Protocols for bridging the peptide to nonpeptide gap in topological similarity searches.

Similarity searches based on chemical descriptors have proven extremely useful in aiding large-scale drug screening. Typically an investigator starts with a "probe", a drug-like molecule with an interesting biological activity, and searches a database to find similar compounds. In some projects, however, the only known actives are peptides, and the investigator needs to identify drug-like actives. 3D similarity methods are able to help in this endeavor but suffer from the necessity of having to specify the active conformation of the probe, something that is not always possible at the beginning of a project. Also, 3D methods are slow and are complicated by the need to generate low-energy conformations. In contrast, topological methods are relatively rapid and do not depend on conformation. However, unmodified topological similarity methods, given a peptide probe, will preferentially select other peptides from a database. In this paper we show some simple protocols that, if used with a standard topological similarity search method, are sufficient to select nonpeptide actives given a peptide probe. We demonstrate these protocols by using 10 peptide-like probes to select appropriate nonpeptide actives from the MDDR database.

Mining the chemical quarry with joint chemical probes: an application of latent semantic structure indexing (LaSSI) and TOPOSIM (Dice) to chemical database mining.

In this study we use a novel similarity search technique called latent semantic structure indexing (LaSSI) with joint chemical probes as queries to mine the MDL drug data report database. LaSSI is based on latent semantic indexing developed for searching textual databases. We use atom pair and topological torsion descriptors in our calculations. The results obtained with LaSSI are compared with another in-house similarity search technique TOPOSIM. The results from the similarity searches using joint chemical probes are significantly better than searches using single chemical probes for both LaSSI and TOPOSIM. The selected molecules are closely related in activity to their queries and are ranked among the top 300 scoring molecules of the 82 860 entries in the database. Our implementation of LaSSI is very fast and efficient in finding active compounds. The results also show that LaSSI consistently retrieves more diverse chemical structures representative of the joint chemical probes in comparison to TOPOSIM. The use of multimolecule topological probes to identify compounds complements the use of searching databases with 3D pharmacophore hypotheses.

Latent semantic structure indexing (LaSSI) for defining chemical similarity.

A novel method for computing chemical similarity from chemical substructure descriptors is described. This new method, called LaSSI, uses the singular value decomposition (SVD) of a chemical descriptor-molecule matrix to create a low-dimensional representation of the original descriptor space. Ranking molecules by similarity to a probe molecule in the reduced-dimensional space has several advantages over analogous ranking in the original descriptor space: matching latent structures is more robust than matching discrete descriptors, choosing the number of singular values provides a rational way to vary the "fuzziness" of the search, and the reduction in the dimensionality of the chemical space increases searching speed. LaSSI also allows the calculation of the similarity between two descriptors and between a descriptor and a molecule.

Chemical similarity searches using latent semantic structural indexing (LaSSI) and comparison to TOPOSIM.

Similarity searches based on chemical descriptors have proven extremely useful in aiding large-scale drug screening. Here we present results of similarity searching using Latent Semantic Structure Indexing (LaSSI). LaSSI uses a singular value decomposition on chemical descriptors to project molecules into a k-dimensional descriptor space, where k is the number of retained singular values. The effect of the projection is that certain descriptors are emphasized over others and some descriptors may count as partially equivalent to others. We compare LaSSI searches to searches done with TOPOSIM, our standard in-house method, which uses the Dice similarity definition. Standard descriptor-based methods such as TOPOSIM count all descriptors equally and treat all descriptors as independent. For this work we use atom pairs and topological torsions as examples of chemical descriptors. Using objective criteria to determine how effective one similarity method is versus another in selecting active compounds from a large database, we find for a series of 16 drug-like probes that LaSSI is as good as or better than TOPOSIM in selecting active compounds from the MDDR database, if the user is allowed to treat k as an adjustable parameter. Typically, LaSSI selects very different sets of actives than does TOPOSIM, so it can find classes of actives that TOPOSIM would miss.

The centroid approximation for mixtures: calculating similarity and deriving structure--activity relationships.

