Is protein folding hierarchic? II. Folding intermediates and transition states

Unfolding of the loggerhead sea turtle (Caretta caretta) myoglobin: A (1)H-NMR and electronic absorbance study

The effect of urea concentration on the backbone solution structure of the cyanide derivative of ferric Caretta caretta myoglobin (at pH 5.4) is reported. By addition of urea, sequential and long-range nuclear Overhauser effects (NOEs) are gradually lost. By using the residual NOE constraints to build the molecular model, a picture of the unfolding pathway was obtained. When the urea concentration is raised to 2.2 M, helices A and B appear largely disordered; helices C, D, and F loose structural constraints at 3.0 M urea. At urea concentration >6 M, the protein appears to be fully unfolded, including the GH hairpin and helix E stabilizing the prosthetic group. Reversible and cooperative denaturation isotherms obtained by following NOE peaks are considerably different from those obtained by monitoring electronic absorption changes. The reversible and cooperative urea-dependent folding-unfolding process of C. caretta myoglobin follows the minimum three-state mechanism N long left and right arrow X long left and right arrow D, where X represents a disordered globin structure (occurring at approximately 4 M urea) that still binds the heme.

The search for local native-like nucleation centers in the unfolded state of beta -sheet proteins

An approach involving the systematic computational conformational analysis of all overlapping hexapeptide segments in the protein sequence has found fragments with the higher than average propensity to adopt the native-like three-dimensional structure and other regular nonrandom structures in the unfolded states of four beta-sheet proteins, namely IFABP (intestinal fatty acid-binding protein), ILBP (ileal fatty acid-binding protein), CRABP I (cellular retinoic acid-binding protein), and CRBP II (cellular retinal binding protein). The native three-dimensional structures of these four proteins are very similar even though they possess as little as approximately 30% sequence similarity. The computational results were validated by comparison with the experimental data of the heteronuclear sequential quantum correlation NMR spectroscopy obtained earlier for IFABP at high urea concentrations. On this basis, a molecular model of the unfolded state of IFABP has been developed. The model presumes a dynamic equilibrium between various nonrandom structures (including the native-like structure) and random coil in the local segments of the protein sequence. The model explains experimental observations obtained earlier for folding of several mutants of IFABP, as well as the observed differences in molecular mechanisms of folding for the four beta-sheet proteins. Because the computational approach itself does not employ any experimentally derived information in advance, it is not necessarily limited to the beta-sheet proteins.

Contact order and ab initio protein structure prediction

Although much of the motivation for experimental studies of protein folding is to obtain insights for improving protein structure prediction, there has been relatively little connection between experimental protein folding studies and computational structural prediction work in recent years. In the present study, we show that the relationship between protein folding rates and the contact order (CO) of the native structure has implications for ab initio protein structure prediction. Rosetta ab initio folding simulations produce a dearth of high CO structures and an excess of low CO structures, as expected if the computer simulations mimic to some extent the actual folding process. Consistent with this, the majority of failures in ab initio prediction in the CASP4 (critical assessment of structure prediction) experiment involved high CO structures likely to fold much more slowly than the lower CO structures for which reasonable predictions were made. This bias against high CO structures can be partially alleviated by performing large numbers of additional simulations, selecting out the higher CO structures, and eliminating the very low CO structures; this leads to a modest improvement in prediction quality. More significant improvements in predictions for proteins with complex topologies may be possible following significant increases in high-performance computing power, which will be required for thoroughly sampling high CO conformations (high CO proteins can take six orders of magnitude longer to fold than low CO proteins). Importantly for such a strategy, simulations performed for high CO structures converge much less strongly than those for low CO structures, and hence, lack of simulation convergence can indicate the need for improved sampling of high CO conformations. The parallels between Rosetta simulations and folding in vivo may extend to misfolding: The very low CO structures that accumulate in Rosetta simulations consist primarily of local up-down beta-sheets that may resemble precursors to amyloid formation.

Pathway heterogeneity in protein folding

We generate ab initio folding pathways in two single-domain proteins, hyperthermophile variant of protein G domain (1gb4) and ubiquitin (1ubi), both presumed to be two-state folders. Both proteins are endowed with the same topology but, as shown in this work, rely to a different extent on large-scale context to find their native folds. First, we demonstrate a generic feature of two-state folders: A downsizing of structural fluctuations is achieved only when the protein reaches a stationary plateau maximizing the number of highly protected hydrogen bonds. This enables us to identify the folding nucleus and show that folding does not become expeditious until a topology is generated that is able to protect intramolecular hydrogen bonds from water attack. Pathway heterogeneity is shown to be dependent on the extent to which the protein relies on large-scale context to fold, rather than on contact order: Proteins that can only stabilize native secondary structure by packing it against scaffolding hydrophobic moieties are meant to have a heterogeneous transition-state ensemble if they are to become successful folders (otherwise, successful folding would be too fortuitous an event.) We estimate mutational Phi values as ensemble averages and deconvolute individual-route contributions to the averaged two-state kinetic picture. Our results find experimental corroboration in the well-studied chymotrypsin inhibitor (CI2), while leading to verifiable predictions for the other two study cases.

The natively helical chain segment 169-188 of Escherichia coli adenylate kinase is formed in the latest phase of the refolding transition

The refolding transition of Escherichia coli adenylate kinase (AK) was investigated by monitoring the refolding kinetics of a selected 20 residue helical segment in the CORE domain of the protein. Residues 169 and 188 were labeled by 1-acetamido-methyl-pyrene, and by bimane, respectively. The experiment combines double-jump stopped-flow fast mixing initiation of refolding and time-resolved Förster energy transfer spectroscopy for monitoring the conformational transitions (double-kinetics experiment). Two kinetic phases were found in the denaturant-induced unfolding of AK. In the first phase, the fluorescence quantum yields of both probes decreased. The distribution of the distances between them transformed from the native state's narrow distribution with the mean distance corresponding to the distance in the crystal structure, to a distribution compatible with an unordered structure. In the second, slow step of denaturation, neither the fluorescence parameters of the probes nor the distance distribution between them changed. This step appeared to be a transformation of the fast-folding species formed in the first phase, to the slow-folding species. Refolding of the fast-folding species of the denatured state of AK was also a two-phase process. During the first fast phase, within less than 5ms, the fluorescence emission of both probes increased, but the distance distribution between the labeled sites was unchanged. Only during the second slow refolding step did the intramolecular distance distribution change from the characteristic of the denatured state to the narrow distribution of the native state. This experiment shows that for the case of the CORE domain of AK, the large helical segment of residues 169-188 was not formed in the first compaction step of refolding. The helical conformation of this segment is established only in the second, much slower, refolding phase, simultaneously with the completion of the native structure.

