A novel small molecule that directly sensitizes the insulin receptor in vitro and in vivo.

Insulin resistance, an important feature of type 2 diabetes, is manifested as attenuated insulin receptor (IR) signaling in response to insulin binding. A drug that promotes the initiation of IR signaling by enhancing IR autophosphorylation should, therefore, be useful for treating type 2 diabetes. This report describes the effect of a small molecule IR sensitizer, TLK16998, on IR signaling. This compound activated the tyrosine kinase domain of the IR beta-subunit at concentrations of 1 micromol/l or less but had no effect on insulin binding to the IR alpha-subunit even at much higher concentrations. TLK16998 alone had no effect on IR signaling in mouse 3T3-L1 adipocytes but, at concentrations as low as 3.2 micromol/l, enhanced the effects of insulin on the phosphorylation of the IR beta-subunit and IR substrate 1, and on the amount of phosphatidylinositol 3-kinase that coimmunoprecipitated with IRS-1. Phosphopeptide mapping revealed that the effect of TLK16998 on the IR was associated with increased tyrosine phosphorylation of the activation loop of the beta-subunit tyrosine kinase domain. TLK16998 also increased the potency of insulin in stimulating 2-deoxy-D-glucose uptake in 3T3-L1 adipocytes, with a detectable effect at 8 micromol/l and a 10-fold increase at 40 micromol/l. In contrast, only small effects were observed on IGF-1-stimulated 2-deoxy-D-glucose uptake. In diabetic mice, TLK16998, at a dose of 10 mg/kg, lowered blood glucose levels for up to 6 h. These results suggest, therefore, that small nonpeptide molecules that directly sensitize the IR may be useful for treating type 2 diabetes.

The structures of human glutathione transferase P1-1 in complex with glutathione and various inhibitors at high resolution.

The human pi-class glutathione S-transferase (hGST P1-1) is a target for structure-based inhibitor design with the aim of developing drugs that could be used as adjuvants in chemotherapeutic treatment. Here we present seven crystal structures of the enzyme in complex with substrate (glutathione) and two inhibitors (S-hexyl glutathione and gamma-glutamyl- (S-benzyl)cysteinyl-D-phenylglycine). The binding of the modified glutathione inhibitor, gamma-glutamyl-(S-benzyl)cysteinyl-D-phenylglycine, has been characterized with the phenyl group stacking against the benzyl moiety of the inhibitor and making interactions with the active-site residues Phe8 and Trp38. The structure provides an explanation as to why this compound inhibits the pi-class GST much better than the other GST classes. The structure of the enzyme in complex with glutathione has been determined to high resolution (1.9 to 2.2 A) in three different crystal forms and at two different temperatures (100 and 288 K). In one crystal form, the direct hydrogen-bonding interaction between the hydroxyl group of Tyr7, a residue involved in catalysis, and the thiol group of the substrate, glutathione, is broken and replaced by a water molecule that mediates the interaction. The hydrogen-bonding partner of the hydroxyl group of Tyr108, another residue implicated in the catalysis, is space-group dependent. A high-resolution (2.0 A) structure of the enzyme in complex with S-hexyl glutathione in a new crystal form is presented. The enzyme-inhibitor complexes show that the binding of ligand into the electrophilic binding site does not lead to any conformational changes of the protein.

Predicting ligand binding to proteins by affinity fingerprinting.

BACKGROUND: There are many ways to represent a molecule's properties, including atomic-connectivity drawings, NMR spectra, and molecular orbital models. Prior methods for predicting the biological activity of compounds have largely depended on these physical representations. Measuring a compound's binding potency against a small reference panel of diverse proteins defines a very different representation of the molecule, which we call an affinity fingerprint. Statistical analysis of such fingerprints provides new insights into aspects of binding interactions that are shared among a wide variety of proteins. These analyses facilitate prediction of the binding properties of these compounds assayed against new proteins. RESULTS: Affinity fingerprints are reported for 122 structurally-diverse compounds using a reference panel of eight proteins that collectively are able to generate unique fingerprints for about 75% of the small organic compounds tested. Application of multivariate regression techniques to this database enables the creation of computational surrogates to represent new proteins that are surprisingly effective at predicting binding potencies. We illustrate this for two enzymes with no previously recognizable similarity to each other or to any of the reference proteins. Fitting of analogous computational surrogates to four other proteins confirms the generality of the method; when applied to a fingerprinted library of 5000 compounds, several sub-micromolar hits were correctly predicted. CONCLUSIONS: An affinity fingerprint database, which provides a rich source of data defining operational similarities among proteins, can be used to test theories of cryptic homology unexpected from current understanding of protein structure. Practical applications to drug design include efficient pre-screening of large numbers of compounds against target proteins using fingerprint similarities, supplemented by a small number of empirical measurements, to select promising compounds for further study.

