Search

Cell-cell recognition is the first stage in many important phenomena such as infection by bacteria and viruses, communication among cells of lower eukaryotes, binding of sperm to egg, etc. [1]. Cell-cell recognition relies on sugar (carbohydrate) specific interactions at the cell surface. Theoretical studies typically involve molecular modeling of sugars and sugar-specific protein receptors. These studies rely on structural information obtained mainly by crystallography and nuclear magnetic resonance, and deposited in the Protein Databank (PDB). Since the main purpose of PDB is to store the structure of proteins and nucleic acids, thus, it is expected that PDB structure files are complete and correctly annotated. Nonetheless, sugars exhibit a structural diversity larger than amino acids or nucleotides, a property which makes them ideal for recognition. At the same time, sugars are characterized by specific and very sensitive structural features such as multiple chiral centers on each ring. Because of these peculiarities, the validation and annotation of sugar structures is not straightforward. Our first goal was to develop a methodology that can identify whether a sugar structure is complete and correctly annotated. Our second goal was then to check all PDB entries containing sugars, and record whatever problems we encounter in the sugar structures. For this purpose we collected all sugar structures which appear as ligands in PDB entries, and compared them to model structures available in Ligand Expo [2], a curated repository of ligand chemical and structural information. In order to perform the comparison we used several tools for structural comparison currently available (SiteBinder [3], Open Babel [4]), as well as two in-house programs. We report here on our findings regarding the complete and correctly annotated sugar structures in PDB, together with the problematic cases.

Background Channels and pores in biomacromolecules (proteins, nucleic acids and their complexes) play significant biological roles, e.g., in molecular recognition and enzyme substrate specificity. Results We present an advanced software tool entitled MOLE 2.0, which has been designed to analyze molecular channels and pores. Benchmark tests against other available software tools showed that MOLE 2.0 is by comparison quicker, more robust and more versatile. As a new feature, MOLE 2.0 estimates physicochemical properties of the identified channels, i.e., hydropathy, hydrophobicity, polarity, charge, and mutability. We also assessed the variability in physicochemical properties of eighty X-ray structures of two members of the cytochrome P450 superfamily. Conclusion Estimated physicochemical properties of the identified channels in the selected biomacromolecules corresponded well with the known functions of the respective channels. Thus, the predicted physicochemical properties may provide useful information about the potential functions of identified channels. The MOLE 2.0 software is available at http://mole.chemi.muni.cz.

The acid dissociation constant p K a is a very important molecular property, and there is a strong interest in the development of reliable and fast methods for p K a prediction. We have evaluated the p K a prediction capabilities of QSPR models based on empirical atomic charges calculated by the Electronegativity Equalization Method (EEM). Specifically, we collected 18 EEM parameter sets created for 8 different quantum mechanical (QM) charge calculation schemes. Afterwards, we prepared a training set of 74 substituted phenols. Additionally, for each molecule we generated its dissociated form by removing the phenolic hydrogen. For all the molecules in the training set, we then calculated EEM charges using the 18 parameter sets, and the QM charges using the 8 above mentioned charge calculation schemes. For each type of QM and EEM charges, we created one QSPR model employing charges from the non-dissociated molecules (three descriptor QSPR models), and one QSPR model based on charges from both dissociated and non-dissociated molecules (QSPR models with five descriptors). Afterwards, we calculated the quality criteria and evaluated all the QSPR models obtained. We found that QSPR models employing the EEM charges proved as a good approach for the prediction of p K a (63% of these models had R2 > 0.9, while the best had R2 = 0.924). As expected, QM QSPR models provided more accurate p K a predictions than the EEM QSPR models but the differences were not significant. Furthermore, a big advantage of the EEM QSPR models is that their descriptors (i.e., EEM atomic charges) can be calculated markedly faster than the QM charge descriptors. Moreover, we found that the EEM QSPR models are not so strongly influenced by the selection of the charge calculation approach as the QM QSPR models. The robustness of the EEM QSPR models was subsequently confirmed by cross-validation. The applicability of EEM QSPR models for other chemical classes was illustrated by a case study focused on carboxylic acids. In summary, EEM QSPR models constitute a fast and accurate p K a prediction approach that can be used in virtual screening.

