A

ccbewg©

K

1

dcit2gFPcgba2G

0d

Molecular and Cellular Endocrinology 260–262 (2007) 73–82

Assembly and structural characterization of an authentic complexbetween human follicle stimulating hormone and a

hormone-binding ectodomain of its receptor

Qing R. Fan a, Wayne A. Hendrickson a,b,∗a Department of Biochemistry and Molecular Biophysics, Columbia University, 630 West 168th Street,

New York, NY 10032, United Statesb Howard Hughes Medical Institute, Columbia University, 630 West 168th Street, New York, NY 10032, United States

Received 24 August 2005; accepted 22 December 2005

bstract

Follicle stimulating hormone (FSH) is secreted from the pituitary gland to regulate reproduction in vertebrates. FSH signals through a G-proteinoupled receptor (FSHR) on the target cell surface. We describe here the strategy to produce a soluble FSH–FSHR complex that involves theo-secretion of a truncated FSHR ectodomain (FSHRHB) and a covalently linked FSH�� heterodimer from baculovirus-infected insect cells. FSHinds to FSHRHB with a high affinity comparable to that for the full-length receptor. The crystal structure of the FSH–FSHRHB complex provides

xplanations for the high affinity and specificity of FSH interaction with FSHR, and it shows an unexpected dimerization of these complexes. Heree also compare the crystal structure with theoretical models of the FSH–FSHR-binding mode. We conclude that the FSH–FSHRHB structureives an authentic representation of FSH binding to intact FSHR.

2006 Elsevier Ireland Ltd. All rights reserved.

sion;

laaifSh(iu

ccf

eywords: Glycoprotein hormones; Gonadotropin receptors; Insect cell expres

. Introduction

Follicle stimulating hormone (FSH) is essential for the repro-uction of vertebrates. Mutations in the hormone and its receptorause failure to develop ovarian follicles in females as well asmpaired spermatogenesis in males, both of which could leado infertility (Simoni et al., 1997; Themmen and Huhtaniemi,000). Together with the luteinizing hormone (LH), chorioniconadotropin (CG) and thyroid stimulating hormone (TSH),SH belongs to the family of glycoprotein hormones (Pierce andarsons, 1981). These hormones are heterodimeric moleculesomprising �- and �-subunits (Pierce and Parsons, 1981). Thelycoprotein hormones exert their actions on target cells byinding to specific G-protein coupled receptors (GPCR), which

re characterized by seven transmembrane helices (Dias et al.,002; Ascoli et al., 2002; Szkudlinski et al., 2002). Unlike mostPCRs, glycoprotein hormone receptors have large extracel-

∗ Corresponding author. Tel.: +1 212 305 3456; fax: +1 212 305 7379.E-mail address: wayne@convex.hhmi.columbia.edu (W.A. Hendrickson).

ssdugHe

303-7207/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.mce.2005.12.055

Crystal structure

ular domains containing leucine-rich repeats (LRRs) (Dias etl., 2002; Ascoli et al., 2002; Szkudlinski et al., 2002). Inter-ction between a glycoprotein hormone and the ectodomain ofts receptor leads to activation of the transmembrane domainor signal transduction (Dias et al., 2002; Ascoli et al., 2002;zkudlinski et al., 2002). Although the crystal structures ofuman CG (Wu et al., 1994; Lapthorn et al., 1994) and FSHFox et al., 2001) have both been solved previously, structuralnformation for any of the glycoprotein hormone receptors hasntil now defied experimental analysis.

We have assembled a complex of human FSH and the extra-ellular hormone-binding domain of its receptor (FSHRHB) byo-expressing the hormone and receptor. The stable complexormed by FSH and FSHRHB reflects the high affinity andpecificity of the interactions between FSH and FSHR on cellurfaces. The structure of the FSH–FSHRHB complex has beenetermined at 2.9 A resolution, which provides a molecular

nderstanding of the receptor–hormone interface and a sug-ested mode of dimerization (Fan and Hendrickson, 2005a).ere we also compare the crystal structure with theoretical mod-

ls of glycoprotein hormone–receptor binding.

7 d Cell

2

2

iGpCpTCTtpdptG

fTTAtCeiTor

2

s1se

eDcTaft8

2

cdaw(5t

2

b�

ua

nCecrI

2

f1a

2F

c

K

wrp1FmtdSpmoFibfF

2

ssht

a

3

3

ceuwC

4 Q.R. Fan, W.A. Hendrickson / Molecular an

. Methods

.1. Construction of transfer vectors for expression in insect cells

The entire FSH� chain including its signal sequence was amplified fromts cDNA clone using the forward primer 5′-TAGGGCGAATTCACCAG-ATGAAGACACTCCAGTTTTTCTTCCTTTTCTGTTGC-3′ and the reverserimer 5′-GCTAGAGGATCCACCACCAGAACCACCACCTTCTTTCATTT-ACCAAAGGAGCAGTAGCTGGG-3′. The sequence encoding the maturerotein of FSH� was amplified with the forward primer 5′-GCTAGAGGA-CCGGTGGTGGTTCTGGTGGTGGTGCTCCT GATGTGCAGGATTGCC-AGAATGC-3′ and the reverse primer 5′-CCTAGGGCGGCCGCTTAT-AAGATTTGTGATAATAACAAGTACTGCAGTG-3′. The PCR product con-aining the FSH� sequence was digested with EcoRI and BamHI, and the PCRroduct of mature FSH� sequence was digested with BamHI and NotI. Theseigested PCR fragments were ligated together into an EcoRI–NotI digestedFastBac Dual plasmid. This joining thus encoded a single-chain FSH wherebyhe C-terminus of FSH� is connected to the N-terminus of FSH� by a 15-residuely-Ser-rich linker dictated by the primer sequences.

FSHRHB (residues 1–268) was amplified from its cDNA clone using theorward primer 5′-GATCCGAGCTCGAGATAATTATGGCCCTGCTCCTGG-CTCTTTGCTGGCA-3′ and the reverse primer 5′-GAGCGGGGTACCT-ATTACTTATCATCATCATCCTTGTAATCGCTGGCTTCCATGAGGGCG-CAAGCTTTTCCAG-3′. A FLAG tag was engineered at the C-terminus of

he FSHR through the reverse primer. Receptor constructs with alternative-terminal truncations of the FSHR ectodomain were amplified similarly. Inach case, the PCR product was digested with XhoI and KpnI and then ligatednto the pFastBac Dual plasmid which already contained the single-chain FSH.he genes for single-chain FSH and for FSHRHB were placed under the controlf the polyhedrin and p10 promoters, respectively. The sequences of both theeceptor and hormone constructs were verified by DNA sequencing.

.2. Insect cell culture and protein purification

Trichoplusia ni (High Five) cells were cultured in monolayer or suspen-ion at 27 ◦C in Grace medium (Life Technologies Inc.) supplemented with0% fetal bovine serum (Life Technologies Inc.). The cells were adapted intoerum-free medium (SF900II SFM from Life Technologies Inc.) for proteinxpression.

