CATH--a hierarchic classification of protein domain structures.

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.

Solution structure of the IRF-2 DNA-binding domain: a novel subgroup of the winged helix-turn-helix family.

BACKGROUND: The transcription of interferon (IFN) and IFN-inducible genes is mainly regulated by the interferon regulatory factor (IRF) family of proteins, which recognize a unique AAGTGA hexamer repeat motif in the regulatory region of IFN genes. A DNA-binding domain of approximately 100 amino acids has been commonly found in the IRF family of proteins, but it has no sequence homology to known DNA-binding motifs. Elucidation of the structures of members of the IRF family is therefore useful to the understanding of the regulation and evolution of the immune system at the structural level. RESULTS: The solution structure of the DNA-binding domain of interferon regulatory factor-2 (IRF-2) has been determined by NMR spectroscopy. It is composed of a four-stranded antiparallel beta sheet and three alpha helices, and its global fold is similar to those of the winged helix-turn-helix (wHTH) family of proteins. A long loop (Pro37-Asp51) is found immediately before the HTH motif, which is not found in other wHTH proteins. The NMR signals of residues in this long loop, as well as the second helix of the HTH motif, are strongly affected upon the addition of the hexamer repeat DNA, suggesting that these structural elements participate in DNA recognition and binding. CONCLUSIONS: The structural similarity of the DNA-binding domain of IRF-2 with those of proteins in the wHTH family shows that the IRF proteins belong to the wHTH family, even though there is no apparent sequence homology among proteins of the two families. The sequential structure alignment program (SSAP) shows that IRF-2 has a slightly different structure from typical wHTH proteins, mainly in the orientation of helix 2. The IRF family of proteins should therefore be categorized into a subfamily of the wHTH family. The evidence here implies that the evolutional pathway of the IRF family is distinct from that of the other wHTH proteins, in other words, the immune system diverged from an evolutional stem at an early stage.

Target-induced conformational adaptation of calmodulin revealed by the crystal structure of a complex with nematode Ca(2+)/calmodulin-dependent kinase kinase peptide.

Calmodulin (CaM) is a ubiquitous calcium (Ca(2+)) sensor which binds and regulates protein serine/threonine kinases along with many other proteins in a Ca(2+)-dependent manner. For this multi-functionality, conformational plasticity is essential; however, the nature and magnitude of CaM's plasticity still remains largely undetermined. Here, we present the 1.8 A resolution crystal structure of Ca(2+)/CaM, complexed with the 27-residue synthetic peptide corresponding to the CaM-binding domain of the nematode Caenorhabditis elegans Ca(2+)/CaM-dependent kinase kinase (CaMKK). The peptide bound in this crystal structure is a homologue of the previously NMR-derived complex with rat CaMKK, but benefits from improved structural resolution. Careful comparison of the present structure to previous crystal structures of CaM complexed with unrelated peptides derived from myosin light chain kinase and CaM kinase II, allow a quantitative analysis of the differences in the relative orientation of the N and C-terminal domains of CaM, defined as a screw axis rotation angle ranging from 156 degrees to 196 degrees. The principal differences in CaM interaction with various peptides are associated with the N-terminal domain of CaM. Unlike the C-terminal domain, which remains unchanged internally, the N-terminal domain of CaM displays significant differences in the EF-hand helix orientation between this and other CaM structures. Three hydrogen bonds between CaM and the peptide (E87-R336, E87-T339 and K75-T339) along with two salt bridges (E11-R349 and E114-K334) are the most probable determinants for the binding direction of the CaMKK peptide to CaM.

Solution structure of calmodulin-W-7 complex: the basis of diversity in molecular recognition.

The solution structure of calcium-bound calmodulin (CaM) complexed with an antagonist, N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W-7), has been determined by multidimensional NMR spectroscopy. The structure consists of one molecule of W-7 binding to each of the two domains of CaM. In each domain, the W-7 chloronaphthalene ring interacts with four methionine methyl groups and other aliphatic or aromatic side-chains in a deep hydrophobic pocket, the site responsible for CaM binding to CaM-dependent enzymes such as myosin light chain kinases (MLCKs) and CaM kinase II. This competitive binding at the same site between W-7 and CaM-dependent enzymes suggests the mechanism by which W-7 inhibits CaM to activate the enzymes. The orientation of the W-7 naphthalene ring in the N-terminal pocket is rotated approximately 40 degrees with respect to that in the C-terminal pocket. The W-7 ring orientation differs significantly from the Trp800 indole ring of smooth muscle MLCK bound to the C-terminal pocket and the phenothiazine ring of trifluoperazine bound to the N or C-terminal pocket. These comparative structural analyses demonstrate that the two hydrophobic pockets of CaM can accommodate a variety of bulky aromatic rings, which provides a plausible structural basis for the diversity in CaM-mediated molecular recognition.

