Identification and application of the concepts important for accurate and reliable protein secondary structure prediction

Characterization of the selectivity filter of the epithelial sodium channel

The epithelial sodium channel (ENaC) is composed of three homologous subunits termed alpha, beta, and gamma. Previous studies suggest that selected residues within a hydrophobic region immediately preceding the second membrane-spanning domain of each subunit contribute to the conducting pore of ENaC. We probed the pore of mouse ENaC by systematically mutating all 24 amino acids within this putative pore region of the alpha-subunit to cysteine and co-expressing these mutants with wild type beta- and gamma-subunits of mouse ENaC in Xenopus laevis oocytes. Functional characteristics of these mutants were examined by two-electrode voltage clamp and single channel recording techniques. Two distinct domains were identified based on the functional changes associated with point mutations. An amino-terminal domain (alpha-Val(569)-alpha-Gly(579)) showed minimal changes in cation selectivity or amiloride sensitivity following cysteine substitution. In contrast, cysteine substitutions within the carboxyl-terminal domain (alpha-Ser(580)-alpha-Ser(592)) resulted in significant changes in cation selectivity and moderately altered amiloride sensitivity. The mutant channels containing alphaG587C or alphaS589C were permeable to K(+), and mutation of a GSS tract (positions alpha587-alpha589) to GYG resulted in a moderately K(+)-selective channel. Our results suggest that the C-terminal portion of the pore region within the alpha-subunit contributes to the selectivity filter of ENaC.

Purification, cloning, and three-dimensional structure prediction of Micrococcus luteus FAD-containing tyramine oxidase

The FAD-containing tyramine oxidase enzyme and gene from the Gram (+) bacterium Micrococcus luteus were isolated, and computer prediction was used to propose a preliminary 3D model of the protein. A 2.8-kb Sau3AI fragment containing the structural gene of tyramine oxidase was cloned from a M. luteus genomic DNA library. The 1332 bp gene encodes a protein of 443 amino acids, with a calculated molecular mass of 49.1 kDa. The enzyme was found to be a homodimer with a molecular weight of 49,000. It oxidizes tyramine, adrenaline, 3-hydroxytyramine, dopamine, and noradrenaline, and was reversibly inhibited by FAD-containing monoamine oxidase A and B specific inhibitors. Sequence comparison show that tyramine oxidase is smaller than other FAD-amine oxidases but that it contains well-conserved amino acid residues reported in all other FAD-amine oxidases. A hypothetical three-dimensional structure of tyramine oxidase has also been proposed based on secondary structure predictions, threading, and comparative modeling.

Topology prediction of Brucella abortus Omp2b and Omp2a porins after critical assessment of transmembrane beta strands prediction by several secondary structure prediction methods

In order to propose a reliable model for Brucella porin topology, several structure prediction methods were evaluated in their ability to predict porin topology. Four porins of known structure were selected as test-cases and their secondary structure delineated. The specificity and sensitivity of 11 methods were separately evaluated. Our critical assessment shows that some secondary structure prediction methods (PHD, Dsc, Sopma) originally designed to predict globular protein structure are useful on porin topology prediction. The overall best prediction is obtained by combining these three "generalist" methods with a transmembrane beta strand prediction technique. This "consensus" method was applied to Brucella porins Omp2b and Omp2a, sharing no sequence homology with any other porin. The predicted topology is a 16-stranded antiparallel beta barrel with Omp2a showing a higher number of negatively charged residue in the exposed loops than Omp2b. Experiments are in progress to validate the proposed topology and the functional hypotheses. The ability of the proposed consensus method to predict topology of complex outer membrane protein is briefly discussed.

Identification of archaeal genes encoding a novel stationary phase-response protein

A novel stationary phase-response protein has been identified in the acid-soluble protein extract of the thermophilic archaeon, Thermococcus zilligii. N-Terminal sequencing data were used to identify likely genes for homologues of the protein in the complete genome sequences of various archaeal species. The corresponding genes were identified and analysed. The genes encode a protein ranging from 83 to 92 amino acids in length, with a calculated pI ranging from 4.6 to 9.7. The amino acid sequences of the genes were highly conserved, even between members belonging to the different archaeal kingdoms. The computed secondary structure of the protein indicates it consists of four large helical regions separated by short coiled regions. We propose this protein as a candidate regulator of gene expression in stationary phase.

Is it better to combine predictions?

We have compared the accuracy of the individual protein secondary structure prediction methods: PHD, DSC, NNSSP and Predator against the accuracy obtained by combing the predictions of the methods. A range of ways of combing predictions were tested: voting, biased voting, linear discrimination, neural networks and decision trees. The combined methods that involve 'learning' (the non-voting methods) were trained using a set of 496 non-homologous domains; this dataset was biased as some of the secondary structure prediction methods had used them for training. We used two independent test sets to compare predictions: the first consisted of 17 non-homologous domains from CASP3 (Third Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction); the second set consisted of 405 domains that were selected in the same way as the training set, and were non-homologous to each other and the training set. On both test datasets the most accurate individual method was NNSSP, then PHD, DSC and the least accurate was Predator; however, it was not possible to conclusively show a significant difference between the individual methods. Comparing the accuracy of the single methods with that obtained by combing predictions it was found that it was better to use a combination of predictions. On both test datasets it was possible to obtain a approximately 3% improvement in accuracy by combing predictions. In most cases the combined methods were statistically significantly better (at P = 0.05 on the CASP3 test set, and P = 0.01 on the EBI test set). On the CASP3 test dataset there was no significant difference in accuracy between any of the combined method of prediction: on the EBI test dataset, linear discrimination and neural networks significantly outperformed voting techniques. We conclude that it is better to combine predictions.

Neural networks predict protein folding and structure: artificial intelligence faces biomolecular complexity

In the genomic era DNA sequencing is increasing our knowledge of the molecular structure of genetic codes from bacteria to man at a hyperbolic rate. Billions of nucleotides and millions of aminoacids are already filling the electronic files of the data bases presently available, which contain a tremendous amount of information on the most biologically relevant macromolecules, such as DNA, RNA and proteins. The most urgent problem originates from the need to single out the relevant information amidst a wealth of general features. Intelligent tools are therefore needed to optimise the search. Data mining for sequence analysis in biotechnology has been substantially aided by the development of new powerful methods borrowed from the machine learning approach. In this paper we discuss the application of artificial feedforward neural networks to deal with some fundamental problems tied with the folding process and the structure-function relationship in proteins.

