Accounting for a mirror-image conformation as a subtle effect in protein folding

By using local (free-energy profiles along the amino acid sequence and ¹³C^α chemical shifts) and global (principal component) analyses to examine the molecular dynamics of protein-folding trajectories, generated with the coarse-grained united-residue force field, for the B domain of staphylococcal protein A, we are able to (i) provide the main reason for formation of the mirror-image conformation of this protein, namely, a slow formation of the second loop and part of the third helix (Asp29–Asn35), caused by the presence of multiple local conformational states in this portion of the protein; (ii) show that formation of the mirror-image topology is a subtle effect resulting from local interactions; (iii) provide a mechanism for how protein A overcomes the barrier between the metastable mirror-image state and the native state; and (iv) offer a plausible reason to explain why protein A does not remain in the metastable mirror-image state even though the mirror-image and native conformations are at least energetically compatible.To perform their functions in living organisms, most proteins must fold from unfolded polypeptides into their functional, unique 3D structures. Understanding protein-folding mechanisms is crucial because misfolded proteins can cause many diseases, including neurodegenerative diseases (1) such as Alzheimer’s, Parkinson, and Huntington diseases. From theoretical and conceptual points of view, it has been suggested that a native protein exists in a thermodynamically stable state with its surroundings (2) and that a study of free-energy landscapes (FELs) holds the key to understanding how proteins fold and function (3, 4).The native structures of some proteins contain a high degree of symmetry that, in addition to the native structure, allows the existence of another, energetically very close to the native conformation, a native-like “mirror-image” structure. One of the representatives of such symmetrical proteins is the 10- to 55-residue fragment of the B domain of staphylococcal protein A [Protein Data Bank (PDB) ID: 1BDD, a three-α-helix bundle] (5). Protein A has been the subject of extensive theoretical (6–18) and experimental (19–23) studies because of its small size, fast folding kinetics, and biological importance. However, the mirror-image topology has never been a subject for discussion except for the earlier work by Olszewski et al. (7) and recent work by Noel et al. (24). The reason for this might be that it has never been detected experimentally and it was observed only in some theoretical studies (7–9, 12, 13, 15, 17, 18, 24) with different force fields. It is of interest to determine how realistic the mirror-image conformation is. Is it an artifact of the simulations or is it a conformation difficult to observe experimentally? Noel et al. (24) showed that the native and mirror-image structures have a similar enthalpic stability and are thermodynamically competitive and that the mirror image can be considered not just a computational annoyance, but as a real conformation competing with the native structure. Moreover, the mirror-image conformation is more entropically favorable than the native conformation (24). By making multiple mutations in the hydrophobic core and the first loop region, Olszewski et al. (7) found that the change in the handedness of the first loop induced by the mutations, the burial of the N cap of the second helix, and repacking of the hydrophobic core are responsible for formation of the mirror-image conformation. However, at the end, the authors stated: “… Whether the conclusion about the possible importance of turns in defining the global topology holds in general or is just specific to the three-helix bundles analyzed here requires additional investigation....” (ref. 7, p. 298).The difficulties for experiments to detect the mirror-image topology arise because the secondary structures of the mirror-image and the native conformation are identical and the native-contact interactions are similar in both conformations (details in Fig. S1 and SI Native and Mirror-Image Structures of Protein A). Hence, with an experimental technique such as circular dichroism, used to estimate the fraction of secondary-structure content, it is almost impossible to distinguish the mirror-image structure from the native structure. It would have been desirable if the mirror-image conformation and its evolution to the native structure could be detected by NMR spectroscopy. Nevertheless, by using local [¹³C^α chemical shift (25) and free-energy profiles (FEPs) along the amino acid sequence (26–28)] and global [principal component (PC) (29)] analyses (SI Materials and Methods), we examined molecular dynamics (MD) trajectories of protein A, generated with the coarse-grained united-residue (UNRES) force field (27, 30–32) (Fig. S2 and SI Materials and Methods). These analyses of the MD trajectories, in which folding from a fully unfolded conformation occurs either almost instantly or through a metastable state formed by the mirror-image topology, enabled us to elucidate the origin of the formation of a mirror-image topology and how the protein emerges from the kinetic trap and folds to the native state.The results presented in this work are based on the analysis of four pairs of MD trajectories at 270 K (in each pair, one trajectory folds directly to the native state and the other folds through the metastable mirror-image state) selected from 96 MD simulations, which we carried out in a broad range of temperatures (details in Materials and Methods). The mirror-image conformation is energetically competitive with the native conformation in the studied trajectories (an illustrative example of two trajectories is in Fig. S3), and these results are in agreement with those of earlier studies (12, 24).     

