Journal of Molecular Biology
Volume 276, Issue 3, 27 February 1998, Pages 657-667
Journal home page for Journal of Molecular Biology

Regular article
Folding kinetics of the SH3 domain of PI3 kinase by real-time NMR combined with optical spectroscopy1

https://doi.org/10.1006/jmbi.1997.1553Get rights and content

Abstract

The refolding kinetics of the chemically denatured SH3 domain of phosphatidylinositol 3′-kinase (PI3-SH3) have been monitored by real-time one-dimensional 1H NMR coupled with a variety of other biophysical techniques. These experiments indicate that the refolding kinetics of PI3-SH3 are biphasic. The slow phase (27 (±8)% amplitude) is due to a population of substantially unfolded molecules with an incorrectly configured cis proline residue. The fast phase (73 (±8)% amplitude) corresponds to the folding of protein molecules with proline residues in a trans configuration in the unfolded state. NMR experiments indicate that the first species populated after the initiation of folding exhibit poor chemical shift dispersion and have spectra very similar to that of the denatured protein in 8 M guanidine hydrochloride. Linear combinations of the first spectrum and of the spectrum of the native protein accurately reconstruct all of the spectra acquired during refolding. Consistent with this, native side-chain and backbone Hα atom packing (NMR), secondary structure (far-UV circular dichroism), burial of aromatic residues (near-UV circular dichroism), intrinsic fluorescence and peptide binding activity are all recovered with effectively identical kinetics. Equilibrium unfolding and folding/unfolding kinetics yield, within experimental error, identical values for the free energy of unfolding (ΔGu-H2O = 3.38 kcal mol−1) and for the slope of the free energy of unfolding versus denaturant concentration (meq = 2.33 kcal mol−1 M−1). Together, these data provide strong evidence that PI3-SH3 folds without significant population of kinetic well-structured intermediates. That PI3-SH3 folds slowly (time constant 2.8 seconds in H2O at 20°C) indicates that slow refolding is not always a consequence of kinetic traps but may be observed even when a protein appears to fold via a simple, two-state mechanism.

Introduction

The mechanisms by which polypeptides fold to their native conformation remains an area of active research in structural biology. Nuclear magnetic resonance (NMR) has proved to be an important technique for studying the complex phenomenon of protein folding because of the structural and dynamical data that it can provide at atomic resolution (for a recent review, see Dyson & Wright, 1996). NMR has been widely used to characterize models of folding intermediates at equilibrium and denatured or unfolded proteins Dobson 1992, Wuthrich 1994, Shortle 1996, Smith et al 1996. Amide-hydrogen exchange methods have played a major role in these studies and have furnished a wealth of information on partly folded and denatured states. Studies on small peptides (Dyson & Wright, 1996) have been helpful in clarifying the role of local structures in early folding events. The rate of interconversion between native and unfolded states at specific residues has been obtained by magnetization transfer and line shape analysis techniques for proteins under equilibrium conditions Dobson and Evans 1984, Evans et al 1989, Roder 1989, Huang and Oas 1995a, Zhang and Forman-Kay 1995. An NMR technique that has been widely used to follow indirectly the kinetics of protein folding is hydrogen-exchange pulse labeling Roder et al 1988, Udgaonkar and Baldwin 1988, Roder 1989, Englander and Mayne 1992, Baldwin 1993, Radford and Dobson 1995. Undoubtedly, pulse labeling and equilibrium NMR experiments have made major contributions to our current understanding of the folding processes.

A powerful complementary approach to these studies of protein folding is to monitor folding kinetics directly by real-time NMR. Time-resolved NMR is the only available technique that can directly probe the packing of side-chains and backbone atoms at individual sites. Real-time NMR provides a means of directly monitoring folding kinetics at multiple specific sites, thus allowing the detection of intermediates and the determination of the cooperativity of the folding process. In spite of the potential of the technique, its application has so far largely been limited to a few examples of slowly folding (or unfolding) proteins, due to its inherent relatively poor time-resolution. Real-time NMR has been used to monitor the folding of thermally denatured ribonuclease A Blum et al 1978, Adler and Scheraga 1988, Akasaka et al 1991, staphylococcal nuclease (Kautz & Fox, 1993), and a collagen-like, triple-helical peptide (Liu et al., 1996). Studies of the refolding of chemically denatured apoplastocyanin have indicated an intermediate trapped by an incorrect trans proline conformation (Koide et al., 1993). When applied to the unfolding of ribonuclease A, real-time NMR suggested the existence of a novel unfolding intermediate (Kiefhaber et al., 1995).

