Advertisement

Dynamic Coupling and Allosteric Networks in the α Subunit of Heterotrimeric G Proteins*

  • Author Footnotes
    1 These authors contributed equally to this work.
    Xin-Qiu Yao
    Footnotes
    1 These authors contributed equally to this work.
    Affiliations
    From the Department of Computational Medicine and Bioinformatics, Ann Arbor, Michigan 48109
    Search for articles by this author
  • Author Footnotes
    1 These authors contributed equally to this work.
    Rabia U. Malik
    Footnotes
    1 These authors contributed equally to this work.
    Affiliations
    Cell and Developmental Biology, Ann Arbor, Michigan 48109

    Department of Genetics, Cell Biology, and Development, University of Minnesota, Minneapolis, Minnesota 55455
    Search for articles by this author
  • Nicholas W. Griggs
    Affiliations
    Pharmacology, University of Michigan, Ann Arbor, Michigan 48109,
    Search for articles by this author
  • Lars Skjærven
    Affiliations
    Department of Biomedicine, University of Bergen, 5020 Bergen, Norway
    Search for articles by this author
  • John R. Traynor
    Affiliations
    Pharmacology, University of Michigan, Ann Arbor, Michigan 48109,
    Search for articles by this author
  • Sivaraj Sivaramakrishnan
    Affiliations
    Department of Genetics, Cell Biology, and Development, University of Minnesota, Minneapolis, Minnesota 55455
    Search for articles by this author
  • Barry J. Grant
    Correspondence
    To whom correspondence may be addressed: 100 Washtenaw Ave., 2017 Palmer Commons Bldg., Ann Arbor, MI 48109-2218. Tel.: 734-647-3113;.
    Affiliations
    From the Department of Computational Medicine and Bioinformatics, Ann Arbor, Michigan 48109
    Search for articles by this author
  • Author Footnotes
    * This work was supported by University of Michigan, National Institutes of Health Grants R01-GM105646 (to S. S.) and R01-DA039997 (to J. R. T.), and American Heart Association Predoctoral Fellowship 14PRE18560010 (to R. U. M.). The authors declare that they have no conflicts of interest with the contents of this article. The content is solely the responsibility of theauthors and does not necessarily represent the official views of the National Institutes of Health.
    This article contains supplemental Table S1 and S2.
    1 These authors contributed equally to this work.
Open AccessPublished:December 24, 2015DOI:https://doi.org/10.1074/jbc.M115.702605
      G protein α subunits cycle between active and inactive conformations to regulate a multitude of intracellular signaling cascades. Important structural transitions occurring during this cycle have been characterized from extensive crystallographic studies. However, the link between observed conformations and the allosteric regulation of binding events at distal sites critical for signaling through G proteins remain unclear. Here we describe molecular dynamics simulations, bioinformatics analysis, and experimental mutagenesis that identifies residues involved in mediating the allosteric coupling of receptor, nucleotide, and helical domain interfaces of Gαi. Most notably, we predict and characterize novel allosteric decoupling mutants, which display enhanced helical domain opening, increased rates of nucleotide exchange, and constitutive activity in the absence of receptor activation. Collectively, our results provide a framework for explaining how binding events and mutations can alter internal dynamic couplings critical for G protein function.