Building native protein conformation from highly approximate backbone torsion angles

Reconstructing a protein in three dimensions from its backbone torsion angles is an ongoing challenge because minor inaccuracies in these angles produce major errors in the structure. As a familiar example, a small change in an elbow angle causes a large displacement at the end of your arm, the longer the arm, the larger the displacement. Even accurate knowledge of the backbone torsions and Psi is insufficient, owing to the small, but cumulative, deviations from ideality in backbone planarity, which, if ignored, also lead to major errors in the structure. Against this background, we conducted a computational experiment to assess whether protein conformation can be determined from highly approximate backbone torsion angles, the kind of information that is now obtained readily from NMR. Specifically, backbone torsion angles were taken from proteins of known structure and mapped into 60 degrees x 60 degrees grid squares, called mesostates. Side-chain atoms beyond the beta -carbon were discarded. A mesostate representation of the protein backbone was then used to extract likely candidates from a fragment library of mesostate pentamers, followed by Monte Carlo-based fragment-assembly simulations to identify stable conformations compatible with the given mesostate sequence. Only three simple energy terms were used to gauge stability: molecular compaction, soft-sphere repulsion, and hydrogen bonding. For the six representative proteins described here, stable conformers can be partitioned into a remarkably small number of topologically distinct clusters. Among these, the native topology is found with high frequency and can be identified as the cluster with the most favorable energy.     

Theoretical and experimental demonstration of the importance of specific nonnative interactions in protein folding

Many experimental and theoretical studies have suggested a significant role for nonnative interactions in protein folding pathways, but the energetic contributions of these interactions are not well understood. We have addressed the energetics and the position specificity of nonnative hydrophobic interactions by developing a continuum coarse-grained chain model with a native-centric potential augmented by sequence-dependent hydrophobic interactions. By modeling the effect of different hydrophobicity values at various positions in the Fyn SH3 domain, we predicted energetically significant nonnative interactions that led to acceleration or deceleration of the folding rate depending on whether they were more populated in the transition state or unfolded state. These nonnative contacts were centered on position 53 in the Fyn SH3 domain, which lies in an exposed position in a 3(10)-helix. The energetic importance of the predicted nonnative interactions was confirmed experimentally by folding kinetics studies combined with double mutant thermodynamic cycles. By attaining agreement of theoretical and experimental investigations, this study provides a compelling demonstration that specific nonnative interactions can significantly influence folding energetics. Moreover, we show that a coarse-grained model with a simple consideration of hydrophobicity is sufficient for the accurate prediction of kinetically important nonnative interactions.     

Effect of glycosylation on protein folding: a close look at thermodynamic stabilization

Glycosylation is one of the most common posttranslational modifications to occur in protein biosynthesis, yet its effect on the thermodynamics and kinetics of proteins is poorly understood. A minimalist model based on the native protein topology, in which each amino acid and sugar ring was represented by a single bead, was used to study the effect of glycosylation on protein folding. We studied in silico the folding of 63 engineered SH3 domain variants that had been glycosylated with different numbers of conjugated polysaccharide chains at different sites on the protein's surface. Thermal stabilization of the protein by the polysaccharide chains was observed in proportion to the number of attached chains. Consistent with recent experimental data, the degree of thermal stabilization depended on the position of the glycosylation sites, but only very weakly on the size of the glycans. A thermodynamic analysis showed that the origin of the enhanced protein stabilization by glycosylation is destabilization of the unfolded state rather than stabilization of the folded state. The higher free energy of the unfolded state is enthalpic in origin because the bulky polysaccharide chains force the unfolded ensemble to adopt more extended conformations by prohibiting formation of a residual structure. The thermodynamic stabilization induced by glycosylation is coupled with kinetic stabilization. The effects introduced by the glycans on the biophysical properties of proteins are likely to be relevant to other protein polymeric conjugate systems that regularly occur in the cell as posttranslational modifications or for biotechnological purposes.     