Recombinant baculoviruses encoding both single-chain FSH and truncatedctodomain constructs of FSHR were produced in High Five cells using bacmidNA transposed with the transfer vectors. High Five cells grown to 2.0 × 106

ells/ml were infected with each baculovirus carrying the FSH and FSHR genes.he cell culture supernatant was harvested 72 h after infection and applied ton anti-FLAG antibody (M2) affinity column. The FSH–FSHR complex elutedrom the antibody column using FLAG peptide was purified further by gel fil-ration column chromatography (Superdex 200, Pharmacia) in 20 mM Tris, pH.0 and 150 mM NaCl.

.3. Enzymatic deglycosylation

The FSH–FSHRHB complex was partially deglycosylated using endogly-osidase F1, F2 and F3, either alone or in combination with one another. Theeglycosylation reactions were carried out at 20 ◦C for 5 days in 50 mM sodiumcetate, pH 4.5, or 50 mM sodium citrate, pH 5.5. The FSH–FSHRHB proteinas separated from deglycosylation enzymes by ion-exchange chromatography

MonoQ, Pharmacia) in 20 mM Tris, pH 8.0, using a salt gradient from 0 to00 mM NaCl. Partially deglycosylated FSH–FSHRHB complex was concen-rated to 9.8 mg/ml in 20 mM Tris, pH 8.0, for crystallization.

.4. Production of free hormones

Recombinant human FSH, CG and LH were produced in insect cells using theaculovirus expression systems. In each case, the genes encoding the hormone- and �-subunits were subcloned into a pFastBac Dual plasmid (Invitrogen)nder separate promoters. The �-construct comprises its native signal sequencend entire mature sequence (residues 1–92). The �-construct comprises the

Shta

ular Endocrinology 260–262 (2007) 73–82

ative signal sequence and entire mature sequence of FSH (residues 1–111),G (1–145) or LH (1–121). A FLAG tag is also engineered at the C-terminalnd of every hormone �-chain to facilitate purification by antibody affinityhromatography. Each hormone is further purified by gel filtration chromatog-aphy. Recombinant hCG expressed in CHO cells was obtained from Serononc. through Joyce Lustbader.

.5. Competition-binding assays

The FSH–FSHRHB complex and excess amount of free FSH, free CG orree LH were incubated in a reaction buffer containing 20 mM Tris, pH 8.0, and50 mM NaCl for 30 min at room temperature. The reaction mixture was thenssayed on a 8–25% native gel.

.6. Estimation of the dissociation constant (Kd) of solubleSH–FSHRHB complex

We estimated the Kd of the soluble FSH–FSHRHB complex based on itsoncentration in cell culture.

d = [H][R]

[HR],

here [H], [R] and [HR] represent the concentrations of free hormone, freeeceptor and their complex, respectively. The yield of purified FSH–FSHRHB

rotein is about 0.1 mg per liter of insect cell culture, which corresponds to.3 nM of the complex. When expressed alone, the production level of freeSH is about 1.6 mg per liter of insect cell culture. Given that the estimatedolecular mass of free FSH is about 40 kDa based on its mobility on SDS page,

he concentration of free hormone in the cultured medium is about 40 nM. Weo not have any direct measure of the amount of free receptor in the medium.ince a FLAG tag is engineered at the receptor C-terminus, all expressed receptorrotein, whether it is in hormone-bound or free form, would be purified from theedium by the antibody affinity column. Upon gel filtration chromatography

f this protein mixture, there is no detectable amount of any properly foldedSHRHB, indicating either that there is very little free receptor or that most of

t has precipitated or formed soluble aggregates. This result has been confirmedy expressing FSHRHB alone in insect cells. If we assume the concentration ofree receptor in the cell culture is at most 1/10 that of the complex, the Kd of theSH–FSHRHB complex would be equal to or lower than approximately 4 nM.

.7. Structural superposition and illustrations

The program LSQMAN (Kleywegt, 1996) was used to determine the optimaltructural alignment between the FSH molecule in the FSH–FSHRHB complextructure (1XWD) and hCG in the theoretical hCG–LHR model (1XUL). Theormone molecules were then superimposed in O (Jones et al., 1991) using theransformation matrix that related their best aligned segments.

Figures were generated using RIBBONS (Carson, 1997), Adobe Illustratornd Adobe Photoshop.

. Results and discussion

.1. Assembly of a soluble FSH–FSHRHB complex

In order to assemble a stable FSH–FSHR complex, weo-expressed the receptor and hormone using the baculovirusxpression system (Fig. 1A). In addition, the �- and �- sub-nits of the heterodimeric hormone were covalently stabilizedith a Gly-Ser-rich linker (GGGSGGGSGGGSGGG) from the-terminus of FSH� to the N-terminus of FSH� (Fig. 1A).

ingle-chain FSH and CG constructs with different linkersave previously been shown to retain receptor-binding activi-ies (Sugahara et al., 1995; Narayan et al., 1995; Schmidt etl., 2001). Genes encoding the extracellular domain of FSHR

Q.R. Fan, W.A. Hendrickson / Molecular and Cellular Endocrinology 260–262 (2007) 73–82 75

Fig. 1. (A) Schematic diagram of the baculovirus expression constructs for human FSHRHB and the covalently linked human FSH� and FSH� subunits. (B) Westernblot of the culture medium of insect cells that were infected with the recombinant virus for FSHR and single-chain FSH. Culture samples were concentrated2 l (DTa

aetc

cerasbmttraoeadwhCgw

fopNeggtcuTstc

a(tccpeStnobaoahmMiios2

hFtbFitonFp

0-fold and run on a 4–15% SDS gel in the presence of 100 mM dithiothreitontibody.

nd the single-chain FSH were inserted into the same vector forxpression. A FLAG tag was engineered at the C-terminus ofhe receptor to facilitate purification of the complex by affinityolumn chromatography.

To dissect the ligand-binding region of FSHR, we testedonstructs with different C-terminal truncations of the FSHRctodomain. The ectodomains of all glycoprotein hormoneeceptors consist of a series of leucine-rich repeats flankedt each end by cysteine clusters. We discovered early in ourtudies that it was important to retain the N-terminal cluster,ut the C-terminal cysteine-rich region is not required for hor-one binding. We found that the hormone-binding domain of

he receptor, FSHRHB, contains the N-terminal cysteine clus-er and the LRRs (residues 1–268). Expression of the truncatedeceptor ectodomain could be detected by Western blot usingn anti-FLAG antibody (Fig. 1B). A slightly shorter constructf FSHR containing residues 1–247 was also secreted as prop-rly folded protein in complex with single-chain FSH. We wereble to obtain crystals of the FSH–FSHR (1–247) complex thatiffracted to 9 A. FSHR constructs that were shortened furtherere hardly secreted even in the presence of FSH. On the otherand, longer constructs that additionally included part of the-terminal cysteine-rich region were secreted as soluble aggre-ates. A construct that encoded the entire ectodomain of FSHRas hardly secreted with FSH.FSHRHB forms a stable complex with FSH when secreted

rom insect cells. Fig. 2A shows the gel filtration chromatogramf co-secreted FSH and FSHRHB. The correctly folded proteineak contains both the receptor and hormone according to-terminal sequencing of the two bands from SDS-PAGE

lectrophoresis (Fig. 2B) as well as the band from a nativeel (Fig. 2C). The retention time for the folded complex byel filtration chromatography corresponds to a 72 kDa pro-ein, which is close to the exact molecular mass of the gly-osylated FSH–FSHRHB complex determined by analyticalltracentrifugation (78 kDa) (Fan and Hendrickson, 2005a).

