Efficient production of fluorophore-labeled CC chemokines for biophysical studies using recombinant enterokinase and recombinant sortase.

Chemokines are important immune system proteins, many of which mediate inflammation due to their function to activate and cause chemotaxis of leukocytes. An important anti-inflammatory strategy is therefore to bind and inhibit chemokines, which leads to the need for biophysical studies of chemokines as they bind various possible partners. Because a successful anti-chemokine drug should bind at low concentrations, techniques such as fluorescence anisotropy that can provide nanomolar signal detection are required. To allow fluorescence experiments to be carried out on chemokines, a method is described for the production of fluorescently labeled chemokines. First, a fusion-tagged chemokine is produced in Escherichia coli, then efficient cleavage of the N-terminal fusion partner is carried out with lab-produced enterokinase, followed by covalent modification with a fluorophore, mediated by the lab-produced sortase enzyme. This overall process reduces the need for expensive commercial enzymatic reagents. Finally, we utilize the product, vMIP-fluor, in binding studies with the chemokine binding protein vCCI, which has great potential as an anti-inflammatory therapeutic, showing a binding constant for vCCI:vMIP-fluor of 0.37 ± 0.006 nM. We also show how a single modified chemokine homolog (vMIP-fluor) can be used in competition assays with other chemokines and we report a Kd for vCCI:CCL17 of 14 µM. This work demonstrates an efficient method of production and fluorescent labeling of chemokines for study across a broad range of concentrations.

Natural and Engineered Isoforms of the Inflammasome Adaptor ASC Form Noncovalent, pH-Responsive Hydrogels.

The protein ASC polymerizes into intricate filament networks to assemble the inflammasome, a filamentous multiprotein complex that triggers the inflammatory response. ASC carries two Death Domains integrally involved in protein self-association for filament assembly. We have leveraged this behavior to create noncovalent, pH-responsive hydrogels of full-length, folded ASC by carefully controlling the pH as a critical factor in the polymerization process. We show that natural variants of ASC (ASC isoforms) involved in inflammasome regulation also undergo hydrogelation. To further demonstrate this general capability, we engineered proteins inspired by the ASC structure that also form hydrogels. We analyzed the structural network of the natural and engineered protein hydrogels using transmission and scanning electron microscopy and studied their viscoelastic behavior using shear rheology. Our results reveal one of the very few examples of hydrogels created by the self-assembly of globular proteins and domains in their native conformation and show that Death Domains can be used alone or as building blocks to engineer bioinspired hydrogels.

Natural and engineered isoforms of the inflammasome adaptor ASC form non-covalent, pH-responsive hydrogels.

The protein ASC polymerizes into intricate filament networks to assemble the inflammasome, a filamentous multiprotein complex that triggers the inflammatory response. ASC carries two Death Domains integrally involved in protein self-association for filament assembly. We have leveraged this behavior to create non-covalent, pH-responsive hydrogels of full-length, folded ASC by carefully controlling the pH as a critical factor in the polymerization process. We show that natural variants of ASC (ASC isoforms) involved in inflammasome regulation also undergo hydrogelation. To further demonstrate this general capability, we engineered proteins inspired in the ASC structure that successfully form hydrogels. We analyzed the structural network of the natural and engineered protein hydrogels using transmission and scanning electron microscopy, and studied their viscoelastic behavior by shear rheology. Our results reveal one of the very few examples of hydrogels created by the self-assembly of globular proteins and domains in their native conformation and show that Death Domains can be used alone or as building blocks to engineer bioinspired hydrogels.

A modular approach to map out the conformational landscapes of unbound intrinsically disordered proteins.

Intrinsically disordered proteins (IDPs) fold upon binding to select/recruit multiple partners, morph around the partner's structure, and exhibit allostery. However, we do not know whether these properties emerge passively from disorder, or rather are encoded into the IDP's folding mechanisms. A main reason for this gap is the lack of suitable methods to dissect the energetics of IDP conformational landscapes without partners. Here we introduce such an approach that we term molecular LEGO, and apply it to NCBD, a helical, molten globule­like IDP, as proof of concept. The approach entails the experimental and computational characterization of the protein, its separate secondary structure elements (LEGO building blocks), and their supersecondary combinations. Comparative analysis uncovers specific, yet inconspicuous, energetic biases in the conformational/folding landscape of NCBD, including 1) strong local signals that define the three native helices, 2) stabilization of helix­helix interfaces via soft pairwise tertiary interactions, 3) cooperative stabilization of a heterogeneous three-helix bundle fold, and 4) a dynamic exchange between sets of tertiary interactions (native and nonnative) that recapitulate the different structures NCBD adopts in complex with various partners. Crucially, a tug of war between sets of interactions makes NCBD gradually shift between structural subensembles as a conformational rheostat. Such conformational rheostatic behavior provides a built-in mechanism to modulate binding and switch/recruit partners that is likely at the core of NCBD's function as transcriptional coactivator. Hence, the molecular LEGO approach emerges as a powerful tool to dissect the conformational landscapes of unbound IDPs and rationalize their functional mechanisms.

