|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 47, 47136-47144, November 21, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Lipid Bilayer Simulations of CXCR4 with Inverse Agonists and Weak Partial Agonists*
¶||
From the
J. G. Brown Modeling Facility, Brown Cancer Center, ¶Department of Medicine, University of Louisville, Louisville, Kentucky 40202, the **Department of Pathology and Institute of Molecular Medicine and Genetics, Medical College of Georgia, Augusta, Georgia 30912, and the Graduate School of Pharmaceutical Sciences, Kyoto University, Kyoto, Japan
Received for publication, July 21, 2003 , and in revised form, September 3, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
CXCR4 is a G protein-coupled receptor (GPCR) that has multiple critical functions in normal and pathologic physiology that include regulation of the metastatic behavior of mammary carcinoma, and utilization as a coreceptor for infection by T-tropic strains of human immunodeficiency virus-1. Molecular dynamic simulations of the rhodopsin-based homology model of CXCR4 were performed in a solvated lipid bilayer to reproduce the microenvironment of this integral membrane protein. The amino acids in CXCR4 necessary for interaction with an inverse agonist, T140, and a weak partial agonist, AMD3100, identified by alanine scanning mutants, were spatially consistent when computationally docked. Whereas T140 binds residues in extracellular domains and regions of the hydrophobic core proximal to the cell surface, amino acids in the central hydrophobic core are critical to binding of AMD3100. The physical localization of T140 binding to CXCR4 by biochemical analyses corroborated the molecular and computational approaches. The structural basis for the interaction of T140 and AMD3100 with CXCR4 confirms that the mechanisms used by these agents are different. This complementary utilization of molecular, physical, and computation analysis provides a powerful approach to elucidate GPCR conformation.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
CXCR4 is a GPCR that exclusively transduces cellular signals for stromal cell-derived factor 1 (SDF-1),1 a member of the CXC branch of the chemokine family also designated CXCL12. SDF-1 is chemotactic for B- and T-lymphocyte subsets and is critical for the migration of progenitors during embryologic development of the central nervous, cardiovascular, and hematopoietic systems (1, 2). CXCR4 has been shown to have a key role in germ cell migration during embryogenesis in zebrafish (3) and blockade of this receptor in humans results in mobilization of hematopoietic stem cells (4). CXCR4 is a front line coreceptor for the CD4-dependent entry of target cells by T-tropic strains of HIV-1 that evolve during the course of HIV-1 infection (5). In addition, the expression of CXCR4 by mammary carcinoma and other malignancies results in the hijacking of its ability to mediate directed migration resulting in the programming of metastatic spread to target organs that secrete SDF-1 (i.e. lymph node, bone marrow, lung, and liver). Blockade of CXCR4 with a monoclonal antibody has been reported to inhibit the metastasis of mammary carcinoma cells in a xenograft model in immunodeficient mice (6). Thus, CXCR4 represents a critical molecular target to disrupt the pathogenesis of HIV-1 infection, to block tumor metastasis, and to mobilize stem cells.
Three CXCR4 antagonists have been described and two (AMD3100 (7) (Fig. 1) and ALX40-4C (8)) have been administered to human subjects. AMD3100 was withdrawn from clinical trials in patients with HIV-1 infections because of cardiac toxicity.2 The mechanism of antagonist action has been determined using human CXCR4 mutants with constitutive activity derived in yeast strains expressing the receptor functionally coupled to the pheromone response pathway (10). AMD3100 and ALX40-4C demonstrated weak partial agonist activity (10). In contrast, T140 (11) (Fig. 1), a 14-residue polypeptide downsized from a naturally occurring horseshoe crab protein, shifted the conformational equilibrium of wild type and constitutively active mutants of CXCR4 to the inactive state, characteristic of an inverse agonist (10). Whereas the mechanism for CXCR4 blockade does not appear to be critical to therapeutic applications for coreceptor antagonists, it is important to develop agents that lack partial activating activity, such as inverse agonists, for therapies to block the metastatic spread of malignancies.
|
Insight into the structure of molecular targets may provide opportunities for rational drug design or strategies to enhance the activity of candidate lead compounds identified by high throughput screening. Although GPCRs are frequent targets for therapeutic drugs, understanding of their structure is quite limited. The only available high resolution GPCR structure has been determined by x-ray crystallographic analysis of bovine rhodopsin (12). Biophysical analyses have elucidated the mechanism for reorientation of transmembrane helices involved in receptor activation and initiation of signaling (13). Whereas a variety of structural and modeling approaches have been attempted for other GPCRs, it has not been possible to establish a reliable understanding of native receptor structure. There is increased interest (14, 15) in the use of homology models of GPCRs and computational analyses have been applied to extend observations of dynamic equilibrium of GPCR conformation. The interpretation of such molecular modeling experiments is constrained by the performance of the dynamic simulations in a vacuum or water, neither of which accurately reproduces the membrane environment of GPCRs.
