FEBS Letters 587 (2013) 2656–2661

journal homepage: www.FEBSLetters .org

Conformational restriction of G-proteins Coupled Receptors (GPCRs)upon complexation to G-proteins: A putative activation mode of GPCRs?

0014-5793/$36.00 � 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.febslet.2013.06.052

Abbreviation: MRMS, Mass-weighted Root Mean Square⇑ Corresponding author. Fax: +33 0411759641.

E-mail address: nicolas.floquet@univ-montp1.fr (N. Floquet).

Maxime Louet a, Esra Karakas a, Alexandre Perret a, David Perahia b, Jean Martinez a, Nicolas Floquet a,⇑a Institut des Biomolécules Max Mousseron (IBMM, CNRS UMR5247), Faculté de Pharmacie, 15 avenue Charles Flahault, BP 14491, 34093 Montpellier Cedex 05, Franceb Laboratoire de Biologie et Pharmacologie Appliquée (LBPA), CNRS UMR8113, École Normale Supérieure de Cachan, 61 Avenue du Président Wilson, 94235 Cachan Cedex, France

a r t i c l e i n f o a b s t r a c t

Article history:Received 18 April 2013Revised 13 June 2013Accepted 29 June 2013Available online 10 July 2013

Edited by Christian Griesinger

Keywords:G-protein Coupled ReceptorG-protein heterotrimer complexNormal Mode AnalysisActivation mechanismUnique motion

GPCRs undergo large conformational changes during their activation. Starting from existing X-raystructures, we used Normal Modes Analyses to study the collective motions of the agonist-boundb2-adrenergic receptor both in its isolated ‘‘uncoupled’’ and G-protein ‘‘coupled’’ conformations.We interestingly observed that the receptor was able to adopt only one major motion in the pro-tein:protein complex. This motion corresponded to an anti-symmetric rotation of both its extra-and intra-cellular parts, with a key role of previously identified highly conserved proline residues.Because this motion was also retrieved when performing NMA on 7 other GPCRs which structureswere available, it is strongly suspected to possess a significant biological role, possibly being the‘‘activation mode’’ of a GPCR when coupled to G-proteins.� 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

G-protein Coupled Receptors (GPCRs) form a large family ofproteins constituted by seven hydrophobic, trans-membrane heli-cal segments (noted TM1 to TM7). Since 2007, many X-ray struc-tures have been solved describing GPCRs in complex with theirligands and/or protein partners. Together, these X-ray structuresconcluded to a common highly conserved fold and binding crevicefor these receptors, despite a low sequence conservation (�25%identity). More surprisingly, co-crystallized ligands of these recep-tors that include agonists, antagonists and inverse agonists all bindto the same orthosteric binding pocket, without any clear differ-ences in their binding modes and in the resulting conformationof the receptor [1]. This is in contradiction with biophysical datathat clearly indicate large conformational re-arrangements of thesereceptors that are directly dependent on the nature of the boundpartner(s) [2]. Recently solved structures of the b2-adrenergicreceptor have confirmed that conformational re-arrangementsoccur upon the complexation of this receptor to intra-cellularG-proteins [3]. The related conformational re-arrangement mostlyinclude a spreading of the trans-membraneous (TM) helices 5 and

6 that permits to the G-protein C-ter helix to penetrate inside thereceptor [4]. Because the putative mechanism of GPCRs activationhas for a long time been primarily associated to a G-proteinrecruitment, these re-arrangements have been logically associatedto the ‘‘inactive’’ and ‘‘active’’ states of GPCRs. However, theserather reflect a ‘‘coupling motion’’ of these receptors. In agreement,and because several recent studies have argued for a pre-couplingbetween GPCRs and G-proteins [2,5,6], one can ask what is the realactivation mechanism/motion of a GPCR after its complexation toG-proteins.

