Substrate specificity of human matriptase-2.

Human matriptase-2 is an enzyme that belongs to the family of type II transmembrane serine proteases. So far there is a limited knowledge regarding its specificity and protein substrate(s). One of the identified natural substrates is hemojuvelin, a protein involved in the control of iron homeostasis. In this work, we describe the synthesis and evaluation of internal quenched substrates using a combinatorial approach. The iterative deconvolution of two libraries to define the specificity of matriptase-2 yielded to the identification of the substrate ABZ-Ile-Arg-Ala-Arg-Ser-Ala-Gly-Tyr(3-NO2)-NH2 with a k(cat)/K(m) value of 4.5 × 10(5) M(-1) × s(-1), i.e. the highest specificity constant reported so far for matriptase-2.

Activity Landscapes, Information Theory, and Structure - Activity Relationships.

Activity landscapes provide a comprehensive description of structure-activity relationships (SARs). An information theoretic assessment of their features, namely, activity cliffs, similarity cliffs, smooth-SAR, and featureless regions, is presented based on the probability of occurrence of these features. It is shown that activity cliffs provide highly informative SARs compared to smooth-SAR regions, although the latter are the basis for most QSAR studies. This follows since small structural changes in the former are coupled with relatively large changes in activity, thus pinpointing specific structural features associated with the changes in activity. In contrast, Smooth-SAR regions are typically associated with relatively small changes in both structure and activity. Surprisingly, similarity cliffs, which occur when both compounds in a compound-pair have approximately equal activities but significantly different structures, are the most prevalent feature of activity landscapes. Hence, from an information theoretic point of view, they are the least informative landscape feature. Nevertheless, similarity cliffs do provide SAR information on potentially new active compound classes, and in that sense they are quite useful in drug discovery programs since they provide alternative possibilities should ADMET or other issues arise during the discovery and earlier preclinical development phases of drug research.

Computational techniques for diversity analysis and compound classification.

Molecular similarity and diversity analysis has played a significant role in computer-aided drug discovery for more than a decade. Compound classification methods have also become increasingly important for the design and organization of compound databases and in silico screening. Here we review these related methodologies and discuss selected applications.

Mini-fingerprints for virtual screening: design principles and generation of novel prototypes based on information theory.

Binary fingerprint representations of molecular structure and properties are convenient computational tools for similarity searching in compound databases and virtual screening (VS). We are investigating the design of relatively simple fingerprints for the identification of molecules having similar biological activity and recognition of remote similarity relationships. Since our designs are considerably shorter than other fingerprints used in VS, we have previously termed them "mini-fingerprints" (MFPs). A key aspect of the design strategy is the identification of suitable molecular descriptors. Whereas our initial fingerprint designs have relied on descriptor combinations that performed well in compound classification according to biological activity, second generation MFPs encode combinations of descriptors with high information content in large compound databases and high frequency of occurrence in drug-like molecules. Thus, the design of these new fingerprints does not depend on the analysis of specific classes of bioactive compounds, but rather on descriptor information content in large compound databases. Systematic evaluation of fingerprint performance in VS test calculations demonstrates that these new prototypes perform better than previously generated MFPs. The analysis described herein provides an example for the development of search tools for VS.

Rational drug discovery revisited: interfacing experimental programs with bio- and chemo-informatics.

Over the past few years, bio- and chemo-informatics have rapidly evolved as related yet distinct disciplines. In drug discovery, it is increasingly recognized that combining and integrating these approaches is crucial for their successful application. In addition, the use of complementary experimental and informatics techniques increases the chances of success in many stages of the discovery process, from the identification of novel targets and elucidation of their functions to the discovery and development of lead compounds with desired properties. This review highlights recent trends that emphasize the role of integrated bio- and chemo-informatics research in drug discovery and discusses representative concepts and methodologies.

Differential Shannon Entropy as a sensitive measure of differences in database variability of molecular descriptors.

A method termed Differential Shannon Entropy (DSE) is introduced to compare differences in information content and variance of molecular descriptors between compound databases. The analysis is based on histograms recording the individual and grouped distributions of molecular descriptors and calculation of Shannon entropy (SE), a formalism originally applied to digital communication. We have recently shown that SE values reflect the nonparametric variability of descriptor settings. Now the analysis has been advanced to assess differences in information content of 143 molecular descriptors in databases containing synthetic compounds, natural products, or drug-like molecules. The DSE metric captures the degree to which descriptor distributions complement or duplicate information contained in molecular databases. In our analysis, we observe significant differences for a number of descriptors and rank them according to their associated DSE values. Using DSE calculations, relative information content of different types of descriptors can be quantified, even if differences are subtle.

