Crystal structure of an Xrcc4-DNA ligase IV complex

A complex of two proteins, Xrcc4 and DNA ligase IV, plays a fundamental role in DNA non-homologous end joining (NHEJ), a cellular function required for double-strand break repair and V(D)J recombination. Here we report the crystal structure of human Xrcc4 bound to a polypeptide that corresponds to the DNA ligase IV sequence linking its two BRCA1 C-terminal (BRCT) domains. In the complex, a single ligase chain binds asymmetrically to an Xrcc4 dimer. The helical tails of Xrcc4 undergo a substantial conformational change relative to the uncomplexed protein, forming a coiled coil that unwinds upon ligase binding, leading to a flat interaction surface. A buried network of charged hydrogen bonds surrounded by extensive hydrophobic contacts explains the observed tightness of the interaction. The strong conservation of residues at the interface between the two proteins provides evidence that the observed mode of interaction has been maintained in NHEJ throughout evolution.

Protein-protein interactions in receptor activation and intracellular signalling

We review here signalling complexes that we have defined using X-ray analysis in our laboratory. They include growth factors and their receptors: nerve growth factor (NGF) and its hetero-hexameric 7S NGF storage complex, hepatocyte growth factor/scatter factor (HGF/SF) NK1 dimers and fibroblast growth factor (FGF1) in complex with its receptor (FGFR2) ectodomain and heparin. We also review our recent structural studies on intracellular signalling complexes, focusing on phosducin transducin GPry, CK2 protein kinase and its complexes, and the cyclin D-dependent kinase, Cdk6, bound to the cell cycle inhibitor p19INK4d. Comparing the structures of these complexes with others we show that the surface area buried in signalling interactions does not always give a good indication of the strength of the interactions. We show that conformational changes are often important in complexes with intermediate buried surface areas of 1500 to 2000 A2, such as Cdk6INK4 interactions. Some interactions involve recognition of continuous epitopes, where there is no necessity for a tertiary structure and very often the binding conformation is induced during the process of interaction, for example phosducin binding to the betagamma subunits (Gtbetagamma) of the heterotrimeric G protein transducin.

Crystal structure of aspartate decarboxylase at 2.2 A resolution provides evidence for an ester in protein self-processing

The structure of L-aspartate-alpha-decarboxylase from E. coli has been determined at 2.2 A resolution. The enzyme is a tetramer with pseudofour-fold rotational symmetry. The subunits are six-stranded beta-barrels capped by small alpha-helices at each end. The active sites are located between adjacent subunits. The electron density provides evidence for catalytic pyruvoyl groups at three active sites and an ester at the fourth. The ester is an intermediate in the autocatalytic self-processing leading to formation of the pyruvoyl group. This unprecedented structure provides novel insights into the general phenomenon of protein processing.

Accommodating sequence changes in beta-hairpins in proteins

A systematic study of homologous beta-hairpins in proteins of known structure reveals how insertions and deletions (herein known as indels) in the sequence are accommodated. The study was made for 12 protein families comprising 50 different structures, in which there were 49 independent hairpins. Each hairpin was classified according to its loop length and hydrogen bonding pattern. Most indels were found to occur in the loops and their frequency decreases rapidly with the size of the indel and approximately halves for each extra residue inserted. In very short loops, critical glycines are the primary determinants of loop structure and conversions between the two classic two-residue hairpin loops (with type I' and II' beta-turns) are quite common. Longer insertions are often accommodated by extending the beta-ladder and forming extra hydrogen bonds. There are also several indels that are not accommodated in the loop, but by forming a beta-bulge in one of the strands. This study should provide a useful aid to modelling hairpins in homologous structures.

