Quaternary structure of a G-protein-coupled receptor heterotetramer in complex with Gi and Gs
- 3.2k Downloads
G-protein-coupled receptors (GPCRs), in the form of monomers or homodimers that bind heterotrimeric G proteins, are fundamental in the transfer of extracellular stimuli to intracellular signaling pathways. Different GPCRs may also interact to form heteromers that are novel signaling units. Despite the exponential growth in the number of solved GPCR crystal structures, the structural properties of heteromers remain unknown.
We used single-particle tracking experiments in cells expressing functional adenosine A1-A2A receptors fused to fluorescent proteins to show the loss of Brownian movement of the A1 receptor in the presence of the A2A receptor, and a preponderance of cell surface 2:2 receptor heteromers (dimer of dimers). Using computer modeling, aided by bioluminescence resonance energy transfer assays to monitor receptor homomerization and heteromerization and G-protein coupling, we predict the interacting interfaces and propose a quaternary structure of the GPCR tetramer in complex with two G proteins.
The combination of results points to a molecular architecture formed by a rhombus-shaped heterotetramer, which is bound to two different interacting heterotrimeric G proteins (Gi and Gs). These novel results constitute an important advance in understanding the molecular intricacies involved in GPCR function.
KeywordsGPCR Heterotetramer Heterotrimeric G protein Single-particle tracking BRET Molecular modeling
G-protein-coupled receptor (GPCR) oligomerization is heavily supported by recent biochemical and structural data [1, 2, 3, 4, 5, 6]. Optical-based techniques are instrumental in studying the dynamics and organization of receptor complexes in living cells . For instance, total internal reflection fluorescence microscopy shows that 30 % of muscarinic M1 receptors exist as dimers (with no evidence of higher oligomers) that undergo interconversion with monomers on a timescale of seconds . Similarly, the β1-adrenergic receptors (β1-AR) are expressed as a mixture of monomers and dimers whereas β2-adrenergic receptors (β2-AR) have a tendency to form dimers and higher-order oligomers . Moreover, the monomer-dimer equilibrium of the chemoattractant N-formyl peptide receptor at a physiological level of expression lies within a timescale of milliseconds . Together, these studies in heterologous systems show that a given GPCR is present in a dynamic equilibrium between monomers, dimers, and higher-order oligomers.
Studies in a broad spectrum of GPCRs [11, 12, 13, 14] show that these receptors may form heteromers. GPCR heteromers are defined as novel signaling units with functional properties different from homomers and they represent a completely new field of study . Innovative crystallographic techniques have permitted researchers to obtain crystal structures of GPCR families A, B, C, and F, bound to either agonists, antagonists, inverse agonists or allosteric modulators; in the form of monomers or homo-oligomers; and in complex with a G protein or with a ß-arrestin . However, crystal structures of GPCR heteromers have not yet been obtained. Here, we propose a quaternary structure of a heteromer, taking into account the molecular stoichiometry and the interacting G proteins. Adenosine A1-A2A receptor (A1R-A2AR) complexes constitute a paradigm in the GPCR heteromer field because A1R is coupled to Gi and A2AR to Gs; that is, they transduce opposite signals in cyclic adenosine monophosphate (cAMP)-dependent intracellular cascades. First described as a concentration-sensing device in striatal glutamatergic neurons , the A1R-A2AR heteromer is thought to function as a Gs/Gi-mediated switching mechanism by which low and high concentrations of adenosine inhibit and stimulate, respectively, glutamate release [17, 18]. The structural basis of this switch is key to understanding heteromer function and the biological advantage behind the GPCR heteromerization phenomenon. Here, we have devised the molecular architecture of the adenosine A1R-A2AR heteromer in complex with G proteins using a combination of microscope-based single-particle tracking, molecular modeling, and energy transfer assays in combination with molecular complementation. The results point to A1 and A2A receptors organizing into a rhombus-shaped heterotetramer that couples to Gi and Gs. The overall structure is very compact and provides interacting interfaces for GPCRs and for G proteins.
