Theory of protein secondary structure and algorithm of its prediction

Genome-Wide Identification and Characterization of Tomato Fatty Acid β-Oxidase Family Genes KAT and MFP

Fatty acids and their derivatives play a variety of roles in living organisms. Fatty acids not only store energy but also comprise membrane lipids and act as signaling molecules. There are three main proteins involved in the fatty acid β-oxidation pathway in plant peroxisomes, including acyl-CoA oxidase (ACX), multifunctional protein (MFP), and 3-ketolipoyl-CoA thiolase (KAT). However, genome-scale analysis of KAT and MFP has not been systemically investigated in tomatoes. Here, we conducted a bioinformatics analysis of KAT and MFP genes in tomatoes. Their physicochemical properties, protein secondary structure, subcellular localization, gene structure, phylogeny, and collinearity were also analyzed. In addition, a conserved motif analysis, an evolutionary pressure selection analysis, a cis-acting element analysis, tissue expression profiling, and a qRT-PCR analysis were conducted within tomato KAT and MFP family members. There are five KAT and four MFP family members in tomatoes, which are randomly distributed on four chromosomes. By analyzing the conserved motifs of tomato KAT and MFP family members, we found that both KAT and MFP members are highly conserved. In addition, the results of the evolutionary pressure selection analysis indicate that the KAT and MFP family members have evolved mainly from purifying selection, which makes them more structurally stable. The results of the cis-acting element analysis show that SlKAT and SlMFP with respect may respond to light, hormones, and adversity stresses. The tissue expression analysis showed that KAT and MFP family members have important roles in regulating the development of floral organs as well as fruit ripening. The qRT-PCR analysis revealed that the expressions of SlKAT and SlMFP genes can be regulated by ABA, MeJA, darkness, NaCl, PEG, UV, cold, heat, and H2O2 treatments. These results provide a basis for the involvement of the SlKAT and SlMFP genes in tomato floral organ development and abiotic stress response, which lay a foundation for future functional study of SlKAT and SlMFP in tomatoes.

Protein folding problem: enigma, paradox, solution

The ability of protein chains to spontaneously form their three-dimensional structures is a long-standing mystery in molecular biology. The most conceptual aspect of this mystery is how the protein chain can find its native, "working" spatial structure (which, for not too big protein chains, corresponds to the global free energy minimum) in a biologically reasonable time, without exhaustive enumeration of all possible conformations, which would take billions of years. This is the so-called "Levinthal's paradox." In this review, we discuss the key ideas and discoveries leading to the current understanding of protein folding kinetics, including folding landscapes and funnels, free energy barriers at the folding/unfolding pathways, and the solution of Levinthal's paradox. A special role here is played by the "all-or-none" phase transition occurring at protein folding and unfolding and by the point of thermodynamic (and kinetic) equilibrium between the "native" and the "unfolded" phases of the protein chain (where the theory obtains the simplest form). The modern theory provides an understanding of key features of protein folding and, in good agreement with experiments, it (i) outlines the chain length-dependent range of protein folding times, (ii) predicts the observed maximal size of "foldable" proteins and domains. Besides, it predicts the maximal size of proteins and domains that fold under solely thermodynamic (rather than kinetic) control. Complementarily, a theoretical analysis of the number of possible protein folding patterns, performed at the level of formation and assembly of secondary structures, correctly outlines the upper limit of protein folding times.

Multifaceted analysis of training and testing convolutional neural networks for protein secondary structure prediction

Protein secondary structure prediction remains a vital topic with broad applications. Due to lack of a widely accepted standard in secondary structure predictor evaluation, a fair comparison of predictors is challenging. A detailed examination of factors that contribute to higher accuracy is also lacking. In this paper, we present: (1) new test sets, Test2018, Test2019, and Test2018-2019, consisting of proteins from structures released in 2018 and 2019 with less than 25% identity to any protein published before 2018; (2) a 4-layer convolutional neural network, SecNet, with an input window of ±14 amino acids which was trained on proteins ≤25% identical to proteins in Test2018 and the commonly used CB513 test set; (3) an additional test set that shares no homologous domains with the training set proteins, according to the Evolutionary Classification of Proteins (ECOD) database; (4) a detailed ablation study where we reverse one algorithmic choice at a time in SecNet and evaluate the effect on the prediction accuracy; (5) new 4- and 5-label prediction alphabets that may be more practical for tertiary structure prediction methods. The 3-label accuracy (helix, sheet, coil) of the leading predictors on both Test2018 and CB513 is 81-82%, while SecNet's accuracy is 84% for both sets. Accuracy on the non-homologous ECOD set is only 0.6 points (83.9%) lower than the results on the Test2018-2019 set (84.5%). The ablation study of features, neural network architecture, and training hyper-parameters suggests the best accuracy results are achieved with good choices for each of them while the neural network architecture is not as critical as long as it is not too simple. Protocols for generating and using unbiased test, validation, and training sets are provided. Our data sets, including input features and assigned labels, and SecNet software including third-party dependencies and databases, are downloadable from dunbrack.fccc.edu/ss and github.com/sh-maxim/ss.