Here we describe a multidisciplinary approach to elucidate the structural conformation of GPCRs that involves molecular dynamic simulations in a solvated lipid bilayer, identification of residues necessary for receptor interaction with antagonists, verification of these findings using biochemical strategies, and computational docking of antagonists and receptors. This iterative strategy should provide insight into mechanisms for GPCR activation and receptor domains critical to interaction with pharmacophores that influence the dynamic equilibrium between active and inactive conformations. This understanding should, in turn, permit the development of a new generation of CXCR4 antagonists by rational drug design.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
Molecular Dynamic Simulations—A lipid bilayer was initially constructed comprised of two layers of 256 (16 x 16 grid) dodecamaltoside molecules in the xy plane aligned to the ±z axis. The homology model of CXCR4, derived from the rhopdopsin crystal structure, was inserted centrally aligned to the z axis and the dodecamaltoside molecules within 3 Å of the protein residues were removed, leaving 375 lipid molecules. The system was solvated in the ±z direction 12 Å away from CXCR4 and water molecules were removed from the lipid bilayer leaving 10,111 TIP3P solvent molecules. The final dimensions of the system were 75 x 75 x 113 Å. The system was carefully equilibrated in three stages: initially the lipid was allowed to gradually interact with the fixed protein with water omitted from the calculation. Second, the water was allowed to gradually interact with unrestrained lipid and fixed protein. Finally, the protein was allowed to gradually interact with the unrestrained water and lipid. The protocol for each stage was: minimization with 100 kcal (mol Å)-1 restraints (5,000 steps steepest descents, 20,000 steps conjugate gradient); molecular dynamics (100 K, 25 ps); minimization with 100 kcal (mol Å)-1 restraints (5,000 steps steepest descent, 20,000 steps conjugate gradient); minimization (5,000 steps steepest descent, 20,000 steps conjugate gradient); molecular dynamics with 100 kcal (mol Å)-1 restraints (100 K, 25 ps); molecular dynamics with 100 kcal (mol Å)-1 restraints (300 K, 25 ps); molecular dynamics (300 K, 75 ps) with gradual removal of restraints. The equilibrium phase ended with 500 ps of unrestrained molecular dynamics to allow for further movement. The component that is fixed had constant 100 kcal (mol Å)-1 restraints, whereas the component that is allowed to gradually interact had the gradual removal of restraints. Molecular dynamics simulations using the AMBER-96 force field and the Sander_classic module were performed in the isothermal isobaric ensemble (p = 1 atmosphere) with the AMBER 6.0 program (although initial benchmarks on this system indicate AMBER 7.0 is significantly faster), using periodic boundary conditions and the PME algorithm. A 2-fs time step was used with all bond distances involving hydrogen atoms frozen using SHAKE. Chloride anions were added randomly for charge neutrality. The production runs were unrestrained and run for 6 ns on a 32 processor SGI Origin2000. The structures were sampled in the last 2 ns.
The simulations including T140 and AMD3100 effectively used steered molecular dynamics for the equilibrium period that started with the fully equilibrated lipid:protein without water, T140 or AMD3100 were added 10 Å away from the CXCR4 and the total system was solvated at 12 Å in the z direction. The third stage of the above protocol was then repeated with water equilibrating and T140 and AMD3100 fixed. The final 500 ps of equilibrium had additional distance restraints (20 kcal (mol Å)-1) of 4 Å for pairs of atoms from two residues known to be important in CXCR4 to the terminal residues in two different orientations of T140 and one of AMD3100. This allowed for the gradual introduction of the ligand with time for receptor reorganization. No distance restraints were used in the production phase. T140, AMD3100, and dodecylmaltoside were parameterized for AMBER and electrostatic potential charges were derived from ab initio calculations (6-311G* basis set).
HIV-1 Envelope-mediated Fusion Assay—The ability of CXCR4 antagonists to block its coreceptor function was determined in a fusion assay using effector QT6 cells programmed to express the envelope glycoprotein from the 89.6 strain of HIV-1 and target QT6 cells transiently expressing CXCR4 alanine scanning mutants and CD4 as previously described (16).
Displacement of SDF-1 Binding—The loss of sensitivity to antagonists by CXCR4 alanine scanning mutants in the coreceptor assay was confirmed by homologous displacement of T140 binding or heterologous displacement of [125I]SDF-1 binding by AMD3100 using standard techniques (17). CHO cell transfectants stably expressing high levels of various CXCR4 alanine scanning mutants, derived as previously described (10), served as target cells for these experiments.
