Dynamic redox environment-intensified disulfide bond shuffling for protein refolding in vitro: molecular simulation and experimental validation

One challenge in protein refolding is to dissociate the non-native disulfide bonds and promote the formation of native ones. In this study, we present a coarse-grained off-lattice model protein containing disulfide bonds and simulate disulfide bond shuffling during the folding of this model protein. Introduction of disulfide bonds in the model protein led to enhanced conformational stability but reduced foldability in comparison to counterpart protein without disulfide bonds. The folding trajectory suggested that the model protein retained the two-step folding mechanism in terms of hydrophobic collapse and structural rearrangement. The disulfide bonds located in the hydrophobic core were formed before the collapsing step, while the bonds located on the protein surface were formed during the rearrangement step. While a reductive environment at the initial stage of folding favored the formation of native disulfide bonds in the hydrophobic core, an oxidative environment at a later stage of folding was required for the formation of disulfide bonds at protein surface. Appling a dynamic redox environment, that is, one that changes from reductive to oxidative, intensified disulfide bond shuffling and thus resulted in improved recovery of the native conformation. The above-mentioned simulation was experimentally validated by refolding hen-egg lysozyme at different urea concentrations and oxidized glutathione/reduced glutathione (GSSG/GSH) ratios, and an optimal redox environment, in terms of the GSSG to GSH ratio, was identified. The implementation of a dynamic redox environment by tuning the GSSG/GSH ratio further improved the refolding yield of lysozyme, as predicted by molecular simulation.     

Native conformational tendencies in unfolded polypeptides: development of a novel method to assess native conformational tendencies in the reduced forms of multiple disulfide-bonded proteins

Oxidative folding is the concomitant formation of the native disulfide bonds and the native tertiary structure from the reduced and unfolded polypeptide. Of interest is the inherent conformational tendency (bias) present in the reduced polypeptide to dictate the formation of the full set of native disulfide bonds. Here, by application of a novel tool, we have been able to assess this "native conformational tendency" present in reduced and unfolded bovine pancreatic ribonuclease A (RNase A). The essence of this method lies in the ability of the oxidant [Pt(en)(2)Cl(2)](2+) (where "en" is ethylenediamine) to oxidize disulfide bonds under conditions in which both reduction and disulfide reshuffling, which are essential for rearranging non-native disulfide bonds, are extremely slow. When applied to RNase A, the method revealed little or no bias toward formation of the full native set of disulfide bonds in the fully reduced protein.     

The native ensemble and folding of a protein molten-globule: functional consequence of downhill folding

The continually emerging functional significance of intrinsic disorder and conformational flexibility in proteins has challenged the long-standing dogma of a well-defined structure contributing to a specific function. Molten-globular states, a class of proteins with significant secondary-structure but a fluid hydrophobic core, is one such example. They have however been difficult to characterize due to the complexity of experimental data and lack of computational avenues. Here, we dissect the folding mechanism of the α-helical molten-globular protein NCBD from three fundamentally different approaches: statistical-mechanical variable barrier model, C(α)-based Gō-model and explicit water all-atom molecular dynamics (MD) simulations. We find that NCBD displays the characteristics of a one-state globally downhill folder but is significantly destabilized. Using simulation techniques, we generate a highly constrained but a heterogeneous native ensemble of the molten-globule for the first time that is consistent with experimental data including small angle X-ray scattering (SAXS), circular dichroism (CD), and nuclear magnetic resonance (NMR). The resulting native ensemble populates conformations reported in other bound-forms providing direct evidence to the mechanism of conformational selection for binding multiple partners in this domain. Importantly, our simulations reveal a connection between downhill folding and large conformational flexibility in this domain that has been evolutionarily selected and functionally exploited resulting in large binding promiscuity. Finally, the multimodel approach we employ here serves as a powerful methodology to study mechanisms and suggests that the thermodynamic features of molten-globules fall within the array of folding mechanisms available to small single-domain proteins.     

