Seminar: Structure of a nanobody-stabilized active state of the β2 adrenoceptor

Rasmussen, S. G. F., et al. (2011)

Presented by Bundit Boonyarit 5410210278

Dept.Chemistry, Fac.Science, Prince of Songkla University

tructure of a nanobody-stabilized active state of the β2 adrenoceptorS

April 23, 2015

Brian K. KobilkaAmerican physiologist

The Nobel Prize in Chemistry 2012 “for studies of G-protein-coupled receptors”

�2

�3

INTRODUCTION

�4

INTRODUCTION

GPCRG-Protein Coupled Receptor

First messengerSignal molecule (such as epinephrine)

Receptor (GPCR)

G protein Adenylyl cyclase

GTP

ATPcAMP Second

messenger

Protein kinase A

Signal transduction

Extracellular

Intracellular (cytoplasm)

Transmembrane

�5

INTRODUCTION

Signal molecule (such as epinephrine)

Receptor (GPCR)

G protein

GTP

NATURE REVIEWS | DRUG DISCOVERY VOLUME 3 | JULY 2004 | 593

T W E N T Y Q U E S T I O N S

Tamas Bartfai. Both inside and outside the pharmaceu-tical industry, the rhodopsin model has been combinedwith what we know about PHARMACOPHORES — which is aknowledge-rich area for several monoamine receptors— to cobble together models. However, such modelshave often been of post factum value; explaining ratherthan predicting results.

Joël Bockaert. The predictions have generally been‘correct’, but the rhodopsin crystal was obtained in thepresence of retinal, which is an inverse agonist. So far, thestructure of an ‘active’ rhodopsin molecule is still lacking.

Arthur Christopoulos. The determination of thecrystal structure of bovine rhodopsin at high resolution59

has certainly been a boon to the GPCR field. One mustbear in mind, however, that rhodopsin has low sequencesimilarity to most other GPCRs, has an inverse agonistincorporated into its structure, and the crystals obtainedwere of the receptor in its inactive state, so we should notnecessarily expect to find high degrees of concordancebetween this structure and other GPCRs. Nevertheless, agood starting point is better than none, and there havebeen very successful predictions for the structures ofsome of the receptors for bioamines and related smallmolecules based on the rhodopsin model; for example,muscarinic M1, dopamine D2, α1-adrenoceptor, hista-mine H3 and adenosine A1 receptors4,63,64. Of course, thisonly refers to the transmembrane domains of the GPCRs;we are still some way off determining the intra- andextracellular loop structures.

Jean-Philippe Pin. The structural predictions havebeen good enough for the general analysis of homo-logous protein structures, even for distantly relatedGPCRs, such as the heptahelical domain of class III(family 3/C) GPCRs, which include metabotropic glutamate, GABAB (γ-amino butyric acid, type B), Ca2+

and some taste and pheromone receptors (see, forexample, REF. 65). However, one should take into con-sideration that the rhodopsin structure corresponds toa fully inactive state of a GPCR (stabilized by an inverseagonist), such that the various active states cannot yetbe predicted with accuracy. Accordingly, such modelsare more useful for the characterization of the bindingsites of inverse agonists. I am still not convinced thatsuch models are accurate enough for a detailed analysisof ligand-binding sites and drug design. Recently,Didier Rognan and collaborators have used such arhodopsin-based model of the vasopressin V2 receptorfor in silico screening of possible new agonists andantagonists66. This approach allowed them to identifyknown agonists and antagonists hidden in the chemicaldatabase, which suggests that it is a promising technique.However, whether new leads can be identified in thisway remains to be shown.

Bernard P. Roques. Few breakthroughs in the field ofGPCR structural determination have emerged from mol-ecular modelling using the rhodopsin structure, exceptsome data about the intra-cytoplasmic loops involved in

Gα

N

C

CPlasma membrane GPCR

β γ

A N

GPCR

Gα Gα

Gα

Gα

Gα

βγ

βγ

GDP

βγ

GTP

GTP

GDP

GTP

GDP

Effector

Pi

βγ

a

b

c

d

e

GPCR

B Agonist

Agonist

Figure 3 | GPCR–Gαα fusion proteins as a model system for the analysis of receptor–G-protein coupling. A | Schematic of a G-protein-coupled receptor (GPCR)–Gα fusion protein.The GPCR carboxyl terminus (C) is fused to the amino terminus (N) of Gα22, ensuring closeproximity and defined stoichiometry of the two coupling partners. GPCRs can activate G proteinslinearly (that is, one GPCR molecule activates one G protein) rather than catalytically (that is, oneGPCR molecule activates several G proteins), which substantiates the relevance of the fusionprotein technique26,85. Fusion proteins also enable the study of coupling between GPCR speciesisoforms58 or intra-species polymorphic forms185 and a particular Gα isoform. Correspondingly,the coupling of a particular receptor to various Gα isoforms can also be analysed26,85. Crosstalkbetween fusion proteins to non-fused Gα and between different fusion protein molecules199,200

must be taken into consideration. B | G-protein cycling. Rate-limiting receptor-promoted GDPdissociation (a) is followed by ternary complex formation (b). The GPCR then catalyses the bindingof GTP to Gα (c), which disrupts the ternary complex, causing dissociation of the G-proteinheterotrimer into Gα and βγ. Both entities regulate the activity of effector systems (d; see alsoTABLES 1, 2 and 3). G-protein activation is terminated by hydrolysis of the Gα-bound GTP to GDPand Pi (e). GPCR–Gα fusions are useful for studying steps b, c, d and e of the G-protein cycle(to our knowledge, GDP dissociation (a) has not been directly studied with this system). Figure prepared by Roland Seifert.

N-terminal segment (Extracellular)7 Transmembrane domain (TM1-TM7) to form TM core3 exoloops3-4 cytoloopsC-terminal segment (Intracellular)

1 2 3 4 5 6 7

GPCRG-Protein Coupled Receptor

EllisClare (2004)

Extracellular

Intracellular (cytoplasm)

�6

INTRODUCTION

β2 adrenoceptor β2 adrenergic receptorOR



different types of mammalian cell has been useful.Using these preparations, a great number of pharmaco-logically useful probes have been created (includingtheir radiolabelled counterparts when necessary), whichhas allowed the determination of receptor distribution,their principal functions and agonist versus antagonistactivity. Expression in cells of both human GPCRs andtheir corresponding rodent proteins is an importantstep towards a rapid physiopharmacological, and puta-tively clinical, study of the target. At this time, there are

Bernard P. Roques. GPCR gene knockouts haveallowed, in many cases, the determination of the maineffects resulting from the TONIC or PHASIC stimulationof the target. A typical example is the knockout of theµ-opioid receptor14, which demonstrated definitivelythe major role of this receptor in analgesia and opioidaddiction, and the putative clinical interest of selectiveδ-opioid receptor stimulation15–17.

In addition, the development of HTS methods byexpressing GPCRs (including orphan receptors) in

Box 1 | G-protein-coupled-receptor families

G-protein-coupled receptors (GPCRs) are the largestfamily of cell-surface receptors, and transduce the signalsmediated by a diverse range of signalling molecules,including ions, biogenic amines, peptides and lipids, as wellas photons, to mediate alterations of intracellular function.GPCRs can be divided into different families on the basisof their structural and genetic characteristics (see GPCRDatabase online). GPCRs in the different families do notshare significant sequence similarity, although they all havethe characteristic seven transmembrane (TM) domains.The figure shows schematic representations of receptormonomers (GPCRs have been shown to exist as dimers oroligomers), and illustrates some key structural aspects ofthe three main GPCR families known at present. Family 1(panel a; also referred to as family A or the rhodopsin-likefamily) is by far the largest subgroup and containsreceptors for odorants, important neurotransmitters, suchas dopamine and serotonin, as well as neuropeptides andglycoprotein hormones. Receptors of family 1 arecharacterized by several highly conserved amino acids(some of which are indicated in the diagram by red circles)and a disulphide bridge that connects the first and secondextracellular loops (ECLs). Most of these receptors alsohave a PALMITOYLATED cysteine in the carboxy-terminal tail,which serves as an anchor to the membrane (orange zig-zag). The recent determination of the crystal structure ofrhodopsin59 has indicated that the transmembranedomains of family 1 receptors are ‘tilted’ and ‘kinked’ asshown, due to the presence of amino acids such as prolinethat distort the helical transmembrane domain. Family 2 orfamily B GPCRs (panel b) are characterized by a relativelylong amino terminus, which contains several cysteines thatform a network of disulphide bridges. Their morphology issimilar to some family 1 receptors, but the palmitoylationsite is missing and the conserved residues and motifs aredifferent from the conserved residues in the family 1receptors. Little is known about the orientation of the TMdomains, but given the divergence in amino-acid sequence,they are likely to be different from family 1 receptors.Ligands for family 2 GPCRs include hormones, such asglucagon, secretin and parathyroid hormone. Family 3(panel c) contains the metabotropic glutamate, the Ca2+-sensing and the GABAB (γ-aminobutyric acid, type B)receptors. These receptors are characterized by a long amino terminus and carboxyl tail. The ligand-bindingdomain (shown in yellow) is located in the amino terminus, which has been shown by the crystal structure of themetabotropic glutamate receptor to form a disulphide-linked dimer103. It is thought to resemble a Venus fly trap,which can open and close with the agonist bound inside. Except for two cysteines in ECL1 and ECL2 that form aputative disulphide bridge, the family 3 receptors do not have any of the features that characterize family 1 and 2receptors. A unique characteristic of these receptors is that the third intracellular loop is short and highlyconserved. At present, little is known about the orientation of the TM domains. Box adapted with permission fromREF. 40 © 2002 Macmillan Magazines Limited. Susan R. George

b Family 2NH2

1 23

45

67

COOH

C

C

C

C CC

C C

C

P

K

5

6

4

2

c Family 3

NH2

13

COOH

NEA

C

7

C

P

W

COOH

a Family 1

P

NH2

CC

DPN

DR Y

C

7

6 5

1

23

4

GPCR family A

EllisClare (2004)

�7

INTRODUCTION


Receptor (GPCR)

G protein

GTP

First messenger



different types of mammalian cell has been useful.Using these preparations, a great number of pharmaco-logically useful probes have been created (includingtheir radiolabelled counterparts when necessary), whichhas allowed the determination of receptor distribution,their principal functions and agonist versus antagonistactivity. Expression in cells of both human GPCRs andtheir corresponding rodent proteins is an importantstep towards a rapid physiopharmacological, and puta-tively clinical, study of the target. At this time, there are

Bernard P. Roques. GPCR gene knockouts haveallowed, in many cases, the determination of the maineffects resulting from the TONIC or PHASIC stimulationof the target. A typical example is the knockout of theµ-opioid receptor14, which demonstrated definitivelythe major role of this receptor in analgesia and opioidaddiction, and the putative clinical interest of selectiveδ-opioid receptor stimulation15–17.