Fingerprint scaling increases the probability of identifying molecules with similar activity in virtual screening calculations.

Results of systematic virtual screening calculations using a structural key-type fingerprint are reported for compounds belonging to 14 activity classes added to randomly selected synthetic molecules. For each class, a fingerprint profile was calculated to monitor the relative occupancy of fingerprint bit positions. Consensus bit patterns were determined consisting of all bits that were always set on in compounds belonging to a specific activity class. In virtual screening calculations, scale factors were applied to each consensus bit position in fingerprints of query molecules. This technique, called "fingerprint scaling", effectively increases the weight of consensus bit positions in fingerprint comparisons. Although overall prediction accuracy was satisfactory using unscaled calculations, scaling significantly increased the number of correct predictions but only slightly increased the rate of false positives. These observations suggest that fingerprint scaling is an attractive approach to increase the probability of identifying molecules with similar activity by virtual screening. It requires the availability of a series of related compounds and can be easily applied to any keyed fingerprint representation that associates bit positions with specific molecular features.

Mini-fingerprints detect similar activity of receptor ligands previously recognized only by three-dimensional pharmacophore-based methods.

Mini-fingerprints (MFPs) are short binary bit string representations of molecular structure and properties, composed of few selected two-dimensional (2D) descriptors and a number of structural keys. MFPs were specifically designed to recognize compounds with similar activity. Here we report that MFPs are capable of detecting similar activities of some druglike molecules, including endothelin A antagonists and alpha(1)-adrenergic receptor ligands, the recognition of which was previously thought to depend on the use of multiple point three-dimensional (3D) pharmacophore methods. Thus, in these cases, MFPs and pharmacophore fingerprints produce similar results, although they define, in terms of their complexity, opposite ends of the spectrum of methods currently used to study molecular similarity or diversity. For each of the studied compound classes, comparison of MFP bit settings identified a consensus or signature pattern. Scaling factors can be applied to these bits in order to increase the probability of finding compounds with similar activity by virtual screening.

Molecular descriptors in chemoinformatics, computational combinatorial chemistry, and virtual screening.

Many contemporary applications in computer-aided drug discovery and chemoinformatics depend on representations of molecules by descriptors that capture their structural characteristics and properties. Such applications include, among others, diversity analysis, library design, and virtual screening. Hundreds of molecular descriptors have been reported in the literature, ranging from simple bulk properties to elaborate three-dimensional formulations and complex molecular fingerprints, which sometimes consist of thousands of bit positions. Knowledge-based selection of descriptors that are suitable for specific applications is an important task in chemoinformatics research. If descriptors are to be selected on rational grounds, rather than guesses or chemical intuition, detailed evaluation of their performance is required. A number of studies have been reported that investigate the performance of molecular descriptors in specific applications and/or introduce novel types of descriptors. Progress made in this area is reviewed here in the context of other computational developments in combinatorial chemistry and compound screening.

Evaluation of descriptors and mini-fingerprints for the identification of molecules with similar activity.

Combinations of 65 preferred 1D/2D molecular descriptors and 143 single structural keys were evaluated for their performance in compound classification focused on biological activity. The analysis was based on principal component analysis of descriptor combinations and facilitated by use of a genetic algorithm and different scoring functions. In these calculations, several descriptor combinations with greater than 95% prediction accuracy were identified. A set of 40 preferred structural keys was incorporated into a small binary fingerprint designed to search databases for compounds with biological activity similar to query molecules. The performance of mini-fingerprints was tested by systematic similarity search calculations in a database consisting of compounds belonging to seven biological activity classes, which had not been used to select effective descriptors. In these blind test calculations, mini-fingerprints correctly identified approximately 54% of compounds sharing similar biological activity and with 1% false positives. Thus, although the design of mini-fingerprints is conceptually simple, they perform well in activity-oriented similarity searching.

Distinguishing between natural products and synthetic molecules by descriptor Shannon entropy analysis and binary QSAR calculations.