A new method for rapid characterization of the folding pathways of multidisulfide-containing proteins

Oxidative folding is a composite process that consists of both the conformational folding to the native three-dimensional structure and the regeneration of the native disulfide bonds of a protein, frequently involving over 100 disulfide intermediate species. Understanding the oxidative folding pathways of a multiple-disulfide-containing protein is a very difficult task that often requires years of devoted research due to the high complexity of the process and the very similar features of the large number of intermediates. Here we developed a method for rapidly delineating the major features of the oxidative folding pathways of a protein. The method examines the temperature dependence of the oxidative folding rate of the protein in combination with reduction pulses. Reduction pulses expose the presence of structured intermediates along the pathways. The correlation between the regeneration rate at different temperatures and the stability of the structured intermediates reveals the role that the intermediates play in determining the pathway. The method was first tested with bovine pancreatic ribonuclease A whose folding pathways were defined earlier. Then, it was explored to discern some of the major features of the folding pathways of its homologue, frog Onconase. The results suggest that the stability of the three-dimensional structure of the native protein is a major determinant of the folding rate in oxidative folding.     

Context-dependence of the contribution of disulfide bonds to beta-hairpin stability

Incorporation of disulfide bonds to stabilize protein and peptide structures is not always a successful strategy. To advance current knowledge on the contribution of disulfide bonds to beta-hairpin stability, a previously reported beta-hairpin-forming peptide was taken as a template to design a series of Cys-containing peptides. The conformational behavior of these peptides in their oxidized, disulfide-cyclized peptides, and reduced, linear peptides, was investigated on the basis of NMR parameters: NOEs, and 1H and 13C chemical shifts. We found that the effect of disulfide bonds on beta-hairpin stability depends on its location within the beta-hairpin structure, being very small or even destabilizing when connecting two hydrogen-bonded facing residues. When the disulfide bond is linking non-hydrogen-bonded facing residues, we estimated that its contribution to the free-energy change of beta-hairpin folding is approximately -1.0 kcal mol(-1). This value is larger than those reported for most beta-hairpin-stabilizing cross-strand side-chain-side-chain interactions, except for some aromatic-aromatic interactions, in particular the Trp-Trp one, and the cation-pi interaction between Trp and the non-natural methylated Arg/Lys. As disulfide bonds are frequently used to stabilize peptide conformations, our conclusions can be useful for peptide, peptidomimetic, and protein design, and may even extend to other chemical cross-links.     

Conformational simulations of aqueous solvated alpha-conotoxin GI and its single disulfide analogues using a polarizable force field model

The solution conformation of alpha-conotoxin GI and its two single disulfide analogues are simulated using a polarizable force field in combination with the molecular fragmentation quantum chemical calculation. The polarizability is explicitly described by allowing the partial charges and fragment dipole moments to be variables, with values coming from the linear-scaling energy-based molecular fragmentation calculations at the B3LYP/6-31G(d) level. In comparison with the full quantum chemical calculations, the fragmentation approaches can yield precise ground-state energies, dipole moments, and static polarizabilities for peptides. The B3LYP/6-31G(d) charges and fragment-centered dipole moments are introduced in calculations of electrostatic terms in both AmberFF03 and OPLS force fields. Our test calculations on the gas-phase glucagon (PDB code: 1gcn) and solvated alpha-conotoxin GI (PDB code: 1not) demonstrate that the present polarization model is capable of describing the structural properties (such as the relative conformational energies, intramolecular hydrogen bonds, and disulfide bonds) with accuracy comparable to some other polarizable force fields (ABEEM/MM and OPLS-PFF) and the quantum mechanics/molecular mechanics (QM/MM) hybrid model. The employment of fragment-centered dipole moments in calculations of dipole-dipole interactions can save computational time in comparison with those polarization models using atom-centered dipole moments without much loss of accuracy. The molecular dynamics simulations using the polarizable force field demonstrate that two single disulfide GI analogues are more flexible and less structured than the native alpha-conotoxin GI, in agreement with NMR experiments. The polarization effect is important in simulations of the folding/unfolding process of solvated proteins.     

