Protein Reviews Series Editor: M. Zouhair Atassi Baylor College of Medicine Houston, Texas
For further volumes: http://www.springer.com/series/6876
Rowen J.Y. Chang • Salvador Ventura Editors
Folding of Disulfide Proteins
Editors Rowen J.Y. Chang Research Center for Protein Chemistry Brown Foundation Inst. of Molecular Medicine, Dept. of Biochemistry and Molecular Biology The University of Texas Houston, TX 77030, USA
[email protected]
Salvador Ventura Dept. de Bioquímica i Biologia Molecular Institut de Biotecnologia i de Biomed. Universitat Autònoma de Barcelona Bellaterra-08193, Spain
[email protected]
ISBN 978-1-4419-7272-9 e-ISBN 978-1-4419-7273-6 DOI 10.1007/978-1-4419-7273-6 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011932496 © Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)
Preface
The formation of selected intramolecular disulfide bonds is one of the most important postraductional modifications of proteins, influencing their folding, stability, and biological function. The folding of disulfide proteins is usually an intricate process in which, after their synthesis at the ribosome in a reduced and unfolded state, polypeptides gain coordinately their native disulfide bonds as well as their unique and stable conformation. In the cell, this process is exquisitely controlled and catalyzed by a complex protein machinery to avoid mispairing of cysteine residues, which might prevent the attaining of functional conformations leading to misfolding and, in some cases, triggering pathological processes. In vitro, the folding of disulfide proteins has constituted the bedrock for a long time on which to understand at the kinetic and structural levels the mechanisms by which a particular amino acid chain folds into a specific functional conformation. The aim of the present monograph is to provide the reader with a detailed view of our current structural and functional understanding of the complex process of protein oxidative folding and of their chemical, biotechnological, and biomedical implications; together with a historical perspective of a field that this year 2011 celebrates its 50th anniversary. The book presents a comprehensive description of the complexity and diversity of folding pathways of different disulfide protein models, including polypeptides with biomedical and biotechnological interest like insulin or plant cyclotides. It is discussed how the cellular machinery and more specifically protein disulfide isomerase promotes and proofreads the formation of native disulfides and how this process can be emulated in vitro exploiting the redox properties of small catalysts. Many therapeutically relevant proteins contain disulfides, which mispairing during recombinant production precludes their commercialization; top experts in the field describe why this occurs and how it can be avoided. The book focuses also on the emerging role played by certain disulfide bonds in the allosteric control of protein function and their implication in health and disease. In addition, the information
v
vi
Preface
accumulated in the present text constitutes a compendium of state of the art technical approaches in protein separation technology, in mass spectrometry and in highresolution structural characterization of folded and unfolded protein states of intrinsic high value for any protein chemist. Houston, TX, USA Bellaterra, Spain
Rowen J.Y. Chang Salvador Ventura
Contents
1 Oxidative Folding: Coupling Conformational Folding and Disulfide Formation.......................................................................... Salvador Ventura and Rowen J.Y. Chang
1
2 The Case of Oxidative Folding of Ribonuclease A: Factors Impacting Fold Maturation of ER-Processed Proteins........... Mahesh Narayan
23
3 Cystine Knot Folding in Cyclotides........................................................ Norelle L. Daly, Christian W. Gruber, Ulf Göransson, and David J. Craik
43
4 In Vitro Folding of Single/Double Chain Insulins and Related Proteins................................................................................ You-Min Feng
63
5 Unfolding and Refolding of Disulfide Proteins Using the Method Disulfide Scrambling................................................ Rowen J.Y. Chang
91
6 Oxidative Protein Folding with Small Molecules . ............................... 109 Watson J. Lees 7 Protein Disulfide Isomerase and the Catalysis of Oxidative Protein Folding................................................................... 133 Hiram F. Gilbert 8 Allosteric Disulfide Bonds....................................................................... 151 Jason W.H. Wong and Philip J. Hogg 9 The Problem of Expression of Multidisulfide Bonded Recombinant Proteins in E. coli............................................................. 183 Silvia A. Arredondo and George Georgiou
vii
viii
Preface
10 NMR-Spectroscopic Investigation of Disulfide Dynamics in Unfolded States of Proteins............................................... 217 Robert Silvers, Kai Schlepckow, Julia Wirmer-Bartoschek, and Harald Schwalbe 11 A Half-Century of Oxidative Folding and Protein Disulphide Formation......................................................... 257 Robert B. Freedman Index.................................................................................................................. 277