Deeper Profiles and Cascaded Recurrent and Convolutional Neural Networks for state-of-the-art Protein Secondary Structure Prediction

Protein Secondary Structure prediction has been a central topic of research in Bioinformatics for decades. In spite of this, even the most sophisticated ab initio SS predictors are not able to reach the theoretical limit of three-state prediction accuracy (88–90%), while only a few predict more than the 3 traditional Helix, Strand and Coil classes. In this study we present tests on different models trained both on single sequence and evolutionary profile-based inputs and develop a new state-of-the-art system with Porter 5. Porter 5 is composed of ensembles of cascaded Bidirectional Recurrent Neural Networks and Convolutional Neural Networks, incorporates new input encoding techniques and is trained on a large set of protein structures. Porter 5 achieves 84% accuracy (81% SOV) when tested on 3 classes and 73% accuracy (70% SOV) on 8 classes on a large independent set. In our tests Porter 5 is 2% more accurate than its previous version and outperforms or matches the most recent predictors of secondary structure we tested. When Porter 5 is retrained on SCOPe based sets that eliminate homology between training/testing samples we obtain similar results. Porter is available as a web server and standalone program at http://distilldeep.ucd.ie/porter/ alongside all the datasets and alignments.

Fuzziness in Protein Interactions-A Historical Perspective

The proposal that coupled folding to binding is not an obligatory mechanism for intrinsically disordered (ID) proteins was put forward 10 years ago. The notion of fuzziness implies that conformational heterogeneity can be maintained upon interactions of ID proteins, which has a functional impact either on regulated assembly or activity of the corresponding complexes. Here I review how the concept has evolved in the past decade, via increasing experimental data providing insights into the mechanisms, pathways and regulatory modes. The effects of structural diversity and transient contacts on protein assemblies have been collected and systematically analyzed (Fuzzy Complexes Database, http://protdyn-database.org). Fuzziness has also been exploited as a framework to decipher molecular organization of higher-order protein structures. Quantification of conformational heterogeneity opens exciting future perspectives for drug discovery from small molecule-ID protein interactions to supramolecular assemblies.

Tom70 enhances mitochondrial preprotein import efficiency by binding to internal targeting sequences

N-terminal matrix-targeting signals (MTSs) are critical for mitochondrial protein import. Backes et al. identified additional internal MTS-like sequences scattered along the sequences of mitochondrial proteins. By binding to Tom70 on the mitochondrial surface, these sequences support the import process.

The biogenesis of mitochondria depends on the import of hundreds of preproteins. N-terminal matrix-targeting signals (MTSs) direct preproteins to the surface receptors Tom20, Tom22, and Tom70. In this study, we show that many preproteins contain additional internal MTS-like signals (iMTS-Ls) in their mature region that share the characteristic properties of presequences. These features allow the in silico prediction of iMTS-Ls. Using Atp1 as model substrate, we show that iMTS-Ls mediate the binding to Tom70 and have the potential to target the protein to mitochondria if they are presented at its N terminus. The import of preproteins with high iMTS-L content is significantly impaired in the absence of Tom70, whereas preproteins with low iMTS-L scores are less dependent on Tom70. We propose a stepping stone model according to which the Tom70-mediated interaction with internal binding sites improves the import competence of preproteins and increases the efficiency of their translocation into the mitochondrial matrix.

Substrates of the chloroplast small heat shock proteins 22E/F point to thermolability as a regulative switch for heat acclimation in Chlamydomonas reinhardtii

We have identified 39 proteins that interact directly or indirectly with high confidence with chloroplast HSP22E/F under heat stress thus revealing chloroplast processes affected by heat. Under conditions promoting protein unfolding, small heat shock proteins (sHsps) prevent the irreversible aggregation of unfolding proteins by integrating into forming aggregates. Aggregates containing sHsps facilitate the access of Hsp70 and ClpB/Hsp104 chaperones, which in ATP-dependent reactions disentangle individual proteins from the aggregates and assist in their refolding to the native state. Chlamydomonas reinhardtii encodes eight different sHsps (HSP22A to H). The goal of this work was to identify chloroplast-targeted sHsps in Chlamydomonas and to obtain a comprehensive list of the substrates with which they interact during heat stress in order to understand which chloroplast processes are disturbed under heat stress. We show that HSP22E and HSP22F are major chloroplast-targeted sHsps that have emerged from a recent gene duplication event resulting from the ongoing diversification of sHsps in the Volvocales. HSP22E/F strongly accumulate during heat stress and form high molecular mass complexes. Using differential immunoprecipitation, mass spectrometry and a stringent filtering algorithm we identified 39 proteins that with high-confidence interact directly or indirectly with HSP22E/F under heat stress. We propose that the apparent thermolability of several of these proteins might be a desired trait as part of a mechanism enabling Chlamydomonas chloroplasts to rapidly react to thermal stress.

Unfolding of CPR3 Gets Initiated at the Active Site and Proceeds via Two Intermediates

Cyclophilin catalyzes the ubiquitous process "peptidyl-prolyl cis-trans isomerization," which plays a key role in protein folding, regulation, and function. Here, we present a detailed characterization of the unfolding of yeast mitochondrial cyclophilin (CPR3) induced by urea. It is seen that CPR3 unfolding is reversible and proceeds via two intermediates, I1 and I2. The I1 state has native-like secondary structure and shows strong anilino-8-naphthalenesulphonate binding due to increased exposure of the solvent-accessible cluster of non-polar groups. Thus, it has some features of a molten globule. The I2 state is more unfolded, but it retains some residual secondary structure, and shows weak anilino-8-naphthalenesulphonate binding. Chemical shift perturbation analysis by ¹H-¹⁵N heteronuclear single quantum coherence spectra reveals disruption of the tertiary contacts among the regions close to the active site in the first step of unfolding, i.e., the N-I1 transition. Both of the intermediates, I1 and I2, showed a propensity to self-associate under stirring conditions, but their kinetic profiles are different; the native protein did not show any such tendency under the same conditions. All these observations could have significant implications for the function of the protein.