Probing the effect of temperature on the backbone dynamics of the human alpha-lactalbumin molten globule

Molten globules are compact, partially folded proteins postulated to be general intermediates in protein folding. Human alpha-lactalbumin (alpha-LA) is a two-domain Ca(2+)-binding protein that partially unfolds at low pH to form a molten globule. NMR spectra of molten globules are characterized by broadened resonances due to conformational fluctuations on microsecond to millisecond time scales. These species are often studied at high temperature where NMR resonances are observed to sharpen. The effect of higher temperatures on fast time-scale backbone dynamics of molten globules has not been investigated previously. Here, 1D (15)N direct-detection and 2D indirect-detection (1)H-(15)N heteronuclear NOE experiments have been used to probe fast time-scale dynamics at low and high temperatures for three disulfide-bond variants of human alpha-LA that form molten globules. Disulfide bonds are found to have a significant effect on backbone dynamics within the beta-domain of the molten globule; within the alpha-domain, dynamics are not significantly influenced by these bonds. At 20 degrees C, backbone mobility is significantly decreased in both domains of the molten globule compared to the mobility at 40-50 degrees C. Heteronuclear NOE values determined at 20 degrees C for the alpha-domain are closely similar to those observed for native alpha-LA, indicating that the alpha-LA molten globule has even more native-like character than suggested by studies conducted at higher temperature. Our results highlight the importance of considering the temperature dependence of the molten globule ensemble when making comparisons between experimental data obtained under different conditions.     

Development of a novel method to populate native disulfide-bonded intermediates for structural characterization of proteins: implications for the mechanism of oxidative folding of RNase A

RNase A, a model protein for oxidative folding studies, has four native disulfide bonds. The roles of des [40-95] and des [65-72], the two native-like structured three-disulfide-bonded intermediates populated between 8 and 25 degrees C during the oxidative folding of RNase A, are well characterized. Recent work focuses on both the formation of these structured disulfide intermediates from their unstructured precursors and on the subsequent oxidation of the structured species to form the native protein. The major obstacles in this work are the very low concentration of the precursor species and the difficulty of isolating some of the structured intermediates. Here, we demonstrate a novel method that enables the native disulfide-bonded intermediates to be populated and studied regardless of whether they have stable structure and/or are present at low concentrations during the oxidative folding or reductive unfolding process. The application of this method enabled us to populate and, in turn, study the key intermediates with two native disulfide bonds on the oxidative folding pathway of RNase A; it also facilitated the isolation of des [58-110] and des [26-84], the other two native-like structured des species whose isolation had thus far not been possible.     

Diaminodiacid Bridges to Improve Folding and Tune the Bioactivity of Disulfide‐Rich Peptides

Disulfide‐rich peptides containing three or more disulfide bonds are promising therapeutic and diagnostic agents, but their preparation is often limited by the tedious and low‐yielding folding process. We found that a single cystine‐to‐diaminodiacid replacement could significantly increase the folding efficiency of disulfide‐rich peptides and thus improve their production yields. The practicality of this strategy was demonstrated by the synthesis and folding of derivatives of the μ‐conotoxin SIIIA, the preclinical hormone hepcidin, and the trypsin inhibitor EETI‐II. NMR and X‐ray crystallography studies confirmed that these derivatives of disulfide‐rich peptide retained the correct three‐dimensional conformations. Moreover, the cystine‐to‐diaminodiacid replacement enabled structural tuning, thereby leading to an EETI‐II derivative with higher bioactivity than the native peptide.     