The prediction of physical and chemical properties of molecules is a very important step in the drug discovery process. QSAR and QSPR models are strong tools for predicting these properties. The models employ descriptors and statistical approaches to provide an estimation of the desired property. An abundance of descriptors and QSAR/QSPR models were published, but the prediction of some properties (i.e., pKa, logP) is still a challenge [1]. For this reason, researchers are perpetually working on identifying new descriptors and analyzing their performance in newly designed models. The process of design, parameterization and evaluation of a QSAR/QSPR model is relatively complicated. Therefore, several software tools for its automation are currently under development [2,3]. These tools are very useful if we wish to apply some of many descriptors which they implement. But if we need to use other descriptors or test our own, we require a different solution. We introduce a new version of the software QSPR Designer that fulfils the above mentioned requirement. Specifically, the user can easily and quickly add his own module for descriptor calculation into QSPR Designer. Alternatively, descriptors can also be obtained from an input file. When the descriptors are available, QSPR Designer allows the user to include them to the model, to validate this model and to perform further tasks. The functionality of QSPR Designer is demonstrated on three case studies, in which the user gradually tests and improves his ideas of how to predict pKa. The first case study is focused on the question whether pKa has any relation with atomic charges. The second case study tests if we can use charges and a simple similarity-based model to predict pKa. The third case study analyzes several QSPR models predicting pKa from charges.

Novel high-throughput experimental techniques produce a large amount of data on the 3D structure of proteins and their structural motifs. These motifs can be used as patterns in drug discovery [1], can help to understand the relationship between a protein’s structure and its function [2] and to classify proteins [3]. In order to extract as much information as possible from this data, new techniques and tools are necessary, and among them fast approaches to perform the multiple superimposition of large sets of protein structural motifs. We report here on the development of such a tool. We have implemented our newly developed multiple superimposition methodology in the web application SiteBinder, which is able to process hundreds of protein structural motifs in a very short time and provides an intuitive and user-friendly interface. We also demonstrate the applicability of SiteBinder using three case studies, focused on biologically important protein motifs. In the first case study, we compared the structures of 67 PA-IIL sugar binding sites containing 9 different sugars and we found that the sugar binding sites of PA-IIL and its mutants have a conserved structure despite their binding different sugars. The second case study focused on more than 300 zinc finger central motifs and revealed that the molecular structure in the vicinity of the Zn atom in Cys2His2 zinc fingers is highly conserved. In the last case study, we superimposed 12 BH3 domains from pro-apoptotic proteins, and found that there is a structural basis for the functional segregation of these proteins into activators and enablers.

Tunnels are access paths connecting the interior of molecular systems with the surrounding environment. The presence of tunnels in proteins influences their reactivity, as they determine the nature and intensity of the interaction that these proteins can take part in. A few examples of systems whose function relies on tunnels include transmembrane proteins involved in small molecule transport and signal transduction, peptide exit channels through which ribosomes release newly synthesized proteins during transcription Knowledge of the location and characteristics of protein tunnels can find immediate applications in rational drug design, protein engineering, enzymology etc. Identification and characterization of tunnels has been the focus of several studies, and various algorithms and software tools have been developed for these purposes [1-4]. These methodologies use special mathematical algorithms to represent and scan the surface of the protein in search for tunnels and the amino acid residues involved. In the presented study we perform a benchmarking study of the most known approaches and software tools for finding tunnels in proteins (Mole, MolAxis, Hollow, etc.). We focused on proteins from the cytochrome P450 family, which are very important from the biological point of view. We provide a critical discussion of the strong and weak points of the analyzed approaches and software tools.

Nowadays, a large amount of experimental and predicted data about the 3D structure of organic molecules and biomolecules is available. Advanced computational methods and high performance computers allow us to obtain large sets of descriptors that can be used to estimate physicochemical properties. It is often of interest to study the correlations between descriptors and properties using multilinear regression and to design, parameterize, and test different QSPR (Quantitative Structure Property Relationship) models. We developed a modular and easily extensible program, called QSPR Designer, which can read or calculate structural properties of atoms and bonds, employ them as QSPR descriptors, and evaluate correlations between the descriptors and the examined physicochemical property of a molecule. Furthermore, the software allows us to effectively design and parameterize QSPR models, calculate physicochemical properties via the models, test the quality of the models, and provide graphs and tables summarizing the results. The performance of the software is demonstrated by a case study on the prediction of pKa. The pKa is of fundamental relevance for chemical, biological and pharmaceutical research, because many important physicochemical properties are pKa dependent. Unfortunately, pKa is also one of the most challenging properties to calculate [1]. Atomic charges have proven very successful descriptors for the prediction of pKa [2]. Charges can be calculated using a variety of methods (HF, MP2, functionals, etc.), population analyses (Mulliken, ESP, NPA, etc.) and basis sets. Consequently, the procedure of charge calculation strongly influences their correlation with pKa [3]. Using the QSPR Designer, we have successfully designed, evaluated, and compared 75 different QSPR models for the prediction of pKa from charges. Our best model predicted the pKa for 143 phenols with a correlation coefficient 0.969, RMSE (root mean square error) 0.416 and the average pKa error 0.329.

