serotonin receptors science 2013

8/12/2019 Serotonin Receptors Science 2013

1/6

DOI: 10.1126/science.1232807, 610 (2013);340Scienceet al.Chong Wang

Structural Basis for Molecular Recognition at Serotonin Receptors

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by, you can order high-quality copies for yourIf you wish to distribute this article to others

here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

):August 19, 2013www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/340/6132/610.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2013/03/20/science.1232807.DC1.htmlcan be found at:Supporting Online Material

http://www.sciencemag.org/content/340/6132/610.full.html#relatedfound at:

can berelated to this articleA list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/340/6132/610.full.html#ref-list-1

, 13 of which can be accessed free:cites 47 articlesThis article

http://www.sciencemag.org/content/340/6132/610.full.html#related-urls3 articles hosted by HighWire Press; see:cited byThis article has been

http://www.sciencemag.org/cgi/collection/biochemBiochemistry

subject collections:This article appears in the following

registered trademark of AAAS.is aScience2013 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrighAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005.(print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by thScience
http://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/content/340/6132/610.full.htmlhttp://www.sciencemag.org/content/340/6132/610.full.htmlhttp://www.sciencemag.org/content/340/6132/610.full.html#relatedhttp://www.sciencemag.org/content/340/6132/610.full.html#relatedhttp://www.sciencemag.org/content/340/6132/610.full.html#ref-list-1http://www.sciencemag.org/content/340/6132/610.full.html#ref-list-1http://www.sciencemag.org/content/340/6132/610.full.html#ref-list-1http://www.sciencemag.org/content/340/6132/610.full.html#ref-list-1http://www.sciencemag.org/content/340/6132/610.full.html#related-urlshttp://www.sciencemag.org/content/340/6132/610.full.html#related-urlshttp://www.sciencemag.org/content/340/6132/610.full.html#related-urlshttp://www.sciencemag.org/content/340/6132/610.full.html#related-urlshttp://www.sciencemag.org/cgi/collection/biochemhttp://www.sciencemag.org/cgi/collection/biochemhttp://www.sciencemag.org/cgi/collection/biochemhttp://www.sciencemag.org/content/340/6132/610.full.html#related-urlshttp://www.sciencemag.org/content/340/6132/610.full.html#ref-list-1http://www.sciencemag.org/content/340/6132/610.full.html#relatedhttp://www.sciencemag.org/content/340/6132/610.full.htmlhttp://www.sciencemag.org/about/permissions.dtlhttp://www.sciencemag.org/about/permissions.dtlhttp://oascentral.sciencemag.org/RealMedia/ads/click_lx.ads/sciencemag/cgi/reprint/L22/373986199/Top1/AAAS/PDF-R-and-D-Systems-Science-130301/SCad2_Science.com_week2.raw/1?x


2/6

24. V. R. Baker, D. Nummedal, Eds., The Channeled

Scabland; a Guide to the Geomorphology of the

Columbia Basin (NASA, Washington, DC, 1978).

Acknowledgments:C. Chazen and T. Watters provided very

helpful reviews of earlier versions of the paper. Comments

and suggestions by four anonymous referees were extremely

beneficial. Thanks are also extended to P. Russell for his

technical assistance. Funding for this work was provided by

the NASA Mars Reconnaissance Orbiter Project. The SHARAD

data and MOLA altimetry are available through NASAsPlanetary Data system.

