A Family-based Approach Reveals the Function of Residues ......UNCORRECTED PROOF A Family-based...

UNCORRECTED PROOF

A Family-based Approach Reveals the Functionof Residues in the Nuclear ReceptorLigand-binding Domain

Simon Folkertsma1, Paula van Noort4, Joost Van Durme1,Henk-Jan Joosten1, Emmanuel Bettler5, Wilco Fleuren1,Laerte Oliveira2, Florence Horn3, Jacob de Vlieg4 and Gerrit Vriend1*

1CMBI, University of NijmegenPO Box 9010, 6500 GLNijmegen, The Netherlands

2Escola Paulista de MedicinaUNIFESP, PO Box 2038804041-990 Sao Paulo, Brazil

3Department of Cellular andMolecular PharmacologyUCSF, P.O. Box 2240, SanFrancisco, CA 94143-2240USA

4Organon NV, PO Box 205340 BH Oss, The Netherlands

5IBCP, Passage du Vercors 769367 Lyon, France

Literature studies, 3D structure data, and a series of sequence analysistechniques were combined to reveal important residues in the structureand function of the ligand-binding domain of nuclear hormone receptors.A structure-based multiple sequence alignment allowed for the seamlesscombination of data from many different studies on different receptorsinto one single functional model.

It was recently shown that a combined analysis of sequence entropyand variability can divide residues in five classes (1) the main function oractive site, (2) support for the main function, (3) signal transduction,(4) modulator or ligand binding and (5) the rest. Mutation data extractedfrom the literature and intermolecular contacts observed in nuclear recep-tor structures were analyzed in view of this classification and showed thatthe main function or active site residues of the nuclear receptor ligand-binding domain are involved in cofactor recruitment. Furthermore, thesequence entropy-variability analysis identified the presence of signaltransduction residues that are located between the ligand, cofactor anddimer sites, suggesting communication between these regulatory bindingsites. Experimental and computational results agreed well for most resi-dues for which mutation data and intermolecular contact data were avail-able. This allows us to predict the role of the residues for which nofunctional data is available yet.

This study illustrates the power of family-based approaches towardsthe analysis of protein function, and it points out the problems and possi-bilities presented by the massive amounts of data that are becoming avail-able in the “omics era”. The results shed light on the nuclear receptorfamily that is involved in processes ranging from cancer to infertility,and that is one of the more important targets in the pharmaceuticalindustry.

q 2004 Published by Elsevier Ltd.

Keywords: nuclear receptors; entropy–variability plots; mutation data;ligand binding residues; structure-based sequence alignment*Corresponding author

Introduction

Nuclear receptors

Nuclear receptors (NRs) are ligand-inducibletranscription factors that regulate processes, suchas homeostasis, differentiation, embryonic develop-ment and organ physiology.1 A total of 48 humanNRs have been identified.2 Their ligands are lipo-philic compounds such as steroids, thyroid hor-mone, vitamin D3, and retinoids. However, the

0022-2836/$ - see front matter q 2004 Published by Elsevier Ltd.

E-mail address of the corresponding author:[email protected]

Abbreviations used: DBD, DNA-binding domain; ER,estrogen receptor; EcR, ecdysone receptor; LBD, ligand-binding domain; LIG, ligand interacting group; LRH-1,liver receptor homolog 1; NR, nuclear receptor; NRMD,nuclear receptor mutation database; PDB, RCSB ProteinData Bank; PPAR, peroxisome proliferator-activatedreceptor; PXR, pregnane X receptor; RXR, retinoid Xreceptor; CR, conserved regions; USP, ultraspiracleprotein; VDR, vitamin D receptor.

YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL

doi:10.1016/j.jmb.2004.05.075 J. Mol. Biol. (2004) not known, xxx–xxx

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

ARTICLE IN PRESS

UNCORRECTED PROOF

endogenous ligands are not yet known for 30% ofthe NRs.3 NRs are implicated in many importantdiseases like cancer,4 diabetes,5 and osteoporosis,6

and, therefore, are targets for pharmaceuticalindustries with similar importance as the Gprotein-coupled receptors (GPCRs), ion channels,or kinases.7

Figure 1 shows the common architecture ofNRs.8 They consist of a variable N-terminaldomain, a conserved DNA-binding domain (DBD)that binds to the hormone response element(HRE) of the target gene, a flexible linker region,and a C-terminal ligand-binding domain (LBD).Some receptors possess an additional far C-terminalregion for which the function is not yet known.The N-terminal region of most NRs has ahormone-independent trans-activation function(AF-1). The hormone-dependent trans-activationfunction (AF-2) is located in the LBD.9

The DBD is highly conserved. It consists of 66residues, and contains two C4 type zinc fingers.10

The LBD consists of about 250 amino acid residues.LBDs are much less conserved than DBDs. Never-theless, LBDs have a common fold that consists of12 helices (numbered H1–H12; see Figure 2)organized in a helical sandwich, and one betasheet that normally consists of two short strands.11

The structure of this sheet is rather variable. Forexample, the pregnane X receptor (PXR) has a betasheet of five strands12 and the peroxisome pro-liferator-activated receptors (PPARs) have fourstrands.13 The region between H1 and H3 is vari-able in length and structure, and is helical in onlya few receptors (e.g. PPARs, the retinoic acid-related orphan receptors (RORs), the vitamin Dreceptor (VDR) and liver receptor homolog 1(LRH-1)). In some X-ray structures, the H2 regionis not observed due to disorder or mobility.

The ligand-binding pocket is very hydrophobicand is mainly formed by helices H3, H5, H6, andH11 and the small sheet. The residues of theligand-dependent activation function 2 (AF-2) arelocated in H12.9 The position of this helix is highlyflexible, and its repositioning upon ligand bindingis essential for recruitment of transcription cofactors,the so-called coactivators or corepressors. The three-dimensional structures of NRs have been deter-mined in four different receptor conformations. At

present these four conformations have not yetbeen solved for one receptor and therefore multiplereceptors are used to illustrate these four confor-mations: apo-form (Figure 2(a)), agonist-bound(Figure 2(b)), and two different types of antagonistbound (Figure 2(c) and (d)).

Most NRs can bind coactivators and corepres-sors that both bind at nearly the same spot, the so-called cofactor-binding site. Coactivators and core-pressors cannot be bound simultaneously. Uponagonist binding, H12 docks at the entrance to theligand-binding pocket in the so-called H12 groovethat is mainly formed by H3 and H11 (Figure 2(b)).This creates a charge clamp14 consisting of the con-served glutamate in H12 and the conserved lysineat the C-terminal end of helix 3. This chargeclamp is crucial for coactivator binding by inter-acting with the two poles of the helix dipole of therecognition helix of the coactivator.15,16 Uponantagonist binding, the ligand prevents H12 fromdocking in the H12 groove, a corepressor occupiespart of this groove and H12 is forced to take analternative position (Figure 2(c)). On the otherhand, antagonists can also force H12 to dock inthe cofactor-binding groove, resulting in repressionof trans-activation (Figure 2(d)). In this positionH12 blocks coactivator binding.17 Some NRs donot fit this mechanistic model. The Estrogen-Related Receptor 3 (ERR3) and LRH-1 constitu-tively activate transcription without any ligandbound.18,19 Similarly, the orphan receptor Nurr1neither has a ligand-binding pocket, nor a cofactor-binding site and is constitutively active.20 Ligand-independent stabilization is important for thetranscription activity of these receptors.19,20

Sequence analysis

The analysis of conservation, correlation andvariability patterns in multiple sequence align-ments is a powerful tool for the study of proteinfamilies. The constraints that structure and func-tion put on the sequence variation caused byevolutionary processes leave their traces in themultiple sequence alignment in many differentways.21–24 The most prominent evolutionary trace,residue conservation, has been evaluated in mul-tiple sequence alignments by means of variability(number of different amino acids found), Shannonentropy, variance-based and score-matrix indices.25–27

Conservation patterns have been used to improveor evaluate multiple sequence alignments,26 orused to define structure fingerprints.28 Conservedresidues sometimes are clustered at “universallyconserved positions”,29 that can form a motifcharacteristic of the fold. These positions also canbe found in the corresponding segments of analogsand their location often coincides with that ofsuper-sites.30 The identification of conservationpatterns in proteins has been used to search forfunction. Some methods are based on energy calcu-lations, and look for surfaces that potentiallyinteract.31–36 Other methods predict functional

Figure 1. Schematic representation of a NR. NRs con-sist of six domains (A–F). The Nterminus (A/B) is vari-able, the DNA binding domain (DBD; C) is the mostconserved region and contains two zinc fingers. Thehinge region (D) is the connection between the DBDand the ligand-binding domain (LBD). The LBD (E) ismainly responsible for ligand binding and dimer for-mation and contains the activation function 2 (AF-2)domain. Some receptors possess an additional F region,for which the function is yet unknown.


