Categorization of metabolome in bacterial systems.

Analyses of biological databases such as those of genome, proteome, metabolome etc., have given insights in organization of biological systems. However, current efforts do not utilize the complete potential of available metabolome data. In this study, metabolome of bacterial systems with reliable annotations are analyzed and a simple method is developed to categorize pathways hierarchically, using rational approach. Ninety-four bacterial systems having for each ≥ 250 annotated metabolic pathways were used to identify a set of common pathways. 42 pathways were present in all bacteria which are termed as Core/Stage I pathways. This set of pathways was used along with interacting compounds to categorize pathways in the metabolome hierarchically. In each metabolome non-interacting pathways were identified including at each stage. The case study of Escherichia coli O157, having 433 annotated pathways, shows that 378 pathways interact directly or indirectly with 41 core pathways while 14 pathways are noninteracting. These 378 pathways are distributed in Stage II (289), Stage III (75), Stage IV (13) and Stage V (1) category. The approach discussed here allows understanding of the complexity of metabolic networks. It has pointed out that core pathways could be most ancient pathways and compounds that interact with maximum pathways may be compounds with high biosynthetic potential, which can be easily identified. Further, it was shown that interactions of pathways at various stages could be one to one, one to many, many to one or many to many mappings through interacting compounds. The granularity of the method discussed being high; the impact of perturbation in a pathway on the metabolome and particularly sub networks can be studied precisely. The categorizations of metabolic pathways help in identifying choke point enzymes that are useful to identify probable drug targets. The Metabolic categorizations for 94 bacteria are available at http://115.111.37.202/mpe/.

Towards a consensus on datasets and evaluation metrics for developing B-cell epitope prediction tools.

A B-cell epitope is the three-dimensional structure within an antigen that can be bound to the variable region of an antibody. The prediction of B-cell epitopes is highly desirable for various immunological applications, but has presented a set of unique challenges to the bioinformatics and immunology communities. Improving the accuracy of B-cell epitope prediction methods depends on a community consensus on the data and metrics utilized to develop and evaluate such tools. A workshop, sponsored by the National Institute of Allergy and Infectious Disease (NIAID), was recently held in Washington, DC to discuss the current state of the B-cell epitope prediction field. Many of the currently available tools were surveyed and a set of recommendations was devised to facilitate improvements in the currently existing tools and to expedite future tool development. An underlying theme of the recommendations put forth by the panel is increased collaboration among research groups. By developing common datasets, standardized data formats, and the means with which to consolidate information, we hope to greatly enhance the development of B-cell epitope prediction tools.

CEP: a conformational epitope prediction server.

CEP server (http://bioinfo.ernet.in/cep.htm) provides a web interface to the conformational epitope prediction algorithm developed in-house. The algorithm, apart from predicting conformational epitopes, also predicts antigenic determinants and sequential epitopes. The epitopes are predicted using 3D structure data of protein antigens, which can be visualized graphically. The algorithm employs structure-based Bioinformatics approach and solvent accessibility of amino acids in an explicit manner. Accuracy of the algorithm was found to be 75% when evaluated using X-ray crystal structures of Ag-Ab complexes available in the PDB. This is the first and the only method available for the prediction of conformational epitopes, which is an attempt to map probable antibody-binding sites of protein antigens.

VirGen: a comprehensive viral genome resource.

VirGen is a comprehensive viral genome resource that organizes the 'sequence space' of viral genomes in a structured fashion. It has been developed with the objective of serving as an annotated and curated database comprising complete genome sequences of viruses, value-added derived data and data mining tools. The current release (v1.1) contains 559 complete genomes in addition to 287 putative genomes of viruses belonging to eight viral families for which the host range includes animals and plants. Viral genomes in VirGen are annotated using sequence-based Bioinformatics approaches. The genomic data is also curated to identify 'alternate names' of viral proteins, where available. VirGen archives the results of comparisons of genomes, proteomes and individual proteins within and between viral species. It is the first resource to provide phylogenetic trees of viral species computed using whole-genome sequence data. The module of predicted B-cell antigenic determinants in VirGen is an attempt to link the genome to its vaccinome. Comparative genome analysis data facilitate the study of genome organization and evolution of viruses, which would have implications in applied research to identify candidates for the design of vaccines and antiviral drugs. VirGen is a relational database and is available at http://bioinfo. ernet.in/virgen/virgen.html.