Cognate DNA Recognition by Engrailed Homeodomain Involves a Conformational Change Controlled via an Electrostatic-Spring-Loaded Latch.

Transcription factors must scan genomic DNA, recognize the cognate sequence of their control element(s), and bind tightly to them. The DNA recognition process is primarily carried out by their DNA binding domains (DBD), which interact with the cognate site with high affinity and more weakly with any other DNA sequence. DBDs are generally thought to bind to their cognate DNA without changing conformation (lock-and-key). Here, we used nuclear magnetic resonance and circular dichroism to investigate the interplay between DNA recognition and DBD conformation in the engrailed homeodomain (enHD), as a model case for the homeodomain family of eukaryotic DBDs. We found that the conformational ensemble of enHD is rather flexible and becomes gradually more disordered as ionic strength decreases following a Debye-Hückel's dependence. Our analysis indicates that enHD's response to ionic strength is mediated by a built-in electrostatic spring-loaded latch that operates as a conformational transducer. We also found that, at moderate ionic strengths, enHD changes conformation upon binding to cognate DNA. This change is of larger amplitude and somewhat orthogonal to the response to ionic strength. As a consequence, very high ionic strengths (e.g., 700 mM) block the electrostatic-spring-loaded latch and binding to cognate DNA becomes lock-and-key. However, the interplay between enHD conformation and cognate DNA binding is robust across a range of ionic strengths (i.e., 45 to 300 mM) that covers the physiologically-relevant conditions. Therefore, our results demonstrate the presence of a mechanism for the conformational control of cognate DNA recognition on a eukaryotic DBD. This mechanism can function as a signal transducer that locks the DBD in place upon encountering the cognate site during active DNA scanning. The electrostatic-spring-loaded latch of enHD can also enable the fine control of DNA recognition in response to transient changes in local ionic strength induced by variate physiological processes.

Lessons about Protein Folding and Binding from Archetypal Folds.

The function of proteins as biological nanomachines relies on their ability to fold into complex 3D structures, bind selectively to partners, and undergo conformational changes on cue. The native functional structures, and the rates of interconversion between conformational states (folded-unfolded, bound-free), are all encoded in the physical chemistry of their amino acid sequence. However, despite extensive research over decades, this code has proven difficult to fully crack, in terms of both prediction and understanding the molecular mechanisms at play.Earlier work on single-domain proteins reported a commonality of slow rates (10-2-102 s-1) and simple behavior in both kinetic and thermodynamic unfolding experiments, which suggested the process was all-or-none and thereby analogous to a chemical reaction (e.g., A ⇄ B). In the absence of a first-principles pre-exponential factor for protein (un)folding dynamics, the rates could only be interpreted in relative terms, e.g., the changes induced by mutation, and hence, neither the height of nor the entropic contribution to the free energy barriers was known. The rates were also many orders of magnitude too slow for direct atomistic simulations, and the computational focus was on predicting rate changes induced by mutation via coarse grained simulations. However, even the effects of mutation proved to be strikingly homogeneous with all experimental data clustering at ∼1/3 of the free energy perturbation recovered on folding and ∼2/3 on unfolding.The implementation of ultrafast kinetic methods turned the field upside down because they allowed researchers to measure the time scales of elementary (un)folding motions, which set the pre-exponential factor for protein conformational transitions at ∼1 µs. In parallel, we and others set out to investigate the simplest possible protein structures capable of autonomous folding, which we defined as archetypal folds. The rationale was to recapitulate the hierarchical organization of protein structure, starting from the bottom up. The study of fold archetypes ended up opening new research avenues in protein (un)folding, but also making unexpected connections with the folding upon binding of intrinsically disordered proteins and suggesting their functioning as conformational rheostats.This Account describes our work on the kinetic, thermodynamic, mechanistic, and functional analysis of fold archetypes. We first discuss the kinetic studies, emphasizing their impact on our understanding of (un)folding rates, of barrierless (downhill) folding, and as benchmarks for atomistic simulations. We continue with the thermodynamic analysis, introducing the differential scanning calorimetry, multiprobe, and NMR approaches that we developed to dissect their gradual, minimally cooperative (un)folding transitions and to probe the underlying mechanisms with unprecedented detail. The last two sections cover single-molecule analyses and some recent, mostly computational, results on the exploration of possible biological and technological roles for the gradual conformational transitions of fold archetypes.