Chemical Cross-linking and Chemical Cleavage—T140 was radioiodinated with chloramine T using standard techniques. The labeled antagonist was cross-linked to CHO transfectants expressing high levels of CXCR4 containing Myc and His epitope tags at the N and C terminus, respectively, using the bifunctional agent ethyl-3-(3-dimethylamino)propylcarbodiimide. The antagonist-receptor complex was pulled down using nickel resin and subjected to cleavage with CNBr or N-chlorosuccinimide. Labeled fragments were resolved in polyacrylamide gels containing SDS and Tricine and detected by autoradiography. Fragments from the N or C terminus were detected by Western blotting with monoclonal antibodies to the Myc or His epitope tags, respectively.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
Structural Studies on the Native Receptor—The strategy applied for computational analysis of CXCR4 conformational structure employed a solvated lipid bilayer to approximate the environment of the plasma membrane. Dodecylmaltoside was selected for the lipid bilayer, as it has been used to solubilize biologically active CXCR4 (18). Two previous CXCR4 modeling experiments performed in vacuo (19) and solvated in water (20) neglected to consider the effect of the membrane microenvironment on the receptor and both employed approaches for generating interhelical loops without consideration of the only available high resolution GPCR crystal structure. Moreover, it is unclear whether these computational strategies (19, 20) can accurately ab initio predict the structure of rhodopsin itself. These models have significantly different interhelical loop regions, both from each other and our model based on the GPCR crystal structure. Nature is highly conservative in using protein folds and therefore we thought that it was a reasonable approach to use information from all domains of the only high resolution GPCR structure available, instead of limiting analysis to the helical regions and ab initio determining new folds for the extra- and intracellular interhelical loops. However, a recent report states that first principle methods can reproduce the crystal structure of rhodopsin and the ligand orientation (21) and were applied to other GPCRs. The optimization of ligand binding used molecular mechanics and the continuum solvation approximation that appeared to work well if the binding site is deep within the receptor and there is not significant movement of the receptor upon binding. These methods may provide an alternative to homology-based models. For conformational movement, such as some ligand binding and signaling mechanisms, it may be necessary to include an environment that mimics the plasma membrane. The starting structures used here were based on a homology model of the rhodopsin crystal structure and allowed to evolve using fully solvated lipid bilayer molecular dynamic simulations.
The molecular dynamic simulations of CXCR4 using AMBER (22) were energetically stable throughout the production trajectory and there was an evolution from the rhodopsin structure to the final CXCR4 conformation. The thickness of the lipid bilayer compressed to 50 Å (from 55 Å) around the receptor associating exclusively with the predicted transmembrane helices (Fig. 2A). The lipid molecules ensheath the external aspect of the bundle of helices that forms the surface of the hydrophobic core of CXCR4 (Fig. 2B), thereby excluding the penetration of water molecules past the polar head groups of the bilayer. The lipid bilayer flattened the outer hydrophobic residues in the core of CXCR4, providing a permissive environment for changes in orientation of the helices involved in signal transduction. The small "eighth" helix in the C-terminal tail inserted in and was stabilized by the lipid bilayer (Fig. 2A).
|
The prototypic topology of GPCRs interhelical loops was preserved although the precise conformation achieved an equilibrium that was different from the initial homology model. The solvent-exposed N-terminal extracellular domain formed a loop that extended over the extracellular loop (E-L) 2 region (Fig. 2C). The polar residues of the small E-L1 and E-L3 loops were exposed to solvent, thus stabilizing the interface with the hydrophobic helices. The C-terminal segment of 30 residues distal to the eighth helix was dynamic and exposed to the solvent in the cytosol. The intracellular loops (I-L) were also exposed to solvent in the cytoplasmic compartment (Fig. 2D). The transmembrane helices were stabilized by a network of 24 hydrogen bonds (45–288, 240–73, 75–153, 75–153, 76–298, 119–84, 298–84, 298–84, 110–171, 113–171, 255–116, 298–119, 203–121, 121–203, 294–123, 73–133, 200–268, 203–255, 208–256, 256–208, 239–302, 263–286, 302–297, 119–294, repeats have two H-bonds) and nonbonded interactions.
The third helix was interrupted by a central kink that involved Asn119, a residue that has been demonstrated to be critical for the regulation of receptor signaling (10). The lipid bilayer maintained the orientation and locations of the transmembrane and eighth helices in the modeled native structure as they are very similar to the rhodopsin crystal structure with little movement or change in the orientation of the transmembrane helices. However, there was a slight inward movement of TM7 that effectively reduces the corresponding retinal binding pocket of rhodopsin, which may account for the lack of activity of retinal-like compounds with CXCR4. Whereas the rhodopsin crystal structure was solved with the complexed retinal in place, the native CXCR4 receptor model did not include such a ligand, and the fact that the CXCR4 model was different from rhodopsin at this site is significant, as the simulation may reproduce the inactive conformation of the receptor. One advantage of the model is that it may be possible to design CXCR4 antagonists that lock the receptor in the inactive conformation based on the model. The major differences between the high resolution rhodopsin structure and the CXCR4 model were predicted to occur in the extracellular domains. These involved the extension of the N terminus over the E-L2 of CXCR4 and the greater length of the CXCR4 E-L3 than the corresponding domain of the rhodopsin structure. The first and second cytoplasmic loops of the model were also very similar to the crystal structure. The C terminus and I-L3 were not well defined in the crystal structure, thus it is not possible to make a meaningful comparison. The importance of an experimentally untested salt bridge, Arg188-Glu277, predicted by a previous solvated CXCR4 model (20) was not supported by our model as they are 22 Å apart, nor was this interaction predicted by the in vacuo model (19).