Crystal structure of the cytosolic C2A-C2B domains of synaptotagmin III. Implications for Ca(+2)-independent snare complex interaction

Synaptotagmins are synaptic vesicle-associated, phospholipid-binding proteins most commonly associated with Ca(+2)-dependent exocytotic and Ca(+2)- independent endocytotic events. Synaptotagmin III is a 63.2-kD member of the synaptotagmin homology group; one of its characteristic properties is the ability to bind divalent cations and accessory proteins promiscuously. In the cytosolic portion of this protein, a flexible seven-amino acid linker joins two homologous C2 domains. The C2A domain binds to phospholipid membranes and other accessory proteins in a divalent cation-dependent fashion. The C2B domain promotes binding to other C2B domains, as well as accessory proteins independent of divalent cations. The 3.2 A crystal structure of synaptotagmin III, residues 295-566, which includes the C2A and C2B domains, exhibits differences in the shape of the Ca(+2)-binding pocket, the electrostatic surface potential, and the stoichiometry of bound divalent cations for the two domains. These observations may explain the disparate binding properties of the two domains. The C2A and the C2B domains do not interact; synaptotagmin, therefore, covalently links two independent C2 domains, each with potentially different binding partners. A model of synaptotagmin's involvement in Ca(+2)-dependent regulation of membrane fusion through its interaction with the SNARE complex is presented.

Analysis of the reaction mechanism and substrate specificity of haloalkane dehalogenases by sequential and structural comparisons

Haloalkane dehalogenases catalyse environmentally important dehalogenation reactions. These microbial enzymes represent objects of interest for protein engineering studies, attempting to improve their catalytic efficiency or broaden their substrate specificity towards environmental pollutants. This paper presents the results of a comparative study of haloalkane dehalogenases originating from different organisms. Protein sequences and the models of tertiary structures of haloalkane dehalogenases were compared to investigate the protein fold, reaction mechanism and substrate specificity of these enzymes. Haloalkane dehalogenases contain the structural motifs of alpha/beta-hydrolases and epoxidases within their sequences. They contain a catalytic triad with two different topological arrangements. The presence of a structurally conserved oxyanion hole suggests the two-step reaction mechanism previously described for haloalkane dehalogenase from Xanthobacter autotrophicus GJ10. The differences in substrate specificity of haloalkane dehalogenases originating from different species might be related to the size and geometry of an active site and its entrance and the efficiency of the transition state and halide ion stabilization by active site residues. Structurally conserved motifs identified within the sequences can be used for the design of specific primers for the experimental screening of haloalkane dehalogenases. Those amino acids which were predicted to be functionally important represent possible targets for future site-directed mutagenesis experiments.

Protein secondary structure prediction based on position-specific scoring matrices

A two-stage neural network has been used to predict protein secondary structure based on the position specific scoring matrices generated by PSI-BLAST. Despite the simplicity and convenience of the approach used, the results are found to be superior to those produced by other methods, including the popular PHD method according to our own benchmarking results and the results from the recent Critical Assessment of Techniques for Protein Structure Prediction experiment (CASP3), where the method was evaluated by stringent blind testing. Using a new testing set based on a set of 187 unique folds, and three-way cross-validation based on structural similarity criteria rather than sequence similarity criteria used previously (no similar folds were present in both the testing and training sets) the method presented here (PSIPRED) achieved an average Q3 score of between 76.5% to 78.3% depending on the precise definition of observed secondary structure used, which is the highest published score for any method to date. Given the success of the method in CASP3, it is reasonable to be confident that the evaluation presented here gives a fair indication of the performance of the method in general.

Multiple sequence alignment: algorithms and applications

Elucidation of interrelationships among sequence, structure, function, and evolution (FESS relationships) of a family of genes or gene products is a central theme of modern molecular biology. Multiple sequence alignment has been proven to be a powerful tool for many fields of studies such as phylogenetic reconstruction, illumination of functionally important regions, and prediction of higher order structures of proteins and RNAs. However, it is far too trivial to automatically construct a multiple alignment from a set of related sequences. A variety of methods for solving this computationally difficult problem are reviewed. Several important applications of multiple alignment for elucidation of the FESS relationships are also discussed. For a long period, progressive methods have been the only practical means to solve a multiple alignment problem of appreciable size. This situation is now changing with the development of new techniques including several classes of iterative methods. Today's progress in multiple sequence alignment methods has been made by the multidisciplinary endeavors of mathematicians, computer scientists, and biologists in various fields including biophysicists in particular. The ideas are also originated from various backgrounds, pure algorithmics, statistics, thermodynamics, and others. The outcomes are now enjoyed by researchers in many fields of biological sciences. In the near future, generalized multiple alignment may play a central role in studies of FESS relationships. The organized mixture of knowledge from multiple fields will ferment to develop fruitful results which would be hard to obtain within each area. I hope this review provides a useful information resource for future development of theory and practice in this rapidly expanding area of bioinformatics.

Single amino acid substitutions globally suppress the folding defects of temperature-sensitive folding mutants of phage P22 coat protein

The amino acid sequence of a polypeptide defines both the folding pathway and the final three-dimensional structure of a protein. Eighteen amino acid substitutions have been identified in bacteriophage P22 coat protein that are defective in folding and cause their folding intermediates to be substrates for GroEL and GroES. These temperature-sensitive folding (tsf) substitutions identify amino acids that are critical for directing the folding of coat protein. Additional amino acid residues that are critical to the folding process of P22 coat protein were identified by isolating second site suppressors of the tsf coat proteins. Suppressor substitutions isolated from the phage carrying the tsf coat protein substitutions included global suppressors, which are substitutions capable of alleviating the folding defects of numerous tsf coat protein mutants. In addition, potential global and site-specific suppressors were isolated, as well as a group of same site amino acid substitutions that had a less severe phenotype than the tsf parent. The global suppressors were located at positions 163, 166, and 170 in the coat protein sequence and were 8-190 amino acid residues away from the tsf parent. Although the folding of coat proteins with tsf amino acid substitutions was improved by the global suppressor substitutions, GroEL remained necessary for folding. Therefore, we believe that the global suppressor sites identify a region that is critical to the folding of coat protein.

