FlexPred: a web-server for predicting residue positions involved in conformational switches in proteins

Conformational switches observed in the protein backbone play a key role in a variety of fundamental biological activities. This paper describes a web-server that implements a pattern recognition algorithm trained on the examples from the Database of Macromolecular Movements to predict residue positions involved in conformational switches. Prediction can be performed at an adjustable false positive rate using a user-supplied protein sequence in FASTA format or a structure in a Protein Data Bank (PDB) file. If a protein sequence is submitted, then the web-server uses sequence-derived information only (such as evolutionary conservation of residue positions). If a PDB file is submitted, then the web-server uses sequence-derived information and residue solvent accessibility calculated from this file.     

Analysing six types of protein-protein interfaces

Non-covalent residue side-chain interactions occur in many different types of proteins and facilitate many biological functions. Are these differences manifested in the sequence compositions and/or the residue-residue contact preferences of the interfaces? Previous studies analysed small data sets and gave contradictory answers. Here, we introduced a new data-mining method that yielded the largest high-resolution data set of interactions analysed. We introduced an information theory-based analysis method. On the basis of sequence features, we were able to differentiate six types of protein interfaces, each corresponding to a different functional or structural association between residues. Particularly, we found significant differences in amino acid composition and residue-residue preferences between interactions of residues within the same structural domain and between different domains, between permanent and transient interfaces, and between interactions associating homo-oligomers and hetero-oligomers. The differences between the six types were so substantial that, using amino acid composition alone, we could predict statistically to which of the six types of interfaces a pool of 1000 residues belongs at 63-100% accuracy. All interfaces differed significantly from the background of all residues in SWISS-PROT, from the group of surface residues, and from internal residues that were not involved in non-trivial interactions. Overall, our results suggest that the interface type could be predicted from sequence and that interface-type specific mean-field potentials may be adequate for certain applications.     

Prediction-based fingerprints of protein-protein interactions

The recognition of protein interaction sites is an important intermediate step toward identification of functionally relevant residues and understanding protein function, facilitating experimental efforts in that regard. Toward that goal, the authors propose a novel representation for the recognition of protein-protein interaction sites that integrates enhanced relative solvent accessibility (RSA) predictions with high resolution structural data. An observation that RSA predictions are biased toward the level of surface exposure consistent with protein complexes led the authors to investigate the difference between the predicted and actual (i.e., observed in an unbound structure) RSA of an amino acid residue as a fingerprint of interaction sites. The authors demonstrate that RSA prediction-based fingerprints of protein interactions significantly improve the discrimination between interacting and noninteracting sites, compared with evolutionary conservation, physicochemical characteristics, structure-derived and other features considered before. On the basis of these observations, the authors developed a new method for the prediction of protein-protein interaction sites, using machine learning approaches to combine the most informative features into the final predictor. For training and validation, the authors used several large sets of protein complexes and derived from them nonredundant representative chains, with interaction sites mapped from multiple complexes. Alternative machine learning techniques are used, including Support Vector Machines and Neural Networks, so as to evaluate the relative effects of the choice of a representation and a specific learning algorithm. The effects of induced fit and uncertainty of the negative (noninteracting) class assignment are also evaluated. Several representative methods from the literature are reimplemented to enable direct comparison of the results. Using rigorous validation protocols, the authors estimated that the new method yields the overall classification accuracy of about 74% and Matthews correlation coefficients of 0.42, as opposed to up to 70% classification accuracy and up to 0.3 Matthews correlation coefficient for methods that do not utilize RSA prediction-based fingerprints. The new method is available at http://sppider.cchmc.org.     

Comprehensive statistical analysis of residues interaction specificity at protein-protein interfaces

We calculated interchain contacts on the atomic level for nonredundant set of 4602 protein-protein interfaces using an unbiased Voronoi-Delaune tessellation method, and made 20x20 residue contact matrixes both for homodimers and heterocomplexes. The area of contacts and the distance distribution for these contacts were calculated on both the residue and the atomic levels. We analyzed residue area distribution and showed the existence of two types of interresidue contacts: stochastic and specific. We also derived formulas describing the distribution of contact area for stochastic and specific interactions in parametric form. Maximum pairing preference index was found for Cys-Cys contacts and for oppositely charged interactions. A significant difference in residue contacts was observed between homodimers and heterocomplexes. Interfaces in homodimers were enriched with contacts between residues of the same type due to the effects of structure symmetry.     

