Polymer principles and protein folding

This paper surveys the emerging role of statistical mechanics and polymer theory in protein folding. In the polymer perspective, the folding code is more a solvation code than a code of local phipsi propensities. The polymer perspective resolves two classic puzzles: (1) the Blind Watchmaker's Paradox that biological proteins could not have originated from random sequences, and (2) Levinthal's Paradox that the folded state of a protein cannot be found by random search. Both paradoxes are traditionally framed in terms of random unguided searches through vast spaces, and vastness is equated with impossibility. But both processes are partly guided. The searches are more akin to balls rolling down funnels than balls rolling aimlessly on flat surfaces. In both cases, the vastness of the search is largely irrelevant to the search time and success. These ideas are captured by energy and fitness landscapes. Energy landscapes give a language for bridging between microscopics and macroscopics, for relating folding kinetics to equilibrium fluctuations, and for developing new and faster computational search strategies.

The effect of multiple binding modes on empirical modeling of ligand docking to proteins

A popular approach to the computational modeling of ligand/receptor interactions is to use an empirical free energy like model with adjustable parameters. Parameters are learned from one set of complexes, then used to predict another set. To improve these empirical methods requires an independent way to study their inherent errors. We introduce a toy model of ligand/receptor binding as a workbench for testing such errors. We study the errors incurred from the two state binding assumption--the assumption that a ligand is either bound in one orientation, or unbound. We find that the two state assumption can cause large errors in free energy predictions, but it does not affect rank order predictions significantly. We show that fitting parameters using data from high affinity ligands can reduce two state errors; so can using more physical models that do not use the two state assumption. We also find that when using two state models to predict free energies, errors are more severe on high affinity ligands than low affinity ligands. And we show that two state errors can be diagnosed by systematically adding new binding modes when predicting free energies: if predictions worsen as the modes are added, then the two state assumption in the fitting step may be at fault.

Predicting the structures of 18 peptides using Geocore

We describe an extensive test of Geocore, an ab initio peptide folding algorithm. We studied 18 short molecules for which there are structures in the Protein Data Bank; chains are up to 31 monomers long. Except for the very shortest peptides, an extremely simple energy function is sufficient to discriminate the true native state from more than 10(8) lowest energy conformations that are searched explicitly for each peptide. A high incidence of native-like structures is found within the best few hundred conformations generated by Geocore for each amino acid sequence. Predictions improve when the number of discrete phi/psi choices is increased.

Native protein fluctuations: the conformational-motion temperature and the inverse correlation of protein flexibility with protein stability

We study the fluctuations of native proteins by exact enumeration using the HP lattice model. The model fluctuations increase with temperature. We observe a low-temperature point, below which large fluctuations are frozen out. This prediction is consistent with the observation by Tilton et al. [R. F. Tilton, Jr., J. C. Dewan, and G. A. Petsko, Biochemistry 31, 2469 (1992)], that the thermal motions of ribonuclease A increase sharply above about 200 K. We also explore protein "flexibility" as defined by Debye-Waller-like factors and solvent accessibilities of core residues to hydrogen exchange. We find that proteins having greater stability tend to have fewer large fluctuations, and hence lower flexibilities. If flexibility is necessary for enzyme catalysis, this could explain why proteins from thermophilic organisms, which are exceptionally stable, may be catalytically inactive at normal temperatures.

NMR determination of the major solution conformation of a peptoid pentamer with chiral side chains

Polymers of N-substituted glycines ("peptoids") containing chiral centers at the alpha position of their side chains can form stable structures in solution. We studied a prototypical peptoid, consisting of five para-substituted (S)-N-(1-phenylethyl)glycine residues, by NMR spectroscopy. Multiple configurational isomers were observed, but because of extensive signal overlap, only the major isomer containing all cis-amide bonds was examined in detail. The NMR data for this molecule, in conjunction with previous CD spectroscopic results, indicate that the major species in methanol is a right-handed helix with cis-amide bonds. The periodicity of the helix is three residues per turn, with a pitch of approximately 6 A. This conformation is similar to that anticipated by computational studies of a chiral peptoid octamer. The helical repeat orients the amide bond chromophores in a manner consistent with the intensity of the CD signal exhibited by this molecule. Many other chiral polypeptoids have similar CD spectra, suggesting that a whole family of peptoids containing chiral side chains is capable of adopting this secondary structure motif. Taken together, our experimental and theoretical studies of the structural properties of chiral peptoids lay the groundwork for the rational design of more complex polypeptoid molecules, with a variety of applications, ranging from nanostructures to nonviral gene delivery systems.