2 Role of Amino Acids in NR-LBDs

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

ARTICLE IN PRESS

UNCORRECTED PROOF

motifs from an analysis of protein interaction sur-faces using principal component analysis,37 analy-sis of physicochemical descriptors of protein–protein interactions,38 motifs in Blocks databases,39

or alignment of hinge regions.40 Evolutionary traceanalyses involve searching for conservation pat-terns in different branches of phylogenetic treesand mapping them onto 3D structures to look forclusters of functionally important residues.24,41–43

These sequence analysis methods normally use asingle measure of variability, and a Shannon-typeentropy term is commonly selected. Many of these

methods are well suited to find functionallyimportant residues but cannot generate a compre-hensive overview of residue functions relative toeach other and relative to the structure.We previously developed a sequence analysis

technique based on the combination of twosequence variability measures. The first is aShannon-type entropy. The second, variability, isthe number of different amino acid types observedat one position in a multiple sequence alignment.A relation between the function of a residue andits location in a plot of entropy versus variability

Figure 2. (a) Apo form of a NR LBD (Figure based on PDB entry 1LBD,74 human retinoid X receptor (hRXRa)). (b)Structure with a bound agonist (3ERD,75 hERa, human estrogen receptor alpha). Act ¼ the nuclear receptor coactivator2 (NCoA-2) box II. (c) Structure with a bound antagonist and corepressor (1KKQ,76 hPPARa). Rep ¼ the nuclear recep-tor co-repressor 2 (N-CoR2). (d) Structure with a bound antagonist (3ERT,75 hERa). Figures were generated withYASARA (http://www.yasara.org). Helices 1–12 are shown as cylinders, the sheet as (red) arrows.


Role of Amino Acids in NR-LBDs 3

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

ARTICLE IN PRESS

http://www.yasara.org



UNCORRECTED PROOF

has been shown.21,22 Fine-tuning of the entropy bythe variability and fine-tuning of the variability bythe entropy allows us to draw many more con-clusions about the role of individual residue posi-tions than is possible using techniques based onvariability or entropy alone.

The method was tested on four protein familiesfor which very many sequences are available andfor which the function of nearly all residueshave been well-established experimentally:globin chains,44 –47 GPCRs,22 ras-like proteins,48 –52

and serine-proteases.53,54 Positions related to themain function, related to co-factor or regulatorbinding, positions in the core of the protein, andpositions not associated with any known functioncluster in the entropy-variability plots. Thismethod requires correct alignments, which canbe difficult to obtain when sequences in thefamily share low sequence identity. To circum-vent this problem a multiple sequence alignmentmethod based on structure superposition wasused for the LBDs.

To corroborate the predicted role of residues inthe LBD, the nuclear receptor mutation data-base55 (NRMD) was used as a source for experi-mental verification. Experimental data onmutants are available for more than 75% of allresidue positions in the NR LBD family. Theseexperimental results were compared with theresidue function as predicted from the sequenceentropy and variability analysis. The excellentagreement between theory and experiment forthe residues for which mutation and structuraldata are available gives confidence in the predic-tion of the functional classification of the residuesin the NR LBD for which no experimental dataare available yet.

Results and Discussion

The structurally conserved residues

A fully automatic structure-based alignment of184 individual LBD domains from 97 PDB filesrevealed 152 structurally conserved residues ofwhich the Ca is within 1.9 A of the average Ca.The fully automatic multiple structure superposi-tion procedure uses rather strict acceptance criteriaupon defining this structurally conserved core.After visual inspection of the superposed LBDdomains, 23 residues were added (mainly locatedat both ends of H1, the second sheet, H6 and theN-terminal part of H7). Helix 12 holds the con-served LXXLL sequence motif, which allows us tounambiguously add a helix of eight amino acidsto the multiple sequence alignment despite itslarge structural heterogeneity. The final, extendedset of 183 aligned residues (conserved regions, CR,

Figure 3(a)) includes all helices and strands exceptthe highly variable, and often even absent helix 2.Details about the automatic structure superposi-tion method are available†. This WWW pageholds a detailed explication of the superpositionmethod,56 the list of files and domains used, theRMS deviation for each residue in the superposi-tion, pictures of each superposed domain, etc.Figure 3 shows the core of 183 residues(Figure 3(a)) and the RMS deviation per residue(Figure 3(b)).

The retinoid X receptors (RXRs) have an inser-tion of a glutamic acid in the middle of H7.57 Forconvenience, this glutamic acid is part of the align-ment, but has not been included in the analyses.Table 1 shows for all regular secondary structureelements their lengths and lists the most conservedresidues. Per structural element, the most con-served residue was labelled with the helix numberfollowed by 50.

Multiple sequence alignment of the NR LBDs

The structure-based alignment of 97 NR LBDswas used to generate an alignment profile. Startingfrom this profile 443 NR sequences could bereliably aligned. This alignment is available fromthe NucleaRDB‡. The sequences of the nematodeCaenorhabditis elegans were not included in themultiple sequence alignment, because their LBDsare too different from all others.58 Also, despite the

Table 1. Numbering and length of secondary structureelements in the CR

Secondarystructureelement Numbering Length

3D number andidentity of most

conservedresidue

Percentidentity

Helix 1 137–156 20 150 E 48Helix 3 328–353 26 350 A 95Loop H3-H4

L341–L346 6 –

Helix 4 447–458 12 450 D 100Helix 5 548–560 13 550 E 75Beta sheet1

B150–B155 6 –

Beta sheet2

B250–B254 5 –

Helix 6 (ifpresent)

647–653 7 650 (A, G, L, M) 18/11/23/21

Helix 7 736–751 16 750 L 46Loop H7-H8

L781–L785 5

Helix 8 848–859 12 850 E 84Helix 9 936–958 23 950 L 88Helix 10/11

1038–1061 24 1050 L 68

Helix 12 1245–1252 8 1250 E 80

The last two digits are the residue number in the secondarystructure element. The helix number, or L or B precedes thesetwo digits for loop or b-strand, respectively.

†http://www.receptors.org/NR/articles/JMB04/structsup.html ‡http://www.receptors.org/NR/



379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

ARTICLE IN PRESS

http://www.receptors.org/NR/articles/JMB04/structsup.html


http://www.receptors.org/NR/

UNCORRECTED PROOF

structure-based alignment, it was not yet possibleto unambiguously align H1, the sheet region, H6,H7 and H12 in all receptors. The multiple sequencealignment shows well-aligned sequence finger-prints in many helices, e.g. the (W/F)A(K/R)motif in the C-terminal end of H3, the conservedDQ in H4 and the conserved glutamate in H12, asdescribed by Wurtz et al.11

The ligand-binding residues

To identify the ligand-binding residues in thestructure-based multiple sequence alignment the3D positions of ligand-binding residues weredetermined in 86 LBD structures. NR LBD ligand-binding pockets vary widely in shape and volume.Consequently, we observe residue positions thatare involved in ligand binding in some NRs butnot in others. The positions of residues that arefrequently (Ligand Interacting Group 1, LIG1),moderately (LIG2), occasionally (LIG3), or rarely(LIG4) involved in ligand contacts are shown inFigure 4 and listed quantitatively§. The datasetconsist of 72 unique ligand–receptor combinations;LIG1 residues generally make at least one contactin 63–72 of these unique complexes, LIG2 in 38–62, LIG3 in 12–37 and LIG4 in 1–11. Of the sixLIG1 residues, two are in H3, two in the middle ofH5, one in the sheet and one in H11 (Figure 4(a)).The ten LIG2 residues are located in the C-terminalof H3, in H5 and in H11 (Figure 4(b)). The LIG3

and LIG4 residue positions generally are furtheraway from the centre of the ligand-binding pocket(Figure 4(c) and (d)).PPAR, for example, can bind fatty acids and

their long hydrophobic tails make contacts inregions of the receptor LBD where steroidsnever bind. Several LIG4 residue positions areinvolved in ligand binding in the PXR receptor.PXR serves as a broad chemical “sensor”12 andcan bind very large compounds (such asrifampicin). Figure 4(e) shows the four residuepositions (B155, 342, 559, 1057) that often forma hydrogen bond with the ligand. Most ligandsare anchored in the binding pocket by at leastone such hydrogen bond.