In addition, the development of HTS methods byexpressing GPCRs (including orphan receptors) in

Box 1 | G-protein-coupled-receptor families

G-protein-coupled receptors (GPCRs) are the largestfamily of cell-surface receptors, and transduce the signalsmediated by a diverse range of signalling molecules,including ions, biogenic amines, peptides and lipids, as wellas photons, to mediate alterations of intracellular function.GPCRs can be divided into different families on the basisof their structural and genetic characteristics (see GPCRDatabase online). GPCRs in the different families do notshare significant sequence similarity, although they all havethe characteristic seven transmembrane (TM) domains.The figure shows schematic representations of receptormonomers (GPCRs have been shown to exist as dimers oroligomers), and illustrates some key structural aspects ofthe three main GPCR families known at present. Family 1(panel a; also referred to as family A or the rhodopsin-likefamily) is by far the largest subgroup and containsreceptors for odorants, important neurotransmitters, suchas dopamine and serotonin, as well as neuropeptides andglycoprotein hormones. Receptors of family 1 arecharacterized by several highly conserved amino acids(some of which are indicated in the diagram by red circles)and a disulphide bridge that connects the first and secondextracellular loops (ECLs). Most of these receptors alsohave a PALMITOYLATED cysteine in the carboxy-terminal tail,which serves as an anchor to the membrane (orange zig-zag). The recent determination of the crystal structure ofrhodopsin59 has indicated that the transmembranedomains of family 1 receptors are ‘tilted’ and ‘kinked’ asshown, due to the presence of amino acids such as prolinethat distort the helical transmembrane domain. Family 2 orfamily B GPCRs (panel b) are characterized by a relativelylong amino terminus, which contains several cysteines thatform a network of disulphide bridges. Their morphology issimilar to some family 1 receptors, but the palmitoylationsite is missing and the conserved residues and motifs aredifferent from the conserved residues in the family 1receptors. Little is known about the orientation of the TMdomains, but given the divergence in amino-acid sequence,they are likely to be different from family 1 receptors.Ligands for family 2 GPCRs include hormones, such asglucagon, secretin and parathyroid hormone. Family 3(panel c) contains the metabotropic glutamate, the Ca2+-sensing and the GABAB (γ-aminobutyric acid, type B)receptors. These receptors are characterized by a long amino terminus and carboxyl tail. The ligand-bindingdomain (shown in yellow) is located in the amino terminus, which has been shown by the crystal structure of themetabotropic glutamate receptor to form a disulphide-linked dimer103. It is thought to resemble a Venus fly trap,which can open and close with the agonist bound inside. Except for two cysteines in ECL1 and ECL2 that form aputative disulphide bridge, the family 3 receptors do not have any of the features that characterize family 1 and 2receptors. A unique characteristic of these receptors is that the third intracellular loop is short and highlyconserved. At present, little is known about the orientation of the TM domains. Box adapted with permission fromREF. 40 © 2002 Macmillan Magazines Limited. Susan R. George

b Family 2NH2

1 23

45

67

COOH

C

C

C

C CC

C C

C

P

K

5

6

4

2

c Family 3

NH2

13

COOH

NEA

C

7

C

P

W

COOH

a Family 1

P

NH2

CC

DPN

DR Y

C

7

6 5

1

23

4

Rhodopsin: 11-cis-retinal

β2 adrenergic receptor: adrenaline

Biogenic amine receptors: serotonin, dopamine, histamine

(agonist)

(agonist)

(agonist)

GPCR family A

EllisClare (2004)

�8


Receptor (GPCR)

G protein

GTP

First messenger

(agonist)

β2 adrenergic receptor: adrenaline

AsthmaPulmonary disease

GPCR family A

INTRODUCTION

Out

In

Ca2+PKA

P

L-type Ca2+channel

Gs

ERK 1,2

Adenylatecyclase

cAMP

PDE

Gi

0

Biol

ogic

al re

spon

se (%

)

Log drug concentration

100

50

Full agonist

Partial agonist

Inverse agonist

Neutral antagonist

+–

G-protein uncouplingby phosphorylation

and arrestin binding

P PP

Desensitized 2AR2AR and G-protein-dependent signalling

PPP

Recycling back to membrane

P

Targeted for degradationin lysosomes

MAP kinasepathway

Gene expression

ATP

Arrestin

G-protein-independent

signallingBasal activity

PKC

GRK

Arrestin

Internalization via clathrin-mediated

endocytosis

inverse agonists reduce the level of basal or constitutive activity below that of the unliganded receptor. The wide spectrum of ligand effica-cies for individual GPCRs shows that efficient energy transfer between the binding pocket and the site of G-protein interaction is dependent on multiple interactions between receptor and hormone, and requires more than simply occupying the binding site. Further, biophysical stud-ies on purified fluorescently labelled β2AR demonstrated that partial and full agonists containing different subsets of functional groups sta-bilize distinct conformational states by engaging with distinct subsets of conformational switches in the receptor12–14. These findings lead to a complex picture of GPCR activation in which a distinct conformation stabilized by a ligand’s structure determines the efficacy towards a spe-cific pathway. Many GPCRs can stimulate multiple signalling systems, and specific ligands can have different relative efficacies to different pathways15. In the extreme case, even opposite activities for different signalling pathways are observed: for β2AR, agonists for the arrestin/MAP kinase pathway are also inverse agonists for the classical Gαs/cAMP/PKA pathway7,16. GPCRs are no longer thought to behave as simple two-state switches. Rather, they are more like molecular rhe-ostats, able to sample a continuum of conformations with relatively closely spaced energies17. Specific ligands achieve varying efficacies for different signalling pathways by stabilizing particular sets of conforma-tions that can interact with specific effectors.

The inactive states of four GPCRs The first insights into the structure of GPCRs came from two-dimensional crystals of rhodopsin18,19. These structures revealed the general architec-ture of the seven transmembrane helices. However, given the conform-ational complexity of ligand-activated GPCRs, it is not surprising that it took so long to obtain three-dimensional crystal structures. As detailed in Box 1, a variety of different protein-modification and engineering approaches have contributed to recent advances in GPCR crystallography. We now have inactive-state structures of four GPCRs for comparison: human β2AR bound to the high-affinity inverse agonists carazolol20–22 and timolol23; avian β1AR bound to the antagonist cyanopindolol24; the human A2A adenosine receptor bound to the antagonist ZM241385 (ref. 25); and bovine rhodopsin26–28 containing the covalently bound inverse agonist 11-cis retinal. The superpositions of different receptors using the homologous transmembrane domains led to root mean squared deviation (r.m.s.d.) values of less than 3 Å. This degree of overlap indicates that these four proteins have a similar overall architecture, yet the diver-gences are still high enough to signify important differences in helical packing interactions (Fig. 2).

Extracellular surfaces and ligand-binding sitesAs might be expected from the functional differences between the recep-tors, the most significant structural divergences lie in the extracellular

Figure 1 | Signal transduction in G-protein-coupled receptors. Diverse signalling pathways regulated by the type 2 beta adrenergic receptor (β2AR). The β2AR can activate two G proteins, Gαs and Gαi (part of the Gs and Gi heterotrimers, respectively), which differentially regulate adenylate cyclase. Adenylate cyclase generates cyclic AMP (cAMP), which activates protein kinase A (PKA), a kinase that regulates the activity of several cellular proteins including the L-type Ca2+ channel and the β2AR. cAMP second messenger levels are downregulated by specific phosphodiesterase proteins (PDEs). Activation of the β2AR also leads to phosphorylation by a G-protein-coupled receptor kinase (GRK) and subsequent coupling to

arrestin. Arrestin is a signalling and regulatory protein that promotes the activation of extracellular signal-regulated kinases (ERK), prevents the activation of G proteins and promotes the internalization of the receptor through clathrin-coated pits. PKC, protein kinase C. The inset shows classification of ligand efficacy for GPCRs. Many GPCRs exhibit basal, agonist-independent activity. Inverse agonists inhibit this activity, and neutral antagonists have no effect. Agonists and partial agonists stimulate biological responses above the basal activity. Efficacy is not directly related to affinity; for example, a partial agonist can have a higher affinity for a GPCR than a full agonist.

357

NATURE|Vol 459|21 May 2009 REVIEW INSIGHT

Kobilka.indd NS.indd 357Kobilka.indd NS.indd 357 13/5/09 16:49:4213/5/09 16:49:42

�9Rosenbaum et al. (2009)

Drug as biological response

Antagonist: any ligand that blocks binding of endogenous agonists to the receptor.

Agonist: a ligand that binds to and activates a receptor and elicits a physiological response.

Basal activity: a physiological response that occurs in the absence of an agonist due to a proportion of the receptor being in the activated state.

Inverse agonist: a ligand that binds to a receptor and inhibits or eliminates a physiological response.

Neutral antagonist: a ligand that neither stimulates receptor activation nor inhibits basal activity.

INTRODUCTION

�10

INTRODUCTION

INACTIVE STATE

BI-167107 (agonist)

Receptor (β2 adrenergic

receptor)

Stimulatory G protein (GS)

GTP

Recent crystal structure provide insights into inactive states of

several GPCRs.

�11

INTRODUCTION

CHALLENGE

Agonist are much less efficient at stabilizing the active state of β2AR, making its difficult to capture this

state in a crystal structure.

ACTIVE STATE

BI-167107 (agonist)


receptor)


GTP

Instability

�12

INTRODUCTION

ACTIVE STATE

BI-167107 (agonist)


receptor)

T4 lysozyme (T4L)

Nanobody (Nb80)

BI-167107 (agonist)


receptor)


GTP

STRATEGY

Develop a binding protein that preferentially binds to and

stabilises an active conformation, acting as a surrogate for Gs

(Nanobody, Nb80).

