Comparative protein modelling by satisfaction of spatial restraints

A model for human cytochrome P450 2D6 based on homology modeling and NMR studies of substrate binding

The cytochrome P450 responsible for the debrisoquine/sparteine polymorphism (P450 2D6) has been produced in large quantities by expression of a modified cDNA in baculovirus. A polyhistidine extension was incorporated at the C-terminus of the expressed protein, which, after purification of the protein on a nickel-agarose column, could be removed proteolytically by treatment with thrombin. Purified yields of P450 2D6 were 2.4 mg from 700 mL of cell culture. The protein had a greater than 90% heme content and was fully active, having no residual absorbance at 420 nm in the reduced CO complex. The quantities produced allowed direct study of the interaction of the substrate codeine with the enzyme by paramagnetic relaxation effects on the NMR spectrum of the substrate. Distances between the heme iron atom and substrate protons were calculated from these experiments, and the orientation of the substrate in the binding pocket was determined. This showed that codeine was bound with the methoxy group of the molecule closest to the heme iron (iron-methyl proton distance of 3.1 +/- 0.1 A), consistent with the observed O-demethylation to morphine. A model of the complex Of P450 2D6 with codeine was built from a multiple sequence and structure alignment of the known crystal structures for P450s, incorporating the experimental constraints derived from the NMR studies. This showed that the overall fold Of P450 2D6 is more similar to that of P450 BM3 than to either P450 cam or P450 terp. Codeine binds to P450 2D6 so that the methoxy group is directly above the A ring of the heme, while the basic nitrogen interacts with the carboxylate of aspartate 301.

Directed evolution of a para-nitrobenzyl esterase for aqueous-organic solvents

Through sequential generations of random mutagenesis and screening, we have directed the evolution of an esterase for deprotection of an antibiotic p-nitrobenzyl ester in aqueous-organic solvents. Because rapid screening directly on the desired antibiotic (loracarbef) nucleus p-nitrobenzyl ester was not feasible, the p-nitrophenyl ester was employed. Catalytic performance on the screening substrate was shown to reasonably mimic enzyme activity toward the desired ester. One p-nitrobenzyl esterase variant performs as well in 30% dimethylformamide as the wildtype enzyme in water, reflecting a 16-fold increase in esterase activity. Random pairwise gene recombination of two positive variants led to a further two-fold improvement in activity. Considering also the increased expression level achieved during these experiments, the net result of four sequential generations of random mutagenesis and the one recombination step is a 50-60-fold increase in total activity. Although the contributions of individual effective amino acid substitutions to enhanced activity are small (< 2-fold increases), the accumulation of multiple mutations by directed evolution allows significant improvement of the biocatalyst for reactions on substrates and under conditions not already optimized in nature. The positions of the effective amino acid substitutions have been identified in a pNB esterase structural model developed based on its homology to acetylcholinesterase and triacylglycerol lipase. None appear to interact directly with the antibiotic substrate, further underscoring the difficulty of predicting their effects in a 'rational' design effort.

Optimum superimposition of protein structures: ambiguities and implications

BACKGROUND

Techniques for comparison and optimum superimposition of protein structures are indispensable tools, providing the basis for statistical analysis, modeling, prediction and classification of protein folds. Observed similarity of structures is frequently interpreted as an indication of evolutionary relatedness. A variety of advanced techniques are available, but so far the important issue of uniqueness of structural superimposition has been largely neglected. We set out to investigate this issue by implementing an efficient algorithm for structure superimposition enabling routine searches for alternative alignments.

RESULTS

The algorithm is based on optimum superimposition of structures and dynamic programming. The implementation is tested and validated using published results. In particular, an automatic classification of all protein folds in a recent release of the protein data bank is performed. The results obtained are closely related to published data. Surprisingly, for many protein pairs alternative alignments are obtained. These alignments are indistinguishable in terms of number of equivalent residues and root mean square error of superimposition, but the respective sets of equivalent residue pairs are completely distinct. Alternative alignments are observed for all protein architectures, including mixed alpha/beta folds.

