and appears to be the ortholog of RXR in Arthropods (19). RXR binds 9-_cis-_RA from Cnidarians to Vertebrates except in Arthropods where its ortholog is unable to do so. The natural ligand for USP is still unknown, although it was proposed that juvenile hormone, a natural isoprenic compound, could be a candidate (20). The two crystal structures of USP that have been recently reported (21, 22) showed the presence in the binding pocket of an unexpected phospholipidic compound trapped during protein extraction and purification. It is proposed that, during evolution, 9-_cis-_RA binding to RXR occurred very early during metazoan evolution, and the secondary loss was specific to Arthropods for the USP (23).

However, 9-_cis-_RA has been very difficult to detect in vivo. Vitamin A (retinol) is acquired via the diet and can be converted into retinal in cells. Retinal is a key metabolite required for the visual process, but also serves as the precursor for all-_trans-_RA biosynthesis. An enzymatic isomerization of the all-_trans-_isomer yields its 9-_cis-_isomer. The regulation of this key enzymatic step controls the relative 9-cis/all-trans ratio within the cell and therefore regulates the RXR and RAR pathways (24, 25). However, this conversion step remains elusive. The crystal structure of a RXR/RAR LBD mutant heterodimer showed that RXR was also able to bind a fatty acid molecule (26). More recently, docosa hexaenoic acid (DHA or C22:6 all-_cis-_Δ,7,10, 13,16,19), a natural derivative of eicosanoids, has been identified as a new potential ligand for this receptor and was shown to activate RXR in mouse brain cells (27). DHA is abundant in mammalian brain cells, where it can constitute up to 30–50% of the total fatty acids under a predominantly membrane phospholipid-associated form. Interestingly, DHA can be released from phospholipids through PLA2 and PLC-dependent mechanisms. During late gestation and early postnatal life, DHA becomes highly enriched in the mammalian brain and is required for brain maturation in rodents and humans. DHA accumulates in the retina, a tissue whose development is affected in RXRα-deficient mice (5). DHA deficiency in rats and humans results in abnormalities similar to those observed in RXRγ knock-out mice.

Here we report the crystal structures of human RXR LBD bound to DHA and BMS 649, a synthetic pharmaceutical RXR-selective ligand close to Targretin (LGD1069 from Ligand Pharmaceuticals, Inc., San Diego, CA) and efficiently used in association with the antiestrogen Tamoxifen to treat mammary carcinoma (28).

RESULTS

Structure Determination and Overall Protein Organization

The hRXRα LBD was cocrystallized with the two agonist ligands DHA or BMS 649 in the presence of a synthetic peptide containing a coactivator consensus sequence LXXLL (Materials and Methods). The limit of diffraction of these crystals is 1.5 Å but, due to technical limitations, data were collected to 1.9 Å and the structures were solved by molecular replacement using entire and truncated RXR structures as search probes. The data and refinement statistics are summarized in Table 1. The experimental maps showed clear electron densities that could be unambiguously assigned to the ligands.

Table 1.

Data Collection and Statistics of Crystallographic Data and Refinement

r.m.s.d is the root-mean square deviation from ideal geometry.

a

Rsym = Σhkl Σi |I hkl,i − 〈I hk,i〉|/Σhkl Σi |I hkl,i| where 〈I hk,i〉 is the average intensity of the multiple hkl, i observations for symmetry-related reflections.

b

Rcryst/free = Σ|_F_obs − _F_calc|/Σ|_F_obs|. _F_obs and _F_calc are observed and calculated structure factors, Rfree is calculated from a randomly chosen 7.5% of reflections (2ς), and Rcryst is calculated for the remaining 92.5% of reflections (2ς).

c

In the triclinic form the two homodimers are constituted with monomers 1 and 2, and 3 and 4, respectively.

Table 1.

Data Collection and Statistics of Crystallographic Data and Refinement

r.m.s.d is the root-mean square deviation from ideal geometry.

a

Rsym = Σhkl Σi |I hkl,i − 〈I hk,i〉|/Σhkl Σi |I hkl,i| where 〈I hk,i〉 is the average intensity of the multiple hkl, i observations for symmetry-related reflections.

b

Rcryst/free = Σ|_F_obs − _F_calc|/Σ|_F_obs|. _F_obs and _F_calc are observed and calculated structure factors, Rfree is calculated from a randomly chosen 7.5% of reflections (2ς), and Rcryst is calculated for the remaining 92.5% of reflections (2ς).

c

In the triclinic form the two homodimers are constituted with monomers 1 and 2, and 3 and 4, respectively.

In all structures, the RXR LBD adopts the canonical conformation of all previously reported agonist-bound NR-LBDs (29), the _trans_-activation C-terminal H12 helix being packed toward the protein core in a hydrophobic groove constituted by helices H3, H5, and H11 (Fig. 1, A and B). To stabilize the active conformation, crystallization trials were carried out in the presence of a synthetic peptide containing the consensus sequence LXXLL, which mimics the NR-interacting motif (30–32). In the structure, the coactivator peptide is tightly bound to the LBD AF-2 surface formed by helices H3, H4, and H12 of the LBD. The polar and hydrophobic interactions between the receptor and the coactivator peptide are similar to those previously described in other LBD-coactivator peptide complex structures (30–32).

Figure 1.

The Structures of the RXR LBD Bound to DHA or BMS 649 Compounds and a Coactivator Peptide The RXR-DHA (A) and RXR-BMS 649 (B) complex structures and a stereoview of the homodimeric association mode of agonist bound holo-RXR LBD (C) are shown. The crystallographic symmetry 2-fold axis is indicated with a line. Helices H10, H11, and H12 are colored in yellow, pink, and red, respectively. The coactivator helical peptide is drawn in blue. The ligand molecules of DHA and BMS 649 are displayed in yellow and red for their carbon and oxygen atoms, respectively.