Comparison of protein fragments identified by limited proteolysis and by computational cutting of proteins

Here we present a comparison between protein fragments produced by limited proteolysis and those identified by computational cutting based on the building block folding model. The principles upon which the two methods are based are different. Limited proteolysis of natively folded proteins occurs at flexible sites and never at the level of chain segments of regular secondary structure such as alpha-helices. Therefore, the targets for limited proteolysis are locally unfolded regions. In contrast, the computational cutting algorithm considers the compactness of the fragments, their nonpolar buried surface area, and their isolatedness, that is, the surface area which was buried prior to the cutting and becomes exposed subsequently. Despite the different criteria, there is an overall correspondence between sites or regions of limited proteolysis with those identified by computational cutting. The computational cutting method has been applied to several model proteins for which detailed limited proteolysis data are available, namely apomyoglobin, cytochrome c, ribonuclease A, alpha-lactalbumin, and thermolysin. As expected, more cuts are obtained computationally than experimentally and the agreement is better when a number of proteolytic enzymes are used. For example, cytochrome c is cleaved by thermolysin at 56-57, 45-46, and at 80-81, and by proteinase K at 48-49 and 50-51. Incubation of the noncovalent and native-like complex of cytochrome c fragments 1-56 and 57-104 with proteinase K yielded the gapped protein species 1-48/57-104 and finally 1-40/57-104. Computational cutting of cytochrome c reproduced the major experimental observations, with cuts at 47, 64-65 or 65-66 and 80-81 and an unstable 32-47 region not assigned to any building block. The next step, not addressed in this work, is to probe the ability of the generated fragments to fold independently. Since both the computational algorithm and limited proteolysis attempt to dissect the protein folding problem, the general agreement between the two procedures is gratifying. This consistency allows us to propose the use of limited proteolysis to produce protein fragments that can adopt an independent folding and, therefore, to study folding intermediates. The results of the present study appear to validate the building block folding model and are in line with the proposal that protein folding is a hierarchical process, where parts constituting local minima of energy fold first, with their subsequent association and mutual stabilization to finally yield the global fold.

How do we probe ubiquitin's pathway heterogeneity?

We identify folding pathways for ubiquitin and assess its extent of transition state (TS) heterogeneity using a kinetically controlled ab initio algorithm that generates a coarse-grained description of torsional dynamics. The algorithm computes the time evolution of backbone-motion constraints, finds optimized conformations within such constraints, evaluates local solvent environments, and rescales accordingly the energetic contributions to determine the transition to the next set of torsional constraints. Native and nonnative structural features are found in the TS ensemble determined from a pool of 72 successful runs whose final folds are within 4-5 A RMSD from native. Certain nonnative features at the TS are shown to be necessary to create a large-scale context that overrides local propensities. Such misfolds undergo a subsequent rearrangement on the downhill side of the energy profile. The effects of tunable bi-histidine metal-binding sites, point mutations, negative Phi-value mutations, and denaturant on kinetics are predicted.

Principles of docking: An overview of search algorithms and a guide to scoring functions

The docking field has come of age. The time is ripe to present the principles of docking, reviewing the current state of the field. Two reasons are largely responsible for the maturity of the computational docking area. First, the early optimism that the very presence of the "correct" native conformation within the list of predicted docked conformations signals a near solution to the docking problem, has been replaced by the stark realization of the extreme difficulty of the next scoring/ranking step. Second, in the last couple of years more realistic approaches to handling molecular flexibility in docking schemes have emerged. As in folding, these derive from concepts abstracted from statistical mechanics, namely, populations. Docking and folding are interrelated. From the purely physical standpoint, binding and folding are analogous processes, with similar underlying principles. Computationally, the tools developed for docking will be tremendously useful for folding. For large, multidomain proteins, domain docking is probably the only rational way, mimicking the hierarchical nature of protein folding. The complexity of the problem is huge. Here we divide the computational docking problem into its two separate components. As in folding, solving the docking problem involves efficient search (and matching) algorithms, which cover the relevant conformational space, and selective scoring functions, which are both efficient and effectively discriminate between native and non-native solutions. It is universally recognized that docking of drugs is immensely important. However, protein-protein docking is equally so, relating to recognition, cellular pathways, and macromolecular assemblies. Proteins function when they are bound to other molecules. Consequently, we present the review from both the computational and the biological points of view. Although large, it covers only partially the extensive body of literature, relating to small (drug) and to large protein-protein molecule docking, to rigid and to flexible. Unfortunately, when reviewing these, a major difficulty in assessing the results is the non-uniformity in the formats in which they are presented in the literature. Consequently, we further propose a way to rectify it here.

Structural restoration of inactive recombinant fish growth hormones by chemical chaperonin and solvent restraint approaches

Recombinant proteins may undergo conformational distortion, leading to aggregation and loss of function, when they are expressed in heterologous systems. The structural and functional restoration of such inactive proteins is highly desirable. We have over-expressed recombinant growth hormones from the fish ayu (Plecoglossus altivelis) and yellow grouper (Epinephelus awoara) by a pET expression system. Both recombinant proteins accumulate as insoluble form in Escherichia coli. We refolded these inactive proteins into the active form using a stepwise refolding process with a dilute denaturing agent as a steric blocker and chemical chaperonin. Optical characterization showed that stable folding intermediates with a helical conformation can be detected in the molten globule state. Moreover, the function of restored recombinant growth hormones was demonstrated by its ability to stimulate proliferation in zebrafish liver cells.

Application of the diffusion-collision model to the folding of three-helix bundle proteins

The diffusion-collision model has been successful in explaining many features of protein folding kinetics, particularly for helical proteins. In the model the folding reaction is described in terms of coupled chemical kinetic (Master) equations of coarse grained entities, called microdomains. Here, the diffusion-collision model is applied to compute the folding kinetics of four three-helix bundle proteins, all of which fold on a time scale of tens of microseconds and appear to have two-state folding. The native structure and the stability of the helical microdomains are used to determine the parameters of the model. The formulation allows computation of the overall rate and determination of the importance of kinetic intermediates. The proteins considered are the B domain of protein A (1BDC), the Engrailed Homeodomain (1ENH), the peripheral sub-unit-binding domain (1EBD C-chain) and the villin headpiece subdomain (1VII). The results for the folding time of protein A, the Engrailed Homeodomain, and 1EBD C-chain are in agreement with experiment, while 1VII is not stable in the present model. In the three proteins that are stable, two-state folding is predicted by the diffusion-collision model. This disagrees with published assertions that multistate kinetics would be obtained from the model. The contact order prediction agrees with experiment for protein A, but yields values that are a factor of 40, 30 and 15 too slow for 1ENH, 1EBD C-chain and 1VII. The effect of mutants on folding is described for protein A and it is demonstrated that significant intermediate concentrations (i.e. deviation from two-state folding) can occur if the stability of some of the helical microdomains is increased. A linear relationship between folding time and the length of the loop between helices B and C in protein A is demonstrated; this is not evident in the contact order description.

