diSBPred: A machine learning based approach for disulfide bond prediction

The protein disulfide bond is a covalent bond that forms during post-translational modification by the oxidation of a pair of cysteines. In protein, the disulfide bond is the most frequent covalent link between amino acids after the peptide bond. It plays a significant role in three-dimensional (3D) ab initio protein structure prediction (aiPSP), stabilizing protein conformation, post-translational modification, and protein folding. In aiPSP, the location of disulfide bonds can strongly reduce the conformational space searching by imposing geometrical constraints. Existing experimental techniques for the determination of disulfide bonds are time-consuming and expensive. Thus, developing sequence-based computational methods for disulfide bond prediction becomes indispensable. This study proposed a stacking-based machine learning approach for disulfide bond prediction (diSBPred). Various useful sequence and structure-based features are extracted for effective training, including conservation profile, residue solvent accessibility, torsion angle flexibility, disorder probability, a sequential distance between cysteines, and more. The prediction of disulfide bonds is carried out in two stages: first, individual cysteines are predicted as either bonding or non-bonding; second, the cysteine-pairs are predicted as either bonding or non-bonding by including the results from cysteine bonding prediction as a feature.The examination of the relevance of the features employed in this study and the features utilized in the existing nearest neighbor algorithm (NNA) method shows that the features used in this study improve about 7.39 % in jackknife validation balanced accuracy. Moreover, for individual cysteine bonding prediction and cysteine-pair bonding prediction, diSBPred provides a 10-fold cross-validation balanced accuracy of 82.29 % and 94.20 %, respectively. Altogether, our predictor achieves an improvement of 43.25 % based on balanced accuracy compared to the existing NNA based approach. Thus, diSBPred can be utilized to annotate the cysteine bonding residues of protein sequences whose structures are unknown as well as improve the accuracy of the aiPSP method, which can further aid in experimental studies of the disulfide bond and structure determination.     

Conotoxin Φ‐MiXXVIIA from the Superfamily G2 Employs a Novel Cysteine Framework that Mimics Granulin and Displays Anti‐Apoptotic Activity

Conotoxins are a large family of disulfide‐rich peptides that contain unique cysteine frameworks that target a broad range of ion channels and receptors. We recently discovered the 33‐residue conotoxin Φ‐MiXXVIIA from Conus miles with a novel cysteine framework comprising three consecutive cysteine residues and four disulfide bonds. Regioselective chemical synthesis helped decipher the disulfide bond connectivity and the structure of Φ‐MiXXVIIA was determined by NMR spectroscopy. The 3D structure displays a unique topology containing two β‐hairpins that resemble the N‐terminal domain of granulin. Similar to granulin, Φ‐MiXXVIIA promotes cell proliferation (EC₅₀ 17.85 μm ) while inhibiting apoptosis (EC₅₀ 2.2 μm ). Additional framework XXVII sequences were discovered with homologous signal peptides that define the new conotoxin superfamily G2. The novel structure and biological activity of Φ‐MiXXVIIA expands the repertoire of disulfide‐rich conotoxins that recognize mammalian receptors.     

Bioinformatics approaches for disulfide connectivity prediction

Protein structure prediction with computational methods has gained much attention in the research fields of protein engineering and protein folding studies. Due to the vastness of conformational space, one of the major tasks is to restrain the flexibility of protein structure and reduce the search space. Many studies have revealed that, with the information of disulfide connectivity available, the search in conformational space can be dramatically reduced and lead to significant improvements in the prediction accuracy. As a result, predicting disulfide connectivity using bioinformatics approaches is of great interest nowadays. In this mini-review, the prediction of disulfide connectivity in proteins will be discussed in four aspects: (1) how the problem formulated and the computational techniques used in the literatures; (2) the effects of the features adopted to encode the information and the biological meanings implied; (3) the problems encountered and limitations of disulfide connectivity prediction; and (4) the practical usages of predicted disulfide bond information in molecular simulation and the prospects in the future.     