      Introduction

      Heterotrimeric G proteins are key mediators of intracellular signaling pathways that control diverse cellular processes ranging from movement and division to differentiation and neuronal activity (
      • Neves S.R.
      • Ram P.T.
      • Iyengar R.
      G protein pathways.
      ). G proteins consist of three subunits: Gα, Gβ, and Gγ. Bound with GDP, Gα forms an inactive complex with its Gβγ subunit partners. Interaction with activated receptor (GPCR)
      The abbreviations used are: GPCR, G protein-coupled receptor; GDI, GDP dissociation inhibitor; RasD, Ras-like domain; HD, helical domain; PCA, principal component analysis; MD, molecular dynamics; A1R, adenosine 1 receptor; Fsk, forskolin; PTX, pertussis toxin; GTPγS, guanosine 5′-3-O-(thio)triphosphate; PL, P-loop.
      promotes the exchange of GDP for GTP on Gα and its separation from Gβγ. Both isolated Gα and Gβγ can then bind and activate or inhibit downstream effectors. GTP hydrolysis deactivates Gα, which subsequently reassociates with Gβγ completing the cycle. This cycle is further regulated by two classes of additional proteins called regulators of G protein signaling. These function as either GTPase-activating proteins (which promote GTP hydrolysis) or GDP dissociation inhibitors (GDIs, which hinder exchange of GDP for GTP) (
      • Sprang S.R.
      G protein mechanisms: insights from structural analysis.
      ). Important conformational transitions occurring at each stage of this regulated cycle have been characterized from extensive crystallographic studies. These include GDP, GTP analogue, Gβγ, GTPase-activating protein, GDI and most recently GPCR bound complex structures of Gα. However, the link between the observed conformations and the atomic level mechanisms involved in coupling receptor association, G protein activation, and effector interaction remain unclear.
      All Gα proteins consist of a catalytic GTP binding Ras-like domain (termed RasD) and a heterotrimeric G protein specific helical domain (HD). Recent principal component analysis (PCA) of 53 available Gα crystallographic structures identified three major conformationally distinct groups (Fig. 1 and Ref.
      • Yao X.-Q.
      • Grant B.J.
      Domain-opening and dynamic coupling in the α-subunit of heterotrimeric G proteins.
      ). These groups correspond to structures with bound GTP analogues, GDP, and GDI (red, green, and blue points in Fig. 1a, respectively). The major variation in the accumulated structures is the concerted displacements of three nucleotide-binding site loops termed the switch regions (SI, SII, and SIII), as well as a relatively small scale (<10°) rotation of the constituent HD and RasD regions. A much larger (127°) clam-shell like displacement of the HD with respect to RasD was reported recently in the crystallographic structure of Gαs (the α subunit of the stimulatory G protein for adenylyl cyclase) in complex with Gβγ and the β2 adrenergic receptor (
      • Rasmussen S.G.
      • DeVree B.T.
      • Zou Y.
      • Kruse A.C.
      • Chung K.Y.
      • Kobilka T.S.
      • Thian F.S.
      • Chae P.S.
      • Pardon E.
      • Calinski D.
      • Mathiesen J.M.
      • Shah S.T.
      • Lyons J.A.
      • Caffrey M.
      • Gellman S.H.
      • Steyaert J.
      • Skiniotis G.
      • Weis W.I.
      • Sunahara R.K.
      • Kobilka B.K.
      Crystal structure of the β2 adrenergic receptor-Gs protein complex.
      ). This conformational change, which effectively exposes the otherwise buried nucleotide binding site, has been linked to GPCR-mediated nucleotide exchange (
      • Rasmussen S.G.
      • DeVree B.T.
      • Zou Y.
      • Kruse A.C.
      • Chung K.Y.
      • Kobilka T.S.
      • Thian F.S.
      • Chae P.S.
      • Pardon E.
      • Calinski D.
      • Mathiesen J.M.
      • Shah S.T.
      • Lyons J.A.
      • Caffrey M.
      • Gellman S.H.
      • Steyaert J.
      • Skiniotis G.
      • Weis W.I.
      • Sunahara R.K.
      • Kobilka B.K.
      Crystal structure of the β2 adrenergic receptor-Gs protein complex.
      ). Evidence for domain opening has also been obtained from recent electron microscopy (
      • Westfield G.H.
      • Rasmussen S.G.
      • Su M.
      • Dutta S.
      • DeVree B.T.
      • Chung K.Y.
      • Calinski D.
      • Velez-Ruiz G.
      • Oleskie A.N.
      • Pardon E.
      • Chae P.S.
      • Liu T.
      • Li S.
      • Woods Jr., V.L.
      • Steyaert J.
      • Kobilka B.K.
      • Sunahara R.K.
      • Skiniotis G.
      Structural flexibility of the Gαs α-helical domain in the β2-adrenoceptor Gs complex.
      ), double electron-electron resonance analysis (
      • Van Eps N.
      • Preininger A.M.
      • Alexander N.
      • Kaya A.I.
      • Meier S.
      • Meiler J.
      • Hamm H.E.
      • Hubbell W.L.
      Interaction of a G protein with an activated receptor opens the interdomain interface in the α subunit.
      ), hydrogen-deuterium exchange mass spectrometry (
      • Chung K.Y.
      • Rasmussen S.G.
      • Liu T.
      • Li S.
      • DeVree B.T.
      • Chae P.S.
      • Calinski D.
      • Kobilka B.K.
      • Woods Jr., V.L.
      • Sunahara R.K.
      Conformational changes in the G protein Gs induced by the β2 adrenergic receptor.
      ), biochemical analysis (
      • Jones J.C.
      • Duffy J.W.
      • Machius M.
      • Temple B.R.
      • Dohlman H.G.
      • Jones A.M.
      The crystal structure of a self-activating G protein α subunit reveals its distinct mechanism of signal initiation.
      ), and molecular dynamics (MD) simulations (
      • Yao X.-Q.
      • Grant B.J.
      Domain-opening and dynamic coupling in the α-subunit of heterotrimeric G proteins.
      ,
      • Jones J.C.
      • Jones A.M.
      • Temple B.R.
      • Dohlman H.G.
      Differences in intradomain and interdomain motion confer distinct activation properties to structurally similar Gα proteins.
      ,
      • Dror R.O.
      • Mildorf T.J.
      • Hilger D.
      • Manglik A.
      • Borhani D.W.
      • Arlow D.H.
      • Philippsen A.
      • Villanueva N.
      • Yang Z.
      • Lerch M.T.
      • Hubbell W.L.
      • Kobilka B.K.
      • Sunahara R.K.
      • Shaw D.E.
      Structural basis for nucleotide exchange in heterotrimeric G proteins.
      ). In addition, the structure of Rasmussen and co-workers (
      • Chung K.Y.
      • Rasmussen S.G.
      • Liu T.
      • Li S.
      • DeVree B.T.
      • Chae P.S.
      • Calinski D.
      • Kobilka B.K.
      • Woods Jr., V.L.
      • Sunahara R.K.
      Conformational changes in the G protein Gs induced by the β2 adrenergic receptor.
      ) together with mass spectrometry results also confirm that both N-terminal β1 strand and C-terminal α5 helix are major interaction sites for receptors. This supports the previously suggested role of these regions in coupling receptor binding and nucleotide dissociation activities (
      • Oldham W.M.
      • Hamm H.E.
      Heterotrimeric G protein activation by G-protein-coupled receptors.
      • Iiri T.
      • Herzmark P.
      • Nakamoto J.M.
      • van Dop C.
      • Bourne H.R.
      Rapid GDP release from Gsα in patients with gain and loss of endocrine function.
      ,
      • Marin E.P.
      • Krishna A.G.
      • Sakmar T.P.
      Rapid activation of transducin by mutations distant from the nucleotide-binding site: evidence for a mechanistic model of receptor-catalyzed nucleotide exchange by G proteins.
      ,
      • Kapoor N.
      • Menon S.T.
      • Chauhan R.
      • Sachdev P.
      • Sakmar T.P.
      Structural evidence for a sequential release mechanism for activation of heterotrimeric G proteins.
      ,
      • Preininger A.M.
      • Funk M.A.
      • Oldham W.M.
      • Meier S.M.
      • Johnston C.A.
      • Adhikary S.
      • Kimple A.J.
      • Siderovski D.P.
      • Hamm H.E.
      • Iverson T.M.
      Helix dipole movement and conformational variability contribute to allosteric GDP release in Gαi subunits.
      ,
      • Thaker T.M.
      • Sarwar M.
      • Preininger A.M.
      • Hamm H.E.
      • Iverson T.M.
      A transient interaction between the phosphate binding loop and switch I contributes to the allosteric network between receptor and nucleotide in Gαi1.
      • Kaya A.I.
      • Lokits A.D.
      • Gilbert J.A.
      • Iverson T.M.
      • Meiler J.
      • Hamm H.E.
      A conserved phenylalanine as relay between the α5 helix and the GDP binding region of heterotrimeric Gi protein α subunit.
      ). Despite these advances, critical questions remain unanswered: How do the distinct conformations evident in the accumulated structures interconvert? And critically, how do distal functional sites responsible for GPCR, nucleotide, and partner protein binding allosterically coordinate their activities? Addressing these questions requires information on protein dynamics, which is not directly available from the accumulated static experimental structures.
      Figure thumbnail gr1
      FIGURE 1.Principal component analysis of crystallographic Gα structures reveals three major conformations. a, projection of 53 available Protein Data Bank crystal structures of Gα (represented by round points; see also ) onto the first two principal components (PC) that together account for 65.4% of the total structural variance. Three conformational clusters correspond to GTP-bound (red), GDP-bound (green), and GDI-bound (blue) structures. Projection of conformations sampled from MD simulations under GTP- and GDI-bound conditions are also shown as transparent red- and blue-shaded areas, respectively. b, superimposition of structures with high variance regions (SI, SII, SIII, and LC) colored by conformational cluster.
      In this study, we aimed to dissect detailed mechanisms of allostery in Gα by employing a combined computational and experimental approach. This entailed long time MD simulations, ensemble-based correlation network analysis of dynamic couplings related to allostery, and experimental mutagenesis and functional assays. This approach identified fluctuations and dynamic residue couplings in functional regions that distinguished GTP, GDP, and GDI states. Network analysis revealed a consistent bilobal dynamic partitioning of the RasD region. Partitioning of residues into these correlated segments was similar in different states but displayed distinct nucleotide dependent coupling strengths between segments. The active GTP state was shown to have the strongest overall couplings, exhibiting “dynamical tightening” with respect to GDP and GDI states. Furthermore, network path analysis delineated the detailed mechanism of dynamic coupling and revealed residues predicted to be involved in mediating the distal (>30 Å) allosteric coupling of receptor, nucleotide, and HD interfaces. Results from mutational simulations further supported the functional relevance of the identified allosteric paths with selected path mutations displaying a dynamic decoupling of distal sites along with an enhanced rate of spontaneous RasD-HD domain opening. Experimental mutagenesis of a number of these sites together with in vitro cAMP and [35S]GTPγS assays indicated that the signaling properties of Gαi can indeed be modulated by these single point mutations that act allosterically. In particular, the novel L32A mutation (numbering based on the α subunit of bovine transducin) was predicted to enhance domain opening and was found to increase nucleotide exchange rates, increase G protein activation, and decouple G protein from receptor activation leading to constitutive activity.

