No abstract

Distinct structural regions have been found in several globular proteins composed of single polypeptide chains. The existence of such regions and the continuity of peptide chain within them, coupled with kinetic arguments, suggests that the early stages of three- For some decades thermodynamic determination of threedimensional protein structure has been a basic guiding principle. However, in the past few years there has been a growing awareness of the possibility that kinetic factors may play a significant role in determining protein structure (1-5). In particular, the argument was advanced by Levinthal (refs. 6,7, see also Appendix) that the time for a random search of all possible structures would be unrealistically long for even a small protein, and that something like a nucleation event must occur to permit structure formation in biologically feasible time. Studies from our laboratory have led to experimental evidence for nucleation in physiologically feasible protein-folding experiments (8-10). Nucleation has also been inferred from kinetic evidence by Tsong et al. (11) in the refolding of ribonuclease with intact disulfide crosslinks. Arguments for the plausibility of nucleation-initiated protein folding have begun to appear in the literature (12, 13).In studying several known protein structures, I have noted the existence of distinct structural regions in several singlechain proteins, and less distinct regions in others. Such observations are not original, having been made in almost every case by the investigators who determined the structure of the particular proteins.In defining what I mean by structural regions, I limit our discussion to protein models that present only the course of the peptide backbone. A "region" is a section of peptide chain that can be enclosed in a compact volume. A continuous region can be completely surrounded by a closed surface (a closed line if two-dimensional, as in Fig. 1), and is characterized by possession of two terminal points. I define the terminal points of a protein region as those points where the peptide chain crosses the enclosing envelope, and also those points where the peptide chain terminates (amino or carboxyl terminus). A region having more than two terminal points is discontinuous. Fig. 1 illustrates continuous and discontinuous regions with several two-dimensional examples. Each of the C regions is seen to have two terminal points, and each of the D 697 regions has four terminal points. The aminoacid residues in a continuous region are one-dimensionally contiguous; the condition of contiguity does not obtain for a discontinuous region. It remains to say how compactness is determined. For the present work, compactness has been decided subjectively, but a quantitative index of compactness is suggested in the Discussion.Examination of protein structures was facilitated by the construction of three-dimensional peptide chain models. Bentwire models were built with the Rubin-Richardson bending tool (14). As a general test of fidelity the physical model...

The denaturation and renaturation of carbonic anhydrase I1 (CAII) has been studied in several laboratories. Both thermodynamic and kinetic evidence support the existence of at least two intermediates between denatured and native protein. Previous studies have shown that on rapid dilution of a CAII solution from 5 M to 1 M guanidinium chloride, aggregation strongly competes with renaturation at higher protein concentrations, suggesting an upper limit for [CAII] of -0.1%. Our experiments show 60% renaturation at 0.4% [CAII] and that aggregate formation is partially reversible. This yield can be substantially increased by several surfactant additives, including simple alkanols as well as micelle-forming surfactants. Effective surfactants (promoters) act by suppressing initial aggregate formation, not by dissolving aggregates. Promoters act on either the first folding intermediate ( I I ) or oligomers thereof. Eight of the 18 surfactants examined showed promoter activity, and no correlation was evident between promoter activity and chemical structure or surface tension lowering. These results indicate discrimination (molecular recognition) by I I and/or its oligomers.

Experimental circular dichroic (CD) spectra of three proteins have been combined with estimates of the content of peptide-chain structural modes obtained from x-ray diffraction studies of the same proteins. Solution of the simultaneous equations at a series of wavelengths permits the construction of a CD spectrum for each of three structural modes: a-helix, j8-structure, and the so-called "random". The CD spectra thus obtained are compared with those obtained from polypeptide models. The a-helical spectra from the two approaches are nearly congruent, the 3-structure spectra are in fair agreement, and the third forms agree qualitatively, but are substantially different quantitatively. Comparisons are made between the present approach and earlier approaches to interpreting protein CD spectra. Certain advantages of the present approach are indicated.In recent years, much use has been made of circular dichroism (CD) and optical rotatory dispersion (ORD) to investigate the modes of polypeptide-chain folding in proteins. Without exception, earlier investigators have used synthetic polypeptides to calibrate the parameters employed in interpreting the ORD and CD of proteins (1-6). The degree of success of these attempts has been judged by comparison of the content of structural modes obtained by CD and/or ORD with the content found from structural solutions from x-ray diffraction studies of individual proteins.There are now several proteins for which the peptidechain structures are known from x-ray investigations. The structural content of these proteins can be linked with CD spectra for the same proteins in a set of simultaneous equations, whose solution yield the CD spectra of the structural modes as they are found in proteins. The same treatment can be applied to ORD data, although we have not done so here. We report the results of application of such a treatment to the x-ray structural and CD parameters of myoglobin, ribonuclease, and lysozyme. We have assumed that each of these three proteins contains only the structural modes commonly called "a-helix", ",8-structure (antiparallel)", and "random". The first two of these are obtained, not without uncertainties, from structural models derived from three-dimensional electron density maps constructed from single-crystal x-ray diffraction. The third, the so-called "random" structure, is surely not random in the usual structural sense of that word. It might better be called "without long-range order" or "remainder", since its fraction is defined as: third form = 1-(a + ().To determine (for a given wavelength) the mean residue ellipticities of the a-helix, (-structure, and the remainder, we write the following set of independent equations.[0]1 = X1 [0] [0]r, the intrinsic residue ellipticities of the three structural modes, at a number of wavelengths. When we plot these parameters against wavelength, we generate the CD spectra of the three structural modes as found in proteins. We used the data of Table 1 and Table 2 to evaluate the coefficients of these equation...

No abstract