Mapping backbone and side-chain interactions in the transition state of a coupled protein folding and binding reaction

Understanding the mechanism of protein folding requires a detailed knowledge of the structural properties of the barriers separating unfolded from native conformations. The S-peptide from ribonuclease S forms its α-helical structure only upon binding to the folded S-protein. We characterized the transition state for this binding-induced folding reaction at high resolution by determining the effect of site-specific backbone thioxylation and side-chain modifications on the kinetics and thermodynamics of the reaction, which allows us to monitor formation of backbone hydrogen bonds and side-chain interactions in the transition state. The experiments reveal that α-helical structure in the S-peptide is absent in the transition state of binding. Recognition between the unfolded S-peptide and the S-protein is mediated by loosely packed hydrophobic side-chain interactions in two well defined regions on the S-peptide. Close packing and helix formation occurs rapidly after binding. Introducing hydrophobic residues at positions outside the recognition region can drastically slow down association.     

Alpha-helical stabilization by side chain shielding of backbone hydrogen bonds

We study atomic models of the thermodynamics of the structural transition of peptides that form alpha-helices. The effect of sequence variation on alpha-helix formation for alanine-rich peptides, Ac-Ala21-methyl amide (A21) and Ac-A5 (AAARA)3A-methyl amide (Fs peptide), is investigated by atomic simulation studies of the thermodynamics of the helix-coil transition in explicit water. The simulations show that the guanidinium group in the Arg side chains in the Fs peptide interacts with the carbonyl group four amino acids upstream in the chain and desolvates backbone hydrogen bonds. This desolvation can be directly correlated with a higher probability of hydrogen bond formation. We find that Fs has higher helical content than A21 at all temperatures. A small modification in the amber force field reproduces the experimental helical content and helix-coil transition temperatures for the Fs peptide.     

The molecular structure of spider dragline silk: folding and orientation of the protein backbone

The design principles of spider dragline silk, nature's high-performance fiber, are still largely unknown, in particular for the noncrystalline glycine-rich domains, which form the bulk of the material. Here we apply two-dimensional solid-state NMR to determine the distribution of the backbone torsion angles (phi,psi) as well as the orientation of the polypeptide backbone toward the fiber at both the glycine and alanine residues. Instead of an "amorphous matrix," suggested earlier for the glycine-rich domains, these new data indicate that all domains in dragline silk have a preferred secondary structure and are strongly oriented, with the chains predominantly parallel to the fiber. As proposed previously, the alanine residues are predominantly found in a beta sheet conformation. The glycine residues are partly incorporated into the beta sheets and otherwise form helical structures with an approximate 3-fold symmetry.     

Unspecific hydrophobic stabilization of folding transition states

Here we present a method for determining the inference of non-native conformations in the folding of a small domain, alpha-spectrin Src homology 3 domain. This method relies on the preservation of all native interactions after Tyr/Phe exchanges in solvent-exposed, contact-free positions. Minor changes in solvent exposure and free energy of the denatured ensemble are in agreement with the reverse hydrophobic effect, as the Tyr/Phe mutations slightly change the polypeptide hydrophilic/hydrophobic balance. Interestingly, more important Gibbs energy variations are observed in the transition state ensemble (TSE). Considering the small changes induced by the H/OH replacements, the observed energy variations in the TSE are rather notable, but of a magnitude that would remain undetected under regular mutations that alter the folded structure free energy. Hydrophobic residues outside of the folding nucleus contribute to the stability of the TSE in an unspecific nonlinear manner, producing a significant acceleration of both unfolding and refolding rates, with little effect on stability. These results suggest that sectors of the protein transiently reside in non-native areas of the landscape during folding, with implications in the reading of phi values from protein engineering experiments. Contrary to previous proposals, the principle that emerges is that non-native contacts, or conformations, could be beneficial in evolution and design of some fast folding proteins.     

Kinetic role of nonnative species in the folding of bovine pancreatic trypsin inhibitor.