Folding, conformational changes, and dynamics of cytochromes C probed by NMR spectroscopy

NMR spectroscopy has become a vital tool for studies of protein conformational changes and dynamics. Oxidized Fe(III)cytochromes c are a particularly attractive target for NMR analysis because their paramagnetism (S = (1)/(2)) leads to high (1)H chemical shift dispersion, even for unfolded or otherwise disordered states. In addition, analysis of shifts induced by the hyperfine interaction reveals details of the structure of the heme and its ligands for native and nonnative protein conformational states. The use of NMR spectroscopy to investigate the folding and dynamics of paramagnetic cytochromes c is reviewed here. Studies of nonnative conformations formed by denaturation and by anomalous in vivo maturation (heme attachment) are facilitated by the paramagnetic, low-spin nature of native and nonnative forms of cytochromes c. Investigation of the dynamics of folded cytochromes c also are aided by their paramagnetism. As an example of this analysis, the expression in Escherichia coli of cytochrome c(552) from Nitrosomonas europaea is reported here, along with analysis of its unusual heme hyperfine shifts. The results are suggestive of heme axial methionine fluxion in N. europaea ferricytochrome c(552). The application of NMR spectroscopy to investigate paramagnetic cytochrome c folding and dynamics has advanced our understanding of the structure and dynamics of both native and nonnative states of heme proteins.     

Enzymes catalyzing protein folding and their cellular functions

In live cells, protein folding often cannot occur spontaneously, but requires the participation of helper proteins - molecular chaperones and foldases. The mechanisms employed by chaperones markedly increase the effectiveness of protein folding, but have no bearing on the rate of this process, whereas foldases actually accelerate protein folding by exerting a direct influence on the rate-limiting steps of the overall reaction. Two types of foldases are known, using different principles of action. Peptidyl-prolyl cis/trans isomerase and protein-disulfide isomerase catalyze the folding of every protein that needs isomerization of prolyl peptide bonds or formation and isomerization of disulfide bonds for proper folding. By contrast, some foldases operating in the periplasm of bacterial cells are specifically designed to help in the folding of substrate proteins whose primary structure does not contain sufficient information for correct folding. In this review, we discuss recent data on the catalytic mechanisms of both types of foldases, focusing specifically on how a catalyst provides the structural information required for the folding of a target protein. Comparative analysis of the mechanisms employed by two different periplasmic foldases is used to substantiate the notion that combinations of a protein which is unable to fold independently and a specific catalyst delivering the necessary steric information are probably designed to achieve some particular biological purposes. The review also covers the problem of participation of peptidyl-prolyl cis/trans isomerase in different cellular functions, highlighting the role of this enzyme in conformational rearrangements of folded native proteins.     

How does the first water shell fold proteins so fast?

First shells of hydration and bulk solvent play a crucial role in the folding of proteins. Here, the role of water in the dynamics of proteins has been investigated using a theoretical protein-solvent model and a statistical physics approach. We formulate a hydration model where the hydrogen bonds between water molecules pertaining to the first shell of the protein conformation may be either mainly formed or broken. At thermal equilibrium, hydrogen bonds are formed at low temperature and are broken at high temperature. To explore the solvent effect, we follow the folding of a large sampling of protein chains, using a master-equation evolution. The dynamics shows a clear mechanism. Above the glass-transition temperature, a large ratio of chains fold very rapidly into the native structure irrespective of the temperature, following pathways of high transition rates through structures surrounded by the solvent with broken hydrogen bonds. Although these states have an infinitesimal probability, they act as strong dynamical attractors and fast folding proceeds along these routes rather than pathways with small transition rates between configurations of much higher equilibrium probabilities. At a given low temperature, a broad jump in the folding times is observed. Below this glass temperature, the pathways where hydrogen bonds are mainly formed become those of highest rates although with conformational changes of huge relaxation times. The present results reveal that folding obeys a double-funnel mechanism.     

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.     