The Homodimeric Association Mode of Agonist-Bound RXR LBD and the Dimerization Interface

In contrast to the situation of holo-RXR LBD bound to 9-_cis-_RA in the absence of coactivator peptide where a monomeric form of the complex was observed (11), all the present structures show a homodimeric association mode for agonist-bound RXR LBD in the presence of the coactivator peptide. Although in the tetragonal crystals of both RXR-DHA and RXR-BMS 649 complexes, the asymmetric unit contains a single monomer, the previously described biological homodimer is rebuilt through the space-group symmetry operators. Comparison with the previously reported homodimeric structures of _apo_-RXR (33) or heterodimeric structures of RXR/RAR (26) or RXR/PPAR (12) shows that the overall dimeric arrangement is identical (Fig. 1C). The present results based on these crystallographic structures are in agreement with extensive biophysical studies carried out on RXR, RAR, and VDR LBDs in solution by small-angle x-ray and neutron scattering and analytical ultracentrifugation that unambiguously demonstrated that holo-RXR LBD was a dimer in solution regardless of protein concentration (17, 34). In the case of the RXR-BMS 649 complex, a triclinic crystal form was also obtained in the same crystallization conditions. In this specific case the asymmetric unit contains two LBD homodimers weakly associated into a tetramer through the H1 helices. This crystal packing-induced tetramer is not reminiscent of that previously reported for the _apo_-RXR LBD (16, 17).

Using the same crystallization conditions, notably the presence of coactivator peptide, we also were able to grow isomorphous tetragonal crystals of RXR LBD in the presence of 9-_cis-_RA. Furthermore, in the absence of peptide we also grew crystals of the RXR-BMS 649 complex isomorphous to those described for the RXR-9-_cis-_RA (35). A close inspection of the packing contacts in the various crystal forms provides a likely explanation for the oligomeric state fluctuations. In the absence of the peptide (monomeric case), the hydrophobic groove formed by helices H3, H4, and H12 that constitutes the target of the LXXLL motif is filled with hydrophobic peptides of related sequence of neighboring molecules (11). The saturation of the hydrophobic groove is required to stabilize the crystal packing (36, 37) but cannot be obtained without disrupting the homodimeric units. Although in solution the monomer-dimer equilibrium is strongly in favor of the second entity, the crystal that competes for the monomers will find enough material if the packing energy is close to that provided by the dimeric association. The comparison of the surface buried upon association provides a reasonable estimate of the energy involved. In the present homodimers the dimerization interface is about 1,100 ± 20 Å2. The bound coactivator peptide releases the constraint by filling the hydrophobic groove and allowing different packings.

The Ligand-Binding Pockets (LBPs)

DHA and BMS 649 molecules are buried in the predominantly hydrophobic pocket that was identified in RXR-9-_cis-_RA complexes (11, 12). Despite the overall similarity of protein-ligand interactions, some differences can be noted. Additional residues, such as V265 on H3, I310 on H5, I324 on the β-strand, and V346 on H7, interact with either DHA or BMS 649 whereas they are not in contact with 9-_cis-_RA. In contrast with the case of the RXR-9-_cis-_RA holostructure, all atoms of both DHA and BMS 649 ligands are in direct van der Waals contact with the protein. Thus, the majority of protein-ligand contacts are van der Waals interactions (Fig. 2A, B, and C). The elongated and L-shaped binding pocket is sealed by R316 of helix H5 on one side and transactivation helix H12 on the other side. Both DHA and BMS 649 contain a carboxylate group that is involved in an ionic interaction with the strictly conserved basic residue R316 of helix H5 and forms a hydrogen bond with the backbone carbonyl amide group of the β-turn residue A327. These two anchoring interactions are observed in all the holostructures of RXRs or RARs that have been solved. The carboxylate moiety also participates in a water-mediated hydrogen bond network involving the backbone carbonyl group of L309 and the side chains of Q275 and R371. As in the case of the RXR-9-_cis-_RA holostructure, no interaction is established between ligands and the transactivation helix H12.

Figure 2.

The LBPs of RXR/DHA and RXR/BMS 649 Complexes A–C, Schematic drawing showing the interactions between the protein and the ligand molecules: 9-_cis-_RA (A) (11 ), DHA (B), and BMS 649 (C). Only contacts closer than 4.2 Å are indicated with dotted lines. Residues in close contact (< 3.7 Å) are underlined. The DHA (D) and BMS 649 (E) molecules are shown in their Fo-Fc electron density omit maps contoured at 3.0 and 2.5 ς, respectively. Water molecules are indicated as red spheres. Direct or water-mediated hydrogen bonds are depicted as magenta dotted lines. Only residues closer than 4.2 Å are shown.

In both structures (Fig. 2, D and E) the ligand molecules adopt a well defined conformation as shown by the clearly defined electron density and the average B factors values, respectively, of 37 Å2 and 14 Å2 observed for ligand atoms of DHA and BMS 649 (Table 1). The respective conformational rigidity of the ligands accounts for the difference. Indeed, the DHA molecule is much more flexible than the rigid and constrained BMS 649 compound. The 9-_cis-_RA molecule represents an intermediate case in terms of rigidity and conformational flexibility. In particular, all three ligands interact extensively with a set of residues delineating the binding cavity. This set includes residues A271,Q275 (both on H3), L326 (on the β-strand), and F313 (on H5) at the entrance of the LBP and residues I268 (on H3) and C432 (on H11) at the level of the hinge and at the bottom of the L-shaped binding cavity.

