Search

This paper explores the structural continuum in CATH and the extent to which superfamilies adopt distinct folds. Although most superfamilies are structurally conserved, in some of the most highly populated superfamilies (4% of all superfamilies) there is considerable structural divergence. While relatives share a similar fold in the evolutionary conserved core, diverse elaborations to this core can result in significant differences in the global structures. Applying similar protocols to examine the extent to which structural overlaps occur between different fold groups, it appears this effect is confined to just a few architectures and is largely due to small, recurring super-secondary motifs (e.g., alphabeta-motifs, alpha-hairpins). Although 24% of superfamilies overlap with superfamilies having different folds, only 14% of nonredundant structures in CATH are involved in overlaps. Nevertheless, the existence of these overlaps suggests that, in some regions of structure space, the fold universe should be seen as more continuous.

With many genomes now sequenced, computational annotation methods to characterize genes and proteins from their sequence are increasingly important. The BioSapiens Network has developed tools to address all stages of this process, and here we review progress in the automated prediction of protein function based on protein sequence and structure.

Motivation: Current computational methods for the prediction of function from structure are restricted to the detection of similarities and subsequent transfer of functional annotation. In a significant minority of cases, global sequence or structural (fold) similarities do not provide clues about protein function. In these cases, one alternative is to detect local binding site similarities. These may still reflect more distant evolutionary relationships as well as unique physico-chemical constraints necessary for binding similar ligands, thus helping pinpoint the function. In the present work, we ask the following question: is it possible to discriminate within a dataset of non-homologous proteins those that bind similar ligands based on their binding site similarities?, Methods: We implement a graph-matching-based method for the detection of 3D atomic similarities introducing some simplifications that allow us to extend its applicability to the analysis of large allatom binding site models. This method, called IsoCleft, does not require atoms to be connected either in sequence or space. We apply the method to a cognate-ligand bound dataset of non-homologous proteins. We define a family of binding site models with decreasing knowledge about the identity of the ligand-interacting atoms to uncouple the questions of predicting the location of the binding site and detecting binding site similarities. Furthermore, we calculate the individual contributions of binding site size, chemical composition and geometry to prediction performance., Results: We find that it is possible to discriminate between different ligand-binding sites. In other words, there is a certain uniqueness in the set of atoms that are in contact to specific ligand scaffolds. This uniqueness is restricted to the atoms in close proximity of the ligand in which case, size and chemical composition alone are sufficient to discriminate binding sites. Discrimination ability decreases with decreasing knowledge about the identity of the ligand-interacting binding site atoms. The decrease is quite abrupt when considering size and chemical composition alone, but much slower when including geometry. We also observe that certain ligands are easier to discriminate. Interestingly, the subset of binding site atoms belonging to highly conserved residues is not sufficient to discriminate binding sites, implying that convergently evolved binding sites arrived at dissimilar solutions., Availability: IsoCleft can be obtained from the authors.

We propose that a necessary condition for a protein to be soluble is the absence of large hydrophobic patches on its solvent-accessible surface, which can cause aggregation to occur. We note that the polar nature of the backbone of all amino acids guarantees a minimum polar content and hence can interrupt such patches. As a result, a carefully conserved detailed atomic placement of residues on the protein surface is not necessary for solubility. In order to demonstrate this, we construct a measure based on the average hydrophobicity of a simply defined patch. We use this measurement to compare surfaces that exhibit a clear difference in their solubility properties, namely, a) the solvent accessible surfaces for a set of homo-dimers and the surfaces buried in their interfaces and b) for a set of monomers the surfaces of fragments of secondary structure which are solvent accessible/inaccessible. Having demonstrated a difference in the first set of distributions, we characterize the solvent accessible surfaces of monomeric proteins. To test if cooperative behavior occurs between the atoms for these surfaces, we construct a set of randomized surfaces, which obey a very simple stereochemical constraint. We find that the observed and randomized distributions are much more similar than the previous sets we examined. This implies that while surfaces of soluble proteins must have sufficient polar content, the relative placement of atoms of one amino acid with respect to the atoms of neighboring amino acid need not be finely tuned, which provides an innate robustness for protein design and folding., ((c) 2005 Wiley Periodicals, Inc.)

The CATH database of protein domain structures (http://www.biochem.ucl.ac.uk/bsm/cath/) currently contains 43,229 domains classified into 1467 superfamilies and 5107 sequence families. Each structural family is expanded with sequence relatives from GenBank and completed genomes, using a variety of efficient sequence search protocols and reliable thresholds. This extended CATH protein family database contains 616,470 domain sequences classified into 23,876 sequence families. This results in the significant expansion of the CATH HMM model library to include models built from the CATH sequence relatives, giving a 10% increase in coverage for detecting remote homologues. An improved Dictionary of Homologous superfamilies (DHS) (http://www.biochem.ucl.ac.uk/bsm/dhs/) containing specific sequence, structural and functional information for each superfamily in CATH considerably assists manual validation of homologues. Information on sequence relatives in CATH superfamilies, GenBank and completed genomes is presented in the CATH associated DHS and Gene3D resources. Domain partnership information can be obtained from Gene3D (http://www.biochem.ucl.ac.uk/bsm/cath/Gene3D/). A new CATH server has been implemented (http://www.biochem.ucl.ac.uk/cgi-bin/cath/CathServer.pl) providing automatic classification of newly determined sequences and structures using a suite of rapid sequence and structure comparison methods. The statistical significance of matches is assessed and links are provided to the putative superfamily or fold group to which the query sequence or structure is assigned.