SEGE: A database on 'intron less/single exonic' genes from eukaryotes.

UNLABELLED: Eukaryotes have both 'intron containing' and 'intron less' genes. Several databases are available for 'intron containing' genes in eukaryotes. In this note, we describe a database for 'intron less' genes from eukaryotes. 'Intron less' eukaryotic genes having prokaryotic architecture will help to understand gene evolution in a much simpler way unlike 'intron containing' genes. AVAILABILITY: SEGE is available at http://intron.bic.nus.edu.sg/seg/ CONTACT: mmeena@ntu.edu.sg

Online identification of viruses.

A computerized animal virus information system is developed in the Sequence Retrieval System (SRS) format. This database is available on the Word Wide Web (WWW) at the site http://bioinfo.ernet.in/www/avis/avis++ +.html. The database has been used to generate large number of identification matrices for each family. The software is developed in C. Unix shell scripts and Hypertext Marked-up Language (HTML) to assign the family to an unknown virus deterministically and to identify the virus probabilistically. It has been shown that such web based virus identification approach provides results with high confidence in those cases where identification matrix uses large number of independent characters. Protein sequence data for animal viruses have been analyzed and oligopeptides specific to each virus family and also specific to each virus species are identified for several viruses. These peptides thus could be used to identify the virus and to assign the virus family with high confidence showing the usefulness of sequence data in virus identification.

Prediction of three-dimensional structure and mapping of conformational epitopes of envelope glycoprotein of Japanese encephalitis virus.

Japanese encephalitis virus (JEV), a mosquito-borne flavivirus, is an important human pathogen. The envelope glycoprotein (Egp), a major structural antigen, is responsible for viral haemagglutination and eliciting neutralising antibodies. The three-dimensional structure of the Egp of JEV was predicted using the knowledge-based homology modeling approach and X-ray structure data of the Egp of tick-borne encephalitis virus as a template (Rey et al., 1995). In the initial stages of optimisation, a distance-dependent dielectric constant of 4r(ij) was used to simulate the solvent effect. The predicted structure was refined by solvating the protein in a 10-A layer of water by explicitly considering 4867 water molecules. Four independent structure evaluation methods report this structure to be acceptable stereochemically and geometrically. The Egp of JEV has an extended structure with seven beta-sheets, two alpha-helices, and three domains. The water-solvated structure was used to delineate conformational and sequential epitopes. These results document the importance of tertiary structure in understanding the antigenic properties of flaviviruses in general and JEV in particular. The conformational epitope prediction method could be used to identify conformational epitopes on any protein antigen with known three-dimensional structure. This is one of the largest proteins whose three-dimensional structure has been predicted using an homology modeling approach and water as a solvent.

Determination of primary amino acid sequence and unique three-dimensional structure of WGH1, a monoclonal human IgM antibody with anti-PR3 specificity.

Transformed B cells making monoclonal IgM-lambda anti-PR3 antibody WGH1 from a patient with Wegener's granulomatosis were used to prepare mRNA and synthesize cDNA. PCR primers for human micro and lambda chains were then employed to amplify heavy- and light-chain V-regions followed by cloning into pCR2-1 vector and sequencing. Molecular modeling of VH regions employed knowledge-based homology modeling to obtain minimum energy conformation. The VH sequence was subgroup III with marked overall homology to VH1.9III. The VHCDR3 region of WGH1 was unique, consisting of 21 amino acid residues which included seven tyrosines as well as three negatively charged aspartic acid residues. The VL region was subgroup II with a negatively charged glutamic acid at position 100 in CDR3. Molecular modeling of VH revealed a major conformational difference in the shape of CDR3 compared with other antibodies for which three-dimensional structures have been determined. Monoclonal antibody WGH1 reacting with PR3 (a highly positively charged molecule) shows a unique reactive cassette within VHCDR3 with a number of negatively charged aspartic acid residues. WGH1 VHCDR3 contains a loop which shows a major projection not usually recorded in other previously studied antibody molecules.