he molecular mass of the FSH–FSHRHB complex is con-istent with a 1:1 binding stoichiometry between the recep-or and hormone, with 32% of the mass being contributed byarbohydrates.

bTas

HB

T). The blot was developed using the M2 anti-FLAG antibody as the primary

Human FSH binds to cell surface FSHR with high affinitynd a dissociation constant (Kd) in the nanomolar (nM) rangeSimoni et al., 1997; Ascoli et al., 2002). We have estimatedhe Kd of the soluble FSH–FSHRHB complex based on its con-entration in cell culture. The yield of purified FSH–FSHRHBomplex was used to calculate the concentration of the com-lex in the medium, and the concentration of free hormone wasstimated from the yield of free FSH when expressed alone.ince FSHRHB was secreted as soluble aggregates, we assumed

hat the concentration of free receptor in the medium couldot possibly exceed 1/10 that of the complex. We have thenbtained an upper limit of 4 nM for the Kd for our recom-inant FSH–FSHRHB complex, which is comparable to theffinity observed for FSH binding to membrane-bound FSHRn cell surfaces. This is in keeping with measurements of theffinity of similar ectodomain constructs with their cognateormones, which typically even exceed the tight binding of hor-one to the intact receptor on membranes (Simoni et al., 1997).oreover, the stability of the FSH–FSHRHB complex through

mmunoaffinity and gel filtration chromatography, as well dur-ng crystallization, favors a high affinity complex, or at leastne with an extremely low rate constant of dissociation, con-istent with studies of others (Simoni et al., 1997; Ascoli et al.,002).

The interaction between the co-secreted FSH and FSHRHB isighly specific as demonstrated by competition-binding assays.irst, in accordance with the preference of FSHR for FSH,

he soluble FSH–FSHRHB complex could not be dissociatedy excess amount of free hCG or LH hormone. As shown inig. 2D for hCG, addition of free hCG did not change the mobil-

ty of hCG or FSH–FSHRHB complex on native gel, indicatinghat hCG could not displace FSH in binding to FSHRHB. Sec-nd, in the presence of excess amount of free FSH, there waso additional binding observed between the free FSH and theSH–FSHRHB complex. Incubation of the FSH–FSHRHB com-lex with excess amount of free FSH did not produce any novel

and that would correspond to a second binding event (Fig. 2E).his indicates that the hormone-binding site on FSHRHB hadlready been occupied by the single-chain FSH that was co-ecreted with FSHRHB. These observations confirm that the tight

76 Q.R. Fan, W.A. Hendrickson / Molecular and Cellular Endocrinology 260–262 (2007) 73–82

F x. Sa( monea free F

at

3F

cittFuberfgbtprwbp

8dFe

tawdtps2

3

mcsv

ig. 2. (A) Gel filtration column chromatography of the FSH–FSHRHB compleC) 8–25% native gel. Samples of purified FSH–FSHRHB (5 �g) and free hordditional interactions upon mixture for: (D) 5 �g of free CG and (E) 7.5 �g of

ssociation between co-expressed FSH and FSHRHB is not dueo non-specific binding.

.2. Partial deglycosylation and crystallization of theSH–FSHRHB complex

The fully glycosylated FSH–FSHRHB (1–268) complexould be crystallized in several different forms, one of whichs shown in Fig. 3C. The best of these crystals only diffractedo 9 A spacings. In order to improve the resolution limits ofhese crystals, we carried out enzymatic deglycosylation of theSH–FSHRHB complex. Partial deglycosylation was observedsing endoglycosidase F1, F2 and F3, either alone or in com-ination with one another. We chose to use a combination ofndoglycosidase F2 and F3 because the cleaved product had aelatively tighter band on native gel when compared with theully glycosylated control (Fig. 3B), indicating a more homo-eneous sample. SDS gels showed that about 20% of the car-ohydrate was removed, and that sugars were mainly beingaken from FSH, which was the top band in the control sam-le (Fig. 3A). As a result of deglycosylation, the hormone and

eceptor condensed into one band due to their similar moleculareight (Fig. 3A). Removal of carbohydrate reduced the solu-ility of the FSH–FSHRHB complex, causing about 50% of therotein to precipitate out of solution. There was an overall loss of

bloa

mples of the indicated peak fraction were assayed on (B) 8–25% SDS gel andligands were also analyzed by 8–25% native gel electrophoresis for possibleSH.

0% from the deglycosylation procedure. Nevertheless, partialeglycosylation did not disrupt the interaction between FSH andSHRHB; the complex stayed together on gel filtration and ion-xchange chromatorgraphy after the endoglycosidase treatment.

The partially deglycosylated FSH–FSHRHB complex crys-allized in a new crystal form (Fig. 3D). These crystals grews clusters of plates which were 5–10 �m thick. Single crystalsere broken off the cluster for measurement of X-ray diffractionata. The best of these crystals diffracted to a limit correspondingo 2.9 A resolution. The structure of the FSH–FSHRHB com-lex was determined by molecular replacement using the knowntructure of FSH as the search model (Fan and Hendrickson,005a).

.3. Structure of the FSH–FSHRHB complex

The structure of the FSH–FSHRHB complex shows the hor-one bound in a handclasp fashion to the concave face of a

urved receptor (Fan and Hendrickson, 2005a) (Fig. 4). Thetructure of FSHR consists of 10 highly irregular repeats, whichary both in length and overall conformation. The hormone-

inding site on FSHR includes all 10 parallel �-strands andoops just C-terminal to these strands. The receptor-binding siten FSH is formed by the C-terminal segments of both FSH�-nd �-subunits as well as the �L2 and �L2 loops. None of

Q.R. Fan, W.A. Hendrickson / Molecular and Cellular Endocrinology 260–262 (2007) 73–82 77

F B comc tert-d d 0.1

ti

d�etaattot

tmathapricvbc

t�dKasa(sT

3

udscabtt

ig. 3. Fully glycosylated (control) and partially deglycosylated FSH–FSHRH

rystal of fully glycosylated FSH–FSHRHB complex grown at 4 ◦C from 16%eglycosylated FSH–FSHRHB complex grown at 20 ◦C from 10% PEG 3350 an

he carbohydrate moieties on the hormone or on the receptor isnvolved in the interaction interface.

The linker peptide used to stabilize the FSH�� heterodimeroes not appear to affect the association between FSH�- and-subunits or the receptor-binding mode of FSH. First of all, thentire linker together with four C-terminal residues of FSH� andwo N-terminal residues of FSH� (five in the second protomer)re not visible in the electron density map, indicating that theyre disordered and highly flexible. From the location of theseermini and as designed, the covalent linker connecting the C-erminus of FSH� and N-terminus of FSH� must span a regionn the FSH surface opposite to the receptor-binding site, andherefore would not interfere with receptor binding.

