CoMFA methodology in structure-activity analysis of hexahydro- and octahydropyrido[1,2-c]pyrimidine derivatives based on affinity towards 5-HT1A, 5-HT2A and alpha1-adrenergic receptors

Structural features of the pyrido[1,2-c]pyrimidine derivatives with arylpiperazine moiety and their affinities towards 5-HT1A, 5-HT2A and alpha1-adrenergic receptors were analyzed using the CoMFA procedure. On the basis of 3D-QSAR models for the 5-HT2A and alpha1-adrenergic receptors, four compounds with expected better affinity/selectivity were proposed and synthesized. The affinities obtained confirm experimentally the usefulness of CoMFA models. Our results suggest that active conformations adopted by the studied molecules when interacting with the receptors are neutral instead of the protonated ones.     

2- and 8-alkynyladenosines: conformational studies and docking to human adenosine A3 receptor can explain their different biological behavior

Adenosine (Ado) derivatives substituted at the C2 position with an alkynyl chain are endowed with high affinity for A(1), A(2A) and A(3) human adenosine receptors, while being less active at the low affinity A(2B) subtype. On the other hand, the introduction of an alkynyl chain at the C8 position of adenosine is detrimental for the affinity and potency at A(1), A(2A), and A(2B) receptors, while is more tolerated by the A(3) receptor. The evaluation of the stimulation of [35S]GTPgammaS binding revealed that 2-alkynyladenosines behave as adenosine receptors agonists while, on the contrary, 8-alkynyladenosines behave as antagonists.With this work we demonstrated, by means of an NMR-based and a computational conformational analysis, that 8-alkynyladenosines, differently from 2-alkynyladenosines, cannot adopt the sugar-base anti conformation required for adenosine receptor activation.Furthermore, using the recently reported X-ray crystal structure of bovine rhodopsin as template, we built a 3D model of the seven transmembrane domains of the human adenosine A(3) receptor with the homology modeling. After identification of the binding site we carried out docking experiments, demonstrating that the two class of molecules have different binding modes that explain their different degree of affinity and the shift of their activity from agonism to antagonism.     

Structural basis for ligand recognition at the benzodiazepine binding site of GABAA alpha 3 receptor, and pharmacophore-based virtual screening approach

Given the heterogeneity of GABA(A) receptor, the pharmacological significance of identifying subtype selective modulators is increasingly being recognized. Thus, drugs selective for GABA(A) alpha(3) receptors are expected to display fewer side effects than the drugs presently in clinical use. Hence we carried out 3D QSAR (three-dimensional quantitative structure-activity relationship) studies on a series of novel GABA(A) alpha(3) subtype selective modulators to gain more insight into subtype affinity. To identify the 3D functional attributes required for subtype selectivity, a chemical feature-based pharmacophore, primarily based on selective ligands representing diverse structural classes was generated. The obtained pseudo receptor model of the benzodiazepine binding site revealed a binding mode akin to "Message-Address" concept. Scaffold hopping was carried out across multi-conformational May Bridge database for the identification of novel chemotypes. Further a focused data reduction approach was employed to choose a subset of enriched compounds based on "Drug likeness" and "Similarity-based" methods. These results taken together could provide impetus for rational design and optimization of more selective and high affinity leads with a potential to have decreased adverse effects.     

Alpha(1) adrenoceptor subtype selectivity. 3D-QSAR models for a new class of alpha(1) adrenoceptor antagonists derived from the novel antipsychotic sertindole

Receptor-binding affinities for the alpha(1) adrenoceptor subtypes alpha(1a), alpha(1b) and alpha(1d) for a series of 39 alpha(1) adrenoceptor antagonists derived from the antipsychotic sertindole are reported. The SAR of the compounds with respect to affinity for the alpha(1a), alpha(1b) and alpha(1d) adrenoceptor subtypes as well as affinity obtained by an alpha(1) assay (rat brain membranes) were investigated using a 3D-QSAR approach based on the GRID/GOLPE methodology. Good statistics (r(2)=0.91-0.96; q(2)=0.65-0.73) were obtained with the combination of the water (OH2) and methyl (C3) probes. The combination of steric repulsion and electrostatic attractions explain the affinities of the included molecules. The adrenergic alpha(1a) receptor seems to be more tolerant to large substituents in the area between the indole 5- and 6-positions compared to the adrenergic alpha(1b) and alpha(1d) receptor subtypes. There seems to be minor differences in the position of areas in the alpha(1b) receptor compared to alpha(1a) and alpha(1d) receptors where electrostatic interaction between the molecules and the receptor (OH2 probe) contribute to increased affinity. These observations may be used in the design of new subtype selective compounds. In addition, the model based on biological data from an alpha(1) assay (rat brain membranes) resembles the model for the alpha(1b) adrenoceptor subtype.     