Glubodies: randomized libraries of glutathione transferase enzymes.

BACKGROUND: The immunoglobulin framework has been mutagenized to engineer recombinant libraries of proteins as potential diagnostics and novel catalysts, although the often shallow binding cleft may limit the utility of this framework for binding diverse small organic molecules. By contrast, the glutathione S-transferase (GST) family of enzymes contains a deep binding cleft, which has evolved to accommodate a broad range of hydrophobic xenobiotics. We set out to determine whether GST molecules with novel ligand-binding characteristics could be produced by random mutagenesis of segments of the binding cleft. RESULTS: We have identified two ligand-recognition segments (LRSs) in human GST P1, which are near the active site in the folded protein, but have characteristics indicating that the integrity of their sequence is not essential for the overall structure or activity of the protein. Libraries of GST P1-derived proteins were produced by substituting randomized sequences for an LRS or inserting random sequences into an LRS. The recombinant proteins in the libraries, collectively designated as 'glubodies,' generally retain enzymatic activity but differ markedly both from each other and from the parent enzyme in sensitivity to inhibition by diverse small organic compounds. In some instances, a glubody is inhibited by completely novel structures. CONCLUSIONS: We have shown that a non-antibody framework can be used to create large libraries of proteins with a wide range of binding specificities for small organic molecules. The glubodies provide a rich source of data for correlating the structural and functional features of proteins relevant to ligand binding. The criteria applied for identifying an LRS in GST P1 are generally applicable to other protein frameworks.

Amino acid preferences at protein binding sites.

An analysis of the amino acid distribution at protein binding sites was carried out using 50 diverse macromolecules for which crystallographic data with a bound ligand are available. The purpose of this study is to determine whether differential trends in amino acid distributions exist at binding sites compared to other regions in the proteins. The results indicate that some residues, particularly Arg, His, Trp and Tyr are substantially more frequent at the binding sites, compared to the number of times these residues are present in proteins generally. These effects go beyond the differences seen comparing surface exposed residues to bulk protein. The resemblance in the residue utilization at the binding sites of unrelated proteins restricts the possible types of interactions with ligands, possibly accounting for the repetition of substructural motifs in chemicals with diverse pharmacological action. Further, the use of these diagnostic features may permit identification of ligand binding pockets in a protein structure deduced from sequence information or from data in the absence of a ligand. Some of these findings complement and extend previously described trends for antibody binding sites.

Molecular electrostatic potential of D1 and D2 dopamine agonists.

The molecular electrostatic potential (MEP) of four selective D1, four nonselective D1/D2, and three selective D2 agonists has been calculated in a three-dimensional grid surrounding the molecules. The local density functional program DMol was used to evaluate the MEP. A comparison of the MEPs of all compounds revealed that while the electrostatic effects may be important for the affinity in both D1- and D2-selective ligands, it only appears to be a subtle modulator of the selectivity. Slight differences were found in the negative regions in the vicinity of the catechol ring that can account for the D1 versus D2 selectivity in the compounds studied.

Toward the design of chemical libraries for mass screening biased against mutagenic compounds.