The interactions between FSH and FSHRHB suggest thathe binding mode is universal among all glycoprotein hor-

ones and receptors. First, the structures of the hormonend the hormone-binding domain of the receptor are expectedo be very similar among their family members because ofigh sequence similarity. Second, the structural elements foundt the receptor–hormone interface of FSH–FSHRHB com-lex have also been implicated in the interactions of othereceptor–hormone pairs. Finally, many residues of FSHRHBnvolved in contacting the common �-subunit of hormone are

onserved among receptor homologues, consistent with a uni-ersal binding mode. The FSH–FSHRHB interface has a highuried charge density, and many residues with complementaryharges contribute to the universal interactions.

dYsw

plex were assayed on: (A) 8–25% SDS gel and (B) 8–25% native gel. (C) Abutanol and 0.1 M sodium citrate, pH 6.0. (D) Thin plate crystals of partiallyM Li2SO4.

The structure of FSH–FSHRHB complex also demonstrateshat receptor specificity is jointly mediated by the hormone- and �-subunits. There are four predominant specificity-etermining sites on the surface of the receptor at residues L55,179, I222 and the structurally adjacent E76 and R101. Inter-

ctions between receptor residues 55 and 179 and the hormoneeat-belt region of FSH� determine the selection between FSHnd TSH versus LH/CG, consistent with mutational analysesDias et al., 1994; Smits et al., 2003; Vischer et al., 2003a). Theites at residues 76, 101 and 222 distinguish between FSH andSH.

.4. Implications of dimeric associations in the crystal

There are two FSH–FSHRHB complexes in the asymmetricnit of the crystal, and these associate as a biologically plausibleimer. This observation provoked us to perform experiments inolution where we found dimers characterized by a dissociationonstant of Kd ∼400 �M (Fan and Hendrickson, 2005a). Thessociated complexes in the crystallographic dimer are relatedy approximate two-fold symmetry (rotation χ= 175.2◦, screwranslation tχ = −1.2 A) with a dimer interface that involves onlyhe outer surface of the two receptor molecules (Fig. 5A). The

irect contacts are mainly hydrophobic and the highly conserved110 residue predominates in the interaction. The weak intrin-

ic affinity of the FSHRHB dimer is expected to be enhancedhen FSHRHB is tethered to the cell membrane. The mode of

78 Q.R. Fan, W.A. Hendrickson / Molecular and Cell

Fig. 4. Ribbon diagram of the crystal structure of human FSH bound to FSHRHB.FSH�- and �-chains are in green and cyan, respectively. FSHRHB is in red. Theobserved N-linked carbohydrates at Asn52, Asn78 of FSH�, Asn7 and Asn24of FSH� and Asn191 of FSHRHB are in yellow. Disulfide bonds are in black.Adapted with permission from Nature (Fan and Hendrickson, 2005a).

it

cbhfa(aotTst

t(hirno7Cabe separated by ∼90 A if each were equivalently engaged with

Fpt(wda

�8h

ig. 5. Dimeric associations. (A–C) Ribbon diagrams of quasi-symmetric crystallogarallel to the symmetry axis, as if looking at the membrane. (D and E) Schematic ilhe complexes shown above and from rhodopsin are shown for the hormone (cyan)black dotted trace). The view is perpendicular to figures directly above, parallel withith outline of an offset 7TM dimer disposed for interaction with one of the bound FSimer with both 7TM domains engaged. (E) Asymmetric principal dimer with only oligned. (F) Symmetric alternative dimer with both 7TM domains engaged.

ular Endocrinology 260–262 (2007) 73–82

nteraction in this principal dimer is plausible for all glycopro-ein hormones, and could exist in uncomplexed receptors.

The crystal packing is such that an alternative dimeric asso-iation of FSH–FSHRHB complexes also exists (suggested to usy Xuliang Jiang). It has three separate areas of contact: eachormone of the dimer contacts the N-terminus of the receptorrom the neighboring complex, and the two receptors also inter-ct with each other through their N-terminal cystein-rich regionsFig. 5C). Because the N-terminal segment is highly variablemong the different glycoprotein hormone receptors, this modef dimerization is unlikely to occur for LHR or TSHR. Nei-her would it be expected to exist in absence of the FSH ligand.his alternative dimer is also much less symmetric, with a sub-tantial screw translation (rotation χ= 175.2◦, screw translationz = −6.6 A).

Compelling evidence has been presented recently for func-ional dimerization of TSHR through the seven-transmembrane7TM) helical portion of the receptor and for an asymmetry inormone binding, with negative cooperativity between two bind-ng sites (Urizar et al., 2005). How might ectodomain dimers beelated to 7TM dimers, and what are the implications for sig-al transduction? We previously noted the puzzling separationf the FSHRHB dimers in light of a prospective association ofTM domains into dimers (Fan and Hendrickson, 2005a). The-termini of FSHRHB domains in the principal dimer are 74 Apart, and the centers of their respective 7TM domains would

L3 tips of FSH in the complex (Fig. 5D). These separations are8 and ∼65 A, respectively, in the alternative dimer. Even here,owever, the �L3 tips would be too far apart for simultaneous

raphic dimers of FSH–FSHRHB complexes, colored as in Fig. 4. The view islustrations of dimeric receptors on the cell membrane. Projected outlines from, hormone-binding domain (orange), linker and 7TM transmembrane domainthe membrane plane. (A) Principal crystallographic dimer. (B) Principal dimerH molecules. (C) Alternative crystallographic dimer. (D) Symmetric principal

ne 7TM domain engaged. Note that dimer axes for the two components are not

Cell

e7habh

atowit(Fitistritdac(m

3

osoeo�

(eawemsoosta�vt

ttiaKttltS

htdFhto

FFF

Q.R. Fan, W.A. Hendrickson / Molecular and

ngagement with a 7TM dimer (Fig. 5F). Actual structures ofTM dimers are not known, but a model of the rhodopin dimeras been fitted to images of retinal disc membranes made bytomic force microscopy (Fotiadis et al., 2004). The distanceetween 11-cis retinal sites in this model is ∼31 A, far less thanormone separations in the ectodomain dimers.

Asymmetry in the dimeric signaling complex provides forpossible resolution of the apparent dimer mismatch. The dis-

ance between FSHR residues 360 at the start of helices TM1f the rhodopsin-based 7TM dimer model for FSHR is 78 A,hich is very similar to the distance between FSHRHB C-termini

n principal dimers and, incidentally, also similar to the separa-ion of C-termini in dimers of glutamate-receptor ectodomainsKunishima et al., 2000). The linker segment that connectsHSRHB to its 7TM domain comprises 98 residues (269–366),

ncluding three disulfide bonds, and it is straightforward to pic-ure a symmetric empty receptor with aligned two-fold axes forts ectodomain and 7TM-domain dimers. This would have 7TMites occluded from possible interaction with the hormone, buthe linker segment may afford sufficient flexibility to permit aelative displacement and possible tilting. Upon hormone bind-ng, the ligand at one site could engage one 7TM domain whilehat bound at the second site would be remote from the 7TMimer (Fig. 5B and E). The linkers are each 42 A long, but theyre differently disposed in this asymmetric model. Such physi-al asymmetry may relate to the observed negative cooperativityUrizar et al., 2005) and it has significant implications for theechanism of G-protein stimulation of GPCRs.