Atomic contact vectors in protein-protein recognition

The ability to analyze and compare protein-protein interactions on the structural level is critical to our understanding of various aspects of molecular recognition and the functional interplay of components of biochemical networks. In this study, we introduce atomic contact vectors (ACVs) as an intuitive way to represent the physico-chemical characteristics of a protein-protein interface as well as a way to compare interfaces to each other. We test the utility of ACVs in classification by using them to distinguish between homodimers and crystal contacts. Our results compare favorably with those reported by other authors. We then apply ACVs to mine the PDB for all known protein-protein complexes and separate transient recognition complexes from permanent oligomeric ones. Getting at the basis of this difference is important for our understanding of recognition and we achieved a success rate of 91% for distinguishing these two classes of complexes. Although accessible surface area of the interface is a major discriminating feature, we also show that there are distinct differences in the contact preferences between the two kinds of complexes. Illustrating the superiority of ACVs as a basic comparison measure over a sequence-based approach, we derive a general rule of thumb to determine whether two protein-protein interfaces are redundant. With this method, we arrive at a nonredundant set of 209 recognition complexes--the largest set reported so far.     

Interaction geometry involving planar groups in protein-protein interfaces

The geometry of interactions of planar residues is nonrandom in protein tertiary structures and gives rise to conventional, as well as nonconventional (X--H...pi, X--H...O, where X = C, N, or O) hydrogen bonds. Whether a similar geometry is maintained when the interaction is across the protein-protein interface is addressed here. The relative geometries of interactions involving planar residues, and the percentage of contacts giving rise to different types of hydrogen bonds are quite similar in protein structures and the biological interfaces formed by protein chains in homodimers and protein-protein heterocomplexes--thus pointing to the similarity of chemical interactions that occurs during protein folding and binding. However, the percentage is considerably smaller in the nonspecific and nonphysiological interfaces that are formed in crystal lattices of monomeric proteins. The C--H...O interaction linking the aromatic and the peptide groups is quite common in protein structures as well as the three types of interfaces. However, as the interfaces formed by crystal contacts are depleted in aromatic residues, the weaker hydrogen bond interactions would contribute less toward their stability.     

Structural segments and residue propensities in protein-RNA interfaces: comparison with protein-protein and protein-DNA complexes

The interface of a protein molecule that is involved in binding another protein, DNA or RNA has been characterized in terms of the number of unique secondary structural segments (SSSs), made up of stretches of helix, strand and non-regular (NR) regions. On average 10-11 segments define the protein interface in protein-protein (PP) and protein-DNA (PD) complexes, while the number is higher (14) for protein-RNA (PR) complexes. While the length of helical segments in PP interaction increases with the interface area, this is not the case in PD and PR complexes. The propensities of residues to occur in the three types of secondary structural elements (SSEs) in the interface relative to the corresponding elements in the protein tertiary structures have been calculated. Arg, Lys, Asn, Tyr, His and Gln are preferred residues in PR complexes; in addition, Ser and Thr are also favoured in PD interfaces.     

面向蛋白质功能位点识别的机器学习平台构建

有关蛋白质功能的研究是解析生命奥秘的基础,机器学习技术在该领域已有广泛应用。利用支持向量机(support vectormachine,SVM)方法,构建一个预测蛋白质功能位点的通用平台。该平台先提取非同源蛋白质序列,再对这些序列进行特征编码(包括序列的基本信息、物化特征、结构信息及序列保守性特征等),以编码好的样本作为训练数据,利用SVM进行训练,得到敏感性、特异性、Matthew相关系数、准确率及ROC曲线等评价指标,反复测试,得到评价指标最优的SVM模型后,便可以用来预测蛋白质序列上的功能位点。该平台除了应用在预测蛋白质功能位点之外,还可以应用于疾病相关单核苷酸多态性(SNP)预测分析、预测蛋白质结构域分析、生物分子间的相互作用等。     

Dissection, residue conservation, and structural classification of protein-DNA interfaces