The recently solved structure of the b2-adrenergic receptorcomplexed to both an agonist molecule and to the Gs heterotri-meric G-protein [3] therefore appears as a good starting materialto decipher such an activation mechanism. Nevertheless, the acti-vation mechanism of these large complexes still requires to be elu-cidated at the molecular scale. Because the activation mechanismof GPCRs involves different key steps that are thought to occuron highly diverse time-scales from the nanosecond to the millisec-ond [7] it cannot, or hardly, be addressed by experiments. This acti-vation mechanism includes three main, consecutive steps: (i)ligand binding in the receptor that promotes stabilization of an ac-tive conformation [8], (ii) GDP:GTP exchange in the Ga subunit ofthe G-protein that is the rate limiting step [9]; and (iii) the subse-quent dissociation of the G-protein into two membrane-anchoredGa:GTP and Gbc subunits [10]. Both GDP release and dissociation

M. Louet et al. / FEBS Letters 587 (2013) 2656–2661 2657

of the G-protein are known to be triggered by the presence of thereceptor [11,12], even if these steps can occur in absence of any li-gand in case of constitutive activity.

Molecular modeling appears to be a method of choice to studyclass A GPCRs as reviewed in recent published papers [13,14].Moreover, because GPCR undergo large conformational changes,the Normal Modes Analysis (NMA), which is a good tool to predictcollective motions, has been used to study GPCRs [15] and G-pro-teins [16] in their isolated conformations. We have shown thatNMA was a powerful technique to study functional motions ofthese membrane proteins. Moreover, several studies proved thatlowest frequencies normal modes are often related to protein func-tions and permit to study motions occurring on large timescales, asit is the case for GPCRs [17,18].

In this study, we first validated our NMA protocol on the iso-lated B2AR receptor. Then, we employed the same NMA techniqueto predict motions that could exist in the GPCR:G-protein complex.

2. Materials and methods

2.1. Building of initial models

Several isolated b-2 adrenergic receptor (B2AR) (Protein DataBank (PDB) ids: 2RH1, [19] 3D4S, [20] 3NY8, [21] 3NY9, [21]3NYA, [21] 3P0G) [22] and the B2AR:Gs-protein complex (PDB id:3SN6) [3] solved by X-ray crystallography were subjected to Nor-mal Mode Analyses (NMA). Seven other receptors were also testedincluding the Beta-1 (2VT4), [23] CXCR4 (3ODU), [24] Dopamine(3PBL), [25] Histamine (3RZE), [26] Adenosine (3EML), [27] Sphin-gosine (3V2W) [28] and Muscarinic (3UON) [29] receptors. Thethird intracellular loop of all receptors was in each case completedby 6 alanine residues, whereas the lacking N-ter and C-ter regionsof receptors were not built to avoid unrealistic folding of loops. Co-crystallized ligands were included in calculations using theCHARMM General Forces Field (CGENFF) [30]. The crystallographicstructure of the GPCR:G-protein complex was modified accordingto other available G-proteins structures. These modifications in-cluded the repositioning of the helical domain of Ga at proximityof the ras-like domain as described in the PDB id 1GP2 X-ray struc-ture [31]. Indeed, this domain was rotated by about 180� in the ini-tial X-ray structure, probably resulting from crystallographicartifacts including antibody co-crystallization. This reconstruction,more compatible with recently published data, [32] allowed tocompare the Normal Modes (NMs) obtained for the complex tothose computed for the isolated G-protein published elsewhere[16].

2.2. Generation of low-energy conformations along the normal modesvectors

Normal Mode Analyses (NMA) were performed with theCHARMM software [33] and the CHARMM27 [34] forces field,excluding CMAP [35] parameters. The energy of each initial struc-ture was first minimized in vacuo by combining Steepest-Descentand Adopted Based Newton–Raphson algorithms to reach a lowenergy gradient of 10�5 kcal mol Å�2. NMA were computed withthe DIMB module as implemented in CHARMM [36]. The first 20lowest frequencies NMs of each initial structure were then usedas constraints to generate low-energy conformers along the normalmode directions. At this step, the CMAP corrections were turned onagain. This method is described in detail and was validated ondifferent biological systems elsewhere [16,37,38].