We have shown previously that during the oxidative folding of bovine pancreatic trypsin inhibitor only intermediates with native disulfide bonds are well populated. Nevertheless, these studies also confirmed the earlier conclusion [Creighton, T. E. (1977) J. Mol. Biol. 113, 275-293] that the rate-limiting transition in the kinetically preferred route for folding involves intramolecular disulfide bond rearrangements. Consequently, intermediates with nonnative disulfide bonds must form transiently during folding. Two specific nonnative species, denoted [30-51; 5-14] and [30-51; 5-38], in which numbers indicate residues participating in a disulfide bond, can be detected at low levels in kinetic folding experiments with bovine pancreatic trypsin inhibitor. By working with purified reversibly trapped intermediates, the role of these two nonnative species has been examined directly. These species are found to be in relatively rapid exchange with each other and with an initially formed native two-disulfide intermediate [30-51; 14-38]. Thus, the low abundance of the two nonnative species detected in kinetic folding experiments reflects primarily their low thermodynamic stability as compared to this native intermediate. To a small extent, these nonnative species form the productive native intermediate [30-51; 5-55], which is the immediate precursor to the native protein. However, an equal amount of [5-55; 14-38], a nonproductive dead-end intermediate, is also produced. Thus, the nonnative species detected during the folding of bovine pancreatic trypsin inhibitor are not committed to forming the productive native intermediate, nor do they serve to direct folding specifically toward a productive route.     

Origin of the neighboring residue effect on peptide backbone conformation

Unfolded peptides in water have some residual structure that may be important in the folding process, and the nature of the residual structure is currently of much interest. There is a neighboring residue effect on backbone conformation, discovered in 1997 from measurements of (3)J(HN alpha) coupling constants. The neighboring residue effect appears also in the "coil library" of Protein Data Bank structures of residues not in alpha-helix and not in beta-structure. When a neighboring residue (i - 1 or i + 1) belongs to class L (aromatic and beta-branched amino acids, FHITVWY) rather than class S (all others, G and P excluded), then the backbone angle of residue i is more negative for essentially all amino acids. Calculated values of peptide solvation (electrostatic solvation free energy, ESF) predict basic properties of the neighboring residue effect. We show that L amino acids reduce the solvation of neighboring peptide groups more than S amino acids. When tripeptides from the coil library are excised to allow solvation, the central residues have more negative values of but less negative values of with L than with S neighbors. The coil library values of <(3)J(HN alpha)>, which vary strikingly among the amino acids, are correlated with the neighboring residue effect seen by ESF. Moreover, values for the "blocking effect" of side chains on the hydrogen exchange rates of peptide NH protons are correlated with ESF values.     

Conformation and orientation of a protein folding intermediate trapped by adsorption

Although adsorption-induced conformational changes of proteins play an essential role during protein adsorption on interfaces, detailed information about these changes is lacking. To further the current understanding of protein adsorption, in this study, the orientation, conformation, and local stability of bovine alpha-lactalbumin (BLA) adsorbed on polystyrene nanospheres is characterized at the residue level by hydrogen/deuterium exchange and 2D NMR spectroscopy. Most of the adsorbed BLA molecules have conformational properties similar to BLA molecules in the acid-induced molten globule state (A state). A folding intermediate of BLA is thus induced and trapped by adsorption of the protein on the hydrophobic interface. Several residues, clustered on one side of the adsorbed folding intermediate of BLA, have altered amide proton exchange protection factors compared to those of the A state of BLA. This side preferentially interacts with the interface and includes residues in helix C, the calcium binding site, and part of the beta-domain. Local unfolding of this interacting part of the adsorbed protein seems to initiate the adsorption-induced unfolding of BLA. Adsorption-induced protein unfolding apparently resembles more the mechanical unfolding of a protein than the global unfolding of a protein as induced by denaturant, pH, or pressure. 2D macromolecular crowding prevented the minority of adsorbed BLA molecules, which arrived late at the interface, to unfold to the A state. Protein adsorption is a novel and challenging approach to probe features of the free energy landscapes accessible to unfolding proteins.     