There and back again: Two views on the protein folding puzzle

The ability of protein chains to spontaneously form their spatial structures is a long-standing puzzle in molecular biology. Experimentally measured folding times of single-domain globular proteins range from microseconds to hours: the difference (10-11 orders of magnitude) is the same as that between the life span of a mosquito and the age of the universe. This review describes physical theories of rates of overcoming the free-energy barrier separating the natively folded (N) and unfolded (U) states of protein chains in both directions: "U-to-N" and "N-to-U". In the theory of protein folding rates a special role is played by the point of thermodynamic (and kinetic) equilibrium between the native and unfolded state of the chain; here, the theory obtains the simplest form. Paradoxically, a theoretical estimate of the folding time is easier to get from consideration of protein unfolding (the "N-to-U" transition) rather than folding, because it is easier to outline a good unfolding pathway of any structure than a good folding pathway that leads to the stable fold, which is yet unknown to the folding protein chain. And since the rates of direct and reverse reactions are equal at the equilibrium point (as follows from the physical "detailed balance" principle), the estimated folding time can be derived from the estimated unfolding time. Theoretical analysis of the "N-to-U" transition outlines the range of protein folding rates in a good agreement with experiment. Theoretical analysis of folding (the "U-to-N" transition), performed at the level of formation and assembly of protein secondary structures, outlines the upper limit of protein folding times (i.e., of the time of search for the most stable fold). Both theories come to essentially the same results; this is not a surprise, because they describe overcoming one and the same free-energy barrier, although the way to the top of this barrier from the side of the unfolded state is very different from the way from the side of the native state; and both theories agree with experiment. In addition, they predict the maximal size of protein domains that fold under solely thermodynamic (rather than kinetic) control and explain the observed maximal size of the "foldable" protein domains.

PREDICTION OF LOOP REGIONS IN PROTEIN SEQUENCE

We suggest an algorithm that inputs a protein sequence and outputs a decomposition of the protein chain into a regular part including secondary structures and a nonregular part corresponding to loop regions. We have analyzed loop regions in a protein dataset of 3,769 globular domains and defined the optimal parameters for this prediction: the threshold between regular and nonregular regions and the optimal window size for averaging procedures using the scale of the expected number of contacts in a globular state and entropy scale as the number of degrees of freedom for the angles φ, ψ, and χ for each amino acid. Comparison with known methods demonstrates that our method gives the same results as the well-known ALB method based on physical properties of amino acids (the percentage of true predictions is 64% against 66%), and worse prediction for regular and nonregular regions than PSIPRED (Protein Structure Prediction Server) without alignment of homologous proteins (the percentage of true predictions is 73%). The potential advantage of the suggested approach is that the predicted set of loops can be used to find patterns of rigid and flexible loops as possible candidates to play a structure/function role as well as a role of antigenic determinants.

Accurate prediction of protein folding rates from sequence and sequence-derived residue flexibility and solvent accessibility

Protein folding rates vary by several orders of magnitude and they depend on the topology of the fold and the size and composition of the sequence. Although recent works show that the rates can be predicted from the sequence, allowing for high-throughput annotations, they consider only the sequence and its predicted secondary structure. We propose a novel sequence-based predictor, PFR-AF, which utilizes solvent accessibility and residue flexibility predicted from the sequence, to improve predictions and provide insights into the folding process. The predictor includes three linear regressions for proteins with two-state, multistate, and unknown (mixed-state) folding kinetics. PFR-AF on average outperforms current methods when tested on three datasets. The proposed approach provides high-quality predictions in the absence of similarity between the predicted and the training sequences. The PFR-AF's predictions are characterized by high (between 0.71 and 0.95, depending on the dataset) correlation and the lowest (between 0.75 and 0.9) mean absolute errors with respect to the experimental rates, as measured using out-of-sample tests. Our models reveal that for the two-state chains inclusion of solvent-exposed Ala may accelerate the folding, while increased content of Ile may reduce the folding speed. We also demonstrate that increased flexibility of coils facilitates faster folding and that proteins with larger content of solvent-exposed strands may fold at a slower pace. The increased flexibility of the solvent-exposed residues is shown to elongate folding, which also holds, with a lower correlation, for buried residues. Two case studies are included to support our findings.

A new artificial antigen of the hepatitis E virus

An artificial antigen composed of 12 small antigenic regions derived from the ORF2 and ORF3 HEV proteins was designed. The gene encoding for this artificial antigen was assembled from synthetic oligonucleotides by a new method called Restriction Enzyme-Assisted Ligation (REAL). The diagnostic relevance of this second generation HEV mosaic protein (HEV MA-II) was demonstrated by testing this antigen against a panel of 142 well defined anti-HEV positive and anti-HEV negative serum samples. The data obtained in this study support the substantial diagnostic potential of this HEV mosaic antigen.

Intrinsically disordered human C/EBP homologous protein regulates biological activity of colon cancer cells during calcium stress

Intrinsically disordered proteins are emerging as substantial functional constituents of mammalian proteomes. Although the abundance of these proteins has been established by bioinformatics approaches, the vast majority have not been characterized structurally or functionally. The C/EBP homologous protein (CHOP) is a proto-oncogene, traditionally shown as a dominant-negative inhibitor of C/EBPs and a transcriptional activator of activating protein-1. We report here the in vitro characterization of CHOP, where our computational analyses and experimental evidences show for the first time that CHOP is an intrinsically disordered protein. Intrinsic fluorescence, NMR spectroscopy, and analytical size-exclusion chromatography studies indicate that CHOP contains extensive disordered regions and self-associate in solution. Interestingly, the disordered N-terminal region has a key role in the oligomerization of CHOP and is vital for its biological activity. We report a novel mechanistic role of CHOP in the inhibition of Wnt/TCF signaling and stimulation of c-Jun and sucrase-isomaltase reporter activity in intestinal colon cancer cells. These findings are discussed in the context of oligomerization of intrinsically disordered proteins as one of the mechanisms through which they exert their biological function.