The solvent-accessible surfaces of the ligand binding cavities and their occupancy reveal specific features of the ligand recognition by RXR (Fig. 3). As shown in Table 2, compared with 9-_cis-_RA, BMS 649 and DHA are larger molecules with molecular volumes of 364, 413, and 427 Å3, respectively; nevertheless, the solvent-accessible volume of the binding cavity in the BMS 649 complex is about 476 Å3 and is thus significantly lower than the value of 528 Å3 calculated in the DHA complex (as a comparison, this value is 504 Å3 in the 9-_cis-_RA complex). This 10% amplitude decrease in LBP volume is mainly due to the reorientation of a single residue, the side chain of residue N306 on helix H5 that moves its polar amide group from a surface-exposed position observed in both the 9-_cis-_RA and DHA complexes (Fig. 2C) to a buried position inside the pocket in the BMS 649 complex (Fig. 2D).

Figure 3.

Comparison of Solvent-Accessible Surfaces in the LBPs of Several RXR/Ligand Complexes Solvent-accessible surfaces computed with the MSMS program are displayed with DINO (59 ) as gray transparent envelopes. The ligands 9-_cis-_RA (A) (11 ), DHA (B), BMS 649 (C), and oleic acid (D) (26 ) are shown by transparency. The arrows indicate the hinge region of the L-shaped binding pocket.

Table 2.

Comparison of the Binding Pocket Properties in RXR-Agonist Ligands Complexes

Ligand volumes were calculated using the MSMS program. Cavity volumes were computed with the VOIDOO program (58 ). Occupancy is defined as the ratio between the ligand volume and the cavity volume. All protein ligand contacts including polar contacts (hydrogen bonds) and van der Waals (vdw) contacts (with a distance cutoff of 4.2 Å) were evaluated using the CCP4 suite of programs.

a and b, Two RXR/9-_cis_-RA complex crystal structures are available, the crystal structure of RXR bound to 9-_cis_-RA (11 ) and the crystal structure of a RXR/PPAR heterodimer with 9-_cis_-RA bound in the RXR subunit (12 ).

c and d, The RXR/DHA and RXR/BMS 649 complexes are described in this study.

e The RAR/9-_cis_-RA structure has been described in the previously reported structure of RAR bound to 9-_cis_-RA (49 ).

The RXR/oleat structure has been described in the previously reported structure of a RXR/RAR heterodimer (26 ). In this latter structure, RXR adopts an antagonist conformation.

Table 2.

Comparison of the Binding Pocket Properties in RXR-Agonist Ligands Complexes

Ligand volumes were calculated using the MSMS program. Cavity volumes were computed with the VOIDOO program (58 ). Occupancy is defined as the ratio between the ligand volume and the cavity volume. All protein ligand contacts including polar contacts (hydrogen bonds) and van der Waals (vdw) contacts (with a distance cutoff of 4.2 Å) were evaluated using the CCP4 suite of programs.

a and b, Two RXR/9-_cis_-RA complex crystal structures are available, the crystal structure of RXR bound to 9-_cis_-RA (11 ) and the crystal structure of a RXR/PPAR heterodimer with 9-_cis_-RA bound in the RXR subunit (12 ).

c and d, The RXR/DHA and RXR/BMS 649 complexes are described in this study.

e The RAR/9-_cis_-RA structure has been described in the previously reported structure of RAR bound to 9-_cis_-RA (49 ).

The RXR/oleat structure has been described in the previously reported structure of a RXR/RAR heterodimer (26 ). In this latter structure, RXR adopts an antagonist conformation.

The calculated values show a clear correlation between the protein-ligand contacts and the occupancy ratio of the LBPs (Table 2). A similar number of 95 and 97 protein-ligand contacts are observed in both DHA and BMS 649 complexes, respectively, compared with the significantly lower number of 77 contacts observed for the 9-_cis-_RA complex. Concomitantly, we observe a remarkable increase in the LBP occupancy ratios by the different agonists that ranges from 72% for 9-_cis-_RA to 80% and 88% for DHA and BMS 649, respectively.

The adaptability of RXR ligands is revealed by the superimposition of the corresponding LBPs that shows the conserved structural features of ligand recognition by RXR (Fig. 4). The position of the carboxylic group is remarkably stable within the three-agonist structures. The hydrophobic channel at the H3/H5 end of the cavity is particularly tight. With the BMS 649 molecule, the benzoate group optimally occupies this part of the binding cavity. The β-ionone ring binding subpocket (delineated by helices H3, H7, and H11, as defined in the 9-_cis-_RA complex), is an exclusively hydrophobic cavity that is fully occupied by DHA (ligand atoms C14-C22) or by BMS 649 (the tetra hydro tetramethyl naptho group). At the level of the hinge of the L-shaped binding pocket, the situation is different: BMS 649 occupies this corner more efficiently due to the above mentioned motion of the side chain of residue N306. The solvent-accessible volumes of the LBPs reveal empty cavities around residues W305/N306 (on helix H5) and residues I268/F313 (on helices H3 and H5,respectively). These cavities were observed when we first reported the RXR-9-_cis-_RA complex structure (11), and these two new agonist complex structures highlight the structural determinants of the molecular specificity of RXRs toward agonist ligands. This hinge region, where these two cavities are located, appears to be the region of maximal flexibility and adaptability within the RXR LBP.

Figure 4.

Comparison of Ligand Positioning in RXR Binding Pocket Reveals a Conserved Set of Molecular Features in RXR-Ligand Recognition The relative orientation and positioning of ligands is presented after superimposition of the protein structures: 9-_cis-_RA (11 ), DHA, BMS 649, and oleic acid (26 ) are depicted in yellow, cyan, green, and red, respectively.