Pathway diversity and concertedness in protein folding: an ab-initio approach

Making use of an ab-initio folding simulator, we generate in vitro pathways leading to the native fold in moderate size single- domain proteins. The assessment of pathway diversity is not biased by any a priori information on the native fold. We focus on two study cases, hyperthermophile variant of protein G domain (1gb4) and ubiquitin (1ubi), with the same topology but different context dependence in their native folds. We demonstrate that a quenching of structural fluctuations is achieved once the proteins find a stationary plateau maximizing the number of highly protected hydrogen bonds. This enables us to identify the folding nucleus and show that folding does not become expeditious until a concerted event takes place generating a topology able to prevent water attack on a maximal number of hydrogen bonds. This result is consistent with the standard nucleation mechanism postulated for two-state folders. Pathway diversity is correlated with the extent of conflict between local structural propensity and large-scale context, rather than with contact order: In highly context-dependent proteins, the success of folding cannot rely on a single fortuitous event in which local propensity is overruled by large-scale effects. We predict mutational Pi values on individual pathways, compute ensemble averages and predict extent of surface burial and percentage of hydrogen bonding on each component of the transition state ensemble, thus deconvoluting individual folding-route contributions to the averaged two-state kinetic picture. Our predicted kinetic isotopic effects find experimental support and lead to further probes. Finally, the molecular redesign potentiality of the method, aimed at increasing folding expediency, is explored.

Compressibility of protein transitions

We review the results of compressibility studies on proteins and low molecular weight compounds that model the hydration properties of these biopolymers. In particular, we present an analysis of compressibility changes accompanying conformational transitions of globular proteins. This analysis, in conjunction with experimental compressibility data on protein transitions, were used to define the changes in the hydration properties and intrinsic packing associated with native-to-molten globule, native-to-partially unfolded, and native-to-fully unfolded transitions of globular proteins. In addition, we discuss the molecular origins of predominantly positive changes in compressibility observed for pressure-induced denaturation transitions of globular proteins. Throughout this review, we emphasize the importance of compressibility data for characterizing protein transitions, while also describing how such data can be interpreted to gain insight into role that hydration and intrinsic packing play in modulating the stability of and recognition between proteins and other biologically important compounds.

Detecting equilibrium cytochrome c folding intermediates by electrospray ionisation mass spectrometry: two partially folded forms populate the molten-globule state

Nanoelectrospray ionization mass spectrometry (nano-ESI-MS) is applied to the characterization of ferric cytochromec (cytc) conformational states under different solvent conditions. The methanol-induced molten-globule state in the pH range 2.6-3.0 is found to be populated by two distinct, partially folded conformers I(A) and I(B). The more compact intermediate I(B) resembles that induced by glycerol in acid-unfolded cytc. The less compact one, I(A), also can be induced by destabilization of the native structure by trifluoroethanol. I(A) and I(B) can be detected, in the absence of additives, around the midpoint of the acid-induced unfolding transition, providing direct evidence for involvement of equilibrium folding intermediates in cytc conformational transitions at low pH. This study shows that mass spectrometry can contribute to the characterization of molten-globule states of proteins by detection of distinct, although poorly populated, conformations involved in a dynamic equilibrium.

Chiral mutagenesis of insulin's hidden receptor-binding surface: structure of an allo-isoleucine(A2) analogue

The hydrophobic core of vertebrate insulins contains an invariant isoleucine residue at position A2. Lack of variation may reflect this side-chain's dual contribution to structure and function: Ile(A2) is proposed both to stabilize the A1-A8 alpha-helix and to contribute to a "hidden" functional surface exposed on receptor binding. Substitution of Ile(A2) by alanine results in segmental unfolding of the A1-A8 alpha-helix, lower thermodynamic stability and impaired receptor binding. Such a spectrum of perturbations, although of biophysical interest, confounds interpretation of structure-activity relationships. To investigate the specific contribution of Ile(A2) to insulin's functional surface, we have employed non-standard mutagenesis: inversion of side-chain chirality in engineered monomer allo-Ile(A2)-DKP-insulin. Although the analogue retains native structure and stability, its affinity for the insulin receptor is impaired by 50-fold. Thus, whereas insulin's core readily accommodates allo-isoleucine at A2, its activity is exquisitely sensitive to chiral inversion. We propose that the Ile(A2) side-chain inserts within a chiral pocket of the receptor as part of insulin's hidden functional surface.

Thermal unfolding molecular dynamics simulation of Escherichia coli dihydrofolate reductase: thermal stability of protein domains and unfolding pathway

Temperature induced unfolding of Escherichia coli dihydrofolate reductase was carried out by using molecular dynamic simulations. The simulations show that the unfolding generally involves an initial end-to-end collapse of the adenine binding domain into partially extended loops, followed by a gradual breakdown of the remaining beta sheet core structure. The core, which consists of beta strands 5-7, was observed to be the most resistant to thermal unfolding. This region, which is made up of part of the N terminus domain and part of the large domain of the E. coli dihydrofolate reductase, may constitute the nucleation site for protein folding and may be important for the eventual formation of both domains. The unfolding of different domains at different stages of the unfolding process suggests that protein domains vary in stability and that the rate at which they unfold can affect the overall outcome of the unfolding pathway. This observation is compared with the recently proposed hierarchical folding model. Finally, the results of the simulation were found to be consistent with a previous experimental study (Frieden, Proc Natl Acad Sci USA 1990;87:4413-4416) which showed that the folding process of E. coli dihydrofolate reductase involves sequential formation of the substrate binding sites.