The basic DNA-binding modules of 128 protein-DNA interfaces have been analyzed. Although these are less planar, like the protein-protein interfaces, the protein-DNA interfaces can also be dissected into core regions in which all the fully-buried atoms are located, and rim regions having atoms with residual accessibilities. The sequence entropy of the core residues is smaller than those in the rim, indicating that the former are better conserved and possibly contribute more towards the binding free energy, as has been implicated in protein-protein interactions. On the protein side, 1014 A(2) of the surface is buried of which 63% belong to the core. There are some differences in the propensities of residues to occur in the core and the rim. In the DNA strands, the nucleotide(s) containing fully-buried atoms in all three components usually occupy central positions of the binding region. A new classification scheme for the interfaces has been introduced based on the composition of secondary structural elements of residues and the results compared with the conventional classification of DNA-binding proteins, as well as the protein class of the molecule. It appears that a common framework may be developed to understand both protein-protein and protein-DNA interactions.     

Insights into protein-protein interfaces using a Bayesian network prediction method

Identifying the interface between two interacting proteins provides important clues to the function of a protein, and is becoming increasing relevant to drug discovery. Here, surface patch analysis was combined with a Bayesian network to predict protein-protein binding sites with a success rate of 82% on a benchmark dataset of 180 proteins, improving by 6% on previous work and well above the 36% that would be achieved by a random method. A comparable success rate was achieved even when evolutionary information was missing, a further improvement on our previous method which was unable to handle incomplete data automatically. In a case study of the Mog1p family, we showed that our Bayesian network method can aid the prediction of previously uncharacterised binding sites and provide important clues to protein function. On Mog1p itself a putative binding site involved in the SLN1-SKN7 signal transduction pathway was detected, as was a Ran binding site, previously characterized solely by conservation studies, even though our automated method operated without using homologous proteins. On the remaining members of the family (two structural genomics targets, and a protein involved in the photosystem II complex in higher plants) we identified novel binding sites with little correspondence to those on Mog1p. These results suggest that members of the Mog1p family bind to different proteins and probably have different functions despite sharing the same overall fold. We also demonstrated the applicability of our method to drug discovery efforts by successfully locating a number of binding sites involved in the protein-protein interaction network of papilloma virus infection. In a separate study, we attempted to distinguish between the two types of binding site, obligate and non-obligate, within our dataset using a second Bayesian network. This proved difficult although some separation was achieved on the basis of patch size, electrostatic potential and conservation. Such was the similarity between the two interacting patch types, we were able to use obligate binding site properties to predict the location of non-obligate binding sites and vice versa.     

Specificity of molecular interactions in transient protein-protein interaction interfaces

In this study, we investigate what types of interactions are specific to their biological function, and what types of interactions are persistent regardless of their functional category in transient protein-protein heterocomplexes. This is the first approach to analyze protein-protein interfaces systematically at the molecular interaction level in the context of protein functions. We perform systematic analysis at the molecular interaction level using classification and feature subset selection technique prevalent in the field of pattern recognition. To represent the physicochemical properties of protein-protein interfaces, we design 18 molecular interaction types using canonical and noncanonical interactions. Then, we construct input vector using the frequency of each interaction type in protein-protein interface. We analyze the 131 interfaces of transient protein-protein heterocomplexes in PDB: 33 protease-inhibitors, 52 antibody-antigens, 46 signaling proteins including 4 cyclin dependent kinase and 26 G-protein. Using kNN classification and feature subset selection technique, we show that there are specific interaction types based on their functional category, and such interaction types are conserved through the common binding mechanism, rather than through the sequence or structure conservation. The extracted interaction types are C(alpha)-- H...O==C interaction, cation...anion interaction, amine...amine interaction, and amine...cation interaction. With these four interaction types, we achieve the classification success rate up to 83.2% with leave-one-out cross-validation at k = 15. Of these four interaction types, C(alpha)--H...O==C shows binding specificity for protease-inhibitor complexes, while cation-anion interaction is predominant in signaling complexes. The amine ... amine and amine...cation interaction give a minor contribution to the classification accuracy. When combined with these two interactions, they increase the accuracy by 3.8%. In the case of antibody-antigen complexes, the sign is somewhat ambiguous. From the evolutionary perspective, while protease-inhibitors and sig-naling proteins have optimized their interfaces to suit their biological functions, antibody-antigen interactions are the happenstance, implying that antibody-antigen complexes do not show distinctive interaction types. Persistent interaction types such as pi...pi, amide-carbonyl, and hydroxyl-carbonyl interaction, are also investigated. Analyzing the structural orientations of the pi...pi stacking interactions, we find that herringbone shape is a major configuration in transient protein-protein interfaces. This result is different from that of protein core, where parallel-displaced configurations are the major configuration. We also analyze overall trend of amide-carbonyl and hydroxyl-carbonyl interactions. It is noticeable that nearly 82% of the interfaces have at least one hydroxyl-carbonyl interactions.     

