Prediction Of Stability Changes Research Articles

AUTO-MUTE: web-based tools for predicting stability changes in proteins due to single amino acid replacements

Utilizing cutting-edge supervised classification and regression algorithms, three web-based tools have been developed for predicting stability changes upon single residue substitutions in proteins with known native structures. Trained models classify independent mutant test sets with accuracies ranging from 87 to 94%. Attributes representing each mutant protein are based on a computational mutagenesis methodology relying on a four-body statistical potential, illustrating a novel integration of both energy-based and machine learning approaches. The servers are written in PHP and hosted on a Linux platform, and they can be freely accessed online along with detailed data sets, documentation and performance results at http://proteins.gmu.edu/automute.

A Fuzzy Query Method Based on Human-Readable Rules for Predicting Protein Stability Changes

Predicting protein mutant stability changes is important for protein design. Many methods have tried to improve prediction accuracy by various models. However, it will be difficult to employ them when the required input information is incomplete. This paper presents a fuzzy query method (named FQ-STAB), which cooperates with a human- readable rule base to predict stability changes upon single mutations. Firstly, we have effectively established a set of classification rules as a knowledge representation from the thermodynamic database. Next, we applied the proposed method on the rules to predict stability changes under the condition without sufficient information. Further, FQ-STAB has been tested on a data set of single point mutants. The results show that it can be applied to the prediction of stability changes using partial input information, and can also provide the explanation capabilities for the predicted outcome. We have developed a web server for predicting protein stability changes upon single mutations by using fuzzy query mechanism and it is available at http: //bioinformatics.myweb.hinet.net/fqstab.htm.

Fast and accurate predictions of protein stability changes upon mutations using statistical potentials and neural networks: PoPMuSiC-2.0

The rational design of proteins with modified properties, through amino acid substitutions, is of crucial importance in a large variety of applications. Given the huge number of possible substitutions, every protein engineering project would benefit strongly from the guidance of in silico methods able to predict rapidly, and with reasonable accuracy, the stability changes resulting from all possible mutations in a protein. We exploit newly developed statistical potentials, based on a formalism that highlights the coupling between four protein sequence and structure descriptors, and take into account the amino acid volume variation upon mutation. The stability change is expressed as a linear combination of these energy functions, whose proportionality coefficients vary with the solvent accessibility of the mutated residue and are identified with the help of a neural network. A correlation coefficient of R = 0.63 and a root mean square error of sigma(c) = 1.15 kcal/mol between measured and predicted stability changes are obtained upon cross-validation. These scores reach R = 0.79, and sigma(c) = 0.86 kcal/mol after exclusion of 10% outliers. The predictive power of our method is shown to be significantly higher than that of other programs described in the literature. http://babylone.ulb.ac.be/popmusic

Invivo folding efficiencies for mutants of the P22 tailspike β-helix protein correlate with predicted stability changes

Parallel β-helices are among the simplest repetitive structural elements in proteins. The folding behavior of β-helix proteins has been studied intensively, also to gain insight on the formation of amyloid fibrils, which share the parallel β-helix as a central structural motif. An important system for investigating β-helix folding is the tailspike protein from the Salmonella bacteriophage P22. The central domain of this protein is a right-handed parallel β-helix with 13 windings. Extensive mutational analyses of the P22 tailspike protein have revealed two main phenotypes: temperature-sensitive-folding (tsf) mutations that reduce the folding efficiency at elevated temperatures, and global suppressor (su) mutations that increase the tailspike folding efficiency. A central question is whether these phenotypes can be understood from changes in the protein stability induced by the mutations. Experimental determination of the protein stability is complicated by the nearly irreversible trimerization of the folded tailspike protein. Here, we present calculations of stability changes with the program FoldX, focusing on a recently published extensive data set of 145 singe-residue alanine mutants. We find that the calculated stability changes are correlated with the experimentally measured in vivo folding efficiencies. In addition, we determine the free-energy landscape of the P22 tailspike protein in a nucleation–propagation model to explore the folding mechanism of this protein, and obtain a processive folding route on which the protein nucleates in the N-terminal region of the helix.