Contributors
Silvia A. Arredondo Department of Chemical Engineering, University of Texas, Austin, TX 78712, USA
[email protected] Rowen J.Y. Chang Research Center for Protein Chemistry, Brown Foundation Institute of Molecular Medicine, 1825 Pressler Street, Houston, TX 77030, USA; Department of Biochemistry and Molecular Biology, The University of Texas, Houston, TX 77030, USA
[email protected] David J. Craik Institute for Molecular Bioscience, Division of Chemistry and Structural Biology, The University of Queensland, QLD 4072, Brisbane, Australia
[email protected] Norelle L. Daly Institute for Molecular Bioscience, Division of Chemistry and Structural Biology, The University of Queensland, QLD 4072, Brisbane, Australia
[email protected] You-Min Feng Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yue-Yang Road, Shanghai 200031, China
[email protected] Robert B. Freedman School of Life Sciences, University of Warwick, Coventry, CV4 7AL, UK
[email protected] Ulf Göransson Division of Pharmacognosy, Department of Medicinal Chemistry, Uppsala University, Biomedical Centre, Box 574, SE-751 23, Uppsala, Sweden
[email protected]
ix
x
Contributors
George Georgiou Department of Chemical Engineering, University of Texas, Austin, TX 78712, USA; Department of Biomedical Engineering, University of Texas, Austin, TX 78712, USA; Department of Molecular Genetics and Microbiology, University of Texas, Austin, TX 78712, USA; Institute for Cell and Molecular Biology, University of Texas, Austin, TX 78712, USA
[email protected] Hiram F. Gilbert Verna and Marrs McLean Department of Biochemistry and Molecular Biology, Baylor College of Medicine, Houston, TX 77030, USA
[email protected] Christian W. Gruber Center for Physiology and Pharmacology, University of Vienna, A-1090 Vienna, Austria
[email protected] Philip J. Hogg Lowy Cancer Research Centre, Prince of Wales Clinical School, University of New South Wales, 2052 Sydney, Australia
[email protected] Watson J. Lees Department of Chemistry and Biochemistry, Florida International University, 11200 SW 8th Street, Miami, FL 33199, USA
[email protected] Mahesh Narayan Department of Chemistry, The University of Texas at El Paso, 500 W. Univ. Ave., El Paso, TX 79968, USA
[email protected] Kai Schlepckow Institute of Organic Chemistry and Chemical Biology, Center for Biomolecular Magnetic Resonance (BMRZ), Goethe University Frankfurt, Max-von-Laue-Straße 7, 60438 Frankfurt/Main, Germany Harald Schwalbe Institute of Organic Chemistry and Chemical Biology, Center for Biomolecular Magnetic Resonance (BMRZ), Goethe University Frankfurt, Max-von-Laue-Straße 7, 60438 Frankfurt/Main, Germany
[email protected] Robert Silvers Institute of Organic Chemistry and Chemical Biology, Center for Biomolecular Magnetic Resonance (BMRZ), Goethe University Frankfurt, Max-von-Laue-Straße 7, 60438 Frankfurt/Main, Germany
[email protected] Salvador Ventura Institut de Biotecnologia i de Biomedicina and Departament de Bioquímica i Biologia Molecular, Universitat Autònoma de Barcelona, Bellaterra-08193, Barcelona, Spain
[email protected] Julia Wirmer-Bartoschek Institute of Organic Chemistry and Chemical Biology, Center for Biomolecular Magnetic Resonance (BMRZ), Goethe University Frankfurt, Max-von-Laue-Straße 7, 60438 Frankfurt/Main, Germany Jason W.H. Wong Lowy Cancer Research Centre, Prince of Wales Clinical School, University of New South Wales, 2052 Sydney, Australia
[email protected]
Chapter 1
Oxidative Folding: Coupling Conformational Folding and Disulfide Formation Salvador Ventura and Rowen J.Y. Chang
Abstract Determining how a string of amino acid residues folds into the biologically active protein conformation remains as one of the most important and challenging tasks in biology. Protein folding is usually a fast reaction in which transient intermediates in the folding pathway are short lived, highly dynamic, and very difficult to be trapped, isolated, and characterized. The technique of oxidative folding applied to study disulfide proteins overcomes some of these problems. During protein oxidative folding, the coupling between conformational folding and disulfide formation together with the possibility to selectively quench the progress of the oxidative reaction permits the isolation and further structural characterization of transient folding intermediates in atomic detail. With its unique chemistry and relatively slow kinetics of disulfide formation, the technique of oxidative folding has facilitated the detailed characterization of the folding pathways of an important number of disulfiderich proteins. The results reveal a high degree of diversity of folding mechanisms, which are mainly manifested by the extent of heterogeneity and native-like structures of their intermediate ensembles. Overall, as we will discuss in this chapter, the study of disulfide-containing polypeptides has contributed significantly to our current knowledge on the molecular basis of protein folding. Keywords Disulfide bonds • Protein folding • Folding intermediates • Oxidative folding • Reductive unfolding • Folding pathways • Protein structure