Downhill (Un)Folding Coupled to Binding as a Mechanism for Engineering Broadband Protein Conformational Transducers.

Canonical proteins fold and function as conformational switches that toggle between their folded (on) and unfolded (off) states, a mechanism that also provides the basis for engineering transducers for biosensor applications. One of the limitations of such transducers, however, is their relatively narrow operational range, limited to ligand concentrations 20-fold below or above their C50. Previously, we discovered that certain fast-folding proteins lose/gain structure gradually (downhill folding), which led us to postulate their operation as conformational rheostats capable of processing inputs/outputs in analog fashion. Conformational rheostats could make transducers with extended sensitivity. Here we investigate this hypothesis by engineering pH transducing into the naturally pH insensitive, downhill folding protein gpW. Particularly, we engineered histidine grafts into its hydrophobic core to induce unfolding via histidine ionization. We designed and tested the effects of ionization via computational modeling and studied experimentally the four most promising single grafts and two double grafts. All tested mutants become reversible pH transducers in the 4-9 range, and their response increases proportionally to how buried the histidine graft is. Importantly, the pH-dependent reversible (un)folding occurs in rheostatic fashion, so the engineered transducers can detect up to 6 orders of magnitude in [H+] for single grafts, and even more for double grafts. Our results demonstrate that downhill (un)folding coupled to binding produces the gradual, analog responses to the ligand (here H+) that are expected of conformational rheostats, and which make them a powerful mechanism for engineering transducers with sensitivity over many orders of magnitude in ligand concentration (broadband).

Engineering protein assemblies with allosteric control via monomer fold-switching.

The macromolecular machines of life use allosteric control to self-assemble, dissociate and change shape in response to signals. Despite enormous interest, the design of nanoscale allosteric assemblies has proven tremendously challenging. Here we present a proof of concept of allosteric assembly in which an engineered fold switch on the protein monomer triggers or blocks assembly. Our design is based on the hyper-stable, naturally monomeric protein CI2, a paradigm of simple two-state folding, and the toroidal arrangement with 6-fold symmetry that it only adopts in crystalline form. We engineer CI2 to enable a switch between the native and an alternate, latent fold that self-assembles onto hexagonal toroidal particles by exposing a favorable inter-monomer interface. The assembly is controlled on demand via the competing effects of temperature and a designed short peptide. These findings unveil a remarkable potential for structural metamorphosis in proteins and demonstrate key principles for engineering protein-based nanomachinery.

Ultrafast folding kinetics of WW domains reveal how the amino acid sequence determines the speed limit to protein folding.

Protein (un)folding rates depend on the free-energy barrier separating the native and unfolded states and a prefactor term, which sets the timescale for crossing such barrier or folding speed limit. Because extricating these two factors is usually unfeasible, it has been common to assume a constant prefactor and assign all rate variability to the barrier. However, theory and simulations postulate a protein-specific prefactor that contains key mechanistic information. Here, we exploit the special properties of fast-folding proteins to experimentally resolve the folding rate prefactor and investigate how much it varies among structural homologs. We measure the ultrafast (un)folding kinetics of five natural WW domains using nanosecond laser-induced temperature jumps. All five WW domains fold in microseconds, but with a 10-fold difference between fastest and slowest. Interestingly, they all produce biphasic kinetics in which the slower phase corresponds to reequilibration over the small barrier (<3 RT) and the faster phase to the downhill relaxation of the minor population residing at the barrier top [transition state ensemble (TSE)]. The fast rate recapitulates the 10-fold range, demonstrating that the folding speed limit of even the simplest all-ß fold strongly depends on the amino acid sequence. Given this fold's simplicity, the most plausible source for such prefactor differences is the presence of nonnative interactions that stabilize the TSE but need to break up before folding resumes. Our results confirm long-standing theoretical predictions and bring into focus the rate prefactor as an essential element for understanding the mechanisms of folding.

Glutton: a tool for generating structural ensembles of partly disordered proteins from chemical shifts.