Compounds are often synthesized and tested as mixtures. We propose the idea that the descriptor representation of a mixture may be approximated as the descriptor average of its individual component molecules. This centroid approximation has several potential advantages: the representation is very compact, calculating similarities and deriving structure-activity relationships (SARs) of mixtures involves very little computation, and existing software can be directly applied to mixtures as if they were single molecules. Here we use the atom pair and topological torsion descriptors. We run several types of simulations using mixtures composed of druglike molecules from the MDL Drug Data Report database. We show that similarity searches using mixtures as queries and/or database entries yield reasonable results, with the caveat that a correction is necessary for mixture-mixture comparisons where at least one of the mixtures contains very diverse molecules. We also show that predictive SARs in the form of trend vectors can be derived from mixtures.

Designing targeted libraries with genetic algorithms.

In combinatorial synthesis, molecules are assembled by linking chemically similar fragments. Because the number of available chemical fragments often greatly exceeds the number that can be used in one synthetic experiment, one needs a rational method for choosing a subset of desirable fragments. If a combinatorial library is to be targeted against a particular biological activity, virtual screening methods can be used to predict which molecules in a virtual library are most likely to be active. When the number of possible molecules in a virtual library is very large, genetic algorithms (GAs) or simulated annealing can be used to quickly find high-scoring molecules by sampling a small subset of the total combinatorial space. We previously demonstrated how a GA can be used to select a subset of fragments for a combinatorial library, and we used topology-based methods of scoring. Here we extend that earlier work in three ways. (1) We demonstrate use of the GA with 3D scoring methods developed in our laboratory. (2) We show that the approach of assembling libraries from fragments in high-scoring molecules is a reasonable one. (3) We compare results from a library-based GA to those from a molecule-based GA.

SQ: a program for rapidly producing pharmacophorically relevent molecular superpositions.

A new method SQ has been developed to provide fast, automatic, and objective pairwise three-dimensional molecular alignments. SQ uses an atom-based clique-matching step followed by an alignment scoring function that has been parametrized to recognize pharmacologically relevant atomic properties. Molecular alignments from SQ are consistent with known drug-receptor interactions. We demonstrate this with six pairs of receptor-ligand complexes from the Brookhaven Protein Data Bank. The SQ-generated alignment of one isolated ligand onto another is shown to approximate the alignment of the ligands when the receptors are superimposed. SQ appears to be better than its predecessor SEAL (Kearsley and Smith, Tetrahedron Comput. Methodol. 1990, 3, 615-633) in this regard. SQ has been tailored so that, given one molecule as a probe, it can be used to routinely scan large chemical databases for which precomputed conformations have been stored. The SQ score, a measure of 3D similarity of each candidate molecule to the probe, can be used to rank compounds for the purposes of chemical screening. We demonstrate this with three probes (a thrombin inhibitor, an HIV protease inhibitor, and a model for angiotensin II). In each case SQ can preferentially select from the MDDR database other compounds with the same activity as the probe. We further show, using the angiotensin example, how SQ can identify topologically diverse compounds with the same activity.

Sesamoid bone interposition complicating reduction of a hallux joint dislocation.

This is the first reported case of sesamoid bone interposition in the interphalangeal joint of the hallux as a complication of closed reduction of a dislocated interphalangeal joint of the hallux. The case also highlights the importance of post-reduction radiographs.

Flexibases: a way to enhance the use of molecular docking methods.

Specially expanded databases containing three-dimensional structures are created to enhance the utility of docking methods to find new leads, i.e., active compounds of pharmacological interest. The expansion is based on the automatic generation of a set of maximally dissimilar conformations. The ligand receptor system of methotrexate and dihydrofolate reductase is used to demonstrate the feasibility of creating flexibases and their utility in docking studies.

Extending the trend vector: the trend matrix and sample-based partial least squares.

