ChemDB update--full-text search and virtual chemical space

ChemDB is a chemical database containing nearly 5M commercially available small molecules, important for use as synthetic building blocks, probes in systems biology and as leads for the discovery of drugs and other useful compounds. The data is publicly available over the web for download and for targeted searches using a variety of powerful methods. The chemical data includes predicted or experimentally determined physicochemical properties, such as 3D structure, melting temperature and solubility. Recent developments include optimization of chemical structure (and substructure) retrieval algorithms, enabling full database searches in less than a second. A text-based search engine allows efficient searching of compounds based on over 65M annotations from over 150 vendors. When searching for chemicals by name, fuzzy text matching capabilities yield productive results even when the correct spelling of a chemical name is unknown, taking advantage of both systematic and common names. Finally, built in reaction models enable searches through virtual chemical space, consisting of hypothetical products readily synthesizable from the building blocks in ChemDB. AVAILABILITY: ChemDB and Supplementary Materials are available at http://cdb.ics.uci.edu. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.     

Challenges for the prediction of macromolecular interactions

Macromolecular interactions are central to most cellular processes. Experimental methods generate diverse data on these interactions ranging from high throughput protein-protein interactions (PPIs) to the crystallised structures of complexes. Despite this, only a fraction of interactions have been identified and therefore predictive methods are essential to fill in the numerous gaps. Many predictive methods use information from related proteins. Accordingly, we review the conservation of interface and ligand binding sites within protein families and their association with conserved residues and Specificity Determining Positions. We then review recent developments in predictive methods for the identification of PPIs, protein interface sites and small molecule ligand binding sites. The challenges that are still faced by the community in these areas are discussed.     

The determination of equilibrium constants for heterogeneous macromolecular interactions

A method has been developed to determine the association constant for a heterogeneous association of the type A + B in equilibrium AB. This method requires knowledge of the two initial concentrations and of the resulting weight-average molecular weight for each data point. Computer simulations using Gaussian-distributed error on the measured parameters show that the researcher can readily determine whether the particular concentration range chosen is appropriate for the strength of binding and therefore how reliable the calculated constant might be. It is also shown that errors in measuring molecular weight have, in general, a more profound effect than do errors in concentration.     

GTP-mediated macromolecular interactions: the common features of different systems

G proteins that serve to transduce external signals in membranes share with protein synthesis factors and tubulin structural and functional features that are common to proteins that participate in reversible GTP-mediated macromolecular interactions. These proteins can bind GTP and GDP with high affinity, adopting different structures depending on whether they are complexed with the nucleotide diphosphate or triphosphate. The GTP.protein complex has high affinity for an acceptor macromolecule (or complex of macromolecules) and interacts with it, affecting its activity. These GTP-binding proteins also possess an intrinsic GTPase activity that is generally stimulated by its interaction with the acceptor. The GTPase activity converts the bound GTP to GDP, switching the configuration of the complexed protein to one of low affinity for the acceptor and causing its dissociation. The protein.GDP complex must exchange its GDP for GTP to allow the protein to acquire the high-affinity structure that can cycle back to the acceptor macromolecule. This exchange of guanine nucleotides requires in several instances exchange factors that can regulate the whole process. A detailed comparison of the features of the different systems is made with respect to structural similarities, regulation by protein phosphorylation, ADP ribosylation by bacterial toxins, and requirements for exchange factors. It is also proposed that there is a similar mechanism that involves ATP/ADP-binding proteins.     

Atomic force microscopy of macromolecular interactions.

The introduction of functional imaging tools and techniques that operate at molecular-length scales has provided investigators with unique approaches to characterizing biomolecular structure and function relationships. Recent advances in the field of scanning probe techniques and, in particular, atomic force microscopy have yielded tantalizing insights into the dynamics of protein self-assembly and the mechanics of protein unfolding.     

Feature space resampling for protein conformational search

De novo protein structure prediction requires location of the lowest energy state of the polypeptide chain among a vast set of possible conformations. Powerful approaches include conformational space annealing, in which search progressively focuses on the most promising regions of conformational space, and genetic algorithms, in which features of the best conformations thus far identified are recombined. We describe a new approach that combines the strengths of these two approaches. Protein conformations are projected onto a discrete feature space which includes backbone torsion angles, secondary structure, and beta pairings. For each of these there is one “native” value: the one found in the native structure. We begin with a large number of conformations generated in independent Monte Carlo structure prediction trajectories from Rosetta. Native values for each feature are predicted from the frequencies of feature value occurrences and the energy distribution in conformations containing them. A second round of structure prediction trajectories are then guided by the predicted native feature distributions. We show that native features can be predicted at much higher than background rates, and that using the predicted feature distributions improves structure prediction in a benchmark of 28 proteins. The advantages of our approach are that features from many different input structures can be combined simultaneously without producing atomic clashes or otherwise physically inviable models, and that the features being recombined have a relatively high chance of being correct. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

A search for specificity in DNA-drug interactions

The GRID force field and a principal component analysis have been used in order to predict the interactions of small chemical groups with all 64 different triplet sequences of B-DNA. Factors that favor binding to guanine-cytosine base pairs have been identified, and a dictionary of ligand groups and their locations is presented as a guide to the design of specific DNA ligands.     

Natural insights for chemical biologists

Chemical biologists studying natural-product pathways encoded in genomes have unearthed new chemistry and insights into the evolution of biologically active metabolites.     