Molecular recognition of docosahexaenoic acid by peroxisome proliferator-activated receptors and retinoid-X receptor alpha

The family of peroxisome proliferator-activated receptors (PPARs) is the molecular target of synthetic antidiabetic and hypolipidemic drugs. The side effects of these drugs are limiting their use in patients with high lipid levels. Natural compounds, like Docosahexaenoic acid (DHA) from fish oil, have beneficial effects in the treatment of metabolic diseases, and several DHA derivatives are known to activate PPAR genes. Experimental studies on affinities of DHA and its derivatives for PPARs are not available. In the present study we are therefore using computational docking, molecular dynamics simulation, and several scoring programs to predict affinities and binding modes of DHA for PPARs and retinoid-X receptor alpha, which is the DNA binding partner of PPARs. The calculations indicated that DHA binds to PPARs and the retinoid-X receptor alpha with high affinity, and that different PPARs exhibited different structural effects on the first four carbons atoms of DHA. Our data indicate that the beneficial health effects of DHA may be obtained by high affinity binding to the PPARs.     

Identification of novel allosteric modulator binding sites in NMDA receptors: A molecular modeling study

The dysfunction of N-methyl-d-Aspartate receptors (NMDARs), a subtype of glutamate receptors, is correlated with schizophrenia, stroke, and many other neuropathological disorders. However, not all NMDAR subtypes equally contribute towards these disorders. Since NMDARs composed of different GluN2 subunits (GluN2A-D) confer varied physiological properties and have different distributions in the brain, pharmacological agents that target NMDARs with specific GluN2 subunits have significant potential for therapeutic applications. In our previous research, we have identified a family of novel allosteric modulators that differentially potentiate and/or inhibit NMDARs of differing GluN2 subunit composition. To further elucidate their molecular mechanisms, in the present study, we have identified four potential binding sites for novel allosteric modulators by performing molecular modeling, docking, and in silico mutations. The molecular determinants of the modulator binding sites (MBS), analysis of particular MBS electrostatics, and the specific loss or gain of binding after mutations have revealed modulators that have strong potential affinities for specific MBS on given subunits and the role of key amino acids in either promoting or obstructing modulator binding. These findings will help design higher affinity GluN2 subunit-selective pharmaceuticals, which are currently unavailable to treat psychiatric and neurological disorders.     

Comparative modeling of a GABAA alpha1 receptor using three crystal structures as templates

We built a model of a GABAA alpha1 receptor (GABAAR) that combines the ligand binding (LBD) and the transmembrane domains (TMD). We used six steps: (1) a four-alpha helical bundle in the crystal structure of bovine cytochrome c oxidase (2OCC) was identified as a template for the TMD of a single subunit. (2) The five pore-forming alpha helices of a bacterial mechanosensitive channel (1MSL) served as a template for the pentameric ion channel. (3) Five copies of the tetrameric template from 2OCC were superimposed on 1MSL to produce a homopentamer containing 20 alpha helices arranged around a funnel-shaped central pore. (4) Five copies of the GABAAR sequence were threaded onto the alpha-helical segments of this template and inter-helical loops were generated to produce the TMD model. (5) A model of the LBD was built by threading the aligned sequence of GABAAR onto the crystal structure of the acetylcholine binding protein (1I9B). (6) The models of the LBD and the TMD were aligned along a common five-fold axis, moved together along that axis until in vdW contact, merged, and then optimized with restrained molecular dynamics. Our model corresponds closely with recently published coordinates of the acetylcholine receptor (1OED) but also explains additional features. Our model reveals structures of loops that were not visible in the cryoelectron micrograph and satisfies most labeling and mutagenesis data. It also suggests mechanisms for ligand binding transduction, ion selectivity, and anesthetic binding.     

Studying the catechol binding cavity in comparative models of human dopamine D2 receptor

Obtaining more structural information of human dopamine D(2) receptor may help in the design of better therapeutic agents against diseases such as Parkinson. In this study attempts have been made to develop a functional model for the catechol binding site of the human dopamine D(2) receptor, with two primary models being postulated based on the presence of a disulfide bridge in the second extracellular loop. The models have been subjected to subsequent molecular dynamics simulation and receptor based virtual screening of catechol structures. During steady state of the simulations, representative models with the reduced disulfide bridge were more capable of discriminating between active and inactive catechol structures. It is postulated that similar conformational changes of the second extracellular loop observed in 5-HT4 and β-adrenergic receptors, might also take place in the human D(2) receptor during its interaction with agonist ligands.     