NMR structure of the Streptomyces metalloproteinase inhibitor, SMPI, isolated from Streptomyces nigrescens TK-23: another example of an ancestral beta gamma-crystallin precursor structure.

The Streptomyces metalloproteinase inhibitor, SMPI, isolated from Streptomyces nigrescens TK-23, is a proteinaceous metalloproteinase inhibitor, and consists of 102 amino acid residues with two disulfide bridges. SMPI specifically inhibits metalloproteinases such as thermolysin. In the present work, the solution structure of SMPI was determined on the basis of 1536 nuclear Overhauser enhancement derived distance restraints and 52 dihedral angle restraints obtained from three-bond spin coupling constants. The final ensemble of 20 NMR structures overlaid onto their mean coordinate with backbone (N, Calpha, C') r.m.s.d. values of 0. 45(+/-0.11) A and 0.57(+/-0.18) A for residues 6 to 99 and the entire 102 residues, respectively. SMPI is essentially composed of two beta-sheets, each consisting of four antiparallel beta-strands. The structure can be considered as two Greek key motifs with 2-fold internal symmetry, a Greek key beta-barrel. One unique structural feature found in SMPI is in its extension between the first and second strands of the second Greek key motif. Interestingly, this extended segment is known to be involved in the inhibitory activity of SMPI. In the absence of sequence similarity, the SMPI structure shows clear similarity to both domains of the eye lens crystallins, both domains of the calcium sensor protein-S, as well as the single-domain yeast killer toxin. The yeast killer toxin structure was thought to be a precursor of the two-domain beta gamma-crystallin proteins, because of its structural similarity to each domain of the beta gamma-crystallins. SMPI thus provides another example of a single-domain protein structure that corresponds to the ancestral fold from which the two-domain proteins in the beta gamma-crystallin superfamily are believed to have evolved.

Structure prediction and modelling.

Cracking the second fundamental code of molecular biology (how the tertiary structure of a protein is determined by its amino acid sequence) remains an elusive goal. However, the impetus to establish credible approximations, if not a definitive solution to this relationship, has never been greater. In the past year significant progress has been made through a series of novel approaches. This review describes the most important developments and outlines how they can be usefully employed by those whose specialization lies outside the field.

Structure prediction and modelling.

Protein structure prediction from sequence remains a major goal in molecular biology. The methods described in this review concentrate on deriving structural information through the detection of similarities between a test sequence and a database of known structures. Such methods are often referred to as knowledge-based strategies reflecting the use of a structural database in the analyses. The past year has seen considerable advances in both the development of automated procedures and their application to protein sequences of outstanding biological interest.

Classification of doubly wound nucleotide binding topologies using automated loop searches.

A classification is presented of doubly wound alpha/beta nucleotide binding topologies, whose binding sites are located in the cleft formed by a topological switch point. In particular, the switch point loop nearest the N-terminus is used to identify specific structural classes of binding protein. This yields seven structurally distinct loop conformations, which are subsequently used as motifs for scanning the Protein Data Bank. The searches, which are effective at identifying functional relationships within a large database of structures, reveal a remarkable and previously unnoticed similarity between the coenzyme binding sites of flavodoxin and tryptophan synthetase, even though there is no sequence or topological similarity between them.

A procedure for detecting structural domains in proteins.

A procedure is described for detecting domains in proteins of known structure. The method is based on the intuitively simple idea that each domain should contain an identifiable hydrophobic core. By applying the algorithm described in the companion paper (Swindells MB, 1995, Protein Sci 4:93-102) to identify distinct cores in multi-domain proteins, one can use this information to determine both the number and the location of the constituent domains. Tests have shown the procedure to be effective on a number of examples, even when the domains are discontinuous along the sequence. However, deficiencies also occur when hydrophobic cores from different domains continue through the interface region and join one another.

A procedure for the automatic determination of hydrophobic cores in protein structures.

An algorithm is described for automatically detecting hydrophobic cores in proteins of known structure. Three pieces of information are considered in order to achieve this goal. These are: secondary structure, side-chain accessibility, and side-chain-side-chain contacts. Residues are considered to contribute to a core when they occur in regular secondary structure and have buried side chains that form predominantly nonpolar contacts with one another. This paper describes the algorithm's application to families of proteins with conserved topologies but low sequence similarities. The aim of this investigation is to determine the efficacy of the algorithm as well as to study the extent to which similar cores are identified within a common topology.

Genome analysis: Assigning protein coding regions to three-dimensional structures.

