Rapid Formation of Non-native Contacts During the Folding of HPr Revealed by Real-time Photo-CIDNP NMR and Stopped-flow Fluorescence Experiments

https://doi.org/10.1016/S0022-2836(03)00507-2Get rights and content

Abstract

We report the combined use of real-time photo-CIDNP NMR and stopped-flow fluorescence techniques to study the kinetic refolding of a set of mutants of a small globular protein, HPr, in which each of the four phenylalanine residues has in turn been replaced by a tryptophan residue. The results indicate that after refolding is initiated, the protein collapses around at least three, and possibly all four, of the side-chains of these residues, as (i) the observation of transient histidine photo-CIDNP signals during refolding of three of the mutants (F2W, F29W, and F48W) indicates a strong decrease in tryptophan accessibility to the flavin dye; (ii) iodide quenching experiments show that the quenching of the fluorescence of F48W is less efficient for the species formed during the dead-time of the stopped-flow experiment than for the fully native state; and (iii) kinetic fluorescence anisotropy measurements show that the tryptophan side-chain of F48W has lower mobility in the dead-time intermediate state than in both the fully denatured and fully native states. The hydrophobic collapse observed for HPr during the early stages of its folding appears to act primarily to bury hydrophobic residues. This process may be important in preventing the protein from aggregating prior to the acquisition of native-like structure in which hydrophobic residues are exposed in order to play their role in the function of the protein. The phenylalanine residue at position 48 is likely to be of particular interest in this regard as it is involved in the binding to enzymes I and II that mediates the transfer of a phosphoryl group between the two enzymes.

Introduction

The manner in which a protein molecule achieves its unique native state during folding is one of the most dramatic examples of the way in which evolutionary pressure has generated molecules with properties that are able to generate diversity and selectivity in biology.1., 2., 3. As well as resulting in the efficient production of functional proteins, the folding process is now recognized as being coupled to a variety of other biological processes including protein trafficking and the regulation of the cell cycle.4 In addition, proteins have evolved to remain soluble in the complex milieu of the crowded biological environment, and to have the ability to interact specifically with their natural partners but not with other molecules. It is increasingly evident that the loss of control and regulation of cellular processes is the origin of a range of pathological conditions. These include amyloidosis and other protein degradation diseases, including many conditions associated with old age such as Alzheimer's and Parkinson's diseases.3., 5.

The manner by which a protein is able to achieve its functional state is beginning to emerge from experimental investigations of the structural changes occurring during protein folding. A more detailed understanding of such processes will shed light not only on how structure is encoded by the fold, but on the principles of the design of biological molecules that enable proteins to fold efficiently rather than to aggregate to species that may prevent completion of the folding process. Although in vivo molecular chaperones and other factors play an important role in controlling aggregation and other aspects of folding, the intrinsic properties of the sequence are likely to be the dominant factor that enables a protein to fold efficiently. It is therefore of particular importance to understand the detailed mechanism of protein folding in vitro, and to identify the fundamental determinants of the folding process.

The combination of kinetic NMR experiments with photo-CIDNP spectroscopy, in which enhanced nuclear spin polarization is generated by a laser flash, represents a powerful approach to characterizing in a direct manner important aspects of the structure of a protein during the folding process.6 The strength of the photo-CIDNP approach lies not only in its time resolution but also in the reduction in spectral crowding arising from the polarization of only a limited number of residues (tyrosine, tryptophan and, under some conditions, histidine and methionine). Moreover, polarization occurs only when these residues are accessible to the photo-excited dye molecules in the solution, enabling the environment of individual residues to be probed directly in real time as folding takes place. This approach has been used to study the folding of lysozyme7 and α-lactalbumin in the absence8 and presence of Ca2+ ions.9 The results of these experiments suggest that for both proteins a relatively disorganised collapsed state is initially formed, and that reorganization then occurs to generate the native structure. Experiments of this type are therefore extremely valuable in studying details of folding processes, although care is needed to ensure that intensity effects are correctly interpreted in such complex systems where more than one CIDNP-active residue is present.10 This consideration prompted us to use the photo-CIDNP approach to study the folding of single-tryptophan mutants of a protein such that the analysis of the CIDNP effects can be carried out for individual tryptophan residues located at selected positions within the overall fold.

The histidine-containing phosphocarrier protein HPr from Escherichia coli is a small 85-residue protein that is well suited to such an approach. The wild-type (WT) protein contains no tyrosine or tryptophan residues, but four phenylalanine residues that are particularly appropriate as mutation sites for replacement by tryptophan. Figure 1shows a representation of the solution structure of the protein, showing that three α-helices are packed against a four-stranded antiparallel β-sheet.11Figure 1 also shows the location in the structure of the four phenylalanine residues. The replacement of each of the phenylalanine residues in turn with single tryptophan residues, resulting in four single-tryptophan mutants F2W, F22W, F29W, and F48W, does not result in substantial structural changes, but the presence of the bulkier side-chain leads to localized rearrangements around the mutated site.31 The tryptophan side-chains are in similar environments to the corresponding phenylalanine side-chains in the WT structure, with that of residue 48 completely solvent accessible, those of residues 2 and 29 partially buried, and that of residue 22 completely buried in the core of the protein, inaccessible to the solvent. Replacement of the phenylalanine residues associated with the hydrophobic core of the protein, particularly those at positions 22 and 29, results in lower stability, most likely as a result of the introduction of the larger sized tryptophan side-chain that disrupts slightly the optimal packing of the core.31Figure 1 shows in addition the two histidine residues in HPr; His15 in the active-site is fully exposed to the solvent, while the His76 is less accessible. These residues were not mutated in the present study but act as reporters of the overall exposure of the tryptophan residues as we show below.