The ability to develop a chemical into a drug depends on multiple factors. Beyond potency and selectivity, ADME/PK and the toxicological profile of the compound play a significant role in its evaluation as a candidate for development. Those factors are being brought into bear earlier in the discovery process and even into the design of libraries for screening. The purpose of our study is the comparative analysis of simple physical characteristics of compounds that have been reported to be mutagens and nonmutagenic ones. The analysis of differences can lead to the development of knowledge-based biases in the libraries designed for massive screening. For each of four Salmonella strains, TA-98, TA-100, TA-1535, and TA-1537, an analysis of the statistical significance of the deviance of the averages for a number of global properties was carried out. The properties studied included parameters, such as topological indices, and bit strings representing the presence or absence of certain chemical moieties. The results suggest that mutagens display a larger number of hydrogen bond acceptor centers for most strains. Moreover, the use of bit strings points to the importance of certain molecular fragments, such a nitro groups, for the outcome of a mutagenicity study. Development of multivariate models based on global molecular properties or bit strings point to a small advantage of the latter for the prediction of mutagenicity. The benefits of the bit strings are in accord with the use of fragment-based approaches for the prediction of carcinogenicity and mutagenicity in methods described in the literature.

LASSOO: a generalized directed diversity approach to the design and enrichment of chemical libraries.

Pharmaceutical discovery relies on the screening of chemical libraries that are as diverse as possible yet constrained in favor of compounds possessing attributes that are normally associated with successful drug candidates. We describe a new algorithm for simultaneously addressing both objectives, providing an effective means to increase structural diversity in a chemical library while maintaining a bias toward compounds that retain the desirable properties of drugs. The LASSOO algorithm exploits differences in descriptor distributions to identify novel compounds that are most dissimilar to the members of an existing screening library and most similar to members of a target library with desirable characteristics. We illustrate the LASSOO technique using publicly available compound databases and bit string descriptors. The architecture of the algorithm is general enough to allow any set of descriptors or similarity measures to be employed, and it is easily adaptable to other means of directing diversity, such as the avoidance of toxicity and/or poor pharmacokinetic properties.

Design, synthesis, and evaluation of latent alkylating agents activated by glutathione S-transferase.

In search of compounds with improved specificity for targeting the important cancer-associated P1-1 glutathione S-transferase (GST) isozyme, new analogs 4 and 5 of the previously reported glutathione S-transferase (GST)-activated latent alkylating agent gamma-glutamyl-alpha-amino-beta-[[[2-[[bis[bis(2-chloroethyl)amino]ph osp horyl]oxy]ethyl]sulfonyl]propionyl]-(R)-(-)-phenylglycine (3) have been designed, synthesized, and evaluated. One of the diastereomers of 4 exhibited good selectivity for GST P1-1. The tetrabromo analog 5 of the tetrachloro compound 3 maintained its specificity and was found to be more readily activated by GSTs than 3. The GST activation concept was further broadened through design, synthesis, and evaluation of a novel latent urethane mustard 8 and its diethyl ester 9. Interestingly, 8 showed very good specificity for P1-1 GST. Cell culture studies were carried out on 4, 5, 8, and 9 using cell lines engineered to have varying levels of GST P1-1 isozyme. New analogs 4 and 5 exhibited increased toxicity to cell lines with overexpressed GST P1-1 isozyme. The urethane mustard 8 and its diethyl ester 9 were found to be not as toxic. However, they too exhibited more toxicity to a cell line engineered to have elevated P1-1 levels, which was in agreement with the observed in vitro specificity of 8 for P1-1 GST isozyme. Mechanistic studies on alkaline as well as enzyme-catalyzed decomposition of latent mustard 3 provided experimental proof for the hypothesis that 3 breaks down into an active phosphoramidate mustard and a reactive vinyl sulfone. The alkylating nature of the decomposition products was further demonstrated by trapping those transient species as relatively stable diethyldithiocarbamic acid adducts. These results substantially extend previous efforts to develop drugs targeting GST and provide a paradigm for development of other latent drugs.

Thermodynamic analysis of binding to the cerebellar type I benzodiazepine receptor.

The temperature dependence of binding to the type I benzodiazepine receptor in rat cerebellum was determined using [3H]Ro15-1788 and regression analysis techniques. The ligands chosen were from diverse chemical families and display different pharmacological properties. Included were the agonists flunitrazepam, CL 218,872, zolpidem and alpidem; the antagonists Ro15-1788 and propyl-beta-carboline-3-carboxylate (beta-CCP); the inverse agonists ethyl-beta-carboline-3-carboxylate (beta-CCE) and methyl-6,7-dimethoxy-beta-carboline-3-carboxylate (DMCM); and the selective muscle relaxant AHR-11797. Assays were performed at 0 degrees C, 20 degrees C and 37 degrees C and the K(i) at each temperature was used to construct a van't Hoff plot for each compound. The binding of all ligands, with the exception of DMCM, was enthalpy-driven. However, enthalpy alone does not determine the rank order of affinity. There was no relationship between the thermodynamic behavior of binding and the observation of agonism, antagonism or inverse agonism, indicating that activation and recognition are distinct steps in this receptor system.