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 of cysteine residues and disulfide bonds in proteins by liquid chromatography/electrospray ionization tandem mass spectrometry

Cysteine residues and disulfide bonds are important for protein structure and function. We have developed a simple and sensitive method for determining the presence of free cysteine (Cys) residues and disulfide bonded Cys residues in proteins (<100 pmol) by liquid chromatography/electrospray ionization tandem mass spectrometry (LC/ESI-MS/MS) in combination with protein database searching using the program Sequest. Free Cys residues in a protein were labeled with PEO-maleimide biotin immediately followed by denaturation with 8 M urea. Subsequently, the protein was digested with trypsin or chymotrypsin and the resulting products were analyzed by capillary LC/ESI-MS/MS for peptides containing modified Cys and/or disulfide bonded Cys residues. Although the MS method for identifying disulfide bonds has been routinely employed, methods to prevent thiol-disulfide exchange have not been well documented. Our protocol was found to minimize the occurrence of the thiol-disulfide exchange reaction. The method was validated using well-characterized proteins such as aldolase, ovalbumin, and beta-lactoglobulin A. We also applied this method to characterize Cys residues and disulfide bonds of beta 1,4-galactosyltransferase (five Cys), and human blood group A and B glycosyltransferases (four Cys). Our results demonstrate that beta 1,4-galactosyltransferase contains one free Cys residue and two disulfide bonds, which is in contrast to work previously reported using chemical methods for the characterization of free Cys residues, but is consistent with recently published results from x-ray crystallography. In contrast to the results obtained for beta 1,4-galactosyltransferase, none of the Cys residues in A and B glycosyltransferases were found to be involved in disulfide bonds.     

"Forbidden" disulfides: their role as redox switches

Seminal studies by Richardson and Thornton defined the constraints imposed by protein structure on disulfide formation and flagged forbidden regions of primary or secondary structure seemingly incapable of forming disulfide bonds between resident cysteine pairs. With respect to secondary structure, disulfide bonds were not found between cysteine pairs: A. on adjacent beta-stands; B. in a single helix or strand; C. on non-adjacent strands of the same beta-sheet. In primary structure, disulfide bonds were not found between cysteine pairs: D. adjacent in the sequence. In the intervening years it has become apparent that all these forbidden regions are indeed occupied by disulfide-bonded cysteines, albeit rather strained ones. It has been observed that sources of strain in a protein structure, such as residues in forbidden regions of the Ramachandran plot and cis-peptide bonds, are found in functionally important regions of the protein and warrant further investigation. Like the Ramachandran plot, the earlier studies by Richardson and Thornton have identified a fundamental truth in protein stereochemistry: "forbidden" disulfides adopt strained conformations, but there is likely a functional reason for this. Emerging evidence supports a role for forbidden disulfides in redox-regulation of proteins.     

Total Synthesis of Human Hepcidin through Regioselective Disulfide‐Bond Formation by using the Safety‐Catch Cysteine Protecting Group 4,4′‐Dimethylsulfinylbenzhydryl

A safety‐catch cysteine protecting group, S‐4,4′‐dimethylsulfinylbenzhydryl (Msbh), was designed and developed to expand the capabilities of synthetic strategies for the regioselective formation of disulfide bonds in cysteine‐rich peptides. The directed regioselective synthesis of human hepcidin, which contains four disulfide bonds, was undertaken and led to a high‐resolution NMR structure under more physiologically relevant conditions than previously. Conversely, hepcidin synthesized with the formerly assigned vicinal disulfide‐bond connectivity displayed significant conformational heterogeneity under similar conditions. The two synthetic forms of human hepcidin induced ferroportin internalization with apparent EC₅₀ values of 2.0 (native fold, 1 ) and 4.4 nM (non‐native fold, 2 ), with 2 undergoing isomerization to 1 in the presence of ferroportin expressing cells.     

A small-molecule catalyst of protein folding in vitro and in vivo

BACKGROUND: The formation of native disulfide bonds between cysteine residues often limits the rate and yield of protein folding. The enzyme protein disulfide isomerase (PDI) catalyzes the interchange of disulfide bonds in substrate proteins. The two -Cys-Gly-His-Cys- active sites of PDI provide a thiol that has a low pKa value and a disulfide bond of high reduction potential (Eo'). RESULTS: A synthetic small-molecule dithiol, (+/-)-trans-1,2-bis(2-mercaptoacetamido)cyclohexane (BMC), has a pKa value of 8.3 and an Eo' value of -0.24 V. These values are similar to those of the PDI active sites. BMC catalyzes the activation of scrambled ribonuclease A, an inactive enzyme with non-native disulfide bonds, and doubles the yield of active enzyme. A monothiol analog of BMC, N-methylmercaptoacetamide, is a less efficient catalyst than BMC. BMC in the growth medium of Saccharomyces cerevisiae cells increases by > threefold the heterologous secretion of Schizosaccharomyces pombe acid phosphatase, which has eight disulfide bonds. This effect is similar to that from the overproduction of PDI in the S. cerevisiae cells, indicating that BMC, like PDI, can catalyze protein folding in vivo. CONCLUSIONS: A small-molecule dithiol with a low thiol pKa value and high disulfide Eo' value can mimic PDI by catalyzing the formation of native disulfide bonds in proteins, both in vitro and in vivo.     