.5. Comparison with theoretical models of the complex

Based on the first structure of a LRR containing protein, thatf porcine ribonuclease inhibitor (Kobe and Deisenhofer, 1993),everal models have been predicted for the extracellular domain

f glycoprotein hormone receptors (Jiang et al., 1995; Kajavat al., 1995; Bhowmick et al., 1996; Kleinau et al., 2004). Sec-ndary structure prediction suggested the presence of alternating-strands and �-helices in the ectodomain of these receptors

t(eo

ig. 6. Ribbon diagrams comparing the predicted hCG–LHRHB interaction model (SH–FSHRHB complex (FSH in blue and FSHRHB in red). The structure of hCG inSH. (A) Top view and (B) side view.

ular Endocrinology 260–262 (2007) 73–82 79

Jiang et al., 1995). This had led to the initial prediction that thectodomain structure of glycoprotein hormone receptors wouldlso feature �� conformations as in the structure of RI, albeitith less developed helical structure (Jiang et al., 1995; Kajava

t al., 1995; Bhowmick et al., 1996) (Fig. 6). Recently, homologyodeling of the TSHR ectodomain based on the NgR structure

uggested that it adopts a LRR fold with extended conformationsn the convex side (Kleinau et al., 2004). The crystal structuref FSHRHB is devoid of any helices despite of the fact thatome residues on the outer circumference have (φ, ψ) angles inhe left-handed helical region. The interstrand segments mostlydopt extended conformations with the presence of several short-strands. The structure of FSHRHB also exhibits different cur-atures for its two parts, which was not anticipated in any of theheoretical models.

Different models have also been proposed for the interac-ion between glycoprotein hormones and the ectodomain ofheir receptors based on mutagenesis data, peptide-binding stud-es and antibody blocking assays (Jiang et al., 1995; Moyle etl., 1995, 2004; Dias and Van Roey, 2001; Sohn et al., 2003;leinau et al., 2004). Most of the theoretical models predicted

hat the hormone would bind to the concave surface of the recep-or solenoid structure, but the relative orientation of the hormoneong axis and the receptor solenoid varied from being parallelo perpendicular (Jiang et al., 1995; Dias and Van Roey, 2001;ohn et al., 2003; Kleinau et al., 2004).

The model by Jiang et al. (1995) of the complex betweenuman CG (hCG) and the LHR ectodomain, as deposited intohe Protein Data Bank (accession code 1XUL), roughly pre-icted the binding mode observed in the crystal structure of theSH–FSHRHB complex. Notably, both have the long axis of theormone perpendicular to the receptor tube (Fig. 6). Superposi-ion of the CG–LHR interaction model with the crystal structuref the FSH–FSHRHB complex shows, however, that the recep-

or was predicted to bind to a wrong surface of the hormoneFig. 6). Moreover, the modeling was not decisive on the ori-ntation of the hormone long axis with respect to the directionf the parallel �-strands of the receptor molecule (Jiang et al.,

hCG in green and LHRHB model in gray) with the crystal structure of humanthe predicted model was superimposed onto the structure of receptor-bound

8 d Cell

1erbTe

trtpsbbi(rwmrc

3c

cvetnaitvtcacmtth

atsmiodcFhtao

Ttss(ecFic

twhaacteda11LTasecsomrttre

momtwoeeodas(22

0 Q.R. Fan, W.A. Hendrickson / Molecular an

995), although the deposited complex did have this correct ori-ntation. This receptor model, based exclusively on RI, also leftesidual helical character in backside strands, which would haveeen removed based on later templates (Kleinau et al., 2004).his example illustrates some of the current limitations of mod-ling in predicting complex protein–protein interactions.

We are not able to perform a direct superposition betweenhe FSH–FSHRHB crystal structure and any other of the theo-etical models because no coordinates have been deposited forhese models. Unlike most theoretical models, which correctlyredicted the hormone-binding site to involve the parallel �-heet on the concave surface of the receptor structure, modelsy Moyle et al. (1995, 2004) suggested that hormone wouldind to the upper rim of the LRR containing structure, contact-ng only the loops that follow the parallel �-strands. Moyle et al.2004) proposed instead that a module formed by the cysteine-ich cluster at the C-terminal end of the receptor ectodomainould contact the parallel �-sheet of the receptor. The Moyleodels missed essentially all of the structural elements of the

eceptor that are involved in hormone-binding in the observedomplex (Fan and Hendrickson, 2005a).

.6. The experimental structure in light of theoreticalhallenges

Moyle has tried to explain the disagreement between therystal structure and his theoretical model by impugning thealidity of the experimental result. At the recent ICGR confer-nce, he suggested in his presentation and in open discussionhat the binding mode observed in the crystal structure is due toon-specific binding of FSH to a truncated FSHR ectodomaint the low expression level of the receptor–hormone complexn cell culture and that the receptor structure in the crystal asruncated may be distorted from naturally greater intrinsic cur-ature. More recently, Moyle et al. (2005) have proposed thathe binding mode observed in the structure of the FSH–FSHRHBomplex may represent the way the hormone would bind ton alternatively spliced variant of the receptor, which lacks theysteine-cluster domain that links the LRR domain to the trans-embrane helical domain, but does not reflect authentic binding

o the full-length receptor. Moyle et al. (2005) has also ques-ioned the relevance of some mutational tests in mapping theormone–receptor interface.

Here we provide additional experimental detail to validate theuthenticity of the recombinant human FSH–FSHRHB complexhat we produced from insect cell culture. We have demon-trated that this soluble complex was secreted into the cultureedium as a homogeneous species. Based on its discrete mobil-

ty on native gels, the complex remained the same speciesver a wide range of concentrations, from about 0.1 mg/ml iniluted fractions off the gel filtration column to 10 mg/ml inrystallization samples. The interaction between the co-secretedSH and FSHRHB had native-like affinity (Kd ≤ 4 nM) and was

ighly specific. The FSH–FSHRHB complex remained intact inhe presence of excess hCG or LH. Moreover, it did not binddditional free FSH, indicating that the bound FSH alreadyccupied the high affinity hormone-binding site on FSHRHB.

bts(

ular Endocrinology 260–262 (2007) 73–82

he proposition that truncation or the crystal lattice has dis-orted the structure of FSHRHB is implausible in light of theimilarity of FSHRHB to other LRR structures, none of whichhow shape changes in different lattices or upon ligand bindingKobe and Deisenhofer, 1993, 1995; Huizinga et al., 2002; Ufft al., 2002; Dumas et al., 2004; Schubert et al., 2002). We con-lude that the binding mode observed in the crystal structure ofSH–FSHRHB complex does authentically represent the specific

nteraction between FSH and its cognate receptor under nativeonditions.