Accurate prediction of stability changes in protein mutants by combining machine learning with structure based computational mutagenesis

Accurate predictive models for the impact of single amino acid substitutions on protein stability provide insight into protein structure and function. Such models are also valuable for the design and engineering of new proteins. Previously described methods have utilized properties of protein sequence or structure to predict the free energy change of mutants due to thermal (DeltaDeltaG) and denaturant (DeltaDeltaG(H2O)) denaturations, as well as mutant thermal stability (DeltaT(m)), through the application of either computational energy-based approaches or machine learning techniques. However, accuracy associated with applying these methods separately is frequently far from optimal. We detail a computational mutagenesis technique based on a four-body, knowledge-based, statistical contact potential. For any mutation due to a single amino acid replacement in a protein, the method provides an empirical normalized measure of the ensuing environmental perturbation occurring at every residue position. A feature vector is generated for the mutant by considering perturbations at the mutated position and it's ordered six nearest neighbors in the 3-dimensional (3D) protein structure. These predictors of stability change are evaluated by applying machine learning tools to large training sets of mutants derived from diverse proteins that have been experimentally studied and described. Predictive models based on our combined approach are either comparable to, or in many cases significantly outperform, previously published results. A web server with supporting documentation is available at http://proteins.gmu.edu/automute.

Protein Stability Prediction: A Poisson−Boltzmann Approach

Most proteins are only marginally stable at physiological temperatures. Thus a common defect due to mutation is the loss of protein stability, resulting in loss of their well-defined structures and functions at their functioning temperatures. Quantification of protein stability change upon mutation has attracted a large number of experimental and theoretical studies. In this work, we have extended the Poisson-Boltzmann theory that is originally used for predicting stability changes of charged mutations to predicting stability changes of all mutations. To achieve this, we have proposed a free energy model covering both electrostatic and hydrophobic interactions. A Gõ-like model for the denatured state that incorporates both nativeness and randomness of the denatured state has been used to calculate the hydrophobic contribution to protein stability. The new model is computationally simple and fast, and performs well for charged and hydrophobic mutations for all four tested proteins. Future directions for extending the method into pH-dependent effect and more accurate prediction for polar mutations are discussed.

Prediction of protein stability upon point mutations

Prediction of protein stability upon amino acid substitution is a challenging problem and it will be helpful for designing stable mutants. We have developed a thermodynamic database for proteins and mutants (ProTherm), which has more than 20000 thermodynamic data along with sequence and structure information, experimental conditions and literature information. It is freely accessible at http://gibk26.bse.kyutech.ac.jp/jouhou/protherm/protherm.html. Utilizing the database, we have analysed the relationship between amino acid properties and protein stability and developed different methods, such as average assignment method, distance and torsion potentials and decision tree models to discriminate the stabilizing and destabilizing mutants, and to predict the stability change upon mutation. Our method could distinguish the stabilizing and destabilizing mutants with an accuracy of 82 and 85% respectively from amino acid sequence and protein three-dimensional structure. We obtained the correlation of 0.70 and 0.87, between the experimental and predicted stability changes upon mutations, from sequence and structure respectively. Furthermore, we have developed different web servers for discrimination and prediction and they are freely accessible at http://bioinformatics.myweb.hinet.net/iptree.htm and http://cupsat.tu-bs.de/.