Supplementary Materialswww.sciencemag.org/cgi/content/full/science.1234787/DC1

Materials and Methods

Supplementary Text

Figs. S1 to S4

Table S1

Reference (25)

4 January 2013; accepted 18 February 2013

Published online 7 March 2013;

10.1126/science.1234787

Structural Basis for MolecularRecognition at Serotonin ReceptorsChong Wang,1*Yi Jiang,1,2*Jinming Ma,1,3*Huixian Wu,1 Daniel Wacker,1 Vsevolod Katritch,1

Gye Won Han,1 Wei Liu,1 Xi-Ping Huang,4 Eyal Vardy,4 John D. McCorvy,4 Xiang Gao,3

X. Edward Zhou,3 Karsten Melcher,3 Chenghai Zhang,2,3 Fang Bai,5 Huaiyu Yang,6 Linlin Yang,6

Hualiang Jiang,6 Bryan L. Roth,4 Vadim Cherezov,1 Raymond C. Stevens,1 H. Eric Xu2,3

Serotonin or 5-hydroxytryptamine (5-HT) regulates a wide spectrum of human physiology through the5-HT receptor family. We report the crystal structures of the human 5-HT1BG proteincoupled receptorbound to the agonist antimigraine medications ergotamine and dihydroergotamine. The structures revealsimilar binding modes for these ligands, which occupy the orthosteric pocket and an extended binding

pocket close to the extracellular loops. The orthosteric pocket is formed by residues conserved in the5-HT receptor family, clarifying the family-wide agonist activity of 5-HT. Compared with the structureof the 5-HT2Breceptor, the 5-HT1Breceptor displays a 3 angstrom outward shift at the extracellular endof helix V, resulting in a more open extended pocket that explains subtype selectivity. Together withdocking and mutagenesis studies, these structures provide a comprehensive structural basis forunderstanding receptor-ligand interactions and designing subtype-selective serotonergic drugs.

The neuromodulator serotonin [5-hydroxy-tryptamine (5-HT)] is essential for di-verse functions at nearly every organ

system in the human body (14). The activity of5-HT is mediated through activation of mem-bers of a large family of 5-HT receptor pro-teins that can be grouped into seven subfamilies

(5-HT1-7) on the basis of sequence homologyand signaling mechanisms (5). Except for the5-HT3receptor, which is a ligand-gated ion chan-nel, the other 12 members are heterotrimericguanine nucleotide binding protein (G protein)coupled receptors (GPCRs). The serotonergicsystem is a target of many widely prescribeddrugs, including atypical antipsychotics, antimi-

graine medications, anxiolytics, and antidepres-sants (1), and the recently approved antiobesitymedication lorcaserin (6, 7). However, clinicaluse of several serotonergic drugs caused unex-pected side effects arising from off-target inter-actions with 5-HT receptor subtypes and relatedreceptors for biogenic amine (1, 4, 8, 9).

The 5-HT1B receptor couples to G proteinalpha subunits Gior Goand is widely expressedin the brain and the cardiovascular system. Inthe CNS, the 5-HT1B receptor functions as aninhibitory presynaptic receptor to modulate therelease of 5-HT and many other neurotrans-mitters (1, 2). The 5-HT1Breceptor is a primarymolecular target for the antimigraine drugs ergot-amine (ERG) and dihydroergotamine (DHE),which are efficacious 5-HT1Breceptor agonists(10). Off-target activation of the related 5-HT2Breceptor is responsible for the valvulopathic ac-tivity of many approved drugs and is the mainreason for their withdrawal (912). We reporttwo crystal structures of the human 5-HT1Bre-ceptor bound to the full agonists, ERG and DHE(tables S1 and S2). Comparison with the structureof the human 5-HT2B receptor bound to ERG(13) reveals critical structural determinants forligand recognition and subtype selectivity andprovides a structural rationale for designing saferand more effective serotonergic drugs.

Crystallization studies of the 5-HT1Breceptorwere done with engineered constructs, 5-HT1B-1and 5-HT1B-2 (14), which crystallized with ERGand DHE at resolutions of 2.7 and 2.8 ,respectively. Due to the high similarity between

these two structures (fig. S2), for brevity we

cus on the structure of the 5-HT1B-1/ERG coplex for analysis and discussion of key structufeatures for ligand recognition and selectivity5-HT1Bversus 5-HT2B receptors.