CONCLUSIONS

Superimposition of protein folds is frequently ambiguous. This has several implications on the interpretation of structural similarity with respect to evolutionary relatedness and it restricts the range of applicability of superimposed structures in statistical analysis. In particular, studies based on the implicit assumption that optimum superimposition of structures is unique are bound to be misleading.

Fold recognition and ab initio structure predictions using hidden Markov models and beta-strand pair potentials

Protein structure predictions were submitted for 9 of the target sequences in the competition that ran during 1994. Targets sequences were selected that had no known homology with any sequence of known structure and were members of a reasonably sized family of related but divergent sequences. The objective was either to recognize a compatible fold for the target sequence in the database of known structures or to predict ab initio its rough 3D topology. The main tools used were Hidden Markov models (HMM) for fold recognition, a beta-strand pair potential to predict beta-sheet topology, and the PHD server for secondary structure prediction. Compatible folds were correctly identified in a number of cases and the beta-strand pair potential was shown to be a useful tool for ab initio topology prediction.

Homology modeling by the ICM method

Five models have been built by the ICM method for the Comparative Modeling section of the Meeting on the Critical Assessment of Techniques for Protein Structure Prediction. The targets have homologous proteins with known three-dimensional structure with sequence identity ranging from 25 to 77%. After alignment of the target sequence with the related three-dimensional structure, the modeling procedure consists of two subproblems: side-chain prediction and loop prediction. The ICM method approaches these problems with the following steps: (1) a starting model is created based on the homologous structure with the conserved portion fixed and the nonconserved portion having standard covalent geometry and free torsion angles; (2) the Biased Probability Monte Carlo (BPMC) procedure is applied to search the subspaces of either all the nonconservative side-chain torsion angles or torsion angles in a loop backbone and surrounding side chains. A special algorithm was designed to generate low-energy loop deformations. The BPMC procedure globally optimizes the energy function consisting of ECEPP/3 and solvation energy terms. Comparison of the predictions with the NMR or crystallographic solutions reveals a high proportion of correctly predicted side chains. The loops were not correctly predicted because imprinted distortions of the backbone increased the energy of the near-native conformation and thus made the solution unrecognizable. Interestingly, the energy terms were found to be reliable and the sampling of conformational space sufficient. The implications of this finding for the strategies of future comparative modeling are discussed.

A critical assessment of comparative molecular modeling of tertiary structures of proteins

In spite of the tremendous increase in the rate at which protein structures are being determined, there is still an enormous gap between the numbers of known DNA-derived sequences and the numbers of three-dimensional structures. In order to shed light on the biological functions of the molecules, researchers often resort to comparative molecular modeling. Earlier work has shown that when the sequence alignment is in error, then the comparative model is guaranteed to be wrong. In addition, loops, the sites of insertions and deletions in families of homologous proteins, are exceedingly difficult to model. Thus, many of the current problems in comparative molecular modeling are minor versions of the global protein folding problem. In order to assess objectively the current state of comparative molecular modeling, 13 groups submitted blind predictions of seven different proteins of undisclosed tertiary structure. This assessment shows that where sequence identity between the target and the template structure is high (> 70%), comparative molecular modeling is highly successful. On the other hand, automated modeling techniques and sophisticated energy minimization methods fail to improve upon the starting structures when the sequence identity is low (approximately 30%). Based on these results it appears that insertions and deletions are still major problems. Successfully deducing the correct sequence alignment when the local similarity is low is still difficult. We suggest some minimal testing of submitted coordinates that should be required of authors before papers on comparative molecular modeling are accepted for publication in journals.

Evaluation of current techniques for ab initio protein structure prediction

The results of a protein structure prediction contest are reviewed. Twelve different groups entered predictions on 14 proteins of known sequence whose structures had been determined but not yet disseminated to the scientific community. Thus, these represent true tests of the current state of structure prediction methodologies. From this work, it is clear that accurate tertiary structure prediction is not yet possible. However, protein fold and motif prediction are possible when the motif is recognizable similar to another known structure. Internal symmetry and the information inherent in an aligned family of homologous sequences facilitate predictive efforts. Novel folds remain a major challenge for prediction efforts.