      Experimental Procedures

      All structure and trajectory analysis was performed with Bio3D version 2.0 (
      • Grant B.J.
      • Rodrigues A.P.
      • ElSawy K.M.
      • McCammon J.A.
      • Caves L.S.
      Bio3d: an R package for the comparative analysis of protein structures.
      ,
      • Skjaerven L.
      • Yao X.Q.
      • Scarabelli G.
      • Grant B.J.
      Integrating protein structural dynamics and evolutionary analysis with Bio3D.
      ). Molecular graphics were generated with VMD version 1.9 (
      • Humphrey W.
      • Dalke A.
      • Schulten K.
      VMD: visual molecular dynamics.
      ).

      Crystallographic Structures Preparation

      Atomic coordinates for all crystallographic structures of the Gi protein family were obtained from the RCSB Protein Data Bank (
      • Berman H.M.
      • Westbrook J.
      • Feng Z.
      • Gilliland G.
      • Bhat T.N.
      • Weissig H.
      • Shindyalov I.N.
      • Bourne P.E.
      The Protein Data Bank.
      ) via sequence search utilities in the Bio3D package. Structures with unresolved residues in the switch regions were excluded from analysis leading to a data set containing 53 structural species (see supplemental Table S1 for full listing). Prior to assessing the variability of the structures, iterated rounds of structural superposition were performed to identify the most structurally invariant region. During this procedure, residues with the largest positional differences (measured as the volume of an ellipsoid determined from the Cartesian coordinates of the Cα atoms) were removed, before each round of superposition, until only invariant core residues remained (
      • Gerstein M.
      • Altman R.B.
      Average core structures and variability measures for protein families: application to the immunoglobulins.
      ). The identified “core” structure was used as the reference frame for the superposition of crystal structures and subsequent MD trajectories prior to further analysis.

      Principal Component Analysis

      PCA was performed for the 53 crystallographic structures of Gα to characterize interconformer relationship. The application of PCA to both distributions of experimental structures and MD trajectories, along with its ability to provide considerable insight into the nature of conformational differences in a range of protein families has been previously discussed (
      • Caves L.S.
      • Evanseck J.D.
      • Karplus M.
      Locally accessible conformations of proteins: multiple molecular dynamics simulations of crambin.
      • Gorfe A.A.
      • Grant B.J.
      • McCammon J.A.
      Mapping the nucleotide and isoform-dependent structural and dynamical features of Ras proteins.
      ,
      • Grant B.J.
      • McCammon J.A.
      • Caves L.S.
      • Cross R.A.
      Multivariate analysis of conserved sequence-structure relationships in kinesins: coupling of the active site and a tubulin-binding sub-domain.
      • van Aalten D.M.
      • Conn D.A.
      • de Groot B.L.
      • Berendsen H.J.
      • Findlay J.B.
      • Amadei A.
      Protein dynamics derived from clusters of crystal structures.
      ). Briefly, PCA is based on the diagonalization of the variance-covariance matrix, Σ, with elements Σij calculated from the Cartesian coordinates of Cα atoms, r, after the superposition of all structures under analysis,
      ij=(riri)(rjrj)
      (Eq.1)


      where i and j enumerate all 3N Cartesian coordinates. The eigenvectors, or principal components, of Σ form a linear basis set matching the distribution of structures. The variance of the distribution along each principal component is given by the corresponding eigenvalue. Projection of the distribution onto the subspace defined by principal components with the largest eigenvalues provides a low dimensional representation of structures facilitating analysis of interconformer relationships (Fig. 1a).

      Molecular Dynamics Simulations

      MD simulations were performed with AMBER12 (
      • Case D.A.
      • Darden T.A.
      • Cheatham T.E.I.I.I.
      • Simmerling C.L.
      • Wang J.
      • Duke R.E.
      • Luo R.
      • Walker R.C.
      • Zhang W.
      • Merz K.M.
      • Roberts B.
      • Hayik S.
      • Roitberg A.
      • Seabra G.
      • Wails J.
      • Goetz A.W.
      • Kolossvary I.
      • Wong K.F.
      • Paesani F.
      • Vanicek J.
      • Wolf R.M.
      • Liu J.
      • Wu X.
      • Brozell S.R.
      • Steinbrecher T.
      • Gohlke H.
      • Cai Q.
      • Ye X.
      • Hsieh M.J.
      • Cui G.
      • Roe D.R.
      • Mathews D.H.
      • Seetin M.G.
      • Salomon-Ferrer R.
      • Sagui C.
      • Babin V.
      • Luchko T.
      • Gusarov S.
      • Kovalenko A.
      • Kollman P.A.
      ) and corresponding force field ff99SB (
      • Hornak V.
      • Abel R.
      • Okur A.
      • Strockbine B.
      • Roitberg A.
      • Simmerling C.
      Comparison of multiple Amber force fields and development of improved protein backbone parameters.
      ). Additional parameters for guanine nucleotides were taken from Meagher et al. (
      • Meagher K.L.
      • Redman L.T.
      • Carlson H.A.
      Development of polyphosphate parameters for use with the AMBER force field.
      ). The Mg2+·GDP-bound transducin crystal structure (Protein Data Bank code 1TAG) was employed as the starting model for GDP-bound simulations. The Mg2+·GTPγS structure (Protein Data Bank code 1TND) was used as the starting model for GTP-bound simulations. The sulfur atom (S1γ) in the GTPγS was replaced with the corresponding oxygen (O1γ) of GTP. In addition, GDP-bound Gαi/GoLoco motif complex structure (Protein Data Bank code 1KJY) was employed as the starting model for GDI-bound simulations. These structures were identified as cluster representatives from PCA. In all systems, Arg and Lys were protonated, whereas Asp and Glu were deprotonated. The protonation states for His residues were determined based on an inspection of the residues local environment and their pKa values as calculated by PROPKA (
      • Li H.
      • Robertson A.D.
      • Jensen J.H.
      Very fast empirical prediction and rationalization of protein pKa values.
      ). Simulation structures were solvated in a truncated cubic box of pre-equilibrated TIP3P water molecules, which extended 12 Å in each dimension from the surface of the solute. Sodium counter ions (Na+) were added to neutralize the systems. Additional ions were not added to mimic physiological ionic strength. This may have the effect of accentuating electrostatic interactions. Energy minimization was performed in four stages, with each stage employing 500 steps of steepest decent followed by 1500 steps of conjugate gradient. First, minimization for solvent only was performed with fixed positions of protein and ligand atoms. Second, side chain and ligand were relaxed with backbone still fixed. Third, all protein and ligand atoms were relaxed with fixed solvent. Fourth, all atoms were free to move without any restraint. Following minimization, 10 ps of MD simulation was performed to heat the system from 0 to 300 K under constant volume periodic boundary conditions. A further 1 ns of equilibration simulation was performed at constant temperature (T = 300 K) and constant pressure (p = 1 bar). Subsequent 80-ns production phase MD was then performed under the same conditions as equilibration. For both energy minimization and MD simulations, the particle mesh Ewald summation method was adopted to treat long range electrostatic interactions. In addition, an 8 Å cutoff was used to truncate the short range nonbonded van der Waals' interactions. Additional operational parameters for MD included a 2-fs time step, removal of the center of mass motion every 1000 steps and update of the nonbonded neighbor list every 25 steps. All hydrogen atoms were constrained using the SHAKE algorithm.