The main fold of the 5-HT1Breceptor cosists of a canonical seven-transmembrane (7TMa-helical bundle (Fig. 1A). The extracellular lo2 (ECL2) that partially covers the ligand bining pocket is stabilized by a C1223.25-C199EC

disulfide bond, highly conserved in GPCRs. Pof the N terminus folds on top of the bindipocket, where Y40 forms hydrogen-bond inactions with ligand binding residue D3527.36 (S5) (15, 16). This feature suggests that the N t

minus could have a role in ligand recognitionthe 5-HT1B receptor by interacting with residuwithin the binding pocket.

The 5-HT1B/ERG complex structure reveaa large ligand binding cavity defined by redues from helices III, V, VI, VII, and ECLcomprising an orthosteric pocket embedded dein the 7TM core and an extended binding pocclose to the extracellular entrance (Fig. 1). ERadopts a binding mode with the ergoline rsystem occupying the orthosteric binding pocand the cyclic tripeptide moiety bound to the uper extended binding pocket (Fig. 2C). In orthosteric pocket, the ergoline scaffold is

chored through the salt-bridge interaction tween its positively charged nitrogen and tcarboxylate of D1293.32, which is fully conservin 5-HT and other monoamine receptors. Tside chain of D1293.32 is further stabilized bhydrogen bond to the hydroxyl of Y3597.43. Schains of C1333.36, I1303.33, W3276.48, F3306

and F3316.52 form a narrow hydrophobic clewhich packs tightly against the nearly plaergoline ring system. In addition, the indole Nhydrogen forms a hydrogen bond with T134(Fig. 2A). Comparison with the ERG-bou5-HT2B receptor structure revealed that the thosteric binding pockets in the two receptors very similar, with the key interactions conserv(Fig. 2, A, D, and E). The only difference is oserved in the region where residues from heV contact ERG. Due to the lack of a side chaat G2215.42, the side chain of F2175.38 in the 5-HTreceptor reaches into the ligand binding pocand packs on top of the ERG indole ring; comparison, the corresponding interactionthe 5-HT1B receptor occurs between the sichain of S2125.42 and ERG, whereas Y2085

does not interact with the ligand due to the oward shift of helix V (Fig. 2, A and D). Sustantial differences are observed in the extend

1Department of Integrative Structural and Computational Biol-ogy, The Scripps Research Institute, 10550 North Torrey PinesRoad, LaJolla, CA92037, USA.2Van AndelResearch Institute/Shanghai Institute of Materia Medica Center, CAS-Key Lab-oratory of Receptor Research, Shanghai Institute of MateriaMedica, Chinese Academy of Sciences, Shanghai 201203,

China.3

Laboratory of Structural Sciences, Van Andel ResearchInstitute, 333 Bostwick Avenue Northeast, Grand Rapids, MI49503, USA. 4National Institute of Mental Health PsychoactiveDrug Screening Program, Department of Pharmacology andDivision of Chemical Biology and Medicinal Chemistry, Univer-sity of North Carolina Chapel Hill Medical School, Chapel Hill,NC 27599, USA. 5Department of Engineering Mechanics, StateKey Laboratory of Technology, Faculty of Chemical, Environ-mental, and BiologicalScience and Technology,Dalian Universityof Technology, Dalian 116023, China. 6State Key Laboratory ofDrug Research, Shanghai Institute of Materia Medica, ChineseAcademy of Sciences, Shanghai 201203, China.

*These authors contributed equally to this work.Corresponding author. E-mail: [email protected] (H.E.X);[email protected] (R.C.S.)