The hormone–receptor interaction observed in the crys-al structure is consistent with extensive mutagenesis studies,hich have identified the parallel �-sheet of the receptor as itsormone-binding site (Bhowmick et al., 1996, 1999; Schmidt etl., 2001; Song et al., 2001; Smits et al., 2002, 2003; Vischer etl., 2003a,b). The structure is also consistent with studies impli-ating specific hormone segments in receptor binding. Namely,he common C-terminus of the �-chain (Chen et al., 1992; Yoot al., 1993; Grossmann et al., 1995; Arnold et al., 1998), theiscriminating seat-belt segment of the �-chain (Keutmann etl., 1989; Santa Coloma and Reichert, 1990; Campbell et al.,991; Dias et al., 1994; Moyle et al., 1994; Grossmann et al.,997), the �L2 loop (Erickson et al., 1990; Leinung et al., 1991;iu and Dias, 1996) and the �L2 loop (Keutmann et al., 1987).he crystal structure also provides an explanation for manyspects of specificity, again consistent with mutational analy-es (Fan and Hendrickson, 2005a,b). In our reading of Moylet al. (1995, 2004) model, wherein the hormone only makesontacts with the rim of the LRR structure and not with the �-heet surface, this theoretical alternative is at odds with the bodyf mutational analysis. Incidentally, the concordance betweenutagenesis results and the crystal structure does not simply

eflect a mode of binding restricted to FSHRHB and alterna-ively spliced receptor variants (Moyle et al., 2005). Most ofhe mutagenesis experiments were performed using full-lengtheceptors on cell surfaces (Bhowmick et al., 1996, 1999; Songt al., 2001; Smits et al., 2002, 2003; Vischer et al., 2003a,b).

Probes of protein–protein interfaces, such as made byutations or by challenges with peptides or antibodies, are

ftentimes not completely incisive. For examples, disruptionsay be due to misfolding rather than intended local perturba-

ion, and chimeras may retain functions of the host protein asell as giving gain-of-function to that of the substitution. It isnly prudent to be cautious in interpreting such experiments, butqually it is foolish to be dismissive of sound evidence. Moylet al. (2005) argues that experiments implicating involvementf the FSH� C-terminus in receptor interactions are likelyue to disturbance of the interface between FSH� and FSH�,nd thereby dismisses this as a corroboration of the crystaltructure. In fact, however, the crystal structures of free hCGWu et al., 1994; Lapthorn et al., 1994), free FSH (Fox et al.,001) and receptor-complexed FSH (Fan and Hendrickson,005a) have widely different FSH� C-terminal conformations

ut virtually identical heterodimeric associations. Moreover,his FSH�:FSH� interface has exceptionally large buriedurface (4080 A2), exceptionally low buried charge density0.37 charges/nm2) and high shape complementary (0.74) with

Cell

3tCiwafIoi

ltit

4

wsahccrwd

cati�atscavtSiem

wintccrbSui

A

GPsJSAYhgwaF

R

A

A

B

B

C

CC

D

D

D

D

E

F

F

Q.R. Fan, W.A. Hendrickson / Molecular and

5 inter-chain hydrogen bonds. There is nothing from the struc-ures to support the assertion of heterodimer perturbabtion from-terminal modification. Moyle also dismisses incisive exper-

ments that implicate 88DSDS91 of FSH� (Dias et al., 1994)ith K179 of FSHR (Smits et al., 2003; Vischer et al., 2003a)

s evidence, consistent with their match in the crystal structure,or specificity determination in a universal mode of binding.nstead, mistakenly we think, he uses the relaxed specificityf the resulting chimeras as evidence for distinctive modes ofnteraction for the different gonadotropins (Moyle et al., 2005).

To be sure, the crystal structure of the FSH–FSHRHB complexeaves many unanswered questions about signal transductionhrough gonadotropin receptors. We feel that additional exper-mental structures have a better chance at further illuminatinghe process than additional theoretical modeling.

. Conclusions

Human FSH forms a homogeneous complex with FSHRHBhen co-expressed in insect cells from a baculovirus con-

truct. The two components are specifically associated with onenother, interacting with an affinity comparable to that of theormone for the intact receptor on cell surfaces. The purifiedomplex separates sharply and symmetrically on gel filtrationolumns and in native gel electrophoresis, and it crystallizeseadily. The crystal structure was refined at 2.9 A resolutionith excellent stereochemical ideality and agreement with theiffraction data.

Like other LRR proteins, FSHRHB is a curved solenoidal tubeharacterized by a tightly packed, continuous hydrophobic corend a flat, parallel �-sheet. This is a robust structural scaffoldhat is unlikely to be perturbed in overall shape by ligand bind-ng or lattice interactions. FSH binds to the slightly concave,-sheet surface of the receptor where it interacts with all 10 par-llel �-strands of FSHRHB. The linker peptide that fuses FSH�o FSH� is disordered but positioned away from the contacturface. The hormone–receptor interface is large and intimatelyomplementary, consistent with the observed high affinity ofssociation. This interface has features compatible with a uni-ersal but specific mode for glycoprotein-hormone binding toheir respective receptors, consistent with many mutational tests.pecific interactions observed at the hormone–receptor interface

n the FSH–FSHRHB crystal structure agree well with structurallements implicated in binding by mutagenesis studies of hor-one binding to intact receptors on cell membranes.We conclude that the crystal structure of FSH complexed

ith FSHRHB provides an authentic representation of the bind-ng of FSH to intact FSHR. This is not to say that there areo other interactions between hormone and receptor; indeed,he hormone does most likely interact with transmembrane heli-al and/or cysteine-cluster linker domains of the receptor in theourse of receptor activation. But just as hormone binding and

eceptor activation have long been known to be largely separa-le (Simoni et al., 1997; Dias et al., 2002; Ascoli et al., 2002;zkudlinski et al., 2002), so too are the structural features thatnderlie these phenomena. Concerns raised about the authentic-ty of the FSH–FSHRHB structure are unfounded.