Motivation: The solubility of a protein is crucial for its function and is therefore an evolutionary constraint. As the solubility of a protein is related to the distribution of polar and hydrophobic residues on its solvent accessible surface, such a constraint should provide a valuable insight into the evolution of protein surfaces. We examine how the surfaces of proteins have evolved by considering how the average hydrophobicities of patches of surface residues vary across homologous proteins. We derive distributions for the average hydrophobicity/philicity of surface patches at a residue-based level-which we refer to as the residue hydrophobic density. This is computed for a set of 28 monomeric proteins and their homologues. The resulting distributions are compared with a set of randomized sequences, with the same residue content., Results: We find that the patches, involving typically more than 10 residues, maintain a more hydrophilic surface than one would expect from a random substitution model, indicating a cooperative behaviour for these surfaces residues in terms of this single variable., Supplementary Information: Additional plots for all of the proteins examined in this paper can be found at: http://www.ebi.ac.uk/~shanahan/PCon/index.html

This paper reports a graph-theoretic program, GRATH, that rapidly, and accurately, matches a novel structure against a library of domain structures to find the most similar ones. GRATH generates distributions of scores by comparing the novel domain against the different types of folds that have been classified previously in the CATH database of structural domains. GRATH uses a measure of similarity that details the geometric information, number of secondary structures and number of residues within secondary structures, that any two protein structures share. Although GRATH builds on well established approaches for secondary structure comparison, a novel scoring scheme has been introduced to allow ranking of any matches identified by the algorithm. More importantly, we have benchmarked the algorithm using a large dataset of 1702 non-redundant structures from the CATH database which have already been classified into fold groups, with manual validation. This has facilitated introduction of further constraints, optimization of parameters and identification of reliable thresholds for fold identification. Following these benchmarking trials, the correct fold can be identified with the top score with a frequency of 90%. It is identified within the ten most likely assignments with a frequency of 98%. GRATH has been implemented to use via a server (http://www.biochem.ucl.ac.uk/cgi-bin/cath/Grath.pl). GRATH's speed and accuracy means that it can be used as a reliable front-end filter for the more accurate, but computationally expensive, residue based structure comparison algorithm SSAP, currently used to classify domain structures in the CATH database. With an increasing number of structures being solved by the structural genomics initiatives, the GRATH server also provides an essential resource for determining whether newly determined structures are related to any known structures from which functional properties may be inferred.

The CATH database of protein domain structures (http://www.biochem.ucl.ac.uk/bsm/cath_new) currently contains 34 287 domain structures classified into 1383 superfamilies and 3285 sequence families. Each structural family is expanded with domain sequence relatives recruited from GenBank using a variety of efficient sequence search protocols and reliable thresholds. This extended resource, known as the CATH-protein family database (CATH-PFDB) contains a total of 310 000 domain sequences classified into 26 812 sequence families. New sequence search protocols have been designed, based on these intermediate sequence libraries, to allow more regular updating of the classification. Further developments include the adaptation of a recently developed method for rapid structure comparison, based on secondary structure matching, for domain boundary assignment. The philosophy behind CATHEDRAL is the recognition of recurrent folds already classified in CATH. Benchmarking of CATHEDRAL, using manually validated domain assignments, demonstrated that 43% of domains boundaries could be completely automatically assigned. This is an improvement on a previous consensus approach for which only 10-20% of domains could be reliably processed in a completely automated fashion. Since domain boundary assignment is a significant bottleneck in the classification of new structures, CATHEDRAL will also help to increase the frequency of CATH updates.

How have proteins taken on the remarkable diversity of biochemical and physiological functions necessary to create and maintain complex organisms? The majority of proteins are organized hierarchically into families and superfamilies, reflecting an ancient and continuing process of gene duplication and divergence. The techniques of molecular phylogenetics, developed to recover the nested hierarchy of taxa from character information in their gene sequences, can also reconstruct the evolutionary relationships among genes and provide a conceptual foundation for comparative evolutionary analysis of proteins and their functions. In this review, I outline the application of phylogenetic approaches to issues in gene family studies, beginning with the inference of phylogeny and the assessment of the two types of homology by which genes in a family can be related: orthology (common descent from a cladogenetic event) and paralogy (common descent from a gene duplication event). I show how the phylogenetic approach makes possible novel kinds of comparative analysis, including detection of exon shuffling, reconstruction of the evolutionary diversification of gene families, tracing of evolutionary change in protein function at the amino acid level, and prediction of structure-function relationships. A marriage of the principles of phylogenetic systematics with the copious sequence data being generated by molecular biology and genomics promises unprecedented insights into the nature of biological organization and the historical processes that created it.

We describe the design and implementation of a fast topology-based method for protein structure comparison. The approach uses the TOPS topological representation of protein structure, aligning two structures using a common discovered pattern and generating measure of distance derived from an insert score. Heavy use is made of a constraint-based pattern-matching algorithm for TOPS diagrams that we have designed and described elsewhere (Bioinformatics 15(4) (1999) 317). The comparison system is maintained at the European Bioinformatics Institute and is available over the Web at tops.ebi.ac.uk/tops. Users submit a structure description in Protein Data Bank (PDB) format and can compare it with structures in the entire PDB or a representative subset of protein domains, receiving the results by email.