Trend vector analysis [Carhart, R.E. et al., J. Chem. Inf. Comput. Sci., 25 (1985) 64], in combination with topological descriptors such as atom pairs, has proved useful in drug discovery for ranking large collections of chemical compounds in order of predicted biological activity. The compounds with the highest predicted activities, upon being tested, often show a several-fold increase in the fraction of active compounds relative to a randomly selected set. A trend vector is simply the one-dimensional array of correlations between the biological activity of interest and a set of properties or 'descriptors' of compounds in a training set. This paper examines two methods for generalizing the trend vector to improve the predicted rank order. The trend matrix method finds the correlations between the residuals and the simultaneous occurrence of descriptors, which are stored in a two-dimensional analog of the trend vector. The SAMPLS method derives a linear model by partial least squares (PLS), using the 'sample-based' formulation of PLS [Bush, B.L. and Nachbar, R.B., J. Comput.-Aided Mol. Design, 7 (1993) 587] for efficiency in treating the large number of descriptors. PLS accumulates a predictive model as a sum of linear components. Expressed as a vector of prediction coefficients on properties, the first PLS component is proportional to the trend vector. Subsequent components adjust the model toward full least squares. For both methods the residuals decrease, while the risk of overfitting the training set increases. We therefore also describe statistical checks to prevent overfitting. These methods are applied to two data sets, a small homologous series of disubstituted piperidines, tested on the dopamine receptor, and a large set of diverse chemical structures, some of which are active at the muscarinic receptor. Each data set is split into a training set and a test set, and the activities in the test set are predicted from a fit on the training set. Both the trend matrix and the SAMPLS approach improve the predictions over the simple trend vector. The SAMPLS approach is superior to the trend matrix in that it requires much less storage and CPU time. It also provides a useful set of axes for visualizing properties of the compounds. We describe a randomization method to determine the optimum number of PLS components that is very much faster for large training sets than leave-one-out cross-validation.

FLOG: a system to select 'quasi-flexible' ligands complementary to a receptor of known three-dimensional structure.

We present a system, FLOG (Flexible Ligands Oriented on Grid), that searches a database of 3D coordinates to find molecules complementary to a macromolecular receptor of known 3D structure. The philosophy of FLOG is similar to that reported for DOCK [Shoichet, B.K. et al., J. Comput. Chem., 13 (1992) 380]. In common with that system, we use a match center representation of the volume of the binding cavity and we use a clique-finding algorithm to generate trial orientations of each candidate ligand in the binding site. Also we use a grid representation of the receptor to assess the fit of each orientation. We have introduced a number of novel features within this paradigm. First, we address ligand flexibility by including up to 25 explicit conformations of each structure in our databases. Nonhydrogen atoms in each database entry are assigned one of seven atom types (anion, cation, donor, acceptor, polar, hydrophobic and other) based on their local bonded chemical environments. Second, we have devised a new grid-based scoring function compatible with this 'heavy atom' representation of the ligands. This includes several potentials (electrostatic, hydrogen bonding, hydrophobic and van der Waals) calculated from the location of the receptor atoms. Third, we have improved the fitting stage of the search. Initial dockings are generated with a more efficient clique-finding algorithm. This new algorithm includes the concept of 'essential points', match centers that must be paired with a ligand atom. Also, we introduce the use of a rapid simplex-based rigid-body optimizer to refine the orientations. We demonstrate, using dihydrofolate reductase as a sample receptor, that the FLOG system can select known inhibitors from a large database of drug-like compounds.

A systematic search for protein signature sequences.

Signature sequences are contiguous patterns of amino acids 10-50 residues long that are associated with a particular structure or function in proteins. These may be of three types (by our nomenclature): superfamily signatures, remnant homologies, and motifs. We have performed a systematic search through a database of protein sequences to automatically and preferentially find remnant homologies and motifs. This was accomplished in three steps: 1. We generated a nonredundant sequence database. 2. We used BLAST3 (Altschul and Lipman, Proc. Natl. Acad. Sci. U.S.A. 87:5509-5513, 1990) to generate local pairwise and triplet sequence alignments for every protein in the database vs. every other. 3. We selected "interesting" alignments and grouped them into clusters. We find that most of the clusters contain segments from proteins which share a common structure or function. Many of them correspond to signatures previously noted in the literature. We discuss three previously recognized motifs in detail (FAD/NAD-binding, ATP/GTP-binding, and cytochrome b5-like domains) to demonstrate how the alignments generated by our procedure are consistent with previous work and make structural and functional sense. We also discuss two signatures (for N-acetyltransferases and glycerol-phosphate binding) which to our knowledge have not been previously recognized.