Entropy–variability plot

Previous studies on four very well characterisedprotein families21,22 (globins, ras-like proteins,GPCRs, and serine-proteases) showed that resi-dues can be clustered in five groups as function oftheir entropy and variability. These groups corre-spond to boxes in the entropy-variability plot asshown in Figure 5, and the following relationbetween the boxes and residue function wasobserved: box11 residues are involved in the mainfunction of the protein; box12 contains the shell ofresidues around the main functional site; box22residues are often spatially located between box12

and box23 residues, these positions are mainlyinvolved in signal transduction between the modu-lator site and main functional site; box23 mainlycontains residues that interact with a modulator;

Figure 3. (a) Structurally conserved regions of the NR LBD mapped on 3ERD.75 Figure was generated with YASARA(http://www.yasara.org). (b) A Ca trace of the structurally averaged LBD residues is shown in red. The three-dimensional blue ellipsoids represent the spread of the positions of the superposed alpha carbons. Details of theellipsoid calculations are given on the website http://www.receptors.org/NR/articles/JMB04/structsup.html

§http://www.receptors.org/NR/articles/JMB04/ligs.html



505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

ARTICLE IN PRESS






http://www.receptors.org/NR/articles/JMB04/ligs.html

http://www.receptors.org/NR/articles/JMB04/ligs.html

UNCORRECTED PROOF

box33 residues mainly are found at the surface ofthe protein away from the main functional site.

Figure 5 shows the entropy variability plot basedon the alignment of 443 NR-LBDs. In the followingsections, the residues in the five boxes will bedescribed box by box. However, it has to be keptin mind that the location of a residue in theentropy variability plot depends on the choice ofsequences and on the alignment. Consequently,residues that are located near the border betweenboxes sometimes can end up in the neighbour boxif just a few sequences are added to the alignment.We have previously observed that residues nearthe borders between two boxes display functionalcharacteristics commensurate with both boxes;21,22

the constant factor is that residues that are at asimilar location in the entropy-variability plot are

also functionally similar. The recipe for definingthe borders was optimised previously based uponexperimental data on globins, ras-like proteins,GPCRs, and serine-proteases.21,22

Functional characterisation of box residues

Figure 6 shows the mapping of ligand-binding,cofactor-binding and dimer-binding residues onthe different boxes.

The important cofactor-binding residues aremainly observed in box11 and box12, indicating thatthe main function of NRs is the recruitment ofcofactors (Figure 6(a)). The most frequent ligand-binding residues (LIG1 and LIG2) fall mainly inbox23 (Figure 6(b)). As expected a large number ofthe less frequent ligand-binding residues (LIG3

Figure 4. Location of the 56 ligand binding positions in the NR LBD.(a) LIG1 (purple, six residues). (b) LIG2 (red, tenresidues). (c) LIG3 (orange, 14 residues). (d) LIG4 (yellow, 26 residues). (e) Main hydrogen bond forming positions(light blue). Ligand (BMD184394) is shown in purple. The depicted structure is the LBD of the human retinoic acidreceptor gamma (hRARg), 1FCX.77



631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

ARTICLE IN PRESS

UNCORRECTED PROOF

and LIG4) tend to be located mainly in box33. Resi-dues involved in homo- and heterodimerisationare located for 50% in box33 (Figure 6(c)). Theseresidues are located at the surface of the LBD.Some of the frequently observed dimerisationresidues, however, are observed in box22 andbox23, suggesting that a dimerisation partner mayalso act as a modulator and that signal transduc-tion between the three binding sites (cofactor,ligand, dimer) plays a role in dimerisation. Arecent article describing the heterodimer of theecdysone receptor (EcR) and ultraspiracle (USP)shows that ligand binding to EcR requires inter-action with USP, allowing the flexible part of theEcR to mould around the ligand.59 It is alsoknown that a ligand of one member of an RXR het-erodimer can affect the activity of the partner LBD,which is known as the “phantom ligand effect”.60,61

Box11

The seven box11 residues are located at the C-terminal end of H3, the loop between H3 andH4, and at the N-terminal end of H4 (Figure7(a)). This region is known as the cofactor-binding site, and the location of box11 residuesin this region confirms that cofactor binding isthe main function of NRs. Three out of sevenbox11 residues make a contact with the cofactorfragment in at least one PDB62 file. Because wedo not observe box11 residues outside the cofactor-

binding pocket, we concluded that this pocketis the entire cofactor-binding region, and thatthe structurally unknown rest of the cofactordoes not make additional important contactswith the LBD.

Box12

Box12 contains nine residues, which all face box11residues and are all located between box11 residuesand either the ligand-binding pocket or the dimerinterface (Figure 7(b)). Our cofactor contact analy-sis showed that three of these box12 residues (347,351 and 456) are frequently (84%, 100% and 74%,respectively) observed in an interaction with acofactor. The predicted function of box12 residuesis assistance to the main function. The location ofthe box12 residues suggests that ligand and dimerpartner are assisting in cofactor recruitmentthrough these box12 residues.

Box22

Box22 contains 23 residues, which form thesecond shell of positions around the box11 andbox12 residues. They are mainly located in the“core” of the protein in-between the cofactor siteand either the ligand-binding pocket or the dimerinterface. The location of these residues in theLBD is in agreement with a potential signal trans-duction role between the three functional sites

Figure 5. Entropy-variability plot for residue positions in the CR. The five boxes (11, 12, 22, 23, and 33) are indicated.The curved line indicates the maximum entropy possible as function of the variability.



757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

875

876

877

878

879

880

881

882

ARTICLE IN PRESS

UNCORRECTED PROOF

(cofactor-binding site; ligand-binding site; dimerinterface site). A few box22 residues are located atthe dimer interface surface (Figure 7(c)). Previousentropy versus variability analyses showed that

box22 residues were most often buried, but whenobserved at the surface, they were almost invari-ably involved in functionally important intermolecular interactions. In haemoglobins, forexample, surface-exposed box22 residues are impli-cated in subunit–subunit interactions21 that areimportant for the cooperativity of oxygen binding.In LBDs, we observe three box22 residues at the sur-face of helix 10. This helix often plays an importantrole in dimer formation.57,63 The fact that we observebox22 residues at the dimer interface surface is inagreement with the known LBD–LBD interactionsthat are functionally very important. This suggeststhat cooperativity can exist in the ligand binding oftwo LBDs that form a dimer.

Box23

The 53 box23 residues comprise almost the entireligand-binding pocket and consist mainly of resi-dues in H3, H5 and H11 (Figure 7(d)). The majorityof the box23 residues are members of the LIGs(Figure 6(b)). This is in good agreement with thepredicted function of modulator binding residues.

Box33

Box33 contains 70 residues. Figure 7(e) showsthat they are located at the surface of the LBD.Many of them are involved in dimerisation (seeFigure 6(c)). Half of the LIG3 and LIG4 residues,only one LIG2 residue, but none of the LIG1 resi-dues are observed in box33 (Figure 6(b)). This con-firms the predicted role of box33 residues as beingsurface-exposed and not essential for the functionof the protein.

To corroborate the function of the various boxresidues, mutation data for the CR residues wereextracted (when available) from the nuclear recep-tor mutant database55 (NRMD). The results ofmutations of box11 residues have been summar-izedk. Five out seven box11 mutations are relatedto a disease, indicating distortion of the main func-tion. Only the box11 residue mutations at positions350 and 454 have not yet been observed to berelated to a disease. However, mutation of residue350 in the human androgen receptor results in lessthan 10% of the amount of ligand bound comparedto the wild-type receptor, while there was nochange in affinity for the longer isoform of coacti-vator ARA70.

