Prediction of protein quaternary structure with feature selection and analysis based on protein biological features

Information of protein quaternary structure can help to understand the biological functions of proteins. Because wet-lab experiments are both time-consuming and costly, we adopt a novel computational approach to assign proteins into 10 kinds of quaternary structures. By coding each protein using its biochemical and physicochemical properties, feature selection was carried out using Incremental Feature Selection (IFS) method. The thus obtained optimal feature set consisted of 97 features, with which the prediction model was built. As a result, the overall prediction success rate is 74.90% evaluated by Jackknife test, much higher than the overall correct rate of a random guess 10% (1/10). The further feature analysis indicates that protein secondary structure is the most contributed feature in the prediction of protein quaternary structure.     

Probing enzyme quaternary structure by combinatorial mutagenesis and selection

Genetic selection provides an effective way to obtain active catalysts from a diverse population of protein variants. We have used this tool to investigate the role of loop sequences in determining the quaternary structure of a domain-swapped enzyme. By inserting random loops of four to seven residues into a dimeric chorismate mutase and selecting for functional variants by genetic complementation, we have obtained and characterized both monomeric and hexameric enzymes that retain considerable catalytic activity. The low percentage of active proteins recovered from these selection experiments indicates that relatively few loop sequences permit a change in quaternary structure without affecting active site structure. The results of our experiments suggest further that protein stability can be an important driving force in the evolution of oligomeric proteins.     

Substrate-induced interconversion of protein quaternary structure isoforms

Human porphobilinogen synthase (PBGS) can exist in two dramatically different quaternary structure isoforms, which have been proposed to be in dynamic equilibrium. The quaternary structure isoforms of PBGS result from two alternative conformations of the monomer; one monomer structure assembles into a high activity octamer, whereas the other monomer structure assembles into a low activity hexamer. The kinetic behavior of these oligomers led to the hypothesis that turnover facilitates the interconversion of the oligomeric structures. The current work demonstrates that the interactions of ligands at the enzyme active site promote the structural interconversion between human PBGS quaternary structure isoforms, favoring formation of the octamer. This observation illustrates that the assembly and disassembly of oligomeric proteins can be facilitated by the protein motions that accompany enzymatic catalysis.     

Prediction of primary structure deamidation rates of asparaginyl and glutaminyl peptides through steric and catalytic effects.

The primary sequence dependence of deamidation has been quantitatively explained on the basis of a simple steric and catalytic model. Application to the known deamidation rates of peptides produces a table of coefficients that permits calculation of the known deamidation rates and prediction of deamidation rates for peptide sequences that have not yet been measured. This work permits a better understanding of deamidation, provides a prediction procedure for protein engineering, and facilitates improved computation of peptide and protein primary, secondary, tertiary, and quaternary structure deamidation rates.     

The quaternary structure of carbonmonoxy hemoglobin ypsilanti

We present a geometric analysis of the allosteric interface in the new Y state quaternary structure observed in liganded mutant hemoglobin Ypsilanti (β99 Asp → Tyr) by Smith, F.R., Lattman, E.E., Carter, C.W., Jr. (Proteins 10:81–91, 1991). The classical T to R quaternary structure change being a rotation of αβ dimers about an axis which is approximately parallel to the dimer axis of pseudosym-metry, the new quaternary structure is obtained by applying to R an additional rotation about an axis orthogonal to the first. This suggests that Y is a modified R state rather than an intermediate on the T to R pathway. Computer docking experiments designed to simulate the quaternary structure change support this suggestion. © 1993 Wiley-Liss, Inc.     

