Homology model of the CB1 cannabinoid receptor: sites critical for nonclassical cannabinoid agonist interaction

Association of cannabimimetic compounds such as cannabinoids, aminoalkylindoles (AAIs), and arachidonylethanolamide (anandamide) with the brain cannabinoid (CB(1)) receptor activates G-proteins and relays signals to regulate neuronal functions. A CB(1) receptor homology model was constructed using the published x-ray crystal structure of bovine rhodopsin (Palczewski et al., Science, 2000, Vol. 289, pp. 739-745) in the conformation most likely to represent the "high-affinity" state for agonist binding to G-protein coupled receptors (GPCRs). A molecular docking approach that combined Monte Carlo and molecular dynamics simulations was used to identify the putative binding conformations of nonclassical cannabinoid agonists, including AC-bicyclic CP47497 and CP55940, and ACD-tricyclic CP55244. Placement of these ligands was based upon the assumption of a critical hydrogen bond between the A-ring OH and the side chain N of Lys192 in transmembrane helix 3. We evaluated two alternative binding conformations, C3-in and C3-out, denoting the directionality of the ligand C3 side chain within the receptor with respect to the inside or the outside of the cell. Assuming both the C3-in or C3-out conformation, the calculated ligand-receptor binding energy (DeltaE(bind)) was correlated with the experimentally observed binding affinity (K(i)) for a series of nonclassical cannabinoid agonists. The C3-in conformation was marginally better than the alternative C3-out conformation in predicting the rank order of the tested nonclassical cannabinoid analogs. Adopting the C3-in conformation due to the greater number of receptor interactions with known pharmacophoric elements of the ligand, key residues were identified comprising the presumed hydrophobic pocket that interacts with the C3 side chain of cannabinoid agonists. Key hydrogen bonds would form between both K3.28(192) and E(258) and the A-ring OH, and between Q(261) and the C-ring C-12 hydroxypropyl. In summary, the present study represents one of the first attempts to construct a homology model of the CB(1) cannabinoid receptor based upon the published bovine rhodopsin x-ray crystal structure and to elucidate the putative ligand binding site for nonclassical cannabinoid agonists. We postulated sites of the CB(1) receptor critical for the ligand interaction, including the hydrophobic pocket interacting with the key pharmacophoric moiety, the C3 side chain. More work is needed to delineate between two alternative (and possibly other) binding conformations of the nonclassical cannabinoid ligands within the CB(1) receptor. The present study provides a consistent framework for further investigation of the CB(1) receptor-ligand interaction and for the study of CB(1) receptor activation.     

Steroid structure and function:I. Conformational transmission in 17alpha-acetoxy progesterone

The molecular conformation of 17alpha-acetoxy progesterone has been determined crystallographically and is compared with that of progesterone. The 17alpha-acetate substituent restricts the flexibility of the progesterone side chain, strains bond lengths in the C- and D-rings, and has long range effects on the A-ring conformation. The A-ring adopts a perfect sofa conformation similar to that observed in one conformational isomer of progesterone. Consequently this progesterone isomer is proposed to be that best suited to binding in the rabbit and human uterus.     

Steroid structure and function V. A-ring conformation in 17-hydroxy-6 alpha-methylprogesterone

The molecular conformation of 17-hydroxy-6 alpha-methylprogesterone has been determined crystallographically and is compared with 17-hydroxy-progesterone, 17-acetoxyprogesterone and 17-acetoxy-6 alpha-methylprogesterone (MPA). The analysis demonstrates that the 6 alpha-methyl substituent is not sufficient by itself to induce inversion of the A-ring. Consequently, the inverted form observed in MPA and proposed to be responsible for high affinity binding to the progesterone receptor appears to be induced by the combined long range influence of 17 alpha-acetoxy substituent and the direct interaction of the 6 alpha-methyl group with the flexible A-ring.     

Model of three-dimensional structure of vitamin D receptor and its binding mechanism with 1alpha,25-dihydroxyvitamin D(3).