Total chemical synthesis of crambin

Crambin is a small (46 amino acids) protein isolated from the seeds of the plant Crambe abyssinica. Crambin has been extensively used as a model protein for the development of advanced crystallography and NMR techniques and for computational folding studies. We set out to establish synthetic access to crambin. Initially, we synthesized the 46 amino acid polypeptide by native chemical ligation of two distinct sets of peptide segments (15 + 31 and 31 + 15 residues). The synthetic polypeptide chain folded in good yield to give native crambin containing three disulfide bonds. The chemically synthesized crambin was characterized by LC-MS and by 2D-NMR. However, the 31-residue peptide segments were difficult to purify, and this caused an overall low yield for the synthesis. To overcome this problem, we synthesized crambin by the native chemical ligation of three segments (15 + 16 + 15 residues). Total synthesis using the ligation of three segments gave more than a 10-fold increase in yield and a protein product of exceptionally high purity. This work demonstrates the efficacy of chemical protein synthesis by the native chemical ligation of three segments and establishes efficient synthetic access to the important model protein crambin for experimental studies of protein folding and stability.     

Synthesis of PAF,an Antifungal Protein from P. chrysogenum,by Native Chemical Ligation: Native Disulfide Pattern and Fold Obtained upon Oxidative Refolding

The folding of disulfide proteins is of considerable interest because knowledge of this may influence our present understanding of protein folding. However, sometimes even the disulfide pattern cannot be unequivocally determined by the available experimental techniques. For example, the structures of a few small antifungal proteins (PAF, AFP) have been disclosed recently using NMR spectroscopy but with some ambiguity in the actual disulfide pattern. For this reason, we carried out the chemical synthesis of PAF. Probing different approaches, the oxidative folding of the synthetic linear PAF yielded a folded protein that has identical structure and antifungal activity as the native PAF. In contrast, unfolded linear PAF was inactive, a result that may have implications concerning its redox state in the mode of action.     

Homogeneous Human Complex‐Type Oligosaccharides in Correctly Folded Intact Glycoproteins: Evaluation of Oligosaccharide Influence On Protein Folding,Stability, and Conformational Properties

The N‐glycosylation of proteins is generated at the consensus sequence NXS/T (where X is any amino acid except proline) by the biosynthetic process, and occurs in the endoplasmic reticulum and Golgi apparatus. In order to investigate the influence of human complex‐type oligosaccharides on counterpart protein conformation, crambin and ovomucoide, which consist of 46 and 56 amino acid residues, respectively, were selected for synthesis of model glycoproteins. These small glycoproteins were intentionally designed to be glycosylated at the α‐helix (crambin: 8 position), β‐sheet (crambin: 2 position) and loop position between the antiparallel β‐sheets (ovomucoide: 28 position), and were synthesized by using a peptide‐segment coupling strategy. After preparation of these glycosylated polypeptide chains, protein folding experiments were performed under redox conditions by using cysteine–cystine. Although the small glycoproteins bearing intentional glycosylation at the α‐helix and β‐sheet exhibited a suitable folding process, glycosylation at the loop position between the antiparallel β‐strands caused multiple products. The conformational differences in the isolated homogeneous glycoproteins compared with non‐glycosylated counterparts were evaluated by circular dichroism (CD) and NMR spectroscopy. These analyses suggested that this intentional N‐glycosylation did not result in large conformational changes in the purified protein structures, including the case of glycosylation at the loop position between the antiparallel β‐strands. In addition to these experiments, the conformational properties of three glycoproteins were evaluated by CD spectroscopy under different temperatures. The oligosaccharides on the protein surface fluctuated considerably; this was dependent on the increase in the solution temperature and was thought to disrupt the protein tertiary structure. Based on the measurement of the CD spectra, however, the glycoproteins bearing three disulfide bonds did not exhibit any change in their protein tertiary structure. These results suggest that the oligosaccharide conformational fluctuations were not disruptive to protein tertiary structure, and the tertiary structure of glycoproteins might be stabilized by the disulfide bond network.     