�13

OBJECTIVE

To study interaction in active state of β2 adrenergic receptor that stabilized by nanobody

�14

EXPERIMENTS

�15

EXPERIMENTS

ACTIVE STATE

BI-167107 (agonist)

Receptor (β2 adrenergic receptor)

from Homo sapiens

T4 lysozyme (T4L) from Enterobacteria phage T4

Nanobody (Nb80) from Lama glama

High-density lipoprotein

(HDL)

BI-167107 (agonist)


receptor)


GTP

Extracellular

Intracellular

Transmembrane

�16

Nanobody (Nb80)

Antibody fragment consisting of a single monomeric variable antibody domain that secreted by a type of white blood cell called a plasma cell

150 kDa

Camelidae Antibody

14 kDaFab region

Fc region

Heavy chain

Light chain

EXPERIMENTS

�17

ACTIVE STATE

BI-167107 (agonist)


from Homo sapiens

T4 lysozyme (T4L) from Enterobacteria phage T4

Nanobody (Nb80) from Lama glama


(HDL)

Nanobody (Nb80)

Small size (14 kDa) Showed G-protein-like properties upon binding to both wild-type β2AR and β2AR-T4L

PDB ID: 3DWT

Camelidae Antibody

14 kDa

EXPERIMENTS

�18

EXPERIMENTS

1 2 3

Preparation of β2AR-T4L and Nb80 for crystallography

Crystallography

Analyses

PDB ID: 3P0G

solubilized b2AR365 (WT receptor truncatedat 365) and b2AR-T4L were each labeled withmonobromobimane, which has been used pre-viously to monitor conformational changes of theb2AR (21). The addition of the agonist isopro-terenol to purified b2AR365 induces a decreasein fluorescence intensity and a shift in the wave-length at maximum intensity (lmax) for the at-tached bimane probe (Fig. 2B and table S2).These changes in intensity and lmax are con-sistent with an agonist-induced increase in polar-ity around bimane. A smaller change is observedwith the partial agonist salbutamol, whereas theinverse agonist ICI-118,551 had little effect. Forthe b2AR-T4L, there are subtle differences in thebaseline spectrum of the bimane-labeled fusionprotein, as might be expected if the environmentaround Cys2656.27 is altered by T4L. However,the full agonist isoproterenol induces a qualita-tively similar decrease in intensity and rightwardshift in lmax. Thus, the presence of the fusedT4L does not prevent agonist-induced confor-mational changes. The partial agonist salbuta-mol induced larger responses in b2AR-T4L thanwere observed in WT b2AR, and there was asmall increase in fluorescence in response to theinverse agonist ICI-118,551. These properties areobserved in CAMs (15, 22) and are consistentwith the higher affinities for agonists and partial

agonists exhibited by b2AR-T4L. Therefore, weconclude that the T4L fusion induces a partialconstitutively active phenotype in the b2AR, prob-ably caused by changes at the cytoplasmic endsof helices V and VI.

Comparison between b2AR-T4L and b2AR-Fab structures. The b2AR-T4L fusion strategyis validated by a comparison of its structure tothat of WT b2AR complexed with a Fab thatrecognizes a three-dimensional epitope consistingof the N- and C-terminal ends of ICL3, deter-mined at an anisotropic resolution of 3.4 Å/3.7 Å(23). Figure 3A illustrates the similarity betweenthe fusion and antibody complex approaches tob2AR crystallization, in that both strategies relyon attachment (covalent or noncovalent, respec-tively) of a soluble protein partner between helicesV and VI. A major difference between the twostructures is that the extracellular loops and thecarazolol ligand could not be modeled in theb2AR-Fab complex, whereas these regions areresolved in the structure of b2AR-T4L. None-theless, it is clear that the T4L insertion does notsubstantially alter the receptor. Superposition ofthe two structures (fig. S4) illustrates that the trans-membrane helices of the receptor componentsare very similar (root mean square deviation =0.8 Å for 154 common modeled transmembraneCa positions versus 2.3 Å between b2AR-T4L

and the 154 equivalent residues in rhodopsin),especially when the modest resolution of theFab complex is taken into account.

There is one major difference between theFab-complex and chimeric-receptor structuresthat can be attributed to the presence of T4L.The cytoplasmic end of helix VI is pulled out-ward as a result of the fusion to the C terminusof T4L, which alters the packing of Phe2646.26

at the end of helix VI (Fig. 3B). In the Fab-complex b2AR, interactions between Phe2646.26

and residues in helix V, helix VI, and ICL2 maybe important in maintaining the b2AR in the basalstate. The loss of these packing interactions inb2AR-T4L could contribute to the higher agonistbinding affinity characteristic of a CAM.

An unexpected difference between the struc-ture of rhodopsin and the b2AR-T4L involvesthe sequence E/DRY (24) found at the cytoplasmicend of helix III in 71% of class A GPCRs. Inrhodopsin, Glu1343.49 and Arg1353.50 form anetwork of hydrogen bond and ionic interac-tions with Glu2476.30 at the cytoplasmic end ofhelix VI. These interactions have been referredto as an “ionic lock” that stabilizes the inactivestate of rhodopsin and other class A members(25). However, the arrangement of the homolo-gous residues is considerably different in b2AR-T4L: Arg1313.50 interacts primarily with Asp1303.49

CD F

FTNQ YA

A I A SS I V

S F V VP L VI M V F

D

Q

Y SVRVFQEAKRQLQKI

YG N

GNSS

EGTN

GSQ

VHY

EQQEKKENKKL

LCE

D L P G T E FD V G HQQ

GGTS P VD I N D

RGG

QQ S

TS

N

LLSDDNC

365

190

230

400

Extracellular

Helix V

Intracellular

I HVV I N V

I F FP L W C

L T FT G M I

G L

E YVI L L NW I GY V N S

G F NL I YR S

QD

N L IR

K

I CT K L A

K H EK L C F

KS

PD

RL

TGHGG

DGR

FR

IA

FQ

EL

L

CL

RRS

SR

SL

G Y A K

P

260

300

330

350

Helix VI

Helix VII

ECL3

ILC3

N

ADWT

LNKY

A

B

C

D

E

12

3

45

T4-Lysozyme

C

ECL2BA

Positive Control:

FLAG-β1AR

D3

C3

D5

D1

NegativeControl:pCDNA3

β2AR-T4L

M1 M1+DAPI

Fig. 1. Design and optimization of the b2AR-T4L fusion protein. (A) Thesequence of the region of the b2AR targeted for insertion of a crystallizabledomain is shown, and the positions of the junctions between the receptorand T4L (red) for various constructs are indicated. The sequences that wereinitially replaced or removed are faded. Red lines are shown after every tenthresidue. ECL, extracellular loop. (B) Immunofluorescence images of HEK293

cells expressing selected fusion constructs. (Left) M1 anti-FLAG signal corre-sponding to antibody bound to the N terminus of the receptor. (Right) Samesignal merged with blue emission from 4´,6´-diamidino-2-phenylindole(nuclear staining for all cells). Plasma membrane staining is observed inthe positive control, D3, and D1, whereas C3 and D5 are retained in theendoplasmic reticulum.

23 NOVEMBER 2007 VOL 318 SCIENCE www.sciencemag.org1268

RESEARCH ARTICLES

�19

EXPERIMENTS

1 Preparation of β2AR-T4L and Nb80 for crystallography

Rosenbaum, D.M. et al.(2004)

solubilized b2AR365 (WT receptor truncatedat 365) and b2AR-T4L were each labeled withmonobromobimane, which has been used pre-viously to monitor conformational changes of theb2AR (21). The addition of the agonist isopro-terenol to purified b2AR365 induces a decreasein fluorescence intensity and a shift in the wave-length at maximum intensity (lmax) for the at-tached bimane probe (Fig. 2B and table S2).These changes in intensity and lmax are con-sistent with an agonist-induced increase in polar-ity around bimane. A smaller change is observedwith the partial agonist salbutamol, whereas theinverse agonist ICI-118,551 had little effect. Forthe b2AR-T4L, there are subtle differences in thebaseline spectrum of the bimane-labeled fusionprotein, as might be expected if the environmentaround Cys2656.27 is altered by T4L. However,the full agonist isoproterenol induces a qualita-tively similar decrease in intensity and rightwardshift in lmax. Thus, the presence of the fusedT4L does not prevent agonist-induced confor-mational changes. The partial agonist salbuta-mol induced larger responses in b2AR-T4L thanwere observed in WT b2AR, and there was asmall increase in fluorescence in response to theinverse agonist ICI-118,551. These properties areobserved in CAMs (15, 22) and are consistentwith the higher affinities for agonists and partial

agonists exhibited by b2AR-T4L. Therefore, weconclude that the T4L fusion induces a partialconstitutively active phenotype in the b2AR, prob-ably caused by changes at the cytoplasmic endsof helices V and VI.

Comparison between b2AR-T4L and b2AR-Fab structures. The b2AR-T4L fusion strategyis validated by a comparison of its structure tothat of WT b2AR complexed with a Fab thatrecognizes a three-dimensional epitope consistingof the N- and C-terminal ends of ICL3, deter-mined at an anisotropic resolution of 3.4 Å/3.7 Å(23). Figure 3A illustrates the similarity betweenthe fusion and antibody complex approaches tob2AR crystallization, in that both strategies relyon attachment (covalent or noncovalent, respec-tively) of a soluble protein partner between helicesV and VI. A major difference between the twostructures is that the extracellular loops and thecarazolol ligand could not be modeled in theb2AR-Fab complex, whereas these regions areresolved in the structure of b2AR-T4L. None-theless, it is clear that the T4L insertion does notsubstantially alter the receptor. Superposition ofthe two structures (fig. S4) illustrates that the trans-membrane helices of the receptor componentsare very similar (root mean square deviation =0.8 Å for 154 common modeled transmembraneCa positions versus 2.3 Å between b2AR-T4L

and the 154 equivalent residues in rhodopsin),especially when the modest resolution of theFab complex is taken into account.

There is one major difference between theFab-complex and chimeric-receptor structuresthat can be attributed to the presence of T4L.The cytoplasmic end of helix VI is pulled out-ward as a result of the fusion to the C terminusof T4L, which alters the packing of Phe2646.26

at the end of helix VI (Fig. 3B). In the Fab-complex b2AR, interactions between Phe2646.26

and residues in helix V, helix VI, and ICL2 maybe important in maintaining the b2AR in the basalstate. The loss of these packing interactions inb2AR-T4L could contribute to the higher agonistbinding affinity characteristic of a CAM.