61 conformers were generated for each of the initial structureswith Mass Weighted Root Mean Square (MRMS) ranging from �3to +3 Å with a step of 0.1 Å. A negative value of MRMS applies to

displacements along the negative direction of the vector (see Ref.[37] for a graphical representation of the MRMS). During the min-imization protocol, the cut-off for non-bonded interactions was setto 10 Å, with a switching function applied between 8 and 10 Å.Minimization was performed in two successive stages. In a firststage a force constant of 1000 kcal mol�1 Å�2 was applied during2000 and 10000 steps of SD and Conjugate Gradient algorithms,respectively. In a second stage, the force constant was increasedto 20000 kcal mol�1 Å�2 for 5000 additional step of conjugate gra-dient to push the system exactly to the desired MRMS value alongthe vector. For both stages, the minimization process was stoppedwhen the energy gradient get lower than 10�2 kcal mol�1 Å�2.Translational and rotational force constants were set, respectivelyto 1000 and 10�5 kcal mol�1 Å�2.

3. Results and discussion

3.1. Validation of the Normal Modes Analysis protocol on the isolatedb-2 adrenergic receptor

Lowest frequencies Normal Modes (NMs) were computed forsix different X-ray structures describing the isolated B2AR in theProtein Data Bank (PDB ids: 2RH1, 3D4S, 3NYA, 3NY8, 3NY9,3P0G). This receptor can be considered as a perfect case study asit has been crystallized in both its uncoupled (2RH1, 3D4S, 3NYA,3NY8, 3NY9) or coupled (3P0G) states with different effectorsincluding agonist (3P0G), inverse agonists (2RH1, 3D4S, 3NY8and 3NY9) or antagonist (3NYA). The ‘‘coupling motion’’ describedby the transition between the 2RH1 and 3P0G structures was theunique significant collective motion that was observed amongthese structures. In previous published studies, this motion wasassociated to the activation of GPCRs. After computation of theNormal Modes of each of the upper mentioned ligand:receptorcomplexes, low-energy conformations were generated along eachof their twenty lowest frequencies NMs. The resulting motionsidentified along these modes were then quantitatively comparedto the ‘‘coupling motion’’ using the same method we described pre-viously [16]. Briefly, this method consists to compute correlationcoefficients between all pairs of generated motions, assuming thatvalues greater than 0.5 (mean + 2 std) depict highly related mo-tions in the Cartesian space. Interestingly, it was concluded thatat least one of the computed lowest frequencies NMs of eachB2AR structure was able to reproduce the 2RH1:3P0G conforma-tional transition. To convince, a Root Mean Square Deviationcomputation is shown in Fig. SI1 and showed that the transitionis effective. This analysis definitively validated our protocol andconfirmed the ability of our approach in identifying GPCR realisticmotions, including the ‘‘coupling motion’’ described by B2ARcrystallographic structures.

3.2. The isolated receptor adopts a large set of differentconformations. . .

A systematic pair by pair comparison of all the motionsdescribed by the twenty lowest frequencies NMs computed forthe isolated B2AR (PDB id: 3P0G) was performed. Interestingly,using a correlation coefficient threshold of 0.5, we observed thatthe isolated receptor can adopt a large set of different motions(see Fig. 1A). Indeed, only three groups of related modes wereformed by (i) the modes 11, 13, 14, 21, 23, 25; (ii) the modes 19,20, 24 and (iii) the modes 17 and 18. All the other modes describedunique motions with correlation coefficients <0.5 with any othermode. Highly similar results were obtained when performing thesame analysis on motions computed for others B2AR X-ray struc-tures, not only the 3P0G structure (data not shown).

Fig. 1. Analysis of the redundancy existing among the 20 lowest frequencies normal modes computed for: (A) The isolated B2AR; (B) The isolated G-protein; (C) Thecomplexed receptor; and (D) The complexed G-protein. Numbers in circles are NM numbers. Values reported on edges correspond to the correlation coefficient computedbetween the two connected modes; only modes with correlation coefficients greater than 0.5 were connected in these graphs.

2658 M. Louet et al. / FEBS Letters 587 (2013) 2656–2661

3.3. . . . As do the isolated G-protein

In a previous study [16], we described the NMs of the isolatedGi-protein heterotrimer. These NMs were analyzed with the samemotion analysis protocol that was applied to the isolated B2AR.Results of this analysis were reported in Fig. 1B. No redundancywas observed among the 20 lowest frequencies NMs.