Comparative modeling of the vitamin D receptor three-dimensional structure and computational docking of 1alpha,25-dihydroxyvitamin D(3) into the putative binding pocket of the two deletion mutant receptors: (207-423) and (120-422, Delta [164-207]) are reported and evaluated in the context of extensive mutagenic analysis and crystal structure of holo hVDR deletion protein published recently. The obtained molecular model agrees well with the experimentally determined structure. Six different conformers of 1alpha,25-dihydroxyvitamin D(3) were used to study flexible docking to the receptor. On the basis of values of conformational energy of various complexes and their consistency with functional activity, it appears that 1alpha,25-dihydroxyvitamin D(3) binds the receptor in its 6-s-trans form. The two lowest energy complexes obtained from docking the hormone into the deletion protein (207-423) differ in conformation of ring A and orientation of the ligand molecule in the VDR pocket. 1alpha,25-Dihydroxyvitamin D(3) possessing the A-ring conformation with axially oriented 1alpha-hydroxy group binds receptor with its 25-hydroxy substituent oriented toward the center of the receptor cavity, whereas ligand possessing equatorial conformation of 1alpha-hydroxy enters the pocket with A ring directed inward. The latter conformation and orientation of the ligand is consistent with the crystal structure of hVDR deletion mutant (118-425, Delta [165-215]). The lattice model of rVDR (120-422, Delta [164-207]) shows excellent agreement with the crystal structure of the hVDR mutant. The complex obtained from docking the hormone into the receptor has lower energy than complexes for which homology modeling was used. Thus, a simple model of vitamin D receptor with the first two helices deleted can be potentially useful for designing a general structure of ligand, whereas the advanced lattice model is suitable for examining binding sites in the pocket.     

From ligand structure to biological activity: modified estratrienes and their estrogenic and antiestrogenic effects in MCF-7 cells

A variety of compounds, including the selective estrogen receptor (ER) modulators tamoxifen and raloxifene, phytoestrogens such as genistein, and xenoestrogens such as bisphenol, bind to the estrogen receptor and elicit biological responses. Structural studies have linked the altered activity of compounds such as 4-hydroxytamoxifen, raloxifene, genistein, and tetrahydrochrysene, which have substantially different structures from estradiol (E2), to differences in the positioning of the critical "helix 12" within the ligand-binding domain (LBD) of the ER-ligand complex. However, subtle permutations of the E2 molecule would also be expected to modulate the pattern of responses within a cell. Forty-two ligands were constructed by the addition or relocation of double bonds, hydroxyl, keto, amino, and nitro substituents throughout the estra-l,3,5(10)-triene (estratriene) ring system. In this review, we summarize the effects of subtle changes in the estratriene molecule on the ability of the receptor complex to stimulate the growth of MCF-7 cells, or affect the expression of four estrogen-regulated genes (progesterone receptor, pS2 protein, cathepsin D, and tissue plasminogen activator), as well as undergo nuclear processing and downregulate ERalpha mRNA. The affinity of these ligands for, and mechanism of their binding with, the ERalpha have been measured, along with their effect on the conformation of the ER-ERE complex. In particular, two A-ring isomers of E2, 2- and 4-hydroxyestratriene-17beta-ol, display gene selective activity within MCF-7 cells which is dependent on complex endogenous promoters, an intact AF-2 and is sensitive to the level of SRC-1. Both of these A-ring isomers function as antiestrogens. Molecular modeling of these two A-ring isomers complexed with the ER ligand-binding domain supports the idea that the conformation of the LBD is affected by subtle changes in the estratriene structure.     