64 Mutation of residue 454 led toimpairment of transcription activation by thehuman VDR. The mutant could bind 1,25-dihydroxyvitamin but was defective in forming aheterodimer with RXR.65

To characterise and verify the potential func-tional role of the residues in the five boxes weanalysed their mutations. 45% of the mutations inthe NRMD are extracted automatically from the

Figure 6. Distribution of residue positions involved inthe three functions of the NR LBD. (a) Cofactor-contactingresidues. COF1 are residue positions found to bind thecofactor in more than 60% of the known cases, COF2 arethe less frequently observed cofactor-binding residue pos-itions (but observed in.30% of the known structures). (b)Ligand-contacting residues (LIG1, LIG2, LIG2, LIG4). (c)Residues involved in dimerisation. DIM1 and DIM2 asCOF1 and COF2 but contacts are dimerisation contacts.(d) Number of CR positions that are not involved in inter-molecular interactions in any of the 97 PDB files studied.(e) As (d) but expressed as a percentage.

khttp://www.receptors.org/NR/articles/JMB04/box11mutations.html



883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

1001

1002

1003

1004

1005

1006

1007

1008

ARTICLE IN PRESS

http://www.receptors.org/NR/articles/JMB04/box11mutations.html

http://www.receptors.org/NR/articles/JMB04/box11mutations.html

UNCORRECTED PROOF

literature by MuteXt,66 the rest are collected auto-matically from several web-based resources{ ormanually extracted from the literature. MuteXtcan identify 64.5% of the NR point mutations witha specificity of 85.8% (specificity is the ratiobetween the true positives and the sum of truepositives and false positives). A total of 375mutations were distributed over five functionalclasses: (a) disease or transcription; (b) cofactorbinding; (c) dimerisation; (d) ligand binding; (e)no effect. Disease and transcription were taken asone category because they are related and forboth it is unclear whether they are caused by the

underlying effects on cofactor, ligand or dimerpartner binding.Figure 8 shows that almost 50% of the mutations

are disease or transcription related. The remaining50% of the mutations are approximately evenlydistributed across the four other categories. 5% ofall mutations were described to have an effect inmultiple categories, and in 50% of all cases differ-ent studies showed different effects for the samemutated position. This seems to suggest that somestudies are conflicting, but we think that moreoften the data actually are complementary. Resi-dues (especially those in box22) often are involvedin multiple aspects of the functioning. If a studyaims at measuring ligand binding of a certain resi-due, then it is likely that effects at transcriptionmight not be observed in that same study. In all

Figure 7. Structural location of the residues in the five boxes. (a) Box11 (purple). (b) Box12 (orange). (c) Box22 (yellow).(d) Box23 (green). (e) Box33 (light blue). In (f) box33 residues (light blue) are shown again but this time accessible resi-dues are indicated with small dots. For easy reference, the coactivator helix of the glucocorticoid receptor-interactingprotein 1 (GRIP-1), as observed in a PDB file of hERa (3ERD75), and the agonist dihydrotestosteron as observed in thePDB file of the human androgen receptor (1I3778) are shown in yellow and purple, respectively.

{http://www.receptors.org/NR/articles/JMB04/mutationsources.html



1009

1010

1011

1012

1013

1014

1015

1016

1017

1018

1019

1020

1021

1022

1023

1024

1025

1026

1027

1028

1029

1030

1031

1032

1033

1034

1035

1036

1037

1038

1039

1040

1041

1042

1043

1044

1045

1046

1047

1048

1049

1050

1051

1052

1053

1054

1055

1056

1057

1058

1059

1060

1061

1062

1063

1064

1065

1066

1067

1068

1069

1070

1071

1072

1073

1074

1075

1076

1077

1078

1079

1080

1081

1082

1083

1084

1085

1086

1087

1088

1089

1090

1091

1092

1093

1094

1095

1096

1097

1098

1099

1100

1101

1102

1103

1104

1105

1106

1107

1108

1109

1110

1111

1112

1113

1114

1115

1116

1117

1118

1119

1120

1121

1122

1123

1124

1125

1126

1127

1128

1129

1130

1131

1132

1133

1134

ARTICLE IN PRESS

http://www.receptors.org/NR/articles/JMB04/mutationsources.html

http://www.receptors.org/NR/articles/JMB04/mutationsources.html

UNCORRECTED PROOF

these cases, we simply counted each experimentalobservation once, so that we have considerablymore mutations than residue positions in thisstudy. The distribution of the mutations of resi-dues over the five categories for the five entropy-variability boxes is depicted in Figure 9 and willbe described per category.

Disease and transcription

Due to the nature of the assays it is often hard toprecisely pinpoint the molecular consequences ofmutations. For example, if a mutation has beendescribed to cause a disease, it is not clear whetherthis results from diminished ligand binding, modi-fied cofactor specificity, loss of dimerisation, etc.The same problems hold if the observed effect isaltered transcription. For this reason disease andtranscription were combined in one category.Figure 9(a) shows that disease and transcriptionrelated mutations are evenly distributed through-out all boxes, suggesting that this category, asextracted from the NRMD mutant descriptioninformation, is too general to correlate with thedetailed functional role of the residues.

Cofactor binding

Mutations with an effect on cofactor binding aremainly located in box12 (Figure 9(b)). Corollary,35% of the mutations in box12 show a cofactor-related effect. This supports very well the previousfinding that box12 residues are involved in thecofactor binding function of NRs. Actually, onewould expect that mutation of box11 residueswould lead to effects on cofactor binding, but thisis not observed. A potential explanation is the factthat box11 residues are involved in the main func-tion of the protein. They are therefore so importantthat their mutation is likely to lead to a completelyinactive protein. This can explain the very smallnumber of data on box11 residues and the absenceof clear effects on cofactor binding.

Dimerisation

Box22 normally contains residues involved insignal transduction between the main functionalsite and modulator sites. Most box22 mutations inNRs cause a change in dimerisation properties(Figure 9(c)). This suggests that dimerisation couldbe a modulator for the main function: cofactorbinding. Thus, dimerisation may be a prerequisitefor specific cofactor binding. Gampe et al.67 alreadydescribed that helix 12 of PPARg is stabilised by

Figure 9. Distribution of the mutations throughouteach entropy-variability box for various mutation cate-gories. (a) Disease and transcription. (b) Cofactor. (c)Dimerisation. (d) Ligand-binding. (e) No effect.

Figure 8. Number of mutations per category for theCR.



1135

1136

1137

1138

1139

1140

1141

1142

1143

1144

1145

1146

1147

1148

1149

1150

1151

1152

1153

1154

1155

1156

1157

1158

1159

1160

1161

1162

1163

1164

1165

1166

1167

1168

1169

1170

1171

1172

1173

1174

1175

1176

1177

1178

1179

1180

1181

1182

1183

1184

1185

1186

1187

1188

1189

1190

1191

1192

1193

1194

1195

1196

1197

1198

1199

1200

1201

1202

1203

1204

1205

1206

1207

1208

1209

1210

1211

1212

1213

1214

1215

1216

1217

1218

1219

1220

1221

1222

1223

1224

1225

1226

1227

1228

1229

1230

1231

1232

1233

1234

1235

1236

1237

1238

1239

1240

1241

1242

1243

1244

1245

1246

1247

1248

1249

1250

1251

1252

1253

1254

1255

1256

1257

1258

1259

1260

ARTICLE IN PRESS

UNCORRECTED PROOF

interaction with residues in its heterodimeric part-ner RXRa. Since this helix is important in cofactorrecruitment, this interaction could create a differ-ent cofactor-binding environment at the surface ofPPARg than without this interaction. On the otherhand, analysis of the 3D structures reveals a largeseries of box23 and box33 residues that are involvedin dimerisation too. So, despite a low conservationof the dimer interface, the residues involved insignalling between the main functional site andthe dimer/modulator site are also observed inbox22, as expected from previous entropy-variabilityanalysis studies.21

Ligand binding

Mutations that result in an effect on ligand bind-ing, are almost all found in box23 (Figure 9(d)).Corollary, nearly 30% of all mutations describedfor box23 disturb ligand binding. The other effectsobserved upon mutation of box23 residues areprobably the result of either noise in the entropy-variability analysis method, or incomplete experi-mental characterisation of the mutants. Forexample, if a series of mutations are analysed fortheir function in cofactor binding, then thedescribed effect can only be observed on cofactorbinding, even if this effect is the indirect result ofmodified ligand interactions.