F

F

ular Endocrinology 260–262 (2007) 73–82 81

cknowledgements

We thank the organizers of the International Conference ononadotropins and Receptors, Drs. Steven Birken and Daviduett, for the invitation to speak, Drs. James Dias, Gilbert Vas-art, Deborah Segaloff, Mario Ascoli, Marco Muda, Xuliangiang, Axel Themmen, Galina Kovalevskaya, Jan Bogerd,abine Costagliola, Tae Ji, Prema Narayan, George Bousfield,lfredo Ulloa-Aguirre, Irving Boime, David Ben-Menahem,ongsheng Li and many other participants at the conference forelpful discussion. We also thank the referees for helpful sug-estions and Xuliang Jiang for discussions about dimers. Thisork was supported in part by NIH grant GM68671. Q.R.F. was

n Agouron Institute fellow of the Jane Coffin Childs Memorialoundation.

eferences

rnold, C.J., Liu, C., Lindau-Shepard, B., Losavio, M.L., Patrascu, M.T., Dias,J.A., 1998. The human follitropin �-subunit C terminus collaborates with a�-subunit cystine noose and an �-subunit loop to assemble a receptor-bindingdomain competent for signal transduction. Biochemistry 37, 1762–1768.

scoli, M., Fanelli, F., Segaloff, D.L., 2002. The lutropin/choriogonadotropinreceptor, a 2002 perspective. Endocr. Rev. 23, 141–174.

howmick, N., Huang, J., Puett, D., Isaacs, N.W., Lapthorn, A.J., 1996. Determi-nation of residues important in hormone binding to the extracellular domainof the luteinizing hormone/chorionic gonadotropin receptor by site-directedmutagenesis and modeling. Mol. Endocrinol. 10, 1147–1159.

howmick, N., Narayan, P., Puett, D., 1999. Identification of ionizable aminoacid residues on the extracellular domain of the lutropin receptor involvedin ligand binding. Endocrinology 140, 4558–4563.

ampbell, R.K., Dean-Emig, D.M., Moyle, W.R., 1991. Conversion of humanchoriogonadotropin into follitropin by protein chemistry. Proc. Natl. Acad.Sci. U.S.A. 88, 760–764.

arson, M., 1997. Ribbons. Methods Enzymol. 277, 493–505.hen, F., Wang, Y., Puett, D., 1992. The carboxy-terminal region of the glyco-

protein hormone �-subunit: contributions to receptor binding and signalingin human chorionic gonadotropin. Mol. Endocrinol. 6, 914–919.

ias, J.A., Zhang, Y., Liu, X., 1994. Receptor binding and functional propertiesof chimeric human follitropin prepared by an exchange between a smallhydrophilic intercysteine loop of human follitropin and human lutropin. J.Biol. Chem. 269, 25289–25294.

ias, J.A., Van Roey, P., 2001. Structural biology of human follitropin and itsreceptor. Arch. Med. Res. 32, 510–519.

ias, J.A., Cohen, B.D., Lindau-Shepard, B., Nechamen, C.A., Peterson, A.J.,Schmidt, A., 2002. Molecular, structural, and cellular biology of follitropinand follitropin receptor. Vitam. Horm. 64, 249–322.

umas, J.J., Kumar, R., McDonagh, T., Sullivan, F., Stahl, M.L., Somers, W.S.,Mosyak, L., 2004. Crystal structure of the wild-type von Willebrand fac-tor A1-glycoprotein Ib� complex reveals conformation differences with acomplex bearing von Willebrand disease mutations. J. Biol. Chem. 279,23327–23334.

rickson, L.D., Rizza, S.A., Bergert, E.R., Charlesworth, M.C., McCormick,D.J., Ryan, R.J., 1990. Synthetic alpha-subunit peptides stimulate testos-terone production in vitro by rat Leydig cells. Endocrinology 126,2555–2560.

an, Q.R., Hendrickson, W.A., 2005a. Structure of human follicle-stimulatinghormone in complex with its receptor. Nature 433, 269–277.

an, Q.R., Hendrickson, W.A., 2005b. Structural biology of glycoprotein hor-mones and their receptors. Endocrine 26, 179–188.

otiadis, D., Liang, Y., Filipek, S., Saperstein, D.A., Engel, A., Palczewski, K.,2004. The G protein-coupled receptor rhodopsin in the native membrane.FEBS Lett. 564, 281–288.

ox, K.M., Dias, J.A., Van Roey, P., 2001. Three-dimensional structure of humanfollicle-stimulating hormone. Mol. Endocrinol. 15, 378–389.

8 d Cell

G

G

H

J

J

K

K

K

K

K

K

K

K

L

L

L

M

M

M

M

N

P

S

S

S

S

S

S

S

S

S

S

T

U

U

V

V

W

2 Q.R. Fan, W.A. Hendrickson / Molecular an

rossmann, M., Szkudlinski, M.W., Zeng, H., Kraiem, Z., Ji, I., Tropea, J.E., Ji,T.H., Weintraub, B.D., 1995. Role of the carboxy-terminal residues of thealpha-subunit in the expression and bioactivity of human thyroid-stimulatinghormone. Mol. Endocrinol. 9, 948–958.

rossmann, M., Szkudlinski, M.W., Wong, R., Dias, J.A., Ji, T.H., Weintraub,B.D., 1997. Substitution of the seat-belt region of the thyroid-stimulatinghormone (TSH) �-subunit with the corresponding regions of choriogo-nandotropin or follitropin confers luteotropic but not follitropic activity tochimeric TSH. J. Biol. Chem. 272, 15532–15540.

uizinga, E.G., Tsuji, S., Romijn, R.A., Schiphorst, M.E., de Groot, P.G., Sixma,J.J., Gros, P., 2002. Structures of glycoprotein Ib� and its complex with vonWillebrand factor A1 domain. Science 297, 1176–1179.

iang, X., Dreano, M., Buckler, D.R., Cheng, S., Ythier, A., Wu, H., Hendrickson,W.A., Tayar, N.E., 1995. Structural predictions for the ligand-binding regionof glycoprotein hormone receptors and the nature of hormone–receptor inter-actions. Structure 3, 1341–1353.

ones, T.A., Zou, J.Y., Cowan, S.W., Kjeldgaard, M., 1991. Improved methodsfor building protein models in electron density maps and the location oferrors in the models. Acta Crystallogr. A 47, 110–119.

ajava, A.V., Vassart, G., Wodak, S.J., 1995. Modeling of the three-dimensionalstructure of proteins with the typical leucine-rich repeats. Structure 3,867–877.

eutmann, H.T., Charlesworth, M.C., Mason, K.A., Ostrea, T., Johnson, L.,Ryan, R.J., 1987. A receptor-binding region in human choriogo nadotropin/lutropin beta subunit. Proc. Natl. Acad. Sci. U. S. A. 84, 2038–2042.

eutmann, H.T., Mason, K.A., Kitzmann, K., Ryan, R.J., 1989. Role of the� 93–100 determinant loop sequence in receptor binding and biologicalactivity of human leutinizing hormone and chorionic gonadotropin. Mol.Endocrinol. 3, 526–531.

leinau, G., Jaschke, H., Neumann, S., Lattig, J., Paschke, R., Krause, G., 2004.Identification of a novel epitope in the thyroid-stimulating hormone receptorectodomain acting as intramolecular signaling interface. J. Biol. Chem. 279,51590–51600.

leywegt, G.J., 1996. Use of non-crystallographic symmetry in protein structurerefinement. Acta Crystallogr. D 52, 842–857.

obe, B., Deisenhofer, J., 1993. Crystal structure of porcine ribonucleaseinhibitor, a protein with leucine-rich repeats. Nature 366, 751–756.

obe, B., Deisenhofer, J., 1995. A structural basis of the interactions betweenleucine-rich repeats and protein ligands. Nature 374, 183–186.