An unexpected difference between the struc-ture of rhodopsin and the b2AR-T4L involvesthe sequence E/DRY (24) found at the cytoplasmicend of helix III in 71% of class A GPCRs. Inrhodopsin, Glu1343.49 and Arg1353.50 form anetwork of hydrogen bond and ionic interac-tions with Glu2476.30 at the cytoplasmic end ofhelix VI. These interactions have been referredto as an “ionic lock” that stabilizes the inactivestate of rhodopsin and other class A members(25). However, the arrangement of the homolo-gous residues is considerably different in b2AR-T4L: Arg1313.50 interacts primarily with Asp1303.49

CD F

FTNQ YA

A I A SS I V

S F V VP L VI M V F

D

Q

Y SVRVFQEAKRQLQKI

YG N

GNSS

EGTN

GSQ

VHY

EQQEKKENKKL

LCE

D L P G T E FD V G HQQ

GGTS P VD I N D

RGG

QQ S

TS

N

LLSDDNC

365

190

230

400

Extracellular

Helix V

Intracellular

I HVV I N V

I F FP L W C

L T FT G M I

G L

E YVI L L NW I GY V N S

G F NL I YR S

QD

N L IR

K

I CT K L A

K H EK L C F

KS

PD

RL

TGHGG

DGR

FR

IA

FQ

EL

L

CL

RRS

SR

SL

G Y A K

P

260

300

330

350

Helix VI

Helix VII

ECL3

ILC3

N

ADWT

LNKY

A

B

C

D

E

12

3

45

T4-Lysozyme

C

ECL2BA

Positive Control:

FLAG-β1AR

D3

C3

D5

D1

NegativeControl:pCDNA3

β2AR-T4L

M1 M1+DAPI

Fig. 1. Design and optimization of the b2AR-T4L fusion protein. (A) Thesequence of the region of the b2AR targeted for insertion of a crystallizabledomain is shown, and the positions of the junctions between the receptorand T4L (red) for various constructs are indicated. The sequences that wereinitially replaced or removed are faded. Red lines are shown after every tenthresidue. ECL, extracellular loop. (B) Immunofluorescence images of HEK293

cells expressing selected fusion constructs. (Left) M1 anti-FLAG signal corre-sponding to antibody bound to the N terminus of the receptor. (Right) Samesignal merged with blue emission from 4´,6´-diamidino-2-phenylindole(nuclear staining for all cells). Plasma membrane staining is observed inthe positive control, D3, and D1, whereas C3 and D5 are retained in theendoplasmic reticulum.

23 NOVEMBER 2007 VOL 318 SCIENCE www.sciencemag.org1268

RESEARCH ARTICLES

�20

EXPERIMENTS


expressed in Sf-9 insect cell cultures infected with β2AR-T4L baculovirus

Rosenbaum, D.M. et al.(2004)

�21

EXPERIMENTS


expressed in the periplasm of Escherichia coli strain WK6

PDB ID: 3DWT

one ends up at this stage with a continuous volume of moltenlipid in direct contact with the Teflon tip of the plunger. Inthis case, the volume of lipid in the syringe can be readdirectly from the ml markings on the barrel. Given that mono-olein has a density of 0.942 g ml!1 (see ref. 34), the weightof lipid in the syringe can be calculated from the recordedvolume. Alternatively, the loaded syringe can be reweighed andthe lipid mass obtained by taring.m CRITICAL STEP Incubate the lipid for just long enoughto melt the lipid for easy handling. Extensive heating canlead to lipid degradation.

6| Attach the coupler, with both ferrules in place, to theopen end of the lipid-loaded syringe. Carefully screw thecoupler into position.m CRITICAL STEP The coupler must be tightened to adegree that the syringe contents do not leak under thepressure of mechanical mixing but not overtightened, which candeform the ferrule and/or break the glued seal between theglass barrel and the steel barrel termination, resulting in leakage during mixing.

7| Use the plunger to force the molten lipid up the barrel and slowly and gently into the narrow bore needle at the core of thecoupler. If the needle is slightly overfilled, the lipid will be seen beading out at its open end. By backing up the plunger slightly,the excess lipid can be withdrawn into the coupler until it is just flush with the tip of the coupler needle. In this case, the syr-inge/coupler unit is fully loaded and ready to be combined with the protein solution-loaded syringe at Step 12.

8| Calculate the volume of protein solution that can/should be used to form the cubic phase at or close to full hydration. Formonoolein at 20 1C, full hydration with water occurs at close to B40% (wt/wt) water14 (Fig. 1). Thus, for example, if theweight of lipid in the lipid-loaded syringe is 21 mg, then ((21/3) " 2 ¼ ) 14 mg, corresponding to B14 ml, of protein solutionwill be required.

9| Centrifuge the protein solution in a 0.5-ml Eppendorf tube at 14,000g for 5–10 min at 4 1C to remove large aggregatesbefore setting up crystallization trials.m CRITICAL STEP The amount of detergent in the protein solution used for in meso crystallization trials should be minimized toavoid destabilizing the bicontinuous mesophase (see also Troubleshooting section).

10| The second 100-ml Hamilton syringe is used to house the protein solution (or control solution required for prescreeninganalyses; see Box 3). Withdraw the plunger in the syringe to the 20–30 ml mark and slowly inject 14 ml of protein solution

p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e

Figure 7 | Ways to set up in meso crystallization trials using commercialplates. The types of plates used include (a) microbatch, (b) sitting drop and(c,d) hanging drop. In d and e, a sandwich is made of the mesophase (red) byplacing a small glass coverslip (hatched) (d) below or (e) above the bolus.The precipitant solution is colored a shaded pale blue, the vacuum grease ispurple and the sealing tape is gray.

BOX 5 | OPTIMIZING CRYSTALLOGENESISThe conditions to be optimized for in meso crystallogenesis are, by and large, those implemented with soluble and membrane proteins by themore standard vapor diffusion and batch techniques. These include buffer, salt, polymer and additive type and concentration, as well as pH andtemperature (seem CRITICAL STEP at Step 36 regarding temperature). In addition to these more standard approaches, optimization in meso canalso be implemented at the level of the composition of the lipid bilayer from which nuclei form and crystals grow. Thus, the identity of the lipidused to form the cubic phase can be changed or a lipid additive can be included. In the former case, there are several alternativemonoacylglycerols to monoolein, the default lipid, that are available commercially (e.g., from Nu-Chek Prep Inc.) and that can be used. Theseinclude monopalmitolein and monovaccenin. To augment this effort, we have a lipid synthesis and purification program in place in thelaboratory and several novel monoacylglycerols have proven useful at the screening stage35,56. From a practical perspective, these alternativelipids are used in exactly the same way as has been described in Steps 1–13 for the default lipid monoolein.

When an additive lipid is to be included as part of the optimization process, care must be taken to ensure that the amount added to thehosting lipid does not destabilize the cubic phase. We have investigated this for several lipids and found that the cubic phase of monoolein canaccommodate up to 20 mol% DOPE, 20 mol% DOPC, 25 mol% cholesterol, 5 mol% DOPS and 7 mol% cardiolipin57. In that study, phaseidentification and microstructure characterization were performed using small-angle X-ray scattering (Box 2). However, optical clarity andtexture between crossed polarizers, as described in Box 2, are other less-demanding but less-definitive ways of making this determination.A protocol for preparing such lipid mixtures, with monoolein as the host, is described in Box 4.

It should be noted that several membrane proteins have lipophilic cofactors and ligands and their mode of interaction with the target can beof interest. They can also stabilize the protein and render the complex more crystallizable. Such ligands can be combined with the lipid beforecubic-phase preparation in the same way that lipid additives are incorporated, as described in Box 4.

716 | VOL.4 NO.5 | 2009 | NATURE PROTOCOLS

PROTOCOL

Sitting drop1

�22

EXPERIMENTS

2 Crystallization

Caffrey, M. and V. Cherezov (2009)







p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e






PROTOCOL

hanging drop2







p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e






PROTOCOL

hanging drop3







p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e






PROTOCOL

4







p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e






PROTOCOL

Sitting drop







p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e






PROTOCOL

�23

EXPERIMENTS

2 Crystallization BI-167107-bound β2AR-T4L and Nb80 in 1:1.2 molar ratio

monoolein containing 10% cholesterol

50 nL protein:lipid drop overlaid with 0.8 µL precipitant solution

hanging drop


protein:lipid 1:1.5 ratio

protein

lipid

�24

EXPERIMENTS

2 Crystallization







p uorG gn ih si lbu

P eru taN 900 2

©na

ture

prot

ocol

s/

moc.erut an.w

ww//:ptth

a b c

d e






PROTOCOL

reservoir solution (36 to 44% PEG 400, 100 mM Tris pH 8.0, 4% DMSO, 1% 1,2,3-heptanetriol)

Crystals grew to full size within 7 to 10 days.

stored in liquid nitrogen with reservoir solution as cryoprotectant


�25

EXPERIMENTS

3 Analyses Bimane fluorescence spectroscopyTo approve nanobody-stabilized β2AR active state

that exhibits G protein-like behaviour towards the b2AR. Tylopoda(camels, dromedaries and llamas) have developed a unique class offunctional antibody molecules that are devoid of light chains16. A nano-body (Nb) is the recombinant minimal-sized intact antigen-bindingdomain of such a camelid heavy chain antibody and is approximately25% the size of a conventional Fab fragment. To generate receptor-specific nanobodies, a llama was immunized with purified agonist-bound b2AR reconstituted at high density into phospholipid vesicles.A library of single-chain nanobody clones was generated and screenedagainst agonist bound receptor. We identified seven clones that recog-nized agonist-bound b2AR. Of these, Nb80 was chosen because itshowed G-protein-like properties upon binding to both wild-typeb2AR and b2AR–T4L, the b2AR–T4 lysozyme fusion protein used toobtain the high-resolution inactive state crystal structure7,9.

We compared the effect of Nb80 with Gs on b2AR structure andagonist binding affinity. b2AR was labelled at the cytoplasmic end oftransmembrane helix 6 (TM6) at Cys 265 with the fluorophore mono-bromobimane and reconstituted into high-density lipoprotein (HDL)particles. TM6 moves relative to TM3 and TM5 upon agonist activa-tion (Fig. 1a), and we have shown previously that the environmentaround bimane covalently linked to Cys 265 changes with both ago-nist binding and G protein coupling, resulting in a decrease in fluor-escence intensity and a red shift in lmax

15. As shown in Fig. 1b, thecatecholamine agonist isoproterenol and Gs both stabilize an active-like conformation, but the effect of Gs is greater in the presence ofisoproterenol, consistent with the cooperative interactions of agonistand Gs on b2AR structure. Nb80 alone has an effect on bimane fluor-escence and lmax of unliganded b2AR that is similar to that of Gs(Fig. 1c). This effect was not observed in b2AR bound to the inverseagonist ICI-118,551. The effect of Nb80 was increased in the presenceof 10 mM isoproterenol. These results show that Nb80 does not recog-nize the inactive conformation of the b2AR, but binds efficiently to

agonist-occupied b2AR and produces a change in bimane fluor-escence that is indistinguishable from that observed in the presenceof Gs and isoproterenol.