Chimeric studies (23) using CXCR4 and CXCR2 have implicated the N-terminal domain (N-ter) for SDF-1 binding and have demonstrated that a segment that includes TM6, E-L3, and TM7 is required for signaling. This is entirely consistent with the proposed model of the native CXCR4 receptor as the N-ter is obviously accessible to SDF-1, as is E-L3, and the segment that encompasses E-L3, the portions of TM6 and TM7, could translate or rotate and/or change conformation to be involved in the signaling mechanism. Alanine scanning of CXCR4 (24) has identified residues important for HIV coreceptor activity and SDF-1 binding, which are rationalized with this model. HIV coreceptor activity requires Tyr7, Tyr12, Asp97, Asp193, which are solvent accessible and available for direct interactions with glycoprotein 120 in our native CXCR4 model. Tyr12 is also involved in stabilization of the N-ter. Asp97 potentially stabilizes the base of E-L1 by interactions with TM1 and TM2. Asp262 is involved in stabilization of E-L3. The residues important for Site I binding of SDF-1 but not signaling, Glu14, Glu15, and Tyr21 are all solvent accessible and available for direct interactions with SDF-1, and Tyr21 is also involved in stabilization of the N-ter. Residues required for both SDF-1 binding and signaling and probably part of Site II are the first residues (2–9) of the N-ter, Asp97, Asp187, and Glu288. The residues in the N-ter are directly accessible for contact with SDF-1, Asp97 could potentially regulate movement of TM1 and TM2, Asp187 is anchoring the E-L2 to the helix bundle by forming a direct interaction with TM7, and Glu288 stabilizes the base of E-L3 on TM7. Asp187 and Glu288 are also consistent with the chimera studies (23).
Mapping of the CXCR4 Antagonists Binding to Validate the Predicted Structure—T140 and AMD3100 are CXCR4 antagonists of known structure. To validate the model, the binding sites for each antagonist, which use different mechanisms for blockade (10), were determined using biochemical and/or molecular biology approaches. This provides a molecular blueprint that will ultimately allow for the validation of the predicted structure with docked antagonists. CXCR4 alanine scanning mutants in individual acidic, basic, and aromatic residues in extracellular domains and the peripheral half of transmembrane spanning helices and proline and glycine residues in the latter segments were tested for sensitivity to inhibition by T140 in fusion assays with the 89.6 envelope glycoprotein. As shown in Fig. 3A, conversion of five CXCR4 residues to alanine, Asp171, Arg188, Tyr190, Gly207, or Asp262, resulted in loss of sensitivity to T140 inhibition of Env-mediated fusion.
|
Effect of Alanine Scanning Mutations on Antagonist Binding—The alanine scanning mutations shown to decrease the sensitivity of CXCR4 coreceptor activity to inhibition by T140 were analyzed for the ability to bind T140 in homologous and heterologous binding experiments. Whereas [125I]T140 bound transfectants stably expressing wild type CXCR4, conversion of Asp171, Arg188, Tyr190, or Asp262 to alanine resulted in loss of T140 binding by transfectants expressing similar levels of the CXCR4 variant (Fig. 3B). Transfectants expressing CXCR4(G207A) also had decreased binding of T140. Since no specific binding of [125I]T140 to transfectants stably expressing CXCR4(D171A) was detected, this data is not shown in the figure. Parallel experiments to determine the ability of T140 to inhibit [125I]SDF-1 binding to CHO transfectants stably expressing CXCR4 wild type or mutants also showed reduction of T140 inhibition (data not shown).
Physical Association of T140 with E-L2 of CXCR4 —To characterize the physical association of T140 with CXCR4, the radiolabeled antagonist was chemically cross-linked to the receptor and the complex was subjected to chemical cleavage at Met or Trp, with cyanogen bromide (CNBr) or N-chlorosuccinimide, respectively. Following cleavage of [125I]T140-CXCR4 complexes with CNBr, an 18-kDa fragment was detected by autoradiography. The inability to adsorb this fragment to nickel resin via the C-terminal His epitope tag indicates that it is derived from an internal fragment that spans Thr73 to Met205, and not from a similar sized fragment from the C terminus (Fig. 3C). Autoradiography of complexes exposed to N-chlorosuccinimide revealed multiple labeled products, which probably result from partial cleavage. A single labeled fragment with Mr of 24,000 was adsorbed to nickel resin, corresponding to a region extending from Ile162 to the C-terminal His tag (data not shown). These findings provide physical evidence that T140 binds to a site between residues Ile162 and Trp195, corresponding to the segment that includes TM4 and E-L2.
Docking of the T140 Structure with the CXCR4 Model— Identification of key residues in the receptor and antagonist (25) by mutagenesis provides sufficient insight into binding mechanisms for application of a computational approach to structural rationalization. The mechanism for the interaction of T140 with CXCR4 was analyzed by steered molecular dynamics using AMBER (22) software and then allowed to develop over 4 ns of unrestrained dynamics. The docking of T140 with CXCR4 was stable and there was no evidence of shifting in the binding or dissociation of the complex. As shown in Fig. 4A, the four key residues at the N and C terminus of T140 necessary for antagonist activity (25) (Arg2, Nal3, Tyr5, Arg14) were predicted to interact directly with CXCR4. Moreover, this region of T140 bound to E-L2 of CXCR4 is consistent with mutagenesis and chemical cross-linking findings.