Oxidation of bovine serum albumin: identification of oxidation products and structural modifications

Albumin is an important plasma antioxidant protein, contributing to protecting mechanisms of cellular and regulatory long‐lived proteins. The metal‐catalyzed oxidation (MCO) of proteins plays an important role during oxidative stress. In this study, we examine the oxidative modification of albumin using an MCO in vitro system. Mass spectrometry, combined with off‐line nano‐liquid chromatography, was used to identify modifications in amino acid residues. We have found 106 different residues oxidatively damaged, being the main oxidized residues lysines, cysteines, arginines, prolines, histidines and tyrosines. Besides protein hydroxyl derivatives and oxygen additions, we detected other modifications such as deamidations, carbamylations and specific amino acid oxidative modifications. The oxidative damage preferentially affects particular subdomains of the protein at different time‐points. Results suggest the oxidative damage occurs first in exposed regions near cysteine disulfide bridges with residues like methionine, tryptophan, lysine, arginine, tyrosine and proline appearing as oxidatively modified. The damage extended afterwards with further oxidation of cysteine residues involved in disulfide bridges and other residues like histidine, phenylalanine and aspartic acid. The time‐course evaluation also shows the number of oxidized residues does not increase linearly, suggesting that oxidative unfolding of albumin occurs through a step‐ladder mechanism. Copyright © 2009 John Wiley & Sons, Ltd.     

On-line sample clean-up and chromatography coupled with electrospray ionization mass spectrometry to characterize the primary sequence and disulfide bond content of recombinant calcium binding proteins

We have used on-line sample clean-up, concentration, and chromatography with electrospray ionization mass spectrometry (ESI-MS), to characterize and determine the presence of disulfide bonds in recombinant full-length rat brain calbindin D28K and two deletion mutants of the protein, one lacking EF-hand 2 (calbindin delta 2) and the other lacking EF-hands 2 and 6 (calbindin delta 2,6). The molecular weights of the expressed proteins dissolved in biological buffers were determined with high accuracy using a low-flow, pressurized chamber infusion system, that allows on-line protein clean-up by removing buffers/salts incompatible with ESI-MS. The molecular weight determinations showed that the amino-terminal methionine residues had been cleaved during the expression and isolation of the recombinant proteins. Approximately 85-90% of the protein sequences were confirmed by on-line HPLC-ESI-MS analysis of peptides generated by a lysyl endoproteinase C digestion. Comparisons of ESI-MS spectra of native and reduced calbindin D28K and delta 2 show that the full length- and delta 2 mutant-protein contain one disulfide bond. Molecular mass determinations of calbindin delta 2,6 showed that this protein contains a highly active cysteine residue that covalently binds a mercaptoethanol group, or forms a homodimer via a disulfide bond. The results show surprising differences amongst the deletion mutants of calbindin D28K with respect to the formation of disulfide bonds. These differences are not readily detected by other techniques and show that ESI-MS is a powerful, rapid method by which to detect disulfide linkages for intact proteins.     

Molecular Dynamics Simulations of Proteins with Chemically Modified Disulfide Bonds

Proteins that are used as therapeutic drugs act in the extracellular microenvironment. They usually have a small number of intramolecular disulfide bonds to help maintain their tertiary structure in the vascular circulation. In general, most cysteine residues are part of a disulfide bond with free sulfhydrals being uncommon. We have studied whether the site-specific chemical reduction of disulfides and the incorporation of a 3-carbon methylene bridge between the cysteines in interferon-α 2a would change the structure of this protein. Bridging of both of the disulfide bonds of interferon-α 2a was studied using two different molecular simulation protocols: (1) molecular dynamics, and (2) stochastic dynamics. We have shown that the disulfide bonds in interferon-α 2a can be reduced and chemically modified without significantly altering the tertiary structure of the protein. This offers the novel possibility of chemically modifying therapeutically important proteins without affecting their biological properties.     