DISCUSSION

Since the initial discovery of 9-_cis-_RA as a high-affinity ligand for RXR, two monocyclic terpenoid compounds, methoprenic acid (38) and phytanic acid (39, 40), have also been identified as RXR ligands (Fig. 5). Unlike 9-_cis-_RA, which also activates RARs, these two noncyclic terpenoids are highly selective for binding and activation of RXRs, albeit at much higher concentrations than 9-_cis-_RA. Methoprenic acid is a metabolite of the pesticide methoprene, a synthetic analog of the juvenile hormone, an insect growth regulator. Phytanic acid is a metabolite of phytol, a chlorophyll derivative obtained in the diet. Several severe diseases, such as Refsum’s disease and Zellweger’s syndrome or neonatal adrenoleukodystrophy, are associated with the inability to catabolize phytanic acid. A simple docking exercise reveals that these molecules can easily adapt to the RXR LBP, but will not fully occupy the cavity. Furthermore, the recent structure of RXR/RAR heterodimer (18) showed the unexpected presence of an oleic acid molecule, a mono unsaturated fatty acid with 18 carbon atoms, in the RXR binding pocket; this fatty acid has been shown to be a partial agonist in in vitro transcriptional assays. Similarly, in the case of the orphan receptor RA-related receptor, as observed for the RXR/RAR heterodimer, the crystal structure of this receptor revealed the presence of a fatty acid molecule identified as stearic acid, a saturated fatty acid with 18 carbon atoms (29). It has been suggested that the ability of RXR to bind fatty acid molecules may underline the potential involvement of RXR in the lipid homeostasis through complex feedback mechanisms in potential association with other NRs (such as PPAR or Farnesoid X receptor).

Figure 5.

Schematic Structures of Identified RXR Ligands, 9-_cis-_RA, Oleic Acid, Docosa Hexaenoic Acid, BMS 649, Methoprene Acid, and Phytanic Acid

A similar observation can be made in the case of PPARs. PPARα and PPARγ bind eicosanoids, polyunsaturated lipids derived from arachidonic acid metabolism (41) such as leukotrienes (e.g. leukotriene B4) (42) or PGs (e.g. 15-deoxy-Δ12,14-PGJ2) (43, 44). In the case of PPARs, various fatty acids including those that vary in both chain length and degree of saturation bind and activate the PPARs at physiological concentrations. Given the established roles of PPARs in lipid and glucid homeostasis, it is suggested that fatty acids or their metabolites are natural PPAR ligands and exert a feedback regulation. Structural analyses of PPARγ and β/δ LBDs reveal that their binding pockets are about three to four times larger than those of other NRs; they are sufficiently large to allow different fatty acids to bind in multiple conformations (45). These findings suggested that PPARs might have evolved to function as lipid sensors and recognize a number of different metabolites rather than a single high-affinity hormone. Therefore, it appears that flexible saturated or unsaturated fatty acids are probably physiologically relevant ligands of many NRs (RXRs, PPARs, and orphan receptors) even if they display lower affinity and activity compared with other compounds. Local high cellular concentrations could compensate their rather low affinity toward RXRs.

The RXR-DHA complex structure shows how the binding pocket is well ordered around the ligand molecule, optimizing protein-ligand contacts and ensuring a specific molecular recognition (Fig. 2D). Indeed, RXR activation by DHA was shown to be sensitive to mutations altering ligand-binding specificity (27, 46, 47). Furthermore, DHA fails to activate other NRs such as TR, VDR, and RAR (27). This is quite remarkable because 9-_cis-_RA binds to both RXRs and RARs but under two subtly different conformations (11). Observation of the RAR LBP shows that it would be impossible for the flexible, but too long, DHA molecule to fit in. As in the case of PPARs, other polyunsaturated fatty acids closely related to DHA such as docosa tetraenoic acid (C22: 4 all-_cis-_Δ7,10,13,16) or arachidonic acid (C20: 4 all-_cis-_Δ5,8,11,14) can activate RXRs but with lower efficiency; on the other hand, other fatty acids such as erucic acid (C20: 1-_cis-_Δ13) fail to activate RXRs (27). The comparison of RXR vs. RAR specificity toward retinoids and eicosanoids illustrates the role of the NR LBP overall geometry in the process of molecular recognition of a specific class of ligands.

The RXR-BMS 649 complex shows that the protein can undergo a small, but significant, conformational fluctuation while maintaining all other contact points of the LBP. The reorientation of one single side-chain residue (N306) is sufficient to reduce the volume of the cavity up to 10% but does not affect the position of any other atom of the pocket. This side chain motion that generates additional stabilizing contacts is most probably the result of an attraction by the partial negative charge of the ligand oxygen atoms (Fig. 5). An artifact due to crystal packing can be excluded because the conformation is identical in the two nonisomorphous crystal forms of the RXR-BMS 649 complex and different in the isomorphous crystals of the DHA complex. As a result of the change of volume, the level of occupancy of the LBP by the BMS 649 ligand is the highest observed so far.

Why, despite a smaller number of protein-ligand contacts, does 9-_cis-_RA exhibit a higher affinity toward RXR than DHA, BMS 649, or oleic acid? The affinity between the receptor and its ligand results from the energy balance between the cost associated with the desolvation and the conformational adaptation of the bound ligand and the gain resulting from the contacts established within the pocket. This has been described in the case of the structures of VDR-superagonist complexes (48). The RXR-9-_cis-_RA complex may thus represent an optimal compromise between an energetically favorable ligand conformation and sufficient protein-ligand contacts.

Concluding Remarks

The RXR agonists analyzed in this work exhibit a striking structural resemblance with 9-_cis_-RA, as their intrinsic flexibility (e.g. DHA) or their structural adequacy (e.g. BMS 649) allow them to fit into the hydrophobic LBP (Fig. 4). Thus, the hypothesis that ligands adapt their conformation to the binding pocket as was shown in the case of RAR-agonist (49, 50) and VDR-superagonist complexes (48) is also valid in the case of RXRs. The mechanism of protein-ligand recognition is a mutual induced fit that leads to an essentially unique conformation of the LBD.

