Prediction of subcellular localization using sequence-biased recurrent networks

Motivation: Targeting peptides direct nascent proteins to their specific subcellular compartment. Knowledge of targeting signals enables informed drug design and reliable annotation of gene products. However, due to the low similarity of such sequences and the dynamical nature of the sorting process, the computational prediction of subcellular localization of proteins is challenging. Results: We contrast the use of feed forward models as employed by the popular TargetP/SignalP predictors with a sequence-biased recurrent network model. The models are evaluated in terms of performance at the residue level and at the sequence level, and demonstrate that recurrent networks improve the overall prediction performance. Compared to the original results reported for TargetP, an ensemble of the tested models increases the accuracy by 6 and 5% on non-plant and plant data, respectively.

Site-directed mutagenesis and kinetic studies of the West Nile virus NS3 protease identify key enzyme-substrate interactions

The flavivirus West Nile virus (WNV) has spread rapidly throughout the world in recent years causing fever, meningitis, encephalitis, and fatalities. Because the viral protease NS2B/NS3 is essential for replication, it is attracting attention as a potential therapeutic target, although there are currently no antiviral inhibitors for any flavivirus. This paper focuses on elucidating interactions between a hexapeptide substrate (Ae-KPGLKR-p-nitroanilide) and residues at S1 and S2 in the active site of WNV protease by comparing the catalytic activities of selected mutant recombinant proteases in vitro. Homology modeling enabled the predictions of key mutations in VWNV NS3 protease at S1 (V115A/F, D129A/ E/N, S135A, Y150A/F, S160A, and S163A) and S2 (N152A) that might influence substrate recognition and catalytic efficiency. Key conclusions are that the substrate P1 Arg strongly interacts with S1 residues Asp-129, Tyr-150, and Ser-163 and, to a lesser extent, Ser-160, and P2 Lys makes an essential interaction with Asn-152 at S2. The inferred substrate-enzyme interactions provide a basis for rational protease inhibitor design and optimization. High sequence conservation within flavivirus proteases means that this study may also be relevant to design of protease inhibitors for other flavivirus proteases.

Topological side-chain classification of beta-turns: Ideal motifs for peptidomimetic development

beta-turns are important topological motifs for biological recognition of proteins and peptides. Organic molecules that sample the side chain positions of beta-turns have shown broad binding capacity to multiple different receptors, for example benzodiazepines. beta-turns have traditionally been classified into various types based on the backbone dihedral angles (phi 2, psi 2, phi 3 and psi 3). Indeed, 57-68% of beta-turns are currently classified into 8 different backbone families (Type I, Type II, Type I', Type II', Type VIII, Type VIa1, Type VIa2 and Type VIb and Type IV which represents unclassified beta-turns). Although this classification of beta-turns has been useful, the resulting beta-turn types are not ideal for the design of beta-turn mimetics as they do not reflect topological features of the recognition elements, the side chains. To overcome this, we have extracted beta-turns from a data set of non-homologous and high-resolution protein crystal structures. The side chain positions, as defined by C-alpha-C-beta vectors, of these turns have been clustered using the kth nearest neighbor clustering and filtered nearest centroid sorting algorithms. Nine clusters were obtained that cluster 90% of the data, and the average intra-cluster RMSD of the four C-alpha-C-beta vectors is 0.36. The nine clusters therefore represent the topology of the side chain scaffold architecture of the vast majority of beta-turns. The mean structures of the nine clusters are useful for the development of beta-turn mimetics and as biological descriptors for focusing combinatorial chemistry towards biologically relevant topological space.

Chemical synthesis and biosynthesis of the cyclotide family of circular proteins

Cyclotides are a recently discovered class of proteins that have a characteristic head-to-tail cyclized backbone stabilized by a knotted arrangement of three disulfide bonds. They are exceptionally resistant to chemical, enzymatic and thermal treatments because of their unique structural scaffold. Cyclotides have a range of bio-activities, including uterotonic, anti-HIV, anti-bacterial and cytotoxic activity but their insecticidal properties suggest that their natural physiological role is in plant defense. They are genetically encoded as linear precursors and subsequently processed to produce mature cyclic peptides but the mechanism by which this occurs remains unknown. Currently most cyclotides are obtained via direct extraction from plants in the Rubiaceae and Violaceae families. To facilitate the screening of cyclotides for structure-activity studies and to exploit them in drug design or agricultural applications a convenient route for the synthesis of cyclotides is vital. In this review the current chemical, recombinant and biosynthetic routes to the production of cyclotides are discussed.