Enhancing the stability and solubility of TEV protease using in silico design

The ability to rationally increase the stability and solubility of recombinant proteins has long been a goal of biotechnology and has significant implications for biomedical research. Poorly soluble enzymes, for example, result in the need for larger reaction volumes, longer incubation times, and more restricted reaction conditions, all of which increase the cost and have a negative impact on the feasibility of the process. Rational design is achieved here by means of the PoPMuSiC program, which performs in silico predictions of stability changes upon single-site mutations. We have used this program to increase the stability of the tobacco etch virus (TEV) protein. TEV is a 27-kDa nuclear inclusion protease with stringent specificity that is commonly used for the removal of solubility tags during protein purification protocols. However, while recombinant TEV can be produced in large quantities, a limitation is its relatively poor solubility (generally approximately 1 mg/mL), which means that large volumes and often long incubation times are required for efficient cleavage. Following PoPMuSiC analysis of TEV, five variants predicted to be more stable than the wild type were selected for experimental analysis of their stability, solubility, and activity. Of these, two were found to enhance the solubility of TEV without compromising its functional activity. In addition, a fully active double mutant was found to remain soluble at concentrations in excess of 40 mg/mL. This modified TEV appears thus as an interesting candidate to be used in recombinant protein technology.

Computational modeling of protein mutant stability: analysis and optimization of statistical potentials and structural features reveal insights into prediction model development

BackgroundUnderstanding and predicting protein stability upon point mutations has wide-spread importance in molecular biology. Several prediction models have been developed in the past with various algorithms. Statistical potentials are one of the widely used algorithms for the prediction of changes in stability upon point mutations. Although the methods provide flexibility and the capability to develop an accurate and reliable prediction model, it can be achieved only by the right selection of the structural factors and optimization of their parameters for the statistical potentials. In this work, we have selected five atom classification systems and compared their efficiency for the development of amino acid atom potentials. Additionally, torsion angle potentials have been optimized to include the orientation of amino acids in such a way that altered backbone conformation in different secondary structural regions can be included for the prediction model. This study also elaborates the importance of classifying the mutations according to their solvent accessibility and secondary structure specificity. The prediction efficiency has been calculated individually for the mutations in different secondary structural regions and compared.ResultsResults show that, in addition to using an advanced atom description, stepwise regression and selection of atoms are necessary to avoid the redundancy in atom distribution and improve the reliability of the prediction model validation. Comparing to other atom classification models, Melo-Feytmans model shows better prediction efficiency by giving a high correlation of 0.85 between experimental and theoretical ΔΔG with 84.06% of the mutations correctly predicted out of 1538 mutations. The theoretical ΔΔG values for the mutations in partially buried β-strands generated by the structural training dataset from PISCES gave a correlation of 0.84 without performing the Gaussian apodization of the torsion angle distribution. After the Gaussian apodization, the correlation increased to 0.92 and prediction accuracy increased from 80% to 88.89% respectively.ConclusionThese findings were useful for the optimization of the Melo-Feytmans atom classification system and implementing them to develop the statistical potentials. It was also significant that the prediction efficiency of mutations in the partially buried β-strands improves with the help of Gaussian apodization of the torsion angle distribution. All these comparisons and optimization techniques demonstrate their advantages as well as the restrictions for the development of the prediction model. These findings will be quite helpful not only for the protein stability prediction, but also for various structure solutions in future.

Predicting protein stability changes from sequences using support vector machines

The prediction of protein stability change upon mutations is key to understanding protein folding and misfolding. At present, methods are available to predict stability changes only when the atomic structure of the protein is available. Methods addressing the same task starting from the protein sequence are, however, necessary in order to complete genome annotation, especially in relation to single nucleotide polymorphisms (SNPs) and related diseases. We develop a method based on support vector machines that, starting from the protein sequence, predicts the sign and the value of free energy stability change upon single point mutation. We show that the accuracy of our predictor is as high as 77% in the specific task of predicting the DeltaDeltaG sign related to the corresponding protein stability. When predicting the DeltaDeltaG values, a satisfactory correlation agreement with the experimental data is also found. As a final blind benchmark, the predictor is applied to proteins with a set of disease-related SNPs, for which thermodynamic data are also known. We found that our predictions corroborate the view that disease-related mutations correspond to a decrease in protein stability. http://gpcr2.biocomp.unibo.it/cgi/predictors/I-Mutant2.0/I-Mutant2.0.cgi