The membrane association domain of RGS16 contains unique amphipathic features that are conserved in RGS4 and RGS5

Regulators of G protein signaling (RGS proteins) modulate G protein-mediated signaling pathways by acting as GTPase-activating proteins for Gi, Gq, and G12 alpha-subunits of heterotrimeric G proteins. Although it is known that membrane association is critical for the biological activities of many RGS proteins, the mechanism underlying this requirement remains unclear. We reported recently that the NH2 terminus of RGS16 is required for its function in vivo. In this study, we show that RGS16 lacking the NH2 terminus is no longer localized to the plasma membrane as is the wild type protein, suggesting that membrane association is important for biological function. The region of amino acids 7-32 is sufficient to confer the membrane-targeting activity, of which amino acids 12-30 are predicted to adopt an amphipathic alpha-helix. Site-directed mutagenesis experiments showed that the hydrophobic residues of the nonpolar face of the helix and the strips of positively charged side chains positioned along the polar/nonpolar interface of the helix are crucial for membrane association. Subcellular fractionation by differential centrifugation followed by conditions that distinguish peripheral membrane proteins from integral ones indicate that RGS16 is a peripheral membrane protein. We show further that RGS16 membrane association does not require palmitoylation. Our results, together with other recent findings, have defined a unique membrane association domain with amphipathic features. We believe that these structural features and the mechanism of membrane association of RGS16 are likely to apply to the homologous domains in RGS4 and RGS5.

Two strategies for sequence comparison: profile-preprocessed and secondary structure-induced multiple alignment

Multiple sequence alignment remains one of the most powerful tools for assessing sequence relateness and the identification of structurally and functionally important protein regions. In this work, two new techniques are introduced to increase the sensitivity of dynamic programming and to enable checks for alignment consistency: Profile-preprocessed and secondary structure-induced alignments. Both strategies are based upon the hierarchical dynamic programming technique and can be applied separately or used in combination. Alignments resulting from the strategies are shown in comparison with the multiple alignment methods CLUSTALX and MULTAL for distant sequence sets of the flavoxin and cupredoxin protein families.

Conversion of an inactive cardiac dihydropyridine receptor II-III loop segment into forms that activate skeletal ryanodine receptors

A 25 amino acid segment (Glu666-Pro691) of the II-III loop of the alpha1 subunit of the skeletal dihydropyridine receptor, but not the corresponding cardiac segment (Asp788-Pro814), activates skeletal ryanodine receptors. To identify the structural domains responsible for activation of skeletal ryanodine receptors, we systematically replaced amino acids of the cardiac II-III loop with their skeletal counterparts. A cluster of five basic residues of the skeletal II-III loop (681RKRRK685) was indispensable for activation of skeletal ryanodine receptors. In the cardiac segment, a negatively charged residue (Glu804) appears to diminish the electrostatic potential created by this basic cluster. In addition, Glu800 in the group of negatively charged residues 798EEEEE802 of the cardiac II-III loop may serve to prevent the binding of the activation domain.

Improved secondary structure predictions for a nicotinic receptor subunit: incorporation of solvent accessibility and experimental data into a two-dimensional representation

Abstract A refined prediction of the nicotinic acetylcholine receptor (nAChR) subunits' secondary structure was computed with third-generation algorithms. The four selected programs, PHD, Predator, DSC, and NNSSP, based on different prediction approaches, were applied to each sequence of an alignment of nAChR and 5-HT3 receptor subunits, as well as a larger alignment with related subunit sequences from glycine and GABA receptors. A consensus prediction was computed for the nAChR subunits through a "winner takes all" method. By integrating the probabilities obtained with PHD, DSC, and NNSSP, this prediction was filtered in order to eliminate the singletons and to more precisely establish the structure limits (only 4% of the residues were modified). The final consensus secondary structure includes nine alpha-helices (24.2% of the residues, with an average length of 13.9 residues) and 17 beta-strands (22.5% of the residues, with an average length of 6.6 residues). The large extracellular domain is predicted to be mainly composed of beta-strands, with only two helices at the amino-terminal end. The transmembrane segments are predicted to be in a mixed alpha/beta topology (with a predominance of alpha-helices), with no known equivalent in the current protein database. The cytoplasmic domain is predicted to consist of two well-conserved amphipathic helices joined together by an unfolded stretch of variable length and sequence. In general, the segments predicted to occur in a periodic structure correspond to the more conserved regions, as defined by an analysis of sequence conservation per position performed on 152 superfamily members. The solvent accessibility of each residue was predicted from the multiple alignments with PHDacc. Each segment with more than three exposed residues was assumed to be external to the core protein. Overall, these data constitute an envelope of structural constraints. In a subsequent step, experimental data relative to the extracellular portion of the complete receptor were incorporated into the model. This led to a proposed two-dimensional representation of the secondary structure in which the peptide chain of the extracellular domain winds alternatively between the two interfaces of the subunit. Although this representation is not a tertiary structure and does not lead to predictions of specific beta-beta interaction, it should provide a basic framework for further mutagenesis investigations and for fold recognition (threading) searches.

Acn9 is a novel protein of gluconeogenesis that is located in the mitochondrial intermembrane space

Previous studies have indicated that the Acn9 protein is involved in gluconeogenesis. Yeast mutants defective in the ACN9 gene display phenotypes identical with mutants defective in metabolic enzymes required for carbon assimilation. These phenotypes include the inability to utilize acetate as a carbon and energy source, elevated levels of enzymes of the glyoxylate cycle, gluconeogenesis and acetyl-CoA mobilization, and a deficiency in de novo synthesis of glucose from ethanol. The ACN9 gene was isolated by functional complementation of the acetate growth defect of an acn9 mutant. The open reading frame corresponds to YDR511w, and encodes a protein of unknown function. Homologs have been identified in human, mouse, and nematode databases. Two mutant alleles were sequenced. The mutations altered amino acid residues that are conserved among members of the new gene family. ACN9 gene expression was slightly repressed by glucose, and the level of the transcript was approximately 100-fold lower than that of glyoxylate or tricarboxylic acid cycle enzymes. A functional epitope-tagged form of Acn9 was expressed to study expression and the subcellular localization of the protein. The tagged protein was localized to the mitochondrial intermembrane space.