Stopped-flow technology can significantly reduce the dead time of real-time NMR experiments and has been used to characterize the folding (Hoeltzli & Frieden, 1996) and unfolding (Hoeltzli & Frieden, 1995) of 19F-labeled dihydrofolate reductase and the unfolding of 19F labeled intestinal fatty acid binding protein (Frieden et al., 1993). A transient intermediate with a 1H NMR spectrum which resembles that of the well-characterized equilibrium molten globule of α-lactalbumin at acid pH has been detected by one-dimensional (1D) NMR stopped-flow experiments (Balbach et al., 1995), and a single two-dimensional (2D) 1H-15N experiment acquired during refolding has been used to show that the protein backbone forms cooperatively (Balbach et al., 1996). More recently, photo-CIDNP (chemically induced dynamic nuclear polarization) experiments coupled to a rapid mixing device have been developed and applied to study the refolding of hen lysozyme (Hore et al., 1997). Here, we report stopped-flow 1D NMR experiments in conjunction with optical spectroscopy to study the renaturation of the SH3 domain of the p85α subunit of the bovine phosphatidylinositol 3′-kinase (PI3-SH3).

The SH3 domain (approximately 60 to 85 residues) is a small, modular domain that often occurs in proteins involved in intra-cellular signal transduction Morton and Campbell 1994, Musacchio et al 1994. Despite the low sequence homology across the SH3 domain family, all SH3 domains exhibit a common fold which has been well characterized both by NMR spectroscopy and X-ray crystallography. In particular, the structure of PI3-SH3 (84 residues) is composed of two perpendicular antiparallel β-sheets of three and two strands, respectively, and two helix-like turns (Booker et al., 1993). SH3 domains provide simple model systems for the study of protein folding, as they are small, monomeric, structurally independent domains and lack complicating factors such as disulfide bonds, prosthetic groups or native cis-proline residues.

Section snippets

Equilibrium unfolding

The equilibrium unfolding of PI3-SH3 was monitored by the change of intrinsic fluorescence upon unfolding (Figure 1). A single transition is observed between the native and denatured states. The fluorescence of the denatured state shows a linear dependence on denaturant concentration as has been reported for other proteins (for example Koide et al 1993, Viguera et al 1994). Data are well described by a two-state model that assumes that the free energy of unfolding (ΔGu) is a linear function of

Discussion

Real-time NMR has been a particularly useful tool to probe the refolding of PI3-SH3. Indeed, this technique showed that (i) four seconds after initiating the refolding, the side-chains and backbone of the protein are not packed and that no burst-phase collapse leading to a highly structured intermediate occurs; (ii) the transition between the first state (or more accurately, ensemble of states) and the native state is cooperative and no highly structured intermediate accumulates during

Protein and synthetic peptides

PI3-SH3 was expressed in Escherichia coli as a fusion protein with the enzyme glutathione S-transferase and purified as described (Booker et al., 1993). The recombinant protein used in this study consists of 84 residues from the SH3 domain of PI3, plus a two-residue N-terminal extension (GS) and a four-residue C-terminal extension (WNSS). After purification, PI3-SH3 was buffer exchanged to 20 mM NH4CO3 (pH 8.0) by dialysis at 4°C, lyophilized and stored at 4°C until further use. The protein was

Acknowledgements

We thank Maureen Pitkeathly for peptide synthesis, Dennis Benjamin for his help with mass spectrometry and Sophie Jackson for kindly providing the cyclophilin A used in this study. This is a contribution from the Oxford Centre for Molecular Sciences which is supported by the UK Biotechnology and Biological Sciences Research Council, the Engineering and Physical Sciences Research Council and the Medical Research Council. This investigation was also supported in part by the Wellcome Trust

References (56)

  • D.R. Shortle

    Structural analysis of non-native states of proteins by NMR methods

    Curr. Opin. Struct. Biol.

    (1996)
  • R.L. Stein

    Mechanism of enzymatic and nonenzymatic prolyl cis-trans isomerization

    Advan. Protein Chem.

    (1993)
  • K. Wüthrich

    NMR assignments as a basis for structural characterization of denatured states

    Curr. Opin. Struct. Biol.

    (1994)
  • H. Yu et al.

    Structural basis for the binding of proline-rich peptides to SH3 domains

    Cell

    (1994)
  • M. Adler et al.