Tertiary structures for the extracellular domains of the epithelial polyimmunoglobulin receptor (secretory component) derived by primary structure comparisons with immunoglobulins

The amino acid sequences of the rabbit receptor and human secretory component (SC) domains have been compared with those of immunoglobin (Ig) domains. Accessible and inaccessible sites of tryptic cleavage in bovine SC have been located by sequence homology. Computerized secondary structure prediction and three dimensional model building have been carried out. The resulting tertiary structures are extremely Ig-like consisting of two superposed beta-pleated sheets. All carbohydrate sites lie at external positions as do tryptic cleavage sites. Potential sites for tryptic hydrolysis that are not cleaved lie at buried or partially buried positions. 6. Inter-beta-sheet contact between domains appears to be highly unlikely so that the quaternary structure is largely determined by longitudinal contacts.     

Comparative analyses of quaternary arrangements in homo‐oligomeric proteins in superfamilies: Functional implications

A comprehensive analysis of the quaternary features of distantly related homo‐oligomeric proteins is the focus of the current study. This study has been performed at the levels of quaternary state, symmetry, and quaternary structure. Quaternary state and quaternary structure refers to the number of subunits and spatial arrangements of subunits, respectively. Using a large dataset of available 3D structures of biologically relevant assemblies, we show that only 53% of the distantly related homo‐oligomeric proteins have the same quaternary state. Considering these homologous homo‐oligomers with the same quaternary state, conservation of quaternary structures is observed only in 38% of the pairs. In 36% of the pairs of distantly related homo‐oligomers with different quaternary states the larger assembly in a pair shows high structural similarity with the entire quaternary structure of the related protein with lower quaternary state and it is referred as “Russian doll effect.” The differences in quaternary state and structure have been suggested to contribute to the functional diversity. Detailed investigations show that even though the gross functions of many distantly related homo‐oligomers are the same, finer level differences in molecular functions are manifested by differences in quaternary states and structures. Comparison of structures of biological assemblies in distantly and closely related homo‐oligomeric proteins throughout the study differentiates the effects of sequence divergence on the quaternary structures and function. Knowledge inferred from this study can provide insights for improved protein structure classification and function prediction of homo‐oligomers. Proteins 2016; 84:1190–1202. © 2016 Wiley Periodicals, Inc.     

Determinants of quaternary association in legume lectins

It is well known that the sequence of amino acids in proteins code for its tertiary structure. It is also known that there exists a relationship between sequence and the quaternary structure of proteins. The question addressed here is whether the nature of quaternary association can be predicted from the sequence, similar to the three-dimensional structure prediction from the sequence. The class of proteins called legume lectins is an interesting model system to investigate this problem, because they have very high sequence and tertiary structure homology, with diverse forms of quaternary association. Hence, we have used legume lectins as a probe in this paper to (1) gain novel insights about the relationship between sequence and quaternary structure; (2) identify the sequence motifs that are characteristic of a given type of quaternary association; and (3) predict the quaternary association from the sequence motif.     

BIPSPI+: Mining Type-Specific Datasets of Protein Complexes to Improve Protein Binding Site Prediction

Computational approaches for predicting protein-protein interfaces are extremely useful for understanding and modelling the quaternary structure of protein assemblies. In particular, partner-specific binding site prediction methods allow delineating the specific residues that compose the interface of protein complexes. In recent years, new machine learning and other algorithmic approaches have been proposed to solve this problem. However, little effort has been made in finding better training datasets to improve the performance of these methods. With the aim of vindicating the importance of the training set compilation procedure, in this work we present BIPSPI+, a new version of our original server trained on carefully curated datasets that outperforms our original predictor. We show how prediction performance can be improved by selecting specific datasets that better describe particular types of protein interactions and interfaces (e.g. homo/hetero). In addition, our upgraded web server offers a new set of functionalities such as the sequence-structure prediction mode, hetero- or homo-complex specialization and the guided docking tool that allows to compute 3D quaternary structure poses using the predicted interfaces. BIPSPI+ is freely available at https://bipspi.cnb.csic.es.     