Molecular modelling of the GABAA ion channel protein

The GABAA ion channel protein is central to the mechanism of action of general anaesthetics and thus to the phenomenon of human consciousness. A molecular model of the alpha1beta2gamma2 gamma-aminobutyric acid type-A (GABAA) ligand-gated ion channel protein has been constructed. The cryo-electron microscopy structure of the nicotinic acetylcholine receptor (nAChR) from Torpedo marmorata and the X-ray crystal structure of the acetylcholine binding protein (AChBP) from Lymnaea stagnalis were used as starting templates for comparative modelling. Features of the modelling approach used in the development of this GABAA model include: (1) multiple sequence alignment of members of the Cys-loop superfamily; (2) the design and implementation of a quasi-ab initio loop modelling algorithm; (3) expansion of the transmembrane domain (TMD) ion pore to model the open-state of the GABAA channel; (4) hydrophobicity analysis of the TMD to refine the structure in regions involved in general anaesthetic binding. The final model of the alpha1beta2gamma2 GABAA protein agrees with available experimental data concerning general anaesthetics.     

Study on the disulfide bond and disulfide loop of native and mutated SOD1 protein

The superoxide anions in the human body are reduced into hydrogen peroxide and molecular oxygen by the metallo enzyme Cu–Zn superoxide dismutase 1. The disulfide bond in SOD1 is essential to maintain the structural stability of protein and its proper folding. A computational study on the disulfide bond with the addition of residues was made using three different level of theories viz., B3LYP/6-31G (d,p), M052X/6-31G (d,p) and MP2/6-31G (d,p). The nature of disulfide bond was found to be unaffected with the additional residues being attached to the termini of cysteine residues. This result was found to be in agreement with the experimental values. The results of Molecular Dynamics simulation illustrate the crinkled appearance caused in the disulfide loop of A4V mutation. The conformational change in the disulfide loop was found to have significant effect on the loss of dimerization, metal binding affinity and overall protein stability. It is also noted that the disulfide loop with more number of residues is found to have no effect on the disulfide bond characteristics, but the disulfide loop with less number of residues is found to have remarkable effect for mutation in any position of the wild type protein.     

双环硫化磷酸酯类化合物的三维定量构效关系研究

为获取双环硫化磷酸酯类化合物对昆虫γ-氨基丁酸(GABA)受体和哺乳动物γ-氨基丁酸受体的亲合选择性定量信息,基于实验所测定的双环硫化磷酸酯类化合物与家蝇及大白鼠GABA受体结合的相对活性数据,以比较分子场分析法进行了三维定量构效关系研究。结果表明,家蝇GABA受体的亲合区域有足够的空间容纳异丙基和异丁基,但大白鼠GABA受体的亲合区域不具有这样的空间。     

双环硫化磷酸酯类化合物的三维定量构效关系研究

为获取双环硫化磷酸酯类化合物的昆虫γ-氨基丁酸（GABA）受体和哺乳动物γ-氨基丁酸受体的亲合选择性定量信息，基于实验所测定的双环硫化磷酸酯类化合物与家蝇及大白鼠GABA受体结合的相对活性数据，以比较分子场分析法进行了三维定量构效关系研究。结果表明，家蝇GABA受体的亲合区域有足够的窨容纳异丙基和异丁基，但大白鼠GABA受体的亲合区域不具有这样的空间。     

In silico investigation into the interactions between murine 5-HT3 receptor and the principle active compounds of ginger (Zingiber officinale)