Modeling of acanthoxin A1, a PLA2 enzyme from the venom of the common death adder (Acanthophis antarcticus)

The phospholipase A2 enzyme, acanthoxin, found in the venom of the common death adder (Acanthophis antarcticus) as with other snake PLA2 enzymes displays neurotoxic activity. It is unclear whether this neurotoxic activity particular to some snake PLA2 enzymes is a result of structural differences solely within the catalytic sites or at a distant location upon the molecules. We have predicted the three-dimensional structure of one of the two predominant isoforms of acanthoxin (A1) using comparative protein modeling techniques. Given the high degree of homology and the availability of a high quality crystallographic structure, notexin was used as a molecular template to construct an all atom model of acanthoxin. The model was made using the program MODELLER3 and then refined with X-PLOR. Comparison between the predicted structure of acanthoxin and several X-ray structures of toxic and nontoxic PLA2 enzymes has led to a testable two-step proposal of neurotoxic PLA2 activity; involving the favorable binding to acceptor molecules followed by enzymatic intrusion upon the target membrane. The electrostatic potentials across the molecular surfaces of toxic and nontoxic PLA2 enzymes were calculated (GRASP) and it was found that the toxic PLA2 enzymes possessed a charge distribution on the noncatalytic surface not identified in the nontoxic PLA2 enzymes. Thus we have identified residues potentially involved in the interaction of the PLA2 enzymes with their acceptor molecules. Furthermore, the proposed acceptor molecule recognition site is distant from the catalytic site which upon binding of the PLA2 to the acceptor molecule may enhance the enzymatic ability of the toxic PLA2 enzymes on particular cell types.

Identification of Tcf4 residues involved in high-affinity beta-catenin binding

The N-termini of members of the T-cell factor (Tcf) and lymphocyte-enhancement factor (Lef) protein families bind to beta-catenin, forming bipartite transcription factors which regulate expression of genes involved in organismal development and the growth of normal and malignant colon epithelium. Elevated levels of Tcf4:beta-catenin are found in colon tumor cells with mutations in the adenomatous polyposis coli (APC) gene. The elevated levels of Tcf4:beta-catenin result in increased transcription of genes, including c-myc, important for the growth of these tumor cells. Here we analyze the interaction between beta-catenin and Tcf4 and show that the N-terminal 53 amino acids of Tcf4 bind with high affinity to beta-catenin. We show that this high-affinity interaction involves multiple contact points including Tcf4 Asp-16, which is essential for beta-catenin binding. In addition to Tcf/Lef family members, beta-catenin binds to APC and cadherins. We found that the binding of beta-catenin to Tcf4, APC, or E-cadherin was mutually exclusive. These results are discussed with regard to how beta-catenin interacts with its binding partners and to the potential for identifying specific, small molecule inhibitors of these interactions.

Solution structure of the proapoptotic molecule BID: a structural basis for apoptotic agonists and antagonists

Members of the BCL2 family of proteins are key regulators of programmed cell death, acting either as apoptotic agonists or antagonists. Here we describe the solution structure of BID, presenting the structure of a proapoptotic BCL2 family member. An analysis of sequence/structure of BCL2 family members allows us to define a structural superfamily, which has implications for general mechanisms for regulating proapoptotic activity. It appears two criteria must be met for proapoptotic function within the BCL2 family: targeting of molecules to intracellular membranes, and exposure of the BH3 death domain. BID's activity is regulated by a Caspase 8-mediated cleavage event, exposing the BH3 domain and significantly changing the surface charge and hydrophobicity, resulting in a change of cellular localization.

Structure and function prediction of the Brucella abortus P39 protein by comparative modeling with marginal sequence similarities

A methodology is proposed to solve a difficult modeling problem related to the recently sequenced P39 protein. This sequence shares no similarity with any known 3D structure, but a fold is proposed by several threading tools. The difficulty in aligning the target sequence on one of the proposed template structures is overcome by combining the results of several available prediction methods and by refining a rational consensus between them. In silico validation of the obtained model and a preliminary cross-check with experimental features allow us to state that this borderline prediction is at least reasonable. This model raises relevant hypotheses on the main structural features of the protein and allows the design of site-directed mutations. Knowing the genetic context of the P39 reading frame, we are now able to suggest a function for the P39 protein: it would act as a periplasmic substrate-binding protein.

Evaluation and improvement of multiple sequence methods for protein secondary structure prediction

A new dataset of 396 protein domains is developed and used to evaluate the performance of the protein secondary structure prediction algorithms DSC, PHD, NNSSP, and PREDATOR. The maximum theoretical Q3 accuracy for combination of these methods is shown to be 78%. A simple consensus prediction on the 396 domains, with automatically generated multiple sequence alignments gives an average Q3 prediction accuracy of 72.9%. This is a 1% improvement over PHD, which was the best single method evaluated. Segment Overlap Accuracy (SOV) is 75.4% for the consensus method on the 396-protein set. The secondary structure definition method DSSP defines 8 states, but these are reduced by most authors to 3 for prediction. Application of the different published 8- to 3-state reduction methods shows variation of over 3% on apparent prediction accuracy. This suggests that care should be taken to compare methods by the same reduction method. Two new sequence datasets (CB513 and CB251) are derived which are suitable for cross-validation of secondary structure prediction methods without artifacts due to internal homology. A fully automatic World Wide Web service that predicts protein secondary structure by a combination of methods is available via http://barton.ebi.ac.uk/.

How representative are the known structures of the proteins in a complete genome? A comprehensive structural census

BACKGROUND

Determining how representative the known structures are of the proteins encoded by a complete genome is important for assessing to what extent our current picture of protein stability and folding is overly influenced by biases in the structure databank (PDB). It is also important for improving database-based methods of structure prediction and genome annotation.