Local structure prediction with local structure-based sequence profiles

MOTIVATION: A large body of experimental and theoretical evidence suggests that local structural determinants are frequently encoded in short segments of protein sequence. Although the local structural information, once recognized, is particularly useful in protein structural and functional analyses, it remains a difficult problem to identify embedded local structural codes based solely on sequence information. RESULTS: In this paper, we describe a local structure prediction method aiming at predicting the backbone structures of nine-residue sequence segments. Two elements are the keys for this local structure prediction procedure. The first key element is the LSBSP1 database, which contains a large number of non-redundant local structure-based sequence profiles for nine-residue structure segments. The second key element is the consensus approach, which identifies a consensus structure from a set of hit structures. The local structure prediction procedure starts by matching a query sequence segment of nine consecutive amino acid residues to all the sequence profiles in the local structure-based sequence profile database (LSBSP1). The consensus structure, which is at the center of the largest structural cluster of the hit structures, is predicted to be the native state structure adopted by the query sequence segment. This local structure prediction method is assessed with a large set of random test protein structures that have not been used in constructing the LSBSP1 database. The benchmark results indicate that the prediction capacities of the novel local structure prediction procedure exceed the prediction capacities of the local backbone structure prediction methods based on the I-sites library by a significant margin. AVAILABILITY: All the computational and assessment procedures have been implemented in the integrated computational system PrISM.1 (Protein Informatics System for Modeling). The system and associated databases for LINUX systems can be downloaded from the website: http://www.columbia.edu/~ay1/.     

Evolution of allosteric models for hemoglobin

We compare various allosteric models that have been proposed to explain cooperative oxygen binding to hemoglobin, including the two-state allosteric model of Monod, Wyman, and Changeux (MWC), the Cooperon model of Brunori, the model of Szabo and Karplus (SK) based on the stereochemical mechanism of Perutz, the generalization of the SK model by Lee and Karplus (SKL), and the Tertiary Two-State (TTS) model of Henry, Bettati, Hofrichter and Eaton. The preponderance of experimental evidence favors the TTS model which postulates an equilibrium between high (r)- and low (t)-affinity tertiary conformations that are present in both the T and R quaternary structures. Cooperative oxygenation in this model arises from the shift of T to R, as in MWC, but with a significant population of both r and t conformations in the liganded T and in the unliganded R quaternary structures. The TTS model may be considered a combination of the SK and SKL models, and these models provide a framework for a structural interpretation of the TTS parameters. The most compelling evidence in favor of the TTS model is the nanosecond - millisecond carbon monoxide (CO) rebinding kinetics in photodissociation experiments on hemoglobin encapsulated in silica gels. The polymeric network of the gel prevents any tertiary or quaternary conformational changes on the sub-second time scale, thereby permitting the subunit conformations prior to CO photodissociation to be determined from their ligand rebinding kinetics. These experiments show that a large fraction of liganded subunits in the T quaternary structure have the same functional conformation as liganded subunits in the R quaternary structure, an experimental finding inconsistent with the MWC, Cooperon, SK, and SKL models, but readily explained by the TTS model as rebinding to r subunits in T. We propose an additional experiment to test another key prediction of the TTS model, namely that a fraction of subunits in the unliganded R quaternary structure has the same functional conformation (t) as unliganded subunits in the T quaternary structure.     

Pyrococcus abyssi alkaline phosphatase: the dimer is the active form

Alkaline phosphatases (APs), E.C. 3.1.3.1, are non-specific phosphomonoesterases optimally active under alkaline conditions. They are classically known to be homodimeric metalloenzymes. This quaternary structure has been considered necessary for activity, although the relationship between quaternary structure and activity is not well understood. Recombinant Pyrococcus abyssi AP was previously isolated and characterized, appearing to have two active quaternary structures on native polyacrylamide gel electrophoresis: a monomer and a homodimer. The purpose of the present work was to determine the actual quaternary structure of P. abyssi AP in solution, by isolating each of the two quaternary forms and establishing the parameters governing the assembly and dissociation of the dimer. pH appeared to be an important parameter: in acidic media, the monomer/dimer ratio shifted towards monomer. Buffer composition also affected the quaternary structure: at the same pH, in potassium phosphate buffer, the two quaternary structures were observed, whereas in tris(hydroxymethyl)aminomethane hydrochloride buffer, only the dimer was observed. Metals bound to the enzyme were found to be involved in the stability of the quaternary structure. Indeed, the P. abyssi AP obtained upon removal of the metals was monomeric. Reactivation of the latter was achieved with variable efficiency. From these experiments, no active monomer could be isolated, leading the conclusion that the active form of P. abyssi AP is the homodimer.     

