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 ...

851possible, but provides a very simple and accurate method if it is. The systematic ge6metry also has the advantage of providing many independent structure factors, unlike the more accurate critical voltage method, which provides only a relationship between structure factors. The intersecting-Kikuchi-line method can also provide useful approximate values of structure factors, and is by far the easiest to apply since it relies on distance measurements taken from film, rather than intensity measurements. AbstractThe anisotropic displacements of selected rigid groups in bovine ribonuclease A have been refined from X-ray diffraction data by the application of the * Present address:

Neutron reflection has been used to study the effects of solution pH and ionic strength on the surface excess and layer thickness of lysozyme layers adsorbed at the air/water interface. All the measurements were made in null reflecting water (NRW) so that all the specular signal arose from the protein layers. At the low ionic strength of 0.02 M, the adsorption was found to reach a maximum at the protein isoelectric point (IP) of pH 11, with the effect of pH on the adsorbed amount and layer thickness being more pronounced at the higher lysozyme concentration. At the low lysozyme concentration of 0.03 g dm-3, the thicknesses of the adsorbed layers are 30 ± 3 Å over almost the entire pH range, close to the short axial length of the globular dimension of lysozyme, and the area per molecule is 1700 ± 200 Å2, suggesting the formation of a sideways-on monolayer. At the high lysozyme concentration of 1 g dm-3, a number of conformational transitions occur within the adsorbed layers with respect to pH and these variations correlate well with the change in the number of net charges within lysozyme with pH, suggesting that the preferred conformation of protein molecules is dominated by the combined effect of steric and electrostatic repulsion within the adsorbed layer. Subsequent measurements at the high ionic strength of 1 M showed no obvious variation in either layer thickness or surface excess with pH or with bulk protein concentration. The thickness was found to be constant at 30 ± 3 Å and the area per molecule to be 1500 ± 100 Å2, corresponding to the formation of a close-packed sideways-on monolayer. These results clearly show that salt addition has screened the charges within lysozyme molecules.

Three bis-benzoxazine monomers based on the aniline derivatives of bisphenol A (BA-a), bisphenol F (BF-a), and 3,3′-thiodiphenol (BT-a) are examined using a variety of spectroscopic, chromatographic, and thermomechanical techniques. The effect on the polymerization of the monomers is compared using two common compounds, 3,3′-thiodiphenol (TDP) and 3,3′-thiodipropionic acid (TDA), at a variety of loadings. It is found that the diacid has a greater effect on reducing the onset of polymerization and increasing cross-link density and Tg for a given benzoxazine. However, the addition of >5 wt % of the diacid had a detrimental effect on the cross-link density, Tg, and thermal stability of the polymer. The kinetics of the polymerization of BA-a were found to be well described using an autocatalytic model for which values of n = 1.64 and m = 2.31 were obtained for the early and later stages of reaction (activation energy = 81 kJ/mol). Following recrystallization the same monomer yielded values n = 1.89, m = 0.89, and Ea = 94 kJ/mol (confirming the influence of higher oligomers on reactivity). The choice of additive (in particular the magnitude of its pKa) appears to influence the nature of the network formation from a linear toward a more clusterlike growth mechanism.

The atomic displacements of many of the atoms in a macromolecular structure can be modelled in terms of group motions described in the harmonic approximation by T, L and S tensors. Relevant groups may be planar side groups of protein chains, units of secondary structure such as o~-helices or whole protein domains. For the TLS parameters to be interpreted, they must be related to the axes of inertia of the rigid groups and, in the case of the T and S tensors, must be calculated with respect to the centre of reaction of the rigid group. A program (TLSANL)is described that analyses these 21 TLS rigid-body displacement parameters and their relation with the principal axes of the rigid body, from the output of the segmented anisotropic refinement of a macromolecular structure, as produced by a program such as RESTRAIN