Factors influencing the stability of cyclotides: Proteins with a circular backbone and cystine knot motif

Cyclotides are a large family of mini-proteins that have the distinguishing features of a head-to-tail cyclised backbone and a cystine knot formed by six conserved cysteine residues. They are present in plants from the Rubiaceae, Violaceae and Cucurbitaceae families. The unique structural features of the cyclotides make them extremely resistant to chemical, thermal and proteolytic degradation. In this article we review recent Studies from our laboratory that dissect the role of the individual structural elements in defining the stability of cyclotides. The resistance of cyclotides to chemical and proteolytic degradation is in large part due to the cystine knot, whereas the thermal stability is I composite of several features including the cystine knot, the cyclic backbone and the hydrogen bonding network. A range of biological activities of cyclotides is critically dependent oil the presence of the cyclic backbone.

The absolute structural requirement for a proline in the P3 '-position of Bowman-Birk protease inhibitors is surmounted in the minimized SFTI-1 scaffold

SFTI-1 is a small cyclic peptide from sunflower seeds that is one of the most potent trypsin inhibitors of any naturally occurring peptide and is related to the Bowman-Birk family of inhibitors (BBIs). BBIs are involved in the defense mechanisms of plants and also have potential as cancer chemopreventive agents. At only 14 amino acids in size, SFTI-1 is thought to be a highly optimized scaffold of the BBI active site region, and thus it is of interest to examine its important structural and functional features. In this study, a suite of 12 alanine mutants of SFTI-1 has been synthesized, and their structures and activities have been determined. SFTI-1 incorporates a binding loop that is clasped together with a disulfide bond and a secondary peptide loop making up the circular backbone. We show here that the secondary loop stabilizes the binding loop to the consequences of sequence variations. In particular, full-length BBIs have a conserved cis-proline that has been shown previously to be required for well defined structure and potent activity, but we show here that the SFTI-1 scaffold can accommodate mutation of this residue and still have a well defined native-like conformation and nanomolar activity in inhibiting trypsin. Among the Ala mutants, the most significant structural perturbation occurred when Asp(14) was mutated, and it appears that this residue is important in stabilizing the trans peptide bond preceding Pro(13) and is thus a key residue in maintaining the highly constrained structure of SFTI-1. This aspartic acid residue is thought to be involved in the cyclization mechanism associated with excision of SFTI-1 from its 58-amino acid precursor. Overall, this mutational analysis of SFTI-1 clearly defines the optimized nature of the SFTI-1 scaffold and demonstrates the importance of the secondary loop in maintaining the active conformation of the binding loop.

The cyclotide family of circular miniproteins: Nature's combinatorial peptide template

The cyclotides are a recently discovered family of miniproteins that contain a head-to-tail cyclized backbone and a knotted arrangement of disulfide bonds. They are approximately 30 amino acids in size and are present in high abundance in plants from the Violaceae, Rubiaceae, and Cucurbitaceae families, with individual plants containing a suite of up to 100 cyclotides. They have a diverse range of biological activities, including uterotonic, anti-HIV, antitumor, and antimicrobial activities, although their natural function is likely that of defending their host plants from pathogens and pests. This review focuses on the structural aspects of cyclotides, which may be thought of as a natural combinatorial peptide template in which a wide range of amino acids is displayed on a compact molecular core made up of the cyclic cystine knot structural motif. Cyclotides are exceptionally stable and are resistant to denaturation via thermal, chemical, or enzymatic treatments. The struclural features that contribute to their remarkable stability are described ill this review. (c) 2006 Wiley Periodicals, Inc.

Insights to substrate binding and processing by West Nile Virus NS3 protease through combined modeling, protease mutagenesis, and kinetic studies

West Nile Virus is becoming a widespread pathogen, infecting people on at least four continents with no effective treatment for these infections or many of their associated pathologies. A key enzyme that is essential for viral replication is the viral protease NS2B-NS3, which is highly conserved among all flaviviruses. Using a combination of molecular fitting of substrates to the active site of the crystal structure of NS3,site-directed enzyme and cofactor mutagenesis, and kinetic studies on proteolytic processing of panels of short peptide substrates, we have identified important enzyme-substrate interactions that define substrate specificity for NS3 protease. In addition to better understanding the involvement of S2, S3, and S4 enzyme residues in substrate binding, a residue within cofactor NS2B has been found to strongly influence the preference of flavivirus proteases for lysine or arginine at P2 in substrates. Optimization of tetrapeptide substrates for enhanced protease affinity and processing efficiency has also provided important clues for developing inhibitors of West Nile Virus infection.