No effect

The mutations that show no effect on transcrip-tion, dimerisation or ligand binding are mainlylocated in box23 and box33 (Figure 9(e)). However,for each published silent mutation, we found atleast one other study in which a mutation at theequivalent position in the same or another mol-ecule did show some effect on some functionalaspect.

No mutation

We could not find mutation information for 39residues in the CR. Most of these positions fall inbox33, which confirms the less important nature ofresidues in this box (Figure 10).

It is, of course, not clear whether these residueshave never been mutated yet, or that they havebeen mutated but showed no effects and thereforewere not published.

Problems with mutation analysis

The heterogeneity of mutation data makes theiranalyses difficult. Some scientists study the effecton ligand binding, others on transcription of aspecific gene, and for some mutations, it is onlyknown that they are related to a specific disease.Obviously, effects that are not analysed cannot bereported, leading to under-reporting of effects thatare difficult to observe experimentally. It is highlylikely that for many mutations only the mostsevere effect is reported. If, for example, a mutationinfluences ligand binding in a way that results in aloss of cofactor binding, then this cofactor influ-ence is reported and not the underlying ligand-binding effect. The data shown in Figure 9 shouldtherefore not be seen as a quantitative measure,but merely as qualitative indications.

Concluding Remarks

We have demonstrated that a family-basedapproach in which theoretical and experimentaldata are combined can reveal much informationabout a protein family. This approach works wellbecause residues located at equivalent positions instructures of homologue proteins have similarfunctions, even when the residue types are differ-ent. This family-based approach requires a highquality multiple sequence alignment that onlyincludes residues located at positions that areequivalent in the structures. Insertions and loopsthat can adopt different local conformations indifferent structures should not be part of thestudy and thus not be part of the alignment. Inaddition, the huge amount of sequences availablein today’s publicly available databases, allows usto introduce new multiple sequence alignmentanalysis techniques such as entropy-variabilityanalysis. This new approach was applied to thenuclear receptor protein family to functionally clas-sify residues in the NR LBD. The experimental andstructural data available for these residues agreedwell with this functional classification, whichgives us faith in the predictive value of the classifi-cation scheme for those residues for which experi-mental data are not (yet) available.This classification scheme showed that the main

function of the nuclear receptors is cofactor recruit-ment. In addition, a more interesting result is theprediction of the existence of signal transductionresidues between ligand-binding pocket, thecofactor-binding pocket and the dimer interface.This idea is supported by the specific location ofthese residues in the LBD and by a study of themutation data. Recent publications by Shulmanet al.68 andKnettles et al.69 and the so-called “phantom

Figure 10. Distribution throughout the entropy-vari-ability boxes, of the residue positions for which nomutant information is available (as percentage of thetotal number of positions for which no mutation infor-mation is available).



1261

1262

1263

1264

1265

1266

1267

1268

1269

1270

1271

1272

1273

1274

1275

1276

1277

1278

1279

1280

1281

1282

1283

1284

1285

1286

1287

1288

1289

1290

1291

1292

1293

1294

1295

1296

1297

1298

1299

1300

1301

1302

1303

1304

1305

1306

1307

1308

1309

1310

1311

1312

1313

1314

1315

1316

1317

1318

1319

1320

1321

1322

1323

1324

1325

1326

1327

1328

1329

1330

1331

1332

1333

1334

1335

1336

1337

1338

1339

1340

1341

1342

1343

1344

1345

1346

1347

1348

1349

1350

1351

1352

1353

1354

1355

1356

1357

1358

1359

1360

1361

1362

1363

1364

1365

1366

1367

1368

1369

1370

1371

1372

1373

1374

1375

1376

1377

1378

1379

1380

1381

1382

1383

1384

1385

1386

ARTICLE IN PRESS

UNCORRECTED PROOF

ligand effects”60,61 confirm the existence of signaltransduction processes. In this study, potential sig-nal transduction residues were identified that willbe the subject of future research. In addition, NRscontain more sequence signals than just those thatwe discussed here. Examples are glycosylation,sumoylation, phosphorylation sites etc. However,those signals often are not conserved throughoutthe entire NR family, and consequently are notvery important from a family-wide perspective.

Finally, the huge amount of data used in thisstudy also made it very clear that many more(automated) literature data-extraction, and sequenceand structure analysis methods are required toclassify residues better, and with more detail.Many more data and results from this familybased study are available at the WWW address‡.

Materials and Methods

Structure superposition-based sequence alignment

Sequences were obtained from SWISS-PROT andTrEMBL70, and the NucleaRDB.71 A total of 97 NR LBDstructures were extracted from the PDB62 and groupedin 26 subfamilies according to the classification of theStructural Classification of Proteins (SCOP) structurecomparison resource.72 Structure superposition56 andsequence alignment were performed using WHAT IF.73

The structure-based sequence alignment is a five-stepprocess that is described in great detail at the WWW†,and will only be briefly summarized here. (1) 184 LBDdomains are collected, validated and stored in theWHAT IF internal database. (2) All domains are structur-ally superposed on one (randomly chosen) domain. (3)All positions that superpose within essentially 1.9 A ofthe average Ca in more than 90% of the domains werestructurally averaged. (4) All domains are superposedagain, but now on this set of averaged residues. (5) Resi-dues that superpose on the same averaged residue inthis second superposition round are placed “underneatheach other” in the final, structure-based sequencealignment.

After visual inspection, a few functionally importantresidues at the edge of structurally conserved partswere included to arrive at the final conserved regions(CR), despite the fact that their r.m.s.d. was .1.9 A. Thefunctionally very important helix 12 was includeddespite the fact that it cannot be structurally superposed.However, the alignment is quite reliable due to theLXXLL-like motif in H12. For each of the 26 subfamilies,a sequence profile was derived from the correspondingsuperposed structures. Residue positions not present inthe CR were removed from these subfamily-specificstructure-based sequence profiles. These subfamily-specific profiles were used to align all sequences withmore than 40% sequence identity. This cutoff was basedon the experience that sequences in one subfamily sel-dom differed more than 60%. The representative struc-tures were superposed and the resulting structure-basedalignments were used to combine the family-specificmultiple sequence alignments into one big alignment.An overall profile was generated from this overall align-ment and all sequences that had not yet been includedwere aligned against this profile. The final multiplesequence alignment did not include all available

sequences, but only the sequences that could be alignedwell.

Numbering scheme for structurallyconserved residues

A common residue-numbering scheme was intro-duced to allow for a coherent discussion of the results.Each residue is labelled with a two-digit number pre-ceded by either the number of the helix, or a B or L forbeta strands and loops, respectively. The two digit num-bers are chosen such that the most conserved residue ineach helix gets number 50.

Analysis of the multiple sequence alignment

Entropy-variability analysis was performed asdescribed.21 For each residue position in the alignment,two parameters were defined: the variability and aShannon-type entropy. The variability, Vp, is defined asthe number of different amino acids observed at positionp in the multiple sequence alignment, in at least 0.5% ofall sequences. The Shannon entropy is given by:

Sp ¼ 2X20

i¼1

fpi £ lnðfpiÞ

in which i loops over the 20 amino acid types, p loopsover all positions in the structural alignment, and fpi isthe weighted frequency of residue type i at position p inthe multiple sequence alignment.21 The residue positionswere clustered in five groups according to their entropyand variability. The recipe for the definition of theborders was optimised to generate a maximal correlationbetween box number and residue function in four pro-tein families for which the function of nearly all residueshas been well-established experimentally: globin chains,GPCRs, ras-like proteins, and serine-proteases.21,22 Thisrecipe divides the entropy axis into three parts. Thelower boundary is at 0.4. The upper boundary is halfwaybetween the lower boundary and the highest observedentropy. The first boundary of the variability axis is atthe highest variability in box11, the second boundarywas at the highest variability in box22. According to thisrecipe, box11 residues are involved in the main function,box12 residues are support for the main function, box22

are signal transducing residues between modulator andmain function, box23 are modulator binding residuesand box33 residues have no important function.