3 MAY 2013 VOL 340 SCIENCE www.sciencemag.org10

REPORTS


3/6

binding pockets. The extended binding pocket ofthe 5-HT1B receptor is broader than that of the5-HT2Breceptor (Fig. 1, B, C, and D) due to the3.0 outward shift of the top of helix V. More-over, in contrast to the relatively bulky M2185.39

of the 5-HT2Breceptor, the corresponding residueof the 5-HT1B receptor is a smaller threonine,which results in a further expansion of this pocket(Fig. 2B). Despite these differences in contactresidues between the two subtypes, the cyclic

tripeptide moiety of ERG maintains a similar over-all orientation relative to the ergoline moiety,likely stabilized by an intramolecular hydrogenbond (Fig. 2, D and E). The conformations ofthe phenyl group of the ERG cyclic tripeptidedo differ. It contacts L3476.58 and V3486.59 in the5-HT2B receptor, whereas in the 5-HT1Brecep-tor, it rotates to occupy a cavity close to T2095.39

at helix V. Correspondingly, M3376.58 turns intothe pocket to contact with the phenyl ring of theligand in the 5-HT1Breceptor. As shown in theaccompanying paper, these differences in ERGinteractions correlate with different functionalstates. At the 5-HT1Breceptor, ERG causes full

activation, whereas ERG induces intermediate Gprotein activation andb-arrestin biased signalingat the 5-HT2Breceptor (13).

Mutations of several residues in the ortho-steric binding site of the 5-HT1Breceptor, includ-ing the conserved D1293.32, abolished the bindingof the radioligand lysergic acid diethylamide(LSD) (table S4). The prototypical hallucinogenLSD has the same ergoline moiety as ERG, andthese interactions within the orthosteric pocket

appear to be the main driving force for bindingof ergolines. In contrast, none of the tested mu-tations of the extended binding pocket residuesconsiderably reduced the binding affinity of ERG,suggesting that the cyclic tripeptide moiety ofERG is accommodated without contributing sub-stantially to the binding affinity. This extendedbinding site partially overlaps with the sites in-ferred in interaction with allosteric modulators inthe M2muscarinic acetylcholine receptor (17, 18),suggesting its potential role in mediating allo-steric modulations at 5-HT receptors (13).

Alignment of all human 5-HT GPCR se-quences shows that the residues in the ortho-

steric binding pockets are much more conservthan those in the extended binding pockets (fS7). This likely reflects an evolutionary prsure to maintain the structure of the orthostebinding pocket for recognition of the endoenous ligand 5-HT (19). Molecular docking5-HT into the orthosteric binding pocket of 5-HT1B and 5-HT2B receptors revealed imptant residues involved in the recognition of 5-H(Fig. 3A), many of which have been previou

implicated by site-directed mutagenesis and mlecular modeling studies (20, 21). 5-HT sharecommon chemical substructures with ergolin(fig. S4) and recognizes the 5-HT receptors isimilar manner as ERG and DHE. D3.32 formsalt bridge with the positively charged amino groof 5-HT, whereas T3.37, which is highly coserved in 5-HT receptors and most other amnergic GPCRs, forms a hydrogen bond with indole N-H hydrogen. This hydrogen bond apears to be important for the recognition 5-HT by 5-HT1Band 5-HT2Breceptors becaumutating T3.37 to alanine reduces the affinity5-HT at both receptors by a factor of more th

I

IIIIIIVV

VI

VII

VIII

A C 5-HT1Breceptor

5-HT2BreceptorD

V

VI

VII

IIIIV

B

3.0

90o

I

IIIII

IV

V

VI

VII

VIII

N-terminus

ICL2

D352

Y40

S-S

ICL1

Y157 L80

Fig. 1. Overall architecture of the 5-HT1B receptor bound to ERG andcomparison of the ligand binding pocket shapes of the 5-HT1Breceptorand the 5-HT2Breceptor.(A) The 5-HT1Breceptor is shown as a light bluecolored ribbon cartoon; the N terminus, ICL1, and ICL2 are highlighted inyellow. Ligand ERG is colored magenta. The disulfide bond between C122and C199 is shown as orange sticks. The Y40 and D3527.36 side chains, whichmediate the interaction between the N terminus and the ligand binding pocket,are shown as green sticks, with a dashed line indicating the hydrogen bondinteraction. L80ICL1 and Y157ICL2 interact with residues of the 7TM bundle sta-

bilizing the local structures of ICL1 and ICL2, respectively. (B) Shown at the bottis the superposition of the 5-HT1B/ERG structure (light blue) and the 5-HT2B/Estructure (white). The ligands are colored magenta for the 5-HT1B receptor and grfor the 5-HT2Breceptor. The top panel shows an extracellular view of the ligabinding sites. The arrow indicates a 3.0 shift (distance measured between thecarbons of T2095.39 in the 5-HT1Breceptor and M218