Evidence for more than two central benzodiazepine receptors in rat spinal cord.

Competitive binding assays were performed in rat spinal cord membranes at 0 degrees C against [3H]Ro15-1788. The displacement curves of Ro15-1788 and alpidem produced pseudo-Hill slopes of 1.0 and 0.5, respectively. In addition, 5-10% of the specifically bound [3H]Ro15-1788 could not be displaced by greater than 100 microM of alpidem. These two pieces of evidence strongly suggest the presence of at least three central benzodiazepine binding sites in the spinal cord.

Molecular models for recognition and activation at the benzodiazepine receptor: a review.

In this minireview an overview description of the different pharmacophores that have been proposed for the benzodiazepine receptor is presented. The methodology as well as the experimental information used in the development of each model will be described. Finally, the new experimental discoveries that are likely to affect the models presented are briefly indicated.

Comments on the design of chemical libraries for screening.

Different representations of molecules, based on distinct sets of properties can yield different perspectives of the issues involved in library design. In particular, different chemical representations can give rise to very different estimates of required library sizes. We provide a preliminary mathematical framework that examines the size of libraries required to adequately sample the spaces corresponding to some commonly used property sets. Introduction of conformational flexibility is also discussed as a means of increasing coverage of chemical libraries, while at the same time considering the thermodynamic consequences of flexibility upon detectable activity. Our theoretical analysis reveals that the property spaces currently in use are extremely large and unlikely to provide adequate discrimination among compounds.

Discriminating D1 and D2 agonists with a hydrophobic similarity index.

Currently, methods for calculating molecular similarity indices have been developed for comparing steric, charge density, and molecular electrostatic potential (MEP) properties. Much of the existing technology may, however, be applied to the quantitative comparison of molecular hydrophobicities. In this article we present an empirical hydrophobic similarity index. We utilize atomic hydrophobic parameters derived from a quantum mechanical semiempirical wavefunction. Hydrophobicity at points on a grid is computed with a recently introduced "molecular lipophilicity potential." The overlap of pairs of molecules is calculated with the metric introduced by Carbó. This approach is applied to a case in which steric and electrostatic criteria have already been shown to be inadequate in rationalizing selectivity, namely, requirements for recognition at the dopamine D1 and D2 receptors. We demonstrate that, for a set of dopamine agonists, D1 ligands show higher similarity in this property that D2 analogs. This indicator of similarity is more successful at accounting for D1 selectivity than previous methods.

Amino acid preferences of small, naturally occurring polypeptides.

An analysis of amino acid composition of small, naturally occurring peptides ranging in size from 3 to 50 residues has been carried out. The purpose of the study is to determine whether differential trends in amino acid usage exist for small peptides compared to larger polypeptides and proteins. Results indicate that Cys, Trp, and Phe are substantially more frequent in peptides compared to their abundance in proteins at large. Aliphatic hydrophobic residues, particularly Leu and Ile, are somewhat underrepresented, while the frequency of Glu is significantly reduced. The shorter peptides are also more frequently neutral and become increasingly charged as their size increases.

Statistical relationships among docking scores for different protein binding sites.

This report describes the existence of statistical relationships among scores computed with the DOCK program for a library of small molecules and a panel of protein binding sites. Multivariate relationships are observed in docking scores computed for a constant set of ligands in different binding sites of proteins that are dissimilar in structure and function. The structural basis for the correlations found among scores is analyzed in terms of size, shape and charge characteristics of the binding sites considered. Interestingly, these results parallel a growing body of evidence demonstrating the promiscuous behavior of small molecules in their interactions with macromolecules that could have an impact in future efforts in drug design.

Investigation of classification methods for the prediction of activity in diverse chemical libraries.