Background Partial atomic charges are a well-established concept, useful in understanding and modeling the chemical behavior of molecules, from simple compounds, to large biomolecular complexes with many reactive sites. Results This paper introduces AtomicChargeCalculator (ACC), a web-based application for the calculation and analysis of atomic charges which respond to changes in molecular conformation and chemical environment. ACC relies on an empirical method to rapidly compute atomic charges with accuracy comparable to quantum mechanical approaches. Due to its efficient implementation, ACC can handle any type of molecular system, regardless of size and chemical complexity, from drug-like molecules to biomacromolecular complexes with hundreds of thousands of atoms. ACC writes out atomic charges into common molecular structure files, and offers interactive facilities for statistical analysis and comparison of the results, in both tabular and graphical form. Conclusions Due to high customizability and speed, easy streamlining and the unified platform for calculation and analysis, ACC caters to all fields of life sciences, from drug design to nanocarriers. ACC is freely available via the Internet at http://ncbr.muni.cz/ACC. Electronic supplementary material The online version of this article (doi:10.1186/s13321-015-0099-x) contains supplementary material, which is available to authorized users.

Partial atomic charges describe the distribution of electron density in a molecule, and therefore they provide clues regarding the chemical behaviour of molecules. Atomic charges are frequently used in molecular modelling applications such as molecular dynamics, docking, conformational searches, binding site prediction, etc. Recently, partial atomic charges have also become popular chemoinformatics descriptors [1]. Partial atomic charges cannot be determined experimentally, and they are also not quantum mechanical observables. For this reason, many different methods have been developed for their calculation. These charge calculation methods can be divided into two main groups, namely quantum mechanical (QM) approaches and empirical approaches. QM approaches provide accurate charges, but they are very slow and therefore not feasible for large sets of molecules. Empirical charges can be calculated quickly and their accuracy is similar to QM, making empirical charges more appropriate for chemoinformatics applications. A very useful empirical charge calculation method is EEM (Electronegativity Equalization Method) [2,3]. This method provides charges comparable to the QM approach for which the given EEM model was parameterized. The weak point of this empirical method, as well as of other empirical methods, is the necessity for parameterization, and also the insufficient coverage of currently available EEM model parameters. In our work, we first analysed, how applicable are currently published EEM parameters in chemoinformatics. Specifically, how many molecules from databases of known organic compounds (Pubchem, ZINC, Drugbank etc.) they can cover. We found, the coverage is about 50-75%. We would like to show a methodology for preparation of parameters with higher coverage (>95% of molecules) and also its results.

The concept of atomic point charges is well established in theoretical chemistry. Atomic point charges have played an important role in understanding and modeling chemical behavior by allowing to extract and quantify information stored in the molecular electron distribution of chemical compounds. Thus, atomic point charges have been used to estimate reactivity indices, dissociation constants, partition coefficients, the electrostatic contribution in molecular dynamics or docking studies. It is therefore desirable to have knowledge of the values of atomic charges in proteins (see, e.g., [1]). Unfortunately, accurate and universally applicable approaches for atomic charge calculation based on quantum mechanics (QM) are very time consuming and thus cannot be employed for large biomolecules like proteins. An alternative is to use empirical charge calculation methods, such as the electronegativity equalization method (EEM) [2], which is very fast and has accuracy comparable to QM. The challenge is to calibrate (i.e., parametrize) this method for proteins. This parameterization can be done using atomic charges calculated by different types of QM approaches. EEM can be as accurate as the QM approach for which EEM was calibrated. In our work, we present the workflow of the EEM calibration process. Afterwards, we calibrate and validate EEM models for 12 types of QM charges, including the newest approaches like iterative Hirshfeld [3]. The accuracy of the obtained EEM models is evaluated on insulin and ubiquitin. We also show two case studies demonstrating the applicability of atomic charges computed via EEM: a small docking study, and the calculation of electrostatic potential based on the EEM charges [4].