3.4. Collective motions in the GPCR:G-protein complex: an activationmotion of GPCRs?

Using the same protocol as that described for the isolatedpartners, low energy conformations were generated along the 20lowest frequencies NMs computed for the B2AR:Gsabc complex.

Our objective was to compare motions predicted in the complexto those predicted in both isolated proteins. The conformationdetermined by Kobilka and co-workers [3] was used, except forthe Ga helical domain of the Gs protein which was put back toits position observed in the inactive structure of Gi (PDB id: 1GP2).

Surprisingly, as compared to the isolated receptor (Fig. 1A), thecomplexed receptor NMs were much more redundant, adoptingnearly one unique motion in the protein:protein complex(Fig. 1C). On the contrary, the complexed G-protein was still ableto explore a large set of intrinsic motions with the formation ofonly three clusters among its 20 lowest frequencies NMs (Fig. 1D).

More importantly, a detailed analysis showed that the receptoradopted a same motion (mean cross-correlation coefficient = 0.67)in 8 of the 10 lowest frequencies NMs of the complex. Because

Fig. 2. (A) Representation of the motion that was retrieved in all isolated receptors and in the G-protein:receptor complexed. This motion describes two inverse rotations ofthe extra- and intra-cellular parts of the GPCR. (B) A DynDom analysis performed on this motion localizes hinge regions at proximity of highly conserved proline residuesknown to be important for the receptor activation process [40]. (C) Variations of the Ca:Ca distances for all pairs of residues of the B2AR along either our predicted activationmotion (map on the right) or the X-ray motion (map on the left) confirming that these two motions are significantly different.

M. Louet et al. / FEBS Letters 587 (2013) 2656–2661 2659

lowest frequencies normal modes are usually related to proteinsfunctions, [17,18] this motion can be strongly suspected to be bio-logically relevant, probably playing a role in the activity of theGPCR:G-protein complex.

This putative ‘‘activation’’ motion of the complexed receptorwas reported in Fig. 2A. A movie showing this motion was alsosupplied as a Supplementary material SI2. This motion de-scribed an inverse rotation of both the extra- and intra-cellularparts of the receptor. An analysis performed with the Dyndomsoftware [39] confirmed the two extra- and intra-cellular partsof the 7 TMs as the two mobile domains (red+blue domains inFig. 2B). The related rotation axis, permitting such a motion,was nearly perpendicular to the putative longitudinal axis ofthe membrane (blue arrow in Fig. 2B). More interestingly, theresulting hinge regions/residues (in green in Fig. 2B) were local-ized at proximity of the highly conserved proline residuesbelonging to TM helices, suggesting a key role of these prolines.This observation was in good agreement with previously pub-lished site-directed mutagenesis experiments demonstratingthe importance of these residues in the activation mechanismof these receptors [40].

This putative activation motion of GPCRs was significantlydifferent from that previously deduced from available X-ray struc-tures and usually associated to the activation of GPCRs. Indeed, andas shown in Fig. 2C, the intrinsic motions of the receptor along our

‘‘activation’’ mode (map on the right ) involved most of the seventransmembrane segments whereas the ‘‘X-ray’’ motion describedonly a spreading of the TM5 and TM6 relatively to other TMs(map on the left).

3.5. The same motion was retrieved in other GPCRs

To confirm the putative important role of this motion in GPCRs,we performed NMA on seven other GPCRs for which an X-raystructure was available in the PDB: Beta-1 (2VT4), CXCR4(3ODU), Dopamine (3PBL), Histamine (3RZE), Adenosine (3EML),Sphingosine (3V2W), Muscarinic (3UON) receptors.

In each case, low energy conformations were once more gener-ated along each of the 20 lowest frequencies NMs. A pair by paircomparison of all obtained motions was subsequently performedto extract representative motions common to all these receptors.As an interesting result, the same anti-symmetric rotation motionas that reported in Fig. 2 was described by at least one of the lowestfrequencies NM computed for each ligand:receptor complex. Thiswas confirmed in the movies SI3, SI4 and SI5 showing that thesame motion was effectively retrieved in each studied receptor(color code: grey (3EML); red (2VT4); blue (3ODU); yellow(3PBL); orange (3RZE); green (3OUN); magenta (3V2W)). This mo-tion was also observed along the Normal Modes computed for theuncoupled B2AR (data not shown).