Here, we show that the aromatic tryptophan side-chains become inaccessible to solvent during the initial stages of folding of HPr, in some cases by forming non-native contacts that have to be broken in order to form the final native state structure. Results from fluorescence anisotropy and quenching studies are fully consistent with these photo-CIDNP observations and provide evidence for a rigid and relatively non-specific collapse of the protein driven by the energetic advantage of sequestering the hydrophobic residues from solvent. At a later stage in folding the acquisition of native-like structure in the protein results in the exposure of certain hydrophobic groups, e.g. the aromatic residue at position 48, where they are able to perform their role in enabling HPr to transfer a phosphoryl group from the cytoplasmic to the membrane-bound enzyme (enzyme I and II, respectively) that results in the phosphorylation and transport of carbohydrates across the cell membranes of bacteria.12

Section snippets

Photo-CIDNP of native and denatured states of mutants and wild-type HPr

Figure 2shows the photo-CIDNP NMR spectra of WT HPr and the four single-tryptophan mutants in their native and fully unfolded (6 M guanidine hydrochloride, GdnDCl) states. The possibility of creating polarisation is limited to the single tryptophan and the two histidine residues, and only occurs when these residues are accessible to the photo-excited flavin molecules. When more than one side-chain is able to react with excited flavin dye (e.g. histidine and tryptophan in native F48W HPr), the

Discussion

The combination of real-time NMR measurements with photo-CIDNP methods represents a powerful approach for characterizing directly structural changes taking place in a protein during folding. The photo-CIDNP results presented here indicate that after the initiation of the refolding reaction, HPr collapses around at least three (2, 29, and 48) out of four tryptophan side-chains, making them less accessible to the flavin dye to an extent that the histidine residues at position 15 and/or 76 can

Protein production

The production, purification and characterization of the four single-tryptophan mutants F2W, F22W, F29W, and F48W and the WT HPr were carried out as described elsewhere.31

Equilibrium photo-CIDNP experiments

Spectra were obtained essentially as described previously.7., 24., 25. Blue–green light (4 W) from a continuous-wave argon ion laser (Spectra Physics Stabilite 2016-05) was chopped into 100 ms pulses by a mechanical shutter controlled from a home-built 600 MHz spectrometer located at the Oxford Center for Molecular Sciences.

Acknowledgements

We thank Dr K. Maeda (Oxford University) for providing us with the Lorentzian fitting algorithm used in this study. This work has been supported by the INTAS project (01-2126), and by the TMR Network “Structure and dynamics of intermediate states in protein folding” (FMRX960013) and RTD project “Transient NMR” (HPRI-CT-1999-50006) of the European Union. The Oxford Centre for Molecular Sciences is supported by EPSRC, BBSRC and MRC. The research of C.M.D. is also supported by a Programme Grant

References (31)

  • D. Canet et al.

    High-sensitivity fluorescence anisotropy detection of protein-folding events: application to alpha-lactalbumin

    Biophys. J.

    (2001)
  • C.M. Dobson

    Getting out of shape

    Nature

    (2002)
  • R.J. Ellis et al.

    Medicine: danger—misfolding proteins

    Nature

    (2002)
  • N.A.J. Van Nuland et al.

    Real-time NMR studies of protein folding

    Accts. Chem. Res.

    (1998)
  • P.J. Hore et al.

    Stopped-flow photo-CIDNP observation of protein folding

    J. Am. Chem. Soc.

    (1997)
  • Cited by (35)

    • Intrinsically disordered proteins may escape unwanted interactions via functional misfolding

      2011, Biochimica et Biophysica Acta - Proteins and Proteomics
      Citation Excerpt :

      Intramolecular non-native interactions are known to perturb the unfolded state ensemble and affect the equilibrium stability of an ordered protein [66,67]. They also might lead to the accumulation of the on-pathway or off-pathway intermediate states [68–70], perturb the protein folding kinetics and modulate folding landscape by affecting the transition state structure and stability [71–75], and, being a major driving force determining the rapid collapse of an unfolded polypeptide chain during the early stages of the protein folding process, serve as factor preventing proteins from aggregation [76]. Furthermore, for ordered proteins with complex topologies, such as knotted proteins, it was proposed that non-native interactions may be of importance for the correct formation of the knots [77].

    • Chapter 3 Photo-CIDNP Spectroscopy

      2009, Annual Reports on NMR Spectroscopy
      Citation Excerpt :

      An ingenious combination of two lasers has been employed where a uv laser first releases Ca2+ ions from a photolabile chelator, thus inducing refolding, and a visible laser then generates CIDNP in the repetitive acquisition scheme.224 Despite the great promise of this in situ approach, at present faster perturbations of the protein state can be achieved with the more common stopped-flow injector devices, which have frequently been used for real-time folding studies.78,225–227 Several direct CIDNP investigations on denatured proteins have been performed,205,212,228,229 but there is the general problem of broad peaks and poor spectral dispersion caused by conformational exchange.

    View all citing articles on Scopus

    These two authors contributed equally to this work.

    Present addresses: D. Canet, GeneProt Inc., 2 Pre-de-la-Fontaine, 1217 Meyrin/GE, Switzerland; C. M. Dobson, Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK.

    View full text