The molecular recognition and discrimination of adenine and guanine ligand moieties in complexes with proteins have been studied using empirical observations on carefully selected crystal structures. The distribution of protein folds that bind these purines has been found to differ significantly from that across the whole PDB, but the most populated architectures and folds are also the most common in three genomes from the three different domains of life. The protein environments around the two nucleic acid bases were significantly different, in terms of the propensities of amino acid residues to be in the binding site, as well as their propensities to form hydrogen bonds to the bases. Plots of the distribution of protein atoms around the two purines clearly show different clustering of hydrogen bond donors and acceptors opposite complimentary acceptors and donors in the rings, with hydrophobic areas below and above the rings. However, the clustering pattern is fuzzy, reflecting the variety of ways that proteins have evolved to recognise the same molecular moiety. Furthermore, an analysis of the conservation of residues in the protein chains binding guanine shows that residues in contact with the base are in general better conserved than the rest of the chain.

Some crystal contacts are biologically relevant, most are not. We assess the utility of combining measures of size and conservation to discriminate between biological and non-biological contacts. Conservation and size information is calculated for crystal contacts in 53 families of homodimers and 65 families of monomers. Biological contacts are shown to be usually conserved and typically the largest contact in the crystal. A range of neural networks accepting different combinations and encodings of this information is used to answer the following questions: (1) is a given crystal contact biological, and (2) given all crystal contacts in a homodimer, which is the biological one? Predictions for (1) are performed on both homodimer and monomer datasets. The best performing neural network combined size and conservation inputs. For the homodimers, it correctly classified 48 out of 53 biological contacts and 364 out of 366 non-biological contacts, giving a combined accuracy of 98.3 %. A more robust performance statistic, the phi-coefficient, which accounts for imbalances in the dataset, gave a value of 0.92. Taking all 535 non-biological contacts from the 65 monomers, this predictor made erroneous classifications only 4.3 % of the time. Predictions for (2) were performed on homodimers only. The best performing network achieved a prediction accuracy of 98.1 % using size information alone. We conclude that in answering question (1) size and conservation combined discriminate biological from non-biological contacts better than either measure alone. For answering question (2), we conclude that in our dataset size is so powerful a discriminant that conservation adds little predictive benefit., (Copyright 2001 Academic Press.)

The recent growth in protein databases has revealed the functional diversity of many protein superfamilies. We have assessed the functional variation of homologous enzyme superfamilies containing two or more enzymes, as defined by the CATH protein structure classification, by way of the Enzyme Commission (EC) scheme. Combining sequence and structure information to identify relatives, the majority of superfamilies display variation in enzyme function, with 25 % of superfamilies in the PDB having members of different enzyme types. We determined the extent of functional similarity at different levels of sequence identity for 486,000 homologous pairs (enzyme/enzyme and enzyme/non-enzyme), with structural and sequence relatives included. For single and multi-domain proteins, variation in EC number is rare above 40 % sequence identity, and above 30 %, the first three digits may be predicted with an accuracy of at least 90 %. For more distantly related proteins sharing less than 30 % sequence identity, functional variation is significant, and below this threshold, structural data are essential for understanding the molecular basis of observed functional differences. To explore the mechanisms for generating functional diversity during evolution, we have studied in detail 31 diverse structural enzyme superfamilies for which structural data are available. A large number of variations and peculiarities are observed, at the atomic level through to gross structural rearrangements. Almost all superfamilies exhibit functional diversity generated by local sequence variation and domain shuffling. Commonly, substrate specificity is diverse across a superfamily, whilst the reaction chemistry is maintained. In many superfamilies, the position of catalytic residues may vary despite playing equivalent functional roles in related proteins. The implications of functional diversity within supefamilies for the structural genomics projects are discussed. More detailed information on these superfamilies is available at http://www.biochem.ucl.ac.uk/bsm/FAM-EC/., (Copyright 2001 Academic Press.)

In order to support the structural genomic initiatives, both by rapidly classifying newly determined structures and by suggesting suitable targets for structure determination, we have recently developed several new protocols for classifying structures in the CATH domain database (http://www.biochem.ucl.ac.uk/bsm/cath). These aim to increase the speed of classification of new structures using fast algorithms for structure comparison (GRATH) and to improve the sensitivity in recognising distant structural relatives by incorporating sequence information from relatives in the genomes (DomainFinder). In order to ensure the integrity of the database given the expected increase in data, the CATH Protein Family Database (CATH-PFDB), which currently includes 25,320 structural domains and a further 160,000 sequence relatives has now been installed in a relational ORACLE database. This was essential for developing more rigorous validation procedures and for allowing efficient querying of the database, particularly for genome analysis. The associated Dictionary of Homologous Superfamilies [Bray,J.E., Todd,A.E., Pearl,F.M.G., Thornton,J.M. and Orengo,C.A. (2000) Protein Eng., 13, 153-165], which provides multiple structural alignments and functional information to assist in assigning new relatives, has also been expanded recently and now includes information for 903 homologous superfamilies. In order to improve coverage of known structures, preliminary classification levels are now provided for new structures at interim stages in the classification protocol. Since a large proportion of new structures can be rapidly classified using profile-based sequence analysis [e.g. PSI-BLAST: Altschul,S.F., Madden,T.L., Schaffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. (1997) Nucleic Acids Res., 25, 3389-3402], this provides preliminary classification for easily recognisable homologues, which in the latest release of CATH (version 1.7) represented nearly three-quarters of the non-identical structures.