Computer-aided virus identification on the World Wide Web.

An attempt has been made to devise computer software that will aid virologists to identify unknown virus isolates using the World Wide Web. Computerized information from the Animal Virus Information System was used to obtain data on various characters of a virus species. Sequence data banks are used to obtain the molecular data. A probabilistic method of virus identification based on Willcox's implementation of Bayes' theorem is implemented. The program provides hints to the users to carry out additional tests required to obtain higher confidence in identification of virus species. Signature peptides of the virus can also be used to confirm identification. The software is implemented on a UNIX machine and is written in C, UNIX shell scripts and HTML to run on the World Wide Web. This is the first species identification software that allows the user to carry out identification online through Internet.

Molecular dynamics simulation of a 13-mer duplex DNA: a PvuII substrate.

Parallel version of AMBER 4.1 was ported and optimised on the Indian parallel supercomputer PARAM OpenFrame built around Sun Ultra Sparc processors. This version of AMBER program was then used to carry out molecular dynamics (MD) simulations on 5'-TGACCAGCTGGTC-3', a substrate for PvuII enzyme. MD simulations in water are carried out under following conditions: (i) unconstrained at 300 K (230 ps); (ii) unconstrained at 283 K (500 ps); (iii) Watson-Crick basepair constrained at 283 K (1 ns); and (iv) Watson-Crick basepair constrained with ions at 283 K (1.2 ns). In all these simulation studies, the molecule was observed to be bending and maximum distortions in the double helix around was seen around the G7:C7' basepair, which is the phosphodiester bond that is cleaved by PvuII. Analysis of MD simulation with ions carried out for 1.2 ns also pointed out that the conformation of double helix alternates between a conformation close to B-form and close to A-form. It is argued that a bent non-standard conformation is recognised by the PvuII enzyme. The maximum bend occurs at the G7:C7' region, weakening the phosphodiester bond and allows His48 to get placed in such a fashion to permit the scission through a general base mechanism. The bending and distortion observed is a property of the sequence which acts as a substrate for PvuII enzyme. This is confirmed by carrying out MD studies on the Dickerson's sequence d(CGCGAATTCGCG)2 as a reference molecule, which practically does not bend or get deformed.

Antigenic determinants reacting with rheumatoid factor: epitopes with different primary sequences share similar conformation.

Polyclonal or monoclonal human IgM rheumatoid factors (RF) react with eight antigenic sites on the CH3 IgG domain, four sites on CH2 and two on human beta 2-microglobulin. All 14 of these RF-reactive epitopes are linear 7-11 amino acid peptides with different primary sequence. We questioned whether RF reactivity with such a variety of epitopes showing no obvious sequence homology might result from conformational similarities shared by various RF-reactive regions. Strong support for this concept was obtained using rabbit antisera as well as mouse mAbs to individual CH3, CH2 or beta 2m RF-reactive peptides. Major cross-reactivity was demonstrated between most of the 14 different CH3, CH2, or beta 2m RF-reactive peptides using individual anti-epitope antibodies. Molecular modelling studies of these peptides showed striking similarities in three-dimensional shape among many RF-reactive peptides. Main-chain atoms rather than side chains seemed to contribute most directly to conformational similarity. Molecular simulation studies on control peptides showed no conformational similarities with RF-reactive peptides. Our studies indicate that autoantibodies such as RF recognize main-chain conformations of reactive epitopes and react with a number of antigenic determinants of quite different primary sequence but similar main chain conformations.

Prediction of conformational states of amino acids using a Ramachandran plot.