RESULTS

The known structures are compared to the proteins encoded by eight complete microbial genomes in terms of simple statistics such as sequence length, composition and secondary structure. The known structures are represented by a collection of nonhomologous domains from the PDB and a smaller list of 'biophysical proteins' on which folding experiments have concentrated. The proteins encoded by the genomes are considered as a whole and divided into various regions, such as known-structure homologue, low complexity (nonglobular), transmembrane or linker. Various tests are performed to assess the significance of the reported differences, in both a practical and a statistical sense.

CONCLUSIONS

The proteins encoded by the genomes are significantly different from those in the PDB. Their sequence lengths, which follow an extreme value distribution, are longer than the PDB proteins and much longer than the biophysical proteins. Their composition differs from the PDB proteins in having more Lys, Ile, Asn and Gln and less Cys and Trp. This is true overall and especially for the regions corresponding to soluble proteins of as yet unknown fold. Secondary-structure prediction on these uncharacterized regions indicates that they contain on average more helical structure than the PDB; differences about this mean are small, with yeast having slightly more sheet structure and Haemophilus influenzae and Helicobacter pylori more helical structure. Further information is available through the GeneCensus system at http://bioinfo.mbb.yale.edu/genome.

RNA polymerase subunit RPB5 plays a role in transcriptional activation

A mutation in RPB5 (rpb5-9), an essential RNA polymerase subunit assembled into RNA polymerases I, II, and III, revealed a role for this subunit in transcriptional activation. Activation by GAL4-VP16 was impaired upon in vitro transcription with mutant whole-cell extracts. In vivo experiments using inducible reporter plasmids and Northern analysis support the in vitro data and demonstrate that RPB5 influences activation at some, but not all, promoters. Remarkably, this mutation maps to a conserved region of human RPB5 implicated by others to play a role in activation. Chimeric human-yeast RPB5 containing this conserved region now can function in place of its yeast counterpart. The defects noted with rpb5-9 are similar to those seen in truncation mutants of the RPB1-carboxyl terminal domain (CTD). We demonstrate that RPB5 and the RPB1-CTD have overlapping roles in activation because the double mutant is synthetically lethal and has exacerbated activation defects at the GAL1/10 promoter. These studies demonstrate that there are multiple activation targets in RNA polymerase II and that RPB5 and the CTD have similar roles in activation.

The crystal structure of pneumococcal surface antigen PsaA reveals a metal-binding site and a novel structure for a putative ABC-type binding protein

BACKGROUND

. The surface protein PsaA of the pathogenic bacterium Streptococcus pneumoniae plays an essential role in its virulence. PsaA is a putative ATP-binding cassette-type (ABC-type) binding protein involved in the uptake of Mn2+ and possibly Zn2+ and is considered to be both a potential drug target and and a candidate vaccine component.

RESULTS

. The structure of PsaA has been determined to 2.0 A resolution using X-ray crystallography and is the first structure obtained for an ABC-type binding protein from a Gram-positive organism. The protein consists of two (beta/alpha)4 domains linked together by a single helix. A metal-binding site is formed in the domain interface by the sidechains of His67, His139, Glu205 and Asp280 and is occupied in the structure.

CONCLUSIONS

. The structural topology of PsaA is fundamentally different from that of other ABC-type binding proteins determined thus far in that PsaA lacks the characteristic 'hinge peptides' involved in conformational change upon solute uptake and release. In our structure, the metal-binding site is probably occupied by Zn2+. The site seems to be well conserved amongst related receptors from both Gram-positive and Gram-negative bacteria.

Characterization of two LGR genes homologous to gonadotropin and thyrotropin receptors with extracellular leucine-rich repeats and a G protein-coupled, seven-transmembrane region

The receptors for LH, FSH, and TSH belong to the large G protein-coupled, seven-transmembrane (TM) protein family and are unique in having a large N-terminal extracellular (ecto-) domain containing leucine-rich repeats important for interaction with the glycoprotein ligands. We have identified two new leucine-rich repeat-containing, G protein-coupled receptors and named them as LGR4 and LGR5, respectively. The ectodomains of both receptors contain 17 leucine-rich repeats together with N- and C-terminal flanking cysteine-rich sequences, compared with 9 repeats found in known glycoprotein hormone receptors. The leucine-rich repeats in LGR4 and LGR5 are arrays of 24 amino acids showing similarity to repeats found in the acid labile subunit of the insulin-like growth factor (IGF)/IGF binding protein complexes as well as slit, decorin, and Toll proteins. The TM region and the junction between ectodomain and TM 1 are highly conserved in LGR4, LGR5, and seven other LGRs from sea anemone, fly, nematode, mollusk, and mammal, suggesting their common evolutionary origin. In contrast to the restricted tissue expression of gonadotropin and TSH receptors in gonads and thyroid, respectively, LGR4 is expressed in diverse tissues including ovary, testis, adrenal, placenta, thymus, spinal cord, and thyroid, whereas LGR5 is found in muscle, placenta, spinal cord, and brain. Hybridization analysis of genomic DNA indicated that LGR4 and LGR5 genes are conserved in mammals. Comparison of overall amino acid sequences indicated that LGR4 and LGR5 are closely related to each other but diverge, during evolution, from the homologous receptor found in snail and the mammalian glycoprotein hormone receptors. The identification and characterization of new members of the LGR subfamily of receptor genes not only allow future isolation of their ligands and understanding of their physiological roles but also reveal the evolutionary relationship of G protein-coupled receptors with leucine-rich repeats.

Artificial neural networks for computer-based molecular design

The theory of artificial neural networks is briefly reviewed focusing on supervised and unsupervised techniques which have great impact on current chemical applications. An introduction to molecular descriptors and representation schemes is given. In addition, worked examples of recent advances in this field are highlighted and pioneering publications are discussed. Applications of several types of artificial neural networks to compound classification, modelling of structure-activity relationships, biological target identification, and feature extraction from biopolymers are presented and compared to other techniques. Advantages and limitations of neural networks for computer-aided molecular design and sequence analysis are discussed.