An analog of 1alpha,25-dihydroxy-19-norvitamin D3 with the 1alpha-hydroxy group fixed in the axial position lacks biological activity in vitro

The relationship between the A-ring chair conformation of vitamin D compounds and their ability to bind the vitamin D receptor (VDR) has long attracted the attention of many researchers. It was established that in the crystalline complexes of hVDRmt with the natural hormone, 1alpha,25-dihydroxyvitamin D(3) (1), and its side-chain analogs the vitamins exist in beta-chair form with an equatorial orientation of 1alpha-OH. However, with all these ligands the interconversion between both A-ring forms would be possible in solution. In an attempt to verify the conformation of vitamin D compounds required for binding the VDR we prepared analog 4, characterized by the presence of an axial 1alpha-hydroxy group. Since the additional ring connecting 3beta-oxygen and C-2 prevents A-ring conformational flexibility, the synthesized vitamin 4 can exist exclusively in the alpha-chair form. The geometrical isomer 5 with a free 3beta-OH group was also obtained. The analog 5 binds very poorly to VDR, whereas the vitamin 4 is practically devoid of binding ability. Both compounds also show very low HL-60-differentiating activity. When tested in vivo in mice the analogs 4 and 5 exhibit significant calcemic responses with analog 4 showing more activity than analog 5.     

Structural requirements of para-alkylphenols to bind to estrogen receptor.

Octyl- and nonylphenols in the environment have been proposed to function as estrogens. To gain insight into their structural essentials in binding to the estrogen receptor, a series of phenols with saturated alkyl groups at the para position, HO-C6H4-CnH2n+1 (n = 0-12), were examined for their ability to displace [3H]17beta-estradiol in the recombinant human estrogen receptor, which was expressed in Sf9 cells using the vaculovirus expression system. All tested para-alkylphenols were found to bind fully to the estrogen receptors in a dose-dependent manner. The interaction of alkylphenols with the receptor became stronger with increase in the number of the alkyl carbons and the activity was maximized with n = 9 of nonylphenol. Phenol (n = 0) exhibited weak but full binding to the receptor, whereas anisole with a protected phenolic hydroxyl group was completely inactive. Also, alkanes such as n-octane, 2,2, 4-trimethylpentane corresponding to tert-octane, and n-nonane exhibited no binding. The results indicate that the binding of para-alkylphenols to the estrogen receptor is due to the effect of covalent bonding of two constituents of the phenol and alkyl groups, which correspond to the A-ring and hydrophobic moiety of the steroid structure, respectively. When alkylphenols were examined for their receptor binding conformation by 1H-NMR measurements and ab initio molecular orbital calculations, it was suggested that nonbranched alkyl groups are in an extended conformation, while branched alkyl groups are in a folded conformation. These results suggest that branched and nonbranched alkyl moieties of alkylphenols interact differently with the lipophilic ligand binding cavity of the estrogen receptor when compared to the binding of 17beta-estradiol.     

Characterization of the estrogenic activities of zearalenone and zeranol in vivo and in vitro

In the present study, we compared the estrogenic activity of zearalenone (ZEN) and zeranol (ZOL) by determining their relative receptor binding affinities for human ERalpha and ERbeta and also by determining their uterotropic activity in ovariectomized female mice. ZOL displayed a much higher binding affinity for human ERalpha and ERbeta than ZEN did. The IC(50) values of ZEN and ZOL for binding to human ERalpha were 240.4 and 21.79nM, respectively, and the IC(50) values for binding to ERbeta were 165.7 and 42.76nM, respectively. In ovariectomized female ICR mice, s.c. administration of ZEN at doses >or=2mg/kg/day for 3 consecutive days significantly increased uterine wet weight compared with the control group, and administration of ZOL increased the uterine wet weight at lower doses (>or=0.5mg/kg/day for 3 days). Based on available X-ray crystal structures of human ERalpha and ERbeta, we have also conducted molecular modeling studies to probe the binding characteristics of ZEN and ZOL for human ERalpha and ERbeta. Our data revealed that ZEN and ZOL were able to occupy the active site of the human ERalpha and ERbeta in a strikingly similar manner as 17beta-estradiol, such that the phenolic rings of ZEN and ZOL occupied the same receptor region as occupied by the A-ring of 17beta-estradiol. The primary reason that ZOL and ZEN is less potent than 17beta-estradiol is likely because 17beta-estradiol could bind to the receptor pocket without significantly changing its conformation, while ZOL or ZEN would require considerable conformational alterations upon binding to the estrogen receptors (ERs).     