Uncoupled analysis of secondary and tertiary protein structure by circular dichroism and electrospray ionization mass spectrometry

Electrospray ionization mass spectrometry (ESI-MS) applied to protein conformational studies is a powerful new method that seems to provide specific information about protein tertiary structure. In this study, we analyzed the effect of trifluoroethanol (TFE) on a myoglobin peptide and cytochrome c (cyt c) at low pH by circular dichroism (CD) and ESI-MS. These experiments show that coil-to-helix transition per se does not affect ESI mass spectra, confirming that this technique is insensitive to the local conformation of the polypeptidic chain and, rather, reports on the tertiary contacts characterizing different protein conformations. This property makes ESI-MS an excellent method, complementary to CD, for the characterization of protein conformational changes. Fluorinated alcohols have been suggested to induce molten globule formation in acid-unfolded cyt c. The experiments described here show that TFE does not induce major changes in the ESI mass spectrum of cyt c at pH 2.2, indicating that no stabilization of compact, globular structures is detectable under the conditions employed. On the other hand, even low concentrations of TFE (2-5%) are shown to destabilize the folded state of the protein around the mid-point of its acid-induced unfolding transition.

Ab initio prediction of helical segments in polypeptides

An ab initio method has been developed to predict helix formation for polypeptides. The approach relies on the systematic analysis of overlapping oligopeptides to determine the helical propensity for individual residues. Detailed atomistic level modeling, including entropic contributions, and solvation/ionization energies calculated through the solution of the Poisson-Boltzmann equation, is utilized. The calculation of probabilities for helix formation is based on the generation of ensembles of low energy conformers. The approach, which is easily amenable to parallelization, is shown to perform very well for several benchmark polypeptide systems, including the bovine pancreatic trypsin inhibitor, the immunoglobulin binding domain of protein G, the chymotrypsin inhibitor 2, the R69 N-terminal domain of phage 434 repressor, and the wheat germ agglutinin.

Exploring the folding landscape of a structured RNA

Structured RNAs achieve their active states by traversing complex, multidimensional energetic landscapes. Here we probe the folding landscape of the Tetrahymena ribozyme by using a powerful approach: the folding of single ribozyme molecules is followed beginning from distinct regions of the folding landscape. The experiments, combined with small-angle x-ray scattering results, show that the landscape contains discrete folding pathways. These pathways are separated by large free-energy barriers that prevent interconversion between them, indicating that the pathways lie in deep channels in the folding landscape. Chemical protection and mutagenesis experiments are then used to elucidate the structural features that determine which folding pathway is followed. Strikingly, a specific long-range tertiary contact can either help folding or hinder folding, depending on when it is formed during the process. Together these results provide an unprecedented view of the topology of an RNA folding landscape and the RNA structural features that underlie this multidimensional landscape.

Structural transitions in neutral and charged proteins in vacuo

In vacuo proteins provide a simple laboratory to explore the roles of sequence, temperature, charge state, and initial configuration in protein folding. Moreover, by the very absence of solvent, the study of anhydrous proteins in vacuo will also help us to understand specific environmental effects. From the experimental viewpoint, these systems are now beginning to be characterized at low resolution. Molecular dynamics (MD) simulations, in combination with tools for protein shape analysis, can complement experiments and provide further insights on the folding-unfolding transitions of these proteins. We review some aspects of this issue by using the results from a detailed MD study of hen egg-white lysozyme. For lysozyme ions, unfolding can be triggered by Coulombic repulsion. In neutral lysozyme, unfolding can be induced by centrifugal forces and also by weakening the monomer-monomer interaction. In both cases, the resulting unfolded transients can be used as initial configurations for relaxation dynamics. All trajectories are analyzed in terms of global molecular shape features of the backbone, including its anisometry and chain entanglement complexity. This strategy allows us to quantify separately the degree of polymer collapse and the evolution of large-scale folding features. Using these last two notions, we discuss some basic questions regarding the nature of the accessible paths associated with unfolding from, and refolding into, compact conformers.

Analysis of multiple folding routes of proteins by a coarse-grained dynamics model

Langevin dynamics of a protein molecule with Go-type potentials is developed and used to analyze long time-scale events in the folding of cytochrome c. Several trajectories are generated, starting from random coil configurations and going to the native state, that are a few angstroms root mean square deviation (RMSD) from the native structure. The dynamics is controlled, to a large scale, by the two terminal helices that are in contact in the native state. These two helices form very early during folding, and depending on the trajectory, they either stabilize rapidly or break and re-form in going over steric barriers. The extended initial chain exhibits a rapid folding transition into a relatively compact shape, after which the helices are reorganized in a highly correlated manner. The time of formation of residue pair contacts strongly points to the hierarchical nature of folding; i.e., secondary structure forms first, followed by rearrangements of larger length scales at longer times. The kinetics of formation of native contacts is also analyzed, and the onset of a stable globular configuration, referred to as the molten globule in the literature, is identified. Predictions of the model are compared with extensive experimental data on cytochrome c.

Coordination topology and stability for the native and binding conformers of chymotrypsin inhibitor 2

We demonstrate that the stabilization of the binding region is accomplished at the expense of a loss in the stability of the rest of the protein. A novel molecular mechanics (MM) approach is introduced to distinguish residue stabilities of proteins in a given conformation. As an example, the relative stabilities of folded chymotrypsin inhibitor 2 (CI2) in unbound form, and CI2 in complex with subtilisin novo is investigated. The conformation of the molecule in the two states is almost identical, with an approximately 0.6-A root-mean-square deviation (RMSD) of the Calpha atoms. On binding, the packing density changes only at the binding loop. However, residue fluctuations in the rest of the protein are greatly altered solely due to those contacts, indicating the effective propagation of perturbation and the presence of remotely controlling residues. To quantify the interplay between packing density, packing order, residue fluctuations, and residue stability, we adopt an MM approach whereby small displacements are inserted at selected residues, followed by energy minimization; the displacement of each residue in response to such perturbations are organized in a perturbation-response matrix L. We define residue stability lambda(i) = summation operator((j)L(ij))/ summation operator((j) L(ji)) as the ratio of the amount of change to which the residue is amenable, to the ability of a given residue to induce change. We then define the free energy associated with residue stability, DeltaG(lambda) = -RT ln lambda. DeltaG(lambda) intrinsically selects the residues that are in the folding core. Upon complexation, the binding loop becomes more resistant to perturbation, in contrast to the alpha-helix that favors change. Although the two forms of CI2 are structurally similar, residue fluctuations differ vastly, and the stability of many residues is altered upon binding. The decrease in entropy introduced by binding is thus compensated by these changes.

