Prediction of protein structural class for low-similarity sequences using support vector machine and PSI-BLAST profile

Knowledge of structural class plays an important role in understanding protein folding patterns. In this study, a simple and powerful computational method, which combines support vector machine with PSI-BLAST profile, is proposed to predict protein structural class for low-similarity sequences. The evolution information encoding in the PSI-BLAST profiles is converted into a series of fixed-length feature vectors by extracting amino acid composition and dipeptide composition from the profiles. The resulting vectors are then fed to a support vector machine classifier for the prediction of protein structural class. To evaluate the performance of the proposed method, jackknife cross-validation tests are performed on two widely used benchmark datasets, 1189 (containing 1092 proteins) and 25PDB (containing 1673 proteins) with sequence similarity lower than 40% and 25%, respectively. The overall accuracies attain 70.7% and 72.9% for 1189 and 25PDB datasets, respectively. Comparison of our results with other methods shows that our method is very promising to predict protein structural class particularly for low-similarity datasets and may at least play an important complementary role to existing methods.     

A novel protein structural classes prediction method based on predicted secondary structure

Knowledge of structural classes plays an important role in understanding protein folding patterns. In this paper, features based on the predicted secondary structure sequence and the corresponding E–H sequence are extracted. Then, an 11-dimensional feature vector is selected based on a wrapper feature selection algorithm and a support vector machine (SVM). Among the 11 selected features, 4 novel features are newly designed to model the differences between α/β class and α + β class, and other 7 rational features are proposed by previous researchers. To examine the performance of our method, a total of 5 datasets are used to design and test the proposed method. The results show that competitive prediction accuracies can be achieved by the proposed method compared to existing methods (SCPRED, RKS-PPSC and MODAS), and 4 new features are demonstrated essential to differentiate α/β and α + β classes. Standalone version of the proposed method is written in JAVA language and it can be downloaded from http://web.xidian.edu.cn/slzhang/paper.html.     

A step-by-step classification algorithm of protein secondary structures based on double-layer SVM model

In this paper, a step-by-step classification algorithm based on double-layer SVM model is constructed to predict the secondary structure of proteins. The most important feature of this algorithm is to improve the prediction accuracy of α+β and α/β classes through transforming the prediction of two classes of proteins, α+β and α/β classes, with low accuracy in the past, into the prediction of all-α and all-β classes with high accuracy. A widely-used dataset, 25PDB dataset with sequence similarity lower than 40%, is used to evaluate this method. The results show that this method has good performance, and on the basis of ensuring the accuracy of other three structural classes of proteins, the accuracy of α+β class proteins is improved significantly.     

Dynameomics: A Consensus View of the Protein Unfolding/Folding Transition State Ensemble across a Diverse Set of Protein Folds

The Dynameomics project aims to simulate a representative sample of all globular protein metafolds under both native and unfolding conditions. We have identified protein unfolding transition state (TS) ensembles from multiple molecular dynamics simulations of high-temperature unfolding in 183 structurally distinct proteins. These data can be used to study individual proteins and individual protein metafolds and to mine for TS structural features common across all proteins. Separating the TS structures into four different fold classes (all proteins, all-α, all-β, and mixed α/β and α + β) resulted in no significant difference in the overall protein properties. The residues with the most contacts in the native state lost the most contacts in the TS ensemble. On average, residues beginning in an α-helix maintained more structure in the TS ensemble than did residues starting in β-strands or any other conformation. The metafolds studied here represent 67% of all known protein structures, and this is, to our knowledge, the largest, most comprehensive study of the protein folding/unfolding TS ensemble to date. One might have expected broad distributions in the average global properties of the TS relative to the native state, indicating variability in the amount of structure present in the TS. Instead, the average global properties converged with low standard deviations across metafolds, suggesting that there are general rules governing the structure and properties of the TS.     

A protein structural class prediction method based on novel features

In this study, a 12-dimensional feature vector is constructed to reflect the general contents and spatial arrangements of the secondary structural elements of a given protein sequence. Among the 12 features, 6 novel features are specially designed to improve the prediction accuracies for α/β and α + β classes based on the distributions of α-helices and β-strands and the characteristics of parallel β-sheets and anti-parallel β-sheets. To evaluate our method, the jackknife cross-validating test is employed on two widely-used datasets, 25PDB and 1189 datasets with sequence similarity lower than 40% and 25%, respectively. The performance of our method outperforms the recently reported methods in most cases, and the 6 newly-designed features have significant positive effect to the prediction accuracies, especially for α/β and α + β classes.     