Sequence-specific polypeptoids: a diverse family of heteropolymers with stable secondary structure

We have synthesized and characterized a family of structured oligo-N-substituted-glycines (peptoids) up to 36 residues in length by using an efficient solid-phase protocol to incorporate chemically diverse side chains in a sequence-specific fashion. We investigated polypeptoids containing side chains with a chiral center adjacent to the main chain nitrogen. Some of these sequences have stable secondary structure, despite the achirality of the polymer backbone and its lack of hydrogen bond donors. In both aqueous and organic solvents, peptoid oligomers as short as five residues give rise to CD spectra that strongly resemble those of peptide alpha-helices. Differential scanning calorimetry and CD measurements show that polypeptoid secondary structure is highly stable and that unfolding is reversible and cooperative. Thermodynamic parameters obtained for unfolding are similar to those obtained for the alpha-helix to coil transitions of peptides. This class of biomimetic polymers may enable the design of self-assembling macromolecules with novel structures and functions.

Protein folding in the landscape perspective: chevron plots and non-Arrhenius kinetics

We use two simple models and the energy landscape perspective to study protein folding kinetics. A major challenge has been to use the landscape perspective to interpret experimental data, which requires ensemble averaging over the microscopic trajectories usually observed in such models. Here, because of the simplicity of the model, this can be achieved. The kinetics of protein folding falls into two classes: multiple-exponential and two-state (single-exponential) kinetics. Experiments show that two-state relaxation times have "chevron plot" dependences on denaturant and non-Arrhenius dependences on temperature. We find that HP and HP+ models can account for these behaviors. The HP model often gives bumpy landscapes with many kinetic traps and multiple-exponential behavior, whereas the HP+ model gives more smooth funnels and two-state behavior. Multiple-exponential kinetics often involves fast collapse into kinetic traps and slower barrier climbing out of the traps. Two-state kinetics often involves entropic barriers where conformational searching limits the folding speed. Transition states and activation barriers need not define a single conformation; they can involve a broad ensemble of the conformations searched on the way to the native state. We find that unfolding is not always a direct reversal of the folding process.

Protein structure and energy landscape dependence on sequence using a continuous energy function

We have recently described a new conformational search strategy for protein folding algorithms called the CGU (convex global underestimator) method. Here we use a simplified protein chain representation and a differentiable form of the Sun/Thomas/Dill energy function to test the CGU method. Standard search methods, such as Monte Carlo and molecular dynamics are slowed by kinetic traps. That is, the computer time depends more strongly on the shape of the energy landscape (dictated by the amino acid sequence) than on the number of degrees of freedom (dictated by the chain length). The CGU method is not subject to this limitation, since it explores the underside of the energy landscape, not the top. We find that the CGU computer time is largely independent of the monomer sequence for different chain folds and scales as O(n4) with chain length. By using different starting points, we show that the method appears to find global minima. Since we can currently find stable states of 36-residue chains in 2.4 hours, the method may be practical for small proteins.

Ligand binding to proteins: the binding landscape model

Models of ligand binding are often based on four assumptions: (1) steric fit: that binding is determined mainly by shape complementarity; (2) native binding: that ligands mainly bind to native states; (3) locality: that ligands perturb protein structures mainly at the binding site; and (4) continuity: that small changes in ligand or protein structure lead to small changes in binding affinity. Using a generalization of the 2D HP lattice model, we study ligand binding and explore these assumptions. We first validate the model by showing that it reproduces typical binding behaviors. We observe ligand-induced denaturation, ANS and heme-like binding, and "lock-and-key" and "induced-fit" specific binding behaviors characterized by Michaelis-Menten or more cooperative types of binding isotherms. We then explore cases where the model predicts violations of the standard assumptions. For example, very different binding modes can result from two ligands of identical shape. Ligands can sometimes bind highly denatured states more tightly than native states and yet have Michaelis-Menten isotherms. Even low-population binding to denatured states can cause changes in global stability, hydrogen-exchange rates, and thermal B-factors, contrary to expectations, but in agreement with experiments. We conclude that ligand binding, similar to protein folding, may be better described in terms of energy landscapes than in terms of simpler mass-action models.

Solvation: how to obtain microscopic energies from partitioning and solvation experiments

Oil-water partitioning, solubilities, and vapor pressure experiments on small-molecule compounds are often used as models to obtain energies for biomolecular modeling. For example, measured partition coefficients, K, are often inserted into the formula -RT in K to obtain quantities thought to represent microscopic contact interaction free energies. We review evidence here that this procedure does not always give microscopically meaningful free energies. Some partitioning processes, particularly involving polymeric solvents such as octanol or hexadecane, are governed not only by translational entropies and contact interactions, but also by free energies resulting from changes in the conformations of the polymer chains upon solute insertion. The Flory-Huggins theory is more suitable for these situations than is the classical approach. We discuss the physical bases for both approaches.

From Levinthal to pathways to funnels

While the classical view of protein folding kinetics relies on phenomenological models, and regards folding intermediates in a structural way, the new view emphasizes the ensemble nature of protein conformations. Although folding has sometimes been regarded as a linear sequence of events, the new view sees folding as parallel microscopic multi-pathway diffusion-like processes. While the classical view invoked pathways to solve the problem of searching for the needle in the haystack, the pathway idea was then seen as conflicting with Anfinsen's experiments showing that folding is pathway-independent (Levinthal's paradox). In contrast, the new view sees no inherent paradox because it eliminates the pathway idea: folding can funnel to a single stable state by multiple routes in conformational space. The general energy landscape picture provides a conceptual framework for understanding both two-state and multi-state folding kinetics. Better tests of these ideas will come when new experiments become available for measuring not just averages of structural observables, but also correlations among their fluctuations. At that point we hope to learn much more about the real shapes of protein folding landscapes.

An iterative method for extracting energy-like quantities from protein structures

We present a method (ENERGI) for extracting energy-like quantities from a data base of protein structures. In this paper, we use the method to generate pairwise additive amino acid "energy" scores. These scores are obtained by iteration until they correctly discriminate a set of known protein folds from decoy conformations. The method succeeds in lattice model tests and in the gapless threading problem as defined by Maiorov and Crippen [Maiorov, V. N. & Crippen, G. M. (1992) J. Mol. Biol. 227, 876-888]. A more challenging test of threading a larger set of test proteins derived from the representative set of Hobohm and Sander [Hobohm, U. & Sander, C. (1994) Protein Sci. 3, 522-524] is used as a "workbench" for exploring how the ENERGI scores depend on their parameter sets.

A fast conformational search strategy for finding low energy structures of model proteins

We describe a new computer algorithm for finding low-energy conformations of proteins. It is a chain-growth method that uses a heuristic bias function to help assemble a hydrophobic core. We call it the Core-directed chain Growth method (CG). We test the CG method on several well-known literature examples of HP lattice model proteins [in which proteins are modeled as sequences of hydrophobic (H) and polar (P) monomers], ranging from 20-64 monomers in two dimensions, and up to 88-mers in three dimensions. Previous nonexhaustive methods--Monte Carlo, a Genetic Algorithm, Hydrophobic Zippers, and Contact Interactions--have been tried on these same model sequences. CG is substantially better at finding the global optima, and avoiding local optima, and it does so in comparable or shorter times. CG finds the global minimum energy of the longest HP lattice model chain for which the global optimum is known, a 3D 88-mer that has only been reachable before by the CHCC complete search method. CG has the potential advantage that it should have nonexponential scaling with chain length. We believe this is a promising method for conformational searching in protein folding algorithms.

Binding of ionic ligands to polyelectrolytes

Ionic ligands can bind to polyelectrolytes such as DNA or charged polysaccharides. We develop a Poisson-Boltzmann treatment to compute binding constants as a function of ligand charge and salt concentration in the limit of low ligand concentration. For flexible chain ligands, such as oligopeptides, we treat their conformations using lattice statistics. The theory predicts the salt dependence and binding free energies, of Mg(2+) ions to polynucleotides, of hexamine cobalt(III) to calf thymus DNA, of polyamines to T7 DNA, of oligolysines to poly(U) and poly(a), and of tripeptides to heparin, a charged polysaccharide. One parameter is required to obtain absolute binding constants, the distance of closest separation of the ligand to the polyion. Some, but not all, of the binding entropies and enthalpies are also predicted accurately by the model.

Statistical potentials extracted from protein structures: how accurate are they?

"Statistical potentials" are energies widely used in computer algorithms to fold, dock, or recognize protein structures. They are derived from: (1) observed pairing frequencies of the 20 amino acids in databases of known protein structures, and (2) approximations and assumptions about the physical process that these quantities measure. Using exact lattice models, we construct a rigorous test of those assumptions and approximations. We find that statistical potentials often correctly rank-order the relative strengths of interresidue interactions, but they do not reflect the true underlying energies because of systematic errors arising from the neglect of excluded volume in proteins. We find that complex residue-residue distance dependences observed in statistical potentials, even those among charged groups, can be largely explained as an indirect consequence of the burial of non-polar groups. Our results suggest that current statistical potentials may have limited value in protein folding algorithms and wherever they are used to provide energy-like quantities.

Comparing folding codes for proteins and polymers

Proteins fold to unique compact native structures. Perhaps other polymers could be designed to fold in similar ways. The chemical nature of the monomer "alphabet" determines the "energy matrix" of monomer interactions-which defines the folding code, the relationship between sequence and structure. We study two properties of energy matrices using two-dimensional lattice models: uniqueness, the number of sequences that fold to only one structure, and encodability, the number of folds that are unique lowest-energy structures of certain monomer sequences. For the simplest model folding code, involving binary sequences of H (hydrophobic) and P (polar) monomers, only a small fraction of sequences fold uniquely, and not all structures can be encoded. Adding strong repulsive interactions results in a folding code with more sequences folding uniquely and more designable folds. Some theories suggest that the quality of a folding code depends only on the number of letters in the monomer alphabet, but we find that the energy matrix itself can be at least as important as the size of the alphabet. Certain multi-letter codes, including some with 20 letters, may be less physical or protein-like than codes with smaller numbers of letters because they neglect correlations among inter-residue interactions, treat only maximally compact conformations, or add arbitrary energies to the energy matrix.

A simple model of chaperonin-mediated protein folding

Chaperonins are oligomeric proteins that help other proteins fold. They act, according to the "Anfinsen cage" or "box of infinite dilution" model, to provide private space, protected from aggregation, where a protein can fold. Recent evidence indicates, however, that proteins are often ejected from the GroEL chaperonin in nonnative conformations, and repeated cycles of binding and ejection are needed for successful folding. Some experimental evidence suggests that GroEL chaperonins can act as folding "catalysts" in an ATP-dependent manner even when no aggregation takes place. This implies that chaperonins must somehow recognize the kinetically trapped intermediate states of a protein. A central puzzle is how a chaperonin can catalyze the folding reaction of a broad spectrum of different proteins. We propose a physical mechanism by which chaperonins can flatten the energy barriers to folding in a nonspecific way. Using a lattice model, we illustrate how a chaperonin could provide a sticky surface that helps pull apart an incorrectly folded protein so it can try again to fold. Depending on the relative sizes of the protein and the chaperonin cavity, folding can proceed both inside and outside the chaperonin. Consistent with experiments, we find that the folding rate and amount of native protein can be considerably enhanced, or sometimes reduced, depending on the amino acid sequence, the chaperonin size, and the binding and ejection rates from the chaperonin.

Folding proteins with a simple energy function and extensive conformational searching

We describe a computer algorithm for predicting the three-dimensional structures of proteins using only their amino acid sequences. The method differs from others in two ways: (1) it uses very few energy parameters, representing hydrophobic and polar interactions, and (2) it uses a new "constraint-based exhaustive" searching method, which appears to be among the fastest and most complete search methods yet available for realistic protein models. It finds a relatively small number of low-energy conformations, among which are native-like conformations, for crambin (1CRN), avian pancreatic polypeptide (1PPT), melittin (2MLT), and apamin. Thus, the lowest-energy states of very simple energy functions may predict the native structures of globular proteins.

Designing amino acid sequences to fold with good hydrophobic cores

We present two methods for designing amino acid sequences of proteins that will fold to have good hydrophobic cores. Given the coordinates of the desired target protein or polymer structure, the methods generate sequences of hydrophobic (H) and polar (P) monomers that are intended to fold to these structures. One method designs hydrophobic inside, polar outside; the other minimizes an energy function in a sequence evolution process. The sequences generated by these methods agree at the level of 60-80% of the sequence positions in 20 proteins in the Protein Data Bank. A major challenge in protein design is to create sequences that can fold uniquely, i.e. to a single conformation rather than to many. While an earlier lattice-based sequence evolution method was shown not to design unique folders, our method generates unique folders in lattice model tests. These methods may also be useful in designing other types of foldable polymer not based on amino acids.

A statistical mechanical model for hydrogen exchange in globular proteins

We develop a statistical mechanical theory for the mechanism of hydrogen exchange in globular proteins. Using the HP lattice model, we explore how the solvent accessibilities of chain monomers vary as proteins fluctuate from their stable native conformations. The model explains why hydrogen exchange appears to involve two mechanisms under different conditions of protein stability: (1) a "global unfolding" mechanism by which all protons exchange at a similar rate, approaching that of the denatured protein, and (2) a "stable-state" mechanism by which protons exchange at rates that can differ by many orders of magnitude. There has been some controversy about the stable-state mechanism: does exchange take place inside the protein by solvent penetration, or outside the protein by the local unfolding of a subregion? The present model indicates that the stable-state mechanism of exchange occurs through an ensemble of conformations, some of which may bear very little resemblance to the native structure. Although most fluctuations are small-amplitude motions involving solvent penetration or local unfolding, other fluctuations (the conformational distant relatives) can involve much larger transient excursions to completely different chain folds.

A simple protein folding algorithm using a binary code and secondary structure constraints

We describe an algorithm to predict tertiary structures of small proteins. In contrast to most current folding algorithms, it uses very few energy parameters. Given the secondary structural elements in the sequence--alpha-helices and beta-strands--the algorithm searches the remaining conformational space of a simplified real-space representation of chains to find a minimum energy of an exceedingly simple potential function. The potential is based only on a single type of favorable interaction between hydrophobic residues, an unfavorable excluded volume term of spatial overlaps and, for sheet proteins, an interstrand hydrogen bond interaction. Where appropriate, the known disulfide bonds are constrained by a square-law potential. Conformations are searched by a genetic algorithm. The model predicts reasonably well the known tertiary folds of seven out of the 10 small proteins we consider. We draw two conclusions. First, for the proteins we tested, this exceedingly simple potential function is no worse than others having hundreds of energy parameters in finding the right general tertiary structures. Second, despite its simplicity, the potential function is not the weak link in this algorithm. Differences between our predicted structures and the correct targets can be ascribed to shortcomings in our search strategy. This potential function may be useful for testing other conformational search strategies.

Models of cooperativity in protein folding

What is the basis for the two-state cooperativity of protein folding? Since the 1950s, three main models have been put forward. 1. In 'helix-coil' theory, cooperativity is due to local interactions among near neighbours in the sequence. Helix-coil cooperativity is probably not the principal basis for the folding of globular proteins because it is not two-state, the forces are weak, it does not account for sheet proteins, and there is no evidence that helix formation precedes the formation of a hydrophobic core in the following pathways. 2. In the 'sidechain packing' model, cooperativity is attributed to the jigsaw-puzzle-like complementary fits of sidechains. This too is probably not the basis of folding cooperativity because exact models and experiments on homopolymers with sidechains give no evidence that sidechain freezing is two-state, sidechain complementarities in proteins are only weak trends, and the molten globule model predicted by this model is far more native-like than experiments indicate. 3. In the 'hydrophobic core collapse' model, cooperativity is due to the assembly of non-polar residues into a good core. Exact model studies show that this model gives two-state behaviour for some sequences of hydrophobic and polar monomers. It is based on strong forces. There is considerable experimental evidence for the kinetics this model predicts: the development of hydrophobic clusters and cores is concurrent with secondary structure formation. It predicts compact denatured states with sizes and degrees of disorder that are in reasonable agreement with experiments.

Principles of protein folding--a perspective from simple exact models

General principles of protein structure, stability, and folding kinetics have recently been explored in computer simulations of simple exact lattice models. These models represent protein chains at a rudimentary level, but they involve few parameters, approximations, or implicit biases, and they allow complete explorations of conformational and sequence spaces. Such simulations have resulted in testable predictions that are sometimes unanticipated: The folding code is mainly binary and delocalized throughout the amino acid sequence. The secondary and tertiary structures of a protein are specified mainly by the sequence of polar and nonpolar monomers. More specific interactions may refine the structure, rather than dominate the folding code. Simple exact models can account for the properties that characterize protein folding: two-state cooperativity, secondary and tertiary structures, and multistage folding kinetics--fast hydrophobic collapse followed by slower annealing. These studies suggest the possibility of creating "foldable" chain molecules other than proteins. The encoding of a unique compact chain conformation may not require amino acids; it may require only the ability to synthesize specific monomer sequences in which at least one monomer type is solvent-averse.