Current Bioinformatics - Volume 7, Issue 2, 2012

Proteins and their interactions play vital roles in living organisms. Elucidating the mechanism of protein folding as well as the recognition of proteins with other molecules (proteins, nucleic acids and carbohydrates and ligands) are intriguing and challenging problems in computational and molecular biology. The problem of protein folding, stability and interactions has been viewed through several perspectives using experimental and computational approaches. Further, Bioinformatics has been successfully applied to enhance our understanding on protein folding, stability and their interactions. The special issue on “Bioinformatics on proteins and complexes” is aimed at providing a recent update on the computational analysis of proteins based on their folding, stability and interactions. It addressed various issues such as sequence-structure similarity, structure prediction, folding rates and stability of proteins. Further, it covered proteinprotein/ protein-RNA interactions, structure based drug design and proteomics analysis. The special issue is broadly classified into three parts; the first part is focused on the aspects of protein folding and stability with six articles, second part is devoted to protein interactions, which has four papers and the last part is dealing with database searching and preprocessing in mass spectrometry based proteomics....

The rapidly growing structure databases enhance the probability of finding identical sequences sharing structural similarity. Structure prediction methods are being used extensively to abridge the gap between known protein sequences and the solved structures which is essential to understand its specific biochemical and cellular functions. In this work, we plan to study the ambiguity between sequence-structure relationships and examine if sequentially identical peptide fragments adopt similar three-dimensional structures. Fragments of varying lengths (five to ten residues) were used to observe the behavior of sequence and its three-dimensional structures. The STAMP program was used to superpose the three-dimensional structures and the two parameters (Sequence Structure Similarity Score (Sc) and Root Mean Square Deviation value) were employed to classify them into three categories: similar, intermediate and dissimilar structures. Furthermore, the same approach was carried out on all the three-dimensional protein structures solved in the two organisms, Mycobacterium tuberculosis and Plasmodium falciparum to validate our results.

Prediction of protein 3D structure from solely its amino acid sequence is one of the most challenging problems in structural bioinformatics, where the 3D structure reconstruction from observed constraints is the key step. In this paper, we propose a novel protocol called Glocal to recover a protein’s 3D coordinates based on a given 2D contact map by combining both global and local optimization schemes achieved by the swarm intelligence of Particle Swarm Optimization (PSO) and the Simulated Annealing (SA) techniques respectively. Our results demonstrate that Glocal can recover the 3D structures with the average RMSD less than 2 Å from the native contact map. Further analysis also shows that Glocal is powerful for handling with noisy contact map with the proposed combination optimization approaches.

The attachment of glycosylphosphatidylinositol (GPI) is one of the most important post-translational modifications of proteins and plays an important role in promoting biochemical activities in eukaryotic cells. GPIanchored proteins (GPI-APs) are characterized by GPI-anchor attachment signals of hydrophobic residues and small residues near the GPI-anchoring site (ω-site). Here, we describe a new method for predicting GPI-APs based on hydropathy profiles and position-specific scores (PSSs) in combination with the back propagation artificial neural network (BP-ANN). First, the sequences of GPI-APs and negative controls were aligned according to residue size in the Cterminal region and the position-specific amino acid propensities were analyzed according to their alignment positions. Next, PSSs were created using the amino acid propensities of GPI-APs and the negative controls, and BP-ANN with a three-layered structure was trained by the PSSs. The accuracy of discriminating GPI-APs from the negative controls was evaluated in a 4-fold cross-validation test and GPI-APs were detected with 92.9% sensitivity and 94.8% specificity. This result shows that our method can predict GPI-APs with high accuracy and a combination of PSSs and BP-ANN can effectively discriminate GPI-APs.

G protein-coupled receptors (GPCRs) are integral membrane proteins with seven trans-membrane helices. Belonging to the largest family of cell surface receptors, GPCRs are among the most frequent targets of therapeutic drugs. Unfortunately, since they are difficult to crystallize and most of them will not dissolve in normal solvents, so far the number of GPCRs with three-dimensional structure determined is very limited. This situation has challenged us to develop automated methods by which one can predict the family and sub-family classes of GPCRs based on the information of their primary sequences alone, so as to facilitate classifying drugs, a technique called “evolutionary pharmacology” often used in pharmaceutical industries for drug development. In the past eight years, various computational methods were proposed. This review is devoted to summarize their development. Meanwhile, the future challenge in this area has also been briefly addressed.

Free energy of unfolding (ΔGu) is the difference between the free energy values of the folded and unfolded structures of a protein. A successful model describing both folding/unfolding rates of proteins should be able to provide considerable insight on free energy of unfolding. In our earlier works, we have shown that Long-range Order (LRO) correlates well with both folding/unfolding rates of two-state proteins. In the present work, we examine the extent to which LRO can be used to predict the free energy of unfolding. For a standard data set of 29 two-state proteins, no significant correlation was observed between ΔGu and LRO. However after grouping the proteins according to their structural class, all-alpha and all-beta proteins showed a better correlation of r = 0.77 and r = 0.89, whereas mixed-class proteins still showed a poor correlation. We have also analyzed the relationship between various other structure derived topological parameters with ΔGu values and the results observed showed that all these parameters also gave a poor correlation with ΔGu values when structural classification was not taken into account. Similar to LRO, after structural classification better improvement in correlation was observed for all-alpha and all-beta proteins and not a single topological parameter showed reasonable correlation with ΔGu values of mixed-class proteins and suggested that understanding ΔGu values of mixed-class proteins remains complicated. Our present work implies that theoretical models to understand stability of proteins can be developed based on their 3-D structures and further experimental/theoretical studies will shed light on these predictions.