S. Ventura (*) Institut de Biotecnologia i de Biomedicina and Departament de Bioquímica i Biologia Molecular, Universitat Autònoma de Barcelona, Bellaterra-08193, Barcelona, Spain e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_1, © Springer Science+Business Media, LLC 2011
1
2
S. Ventura and R.J.Y. Chang
Abbreviations (des) species 1S 2S 3D structure 3S BPTI Cys–Cys DTTox DTTred EGF GSH GSSG H/D exchange LCI LDTI MCoTI-II MS PDI RNase A RP-HPLC S- SH TAP TCEP TCI
A folding intermediate lacking one disulfide bond One-disulfide intermediates Two-disulfide intermediates Three-dimensional structure Three-disulfide intermediates Bovine pancreatic trypsin inhibitor Cystine Oxidized dithiothreitol Reduced dithiothreitol Epidermal growth factor Reduced glutathione Oxidized glutathione Hydrogen to deuterium exchange Leech carboxypeptidase inhibitor Leech-derived trypsin inhibitor Momordica cochinchinensis trypsin inhibitor II Mass spectrometry Protein disulfide isomerase Bovine pancreatic ribonuclease A Reversed phase high-performance liquid chromatography Thiolate group Thiol group Tick anticoagulant peptide Tris(2-carboxyethyl)phosphine Tick carboxypeptidase inhibitor
1.1 Introduction: Protein Folding In the cell, proteins are synthesized as sequential strings of amino acids. However, these linear chains are in most cases inactive, since a protein becomes functional only when it folds into its specific and unique three-dimensional (3D) structure. Many proteins have been shown to attain spontaneously their functional conformation from an initially unfolded state in vitro, without needing the assistance of the cellular protein quality control (Daggett and Fersht 2003). This implies that most of the information required to form the extremely sophisticated 3D structures of proteins is imprinted somehow in their primary sequences (Anfinsen 1972). As any protein amino acid sequence can be univocally deduced from the corresponding gene, it can be inferred that the genome sequence of an organism contains all the information required to know its complete repertoire of protein structures and their associated functions (Eisenberg et al. 2000). However, although an overwhelming effort has been dedicated to understand the physicochemical determinants and the
1 Protein Oxidative Folding
3
sequence of molecular events that allow a particular amino acid chain to fold into a specific functional conformation, we are still far from being able to assign gene sequences directly to protein structures (Sanchez et al. 2000). The connection between protein sequence and function is lost, in many cases, in the black-box of protein folding (Nishimura et al. 2005). Therefore, it is clear that solving the socalled folding problem constitutes one of the most important and challenging tasks in biochemistry (Creighton 1988a; Gruebele 1999; Fetrow et al. 2002; Daggett and Fersht 2003; Dill et al. 2007). Protein folding obeys the laws of thermodynamics, which means that a protein always folds so that it achieves the conformation with the lowest possible free energy, which in most cases coincides with the functional state. Proteins are able to find the right functional structure, out of an astronomical number of potential 3D conformations in which they could randomly fold, in an extremely fast manner. Local structures can occur in 10–100 ns, loops can fold in 500 ns, and the minimum time required to complete the folding of a small protein is in the order of 1 ms (Gruebele 1999). Cyrus Levinthal in 1969 calculated that finding the right combination in the vast conformational universe by simple trial and error would be completely incompatible with proteins folding in a time scale that is biologically relevant (Honig et al. 1976). Therefore, protein folding must be somehow a directed process, which suggests the existence of defined pathways in which protein folding can occur rapidly. This idea was stimulating, because it implied that the characterization of a defined number of intermediates that populate along folding pathways would allow reconnection of a protein fold to its sequence (Kim and Baldwin 1982; Karplus and Weaver 1994; Ptitsyn 1991). Therefore, a large amount of experimental work has been pursued on the identification and characterization of the ensemble of folding intermediates that are generated during the folding reaction (Dobson and Evans 1988; Weissman 1995). However, due to the extremely fast time scales of protein folding, their transient nature and their highly flexible conformations, folding intermediates have been much more difficult to characterize that initially thought (Baldwin 1994). This is especially true for the initial folding intermediates, which are in fact the most informative species in terms of sequence/conformation relationship (Bryngelson et al. 1995). As it will be demonstrated in this and in subsequent chapters of this book, the study of the folding properties of disulfide-containing proteins has contributed much of our present knowledge about the way the intermediates shape the folding reactions of polypeptides (Creighton 1986).