Evaluation of comparative protein modeling by MODELLER

We evaluate 3D models of human nucleoside diphosphate kinase, mouse cellular retinoic acid binding protein I, and human eosinophil neurotoxin that were calculated by MODELLER, a program for comparative protein modeling by satisfaction of spatial restraints. The models have good stereochemistry and are at least as similar to the crystallographic structures as the closest template structures. The largest errors occur in the regions that were not aligned correctly or where the template structures are not similar to the correct structure. These regions correspond predominantly to exposed loops, insertions of any length, and non-conserved side chains. When a template structure with more than 40% sequence identity to the target protein is available, the model is likely to have about 90% of the mainchain atoms modeled with an rms deviation from the X-ray structure of approximately 1 A, in large part because the templates are likely to be that similar to the X-ray structure of the target. This rms deviation is comparable to the overall differences between refined NMR and X-ray crystallography structures of the same protein.

Comparative protein modeling by satisfaction of spatial restraints

Approximately one third of known protein sequences are related to at least one known protein structure. As a result, an order of magnitude more sequences can be modeled by comparative modeling than there are experimentally determined protein structures. A large fraction of these models has an accuracy approaching that of a low resolution X-ray structure or a medium resolution nuclear magnetic resonance structure. The number of applications where homology modeling has been proven useful is growing rapidly.

Predicting the structure of the light-harvesting complex II of Rhodospirillum molischianum

We attempted to predict through computer modeling the structure of the light-harvesting complex II (LH-II) of Rhodospirillum molischianum, before the impending publication of the structure of a homologous protein solved by means of X-ray diffraction. The protein studied is an integral membrane protein of 16 independent polypeptides, 8 alpha-apoproteins and 8 beta-apoproteins, which aggregate and bind to 24 bacteriochlorophyll-a's and 12 lycopenes. Available diffraction data of a crystal of the protein, which could not be phased due to a lack of heavy metal derivatives, served to test the predicted structure, guiding the search. In order to determine the secondary structure, hydropathy analysis was performed to identify the putative transmembrane segments and multiple sequence alignment propensity analyses were used to pinpoint the exact sites of the 20-residue-long transmembrane segment and the 4-residue-long terminal sequence at both ends, which were independently verified and improved by homology modeling. A consensus assignment for the secondary structure was derived from a combination of all the prediction methods used. Three-dimensional structures for the alpha- and the beta-apoprotein were built by comparative modeling. The resulting tertiary structures are combined, using X-PLOR, into an alpha beta dimer pair with bacteriochlorophyll-a's attached under constraints provided by site-directed mutagenesis and spectral data. The alpha beta dimer pairs were then aggregated into a quaternary structure through further molecular dynamics simulations and energy minimization. The structure of LH-II so determined is an octamer of alpha beta heterodimers forming a ring with a diameter of 70 A.

Packaging of proteases and proteoglycans in the granules of mast cells and other hematopoietic cells. A cluster of histidines on mouse mast cell protease 7 regulates its binding to heparin serglycin proteoglycans

Mouse mast cell protease 7 (mMCP-7) is a tryptase stored in the secretory granules of mast cells. At the granule pH of 5.5, mMCP-7 is fully active and is bound to heparin-containing serglycin proteoglycans. to understand the interaction of mMCP-7 with heparin inside and outside the mast cell, this trytase was first studied by comparative protein modeling. The "pro" form of mMCP-7 was then expressed in insect cells and studied by site-directed mutagenesis. Although mMCP-7 lacks known linear sequences of amino acis that interact with heparin, the three-dimensional model of mMCP-7 revealed an area on the surface of the folded protein away from the substrate-binding site that exhibits a strong positive electrostatic potential at the acidic pH of the granule. In agreement with this calculation, recombinant pro-mMCP-7 bound to a heparin-affinity column at pH 5.5 and readily dissociated from the column at pH > 6.5. Site-directed mutagenesis confirmed the prediction that the conversion of His residues 8,68, and 70 in the positively charged region into Glu prevents the binding of pro-mMCP-7 to heparin. Because the binding requires positively charged His residues, native mMCP-7 is able to dissociate from the protease/proteoglycan macromolecular complex when the complex is exocytosed from bone marrow-derived mast cells into a neutral pH environment. Many hematopoietic effector cells store positively charged proteins in granules that contain serglycin proteoglycans. The heparin/mMCP-7 interaction, which depends on the tertiary structure of the tryptase, may be representative of a general control mechanism by which hematopoietic cells maximize storage of properly folded, enzymatically active proteins in their granules.