Dissociation and unfolding of GCN4 leucine zipper in the presence of sodium dodecyl sulfate

The dissociation and unfolding behavior of the GCN4 leucine zipper has been studied using SDS titration. Circular dichroism (CD) spectra showed that the alpha-helix content of the leucine zipper (20 microM) decreased during the sodium dodecyl sulfate (SDS) titration. However, the alpha-helix content of the leucine zipper still remained significant in the presence of 1 mM SDS, with little change detected when the SDS concentration further increased to 2 mM. The dimer dissociation of the leucine zipper is also a co-operative process during SDS titration; with no dimer remaining when SDS concentration reached 1 mM, as shown by electrophoresis and the the theta(222)/theta(208) ratio. Our results indicate that SDS efficiently induces leucine zipper dimer dissociation with the monomers still partially folded. The experimental results provide important evidence for the previous model that partial helix formation precedes dimerization in coiled coil folding.

Dynamic regimes and correlated structural dynamics in native and denatured alpha-lactalbumin

Understanding the mechanisms of protein folding requires knowledge of both the energy landscape and the structural dynamics of a protein. We report a neutron-scattering study of the nanosecond and picosecond dynamics of native and the denatured alpha-lactalbumin. The quasielastic scattering intensity shows that there are alpha-helical structure and tertiary-like side-chain interactions fluctuating on sub-nanosecond time-scales under extremely denaturing conditions and even in the absence of disulfide bonds. Based on the length-scale dependence of the decay rate of the measured correlation functions, the nanosecond dynamics of the native and the variously denatured proteins have three dynamic regimes. When 0.05<Q<0.5 A(-1) (where the scattering vector, Q, is inversely proportional to the length-scale), the decay rate, Gamma, shows a power law relationship, Gamma proportional to Q(2.42+/-0.08), that is analogous to the dynamic behavior of a random coil. However, when 0.5<Q<1.0 A(-1), the decay rate exhibits a Gamma proportional to Q(1.0+/-0.2) relationship. The effective diffusion constant of the protein decreases with increasing Q, a striking dynamic behavior that is not found in any chain-like macromolecule. We suggest that this unusual dynamics is due to the presence of a strongly attractive force and collective conformational fluctuations in both the native and the denatured states of the protein. Above Q>1.0 A(-1) is a regime that displays the local dynamic behavior of individual residues, Gamma proportional to Q(1.8+/-0.3). The picosecond time-scale dynamics shows that the potential barrier to side-chain proton jump motion is reduced in the molten globule and in the denatured proteins when compared to that of the native protein. Our results provide a dynamic view of the native-like topology established in the early stages of protein folding.

Protein folding theory: from lattice to all-atom models

This review focuses on recent advances in understanding protein folding kinetics in the context of nucleation theory. We present basic concepts such as nucleation, folding nucleus, and transition state ensemble and then discuss recent advances and challenges in theoretical understanding of several key aspects of protein folding kinetics. We cover recent topology-based approaches as well as evolutionary studies and molecular dynamics approaches to determine protein folding nucleus and analyze other aspects of folding kinetics. Finally, we briefly discuss successful all-atom Monte-Carlo simulations of protein folding and conclude with a brief outlook for the future.

Semiempirical prediction of protein folds

We introduce a semiempirical approach to predict ab initio expeditious pathways and native backbone geometries of proteins that fold under in vitro renaturation conditions. The algorithm is engineered to incorporate a discrete codification of local steric hindrances that constrain the movements of the peptide backbone throughout the folding process. Thus, the torsional state of the chain is assumed to be conditioned by the fact that hopping from one basin of attraction to another in the Ramachandran map (local potential energy surface) of each residue is energetically more costly than the search for a specific (Phi, Psi) torsional state within a single basin. A combinatorial procedure is introduced to evaluate coarsely defined torsional states of the chain defined "modulo basins" and translate them into meaningful patterns of long range interactions. Thus, an algorithm for structure prediction is designed based on the fact that local contributions to the potential energy may be subsumed into time-evolving conformational constraints defining sets of restricted backbone geometries whereupon the patterns of nonbonded interactions are constructed. The predictive power of the algorithm is assessed by (a) computing ab initio folding pathways for mammalian ubiquitin that ultimately yield a stable structural pattern reproducing all of its native features, (b) determining the nucleating event that triggers the hydrophobic collapse of the chain, and (c) comparing coarse predictions of the stable folds of moderately large proteins (N approximately 100) with structural information extracted from the protein data bank.

Origin of apparent fast and non-exponential kinetics of lysozyme folding measured in pulsed hydrogen exchange experiments

Folding of lysozyme at pH 5.2 is a complex processes. After rapid collapse (<1 ms) kinetic partitioning into a slow and fast folding pathway occurs. The fast pathway leads directly to the native structure (N), whereas the slow pathway goes through a partially folded intermediate (I(1)) with native-like secondary structure in the alpha-domain. This mechanism is in agreement with data from a large number of spectroscopic probes, from changes in the radius of gyration and from measurements on the time-course of the populations of the different species. Results from pulsed hydrogen exchange experiments, in contrast, revealed that the secondary structure of I(1) and of N is formed significantly faster than changes in spectroscopic properties occur and showed large variations in the protection kinetics of individual amide sites. We investigated the molecular origin of the rapid amide protection by quantitatively simulating all kinetic processes during the pulse-labeling experiments. Absorbance and fluorescence-detected folding kinetics showed that the early events in lysozyme folding are accelerated under exchange conditions (pH 9.2) and that a change in folding mechanism occurs due to the transient population of an additional intermediate (I(2)). This leads to kinetic competition between exchange and folding during the exchange pulse and to incomplete labeling of amide sites with slow intrinsic exchange rates. As a result, apparently faster and non-exponential kinetics of amide protection are measured in the labeling experiments. Our results further suggest that collapsed lysozyme (C) and I(1) have five and ten-times reduced free exchange rates, respectively, due to limited solvent accessibility.

Characterization of the folding kinetics of a three-helix bundle protein via a minimalist Langevin model

We use a simple off-lattice Langevin model of protein folding to characterize the folding and unfolding of a fast-folding, 46 residue three-helix bundle. Under conditions at which the C-terminal helix is 30 % stable, we observe a clear three-state folding mechanism. In the on-pathway intermediate state, the middle and C-terminal helices are folded and in contact with each other, while the N-terminal region remains disordered. Nevertheless, under these conditions this intermediate is thermodynamically unstable relative to its unfolded state. The first and highest folding barrier corresponds to the organization of the hinge between the middle and C-terminal helices. A subsequent major barrier corresponds to the organization of the hinge between the middle and N-terminal helices. Hyperstabilizing the hinge regions leads to twice the folding rate that is obtained from hyperstabilizing the helices, even though much fewer contacts are involved in hinge hyperstabilization than in helix hyperstabilization. Unfolding follows single-exponential kinetics, even at temperatures only slightly above the folding transition temperature.

