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.

Cooperativity in protein-folding kinetics.

How does a protein find its native state without a globally exhaustive search? We propose the "HZ" (hydrophobic zipper) hypothesis: hydrophobic contacts act as constraints that bring other contacts into spatial proximity, which then further constrain and zip up the next contacts, etc. In contrast to helix-coil cooperativity, HZ-heteropolymer collapse cooperativity is driven by nonlocal interactions, causes sheet and irregular conformations in addition to helices, leads to secondary structures concurrently with early hydrophobic core formation, is much more sequence dependent than helix-coil processes, and involves compact intermediate states that have much secondary--but little tertiary--structure. Hydrophobic contacts in the 1992 Protein Data Bank have the type of "topological localness" predicted by the hypothesis. The HZ paths for amino acid sequences that mimic crambin and bovine pancreatic trypsin inhibitor are quickly found by computer; the best configurations thus reached have single hydrophobic cores that are within about 3 kcal/mol of the global minimum. This hypothesis shows how proteins could find globally optimal states without exhaustive search.

Modeling compact denatured states of proteins.

We propose a model for the conformations of compact denatured states of globular proteins: that they are broad ensembles of chain backbone conformations that involve common localized hydrophobic clustering and helical contacts, depending on the amino acid sequence. We construct representative ensembles for chain lengths up to 136 monomers on three-dimensional cubic lattices using the "hydrophobic zippers" method (Fiebig & Dill, 1993). We find that model conformations with radii of gyration about 20% larger than native conformations commonly have bimodal distributions of P(r), of the pairwise interatomic distances, r, and Kratky plots in agreement with recent small-angle X-ray scattering (Sosnick & Trewhella, 1992; Flanagan et al., 1992; Kataoka et al., 1993; Flanagan et al., 1993) experiments on three different proteins. We also find that the lattice model of the Shortle 1-136 fragment of staphylococcal nuclease does not appear capable of forming a single hydrophobic core by hydrophobic zippering, consistent with experiments.

The concept of a random coil. Residual structure in peptides and denatured proteins.

Non-native states of proteins are of increasing interest because of their relevance to issues such as protein folding, translocation and stability. A framework for interpreting the wealth of experimental data for non-native states emerging from rapid advances in experimental techniques involves comparison with a "random coll' state, which possesses no structure except that inherent in the local interactions. We review here the concept of a random coil, from its global to its local properties. In particular, we focus on the description of a random coil in terms of statistical distributions in psi, phi space. We show that such a model, in combination with experimental data, provides insight into the structural properties of polypeptide chains and has significance for understanding protein folding and for molecular design.

Folding intermediates of SNARE complex assembly.

SNARE (soluble NSF attachment protein receptor) proteins assemble into a stable complex essential for vesicle-membrane fusion. To further understand SNARE function we have used solution nuclear magnetic resonance (NMR) spectroscopy to characterize three assembly states of a yeast SNARE complex: first, the 'closed' conformation of Sso1; second, the binary complex of Sso1 and Sec9; and third, the ternary complex of Sso1, Sec9 and Snc1. Sec9 and Snc1 are unstructured in isolation. Sso1 likely consists of a four helix bundle formed by part of the C-terminal Hcore domain and the N-terminal H(A)H(B)H(C) domain, and this bundle is flanked on both sides by large flexible regions. Sso1 switches to an 'open' state when its Hcore domain binds Sec9. Conformational switching of the Hcore domain, via H(A)H(B)H(C), may provide a key regulatory mechanism in SNARE assembly. Formation of binary and ternary complexes induces additional alpha-helical structure in previously unstructured regions. Our data suggest a directed assembly process beginning distal to the membrane surfaces and proceeding toward them, bringing membranes into close proximity and possibly leading to membrane fusion.

A test of lattice protein folding algorithms.

We report a blind test of lattice-model-based search strategies for finding global minima of model protein chains. One of us (E.I.S.) selected 10 compact conformations of 48-mer chains on the three-dimensional cubic lattice and used their inverse folding algorithm to design HP (H, hydrophobic; P, polar) sequences that should fold to those "target" structures. The sequences, but not the structures, were sent to the UCSF group (K.Y., K.M.F., P.D.T., H.S.C., and K.A.D.), who used two methods to attempt to find the globally optimal conformations: "hydrophobic zippers" and a constraint-based hydrophobic core construction (CHCC) method. The CHCC method found global minima in all cases, and the hydrophobic zippers method found global minima in some cases, in minutes to hours on workstations. In 9 out of 10 sequences, the CHCC method found lower energy conformations than the 48-mers were designed to fold to. Thus the search strategies succeed for the HP model but the design strategy does not. For every sequence the global energy minimum was found to have multiple degeneracy with 10(3) to 10(6) conformations. We discuss the implications of these results for (i) searching conformational spaces of simple models of proteins and (ii) how these simple models relate to proteins.

Structural and dynamical properties of a denatured protein. Heteronuclear 3D NMR experiments and theoretical simulations of lysozyme in 8 M urea.

Oxidized and reduced hen lysozyme denatured in 8 M urea at low pH have been studied in detail by NMR methods. 15N correlated NOESY and TOCSY experiments have provided near complete sequential assignment for both 1H and 15N resonances. Over 900 NOEs, including 130 (i, i + 2) and 23 (i, i + 3) NOEs, could be identified by analysis of the NOESY spectra of the denatured states, and 3J(HN, Halpha) coupling constants and 15N relaxation rates have been measured. The coupling constant and NOE data were analyzed by comparisons with theoretical predictions from a random coil polypeptide model based on amino acid specific phi,psi distributions extracted from the protein data bank. There is significant agreement between predicted and experimental NMR parameters suggesting that local conformations of the denatured states are largely determined by short-range interactions within the polypeptide chain. This result is supported by the observation that the chemical shift, coupling constant, and NOE data are little affected by whether or not the four disulfide bridge cross-links are formed in the denatured protein. The relaxation data, however, show significant differences between the oxidized and reduced protein. Analysis of the relaxation data in terms of simple dynamics models provides evidence for weak clustering of hydrophobic groups near tryptophan residues and increased barriers to motion in the more compact conformers formed when the polypeptide chain is cross-linked by the disulfide bridges. Using this information, a structural description of these denatured states is given in terms of an ensemble of conformers, which have a complex relationship between their local and global characteristics.