No abstract

A systematic technique for protein modelling that is applicable to the design of drugs, peptide vaccines and novel proteins is described. Our approach is knowledge-based, depending on the structures of homologous or analogous proteins and more generally on a relational data base of protein three-dimensional structures. The procedure simultaneously aligns the known tertiary structures, selects fragments from the structurally conserved regions on the basis of sequence homology, aligns these with the 'average structure' or 'framework', builds on the loops selected from homologous proteins or a wider database, substitutes sidechains and energy minimises the resultant model. Applications to modelling an homologous structure, tissue plasminogen activator on the basis of another serine proteinase, and to modelling an analogous protein, HIV viral proteinase on the basis of aspartic proteinases, are described. The converse problem of ah initio design is also addressed: this involves the selection of an amino acid sequence to give a particular tertiary structure, in this case a symmetrical domain of two Greekkey motifs.The characterisation, production and engineering of new proteins by recombinant DNA technology has revolutionised approaches to the design of novel molecules of industrial, agricultural and clinical interest. However, systematic design requires a blueprint or plan which must describe the molecular architecture. For this reason protein crystallography and modelling have become central to rational approaches to : (a) modelling of protein receptors for drugs, herbicides and insecticides; (b) design of protein and peptide vaccines; (c) protein engineering of novel molecules.Unfortunately a detailed experimental definition of the three-dimensional structure of a protein is time-consuming and expensive, and its success is usually dependent on the production of well ordered crystals suitable for X-ray diffraction although proteins may be studied by two-dimensional NMR techniques. On the other hand, molecular modelling has been unsuccessful when approached using ab initio techniques of simulation and prediction, At Birkbeck we are developing a systematic approach to molecular modelling and design that is dependent on structural homology and analogy. It utilises a knowledge-based approach derived from structures defined by X-ray analysis. It seeks to develop a relational data base of protein three-dimensional structures and a set of rules that can be used to assemble amino acid sequences, secondary structures, loops, domains and tertiary structures of unknown proteins in an hierarchical manner [I, 21.The value of an homologous protein structure in modelling was first recognised by Browne et al. [3] in the construc- tion of a model of a-lactalbumin based on the threedimensional structure of lysozyme. Similar methods have been used to define structural features of a range of proteins including insulin-like growth factors, serine proteinases, renins and histocompatibility antigens (see [2] for review). However, these ...

A single NMR-derived protein structure is usually deposited as an ensemble containing many structures, each consistent with the restraint set used. The number of NMR-derived structures deposited in the Protein Data Bank (PDB) is increasing rapidly. In addition, many of the structures deposited in an ensemble exhibit variation in only some regions of the structure, often with the majority of the structure remaining largely invariant across the family of structures. Therefore it is useful to determine the set of atoms whose positions are 'well defined' across an ensemble (also known as the 'core' atoms). We have developed a computer program, NMRCORE, which automatically defines (i) the core atoms, and (ii) the rigid body(ies), or domain(s), in which they occur. The program uses a sorted list of the variances in individual dihedral angles across the ensemble to define the core, followed by the automatic clustering of the variances in pairwise inter-atom distances across the ensemble to define the rigid body(ies) which comprise the core. The program is freely available via the World Wide Web (http://neon.chem.le.ac.uk/nmrcore/).

No abstract

No abstract

No abstract

Rapid advances in site-directed mutagenesis and total gene synthesis combined with new expression systems in prokaryotic and eukaryotic cells have provided the molecular biologist with tools for modification of existing proteins to improve catalytic activity, stability and selectivity, for construction of chimeric molecules and for synthesis of completely novel molecules that may be endowed with some useful activity. Such protein engineering can be seen as a cycle in which the structures of engineered molecules are studied by X-ray analysis and two-dimensional nuclear magnetic resonance. The results are used in the improvement of the design by using knowledge-based procedures that exploit facts, rules and observations about proteins of known three-dimensional structure.

The validation, enrichment and organization of the data stored in PDB ®les is essential for those data to be used accurately and ef®ciently for modelling, experimental design and the determination of molecular interactions. The Iditis protein structure database has been designed to allow the widest possible range of queries to be performed across all available protein structures. The Iditis database is the most comprehensive protein structure resource currently available, and contains over 500 ®elds of information describing all publicly deposited protein structures. A custom-written database engine and graphical user interface provide a natural and simple environment for the construction of searches for complex sequence-and structure-based motifs. Extensions and specialized interfaces allow the data generated by the database to used in conjunction with a wide range of applications.