Modeling mutations and homologous proteins

A protein sequence with at lease 40% identity to a known structure can now be modelled automatically, with an accuracy approaching that o fa low-resolution X-ray structure or a medium-resolution nuclear magnetic resonance structure. In general, these models have goods stereochemistry and an overall structural accuracy that is as high as the similarity between the template and the actual structure being predicted. As a result, the number of sequences that can be modelled is an order of magnitude larger then the number of experimentally determined protein structures. In addition, evaluation techniques are available that can estimated errors in different regions of the model. Thus, the number of applications where homology modelling is proving useful is growing rapidly.

Predicting the helix packing of globular proteins by self-correcting distance geometry

A new self-correcting distance geometry method for predicting the three-dimensional structure of small globular proteins was assessed with a test set of 8 helical proteins. With the knowledge of the amino acid sequence and the helical segments, our completely automated method calculated the correct backbone topology of six proteins. The accuracy of the predicted structures ranged from 2.3 A to 3.1 A for the helical segments compared to the experimentally determined structures. For two proteins, the predicted constraints were not restrictive enough to yield a conclusive prediction. The method can be applied to all small globular proteins, provided the secondary structure is known from NMR analysis or can be predicted with high reliability.

A self consistent mean field approach to simultaneous gap closure and side-chain positioning in homology modelling

A new computational procedure which simultaneously provides gap closure and side-chain positioning in homology modelling is described. It uses a database search scheme to generate fragments to model gaps, a rotamer library to define side-chain conformations, and iteratively refines a conformational matrix CM, such that its elements CM(i,j,o) and CM(i,j,k) give the probabilities that the backbone of residue i adopts the conformation described by fragment j and that its side-chain adopts the conformation of its possible rotamer k. Each residue experiences the average of all possible environments, weighted by their respective probabilities. The method converges, thereby deserving the name of 'self consistent mean field' approach.

Protein structure prediction: recognition of primary, secondary, and tertiary structural features from amino acid sequence

This review attempts a critical stock-taking of the current state of the science aimed at predicting structural features of proteins from their amino acid sequences. At the primary structure level, methods are considered for detection of remotely related sequences and for recognizing amino acid patterns to predict posttranslational modifications and binding sites. The techniques involving secondary structural features include prediction of secondary structure, membrane-spanning regions, and secondary structural class. At the tertiary structural level, methods for threading a sequence into a mainchain fold, homology modeling and assigning sequences to protein families with similar folds are discussed. A literature analysis suggests that, to date, threading techniques are not able to show their superiority over sequence pattern recognition methods. Recent progress in the state of ab initio structure calculation is reviewed in detail. The analysis shows that many structural features can be predicted from the amino acid sequence much better than just a few years ago and with attendant utility in experimental research. Best prediction can be achieved for new protein sequences that can be assigned to well-studied protein families. For single sequences without homologues, the folding problem has not yet been solved.

Derivation of rules for comparative protein modeling from a database of protein structure alignments

We describe a database of protein structure alignments as well as methods and tools that use this database to improve comparative protein modeling. The current version of the database contains 105 alignments of similar proteins or protein segments. The database comprises 416 entries, 78,495 residues, 1,233 equivalent entry pairs, and 230,396 pairs of equivalent alignment positions. At present, the main application of the database is to improve comparative modeling by satisfaction of spatial restraints implemented in the program MODELLER (Sali A, Blundell TL, 1993, J Mol Biol 234:779-815). To illustrate the usefulness of the database, the restraints on the conformation of a disulfide bridge provided by an equivalent disulfide bridge in a related structure are derived from the alignments; the prediction success of the disulfide dihedral angle classes is increased to approximately 80%, compared to approximately 55% for modeling that relies on the stereochemistry of disulfide bridges alone. The second example of the use of the database is the derivation of the probability density function for comparative modeling of the cis/trans isomerism of the proline residues; the prediction success is increased from 0% to 82.9% for cis-proline and from 93.3% to 96.2% for trans-proline. The database is available via electronic mail.