Ab initio electronic structure of the progestogen norethisterone and its 5 alpha-derivatives

The steroid 17 alpha-ethynyl-19-nor-4-androsten-17 beta-ol, 3-one (Norethisterone; NET) and its 5 alpha-dihydro (5 alpha-NET), 3 alpha- and 3 beta-tetrahydro derivatives (3 alpha,5 alpha- and 3 beta,5 alpha-NET), were comparatively studied by the ab initio quantum mechanics theory. Additionally, 5 alpha-androstan-3 beta,17 beta-diol (ADIOL) was also studied. The Hartree-Fock method and the 6-31G(*) basis set were used to obtain the lowest energy conformation, geometries, electronic structure and physicochemical properties of the steroids. The results showed bond distances and valence angles similar among all steroids, but some differences in dihedral angles in the A-B-ring system were observed. The electronic structure analysis showed that NET has both frontier orbitals that is, the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) located at the C4-C5 pi-bond. In A-ring reduced derivatives, the HOMO was found at the 17 beta-OH and ethynyl groups. In the case of 5 alpha-NET, the LUMO was confined to the A-ring and its C3 carbonyl group while the two NET tetrahydro-reduced derivatives showed the LUMO at the 17 beta-OH and ethynyl groups. The energy changes of the rotational barrier of the 17 beta-OH group suggest that its movement is somewhat restricted by the 17 alpha-ethynyl group. Interestingly both groups at C17 form a single electrostatic potential with high electronic density. On the other side, the 19-nor condition increases the A-ring mobility. However, the 3 beta-OH group of 3 beta,5 alpha-NET may rotate without significant energy differences as compared to the same group in ADIOL. The electronic structure of NET and its A-ring reduced derivatives explains in some extent their interaction with androgen and progesterone receptors as well as their selectivity for the estrogen alpha-receptor.     

The hydroxylamine reaction of sensory rhodopsin II: light-induced conformational alterations with C13=C14 nonisomerizable pigment

Sensory rhodopsin II, a repellent phototaxis receptor from Natronomonas (Natronobacterium) pharaonis (NpSRII), forms a complex with its cognate transducer (NpHtrII). In micelles the two proteins form a 1:1 heterodimer, whereas in membranes they assemble to a 2:2 complex. Similarly to other retinal proteins, sensory rhodopsin II undergoes a bleaching reaction with hydroxylamine in the dark which is markedly catalyzed by light. The reaction involves cleavage of the protonated Schiff base bond which covalently connects the retinal chromophore to the protein. The light acceleration reflects protein conformation alterations, at least in the retinal binding site, and thus allows for detection of these changes in various conditions. In this work we have followed the hydroxylamine reaction at different temperatures with and without the cognate transducer. We have found that light irradiation reduces the activation energy of the hydroxylamine reaction as well as the frequency factor. A similar effect was found previously for bacteriorhodopsin. The interaction with the transducer altered the light effect both in detergent and membranes. The transducer interaction decreased the apparent light effect on the energy of activation and the frequency factor in detergent but increased it in membranes. In addition, we have employed an artificial pigment derived from a retinal analog in which the critical C13=C14 double bond is locked by a rigid ring structure preventing its isomerization. We have observed light enhancement of the reaction rate and reduction of the energy of activation as well as the frequency factor, despite the fact that this pigment does not experience C13=C14 double bond isomerization. It is suggested that retinal excited state polarization caused by light absorption of the "locked" pigment polarizes the protein and triggers relatively long-lived protein conformational alterations.     