(phi, psi) data from crystal structures of 221 proteins having high resolution and sequence similarity cut-off at the 25% level were analysed by dividing the Ramachandran plot in three regions representing three conformational states: (i) conformational state 1: conformations in the (phi, psi) range from (-140 degrees, -100 degrees) to (0 degrees, 0 degrees); (ii) conformational state 2: conformations with (phi, psi) from (-180 degrees, 80 degrees) to (0 degrees, 180 degrees); and (iii) conformational state 3: all the remaining conformations in the (phi, psi) plane which are not included in the above two conformational states. Normalized probability values of the occurrence of single amino acid residues in conformational regions 1-3 and similar values for dipeptides were calculated. Comparisons of single residue and dipeptide normalized probability values have shown that short-range interactions, although strong, destabilize conformational states of only 44 dipeptides out of the 400 x 9 possible states. However, dipeptide frequency values provide better resolving power than single-residue potentials when used to predict conformational states of residues in a protein from its primary structure. The simple approach used in the present study to predict conformational states yields an accuracy of > 70% for 14 proteins and an accuracy in the range of 50-70% for 247 proteins. Thus these studies point out yet another use of the Ramachandran plot and the role of tertiary interactions in protein folding.

Contextual constraints in the choice of synonymous codons.

From EMBL Nucleotide Sequence Database, protein coding sequences of all E. coli and its DNA phages, were extracted using our computer programme. Same programme has been used to form a database of sequence of oligonucleotides of length 18 nucleotides on both sides of each of the 61 codons. From analysis of this database and study of variations in twist parameter (Tw) values, as an indicator of sequence dependent variations in B-DNA helix, a method is developed to fix the codon among the set of synonymous codons. The accuracy of the method was checked on enlarged data set by adding data from more prokaryotes. Our method assign the codon 85-90% times correctly if the selection has to be made between codons having different sequence in terms of R and Y. The accuracy of the method is somewhat lower when choice of the codon has to be made between codons having same codes in terms of R and Y. This study points out that the major factors which decide the choice of a codon from a set of synonymous codons are contextual constraints arising from flanking regions.

Analysis of computer-predicted antibody inducing epitope on Japanese encephalitis virus.

Theoretical methods to delineate antibody inducing epitopes have been employed to predict antigenic determinants on envelope glycoprotein (gpE) of Japanese encephalitis (JE), West Nile (WN) and Dengue (DEN) I-IV viruses. A predicted region on JE virus gpE 74CPTTGEAHNEKRAD87 was synthesized, conjugated to KLH (KLH-peptide) and used in immunization of mice. A mouse monoclonal antibody (MoAb IVB4) reactive to the peptide was also found to react with native JE virus gpE. Characterization of the idiotypic (ID) determinants with the help of polyclonal domain-specific anti-ID antibodies revealed that polyclonal anti-KLH-peptide antibodies and MoAb IVB4 are flavivirus-cross-reactive to Hx and NHx domains, respectively. The region 74-87 in JE virus gpE has been mapped as a linking area between Hx and NHx domains. Reactivity of the peptide with sera from JE patients and vaccinees also indicated the feasibility of using predicted peptides for diagnostic and prophylastic purposes.

Multiple alignment of sequences on parallel computers.

A software package that allows one to carry out multiple alignment of protein and nucleic acid sequences of almost unlimited length and number of sequences is developed on C-DAC parallel computer--a transputer-based machine. The farming approach is used for data parallelization. The speed gains are almost linear when the number of transputers is increased from 4 to 64. The software is used to carry out multiple alignment of 100 sequences each of alpha-chain and beta-chain of hemoglobin and 83 cytochrome c sequences. The signature sequence of cytochrome c was found to be PGTKMXF. The single parameter, multiple alignment score, S, has been used to categorize proteins in different subfamilies and groups.

Sequence alignment approach to pick up conformationally similar protein fragments.