High-accuracy prediction of protein structural class for low-similarity sequences based on predicted secondary structure

Information on the structural classes of proteins has been proven to be important in many fields of bioinformatics. Prediction of protein structural class for low-similarity sequences is a challenge problem. In this study, 11 features (including 8 re-used features and 3 newly-designed features) are rationally utilized to reflect the general contents and spatial arrangements of the secondary structural elements of a given protein sequence. To evaluate the performance of the proposed method, jackknife cross-validation tests are performed on two widely used benchmark datasets, 1189 and 25PDB with sequence similarity lower than 40% and 25%, respectively. Comparison of our results with other methods shows that our proposed method is very promising and may provide a cost-effective alternative to predict protein structural class in particular for low-similarity datasets.     

Prediction of mitochondrial proteins of malaria parasite using bi-profile Bayes feature extraction

Mitochondrial proteins of Plasmodium falciparum are considered as attractive targets for anti-malarial drugs, but the experimental identification of these proteins is a difficult and time-consuming task. Computational prediction of mitochondrial proteins offers an alternative approach. However, the commonly used subcellular location prediction methods are unsuited for P. falciparum mitochondrial proteins whereas the organism and organelle-specific methods were constructed on the basis of a rather small dataset. In this study, a novel dataset termed PfM233, which included 108 mitochondrial and 125 non-mitochondrial proteins with sequence similarity below 25%, was established and the methods for predicting mitochondrial proteins of P. falciparum were described. Both bi-profile Bayes and split amino acid composition were applied to extract the features from the N- and C-terminal sequences of these proteins, which were then used to construct two SVM based classifiers (PfMP-N25 and PfMP-30). Using PfM233 as the dataset, PfMP-N25 and PfMP-30 achieved accuracies (MCCs) of 90.13% (0.80) and 90.99% (0.82). When tested with the commonly used 40 mitochondrial proteins in PfM175 and the 108 mitochondrial proteins in PfM233, these two methods obviously outperformed the existing general, organelle-specific and organism and organelle-specific methods.     

Origins of structural diversity within sequentially identical hexapeptides

Efforts to predict protein secondary structure have been hampered by the apparent structural plasticity of local amino acid sequences. Kabsch and Sander (1984, Proc. Natl. Acad. Sci. USA 81, 1075–1078) articulated this problem by demonstrating that identical pentapeptide sequences can adopt distinct structures in different proteins. With the increased size of the protein structure database and the availability of new methods to characterize structural environments, we revisit this observation of structural plasticity. Within a set of proteins with less than 50% sequence identity, 59 pairs of identical hexapeptide sequences were identified. These local structures were compared and their surrounding structural environments examined. Within a protein structural class (α/α, β/β, α/β, α + β), the structural similarity of sequentially identical hexapeptides usually is preserved. This study finds eight pairs of identical hexapeptide sequences that adopt β-strand structure in one protein and α-helical structure in the other. In none of the eight cases do the members of these sequence pairs come from proteins within the same folding class. These results have implications for class dependent secondary structure prediction algorithms.     

Novel approach for selecting the best predictor for identifying the binding sites in DNA binding proteins

Protein–DNA complexes play vital roles in many cellular processes by the interactions of amino acids with DNA. Several computational methods have been developed for predicting the interacting residues in DNA-binding proteins using sequence and/or structural information. These methods showed different levels of accuracies, which may depend on the choice of data sets used in training, the feature sets selected for developing a predictive model, the ability of the models to capture information useful for prediction or a combination of these factors. In many cases, different methods are likely to produce similar results, whereas in others, the predictors may return contradictory predictions. In this situation, a priori estimates of prediction performance applicable to the system being investigated would be helpful for biologists to choose the best method for designing their experiments. In this work, we have constructed unbiased, stringent and diverse data sets for DNA-binding proteins based on various biologically relevant considerations: (i) seven structural classes, (ii) 86 folds, (iii) 106 superfamilies, (iv) 194 families, (v) 15 binding motifs, (vi) single/double-stranded DNA, (vii) DNA conformation (A, B, Z, etc.), (viii) three functions and (ix) disordered regions. These data sets were culled as non-redundant with sequence identities of 25 and 40% and used to evaluate the performance of 11 different methods in which online services or standalone programs are available. We observed that the best performing methods for each of the data sets showed significant biases toward the data sets selected for their benchmark. Our analysis revealed important data set features, which could be used to estimate these context-specific biases and hence suggest the best method to be used for a given problem. We have developed a web server, which considers these features on demand and displays the best method that the investigator should use. The web server is freely available at http://www.biotech.iitm.ac.in/DNA-protein/. Further, we have grouped the methods based on their complexity and analyzed the performance. The information gained in this work could be effectively used to select the best method for designing experiments.     