Figure 1d and e shows the effect of Gs and Nb80 on agonist affinityfor b2AR. b2AR was reconstituted into HDL particles and agonistcompetition binding experiments were performed in the absence orpresence of Nb80 and Gs. In the absence of either protein, isoproterenolhas an inhibition constant (Ki) of 107 nM. In the presence of Gs twoaffinity states are observed, because not all of the b2AR is coupled to Gs.In the Gs-coupled state the affinity of isoproterenol increases by 100-fold (Ki 5 1.07 nM) (Fig. 1d and Supplementary Table 1). Similarly, inthe presence of Nb80 the affinity of isoproterenol increases by 95-fold(Ki 5 1.13 nM) (Fig. 1e and Supplementary Table 1). In contrast, Nb80had little effect on b2AR binding to the inverse agonist ICI-118,551(Supplementary Fig. 1 and Supplementary Table 1). These binding dataindicate that Nb80 stabilizes a conformation in wild-type b2AR that isvery similar to that stabilized by Gs, such that the energetic coupling ofagonist and Gs binding is faithfully mimicked by Nb80.

The high-resolution structure of the inactive state of the b2AR wasobtained with a b2AR–T4L fusion protein. We showed previously thatb2AR–T4L has a higher affinity for isoproterenol than wild-typeb2AR7.Nevertheless, in the presence of Nb80 the affinity increased by 60-fold,resulting in an affinity (Ki 5 0.56 nM) comparable to that of wild-typeb2AR bound to Nb80 (Fig. 1f and Supplementary Table 1). Althoughwe cannot study G protein coupling in b2AR–T4L due to steric hind-rance by T4L, the results show that T4L does not prevent binding ofNb80, and the nearly identical Ki values for agonist binding to wild-typeb2AR and b2AR–T4L in the presence of Nb80 indicate that Nb80stabilizes a similar conformation in these two proteins. The most likelyexplanation for the ability of Nb80 to bind to b2AR–T4L whereas Gsdoes not is the difference in size of these two proteins. Nb80 is approxi-mately 14 kDa whereas the Gs heterotrimer is approximately 90 kDa.

425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fluo

resc

ence

inte

nsity

(n

orm

aliz

ed to

unl

igan

ded)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI

mBB-β2AR/HDL with Gs

mBB-β2AR/HDL with Nb80

–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4

β2AR/HDL β2AR–T4L/HDLβ2AR/HDL

+ Nb80 Control Control + Nb80 + Gs + GTPγS

+ Gs

0

20

40

60

80

100

0

20

40

60

80

100

Log ISO concentration (M) Log ISO concentration (M) Log ISO concentration (M)

b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6

Monobromobimane (mBBr)

Wavelength (nm)

Figure 1 | Effect of Nb80 on b2AR structure and function. a, The cartoonillustrates the movement of the environmentally-sensitive bimane probeattached to Cys 2656.27 in the cytoplasmic end of TM6 from a more buried,hydrophobic environment to a more polar, solvent-exposed position duringreceptor activation that results in a decrease in fluorescence in Fig. 1b–c andSupplementary Fig. 2c, d. b, c, Fluorescence emission spectra showing ligand-induced conformational changes of monobromobimane-labelled b2ARreconstituted into high density lipoprotein particles (mBB-b2AR/HDL) in theabsence (black solid line) or presence of full agonist isoproterenol (ISO, green

wide dashed line), inverse agonist ICI-118,551 (ICI, black dashed line), Gsheterotrimer (red solid line), nanobody-80 (Nb80, blue solid lines), andcombinations of Gs with ISO (red wide dashed line), Nb80 with ISO (blue widedashed line), and Nb80 with ICI (blue dashed line). d2f, Ligand binding curvesfor ISO competing against [3H]-dihydroalprenolol ([3H]-DHA) for d, b2AR/HDL reconstituted with Gs heterotrimer in the absence or presence GTPcS;e, b2AR/HDL in the absence and presence of Nb80; and f, b2AR–T4L/HDL inthe absence and presence of Nb80. Error bars represent standard errors.

RESEARCH ARTICLE

1 7 6 | N A T U R E | V O L 4 6 9 | 1 3 J A N U A R Y 2 0 1 1

Macmillan Publishers Limited. All rights reserved©2011

change conformation to activation

Cys265 on TM6 of β2AR

�26

EXPERIMENTS

3 Analyses Bimane fluorescence spectroscopy


T4 lysozyme (T4L)

Nanobody (Nb80)


(HDL)

Ligand


receptor)


GTP

Extracellular

Transmembrane

Intracellular

Ligand

mBB-β2AR/HDL with GsUnligand

Gs Isoproterenol (ISO) (agonist)

Gs + ISO

mBB-β2AR/HDL with Nb80ICI-188,551 (ICI) (inverse agonist) Nb80 + ICI Nb80 Nb80 + ISO

To approve nanobody-stabilized β2AR active state

�27

EXPERIMENTS

3 Analyses Scintillation counter (with [3H]-DHA and ISO)


T4 lysozyme (T4L)

Nanobody (Nb80)


(HDL)

Ligand


receptor)


GTP

Extracellular

Transmembrane

Intracellular

Ligand

β2AR/HDL

Gs Gs + GTP

Nb80 Control

β2AR/HDL and β2AR-T4L/HDL

To approve nanobody-stabilized β2AR active state

�28

EXPERIMENTS

3 Analyses X-ray crystallographyTo study agonist-stabilized changes in the β2AR


T4 lysozyme (T4L)

Nanobody (Nb80)


(HDL)

Ligand


receptor)


GTP

Extracellular

Transmembrane

Intracellular

Ligand

Opsin (GPCR family A) - inverse agonist carazolol-bound β2AR-T4L (β2AR-Cz) - BI-167107 agonist-bound and Nb80-stabilized β2AR-T4L (β2AR-Nb80)

�29

RESULTS


from Homo sapiens

�30

RESULTS

by bimane fluorescence spectroscopy


GTP

Extracellular

Intracellular

Ligand

Isoproterenol (ISO)

(agonist)


(HDL)







425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE



Nanobody-stabilized β2AR active state

�31

RESULTS


Extracellular

Intracellular

(agonist)


T4 lysozyme (T4L)

Nanobody (Nb80)

Ligand(inverse agonist)

Isoproterenol (ISO) ICI-188,551 (ICI)


(HDL)







425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE




�32

RESULTS








425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE



Isoproterenol (ISO) ICI-188,551 (ICI)

(agonist)







425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE



(inverse agonist)


�33

RESULTS


Isoproterenol (ISO)







425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fluo

resc

ence

inte

nsity

(n

orm

aliz

ed to

unl

igan

ded)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE









425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE









425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE



Scintillation counter (with [3H]-DHA and ISO)

�34

RESULTS

Agonist-stabilised changes in the β2ARX-ray crystallography

Efforts to crystallize β2AR-Gs complex and β2AR-T4L bound to

BI-167107 and other agonists failed to produce crystals of

sufficient quality for structure to determination.

Attempted to crystallize BI-167107 bound to β2AR and β2AR-T4L in complex

with Nb80

�35

RESULTS

Extracellular

Intracellular


T4 lysozyme (T4L)

Nanobody (Nb80)

Ligand


(HDL)


PDB ID: 3P0G

BI-167107 (agonist)

β2AR

Nanobody

High affinity b2AR agonistTo stabilize further the active state of the b2AR, we screened over 50commercial and proprietary b2AR ligands. Of these, BI-167107(Boehringer Ingelheim) had the most favourable efficacy, affinity andoff-rate profile. BI-167107 is a full agonist that binds to the b2AR with adissociation constant Kd of 84 pM (Supplementary Fig. 2a and b). Asshown in Supplementary Fig. 2c and d, BI-167107 induces a largerchange in the fluorescence intensity and lmax of bimane bound toCys 265 than does the agonist isoproterenol. Moreover, the rate ofdissociation of BI-167107 was extremely slow. Displacement of BI-167107 with an excess of the neutral antagonist alprenolol required150 h to complete, compared with 5 s for isoproterenol.

Crystallization of b2AR–T4L–Nb80 complexThe b2AR was originally crystallized bound to the inverse agonistcarazolol using two different approaches. The first crystals wereobtained from b2AR bound to a Fab fragment that recognized anepitope composed of the amino and carboxyl terminal ends of thethird intracellular loop connecting TMs 5 and 6 (ref. 8). In the secondapproach, the third intracellular loop was replaced by T4 lysozyme(b2AR–T4L)7. Efforts to crystallize b2AR–Fab complex and b2AR–T4L bound to BI-167107 and other agonists failed to produce crystalsof sufficient quality for structure determination. We thereforeattempted to crystallize BI-167107 bound to b2AR and b2AR–T4L

in complex with Nb80. Although crystals of both complexes wereobtained in lipid bicelles and lipidic cubic phase (LCP), high-resolutiondiffraction was only obtained from crystals ofb2AR–T4L–Nb80 grownin LCP. These crystals grew at pH 8.0 in 39–44% PEG400, 100 mMTris, 4% DMSO and 1% 1,2,3-heptanetriol.

A merged data set at 3.5 A was obtained from 23 crystals(Supplementary Table 2). The structure was solved by molecularreplacement using the structure of the carazolol-bound b2AR and ananobody as search models. Supplementary Fig. 3a shows the packingof the b2AR–T4L–Nb80 complex in the crystal lattice. The receptorhas interactions with lattice neighbours in several directions, and isrelatively well ordered (Supplementary Fig. 3a and b), with readilyinterpretable electron density for most of the polypeptide. Nb80 bindsto the cytoplasmic end of the b2AR, with the third complementarity-determining region (CDR) loop projecting into the core of the recep-tor (Fig. 2a, and Supplementary Fig. 4).