Native-like beta-hairpin retained in the cold-denatured state of bovine beta-lactoglobulin

Bovine beta-lactoglobulin is denatured by increased temperature (heat denaturation) and by decreased temperature (cold-denaturation) in the presence of 4 M urea at pH 2.5. We characterized the structure of the cold-denatured state of beta-lactoglobulin using circular dichroism (CD), small-angle X-ray scattering (SAXS) and heteronuclear nuclear magnetic resonance (NMR). CD and SAXS indicated that the cold-denatured state, in comparison with the highly denatured state induced by urea, is rather compact, retaining some secondary structure, but no tertiary structure. The location of the residual structures in the cold-denatured state and their stability were characterized by 1H/2H exchange combined with heteronuclear NMR. The results indicated that the residues adjacent to the disulfide bond (C106-C119) connecting beta-strands G and H had markedly high protection factors, suggesting the presence of a native-like beta-hairpin stabilized by the disulfide bond. Since this beta-hairpin is conserved between different conformational states, including the kinetic refolding intermediate, it should be of paramount importance for the folding and stability of beta-lactoglobulin. On the other hand, the non-native alpha-helix suggested for the folding intermediate was not detected in the cold-denatured state. The 1H/2H exchange experiments showed that the protection factors of a mixture of the native and cold-denatured states is strongly biased by that of the labile cold-denatured state, consistent with a two-process model of the exchange.

Thermodynamical implications of a protein model with water interactions

We refine a protein model that reproduces fundamental aspects of protein thermodynamics. The model exhibits two transitions, hot and cold unfolding. The number of relevant parameters is reduced to three: (1) binding energy of folding relative to the orientational energy of bound water, (2) ratio of degrees of freedom between the folded and unfolded protein chain, and (3) the number of water molecules that can access the hydrophobic parts of the protein interior upon unfolding. By increasing the number of water molecules in the model, the separation between the two peaks in the heat capacity curve comes closer, which is more consistent with experimental data. In the end we show that if we, as a speculative assumption, assign only two distinct energy levels for the bound water molecules, better correspondence with experiments can be obtained.

A theoretical study on the origin of cooperativity in the formation of 3(10)- and alpha-helices

By using a simple repeating unit method, we have conducted a theoretical study which delineates the preferences for beta-strand, 2(7)-ribbon, 3(10)-helix, and alpha-helix formation for a series of polyglycine models up to 14 amino acid residues (Ac-(Gly)(n), n = 0, 1, 2,., 14). Interactions among residues, which result in cooperativity, are clearly indicated by variations in calculated energies of the residues. Whereas no cooperativity is found in the formation of beta-strands and 2(7)-ribbons, there is a significant cooperativity in the formation of 3(10)- and alpha-helices, especially for the latter. In the case of alpha-helices, the 14th residue is more stable than the 3rd by about 3 kcal/mol. A good correlation between calculated residue energy and residue dipole moment was uncovered, indicating the importance of long-range electrostatic interactions to the cooperativity. The results of our calculations are compared with those of the AMBER and PM3 methods, and indicate that both methods, AMBER and PM3, need further development in the cooperative view of electrostatic interactions. The result should be of importance in providing insight into protein folding and formation of helical structures in a variety of polymeric compounds. This also suggests a strategy for the development of more consistent molecular mechanics force fields.

Cloning and characterization of adenylate kinase from Chlamydia pneumoniae

Chlamydiae proliferate only within the infected host cells and are thought to be "energy parasites," because they take up ATP from the host cell as an energy source. In the present study, we isolated from Chlamydia pneumoniae the gene encoding adenylate kinase (AK). Using the enzyme produced in Escherichia coli, its properties were characterized. K(m) values for AMP and for ADP of the purified C. pneumoniae AK (AKcpn) were each 330 microm, which is significantly higher than the reported values of other AKs, whereas K(m) for ATP was 24 microm, which was rather lower than others. AKcpn contains 1 g atom of zinc/mol of 24,000-dalton protein. Mass spectrometric analysis of AKcpn and analysis of properties of mutated AKcpn strongly suggested that zinc is associated with four cysteine residues in the LID domain of the enzyme. The apo-AKcpn that lost zinc retained AK activity, although K(m) for AMP of apo-AKcpn increased about 2-fold and V(max) decreased about one-half from that of holo-AKcpn. The apo-AKcpn was more thermolabile and sensitive to trypsin digestion than the holo-AKcpn. Moreover, the recovery in vitro of the AK activity during the renaturation process of the denatured apo-AKcpn was dependent on zinc. A mutated protein in which cysteine residues in the LID domain were substituted by other amino acids lost both zinc and enzyme activity. The mutated protein was more sensitive to protease than the apo-AKcpn. These results indicate that zinc in AKcpn, although not essential for the catalysis, stabilizes the enzyme and probably plays a crucial role in proper folding of the protein. Furthermore, the catalytic properties of AKcpn suggest a distinctive regulatory mechanism in the metabolism compared with AKs in other organisms.

Protein folding intermediates and pathways studied by hydrogen exchange

In order to solve the immensely difficult protein-folding problem, it will be necessary to characterize the barriers that slow folding and the intermediate structures that promote it. Although protein-folding intermediates are not accessible to the usual structural studies, hydrogen exchange (HX) methods have been able to detect and characterize intermediates in both kinetic and equilibrium modes--as transient kinetic folding intermediates on a subsecond time scale, as labile equilibrium molten globule intermediates under destabilizing conditions, and as infinitesimally populated intermediates in the high free-energy folding landscape under native conditions. Available results consistently indicate that protein-folding landscapes are dominated by a small number of discrete, metastable, native-like partially unfolded forms (PUFs). The PUFs appear to be produced, one from another, by the unfolding and refolding of the protein's intrinsically cooperative secondary structural elements, which can spontaneously create stepwise unfolding and refolding pathways. Kinetic experiments identify three kinds of barrier processes: (a) an initial intrinsic search-nucleation-collapse process that prepares the chain for intermediate formation by pinning it into a condensed coarsely native-like topology; (b) smaller search-dependent barriers that put the secondary structural units into place; and (c) optional error-dependent misfold-reorganization barriers that can cause slow folding, intermediate accumulation, and folding heterogeneity. These conclusions provide a coherent explanation for the grossly disparate folding behavior of different globular proteins in terms of distinct folding pathways.