Evolutionary information derived from the large number of available protein sequences and structures could powerfully guide both analysis and prediction of protein-protein interfaces. To test the relevance of this information, we assess the conservation of residues at protein-protein interfaces compared with other residues on the protein surface. Six homodimer families are analyzed: alkaline phosphatase, enolase, glutathione S-transferase, copper-zinc superoxide dismutase, Streptomyces subtilisin inhibitor, and triose phosphate isomerase. For each family, random simulation is used to calculate the probability (P value) that the level of conservation observed at the interface occurred by chance. The results show that interface conservation is higher than expected by chance and usually statistically significant at the 5% level or better. The effect on the P values of using different definitions of the interface and of excluding active site residues is discussed.

This review presents a summary of current approaches to extract functional information from structural data on proteins and their complexes. While structural homologs may reveal possible biochemical functions (which may be hidden at the sequence level), elucidating the exact biological role of a protein in vivo will only be possible by including other results, such as data on expression and localization.

Scores calculated from intermolecular contacts of proteins in the crystalline state are used to differentiate monomeric and homodimeric proteins, by classification into two categories separated by a cut-off score value. The generalized classification error is estimated by using bootstrap re-sampling on a nonredundant set of 172 water-soluble proteins whose prevalent quaternary state in solution is known to be either monomeric or homodimeric. A statistical potential, based on atom-pair frequencies across interfaces observed with homodimers, is found to yield an error rate of 12.5%. This indicates a small but significant improvement over the measure of solvent accessible surface area buried in the contact interface, which achieves an error rate of 15.4%. A further modification of the latter parameter relating the two most extensive contacts of the crystal results in an even lower error rate of 11.1%.

In this paper we survey a number of functional classification schemes applicable to genomes. We present the concepts of depth, breadth and resolution as descriptors of the schemes' scope and architecture and compare selected classifications according to these criteria. We also generate a 'Combined Scheme' against which we map six classifications which we believe are representative of the range currently available. The mapping allows the generation of 'FuncWheels', which are graphical representations of hierarchical classification schemes. They are used to illustrate similarities and differences in functional space coverage. This survey highlights many issues related to the design and implementation of gene product functional classifications, which are discussed in the light of emerging 'second-generation' schemes.

A consensus approach has been developed for identifying distant structural homologues. This is based on the CATH Dictionary of Homologous Superfamilies (DHS), a database of validated multiple structural alignments annotated with consensus functional information for evolutionary protein superfamilies (URL: http://www. biochem.ucl.ac.uk/bsm/dhs). Multiple structural alignments have been generated for 362 well-populated superfamilies in the CATH structural domain database and annotated with secondary structure, physicochemical properties, functional sequence patterns and protein-ligand interaction data. Consensus functional information for each superfamily includes descriptions and keywords extracted from SWISS-PROT and the ENZYME database. The Dictionary provides a powerful resource to validate, examine and visualize key structural and functional features of each homologous superfamily. The value of the DHS, for assessing functional variability and identifying distant evolutionary relationships, is illustrated using the pyridoxal-5'-phosphate (PLP) binding aspartate aminotransferase superfamily. The DHS also provides a tool for examining sequence-structure relationships for proteins within each fold group.

An analysis of the characteristic properties of sugar binding sites was performed on a set of 19 sugar binding proteins. For each site six parameters were evaluated: solvation potential, residue propensity, hydrophobicity, planarity, protrusion and relative accessible surface area. Three of the parameters were found to distinguish the observed sugar binding sites from the other surface patches. These parameters were then used to calculate the probability for a surface patch to be a carbohydrate binding site. The prediction was optimized on a set of 19 non-homologous carbohydrate binding structures and a test prediction was carried out on a set of 40 protein-carbohydrate complexes. The overall accuracy of prediction achieved was 65%. Results were in general better for carbohydrate-binding enzymes than for the lectins, with a rate of success of 87%.

The physical and chemical properties of domain-domain interactions have been analysed in two-domain proteins selected from the protein classification, CATH. The two-domain structures were divided into those derived from (i) monomeric proteins, or (ii) oligomeric or complexed proteins. The size, polarity, hydrogen bonding and packing of the intra-chain domain interface were calculated for both sets of two-domain structures. The results were compared with inter-chain interface parameters from permanent and non-obligate protein-protein complexes. In general, the intra-chain domain and inter-chain interfaces were remarkably similar. Many of the intra-chain interface properties are intermediate between those calculated for permanent and non-obligate inter-chain complexes. Residue interface propensities were also found to be very similar, with hydrophobic residues playing a major role, together with positively charged arginine residues. In addition, the residue composition of the domain interfaces were found to be more comparable with domain surfaces than domain cores. The implications of these results for domain swapping and protein folding are discussed.