Results and discussion
Reciprocal restriction of adenosine receptor motion in the plasma membrane
Stoichiometry of A1 and A2A receptor heterocomplexes
The stoichiometry of the fluorescent receptors on the cell surface can be calculated from the brightness distribution of the individual particles  (see “Methods”). In cells expressing A1R-GFP, we found the majority of clusters to consist of either two (~47 %) or four (~34 %) receptors, and clusters with one or three receptors were scarce (~10 % and ~9 %, respectively) (Additional file 2: Figure S2A and black bars in Additional file 2: Figure S2C). In the case of A2AR-mCherry, the stoichiometry analysis showed that the clusters mostly expressed trimers (45 %), with dimers (29 %) and tetramers (12 %) the second and third most common populations (Additional file 2: Figure S2D and black bars in Additional file 2: Figure S2F). Remarkably, this stoichiometry for either A1 or A2A receptors was altered when the partner receptor was also expressed. In cells co-expressing A1R-GFP and A2AR-mCherry, the dimer population increased (57 % for A1R-GFP and 49 % for A2AR-mCherry, blue bars in Additional file 2: Figures S2C, F) and became the predominant species (Additional file 2: Figures S2B, C, E, F).
In order to focus the analysis on heteromer complexes, we identified clusters containing both receptors (individual yellow dots in Fig. 1g, displaying both GFP and mCherry fluorescence). In ~1000 analyzed co-localized clusters that consisted of a mixture of A1-GFP and A2A-Cherry (yellow dots in Fig. 1g), we found a similar high amount of dimers of A1R (75 %, left panel in Fig. 1h and green bar in Fig. 1i) and A2AR (74 %, right panel in Fig. 1h and red bar in Fig. 1i). Trimers and tetramers of A1R, and monomers and tetramers of A2AR, were in the minority or negligible (see Fig. 1h, i). In summary, given that the percentage of dimers of either A1R-GFP or A2AR-mCherry in the yellow dots (which show co-localization of the two receptors) was similar and high (~75 %), the heterotetramer containing two A1Rs and two A2ARs must have been the most predominant species. To our knowledge, this is the first stoichiometry data for a GPCR heteromer in living cells.
Arrangement of G proteins interacting with A1 and A2A receptors
Molecular model of Gi and Gs bound to the A1R-A2AR heterotetramer
We next evaluated, using computational tools, whether the proposed A1R-A2AR heterotetramer could couple to both Gi and Gs proteins. Clearly, the external protomers of the proposed A1R-A2AR heterotetramer can bind to Gi and Gs proteins (Fig. 5d). This model positions the α-subunits of Gi and Gs in close contact, facing the interior of the tetrameric complex, while the N-terminal α-helices of αi and αs point outside the complex. The N-terminal α-helices of the γ-subunits are in close proximity, facing the inside (Additional file 6: Figure S6), which explains the significant energy transfer observed between γ-Rluc and γ-YFP (Fig. 3, bar b). The model provides experimental insights into the structural arrangement of heteromers consisting of two GPCRs and coupled to two G proteins, the possibility of which has recently been discussed . We used MD simulations to study the stability of this complex. Additional file 7: Figure S7 shows root-mean-square deviations (rmsd) on protein α-carbons throughout the MD simulation, as well as key intermolecular distances among protomers and G proteins. Clearly, both the A1R protomer bound to Gi and the A1R protomer that does not interact with it maintained a close structural similarity (rmsd ≈ 0.3 nm) relative to the initial structures. Similar results were obtained for the A2AR protomers (bound and unbound to Gs) (Additional file 7: Figure S7A). The fact that rmsd values of the whole system, formed by the A1R-A2AR heterotetramer bound to Gi and Gs, are of the order of 0.6 nm indicates that the initial structural model is maintained during the MD simulation (Additional file 7: Figure S7A). As a consequence, selected intermolecular distances among protomers and G proteins remain constant during the MD simulation (Additional file 7: Figure S7B). A key aspect in the assembly of the heterotetramer is the TM interfaces for homodimerization (TM4/5) and heterodimerization (TM5/6). Additional file 8: Figure S8B shows rmsd values of the four-helix bundle forming the TM4/5 and TM5/6 interfaces, the initial and final snapshots of these bundles, and the evolution of the A1R-A2AR heterotetramer during the MD simulation. Clearly, the rather small structural variations of these four-helix bundles, also reflected by rmsd <0.3 nm, suggest a stable complex. Notably, the TM5/6 four-helix bundle seems more stable than the TM4/5 bundles, as shown by its lower rmsd value. Additional file 8: Figure S8B, C depicts contact maps of the TM4/5 and TM5/6 interfaces, as well as the evolution of the network of hydrophobic interactions within these interfaces during the MD simulation.