MOTIVATION: Many proteins are partially disordered in physiological conditions and only fold, fully or partially, upon binding. Their structural analysis is challenging because the accessible information, typically chemical shifts (CS) from nuclear magnetic resonance experiments, are averages over broad ensembles of conformations. We aim to develop a database for the analysis of such data in terms of conformational distributions of the protein backbone rather than of individual high-resolution structures. RESULTS: Glutton is the largest available database linking CS and protein 3D structures (5270 entries organized in three levels) and is searchable via a python script. It generates statistical distributions of ϕ-ψ dihedral angles based on CS or vice versa. Such ϕ-ψ distributions are used to calculate structural ensembles of partially disordered proteins from their CS. For folded proteins, such ensembles are excellent starting points for further refinement with additional experimental restraints (structure determination) or computational methods (structure prediction). AVAILABILITY AND IMPLEMENTATION: Glutton is freely available at https://github.com/YeeHo/Glutton. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Limited cooperativity in protein folding.

Theory and simulations predict that the structural concert of protein folding reactions is relatively low. Experimentally, folding cooperativity has been difficult to study, but in recent years we have witnessed major advances. New analytical procedures in terms of conformational ensembles rather than discrete states, experimental techniques with improved time, structural, or single-molecule resolution, and combined thermodynamic and kinetic analysis of fast folding have contributed to demonstrate a general scenario of limited cooperativity in folding. Gradual structural disorder is already apparent on the unfolded and native states of slow, two-state folding proteins, and it greatly increases in magnitude for fast folding domains. These results demonstrate a direct link between how fast a single-domain protein folds and unfolds, and how cooperative (or structurally diverse) is its equilibrium unfolding process. Reducing cooperativity also destabilizes the native structure because it affects unfolding more than folding. We can thus define a continuous cooperativity scale that goes from the 'pliable' two-state character of slow folders to the gradual unfolding of one-state downhill, and eventually to intrinsically disordered proteins. The connection between gradual unfolding and intrinsic disorder is appealing because it suggests a conformational rheostat mechanism to explain the allosteric effects of folding coupled to binding.

Gradual disordering of the native state on a slow two-state folding protein monitored by single-molecule fluorescence spectroscopy and NMR.

Theory predicts that folding free energy landscapes are intrinsically malleable and as such are expected to respond to perturbations in topographically complex ways. Structural changes upon perturbation have been observed experimentally for unfolded ensembles, folding transition states, and fast downhill folding proteins. However, the native state of proteins that fold in a two-state fashion is conventionally assumed to be structurally invariant during unfolding. Here we investigate how the native and unfolded states of the chicken α-spectrin SH3 domain (a well characterized slow two-state folder) change in response to chemical denaturants and/or temperature. We can resolve the individual properties of the two end-states across the chemical unfolding transition employing single-molecule fluorescence spectroscopy (SM-FRET) and across the thermal unfolding transition by NMR because SH3 folds-unfolds in the slow chemical exchange regime. Our results demonstrate that α-spectrin SH3 unfolds in a canonical way in the sense that it converts between the native state and an unfolded ensemble that expands in response to chemical denaturants. However, as conditions become increasingly destabilizing, the native state also expands gradually, and a large fraction of its native intramolecular hydrogen bonds break up. This gradual disordering of the native state takes place in times shorter than the 100 µs resolution of our SM-FRET experiments. α-Spectrin SH3 thus showcases the extreme plasticity of folding landscapes, which extends to the native state of slow two-state proteins. Our results point to the idea that folding mechanisms under physiological conditions might be quite different from those obtained by linear extrapolation from denaturing conditions. Furthermore, they highlight a pressing need for re-evaluating the conventional procedures for analyzing and interpreting folding experiments, which may be based on too-simplistic assumptions.

Protein folding at atomic resolution: analysis of autonomously folding supersecondary structure motifs by nuclear magnetic resonance.

The study of protein folding has been conventionally hampered by the assumption that all single-domain proteins fold by an all-or-none process (two-state folding) that makes it impossible to resolve folding mechanisms experimentally. Here we describe an experimental method for the thermodynamic analysis of protein folding at atomic resolution using nuclear magnetic resonance (NMR). The method is specifically developed for the study of small proteins that fold autonomously into basic supersecondary structure motifs, and that do so in the sub-millisecond timescale (folding archetypes). From the NMR experiments we obtain hundreds of atomic unfolding curves that are subsequently analyzed leading to the determination of the characteristic network of folding interactions. The application of this approach to a comprehensive catalog of elementary folding archetypes holds the promise of becoming the first experimental approach capable of unraveling the basic rules connecting protein structure and folding mechanism.