Extraction of contact residues from structures

Residues that are in contact with a ligand, cofactor ordimeric partner, were extracted from the X-ray structuresusing WHAT IF.73 A contact was defined as two atomsfor which the distance between the Van der Waals’ sur-faces is less than 1.0 A (ligand) or 0.25 A (cofactor anddimer). A total of 86 LBDs out of 97 PDB files contain aligand and were used to analyse the residues in theLBD that form a contact with the ligand. Ions, cosolventsand compounds that did not bind in the ligand-bindingpocket were excluded from the analysis by means of anexclusion dictionary, minimum size and visual inspec-tion. Four sets of residues, so-called Ligand InteractingGroups (LIGs), were defined based on the frequency ofcontacts with the ligand. LIG1, LIG2, LIG3 and LIG4residues showed to have .1000, .500, .100, and 1–100contacts, respectively, integrated over all structures.



1387

1388

1389

1390

1391

1392

1393

1394

1395

1396

1397

1398

1399

1400

1401

1402

1403

1404

1405

1406

1407

1408

1409

1410

1411

1412

1413

1414

1415

1416

1417

1418

1419

1420

1421

1422

1423

1424

1425

1426

1427

1428

1429

1430

1431

1432

1433

1434

1435

1436

1437

1438

1439

1440

1441

1442

1443

1444

1445

1446

1447

1448

1449

1450

1451

1452

1453

1454

1455

1456

1457

1458

1459

1460

1461

1462

1463

1464

1465

1466

1467

1468

1469

1470

1471

1472

1473

1474

1475

1476

1477

1478

1479

1480

1481

1482

1483

1484

1485

1486

1487

1488

1489

1490

1491

1492

1493

1494

1495

1496

1497

1498

1499

1500

1501

1502

1503

1504

1505

1506

1507

1508

1509

1510

1511

1512

ARTICLE IN PRESS

UNCORRECTED PROOF

Cofactor contacting residues were similarly identifiedusing 16 structures, i.e., 15 coactivators and one corepres-sor structure. Residues involved in homo- or hetero-dimer contacts were extracted using the same methodfor 22 dimer structures. Functional dimers were identi-fied by visual inspection, and the literature.

Nuclear receptor mutation database (NRMD)

Information about mutations in NRs was extractedfrom the mutation data in the NucleaRDB, databasesand lists on the Internet{, and scientific papers.

Most of these mutation data were extracted from fulltext articles (using fully automatic methods66) and col-lected in the NRMD. The database contains 1095mutation entries divided over 41 NRs, of which 375 fallin the CR. Each NRMD entry contains one mutationand, if available, its effect.

Mutation effects

The effects of mutations at residue positions in the CRwere classified according to the description given in theNRMD. The five categories were: (a) disease and tran-scription, (b) cofactor binding, (c) dimerisation, (d)ligand binding, and (e) no effect. Disease and transcrip-tion were grouped in one category, for both it is unclearwhether they are caused by the underlying effects oncofactor, ligand or dimer partner binding. If a mutationled to an effect on more than one function, all wereused in the analysis. If a position was mutated togetherwith other positions (double or triple mutants) the result-ing effect was used for all mutated positions.

Acknowledgements

We thank Unilever and Organon for financialsupport and Henk Stunnenberg for critically read-ing this manuscript.

References

1. Mangelsdorf, D. J., Thummel, C., Beato, M., Herrlich,P., Schutz, G., Umesono, K. et al. (1995). The nuclearreceptor superfamily: the second decade. Cell, 83,835–839.

2. NRNC (1999). A unified nomenclature system for thenuclear receptor superfamily. Cell, 97, 161–163.

3. Kliewer, S. A., Lehmann, J. M. & Willson, T. M.(1999). Orphan nuclear receptors: shifting endocrin-ology into reverse. Science, 284, 757–760.

4. Huang, H. & Tindall, D. J. (2002). The role of theandrogen receptor in prostate cancer. Crit. Rev. Eukaryot.Gene Expr. 12, 193–207.

5. Jones, A. B. (2001). Peroxisome proliferator-activatedreceptor (PPAR) modulators: diabetes and beyond.Med. Res. Rev. 21, 540–552.

6. Bonnelye, E., Kung, V., Laplace, C., Galson, D. L. &Aubin, J. E. (2002). Estrogen receptor-related receptoralpha impinges on the estrogen axis in bone: poten-tial function in osteoporosis. Endocrinology, 143,3658–3670.

7. Hopkins, A. L. & Groom, C. R. (2002). The druggablegenome. Nature Rev. Drug Discov. 1, 727–730.

8. Katzenellenbogen, J. A. & Katzenellenbogen, B. S.(1996). Nuclear hormone receptors: ligand-activatedregulators of transcription and diverse cellresponses. Chem. Biol. 3, 529–536.

9. Danielian, P. S., White, R., Lees, J. A. & Parker, M. G.(1992). Identification of a conserved region requiredfor hormone dependent transcriptional activation bysteroid hormone receptors. EMBO J. 11, 1025–1033.

10. Evans, R. M. (1988). The steroid and thyroid hor-mone receptor superfamily. Science, 240, 889–895.

11. Wurtz, J. M., Bourguet, W., Renaud, J. P., Vivat, V.,Chambon, P., Moras, D. et al. (1996). A canonicalstructure for the ligand-binding domain of nuclearreceptors. Nature Struct. Biol. 3, 206.

12. Watkins, R. E., Wisely, G. B., Moore, L. B., Collins,J. L., Lambert, M. H., Williams, S. P. et al. (2001). Thehuman nuclear xenobiotic receptor PXR: structuraldeterminants of directed promiscuity. Science, 292,2329–2333.

13. Kliewer, S. A., Xu, H. E., Lambert, M. H. & Willson,T. M. (2001). Peroxisome proliferator-activated recep-tors: from genes to physiology. Recent Prog. Horm.Res. 56, 239–263.

14. Bourguet, W., Germain, P. & Gronemeyer, H. (2000).Nuclear receptor ligand-binding domains: three-dimensional structures, molecular interactions andpharmacological implications. Trends Pharmacol. Sci.21, 381–388.

15. Darimont, B. D., Wagner, R. L., Apriletti, J. W., Stall-cup, M. R., Kushner, P. J., Baxter, J. D. et al. (1998).Structure and specificity of nuclear receptor–coacti-vator interactions. Genes Dev. 12, 3343–3356.

16. Nolte, R. T., Wisely, G. B., Westin, S., Cobb, J. E.,Lambert, M. H., Kurokawa, R. et al. (1998). Ligandbinding and co-activator assembly of the peroxisomeproliferator-activated receptor-gamma. Nature, 395,137–143.

17. Brzozowski, A. M., Pike, A. C., Dauter, Z., Hubbard,R. E., Bonn, T., Engstrom, O. et al. (1997). Molecularbasis of agonism and antagonism in the oestrogenreceptor. Nature, 389, 753–758.

18. Greschik, H., Wurtz, J. M., Sanglier, S., Bourguet, W.,van Dorsselaer, A., Moras, D. et al. (2002). Structuraland functional evidence for ligand-independenttranscriptional activation by the estrogen-relatedreceptor 3. Mol. Cell. 9, 303–313.

19. Sablin, E. P., Krylova, I. N., Fletterick, R. J. &Ingraham, H. A. (2003). Structural basis for ligand-independent activation of the orphan nuclear recep-tor LRH-1. Mol. Cell. 11, 1575–1585.

20. Wang, Z., Benoit, G., Liu, J., Prasad, S., Aarnisalo, P.,Liu, X. et al. (2003). Structure and function of Nurr1identifies a class of ligand-independent nuclearreceptors. Nature. 423, 555–560.

21. Oliveira, L., Paiva, P. B., Paiva, A. C. & Vriend, G.(2003). Identification of functionally conserved resi-dues with the use of entropy–variability plots. Pro-teins: Struct. Funct. Genet. 52, 544–552.