For more than a decade, experimental evidence has supported the occurrence of homo-oligomers and hetero-oligomers of GPCRs . However, our basic understanding of what makes heteromers different from homomers remains unknown. Our results, studying adenosine receptors as a model heteromer, point to three important new findings. First, the predominant stoichiometry in cells expressing A1R-A2AR heteromers is 2:2; that is, a dimer of dimers (tetramer). Second, two different heterotrimeric G proteins can couple to heteromers, the overall complex constituting a functional unit. Third, the molecular orientation within the heteromer complex affords various qualitatively different interfaces; the two more relevant are the inter-protomer heteromeric interface and the inter-G-protein interface. Presumably, the two interfaces provide the key characteristic of heteromers: the ability of one protomer/G-protein complex to influence the signaling of the other. Surely, allosteric effects occurring between heteroreceptors and between Gs and Gi proteins are due to conformational changes transmitted along the intimately interacting molecules in the complex. In our controlled cell transfection system, which expressed a low density of receptors, minor species formed by monomers and trimers were found in addition to a predominance of tetramers in the plasma membrane, strongly supporting the occurrence of an in vivo dynamic distribution of receptors.
Adenosine was, from an evolutionary point of view, one of the first extracellular regulators given that it is involved in energy and nucleic acid metabolisms. Adenosine A1 and A2A receptors are expressed in almost every mammalian organ and tissue. In the heart, where adenosine plays a key role in both inotropic and chronotropic regulation, A1R-mediated cardioprotection did not occur in A2AR knockout mice, suggesting an interaction between A1 and A2A receptors. In neurons, A1 and A2A receptors show co-localization, leading to inter-receptor interactions unveiled by pharmacological treatments. For instance, Okada et al.  showed that cAMP-dependent protein kinase A plays a role in the regulation of hippocampal serotonin release mediated by both A1 and A2A receptors. Similarly, the control of γ-amino butyric acid transport in astrocytes was attributed to the expression of A1R-A2AR heteromers and to a specific mechanism by which the heteromer signals via Gi or Gs depending on the concentration of adenosine . The structural basis of the differential signaling by the heteromer/G-protein macromolecular complex likely implies communication at the receptor-receptor level but also between Gs and Gi. Because the binding of two G proteins to a heterodimer is not feasible due to steric clashes , our finding that the A1R-A2AR heterotetramer may bind to both Gs and Gi provides a structural framework to interpret experimental data.
Total internal reflection microscopy and single-particle data analysis
Single-particle imaging and tracking were performed on a Nikon Total Internal Reflection Fluorescence (TIRF) system, as detailed in Additional file 11: Supplementary Methods. Typically, 500 readouts of a 512 × 512-pixel region, the full array of the CCD chip, were acquired. For single-particle data analysis, parameters were calculated by applying the equations described in Additional file 11: Supplementary Methods.
Cell culture and transient transfection
HEK-293T cells were grow at 37 °C in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Thermo Fischer Scientific, Madrid, Spain) supplemented with 2 mM L-glutamine, 100 U/ml penicillin/streptomycin, and 5 % (v/v) heat-inactivated fetal bovine serum (FBS) (all supplements were from Invitrogen, Paisley, UK). Cells were transiently transfected with cDNA corresponding to receptors, fusion proteins, A2AR mutants, or G-protein minigene vectors obtained as detailed in an expanded view by the polyethylenimine (PEI; SigmaAldrich, Cerdanyola del Vallès, Spain) method. Sample protein concentration was determined using a Bradford assay kit (Bio-Rad, Munich, Germany) using bovine serum albumin dilutions as standards. For single-particle imaging, cells were seeded into six-well plates containing glass coverslips (No. 1, round, 24 mm; Assistent, Sondheim, Germany) or into the Lab-Tek Chambered #1.0 Borosilicate Coverglass System (Nunc, Thermo Fisher Scientific, Schwerte, Germany). Cell transient transfections were performed with Lipofectamine™ 2000 (Invitrogen, Life Technologies, Darmstadt, Germany) or FuGENE 6 (Roche Applied Science, Indianapolis, IN, USA) and the application of 0.1–0.2 μg plasmid DNA per well. Before each experiment, cells were washed three times with 200 μL phenol red-free DMEM.