Hydrophobic folding units at protein-protein interfaces: implications to protein folding and to protein-protein association.

A hydrophobic folding unit cutting algorithm, originally developed for dissecting single-chain proteins, has been applied to a dataset of dissimilar two-chain protein-protein interfaces. Rather than consider each individual chain separately, the two-chain complex has been treated as a single chain. The two-chain parsing results presented in this work show hydrophobicity to be a critical attribute of two-state versus three-state protein-protein complexes. The hydrophobic folding units at the interfaces of two-state complexes suggest that the cooperative nature of the two-chain protein folding is the outcome of the hydrophobic effect, similar to its being the driving force in a single-chain folding. In analogy to the protein-folding process, the two-chain, two-state model complex may correspond to the formation of compact, hydrophobic nuclei. On the other hand, the three-state model complex involves binding of already folded monomers, similar to the association of the hydrophobic folding units within a single chain. The similarity between folding entities in protein cores and in two-state protein-protein interfaces, despite the absence of some chain connectivities in the latter, indicates that chain linkage does not necessarily affect the native conformation. This further substantiates the notion that tertiary, non-local interactions play a critical role in protein folding. These compact, hydrophobic, two-chain folding units, derived from structurally dissimilar protein-protein interfaces, provide a rich set of data useful in investigations of the role played by chain connectivity and by tertiary interactions in studies of binding and of folding. Since they are composed of non-contiguous pieces of protein backbones, they may also aid in defining folding nuclei.     

Binding interface prediction by combining protein–protein docking results

We developed a method called residue contact frequency (RCF), which uses the complex structures generated by the protein–protein docking algorithm ZDOCK to predict interface residues. Unlike interface prediction algorithms that are based on monomers alone, RCF is binding partner specific. We evaluated the performance of RCF using the area under the precision‐recall (PR) curve (AUC) on a large protein docking Benchmark. RCF (AUC = 0.44) performed as well as meta‐PPISP (AUC = 0.43), which is one of the best monomer‐based interface prediction methods. In addition, we test a support vector machine (SVM) to combine RCF with meta‐PPISP and another monomer‐based interface prediction algorithm Evolutionary Trace to further improve the performance. We found that the SVM that combined RCF and meta‐PPISP achieved the best performance (AUC = 0.47). We used RCF to predict the binding interfaces of proteins that can bind to multiple partners and RCF was able to correctly predict interface residues that are unique for the respective binding partners. Furthermore, we found that residues that contributed greatly to binding affinity (hotspot residues) had significantly higher RCF than other residues. Proteins 2014; 82:57–66. © 2013 Wiley Periodicals, Inc.     

Predicting the helix‐helix interactions from correlated residue mutations

Helix‐helix interactions are crucial in the structure assembly, stability and function of helix‐rich proteins including many membrane proteins. In spite of remarkable progresses over the past decades, the accuracy of predicting protein structures from their amino acid sequences is still far from satisfaction. In this work, we focused on a simpler problem, the prediction of helix‐helix interactions, the results of which could facilitate practical protein structure prediction by constraining the sampling space. Specifically, we started from the noisy 2D residue contact maps derived from correlated residue mutations, and utilized ridge detection to identify the characteristic residue contact patterns for helix‐helix interactions. The ridge information as well as a few additional features were then fed into a machine learning model HHConPred to predict interactions between helix pairs. In an independent test, our method achieved an F‐measure of ～60% for predicting helix‐helix interactions. Moreover, although the model was trained mainly using soluble proteins, it could be extended to membrane proteins with at least comparable performance relatively to previous approaches that were generated purely using membrane proteins. All data and source codes are available at http://166.111.152.91/Downloads.html or https://github.com/dpxiong/HHConPred .     