      Correlation Network Construction

      Network analysis of correlated motions was employed to identify protein segments with coupled dynamics. A weighted network graph was constructed where each node represents an individual residue and the weight of the connection between nodes, i and j, represents their respective cross-correlation value, cij (
      • Ichiye T.
      • Karplus M.
      Collective motions in proteins: a covariance analysis of atomic fluctuations in molecular dynamics and normal mode simulations.
      ). This well established cross-correlation approach is based on linear atomic displacements during the course of simulations. Our evaluation of more recently developed nonlinear mutual information of dihedral angle changes (
      • Dubay K.H.
      • Bothma J.P.
      • Geissler P.L.
      Long-range intra-protein communication can be transmitted by correlated side-chain fluctuations alone.
      ,
      • Wu S.
      • Jun Lee C.
      • Pedersen L.G.
      Analysis on long-range residue-residue communication using molecular dynamics.
      ) indicated that prohibitively longer simulation and analysis time were likely required for the generation of robust networks (data not shown). We used a network construction method similar to that introduced by Luthey-Schulten and co-workers (
      • Sethi A.
      • Eargle J.
      • Black A.A.
      • Luthey-Schulten Z.
      Dynamical networks in tRNA:protein complexes.
      ). However, instead of employing a [4.5 Å] contact map of non-neighboring residues to define network edges (that are then weighted by a single correlation matrix), our network edges were constructed based on the minimum Cα-Cα cross-correlation value between all residues across five replicate simulations. Specifically, cross-correlations were calculated for each trajectory after mass-weighted superposition. Network edges were added for (i) residue pairs with |cij| ≥ 0.6 in all simulations and (ii) residues satisfying |cij| ≥ 0.6 in at least one simulation and with a Cα-Cα distance dij ≤ 10 Å for at least 75% of total simulation frames. Edge weights were calculated as −log(|<cij>|), where <·> denotes the average across simulations. Networks constructed with a cij cutoff between 0.5 and 0.7 yielded equivalent networks with similar community structure (data not shown). This procedure was found to reduce potentially false positive couplings that exist when using only a single trajectory, as well as minimize the arbitrary exclusion of consistent strong couplings that are just beyond a given distance cutoff.

      Network Community and Centrality Analysis

      For each correlation network, hierarchical clustering was performed to generate aggregate nodal clusters, or communities, that are highly intraconnected but loosely interconnected, using a betweenness clustering algorithm similar to that introduced by Girvan and Newman (
      • Girvan M.
      • Newman M.E.
      Community structure in social and biological networks.
      ). However, instead of using the partition with the maximum modularity score, as is common with unweighted networks, we took the partition closest to the maximal modularity value that resulted in the smallest number of overall communities (i.e. the earliest high scoring partition). This avoided the common situation where many small communities with equally high scoring modularity values were generated. Using this approach networks under different states showed a largely consistent community partition, with differences localized to the nucleotide binding P-loop (PL), SI, SII, and α1 regions that were observed to repartition between major communities in the different states (data not shown). Consensus communities that abstracted these regions to new separate communities to facilitate further comparisons were derived from partitioning these regions at the boundaries of their known conserved sequence motifs. Intercommunity correlations were then calculated as the sum of the mean correlation values across simulation replicates associated with all the intercommunity edges. A standard t test was also performed to measure the significance of the mean difference between intercommunity correlations of distinct states.
      Node centralities that assess the density of connections per node were calculated as follows,
      xi=1λjGAijxj
      (Eq.2)


      where xi is the centrality of node i, Aij is the ijth entry of the adjusted adjacent matrix A, λ is a constant to be determined, and G indicates all nodes. Aij is not 0 if node i and j are linked, and it is equal to e−wij, where wij is the edge weight. Solving Equation 2 for every i (iG) is equivalent to finding the eigenvalues and eigenvectors of matrix A. Node centralities can then be obtained from the eigenvector with the largest eigenvalue, after scaling all entries with the largest entry set to be 1.

      Network Path Analysis

      Given a pair of nodes treated as “source” and “sink,” respectively, optimal (shortest) and suboptimal (close to but longer than optimal) connecting network paths were identified using the algorithm in Ref.
      • Yen J.Y.
      Finding K shortest loopless paths in a network.
      . Five hundred paths were collected for each source/sink pair in each network. Comparative path length distributions indicating the strength of correlated motions under distinct conditions were then calculated. In addition, normalized node degeneracy, i.e. the fraction of the number of paths going through each node, was calculated. Residues with high node degeneracy (≥0.1 or 50 paths) in any network were specified as “on-path” residues and were subjected to further analysis including, for select cases, mutagenesis simulations and experimental characterization.

      Molecular Cloning

      cDNA for human adenosine A1 receptor (A1R) and Gαi2 isoform 1 were acquired from DNASU Plasmid Repository and Open Biosystems, respectively. N-terminal Flag-tagged Gαi2 or A1R-Gαi2 fusions and N-terminal mCerulean-tagged Gαi2 were cloned into pBiex1 and pCDNA5/FRT plasmids, respectively. A1R and Gαi2 were fused together using the previously described SPASM technique (
      • Malik R.U.
      • Ritt M.
      • DeVree B.T.
      • Neubig R.R.
      • Sunahara R.K.
      • Sivaramakrishnan S.
      Detection of G protein-selective G protein-coupled receptor (GPCR) conformations in live cells.
      ). Briefly, A1R-Gαi2 fusions were cloned from N to C terminus as follows: A1R, mCitrine, ER/K linker, mCerulean, and Gαi2. Repeating (Gly-Ser-Gly)4 sequences were inserted in between domains to ensure rotational freedom. Gαi2 and Gαt (Bos taurus) residues were aligned to identify the conserved Leu32, Phe195, and Asp333 residues in Gαi2. L36A, F200L, or D338A mutations in Gαi2 were induced via PCR using oligonucleotide-directed mutagenesis (QuikChange site-directed mutagenesis kit; Stratagene). All constructs were confirmed via sequencing.

      Mammalian Cell Culture

      HEK293T-Flp-in (Invitrogen) cells were cultured in DMEM supplemented with 10% FBS (v/v), 4.5 g/liter d-glucose, 1% GlutaMAX, 20 mm HEPES, pH 7.5, at 37 °C in a humidified atmosphere at 5% CO2. The cells were plated at ∼30% confluence into 6-well tissue cultured treated dishes. 16–20 h later, the cells were transfected with indicated construct using XtremeGene HP DNA transfection reagent. Where indicated, 24 h post-transfection, cells were incubated with 100 ng/ml pertussis toxin (PTX) for 20–24 h. Experiments were conducted when fusions or mCerulean-Gαi2 constructs expressed predominately at the plasma membrane with minimal internal localization as evaluated at 20× and 40× magnification on a Nikon tissue culture microscope with fluorescence detection. Experiments were performed at equivalent fusion expression at a cell density of 2 × 106 cells/ml. Fusion expression was quantified by measuring mCitrine and mCerulean fluorescence by exciting cells at 490 and 430 nm, respectively. Excitation and emission bandpass were correspondingly set to 8 and 4 nm. For mCerulean tagged Gαi2 experiments, the cells were harvested at similar wild type and mutant Gαi2 expression levels. It should be noted that wild type, F195L, and D333A expressed twice as much as L32A mutant as indicated by mCerulean fluorescence counts. Fusion integrity was evaluated by measuring mCitrine to mCerulean emission ratio. This ratio was held between 1.7 and 2.1 because mCitrine is twice as bright as mCerulean. All fluorescence measurements were conducted using FlouroMax-4 fluorometer (Horiba Scientific) in an optical glass cuvette.