    Structural studies of a folding intermediate of bovine pancreatic ribonuclease A by continuous recycled flow

    Biochemistry

    (1988)
  • K. Akasaka et al.

    Temperature-jump NMR study of protein foldingRibonuclease A at low pH

    J. Biomol. NMR

    (1991)
  • P. Alexander et al.

    Kinetic analysis of folding and unfolding the 56 amino acid IgG-binding domain of streptococcal protein G

    Biochemistry

    (1992)
  • A.T. Alexandrescu et al.

    Structure and dynamics of the acid-denatured molten globule state of α-lactalbumina two dimensional study

    Biochemistry

    (1993)
  • J. Balbach et al.

    Following protein folding in real time using NMR spectroscopy

    Nature Struct. Biol.

    (1995)
  • J. Balbach et al.

    Protein folding monitored at individual residues during a two-dimensional NMR experiment

    Science

    (1996)
  • A. Chaffotte et al.

    A possible initial folding intermediatethe C-terminal proteolytic domain of tryptophan synthase β chains folds in less than 4 milliseconds into a condensed state with non-native-like secondary structure

    Biochemistry

    (1992)
  • C.M. Dobson et al.

    Protein folding kinetics from magnetization transfer nuclear magnetic resonance

    Biochemistry

    (1984)
  • H.J. Dyson et al.

    Insights into protein folding from NMR

    Annu. Rev. Phys. Chem.

    (1996)
  • M. Engelhard et al.

    Kinetics of interaction of partially folded proteins with a hydrophobic dyeevidence that molten globule character is maximal in early folding intermediates

    Protein Sci.

    (1995)
  • S.W. Englander et al.

    Protein folding studied using hydrogen-exchange labeling and two-dimensional NMR

    Annu. Rev. Biophys. Biomol. Chem.

    (1992)
  • P.A. Evans et al.

    A magnetization-transfer nuclear magnetic resonance study of the folding of staphylococcal nuclease

    Biochemistry

    (1989)
  • C. Frieden et al.

    NMR and protein folding

    Protein Sci.

    (1993)
  • J.I. Guijarro et al.

    Protein folding intermediates with rapidly exchangeable amide protons contain authentic hydrogen-bonded secondary structures

    Biochemistry

    (1995)
  • Cited by (128)

    • Solution NMR views of dynamical ordering of biomacromolecules

      2018, Biochimica et Biophysica Acta - General Subjects
      Citation Excerpt :

      Spectra collected over time can display chemical shift changes and/or changes to peak intensity. This has been applied to study slow turnover of substrate by enzymes [148], folding processes [149–151], the monitoring of chemical synthesis [152,153], and for identifying structural changes in materials [154,155]. Real-time NMR has also benefited from fast acquisition [156] and hyperpolarization [157] techniques to not only improve the rate at which data can be collected, but to also increase the signal-to-noise of the data.

    • A feature analysis of lower solubility proteins in three eukaryotic systems

      2015, Journal of Proteomics
      Citation Excerpt :

      This is potentially problematic, since proteins, by their nature, have a high propensity for aggregation, especially if damaged or improperly folded. The proper assembly of a natively folded protein is not trivial, and even after the native conformation has been attained, unfolding and aggregation can still occur through instability incurred by translational errors or mutations, or be caused by a wide variety of possible stressors, such as high temperature, high pressure and low pH [1–3]. Protein aggregation can be harmful to the cell, both through the potential loss-of-function due to the lack of available functional protein [4,5], as well as through toxic gain-of-function, such as the formation of β-sheet-rich aggregates [6].

    • Inhibition mechanism of hydroxypropyl methylcellulose acetate succinate on drug crystallization in gastrointestinal fluid and drug permeability from a supersaturated solution

      2014, European Journal of Pharmaceutical Sciences
      Citation Excerpt :

      The changes in the molecular state of a drug during supersaturation and drug permeation cannot be precisely evaluated without a real-time monitoring technique. Nuclear magnetic resonance (NMR) techniques enable the real-time monitoring of molecular state changes over a short period (Guijarro et al., 1998; Schanda and Brutscher, 2005). This technique can be a powerful tool for evaluating the dynamic supersaturated state.

    View all citing articles on Scopus
    1

    Edited by P. E. Wright

    2

    Present addresses: C. J. Morton, Department of Chemistry, University of Melbourne, Parkville 3052, Australia; K. W. Plaxco, Department of Biochemistry, University of Washington, Seattle, Washington 98195, USA

    View full text