A Family-based Approach Reveals the Function of Residues ......UNCORRECTED PROOF A Family-based...
Transcript of 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.
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
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.
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
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
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/
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
4 Role of Amino Acids in NR-LBDs
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
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
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
Role of Amino Acids in NR-LBDs 5
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
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
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
6 Role of Amino Acids in NR-LBDs
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.
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
Role of Amino Acids in NR-LBDs 7
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
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
8 Role of Amino Acids in NR-LBDs
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
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
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
Role of Amino Acids in NR-LBDs 9
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
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.
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
10 Role of Amino Acids in NR-LBDs
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).
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
Role of Amino Acids in NR-LBDs 11
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.
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
12 Role of Amino Acids in NR-LBDs
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
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
Role of Amino Acids in NR-LBDs 13
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
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
14 Role of Amino Acids in NR-LBDs
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)
YJMBI—56346—22/6/2004—KAREN.HADLEY—109475 – pp. 1–15/TL
Role of Amino Acids in NR-LBDs 15
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