Computational prediction of the three-dimensional structures for the Caenorhabditis elegans tubulin family

In this article we characterize, from a structural point of view, all 16 members of the tubulin gene family of Caenorhabditis elegans (9 alpha-tubulins, 6 beta-tubulins, and 1 gamma-tubulin). We obtained their tertiary structures by computationally modifying the X-ray crystal structure of the pig brain alpha/beta-tubulin dimer published by Nogales et al. [Nature (London) 1998;391:199-203]. Our computational protocol involves changing the amino acids (with MIDAS; Jarvis et al., UCSF MIDAS. University of California, San Francisco, 1986) in the 3D structure of pig brain alpha/beta-tubulin dimer followed by geometry optimization with the AMBER force field (Perlman et al., AMBER 4. University of California, San Francisco, 1990). We subsequently analyze and compare the resulting structures in terms of the differences in their secondary and tertiary structures. In addition, we compare the pattern of hydrogen bonds and hydrophobic contacts in the guanosine triphosphate (GTP)-binding site for all members of the tubulin family. Our computational results show that, except for gamma-tubulin, all members of the C. elegans tubulin family have similar secondary and 3D structures and that the change in the pattern of hydrogen bonds in the GTP-binding site may be used to assess the relative stability of different alpha/beta-tubulin dimers formed by monomers of the tubulin family.     

Investigation of structural requirements of anticancer activity at the paclitaxel/tubulin binding site using CoMFA and CoMSIA

CoMFA and CoMSIA analysis were utilized in this investigation to define the important interacting regions in paclitaxel/tubulin binding site and to develop selective paclitaxel-like active compounds. The starting geometry of paclitaxel analogs was taken from the crystal structure of docetaxel. A total of 28 derivatives of paclitaxel were divided into two groups-a training set comprising of 19 compounds and a test set comprising of nine compounds. They were constructed and geometrically optimized using SYBYL v6.6. CoMFA studies provided a good predictability (q(2)=0.699, r(2)=0.991, PC=6, S.E.E.=0.343 and F=185.910). They showed the steric and electrostatic properties as the major interacting forces whilst the lipophilic property contribution was a minor factor for recognition forces of the binding site. These results were in agreement with the experimental data of the binding activities of these compounds. Five fields in CoMSIA analysis (steric, electrostatic, hydrophobic, hydrogen-bond acceptor and donor properties) were considered contributors in the ligand-receptor interactions. The results obtained from the CoMSIA studies were: q(2)=0.535, r(2)=0.983, PC=5, S.E.E.=0.452 and F=127.884. The data obtained from both CoMFA and CoMSIA studies were interpreted with respect to the paclitaxel/tubulin binding site. This intuitively suggested where the most significant anchoring points for binding affinity are located. This information could be used for the development of new compounds having paclitaxel-like activity with new chemical entities to overcome the existing pharmaceutical barriers and the economical problem associated with the synthesis of the paclitaxel analogs. These will boost the wide use of this useful class of compounds, i.e. in brain tumors as the most of the present active compounds have poor blood-brain barrier crossing ratios and also, various tubulin isotypes has shown resistance to taxanes and other antimitotic agents.     

Applying linear interaction energy method for binding affinity calculations of podophyllotoxin analogues with tubulin using continuum solvent model and prediction of cytotoxic activity

Podophyllotoxin and its analogues have important therapeutic value in the treatment of cancer, due to their ability to induce apoptosis in cancer cells in a proliferation-independent manner. These ligands bind to colchicine binding site of tubulin near the α- and β-tubulin interface and interfere with tubulin polymerization. The binding free energies of podophyllotoxin-based inhibitors of tubulin were computed using a linear interaction energy (LIE) method with a surface generalized Born (SGB) continuum solvation model. A training set of 76 podophyllotoxin analogues was used to build a binding affinity model for estimating the free energy of binding for 36 inhibitors (test set) with diverse structural modifications. The average root mean square error (RMSE) between the experimental and predicted binding free energy values was 0.56 kcal/mol which is comparable to the level of accuracy achieved by the most accurate methods, such as free energy perturbation (FEP) or thermodynamic integration (TI). The squared correlation coefficient between experimental and SGB–LIE estimates for the free energy for the test set compounds is also significant (R² = 0.733). On the basis of the analysis of the binding energy, we propose that the three-dimensional conformation of the A, B, C and D rings is important for interaction with tubulin. On the basis of this insight, 12 analogues of varying ring modification were taken, tested with LIE methodology and then validated with their experimental potencies of tubulin polymerization inhibition. Low levels of RMSE for the majority of inhibitors establish the structure-based LIE method as an efficient tool for generating more potent and specific inhibitors of tubulin by testing rationally designed lead compounds based on podophyllotoxin derivatization.     