Physicochemical and immunological characterization of the type E botulinum neurotoxin binding protein purified fromClostridium botulinum

Type E botulinum neurotoxin is produced byClostridium botulinum along with a neurotoxin binding protein which helps protect the neurotoxin from adversepH, temperature, and proteolytic conditions. The neurotoxin binding protein has been purified as a 118-kDa protein. Secondary structure content of the neurotoxin binding protein as revealed by far-UV circular dichroism spectroscopy was 19% α-helix, 50%β-sheets, 28% random coils, and 3%β-turns. This compared to 22% α-helix, 44%β-sheets, 34% random coils, and noβ-turns of the type E botulinum neurotoxin. The complex of the two proteins revealed 25%α-helix, 45%β-sheets, 27% random coils, and 3%β-turns, suggesting a significant alteration at least in theα-helical folding of the two proteins upon their interaction. Tyrosine topography is altered considerably (28%) when the neurotoxin and its binding protein are separated, indicating strong interaction between the two proteins. Gel filtration results suggested that type E neurotoxin binding protein clearly complexes with type E neurotoxin. The interaction is favored at lowpH as indicated by an initial binding rate of 8.4 min^?1 atpH 5.7 compared to 4.0 min^?1 atpH 7.5 as determined using a fiber optic-based biosensor. The neurotoxin and its binding protein apparently are of equivalent antigenicity, as both reacted equally on enzyme-linked immunosorbent assay to polyclonal antibodies raised against the toxoid of their complex.     

Experimental Evidence for the Existence of a Stable Half-Barrel Subdomain in the (β/α)8-Barrel Fold

The (β/α)₈-barrel is one of the most common folds functioning as enzymes. The emergence of two (β/α)₈-barrel enzymes involved in histidine biosynthesis, each of which has a twofold symmetric structure, has been proposed to be a consequence of tandem duplication and fusion of a (β/α)₄-half-barrel. However, little evidence has been found for the existence of an ancestral half-barrel in the evolution of other (β/α)₈-barrel proteins. In order to detect remnants of an ancestral half-barrel in the (β/α)₈-barrel structure of Escherichia coli N-(5′-phosphoribosyl)anthranilate isomerase, we engineered three potential half-barrel units, (β/α)_1-4, (β/α)_3-6, and (β/α)_5-8. Among these three arrangements, only (β/α)_3-6 is stable; it exists in equilibrium between monomeric and dimeric forms. Thus, the central segment of N-(5′-phosphoribosyl)anthranilate isomerase from E. coli can serve as a half-barrel precursor. A tandem duplication of (β/α)_3-6 yielded predominantly monomeric structures that were quite stable. This result exemplified that the structural characteristics of noncovalently assembled half-barrels could be improved by duplication and fusion. Moreover, our results may provide information regarding the local structural units that encompass interactions important for the early folding events of this ubiquitous protein conformation.     

Activators of G proteins inhibit GSK-3β and stabilize β-Catenin in Xenopus oocytes

Frizzled proteins, the receptors for Wnt ligands have seven hydrophobic transmembrane domains, a structural feature of G protein coupled receptors. Therefore a role for G proteins in the regulation of Wnt signaling has been proposed. Here I have used Xenopus oocytes to study the role of heterotrimeric G proteins in the regulation of GSK-3β and β-Catenin, two essential components of the canonical Wnt pathway. In these cells, general activators of G proteins such as GTPγ-S and AlF4⁻ increase β-Catenin stability and decrease GSK-3β mediated phosphorylation of the microtubule associated protein, Tau. Among several members of Gα proteins tested, expression of a constitutively active mutant of Gαq (GαqQL) led to a significant increase in accumulation of β-Catenin. The stabilization of β-Catenin mediated by Gαq was reversed by a Gαq specific inhibitor, Gp-antagonist 2A, but not by a specific blocking peptide for Gαs. Expression of GαqQL also inhibited GSK-3β-mediated tau phosphorylation in Xenopus oocytes. These results support a role for the Gq class of G proteins in the regulation of Wnt/β-Catenin signal transduction.     