5.39 in the 5-HT2Breceptat the extracellular end of helix V. (CandD) The surface representation of ligand binding pockets of the 5-HT1B receptor and the 5-HT2B receptor shown in transparent pink and transparent green, respectively.

www.sciencemag.org SCIENCE VOL 340 3 MAY 2013

REP


4/6

10 (table S4). The indole ring points toward res-idues on helix V: S2125.42 and A2165.46 at the5-HT1Breceptor, G221

5.42 and A2255.46 at the5-HT2B receptor. This ligand-receptor interfaceis less polar in the 5-HT receptor family com-pared with those of other biogenic amine recep-tors (table S5), thereby perfectly matching theproperty of the 5-HT indole head group, whichis less polar than those of other aminergic recep-tor native agonists, such as epinephrine, dopa-

mine, and histamine.LSD promiscuously binds to 5-HT receptors:

It has agonist activity at most 5-HT receptors(table S6), whereas it is a potent antagonist at5-HT7Areceptor (13). Docking of LSD into the5-HT1B and 5-HT2B receptor structures sug-

gests that binding of the LSD ergoline moietyin the orthosteric binding pocket is essentiallyidentical to that of the ERG ergoline moiety(Fig. 3B). Many single point mutations withinthe orthosteric binding pocket reduce or abolishthe binding of LSD at both 5-HT1Band 5-HT2Breceptors (table S4). The conservation of the ortho-steric binding pocket provides an atomic-level ex-planation for the observed promiscuous interactionsbetween 5-HT receptors and drugs like LSD.

Triptans are 5-HT analogs that are amongthe most frequently prescribed antimigraine med-ications that act primarily through 5-HT1Band5-HT1D receptors (1, 2, 22). A common struc-tural feature of triptans is the large substitutiongroup at the 5position of the indole ring. Func-

tional assays indicate that triptans act as potagonists of the 5-HT1Breceptor but not theHT2B receptor (table S7). To address the strutural determinants of this selectivity, we performdocking simulations at both 5-HT1Band 5-HTreceptors (fig. S8). Whereas triptans were waccommodated in the 5-HT1Breceptor bindpocket, the narrower extended binding pocin the helix V region of the 5-HT2B recepforced the ligands to adopt unfavorable po

tions (Fig. 3C and fig. S8C). Among the testtriptans, donitriptan and eletriptan showed ratively higher potency in Gq-mediated signalat the 5-HT2B receptor (table S7). These twtriptans have longer and more flexible linkbetween 5substituents and the indole ring, e

Fig. 2. Comparison of ligand-receptor interactions in 5-HT1B/ERG and5-HT2B/ERG structures.(AtoC) Superposition of the ligand binding pocketsof 5-HT1Breceptor (light blue) and 5-HT2B receptor (white). The carbonsof ligand ERG in the 5-HT1Band the 5-HT2Breceptor structures are shownas magenta and green, respectively. (A) Residues forming the orthostericbinding sites are shown as sticks and labeled in blue for 5-HT1Breceptorand black for 5-HT2Breceptor. The salt-bridge interactions between D

3.32

and ERGas well as the hydrogen bond interactions between T3.37 andERG, and between Y7.43 and D3.32are shown as red dashed lines for5-HT1Breceptor and green dashed lines for 5-HT2Breceptor. (B) Substantialdifferences in the extended ligand binding pockets result in different

conformations of the ligand ERG. (C) The overall binding pockets with torthosteric and the extended ligand binding sites shaded red and blurespectively. (D and E) Diagram representation of ligand interactionsthe binding pockets of 5-HT1B and 5-HT2B receptors, respectively. Intmolecular hydrogen bonds within ERG are indicated as dashed linResidues in the orthosteric binding pockets are shown in red boxes, aextended binding pocket residues are shown in blue boxes. The hydrogbond interaction between T3.37 and ERG and the salt-bridge interactiobetween D3.32 and ERG are indicated by red dashed lines. In (A), (B), a(D), Y2085.38 and P3386.59 of the 5-HT1Breceptor, which do not interwith ERG, are labeled in gray.