Applying linear interaction energy method for binding affinity calculations of podophyllotoxin analogues with tubulin using continuum solvent model and prediction of cytotoxic activity

Podophyllotoxin and its analogues have important therapeutic value in the treatment of cancer, due to their ability to induce apoptosis in cancer cells in a proliferation-independent manner. These ligands bind to colchicine binding site of tubulin near the α- and β-tubulin interface and interfere with tubulin polymerization. The binding free energies of podophyllotoxin-based inhibitors of tubulin were computed using a linear interaction energy (LIE) method with a surface generalized Born (SGB) continuum solvation model. A training set of 76 podophyllotoxin analogues was used to build a binding affinity model for estimating the free energy of binding for 36 inhibitors (test set) with diverse structural modifications. The average root mean square error (RMSE) between the experimental and predicted binding free energy values was 0.56 kcal/mol which is comparable to the level of accuracy achieved by the most accurate methods, such as free energy perturbation (FEP) or thermodynamic integration (TI). The squared correlation coefficient between experimental and SGB–LIE estimates for the free energy for the test set compounds is also significant (R² = 0.733). On the basis of the analysis of the binding energy, we propose that the three-dimensional conformation of the A, B, C and D rings is important for interaction with tubulin. On the basis of this insight, 12 analogues of varying ring modification were taken, tested with LIE methodology and then validated with their experimental potencies of tubulin polymerization inhibition. Low levels of RMSE for the majority of inhibitors establish the structure-based LIE method as an efficient tool for generating more potent and specific inhibitors of tubulin by testing rationally designed lead compounds based on podophyllotoxin derivatization.     

A possible model of benzimidazole binding to beta-tubulin disclosed by invoking an inter-domain movement

Although it is well established that benzimidazole (BZMs) compounds exert their therapeutic effects through binding to helminth beta-tubulin and thus disrupting microtubule-based processes in the parasites, the precise location of the benzimidazole-binding site on the beta-tubulin molecule has yet to be determined. In the present study, we have used previous experimental data as cues to help identify this site. Firstly, benzimidazole resistance has been correlated with a phenylalanine-to-tyrosine substitution at position 200 of Haemonchus contortus beta-tubulin isotype-I. Secondly, site-directed mutagenesis studies, using fungi, have shown that other residues in this region of the protein can influence the interaction of benzimidazoles with beta-tubulin. However, the atomic structure of the alphabeta-tubulin dimer shows that residue 200 and the other implicated residues are buried within the protein. This poses the question: how might benzimidazoles interact with these apparently inaccessible residues? In the present study, we present a mechanism by which those residues generally believed to interact with benzimidazoles may become accessible to the drugs. Furthermore, by docking albendazole-sulphoxide into a modelled H. contortus beta-tubulin molecule we offer a structural explanation for how the mutation conferring benzimidazole resistance in nematodes may act, as well as a possible explanation for the species-specificity of benzimidazole anthelmintics.     

Empirical rules facilitate the search for binding sites on protein surfaces

Computational surface screening of 3D protein structures is a valuable means of finding possible docking sites for substrates, effectors and similar molecules. It can be improved by considering properties of molecules which are known to bind to protein surfaces, and thus reflect the required properties of binding sites. In-depth studies are available on drugs and lead compounds as binding partners with statistically assured properties. Here we present a simple strategy for finding binding sites, which is based on the empirical rule-of-five by Lipinski et al. for oral drugs and the rule-of-three by Congreve et al. for leads. The fast automated search with the new C-code TRIDOCK yields a preliminary set of sites, thus facilitating further investigation by visual, comparative and quantitative work. Possible binding sites are tagged by pseudo-atoms added to the structure file for molecular graphical evaluation. Usually, the strategy yields not just a few single sites, but an accumulation of several sites in known substrate binding pockets. Clusters are also found at known or putative protein-protein docking interfaces. A comparison of the activated and inactivated form of the GTPase Ras reveals clear differences and identifies a niche, which is possibly a suitable new target for compounds that bind specifically to activated Ras.     