Using principal component analysis and support vector machine to predict protein structural class for low-similarity sequences via PSSM

The accurate identification of protein structure class solely using extracted information from protein sequence is a complicated task in the current computational biology. Prediction of protein structural class for low-similarity sequences remains a challenging problem. In this study, the new computational method has been developed to predict protein structural class by fusing the sequence information and evolution information to represent a protein sample. To evaluate the performance of the proposed method, jackknife cross-validation tests are performed on two widely used benchmark data-sets, 1189 and 25PDB with sequence similarity lower than 40 and 25%, respectively. Comparison of our results with other methods shows that the proposed method by us is very promising and may provide a cost-effective alternative to predict protein structural class in particular for low-similarity data-sets.     

Binding and Cleavage of E. coli HUβ by the E. coli Lon Protease

The Escherichia coli Lon protease degrades the E. coli DNA-binding protein HUβ, but not the related protein HUα. Here we show that the Lon protease binds to both HUβ and HUα, but selectively degrades only HUβ in the presence of ATP. Mass spectrometry of HUβ peptide fragments revealed that region K18-G22 is the preferred cleavage site, followed in preference by L36-K37. The preferred cleavage site was further refined to A20-A21 by constructing and testing mutant proteins; Lon degraded HUβ-A20Q and HUβ-A20D more slowly than HUβ. We used optical tweezers to measure the rupture force between HU proteins and Lon; HUα, HUβ, and HUβ-A20D can bind to Lon, and in the presence of ATP, the rupture force between each of these proteins and Lon became weaker. Our results support a mechanism of Lon protease cleavage of HU proteins in at least three stages: binding of Lon with the HU protein (HUβ, HUα, or HUβ-A20D); hydrolysis of ATP by Lon to provide energy to loosen the binding to the HU protein and to allow an induced-fit conformational change; and specific cleavage of only HUβ.     

Using principal component analysis and support vector machine to predict protein structural class for low-similarity sequences via PSSM

The accurate identification of protein structure class solely using extracted information from protein sequence is a complicated task in the current computational biology. Prediction of protein structural class for low-similarity sequences remains a challenging problem. In this study, the new computational method has been developed to predict protein structural class by fusing the sequence information and evolution information to represent a protein sample. To evaluate the performance of the proposed method, jackknife cross-validation tests are performed on two widely used benchmark data-sets, 1189 and 25PDB with sequence similarity lower than 40 and 25%, respectively. Comparison of our results with other methods shows that the proposed method by us is very promising and may provide a cost-effective alternative to predict protein structural class in particular for low-similarity data-sets.     

Remote protein homology detection using recurrence quantification analysis and amino acid physicochemical properties

Remote homology detection refers to the detection of structure homology in evolutionarily related proteins with low sequence similarity. Supervised learning algorithms such as support vector machine (SVM) are currently the most accurate methods. In most of these SVM-based methods, efforts have been dedicated to developing new kernels to better use the pairwise alignment scores or sequence profiles. Moreover, amino acids’ physicochemical properties are not generally used in the feature representation of protein sequences. In this article, we present a remote homology detection method that incorporates two novel features: (1) a protein's primary sequence is represented using amino acid's physicochemical properties and (2) the similarity between two proteins is measured using recurrence quantification analysis (RQA). An optimization scheme was developed to select different amino acid indices (up to 10 for a protein family) that are best to characterize the given protein family. The selected amino acid indices may enable us to draw better biological explanation of the protein family classification problem than using other alignment-based methods. An SVM-based classifier will then work on the space described by the RQA metrics. The classification scheme is named as SVM-RQA. Experiments at the superfamily level of the SCOP1.53 dataset show that, without using alignment or sequence profile information, the features generated from amino acid indices are able to produce results that are comparable to those obtained by the published state-of-the-art SVM kernels. In the future, better prediction accuracies can be expected by combining the alignment-based features with our amino acids property-based features. Supplementary information including the raw dataset, the best-performing amino acid indices for each protein family and the computed RQA metrics for all protein sequences can be downloaded from http://ym151113.ym.edu.tw/svm-rqa.     