      Quantification of cAMP Levels

      Protocol was conducted as previously described using the cAMP Glo luminescence based assay (Promega) (
      • Gabriel D.
      • Vernier M.
      • Pfeifer M.J.
      • Dasen B.
      • Tenaillon L.
      • Bouhelal R.
      High throughput screening technologies for direct cyclic AMP measurement.
      ). Briefly, 28–30 h post-transfection, the cells were spun down (300 g, 3 min), resuspended at 1 × 106 cells/ml in PBS solution supplemented with 0.02% glucose and 800 μm ascorbic acid and aliquoted into 96-well round bottom opaque microplates. The cells were treated with 0.25 mm 3-isobutyl-1-methylxanthine, 1 μm forskolin, or 10 μm forskolin in the presence or absence of A1R agonist (12.5 nm N6-cyclopentyladenosine) for 5 min at 37 °C. After incubation with indicated compounds, the cells were lysed, and protocol was followed according to the manufacturer's recommendations. Luminescence was recorded using a microplate luminometer (SpectraMax M5; Molecular Devices). cAMP levels (relative luminescence units) were calculated by subtracting the untransfected untreated background from the indicated conditions. Each experiment had four repeats per condition and was repeated at least three times (n > 12).

      Insect Cell Culture and Protein Purification

      Sf9 cells were cultured and maintained in suspension with shaking at 28 °C in Sf900-II medium (Life Technologies) containing 1% Antibiotic-Antimycotic (Life Technologies). Constructs were transiently transfected into Sf9 cells using Escort IV transfection reagent (Sigma-Aldrich) in antibiotic-free medium. The cells were lysed 3 days post-transfection in HEPES lysis buffer (20 mm HEPES, pH 7.5, 0.5% Igepal, 4 mm MgCl2, 200 mm NaCl, 7% sucrose, 5 mm DTT, 50 μg/ml PMSF, 5 μg/ml aprotinin, 5 μg/ml leupeptin). Lysates were clarified by ultracentrifugation (175,000 × g, 4 °C, 45 min) and bound to anti-FLAG M2 affinity resin (Sigma-Aldrich). Resin was washed with HEPES wash buffer (20 mm HEPES, pH 7.5, 150 mm KCl, 5 mm MgCl2, 5 mm DTT, 50 μg/ml PMSF, 5 μg/ml aprotinin, 5 μg/ml leupeptin) and eluted using FLAG peptide (Sigma-Aldrich). Protein concentration and integrity were assessed by SDS-PAGE and Coomassie staining in comparison to BSA standards.

      Nucleotide Exchange and [35S]GTPγS Incorporation

      Binding of [35S]GTPγS to purified Gαi protein was based on the method outlined in Ref.
      • Slepak V.Z.
      • Wilkie T.M.
      • Simon M.I.
      Mutational analysis of G protein α subunit Goα expressed in Escherichia coli.
      . Briefly, the assay mixture contained 0.5 μg of purified Gαi protein in TED buffer (20 mm Tris-HCl, pH 7.4, 1 mm DTT, 1 mm EDTA, 0.1 mm MgCl2) with 2 μm GTPγS and 0.1 nm [35S]GTPγS, unless otherwise stated, in a total volume of 200 μl. Experiments were performed at room temperature (25 °C), unless stated otherwise, and 50-μl samples were withdrawn and diluted 1:4 in ice-cold TED buffer to stop the reaction at various time points. Aliquots were vacuum-filtered through GF/C filters, and the amount of bound radioactivity was quantified by scintillation counting. Data were analyzed using GraphPad Prism as one-phase association curves.