Crystal structure data of globular proteins were used to prepare (phi, psi) probability maps of 20 proteinous amino acids. These maps were compared grid-wise with each other and a conformational similarity index was calculated for each pair of amino acids. A weight matrix, called Conformational Similarity Weight (CSW) matrix, was prepared using the conformational similarity index. This weight matrix was used to align sequences of 21 pairs of proteins whose crystal structures are known. The aligned regions with more than seven contiguous amino acids were further analysed by plotting average weight (W) values of overlapping hepatapeptides in these regions and carrying out curve fitting by Fourier series having TEN harmonics. The protein fragments corresponding to the half-linewidth of peaks were predicted as fragments having similar conformation in the protein pair under consideration. Such an approach allows us to pick up conformationally similar protein fragments with more than 67% accuracy.

Computerization of virus data and its usefulness in virus classification.

Data on 537 Arboviruses and 180 other viruses have been collected and coded in two different formats. These data include information not only regarding the taxonomy and history of isolation, but also regarding the properties of biomacromolecules, proteins and nucleic acids. Information on antigenic relationships, histopathology and experimental viremia is also included. This information is stored in formats which allow the manipulation and analysis of data by dBASE III PLUS and MICRO-IS. A set of programs was written for interconversion and editing purposes. Transmission electron micrographs are scanned and stored. This stored information can be used in viral classification as shown by carrying out analysis of data on the Bunyaviridae family.

Analysis of inverted repeats in primary structure of proteins.

A computer program has been developed to locate exact inverted repeating subsequences present anywhere in the given primary structure of proteins or nucleic acids. The output is amenable to protein sequence/nucleic acid query (PSQ/NAQ) packages. Our analysis has shown that there is a large number of proteins which have inverted repeats of more than four amino acid residues in length. However, the number is small when conditions such as the existence of more than 20 inverted repeats in given sequence or the existence of inverted repeats having more than five different types of amino acids are applied.

A protein secondary structure database (PSS).

A protein secondary structure database (PSS) has been designed to correlate the Protein Sequence Database of the PIR-International with the atomic coordinates and bond connectivities database of the Protein Data Bank in the Brookhaven National Laboratory. The present database includes secondary structures determined by X-ray diffraction analysis, but not predicted structures. The database currently contains data from both the Protein Sequence Database and the Protein Data Bank Database, and will encompass the NMR database in the future. The main characteristics of the database are as follows: (1) the secondary structures, sites, regions and domains of structural interest are displayed together with protein primary structures; and (2) the secondary structure of a desired length of peptide fragment is displayed upon request, as are the peptide fragment(s) that correspond to a defined secondary structure. This database also has software to indicate amino acid pairs having hydrogen bonds and to count the occurrence frequency of each pair as well as the conformational parameters widely used in semi-empirical methods of secondary structure prediction.

Recognition of helper T cell epitopes in envelope (E) glycoprotein of Japanese encephalitis, west Nile and Dengue viruses.

Helper T (Th) cell antigenic sites were predicted from the primary amino acid sequence (approximately 500 in length) of the envelope (E) glycoprotein (gp) of Japanese encephalitis (JE), West Nile (WN) and Dengue (DEN) I-IV flaviviruses. Prediction of Th epitopes was done by analyzing the occurrence of amphipathic segments, Rothbard-Taylor tetra/pentamer motifs and presence of alpha helix-preferring amino acids. The simultaneous occurrence of all these parameters in segments of E gp were used as criteria for prediction as Th epitopes. Only one cross reactive epitope was predicted in the C-terminal region of the E gp predicted segments of all flaviviruses analyzed. This region is one of the longest amphipathic stretch (approximately from 420 to 455) and also has a fairly large amphipathic score. Based on the predicted findings three selected peptides were synthesized and analyzed for their ability to induce in vitro T cell proliferative response in different inbred strains of mice (Balb/c, C57BL6, C3H/HeJ). Synthetic peptide I and II prepared from C-terminal region gave a cross reactive response to JE, WN and Den-II in Balb/c and C3H/HeJ mice. Synthetic peptide III prepared from N-terminal region gave a proliferative response to DEN-II in Balb/c strain only, indicating differential antigen presentation.