Simplified AutoDock force field for hydrated binding sites

A set of high quality structures of protein-ligand complexes with experimentally determined binding affinities has been extracted from the Protein Data Bank and used to test and recalibrate AutoDock force field. Since for some binding sites water molecules are crucial for bridging the receptor-ligand interactions, they have to be included in the analysis. To simplify the process of incorporating water molecules into the binding sites and make it less ambiguous, new simple water model was created. After recalibration of the force field on the new dataset much better correlation between the computed and experimentally determined binding affinities was achieved and the quality of pose prediction improved even more.     

Quantitative prediction of MHC-II peptide binding affinity using relevance vector machine

Peptide-MHC binding is an important prerequisite event and has immediate consequences to immune response. Those peptides binding to MHC molecules can activate the T-cell immunity, and they are useful for understanding the immune mechanism and developing vaccines for diseases. Accurate prediction of the binding between peptides and MHC-II molecules has long been a challenge in bioinformatics. Recently, instead of differentiating peptides as binder or non-binder, researchers are more interested in making predictions directly on peptide binding affinities. In this paper, we investigate the use of relevance vector machine to quantitatively predict the binding affinities between MHC-II molecules and peptides. In our scheme, a new encoding scheme is used to generate the input vectors, and then by using relevance vector machine we develop the prediction models on the basis of binding cores, which are recognized in an iterative self-consistent way. When applied to three MHC-II molecules DRB1*0101, DRB1*0401 and DRB1*1501, our method produces consistently better performance than several popular quantitative methods, in terms of cross-validated squared error, cross-validated correlation coefficient, and area under ROC curve. All evidences indicate that our method is an effective tool for MHC-II binding affinity prediction.     

Improving the analysis of phylogenetic data

Methods of phylogenetic analysis are presented that result in corrections of highly biased data sets, particularly those in which there are great differences between mutation and/or substitution rates from one nucleotide site to another along a DNA sequence. Two approaches are discussed. In the first, pairwise comparisons of a set of sequences are used to determine whether the most recent substitutions take place at the sites that are most polymorphic—that is, where the mutational “hot spots” are located. In the second, a “topiary pruning” method is used to remove selectively the bases in the data set that are most likely to occupy these hot spots and therefore to result in homoplastic substitutions. The two methods combined yield new and substantially older estimates of the time at which the mitochondrial Eve lived, and increase the likelihood that she lived in Africa. In these data, transversions provide a more satisfactory yardstick for phylogenetic analysis than transitions, because there is no detectable tendency for transversions to occur at mutational hot spots.     

Molecular modelling and competition binding study of Br-noscapine and colchicine provide insight into noscapinoid-tubulin binding site

We have previously discovered the tubulin-binding anti-cancer properties of noscapine and its derivatives (noscapinoids). Here, we present three lines of evidence that noscapinoids bind at or near the well studied colchicine binding site of tubulin: (1) in silico molecular docking studies of Br-noscapine and noscapine yield highest docking score with the well characterised colchicine-binding site from the co-crystal structure; (2) the molecular mechanics-generalized Born/surface area (MM-GB/SA) scoring results ΔΔG(bind-cald) for both noscapine and Br-noscapine (3.915 and 3.025 kcal/mol) are in reasonably good agreement with our experimentally determined binding affinity (ΔΔG(bind-Expt) of 3.570 and 2.988 kcal/mol, derived from K(d) values); and (3) Br-noscapine competes with colchicine binding to tubulin. The simplest interpretation of these collective data is that Br-noscapine binds tubulin at a site overlapping with, or very close to colchicine-binding site of tubulin. Although we cannot rule out a formal possibility that Br-noscapine might bind to a site distinct and distant from the colchicine-binding site that might negatively influence the colchicine binding to tubulin.     

Theoretical studies on binding modes of copper-based nucleases with DNA

In the present work, molecular simulations were performed for the purpose of predicting the binding modes of four types of copper nucleases (a total 33 compounds) with DNA. Our docking results accurately predicted the groove binding and electrostatic interaction for some copper nucleases with B-DNA. The intercalation modes were also reproduced by “gap DNA”. The obtained results demonstrated that the ligand size, length, functional groups and chelate ring size bound to the copper center could influence the binding affinities of copper nucleases. The binding affinities obtained from the docking calculations herein also replicated results found using MM-PBSA approach. The predicted DNA binding modes of copper nucleases with DNA will ultimately help us to better understand the interaction of copper compounds with DNA.     