      Discussion

      Our extensive MD simulations predicted nucleotide-dependent modes of motion and internal dynamic coupling of functional regions including SI, SII, SIII, PL, and HD. Correlation network analysis characterized the conserved bilobal dynamic substructure of RasD reminiscent of that observed in Ras itself. Nucleotide turnover led to a modulation of the coupling between these substructures with an overall dynamical tightening in the GTP state and enhanced HD-RasD couplings in the GDI state. Network path analysis and subsequent mutant simulations highlighted residues of potential importance for the coordination of receptor and nucleotide-binding site to the RasD-HD interface. In particular, the on-path mutation D333A displayed disrupted dynamic coupling between distal functional sites, whereas L32A, F332A, and I339A led to an enhanced helical domain opening. Experimental characterization of D333A and L32A revealed constitutive activity in the absence of receptor supporting the functional relevance of these allosteric mutations. Mutations of the additionally highlighted Phe332 and Ile339 have been shown previously to result in enhanced spontaneous rates of nucleotide exchange (
      • Marin E.P.
      • Krishna A.G.
      • Sakmar T.P.
      Rapid activation of transducin by mutations distant from the nucleotide-binding site: evidence for a mechanistic model of receptor-catalyzed nucleotide exchange by G proteins.
      ,
      • Kaya A.I.
      • Lokits A.D.
      • Gilbert J.A.
      • Iverson T.M.
      • Meiler J.
      • Hamm H.E.
      A conserved phenylalanine as relay between the α5 helix and the GDP binding region of heterotrimeric Gi protein α subunit.
      ).
      Recent alanine scanning mutagenesis and thermostability assays by Sun et al. (
      • Sun D.
      • Flock T.
      • Deupi X.
      • Maeda S.
      • Matkovic M.
      • Mendieta S.
      • Mayer D.
      • Dawson R.J.
      • Schertler G.F.
      • Babu M.M.
      • Veprintsev D.B.
      Probing Gαi1 protein activation at single-amino acid resolution.
      ) identified clusters of hydrophobic residues that confer differential stability to GDP-, GTPγS-, and rhodopsin-bound Gαi1. In particular, residues in α5 (most notably Phe336, which is equivalent to our Phe332), α1 (including Ile49, Gln52, and Met53), and β1 (Leu38 equivalent to Leu34, which is a neighboring residue to our Leu32) were found to confer greater thermal stability to GDP- and GTPγS-bound states. Importantly, F336A was revealed to be the only substitution that resulted in both loss of stability and altered nucleotide binding kinetics. Using independent biophysical modeling, we report here that this substitution results in increased domain opening rates. This provides a clear structural dynamic perspective consistent with the previous finding that this mutation increases the rate of nucleotide release (
      • Kaya A.I.
      • Lokits A.D.
      • Gilbert J.A.
      • Iverson T.M.
      • Meiler J.
      • Hamm H.E.
      A conserved phenylalanine as relay between the α5 helix and the GDP binding region of heterotrimeric Gi protein α subunit.
      ). In a similar manner, we report that the novel L32A mutation results in enhanced helical domain opening, increased nucleotide binding rates, and constitutive activity in the absence of receptor.
      Using differential contact analysis of Gα crystal structure subsets, Flock et al. (
      • Flock T.
      • Ravarani C.N.
      • Sun D.
      • Venkatakrishnan A.J.
      • Kayikci M.
      • Tate C.G.
      • Veprintsev D.B.
      • Babu M.M.
      Universal allosteric mechanism for Gα activation by GPCRs.
      ) recently suggested that structural contacts between α1 and α5 act as a “hub” for Gα allosteric activation. In this model, activation is mediated by the breaking of contacts between α5 and α1, leading to an increased flexibility of α1 that promotes GDP release. The allosteric importance of α5 was also highlighted in recent long time scale MD simulations by Dror et al. (
      • Dror R.O.
      • Mildorf T.J.
      • Hilger D.
      • Manglik A.
      • Borhani D.W.
      • Arlow D.H.
      • Philippsen A.
      • Villanueva N.
      • Yang Z.
      • Lerch M.T.
      • Hubbell W.L.
      • Kobilka B.K.
      • Sunahara R.K.
      • Shaw D.E.
      Structural basis for nucleotide exchange in heterotrimeric G proteins.
      ). These simulations suggested that α5 displacement upon receptor binding results in an increased flexibility of the β6-α5 loop. This loop, located at the N terminus of α5, coordinates the guanosine moiety of a bound nucleotide. Interestingly, Flock et al. state that residues contacting the guanosine moiety, including the β6-α5 “are not extensively reorganized during Gα activation.” In both the Flock et al. and Dror et al. models, α5 acts as the primary initial conduit of information transfer between the receptor and nucleotide binding sites. The models differ in that Dror et al. propose that flexibility differences of the β6-α5 loop complete the connection to the guanosine moiety, whereas Flock et al. propose that increased α1 dynamics is the primary determinant of allosteric coupling. Our path analysis results support the importance of α5 in general and the β6-α5 loop in particular (Dror et al. model). However, we provide new evidence for a dominant alternate allosteric coupling route through β1 that directly links from receptor to the phosphate coordinating P-loop. Both the C-terminal of α5 and the N-terminal β1 are known GPCR binding interfaces. The increased dynamics of both regions upon receptor binding were also evident in earlier hydrogen/deuterium exchange data (
      • Chung K.Y.
      • Rasmussen S.G.
      • Liu T.
      • Li S.
      • DeVree B.T.
      • Chae P.S.
      • Calinski D.
      • Kobilka B.K.
      • Woods Jr., V.L.
      • Sunahara R.K.
      Conformational changes in the G protein Gs induced by the β2 adrenergic receptor.
      ). Moreover, our analysis of the structural dynamic effects of mutations in these regions reveals the novel role of β1 together with β2, β3, P-loop, and SI in the regulation of domain opening that is critical for nucleotide exchange.
      More frequent RasD-HD domain separation has previously been suggested to underlie the self-activation of the G protein GPA1 from Arabidopsis thaliana (
      • Jones J.C.
      • Jones A.M.
      • Temple B.R.
      • Dohlman H.G.
      Differences in intradomain and interdomain motion confer distinct activation properties to structurally similar Gα proteins.
      ). GPA1 is permanently activated, has enhanced nucleotide exchange rates, and displays enhanced domain opening in simulations relative to Gαi. Intriguingly, investigations of chimeric proteins established that the HD αA helix of GPA1 is almost entirely responsible for this enhanced activation. We note that the αA helix spans the two major HD communities (Fig. 3, b and c) and that perturbations to αA have the potential to effect dynamic couplings in the entire HD region. Collectively, our mutational analysis and the GPA1 chimeric analysis indicate that sites distant from regions involved in binding to receptors, effectors, and nucleotides can perturb the structural dynamics and function of G proteins.
      Our results indicate that network analysis of dynamic couplings from multiple replicate MD simulations is a promising method to delineate features of protein allostery. Similar network approaches have been successfully applied to a number of important biological systems (
      • Yao X.-Q.
      • Grant B.J.
      Domain-opening and dynamic coupling in the α-subunit of heterotrimeric G proteins.
      ,
      • Sethi A.
      • Eargle J.
      • Black A.A.
      • Luthey-Schulten Z.
      Dynamical networks in tRNA:protein complexes.
      ,
      • Vanwart A.T.
      • Eargle J.
      • Luthey-Schulten Z.
      • Amaro R.E.
      Exploring residue component contributions to dynamical network models of allostery.
      • Gasper P.M.
      • Fuglestad B.
      • Komives E.A.
      • Markwick P.R.
      • McCammon J.A.
      Allosteric networks in thrombin distinguish procoagulant vs. anticoagulant activities.
      ,
      • Scarabelli G.
      • Grant B.J.
      Mapping the structural and dynamical features of kinesin motor domains.
      ,
      • Scarabelli G.
      • Grant B.J.
      Kinesin-5 allosteric inhibitors uncouple the dynamics of nucleotide, microtubule, and neck-linker binding sites.
      • Guo J.
      • Pang X.
      • Zhou H.-X.
      Two pathways mediate interdomain allosteric regulation in Pin1.
      ). The major improvement in our current implementation versus our previous work (
      • Yao X.-Q.
      • Grant B.J.
      Domain-opening and dynamic coupling in the α-subunit of heterotrimeric G proteins.
      ,
      • Scarabelli G.
      • Grant B.J.
      Mapping the structural and dynamical features of kinesin motor domains.
      ,
      • Scarabelli G.
      • Grant B.J.
      Kinesin-5 allosteric inhibitors uncouple the dynamics of nucleotide, microtubule, and neck-linker binding sites.
      ) and that of others is the use of many multiple replicate trajectories instead of results from single simulations. This reduces statistical errors in the calculated cross-correlation matrix and resulting correlation network and importantly allows for a more robust statistical assessment of within state and between state dynamic coupling differences. It is important to note that this widely applicable approach provides structural and dynamic insights that are not immediately available from accumulated crystal structures or individual pairs of trajectories. Furthermore, combining this approach with targeted computational and experimental mutagenesis lays the foundation for dissecting the dynamic consequences of disease-associated mutations and the potential generality of allosteric coupling mechanisms in related GTPase and ATPase systems.

      Author Contributions

      B. J. G., X.-Q. Y., S. S., and J. R. T. designed the study. X.-Q. Y. and B. J. G. performed and analyzed the MD simulations. R. M. and N. W. G. performed and analyzed the experiment. X.-Q. Y., B. J. G., and L. S. developed methods for the computational modeling and analysis. All authors reviewed the results and wrote the manuscript.

      Acknowledgments

      We thank Drs. G. Scarabelli and J. Tesmer for valuable discussions.