Revisiting the contribution of negative charges on the chaperonin cage wall to the acceleration of protein folding

Chaperonin GroEL mediates the folding of protein encapsulated in a GroES-sealed cavity (cage). Recently, a critical role of negative charge clusters on the cage wall in folding acceleration was proposed based on experiments using GroEL single-ring (SR) mutants SR1 and SRKKK2 [Tang YC, et al. (2006) Cell 125:903–914; Chakraborty K, et al. (2010) Cell 142:112–122]. Here, we revisited these experiments and discovered several inconsistencies. (i) SR1 was assumed to bind to GroES stably and to mediate single-round folding in the cage. However, we show that SR1 repeats multiple turnovers of GroES release/binding coupled with ATP hydrolysis. (ii) Although the slow folding observed for a double-mutant of maltose binding protein (DMMBP) by SRKKK2 was attributed to mutations that neutralize negative charges on the cage wall, we found that the majority of DMMBP escape from SRKKK2 and undergo spontaneous folding in the bulk medium. (iii) An osmolyte, trimethylamine N-oxide, was reported to accelerate SRKKK2-mediated folding of DMMBP by mimicking the effect of cage-wall negative charges of WT GroEL and ordering the water structure to promote protein compaction. However, we demonstrate that in-cage folding by SRKKK2 is unaffected by trimethylamine N-oxide. (iv) Although it was reported that SRKKK2 lost the ability to assist the folding of ribulose-1,5-bisphosphate carboxylase/oxygenase, we found that SRKKK2 retains this ability. Our results argue against the role of the negative charges on the cage wall of GroEL in protein folding. Thus, in chaperonin studies, folding kinetics need to be determined from the fraction of the real in-cage folding.     

Algorithm discovery by protein folding game players

Foldit is a multiplayer online game in which players collaborate and compete to create accurate protein structure models. For specific hard problems, Foldit player solutions can in some cases outperform state-of-the-art computational methods. However, very little is known about how collaborative gameplay produces these results and whether Foldit player strategies can be formalized and structured so that they can be used by computers. To determine whether high performing player strategies could be collectively codified, we augmented the Foldit gameplay mechanics with tools for players to encode their folding strategies as "recipes" and to share their recipes with other players, who are able to further modify and redistribute them. Here we describe the rapid social evolution of player-developed folding algorithms that took place in the year following the introduction of these tools. Players developed over 5,400 different recipes, both by creating new algorithms and by modifying and recombining successful recipes developed by other players. The most successful recipes rapidly spread through the Foldit player population, and two of the recipes became particularly dominant. Examination of the algorithms encoded in these two recipes revealed a striking similarity to an unpublished algorithm developed by scientists over the same period. Benchmark calculations show that the new algorithm independently discovered by scientists and by Foldit players outperforms previously published methods. Thus, online scientific game frameworks have the potential not only to solve hard scientific problems, but also to discover and formalize effective new strategies and algorithms.     

Engineering stable peptide toxins by means of backbone cyclization: stabilization of the alpha-conotoxin MII

Conotoxins (CTXs), with their exquisite specificity and potency, have recently created much excitement as drug leads. However, like most peptides, their beneficial activities may potentially be undermined by susceptibility to proteolysis in vivo. By cyclizing the alpha-CTX MII by using a range of linkers, we have engineered peptides that preserve their full activity but have greatly improved resistance to proteolytic degradation. The cyclic MII analogue containing a seven-residue linker joining the N and C termini was as active and selective as the native peptide for native and recombinant neuronal nicotinic acetylcholine receptor subtypes present in bovine chromaffin cells and expressed in Xenopus oocytes, respectively. Furthermore, its resistance to proteolysis against a specific protease and in human plasma was significantly improved. More generally, to our knowledge, this report is the first on the cyclization of disulfide-rich toxins. Cyclization strategies represent an approach for stabilizing bioactive peptides while keeping their full potencies and should boost applications of peptide-based drugs in human medicine.     

High-resolution protein folding with a transferable potential

A generalized computational method for folding proteins with a fully transferable potential and geometrically realistic all-atom model is presented and tested on seven helix bundle proteins. The protocol, which includes graph-theoretical analysis of the ensemble of resulting folded conformations, was systematically applied and consistently produced structure predictions of approximately 3 A without any knowledge of the native state. To measure and understand the significance of the results, extensive control simulations were conducted. Graph theoretic analysis provides a means for systematically identifying the native fold and provides physical insight, conceptually linking the results to modern theoretical views of protein folding. In addition to presenting a method for prediction of structure and folding mechanism, our model suggests that an accurate all-atom amino acid representation coupled with a physically reasonable atomic interaction potential and hydrogen bonding are essential features for a realistic protein model.     