In contrast to the situation in RARs, the three RXR isotypes (α, β, and γ) exhibit a strict conservation of all the amino acids delineating the binding pocket. Thus, for generating RXR isotype-specific compounds, ligands would probably have to reach residues outside the pocket. This second shell of residues could be made accessible through induced conformational changes affecting the LBP (i.e. at the level of the residue N306) as revealed by the structure of the RXR-BMS649 complex. Although this hypothesis can be questioned, the role of the second shell of residues delineating the binding pockets of proteins has been extensively studied, especially in the case of human immunodeficiency virus type 1 protease-inhibitor complexes (51, 52). Such residues have been shown to play a role in the protein-ligand interactions both at the level of affinity and specificity.

In summary, all physiological, genetic, biochemical, and structural data suggest that RXR could be an opportunistic NR that can bind, with differential affinity and transcriptional activity, several ligands such as 9-_cis-_RA or fatty acids derived from arachidonic acid metabolism depending on the cellular and metabolic context and the local concentration and supply in each potential ligand. Thus, RXR, alone or in association with a heterodimeric partner, could act as a metabolite intracellular sensor.

MATERIALS AND METHODS

Protein Expression and Purification

The human RXRα LBD (residues Thr223 to Thr462) was cloned as an N-terminal hexahistidine-tagged fusion protein in pET15b expression vector and overproduced in Escherichia coli BL21 (DE3) strain. Cells were grown in 2× LB medium and subsequently induced for 5–7 h with 0.8 mm isopropyl-β-d-thiogalactopyranoside at 20 C. In contrast with the previously reported purification of RXRα LBD-9-_cis-_RA complex (35), we purified the apoprotein. Purification procedure included an affinity chromatography step on a cobalt-chelating resin followed by preparative gel filtration to remove apo-tetrameric RXRα LBD species and to keep only the apohomodimer species of RXRα LBD (17). After tag removal by thrombin digestion, protein was further purified on an analytical gel filtration. Apoprotein was concentrated and incubated with a 9:1-fold excess of ligand (alcoholic solutions of DHA or BMS 649) and a 3:1-fold excess of synthetic peptide harboring the second GRIP1 NR-box (30) KHKILHRLLQDSS. Because of ligand photosensitivity, all further manipulations were carried out in dimmed light. Purity and homogeneity of RXRα LBD-ligand complex were assessed by SDS and native PAGE and denaturant and native electrospray ionization mass spectrometry.

Crystallization

Crystals of hRXRα LBD-ligand-peptide complexes were obtained at 22 C and 4 C using vapor diffusion technique in hanging and sitting drops. Reservoir solutions contained 100 mm piperazine-_N,N_′-bis-(2-ethanesulfonic acid), pH 7.0, or bis-Tris, pH 6.5, 5–20% PEG4000, and 5–20% glycerol. Crystals with DHA were easier to obtain and single monocrystals could be grown at 22 C by spontaneous nucleation. By contrast, crystals with BMS 649 appeared to be more fragile and were exclusively grown at 4 C. These crystals belong to the tetragonal space group P43212 with unit cell parameters a = b = 64.4 Å, c = 111.4–112.8 Å, for DHA and BMS 649 complexes, respectively, with one monomer per asymmetric unit and a solvent content of 47%. In the case of the crystals with BMS 649, two crystal forms were simultaneously observed in the same crystallization drops: the tetragonal form and a triclinic form with unit cell parameters a = 47.1 Å, b = 64.7 Å, c = 94.8 Å, and α = 110.0° angle, β = 92.9° angle, γ = 90.0° angle with two homodimers per asymmetric unit and a solvent content of 45%.

Data Collection, Structure Determination, and Refinement

Crystals were flash frozen in liquid ethane using the reservoir solution containing 25% glycerol as a cryoprotectant. One single native data set was collected for each complex on beam line ID14-EH1 at the European Synchrotron Radiation Facility (Grenoble, France). For all complexes and crystal forms, crystals diffracted to at least 1.9Å. Data were processed using DENZO and SCALEPACK programs (53). The structures were solved by molecular replacement using three model probes: the entire or H11-H12 truncated holo wild-type RXRαLBD monomer bound to the natural agonist 9-_cis-_RA (11) and the entire holo-F313A mutant RXRα LBD monomer bound to oleic acid (26). Molecular replacement calculations were carried out on the tetragonal form data sets between 30 and 3.5 Å using the AmoRe program (54). All models gave unambiguous solutions, the best model being the truncated agonist bound holostructure with a correlation coefficient of 37.8% and an R-factor of 48% after fast rigid body refinement. The first map allowed an unambiguous modeling of the coactivator peptide and the repositioning and rebuilding of helix H12 and H11-H12 loop. Iterative cycles of rigid body refinement and torsion angle molecular dynamics at 5,000 K in CNS (55) interspersed with model building in O6 (56) yielded the complete structures. Although their electron densities were clear, the ligand molecules were only included at the last stages of the refinement. Anisotropic scaling and a bulk solvent correction were used, and individual B atomic factors were refined anisotropically. Before a last refinement step, solvent molecules were added according to unassigned peaks in an Fo-FcFourier difference map contoured at 2.5 ς. The final models consist of residues 229–458, the ligands, and 243 (for DHA) or 199 (for BMS 649) water molecules. In both structures, the H1–H3 connecting region and the last four amino acids are not visible.

For the triclinic form, molecular replacement calculations were carried out between 30 and 3.5 Å using the structure solved with the tetragonal crystals. Both monomeric and homodimeric models yielded unambiguous solutions. The refinement was initially carried out in the 30–2.6 Å resolution range and then extended to 1.9 Å resolution. The quality of the density maps allowed us to reliably build a highly structured portion of the H1–H3 loop in only two of the four monomers. Initially, each monomer was successively fixed during the refinement, and restrained noncrystallographic constraints were applied to the helical parts of the structure; at the last cycle of refinement, noncrystallographic symmetry constraints were removed, and all four monomers were refined simultaneously. Anisotropic scaling and a bulk solvent correction were used and individual B atomic factors were refined anisotropically. Before the last refinement step, solvent molecules were added according to unassigned peaks in an Fo-FcFourier difference map contoured at 2.5 ς. The final model contains 829 water molecules and four monomers consisting of residues 229–458 and the ligand.