Origin of tetracycline efflux proteins: conclusions from nucleotide sequence analysis.

The sequences of six tetracycline efflux proteins and three transport proteins which have some resemblance to them were compared. The tetracycline efflux proteins fall into three families: (i) those encoded by pBR322, RP1, and Tn10 (Escherichia coli); (ii) pT181 (Staphylococcus aureus) and pTHT15 (Bacillus subtilis); and (iii) tet347 (Streptomyces rimosus). There is global sequence homology within each of the first two families, but there is none between the families. The pT181/pTHT15 family shares close homology with the N-terminal half of the methylenomycin A efflux protein (Streptomyces coelicor), while tet347 resembles the C-terminal half. Portions of the N-terminal half of the Tn10-encoded protein show significant resemblance to portions in the N-terminal half of the pT181/pTHT15 family, but this sometimes occurs among transport proteins which do not have a common substrate. Tetracycline efflux proteins, therefore, appear to have arisen on at least two, or possibly three, separate occasions, probably from other transport proteins.

Role of lysine-54 in determining cofactor specificity and binding in human dihydrofolate reductase.

Lysine-54 of human dihydrofolate reductase (hDHFR) appears to be involved in the interaction with the 2'-phosphate of NADPH and is conserved as a basic residue in other species. Studies have suggested that in Lactobacillus casei dihydrofolate reductase Arg-43, the homologous residue at this position, plays an important role in the binding of NADPH and in the differentiation of Km values for NADPH and NADH. A Lys-54 to Gln-54 mutant (K54Q) of hDHFR has been constructed by oligodeoxynucleotide-directed mutagenesis in order to study the role of Lys-54 in differentiating Km and Kcat values for NADPH and NADH as well as in other functions of hDHFR. The purpose of this paper is to delineate in quantitative terms the magnitude of the effect of the Lys-54 to Gln-54 replacement on the various kinetic parameters of hDHFR. Such quantitative effects cannot be predicted solely on the basis of X-ray structures. The Km for NADPH for the K54Q mutant enzyme is 58-fold higher, while the Km for NADH for K54Q is only 3.9-fold higher than that of the wild type, indicating that the substitution of Lys-54 with Gln-54 decreases the apparent affinity of the enzyme for NADPH dramatically, but has a lesser effect on the apparent affinity for NADH.(ABSTRACT TRUNCATED AT 250 WORDS)

Searching for pharmacophores in large coordinate data bases and its use in drug design.

Pharmacophores, three-dimensional arrangements of chemical groups essential for biological activity, are being proposed in increasing numbers. We have developed a system to search data bases of three-dimensional coordinates for compounds that contain a particular pharmacophore. The coordinates can be derived from experiment (e.g., Cambridge Crystal Database) or be generated from data bases of connection tables (e.g., Cyanamid Laboratories proprietary compounds) via the program CONCORD. We discuss the results of searches for three sample pharmacophores. Two have been proposed by others based on the conformational analysis of active compounds, and one is inferred from the crystal structure of a protein-ligand complex. These examples show that such searches can identify classes of compounds that are structurally different from the compounds from which the pharmacophore was derived but are known to have the appropriate biological activity. Occasionally, the searches find bond "frameworks" in which the important groups are rigidly held in the proper geometry. These may suggest new structural classes for synthesis.

Using shape complementarity as an initial screen in designing ligands for a receptor binding site of known three-dimensional structure.