The folding of GroEL-bound barnase as a model for chaperonin-mediated protein folding.

We have analyzed the pathway of folding of barnase bound to GroEL to resolve the controversy of whether proteins can fold while bound to chaperonins (GroEL or Cpn60) or fold only after their release into solution. Four phases in the folding were detected by rapid-reaction kinetic measurements of the intrinsic fluorescence of both wild type and barnase mutants. The phases were assigned from their rate laws, sensitivity to mutations, and correspondence to regain of catalytic activity. At high ratios of denatured barnase to GroEL, 4 mol of barnase rapidly bind per 14-mer of GroEL. At high ratios of GroEL to barnase, 1 mol of barnase binds with a rate constant of 3.5 x 10(7) s-1.M-1. This molecule then refolds with a low rate constant that changes on mutation in parallel with the rate constant for the folding in solution. This rate constant corresponds to the regain of the overall catalytic activity of barnase and increases 15-fold on the addition of ATP to a physiologically relevant value of approximately 0.4 s-1. The multiply bound molecules of barnase that are present at high ratios of GroEL to barnase fold with a rate constant that is also sensitive to mutation but is 10 times higher. If the 110-residue barnase can fold when bound to GroEL and many moles can bind simultaneously, then smaller parts of large proteins should be able to fold while bound.     

Sequential folding of a bifunctional allosteric protein.

Aspartokinase I-homoserine dehydrogenase I (EC 2.7.2.4 and EC 1.1.1.3) a bifunctional and allosteric enzyme, has been renatured from its unfolded and separated polypeptide chains. Folding was measured by the reappearance of each of the two enzymatic activities, kinase and dehydrogenase, and of their allosteric inhibition by the same effector, threonine. The various observed properties yield different kinetics of folding, which shows the presence of intermediates having only some of the functional features of the native enzyme. Apparently, three successive steps can be detected during the folding of aspartokinase I-homoserine dehydrogenase I: first, a monomolecular step leads to a monomeric species with the kinase activity; then an association step leads to a dimeric species with the kinase and dehydrogenase activities, and a threonine-sensitive dehydrogenase; finally, a second association step leads to a tetrameric species with the two activities, both sensitive to threonine. The folding of this large protein appears as a sequential process during which the functional properties are regained successively, as the protein structure becomes more complex. During this process, the two regions of each polypeptide chain respectively responsible for the kinase and dehydrogenase activities seem to acquire their native conformation rather independently of each other.     

Structure and folding of a designed knotted protein

A very small number of natural proteins have folded configurations in which the polypeptide backbone is knotted. Relatively little is known about the folding energy landscapes of such proteins, or how they have evolved. We explore those questions here by designing a unique knotted protein structure. Biophysical characterization and X-ray crystal structure determination show that the designed protein folds to the intended configuration, tying itself in a knot in the process, and that it folds reversibly. The protein folds to its native, knotted configuration approximately 20 times more slowly than a control protein, which was designed to have a similar tertiary structure but to be unknotted. Preliminary kinetic experiments suggest a complicated folding mechanism, providing opportunities for further characterization. The findings illustrate a situation where a protein is able to successfully traverse a complex folding energy landscape, though the amino acid sequence of the protein has not been subjected to evolutionary pressure for that ability. The success of the design strategy--connecting two monomers of an intertwined homodimer into a single protein chain--supports a model for evolution of knotted structures via gene duplication.     

Fast folding of a helical protein initiated by the collision of unstructured chains

To examine whether helix formation necessarily precedes chain collision, we have measured the folding of a fully helical coiled coil that has been specially engineered to have negligible intrinsic helical propensity but high overall stability. The folding rate approaches the diffusion-limited value and is much faster than possible if folding is contingent on precollision helix formation. Therefore, the collision of two unstructured chains is the initial step of the dominant kinetic pathway, whereas helicity exerts its influence only at a later step. Folding from an unstructured encounter complex may be efficient and robust, which has implications for any biological process that couples folding to binding.