Fig. 3. Normal modes 8, 12, 13 and 15 of the receptor:G-protein complex. In all cases, the motion of the receptor was the same as that reported in Fig. 2 and corresponded tothe mode 8 computed for the isolated receptor. Coupled motions of the G-protein included modes 8, 9, 10 and 11 computed for the isolated G-protein those we previouslydiscussed a putative role in their activation [41].

2660 M. Louet et al. / FEBS Letters 587 (2013) 2656–2661

3.6. On the putative role of this motion in the activation of theprotein:protein complex

As shown in Fig. 3, the motion identified for the complexedB2AR was coupled to different motions of the G-protein. Amongthe ten lowest frequencies normal modes of the protein:proteincomplex, the modes 8, 12, 13 and 15 were of particular interest be-cause the anti-symetric rotation motion of the receptor was cou-pled to four of the five lowest frequencies modes 8, 9, 10 and 11previously identified for the isolated G-protein [16]. In this previ-ous study, we described how these G-protein NM could be linkedto its activation, especially to the GDP release and, in anotherstudy, the G-protein dissociation [41]. These observations clearlysuggest that the overall dynamical properties of the complex areprobably controlled by the heterotrimeric G-protein and not bythe receptor as it could be thought until now. This is in agreementwith recent biophysical data from our laboratory which suggestedthat complexation of the receptor to its favorite intra-cellular part-ners selected a particular conformation/motion [2]. Our data showhow this interaction between the receptor and the G-protein couldlead to the selection of one particular motion of the receptor. Inter-estingly, we observed that the motion characterized for the recep-tor was also retrieved in all other isolated GPCR whatever thebonded ligand, thus suggesting a common mechanism among allGPCRs. However, how ligands of diverse natures could thermody-namically affect this anti-symetric rotation motion still remainsto be elucidated by more quantitative (free energy) calculationsthat are actually under progress.

Analyses of the NMs computed for the complex showed amechanical relationship between the two receptor TM helices 5and 6 and the G-protein a subunit. Hinge regions of these motionsincluded the C-ter helix of Ga that penetrates into the receptor andthe switch II region located at the interface between the Ga and theGb subunits. These two regions have been shown to be directlyimplicated in coupling and activation of the GPCR:G-protein com-plexes [12,42]. It indicates how a mechanical information could betransmitted from the extra- to the intra-cellular part of these pro-tein:protein complexes, until reaching the inter-subunit interfaceof the G-protein and suggests a possible role of the receptor inG-protein dissociation. Most of these modes also showed an in-ter-domain dynamics of Ga, between the ras-like and the helicaldomains. Such a dynamic is known to be required for GDP release[43].

Obviously NMs calculations were performed in vacuo, the ob-tained intrinsic motions were interestingly compatible with thepresence of the membrane, as no predicted steric clash betweenthe protein and the putative position of the phospholipid bilayerwas observed.

Acknowledgments

We thank J. Marie, J.L. Banères, D. Gagne, C. M’Kadmi, J.C. Gal-leyrand, and J.A. Fehrentz for constructing remarks. We thank the‘‘Agence Nationale de la Recherche’’ for financial support (projectsGHSR-Sig and BIP:BIP). This work was also realized with the sup-port of HPC@LR, a Center of competence in High-Performance

M. Louet et al. / FEBS Letters 587 (2013) 2656–2661 2661

Computing from the Languedoc-Roussillon region, funded by theLanguedoc-Roussillon region, the Europe and the Université Mont-pellier 2 Sciences et Techniques. The HPC@LR Center is equippedwith an IBM hybrid Supercomputer.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.febslet.2013.06.052.

References

[1] Wang, C.I.A. and Lewis, R.J. (2013) Emerging opportunities for allostericmodulation of G-protein coupled receptors. Biochem. Pharmacol. 85, 153–162.

[2] Mary, S. et al. () Ligands and signaling proteins govern the conformationallandscape explored by a G protein-coupled receptor. PNAS, http://dx.doi.org/10.1073/pnas.1119881109.