Agonist-stabilized changes in the b2ARFigure 2 b–d compares the inactive b2AR structure (from the carazo-lol bound b2AR–T4L structure) with the agonist-bound b2AR com-ponent of the b2AR–T4L–Nb80 complex. The largest differences arefound at the cytoplasmic face of the receptor, with outward displace-ment of TM5 and TM6 and an inward movement of TM7 and TM3 inthe b2AR–T4L–Nb80 complex relative to the inactive structure. There

a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å

β2AR–Nb80 β2AR–Nb80

D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus

β2AR–Nb80β2AR–CzNb80

Figure 2 | Comparison of the agonist-Nb80 stabilized crystal structures ofthe b2AR with inverse agonist bound b2AR and opsin. The structure ofinverse agonist carazolol-bound b2AR–T4L (b2AR–Cz) is shown in blue withthe carazolol in yellow. The structure of BI-167107 agonist-bound and Nb80-stabilized b2AR–T4L (b2AR–Nb80) is shown in orange with BI-167107 ingreen. These two structures were aligned using the PyMOL align function.a, Side view of the b2AR–Nb80 complex with b2AR in orange and CDRs ofNb80 in light blue (CDR1) and blue (CDR3). b, Side view of the superimposedstructures showing significant structural changes in the intracellular and Gprotein facing part of the receptors. c, Comparison of the extracellular ligand

binding domains showing modest structural changes. d, Cytoplasmic viewshowing the ionic lock interaction between Asp 3.49 and Arg 3.50 of the DRYmotif in TM3 is broken in the b2AR–Nb80 structure. The intracellular end ofTM6 is moved outward and away from the core of the receptor. The arrowindicates an 11.4 A change in distance between the a-carbon of Glu 6.30 in thestructures of b2AR–Cz and b2AR–Nb80. The intracellular ends of TM3 andTM7 move towards the core by 4 and 2.5 A, respectively, while TM5 movesoutward by 6 A. e, The b2AR–Nb80 structure superimposed with the structureof opsin crystallized with the C-terminal peptide of Gt (transducin)2. PyMOL(http://www.pymol.org) was used for the preparation of all structure figures.

ARTICLE RESEARCH

1 3 J A N U A R Y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 1 7 7


�36

RESULTS

Extracellular

Intracellular


T4 lysozyme (T4L)

Nanobody (Nb80)

Ligand


(HDL)

Agonist-stabilised changes in the β2AR

BI-167107 (agonist)

β2AR-Nb80 (active)

carazolol (inverse agonist)

β2AR-Cz (inactive)

X-ray crystallography






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



�37

RESULTS

Extracellular

Intracellular


T4 lysozyme (T4L)

Nanobody (Nb80)

Ligand


(HDL)

Agonist-stabilised changes in the β2AR

β2AR-Nb80 (active)

β2AR-Cz (inactive)

11.4Å

6Å 2.5Å

4Å

β2AR can interact with Nb80 (change conformation from inactive to active state

X-ray crystallography

agonist has a longer alkyl substituent on the amine, which ends with aphenyl ring that lies in a hydrophobic pocket formed by Trp 1093.28,Phe 1935.32 and Ile 3097.36.

The greatest difference between inactive and active structures in theligand-binding site is an inward bulge of TM5 centred aroundSer 2075.46, whose Ca position shifts by 2.1 A (Fig. 4a). In addition,there are smaller inward movements of TM6 and TM7. The basalactivity shown by the b2AR indicates that the protein structure sur-rounding the binding pocket is relatively dynamic in the absence ofligand, such that it samples active and inactive conformations. Thepresence of Pro 2115.50 in the following turn, which cannot form ahydrogen bond with the backbone at Ser 2075.46, is likely to lower thebarrier to the transition between the conformations observed in thepresence of carazolol and BI-167107. There are extensive interactionsbetween the carbonyl oxygen, amine and hydroxyl groups on theheterocycle of BI-167107 and Ser 2035.42 and 2075.46 in TM5, as wellas Asn 2936.55 in TM6 and Tyr 3087.35 in TM7. In contrast, there isonly one polar interaction between the nitrogen in the heterocycle ofcarazolol and Ser 2035.42. Interactions of Ser 2035.42, Ser 2045.43 andSer 2075.46 with catecholamine hydroxyls have been proposed, on thebasis of mutagenesis studies showing that these serines are importantfor agonist binding and activation18,19. Whereas Ser 2045.43 does notinteract directly with the ligand, it forms a hydrogen bond withAsn 2936.55 on TM6, which is in turn linked to Tyr 3087.35 of extra-cellular loop 3 (ECL3) (Fig. 3a). This tyrosine packs againstPhe 1935.32 of ECL2, and both residues move to close off the ligand-binding site from the extracellular space.

Asn 2936.55 contributes to enantiomeric selectivity for catecholamineagonists20. The b-OH of BI-167107 does not interact with Asn 2936.55,but forms hydrogen bonds with Asp 1133.32 and Asn 3127.39, similar towhat is observed for carazolol in the inactive structure. The chirality oftheb-OH influences the spatial position of the aromatic ring system inb2AR ligands, so the effect of Asn 2936.55 on b-OH enantiomericselectivity may arise from its direct interaction with the aromatic ringsystem of the ligand, as well as its positioning of Ser 2045.43 andTyr 3087.35, which also interact with this portion of the ligand.However, BI-167107 is not a catecholamine, and it is possible thatthe b-OH of catecholamine agonists, such as adrenaline and noradre-naline, has a direct interaction with Asn 2936.55, because mutation ofAsn 2936.55 has a stronger influence on the preference for the chiralityof the b-OH of catecholamine agonists, compared with non-catecholagonists and antagonists20.

Trp 6.48 is highly conserved in Family A GPCRs, and it has beenproposed that its rotameric state has a role in GPCR activation (rotamer

toggle switch)21. We observe no change in the side chain rotamer ofTrp 2866.48 in TM6 (Fig. 4a), which lies near the base of the ligand-binding pocket, although its position shifts slightly in concert withrearrangements of nearby residues Ile 1213.40 and Phe 2826.44.Although there is spectroscopic evidence for changes in the environ-ment of Trp 6.48 upon activation of rhodopsin22, a rotamer change is notobserved in the crystal structures of rhodopsin and low-pH opsin.Moreover, recent mutagenesis experiments on the serotonin 5HT4receptor demonstrate that Trp 6.48 is not required for activation of thisreceptor by serotonin23. These observations indicate that, althoughchanges in hydrophobic packing alter the conformation of the receptorin this region, changes in the Trp 6.48 rotamer do not occur as part of theactivation mechanism.

It is interesting to speculate how the small changes around theagonist-binding pocket are coupled to much larger structural changesin the cytoplasmic regions of TMs 5, 6 and 7 that facilitate binding ofNb80 and Gs. A potential conformational link is shown in Fig. 4.Agonist interactions with Ser 2035.42 and 2075.46 stabilize a receptorconformation that includes a 2.1-A inward movement of TM5 at posi-tion 2075.46 and 1.4-A inward movement of the conserved Pro 2115.50

relative to the inactive, carazolol-bound structure. In the inactive state,the relative positions of TM5, TM3, TM6 and TM7 are stabilized byinteractions between Pro 2115.50, Ile 1213.40, Phe 2826.44 and Asn 3187.45.The position of Pro 2115.50 observed in the agonist structure is incom-patible with this network of interactions, and Ile 1213.40 and Phe 2826.44

are repositioned, with a rotation of TM6 around Phe 2826.44 leading toan outward movement of the cytoplasmic end of TM6.

Although some of the structural changes observed in the cytoplas-mic ends of transmembrane domains of the b2AR–T4L–Nb80 com-plex arise from specific interactions with Nb80, the fact that Nb80 andGs induce or stabilize similar structural changes in the b2AR, as deter-mined by fluorescence spectroscopy and by agonist binding affinity,suggests that Nb80 and Gs recognize similar agonist-stabilized con-formations. The observation that the transmembrane domains of rho-dopsin and the b2AR undergo similar structural changes uponactivation provides further support that the agonist-bound b2AR–T4L–Nb80 represents an active conformation and is consistent witha conserved mechanism of G protein activation.

However, the mechanism by which agonists induce or stabilizethese conformational changes likely differs for different ligands andfor different GPCRs. The conformational equilibria of rhodopsin andb2AR differ, as shown by the fact that rhodopsin appears to adopt afully active conformation in the absence of a G protein24 whereasb2AR cannot15. Thus, the energetics of activation and conformational

TM5TM5

TM7 TM7

TM6TM6

TM3 TM3

Pro 211

Pro 211

Asn 318 Asn 318

Phe 282Phe 282

Ile 121Ile 121

Carazolol BI-167107β2AR–Nb80

1

2

34

5

Pro 211

Asn 318Phe 282

Ile 121

Ser 207

Ser 203

2.1 Å

Trp 286

b ca

β2AR–Cz

Figure 4 | Rearrangement of transmembrane segment packing interactionsupon agonist binding a, The BI-167107- and carazolol-bound structures aresuperimposed to show structural differences propagating from the ligand-binding pocket. BI-167107 and carazolol are shown in green and yellow,respectively. b, Packing interactions that stabilize the inactive state are observed

between Pro 211 in TM5, Ile 121 in TM3, Phe 282 in TM6 and Asn 318 in TM7.c, The inward movement of TM5 upon agonist binding destabilizes the packingof Ile 121 and Pro 211, resulting in a rearrangement of interactions betweenIle 121 and Phe 282. These changes contribute to a rotation and outwardmovement of TM6 and an inward movement of TM7.

ARTICLE RESEARCH








a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



β2AR-Nb80 (active)

β2AR-Cz (inactive)

11.4Å

2.5Å

RESULTS


�38inward bulge of TM5

centred around Ser207 2.1Å






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



β2AR-Nb80 (active)

β2AR-Cz (inactive)

11.4Å

2.5Å

agonist has a longer alkyl substituent on the amine, which ends with aphenyl ring that lies in a hydrophobic pocket formed by Trp 1093.28,Phe 1935.32 and Ile 3097.36.

The greatest difference between inactive and active structures in theligand-binding site is an inward bulge of TM5 centred aroundSer 2075.46, whose Ca position shifts by 2.1 A (Fig. 4a). In addition,there are smaller inward movements of TM6 and TM7. The basalactivity shown by the b2AR indicates that the protein structure sur-rounding the binding pocket is relatively dynamic in the absence ofligand, such that it samples active and inactive conformations. Thepresence of Pro 2115.50 in the following turn, which cannot form ahydrogen bond with the backbone at Ser 2075.46, is likely to lower thebarrier to the transition between the conformations observed in thepresence of carazolol and BI-167107. There are extensive interactionsbetween the carbonyl oxygen, amine and hydroxyl groups on theheterocycle of BI-167107 and Ser 2035.42 and 2075.46 in TM5, as wellas Asn 2936.55 in TM6 and Tyr 3087.35 in TM7. In contrast, there isonly one polar interaction between the nitrogen in the heterocycle ofcarazolol and Ser 2035.42. Interactions of Ser 2035.42, Ser 2045.43 andSer 2075.46 with catecholamine hydroxyls have been proposed, on thebasis of mutagenesis studies showing that these serines are importantfor agonist binding and activation18,19. Whereas Ser 2045.43 does notinteract directly with the ligand, it forms a hydrogen bond withAsn 2936.55 on TM6, which is in turn linked to Tyr 3087.35 of extra-cellular loop 3 (ECL3) (Fig. 3a). This tyrosine packs againstPhe 1935.32 of ECL2, and both residues move to close off the ligand-binding site from the extracellular space.