X-ray analyses of peptide-inhibitor complexes define the structural basis of specificity for human and mouse renins

X-ray analyses have defined the three-dimensional structures of crystals of mouse and human renins complexed with peptide inhibitors at resolutions of 1.9 and 2.8 A, respectively. The exquisite specificity of renin arises partly from ordered loop regions at the periphery of the binding cleft. Although the pattern of main-chain hydrogen bonding in other aspartic proteinase inhibitor complexes is conserved in renins, differences in the positions of secondary structure elements (particularly helices) also lead to improved specificity in renins for angiotensinogen substrates.

Crystallization and preliminary X-ray analysis of complexes of peptide inhibitors with human recombinant and mouse submandibular renins

Inhibitor-complexed crystals of mouse and human renins suitable for X-ray analysis have been prepared. The mouse renin is complexed with a non-hydrolysable decapeptide analogue of rat angiotensinogen containing a hydroxyethylene isostere in place of the scissile bond. The crystals are monoclinic, space group P2(1) with cell dimensions a = 78.3 A, b = 117.8 A, c = 85.9 A, beta = 101.18 degrees containing four molecules per asymmetric unit. The human renin is fully glycosylated and complexed with a tetrapeptide containing norstatine. The complex crystallises in the cubic space group P2(1)3 with a = 143.1 A and has two molecules in the asymmetric unit. The rotation function of the mouse renin complex indicates pseudo 222 symmetry while that of human renin indicates a pseudo 2-fold axis. Full structural analyses of the two complexes are underway.

Conformation of beta-hairpins in protein structures. A systematic classification with applications to modelling by homology, electron density fitting and protein engineering

A systematic classification of beta-hairpin structures which takes into account the polypeptide chain length and hydrogen bonding between the two antiparallel beta-strands is described. We have used this classification of beta-hairpin structures and their specific sequence pattern to derive rules which demonstrate its usefulness in assisting modelling beta-hairpins. These rules can be applied to comparative model building, modelling into electron density and in the prediction of conformation of beta-hairpins to aid protein engineering.

18th Sir Hans Krebs lecture. Knowledge-based protein modelling and design

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 ab 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 Greek-key motifs.

High resolution X-ray analyses of renin inhibitor-aspartic proteinase complexes

Inhibitors of the conversion of angiotensinogen to the vasoconstrictor angiotensin II have considerable value as antihypertensive agents. For example, captopril and enalapril are clinically useful as inhibitors of angiotensin-converting enzyme. This has encouraged intense activity in the development of inhibitors of kidney renin, which is a very specific aspartic proteinase catalysing the first and rate limiting step in the conversion of angiotensinogen to angiotensin II. The most effective inhibitors such as H-142 and L-363,564 have used non-hydrolysable analogues of the proposed transition state, and partial sequences of angiotensinogen (Table 1). H-142 is effective in lowering blood pressure in humans but has no significant effect on other aspartic proteinases such as pepsin in the human body (Table 1). At present there are no crystal structures available for human or mouse renins although three-dimensional models demonstrate close structural similarity to other spartic proteinases. We have therefore determined by X-ray analysis the three-dimensional structures of H-142 and L-363,564 complexed with the aspartic proteinase endothiapepsin, which binds these inhibitors with affinities not greatly different from those measured against human renin (Table 1). The structures of these complexes and of that between endothiapepsin and the general aspartic proteinase inhibitor, H-256 (Table 1) define the common hydrogen bonding schemes that allow subtle differences in side-chain orientations and in the positions of the transition state analogues with respect to the active-site aspartates.

Knowledge-based prediction of protein structures and the design of novel molecules

Prediction of the tertiary structures of proteins may be carried out using a knowledge-based approach. This depends on identification of analogies in secondary structures, motifs, domains or ligand interactions between a protein to be modelled and those of known three-dimensional structures. Such techniques are of value in prediction of receptor structures to aid the design of drugs, herbicides or pesticides, antigens in vaccine design, and novel molecules in protein engineering.