Computational studies reveal phosphorylation-dependent changes in the unstructured R domain of CFTR

The cystic fibrosis transmembrane conductance regulator (CFTR) is a cAMP-dependent chloride channel that is mutated in cystic fibrosis, an inherited disease of high morbidity and mortality. The phosphorylation of its approximately 200 amino acid R domain by protein kinase A is obligatory for channel gating under normal conditions. The R domain contains more than ten PKA phosphorylation sites. No individual site is essential but phosphorylation of increasing numbers of sites enables progressively greater channel activity. In spite of numerous studies of the role of the R domain in CFTR regulation, its mechanism of action remains largely unknown. This is because neither its structure nor its interactions with other parts of CFTR have been completely elucidated. Studies have shown that the R domain lacks well-defined secondary structural elements and is an intrinsically disordered region of the channel protein. Here, we have analyzed the disorder pattern and employed computational methods to explore low-energy conformations of the R domain. The specific disorder and secondary structure patterns detected suggest the presence of molecular recognition elements (MoREs) that may mediate phosphorylation-regulated intra- and inter-domain interactions. Simulations were performed to generate an ensemble of accessible R domain conformations. Although the calculated structures may represent more compact conformers than occur in vivo, their secondary structure propensities are consistent with predictions and published experimental data. Equilibrium simulations of a mimic of a phosphorylated R domain showed that it exhibited an increased radius of gyration. In one possible interpretation of these findings, by changing its size, the globally unstructured R domain may act as an entropic spring to perturb the packing of membrane-spanning sequences that constitute the ion permeability pathway and thereby activate channel gating.

The relationship between n-gram patterns and protein secondary structure

An n-gram pattern (NP{n,m}) in a protein sequence is a set of n residues and m wildcards in a window of size n+m. Each window of n+m amino acids is associated with a collection of NP{n,m} patterns based on the combinatorics of n+m objects taken m at a time. NP{n,m} patterns that are shared between sequences reflect evolutionary relationships. Recently the authors developed an alignment-independent protein classification algorithm based on shared NP{4,2} patterns that compared favorably to PSI-BLAST. Theoretically, NP{4,2} patterns should also reflect secondary structure propensity since they contain all possible n-grams for 1 < or = n < or = 4 and a window of 6 residues is wide enough to capture periodicities in the 2 < or = n < or = 5 range. This sparked interest in differentiating the information content in NP{4,2} patterns related to evolution from the content related to local propensity. The probability of alpha-, beta-, and coil components was determined for every NP{4,2} pattern over all the chains in the Protein Data Bank (PDB). An algorithm exclusively based on the Z-values of these distributions was developed, which accurately predicted 71-76% of alpha-helical segments and 62-67% of beta-sheets in rigorous jackknife tests. This provided evidence for the strong correlation between NP{4,2} patterns and secondary structure. By grouping PDB chains into subsets with increasing levels of sequence identity, it was also possible to separate the evolutionary and local propensity contributions to the classification process. The results showed that information derived from evolutionary relationships was more important for beta-sheet prediction than alpha-helix prediction.

Application of expert networks for predicting proteins secondary structure

The present study utilizes expert neural networks for the prediction of proteins secondary structure. We use three independent networks, one for each structure (alpha, beta and coil) as the first-level processing unit; decision upon the chosen structure for each residue is carried out by a second-level, post-processing unit, which utilizes the Chou and Fasman frequency values Falpha and Fbeta in order to strengthen and/or deplete the probability of the specific structure under investigation. The highest prediction case was 76%. Our method requires primitive computational means and a relatively small training set, while still been comparable to previous work. It is not meant to be an alternative to the determination of secondary structure by means of free energy minimization, integration of dynamic equations of motion or crystallography, which are expensive, time-consuming and complicated, but to provide additional constrains, which might be considered and incorporated into larger computing setups in order to reduce the initial search space for the above methods.

Prediction of folding transition-state position (βT) of small, two-state proteins from local secondary structure content

Folding kinetics of proteins is governed by the free energy and position of transition states. But attempts to predict the position of folding transition state on reaction pathway from protein structure have been met with only limited success, unlike the folding-rate prediction. Here, we find that the folding transition-state position is related to the secondary structure content of native two-state proteins. We present a simple method for predicting the transition-state position from their alpha-helix, turn and polyproline secondary structures. The method achieves 81% correlation with experiment over 24 small, two-state proteins, suggesting that the local secondary structure content, especially for content of alpha-helix, is a determinant of the solvent accessibility of the transition state ensemble and size of folding nucleus.

An α-Helical Burst in the src SH3 Folding Pathway

Src SH3 is a small all-beta-sheet protein composed of a single domain. We studied the folding behavior of src SH3 at various conditions by circular dichroism (CD), fluorescence, and X-ray solution scattering methods. On the src SH3 folding pathway, an alpha-helix-rich intermediate appeared not only at subzero temperatures but also above 0 degrees C. The fraction of alpha-helix in the kinetically observed intermediate is ca. 26% based on the kinetic CD experiment. X-ray solution scattering revealed that the intermediate was compact but not fully packed. The analysis of CD implies that the amplitude of the burst phase is proportional to the helical fraction calculated according to the helix-coil transition theory. This strongly suggests that the initial folding core is formed by the collapse of much less stably existing alpha-helices.