[3] Rasmussen, S.G.F. et al. (2011) Crystal structure of the b2 adrenergic receptor-Gs protein complex. Nature 477, 549–555.

[4] Deupi, X. and Standfuss, J. (2011) Structural insights into agonist-inducedactivation of G-protein-coupled receptors. Curr. Opin. Struct. Biol. 21, 541–551.

[5] Drastichova, Z. and Novotny, J. (2012) Identification of a preassembled TRHreceptor-G(q/11) protein complex in HEK293 cells. Cell Struct. Funct. 37, 1–12.

[6] Galés, C. et al. (2006) Probing the activation-promoted structuralrearrangements in preassembled receptor–G protein complexes. Nat. Struct.Mol. Biol. 13, 778–786.

[7] Struts, A.V., Salgado, G.F.J., Martínez-Mayorga, K. and Brown, M.F. (2011)Retinal dynamics underlie its switch from inverse agonist to agonist duringrhodopsin activation. Nat. Struct. Mol. Biol. 18, 392–394.

[8] Rosenbaum, D.M., Rasmussen, S.G.F. and Kobilka, B.K. (2009) The structure andfunction of G-protein-coupled receptors. Nature 459, 356–363.

[9] Higashijima, T., Ferguson, K.M., Smigel, M.D. and Gilman, A.G. (1987) Theeffect of GTP and Mg2+ on the GTPase activity and the fluorescent properties ofGo. J. Biol. Chem. 262, 757–761.

[10] Tuteja, N. (2009) Signaling through G protein coupled receptors. Plant SignalBehav. 4, 942–947.

[11] Johnston, C.A. and Siderovski, D.P. (2007) Receptor-mediated activation ofheterotrimeric G-proteins: current structural insights. Mol. Pharmacol. 72,219–230.

[12] Hamm, H. E., Kaya, A. I., Gilbert, J. A., 3rd. and Preininger, A. M. (2013) In press,Linking receptor activation to changes in Sw I and II of Ga proteins. J. Struct.Biol.

[13] Shan, J., Khelashvili, G., Mondal, S., Mehler, E.L. and Weinstein, H. (2012)Ligand-dependent conformations and dynamics of the serotonin 5-HT(2A)receptor determine its activation and membrane-driven oligomerizationproperties. PLoS Comput. Biol. 8, e1002473.

[14] Taddese, B., Simpson, L.M., Wall, I.D., Blaney, F.E. and Reynolds, C.A. (2013)Modeling active GPCR conformations. Methods Enzymol. 522, 21–35.

[15] Floquet, N. et al. (2010) Activation of the ghrelin receptor is described by aprivileged collective motion: a model for constitutive and agonist-inducedactivation of a sub-class a g-protein coupled receptor (GPCR). J. Mol. Biol. 395,769–784.

[16] Louet, M., Perahia, D., Martinez, J. and Floquet, N. (2011) A concertedmechanism for opening the GDP binding pocket and release of thenucleotide in hetero-trimeric G-proteins. J. Mol. Biol. 411, 298–312.

[17] Bahar, I., Lezon, T.R., Bakan, A. and Shrivastava, I.H. (2010) Normal modeanalysis of biomolecular structures: functional mechanisms of membraneproteins. Chem. Rev. 110, 1463–1497.

[18] Bahar, I., Lezon, T.R., Yang, L.-W. and Eyal, E. (2010) Global dynamics of proteins:bridging between structure and function. Annu. Rev. Biophys. 39, 23–42.

[19] Cherezov, V. et al. (2007) High-resolution crystal structure of an engineeredhuman beta2-adrenergic G protein-coupled receptor. Science 318, 1258–1265.

[20] Hanson, M.A. et al. (2008) A specific cholesterol binding site is established bythe 2.8 Å structure of the human beta2-adrenergic receptor. Structure 16,897–905.

[21] Wacker, D. et al. (2010) Conserved binding mode of human beta2 adrenergicreceptor inverse agonists and antagonist revealed by X-ray crystallography. J.Am. Chem. Soc. 132, 11443–11445.