Protein thermostabilization has been the focus of recent research due to growing interest in the production of enzymes that can operate at temperatures that are industrially beneficial. Understanding the determinants of thermostabilization at the level of sequence and structure is important to design such enzymes. A bioinformatical approach was used to determine the extent by which reduced amino acid alphabets (RAAA) with n-grams (subsequences of length n) that were subjected to a t-test-based feature selection procedure can be used to discriminate proteins from thermophiles and mesophiles. Classification performance of 65 different protein alphabets with 3 different n-gram sizes was systematically evaluated using support vector machines in a test set that contained 707 proteins from mesophilic Xylella fastidosa and thermophilic Aquifex aeolicus. A classification accuracy of 91.796% was achieved with Hsdm16 RAAA with 13 features: EK-ILV-ST-A-G-F-H-Q-N-R-M-W-Y. The t-test-based feature selection procedure reduced the classification time without significantly affecting classification accuracy. The overall combination of methods in this paper is useful and computationally fast for classifying protein sequences from thermophiles and mesophiles using sequence information alone.

The structural analysis of protein-protein interactions and the prediction of their functional properties are important areas in modern structural bioinformatics. First, we review concepts for classifying protein-protein interactions, and for analyzing the geometry and composition of binding interfaces. Next, computational methods are discussed that allow predicting hot-spot residues and the kinetics and the rmodynamics of binding. Then, we focus on the mode of action of small molecule effectors that may either act as competitive antagonists of protein binders or as allosteric modulators. Here, we emphasize the roles of pre-formed or transiently open ligand-binding pockets at protein-protein interfaces. The presentation is rounded up by an overview over databases on protein-protein and protein-small-molecule interactions.

It has been well established that the elastic character of DNA plays an important role in protein-DNA binding specificity. In this work, we have analyzed the role of elasticity to understand the binding specificity of protein-RNA complexes. We have developed a sequence dependent stiffness scale for the trinucleotides in RNA and revealed the similarities and differences compared with DNA. We found that the stiffness of 15 trinucleotides has inverse effects and nine nucleotides are significantly different between RNA and DNA. The analysis on the relationship between RNA stiffness and RNA-binding specificity shows that the influence of elasticity is minimal in protein-RNA recognition, whereas it plays an important role in protein-DNA binding specificity. We observed a moderate correlation between stiffness and dissociation constant in U1A RBD1 protein and PP7 coat protein, whereas the correlation is poor for many other complexes. These results show that along with RNA stiffness, other interactions, such as shape complementarity, electrostatic interactions, hydrogen bonds and direct contacts between RNA and protein atoms are important for protein- RNA recognition.

Background: Metal atoms are involved in many biological mechanisms, such as protein structure stability, apoptosis and aging. Therefore, identifying metal-binding sites in proteins is an important issue in helping biologists better understand the workings of these mechanisms. Methods: We propose a method based on Position Specific Scoring Matrix (PSSM) profiles and additional information (conservation score and solvent accessible surface area (ASA)) to identify the metal-binding residues in proteins. Results: We have selected a non-redundant set of 262 metal-binding proteins and 617 disulfide proteins as the independent test set. The proposed method can predict metal-binding sites at 51.0% recall and 73.4% precision. Comparing with the previous work of A. Passerini et al., the proposed method can improve over 7% of precision with the same level of recall on the independent dataset. Conclusions: We have developed a novel approach based on PSSM profiles and additional properties for identifying metal-binding sites from proteins. The proposed approach achieved a significant improvement with newly discovered metal-binding proteins and disulfide proteins.

Viral diseases cause severe damage to human lives than any other microbes. Hepatitis is the inflammation of liver and currently six strains of viral hepatitis are identified. Infection by Hepatitis B Virus (HBV) and Hepatitis C Virus (HCV) causes serious mortality, morbidity and becomes a global health problem. Human Immunodeficiency Virus (HIV) is increasing in the world, with an estimation of 5.7 million cases of HIV infection in India. In addition to these viruses, Dengue virus, which belongs to the family Flaviviridae has also emerged as a global threat to humans and is a major emerging pathogen for which the development of vaccine and anti-viral therapy has seen a little success. The NS3 viral protease is a potential target for anti-viral drugs, since it is required for viral replication. As Dengue hemorrhage diseases are the life-threatening ones, attempts are being made worldwide to design inhibitors for DENV-2 NS2B-NS3 protease, DENV-4 NS3 protease-helicase as targets. In view of the above viral threats to human life, attempts are being made to come out with anti-viral compounds from natural resources and also from synthetic routes. Natural sources include compounds reported from Neem (Azadirachta indica), Bael (Aegle marmelos), Murraya koenigii, Heliopsis scabra, Taiwania cryptomerioides, edible fishes, and crab. Synthetic peptides and organic compounds are also attempted as inhibitors. Viral proteins are retrieved from Protein Data Bank (PDB) and docked with these lead compounds and the results are analyzed. All docking studies have been carried out using Schrödinger USA suite of programs 2009.

Mass spectrometry based proteomics analysis can produce many thousands of spectra in a single experiment, and much of this data, frequently greater than 50%, cannot be properly evaluated computationally. Therefore a number of strategies have been developed to aid the processing of mass spectra and typically focus on the identification and elimination of noise, which can provide an immediate improvement in the analysis of large data streams. This is mostly carried out with proprietary software. Here we review the current main principles underlying the preprocessing of mass spectrometry data give an overview of the publicly available tools.

Bottom-up proteomics (mass spectrometry analysis of peptides obtained by proteolysis and separated by liquid chromatography, (LC-MS/MS)) is one of the most frequently used techniques for identifying and characterizing proteins in biological samples. A key element of the analysis is database searching when the mass spectra of the peptides are compared with a database of theoretically computed (or experimental) peptide spectra. Here we discuss the main computational approaches to spectrum database searching and the statistical analysis of the results.