Support vector machines for prediction of dihedral angle regions

MOTIVATION

Most secondary structure prediction programs target only alpha helix and beta sheet structures and summarize all other structures in the random coil pseudo class. However, such an assignment often ignores existing local ordering in so-called random coil regions. Signatures for such ordering are distinct dihedral angle pattern. For this reason, we propose as an alternative approach to predict directly dihedral regions for each residue as this leads to a higher amount of structural information.

RESULTS

We propose a multi-step support vector machine (SVM) procedure, dihedral prediction (DHPRED), to predict the dihedral angle state of residues from sequence. Trained on 20,000 residues our approach leads to dihedral region predictions, that in regions without alpha helices or beta sheets is higher than those from secondary structure prediction programs.

AVAILABILITY

DHPRED has been implemented as a web service, which academic researchers can access from our webpage http://www.fz-juelich.de/nic/cbb

Direct correlation between proteins' folding rates and their amino acid compositions: An ab initio folding rate prediction

Discovering the mechanism of protein folding, in molecular biology, is a great challenge. A key step to this end is to find factors that correlate with protein folding rates. Over the past few years, many empirical parameters, such as contact order, long-range order, total contact distance, secondary structure contents, have been developed to reflect the correlation between folding rates and protein tertiary or secondary structures. However, the correlation between proteins' folding rates and their amino acid compositions has not been explored. In the present work, we examined systematically the correlation between proteins' folding rates and their amino acid compositions for two-state and multistate folders and found that different amino acids contributed differently to the folding progress. The relation between the amino acids' molecular weight and degeneracy and the folding rates was examined, and the role of hydrophobicity in the protein folding process was also inspected. As a consequence, a new indicator called composition index was derived, which takes no structure factors into account and is merely determined by the amino acid composition of a protein. Such an indicator is found to be highly correlated with the protein's folding rate (r > 0.7). From the results of this work, three points of concluding remarks are evident. (1) Two-state folders and multistate folders have different rate-determining amino acids. (2) The main determining information of a protein's folding rate is largely reflected in its amino acid composition. (3) Composition index may be the best predictor for an ab initio protein folding rate prediction directly from protein sequence from the standpoint of practical application.

Intrinsic disorder is a key characteristic in partners that bind 14-3-3 proteins

Proteins named 14-3-3 can bind more than 200 different proteins, mostly (but not exclusively) when they are at a phosphorylated state. These partner proteins are involved in different cellular processes, such as cell signaling, transcription factors, cellular morphology, and metabolism; this suggests pleiotropic functionality for 14-3-3 proteins. Recent efforts to establish a rational classification of 14-3-3 binding partners showed neither structural nor functional relatedness in this group of proteins. Using three natural predictors of disorder in proteins, and the structural available information, we show that >90% of 14-3-3 protein partners contain disordered regions. This percentage is significantly high when compared with recent studies on cell signaling and cancer-related proteins or RNA chaperons. More important, almost all 14-3-3-binding sites are inside disordered regions, this reinforcing the importance of structural disorder in this class of proteins. We also propose that a disorder-to-order transition occurs in the binding of 14-3-3 proteins with their partners. We discuss the consequences of the latter for consensus binding sequences, specificity, affinity, and thermodynamic control.

Amino acid sequence predicts folding rate for middle-size two-state proteins

The significant correlation between protein folding rates and the sequence-predicted secondary structure suggests that folding rates are largely determined by the amino acid sequence. Here, we present a method for predicting the folding rates of proteins from sequences using the intrinsic properties of amino acids, which does not require any information on secondary structure prediction and structural topology. The contribution of residue to the folding rate is expressed by the residue's Omega value. For a given residue, its Omega depends on the amino acid properties (amino acid rigidity and dislike of amino acid for secondary structures). Our investigation achieves 82% correlation with folding rates determined experimentally for simple, two-state proteins studied until the present, suggesting that the amino acid sequence of a protein is an important determinant of the protein-folding rate and mechanism.

A hybrid genetic-neural system for predicting protein secondary structure

BACKGROUND

Due to the strict relation between protein function and structure, the prediction of protein 3D-structure has become one of the most important tasks in bioinformatics and proteomics. In fact, notwithstanding the increase of experimental data on protein structures available in public databases, the gap between known sequences and known tertiary structures is constantly increasing. The need for automatic methods has brought the development of several prediction and modelling tools, but a general methodology able to solve the problem has not yet been devised, and most methodologies concentrate on the simplified task of predicting secondary structure.

RESULTS

In this paper we concentrate on the problem of predicting secondary structures by adopting a technology based on multiple experts. The system performs an overall processing based on two main steps: first, a "sequence-to-structure" prediction is enforced by resorting to a population of hybrid (genetic-neural) experts, and then a "structure-to-structure" prediction is performed by resorting to an artificial neural network. Experiments, performed on sequences taken from well-known protein databases, allowed to reach an accuracy of about 76%, which is comparable to those obtained by state-of-the-art predictors.

CONCLUSION

The adoption of a hybrid technique, which encompasses genetic and neural technologies, has demonstrated to be a promising approach in the task of protein secondary structure prediction.