We report the latest release (version 1.6) of the CATH protein domains database (http://www.biochem.ucl. ac.uk/bsm/cath ). This is a hierarchical classification of 18 577 domains into evolutionary families and structural groupings. We have identified 1028 homo-logous superfamilies in which the proteins have both structural, and sequence or functional similarity. These can be further clustered into 672 fold groups and 35 distinct architectures. Recent developments of the database include the generation of 3D templates for recognising structural relatives in each fold group, which has led to significant improvements in the speed and accuracy of updating the database and also means that less manual validation is required. We also report the establishment of the CATH-PFDB (Protein Family Database), which associates 1D sequences with the 3D homologous superfamilies. Sequences showing identifiable homology to entries in CATH have been extracted from GenBank using PSI-BLAST. A CATH-PSIBLAST server has been established, which allows you to scan a new sequence against the database. The CATH Dictionary of Homologous Superfamilies (DHS), which contains validated multiple structural alignments annotated with consensus functional information for evolutionary protein superfamilies, has been updated to include annotations associated with sequence relatives identified in GenBank. The DHS is a powerful tool for considering the variation of functional properties within a given CATH superfamily and in deciding what functional properties may be reliably inherited by a newly identified relative.

The evolution of proteins and their functions is reviewed from a structural perspective in the light of the current database. Protein domain families segregate unequally between the three major classes, the 32 different architectures and almost 700 folds observed to date. We find that the number of new topologies is still increasing, although 25 new structures are now determined for each new topology. The corresponding analysis and classification of function is only just beginning, fuelled by the genome data. The structural data revealed unexpected conservations and divergence of function both within and between families. The next five years will see the compilation of a definitive dictionary of protein families and their related functions, based on structural data which reveals relationships hidden at the sequence level. Such information will provide the foundation to build a better understanding of the molecular basis of biological complexity and hopefully to facilitate rational molecular design., (Copyright 1999 Academic Press.)

Several databases of protein structural families now exist-organised according to both evolutionary relationships and common folding arrangements. Although these lag behind sequence databases in size, the prospect of structural genomics initiatives means that they may soon include representatives of many of the sequence families. To some extent, functional information can be derived from structural similarity. For some structural families, their function is highly conserved, whereas, for others, it can only be inherited or derived on the basis of additional information (e.g. sequence patterns, common residue clusters and characteristic surface properties).

A program is described for automatically generating schematic linear representations of protein chains in terms of their structural domains. The program requires the co-ordinates of the chain, the domain assignment, PROSITE information and a file listing all intermolecular interactions in the protein structure. The output is a PostScript file in which each protein is represented by a set of linked boxes, each box corresponding to all or part of a structural domain. PROSITE motifs and residues involved in ligand interactions are highlighted. The diagrams allow immediate visualization of the domain arrangement within a protein chain, and by providing information on sequence motifs, and metal ion, ligand and DNA binding at the domain level, the program facilitates detection of remote evolutionary relationships between proteins.

An interactive program is described for calculating the second virial coefficient contribution to the thermodynamic nonideality of solutions of rigid macromolecules based on their triaxial dimensions. The FORTRAN-77 program, available in precompiled form for the PC, is based on theory for the covolume of triaxial ellipsoid particles [Rallison, J. M., and S.E Harding. (1985). J. Colloid Interface Sci. 103:284-289]. This covolume has the potential to provide a magnitude for the second virial coefficient of macromolecules bearing no net charge. Allowance for a charge-charge contribution is made via an expression based on Debye-Hückel theory and uniform distribution of the net charge over the surface of a sphere with dimensions governed by the Stokes radius of the macromolecule. Ovalbumin, ribonuclease A, and hemoglobin are used as model systems to illustrate application of the COVOL routine.

A systematic variation with resolution of the mean values of the gauche-, trans and gauche+ chi1 rotamers in protein structures determined by X-ray crystallography has been observed. Further analysis revealed that these correlations differ considerably between residue types, being highly significant for some residue types (e.g. Ser, Thr, Leu, Lys) and absent for others (e.g. aromatics). For the individual residue types which exhibited the trend most strongly, these changes were accompanied by corresponding systematic variations in the percentage relative populations in the three energy wells. Examination of a uniformly sized subset of monomers showed that this effect, while attenuated, was still present, and was thus not entirely a consequence of the change in size and surface area which also correlates with resolution. An analysis of B values in the disfavoured high-energy barrier region between the rotameric wells showed a pronounced tendency towards larger than average values. As a plausible hypothesis, it is suggested here that these observations can be accounted for by the presence of multiple rotameric states. The averaged electron density produced by dual occupancy at low resolution giving an averaged conformation is resolved at high resolution into its individual components.

A detailed analysis of the DNA-binding sites of 26 proteins is presented using data from the Nucleic Acid Database (NDB) and the Protein Data Bank (PDB). Chemical and physical properties of the protein-DNA interface, such as polarity, size, shape, and packing, were analysed. The DNA-binding sites shared common features, comprising many discontinuous sequence segments forming hydrophilic surfaces capable of direct and water-mediated hydrogen bonds. These interface sites were compared to those of protein-protein binding sites, revealing them to be more polar, with many more intermolecular hydrogen bonds and buried water molecules than the protein-protein interface sites. By looking at the number and positioning of protein residue-DNA base interactions in a series of interaction footprints, three modes of DNA binding were identified (single-headed, double-headed and enveloping). Six of the eight enzymes in the data set bound in the enveloping mode, with the protein presenting a large interface area effectively wrapped around the DNA.A comparison of structural parameters of the DNA revealed that some values for the bound DNA (including twist, slide and roll) were intermediate of those observed for the unbound B-DNA and A-DNA. The distortion of bound DNA was evaluated by calculating a root-mean-square deviation on fitting to a canonical B-DNA structure. Major distortions were commonly caused by specific kinks in the DNA sequence, some resulting in the overall bending of the helix. The helix bending affected the dimensions of the grooves in the DNA, allowing the binding of protein elements that would otherwise be unable to make contact. From this structural analysis a preliminary set of rules that govern the bending of the DNA in protein-DNA complexes, are proposed., (Copyright 1999 Academic Press.)