22. Oliveira, L., Paiva, P. B., Paiva, A. C. & Vriend, G.(2003). Sequence analysis reveals how G protein-coupled receptors transduce the signal to the Gprotein. Proteins: Struct. Funct. Genet. 52, 553–560.

23. Bork, P. & Gibson, T. J. (1996). Applying motif andprofile searches. Methods Enzymol. 266, 162–184.

24. Lichtarge, O., Bourne, H. R. & Cohen, F. E. (1996). Anevolutionary trace method defines binding surfacescommon to protein families. J. Mol. Biol. 257,342–358.

25. Mirny, L. & Shakhnovich, E. (2001). Evolutionary



1513

1514

1515

1516

1517

1518

1519

1520

1521

1522

1523

1524

1525

1526

1527

1528

1529

1530

1531

1532

1533

1534

1535

1536

1537

1538

1539

1540

1541

1542

1543

1544

1545

1546

1547

1548

1549

1550

1551

1552

1553

1554

1555

1556

1557

1558

1559

1560

1561

1562

1563

1564

1565

1566

1567

1568

1569

1570

1571

1572

1573

1574

1575

1576

1577

1578

1579

1580

1581

1582

1583

1584

1585

1586

1587

1588

1589

1590

1591

1592

1593

1594

1595

1596

1597

1598

1599

1600

1601

1602

1603

1604

1605

1606

1607

1608

1609

1610

1611

1612

1613

1614

1615

1616

1617

1618

1619

1620

1621

1622

1623

1624

1625

1626

1627

1628

1629

1630

1631

1632

1633

1634

1635

1636

1637

1638

ARTICLE IN PRESS

UNCORRECTED PROOF

conservation of the folding nucleus. J. Mol. Biol. 308,123–129.

26. Pei, J. & Grishin, N. V. (2001). AL2CO: calculation ofpositional conservation in a protein sequence align-ment. Bioinformatics, 17, 700–712.

27. Shenkin, P. S., Erman, B. & Mastrandrea, L. D. (1991).Information-theoretical entropy as a measure ofsequence variability. Proteins: Struct. Funct. Genet. 11,297–313.

28. Zuckerkandl, E. & Pauling, L. (1965). Evolutionarydivergence and convergence in proteins Evolving Genesand Proteins, Academic Press, New York pp. 97–166..

29. Mirny, L. A. & Shakhnovich, E. I. (1999). Universallyconserved positions in protein folds: reading evolu-tionary signals about stability, folding kinetics andfunction. J. Mol. Biol. 291, 177–196.

30. Russell, R. B., Sasieni, P. D. & Sternberg, M. J. (1998).Supersites within superfolds. Binding site similarityin the absence of homology. J. Mol. Biol. 282, 903–918.

31. Kuntz, I. D., Blaney, J. M., Oatley, S. J., Langridge, R.& Ferrin, T. E. (1982). A geometric approach tomacromolecule–ligand interactions. J. Mol. Biol. 161,269–288.

32. DesJarlais, R. L., Sheridan, R. P., Seibel, G. L., Dixon,J. S., Kuntz, I. D. & Venkataraghavan, R. (1988).Using shape complementarity as an initial screen indesigning ligands for a receptor binding site ofknown three-dimensional structure. J. Med. Chem.31, 722–729.

33. Honig, B. & Nicholls, A. (1995). Classical electro-statics in biology and chemistry. Science, 268,1144–1149.

34. Miranker, A. & Karplus, M. (1991). Functionalitymaps of binding sites: a multiple copy simultaneoussearch method. Proteins: Struct. Funct. Genet. 11,29–34.

35. Lamb, M. L. & Jorgensen, W. L. (1997). Compu-tational approaches to molecular recognition. Curr.Opin. Chem. Biol. 1, 449–457.

36. Wang, W., Donini, O., Reyes, C. M. & Kollman, P. A.(2001). Biomolecular simulations: recent develop-ments in force fields, simulations of enzyme catalysis,protein–ligand, protein–protein, and protein–nucleicacid noncovalent interactions. Annu. Rev. Biophys. Bio-mol. Struct. 30, 211–243.

37. Casari, G., Sander, C. & Valencia, A. (1995). Amethod to predict functional residues in proteins.Nature Struct. Biol. 2, 171–178.

38. Jones, S. & Thornton, J. M. (1997). Prediction ofprotein-protein interaction sites using patch analysis.J. Mol. Biol. 272, 133–143.

39. Pietrokovski, S., Henikoff, J. G. & Henikoff, S. (1996).The Blocks database–a system for protein classifi-cation. Nucl. Acids Res. 24, 197–200.

40. Shatsky, M., Nussinov, R. & Wolfson, H. J. (2002).Flexible protein alignment and hinge detection. Pro-teins: Struct. Funct. Genet. 48, 242–256.

41. Sali, A., Overington, J. P., Johnson, M. S. & Blundell,T. L. (1990). From comparisons of protein sequencesand structures to protein modelling and design.Trends Biochem. Sci. 15, 235–240.

42. Innis, C. A., Shi, J. & Blundell, T. L. (2000). Evolution-ary trace analysis of TGF-beta and related growthfactors: implications for site-directed mutagenesis.Protein Eng. 13, 839–847.

43. Landgraf, R., Xenarios, I. & Eisenberg, D. (2001).Three-dimensional cluster analysis identifies inter-

faces and functional residue clusters in proteins.J. Mol. Biol. 307, 1487–1502.

44. Nobbs, C. L., Watson, H. C. & Kendrew, J. C. (1966).Structure of deoxymyoglobin: a crystallographicstudy. Nature, 209, 339–341.

45. Perutz, M. F. (1969). The Croonian Lecture, 1968. Thehaemoglobin molecule. Proc. R. Soc. Lond. ser. B: Biol.Sci. 173, 113–140.

46. Perutz, M. F. (1970). Stereochemistry of cooperativeeffects in haemoglobin. Nature, 228, 726–739.

47. Royer, W. E., Jr, Hendrickson, W. A. & Chiancone, E.(1990). Structural transitions upon ligand binding ina cooperative dimeric hemoglobin. Science, 249,518–521.

48. Pai, E. F., Kabsch, W., Krengel, U., Holmes, K. C.,John, J. & Wittinghofer, A. (1989). Structure of theguanine-nucleotide-binding domain of the Ha-rasoncogene product p21 in the triphosphate confor-mation. Nature, 341, 209–214.

49. Pai, E. F., Krengel, U., Petsko, G. A., Goody, R. S.,Kabsch, W. & Wittinghofer, A. (1990). Refined crystalstructure of the triphosphate conformation of H-rasp21 at 1.35 A resolution: implications for the mech-anism of GTP hydrolysis. EMBO J. 9, 2351–2359.

50. Takai, Y., Sasaki, T. & Matozaki, T. (2001). Small GTP-binding proteins. Physiol. Rev. 81, 153–208.

51. Huang, L., Hofer, F., Martin, G. S. & Kim, S. H.(1998). Structural basis for the interaction of Raswith RalGDS. Nature Struct. Biol. 5, 422–426.

52. Pacold, M. E., Suire, S., Perisic, O., Lara-Gonzalez, S.,Davis, C. T., Walker, E. H. et al. (2000). Crystal struc-ture and functional analysis of Ras binding to itseffector phosphoinositide 3-kinase gamma. Cell, 103,931–943.

53. Ruhlmann, A., Kukla, D., Schwager, P., Bartels, K. &Huber, R. (1973). Structure of the complex formedby bovine trypsin and bovine pancreatic trypsininhibitor. Crystal structure determination and stereo-chemistry of the contact region. J. Mol. Biol. 77,417–436.

54. Huber, R., Kukla, D., Bode, W., Schwager, P., Bartels,K., Deisenhofer, J. et al. (1974). Structure of the com-plex formed by bovine trypsin and bovine pancreatictrypsin inhibitor. II. Crystallographic refinement at1.9 A resolution. J. Mol. Biol. 89, 73–101.

55. Van Durme, J. J., Bettler, E., Folkertsma, S., Horn, F. &Vriend, G. (2003). NRMD: Nuclear Receptor Muta-tion Database. Nucl. Acids Res. 31, 331–333.