DNA sequences encoding amino acid residues 1–155 and 155–238 of YFP Venus protein, and amino acids residues 1–229 and 230–311 of RLuc8 protein were subcloned in the pcDNA3.1 vector to obtain the YFP Venus and RLuc8 hemi-truncated proteins. The human cDNAs for adenosine receptors, A2AR and A1R, cloned into pcDNA3.1, were amplified without their stop codons using sense and antisense primers harboring unique EcoRI and BamHI sites to clone receptors into the pcDNA3.1RLuc vector (pRLuc-N1; PerkinElmer, Wellesley, MA, USA), and EcoRI and KpnI to clone A2AR, A1R, or GHS1a into the pEYFP-N1 vector (enhanced yellow variant of GFP; Clontech, Heidelberg, Germany). Gαs cloned into the SFV1 vector, Gαi cloned into the pcDNA3.1 vector, or Gγ cloned into the pEYFP-C1 vector were amplified without their stop codons using sense and antisense primers harboring unique HindIII and BamHI sites to clone them into the pcDNA3.1-Rluc vector, or EcoRI and KpnI to clone Gαs into the pEYFP-N1 vector. The amplified fragments were subcloned to be in-frame with restriction sites of the pcDNA3.1RLuc or pEYFP-N1 vectors to give plasmids that expressed proteins fused to RLuc or YFP on the N-terminal end (Gαs-RLuc, Gαi-RLuc, Gγ-RLuc, Gαs-YFP, and Gγ-YFP) or the C-terminal end (A1R-RLuc, A2AR-RLuc, A1R-YFP, A2AR-YFP, and GHS1a-YFP). The human cDNAs for A1R or GHS1a were subcloned into pcDNA3.1-nRLuc8 or pcDNA3.1-nVenus to give plasmids that expressed A1R or GHS1a fused to either nRLuc8 or nYFP Venus on the C-terminal end of the receptor (A1R-nRLuc8 and A1R-nVenus or GHS1a-nRLuc8 and GHS1a-nVenus). The cDNAs for human A2A or GHS1a receptors were subcloned into pcDNA3.1-cRLuc8 or pcDNA3.1-cVenus to give plasmids that expressed receptors fused to either cRLuc8 or cYFP Venus on the C-terminal end of the receptor (A2AR-cRLuc8 and A2AR-cVenus or GHS1a-cRLuc8 and GHS1a-cVenus). Expression of constructs was tested by confocal microscopy and the receptor-fusion protein functionality by measuring ERK1/2 phosphorylation and cAMP production, as described previously [13, 14, 17, 29].
“Minigene” plasmid vectors are constructs designed to express relatively short polypeptide sequences following their transfection into mammalian cells. Here, we used minigene constructs encoding the carboxyl-terminal 11-amino acid residues from Gα subunits of Gi1/2 (Gi minigene) or Gs (Gs minigene) G proteins; the resulting peptides inhibit G-protein coupling to the receptor and consequently inhibit the receptor-mediated cellular responses as previously described . The cDNA encoding the last 11 amino acids of human Gα subunit corresponding to Gi1/2 (I K N N L K D C G L F) or Gs (Q R M H L R Q Y E L L), inserted in a pcDNA3.1 plasmid vector, were generously provided by Dr Heidi Hamm.
Energy transfer assays
For BRET and complementation BRET assays, HEK-293T cells were transiently cotransfected with a constant amount of cDNA encoding for proteins fused to RLuc, nRLuc8, or cRLuc8, and with increasing amounts of the cDNA corresponding to proteins fused to YFP, nYFP Venus, or cYFP Venus (see figure legends). To quantify protein-YFP expression or protein-reconstituted YFP Venus expression, cells (20 μg protein) were distributed in 96-well microplates (black plates with a transparent bottom) and fluorescence was read in a FLUOstar OPTIMA Fluorimeter (BMG Labtechnologies, Offenburg, Germany) equipped with a high-energy xenon flash lamp, using a 10 nm bandwidth excitation filter at 400 nm reading. Protein fluorescence expression was determined as the fluorescence of the sample minus the fluorescence of cells expressing the BRET donor alone. For BRET measurements, the equivalent of 20 μg of cell suspension were distributed in 96-well microplates (Corning 3600, white plates; Sigma) and 5 μM coelenterazine h (Molecular Probes, Eugene, OR, USA) was added. After 1 min for BRET or after 5 min for BRET with bimolecular fluorescence complementation, the readings were collected using a Mithras LB 940 that allows the integration of the signals detected in the short-wavelength filter at 485 nm (440–500 nm) and the long-wavelength filter at 530 nm (510–590 nm). To quantify protein-RLuc or protein-reconstituted RLuc8 expression, luminescence readings were also performed 10 min after adding 5 μM coelenterazine h. The net BRET was defined as [(long-wavelength emission)/(short-wavelength emission)] – Cf, where Cf corresponds to [(long-wavelength emission)/(short-wavelength emission)] for the donor construct expressed alone in the same experiment. BRET is expressed as miliBRET units (mBU; net BRET × 1000).