1.2 Disulfide Bonds of Proteins The covalent link of cysteine residues by disulfide bonds constitutes an important and in many cases an essential structural feature of numerous proteins (Ventura 2008; Abkevich and Shakhnovich 2000; Darby and Creighton 1997). Disulfide bonds are thought to serve several functions. They thermodynamically stabilize the native conformation of proteins. By cross-linking sequentially distant regions of
4
S. Ventura and R.J.Y. Chang
the polypeptide chain, they decrease the entropy of the unfolded ensemble, making it less favorable compared with the folded conformation. The increase in stability is directly related to the length of the polypeptide loop the disulfide bond closes, the longer the loop the larger the conformational stabilization (Poland and Scheraga 1965; Arolas et al. 2009; Camacho and Thirumalai 1995). In some cases, disulfide bonds also act enthalpically through the stabilization of favorable local interactions in the native conformation (Wedemeyer et al. 2000b). The extra stability provided by disulfides is especially important for protecting secreted proteins from oxidants and proteases present in the harsh extracellular environment, thus preventing their inactivation and increasing their half-life (Zavodszky et al. 2001). In addition to their ability to form chemically inert structural disulfide bonds, cysteine residues can react with and be modified by their local environment. Accordingly, increasing evidence indicates that certain disulfide bonds are actually relevant from a functional point of view. Some are found in the active site of thiol/disulfide oxidoreductases and play a catalytic role in the formation, disruption, and exchange of protein disulfides, whereas others act as pH or redox sensors promoting structural changes that inhibit or activate the protein in which they reside (Wouters et al. 2010; Hogg 2003, 2009). The complement of disulfide bonds of a functional protein results from a complicated succession of covalent reactions, including oxidation or disulfide formation, reduction or disulfide disruption, and isomerization or disulfide reshuffling (Creighton 1986, 1997; Wedemeyer et al. 2000a). In vitro, the formation of a protein disulfide bond consists, in most of the cases, in two consecutive thiol/disulfide exchanges with a redox agent. In a first stage, the redox reagent forms a mixed disulfide with the free thiolate in the reduced protein. The mixed disulfide is subsequently attacked by another free cysteine in the protein, resulting in the formation of an intramolecular disulfide and the release of the reduced form of the redox agent (Fig. 1.1). In the absence of chemical reagents, molecular oxygen may act itself as a redox agent and promote the spontaneous formation of disulfide bonds in aqueous solutions. Also, free protein thiolate groups can attack intramolecularly preformed disulfide bonds, leading to an isomerization reaction that results in new disulfide connectivity in the protein. In proteins, the formation and disruption of disulfide bonds cannot be seen simply as chemical reactions, since the forward and backward reaction rates are sharply modulated by the effective concentration of thiolate anions as well as by the reactivity, proximity, and accessibility of both free cysteine residues and disulfide bonds. These specific features render disulfide bonds formation and disruption useful tools to approximate the conformational properties of folding intermediates and native states. Oxidation, reduction, and reshuffling reactions are impeded when the reactive groups are buried into stable and compact protein structure (Narayan et al. 2000; Wedemeyer et al. 2000a, 2002). Therefore, in polypeptides the equilibrium constant (Keq) for a thiol/disulfide exchange reaction is in fact a sensor of the accessibility of the involved groups (Wedemeyer et al. 2002). The Keq for such reactions can vary in more than eight orders of magnitude depending on whether they occur in an unfolded context or inside a compact globular structure (Darby and Creighton 1993). Therefore, they
1 Protein Oxidative Folding
5
Fig. 1.1 Chemistry of a thiol–disulfide exchange reaction. A reactive thiolate group in the reduced form of a protein attacks the accessible disulfide bond of a (cyclic) redox compound. This results in the formation of a mixed disulfide precursor in which the protein and the redox reagent are covalently bound. The precursor is subsequently resolved by the attack of another free cysteine in the protein, which results in the formation of an intramolecular disulfide bond and the release of the reduced (linear) form of the redox reagent and the oxidized protein as final products of the chemical reaction
indirectly report on the conformational stability of the structural elements in their vicinity, in a similar manner to H/D exchange experiments (Darby and Creighton 1997). The rate of disulfide bond formation depends also on the spatial proximity of the two reactive groups. Proximity in this context refers to the probability of the two free thiolates coming within the minimal distance required for their covalent bonding (Wedemeyer et al. 2000b). Thus, disulfide formation kinetics provide us with a kind of molecular ruler to measure relative distances between residues in secondary, tertiary, or quaternary structures (Creighton 1997; Welker et al. 2001a). Finally, disulfide formation depends on the reactivity of the involved chemical groups. The thiolate form (S-) is the reactive form of cysteine whereas the thiol form (SH) cannot establish disulfide bonds. Therefore, both the specific cysteine’s pKa and the pH of the local environment control thiol/disulfide exchange reaction. The pKa of cysteine in an unfolded protein background is around 8.7 but this says little about the effective pKa of a given cysteine in a globular domain where the electrostatic environment is significantly different. As a general rule, it can be assumed that basic environments promote disulfide formation/disruption reactions whereas acidic conditions impede them. As we will see later this behavior allows the selective quenching of oxidative folding reactions.