3 MAY 2013 VOL 340 SCIENCE www.sciencemag.org12

REPORTS


5/6

abling them to fit better in the narrow bindingpocket of the 5-HT2Breceptor. The M218

5.39Amutation of the 5-HT2Breceptor, which increasesthe space in the binding pocket, substantially en-hanced the potency of donitriptan and eletriptan(fig. S9), although their potency was still lowerthan that at the 5-HT1Breceptor. Thus, the broaderopening near the extracellular end of helix V inthe 5-HT1Breceptor appears to be important forselectively accommodating the 5substituents of

triptan ligands, whereas the narrower pocket inthe 5-HT2Breceptor shows reduced binding tothese compounds.

Norfenfluramine is the active metabolite offenfluramine, which is one of the two compo-nents of the infamous Fen-Phen antiobesity

cocktail. The antiobesity effect of norfenfluramineoccurs mainly through activation of the 5-HT2Creceptor (1), but due to its high potency and ef-ficacy as a 5-HT2Breceptor agonist, it can causelife-threatening side effects, including pulmonaryhypertension and heart valve disease (9, 2325).To elucidate the structural basis for norfenflur-amines subtype selectivity, we simulated its bind-ing to the 5-HT1Band 5-HT2Breceptor structures(Fig. 3D). Norfenfluramine tightly fit the ortho-

steric binding pocket of the 5-HT2B receptor,with F2175.38 and M2185.39 forming a hydropho-bic cap that interacted with the trifluoromethylgroup. The F2175.38A mutation (table S8) andthe M2185.39V mutation (26) both reduce thepotency of norfenfluramine compared with that

for the wild-type 5-HT2Breceptor. In the dockmodel of the 5-HT1Breceptor, these close cotacts are missing; thus, the potency of norfenflamine is reduced compared with that at t5-HT2B receptor (table S8).

Species-specific differences in the ligand bining properties of 5-HT1Band other 5-HT rectors from rodents and humans have impeded extrapolation of findings from animal modelshumans (27, 28). The rodent 5-HT1Brecepto

for instance, have a much higher binding affity than the human 5-HT1B receptor for certadrenergic compounds caused by a differenat position 7.39 (N351 for rat and mouse, T3for human) (27, 29). Modeling of the T35mutation into the binding pocket of hum

Fig. 3. Docking of the promiscuous (5-HT and LSD) and selective(sumatriptan and norfenfluoramine) ligands into the binding pocketsof the 5-HT1Band 5-HT2Breceptor structures.Docking of 5-HT (A), LSD(B), sumatriptan (C), and norfenfluramine (D) into the ligand binding pocketsof 5-HT1B(receptor colored light blue; ligands colored magenta) and 5-HT2B(receptor colored white; ligands colored green) receptors. In (A) and (B), thepolar interactions between Y7.43 and D3.32, and between D3.32 and the ligands

and T3.37 and the ligands are shown as dashed lines. In (A), the nonpointeractions between the indole ring of 5-HT and residues at positions 5.and 5.46 are shown as dotted lines. In (C), steric hindrance from M218forced a reorientation of the sulfonamide group and a shift of the indole costructure of sumatriptan when docked into the 5-HT2Breceptor. In (D), in t5-HT2B receptor, F217

