Analysis of the folding of hen lysozyme shows that the protein does not become organized in a single cooperative event but that different parts of the structure become stabilized with very different kinetics. In particular, in most molecules the alpha-helical domain folds faster than the beta-sheet domain. Furthermore, different populations of molecules fold by kinetically distinct pathways. Thus, folding is not a simple sequential assembly process but involves parallel alternative pathways, some of which may involve substantial reorganization steps.

NMR spectroscopy has been used to investigate the structure of a partially folded state of a protein, the molten globule or A-state of alpha-lactalbumin. The 1H NMR spectrum of this species differs substantially from those of both the native and fully unfolded states, reflecting the intermediate level of order. The resolution in the spectrum is limited by the widespread overlap and substantial line widths of many of the resonances. Methods have therefore been developed that exploit the well-resolved spectrum of the native protein to probe indirectly the A-state. A number of resonances of the A-state have been found to be substantially shifted from their positions in the spectrum of the unfolded state and have been identified through magnetization transfer with the native state, under conditions where the two states are interconverting. The most strongly perturbed residues in the A-state were found to be among those that form a hydrophobic core to the native structure. A number of amides were found to be highly protected from solvent exchange in the A-state. These have been identified through pH-jump experiments, which label them in the spectrum of the native protein. They were found to occur mainly in segments that are helical in the native structure. These results enable a model of the A-state to be proposed in which significant conformational freedom exists but where specific elements of native-like structure are preserved.

No abstract

Two-dimensional 1H-NMR spectroscopy has been used to study the acid-denatured molten globule (A-state) of alpha-lactalbumin. The NMR spectra show that chemical shift dispersion is limited but significantly greater than that expected for a random coil conformation. The small chemical shift dispersion of side-chain resonances in the A-state together with line broadening associated with conformational averaging indicates that most of the long-range tertiary structure in the A-state is likely to be nonspecific. Side-chain resonances in the A-state are generally shifted somewhat upfield of random coil values; this and the observation of a large number of interresidue NOEs, however, indicate that some side-chain interactions, at least at the level of hydrophobic clustering, exist in the A-state. Analysis of NOESY spectra shows no evidence for an ordered structure for either of the two major clusters of aromatic residues which in the native structure make up part of the hydrophobic core of the helical domain of the native protein. A new aromatic cluster in the A-state which results from rearrangement of the side chains of Tyr103, Trp104, and His107 from their native state positions was, however, detected by a number of well-defined interresidue NOE effects. Similar NOE patterns are observed in a peptide corresponding to residues 101-110 of alpha-lactalbumin in trifluoroethanol, suggesting that the non-native structure in the 101-110 region of the A-state is not dependent on specific interactions with the rest of the chain. Trapping experiments indicate that amide protons from regions of the sequence which in the native state are helical are among those strongly protected from solvent exchange in the A-state; those from one of the helices (the C helix) were specifically identified. Taken together, these results reinforce a model of the A-state which has stable regions of localized secondary structure but a largely disordered tertiary structure.

The refolding kinetics of hen lysozyme have been studied using a range of fluorescent probes. These experiments have provided new insight into the nature of intermediates detected in our recent hydrogen-exchange labeling studies [Radford, S.E., et al. (1992) Nature 358, 302-307], which were performed under the same conditions. Protection from exchange results primarily from the development of stabilizing side-chain interactions, and the fluorescence studies reported here have provided a new perspective on this aspect of the refolding process. The intrinsic fluorescence of the six tryptophan residues and its susceptibility to quenching by iodide have been used to monitor the development of hydrophobic structure, and these studies have been complemented by experiments involving binding to a fluorescent hydrophobic dye 1-anilino-naphthalenesulfonic acid (ANS). Formation of fixed tertiary interactions of aromatic residues has been monitored by near-UV circular dichorism, while development of a competent active site has been probed by binding to a competitive inhibitor bearing a fluorescent label, 4-methylumbelliferyl-N,N'-diacetyl-beta-chitobiose. The combination of these techniques has enabled us to monitor the development both of the hydrophobic core of the protein and of interactions between the two folding domains. If the behavior of the tryptophans is representative of the hydrophobic residues of the protein in general, it seems that collapse is already substantial in species formed within the first few milliseconds of refolding and is highly developed in later intermediates which nonetheless appear to lack many fixed tertiary interactions.(ABSTRACT TRUNCATED AT 250 WORDS)

The hydrogen exchange kinetics of 68 individual amide protons in the native state of hen lysozyme have been measured at pH 7.5 and 30 degrees C by 2D NMR methods. These constitute the most protected subset of amides, with exchange half lives some 10(5)-10(7) times longer than anticipated from studies of small model peptides. The observed distribution of rates under these conditions can be rationalized to a large extent in terms of the hydrogen bonding of individual amides and their burial from bulk solvent. Exchange rates have also been measured in a reversibly denatured state of lysozyme; this was made possible under very mild conditions, pH 2.0 35 degrees C, by lowering the stability of the native state through selective cleavage of the Cys-6-Cys-127 disulfide cross-link (CM6-127 lysozyme). In this state the exchange rates for the majority of amides approach, within a factor of 5, the values anticipated from small model peptides. For a few amides, however, there is evidence for significant retardation (up to nearly 20-fold) relative to the predicted rates. The pattern of protection observed under these conditions does not reflect the behavior of the protein under strongly native conditions, suggesting that regions of native-like structure do not persist significantly in the denatured state of CM6-127 lysozyme. The pattern of exchange rates from the native protein at high temperature, pH 3.8 69 degrees C, resembles that of the acid-denatured state, suggesting that under these conditions the exchange kinetics are dominated by transient global unfolding. The rates of folding and unfolding under these conditions were determined independently by magnetization transfer NMR methods, enabling the intrinsic exchange rates from the denatured state to be deduced on the basis of this model, under conditions where the predominant equilibrium species is the native state. Again, in the case of most amides these rates showed only limited deviation from those predicted by a simple random coil model. This reinforces the view that these denatured states of lysozyme have little persistent residual order and contrasts with the behavior found for compact partially folded states of proteins, including an intermediate detected transiently during the refolding of hen lysozyme.

Nuclear magnetic resonance (NMR) studies have shown that two distinct folded conformations of staphylococcal nuclease coexist in solution and that these two states can interconvert directly without passing through an unfolded state. These experiments have also revealed that the two forms have very different folding kinetics, although the possibility that one component is an obligatory intermediate for the folding of the other form could be discounted. Here we report NMR data which show that alternative unfolded states are also distinguishable. These observations led us to hypothesize that cis/trans isomerism at a single peptide bond between a proline and its preceding residue might be the origin of the conformational multiplicity. Proline 117 was identified as a likely candidate for the site concerned and a mutant protein, in which Pro 117 was replaced by Gly, was constructed in order to test this. Alternative conformations are not observed in the spectrum of this mutant, lending powerful support to this hypothesis.

No abstract