No abstract

The diffusion-collision model of protein folding is assessed. A description is given of the qualitative aspects and quantitative results of the diffusion-collision model and their relation to available experimental data. We consider alternative mechanisms for folding and point out their relationship to the diffusion-collision model. We show that the diffusion-collision model is supported by a growing body of experimental and theoretical evidence, and we outline future directions for developing the model and its applications.Keywords: diffusion-collision model; folding models; folding pathways; kinetic intermediates; microdomains; protein folding dynamicsThe biological activity of a globular protein is determined by its average 3-dimensional structure (the native conformation) and by the internal flexibility of this structure (Brooks et al., 1988). It has been shown (Anfinsen, 1973) for a number of proteins that the native structure can fold spontaneously under appropriate conditions without any information other than that contained in the linear sequence of amino acid residues. Because the purpose of the genetic code is to specify the sequence of amino acids, the prediction of the structure of a protein from its amino acid sequence is an essential part of molecular biology. There are now about 1,700 proteins whose structures have been determined (Bernstein et al., 1977) by X-ray crystallography and by NMR. Although these structures have served as the basis of much of our understanding of proteins, they represent an insignificant fraction of the total number of different proteins (on the order of lo9) in living systems. It will not be possible to determine all of their 3-dimensional structures by experimental methods, though the problem would be simplified if the naturally occurring proteins are constructed from a much more limited number of structural elements (Dorit et al., 1990 ture. A rational approach to protein design and engineering thus has as one of its elements the prediction of the structure of a protein from its amino acid sequence.There are 2 aspects to the prediction problem. One is thermodynamic in character and concerns the prediction of the native structure of a protein from its sequence; the other is dynamical in character and concerns the prediction of the mechanism by which a denatured protein folds to the native conformation in solution or in vivo. The dynamic aspect of the folding problem is considered in this review. It is often phrased in terms of the Levinthal paradox (Levinthal, 1966(Levinthal, , 1968, which corresponds to the realization that a random search of all possible structures would take longer than the age of the universe to find the native conformation. There would be no search problem if each of the amino acids could find its native conformation independently of the others, or if only nearest-neighbor interactions were involved. This would reduce the protein folding problem to an analog of the helix-coil transition. What makes the search problem difficult is that long...

SynopsisThe basic equations for the elementary step in the diffusion-collision-coalescence model of protein folding are derived for the case of two radially diffusing spherical microdomains. Refinements and biological implications of the mechanism are considered; included are detailed discussions of the parameters of the model, the possibilities of rotational diffusion and surface diffusion in one or two dimensions, the nature of the microdomains, and the application of the model to protein unfolding.