Finding novel leads from which to design drug molecules has traditionally been a matter of screening and serendipity. We present a method for finding a wide assortment of chemical structures that are complementary to the shape of a macromoleculer receptor site whose X-ray crystallographic structure is known. Each of a set of small molecules from the Cambridge Crystallographic Database (Allen; et al. J. Chem. Doc. 1973, 13, 119) is individually docked to the receptor in a number of geometrically permissible orientations with use of the docking algorithm developed by Kuntz et al. (J. Mol. Biol. 1982, 161, 269). The orientations are evaluated for goodness-of-fit, and the best are kept for further examination using the molecular mechanics program AMBER (Weiner; Kollman J. Comput. Chem. 1981, 106, 765). The shape-search algorithm finds known ligands as well as novel molecules that fit the binding site being studied. The highest scoring orientations of known ligands resemble binding modes generated by interactive modeling or determined crystallographically. We describe the application of this procedure to the binding sites of papain and carbonic anhydrase. While the compounds recovered from the Cambridge Crystallographic Database are not, themselves, likely to be inhibitors or substrates of these enzymes, we expect that the structures from such searches will be useful in the design of active compounds.

Designing novel nicotinic agonists by searching a database of molecular shapes.

We introduce an approach by which novel ligands can be designed for a receptor if a pharmacophore geometry has been established and the receptor-bound conformations of other ligands are known. We use the shape-matching method of Kuntz et al. [J. Mol. Biol., 161 (1982) 269-288] to search a database of molecular shapes for those molecules which can fit inside the combined volume of the known ligands and which have interatomic distances compatible with the pharmacophore geometry. Some of these molecules are then modified by interactive modeling techniques to better match the chemical properties of the known ligands. Our shape database (about 5000 candidate molecules) is derived from a subset of the Cambridge Crystallographic Database [Allen et al., Acta Crystallogr., Sect. B,35 (1979) 2331-2339]. We show, as an example, how several novel designs for nicotinic agonists can be derived by this approach, given a pharmacophore model derived from known agonists [Sheridan et al., J. Med. Chem., 29 (1986) 889-906]. This report complements our previous report [DesJarlais et al., J. Med. Chem., in press], which introduced a similar method for designing ligands when the structure of the receptor is known.

Docking flexible ligands to macromolecular receptors by molecular shape.

We present a method to explore the interaction of flexible ligands with receptors of known geometry on the basis of molecular shape. This method is an extension of that described by Kuntz et al. (J. Mol. Biol. 1982, 161, 269). The shape of a binding site on a macromolecular receptor is represented as a set of overlapping spheres. Each ligand is divided into a small set of large rigid fragments that are docked separately into the binding site and then rejoined later in the calculation. The division of ligands into separate fragments allows a degree of flexibility at the position that joins them. The rejoined fragments are then energy minimized in the receptor site. We illustrate the method with two test cases: dihydrofolate reductase/methotrexate and prealbumin/thyroxine. For each test case, the method finds binding geometries for the ligand near that observed crystallographically as well as others that provide good steric fit with the receptor.

The ensemble approach to distance geometry: application to the nicotinic pharmacophore.

We develop an extension of conventional distance geometry techniques that treats two or more molecules as a single "ensemble". This extension can be used to find a common pharmacophore, i.e., the spatial arrangement of essential groups, from a small set of biologically active molecules. The approach can generate, in one step, coordinates for the set of molecules in their "active" conformations such that their essential groups are superimposed. As an example, we show how the nicotinic pharmacophore can be deduced from a set of four nicotinic agonists: nicotine, cytisine, ferruginine methiodide, and muscarone. Three essential groups in each agonist are chosen: the cationic center (A), an electronegative atom (B), and an atom (C) that forms a dipole with B. There is only one pharmacophore possible for the superposition of these essential groups: a triangle with sides 4.8 A (A-B), 4.0 A (A-C), and 1.2 A (B-C). The pharmacophore triangle, which is consistent with previous models in the literature, can also be achieved by the agonist trans-3,3'-bis[(trimethylammonio)methyl]azobenzene and the antagonists strychnine, trimethaphan, and dihydro-beta-erythroidine. An examination of the common volumes of agonists suggests a specific disposition of molecular volume relative to the pharmacophore triangle. We discuss the relative strengths and drawbacks of the ensemble approach vs. other conformational search methods.