Characterization of bovine surfactant proteins B and C by electrospray ionization mass spectrometry

Bovine surfactant proteins B (SP-B) and C (SP-C) were analyzed by nano-electrospray ionization mass spectrometry (nano-ESI-MS). The observed molecular masses showed discrepancies compared to the calculated molecular masses using the published amino acid sequences. The number of cysteine residues in the published bovine SP-B amino acid sequences also failed to match the observed mass shift upon reduction of the SP-B dimer. To determine the amino acid sequences of two proteins, SP-B was first digested with trypsin and analyzed by liquid chromatography/tandem mass spectrometry (LC/MS/MS), while SP-C was analyzed by MS/MS in its intact form. The amino acid sequence of bovine SP-B determined here matches the observed molecular mass. The sequence is almost identical to the sheep SP-B except for two amino acid residues, consistent with the proximity of the two species. The correct sequence contains seven cysteine residues. Bovine SP-B exists as dimers and all cysteines are oxidized to form disulfide bonds in physiological conditions, which is in agreement with the observed mass shift upon reduction of the SP-B dimer. These cysteine residues are completely conserved across all species indicating their importance for the biological functions of this surfactant protein. The sequence of SP-C determined here also reveals an L to V substitution at its position 22 compared with the published bovine SP-B sequence.     

Multiple Disulfide-Bonded States of Native Proteins: Estimate of Number Using Probabilities of Disulfide Bond Formation

The polypeptide backbone of proteins is held together by two main types of covalent bonds: the peptide bonds that link the amino acid residues and the disulfide bonds that link pairs of cysteine amino acids. Disulfide bonds form as a protein folds in the cell and formation was assumed to be complete when the mature protein emerges. This is not the case for some secreted human blood proteins. The blood clotting protein, fibrinogen, and the protease inhibitor, α2-macroglobulin, exist in multiple disulfide-bonded or covalent states in the circulation. Thousands of different states are predicted assuming no dependencies on disulfide bond formation. In this study, probabilities for disulfide bond formation are employed to estimate numbers of covalent states of a model polypeptide with reference to α2-macroglobulin. When disulfide formation is interdependent in a protein, the number of covalent states is greatly reduced. Theoretical estimates of the number of states will aid the conceptual and experimental challenges of investigating multiple disulfide-bonded states of a protein.     

Structural and functional diversity of glutaredoxins in yeast

Glutaredoxins are defined as thiol disulfide oxidoreductases that reduce disulfide bonds employing reduced glutathione as electron donor. They constitute a complex family of proteins with a diversity of enzymatic and functional properties. Thus, dithiol glutaredoxins are able to reduce disulfide bonds and deglutathionylate mixed disulfides between glutathione and cysteine protein residues. They could act regulating the redox state of sulfhydryl residues of specific proteins, while thioredoxins (another family of thiol disulfide oxidoreductases which employ NADPH as electron donor) would be the general sulfhydryl reductants. Some dithiol glutaredoxins such as human Grx2 form dimers bridged by one iron-sulfur cluster, which acts as a sensor of oxidative stress, therefore regulating the activity of the glutaredoxin. The ability to interact with iron-sulfur clusters as ligands is also characteristic of monothiol glutaredoxins with a CGFS-type active site. These do not display thiol oxidoreductase activity, but have roles in iron homeostasis. The three members of this subfamily in Saccharomyces cerevisiae participate in the synthesis of the iron-sulfur clusters in mitochondria (Grx5), or in signalling the iron status inside the cell for regulation of iron uptake and intracellular iron relocalization (Grx3 and Grx4). Such a role in iron metabolism seems to be evolutionary conserved. Fungal cells also contain membrane-associated glutaredoxins structurally and enzymatically similar to dithiol glutaredoxins, which may act as redox regulators at the early stages of the protein secretory machinery.     

Helix‐coil and beta sheet‐coil transitions in a simplified,yet realistic protein model

A reduced model of polypeptide chains and protein stochastic dynamics is employed in Monte Carlo studies of the coil‐globule transition. The model assumes a high‐resolution lattice representation of protein conformational space. The interaction scheme is derived from a statistical analysis of structural regularities seen in known three‐dimensional protein structures. It is shown that model polypeptides containing residues that have strong propensities towards locally expanded conformations collapse to β‐like globular conformations, while polypeptides containing residues with helical propensities form globules of closely packed helices. A more cooperative transition is observed for β‐type systems. It is also demonstrated that hydrogen bonding is an important factor for protein cooperativity, although, for systems with suppressed hydrogen bond interactions, a higher cooperativity of β‐type proteins is also observed.     