Using fragment lengths from incomplete digestion by multiply cleaving enzymes to map antibody binding sites on a protein.

The problem of mapping the positions of the unique binding sites of several monoclonal antibodies on a linear protein structure is considered. Data giving the incidence of binding of individual antibodies to fragments of the protein obtained from it by the incomplete chemical or enzymatic digestion are used to formulate a series of linear programming problems. The solution to these problems shows which orderings of binding sites are possible, and gives upper and lower bounds for the relative positions of the sites.     

Towards the identification of the binding site of benzimidazoles to β-tubulin of Trichinella spiralis: Insights from computational and experimental data

Benzimidazole-2-carbamate derivatives (BzC) are among the most important broad-spectrum anthelmintic drugs for the treatment of nematode infections. BzC selectively bind to the β-tubulin monomer and inhibit microtubule polymerization. However, the crystallographic structure of the nematode tubulin and the mechanism of action are still unknown. Moreover, the relation between the mechanism of action and the binding site of BzC has not yet been explained accurately. By using the amino acid sequence of Trichinella spiralis β-tubulin as a basis and by applying homology modeling techniques, we were able to build a 3D structure of this protein. In order to identify a binding site for BzC, molecular docking and molecular dynamics calculations were carried out with this model. The results were in good agreement with the most common amino acid mutations associated with drug resistance (F167Y, E198A and F200Y) and with the experimental results of competitive inhibition of colchicine binding to tubulin. Besides, Glu198, Thr165, Cys239 and Gln134 were identified as important amino acids in the binding process since they directly interact with BzC in the formation of hydrogen bonds. The results presented in this paper are a step further towards the understanding, at the molecular level, of the mode of action of anthelmintic drugs. These results constitute valuable information for the design or improvement of more potent and selective molecules.     

Elucidation of the Na+, K+-ATPase digitalis binding site

Despite controversy over their use and the potential for toxic side effects, cardiac glycosides have remained an important clinical component for the treatment for congestive heart failure (CHF) and supraventricular arrhythmias since the effects of Digitalis purpurea were first described in 1785. While there is a wealth of information available with regard to the effects of these drugs on their pharmacological receptor, the Na(+), K(+)-ATPase, the exact molecular mechanism of digitalis binding and inhibition of the enzyme has remained elusive. In particular, the absence of structural knowledge about Na(+), K(+)-ATPase has thwarted the development of improved therapeutic agents with larger therapeutic indices via rational drug design approaches. Here, we propose a binding mode for digoxin and several analogues to the Na(+), K(+)-ATPase. A 3D-structural model of the extracellular loop regions of the catalytic alpha1-subunit of the digitalis-sensitive sheep Na(+), K(+)-ATPase was constructed from the crystal structure of an E(1)Ca(2+) conformation of the SERCA1a and a consensus orientation for digitalis binding was inferred from the in silico docking of a series of steroid-based cardiotonic compounds. Analyses of species-specific enzyme affinities for ouabain were also used to validate the model and, for the first time, propose a detailed model of the digitalis binding site.     

Prediction of protein-ligand binding affinities using multiple instance learning

Accurate prediction of protein-ligand binding affinities for lead optimization in drug discovery remains an important and challenging problem on scoring functions for docking simulation. In this paper, we propose a data-driven approach that integrates multiple scoring functions to predict protein-ligand binding affinity directly. We then propose a new method called multiple instance regression based scoring (MIRS) that incorporates unbound ligand conformations using multiple scoring functions. We evaluated the predictive performance of MIRS using 100 protein-ligand complexes and their binding affinities. The experimental results showed that MIRS outperformed the 11 conventional scoring functions including LigScore, PLP, AutoDock, G-Score, D-Score, LUDI, F-Score, ChemScore, X-Score, PMF, and DrugScore. In addition, we confirmed that MIRS performed well on binding pose prediction. Our results reveal that it is indispensable to incorporate unbound ligand conformations in both binding affinity prediction and binding pose prediction. The proposed method will accelerate efficient lead optimization on structure-based drug design and provide a new direction to designing of new scoring score functions.     