An amino acid code for protein folding

Direct structural information obtained for many proteins supports the following conclusions. The amino acid sequences of proteins can stabilize not only the final native state but also a small set of discrete partially folded native-like intermediates. Intermediates are formed in steps that use as units the cooperative secondary structural elements of the native protein. Earlier intermediates guide the addition of subsequent units in a process of sequential stabilization mediated by native-like tertiary interactions. The resulting stepwise self-assembly process automatically constructs a folding pathway, whether linear or branched. These conclusions are drawn mainly from hydrogen exchange-based methods, which can depict the structure of infinitesimally populated folding intermediates at equilibrium and kinetic intermediates with subsecond lifetimes. Other kinetic studies show that the polypeptide chain enters the folding pathway after an initial free-energy-uphill conformational search. The search culminates by finding a native-like topology that can support forward (native-like) folding in a free-energy-downhill manner. This condition automatically defines an initial transition state, the search for which sets the maximum possible (two-state) folding rate. It also extends the sequential stabilization strategy, which depends on a native-like context, to the first step in the folding process. Thus the native structure naturally generates its own folding pathway. The same amino acid code that translates into the final equilibrium native structure-by virtue of propensities, patterning, secondary structural cueing, and tertiary context-also produces its kinetic accessibility.

Molecular dynamics simulation of Escherichia coli dihydrofolate reductase and its protein fragments: relative stabilities in experiment and simulations

We have carried out molecular dynamics simulations of the native dihydrofolate reductase from Escherichia coli and several of its folded protein fragments at standard temperature. The simulations have shown fragments 1--36, 37--88, and 89--159 to be unstable, with a C(alpha)RMSD (C(alpha) root mean squared deviation) >5 A after 3.0 nsec of simulation. The unfolding of fragment 1--36 was immediate, whereas fragments 37--88 and 89--159 gradually unfolded because of the presence of the beta-sheet core structure. In the absence of residues 1--36, the two distinct domains comprising fragment 39--159 associated with each other, resulting in a stable conformation. This conformation retained most of its native structural elements. We have further simulated fragments derived from computational protein cutting. These were also found to be unstable, with the exception of fragment 104--159. In the absence of alpha(4), the loose loop region of residues 120--127 exhibited a beta-strand-like behavior, associating itself with the beta-sheet core of the protein fragment. The current study suggests that the folding of dihydrofolate reductase involves cooperative folding of distinct domains which otherwise would have been unstable as independent folded units in solution. Finally, the critical role of residues 1--36 in allowing the two distinct domains of fragment 104--159 to fold into the final native conformation is discussed.

A systematic study of the vibrational free energies of polypeptides in folded and random states

Molecular vibrations, especially low frequency motions, may be used as an indication of the rigidity or the flatness of the protein folding energy landscape. We have studied the vibrational properties of native folded as well as random coil structures of more than 60 polypeptides. The picture we obtain allows us to perceive how and why the energy landscape progressively rigidifies while still allowing potential flexibility. Compared with random coil structures, both alpha-helices and beta-hairpins are vibrationally more flexible. The vibrational properties of loop structures are similar to those of the corresponding random coil structures. Inclusion of an alpha-helix tends to rigidify peptides and so-called building blocks of the structure, whereas the addition of a beta-structure has less effect. When small building blocks coalesce to form larger domains, the protein rigidifies. However, some folded native conformations are still found to be vibrationally more flexible than random coil structures, for example, beta(2)-microglobulin and the SH3 domain. Vibrational free energy contributes significantly to the thermodynamics of protein folding and affects the distribution of the conformational substates. We found a weak correlation between the vibrational folding energy and the protein size, consistent with both previous experimental estimates and theoretical partition of the heat capacity change in protein folding.

Pathways in two-state protein folding

Thermodynamic measurements of proteins indicate that the folding to the native state takes place either through stable intermediates or through a two-state process without intermediates. The rather short folding times of proteins indicate that folding is guided through some sequence of contact bindings. We discuss the possibility of reconciling a two-state folding event with a sequential folding process in a schematic model of protein folding. We propose a new dynamical transition temperature that is lower than the temperature at which proteins in equilibrium unfold. This is in qualitative agreement with observations of in vivo protein folding activity quantified by chaperone concentration in Escherichia coli. Finally, we discuss our framework in connection with the unfolding of proteins at low temperatures.

Formation of hydrogen bonds precedes the rate-limiting formation of persistent structure in the folding of ACBP

A burst phase in the early folding of the four-helix two-state folder protein acyl-coenzyme A binding protein (ACBP) has been detected using quenched-flow in combination with site-specific NMR-detected hydrogen exchange. Several of the burst phase structures coincide with a structure consisting of eight conserved hydrophobic residues at the interface between the two N and C-terminal helices. Previous mutation studies have shown that the formation of this structure is rate limiting for the final folding of ACBP. The burst phase structures observed in ACBP are different from the previously reported collapsed types of burst phase intermediates observed in the folding of other proteins.

Binding and folding: in search of intramolecular chaperone-like building block fragments

We propose an intramolecular chaperone which catalyzes folding and neither dissociates nor is cleaved. This uncleaved foldase is an intramolecular chain-linked chaperone, which constitutes a critical building block of the structure. Macroscopically, all molecular chaperones facilitate folding reactions and manifest similar energy landscapes. However, microscopically they differ. While intermolecular chaperones catalyze folding by unfolding misfolded conformations or prevent misfolding, the chain-linked cleaved (proregion) and uncleaved intramolecular chaperone-like building blocks suggested here, catalyze folding by binding to, stabilizing and increasing the populations of native conformations of adjacent building block fragments. In both, the more stable the intramolecular chaperone fragment region, the faster is the folding rate. Hence, mechanistically, intramolecular chaperones and chaperone-like segments are similar. Both play a dual role, in folding and in protein function. However, while the functional role of the proregions is inhibitory, necessitating their cleavage, the function of the uncleaved intramolecular chaperone-like building blocks does not require their subsequent removal. On the contrary, it requires that they remain in the structure. This may lead to the difference in the type of control they are under: proteins folding with the assistance of the proregion have been shown to be under kinetic control. It has been suggested that kinetically controlled folding reactions, with the proregion catalyst removed, lend longevity under harsh conditions. On the other hand, proteins with uncleaved intramolecular chaperone-like building blocks, with their 'foldases' still attached, are largely under thermodynamic control, consistent with the control observed in most protein folding reactions. We propose that an uncleaved intramolecular chaperone-like fragment occurs frequently in proteins. We further propose that such proteins would be prone to changing conditions and in particular, to mutations in this critical building block region. We describe the features qualifying it for its proposed chaperone-like role, compare it with inter- and intramolecular chaperones and review current literature in this light. We further propose a mechanism showing how it lowers the barrier heights, leading to faster folding reaction rates. Since these fragments constitute an intergal part of the protein structure, we call these critical building blocks intramolecular, chaperone-like fragments, to clarify, distinguish and adhere to the definition of the transiently associating chaperones. The new mechanism presented here differs from the concept of 'folding nuclei'. While the concept of folding nuclei focuses on a non-sequential distribution of the folding information along the entire protein chain, the chaperone-like building block fragments proposition focuses on a segmental distribution of the folding information. This segmental distribution controls the distributions of the populations throughout the hierarchical folding processes.