Discovery of octahydropyrrolo [3,2-b] pyridin derivative as a highly selective Type I inhibitor of FGFR3 over VEGFR2 by high-throughput virtual screening

Although the aberrant activity of fibroblast growth factor receptor 3 (FGFR3) is implicated in various cancers, the reported kinase inhibitors of FGFR3 tend to cause side effects resulting from the inhibitory activity on vascular endothelial growth factor receptor 2 (VEGFR2). Therefore, it is necessary to find a novel high-selective inhibitor of FGFR3 over VEGFR2 from the small-molecule compound database. In this study, integrated virtual screening protocols were established to screen for selective inhibitors of FGFR3 over VEGFR2 in Drugbank and Asinex databases by combining three-dimensional pharmacophore model, molecular docking, molecular dynamics (MD) simulation, and molecular mechanics Poisson–Boltzmann surface area (MMPBSA) calculations. Finally, it is found that Asinex-5082, as an octahydropyrrolo[3,2-b] pyridin derivative, has larger binding free energy with FGFR3 (−39.3 kcal/mol) than reference drug Erdafitinib (−29.9 kcal/mol), while cannot bind with VEGFR2, resulting in considerable inhibitory selectivity. This is because Asinex-5082, unlike Erdafitinib, has not m-dimethoxybenzene with large steric hindrance, thus can enter the larger ATP-binding pocket of FGFR3 with DFG-in conformation to form hydrophobic interaction with residues Met529, Ile539, and Tyr557 as well as hydrogen bond with Ala558. On the other hand, due to the fact that the benzodioxane and N-heterocyclic rings are connected by carbonyl (C=O), Asinex-5082 cannot rotate freely so as to enter the smaller ATP binding pocket of VEGFR2 on the DFG-out conformation. The lead molecule Asinex-5082 may facilitate the rational design and development of novel selective inhibitors of FGFR3 over VEGFR2 as anticancer drugs.     

Structure-activity studies of non-steroidal aromatase inhibitors: the crystal and molecular structures of CGS 16949A and CGS 18320B.

The crystal and molecular structures of 4-(5,6,7,8-tetrahydroimidazo[1,5-a]pyridin-5-yl)benzonitrile hydrochloride (CGS 16949A) and bis(p-cyanophenyl)imidazo-1-yl methane hemisuccinate (CGS 18320B) have been determined as part of structure-activity relationship studies of non-steroidal aromatase inhibitors. CGS 18320B crystallizes with two inhibitor molecules in the asymmetric unit that are similar in conformation. The cyanophenyl groups and the imidazole moieties in the CGS 18320B molecules display a propellor-like arrangement. The orientation of the imidazole ring in CGS 16949A, which is constrained by the piperidine ring, differs by about 80 degrees from the orientations in both CGS 18320B molecules. The conformations of both compounds are consistent with the proposed model (Banting et al. (1988) J. Enz. Inhibit., 2, 216) for inhibitor binding by positioning of the cyanophenyl group in the steroid A-ring binding site and interaction of the imidazole nitrogen with the iron of the haem.     

Deconstructing estradiol: Removal of B-ring generates compounds which are potent and subtype-selective estrogen receptor agonists

Estradiol and related estrogens have been widely used as supplements to relieve menopausal symptoms, but they lead to an increased risk of breast and endometrial cancer. Here we report the synthesis of a new family of compounds where we have removed the B-ring from the steroid ABCD structure, and functionalized the A-ring. These A-CD compounds show a preferential affinity for the estrogen receptor subtype ERβ. Some show binding affinities which are greater than estradiol. The presence of electron-withdrawing substituents on the A-ring should reduce the tendency of these compounds to form carcinogenic metabolites, so they might lead to a safer approach to hormone replacement therapy.     

Biological activity and ligand binding mode to the progesterone receptor of A-homo analogues of progesterone

The biological activity of two seven-membered A-ring (A-homo) analogues of progesterone was evaluated by transactivation assays in Cos-1 cells and by determination of Bcl-x(L) expression levels in T47D cells. The results show that both compounds act as selective progesterone receptor (PR) agonists but lack mineralocorticoid receptor (MR) activity. Molecular modelling using semiempirical AM1 and ab initio HF/6-31G** calculations, showed that the A-ring of the A-homo steroids may adopt five different conformations, although only three correspond to low energy conformers. The low energy conformers of each analogue were introduced into the ligand binding pocket of the PR ligand binding domain (LBD) obtained from the PR LBD-progesterone crystal structure. The steroid binding mode was then analyzed using 10 ns of molecular dynamics (MD) simulation. The PR LBD-progesterone complex was also simulated as a control system. The MD results showed that both A-homo steroids have one conformer that may be properly recognized by the PR, in agreement with the observed progestagen activity. Moreover, the simulation revealed the importance of a water molecule in the formation of a hydrogen bonding network among specific receptor residues and the steroid A-ring carbonyl.     