|
The dynamic binding site of T140, shown in Fig. 4, B–D, is composed of the N terminus, which interacts with E-L2 of CXCR4, and the C terminus is oriented inferiorly and interacts with residues in TM4 and E-L2 and E-L3. The Cys residues forming the structural disulfide link and the residues in the
hairpin turn of T140 (Lys7, Lys8, Pro9, and Tyr10) did not have direct contacts with CXCR4 within 3.0 Å. The four key T140 residues all had multiple interactions with residues in CXCR4, including amino acids in the N-ter, TM4, E-L2, TM5, and E-L3. Specific interactions between key T140 residues and amino acids in CXCR4 are shown in Fig. 4A. The Arg14 residue of T140 forms strong hydrogen bonds with Asp171 of CXCR4. The Tyr190 of CXCR4 is involved in a hydrophobic network of stabilization that interacts with Nal3 and Tyr5 of T140 and Phe174 and Phe201 of CXCR4. The fact that individually Phe174 and Phe201 did not appear to be crucial for binding may be because of partial stabilization by single mutants and or with the mutated alanine, whereas a double mutant is predicted to reduce binding. Arg188 of CXCR4 lies at the base of the active site and is involved in stabilizing the secondary structure of the hydrophobic network of interactions of Nal3 and Tyr5 of T140. Asp262 of CXCR4 is crucial for the structure of the E-L3 with multiple hydrogen bonds (CXCR4 residues 258–259 and 263–266). Gly207 of CXCR4 is close to Arg188 and steric change in this position may alter the base of the binding site because of a change in the helical nature of TM5. Arg2 of T140 is extended near the N-ter of CXCR4 and directly interacts with Asp10 and Asn11 of CXCR4. There are three effects that a single mutant residue can effect binding of a ligand to a receptor. The first is that it changes a direct interaction, such as a salt bridge, or hydrogen bond, or hydrophobic stabilization to the ligand, which we observe with Arg2, Nal3, Tyr5, and Arg14 of T140 and Asp171 of CXCR4. The second is that it structurally changes the binding site, either in shape or flexibility, which would account for Arg188 and Asp262 and possibly Gly207 of CXCR4. The third effect is related to the second, in that it hinders movement distal to the binding site, thereby altering binding or the signaling mechanism, which could also account for Gly207 of CXCR4.
Identification of Critical Residues for AMD3100 Binding— Parallel analysis of the CXCR4 alanine scanning mutants for loss of sensitivity to AMD3100 revealed a critical role for Asp262 and Glu288 (Fig. 5A). Conversion of Phe189 or Tyr190 individually to alanine did not alter the sensitivity of CXCR4 coreceptor activity to inhibition by AMD3100. A CXCR4 mutant in which both were converted to alanine retained coreceptor function that was insensitive to AMD3100.
|
Heterologous displacement experiments were performed to characterize the effect of alanine scanning mutations on AMD3100 binding. As shown in Fig. 5B, substitution of alanine for Asp262 or Glu288 dramatically reduced the ability of AMD3100 to displace [125I]SDF-1 binding to stable transfectants. The low level expression of CXCR4(F189A,Y190A) by CHO transfectants precluded ligand binding experiments.
Docking of AMD3100 into the CXCR4 Model—Computational docking of AMD3100 with CXCR4 was performed using steered molecular dynamics, as described for T140. The final model was stable and the spatial orientation of the two positively charged bicyclam rings associated with Asp262 and Glu288 and the phenylenebis(methylene) linker associated with Phe189 and Tyr190 was consistent with experimental mutagenesis data (Fig. 6A). AMD3100 bound to the base of E-L3 and was bounded superiorly by this loop (Fig. 6, B and C). The bicyclam rings extended into the hydrophobic core and outside the bundle of helices. The posterior boundary of the binding pocket consisted of hydrophobic residues from E-L2. The nonbonded interactions of AMD3100, within 3.0 Å, which are primarily hydrophobic, included the following residues: Ala180, Asp187, Tyr190, Pro191, Tyr255, Tyr256, Ile261, Asp262, Phe264, Ile265, Leu266, Ile284, Thr287, and Glu288 (Fig. 6A). Asp262 and Glu288 formed hydrogen bonds to the bicyclam rings. Glu288 also formed hydrogen bonds stabilizing the base of E-L3, which has the effect of forcing the N-ter and E-L2 together (Fig. 6D).
|
The effect of AMD3100 binding on the conformation of the CXCR4 model was different from that of T140. In comparison to the native model, there was greater movement of the transmembrane helices upon AMD3100 binding than was associated with T140 binding. Whereas the intracellular domains of CXCR4 complexed with T140 were similar to those of the native receptor, AMD3100 binding was associated with a change in orientation of the lower portions of the TM helices and cytoplasmic domains (Fig. 6D). The altered orientation provides a potential conformational rationalization for the finding that AMD3100 is a weak partial agonist, whereas T140 binding does not alter the conformation of the native receptor, which by analogy to the crystal structure of rhodopsin bound to retinal, is in the inactive state.