Plant Antimicrobial Peptides Snakin‐1 and Snakin‐2: Chemical Synthesis and Insights into the Disulfide Connectivity

Antimicrobial peptides and proteins represent an important class of plant defensive compounds against pathogens and provide a rich source of lead compounds in the field of drug discovery. We describe the effective preparation of the cysteine‐rich snakin‐1 and ‐2 antimicrobial peptides by using a combination of solid‐phase synthesis and native chemical ligation. A subsequent cysteine/cystine mediated oxidative folding to form the six internal disulfide bonds concurrently gave the folded proteins in 40–50 % yield. By comparative evaluation of mass spectrometry, HPLC, biological data and trypsin digest mapping of folded synthetic snakin‐2 compared to natural snakin‐2, we demonstrated that synthetic snakin‐2 possesses full antifungal activity and displayed similar chromatographic behaviour to natural snakin‐2. Trypsin digest analysis allowed tentative assignment of three of the purported six disulfide bonds.     

Connectivity and binding‐site recognition: Applications relevant to drug design

Here, we describe a family of methods based on residue–residue connectivity for characterizing binding sites and apply variants of the method to various types of protein–ligand complexes including proteases, allosteric‐binding sites, correctly and incorrectly docked poses, and inhibitors of protein–protein interactions. Residues within ligand‐binding sites have about 25% more contact neighbors than surface residues in general; high‐connectivity residues are found in contact with the ligand in 84% of all complexes studied. In addition, a k‐means algorithm was developed that may be useful for identifying potential binding sites with no obvious geometric or connectivity features. The analysis was primarily carried out on 61 protein–ligand structures from the MEROPS protease database, 250 protein–ligand structures from the PDBSelect (25%), and 30 protein–protein complexes. Analysis of four proteases with crystal structures for multiple bound ligands has shown that residues with high connectivity tend to have less variable side‐chain conformation. The relevance to drug design is discussed in terms of identifying allosteric‐binding sites, distinguishing between alternative docked poses and designing protein interface inhibitors. Taken together, this data indicate that residue–residue connectivity is highly relevant to medicinal chemistry. © 2010 Wiley Periodicals, Inc. J Comput Chem, 2010     

Site‐Selective Disulfide Modification of Proteins: Expanding Diversity beyond the Proteome

The synthetic transformation of polypeptides with molecular accuracy holds great promise for providing functional and structural diversity beyond the proteome. Consequently, the last decade has seen an exponential growth of site‐directed chemistry to install additional features into peptides and proteins even inside living cells. The disulfide rebridging strategy has emerged as a powerful tool for site‐selective modifications since most proteins contain disulfide bonds. In this Review, we present the chemical design, advantages and limitations of the disulfide rebridging reagents, while summarizing their relevance for synthetic customization of functional protein bioconjugates, as well as the resultant impact and advancement for biomedical applications.     

An information‐theoretic approach to the prediction of protein structural class

An information‐theoretical approach, which combines a sequence decomposition technique and a fuzzy clustering algorithm, is proposed for prediction of protein structural class. This approach could bypass the process of selecting and comparing sequence features as done previously. First, distances between each pair of protein sequences are estimated using a conditional decomposition technique in information theory. Then, the fuzzy k‐nearest neighbor algorithm is used to identify the structural class of a protein given as set of sample sequences. To verify the strength of our method, we choose three widely used datasets constructed by Chou and Zhou. It is shown by the Jackknife test that our approach represents an improvement in the prediction of accuracy over existing methods. © 2009 Wiley Periodicals, Inc. J Comput Chem, 2010     

4,4'-dithiobisdipicolinic acid: a small and convenient lanthanide binding tag for protein NMR spectroscopy

Pseudocontact shifts (PCS) from paramagnetic lanthanide ions present powerful long-range structure restraints for studies of proteins by nuclear magnetic resonance spectroscopy. To elicit PCSs, the lanthanide must be attached site-specifically to the target protein. In addition, it needs to be attached rigidly to avoid averaging of the PCSs due to mobility with respect to the protein and it must not interfere with the function of the protein. Here, we present a dipicolinic acid reagent that spontaneously forms a disulfide bond with thiol groups of accessible cysteine residues. A minimal number of rotatable bonds between the cysteine side chain and the tag helps to minimise mobility. Combined with the small size of the tag and quantitative tagging yields, these features make it a highly attractive tool for generating structure restraints by paramagnetic lanthanides.