Motivation: TOPS cartoons are a schematic ion of protein three-dimensional structures in two dimensions, and are used for understanding and manual comparison of protein folds. Recently, an algorithm that produces the cartoons automatically from protein structures has been devised and cartoons have been generated to represent all the structures in the structural databank. There is now a need to be able to define target topological patterns and to search the database for matching domains., Results: We have devised a formal language for describing TOPS diagrams and patterns, and have designed an efficient algorithm to match a pattern to a set of diagrams. A pattern-matching system has been implemented, and tested on a database derived from all the current entries in the Protein Data Bank (15,000 domains). Users can search on patterns selected from a library of motifs or, alternatively, they can define their own search patterns., Availability: The system is accessible over the Web at http://tops.ebi.ac.uk/tops

Pattern matches for each of the sequence patterns in PROSITE, a database of sequence patterns, were searched in all protein sequences in the Brookhaven Protein Data Bank (PDB). The three-dimensional structures of the pattern matches for the 20 patterns with the largest numbers of hits were analysed. We found that the true positives have a common three-dimensional structure for each pattern; the structures of false positives, found for six of the 20 patterns, were clearly different from those of the true positives. The results suggest that the true pattern matches each have a characteristic common three-dimensional structure, which could be used to create a template to define a three-dimensional functional pattern., (Copyright 1998 Academic Press.)

We report the latest release (version 1.4) of the CATH protein domains database (http://www.biochem.ucl.ac.uk/bsm/cath). This is a hierarchical classification of 13 359 protein domain structures into evolutionary families and structural groupings. We currently identify 827 homologous families in which the proteins have both structual similarity and sequence and/or functional similarity. These can be further clustered into 593 fold groups and 32 distinct architectures. Using our structural classification and associated data on protein functions, stored in the database (EC identifiers, SWISS-PROT keywords and information from the Enzyme database and literature) we have been able to analyse the correlation between the 3D structure and function. More than 96% of folds in the PDB are associated with a single homologous family. However, within the superfolds, three or more different functions are observed. Considering enzyme functions, more than 95% of clearly homologous families exhibit either single or closely related functions, as demonstrated by the EC identifiers of their relatives. Our analysis supports the view that determining structures, for example as part of a 'structural genomics' initiative, will make a major contribution to interpreting genome data.

The CATH database of protein domain structures classifies structures according to their (C)lass, (A)rchitecture, (T)opology or fold and (H)omologous family (http://www.biochem.ucl.ac.uk/bsm/cath). Although the protocol used is mostly automatic, manual inspection is used to check assignments at some critical stages, such as the detection of very distantly related homologues and anologues and the assignment of novel architectures. Described in this article is a recently established facility to search the database with the coordinates of a newly determined structure. The CATH server first locates domain boundaries and then uses automatic sequence and structure comparison methods to assign this new structure to one or more of the domain families within CATH. Diagnostic reports are generated, together with multiple structural alignments for close relatives. The Server can be accessed over the World Wide Web (WWW) and mirror sites are planned to improve access.

In a similar manner to sequence database searching, it is also possible to compare three-dimensional protein structure. Such methods can be extremely useful because a structural similarity may represent a distant evolutionary relationship that is undetectable by sequence analysis. In this review, we summarise the most popular structure comparison methods, show how they can be used for database searching, and then describe some of the most advanced attempts to develop comprehensive protein structure classifications. With such data, it is possible to identify distant evolutionary relationships, provide libraries of unique folds for structure prediction, estimate the total number of folds that exist, and investigate the preference for certain types of structures over others.

The objectives of this study were to determine the time course of the stimulatory effect of dexamethasone on serum leptin and whether it depends on food intake. Dexamethasone (4mg) was administered I.V. over 1 minute to healthy human volunteers (n=8) under fasting and feeding conditions (2000 kcal given at three meals over 7 hours). At 10 hours, serum leptin levels were increased only in the fed subjects (delta leptin 10.6+/-1.6 vs -2.4+/-1.9 ng/ml, p=0.01, n=8). To assess the interactive effect of food and dexamethasone on serum leptin, a subgroup (n=4) was studied under 4 conditions: 1) dexamethasone/fast; 2) dexamethasone/food; 3) saline/fast; 4) saline/food. Serum leptin declined from baseline under the fasting conditions, with or without dexamethasone. Feeding prevented the drop in serum leptin. In the dexamethasone/food condition, leptin levels rose from baseline after 7 hours and doubled after 10 hours (p<0.05). The rise in serum leptin was significantly greater in the food/dexamethasone condition compared to all other conditions (p<0.05). In summary, dexamethasone has no independent effect on serum leptin in the absence of food intake. Rather, dexamethasone appears to potentiate the food-induced increase in serum leptin. This synergism may be mediated by insulin and/or other factors associated with food ingestion.