Whereas Asp171 has previously been implicated in AMD3100 binding to CXCR4, mutation of this residue did not alter AMD3100 inhibition of coreceptor activity. CXCR4(D171A) does have dramatically decreased SDF-1 binding (data not shown), complicating interpretation of the displacement analysis (9). Our molecular modeling provides a spatial rationalization of the involvement of Asp262 and Glu288 in AMD3100 binding to the two bicyclam rings that is also consistent with the interaction of the hydrophobic interaction of the phenylenebis(methylene) linker with the contiguous aromatic residues Phe189 and Tyr190. It would appear that one aromatic residue, Tyr190 or Phe189, is necessary for hydrophobic interaction with the phenylenebis(methylene) linker of AMD3100 and either can accommodate this as each single mutant. However, the double mutation provides no such stabilization. In our model the bicyclam rings of AMD3100 were not in the extended conformation and were close together. Although an extended conformation of AMD3100 does fit into a previous model simulation (9), that model included only the transmembrane helical regions with Asp171 and Asp262. In our system, there would have to be significant conformational change of the extracellular loop regions to allow for this interaction.
A series of AMD3100 compounds (Fig. 1) have been previously tested for activity and only one, AMD3106, is comparable with the parent compound, whereas the others have decreased affinity ranging to that of the monocyclam. Our model can be used to rationalize the observed activity range. The 3-fold decreased activity of AMD2763, which possesses a short propyl linker that holds the two cyclams close together, does not have the aromatic stabilization of AMD3100. AMD2849, with the hexyl linker, has decreased activity similar to that of the monocyclam: the cyclam rings are more flexible in conformational space, are not held in close proximity, and do not have the aromatic stabilization, thus decreasing stabilization in our model. AMD3106, with the 2,4-disubstituted pyridine linker, has similar activity to AMD3100 and it is possible for the aromatic stabilization to occur between the linker and Tyr190 and/or Phe189, whereas holding the cyclam rings in a similar proximity, as in AMD3100. An isomer of AMD3106, AMD3108, has interactions of the 2,4-disubstituted pyridine linker with the cyclam ring that hold it in an unfavorable conformation. This causes decreased activity similar to the monocyclam, the same argument is applicable in our model.
The conformation of the extracellular loops and upper regions of the TM helices of CXCR4 bound to AMD3100 are similar to those of the native receptor, with the exception of minor movement for the optimization of antagonist binding. TM3 was straightened out and there was significant movement of the cytoplasmic loops, which appeared to extend farther into the aqueous environment of the cytosol. The largest change was in the C-terminal tail, residues 344–352, which was different from the native and T140-bound CXCR4 receptor (Fig. 6D). However, the C terminus is dynamic and not well defined in the rhodopsin crystal structure. I-L2 and I-L3 were in closer proximity than in the native receptor (Fig. 6D).
Insights into CXCR4 and Structure-based Drug Design— CXCR4 is a biologically and medically important receptor in HIV entry into target cells, and cancer metastasis. Thus, it is crucial for structure-based drug design efforts to have a reliable structure for the target. Traditional methods, such as x-ray crystallography and NMR, are just beginning to be applied successfully to membrane proteins, as there are several problems to overcome, such as expression, purification, solubility, crystallization, and isotopic labeling. The interdisciplinary strategy of molecular and computational biology provides information that is currently unavailable using these standard structural approaches and incorporates insight into the relationships between the target GPCR and early generation antagonists. We have effectively computationally modeled CXCR4 in the most realistic environment for a GPCR, a solvated lipid bilayer that mimics a cellular membrane. The simulations were validated by being entirely consistent with current experimental molecular biology data and provide fundamental insights into the receptor and its dynamic nature. Molecular simulations allow dynamic conformational movement, such as is involved in the signaling mechanism of CXCR4, and the methods employed here rationalized the different agonist and antagonist binding modes, providing a realistic platform for specific structure-based drug design. Previous studies (9, 11) have identified two classes of compounds that bind to CXCR4, represented in this work by AMD3100 and T140. The simulations presented here provide a rationalization of the experimental mapping and binding data. From these models it is possible to generate new hypotheses that can be experimentally tested for further validation. Ultimately the accuracy of these predictions will be determined when the high resolution structure has been established. Further insight into the mechanism will come from variants in which the active and inactive conformations are stabilized (10).
This is the first fully solvated lipid bilayer simulation of CXCR4 to be reported, which importantly, did not rely on National Supercomputing resources. Thus, it is feasible for individual investigators to study GPCR structure using this approach. As our understanding and the methodology of simulating membrane proteins with multiple domains increases, the reliability of these simulations will also increase, making this receptor superfamily more accessible to structure-based drug design efforts.
|
FOOTNOTES |
|---|
* This work was supported by National Institutes of Health Grant R01 AI41346 (to S. C. P.), the Georgia Cancer Coalition (to S. C. P.), United States Department of Defense Grant DAMD17-02-1-0446 (to J. O. T.), and the Philip Morris External Research Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|| To whom correspondence may be addressed: 323 Brown Cancer Center, University of Louisville, 529 S. Jackson St., Louisville, KY 40202. Tel.: 502-852-2194; Fax: 502-852-2195; E-mail: john.trent{at}louisville.edu. To whom correspondence may be addressed. E-mail: speiper{at}mail.mcg.edu.