Homogeneous human complex-type oligosaccharides in correctly folded intact glycoproteins: evaluation of oligosaccharide influence on protein folding, stability, and conformational properties

The N-glycosylation of proteins is generated at the consensus sequence NXS/T (where X is any amino acid except proline) by the biosynthetic process, and occurs in the endoplasmic reticulum and Golgi apparatus. In order to investigate the influence of human complex-type oligosaccharides on counterpart protein conformation, crambin and ovomucoide, which consist of 46 and 56 amino acid residues, respectively, were selected for synthesis of model glycoproteins. These small glycoproteins were intentionally designed to be glycosylated at the α-helix (crambin: 8?position), β-sheet (crambin: 2?position) and loop position between the antiparallel β-sheets (ovomucoide: 28?position), and were synthesized by using a peptide-segment coupling strategy. After preparation of these glycosylated polypeptide chains, protein folding experiments were performed under redox conditions by using cysteine-cystine. Although the small glycoproteins bearing intentional glycosylation at the α-helix and β-sheet exhibited a suitable folding process, glycosylation at the loop position between the antiparallel β-strands caused multiple products. The conformational differences in the isolated homogeneous glycoproteins compared with non-glycosylated counterparts were evaluated by circular dichroism (CD) and NMR spectroscopy. These analyses suggested that this intentional N-glycosylation did not result in large conformational changes in the purified protein structures, including the case of glycosylation at the loop position between the antiparallel β-strands. In addition to these experiments, the conformational properties of three glycoproteins were evaluated by CD spectroscopy under different temperatures. The oligosaccharides on the protein surface fluctuated considerably; this was dependent on the increase in the solution temperature and was thought to disrupt the protein tertiary structure. Based on the measurement of the CD spectra, however, the glycoproteins bearing three disulfide bonds did not exhibit any change in their protein tertiary structure. These results suggest that the oligosaccharide conformational fluctuations were not disruptive to protein tertiary structure, and the tertiary structure of glycoproteins might be stabilized by the disulfide bond network.     

Performance assessment of different constraining potentials in computational structure prediction for disulfide-bridged proteins

The presence of disulfide bonds in proteins has very important implications on the three-dimensional structure and folding of proteins. An adequate treatment of disulfide bonds in de-novo protein simulations is therefore very important. Here we present a computational study of a set of small disulfide-bridged proteins using an all-atom stochastic search approach and including various constraining potentials to describe the disulfide bonds. The proposed potentials can easily be implemented in any code based on all-atom force fields and employed in simulations to achieve an improved prediction of protein structure. Exploring different potential parameters and comparing the structures to those from unconstrained simulations and to experimental structures by means of a scoring function we demonstrate that the inclusion of constraining potentials improves the quality of final structures significantly. For some proteins (1KVG and 1PG1) the native conformation is visited only in simulations in presence of constraints. Overall, we found that the Morse potential has optimal performance, in particular for the β-sheet proteins.     

Accelerated molecular dynamics simulations of protein folding

Folding of four fast‐folding proteins, including chignolin, Trp‐cage, villin headpiece and WW domain, was simulated via accelerated molecular dynamics (aMD). In comparison with hundred‐of‐microsecond timescale conventional molecular dynamics (cMD) simulations performed on the Anton supercomputer, aMD captured complete folding of the four proteins in significantly shorter simulation time. The folded protein conformations were found within 0.2–2.1 Å of the native NMR or X‐ray crystal structures. Free energy profiles calculated through improved reweighting of the aMD simulations using cumulant expansion to the second‐order are in good agreement with those obtained from cMD simulations. This allows us to identify distinct conformational states (e.g., unfolded and intermediate) other than the native structure and the protein folding energy barriers. Detailed analysis of protein secondary structures and local key residue interactions provided important insights into the protein folding pathways. Furthermore, the selections of force fields and aMD simulation parameters are discussed in detail. Our work shows usefulness and accuracy of aMD in studying protein folding, providing basic references in using aMD in future protein‐folding studies. © 2015 Wiley Periodicals, Inc.