Beta-hairpin families in globular proteins

Beta-hairpins, one of the simplest supersecondary structures, are widespread in globular proteins, and have often been suggested as possible sites for nucleation. Here we consider the conformation and sequences of the loop regions of beta-hairpins by analysing proteins of known structure. We find that the 'tight' beta-hairpins, classified by the length and conformations of their loop regions, form distinct families and that the loop regions of the family members have sequences which are characteristic of that family. The two-residue hairpin loops include almost entirely I' or II' beta-turns, in contrast to the general preference for type I and type II turns. These findings are being used to help define templates or consensus sequences to be incorporated into our existing supersecondary structure prediction algorithm. This information can also be used in model-building homologous proteins.

Computer graphics modelling and the specificity of renins

Mouse submaxillary, mouse kidney and human renins have been modelled using the three-dimensional structure of the homologous endothiapepsin, which has been defined by high-resolution X-ray analysis, and the amino acid sequences derived from protein, complementary DNA or gene sequencing. These computer graphics studies have shown that all renins can adopt three-dimensional structures similar to those of other aspartic proteinases with small insertions and deletions at the surface and often at beta-turns. The catalytically essential aspartates lie at the centre of a deep and extended cleft, and differences in the subsites, especially at S3' have been correlated with the known specificities of the renins. The models have shown that various residues on the surfaces of the renins adjacent to the active site cleft may play a part in recognizing and binding angiotensinogen. Reasons for a neutral pH optima have also been suggested.

Computer graphics modelling of human renin. Specificity, catalytic activity and intron-exon junctions

A model has been constructed using computer graphics for human renin based on the sequence derived from that of the gene and the 3-dimensional structure defined at high resolution for other homologous aspartic proteinases. Human renin can adopt a 3-dimensional structure close to that of other aspartic proteinases, in which amino acids corresponding to intron-exon junctions in the gene are at surface regions in the 3-dimensional structure. As expected, the essential catalytic residues are retained and the nearby residue 304 is alanine as in the mouse sequence, supporting the idea that Asp 304 of other aspartic proteinases may contribute to the low pH of their optimal activity. There are interesting differences at subsite S3' which may contribute to the specificity of human renin. Certain residues at the surface of the enzyme adjacent to the active site cleft are unique to renins and may play a role in recognition and binding of angiotensinogen.

Three-dimensional structure, specificity and catalytic mechanism of renin

Renin is an aspartyl proteinase that catalyses the first, and rate-limiting, step in the conversion of angiotensinogen to the hormone angiotensin II. The catalysis is highly specific, and plays an important physiological part in the regulation of blood pressure. For this reason inhibitors of renin are of potential value in the treatment of certain forms of hypertension. Although progress has been made in the design of inhibitors for clinical use by modification of angiotensinogen sequences, and as pepstatin analogues or with reduced peptide bonds, we have now provided the basis for a more rational approach by the use of interactive computer graphics techniques to build a three-dimensional model of renin. The model is based on the three-dimensional structure of endothia pepsin and the primary structure of mouse renin, which is very similar to that of the human enzyme. We show that renin may have a three-dimensional structure similar to that of other aspartyl proteinases.

Amino and carboxy-terminal regions in globular proteins

The structural, dynamic and functional aspects of amino and carboxy-terminal regions in proteins of known structure have been analysed. Terminal regions are usually located on the surface of the protein, accessible to solvent, and are often flexible. There is a significant preference for terminal regions in single domain proteins, and within individual domains of larger proteins, to be in close proximity. This partially reflects the compact globular nature of proteins, but the preference for spatial proximity is stronger in native proteins than in randomly generated structures. In addition in multi-domain and multi-subunit proteins we find that the terminal regions are commonly involved in the interface between domains and subunits. In the 18 multi-domain structures analysed, 19 terminal regions provide a link between domains. Subunit links are also frequently observed. In contrast, the distribution of active site residues along the sequence, indicates that the terminal regions are less frequently involved in activity. These data suggest that in many globular proteins the terminal regions fulfil a structural role, stabilizing the tertiary or quaternary structure to provide a framework for the active site.