Computational model of the A1R-A2AR tetramer in complex with Gi and Gs
The crystal structure of inactive A2AR [PDB:4EIY]  was used for the construction of human A2AR [UniProt:P29274] and A1R [UniProt:P30542] homology models using Modeller 9.12 . These receptors share 51 % of sequence identity and 62 % of sequence similarity, excluding the C-terminal after helix 8. Intracellular loop 3 (ICL3) of A2AR (Lys209–Gly218) and A1R (Asn212–Ser219) were modeled using Modeller 9.12  using ICL3 of squid rhodopsin [PDB:2Z73] as a template. The C-terminus tails of A1R, containing 16 amino acids (Pro311–Asp326), and of A2AR, containing 102 amino acids (Gln311–Ser412), were modeled as suggested for the oxoeicosanoid receptor (OXER)  (see Additional file 9: Figure S9 for details). The N-terminus of A1R and A2AR were not included in the model. The “active” conformations of A1R bound to Gi and A2AR bound to Gs were modeled using the crystal structure of β2-AR in complex with Gs [PDB:3SN6] . The globular α-helical domain of the α-subunit was modeled in the “closed” conformation , using the crystal structure of [AlF4 −]-activated Gi [PDB:1AGR]. The location of YFP [PDB:2RH7] attached to the C-tail of A2AR was determined as suggested for the OXER  (see Additional file 9: Figure S9 for details). Rluc [PDB:2PSD] and YFP were fused to the to the N-terminus of the α-subunits and γ-subunits of Gi and Gs by a covalent bond. The structures of adenosine receptor oligomers were modeled via the TM4/5 interface for homodimerization, using the oligomeric structure of the β1-AR [PDB:4GPO] , or via the TM5/6 interface for heterodimerization, using the structure of the μ-OR [PDB:4DKL] . The Gi-bound A1R and Gs-bound A2AR protomers were rotated 10° to avoid the steric clash of the N-terminal helix of Gi and Gs with the C-terminal helix (Hx8) of Gs-unbound A2AR and Gi-unbound A1R, respectively. This computational model, without Rluc and YFP, was placed in a rectangular box containing a lipid bilayer (814 molecules of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine - POPC -) with explicit solvent (102,973 water molecules) and a 0.15 M concentration of Na+ and Cl− (1762 ions). This initial complex was energy-minimized and subsequently subjected to a 10 ns MD equilibration, with positional restraints on protein coordinates. These restraints were released and 500 ns of MD trajectory were produced at constant pressure and temperature (see Additional file 10: Movie M1). Computer simulations were performed with the GROMACS 4.6.3 simulation package , using the AMBER99SB force field as implemented in GROMACS and Berger parameters for POPC lipids. This procedure has been previously validated .
Availability of data and materials
The crystal structures 4EIY, 2Z73, 3SN6, 1AGR, 2RH7, 2PSD, 4GPO, and 4DKL are available from PDB (http://www.rcsb.org). All other relevant data are within the paper and its Additional files.
We acknowledge the technical help provided by Jasmina Jiménez (CIBERNED, University of Barcelona). This study was supported by grants from the Spanish Ministerio de Ciencia y Tecnología (SAF2009-07276, SAF2010-18472, SAF2011-23813, SAF2013-48271-C2-2-R; those grants may include FEDER funds), the Max Planck Society, the German Cancer Research Center, and the German Ministry for Education and Research (BMBF). PJM and LP participate in the European COST Action CM1207 (GLISTEN). Authors gratefully acknowledge the computer resources provided by the Barcelona Supercomputing Center - Centro Nacional de Supercomputación.
Assembly of adenosine A1 and A2A receptors in complex with two G proteins and MD simulation of the system. The assembly of Gs and Gi bound to the adenosine A1R-A2AR heterotetramer was subjected to 500 ns of MD simulation in a rectangular box containing the system, the lipid bilayer, explicit solvent, and ions. A1R protomers are in orange and red, A2AR protomers in light and dark green, Gα in white, Gβ in gray, and Gγ in purple. For easier visualization of protomer-protomer interfaces, TMs 4 and 5 are highlighted in blue and white, respectively. (MPEG 87870 kb)
- 32.Blattermann S, Peters L, Ottersbach PA, Bock A, Konya V, Weaver CD, et al. A biased ligand for OXE-R uncouples Gα and Gβγ signaling within a heterotrimer. Nat Chem Biol. 2012;8:631–8. doi: 10.1038/nchembio.962. Epub 2012 May 27.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.