Gingerols and shogaols are the primary non-volatile actives within ginger (Zingiber officinale). These compounds have demonstrated in vitro to exert 5-HT₃ receptor antagonism which could benefit chemotherapy-induced nausea and vomiting (CINV). The site and mechanism of action by which these compounds interact with the 5-HT₃ receptor is not fully understood although research indicates they may bind to a currently unidentified allosteric binding site. Using in silico techniques, such as molecular docking and GRID analysis, we have characterized the recently available murine 5-HT₃ receptor by identifying sites of strong interaction with particular functional groups at both the orthogonal (serotonin) site and a proposed allosteric binding site situated at the interface between the transmembrane region and the extracellular domain. These were assessed concurrently with the top-scoring poses of the docked ligands and included key active gingerols, shogaols and dehydroshogaols as well as competitive antagonists (e.g. setron class of pharmacologically active drugs), serotonin and its structural analogues, curcumin and capsaicin, non-competitive antagonists and decoys. Unexpectedly, we found that the ginger compounds and their structural analogs generally outscored other ligands at both sites. Our results correlated well with previous site-directed mutagenesis studies in identifying key binding site residues. We have identified new residues important for binding the ginger compounds. Overall, the results suggest that the ginger compounds and their structural analogues possess a high binding affinity to both sites. Notwithstanding the limitations of such theoretical analyses, these results suggest that the ginger compounds could act both competitively or non-competitively as has been shown for palonosetron and other modulators of CYS loop receptors.     

Development of CoMFA models of affinity and selectivity to angiotensin II type-1 and type-2 receptors

The renin-angiotensin system (RAS) is of major importance in cardiovascular and renal regulation and has been an attractive target in drug discovery for a long time. The main receptors involved in the RAS are the Angiotensin type-1 (AT(1)) and type-2 (AT(2)) receptors, which are both activated by the endogenous octapeptide angiotensin II (AngII). This study describes the development of 3D-QSAR models for AT(1) and AT(2) receptor affinity and AT(1)/AT(2) receptor selectivity using CoMFA. A data set of 244 compounds, based on the triazolinone and quinazolinone structural classes was compiled from the literature. Before CoMFA could be performed, an alignment rule for the two structural classes was defined using the pharmacophore-searching program DISCOtech. Models were validated using a test set obtained by dividing the data set into a training set and test set using hierarchical clustering, based on the CoMFA fields, AT(1)-, AT(2)-receptor affinities, and AT(1)/AT(2) selectivity values. Predictive models with good statistics could be developed both for AT(1) and AT(2) receptor affinity as well as selectivity towards these receptors.     

BMP-2 and BMP-9 binding specificities with ALK-3 in aqueous solution with dynamics

Signal ligands of the transforming growth factor-β (TGF-β) superfamily include the bone morphogenetic proteins (BMPs). BMPs bind to type I and type II serine-threonine kinase receptors and trigger the transphosphorylation cascade, wherein the active type II receptor phosphorylates the inactive type I receptor. This process further activates the cytoplasmic effectors of the pathway, such as SMAD proteins, which are homologs of both the Drosophila protein MAD (mothers against decapentaplegic) and the Caenorhabditis elegans protein SMA (small body size). Even though biological and medicinal studies have been performed on these complex species, we currently do not know the underlying molecular mechanisms of the signal ligand interactions with the receptors. Detailed understanding of these interactions increases our knowledge about these proteins, and also can provide the lacking information for successful mutation experiments. This study focuses on the computational analysis of binding affinities and structural binding specificities of two different types of BMPs (BMP-2 and BMP-9) to the activin receptor-like kinases (ALK-3) in solution. For studying the binding characteristics of BMP-2 or BMP-9 with ALK-3 in aqueous solution, we performed extensive molecular dynamics simulations coupled with thermodynamic calculations. The calculated thermodynamic properties show that the BMP-2/ALK-3 complex is thermodynamically more stable than a possible BMP-9/ALK-3 species in aqueous solution. The binding free energies indicate that ALK-3 preferably binds to BMP-2 instead of BMP-9. The structural analysis shows that ALK-3 binding with BMP-2 occurs in a perfectly symmetry pathway, whereas this symmetry is lost for possible ALK-3 interactions with BMP-9. The Phe49 to Val70 loop region of BMP-2 presents strong inter-molecular interactions with ALK-3. On the other hand, BMP-9 presents weaker interactions with ALK-3 via a non-continuous sequence. ALK-3-binding region of BMP-2 corresponds to the region predicted to be flexible by our intrinsic disorder analysis, whereas the related region of BMP-9 is expected to be noticeably less flexible. This study proposes that mutating the BMP-9 with the partial Phe49 to Val70 sequence of BMP-2 can help to increase the reactivity of BMP-9 towards stable ALK-3 binding, which in turn has the potential to develop new signaling pathways for improving the formation of tissues and to prevent or treat severe diseases. Furthermore, this study also demonstrates the usefulness of theoretical physical chemistry tools, such as molecular dynamics simulations and the ProtMet simulation software package in the structural characterization of the TGF-β superfamily proteins.     