Automated comparative modelling of protein structures

Most automated methods for protein modelling rely on the assembly of rigid fragments from known three-dimensional structures. Although the modelling of side chains has received attention recently, loop regions continue to be neglected, and these conformations contribute most errors to models. Comparative modelling procedures using distance restraints are particularly useful for modelling distantly related proteins with a common fold. Progress has also recently been made in the automatic evaluation of model structures.

The recognition of protein structure and function from sequence: adding value to genome data

The explosion of DNA sequence data from genome projects presents many challenges. For instance, we must extend our current knowledge of protein structure and function so that it can be applied to these new sequences. The derivation of rules for the relationships between sequence and structure allow us to recognize a common fold by the use of tertiary templates. New techniques enable us to begin to meet the challenge of rule-based modelling of distantly related proteins. This paper describes an integrated and knowledge-based approach to the prediction of protein structure and function which can maximize the value of sequence information.

Pattern recognition and self-correcting distance geometry calculations applied to myohemerythrin

A topological list, consisting of segments of regular secondary structures and a list of buried and solvent accessible residues, is automatically predicted from multiple aligned sequences in a protein family. This topological list is translated into geometric constraints for distance geometry calculation in torsion angle space. A new self-correcting distance geometry method detects and eliminates false distance constraints. In an application to the four-helix bundle protein, myohem-erythrin, the right-handed global fold was correctly reproduced with a root-mean-square deviation of 2.6 A, when the topological list was derived from the X-ray structure. A predicted topological list, coupled with constraints from the residues in the active site of myohemerythrin, predicted the correct fold with a root-mean-square deviation of 4 A for backbone atoms.

Structure-based identification and clustering of protein families and superfamilies

We describe an approach to protein structure comparison designed to detect distantly related proteins of similar fold, where the procedure must be sufficiently flexible to take into account the elasticity of protein folds without losing specificity. Protein structures are represented as a series of secondary structure elements, where for each element a local environment describes its relations with the elements that surround it. Secondary structures are then aligned by comparing their features and local environments. The procedure is illustrated with searches of a database of 468 protein structures in order to identify proteins of similar topology to porcine pepsin, porphobilinogen deaminase and serum amyloid P-component. In all cases the searches correctly identify protein structures of similar fold as the search proteins. Multiple cross-comparisons of protein structures allow the clustering of proteins of similar fold. This is exemplified with a clustering of alpha/beta- and beta-class protein structures. We discuss applications of the comparison and clustering of three-dimensional protein structures to comparative modelling and structure-based protein design.

Magnifying endoscopic observation of the gastric mucosa, particularly in patients with atrophic gastritis

The gastric mucosal surface was observed using the magnifying fibergastroscope (FGS-ML), and the fine gastric mucosal patterns, which were even smaller than one unit of gastric area, were examined at a magnification of about 30. For simplicification, we classified these patterns by magnifying endoscopy in the following ways; FP, FIP, FSP, SP and MP, modifying Yoshii's classification under the dissecting microscope. The FIP, which was found to have round and long elliptical gastric pits, is a new addition to our endoscopic classification. The relationship between the FIP and the intermediate zone was evaluated by superficial and histological studies of surgical and biopsy specimens. The width of the band of FIP seems to be related to the severity of atrophic gastritis. Also, the transformation of FP to FIP was assessed by comparing specimens taken from the resected and residual parts of the stomach, respectively. Moreover, it appears that severe gastritis occurs in the gastric mucosa which shows a FIP. Therefore, we consider that the FIP indicates the position of the atrophic border.