1 The abbreviations used are: SDF-1, stromal cell-derived factor 1; HIV, human immunodeficiency virus; GPCR, G protein-coupled receptor; CHO, Chinese hamster ovary; TM, transmembrane; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine; N-ter, N-terminal domain.
2 www.AnorMED.com, July 8, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
- Tachibana, K., Hirota, S., Iizasa, H., Yoshida, H., Kawabata, K., Kataoka, Y., Kitamura, Y., Matsushima, K., Yoshida, N., Nishikawa, S., Kishimoto, T., and Nagasawa, T. (1998) Nature 393, 591-594[CrossRef][Medline] [Order article via Infotrieve]
- Zou, Y. R., Kottmann, A. H., Kuroda, M., Taniuchi, I., and Littman, D. R. (1998) Nature 393, 595-599[CrossRef][Medline] [Order article via Infotrieve]
- Knaut, H., Werz, C., Geisler, R., and Nusslein-Volhard, C. (2003) Nature 421, 279-282[CrossRef][Medline] [Order article via Infotrieve]
- Lapidot, T., and Petit, I. (2002) Exp. Hematol. 30, 973-981[CrossRef][Medline] [Order article via Infotrieve]
- Feng, Y., Broder, C. C., Kennedy, P. E., and Berger, E. A. (1996) Science 272, 872-877[Abstract]
- Muller, A., Homey, B., Soto, H., Ge, N., Catron, D., Buchanan, M. E., McClanahan, T., Murphy, E., Yuan, W., Wagner, S. N., Barrera, J. L., Mohar, A., Verastegui, E., and Zlotnik, A. (2001) Nature 410, 50-56[CrossRef][Medline] [Order article via Infotrieve]
- Hendrix, C. W., Flexner, C., MacFarland, R. T., Giandomenico, C., Fuchs, E. J., Redpath, E., Bridger, G., and Henson, G. W. (2000) Antimicrob. Agents Chemother. 44, 1667-1673
[Abstract/Free Full Text] - Doranz, B. J., Filion, L. G., Diaz-Mitoma, F., Sitar, D. S., Sahai, J., Baribaud, F., Orsini, M. J., Benovic, J. L., Cameron, W., and Doms, R. W. (2001) AIDS Res. Hum. Retroviruses 17, 475-486[CrossRef][Medline] [Order article via Infotrieve]
- Gerlach, L. O., Skerlj, R. T., Bridger, G. J., and Schwartz, T. W. (2001) J. Biol. Chem. 276, 14153-14160
[Abstract/Free Full Text] - Zhang, W. B., Navenot, J. M., Haribabu, B., Tamamura, H., Hiramatu, K., Omagari, A., Pei, G., Manfredi, J. P., Fujii, N., Broach, J. R., and Peiper, S. C. (2002) J. Biol. Chem. 277, 24515-24521
[Abstract/Free Full Text] - Tamamura, H., Xu, Y., Hattori, T., Zhang, X., Arakaki, R., Kanbara, K., Omagari, A., Otaka, A., Ibuka, T., Yamamoto, N., Nakashima, H., and Fujii, N. (1998) Biochem. Biophys. Res. Commun. 253, 877-882[CrossRef][Medline] [Order article via Infotrieve]
- Palczewski, K., Kumasaka, T., Hori, T., Behnke, C. A., Motoshima, H., Fox, B. A., Le Trong, I., Teller, D. C., Okada, T., Stenkamp, R. E., Yamamoto, M., and Miyano, M. (2000) Science 289, 739-745
[Abstract/Free Full Text] - Altenbach, C., Klein-Seetharaman, J., Cai, K., Khorana, H. G., and Hubbell, W. L. (2001) Biochemistry 40, 15493-15500[CrossRef][Medline] [Order article via Infotrieve]
- Kirkpatrick, P. (2003) Nat. Rev. Drug Dis. 2, 171
- Bissantz, C., Bernard, P., Hibert, M., and Rognan, D. (2003) Proteins 50, 5-25[CrossRef][Medline] [Order article via Infotrieve]
- Wang, Z. X., Berson, J. F., Zhang, T. Y., Cen, Y. H., Sun, Y., Sharron, M., Lu, Z. H., and Peiper, S. C. (1998) J. Biol. Chem. 273, 15007-15015
[Abstract/Free Full Text] - Navenot, J.-M., Wang, Z.-X., Trent, J. O., Murray, J., Hu, Q.-X., DeLeeuw, L., Moore, P., Chang, Y., and Peiper, S. C. (2001) J. Mol. Biol. 313, 1181-1193[CrossRef][Medline] [Order article via Infotrieve]
- Staudinger, R., and Bandres, J. C. (2000) Biochem. Biophys. Res. Commun. 274, 153-156[CrossRef][Medline] [Order article via Infotrieve]
- Zhou, N., Luo, Z., Luo, J., Liu, D., Hall, J. W., Pomerantz, R. J., and Huang, Z. (2001) J. Biol. Chem. 276, 42826-42833
[Abstract/Free Full Text] - Huang, X., Shen, J., Cui, M., Shen, L., Luo, X., Ling, K., Pei, G., Jiang, H., and Chen, K. (2003) Biophys. J. 84, 171-184[Medline] [Order article via Infotrieve]
- Vaidehi, N., Floriano, W. B., Trabanino, R., Hall, S. E., Freddolino, P., Choi, E. J., Zamanakos, G., and Goddard, W. A., III (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 12622-12627