Protein Folding Rates Estimated from Contact Predictions

Folding rates of small single-domain proteins that fold through simple two-state kinetics can be estimated from details of the three-dimensional protein structure. Previously, predictions of secondary structure had been exploited to predict folding rates from sequence. Here, we estimate two-state folding rates from predictions of internal residue-residue contacts in proteins of unknown structure. Our estimate is based on the correlation between the folding rate and the number of predicted long-range contacts normalized by the square of the protein length. It is well known that long-range order derived from known structures correlates with folding rates. The surprise was that estimates based on very noisy contact predictions were almost as accurate as the estimates based on known contacts. On average, our estimates were similar to those previously published from secondary structure predictions. The combination of these methods that exploit different sources of information improved performance. It appeared that the combined method reliably distinguished fast from slow two-state folders.

Three Dimensional Structure of Mammalian Tachykinin Peptide Neurokinin B Bound to Lipid Micelles

Neurokinin B (NKB), a decapeptide of mammalian origin exhibits a variety of biological activities such as regulatory functions in reproduction, pre-eclampsia and neuroprotection in Alzheimer's disease. In order to gain insight into structure-function relationship, three-dimensional structure of NKB has been investigated using CD spectropolarimetry and two-dimensional proton nuclear magnetic resonance (2D 1H-NMR) spectroscopy in aqueous and membrane mimetic solvents. Unambiguous NMR assignments of resonances have been made with the aid of correlation spectroscopy (DQF-COSY and TOCSY) experiments and Nuclear Overhauser Effect Spectroscopy (NOESY) experiments. Distance constraints obtained from the NMR data have been used to generate a family of structures, which have been refined using restrained energy minimization and dynamics. Our data show that a helical structure is induced in NKB, in presence of perdeuterated dodecyl phosphocholine (DPC) micelles, a membrane model system. Further, the conformation adopted by NKB in presence of DPC micelles represents a structural motif typical of neurokinin-3 selective agonists.

Preformed structural elements feature in partner recognition by intrinsically unstructured proteins

Intrinsically unstructured proteins (IUPs) are devoid of extensive structural order but often display signs of local and limited residual structure. To explain their effective functioning, we reasoned that such residual structure can be crucial in their interactions with their structured partner(s) in a way that preformed structural elements presage their final conformational state. To check this assumption, a database of 24 IUPs with known 3D structures in the bound state has been assembled and the distribution of secondary structure elements and backbone torsion angles have been analysed. The high proportion of residues in coil conformation and with phi, psi angles in the disallowed regions of the Ramachandran map compared to the reference set of globular proteins shows that IUPs are not fully ordered even in their bound form. To probe the effect of partner proteins on IUP folding, inherent conformational preferences of IUP sequences have been assessed by secondary structure predictions using the GOR, ALB and PROF algorithms. The accuracy of predicting secondary structure elements of IUPs is similar to that of their partner proteins and is significantly higher than the corresponding values for random sequences. We propose that strong conformational preferences mark regions in IUPs (mostly helices), which correspond to their final structural state, while regions with weak conformational preferences represent flexible linkers between them. In our interpretation, preformed elements could serve as initial contact points, the binding of which facilitates the reeling of the flexible regions onto the template. This finding implies that IUPs draw a functional advantage from preformed structural elements, as they enable their facile, kinetically and energetically less demanding, interaction with their physiological partner.

Prediction of protein folding rates from the amino acid sequence-predicted secondary structure

We present a method for predicting folding rates of proteins from their amino acid sequences only, or rather, from their chain lengths and their helicity predicted from their sequences. The method achieves 82% correlation with experiment over all 64 "two-state" and "multistate" proteins (including two artificial peptides) studied up to now.

Class-specific correlations between protein folding rate, structure-derived, and sequence-derived descriptors

Small single-domain proteins that fold by simple two-state kinetics have been shown to exhibit a wide variation in their folding rates. It has been proposed that folding mechanisms in these proteins are largely determined by the native-state topology, and a significant correlation between folding rate and measures of the average topological complexity, such as relative contact order (RCO), has been reported. We perform a statistical analysis of folding rate and RCO in all three major structural classes (alpha, beta, and alpha/beta) of small two-state proteins and of RCO in groups of analogous and homologous small single-domain proteins with the same topology. We also study correlation between folding rate and the average physicochemical properties of amino acid sequences in two-state proteins. Our results indicate that 1) helical proteins have statistically distinguishable, class-specific folding rates; 2) RCO accounts for essentially all the variation of folding rate in helical proteins, but for only a part of the variation in beta-sheet-containing proteins; and 3) only a small fraction of the protein topologies studied show a topology-specific RCO. We also report a highly significant correlation between the folding rate and average intrinsic structural propensities of protein sequences. These results suggest that intrinsic structural propensities may be an important determinant of the rate of folding in small two-state proteins.

Noncovalent Cross-links in Context with Other Structural and Functional Elements of Proteins

Proteins are heteropolymers with evolutionary selected native sequences of residues. These native sequences code for unique and stable 3D structures indispensable for biochemical activity and for proteolysis resistance, the latter which guarantees an appropriate lifetime for the protein in the protease rich cellular environment. Cross-links between residues close in space but far in the primary structure are required to maintain the folded structure of proteins. Some of these cross-links are covalent, most frequently disulfide bonds, but the majority of the cross-links are sets of cooperative noncovalent long-range interactions. In this paper we focus on special clusters of noncovalent long-range interactions: the Stabilization Centers (SCs). The relation between the SCs and secondary structural elements as well as the relation between SCs and functionally important regions of proteins are presented to show a detailed picture of these clusters, which are believed to be primarily responsible for major aspects of protein stability.