5.38 and M2185.39 form closed contacts with trifluoromethyl group of the ligand, which are absent in the 5-HT1Brecept

www.sciencemag.org SCIENCE VOL 340 3 MAY 2013

REP


6/6

5-HT1Breceptor revealed that N3557.39, together

with D1293.32 and Y3597.43, form a polar inter-action network that anchors the propanolaminemoiety of adrenergic antagonists, propranololand cyanopindolol, thereby mimicking the rec-ognition modes observed in both the b1and b2adrenergic receptors (Fig. 4). Hydrophobic in-teractions of I1303.33, W3276.48, F3306.51, andF3316.52 with the aromatic rings of propranololand cyanopindolol are largely conserved between5-HT1Band b-adrenergic receptors (30, 31), al-lowing high-affinity binding of these antago-nists in the human 5-HT1BT355

7.39N and rodent5-HT1Breceptors. These findings thereby providea structural explanation for the pharmacolog-ical differences between these rodent and humanGPCRs.

In conclusion, comparative analysis of the5-HT1Band 5-HT2Breceptor structures and func-tions, together with specific mutagenesis studies,provided a comprehensive framework for under-standing ligand promiscuity and selectivity, there-

by aiding development of safer and more effectivemedications that target the GPCR superfamily.

References and Notes1. M.Berger,J. A.Gray,B. L.Roth,Annu.Rev.Med. 60, 355(2009).

2. D. Hoyer et al., Pharmacol. Rev. 46, 157 (1994).

3. B. L. Roth,S. M. Hanizavareh, A. E. Blum, Psychopharmacology

174, 17 (2004).

4. B. L. Roth, D. J. Sheffler, W. K. Kroeze,Nat. Rev. Drug

Discov. 3, 353 (2004).

5. W. K. Kroeze, K. Kristiansen, B. L. Roth, Curr. Top. Med.

Chem.2, 507 (2002).6. M. M. Goldenberg,Pharm. Ther. 37, 499 (2012).

7. B. M. Smith et al., J. Med. Chem. 51, 305 (2008).

8. J. Besnard et al., Nature 492, 215 (2012).

9. B. L. Roth, N. Engl. J. Med. 356 , 6 (2007).

10. C. M. Villaln, D. Centurin, L. F. Valdivia, P. de Vries,

P. R. Saxena, Curr. Vasc. Pharmacol. 1, 71 (2003).

11. Affected drugs include the antimigraine drugs ERG,

DHE, and methysergide (usage limited); the appetite

suppressant drugs fenfluramine, dex-fenfluramine, and

aminorex (all withdrawn); and some drugs used for

treating Parkinsons disease: pergolide and cabergoline(both withdrawn).

12. R. B. Rothman, M. H. Baumann,Pharmacol. Biochem.

Behav. 71, 825 (2002).

13. D. Wacker et al., Science 340, 615 (2013).

14. Materials and methods are available as supplementa

materials on Science Online.15. Superscripts refer to the Ballesteros-Weinstein

numbering, in which the most conserved among clas

GPCR residues in each transmembrane helix are

designated x.50, where x is the helix number.

16. J. A. Ballesteros, H. Weinstein,Methods Neurosci. 25

366 (1995).

17. Residues Y177ECL2, W4227.35, and T4237.36 in

muscarinic acetylcholine receptor M2 are reported to

involved in mediating the action of allosteric modulators

Corresponding residues in the 5-HT1BreceptorV200ECL

F3517.35, and D3527.36are part of the extended binding 18. K. Haga et al., Nature 482, 547 (2012).

19. V. Katritch, V. Cherezov, R. C. Stevens,Trends Pharmacol.

33, 17 (2012).

20. M. R. Braden, D.E. Nichols,Mol. Pharmacol. 72, 1200 (20

21. D. A. Shapiro, K. Kristiansen, W. K. Kroeze, B. L. Rot

Mol. Pharmacol.58, 877 (2000).

22. S. Mehrotraet al., Naunyn Schmiedebergs Arch.

Pharmacol.378 , 371 (2008).