Molecular descriptors were identified by Shannon entropy analysis that correctly distinguished, in binary QSAR calculations, between naturally occurring molecules and synthetic compounds. The Shannon entropy concept was first used in digital communication theory and has only very recently been applied to descriptor analysis. Binary QSAR methodology was originally developed to correlate structural features and properties of compounds with a binary formulation of biological activity (i.e., active or inactive) and has here been adapted to correlate molecular features with chemical source (i.e., natural or synthetic). We have identified a number of molecular descriptors with significantly different Shannon entropy and/or "entropic separation" in natural and synthetic compound databases. Different combinations of such descriptors and variably distributed structural keys were applied to learning sets consisting of natural and synthetic molecules and used to derive predictive binary QSAR models. These models were then applied to predict the source of compounds in different test sets consisting of randomly collected natural and synthetic molecules, or, alternatively, sets of natural and synthetic molecules with specific biological activities. On average, greater than 80% prediction accuracy was achieved with our best models. For the test case consisting of molecules with specific activities, greater than 90% accuracy was achieved. From our analysis, some chemical features were identified that systematically differ in many naturally occurring versus synthetic molecules.

Searching for molecules with similar biological activity: analysis by fingerprint profiling.

We have recently developed a mini-fingerprint (MFP) representation for small molecules that performs well in database searches for compounds with similar biological activity. The MFP consists of only 54 bit positions that account for numerical ranges of three two-dimensional (2D) descriptors or the presence or absence of defined structural fragments. Here we present an analysis method, termed fingerprint profiling, to systematically compare bit patterns of compounds belonging to different biological activity classes. Some but not all bit positions were variably occupied in seven different activity classes and responsible for the detection of structure-activity differences. The analysis has made it possible to rank bit positions and encoded molecular descriptors according to their importance for our similarity search calculations. Fingerprint profiling can be applied to any keyed bit string representation and should be helpful, for example, to analyze descriptor distributions in large compound databases.

Cell surface receptors and their ligands: in vitro analysis of CD6-CD166 interactions.

CD6 is a cell surface receptor belonging to the scavenger receptor cysteine-rich (SRCR) protein superfamily (SRCRSF). It specifically binds activated leukocyte cell adhesion molecule (ALCAM, CD166), a member of the immunoglobulin (Ig) superfamily (IgSF). CD166 was among the first molecules identified as a ligand for an SRCRSF receptor, and the CD6-CD166 interaction was the first interaction characterized involving SRCRSF and IgSF proteins. We focus here on what has been learned about the specifics of the CD6-CD166 interaction from in vitro analysis. The studies are thought to provide an instructive example for the analysis of interactions between single-path transmembrane cell surface proteins. Using soluble recombinant forms, the extracellular binding domains of receptor and ligand have been identified and characterized in a variety of assay systems. Both CD6 and CD166 have been subjected to intense mutagenesis and monoclonal antibody (mAb) binding studies and residues critical for their interaction have been identified. The availability of structural prototypes of both superfamilies has made it possible to map the binding site in CD166 and, more recently, in CD6 and compare these regions to epitopes of mAbs that block, or do not block, the interaction. In addition, the molecular basis of observed cross-species receptor-ligand interactions could be rationalized. These studies illustrate the value of structural templates for the interpretation of sequence and mutagenesis analyses. Proteins 2000;40:420-428.

Variability of molecular descriptors in compound databases revealed by Shannon entropy calculations

A method is introduced to calculate and compare the variability of molecular descriptors in compound databases. Descriptor variability analysis is based on histograms recording the distribution of molecular descriptors and calculation of Shannon entropy (SE), a metric originally applied in digital communication. SE values reflect the variability of descriptor settings. We have calculated a total of 92 molecular descriptors in the ACD and NCI databases and ranked them according to their variability. Significant differences in entropy are observed for a number of descriptors. However, the most variable descriptors are similar in the ACD and NCI databases. Such high-entropy descriptors are preferred tools to discriminate between compounds or account for the diversity of chemical libraries.

Molecular descriptors for effective classification of biologically active compounds based on principal component analysis identified by a genetic algorithm.

We have evaluated combinations of 111 descriptors that were calculated from two-dimensional representations of molecules to classify 455 compounds belonging to seven biological activity classes using a method based on principal component analysis. The analysis was facilitated by application of a genetic algorithm. Using scoring functions that related the number of compounds in pure classes (i.e., compounds with the same biological activity), singletons, and mixed classes, effective descriptor sets were identified. A combination of only four molecular descriptors accounting for aromatic character, hydrogen bond acceptors, estimated polar van der Waals surface area, and a single structural key gave overall best results. At this performance level, approximately 91% of the compounds occurred in pure classes and mixed classes were absent. The results indicate that combinations of only a few critical descriptors are preferred to partition compounds according to their biological activity, at least in the test cases studied here.