With the aim of bridging the gap between protein sequence and structural analyses, we have developed a tool to aid the identification of new protein sequences by recognizing distant homologues using structural information. The tool generates sequence annotated by structure (SAS) files, applying structural information derived from structural analyses to a given protein sequence. A World Wide Web interface allows a given sequence to be submitted either for structural annotation or, where its structure is unknown, for search and alignment against sequences of known structure. In both cases, SAS will colour residues in the sequence of known structure according to a selection of properties, including secondary structure, interatomic contacts and active site information. SAS can also be used to inspect properties of a single structure.

The growing number of protein structures solved at atomic resolution holds the promise of further improvements in geometry-based validation parameters. Additionally, the estimated standard uncertainties of the atomic coordinates have been computed for a number of X-ray structures, providing a measure of the coordinate precision. In NMR spectroscopy, a measure analogous to the crystallographic R-factor has been developed.

Background: The recent rapid increase in the number of available three-dimensional protein structures has further highlighted the necessity to understand the relationship between biological function and structure. Using structural classification schemes such as SCOP, CATH and DALI, it is now possible to explore global relationships between protein fold and function, something which was previously impractical., Results: Using a relational database of CATH data we have generated fold distributions for arbitrary selections of proteins automatically. These distributions have been examined in the light of protein function and bound ligand. Different enzyme classes are not clearly reflected in distributions of protein class and architecture, whereas the type of bound ligand has a much more dramatic effect., Conclusions: The availability of structural classification data has enabled this novel overview analysis. We conclude that function at the top level of the EC number enzyme classification is not related to fold, as only a very few specific residues are actually responsible for enzyme activity. Conversely, the fold is much more closely related to ligand type.

Despite the importance of limited proteolysis in biological systems it is often difficult to rationalize why a proteinase hydrolyses a particular bond, given a simple sequence specificity alone. Understanding of the structural properties limiting the proteolysis represents a first step on the pathway to control and manipulation of this phenomena. An expanded set of nick-sites in proteins of known tertiary structure, cut by both narrow and broad specificity proteinases, has been generated yielding a robust data set of strictly limited sites. A critical evaluation of an expanded set of conformational parameters revealed a strong correlation with limited proteolytic sites, although they are only modest predictors in isolation. The overall predictive power is significantly improved when the conformational parameters are combined in a weighted predictive scheme that permits their relative importance to be compared via a Metropolis search protocol. A subset of the parameters performs equally well demonstrating the key determinants of susceptibility. The derived predictive algorithm has been made available via the internet. Its utility for predicting other surface-correlated features is also discussed.

The hormone leptin acts on the brain to regulate feeding, metabolism, and reproduction; however, its cellular targets and molecular mechanisms of action remain to be fully elucidated. The melanocortins, which are derived from the precursor proopiomelanocortin (POMC), are also implicated in the physiological regulation of body weight. POMC-containing neurons express the leptin receptor, and thus it is conceivable that the POMC gene itself may be part of the signaling pathway involved in leptin's action on the brain. Using in situ hybridization and computerized image analysis, we tested the hypothesis that the POMC gene is a target for regulation by leptin by comparing cellular levels of POMC mRNA in the hypothalamus among groups of leptin-deficient (ob/ob) mice, leptin-treated ob/ob mice, and wild-type controls. POMC mRNA levels were significantly reduced throughout the arcuate nucleus in vehicle-treated ob/ob mice relative to wild-type controls, whereas POMC mRNA levels in leptin-treated ob/ob mice were indistinguishable from wild-type controls. These observations suggest that one or more products of POMC serve as an integrative link between leptin and the central mechanisms governing body weight regulation and reproduction.

A method for defining and analysing a series of residue patches on the surface of protein structures is used to predict the location of protein-protein interaction sites. Each residue patch is analysed for six parameters; solvation potential, residue interface propensity, hydrophobicity, planarity, protrusion and accessible surface area. The method involves the calculation of a relative combined score that gives the probability of a surface patch forming protein-protein interactions. Predictions are made for the known structures of protomers from 28 homo-dimers, large protomers from 11 hetero-complexes, small protomers from 14 hetero-complexes, and antigens from six antibody-antigen complexes. The predictions are successful for 66% (39/59) of the structures and the remainder can usually be rationalized in terms of additional interaction sites., (Copyright 1997 Academic Press Limited.)

Protein-protein interaction sites in complexes of known structure are characterised using a series of parameters to evaluate what differentiates them from other sites on the protein surface. Surface patches are defined in protomers from a data set of 28 homo-dimers, 20 different hetero-complexes (segregated into large and small protomers), and antigens from six antibody-antigen complexes. Six parameters (solvation potential, residue interface propensity, hydrophobicity, planarity, protrusion and accessible surface area) are calculated for the observed interface patch and all other surface patches defined on each protein. A ranking of the observed interface, relative to all other possible patches, is calculated. With this approach it becomes possible to analyse the distribution of the rankings of all the observed patches, relative to all other surface patches, for each data set. For each type of complex, none of the parameters were definitive, but the majority showed trends for the observed interface to be distinguished from other surface patches., (Copyright 1997 Academic Press Limited.)