      References

        • Neves S.R.
        • Ram P.T.
        • Iyengar R.
        G protein pathways.
        Science. 2002; 296: 1636-1639
        • Sprang S.R.
        G protein mechanisms: insights from structural analysis.
        Annu. Rev. Biochem. 1997; 66: 639-678
        • Yao X.-Q.
        • Grant B.J.
        Domain-opening and dynamic coupling in the α-subunit of heterotrimeric G proteins.
        Biophys. J. 2013; 105: L08-L10
        • Rasmussen S.G.
        • DeVree B.T.
        • Zou Y.
        • Kruse A.C.
        • Chung K.Y.
        • Kobilka T.S.
        • Thian F.S.
        • Chae P.S.
        • Pardon E.
        • Calinski D.
        • Mathiesen J.M.
        • Shah S.T.
        • Lyons J.A.
        • Caffrey M.
        • Gellman S.H.
        • Steyaert J.
        • Skiniotis G.
        • Weis W.I.
        • Sunahara R.K.
        • Kobilka B.K.
        Crystal structure of the β2 adrenergic receptor-Gs protein complex.
        Nature. 2011; 477: 549-555
        • Westfield G.H.
        • Rasmussen S.G.
        • Su M.
        • Dutta S.
        • DeVree B.T.
        • Chung K.Y.
        • Calinski D.
        • Velez-Ruiz G.
        • Oleskie A.N.
        • Pardon E.
        • Chae P.S.
        • Liu T.
        • Li S.
        • Woods Jr., V.L.
        • Steyaert J.
        • Kobilka B.K.
        • Sunahara R.K.
        • Skiniotis G.
        Structural flexibility of the Gαs α-helical domain in the β2-adrenoceptor Gs complex.
        Proc. Natl. Acad. Sci. U.S.A. 2011; 108: 16086-16091
        • Van Eps N.
        • Preininger A.M.
        • Alexander N.
        • Kaya A.I.
        • Meier S.
        • Meiler J.
        • Hamm H.E.
        • Hubbell W.L.
        Interaction of a G protein with an activated receptor opens the interdomain interface in the α subunit.
        Proc. Natl. Acad. Sci. U.S.A. 2011; 108: 9420-9424
        • Chung K.Y.
        • Rasmussen S.G.
        • Liu T.
        • Li S.
        • DeVree B.T.
        • Chae P.S.
        • Calinski D.
        • Kobilka B.K.
        • Woods Jr., V.L.
        • Sunahara R.K.
        Conformational changes in the G protein Gs induced by the β2 adrenergic receptor.
        Nature. 2011; 477: 611-615
        • Jones J.C.
        • Duffy J.W.
        • Machius M.
        • Temple B.R.
        • Dohlman H.G.
        • Jones A.M.
        The crystal structure of a self-activating G protein α subunit reveals its distinct mechanism of signal initiation.
        Sci. Signal. 2011; 4: ra8
        • Jones J.C.
        • Jones A.M.
        • Temple B.R.
        • Dohlman H.G.
        Differences in intradomain and interdomain motion confer distinct activation properties to structurally similar Gα proteins.
        Proc. Natl. Acad. Sci. U.S.A. 2012; 109: 7275-7279
        • Dror R.O.
        • Mildorf T.J.
        • Hilger D.
        • Manglik A.
        • Borhani D.W.
        • Arlow D.H.
        • Philippsen A.
        • Villanueva N.
        • Yang Z.
        • Lerch M.T.
        • Hubbell W.L.
        • Kobilka B.K.
        • Sunahara R.K.
        • Shaw D.E.
        Structural basis for nucleotide exchange in heterotrimeric G proteins.
        Science. 2015; 348: 1361-1365
        • Oldham W.M.
        • Hamm H.E.
        Heterotrimeric G protein activation by G-protein-coupled receptors.
        Nat. Rev. Mol. Cell Biol. 2008; 9: 60-71
        • Iiri T.
        • Herzmark P.
        • Nakamoto J.M.
        • van Dop C.
        • Bourne H.R.
        Rapid GDP release from Gsα in patients with gain and loss of endocrine function.
        Nature. 1994; 371: 164-168
        • Marin E.P.
        • Krishna A.G.
        • Sakmar T.P.
        Rapid activation of transducin by mutations distant from the nucleotide-binding site: evidence for a mechanistic model of receptor-catalyzed nucleotide exchange by G proteins.
        J. Biol. Chem. 2001; 276: 27400-27405
        • Kapoor N.
        • Menon S.T.
        • Chauhan R.
        • Sachdev P.
        • Sakmar T.P.
        Structural evidence for a sequential release mechanism for activation of heterotrimeric G proteins.
        J. Mol. Biol. 2009; 393: 882-897
        • Preininger A.M.
        • Funk M.A.
        • Oldham W.M.
        • Meier S.M.
        • Johnston C.A.
        • Adhikary S.
        • Kimple A.J.
        • Siderovski D.P.
        • Hamm H.E.
        • Iverson T.M.
        Helix dipole movement and conformational variability contribute to allosteric GDP release in Gαi subunits.
        Biochemistry. 2009; 48: 2630-2642
        • Thaker T.M.
        • Sarwar M.
        • Preininger A.M.
        • Hamm H.E.
        • Iverson T.M.
        A transient interaction between the phosphate binding loop and switch I contributes to the allosteric network between receptor and nucleotide in Gαi1.
        J. Biol. Chem. 2014; 289: 11331-11341
        • Kaya A.I.
        • Lokits A.D.
        • Gilbert J.A.
        • Iverson T.M.
        • Meiler J.
        • Hamm H.E.
        A conserved phenylalanine as relay between the α5 helix and the GDP binding region of heterotrimeric Gi protein α subunit.
        J. Biol. Chem. 2014; 289: 24475-24487
        • Grant B.J.
        • Rodrigues A.P.
        • ElSawy K.M.
        • McCammon J.A.
        • Caves L.S.
        Bio3d: an R package for the comparative analysis of protein structures.
        Bioinformatics. 2006; 22: 2695-2696
        • Skjaerven L.
        • Yao X.Q.
        • Scarabelli G.
        • Grant B.J.
        Integrating protein structural dynamics and evolutionary analysis with Bio3D.
        BMC Bioinformatics. 2014; 15: 399
        • Humphrey W.
        • Dalke A.
        • Schulten K.
        VMD: visual molecular dynamics.
        J. Mol. Graph. 1996; 14: 33-38
        • Berman H.M.
        • Westbrook J.
        • Feng Z.
        • Gilliland G.
        • Bhat T.N.
        • Weissig H.
        • Shindyalov I.N.
        • Bourne P.E.
        The Protein Data Bank.
        Nucleic Acids Res. 2000; 28: 235-242
        • Gerstein M.
        • Altman R.B.
        Average core structures and variability measures for protein families: application to the immunoglobulins.
        J. Mol. Biol. 1995; 251: 161-175
        • Caves L.S.
        • Evanseck J.D.
        • Karplus M.
        Locally accessible conformations of proteins: multiple molecular dynamics simulations of crambin.
        Protein Sci. 1998; 7: 649-666
        • Gorfe A.A.
        • Grant B.J.
        • McCammon J.A.
        Mapping the nucleotide and isoform-dependent structural and dynamical features of Ras proteins.
        Structure. 2008; 16: 885-896
        • Grant B.J.
        • McCammon J.A.
        • Caves L.S.
        • Cross R.A.
        Multivariate analysis of conserved sequence-structure relationships in kinesins: coupling of the active site and a tubulin-binding sub-domain.
        J. Mol. Biol. 2007; 368: 1231-1248
        • van Aalten D.M.
        • Conn D.A.
        • de Groot B.L.
        • Berendsen H.J.
        • Findlay J.B.
        • Amadei A.
        Protein dynamics derived from clusters of crystal structures.
        Biophys. J. 1997; 73: 2891-2896