Cycloartane-type glycosides from Astragalus amblolepis

Five cycloartane-type triterpene glycosides were isolated from the methanol extract of the roots of Astragalus amblolepis Fischer along with one known saponin, 3-O-β-D-xylopyranosyl-16-O-β-D-glucopyranosyl-3β,6α,16β,24(S),25-pentahydroxy-cycloartane. Structures of the compounds were established as 3-O-β-D-xylopyranosyl-25-O-β-D-glucopyranosyl-3β,6α,16β,24(S),25-pentahydroxy-cycloartane, 3-O-[β-D-glucuronopyranosyl-(1 → 2)-β-D-xylopyranosyl]-25-O-β-D-glucopyranosyl-3β,6α,16β,24(S),25-pentahydroxy-cycloartane, 3-O-β-D-xylopyranosyl-24,25-di-O-β-D-glucopyranosyl-3β,6α,16β,24(S),25-pentahydroxy-cycloartane, 6-O-α-L-rhamnopyranosyl-16,24-di-O-β-D-glucopyranosyl-3β,6α,16β,24(S),25-pentahydroxy-cycloartane, 6-O-α-L-rhamnopyranosyl-16,25-di-O-β-D-glucopyranosyl-3β,6α,16β,24(S),25-pentahydroxy-cycloartane by using 1D and 2D-NMR techniques and mass spectrometry. To the best of our knowledge, the glucuronic acid moiety in cycloartanes is reported for the first time.     

Dodecyl Maltoside Protects Membrane Proteins In Vacuo

Molecular dynamics simulations have been used to characterize the effects of transfer from aqueous solution to a vacuum to inform our understanding of mass spectrometry of membrane-protein-detergent complexes. We compared two membrane protein architectures (an α-helical bundle versus a β-barrel) and two different detergent types (phosphocholines versus an alkyl sugar) with respect to protein stability and detergent packing. The β-barrel membrane protein remained stable as a protein-detergent complex in vacuum. Zwitterionic detergents formed conformationally destabilizing interactions with an α-helical membrane protein after detergent micelle inversion driven by dehydration in vacuum. In contrast, a nonionic alkyl sugar detergent resisted micelle inversion, maintaining the solution-phase conformation of the protein. This helps to explain the relative stability of membrane proteins in the presence of alkyl sugar detergents such as dodecyl maltoside.     

Geometrical Membrane Curvature as an Allosteric Regulator of Membrane Protein Structure and Function

Transmembrane proteins are embedded in cellular membranes of varied lipid composition and geometrical curvature. Here, we studied for the first time the allosteric effect of geometrical membrane curvature on transmembrane protein structure and function. We used single-channel optical analysis of the prototypic transmembrane β-barrel α-hemolysin (α-HL) reconstituted on immobilized single small unilamellar liposomes of different diameter and therefore curvature. Our data demonstrate that physiologically abundant geometrical membrane curvatures can enforce a dramatic allosteric regulation (1000-fold inhibition) of α-HL permeability. High membrane curvatures (1/diameter ∼1/40 nm⁻¹) compressed the effective pore diameter of α-HL from 14.2 ± 0.8 Å to 11.4 ± 0.6 Å. This reduction in effective pore area (∼40%) when combined with the area compressibility of α-HL revealed an effective membrane tension of ∼50 mN/m and a curvature-imposed protein deformation energy of ∼7 k_BT. Such substantial energies have been shown to conformationally activate, or unfold, β-barrel and α-helical transmembrane proteins, suggesting that membrane curvature could likely regulate allosterically the structure and function of transmembrane proteins in general.     

Structural Determination of Aβ25-35 Micelles by Molecular Dynamics Simulations

Amyloid-β (Aβ) peptides and other amyloidogenic proteins can form a wide range of soluble oligomers of varied morphologies at the early aggregation stage, and some of these oligomers are biologically relevant to the pathogenesis of Alzheimer's disease. Spherical micelle-like oligomers have been often observed for many different types of amyloids. Here, we report a hybrid computational approach to systematically construct, search, optimize, and rank soluble micelle-like Aβ_25-35 structures with different side-chain packings at the atomic level. Simulations reveal for the first time, to our knowledge, that two Aβ micelles with antiparallel peptide organization and distinct surface hydrophobicity display high structural stability. Stable micelles experience a slow secondary structural transition from turn to α-helix. Energetic analysis coupled with computational mutagenesis reveals that van der Waals and solvation energies play a more pronounced role in stabilizing the micelles, whereas the electrostatic energies present a stable but minor energetic contribution to peptide assemblies. Modeled Aβ micelles with shapes and dimensions similar to those of experimentally derived spherical structures also provide detailed information about the roles of structural dynamics and transition in the formation of amyloid fibrils. The strong binding affinity of our micelles to antibodies implies that micelles may be a biologically relevant species.