unishima, N., Shimada, Y., Tsuji, Y., Sato, T., Yamamoto, M., Kumasaka, T.,Nakanishi, S., Jingami, H., Morikawa, K., 2000. Structural basis of gluta-mate recognition by a dimeric metabotropic glutamate receptor. Nature 407,971–977.

apthorn, A.J., Harris, D.C., Littlejohn, A., Lustbader, J.W., Canfield, R.E.,Machin, K.J., Morgan, F.J., Isaacs, N.W., 1994. Crystal structure of humanchorionic gonadotropin. Nature 369, 455–461.

einung, M.C., Reed, D.K., McCormick, D.J., Ryan, R.J., Morris, J.C., 1991.Further characterization of the receptor-binding region of the thyroid-stimulating hormone � subunit: a comprehensive synthetic peptide study ofthe �-subunit 26–46 sequence. Proc. Natl. Acad. Sci. U.S.A. 88, 9707–9711.

iu, C., Dias, J.A., 1996. Long loop residues 33–58 in the human glycopro-tein hormone common � subunit contain structural components for subunitheterodimerization and human follitropin–receptor binding. Arch. Biochem.Biophys. 329, 127–135.

oyle, W.R., Campbell, R.K., Myers, R.V., Bernard, M.P., Han, Y., Wang, X.,1994. Co-evolution of ligand–receptor pairs. Nature 368, 251–255.

oyle, W.R., Campbell, R.K., Rao, S.N.V., Ayad, N.G., Bernard, M.P., Han,Y., Wang, Y., 1995. Model of human chorionic gonadotropin and lutropinreceptor interaction that explains signal transduction of the glycoproteinhormones. J. Biol. Chem. 270, 20020–20031.

oyle, W.R., Xing, Y., Lin, W., Cao, D., Myers, R.V., Kerrigan, J.E., Bernard,

M.P., 2004. Model of glycoprotein hormone receptor ligand binding andsignaling. J. Biol. Chem. 279, 44442–44459.

oyle, W.R., Lin, W., Myers, R.V., Cao, D., Kerrigan, J.E., Bernard, M.P.,2005. Models of glycoprotein hormone receptor interaction. Endocrine 26,189–205.

Y

ular Endocrinology 260–262 (2007) 73–82

arayan, P., Wu, C., Puett, D., 1995. Functional expression of yokedhuman chorionic gonadotropin in baculovirus-infected insect cells. Mol.Endocrinol. 9, 1720–1726.

ierce, J.G., Parsons, T.F., 1981. Glycoprotein hormones: structure and function.Annu. Rev. Biochem. 50, 465–495.

anta Coloma, T.A., Reichert, L.E.J., 1990. Identification of a follicle-stimulating hormone receptor-binding region in hFSH-�-(81–95) using syn-thetic peptides. J. Biol. Chem. 265, 5037–5042.

chmidt, A., MacColl, R., Lindau-Shepard, B., Buckler, D.R., Dias, J.A., 2001.Hormone-induced conformational change of the purified soluble hormonebinding domain of follitropin receptor complexed with single chain fol-litropin. J. Biol. Chem. 276, 23373–23381.

chubert, W.D., Urbanke, C., Ziehm, T., Beier, V., Machner, M.P., Domann,E., Wehland, J., Chakraborty, T., Heinz, D.W., 2002. Structure of Internalin,a major invasion protein of Listeria monocytogenes, in complex with itshuman receptor E-cadherin. Cell 111, 825–836.

imoni, M., Gromoll, J., Nieschlag, E., 1997. The follicle-stimulating hormonereceptor: biochemistry, molecular biology, physiology, and pathophysiology.Endocr. Rev. 18, 739–773.

mits, G., Govaerts, C., Nubourgh, I., Pardo, L., Vassart, G., Costagliola, S.,2002. Lysine 183 and glutamic acid 157 of the TSH receptor: two interactingresidues with a key role in determining specificity toward TSH and humanCG. Mol. Endocrinol. 16, 722–735.

mits, G., Campillo, M., Govaerts, C., Janssens, V., Richter, C., Vassart, G.,Pardo, L., Costagliola, S., 2003. Glycoprotein hormone receptors: determi-nants in leucine-rich repeats responsible for ligand specificity. EMBO J. 22,2692–2703.

ohn, J., Youn, H., Jeoung, M., Koo, Y., Yi, C., Ji, I., Ji, T.H., 2003. Orientationof follicle-stimulating hormone (FSH) subunits complexed with the FSHreceptor. J. Biol. Chem. 278, 47868–47876.

ong, Y.S., Ji, I., Beauchamp, J., Isaacs, N.W., Ji, T.H., 2001. Hormone interac-tions to leu-rich repeats in the gonadotropin receptors. J. Biol. Chem. 276,3426–3435.

ugahara, T., Pixley, M.R., Minami, S., Perlas, E., Ben-Menahem, D.,Hsueh, A.J., Boime, I., 1995. Biosynthesis of a biologically activesingle peptide chain containing the human common � and chorionicgonadotropin � subunits in tandem. Proc. Natl. Acad. Sci. U.S.A. 92, 2041–2045.

zkudlinski, M.W., Fremont, V., Ronin, C., Weintraub, B.D., 2002.Thyroid-stimulating hormone and thyroid-stimulating hormone receptorstructure–function relationships. Physiol. Rev. 82, 473–502.

hemmen, A.P.N., Huhtaniemi, I.T., 2000. Mutations of gonadotropins andgonadotropin receptors: elucidating the physiology and pathophysiology ofpituitary–gonadal function. Endocr. Rev. 21, 551–583.

ff, S., Clemetson, J.M., Harrison, T., Clemetson, K.J., Emsley, J., 2002. Crys-tal structure of the platelet glycoprotein Ib� N-terminal domain revealsan unmasking mechanism for receptor activation. J. Biol. Chem. 277,35657–35663.

rizar, E., Montanelli, L., Loy, T., Bonomi, M., Swillens, S., Gales, C., Bou-vier, M., Smits, G., Vassart, G., Costagliola, S., 2005. Glycoprotein hormonereceptors: link between receptor homodimerization and negative coopera-tivity. EMBO J. 24, 1954–1964.

ischer, H.F., Granneman, J.C.M., Bogerd, J., 2003a. Opposite contribu-tion of two ligand-selective determinants in the N-terminal hormone-binding exodomain of human gonadotropin receptors. Mol. Endocrinol. 17,1972–1981.

ischer, H.F., Granneman, J.C.M., Noordam, M.J., Mosselman, S., Bogerd, J.,2003b. Ligand selectivity of gonadotropin receptors. J. Biol. Chem. 278,15505–15513.

u, H., Lustbader, J.W., Liu, Y., Canfield, R.E., Hendrickson, W.A.,1994. Structure of human chorionic gonadotropin at 2.6 A resolution

from MAD analysis of the selenomethionyl protein. Structure 2, 545–558.

oo, J., Zeng, H., Ji, I., Murdoch, W.J., Ji, T.H., 1993. COOH-terminal aminoacids of the � subunit play common and different roles in human choriogo-nadotropin and follitropin. J. Biol. Chem. 268, 13034–13042.