[Abstract/Free Full Text] - Cornell, W. D., Cieplak, P., Bayly, C. I., Gould, I. R., Merz, K. M., Ferguson, D. M., Spellmeyer, D. C., Fox, T., Caldwell, J. W., and Kollman, P. A. (1995) J. Am. Chem. Soc. 117, 5179-5197[CrossRef]
- Doranz, B. J., Orsini, M. J., Turner, J. D., Hoffman, T. L., Berson, J. F., Hoxie, J. A., Peiper, S. C., Brass, L. F., and Doms, R. W. (1999) J. Virol. 73, 2752-2761
[Abstract/Free Full Text] - Brelot, A., Heveker, N., Montes, M., and Alizon, M. (2000) J. Biol. Chem. 275, 23736-23744
[Abstract/Free Full Text] - Tamamura, H., Omagari, A., Oishi, S., Kanamoto, T., Yamamoto, N., Peiper, S. C., Nakashima, H., Otaka, A., and Fujii, N. (2000) Bioorg. Med. Chem. Lett. 10, 2633-2637[CrossRef][Medline] [Order article via Infotrieve]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. S. Y. Wong, V. Bodart, M. Metz, J. Labrecque, G. Bridger, and S. P. Fricker Comparison of the Potential Multiple Binding Modes of Bicyclam, Monocylam, and Noncyclam Small-Molecule CXC Chemokine Receptor 4 Inhibitors Mol. Pharmacol., December 1, 2008; 74(6): 1485 - 1495. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Lin, A. Bertolotti-Ciarlet, B. Haggarty, J. Romano, K. M. Nolan, G. J. Leslie, A. P.-O. Jordan, C.-c. Huang, P. D. Kwong, R. W. Doms, et al. Replication-Competent Variants of Human Immunodeficiency Virus Type 2 Lacking the V3 Loop Exhibit Resistance to Chemokine Receptor Antagonists J. Virol., September 15, 2007; 81(18): 9956 - 9966. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Abraham, K. Biyder, M. Begin, H. Wald, I. D. Weiss, E. Galun, A. Nagler, and A. Peled Enhanced Unique Pattern of Hematopoietic Cell Mobilization Induced by the CXCR4 Antagonist 4F-Benzoyl-TN14003 Stem Cells, September 1, 2007; 25(9): 2158 - 2166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Basu, V. R. Jala, S. Mathis, S. T. Rajagopal, A. Del Prete, P. Maturu, J. O. Trent, and B. Haribabu Critical Role for Polar Residues in Coupling Leukotriene B4 Binding to Signal Transduction in BLT1 J. Biol. Chem., March 30, 2007; 282(13): 10005 - 10017. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Yang, E. Jackson, B. M. Woerner, A. Perry, D. Piwnica-Worms, and J. B. Rubin Blocking CXCR4-Mediated Cyclic AMP Suppression Inhibits Brain Tumor Growth In vivo Cancer Res., January 15, 2007; 67(2): 651 - 658. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Burger and T. J. Kipps CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment Blood, March 1, 2006; 107(5): 1761 - 1767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. Mkrtchyan, R. M. Markosyan, M. T. Eadon, J. P. Moore, G. B. Melikyan, and F. S. Cohen Ternary Complex Formation of Human Immunodeficiency Virus Type 1 Env, CD4, and Chemokine Receptor Captured as an Intermediate of Membrane Fusion J. Virol., September 1, 2005; 79(17): 11161 - 11169. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Burger, T. Hartmann, M. Krome, J. Rawluk, H. Tamamura, N. Fujii, T. J. Kipps, and J. A. Burger Small peptide inhibitors of the CXCR4 chemokine receptor (CD184) antagonize the activation, migration, and antiapoptotic responses of CXCL12 in chronic lymphocytic leukemia B cells Blood, September 1, 2005; 106(5): 1824 - 1830. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Percherancier, Y. A. Berchiche, I. Slight, R. Volkmer-Engert, H. Tamamura, N. Fujii, M. Bouvier, and N. Heveker Bioluminescence Resonance Energy Transfer Reveals Ligand-induced Conformational Changes in CXCR4 Homo- and Heterodimers J. Biol. Chem., March 18, 2005; 280(11): 9895 - 9903. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Galliera, V. R. Jala, J. O. Trent, R. Bonecchi, P. Signorelli, R. J. Lefkowitz, A. Mantovani, M. Locati, and B. Haribabu {beta}-Arrestin-dependent Constitutive Internalization of the Human Chemokine Decoy Receptor D6 J. Biol. Chem., June 11, 2004; 279(24): 25590 - 25597. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |