Escherichia coli dihydrofolate reductase (DHFR) catalyzes the reduction of dihydrofolate to tetrahydrofolate. During the catalytic cycle, DHFR undergoes conformational transitions between the closed (CS) and occluded (OS) states that, respectively, describe whether the active site is closed or occluded by the Met20 loop. The CS-->OS and the reverse transition may be viewed as allosteric transitions. Using a sequence-based approach, we identify a network of residues that represents the allostery wiring diagram. Many of the residues in the allostery wiring diagram, which are dispersed throughout the adenosine-binding domain as well as the loop domain, are not conserved. Several of the residues in the network have been previously shown by NMR experiments, mutational studies, and molecular dynamics simulations to be linked to equilibration conformational fluctuations of DHFR. To further probe the nature of events that occur during conformational fluctuations, we use a self-organized polymer model to monitor the kinetics of the CS-->OS and the reverse transitions. During the CS-->OS transition, coordinated changes in a number of residues in the loop domain enable the Met20 loop to slide along the alpha-helix in the adenosine-binding domain. Sliding is triggered by pulling of the Met20 loop by the betaG-betaH loop and the pushing action of the betaG-betaH loop. The residues that facilitate the Met20 loop motion are part of the network of residues that transmit allosteric signals during the CS-->OS transition. Replacement of M16 and G121, whose C(alpha) atoms are about 4.3 A in the CS, by a disulfide cross-link impedes that CS-->OS transition. The order of events in the OS-->CS transition is not the reverse of the forward transition. The contact Glu18-Ser49 in the OS persists until the sliding of the Met20 loop is nearly complete. The ensemble of structures in the transition state in both the allosteric transitions is heterogeneous. The most probable transition-state structure resembles the OS (CS) in the CS-->OS (OS-->CS) transition, which is in accord with the Hammond postulate. Structures resembling the OS (CS) are present as minor ( approximately 1-3%) components in equilibrated CS (OS) structures.
The Escherichia coli chaperonin GroEL, which helps proteins to fold, consists of two heptameric rings stacked back-to-back. During the reaction cycle GroEL undergoes a series of allosteric transitions triggered by ligand (substrate protein, ATP, and the cochaperonin GroES) binding. Based on an elastic network model of the bullet-shaped double-ring chaperonin GroEL-(ADP)(7)-GroES structure (R''T state), we perform a normal mode analysis to explore the energetically favorable collective motions encoded in the R''T structure. By comparing each normal mode with the observed conformational changes in the R''T --> TR'' transition, a single dominant normal mode provides a simple description of this highly intricate allosteric transition. A detailed analysis of this relatively high-frequency mode describes the structural and dynamic changes that underlie the positive intra-ring and negative inter-ring cooperativity. The dynamics embedded in the dominant mode entails highly concerted structural motions with approximate preservation of sevenfold symmetry within each ring and negatively correlated ones between the two rings. The dominant normal mode (in comparison with the other modes) is robust to parametric perturbations caused by sequence variations, which validates its functional importance. Response of the dominant mode to local changes that mimic mutations using the structural perturbation method technique leads to a wiring diagram that identifies a network of key residues that regulate the allosteric transitions. Many of these residues are located in intersubunit interfaces, and may therefore play a critical role in transmitting allosteric signals between subunits.
Characterization of the structures of the transition state ensemble is a key step in describing the folding reaction. Using two variants of a coarse-grained model of the three-stranded beta-sheet WW domain and a fully automated progress variable clustering (PVC) algorithm, we have dissected the effect of macromolecular crowding and confinement on the changes in the transition state structures in comparison to bulk. Each amino acid is represented using a Calpha atom and a side chain. The distance between the Calpha atom and center of mass of the side chain is taken to be its effective van der Waals radius. For the bulk case, we predict using the PVC algorithm, which does not assume knowledge of the underlying folding reaction coordinate, that there are two classes of structures in the transition state ensemble (TSE). The structures in both of the classes are compact. The dominant cluster is more structured than the structures in the less populated class. In accord with bulk experiments, the residues in strands beta2 and beta3 and the interactions at the beta2-beta3 interface are structured. When only excluded volume interactions between the crowding particles and the WW domain are taken into account or when the protein is confined to an inert spherical pore, the overall structure of the TSE does not change dramatically. However, in this entropy dominated regime, the width of the TSE decreases and the structures become more oblate and less spherical as the volume fraction of crowding particle increases or when the pore radius decreases. It suggests that the shape changes, which are computed using the moment of inertia tensor, may represent the slow degrees of freedom during the folding process. When non-native interactions between side chains and interactions with the cavity of the pores are taken into account, the TSE becomes considerably broader. Although the topology in the transition has a fold similar to the native state, the structures are far more plastic than in the bulk. The TSE is sensitive to the size of the pore as well as interactions between the pore and the protein. The differences between the two cases (confinement in an inert pore and when pore-protein interactions are considered) arise due to the increased importance of enthalpic interactions in the transition state as the strength of the protein-pore interaction increases.
Nanomanipulation of biomolecules by using single-molecule methods and computer simulations has made it possible to visualize the energy landscape of biomolecules and the structures that are sampled during the folding process. We use simulations and single-molecule force spectroscopy to map the complex energy landscape of GFP that is used as a marker in cell biology and biotechnology. By engineering internal disulfide bonds at selected positions in the GFP structure, mechanical unfolding routes are precisely controlled, thus allowing us to infer features of the energy landscape of the wild-type GFP. To elucidate the structures of the unfolding pathways and reveal the multiple unfolding routes, the experimental results are complemented with simulations of a self-organized polymer (SOP) model of GFP. The SOP representation of proteins, which is a coarse-grained description of biomolecules, allows us to perform forced-induced simulations at loading rates and time scales that closely match those used in atomic force microscopy experiments. By using the combined approach, we show that forced unfolding of GFP involves a bifurcation in the pathways to the stretched state. After detachment of an N-terminal alpha-helix, unfolding proceeds along two distinct pathways. In the dominant pathway, unfolding starts from the detachment of the primary N-terminal beta-strand, while in the minor pathway rupture of the last, C-terminal beta-strand initiates the unfolding process. The combined approach has allowed us to map the features of the complex energy landscape of GFP including a characterization of the structures, albeit at a coarse-grained level, of the three metastable intermediates.
Allosteric interactions between residues that are spatially apart and well separated in sequence are important in the function of multimeric proteins as well as single-domain proteins. This observation suggests that, among the residues that are involved in long-range communications, mutation at one site should affect interactions at a distant site. By adopting a sequence-based approach, we present an automated approach that uses a generalization of the familiar sequence entropy in conjunction with a coupled two-way clustering algorithm, to predict the network of interactions that trigger allosteric interactions in proteins. We use the method to identify the subset of dynamically important residues in three families, namely, the small PDZ family, G protein-coupled receptors (GPCR), and the Lectins, which are cell-adhesion receptors that mediate the tethering and rolling of leukocytes on inflamed endothelium. For the PDZ and GPCR families, our procedure predicts, in agreement with previous studies, a network containing a small number of residues that are involved in their function. Application to the Lectin family reveals a network of residues interspersed throughout the C-terminal end of the structure that are responsible for binding to ligands. Based on our results and previous studies, we propose that functional robustness requires that only a small subset of distantly connected residues be involved in transmitting allosteric signals in proteins.
Dynamics of tRNA was studied using neutron scattering spectroscopy. Despite vast differences in the architecture and backbone structure of proteins and RNA, hydrated tRNA undergoes the dynamic transition at the same temperature as hydrated lysozyme. The similarity of the dynamic transition in RNA and proteins supports the idea that it is solvent induced. Because tRNA essentially has no methyl groups, the results also suggest that methyl groups are not the main contributor of the dynamic transition in biological macromolecules. However, they may explain strong differences in the dynamics of tRNA and lysozyme observed at low temperatures.
We analyze the dependence of cooperativity of the thermal denaturation transition and folding rates of globular proteins on the number of amino acid residues, N, using lattice models with side chains, off-lattice Go models, and the available experimental data. A dimensionless measure of cooperativity, Omega(c) (0 < Omega(c) < infinity), scales as Omega(c) approximately N(zeta). The results of simulations and the analysis of experimental data further confirm the earlier prediction that zeta is universal with zeta = 1 + gamma, where exponent gamma characterizes the susceptibility of a self-avoiding walk. This finding suggests that the structural characteristics in the denaturated state are manifested in the folding cooperativity at the transition temperature. The folding rates k(F) for the Go models and a dataset of 69 proteins can be fit using k(F) = k(F)0 exp(-cN(beta)). Both beta = 1/2 and 2/3 provide a good fit of the data. We find that k(F) = k(F)0 exp(-cN(1/2)), with the average (over the dataset of proteins) k(F)0 approximately (0.2 micros)(-1) and c approximately 1.1, can be used to estimate folding rates to within an order of magnitude in most cases. The minimal models give identical N dependence with c approximately 1. The prefactor for off-lattice Go models is nearly 4 orders of magnitude larger than the experimental value.
Nanomanipulation of individual RNA molecules, using laser optical tweezers, has made it possible to infer the major features of their energy landscape. Time-dependent mechanical unfolding trajectories, measured at a constant stretching force (f(S)) of simple RNA structures (hairpins and three-helix junctions) sandwiched between RNA/DNA hybrid handles show that they unfold in a reversible all-or-none manner. To provide a molecular interpretation of the experiments we use a general coarse-grained off-lattice Gō-like model, in which each nucleotide is represented using three interaction sites. Using the coarse-grained model we have explored forced-unfolding of RNA hairpin as a function of f(S) and the loading rate (r(f)). The simulations and theoretical analysis have been done both with and without the handles that are explicitly modeled by semiflexible polymer chains. The mechanisms and timescales for denaturation by temperature jump and mechanical unfolding are vastly different. The directed perturbation of the native state by f(S) results in a sequential unfolding of the hairpin starting from their ends, whereas thermal denaturation occurs stochastically. From the dependence of the unfolding rates on r(f) and f(S) we show that the position of the unfolding transition state is not a constant but moves dramatically as either r(f) or f(S) is changed. The transition-state movements are interpreted by adopting the Hammond postulate for forced-unfolding. Forced-unfolding simulations of RNA, with handles attached to the two ends, show that the value of the unfolding force increases (especially at high pulling speeds) as the length of the handles increases. The pathways for refolding of RNA from stretched initial conformation, upon quenching f(S) to the quench force f(Q), are highly heterogeneous. The refolding times, upon force-quench, are at least an order-of-magnitude greater than those obtained by temperature-quench. The long f(Q)-dependent refolding times starting from fully stretched states are analyzed using a model that accounts for the microscopic steps in the rate-limiting step, which involves the trans to gauche transitions of the dihedral angles in the GAAA tetraloop. The simulations with explicit molecular model for the handles show that the dynamics of force-quench refolding is strongly dependent on the interplay of their contour length and persistence length and the RNA persistence length. Using the generality of our results, we also make a number of precise experimentally testable predictions.
Mechanical folding trajectories for polyproteins starting from initially stretched conformations generated by single-molecule atomic force microscopy experiments [Fernandez, J. M. & Li, H. (2004) Science 303, 1674-1678] show that refolding, monitored by the end-to-end distance, occurs in distinct multiple stages. To clarify the molecular nature of folding starting from stretched conformations, we have probed the folding dynamics, upon force quench, for the single I27 domain from the muscle protein titin by using a C(alpha)-Go model. Upon temperature quench, collapse and folding of I27 are synchronous. In contrast, refolding from stretched initial structures not only increases the folding and collapse time scales but also decouples the two kinetic processes. The increase in the folding times is associated primarily with the stretched state to compact random coil transition. Surprisingly, force quench does not alter the nature of the refolding kinetics, but merely increases the height of the free-energy folding barrier. Force quench refolding times scale as tau(F) approximately tau(F)(0)exp(f(q)Deltax(f)/k(B)T), where Deltax(f) approximately 0.6 nm is the location of the average transition state along the reaction coordinate given by end-to-end distance. We predict that tau(F) and the folding mechanism can be dramatically altered by the initial and/or final values of force. The implications of our results for design and analysis of experiments are discussed.
We have studied the stability and the yield of the folded WW domains in a spherical nanopore to provide insights into the changes in the folding characteristics due to interactions of the polypeptide (SP) with the walls of the pore. Using different models for the interactions between the nanopore and the polypeptide chain we have obtained results that are relevant to a broad range of experiments. (a) In the temperature and the strength of the SP-pore interaction plane (lambda), there are four "phases," namely, the unfolded state, the native state, the molten globule phase (MG), and the surface interaction-stabilized (SIS) state. The MG and SIS states are populated at moderate and large values of lambda, respectively. For a fixed pore size, the folding rates vary non-monotonically as lambda is varied with a maximum at lambda approximately 1 at which the SP-nanopore interaction is comparable to the stability of the native state. At large lambda values, the WW domain is kinetically trapped in the SIS states. Using multiple sequence alignment, we conclude that similar folding mechanism should be observed in other WW domains as well. (b) To mimic the changes in the nature of the allosterically driven SP-GroEL interactions we consider two models for the dynamic Anfinsen cage (DAC). In DAC1, the SP-cavity interaction cycles between hydrophobic (lambda>0) and hydrophilic (lambda=0) with a period tau. The yield of the native state is a maximum for an optimum value of tau=tau(OPT). At tau=tau(OPT), the largest yield of the native state is obtained when tau(H) approximately tau(P) where tau(H)(tau(P)) is the duration for which the cavity is hydrophobic (hydrophilic). Thus, in order to enhance the native state yield, the cycling rate, for a given loading rate of the GroEL nanomachine, should be maximized. In DAC2, the volume of the cavity is doubled (as happens when ATP and GroES bind to GroEL) and the SP-pore interaction simultaneously changes from hydrophobic to hydrophilic. In this case, we find greater increase in yield of the native state compared to DAC1 at all values of tau.
The self-assembly of RNA structure depends on the interactions of counterions with the RNA and with each other. Comparison of various polyamines showed that the tertiary structure of the Tetrahymena ribozyme is more stable when the counterions are small and highly charged. By monitoring the folding kinetics of the ribozyme as a function of polyamine concentration, we now find that the charge density of the counterions determines the positions of the folding transition states. The transition state ensemble (TSE) between U and N moves away from the native state as the counterion valence and charge density increase, as predicted by the Hammond postulate. The TSE is broader and less structured when the RNA is refolded in polyamines rather than Mg2+. That the charge density of the counterions determines the plasticity of the TSE demonstrates the importance of interactions among condensed counterions for the self-assembly of RNA structures. We propose that the major barrier to RNA folding is dominated by entropy changes when counterion charge density is low and enthalpy differences when it is high.
A number of situations such as protein folding in confined spaces, lubrication in tight spaces, and chemical reactions in confined spaces require an understanding of water-mediated interactions. As an illustration of the profound effects of confinement on hydrophobic and ionic interactions, we investigate the solvation of methane and methane decorated with charges in spherically confined water droplets. Free energy profiles for a single methane molecule in droplets, ranging in diameter (D) from 1 to 4 nm, show that the droplet surfaces are strongly favorable as compared to the interior. From the temperature dependence of the free energy in D = 3 nm, we show that this effect is entropically driven. The potentials of mean force (PMFs) between two methane molecules show that the solvent separated minimum in the bulk is completely absent in confined water, independent of the droplet size since the solute particles are primarily associated with the droplet surface. The tendency of methanes with charges (M(q+) and M(q-) with q(+) = |q(-)| = 0.4e, where e is the electronic charge) to be pinned at the surface depends dramatically on the size of the water droplet. When D = 4 nm, the ions prefer the interior whereas for D < 4 nm the ions are localized at the surface, but with much less tendency than for methanes. Increasing the ion charge to e makes the surface strongly unfavorable. Reflecting the charge asymmetry of the water molecule, negative ions have a stronger preference for the surface compared to positive ions of the same charge magnitude. With increasing droplet size, the PMFs between M(q+) and M(q-) show decreasing influence of the boundary owing to the reduced tendency for surface solvation. We also show that as the solute charge density decreases the surface becomes less unfavorable. The implications of our results for the folding of proteins in confined spaces are outlined.
Loop formation between monomers in the interior of semiflexible chains describes elementary events in biomolecular folding and DNA bending. We calculate analytically the interior distance distribution function for semiflexible chains using a mean field approach. Using the potential of mean force derived from the distance distribution function we present a simple expression for the kinetics of interior looping by adopting Kramers theory. For the parameters, that are appropriate for DNA, the theoretical predictions in comparison with the case are in excellent agreement with explicit Brownian dynamics simulations of wormlike chain (WLC) model. The interior looping times (tauIC) can be greatly altered in the cases when the stiffness of the loop differs from that of the dangling ends. If the dangling end is stiffer than the loop then tauIC increases for the case of the WLC with uniform persistence length. In contrast, attachment of flexible dangling ends enhances rate of interior loop formation. The theory also shows that if the monomers are charged and interact via screened Coulomb potential then both the cyclization (tauc) and interior looping (tauIC) times greatly increase at low ionic concentration. Because both tauc and tauIC are determined essentially by the effective persistence length [lp(R)] we computed lp(R) by varying the range of the repulsive interaction between the monomers. For short range interactions lp(R) nearly coincides with the bare persistence length which is determined largely by the backbone chain connectivity. This finding rationalizes the efficacy of describing a number of experimental observations (response of biopolymers to force and cyclization kinetics) in biomolecules using WLC model with an effective persistence length.
By representing the high-resolution crystal structures of a number of enzymes using the elastic network model, it has been shown that only a few low-frequency normal modes are needed to describe the large-scale domain movements that are triggered by ligand binding. Here we explore a link between the nearly invariant nature of the modes that describe functional dynamics at the mesoscopic level and the large evolutionary sequence variations at the residue level. By using a structural perturbation method (SPM), which probes the residue-specific response to perturbations (or mutations), we identify a sparse network of strongly conserved residues that transmit allosteric signals in three structurally unrelated biological nanomachines, namely, DNA polymerase, myosin motor, and the Escherichia coli chaperonin. Based on the response of every mode to perturbations, which are generated by interchanging specific sequence pairs in a multiple sequence alignment, we show that the functionally relevant low-frequency modes are most robust to sequence variations. Our work shows that robustness of dynamical modes at the mesoscopic level is encoded in the structure through a sparse network of residues that transmit allosteric signals.
We present, to our knowledge, a new theory that takes internal dynamics of proteins into account to describe forced-unfolding and force-quench refolding in single molecule experiments. In the current experimental setup (using either atomic force microscopy or laser optical tweezers) the distribution of unfolding times, P(t), is measured by applying a constant stretching force f(S) from which the apparent f(S)-dependent unfolding rate is obtained. To describe the complexity of the underlying energy landscape requires additional probes that can incorporate the dynamics of tension propagation and relaxation of the polypeptide chain upon force quench. We introduce a theory of force correlation spectroscopy to map the parameters of the energy landscape of proteins. In force correlation spectroscopy, the joint distribution P(T, t) of folding and unfolding times is constructed by repeated application of cycles of stretching at constant f(S) separated by release periods T during which the force is quenched to f(Q) < f(S). During the release period, the protein can collapse to a manifold of compact states or refold. We show that P(T, t) at various f(S) and f(Q) values can be used to resolve the kinetics of unfolding as well as formation of native contacts. We also present methods to extract the parameters of the energy landscape using chain extension as the reaction coordinate and P(T, t). The theory and a wormlike chain model for the unfolded states allows us to obtain the persistence length l(p) and the f(Q)-dependent relaxation time, giving us an estimate of collapse timescale at the single molecular level, in the coil states of the polypeptide chain. Thus, a more complete description of landscape of protein native interactions can be mapped out if unfolding time data are collected at several values of f(S) and f(Q). We illustrate the utility of the proposed formalism by analyzing simulations of unfolding-refolding trajectories of a coarse-grained protein (S1) with beta-sheet architecture for several values of f(S), T, and f(Q) = 0. The simulations of stretch-relax trajectories are used to map many of the parameters that characterize the energy landscape of S1.
We have used a bioinformatic approach to predict the natural substrate proteins for the Escherichia coli chaperonin GroEL based on two simple criteria. Natural substrate proteins should contain binding motifs similar in sequence to the mobile loop peptide of GroES that displaces the binding motif during the chaperonin cycle. Secondly, each substrate protein should contain multiple copies of the binding motif so that the chaperonin can perform "work" on the substrate protein. To validate these criteria, we have used a database of 252 proteins that have been experimentally shown to interact with the chaperonin machinery in vivo. More than 80% are identified by these criteria. The binding motifs of all 79 proteins in the database with a known three-dimensional structure are buried (<50% solvent-accessible surface area) in the native state. Our results show that the binding motifs are inaccessible in the native state but become solvent-exposed in unfolded state, thus enabling GroEL to distinguish between unfolded and native states. The structures of the binding motif in the native states of the substrate proteins include alpha-helices, beta-strands, and random coils. The diversity of secondary structures implies that there are large and varied conformational transitions in the recognition motifs after their displacement by the mobile loops of GroES.
Systematic studies based on 1H NMR and 13C NMR indicated that the alkylthio group behaves as a weak electron-withdrawing group in a simple aniline system like 2-butylthioaniline, while the same alkylthio group clearly acted as a resonance electron-donating group in higher conjugated aniline trimer systems, like butylthio-substituted PDA (mono-PDA) and dibutylthio-substituted PDA (2,6-diPDA). The formation of 2,6-diPDA as the major byproduct during the preparation of mono-PDA from PDI and butane-1-thiol provided additional support for the resonance electron donating nature of the butylthio group in these aniline trimer systems. Furthermore, CV studies also clearly indicated that the redox potential E degrees (vs. SCE) of the aniline trimer systems decreased with the increase in the number of butylthio groups, further confirming the electron-donating nature of the butylthio group in these higher conjugated trimer systems.
Atomic force measurements of unbinding rates (or off-rates) of ligands bound to a class of cell adhesion molecules from the selectin family show a transition from catch to slip bonds as the value of external force (f) is increased. At low forces (<10 pN), the unbinding rates decrease (catch regime), while, at high forces, the rates increase in accord with the Bell model (slip regime). The energy landscape underlying the catch-slip transition can be captured by a two-state model that considers the possibility of redistribution of population from the force-free bound state to the force-stabilized bound state. The excellent agreement between theory and experiments is used to extract the parameters characterizing the energy landscape of the complex by fitting the calculated curves to lifetime data (obtained at constant f) for the monomeric form of PSGL-1 (sPSGL-1). We used the constant force parameters to predict the distributions of unbinding times and unbinding forces as a function of the loading rate. The general two-state model, which also correctly predicts the absence of catch bonds in the binding of antibodies to selectins, is used to resolve the energy landscape parameters characterizing adhesive interactions of P- and L-selectins with physiological ligands such as sPSGL-1 and endoglycan and antibodies such as G1 and DREG56. Despite high sequence similarity, the underlying shapes of the energy landscape of P-selectin and L-selectin interacting with sPSGL-1 are markedly different. The underlying energy landscape of the selectin cell adhesion complex is sensitive to the nature of the ligand. The unified description of selectins bound to physiological ligands and antibodies in conjunction with experimental data can be used to extract the key parameters that describe the dynamics of cell adhesion complexes.
The chaperonin GroEL-GroES, a machine that helps proteins to fold, cycles through a number of allosteric states, the T state, with high affinity for substrate proteins, the ATP-bound R state, and the R" (GroEL-ADP-GroES) complex. Here, we use a self-organized polymer model for the GroEL allosteric states and a general structure-based technique to simulate the dynamics of allosteric transitions in two subunits of GroEL and the heptamer. The T --> R transition, in which the apical domains undergo counterclockwise motion, is mediated by a multiple salt-bridge switch mechanism, in which a series of salt-bridges break and form. The initial event in the R -->R" transition, during which GroEL rotates clockwise, involves a spectacular outside-in movement of helices K and L that results in K80-D359 salt-bridge formation. In both the transitions there is considerable heterogeneity in the transition pathways. The transition state ensembles (TSEs) connecting the T, R, and R" states are broad with the TSE for the T --> R transition being more plastic than the R --> R" TSE.
In the amyloid fibrils formed from long fragments of the amyloid beta-protein (Abeta-protein), the monomers are arranged in parallel and lie perpendicular to the fibril axis. The structure of the monomers satisfies the amyloid self-organization principle; namely, the low free energy state of the monomer maximizes the number of intra- and interpeptide contacts and salt bridges. The formation of the intramolecular salt bridge between Asp(D)23 and Lys(K)28 ensures that unpaired charges are not buried in the low-dielectric interior. We have investigated, using all-atom molecular dynamics simulations in explicit water, whether the D23-K28 interaction forms spontaneously in the isolated Abeta10-35 monomer. To validate the simulation protocol, we show, using five independent trajectories spanning a total of 100 ns, that the pKa values of the titratable groups are in good agreement with experimental measurements. The computed free energy disconnectvity graph shows that broadly the ensemble of compact random coil conformations can be clustered into four basins that are separated by free energy barriers ranging from 0.3 to 2.7 kcal/mol. There is significant residual structure in the conformation of the peptide in each of the basins. Due to the desolvation penalty, the structural motif with a stable turn involving the residues VGSN and a preformed D23-K28 contact is a minor component of the simulated structures. The extent of solvation of the peptides in the four basins varies greatly, which underscores the dynamical fluctuations in the monomer. Our results suggest that the early event in the oligomerization process must be the expulsion of discrete water molecules that facilitates the formation of interpeptide-interaction-driven stable structures with an intramolecular D23-K28 salt bridge and an intact VGSN turn.