Trapping the monomeric alpha-helical state during unfolding of coiled-coils by reversed-phase liquid chromatography

Reversed-phase liquid chromatography (RPLC) offers a unique opportunity to monitor the transition from the native state (N) to the structural intermediate state (I) for proteins whose secondary structure is comprised entirely of amphipathic helices, such as coiled-coils. During RPLC, the hydrophobicity of the stationary phase and mobile phase results in the unfolding of the tertiary/quaternary structure of coiled-coils but retains alpha-helical secondary structure and thus isolates the I state. A set of five peptides, alphaalpha-36, betabeta-36, alphabeta-36, gammadelta-36 and omegaomega-36, was generated by shuffling guest hydrophobes at equivalent sites in a symmetric host frame. In one of the peptides, omegaomega-36, all the alpha-glutamic residues in the host frame were replaced by gamma-glutamic residues. alphaalpha-36, betabeta-36, alphabeta-36, gammadelta-36 form two-stranded coiled-coils of identical helical content and unfold as a two-state transition during temperature denaturation while the fifth peptide, omegaomega-36, is a random coil and cannot be induced in to an alpha-helical structure even in the presence of a helix inducing solvent, 50% trifluoroethanol. By comparing the stability order of the four coiled-coils in the N-->I transition (measured by RPLC studies) with that in the N-->D (denatured state) transition (measured by calorimetry), it is concluded that there is a direct correlation between the relative stabilities of these peptides in these two unfolding transitions. This result supports a hierarchical folding mechanism for coiled-coils.

Generating ensemble averages for small proteins from extended conformations by Monte Carlo simulations

Since it is not feasible to determine the structure of every protein by experiment, algorithms delivering the folded conformation of a protein solely from its amino acid sequence are desirable. Here the diffusion-process controlled-Monte Carlo approach has been applied to generating ensemble averages for three small proteins with 31, 36, and 46 residues. Starting from extended conformations and using an energy model that was developed on other protein models, the simulations find nativelike structures deviating by 3 A rms from the experimental structures for the main chain atoms. The balance between long-range and short-range interactions is discussed briefly in the context of stability and folding.

Stepwise formation of alpha-helices during cytochrome c folding

Two models have been proposed to describe the folding pathways of proteins. The framework model assumes the initial formation of the secondary structures whereas the hydrophobic collapse model supposes their formation after the collapse of backbone structures. To differentiate between these models for real proteins, we have developed a novel CD spectrometer that enables us to observe the submillisecond time frame of protein folding and have characterized the timing of secondary structure formation in the folding process of cytochrome c (cyt c). We found that approximately 20% of the native helical content was organized in the first phase of folding, which is completed within milliseconds. Furthermore, we suggest the presence of a second intermediate, which has alpha-helical content resembling that of the molten globule state. Our results indicate that many of the alpha-helices are organized after collapse in the folding mechanism of cyt c.

Protein aging hypothesis of Alzheimer disease

Alzheimer disease (AD), the most common form of aging-related neurodegenerative disorders, is associated with formation of fibrillar deposits of amyloid beta-protein (Abeta). While the direct involvement of Abeta in AD has been well documented, the relations between Abeta production, amyloid formation, and neurodegeneration remain unknown. We propose that AD is initiated by a protein aging-related structural transformation in soluble Abeta. We hypothesize that spontaneous chemical modification of aspartyl residues in Abeta to transient succinimide induces a non-native conformation in a fraction of soluble Abeta, rendering it amyloidogenic and neurotoxic. Conformationally altered Abeta is characterized by increased stability in solution and the presence of a non-native beta-turn that determines folding of Abeta in solution and the structure of Abeta subunits incorporated into amyloid fibrils. While the soluble 'non-native' Abeta is both the factor triggering the neurodegenerative cascade and the precursor of amyloid plaques, these two events result from interaction of Abeta with different sets of cellular components and need not coincide in space and time. Extensive literature data and experimental evidence are provided in support of this hypothesis.

Protein folding and stability investigated by fluorescence, circular dichroism (CD), and nuclear magnetic resonance (NMR) spectroscopy: the flavodoxin story

In this review, the experimental results obtained on the folding and stability of Azotobacter vinelandii flavodoxin are summarised. By doing so, three main spectroscopic techniques used to investigate protein folding and stability are briefly introduced. These techniques are: circular dichroism (CD) spectroscopy, fluorescence emission spectroscopy, and nuclear magnetic resonance (NMR) spectroscopy in combination with the hydrogen exchange methodology. Results on the denaturant-induced and thermal equilibrium unfolding of apoflavodoxin from A. vinelandii, i.e. flavodoxin in the absence of the riboflavin-5'-monophosphate (FMN) cofactor, are discussed. A scheme for the equilibrium unfolding of apoflavodoxin is presented which involves a relatively stable molten globule-like intermediate. Denaturant-induced apoflavodoxin (un)folding as followed at the residue-level by NMR shows that the transition of native A. vinelandii apoflavodoxin to its molten globule state is highly co-operative. However, the unfolding of the molten globule to the unfolded state of the protein is non-co-operative. A comparison of the folding of A. vinelandii flavodoxin with the folding of flavodoxin from Anaboena PCC 7119 is made. The local stabilities of apo- and holoflavodoxin from A. vinelandii as measured by NMR spectroscopy are compared. Both Che Y and cutinase, which have no sequence homology with apoflavodoxin but which share the flavodoxin-like topology, have stabilisation centres different from that of apoflavodoxin from A. vinelandii. The stable centres of structurally similar proteins can thus reside in different parts of the same protein topology. Insight in the variations in (local) unfolding processes of structurally similar proteins can be used to stabilise proteins with a flavodoxin-like fold. Finally, the importance of some recent experimental and theoretical developments for the study of flavodoxin folding is briefly discussed.