Comparative performance evaluation of hot spot contention betweenMIN-based and ring-based shared-memory architectures

Hot spot contention on a network-based shared-memory architecture occurs when a large number of processors try to access a globally shared variable across the network. While multistage interconnection network (MIN) and hierarchical ring (HR) structures are two important bases on which to build large scale shared-memory multiprocessors, the different interconnection networks and cache/memory systems of the two architectures respond very differently to network bottleneck situations. In this paper, we present a comparative performance evaluation of hot spot effects on the MIN-based and HR-based shared-memory architectures. Both nonblocking MIN-based and HR-based architectures are classified, and analytical models are described for understanding network differences and for evaluating hot spot performance on both architectures. The analytical comparisons indicate that HR-based architectures have the potential to handle various contentions caused by hot spots more efficiently than MIN-based architectures. Intensive performance measurements on hot spots have been conducted on the BBN TC2000 (MIN-based) and the KSR1 (HR-based) machines. Performance experiments were also conducted on the practical experience of hot spots with respect to synchronization lock algorithms. The experimental results are consistent with the analytical models, and present practical observations and an evaluation of hot spots on the two types of architectures     

Classification and comparison of ligand-binding sites derived from grid-mapped knowledge-based potentials

We describe the application of knowledge-based potentials implemented in the MOE program to compare the ligand-binding sites of several proteins. The binding probabilities for a polar and a hydrophobic probe are calculated on a grid to allow easy comparison of binding sites of superimposed related proteins. The method is fast and simple enough to simultaneously use structural information of multiple proteins of a target family. The method can be used to rapidly cluster proteins into subfamilies according to the similarity of hydrophobic and polar fields of their ligand-binding sites. Regions of the binding site which are common within a protein family can be identified and analysed for the design of family-targeted libraries or those which differ for improvement of ligand selectivity. The field-based hierarchical clustering is demonstrated for three protein families: the ligand-binding domains of nuclear receptors, the ATP-binding sites of protein kinases and the substrate binding sites of proteases. More detailed comparisons are presented for serine proteases of the chymotrypsin family, for the peroxisome proliferator-activated receptor subfamily of nuclear receptors and for progesterone and androgen receptor. The results are in good accordance with structure-based analysis and highlight important differences of the binding sites, which have been also described in the literature.     

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.     

Probing the binding site characteristics of HSA: A combined molecular dynamics and cheminformatics investigation

Human serum albumin is a remarkable protein found in high concentrations in the body. It contains at least seven distinct fatty acid binding sites and two principle sites for drugs. Its primary function is to act as a fatty acid transport system, but it also shows the capacity to bind a diverse range of acidic, neutral and zwitterionic drug molecules. In this paper we investigate the ligand binding selectivity of HSA using cheminformatics analyses and molecular dynamics simulations. We compare and contrast the known ligand binding specificities as obtained from X-ray structural data using PCA, with additional direct analyses of the seven key binding pockets using analyses derived from molecular simulations. We assess both the fatted and defatted states of HSA using 100 ns simulations of the APO and HOLO forms, as well as structures containing one, three and seven myristic acid molecules. We find that differences in fatty acid binding can have a dramatic effect on the flexibility of the protein and also the pocket characteristics. We discuss how the remarkable selectivity of the HSA pockets towards both endogenous fatty acids and exogenous drug molecules is not simply controlled by bulk property effects such as ionization state and lipophilicity.     

Identification of a putative binding site for 5-alkyl-benzothiadiazides in the AMPA receptor dimer interface

Crystal structures of three different allosteric modulators co-crystallized with the iGluR2 ligand-binding domain are currently available. The modulators, cyclothiazide, aniracetam and CX614, bind at overlapping binding sites in the dimer interface between two iGluR2 subunits. However, pharmacological data indicate that there are one or more additional binding sites for this class of compounds. Based on differences in structure-activity relationship data we show that 5-alkyl-benzothiadiazide (5ABTD) modulators and a series of close analogs of cyclothiazide, despite having a common core structure, do not have the same binding site. In the present work, a new potential binding site for allosteric modulators has been identified in the dimer interface of the iGluR2 ligand-binding domain. By comparing different iGluR2 crystal structures including different co-crystallized agonists, this cavity is shown to be a structurally conserved part of the dimer interface. The cavity is characterized with respect to shape and potential favorable interactions with ligands and docking is used to find a reasonable binding mode for the core structure of the 5ABTDs. The extensive structure-activity data available for this series of compounds are in agreement with the proposed binding mode, supporting the conclusion that the identified cavity most likely is the binding site for the 5ABTDs.