Downhill protein folding modules as scaffolds for broad-range ultrafast biosensors.

Conformational switches are macromolecules that toggle between two states (active/inactive or folded/unfolded) upon specific binding to a target molecule. These molecular devices provide an excellent scaffold for developing real-time biosensors. Here we take this concept one step beyond to build high-performance conformational rheostat sensors. The rationale is to develop sensors with expanded dynamic range and faster response time by coupling a given signal to the continuous (rather than binary) unfolding process of one-state downhill folding protein modules. As proof of concept we investigate the pH and ionic-strength sensing capabilities of the small α-helical protein BBL. Our results reveal that such a pH/ionic-strength sensor exhibits a linear response over 4 orders of magnitude in analyte concentration, compared to the 2 orders of magnitude for switches, and nearly concentration-independent microsecond response times.

The effect of electrostatics on the marginal cooperativity of an ultrafast folding protein.

Proteins fold up by coordinating the different segments of their polypeptide chain through a network of weak cooperative interactions. Such cooperativity results in unfolding curves that are typically sigmoidal. However, we still do not know what factors modulate folding cooperativity or the minimal amount that ensures folding into specific three-dimensional structures. Here, we address these issues on BBL, a small helical protein that folds in microseconds via a marginally cooperative downhill process (Li, P., Oliva, F. Y., Naganathan, A. N., and Muñoz, V. (2009) Proc. Natl. Acad. Sci. USA. 106, 103-108). Particularly, we explore the effects of salt-induced screening of the electrostatic interactions in BBL at neutral pH and in acid-denatured BBL. Our results show that electrostatic screening stabilizes the native state of the neutral and protonated forms, inducing complete refolding of acid-denatured BBL. Furthermore, without net electrostatic interactions, the unfolding process becomes much less cooperative, as judged by the broadness of the equilibrium unfolding curve and the relaxation rate. Our experiments show that the marginally cooperative unfolding of BBL can still be made twice as broad while the protein retains its ability to fold into the native three-dimensional structure in microseconds. This result demonstrates experimentally that efficient folding does not require cooperativity, confirming predictions from theory and computer simulations and challenging the conventional biochemical paradigm. Furthermore, we conclude that electrostatic interactions are an important factor in determining folding cooperativity. Thus, electrostatic modulation by pH-salt and/or mutagenesis of charged residues emerges as an attractive tool for tuning folding cooperativity.

A designed protein as experimental model of primordial folding.

How do proteins accomplish folding during early evolution? Theoretically the mechanism involves the selective stabilization of the native structure against all other competing compact conformations in a process that involves cumulative changes in the amino acid sequence along geological timescales. Thus, an evolved protein folds into a single structure at physiological temperature, but the conformational competition remains latent. For natural proteins such competition should emerge only near cryogenic temperatures, which places it beyond experimental testing. Here, we introduce a designed monomeric miniprotein (FSD-1ss) that within biological temperatures (330-280 K) switches between simple fast folding and highly complex conformational dynamics in a structurally degenerate compact ensemble. Our findings demonstrate the physical basis for protein folding evolution in a designed protein, which exhibits poorly evolved or primordial folding. Furthermore, these results open the door to the experimental exploration of primitive folding and the switching between alternative protein structures that takes place in evolutionary branching points and prion diseases, as well as the benchmarking of de novo design methods.

Exploiting the downhill folding regime via experiment.

Traditionally, folding experiments have been directed at determining equilibrium and relaxation rate constants of proteins that fold with two-state-like kinetics. More recently, the combination of free energy surface approaches inspired by theory with the discovery of proteins that fold in the downhill regime has greatly widened the battlefield for experimentalists. Downhill folding proteins cross very small or no free energy barrier at all so that all relevant partially folded conformations become experimentally accessible. From these combined efforts we now have tools to estimate the height of thermodynamic and kinetic folding barriers. Procedures to measure with atomic resolution the structural heterogeneity of conformational ensembles at varying unfolding degrees are also available. Moreover, determining the dynamic modes driving folding and how they change as folding proceeds is finally at our fingertips. These developments allow us to address via experiment fundamental questions such as the origin of folding cooperativity, the relationship between structure and stability, or how to engineer folding barriers. Moreover, the level of detail attained in this new breed of experiments should provide powerful benchmarks for computer simulations of folding and force-field refinement.