Characterization and prediction of linker sequences of multi-domain proteins by a neural network

In this paper, we describe a neural network analysis of sequences connecting two protein domains (domain linkers). The neural network was trained to distinguish between domain linker sequences and non-linker sequences, using a SCOP-defined domain library. The analysis indicated that a significant difference existed between domain linkers and non-linker regions, including intra-domain loop regions. Moreover, the resulting Hinton diagram showed a position-dependent amino acid preference of the domain linker sequences, and implied their non-random nature. We then applied the neural network to predict domain linkers in multi-domain protein sequences. As the result of a Jack-knife test, 58% of the predicted regions matched actual linker regions (specificity), and 36% of the SCOP-derived domain linkers were predicted (sensitivity). This prediction efficiency is superior to simpler methods derived from secondary structure prediction that assume that long loop regions are putative domain linkers. Altogether, these results suggest that domain linkers possess local characteristics different from those of loop regions.

Comparative studies on the properties of the extrinsic manganese-stabilizing protein from higher plants and of a synthetic peptide of its C-terminus

The present study describes a comparative analysis on the fluorescence properties of the manganese-stabilizing protein (MSP), a synthetic peptide corresponding to its C terminus and a 7:1 (molar ratio) mixture of N-acetyl-tyrosine and N-acetyl-tryptophan, respectively, together with reconstitution experiments of oxygen evolution in MSP-depleted photosystem II (PS II) membrane fragments. It is found: (i) at neutral pH, the fluorescence from Trp(241) is strongly diminished in MSP solutions, whereas it highly dominates the overall emission from the C-terminus peptide; (ii) at alkaline pH, the emission of Tyr and Trp is quenched in both, MSP and C-terminus peptide, with increasing pH but the decline curve is shifted by about two pH units towards the alkaline region in MSP; (iii) a drastically different pattern emerges in the 7:1 mixture where the Trp emission even slightly increases at high pH; (iv) the anisotropy of the fluorescence emission is wavelength-independent (310-395 nm) and indicative of one emitter type (Trp) in the C-terminus peptide and of two emitter types (Tyr, Trp) in MSP; and (v) in MSP-depleted PS II membrane fragments the oxygen evolution is restored (up to 85% of untreated control) by rebinding of MSP but not by the C-terminus peptide, however, the presence of the latter diminishes the restoration effect of MSP. A quenching mechanism of Trp fluorescence by a next neighbored tyrosinate in the peptide chain is proposed and the relevance of the C terminus of MSP briefly discussed.

Lysenin, a unique sphingomyelin-binding protein

Sphingomyelin plays complex structural and signaling functions in the plasma membrane. Of special interest is that hydrolysis of sphingomyelin to ceramide can modulate dynamics of membrane rafts, which serve as signaling platforms for various receptors. This review is focused on a recently discovered sphingomyelin-binding protein, lysenin, which can be used as a unique probe to trace distribution and turnover of sphingomyelin in cellular membranes. We analyze the primary and secondary structures of lysenin with respect to its interaction with the plasma membrane. The specificity of lysenin binding to sphingomyelin, revealed by both biochemical and cytochemical approaches, is discussed.

A pentapeptide-based method for protein secondary structure prediction

We present a new method for protein secondary structure prediction, based on the recognition of well-defined pentapeptides, in a large databank. Using a databank of 635 protein chains, we obtained a success rate of 68.6%. We show that progress is achieved when the databank is enlarged, when the 20 amino acids are adequately grouped in 10 sets and when more pentapeptides are attributed one of the defined conformations, alpha-helices or beta-strands. The analysis of the model indicates that the essential variable is the number of pentapeptides of well-defined structure in the database. Our model is simple, does not rely on arbitrary parameters and allows the analysis in detail of the results of each chosen hypothesis.

EVA: large-scale analysis of secondary structure prediction

EVA is a web-based server that evaluates automatic structure prediction servers continuously and objectively. Since June 2000, EVA collected more than 20,000 secondary structure predictions. The EVA sets sufficed to conclude that the field of secondary structure prediction has advanced again. Accuracy increased substantially in the 1990s through using evolutionary information taken from the divergence of proteins in the same structural family. Recently, the evolutionary information resulting from improved searches and larger databases has again boosted prediction accuracy by more than 4% to its current height around 76% of all residues predicted correctly in one of the three states: helix, strand, or other. The best current methods solved most of the problems raised at earlier CASP meetings: All good methods now get segments right and perform well on strands. Is the recent increase in accuracy significant enough to make predictions even more useful? We believe the answer is affirmative. What is the limit of prediction accuracy? We shall see. All data are available through the EVA web site at [cubic.bioc.columbia.edu/eva/]. The raw data for the results presented are available at [eva]/sec/bup_common/2001_02_22/.

Use of molecular mechanics for secondary structure prediction. Is it possible to reveal alpha-helix?

A new approach to predicting protein standard conformations is suggested. The idea consists in modeling by molecular mechanics tools a continuous alpha-helical conformation for the whole protein. The profile of energy along the model alpha-helix reveals minima corresponding to real alpha-helical segments in the native protein. The 3/10-helices and beta-turns including a local alpha-helical conformation may be detected as well. All alpha-helical segments in the test sample are delineated; mean residue by residue accuracy Q(3alpha) is 79%. This non-statistical approach can shed light on the physical grounds of alpha-helix formation.