According to PROCHECK (57), 94% of all residues in all models are in the most allowed main-chain torsion angle Ramachandran regions and 6% in the additionally allowed regions.

Acknowledgments

This work was supported by the Université Louis Pasteur de Strasbourg, the Centre National de la Recherche Scientifique, the Institut National de la Santé et de la Recherche Médicale and Bristol-Myers-Squibb. P.F.E. was the recipient of a fellowship from Association pour la Recherche sur le Cancer.

We are grateful to Drs. Giuseppe Tocchini-Valentini and Valérie Lamour for useful suggestions and help with the generation of the parameter and topology files for the ligands. We wish to thank Julien Lescar, our local contact at European Synchrotron Radiation Facility, Gilbert Bey, and Denis Zeyer for help during data collection. We thank Jean-Paul Renaud for critical reading of the manuscript and Souphatta Sasorith for advice with volume calculations.

Present adress: Macromolecular Structure Group, University of California San Francisco, Department of Biochemistry and Biophysics, School of Medicine, 513 Parnassus Avenue, Box 0448, San Francisco, California 94143.

Abbreviations:

• DHA,
• LBD,
• LBP,
• NR,
• USP,

1

Gronemeyer

H

,

Laudet

V

1995

Transcription factors 3: nuclear receptors.

Protein Profile

2

:

1173

–

1308

2

Steinmetz

AC

,

Renaud

JP

,

Moras

D

2001

Binding of ligands and activation of transcription by nuclear receptors.

Annu Rev Biophys Biomol Struct

30

:

329

–

359

3

Weatherman

RV

,

Fletterick

RJ

,

Scanlan

TS

1999

Nuclear-receptor ligands and ligand-binding domains.

Annu Rev Biochem

68

:

559

–

581

4

Egea

PF

,

Klaholz

BP

,

Moras

D

2000

Ligand-protein interactions in nuclear receptors of hormones.

FEBS Lett

476

:

62

–

67

5

Kastner

P

,

Mark

M

,

Chambon

P

1995

Nonsteroid nuclear receptors: what are genetic studies telling us about their roles in real life?

Cell

83

:

859

–

869

6

Chambon

P

1996

A decade of molecular biology of retinoic acid receptors.

FASEB J

10

:

940

–

954

7

Heyman

RA

,

Mangelsdörf

DJ

,

Dyck

JA

,

Stein

RB

,

Eichele

G

,

Evans

RM

,

Thaller

C

1992

9-_cis-_Retinoic acid is a high affinity ligand for the retinoid X receptor.

Cell

68

:

397

–

406

8

Levin

AA

,

Sturzenbecker

LJ

,

Kazmer

S

,

Bosakowski

T

,

Huselton

C

,

Allenby

G

,

Speck

J

,

Kratzeisen

C

,

Rosenberger

M

,

Lovey

A

1992

9-_cis-_Stereoisomer of retinoic acid binds and activates the nuclear receptor RXRα.

Nature

355

:

359

–

361

9

Allegretto

EA

,

McClurg

MR

,

Lazarchik

SB

,

Clemm

DL

,

Kerner

SA

,

Elgort

MG

,

Boehm

MF

,

White

SK

,

Pike

JW

,

Heyman

RA

1993

Transactivation properties of retinoic acid and retinoid X receptors in mammalian cells and yeast.

J Biol Chem

268

:

26625

–

26633

10

Allenby

G

,

Bocquel

M-T

,

Saunders

M

,

Kazmer

S

,

Speck

J

,

Rosenberger

M

,

Lovey

A

,

Kastner

P

,

Grippo

J

,

Chambon

P

1993

Retinoic acid receptors and retinoid X receptors: interactions with endogenous retinoic acid.

Biochemistry

90

:

30

–

34

11

Egea

PF

,

Mitschler

A

,

Rochel

N

,

Ruff

M

,

Chambon

P

,

Moras

D

2000

Crystal structure of the human RXRα ligand-binding domain bound to its natural ligand: 9-_cis-_retinoic acid.

EMBO J

19

:

2592

–

2601

12

Gampe

RT

,

Montana

VG

,

Lambert

MH

,

Miller

AB

,

Bledsoe

RK

,

Milburn

MV

,

Kliewer

SA

,

Willson

TM

,

Xu

E

2000

Asymmetry in the PPARγ/RXRαcrystal structure reveals the molecular basis of heterodimerization among nuclear receptors.

Mol Cell

5

:

545

–

555

13

Willy

PJ

,

Mangelsdörf

DJ

1997

Nuclear orphans receptors: the search for novel ligands and signalling pathways

. In:

O’Malley

B

, ed. Hormones and signaling.

New York

:

Academic Press

;

307

–

358

14

Mangelsdörf

DJ

,

Evans

RM

1995

The RXR heterodimers and orphans receptors.

Cell

83

:

841

–

850

15

Kersten

S

,

Reczek

PR

,

Noy

N

1997

The tetramerization region of the retinoid X receptor is important for transcriptional activation by the receptor.

J Biol Chem

272

:

29759

–

29768

16

Gampe

RT

,

Montana

VG

,

Lambert

MH

,

Wisely

GB

,

Milburn

MV

,

Xu

EH

2000

Structural basis for autorepression of retinoid X receptor by tetramer formation and the AF-2 helix.