Alpha-selenoconotoxins, a new class of potent alpha(7) neuronal nicotinic receptor antagonists

Disulfide bonds are important structural motifs that play an essential role in maintaining the conformational stability of many bioactive peptides. Of particular importance are the conotoxins, which selectively target a wide range of ion channels that are implicated in numerous disease states. Despite the enormous potential of conotoxins as therapeutics, their multiple disulfide bond frameworks are inherently unstable under reducing conditions. Reduction or scrambling by thiol-containing molecules such as glutathione or serum albumin in intracellular or extracellular environments such as blood plasma can decrease their effectiveness as drugs. To address this issue, we describe a new class of selenoconotoxins where cysteine residues are replaced by selenocysteine to form isosteric and non-reducible diselenide bonds. Three isoforms of alpha-conotoxin ImI were synthesized by t-butoxycarbonyl chemistry with systematic replacement of one([ Sec(2,8)] ImI or [Sec(3,12)] ImI), or both([Sec(2,3,8,12)] ImI) disulfide bonds with a diselenide bond. Each analogue demonstrated remarkable stability to reduction or scrambling under a range of chemical and biological reducing conditions. Three-dimensional structural characterization by NMR and CD spectroscopy indicates conformational preferences that are very similar to those of native ImI, suggesting fully isomorphic structures. Additionally, full bioactivity was retained at the alpha(7) nicotinic acetylcholine receptor, with each seleno-analogue exhibiting a dose-response curve that overlaps with wild-type ImI, thus further supporting an isomorphic structure. These results demonstrate that selenoconotoxins can be used as highly stable scaffolds for the design of new drugs.

Multilevel modelling for inference of genetic regulatory networks

Time-course experiments with microarrays are often used to study dynamic biological systems and genetic regulatory networks (GRNs) that model how genes influence each other in cell-level development of organisms. The inference for GRNs provides important insights into the fundamental biological processes such as growth and is useful in disease diagnosis and genomic drug design. Due to the experimental design, multilevel data hierarchies are often present in time-course gene expression data. Most existing methods, however, ignore the dependency of the expression measurements over time and the correlation among gene expression profiles. Such independence assumptions violate regulatory interactions and can result in overlooking certain important subject effects and lead to spurious inference for regulatory networks or mechanisms. In this paper, a multilevel mixed-effects model is adopted to incorporate data hierarchies in the analysis of time-course data, where temporal and subject effects are both assumed to be random. The method starts with the clustering of genes by fitting the mixture model within the multilevel random-effects model framework using the expectation-maximization (EM) algorithm. The network of regulatory interactions is then determined by searching for regulatory control elements (activators and inhibitors) shared by the clusters of co-expressed genes, based on a time-lagged correlation coefficients measurement. The method is applied to two real time-course datasets from the budding yeast (Saccharomyces cerevisiae) genome. It is shown that the proposed method provides clusters of cell-cycle regulated genes that are supported by existing gene function annotations, and hence enables inference on regulatory interactions for the genetic network.

Small molecules that mimic components of bioactive protein surfaces

Small molecules designed to mimic specific structural components of a protein (peptide strands, sheets, turns, helices, or amino acids) can be expected to display agonist or antagonist biological responses by virtue of interacting with the same receptors that recognize the protein. Here we describe some minimalist approaches to structural mimetics of amino acids and of strand, turn, or helix segments of proteins. The designed molecules show potent and selective inhibition of protease, transferase, and phospholipase enzymes, or antagonism of G-protein coupled or transcriptional receptors, and have potent anti-tumour, anti-inflammatory, or antiviral activity.