Ligand-selective interdomain conformations of estrogen receptor-alpha

Selective estrogen receptor modulators (SERMs) inhibit estrogen activation of the estrogen receptor (ER) in some tissues but activate ER in other tissues. These tissue-selective actions suggest that SERMs may be identified with tissue specificities that would improve the safety of breast cancer and hormone replacement therapies. The identification of an improved SERM would be aided by understanding the effects of each SERM on the structure and interactions of ER. To date, the inability to obtain structures of the full-length ER has limited our structural characterization of SERM action to their antiestrogenic effects on the isolated ER ligand binding domain. We studied the effects of estradiol and the clinically useful SERMs 4-hydroxytamoxifen and fulvestrant on the conformation of the full-length ERalpha dimer complex by comparing, in living human breast cancer cells, the amounts of energy transfer between fluorophores attached to different domains of ERalpha. Estradiol, 4-hydroxytamoxifen, and fulvestrant all promoted the rapid formation of ERalpha dimers with equivalent interaction kinetics. The amino- and carboxyl-terminal ERalpha domains both contain activation functions differentially affected by these ligands, but the positions of only the carboxyl termini differed upon binding with estradiol, 4-hydroxytamoxifen, or fulvestrant. The association of a specific ERalpha dimer conformation with the binding of ligands of different clinical effect will assist the identification of a SERM with optimal tissue-selective estrogenic and antiestrogenic activities. These studies also provide a roadmap for dissecting important structural and kinetic details for any protein complex from the quantitative analysis of energy transfer.     

Many-atom van der Waals interactions lead to direction-sensitive interactions of covalent bonds

Strict physical theory and numerical calculations show that a specific coupling of many-atom van der Waals interactions with covalent bonding can significantly (half as much) increase the strength of attractive dispersion interactions when the direction of interaction coincides with the direction of the covalent bond, and decrease this strength when the direction of interaction is perpendicular to the direction of the covalent bond. The energy effect is comparable to that caused by the replacement of atoms (e.g. N by C or O) in conventional pairwise van der Waals interactions. Analysis of protein structures shows that they bear an imprint of this effect. This means that many-atom van der Waals interactions cannot be ignored in refinement of protein structures, in simulations of their folding, and in prediction of their binding affinities.     

Binding, X-ray and NMR studies of the three A-ring isomers of natural estradiol

The effect of the position of the phenolic hydroxyl on the conformations of the three A-ring isomers of estradiol, namely, estra-1,3,5(10)-trien-1,17 beta-diol (10), estra-1,3,5(10)-trien-2,17 beta-diol (3), and estra-1,3,5(10)-trien-4,17 beta-diol (6), has been analyzed by X-ray crystallography. The results of these analyses were correlated with the absorptions of the angular methyl groups in the [1H]NMR spectra of these isomers and natural estradiol (E2). It was observed that the changes in chemical shift of protons at C18 corresponded to skeletal modifications in the steroid structure which changed the anisotropic effect of the hydroxyl group at C17. Examination of the affinity of these A-ring isomers of E2 for the estrogen receptor has shown the 2-hydroxylated isomer 3 to retain 1/5th the affinity of E2 for its binding protein. The 1- and 4-hydroxylated derivatives (10 and 6, respectively) bound to a much lesser extent. The receptor affinities of these estrogen analogues may be related to the angle between the 18-methyl and the 17 beta-hydroxyl groups (or the dihedral angle between the planar A-ring and the angular C18 methyl) as well as the position of the A-ring hydroxyl group.