Mutant alleles of the MRS2 gene of yeast nuclear DNA suppress mutations in the catalytic core of a mitochondrial group II intron

Previous studies show that some yeast strains carrying point mutations of domain 5 that block splicing of a mitochondrial group II intron yield spontaneous revertants in which splicing is partially restored by dominant mutations of nuclear genes. Here we cloned and sequenced the suppressor allele of one such gene, and found it to be a missense mutation of the MRS2 gene (MRS2-L232F). The MRS2 gene was first implicated in group II intron splicing by the finding that overexpression of the wild-type gene weakly suppresses the splicing defect of a mutation of another intron. Tetrad analysis showed that independently isolated suppressors of two other domain 5 mutations are also allelles of the MRS2 gene and DNA sequencing identified a new missense mutation in each strain (MRS2-T230I and MRS2-L213M). All three suppressor mutations cause a temperature-sensitive respiration defect that is dominant negative in heterozygous diploids, but those strains splice the mutant intron at the elevated temperature. The three mutations are in a domain of the protein that is likely to be a helix-turn-helix region, so that effects of the mutations on protein-protein interactions may contribute to these phenotypes. These mutations suppress the splicing defect of many, but not all, of the available splicing defective mutations of aI5gamma, including mutations of several intron domains. Protein and RNA blot experiments show that the level of the protein encoded by the MRS2 gene, but not the mRNA, is elevated by these mutations. Interestingly, overexpression of the wild-type protein restores much lower levels of splicing than were obtained with similar elevated levels of the mutated Mrs2 proteins. The splicing phenotypes of these strains suggest a direct role for Mrs2 protein on group II intron splicing, but an indirect effect is not yet ruled out.

Protein structure prediction

Genome sequencing projects continue to provide a flood of new protein sequences, and prediction methods remain an important means of adding structural information. Recently, there have been advances in secondary structure prediction, which feed, in turn, into improved fold recognition algorithms. Finally, there have been technical improvements in comparative modelling, and studies of the expected accuracy of three-dimensional structural models built by this method.

First partial three-dimensional model of human monoamine oxidase A

A survey of the major known structural aspects of monoamine oxidase (MAO) is given and a first partial model of human MAO A is presented. This 3D model has been established using secondary structure predictions and fold recognition methods. It shows two alpha/beta domains (the FAD-binding N-terminal and central domains) and an alpha+beta domain. The C-terminal region is predicted to be responsible for anchoring the protein into the mitochondrial membrane and was not modeled. The covalent binding of the flavin cofactor to a cysteine residue is well predicted. The model is validated with experimental data from the literature and should be useful in designing new experimental studies (site-directed mutagenesis, chemical modification, specific antibodies). This first step towards the 3D structure of monoamine oxidase should contribute to a better understanding of the mechanisms of action and inhibition of this drug target in the treatment of clinical depression.

Enhanced antigenicity of a four-contact-residue epitope of the measles virus hemagglutinin protein by phage display libraries: evidence of a helical structure in the putative active site

Antigenicity and conformational propensities of synthetic peptides corresponding to the sequential epitope H236-255 of the measles virus hemagglutinin protein were investigated. This epitope corresponds to the neutralising and protective monoclonal antibody BH129 and includes Arg243, implicated in CD46-down-regulation and Arg253 that has been mapped to the putative enzymatic site. Fine mapping with truncation-, elongation-, Gly- and Ala-substitution analogues defined EL-QL as the critical residues of the minimal epitope S244ELSQL249. CD spectra of peptides, comparison with the 3D-structure of homologous sequences, and prediction algorithms suggested a helical structure with the contact residues E245L-QL249 located on the protein surface. Mimotopes obtained with a 6-mer phage display library contained a consensus Pro (important for binding) instead of Ser247 of the wild-type sequence (irrelevant for binding). The kink induced by Pro seemed to be essential to bring the 4 contact-residues in the mimotopes and in the corresponding short peptides together. CD analysis and prediction algorithms suggested that non-helical conformations of the phage insert and of the peptides may favourably mimic the antigenic helical turns of the wild-type sequence, resulting in an up to 135 times higher antigenicity of the mAb towards the mimotope peptides.

The Fanconi anemia complementation group A protein contains a peroxidase domain

Computational analysis of the Fanconi anemia (FA) complementation group A protein suggests that it contains a peroxidase domain. FA proteins may be part of a general mechanism that protects cells from oxidative damage.

Isolation and characterization of two knotted-like homeobox genes from tomato

Homeobox genes are known to play a role in developmental regulation. The knotted-like homeobox (knox) genes fall into two classes. The class I knox genes like knl, stml, and knatl are involved in maintaining meristem identity in cells. The function of class II knox genes is at yet undetermined. We have characterized two knox genes from tomato. LeT6 and LeT12 map to distinct chromosome locations that are different from the location for a recently cloned knox gene from tomato, tknl, confirming that plant homeobox genes are not clustered on chromosomes. These genes have a distinct expression pattern. Unlike other class I knl-like genes, LeT6 is expressed in developing lateral organs and developing ovaries in flowers. LeT12 is more ubiquitously expressed in the mature plant. RNA in situ localization data suggest that both these genes may have a role to play in formative events in ovule and embryo morphogenesis.

A family of human receptors structurally related to Drosophila Toll

The discovery of sequence homology between the cytoplasmic domains of Drosophila Toll and human interleukin 1 receptors has sown the conviction that both molecules trigger related signaling pathways tied to the nuclear translocation of Rel-type transcription factors. This conserved signaling scheme governs an evolutionarily ancient immune response in both insects and vertebrates. We report the molecular cloning of a class of putative human receptors with a protein architecture that is similar to Drosophila Toll in both intra- and extracellular segments. Five human Toll-like receptors--named TLRs 1-5--are probably the direct homologs of the fly molecule and, as such, could constitute an important and unrecognized component of innate immunity in humans. Intriguingly, the evolutionary retention of TLRs in vertebrates may indicate another role--akin to Toll in the dorsoventralization of the Drosophila embryo--as regulators of early morphogenetic patterning. Multiple tissue mRNA blots indicate markedly different patterns of expression for the human TLRs. By using fluorescence in situ hybridization and sequence-tagged site database analyses, we also show that the cognate Tlr genes reside on chromosomes 4 (TLRs 1, 2, and 3), 9 (TLR4), and 1 (TLR5). Structure prediction of the aligned Toll-homology domains from varied insect and human TLRs, vertebrate interleukin 1 receptors and MyD88 factors, and plant disease-resistance proteins recognizes a parallel beta/alpha fold with an acidic active site; a similar structure notably recurs in a class of response regulators broadly involved in transducing sensory information in bacteria.