        • Case D.A.
        • Darden T.A.
        • Cheatham T.E.I.I.I.
        • Simmerling C.L.
        • Wang J.
        • Duke R.E.
        • Luo R.
        • Walker R.C.
        • Zhang W.
        • Merz K.M.
        • Roberts B.
        • Hayik S.
        • Roitberg A.
        • Seabra G.
        • Wails J.
        • Goetz A.W.
        • Kolossvary I.
        • Wong K.F.
        • Paesani F.
        • Vanicek J.
        • Wolf R.M.
        • Liu J.
        • Wu X.
        • Brozell S.R.
        • Steinbrecher T.
        • Gohlke H.
        • Cai Q.
        • Ye X.
        • Hsieh M.J.
        • Cui G.
        • Roe D.R.
        • Mathews D.H.
        • Seetin M.G.
        • Salomon-Ferrer R.
        • Sagui C.
        • Babin V.
        • Luchko T.
        • Gusarov S.
        • Kovalenko A.
        • Kollman P.A.
        AMBER 12. University of California, San Francisco, CA2012
        • Hornak V.
        • Abel R.
        • Okur A.
        • Strockbine B.
        • Roitberg A.
        • Simmerling C.
        Comparison of multiple Amber force fields and development of improved protein backbone parameters.
        Proteins. 2006; 65: 712-725
        • Meagher K.L.
        • Redman L.T.
        • Carlson H.A.
        Development of polyphosphate parameters for use with the AMBER force field.
        J. Comput. Chem. 2003; 24: 1016-1025
        • Li H.
        • Robertson A.D.
        • Jensen J.H.
        Very fast empirical prediction and rationalization of protein pKa values.
        Proteins. 2005; 61: 704-721
        • Ichiye T.
        • Karplus M.
        Collective motions in proteins: a covariance analysis of atomic fluctuations in molecular dynamics and normal mode simulations.
        Proteins. 1991; 11: 205-217
        • Dubay K.H.
        • Bothma J.P.
        • Geissler P.L.
        Long-range intra-protein communication can be transmitted by correlated side-chain fluctuations alone.
        PLoS Comput. Biol. 2011; 7: e1002168
        • Wu S.
        • Jun Lee C.
        • Pedersen L.G.
        Analysis on long-range residue-residue communication using molecular dynamics.
        Proteins. 2014; 82: 2896-2901
        • Sethi A.
        • Eargle J.
        • Black A.A.
        • Luthey-Schulten Z.
        Dynamical networks in tRNA:protein complexes.
        Proc. Natl. Acad. Sci. U.S.A. 2009; 106: 6620-6625
        • Girvan M.
        • Newman M.E.
        Community structure in social and biological networks.
        Proc. Natl. Acad. Sci. U.S.A. 2002; 99: 7821-7826
        • Yen J.Y.
        Finding K shortest loopless paths in a network.
        Management Science Series A-Theory. 1971; 17: 712-716
        • Malik R.U.
        • Ritt M.
        • DeVree B.T.
        • Neubig R.R.
        • Sunahara R.K.
        • Sivaramakrishnan S.
        Detection of G protein-selective G protein-coupled receptor (GPCR) conformations in live cells.
        J. Biol. Chem. 2013; 288: 17167-17178
        • Gabriel D.
        • Vernier M.
        • Pfeifer M.J.
        • Dasen B.
        • Tenaillon L.
        • Bouhelal R.
        High throughput screening technologies for direct cyclic AMP measurement.
        Assay Drug Dev. Technol. 2003; 1: 291-303
        • Slepak V.Z.
        • Wilkie T.M.
        • Simon M.I.
        Mutational analysis of G protein α subunit Goα expressed in Escherichia coli.
        J. Biol. Chem. 1993; 268: 1414-1423
        • Grishina G.
        • Berlot C.H.
        A surface-exposed region of Gsα in which substitutions decrease receptor-mediated activation and increase receptor affinity.
        Mol. Pharmacol. 2000; 57: 1081-1092
        • Nishina H.
        • Nimota K.
        • Kukimoto I.
        • Maehama T.
        • Takahashi K.
        • Hoshino S.
        • Kanaho Y.
        • Katada T.
        Significance of Thr182 in the nucleotide-exchange and GTP-hydrolysis reactions of the α subunit of GTP-binding protein Gi2.
        J. Biochem. 1995; 118: 1083-1089
        • Thomas T.C.
        • Schmidt C.J.
        • Neer E.J.
        G-protein αo subunit: mutation of conserved cysteines identifies a subunit contact surface and alters GDP affinity.
        Proc. Natl. Acad. Sci. U.S.A. 1993; 90: 10295-10299
        • Lerman B.B.
        • Dong B.
        • Stein K.M.
        • Markowitz S.M.
        • Linden J.
        • Catanzaro D.F.
        Right ventricular outflow tract tachycardia due to a somatic cell mutation in G protein subunit αi2.
        J. Clin. Invest. 1998; 101: 2862-2868
        • Nesbit M.A.
        • Hannan F.M.
        • Howles S.A.
        • Babinsky V.N.
        • Head R.A.
        • Cranston T.
        • Rust N.
        • Hobbs M.R.
        • Heath 3rd, H.
        • Thakker R.V.
        Mutations affecting G-protein subunit α11 in hypercalcemia and hypocalcemia.
        N. Engl. J. Med. 2013; 368: 2476-2486
        • Alexander N.S.
        • Preininger A.M.
        • Kaya A.I.
        • Stein R.A.
        • Hamm H.E.
        • Meiler J.
        Energetic analysis of the rhodopsin-G-protein complex links the α5 helix to GDP release.
        Nat. Struct. Mol. Biol. 2014; 21: 56-63
        • Sun D.
        • Flock T.
        • Deupi X.
        • Maeda S.
        • Matkovic M.
        • Mendieta S.
        • Mayer D.
        • Dawson R.J.
        • Schertler G.F.
        • Babu M.M.
        • Veprintsev D.B.
        Probing Gαi1 protein activation at single-amino acid resolution.
        Nat. Struct. Mol. Biol. 2015; 22: 686-694
        • Locht C.
        • Coutte L.
        • Mielcarek N.
        The ins and outs of pertussis toxin.
        FEBS J. 2011; 278: 4668-4682
        • Harrison C.
        • Traynor J.R.
        The [35S]GTPgammaS binding assay: approaches and applications in pharmacology.
        Life Sci. 2003; 74: 489-508
        • Flock T.
        • Ravarani C.N.
        • Sun D.
        • Venkatakrishnan A.J.
        • Kayikci M.
        • Tate C.G.
        • Veprintsev D.B.
        • Babu M.M.
        Universal allosteric mechanism for Gα activation by GPCRs.
        Nature. 2015; 524: 173-179
        • Vanwart A.T.
        • Eargle J.
        • Luthey-Schulten Z.
        • Amaro R.E.
        Exploring residue component contributions to dynamical network models of allostery.
        J. Chem. Theory Comput. 2012; 8: 2949-2961
        • Gasper P.M.
        • Fuglestad B.
        • Komives E.A.
        • Markwick P.R.
        • McCammon J.A.
        Allosteric networks in thrombin distinguish procoagulant vs. anticoagulant activities.
        Proc. Natl. Acad. Sci. U.S.A. 2012; 109: 21216-21222
        • Scarabelli G.
        • Grant B.J.
        Mapping the structural and dynamical features of kinesin motor domains.
        PLoS Comput. Biol. 2013; 9: e1003329
        • Scarabelli G.
        • Grant B.J.
        Kinesin-5 allosteric inhibitors uncouple the dynamics of nucleotide, microtubule, and neck-linker binding sites.
        Biophys. J. 2014; 107: 2204-2213
        • Guo J.
        • Pang X.
        • Zhou H.-X.
        Two pathways mediate interdomain allosteric regulation in Pin1.
        Structure. 2015; 23: 237-247