I-Mutant2.0: predicting stability changes upon mutation from the protein sequence or structure

I-Mutant2.0 is a support vector machine (SVM)-based tool for the automatic prediction of protein stability changes upon single point mutations. I-Mutant2.0 predictions are performed starting either from the protein structure or, more importantly, from the protein sequence. This latter task, to the best of our knowledge, is exploited for the first time. The method was trained and tested on a data set derived from ProTherm, which is presently the most comprehensive available database of thermodynamic experimental data of free energy changes of protein stability upon mutation under different conditions. I-Mutant2.0 can be used both as a classifier for predicting the sign of the protein stability change upon mutation and as a regression estimator for predicting the related ΔΔG values. Acting as a classifier, I-Mutant2.0 correctly predicts (with a cross-validation procedure) 80% or 77% of the data set, depending on the usage of structural or sequence information, respectively. When predicting ΔΔG values associated with mutations, the correlation of predicted with expected/experimental values is 0.71 (with a standard error of 1.30 kcal/mol) and 0.62 (with a standard error of 1.45 kcal/mol) when structural or sequence information are respectively adopted. Our web interface allows the selection of a predictive mode that depends on the availability of the protein structure and/or sequence. In this latter case, the web server requires only pasting of a protein sequence in a raw format. We therefore introduce I-Mutant2.0 as a unique and valuable helper for protein design, even when the protein structure is not yet known with atomic resolution. Availability: .

Can Contact Potentials Reliably Predict Stability of Proteins?

The simplest approximation of interaction potential between amino acid residues in proteins is the contact potential, which defines the effective free energy of a protein conformation by a set of amino acid contacts formed in this conformation. Finding a contact potential capable of predicting free energies of protein states across a variety of protein families will aid protein folding and engineering in silico on a computationally tractable time-scale. We test the ability of contact potentials to accurately and transferably (across various protein families) predict stability changes of proteins upon mutations. We develop a new methodology to determine the contact potentials in proteins from experimental measurements of changes in protein's thermodynamic stabilities (ΔΔ G) upon mutations. We apply our methodology to derive sets of contact interaction parameters for a hierarchy of interaction models including solvation and multi-body contact parameters. We test how well our models reproduce experimental measurements by statistical tests. We evaluate the maximum accuracy of predictions obtained by using contact potentials and the correlation between parameters derived from different data-sets of experimental (ΔΔ G) values. We argue that it is impossible to reach experimental accuracy and derive fully transferable contact parameters using the contact models of potentials. However, contact parameters may yield reliable predictions of ΔΔ G for datasets of mutations confined to the same amino acid positions in the sequence of a single protein.

Effective scoring function for protein sequence design.

We have developed an effective scoring function for protein design. The atomic solvation parameters, together with the weights of energy terms, were optimized so that residues corresponding to the native sequence were predicted with low energy in the training set of 28 protein structures. The solvation energy of non-hydrogen-bonded hydrophilic atoms was considered separately and expressed in a nonlinear way. As a result, our scoring function predicted native residues as the most favorable in 59% of the total positions in 28 proteins. We then tested the scoring function by comparing the predicted stability changes for 103 T4 lysozyme mutants with the experimental values. The correlation coefficients were 0.77 for surface mutations and 0.71 for all mutations. Finally, the scoring function combined with Monte Carlo simulation was used to predict favorable sequences on a fixed backbone. The designed sequences were similar to the natural sequences of the family to which the template structure belonged. The profile of the designed sequences was helpful for identification of remote homologues of the native sequence.