Three-dimensional structure of HIV-1 Rev protein filaments

The HIV-1 Rev protein facilitates the export of incompletely spliced and unspliced viral mRNAs from the nucleus. Rev polymerizes into two types of filaments in vitro. In the presence of RNA, Rev forms poorly ordered structures, while in the absence of RNA it polymerizes into regular hollow filaments. We have determined the helical structure of the latter filaments by analysis of cryo-electron micrographs, taking into account STEM measurements of mass-per-unit-length. They are made up of Rev dimers, arranged in a six-start helix, with 31 dimers in 2 turns, a pitch angle of 45 degrees, and an interstrand spacing of 3.8 nm. Three-dimensional reconstruction at 2.1 nm resolution reveals a smooth outer surface and a featured inner surface, with outer and inner diameters of approximately 14.8 and approximately 10.4 nm, respectively. The Rev dimer has a "top-hat" shape with a cylinder approximately 3.2 nm in diameter and approximately 2.2 nm high, pointing inward: the thinner rim areas pack together to form the filament wall. Raman spectroscopy shows polymerized Rev to have approximately 54% alpha-helix and 20-24% beta-sheet content. Electron microdiffraction of aligned filaments reveals a broad meridional reflection at approximately (0.51 nm(-1, suggesting approximate alignment of the alpha-helices with the filament axis. Based on these data, a molecular model for the Rev filament is proposed.

Porcine cerebroside sulfate activator (saposin B) secondary structure: CD, FTIR, and NMR studies

Cerebroside sulfate activator protein (CSAct or saposin B) is one of a group of heat stable, low-molecular-weight proteins that appear to share a common structural motif. These have been referred to as saposin-like proteins and are thought to share a multiple amphipathic helical barrel structure with a conserved pattern of disulfide linkages. Porcine kidney CSAct was prepared in high purity and consisted of three major glycosylated subforms. The protein was studied by physical-chemical methods and evaluated by various methods for structural prediction. All suggest that CSAct has high amounts of alpha-helical conformation and little if any beta-sheet. Circular dichroism (CD) studies indicate 45-50% helical conformation depending on buffer and temperature. There was only a moderate loss in helical content with increasing temperature and no indication of thermal denaturation. Fourier transform infrared spectroscopy (FTIR) measurements on deuterium hydrated self-films also indicated a predominantly helical structure. Helical axis orientation was investigated by both oriented CD and FTIR dichroism, which suggested that the helical axes were roughly parallel and oriented along the axis of the surface on which the self-films had been deposited. One-dimensional nuclear magnetic resonance spectra showed large chemical shift dispersion, indicating a defined tertiary structure with little variation between 6 and 85 degrees C. NOESY spectra failed to show the strong NOE cross peaks expected for a highly helical conformation. This may indicate short-term conformational flexibility within the helices or molecular aggregation at the high protein concentrations employed. These observations are consistent with the 3-4-helix bundle motif suggested for saposin-like proteins by various predictive algorithms.

A structural census of genomes: comparing bacterial, eukaryotic, and archaeal genomes in terms of protein structure

Representative genomes from each of the three kingdoms of life are compared in terms of protein structure, in particular, those of Haemophilus influenzae (a bacteria), Methanococcus jannaschii (an archaeon), and yeast (a eukaryote). The comparison is in the form of a census (or comprehensive accounting) of the relative occurrence of secondary and tertiary structures in the genomes, which particular emphasis on patterns of supersecondary structure. Comparison of secondary structure shows that the three genomes have nearly the same overall secondary-structure content, although they differ markedly in amino acid composition. Comparison of super-secondary structure, using a novel "frequent-words" approach, shows that yeast has a preponderance of consecutive strands (e.g. beta-beta-beta patterns), Haemophilus, consecutive helices (alpha-alpha-alpha), and Methanococcus, alternating helix-strand structures (beta-alpha-beta). Yeast also has significantly more helical membrane proteins than the other two genomes, with most of the differences concentrated in proteins containing two transmembrane segments. Comparison of tertiary structure (by sequence matching and domain-level clustering) highlights the substantial duplication in each genome (approximately 30% to 50%), with the degree of duplication following similar patterns in all three. Many sequence families are shared among the genomes, with the degree of overlap between any two genomes being roughly similar. In total, the three genomes contain 148 of the approximately 300 known protein folds. Forty-five of these 148 that are present in all three genomes are especially enriched in mixed super-secondary structures (alpha/beta). Moreover, the five most common of these 45 (the "top-5") have a remarkably similar super-secondary structure architecture, containing a central sheet of parallel strands with helices packed onto at least one face and beta-alpha-beta connections between adjacent strands. These most basic molecular parts, which, presumably, were present in the last common ancestor to the three Kingdoms, include the TIM-barrel, Rossmann, flavodoxin, thiamin-binding, and P-loop-hydrolase folds.

The solution structure of the pleckstrin homology domain of human SOS1. A possible structural role for the sequential association of diffuse B cell lymphoma and pleckstrin homology domains

A large subset of pleckstrin homology (PH) domains are immediately to the C terminus of diffuse B cell lymphoma (Dbl) homology (DbH) domains. Dbl domains are generally considered to be GTPase-exchange factors; many are proto-oncogenes. PH domains appear to function as membrane-recruitment factors, or have specific protein-protein interactions. Since dual domain (DbH/PH) constructs are known to have significant properties in other pathways, it is possible that a defined interdomain relationship is required for DbH/PH function. We determined the solution structure of the human SOS1 PH domain for a construct partially extended into the preceding DbH domain. There are specific structural contacts between the PH and the vestigial DbH domain. This appears to involve structural elements common to this subfamily of PH domains, and to DbH domains. The human SOS1 PH domain binds to inositol 1,4,5-triphosphate with a approximately 60 mu M affinity. Using chemical shift titration, the binding site is identified to be essentially identical to that observed crystallographically for the inositol 1,4,5-triphosphate complex with the PH domain of phospholipase Cdelta. This site may serve as an interdomain regulator of DbH or other domains' functions. While the overall fold of the human SOS1 PH domain is similar to other PH domains, the size and position of the intrastrand loops and the presence of an N-terminal alpha-helix of the vestigial DbH domain suggest that the subfamily of PH domains associated with DbH domains may be a well defined structural group in which the PH domain is a membrane recruiter and modulator.