Asn 2936.55 contributes to enantiomeric selectivity for catecholamineagonists20. The b-OH of BI-167107 does not interact with Asn 2936.55,but forms hydrogen bonds with Asp 1133.32 and Asn 3127.39, similar towhat is observed for carazolol in the inactive structure. The chirality oftheb-OH influences the spatial position of the aromatic ring system inb2AR ligands, so the effect of Asn 2936.55 on b-OH enantiomericselectivity may arise from its direct interaction with the aromatic ringsystem of the ligand, as well as its positioning of Ser 2045.43 andTyr 3087.35, which also interact with this portion of the ligand.However, BI-167107 is not a catecholamine, and it is possible thatthe b-OH of catecholamine agonists, such as adrenaline and noradre-naline, has a direct interaction with Asn 2936.55, because mutation ofAsn 2936.55 has a stronger influence on the preference for the chiralityof the b-OH of catecholamine agonists, compared with non-catecholagonists and antagonists20.

Trp 6.48 is highly conserved in Family A GPCRs, and it has beenproposed that its rotameric state has a role in GPCR activation (rotamer

toggle switch)21. We observe no change in the side chain rotamer ofTrp 2866.48 in TM6 (Fig. 4a), which lies near the base of the ligand-binding pocket, although its position shifts slightly in concert withrearrangements of nearby residues Ile 1213.40 and Phe 2826.44.Although there is spectroscopic evidence for changes in the environ-ment of Trp 6.48 upon activation of rhodopsin22, a rotamer change is notobserved in the crystal structures of rhodopsin and low-pH opsin.Moreover, recent mutagenesis experiments on the serotonin 5HT4receptor demonstrate that Trp 6.48 is not required for activation of thisreceptor by serotonin23. These observations indicate that, althoughchanges in hydrophobic packing alter the conformation of the receptorin this region, changes in the Trp 6.48 rotamer do not occur as part of theactivation mechanism.

It is interesting to speculate how the small changes around theagonist-binding pocket are coupled to much larger structural changesin the cytoplasmic regions of TMs 5, 6 and 7 that facilitate binding ofNb80 and Gs. A potential conformational link is shown in Fig. 4.Agonist interactions with Ser 2035.42 and 2075.46 stabilize a receptorconformation that includes a 2.1-A inward movement of TM5 at posi-tion 2075.46 and 1.4-A inward movement of the conserved Pro 2115.50

relative to the inactive, carazolol-bound structure. In the inactive state,the relative positions of TM5, TM3, TM6 and TM7 are stabilized byinteractions between Pro 2115.50, Ile 1213.40, Phe 2826.44 and Asn 3187.45.The position of Pro 2115.50 observed in the agonist structure is incom-patible with this network of interactions, and Ile 1213.40 and Phe 2826.44

are repositioned, with a rotation of TM6 around Phe 2826.44 leading toan outward movement of the cytoplasmic end of TM6.

Although some of the structural changes observed in the cytoplas-mic ends of transmembrane domains of the b2AR–T4L–Nb80 com-plex arise from specific interactions with Nb80, the fact that Nb80 andGs induce or stabilize similar structural changes in the b2AR, as deter-mined by fluorescence spectroscopy and by agonist binding affinity,suggests that Nb80 and Gs recognize similar agonist-stabilized con-formations. The observation that the transmembrane domains of rho-dopsin and the b2AR undergo similar structural changes uponactivation provides further support that the agonist-bound b2AR–T4L–Nb80 represents an active conformation and is consistent witha conserved mechanism of G protein activation.

However, the mechanism by which agonists induce or stabilizethese conformational changes likely differs for different ligands andfor different GPCRs. The conformational equilibria of rhodopsin andb2AR differ, as shown by the fact that rhodopsin appears to adopt afully active conformation in the absence of a G protein24 whereasb2AR cannot15. Thus, the energetics of activation and conformational

TM5TM5

TM7 TM7

TM6TM6

TM3 TM3

Pro 211

Pro 211

Asn 318 Asn 318

Phe 282Phe 282

Ile 121Ile 121

Carazolol BI-167107β2AR–Nb80

1

2

34

5

Pro 211

Asn 318Phe 282

Ile 121

Ser 207

Ser 203

2.1 Å

Trp 286

b ca

β2AR–Cz

Figure 4 | Rearrangement of transmembrane segment packing interactionsupon agonist binding a, The BI-167107- and carazolol-bound structures aresuperimposed to show structural differences propagating from the ligand-binding pocket. BI-167107 and carazolol are shown in green and yellow,respectively. b, Packing interactions that stabilize the inactive state are observed

between Pro 211 in TM5, Ile 121 in TM3, Phe 282 in TM6 and Asn 318 in TM7.c, The inward movement of TM5 upon agonist binding destabilizes the packingof Ile 121 and Pro 211, resulting in a rearrangement of interactions betweenIle 121 and Phe 282. These changes contribute to a rotation and outwardmovement of TM6 and an inward movement of TM7.

ARTICLE RESEARCH



RESULTS


β2AR-Cz (inactive)

β2AR-Nb80 (active)

1.4Å inward movement

�39

also� proposed� that� interactions� between� the� aromatic� ring� of� catecholamineagonists� and Phe2906.52� in� TM6� play� a� role� in� the� stabilization of� the� activeform� of� this� switch.� While� this� mechanism� was� initially� defined� for� catechol-amine� receptors,� this� sequence� motif� is� highly� conserved� in� amine� and� opsinreceptors,� so� it� is� expected� that� this� step� in� the� activation� mechanism� will� beconserved� within� these� families.

2.� Ionic� Lock

Another� molecular� switch,� the� ionic� lock,� involves� the� interaction� betweenGlu6.�30�,� highly� conserved� in� amine� and� opsin� receptors� (>�93%),� and� theAsp3.�49�/Arg3.50� pair,� in� the� highly� conserved� (D/E)RY�motif� found� in� virtuallyall� Class� A� GPCRs� (Ballesteros� et� al.,� 2001a)� (Fig.� 7).� This� ionic� interaction� isproposed� to� hold� together� the� cytoplasmic� ends� of� TM3� and� TM6 in theresting� state� of� different� amine� receptors� (Ballesteros� et� al.,� 2�00�1a�;� Greasleyet al., 2002; Shapiro et al., 2002).� This� interaction� is� also� observed� in� the� crystalstructures of inactive rhodopsin (Li et al., 2004; Okada, 2004; Okada et al.,2002; Palczewski et al., 2000; Teller et al., 2001), and disruption of thisinteraction during activation is suggested by various biophysical (Farrenset� al., 1996;� G�et�he�r� et� al.,� 1997b),� biochemical� (Arnis� et� al.,� 1994;� Ghanouniet al., 2000; Sheikh et al., 1996, 1999), andmutagenesis (Alewijnse et al., 2000;

F�ig.� 7.� The� ionic� lock� stabilizes interactions� between� the� cytoplasmic� ends� of TM3and� TM6 in the inactive state. Agonist binding disrupts these interactions.

152 DEUPI AND KOBILKA






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



β2AR-Nb80 (active)

β2AR-Cz (inactive)

11.4Å

RESULTS


�40

ionic lock interaction between Asp130 and Arg131 of the DRY motif in TM3 is broken in

the β2AR-Nb80 structure

RESULTS


�41

Agonist interaction with Ser203 and Ser207 stabilize a receptor conformation that includes a 2.1Å inward movement in TM5

are relatively small changes in the extracellular surface (Fig. 2c). Thesecond intracellular loop (ICL2) between TM3 and TM4 adopts atwo-turn alpha helix (Fig. 2d), similar to that observed in the turkeyb1AR structure11. The absence of this helix in the inactive b2AR struc-ture may reflect crystal lattice contacts involving ICL2.

Figure 2a and Supplementary Fig. 4a–c show details of interactionof Nb80 with the cytoplasmic side of the b2AR. An eight-amino-acidsequence of CDR3 penetrates into a hydrophobic pocket formed byamino acids from TM segments 3, 5, 6 and 7. A four-amino-acidsequence of CDR1 provides additional stabilizing interactions withcytoplasmic ends of TM segments 5 and 6. CDR3 occupies a positionsimilar to the carboxyl terminal peptide of transducin in opsin2

(Supplementary Fig. 4c, d). The majority of interactions betweenNb80 and the b2AR are mediated by hydrophobic contacts.

When comparing the agonist- and inverse agonist-bound struc-tures, the largest change is observed in TM6, with an 11.4-A move-ment of the helix at Glu 2686.30 (part of the ionic lock) (superscripts inthis form indicate Ballesteros–Weinstein numbering for conservedGPCR residues17) (Fig. 2d). This large change is effected by a smallclockwise rotation of TM6 in the turn preceding the conservedPro 2886.50, enabled by the interrupted backbone hydrogen bondingat the proline and repacking of Phe 2826.44 (see below), which swingsthe helix outward.

The changes in agonist-bound b2AR–T4L–Nb80 relative to theinactive carazolol-bound b2AR–T4L are remarkably similar to those

observed between rhodopsin and opsin2,3 (Fig. 2e). The salt bridge inthe ionic lock between highly conserved Arg 1313.50 and Asp/Glu 1303.49 is broken. In opsin, Arg 1353.50 interacts with Tyr 2235.58

in TM5 and a backbone carbonyl of the transducin peptide.Arg 1313.50 of b2AR likewise interacts with a backbone carbonyl ofCDR3 of Nb80. However, Nb80 precludes an interaction betweenArg 1313.50 and Tyr 2195.58, even though the tyrosine occupies a similarposition in opsin and agonist-bound b2AR–T4L–Nb80. As in opsin,Tyr 3267.53 of the highly conserved NPxxY sequence moves into thespace occupied by TM6 in the inactive state. In carazolol-boundb2AR–T4L we observed a network of hydrogen bonding interactionsinvolving highly conserved amino acids in TMs 1, 2, 6 and 7 andseveral water molecules7. Although the resolution of the b2AR–T4L–Nb80 structure is inadequate to detect water molecules, it is clearthat the structural changes we observe would substantially alter thisnetwork.