Mapping the functional domains of elongation factor-2 kinase

A new class of eukaryotic protein kinases that are not homologous to members of the serine/threonine/tyrosine protein kinase superfamily was recently identified [Futey, L. M., et al. (1995) J. Biol. Chem. 270, 523-529; Ryazanov, A. G., et al. (1997) Proc. Natl. Acad. Sci. U.S.A. 94, 4884-4889]. This class includes eukaryotic elongation factor-2 kinase, Dictyostelium myosin heavy chain kinases A, B, and C, and several mammalian putative protein kinases that are not yet fully characterized [Ryazanov, A. G., et al. (1999) Curr. Biol. 9, R43-R45]. eEF-2 kinase is a ubiquitous protein kinase that phosphorylates and inactivates eukaryotic translational elongation factor-2, and thus can modulate the rate of polypeptide chain elongation during translation. eEF-2 was the only known substrate for eEF-2 kinase. We demonstrate here that eEF-2 kinase can efficiently phosphorylate a 16-amino acid peptide, MH-1, corresponding to the myosin heavy chain kinase A phosphorylation site in Dictyostelium myosin heavy chains. This enabled us to develop a rapid assay for eEF-2 kinase activity. To localize the functional domains of eEF-2 kinase, we expressed human eEF-2 kinase in Escherichia coli as a GST-tagged fusion protein, and then performed systematic in vitro deletion mutagenesis. We analyzed eEF-2 kinase deletion mutants for the ability to autophosphorylate, and to phosphorylate eEF-2 as well as a peptide substrate, MH-1. Mutants with deletions between amino acids 51 and 335 were unable to autophosphorylate, and were also unable to phosphorylate eEF-2 and MH-1. Mutants with deletions between amino acids 521 and 725 were unable to phosphorylate eEF-2, but were still able to autophosphorylate and to phosphorylate MH-1. The kinases with deletions between amino acids 2 and 50 and 336 and 520 were able to catalyze all three reactions. In addition, the C-terminal domain expressed alone (amino acids 336-725) binds eEF-2 in a coprecipitation assay. These results suggest that eEF-2 kinase consists of two domains connected by a linker region. The amino-terminal domain contains the catalytic domain, while the carboxyl-terminal domain contains the eEF-2 targeting domain. The calmodulin-binding region is located between amino acids 51 and 96. The amino acid sequence of the carboxyl-terminal domain of eEF-2 kinase displays similarity to several proteins, all of which contain repeats of a 36-amino acid motif that we named "motif 36".

Cascaded multiple classifiers for secondary structure prediction

We describe a new classifier for protein secondary structure prediction that is formed by cascading together different types of classifiers using neural networks and linear discrimination. The new classifier achieves an accuracy of 76.7% (assessed by a rigorous full Jack-knife procedure) on a new nonredundant dataset of 496 nonhomologous sequences (obtained from G.J. Barton and J.A. Cuff). This database was especially designed to train and test protein secondary structure prediction methods, and it uses a more stringent definition of homologous sequence than in previous studies. We show that it is possible to design classifiers that can highly discriminate the three classes (H, E, C) with an accuracy of up to 78% for beta-strands, using only a local window and resampling techniques. This indicates that the importance of long-range interactions for the prediction of beta-strands has been probably previously overestimated.

Conformational selection of inhibitors and substrates by proteolytic enzymes: implications for drug design and polypeptide processing

Inhibitors of proteolytic enzymes (proteases) are emerging as prospective treatments for diseases such as AIDS and viral infections, cancers, inflammatory disorders, and Alzheimer's disease. Generic approaches to the design of protease inhibitors are limited by the unpredictability of interactions between, and structural changes to, inhibitor and protease during binding. A computer analysis of superimposed crystal structures for 266 small molecule inhibitors bound to 48 proteases (16 aspartic, 17 serine, 8 cysteine, and 7 metallo) provides the first conclusive proof that inhibitors, including substrate analogues, commonly bind in an extended beta-strand conformation at the active sites of all these proteases. Representative superimposed structures are shown for (a) multiple inhibitors bound to a protease of each class, (b) single inhibitors each bound to multiple proteases, and (c) conformationally constrained inhibitors bound to proteases. Thus inhibitor/substrate conformation, rather than sequence/composition alone, influences protease recognition, and this has profound implications for inhibitor design. This conclusion is supported by NMR, CD, and binding studies for HIV-1 protease inhibitors/substrates which, when preorganized in an extended conformation, have significantly higher protease affinity. Recognition is dependent upon conformational equilibria since helical and turn peptide conformations are not processed by proteases. Conformational selection explains the resistance of folded/structured regions of proteins to proteolytic degradation, the susceptibility of denatured proteins to processing, and the higher affinity of conformationally constrained 'extended' inhibitors/substrates for proteases. Other approaches to extended inhibitor conformations should similarly lead to high-affinity binding to a protease.

A graphic approach to evaluate algorithms of secondary structure prediction

Algorithms of secondary structure prediction have undergone the developments of nearly 30 years. However, the problem of how to appropriately evaluate and compare algorithms has not yet completely solved. A graphic method to evaluate algorithms of secondary structure prediction has been proposed here. Traditionally, the performance of an algorithm is evaluated by a number, i.e., accuracy of various definitions. Instead of a number, we use a graph to completely evaluate an algorithm, in which the mapping points are distributed in a three-dimensional space. Each point represents the predictive result of the secondary structure of a protein. Because the distribution of mapping points in the 3D space generally contains more information than a number or a set of numbers, it is expected that algorithms may be evaluated and compared by the proposed graphic method more objectively. Based on the point distribution, six evaluation parameters are proposed, which describe the overall performance of the algorithm evaluated. Furthermore, the graphic method is simple and intuitive. As an example of application, two advanced algorithms, i.e., the PHD and NNpredict methods, are evaluated and compared. It is shown that there is still much room for further improvement for both algorithms. It is pointed out that the accuracy for predicting either the alpha-helix or beta-strand in proteins with higher alpha-helix or beta-strand content, respectively, should be greatly improved for both algorithms.