Background: Protein evolution gives rise to families of structurally related proteins, within which sequence identities can be extremely low. As a result, structure-based classifications can be effective at identifying unanticipated relationships in known structures and in optimal cases function can also be assigned. The ever increasing number of known protein structures is too large to classify all proteins manually, therefore, automatic methods are needed for fast evaluation of protein structures., Results: We present a semi-automatic procedure for deriving a novel hierarchical classification of protein domain structures (CATH). The four main levels of our classification are protein class (C), architecture (A), topology (T) and homologous superfamily (H). Class is the simplest level, and it essentially describes the secondary structure composition of each domain. In contrast, architecture summarises the shape revealed by the orientations of the secondary structure units, such as barrels and sandwiches. At the topology level, sequential connectivity is considered, such that members of the same architecture might have quite different topologies. When structures belonging to the same T-level have suitably high similarities combined with similar functions, the proteins are assumed to be evolutionarily related and put into the same homologous superfamily., Conclusions: Analysis of the structural families generated by CATH reveals the prominent features of protein structure space. We find that nearly a third of the homologous superfamilies (H-levels) belong to ten major T-levels, which we call superfolds, and furthermore that nearly two-thirds of these H-levels cluster into nine simple architectures. A database of well-characterised protein structure families, such as CATH, will facilitate the assignment of structure-function/evolution relationships to both known and newly determined protein structures.

The timing of puberty onset in mammals is tightly coupled to the animal's nutritional and metabolic state. We conducted two experiments to test the hypothesis that leptin acts as a metabolic signal for the onset of puberty. In the first experiment, we administered leptin (6.3 micrograms/g twice daily) to a group of normal prepubertal female rats and compared their rate of sexual maturation to that of two control groups. The group of leptin-treated animals and one group of control animals were allowed to eat ad lib, while the other group of control animals was pair-fed to the leptin-treated group. Food intake in the leptin-treated group was reduced to approximately 80% of the ad lib-fed control group, resulting in retarded growth in both leptin-treated and pair-fed animals. All measured indices of pubertal maturation-age at vaginal opening, age at first estrus, ovarian weight, ovulatory index (corpora lutea/ovarian section), uterine weight, and uterine cross-sectional area-were significantly delayed in the pair-fed group but not different between the leptin-treated group and ad lib-fed controls. The second experiment was similar to the first, except that both the leptin-treated group and the pair-fed group were fed at 70% of the ad lib-fed controls. Under these conditions, leptin only partially reversed the delay in sexual maturation, as reflected by the age at vaginal opening and first estrus. These results suggest that leptin is not the primary signal that initiates the onset of puberty but that instead, it acts in a permissive fashion, as a metabolic gate, to allow pubertal maturation to proceed-if and when metabolic resources are deemed adequate; moreover, these observations suggest that other metabolic factors, besides leptin, influence the timing of puberty onset under conditions of more severe dietary stress.

We have surveyed 393 N-termini of alpha-helices and 156 N-termini of 3(10)-helices in 85 high resolution, non-homologous protein crystal structures for N-cap side-chain rotamer preferences, hydrogen bonding patterns, and solvent accessibilities. We find very strong rotamer preferences that are unique to N-cap sites. The following rules are generally observed for N-capping in alpha-helices: Thr and Ser N-cap side chains adopt the gauche - rotamer, hydrogen bond to the N3 NH and have psi restricted to 164 +/- 8 degrees. Asp and Asn N-cap side chains either adopt the gauche - rotamer and hydrogen bond to the N3 NH with psi = 172 +/- 10 degrees, or adopt the trans rotamer and hydrogen bond to both the N2 and N3 NH groups with psi = 1-7 +/- 19 degrees. With all other N-caps, the side chain is found in the gauche + rotamer so that the side chain does not interact unfavorably with the N-terminus by blocking solvation and psi is unrestricted. An i, i + 3 hydrogen bond from N3 NH to the N-cap backbone C = O in more likely to form at the N-terminus when an unfavorable N-cap is present. In the 3(10)-helix Asn and Asp remain favorable N-caps as they can hydrogen bond to the N2 NH while in the trans rotamer; in contrast, Ser and Thr are disfavored as their preferred hydrogen bonding partner (N3 NH) is inaccessible. This suggests that Ser is the optimum choice of N-cap when alpha-helix formation is to be encouraged while 3(10)-helix formation discouraged. The strong energetic and structural preferences found for N-caps, which differ greatly from positions within helix interiors, suggest that N-caps should be treated explicitly in any consideration of helical structure in peptides or proteins.

The work described here is the result of a survey of the peptide omega angles in the Cambridge Structural Database of small molecules, which was carried out to establish "ideal" or "target" values for their distribution. We have shown that substantial deviations from planarity can be tolerated with a standard deviation in the angle of up to 6 degrees about a mean value for the trans peptide that is less than 180 degrees . The distortion can arise from pyramidalization at the amino nitrogen atom as well as simple twist about the peptide bond. We include an analysis of omega angles in the existing database of protein structure (PDB) and show that their distributions can depend on the refinement method used, but no correlation with resolution is evident. A surprising finding was a systematic variation of omega in phi,psi space in proteins as well as in the linear and cyclic peptides. This is particularly manifest as a consistent difference between the mean omega values in chains of left and right-hand chirality. This dichotomy is observed for all the standard amino acids and is especially striking in the absence of secondary structure. The phenomenon is discussed in the context of theoretical work on peptide analogues, and the implications for protein conformation and structure are briefly considered.