Functional inferences from blind ab initio protein structure predictions

Ab initio protein structure prediction methods have improved dramatically in the past several years. Because these methods require only the sequence of the protein of interest, they are potentially applicable to the open reading frames in the many organisms whose sequences have been and will be determined. Ab initio methods cannot currently produce models of high enough resolution for use in rational drug design, but there is an exciting potential for using the methods for functional annotation of protein sequences on a genomic scale. Here we illustrate how functional insights can be obtained from low-resolution predicted structures using examples from blind ab initio structure predictions from the third and fourth critical assessment of structure prediction (CASP3, CASP4) experiments.     

Improved performance in protein secondary structure prediction by inhomogeneous score combination.

MOTIVATION: In many fields of pattern recognition, combination has proved efficient to increase the generalization performance of individual prediction methods. Numerous systems have been developed for protein secondary structure prediction, based on different principles. Finding better ensemble methods for this task may thus become crucial. Furthermore, efforts need to be made to help the biologist in the post-processing of the outputs. RESULTS: An ensemble method has been designed to post-process the outputs of discriminant models, in order to obtain an improvement in prediction accuracy while generating class posterior probability estimates. Experimental results establish that it can increase the recognition rate of protein secondary structure prediction methods that provide inhomogeneous scores, even though their individual prediction successes are largely different. This combination thus constitutes a help for the biologist, who can use it confidently on top of any set of prediction methods. Moreover, the resulting estimates can be used in various ways, for instance to determine which areas in the sequence are predicted with a given level of reliability. AVAILABILITY: The prediction is freely available over the Internet on the Network Protein Sequence Analysis (NPS@) WWW server at http://pbil.ibcp.fr/NPSA/npsa_server.ht ml. The source code of the combiner can be obtained on request for academic use.     

An information measure of the quality of protein secondary structure prediction.

We describe an information-theory-based measure of the quality of secondary structure prediction (RELINFO). RELINFO has a simple yet intuitive interpretation: it represents the factor by which secondary structure choice at a residue has been restricted by a prediction scheme. As an alternative interpretation of secondary structure prediction, RELINFO complements currently used methods by providing an information-based view as to why a prediction succeeds and fails. To demonstrate this score's capabilities, we applied RELINFO to an analysis of a large set of secondary structure predictions obtained from the first five rounds of the Critical Assessment of Structure Prediction (CASP) experiment. RELINFO is compared with two other common measures: percent correct (Q3) and secondary structure overlap (SOV). While the correlation between Q3 and RELINFO is approximately 0.85, RELINFO avoids certain disadvantages of Q3, including overestimating the quality of a prediction. The correlation between SOV and RELINFO is approximately 0.75. The valuable SOV measure unfortunately suffers from a saturation problem, and perhaps has unfairly given the general impression that secondary structure prediction has reached its limit since SOV hasn't improved much over the recent rounds of CASP. Although not a replacement for SOV, RELINFO has greater dispersion. Over the five rounds of CASP assessed here, RELINFO shows that predictions targets have been more difficult in successive CASP experiments, yet the predictions quality has continued to improve measurably over each round. In terms of information, the secondary structure prediction quality has almost doubled from CASP1 to CASP5. Therefore, as a different perspective of accuracy, RELINFO can help to improve prediction of protein secondary structure by providing a measure of difficulty as well as final quality of a prediction.     