The Ov20 protein of the parasitic nematode Onchocerca volvulus. A structurally novel class of small helix-rich retinol-binding proteins

Ov20 is a major antigen of the parasitic nematode Onchocerca volvulus, the causative agent of river blindness in humans, and the protein is secreted into the tissue occupied by the parasite. DNA encoding Ov20 was isolated, and the protein was expressed in Escherichia coli. Fluorescence-based ligand binding assays show that the protein contains a high affinity binding site for retinol, fluorescent fatty acids (11-((5-dimethylaminonaphthalene-1-sulfonyl)amino)undecanoic acid, dansyl-DL-alpha-aminocaprylic acid, and parinaric acid) and, by competition, oleic and arachidonic acids, but not cholesterol. The fluorescence emission of dansylated fatty acids is significantly blue-shifted upon binding in comparison to similarly sized beta-sheet-rich mammalian retinol- and fatty acid-binding proteins. Secondary structure prediction algorithms indicate that a alpha-helix predominates in Ov20, possibly in a coiled coil motif, with no evidence of beta structures, and this was confirmed by circular dichroism. The protein is highly stable in solution, requiring temperatures in excess of 90 degrees C or high denaturant concentrations for unfolding. Ov20 therefore represents a novel class of small retinol-binding protein, which appears to be confined to nematodes. The retinol binding activity of Ov20 could possibly contribute to the eye defects associated with onchocerciasis and, because there is no counterpart in mammals, represents a strategic target for chemotherapy.

Modeling the paramyxovirus hemagglutinin-neuraminidase protein

The paramyxovirus hemagglutinin-neuraminidase (HN) protein exhibits neuraminidase activity and has an active site functionally similar to that in influenza neuraminidases. Earlier work identified conserved amino acids among HN sequences and proposed similarity between HN and influenza neuraminidase sequences. In this work we identify the three-dimensional fold and develop a more detailed model for the HN protein, in the process we examine a variety of protein structure prediction methods. We use the known structures of viral and bacterial neuraminidases as controls in testing the success of protein structure prediction and modeling methods, including knowledge-based threading, discrete three-dimensional environmental profiles, hidden Markov models, neural network secondary structure prediction, pattern matching, and hydropathy plots. The results from threading show that the HN protein sequence has a 6 beta-sheet propellor fold and enable us to assign the locations of the individual beta-strands. The three-dimensional environmental profile and hidden Markov model methods were not successful in this work. The model developed in this work helps to understand better the biological function of the HN protein and design inhibitors of the enzyme and serves as an assessment of some protein structure prediction methods, especially after the x-ray crystallographic solution of its structure.

Predicting protein secondary structure with probabilistic schemata of evolutionarily derived information

We demonstrate the applicability of our previously developed Bayesian probabilistic approach for predicting residue solvent accessibility to the problem of predicting secondary structure. Using only single-sequence data, this method achieves a three-state accuracy of 67% over a database of 473 non-homologous proteins. This approach is more amenable to inspection and less likely to overlearn specifics of a dataset than "black box" methods such as neural networks. It is also conceptually simpler and less computationally costly. We also introduce a novel method for representing and incorporating multiple-sequence alignment information within the prediction algorithm, achieving 72% accuracy over a dataset of 304 non-homologous proteins. This is accomplished by creating a statistical model of the evolutionarily derived correlations between patterns of amino acid substitution and local protein structure. This model consists of parameter vectors, termed "substitution schemata," which probabilistically encode the structure-based heterogeneity in the distributions of amino acid substitutions found in alignments of homologous proteins. The model is optimized for structure prediction by maximizing the mutual information between the set of schemata and the database of secondary structures. Unlike "expert heuristic" methods, this approach has been demonstrated to work well over large datasets. Unlike the opaque neural network algorithms, this approach is physicochemically intelligible. Moreover, the model optimization procedure, the formalism for predicting one-dimensional structural features and our previously developed method for tertiary structure recognition all share a common Bayesian probabilistic basis. This consistency starkly contrasts with the hybrid and ad hoc nature of methods that have dominated this field in recent years.

A Z-DNA binding domain present in the human editing enzyme, double-stranded RNA adenosine deaminase

Editing of RNA changes the read-out of information from DNA by altering the nucleotide sequence of a transcript. One type of RNA editing found in all metazoans uses double-stranded RNA (dsRNA) as a substrate and results in the deamination of adenosine to give inosine, which is translated as guanosine. Editing thus allows variant proteins to be produced from a single pre-mRNA. A mechanism by which dsRNA substrates form is through pairing of intronic and exonic sequences before the removal of noncoding sequences by splicing. Here we report that the RNA editing enzyme, human dsRNA adenosine deaminase (DRADA1, or ADAR1) contains a domain (Zalpha) that binds specifically to the left-handed Z-DNA conformation with high affinity (KD = 4 nM). As formation of Z-DNA in vivo occurs 5' to, or behind, a moving RNA polymerase during transcription, recognition of Z-DNA by DRADA1 provides a plausible mechanism by which DRADA1 can be targeted to a nascent RNA so that editing occurs before splicing. Analysis of sequences related to Zalpha has allowed identification of motifs common to this class of nucleic acid binding domain.

Progress in protein structure prediction

If protein structure prediction methods are to make any impact on the impending onerous task of analyzing the large numbers of unknown protein sequences generated by the ongoing genome-sequencing projects, it is vital that they make the difficult transition from computational 'gedankenexperiments' to practical software tools. This has already happened in the field of comparative modelling and is currently happening in the threading field. Unfortunately, there is little evidence of this transition happening in the field of ab initio tertiary-structure prediction.