Exploring buffer space for molecular interactions

The interaction of two molecules binding to each other is described by two rate parameters, the association rate parameter k(a) and the dissociation rate parameter k(d). Under standardized conditions these kinetic parameters can be determined by analysis of their interaction in an affinity-based biosensor system, such as BIACORE(R) 3000. The association rate describes the collision frequency and the attraction between two molecules and the dissociation rate describes the stability of the molecular complex. By comparing the affinity of different molecules (calculated as the quotient of the association rate parameter and the dissociation rate parameter), an estimation of specificity can be obtained. For dissociation, two different aspects can be considered-the practical aspect, where one is interested in separating the two molecules, and the informative aspect, where one is interested in the reasons for the dissociation event. This review focuses on a method that was designed to solve the practical problem of regeneration, but eventually produced a considerable amount of information about the interactions themselves.     

The analysis of macromolecular interactions by sedimentation equilibrium

The study of macromolecular interactions by sedimentation equilibrium is a highly technical method that requires great care in both the experimental design and data analysis. The complexity of the interacting system that can be analyzed is only limited by the ability to deconvolute the exponential contributions of each of the species to the overall concentration gradient. This is achieved in part through the use of multi-signal data collection and the implementation of soft mass conservation. We illustrate the use of these constraints in SEDPHAT through the study of an A+B+B?AB+B?ABB system and highlight some of the technical challenges that arise. We show that both the multi-signal analysis and mass conservation result in a precise and robust data analysis and discuss improvements that can be obtained through the inclusion of data from other methods such as sedimentation velocity and isothermal titration calorimetry.     

Experimental tests of charge conservation in macromolecular interactions

A combination of equilibrium dialysis and ultrafiltration has been used to demonstrate the conservation of charge in the interaction between bovine serum albumin and methyl orange in Tris-HCl buffer, pH 7.4, I = 0.05 M; and also in the dimerization of alpha-chymotrypsin in acetate/chloride buffer, pH 3.9, I = 0.11 M, containing various concentrations of indole (0-10 mM) in order to displace the equilibrium position towards monomer. In the former study the magnitude of the negative charge on the albumin was shown to increase linearly with the number of molecules of methyl orange bound to the protein, the observed slope (0.96 +/- 0.08) of this relationship being in excellent agreement with that predicted on the basis of charge conservation for attachment of the univalent, negatively charged methyl orange ligand. In the study of alpha-chymotrypsin, the net charge (expressed per monomeric enzyme unit) was +10 in solutions in which the mole fraction of monomer varied between 0.47 and 0.88, the extent of this range having been established by means of constituent association equilibrium constants obtained from sedimentation equilibrium studies.     

Prediction and design of macromolecular structures and interactions

In this article, I summarize recent work from my group directed towards developing an improved model of intra and intermolecular interactions and applying this improved model to the prediction and design of macromolecular structures and interactions. Prediction and design applications can be of great biological interest in their own right, and also provide very stringent and objective tests which drive the improvement of the model and increases in fundamental understanding. I emphasize the results from the prediction and design tests that suggest progress is being made in high-resolution modelling, and that there is hope for reliably and accurately computing structural biology.     

Analysis of sequence-reactivity space for protein-protein interactions

Sequence–reactivity space is defined by the relationships between amino acid type choices at some residue positions in a protein and the reactivities of the resulting variants. We are studying Kazal superfamily serine proteinase inhibitors, under substitution of any combination of residue types at 10 binding‐region positions. Reactivities are defined by the standard free energy of association for an inhibitor against an enzyme, and we are interested in both the strength (the free energy value) and specificity (relative free energy values for one inhibitor against different enzymes). Characterizing the structure of such a space poses several interesting questions: (1) How many sequences achieve particular strength and specificity characteristics? (2) What is the best such sequence? (3) What are some nearly‐as‐good alternatives? (4) What are their common residue type characteristics (e.g., conservation and correlation)? Although these problems are all highly combinatorial in nature, this article develops an efficient, integrated mechanism to address them under a data‐driven model that predicts reactivity for given sequences. We employ sampling and a novel deterministic distribution propagation algorithm, in order to determine both the reactivity distribution and sequence composition statistics; integer programming and a novel branch‐and‐bound search algorithm, in order to optimize sequences and enumerate near‐optimal sequences; and correlation‐based sequence decomposition, in order to identify sequence motifs. We demonstrate the value of our mechanism in analyzing the Kazal superfamily sequence–reactivity space, providing insights into the underlying biochemistry and suggesting hypotheses for further experimental consideration. In general, our mechanism offers a valuable tool for investigating the available degrees of freedom in protein design within a combined computational–experimental framework. Proteins 2005. © 2004 Wiley‐Liss, Inc.     

The determination of equilibrium constants for heterogeneous macromolecular interactions. Systems forming 2:1 complexes

In developing a method for analyzing the heterogeneous association nA + mB in equilibrium AnBm, we have specifically investigated the case of n = 2, m = 1 for both the specific case of no appreciable intermediates and the more general case allowing intermediates. Computer-simulated three-dimensional surfaces of the 2:1 model generated from total concentrations of species A and B and the resulting weight-average molecular weights were analyzed with a Gauss-Newton nonlinear least-squares minimization routine. The surfaces generated included normalized random error of varying standard deviations imposed upon both the concentrations and weight-average molecular weights. For comparison purposes, these surfaces were analyzed not only by using the correct 2:1 model, but also by an incorrect (1:1) model and by the other (incorrect) 2:1 model. Except for those situations where the 'experimental' noise was consistently higher than the concentration of one of the species, correct K values were obtained and the correct model was easily distinguished from the incorrect model. The computer routine similarly distinguished between data correctly described as 1:1 and the same data incorrectly analyzed as either 2:1 model. For those cases in which a microscopic Ki value predicts an association such that all species involved for that particular Ki are in appreciable amounts, the Ki value is returned correctly. Correct overall equilibrium constants are also converged upon as long as adequate amounts of A2B, B and A are present.