23. H. M. Connollyet al., N. Engl. J. Med. 337, 581 (19

24. R. B. Rothman, Ann. Intern. Med. 136, 779 (2002).

25. R. B. Rothmanet al., Circulation 102, 2836 (2000).

26. V. Setola, M. Dukat, R. A. Glennon, B. L. Roth, Mol.

Pharmacol.68 , 20 (2005).

27. N. Adham et al., Neuropharmacology33, 387 (1994

28. D. L. Nelson, V. L. Lucaites, J. E. Audia, J. S. Nissen,

D. B. Wainscott,J. Pharmacol. Exp. Ther.265, 1272 (19

29. D. Oksenberget al., Nature 360 , 161 (1992).30. V. Cherezovet al., Science 318 , 1258 (2007).

31. T. Warneet al., Nature 454, 486 (2008).

Acknowledgments:This work was supported by National

Institute of General Medical Sciences (NIGMS) PSI:Biology

grant U54 GM094618 for biological studies and structur

production (target GPCR-118) (V.K., V.C., and R.C.S.); NI

Common Fund in Structural Biology grant P50 GM073197

technology development (V.C. and R.C.S.); the Jay and B

Van Andel Foundation, R01 DK071662, Ministry of Scien

and Technology (China) grants 2012ZX09301001-005 an

2012CB910403, and Amway (China) (H.E.X.); R01 MH618

U19 MH82441, R01 DA27170, and the National Institute

of Mental Health Psychoactive Drug Screening Program

(X.-P.H., E.V., J.D.M., and B.L.R.); and the Michael Hooke

Chair of Pharmacology (B.L.R.). D.W. is supported by a

Boehringer Ingelheim Fonds Ph.D. Fellowship. We thank

J. Velasquez for help on molecular biology; T. Trinh, K. A

and M. Chu for help on baculovirus expression; I. Wilson

for careful review and scientific feedback on the manusc

K. Kadyshevskaya for assistance with figure preparation;

A. Walker for assistance with manuscript preparation;

J. Smith, R. Fischetti, and N. Sanishvili for assistance in

development and use of the minibeam and beamtime at

GM/CA-CAT beamline 23-ID at the Advanced Photon Sour

(APS), which is supported by National Cancer Institute gr

Y1-CO-1020 and NIGMS grant Y1-GM-1104; and Z. Wawrza

and J. S. Brunzelle for assistance in data collection at the

beamlines of sector 21 (Life Sciences Collaborative Access Team

which is funded in part by the Michigan Economic Developm

Corporation and Michigan Technology Tri-Corridor Grant

085P1000817. Use of the APS was supported by the Offic

of Science of the U.S. Department of Energy. The coordina

and the structure factors have been deposited in the Prote

Data Bank (PDB) under accession codes 4IAR (5-HT1B/ERGand 4IAQ (5-HT1B/DHE). R.C.S. is a founder and paid

consultant for Receptos, a GPCR structurebased drugdiscovery company.

Supplementary Materialswww.sciencemag.org/cgi/content/full/science.1232807/DC1

Materials and Methods

Figs. S1 to S9

Tables S1 to S8

References (3253)

15 November 2012; accepted 27 February 2013

Published online 21 March 2013;

10.1126/science.1232807

Fig. 4. Structural basis for differences in the pharmacological properties between human androdent 5-HT1B receptors. The high-affinity b-adrenergic antagonists propranolol (A) and cyano-pindolol (B), both shown in yellow-colored carbons, are docked in the model based on the human5-HT1B/ERG structure with T

7.39 mutated to N, as found in 5-HT1Brat and mouse orthologs. The N7.39

side chain (magenta carbons) remained flexible in the docking procedure. The hydrogen bond net-

work involving N7.39, D3.32, and Y7.43 and propanolamine moieties of the ligands is shown as orange dots.Carazolol (green carbons) from the superimposed b2-adrenergic receptor structure (PDB ID: 2RH1) is shownfor comparison.

3 MAY 2013 VOL 340 SCIENCE www sciencemag org14

REPORTS

serotonin receptors science 2013

Documents