We describe the results of a procedure for maximizing the number of sequences that can be reliably linked to a protein of known three-dimensional structure. Unlike other methods, which try to increase sensitivity through the use of fold recognition software, we only use conventional sequence alignment tools, but apply them in a manner that significantly increases the number of relationships detected. We analyzed 11 genomes and found that, depending on the genome, between 23 and 32% of the ORFs had significant matches to proteins of known structure. In all cases, the aligned region consisted of either >100 residues or >50% of the smaller sequence. Slightly higher percentages could be attained if smaller motifs were also included. This is significantly higher than most previously reported methods, even those that have a fold-recognition component. We survey the biochemical and structural characteristics of the most frequently occurring proteins, and discuss the extent to which alignment methods can realistically assign function to gene products.

Domain assignment for protein structures using a consensus approach: characterization and analysis.

A consensus approach for the assignment of structural domains in proteins is presented. The approach combines a number of previously published algorithms, and takes advantage of the elevated accuracy obtained when assignments from the individual algorithms are in agreement. The consensus approach is tested on a data set of 55 protein chains, for which domain assignments from four automated methods were known, and for which crystallographers assignments had been reported in the literature. Accuracy was found to increase in this test from 72% using individual algorithms to 100% when all four methods were in agreement. However a consensus prediction using all four methods was only possible for 52% of the dataset. The consensus approach [using three publicly available domain assignment algorithms (PUU, DETECTIVE, DOMAK)] was then used to make domain assignments for a data set of 787 protein chains from the Protein Data Bank. Analysis of the assignments showed 55.7% of assignments could be made automatically, and of these, 13.5% were multi-domain proteins. Of the remaining 44.3% that could not be assigned by the consensus procedure 90.4% had their domain boundaries assigned correctly by at least one of the algorithms. Once identified, these domains were analyzed for trends in their size and secondary structure class. In addition, the discontinuity of each domain along the protein chain was considered.

Protein clefts in molecular recognition and function.

One of the primary factors determining how proteins interact with other molecules is the size of clefts in the protein's surface. In enzymes, for example, the active site is often characterized by a particularly large and deep cleft, while interactions between the molecules of a protein dimer tend to involve approximately planar surfaces. Here we present an analysis of how cleft volumes in proteins relate to their molecular interactions and functions. Three separate datasets are used, representing enzyme-ligand binding, protein-protein dimerization and antibody-antigen complexes. We find that, in single-chain enzymes, the ligand is bound in the largest cleft in over 83% of the proteins. Usually the largest cleft is considerably larger than the others, suggesting that size is a functional requirement. Thus, in many cases, the likely active sites of an enzyme can be identified using purely geometrical criteria alone. In other cases, where there is no predominantly large cleft, chemical interactions are required for pinpointing the correct location. In antibody-antigen interactions the antibody usually presents a large cleft for antigen binding. In contrast, protein-protein interactions in homodimers are characterized by approximately planar interfaces with several clefts involved. However, the largest cleft in each subunit still tends to be involved.

Structural similarity between the pleckstrin homology domain and verotoxin: the problem of measuring and evaluating structural similarity.

An unexpected structural similarity is described between the pleckstrin homology (PH) domain and verotoxin. This similarity has escaped detection primarily due to the differences in topology that exist between the two proteins. By comparing this result with two previously reported similarities for the PH domain, one with the lipocalins and another with the FK506 binding protein, we discuss the problems of measuring and assessing structural similarities.

Prediction of a novel topology in the N-terminal, 14 kDa fragment of Ada protein.

Previously determined protein structures have been analysed, in order to find folding motifs similar to that proposed by NMR spectroscopy, for the N-terminal, 14 kDa fragment of the Ada protein. The analyses reveal only limited similarities with the NMR-derived structural data and strongly suggest that this region of the Ada protein adopts a previously unobserved topology. Characteristic structural features, which arise from the inferred chain connectivity, are examined through comparisons with other structures. Using this information, the topology of the Ada protein 14 kDa fragment has been predicted in order to provide structural data not yet attainable from NMR experiments.

A study of structural determinants in the interleukin-1 fold.

The structures of interleukin-1 beta, basic fibroblast growth factor and Erythrina trypsin inhibitor have been analysed in order to determine whether the hydrophobic core remains conserved, even when the structures have extremely low sequence similarities. We find that there are significant differences in the way each protein achieves a satisfactory arrangement of core residues and that positions which contribute to the core of one structure are not guaranteed to contribute to the integrity of another. Furthermore, the side-chain packing arrangements of these core residues vary significantly between the three structures. During this analysis the side-chain rotamers for three independently determined interleukin-1 beta structures were also compared. It was found that although buried residues are generally in agreement the remaining residues frequently occupy different rotamers in the three structures. This suggests that although meaningful studies are possible for buried side-chains the results obtained from equivalent analyses of accessible residues should be treated with caution. These results are discussed with specific reference to the optimization of side-chain packing in proteins of known structure.