Dynamics, energetics, and structure in protein folding.

For many decades, protein folding experimentalists have worked with no information about the time scales of relevant protein folding motions and without methods for estimating the height of folding barriers. Protein folding experiments have been interpreted using chemical models in which the folding process is characterized as a series of equilibria between two or more distinct states that interconvert with activated kinetics. Accordingly, the information to be extracted from experiments was circumscribed to apparent equilibrium constants and relative folding rates. Recent developments are changing this situation dramatically. The combination of fast-folding experiments with the development of analytical methods more closely connected to physical theory reveals that folding barriers in native conditions range from minimally high (approximately 14RT for the very slow folder AcP) to nonexistent. While slow-folding (i.e., > or = 1 ms) single-domain proteins are expected to fold in a two-state fashion, microsecond-folding proteins should exhibit complex behavior arising from crossing marginal or negligible folding barriers. This realization opens a realm of exciting opportunities for experimentalists. The free energy surface of a protein with a marginal (or no) barrier can be mapped using equilibrium experiments, which could resolve energetic factors from structural factors in folding. Kinetic experiments on these proteins provide the unique opportunity to measure folding dynamics directly. Furthermore, the complex distributions of time-dependent folding behaviors expected for these proteins might be accessible to single-molecule measurements. Here, we discuss some of these recent developments in protein folding, emphasizing aspects that can serve as a guide for experimentalists interested in exploiting this new avenue of research.

Atom-by-atom analysis of global downhill protein folding.

Protein folding is an inherently complex process involving coordination of the intricate networks of weak interactions that stabilize native three-dimensional structures. In the conventional paradigm, simple protein structures are assumed to fold in an all-or-none process that is inaccessible to experiment. Existing experimental methods therefore probe folding mechanisms indirectly. A widely used approach interprets changes in protein stability and/or folding kinetics, induced by engineered mutations, in terms of the structure of the native protein. In addition to limitations in connecting energetics with structure, mutational methods have significant experimental uncertainties and are unable to map complex networks of interactions. In contrast, analytical theory predicts small barriers to folding and the possibility of downhill folding. These theoretical predictions have been confirmed experimentally in recent years, including the observation of global downhill folding. However, a key remaining question is whether downhill folding can indeed lead to the high-resolution analysis of protein folding processes. Here we show, with the use of nuclear magnetic resonance (NMR), that the downhill protein BBL from Escherichia coli unfolds atom by atom starting from a defined three-dimensional structure. Thermal unfolding data on 158 backbone and side-chain protons out of a total of 204 provide a detailed view of the structural events during folding. This view confirms the statistical nature of folding, and exposes the interplay between hydrogen bonding, hydrophobic forces, backbone conformation and side-chain entropy. From the data we also obtain a map of the interaction network in this protein, which reveals the source of folding cooperativity. Our approach can be extended to other proteins with marginal barriers (less than 3RT), providing a new tool for the study of protein folding.

Detection and characterization of partially unfolded oligomers of the SH3 domain of alpha-spectrin.

For the purpose of equilibrium and kinetic folding-unfolding studies, the SH3 domain of alpha-spectrin (spc-SH3) has long been considered a classic two-state folding protein. In this work we have indeed observed that the thermal unfolding curves of spc-SH3 measured at pH 3.0 by differential scanning calorimetry, circular dichroism, and NMR follow apparently the two-state model when each unfolding profile is considered individually. Nevertheless, we have found that protein concentration has a marked effect upon the thermal unfolding profiles. This effect cannot be properly explained in terms of the two-state unfolding model and can only be interpreted in terms of the accumulation of intermediate associated states in equilibrium with the monomeric native and unfolded states. By chemical cross-linking and pulsed-field gradient NMR diffusion experiments we have been able to confirm the existence of associated states formed during spc-SH3 unfolding. A three-state model, in which a dimeric intermediate state is assumed to be significantly populated, provides the simplest interpretation of the whole set of thermal unfolding data and affords a satisfactory explanation for the concentration effects observed. Whereas at low concentrations the population of the associated intermediate state is negligible and the unfolding process consequently takes place in a two-state fashion, at concentrations above approximately 0.5 mM the population of the intermediate state becomes significant at temperatures between 45 degrees C and 80 degrees C and reaches up to 50% at the largest concentration investigated. The thermodynamic properties of the intermediate state implied by this analysis fall in between those of the unfolded state and the native ones, indicating a considerably disordered conformation, which appears to be stabilized by oligomerization.