56. Vriend, G. & Sander, C. (1991). Detection of commonthree-dimensional substructures in proteins. Proteins:Struct. Funct. Genet. 11, 52–58.

57. Gampe, R. T., Jr, Montana, V. G., Lambert, M. H.,Wisely, G. B., Milburn, M. V. & Xu, H. E. (2000).Structural basis for autorepression of retinoid Xreceptor by tetramer formation and the AF-2 helix.Genes Dev. 14, 2229–2241.

58. Enmark, E. & Gustafsson, J. A. (2001). Comparingnuclear receptors in worms, flies and humans. TrendsPharmacol. Sci. 22, 611–615.

59. Billas, I. M., Iwema, T., Garnier, J. M., Mitschler, A.,Rochel, N. & Moras, D. (2003). Structural adapt-ability in the ligand-binding pocket of the ecdysonehormone receptor. Nature, 426, 91–96.

60. Schulman, I. G., Li, C., Schwabe, J. W. & Evans, R. M.(1997). The phantom ligand effect: allosteric controlof transcription by the retinoid X receptor. GenesDev. 11, 299–308.

61. Willy, P. J. & Mangelsdorf, D. J. (1997). Uniquerequirements for retinoid-dependent transcriptional



1639

1640

1641

1642

1643

1644

1645

1646

1647

1648

1649

1650

1651

1652

1653

1654

1655

1656

1657

1658

1659

1660

1661

1662

1663

1664

1665

1666

1667

1668

1669

1670

1671

1672

1673

1674

1675

1676

1677

1678

1679

1680

1681

1682

1683

1684

1685

1686

1687

1688

1689

1690

1691

1692

1693

1694

1695

1696

1697

1698

1699

1700

1701

1702

1703

1704

1705

1706

1707

1708

1709

1710

1711

1712

1713

1714

1715

1716

1717

1718

1719

1720

1721

1722

1723

1724

1725

1726

1727

1728

1729

1730

1731

1732

1733

1734

1735

1736

1737

1738

1739

1740

1741

1742

1743

1744

1745

1746

1747

1748

1749

1750

1751

1752

1753

1754

1755

1756

1757

1758

1759

1760

1761

1762

1763

1764

ARTICLE IN PRESS

UNCORRECTED PROOF

activation by the orphan receptor LXR. Genes Dev. 11,289–298.

62. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G.,Bhat, T. N., Weissig, H. et al. (2000). The ProteinData Bank. Nucl. Acids Res. 28, 235–242.

63. Bourguet, W., Vivat, V., Wurtz, J. M., Chambon, P.,Gronemeyer, H. & Moras, D. (2000). Crystal structureof a heterodimeric complex of RAR and RXR ligand-binding domains. Mol. Cell, 5, 289–298.

64. Alen, P., Claessens, F., Schoenmakers, E., Swinnen,J. V., Verhoeven, G., Rombauts, W. et al. (1999). Inter-action of the putative androgen receptor-specificcoactivator ARA70/ELE1alpha with multiple steroidreceptors and identification of an internally deletedELE1beta isoform. Mol. Endocrinol. 13, 117–128.

65. Whitfield, G. K., Hsieh, J. C., Nakajima, S., MacDonald,P. N., Thompson, P. D., Jurutka, P. W. et al. (1995). Ahighly conserved region in the hormone-bindingdomain of the human vitamin D receptor containsresidues vital for heterodimerization with retinoid Xreceptor and for transcriptional activation. Mol.Endocrinol. 9, 1166–1179.

66. Horn, F., Lau, A. L. & Cohen, F. E. (2004). Automatedextraction of mutation data from the literature: appli-cation of MuteXt to G protein-coupled receptorsand nuclear hormone receptors. Bioinformatics, 20,557–568.

67. Gampe, R. T., Jr, Montana, V. G., Lambert, M. H.,Miller, A. B., Bledsoe, R. K., Milburn, M. V. et al.(2000). Asymmetry in the PPARgamma/RXRalphacrystal structure reveals the molecular basis of het-erodimerization among nuclear receptors. Mol. Cell,5, 545–555.

68. Shulman, A. I., Larson, C., Mangelsdorf, D. J. &Ranganathan, R. (2004). Structural determinants ofallosteric ligand activation in RXR heterodimers.Cell, 116, 417–429.

69. Nettles, K. W., Sun, J., Radek, J. T., Sheng, S.,Rodriguez, A. L., Katzenellenbogen, J. A. et al.(2004). Allosteric control of ligand selectivitybetween estrogen receptors alpha and beta: impli-

cations for other nuclear receptors. Mol. Cell. 13,317–327.

70. Boeckmann, B., Bairoch, A., Apweiler, R., Blatter,M. C.,Estreicher, A., Gasteiger, E. et al. (2003). The SWISS-PROT protein knowledgebase and its supplementTrEMBL in 2003. Nucl. Acids Res. 31, 365–370.

71. Horn, F., Vriend, G. & Cohen, F. E. (2001). Collectingand harvesting biological data: the GPCRDB andNucleaRDB information systems. Nucl. Acids Res. 29,346–349.

72. Murzin, A. G., Brenner, S. E., Hubbard, T. & Chothia,C. (1995). SCOP: a structural classification of proteinsdatabase for the investigation of sequences andstructures. J. Mol. Biol. 247, 536–540.

73. Vriend, 7.3. (1990). WHAT IF: a molecular modelingand drug design program. J. Mol. Graph. 8, 52–56.see also p. 29..

74. Bourguet, W., Ruff, M., Chambon, P., Gronemeyer, H.& Moras, D. (1995). Crystal structure of the ligand-binding domain of the human nuclear receptorRXR-alpha. Nature, 375, 377–382.

75. Shiau, A. K., Barstad, D., Loria, P. M., Cheng, L.,Kushner, P. J., Agard, D. A. et al. (1998). The struc-tural basis of estrogen receptor/coactivator recog-nition and the antagonism of this interaction bytamoxifen. Cell, 95, 927–937.

76. Xu, H. E., Stanley, T. B., Montana, V. G., Lambert,M. H., Shearer, B. G., Cobb, J. E. et al. (2002). Struc-tural basis for antagonist-mediated recruitment ofnuclear co-repressors by PPARalpha. Nature, 415,813–817.

77. Klaholz, B. P., Mitschler, A. & Moras, D. (2000).Structural basis for isotype selectivity of thehuman retinoic acid nuclear receptor. J. Mol. Biol.302, 155–170.

78. Sack, J. S., Kish, K. F., Wang, C., Attar, R. M., Kiefer,S. E., An, Y. et al. (2001). Crystallographic structuresof the ligand-binding domains of the androgenreceptor and its T877A mutant complexed with thenatural agonist dihydrotestosterone. Proc. Natl Acad.Sci. USA, 98, 4904–4909.

Edited by J. Thornton

(Received 29 December 2003; received in revised form 5 May 2004; accepted 18 May 2004)



1765

1766

1767

1768

1769

1770

1771

1772

1773

1774

1775

1776

1777

1778

1779

1780

1781

1782

1783

1784

1785

1786

1787

1788

1789

1790

1791

1792

1793

1794

1795

1796

1797

1798

1799

1800

1801

1802

1803

1804

1805

1806

1807

1808

1809

1810

1811

1812

1813

1814

1815

1816

1817

1818

1819

1820

1821

1822

1823

1824

1825

1826

1827

1828

1829

1830

1831

1832

1833

1834

1835

1836

1837

1838

1839

1840

1841

1842

1843

1844

1845

1846

1847

1848

1849

1850

1851

1852

1853

1854

1855

1856

1857

1858

1859

1860

1861

1862

1863

1864

1865

1866

1867

1868

1869

1870

1871

1872

1873

1874

1875

1876

1877

1878

1879

1880

1881

1882

1883

1884

1885

1886

1887

1888

1889

1890

ARTICLE IN PRESS

A Family-based Approach Reveals the Function of Residues ......UNCORRECTED PROOF A Family-based...

Documents

Transcript of A Family-based Approach Reveals the Function of Residues ......UNCORRECTED PROOF A Family-based...