Exploiting sequence and structure homologs to identify protein-protein binding sites

A rapid increase in the number of experimentally derived three-dimensional structures provides an opportunity to better understand and subsequently predict protein-protein interactions. In this study, structurally conserved residues were derived from multiple structure alignments of the individual components of known complexes and the assigned conservation score was weighted based on the crystallographic B factor to account for the structural flexibility that will result in a poor alignment. Sequence profile and accessible surface area information was then combined with the conservation score to predict protein-protein binding sites using a Support Vector Machine (SVM). The incorporation of the conservation score significantly improved the performance of the SVM. About 52% of the binding sites were precisely predicted (greater than 70% of the residues in the site were identified); 77% of the binding sites were correctly predicted (greater than 50% of the residues in the site were identified), and 21% of the binding sites were partially covered by the predicted residues (some residues were identified). The results support the hypothesis that in many cases protein interfaces require some residues to provide rigidity to minimize the entropic cost upon complex formation.     

Protein stability and in vivo concentration of missense mutations in phenylalanine hydroxylase

A previous computational analysis of missense mutations linked to monogenic disease found a high proportion of missense mutations affect protein stability, rather than other aspects of protein structure and function. The purpose of this study is to relate the presence of such stability damaging missense mutations to the levels of a particular protein present under "in vivo" like conditions, and to test the reliability of the computational methods. Experimental data on a set of missense mutations of the enzyme phenylalanine hydroxylase (PAH) associated with the monogenic disease phenylketonuria (PKU) have been compared with the expected in vivo impact on protein function, obtained using SNPs3D, an in silico analysis package. A high proportion of the PAH mutations are predicted to be destabilizing. The overall agreement between predicted stability impact and experimental evidence for lower protein levels is in accordance with the estimated error rates of the methods. For these mutations, destabilization of protein three-dimensional structure is the major molecular mechanism leading to PKU, and results in a substantial reduction of in vivo PAH protein concentration. Although of limited scale, the results support the view that destabilization is the most common mechanism by which missense mutations cause monogenic disease. In turn, this conclusion suggests the general therapeutic strategy of developing drugs targeted at restoring wild type stability.     

Using physical potentials and learned models to distinguish native binding interfaces from de novo designed interfaces that do not bind

Protein–protein interactions are a fundamental aspect of many biological processes. The advent of recombinant protein and computational techniques has allowed for the rational design of proteins with novel binding capabilities. It is therefore desirable to predict which designed proteins are capable of binding in vitro. To this end, we have developed a learned classification model that combines energetic and non‐energetic features. Our feature set is adapted from specialized potentials for aromatic interactions, hydrogen bonds, electrostatics, shape, and desolvation. A binding model built on these features was initially developed for CAPRI Round 21, achieving top results in the independent assessment. Here, we present a more thoroughly trained and validated model, and compare various support‐vector machine kernels. The Gaussian kernel model classified both high‐resolution complexes and designed nonbinders with 79–86% accuracy on independent test data. We also observe that multiple physical potentials for dielectric‐dependent electrostatics and hydrogen bonding contribute to the enhanced predictive accuracy, suggesting that their combined information is much greater than that of any single energetics model. We also study the change in predictive performance as the model features or training data are varied, observing unusual patterns of prediction in designed interfaces as compared with other data types. Proteins 2013; 81:1919–1930. © 2013 Wiley Periodicals, Inc.     

Evaluating the absolute quality of a single protein model using structural features and support vector machines

Knowing the quality of a protein structure model is important for its appropriate usage. We developed a model evaluation method to assess the absolute quality of a single protein model using only structural features with support vector machine regression. The method assigns an absolute quantitative score (i.e. GDT‐TS) to a model by comparing its secondary structure, relative solvent accessibility, contact map, and beta sheet structure with their counterparts predicted from its primary sequence. We trained and tested the method on the CASP6 dataset using cross‐validation. The correlation between predicted and true scores is 0.82. On the independent CASP7 dataset, the correlation averaged over 95 protein targets is 0.76; the average correlation for template‐based and ab initio targets is 0.82 and 0.50, respectively. Furthermore, the predicted absolute quality scores can be used to rank models effectively. The average difference (or loss) between the scores of the top‐ranked models and the best models is 5.70 on the CASP7 targets. This method performs favorably when compared with the other methods used on the same dataset. Moreover, the predicted absolute quality scores are comparable across models for different proteins. These features make the method a valuable tool for model quality assurance and ranking. Proteins 2009. © 2008 Wiley‐Liss, Inc.