In contrast to the relatively large changes observed in the cytoplas-mic domains of b2AR–T4L–Nb80, the changes in the agonist-bindingpocket are fairly subtle. Figure 3 shows a comparison of the bindingpockets of the inverse agonist- and agonist-bound structures. An omitmap of the ligand-binding pocket is provided in Supplementary Fig. 5.Many of the interactions between the agonist BI-167107 and theb2ARare similar to those observed with the inverse agonist carazolol. Thealkylamine and the b-OH of both ligands form polar interactions withAsp 1133.32 in TM3, and with Asn 3127.39 and Tyr 3167.43 in TM7. The

S203S207

S204Y308

N293

N312

D113

Y316

TM5

TM3

TM6

TM5

TM3

TM7

TM4

TM6 TM7

a b

β2AR–Czβ2AR–Nb80 CarazololBI-167107

F290

F193

V117

W109

V114

I309

S203S207

S204

Y308N293

N312

D113

Y316

TM4

F290

F193

V117

W109

I309

OH

NO

HN

HO

S204 5.43

S207 5.46

Y308 7.35

S203 5.42

N293 6.55

V117 3.36

F290 6.52

A200 5.39

OH

F193 5.32

F289 6.51

S204 5.43

S207 5.46

Y308 7.35

Hydrophobic contacts

Polar interactions

V114 3.33

T118 3.37

Mutation disrupts antagonist and agonist binding Mutation disrupts agonist binding

19

17

N293 6.55

V117 3.36 W286 6.48

F290 6.52

A200 5.39

Y199 5.38

F193 5.32 W109 3.28

F289 6.51

OH

BI-167107 Carazolol

OHN

I309 7.36

W109 3.28

O

c d

OH

O

H2N

O

H2N

OO

HO S203 5.42OH

O

NH2

OO

HO

D113 3.32

N312 7.39

Y316 7.43

D113 3.32

N312 7.39

Y316 7.43

H2 NH2

Figure 3 | Ligand binding pocket of BI-167107 and carazolol-bound b2ARstructures. a, b, Extracellular views of the agonist BI-167107-bound (a) andcarazolol-bound (b) structures, respectively. Residues within 4 A of one or bothligands are shown as sticks. In all panels, red and blue represent oxygen andnitrogen, respectively. c, d, Schematic representation of the interactionsbetween theb2AR and the ligands BI-167107 (c) and carazolol (d). The residues

shown here have at least one atom within 4 A of the ligand in the crystalstructures. Mutations of amino acids in orange boxes have been shown todisrupt both antagonist and agonist binding. Mutations of amino acids in blueboxes have been shown to disrupt agonist binding. Green lines indicatepotential hydrophobic interactions and orange lines indicate potential polarinteractions.

RESEARCH ARTICLE



Y3087.35

N2936.55

S2075.46

S2035.42

S2045.43

N3127.39

D1133.32

N3127.39

D1133.32

V1143.33

S2075.46

S2035.42

S2045.43

Y3087.35

N2936.55






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



β2AR-Nb80 (active)

β2AR-Cz (inactive)

Arg131

Asp130BROKEN






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



β2AR-Nb80 (active)

β2AR-Cz (inactive)4Å


�42

ionic lock interaction between Asp130 and Arg131 of the DRY motif in TM3 is broken in

the β2AR-Nb80 structure

4Å

RESULTS






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH








a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



β2AR-Nb80 (active)

β2AR-Cz (inactive)4Å


�43the ionic lock between highly conserved

Asp130 and Arg/Glu131 in TM3 is brokenβ2AR-Nb80 (active)

Opsin (active)

BROKEN

RESULTS









S203S207

S204Y308

N293

N312

D113

Y316

TM5

TM3

TM6

TM5

TM3

TM7

TM4

TM6 TM7

a b


F290

F193

V117

W109

V114

I309

S203S207

S204

Y308N293

N312

D113

Y316

TM4

F290

F193

V117

W109

I309

OH

NO

HN

HO

S204 5.43

S207 5.46

Y308 7.35

S203 5.42

N293 6.55

V117 3.36

F290 6.52

A200 5.39

OH

F193 5.32

F289 6.51

S204 5.43

S207 5.46

Y308 7.35


Polar interactions

V114 3.33

T118 3.37


19

17

N293 6.55

V117 3.36 W286 6.48

F290 6.52

A200 5.39

Y199 5.38

F193 5.32 W109 3.28

F289 6.51

OH

BI-167107 Carazolol

OHN

I309 7.36

W109 3.28

O

c d

OH

O

H2N

O

H2N

OO

HO S203 5.42OH

O

NH2

OO

HO

D113 3.32

N312 7.39

Y316 7.43

D113 3.32

N312 7.39

Y316 7.43

H2 NH2



RESEARCH ARTICLE



Y3087.35

N2936.55

S2075.46

S2035.42

S2045.43

N3127.39

D1133.32

N3127.39

D1133.32

V1143.33

S2075.46

S2035.42

S2045.43

Y3087.35

N2936.55


Alkylamine and β-OH form polar interactions with Asp113 in TM3 and Asn312 and Tyr316 in TM7

�44

RESULTS









S203S207

S204Y308

N293

N312

D113

Y316

TM5

TM3

TM6

TM5

TM3

TM7

TM4

TM6 TM7

a b


F290

F193

V117

W109

V114

I309

S203S207

S204

Y308N293

N312

D113

Y316

TM4

F290

F193

V117

W109

I309

OH

NO

HN

HO

S204 5.43

S207 5.46

Y308 7.35

S203 5.42

N293 6.55

V117 3.36

F290 6.52

A200 5.39

OH

F193 5.32

F289 6.51

S204 5.43

S207 5.46

Y308 7.35


Polar interactions

V114 3.33

T118 3.37


19

17

N293 6.55

V117 3.36 W286 6.48

F290 6.52

A200 5.39

Y199 5.38

F193 5.32 W109 3.28

F289 6.51

OH

BI-167107 Carazolol

OHN

I309 7.36

W109 3.28

O

c d

OH

O

H2N

O

H2N

OO

HO S203 5.42OH

O

NH2

OO

HO

D113 3.32

N312 7.39

Y316 7.43

D113 3.32

N312 7.39

Y316 7.43

H2 NH2



RESEARCH ARTICLE



Y3087.35

N2936.55

S2075.46

S2035.42

S2045.43

N3127.39

D1133.32

N3127.39

D1133.32

V1143.33

S2075.46

S2035.42

S2045.43

Y3087.35

N2936.55


Hydrophobic pocket in agonist ligand formed by Trp109, Phe193, Ile309

�45

RESULTS

�46

CONCLUSION

CONCLUSION

�47







425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE



(agonist)







425 450 475 500

0.4

0.6

0.8

1.0

425 450 475 500

0.4

0.6

0.8

1.0

Wavelength (nm)

Fl

uore

scen

ce in

tens

ity

(nor

mal

ized

to u

nlig

ande

d)

Gs + ISO

ISO

Gs

Unliganded

Nb80 + ISO

Nb80

Nb80 + ICI

ICI



–12–11–10 –9 –8 –7 –6 –5 –4

0

20

40

60

80

100

[3H

]-D

HA

bin

ding

–12–11–10 –9 –8 –7 –6 –5 –4–12–11–10–9 –8 –7 –6 –5 –4



+ Gs

0

20

40

60

80

100

0

20

40

60

80

100


b ca

d e f

Activation

TM6

TM5

TM3

TM5

TM3

TM6


Wavelength (nm)



RESEARCH ARTICLE



(inverse agonist)






a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH








a

d

b c

TM5

TM6

C terminus

N terminus

TM7

90º

e

TM3 (DRY)

TM5

TM6 TM7 (NPxxY)

TM1TM2

TM4

11.4 Å


D/E3.49

R3.50

Y7.53Y5.58

E6.30

Y3.51

β2AR–Cz Opsin

N terminus




ARTICLE RESEARCH



11.4Å

�48

REFERENCES

(1) Caffrey, M. and V. Cherezov, Crystallizing membrane proteins using lipidic mesophases. Nat. Protocols, 2009. 4(5): p. 706-731.

(2) EllisClare, The state of GPCR research in 2004. Nat Rev Drug Discov, 2004. 3(7): p. 577-626. (3) Hoelder, S., P.A. Clarke, and P. Workman, Discovery of small molecule cancer drugs: Successes, challenges and

opportunities. Molecular Oncology, 2012. 6(2): p. 155-176. (4) Rasmussen, S.G.F., et al., Crystal structure of the human β2 adrenergic G-protein-coupled receptor. Nature, 2007.

450(7168): p. 383-387. (5) Rosenbaum, D.M., et al., GPCR Engineering Yields High-Resolution Structural Insights into β2-Adrenergic Receptor

Function. Science, 2007. 318(5854): p. 1266-1273. (6) Rosenbaum, D.M., S.G.F. Rasmussen, and B.K. Kobilka, The structure and function of G-protein-coupled receptors.

Nature, 2009. 459(7245): p. 356-363. (7) Terstappen, G.C. and A. Reggiani, In silico research in drug discovery. Trends in Pharmacological Sciences, 2001.

22(1): p. 23-26. (8) Ulrik, G., Uncovering Molecular Mechanisms Involved in Activation of G Protein-Coupled Receptors. Endocrine

Reviews, 2000. 21(1): p. 90-113. (9) Venkatakrishnan, A.J., et al., Molecular signatures of G-protein-coupled receptors. Nature, 2013. 494(7436): p.

185-194. (10) Vincke, C., et al., General Strategy to Humanize a Camelid Single-domain Antibody and Identification of a Universal

Humanized Nanobody Scaffold. Journal of Biological Chemistry, 2009. 284(5): p. 3273-3284. (11) Yao, X.J., et al., The effect of ligand efficacy on the formation and stability of a GPCR-G protein complex.

Proceedings of the National Academy of Sciences, 2009. 106(23): p. 9501-9506. (12) Zhao, Q. and B.-l. Wu, Ice breaking in GPCR structural biology. Acta Pharmacol Sin, 2012. 33(3): p. 324-334.

Rasmussen, S. G. F., et al. (2011)

Presented by Bundit Boonyarit 5410210278

Dept.Chemistry, Fac.Science, Prince of Songkla University

tructure of a nanobody-stabilized active state of the β2 adrenoceptorS

April 23, 2015