CONTRAfold: RNA secondary structure prediction without physics-based models

MOTIVATION: For several decades, free energy minimization methods have been the dominant strategy for single sequence RNA secondary structure prediction. More recently, stochastic context-free grammars (SCFGs) have emerged as an alternative probabilistic methodology for modeling RNA structure. Unlike physics-based methods, which rely on thousands of experimentally-measured thermodynamic parameters, SCFGs use fully-automated statistical learning algorithms to derive model parameters. Despite this advantage, however, probabilistic methods have not replaced free energy minimization methods as the tool of choice for secondary structure prediction, as the accuracies of the best current SCFGs have yet to match those of the best physics-based models. RESULTS: In this paper, we present CONTRAfold, a novel secondary structure prediction method based on conditional log-linear models (CLLMs), a flexible class of probabilistic models which generalize upon SCFGs by using discriminative training and feature-rich scoring. In a series of cross-validation experiments, we show that grammar-based secondary structure prediction methods formulated as CLLMs consistently outperform their SCFG analogs. Furthermore, CONTRAfold, a CLLM incorporating most of the features found in typical thermodynamic models, achieves the highest single sequence prediction accuracies to date, outperforming currently available probabilistic and physics-based techniques. Our result thus closes the gap between probabilistic and thermodynamic models, demonstrating that statistical learning procedures provide an effective alternative to empirical measurement of thermodynamic parameters for RNA secondary structure prediction. AVAILABILITY: Source code for CONTRAfold is available at http://contra.stanford.edu/contrafold/.     

Rapid knot detection and application to protein structure prediction

MOTIVATION: Knots in polypeptide chains have been found in very few proteins, and consequently should be generally avoided in protein structure prediction methods. Most effective structure prediction methods do not model the protein folding process itself, but rather seek only to correctly obtain the final native state. Consequently, the mechanisms that prevent knots from occurring in native proteins are not relevant to the modeling process, and as a result, knots can occur with significantly higher frequency in protein models. Here we describe Knotfind, a simple algorithm for knot detection that is fast enough for structure prediction, where tens or hundreds of thousands of conformations may be sampled during the course of a prediction. We have used this algorithm to characterize knots in large populations of model structures generated for targets in CASP 5 and CASP 6 using the Rosetta homology-based modeling method. RESULTS: Analysis of CASP5 models suggested several possible avenues for introduction of knots into these models, and these insights were applied to structure prediction in CASP 6, resulting in a significant decrease in the proportion of knotted models generated. Additionally, using the knot detection algorithm on structures in the Protein Data Bank, a previously unreported deep trefoil knot was found in acetylornithine transcarbamylase. AVAILABILITY: The Knotfind algorithm is available in the Rosetta structure prediction program at http://www.rosettacommons.org.     

Quaternary structure of bovine α-crystallin: influence of temperature

The tertiary and quaternary structure of α-crystallin is still a matter of controversy. We have characterized the native α-crystallin quaternary structure by isolating it at the in vivo temperature and solvent conditions. It can be represented by a distribution of expanded particles with a weight average molar mass of 550 000 g/mol. On decreasing (to 4°C) or increasing (up to 50°C) the temperature, the size distribution increases to larger particles. Only at lower temperatures (4°C), a stable population of particles is obtained with weight average molar mass of 700 000 g/mol. In all conditions, α-crystallin behaves as a very expanded particle with a maximum hydrodynamic volume of 3.15 ml/g. The transitions in quaternary structure are rather slow: it takes several hours to evolve from a population of aggregates, characteristic for given solvent conditions, to another distribution in size and quaternary structure on changing the environment. The quaternary structure of α-crystallin is an uncharacteristic parameter of the particle: a broad distribution of values can be obtained on changing the environment. Any realistic model should include this property. Our studies favor an open loose structure, where peptides can be added or removed without drastic changes of secondary and tertiary structure of the peptides.