Genes Dev

14

:

2229

–

2241

17

Egea

PF

,

Rochel

N

,

Birck

C

,

Vachette

P

,

Timmins

PA

,

Moras

D

2001

Effects of ligand binding on the association properties and conformation in solution of retinoic acid receptors RXR and RAR.

J Mol Biol

307

:

557

–

576

18

Laudet

V

1997

Evolution of the nuclear receptor superfamily: early diversification from an ancestral orphan receptor.

J Mol Endocrinol

19

:

207

–

226

19

Oro

AE

,

McKeown

MK

,

Evans

RM

1990

Relationship between the product of the Drosophila ultraspiracle locus and the vertebrate retinoid X receptor.

Nature

347

:

298

–

301

20

Jones

GA

,

Sharp

PA

1997

Ultraspiracle: an invertebrate nuclear receptor for juvenile hormones.

Proc Natl Acad Sci USA

94

:

13499

–

13503

21

Billas

IML

,

Moulinier

L

,

Rochel

N

,

Moras

D

2001

Crystal structure of the ligand-binding domain of the ultraspiracle protein USP, the ortholog of retinoid X receptors in insects.

J Biol Chem

10

:

7465

–

7474

22

Clayton

GM

,

Peak-Chew

SY

,

Evans

RM

,

Schwabe

JWR

2001

The structure of the ultraspiracle ligand-binding domain reveals a nuclear receptor locked in an inactive conformation.

Proc Natl Acad Sci USA

98

:

1549

–

1554

23

Escriva

H

,

Delaunay

F

,

Laudet

V

2000

Ligand binding and nuclear receptor evolution.

Bioessays

22

:

717

–

727

24

Mangelsdörf

DJ

,

Umesono

K

,

Evans

RM

1994

The retinoids

. In:

Sporn

MB

,

Roberts

AB

,

Goodman

DS

, eds. Biology, chemistry, and medicine.

New York

:

Raven Press Ltd.

;

319

–

349

25

Parker

RS

1996

Absorption, metabolism, and transport of carotenoids.

FASEB J

10

:

542

–

551

26

Bourguet

W

,

Vivat

V

,

Wurtz,

J-M

,

Chambon

P

,

Gronemeyer

H

,

Moras

D

2000

Crystal structure of a heterodimeric complex of RAR and RXR ligand-binding domains.

Mol Cell

5

:

289

–

298

27

Mata de

Urquiza

A

,

Liu

S

,

Sjöberg

M

,

Zetterström

RH

,

Griffiths

W

,

Sjövall

J

,

Perlmann

T

2000

Docosahexaenoic acid, a ligand for the retinoid X receptor in mouse brain.

Science

290

:

2140

–

2144

28

Bischoff

ED

,

Gottardis

MM

,

Moon

TE

,

Heyman

RA

,

Lamph

WW

1998

Beyond tamoxifen: the retinoid X receptor selective ligand LGD 1069 (Targretin) causes complete regression of mammary carcinoma.

Cancer Res

58

:

479

–

484

29

Wurtz

J-M

,

Bourguet

W

,

Renaud

J-P

,

Vivat

V

,

Chambon

P

,

Moras

D

,

Gronemeyer

H

1996

A canonical structure for the ligand-binding domain of nuclear receptors.

Nat Struct Biol

3

:

87

–

94

30

Shiau

AK

,

Barstad

D

,

Loria

PM

,

Cheng

L

,

Kushner

PJ

,

Agard

DA

,

Greene

GL

1998

The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen.

Cell

95

:

927

–

937

31

Darimont

BD

,

Wagner

RL

,

Apriletti

JW

,

Stallcup

MR

,

Kushner

PJ

,

Baxter

JD

,

Fletterick

RJ

,

Yamamoto

KR

1998

Structure and specificity of nuclear receptor-coactivator interactions.

Genes Dev

12

:

3343

–

3356

32

Nolte

RT

,

Wisely

GB

,

Westin

S

,

Cobb

JE

,

Lambert

MH

,

Kurokawa

R

,

Rosenfeld

MG

,

Willson

TM

,

Glass

CK

,

Millburn

MV

1998

Ligand binding and coactivator assembly of peroxysome proliferator activated receptor γ.

Nature

395

:

137

–

143

33

Bourguet

W

,

Ruff

M

,

Chambon

P

,

Gronemeyer

H

,

Moras,

D

1995

Crystal structure of the ligand binding domain of the human nuclear receptor RXRα.

Nature

375

:

377

–

382

34

Rochel

N

,

Tocchini-Valentini

G

,

Egea

PF

,

Juntunen

K

,

Garnier

J-M

,

Vihko

P

,

Moras

D

2001

Biochemical and biophysical characterization of the ligand binding domain of the vitamin D nuclear receptor.

Eur J Biochem

268

:

971

–

979

35

Egea

PF

,

Moras

D

2001

Purification and crystallization of human RXRαligand binding domain/9-_cis-_retinoic acid complex.

Acta Crystallogr D

57

:

434

–

437

36

Klaholz

BP

,

Moras

D

2000

Structural role of a detergent molecule in retinoic acid nuclear receptor crystals.

Acta Crystallogr D

56

:

933

–

935

37

Stehlin

C,

Wurtz

J-M,

Steinmetz

A,

Greiner

E,

Schuele

R,

Moras

D,

Renaud

J-P

2001

X-ray structure of the orphan nuclear receptor RORβ ligand-binding domain in the active conformation.

EMBO J

20

–21:

5822

–

5831

38

Harmon

MA

,

Boehm

MF

,

Heyman

RA

,

Mangelsdörf

DJ

1995

Activation of mammalian retinoid X receptor by the insect growth regulator methoprene.