Molecular simulation study of the unbinding of α-conotoxin [ϒ4E]GID at the α7 and α4β2 neuronal nicotinic acetylcholine receptors

The α7 and α4β2 neuronal nicotinic receptors belonging to the family of ligand-gated ion channels are most prevalent in the brain, and are implicated in various neurodegenerative disorders. α-conotoxin GID (and its analogue [ϒ4E]GID) specifically inhibits these subtypes, with more affinity towards the human α7 (hα7) subtype, and is valuable in understanding the physiological roles of these receptors. In this study, we use umbrella-sampling molecular dynamics simulations to understand the mechanism of interaction between [ϒ4E]GID and the agonist binding pockets of the α4β2 and the hα7 receptors, and to estimate their relative binding affinities (ΔG_bind). The obtained ΔG_bind values indicate stronger interaction with the hα7 receptor, in agreement with previous experimental studies. Simulations also revealed different unbinding pathways between the two receptor subtypes, enabling identification of a number of interactions at locations far from the orthosteric binding site which may explain the difference in [ϒ4E]GID potency. The pathways identified will help in the design of novel conotoxins with increased potency at α4β2, for which there is currently no known highly potent conotoxin inhibitor. Computational mutational free energy analyses also revealed a number of possible single-site mutations to GID which might enhance its selective binding to α4β2 over α7.     

Interactions of acetylcholine binding site residues contributing to nicotinic acetylcholine receptor gating: Role of residues Y93, Y190, K145 and D200

The nicotinic acetylcholine receptor exhibits multiple conformational states, resting (channel closed), active (channel open) and desensitized (channel closed). The resting state may be distinguished from the active and desensitized states by the orientation of loop C in the extracellular ligand binding domain (LBD). Homology modeling was used to generate structures of the Torpedo californica α₂βδγ nAChR that initially represent the resting state (loop C open) and the desensitized state (loop C closed). Molecular dynamics (MD) simulations were performed on the extracellular LBD on each nAChR conformational state, with and without the agonist anabaseine present in each binding site (the αγ and the αδ sites). Three MD simulations of 10 ns each were performed for each of the four conditions. Comparison of dynamics revealed that in the presence of agonist, loop C was drawn inward and attains a more stable conformation. Examination of side-chain interactions revealed that residue αY190 exhibited hydrogen-bonding interactions either with residue αY93 in the ligand binding site or with residue αK145 proximal to the binding site. αK145 also exhibited side chain (salt bridge) interactions with αD200 and main chain interactions with αY93. Residues αW149, αY198, γY116/δT119, γL118/δL121 and γL108/δL111 appear to play the role of stabilizing ligand in the binding site. In MD simulations for the desensitized state, the effect of ligand upon the interactions among αK145, αY190, and αY93 as well as ligand-hydrogen-bonding to αW149 were more pronounced at the αγ interface than at the αδ interface. Differences in affinity for the desensitized state were determined experimentally to be 10-fold. The changes in side chain interactions observed for the two conformations and induced by ligand support a model wherein hydrogen bond interactions between αD200 and αY93 are broken and rearrange to form a salt-bridge between αK145 and αD200 and hydrogen bond interactions between αY93 and αY190 and between αK145 and αY190.     

Prediction of the 3-D structure of rat MrgA G protein-coupled receptor and identification of its binding site

Mrg receptors are orphan G protein-coupled receptors (GPCRs) located mainly at the specific set of sensory neurons in the dorsal root ganglia, suggesting a role in nociception. We report here the 3-D structure of rat MrgA (rMrgA) receptor [obtained from homology modeling to the recently validated predicted structures of mouse MrgA1 and MrgC11] and the structure of adenine (a known agonist, K(i)=18nM) bound to rMrgA. This predicted binding site is located within transmembrane helical domains (TMs) 3, 4, 5 and 6, with Asn residues in TM3 and TM4 identified as the key residues for adenine binding. Here the side chain of Asn88 (TM3) forms two pairs of hydrogen bonds with N3 and N9 of adenine while Asn146 (TM4) makes two pairs of hydrogen bonds with N1 and N6 of adenine. These interactions lock adenine tightly in the binding pocket. We also predict the binding site of guanine (not an agonist) and seven other derivatives. Guanine cannot make the hydrogen bond to Asn146 (TM4), leading to binding too weak to be observed experimentally. The predicted binding affinity for other adenine derivatives correlates with the availability of the hydrogen bonds to these two Asn residues. These results validate the predicted structure for rat MrgA and suggest mutation experiments that could further validate the structure. Moreover, the predicted structure and binding site should be useful for seeking other small molecule agonists and antagonists.