Classification methods based on linear discriminant analysis, recursive partitioning, and hierarchical agglomerative clustering are examined for their ability to separate active and inactive compounds in a diverse chemical database. Topology-based descriptions of chemical structure from the Molconn-X and ISIS programs are used in conjunction with these classification techniques to identify ACE inhibitors, beta-adrenergic antagonists, and H2 receptor antagonists. Overall, discriminant analysis misclassifies the smallest number of active compounds, while recursive partitioning yields the lowest rate of misclassification among inactives. Binary structural keys from the ISIS package are found to generally outperform the whole-molecule Molconn-X descriptors, especially for identification of inactive compounds. For all targets and classification methods, sensitivity toward active compounds is increased by making repetitive classification using training sets that contain equal numbers of actives and inactives. These balanced training sets provide an average numerical class membership score which may be used to select subsets of compounds that are enriched in actives.

Bioactive diversity and screening library selection via affinity fingerprinting.

The Similarity Principle provides the conceptual framework behind most modern approaches to library sampling and design. However, it is often the case that compounds which appear to be very similar structurally may in fact exhibit quite different activities toward a given target. Conversely, some targets recognize a wide variety of molecules and thus bind compounds that have markedly different structures. Affinity fingerprints largely overcome the difficulties associated with selecting compounds on the basis of structure alone. By describing each compound in terms of its binding affinity to a set of functionally dissimilar proteins, fundamental factors relevant to binding and biological activity are automatically encoded. We demonstrate how affinity fingerprints may be used in conjunction with simple algorithms to select active-enriched diverse training sets and to efficiently extract the most active compounds from a large library.

Molecular determinants of non-specific recognition of delta, mu, and kappa opioid receptors.

Identification of the molecular determinants of recognition common to all three opioid receptors embedded in a single three-dimensional (3D) non-specific recognition pharmacophore has been carried out. The working hypothesis that underlies the computational study reported here is that ligands that bind with significant affinity to all three cloned opioid receptors, delta, mu, and kappa, but with different combinations of activation and inhibition properties at these receptors, could be promising behaviorally selective analgesics with diminished side effects. The study presented here represents the first step towards the rational design of such therapeutic agents. The common 3D pharmacophore developed for recognition of delta, mu, and kappa opioid receptors was based on the receptor affinities determined for 23 different opioid ligands that display no specificity for any of the receptor subtypes. The pharmacophore centers identified are a protonated amine, two hydrophobic groups, and the centroid of an aromatic group in a geometric arrangement common to all 23, non-specific, opioid ligands studied. Using this three-dimensional pharmacophore as a query for searching 3D structural databases, novel compounds potentially involved in non-specific recognition of delta, mu, and kappa opioid receptors were retrieved. These compounds can be valuable candidates for novel behaviorally selective analgesics with diminished or no side effects, and thus with potential therapeutic usefulness.

Strategies for indirect computer-aided drug design.

This review is intended to describe some of the methods and procedures used for computer-aided drug design when the structure of the macromolecular target is unknown, as is the case for CNS active drugs. Strategies and methods used in computer-aided design of drugs in such instances must be "indirect," i.e., focusing on the characterization of the ligands themselves. This situation is different from one in which the three-dimensional structure of the macromolecular target for a drug is known, for example, for drugs that are enzyme inhibitors, allowing "direct" characterization of ligand-receptor interactions. Two qualitatively different "indirect" approaches are described here. One, called 2D-QSAR, is briefly reviewed. It is based on delineating regression relationships between a specified biological end point and properties of the compounds eliciting it. The other, based on pharmacophore development, constitutes the main part of this review. Several levels of pharmacophore development are described, which differ in the extent to which they encompass fundamental molecular properties that are determinants of receptor recognition and activation. The strengths and limitations of each procedure are discussed and illustrated by examples. Two methods for obtaining model receptor structures are then briefly described. Both rely on the prior success of the indirect methods in obtaining ligand properties that modulate receptor recognition and activation. These emerging capabilities have the potential to bridge the gap between indirect and direct methods of drug design, since, if successful, the design process can continue in a direct mode using explicit characterization of drug-receptor interactions. Strategies for hypothesis validation and use of hypothesis for drug design and discovery are also briefly reviewed. The final sections of this review describe specific computational tools such as molecular mechanics and quantum mechanical methods used to characterize and identify relevant molecular properties and indicate some areas for future development of computational chemistry methods that could increase its effectiveness in the design of novel drugs.