GANN: Genetic algorithm neural networks for the detection of conserved combinations of features in DNA

Background: The multitude of motif detection algorithms developed to date have largely focused on the detection of patterns in primary sequence. Since sequence-dependent DNA structure and flexibility may also play a role in protein-DNA interactions, the simultaneous exploration of sequence-and structure-based hypotheses about the composition of binding sites and the ordering of features in a regulatory region should be considered as well. The consideration of structural features requires the development of new detection tools that can deal with data types other than primary sequence. Results: GANN ( available at http://bioinformatics.org.au/gann) is a machine learning tool for the detection of conserved features in DNA. The software suite contains programs to extract different regions of genomic DNA from flat files and convert these sequences to indices that reflect sequence and structural composition or the presence of specific protein binding sites. The machine learning component allows the classification of different types of sequences based on subsamples of these indices, and can identify the best combinations of indices and machine learning architecture for sequence discrimination. Another key feature of GANN is the replicated splitting of data into training and test sets, and the implementation of negative controls. In validation experiments, GANN successfully merged important sequence and structural features to yield good predictive models for synthetic and real regulatory regions. Conclusion: GANN is a flexible tool that can search through large sets of sequence and structural feature combinations to identify those that best characterize a set of sequences.

Selecting optimal instantiations of data models - Theory and validation of an ex ante approach

The schema of an information system can significantly impact the ability of end users to efficiently and effectively retrieve the information they need. Obtaining quickly the appropriate data increases the likelihood that an organization will make good decisions and respond adeptly to challenges. This research presents and validates a methodology for evaluating, ex ante, the relative desirability of alternative instantiations of a model of data. In contrast to prior research, each instantiation is based on a different formal theory. This research theorizes that the instantiation that yields the lowest weighted average query complexity for a representative sample of information requests is the most desirable instantiation for end-user queries. The theory was validated by an experiment that compared end-user performance using an instantiation of a data structure based on the relational model of data with performance using the corresponding instantiation of the data structure based on the object-relational model of data. Complexity was measured using three different Halstead metrics: program length, difficulty, and effort. For a representative sample of queries, the average complexity using each instantiation was calculated. As theorized, end users querying the instantiation with the lower average complexity made fewer semantic errors, i.e., were more effective at composing queries. (c) 2005 Elsevier B.V. All rights reserved.

Discovery of cyclotide-like protein sequences in graminaceous crop plants: Ancestral precursors of circular proteins?

Cyclotides are peptides from plants of the Rubiaceae and Violaceae families that have the unusual characteristic of a macrocylic backbone. They are further characterized by their incorporation of a cystine knot in which two disulfides, along with the intervening backbone residues, form a ring through which a third disulfide is threaded. The cyclotides have been found in every Violaceae species screened to date but are apparently present in only a few Rubiaceae species. The selective distribution reported so far raises questions about the evolution of the cyclotides within the plant kingdom. In this study, we use a combined bioinformatics and expression analysis approach to elucidate the evolution and distribution of the cyclotides in the plant kingdom and report the discovery of related sequences widespread in the Poaceae family, including crop plants such as rice ( Oryza sativa), maize ( Zea mays), and wheat ( Triticum aestivum), which carry considerable economic and social importance. The presence of cyclotide-like sequences within these plants suggests that the cyclotides may be derived from an ancestral gene of great antiquity. Quantitative RT-PCR was used to show that two of the discovered cyclotide-like genes from rice and barley ( Hordeum vulgare) have tissue-specific expression patterns.

Prediction of HLA-DQ3.2 beta ligands: evidence of multiple registers in class II binding peptides

Motivation: While processing of MHC class II antigens for presentation to helper T-cells is essential for normal immune response, it is also implicated in the pathogenesis of autoimmune disorders and hypersensitivity reactions. Sequence-based computational techniques for predicting HLA-DQ binding peptides have encountered limited success, with few prediction techniques developed using three-dimensional models. Methods: We describe a structure-based prediction model for modeling peptide-DQ3.2 beta complexes. We have developed a rapid and accurate protocol for docking candidate peptides into the DQ3.2 beta receptor and a scoring function to discriminate binders from the background. The scoring function was rigorously trained, tested and validated using experimentally verified DQ3.2 beta binding and non-binding peptides obtained from biochemical and functional studies. Results: Our model predicts DQ3.2 beta binding peptides with high accuracy [area under the receiver operating characteristic (ROC) curve A(ROC) > 0.90], compared with experimental data. We investigated the binding patterns of DQ3.2 beta peptides and illustrate that several registers exist within a candidate binding peptide. Further analysis reveals that peptides with multiple registers occur predominantly for high-affinity binders.