Proc Natl Acad Sci USA

92

:

6157

–

6160

39

Kitareewan

S

,

Burka

L

,

Tomer

B

,

Parker

CE

,

Deterding

LJ

,

Stevens,

RD

,

Forman

BM

,

Mais

DE

,

Heyman

RA

,

McMorris

T

,

Weinberger

C

1996

Phytol metabolites are circulating dietary factors that activate the nuclear receptor RXR.

Mol Cell Biol

7

:

1153

–

1166

40

LeMotte

PK

,

Keidek

S

,

Apfel

CM

1996

Phytanic acid is a retinoid X receptor ligand.

Eur J Biochem

236

:

328

–

333

41

Yu

K

,

Bayona

W

,

Kallen

CB

,

Harding

HP

,

Ravera

CP

,

McMahon

G

,

Brown

M

,

Lazar

MA

1995

Differential activation of peroxisome proliferator-activated receptors by eicosanoids.

J Biol Chem

270

:

23975

–

23983

42

Devchand

PR

,

Keller

H

,

Peters

JM

,

Vazquez

M

,

Gonzalez

FJ

,

Wahli

W

1996

The PPARα-leukotriene B4 pathway to inflammation control.

Nature

384

:

39

–

43

43

Kliewer

SA

,

Lenhard

JM

,

Wilson

TM

,

Patel

I

,

Morris

DC

,

Lhman

JM

1995

A prostaglandin J2metabolite binds peroxisome proliferator-activated receptor γ and promotes adipocyte differentiation.

Cell

83

:

813

–

819

44

Forman

BM

,

Tontonoz

P

,

Chen

J

,

Brun

RP

,

Spiegelman

BM

,

Evans

RM

1995

15 Deoxy Δ,14 prostaglandin J2 is a ligand for the adipocyte determination fatcor PPARγ.

Cell

83

:

803

–

812

45

Xu

EH

,

Lambert

MH

,

Montana

VG

,

Parks

DJ

,

Blanchard

SG

,

Brown

PJ

,

Stermbach

DD

,

Lehman

JM

,

Wisely

GB

,

Willson

TM

,

Kmiewer

SA

,

Milburn

MV

1999

Molecular recognition of fatty acids by peroxysome proliferator-activated receptors.

Mol Cell

3

:

397

–

403

46

Vivat

V

,

Zechel

C

,

Wurtz

J-M

,

Bourguet

W

,

Kagechika

H

,

Umemiya

H

,

Shudo

K

,

Moras

D

,

Gronemeyer

H

,

Chambon

P

1997

A mutation mimicking ligand-induced conformational change yields a constitutive RXR that senses allosteric effects in heterodimers.

EMBO J

16

:

5697

–

5709

47

Peet

DJ

,

Doyle

DF

,

Corey

DR

,

Mangelsdörf

DJ

1998

Engineering novel specificities for ligand-activated transcription in the nuclear hormone receptor RXR.

Chem Biol

5

:

13

–

21

48

Tocchini-Valentini

G

,

Rochel

N

,

Wurtz

J-M

,

Mitschler

A

,

Moras

D

2001

Crystal structures of the vitamin D receptor complexed to super agonist 20-epi ligands.

Proc Natl Acad Sci

98

:

5491

–

5496

49

Klaholz

BP

,

Renaud

J-P

,

Mitschler

A

,

Zusi

C

,

Chambon

P

,

Gronemeyer

H

,

Moras

D

1998

Conformational adaptation of agonists to the human nuclear receptor RARγ.

Nat Struct Biol

5

:

199

–

201

50

Klaholz

BP

,

Mitschler

A

,

Belema

M

,

Moras

D

2000

Enantiomer discrimination illustrated by high-resolution crystal structures of the human nuclear receptor hRARγ.

Proc Natl Acad Sci USA

97

:

6322

–

6327

51

Hong

L

,

Zhang

XC

,

Harstuck

JA

,

Tang

J

2000

Crystal structure of an in vivo HIV-1 protease inhibitor in complex with saquinavir: insights into the mechanism of drug resistance.

Protein Sci

9

:

1898

–

1904

52

Boden

D

,

Markowitz

M

1998

Resistance to human immunodeficiency virus type 1 protease inhibitors.

Antimicrob Agents Chemother

42

:

2775

–

2783

53

Otwinowski

Z

,

Minor

W

1996

Processing x-ray data collected in oscillation mode.

Methods Enzymol

276

:

307

–

326

54

Navaza,

J

1994

AmoRe: an automated package for molecular replacement.

Acta Crystallogr A

50

:

157

–

163

55

Brünger

AT

,

Adams

PD

,

Clore

GM

,

DeLano

WL

,

Gros

P

,

Grosse-Kunstleve

RW

,

Jiang

JS

,

Kuszewski

J

,

Nilges

M

,

Pannu

NS

,

Read

RJ

,

Rice

LM

,

Simonson

T

,

Warren

GL

1998

Crystallography & NMR system: a new software suite for macromolecular structure determination.

Acta Crystallogr D

54

:

905

–

921

56

Jones

TA

,

Zou

JY

,

Cowan

SW

,

Kjeldgaard

M

1991

Improved methods for building proteins models in electron density maps and the location of errors in these models.

Acta Crystallogr A

47

:

110

–

119

57

Laskowski

RA

,

MacArthur

MW

,

Moss

DS

,

Thornton

JM

1993

PROCHECK: a program to check the stereochemical quality of protein structure coordinates.

J Appl Crystallogr

26

:

283

–

291

58

Kleywegt

G

,

Jones

TA

1994

Detection, delineation, measurement and display of cavities in macromolecular structures.

Acta Crystallogr D

50

:

178

–

185

Copyright © 2002 by The Endocrine Society

Citations

Views

Altmetric

Metrics

Total Views 1,937

1,272 Pageviews

665 PDF Downloads

Since 1/1/2017

×

Email alerts

Related articles in PubMed

Citing articles via

• Latest

• Most Read

• Most Cited

More from Oxford Academic