Prediction of stability changes upon single-site mutations using database-derived potentials

One of the purposes of studying protein stability changes upon mutations is to get information about the dominating interactions that drive folding and stabilise the native structure. With this in mind, we present a method that predicts folding free-energy variations caused by point mutations using combinations of two types of database-derived potentials, i.e. backbone torsion-angle potentials and distance potentials, describing local and non-local interactions along the chain, respectively. The method is applied to evaluate the folding free-energy changes of 344 single-site mutations introduced in six different proteins and a synthetic peptide. We found that the relative importance of local versus non-local interactions along the chain is essentially a function of the solvent accessibility of the mutated residues. For the subset of totally buried residues, the optimal potential is the sum of a distance potential and a torsion potential weighted by a factor of 0.4. This combination yields a correlation coefficient between measured and computed changes in folding free energy of 0.80. For mutations of partially buried residues, the best potential is the sum of a torsion potential and a distance potential weighted by 0.7. For fully accessible residues, the torsion potentials taken alone perform best, reaching correlation coefficients of 0.87 on all but 10 mutations; the excluded mutations seem to modify the backbone structure or to involve interactions that are atypical for the surface. These results show that the relative weight of non-local interactions along the sequence decreases as the solvent accessibility of the mutated residue increases, and vanishes at the protein surface. On the contrary, the weight of local interactions increases with solvent accessibility. The latter interactions are nevertheless never negligible, even for the most buried residues.

Prediction of the stability of protein mutants based on structural environment-dependent amino acid substitution and propensity tables.

An approach to the prediction of mutant stability is described using knowledge of amino acid replacements that are tolerated within the families of homologous proteins of known 3-D structure. Amino acid variations in families of homologous proteins are converted to propensity and substitution tables; these provide quantitative information about the existence of an amino acid in a structural environment and the probability of replacement by any other amino acid. The tables are used to calculate a 'stability difference score', analogous to the difference in free energy between a mutant and the wild type. The method has been developed and tested using the high-resolution structures for T4 lysozyme and 159 site-specific mutants. We show that differences in stability scores are correlated with experimentally observed free energy differences and differences in melting temperature. Blind tests, using only structural information derived from the parent wild-type crystal structures, on a combined set of 83 staphylococcal nuclease and 68 barnase mutants showed a correlation of 0.80 in the predicted stability changes with experimental thermodynamic data. Approximately 86% of the predictions were correctly classified as destabilizing or stabilizing.

Stability Changes upon Mutation of Solvent- accessible Residues in Proteins Evaluated by Database-derived Potentials

The stability changes in peptides and proteins caused by the substitution of a single amino acid, which can be measured experimentally by the change in folding free energy, are evaluated here using effective potentials derived from known protein structures. The analysis is focused on mutations of residues that are accessible to the solvent. These represent in total 106 mutations, introduced at different sites in barnase, bacteriophage T4 lysozyme and chymotrypsin inhibitor 2, and in a synthetic helical peptide. Assuming that the mutations do not modify the backbone structure, the changes in folding free energies are computed using various types of database-derived potentials and are compared with the measured ones. Distance-dependent residue – residue potentials are found to be inadequate for estimating the stability changes caused by these mutations, as they are dominated by hydrophobic interactions, which do not play an essential role at the protein surface. On the contrary, the potentials based on backbone torsion angle propensities yield quite good results. Indeed, for a subset of 96 out of the 106 mutations, the computed and measured changes in folding free energy correlate with a linear correlation coefficient of 0.87. Moreover, the ten mutations that are excluded from the correlation either seem to cause modifications of the backbone structure or to involve strong hydrophobic interactions, which are atypical for solvent-accessible residues. We find furthermore that raising the ionic strength of the solvent used for measuring the changes in folding free energies improves the correlation, as it tends to mask the electrostatic interactions. When adding to these 106 mutations 44 mutations performed in staphylococcal nuclease and chemotactic protein, which were first discarded because some of them were suspected to affect the backbone conformation or the denatured state, the correlation between measured and computed folding free energy changes remains quite good: the correlation coefficient is 0.86 for 135 out of the 150 mutations. The success of the backbone torsion potentials in predicting stability changes indicates that the approximations made for deriving these potentials are adequate. It suggests moreover that the local interactions along the chain dominate at the protein surface.