[22] Rasmussen, S.G.F. et al. (2011) Structure of a nanobody-stabilized active stateof the b(2) adrenoceptor. Nature 469, 175–180.

[23] Warne, T. et al. (2008) Structure of a b1-adrenergic G-protein-coupledreceptor. Nature 454, 486–491.

[24] Wu, B. et al. (2010) Structures of the CXCR4 chemokine GPCR with small-molecule and cyclic peptide antagonists. Science 330, 1066–1071.

[25] Chien, E.Y.T. et al. (2010) Structure of the human dopamine D3 receptor incomplex with a D2/D3 selective antagonist. Science 330, 1091–1095.

[26] Shimamura, T. et al. (2011) Structure of the human histamine H1 receptorcomplex with doxepin. Nature 475, 65–70.

[27] Jaakola, V.-P. et al. (2008) The 2.6 angstrom crystal structure of a human A2Aadenosine receptor bound to an antagonist. Science 322, 1211–1217.

[28] Hanson, M.A. et al. (2012) Crystal structure of a lipid G protein-coupledreceptor. Science 335, 851–855.

[29] Haga, K. et al. (2012) Structure of the human M2 muscarinic acetylcholinereceptor bound to an antagonist. Nature 482, 547–551.

[30] Vanommeslaeghe, K. et al. (2010) CHARMM general force field: a force field fordrug-like molecules compatible with the CHARMM all-atom additivebiological force fields. J. Comput. Chem. 31, 671–690.

[31] Wall, M.A. et al. (1995) The structure of the G protein heterotrimer Gi alpha 1beta 1 gamma 2. Cell 83, 1047–1058.

[32] Van Eps, N. et al. (2011) Interaction of a G protein with an activated receptoropens the interdomain interface in the alpha subunit. Proc. Natl. Acad. Sci. USA108, 9420–9424.

[33] Brooks, B.R. et al. (1983) CHARMM: a program for macromolecular energy,minimization, and dynamics calculations. J. Comput. Chem. 4, 187–217.

[34] MacKerell, A.D. et al. (2002) Encyclopedia of Computational Chemistry, JohnWiley & Sons, Ltd. vol. 1.

[35] Buck, M., Bouguet-Bonnet, S., Pastor, R.W. and MacKerell, A.D. (2006)Importance of the CMAP correction to the CHARMM22 protein force field:dynamics of hen lysozyme. Biophys. J. 90, L36–L38.

[36] Mouawad, L. and Perahia, D. (1993) Diagonalization in a mixed basis: amethod to compute low-frequency normal modes for large macromolecules.Biopolymers 33, 599–611.

[37] Floquet, N. et al. (2006) Normal mode analysis as a prerequisite for drugdesign: application to matrix metalloproteinases inhibitors. FEBS Lett. 580,5130–5136.

[38] Floquet, N. et al. (2009) Collective motions in glucosamine-6-phosphatesynthase: influence of ligand binding and role in ammonia channellingand opening of the fructose-6-phosphate binding site. J. Mol. Biol. 385,653–664.

[39] Hayward, S. and Berendsen, H.J.C. (1998) Systematic analysis of domainmotions in proteins from conformational change: new results on citratesynthase and T4 lysozyme. Proteins 30, 144–154.

[40] Van-Arnam, E.B., Lester, H.A. and Dougherty, D.A. (2011) Dissecting thefunctions of conserved prolines within transmembrane helices of the D2dopamine receptor. ACS Chem. Biol. 6, 1063–1068.

[41] Louet, M., Charlier, L., Martinez, J. and Floquet, N. (2012) Dissociation ofmembrane-anchored heterotrimeric G-protein induced by Ga subunit bindingto GTP. J. Chem. Inf. Model. 52, 3022–3027.

[42] Slessareva, J.E. et al. (2003) Closely related G-protein-coupled receptors usemultiple and distinct domains on G-protein a-subunits for selective coupling.J. Biol. Chem. 278, 50530–50536.

[43] Remmers, A.E., Engel, C., Liu, M. and Neubig, R.R. (1999) Interdomaininteractions regulate GDP release from heterotrimeric G proteins.Biochemistry 38, 13795–13800.