6
S. Ventura and R.J.Y. Chang
Overall, disulfide bonds display unique chemical and structural characteristics that allow us to use them as privileged sensors or reporters of the progress of protein folding pathways and of the structural properties of the intermediate or final species in the folding reaction (Wedemeyer et al. 2000a). (a) The reduction and oxidation of disulfide bonds follow a two-state mechanism and correspond to a specific and localized structural change. (b) The oxidation/reduction rates can be tuned experimentally by modifying the reaction conditions without affecting importantly secondary or tertiary protein contacts. (c) Disulfide bonds stabilize folding intermediates allowing their purification and subsequent structural characterization.
1.3 Experiments of Reductive Unfolding and Oxidative Folding Reactions of Disulfide Proteins 1.3.1 Trapping, Separation, and Characterization of Folding Intermediates The study of oxidative folding and the reductive unfolding pathways of disulfidecontaining protein requires fast and irreversible quenching methods to allow effective trapping of different conformational species that populate along the pathway in a time-course manner (Fig. 1.2). Creighton (1986) pioneered such approach by irreversibly alkylating the free cysteines present in these partially folded species using iodoacetate. Nevertheless, later on, it was observed that reshuffling of intermediates during the trapping step with iodoacetate can still occur, as observed for the oxidative folding of bovine pancreatic ribonuclease A (RNase A) and bovine pancreatic trypsin inhibitor (BPTI) (Weissman and Kim 1991). Although this undesired rearrangement could be reduced in the presence of high concentrations of iodoacetate, this might promote the modification of residues other than cysteine, like histidine, lysine, and methionine, and therefore the disruption or the establishment of anomalous intramolecular interactions. The use of faster and more selective molecules to block free protein thiols, such as 2-aminoethyl methanethiosulfonate, 1-cyano-4dimethylaminopyridinium tetrafluoroborate, and trans-dichloro(diethylenediamine) platinum (IV) ion, has come to solve this problem (Wu et al. 1998; Narayan et al. 2003a, b; Welker et al. 2004). However, the irreversible binding of chemical modifiers to free cysteines usually distorts sterically the conformation of folding intermediates. Quenching of folding intermediates by acidification using aqueous trifluoroacetic acid or acetic acid at pH 2.0–3.5, has become almost the standard trapping method because it is a extremely rapid process that permits the thermodynamic, kinetic, and structural characterization of unmodified species and in addition allows to witness how a particular isolated intermediate explores the folding landscape toward the native structure after readjusting the pH in stop–go studies (Scheraga et al. 1987). It is important to note that because the reactivity of thiols decreases one order of magnitude for each pH unit below their effective pKa,
1 Protein Oxidative Folding
7
Fig. 1.2 Scheme of the succession of experimental techniques used to characterize the oxidative folding reaction of a disulfide-rich protein
a cidification strongly decreases the rate of thiol–disulfide exchange reactions but do not abolish them completely (Mamathambika and Bardwell 2008). The quenching method has been shown to have a strong influence on the apparent oxidative folding pathway of a disulfide-containing protein. This is best exemplified by the discrepancy of BPTI folding pathway elucidated by Creighton and Kim’s groups on the heterogeneity and native-like structure of BPTI folding intermediates. The difference is likely resulted from the dissimilar quenching method used in these studies, in which one group uses alkylation quenching (Creighton 1990) and the other uses acid quenching (Weissman and Kim 1991). Following sample quenching, the identity and/or conformational properties of the different intermediate species in the usually complex mixture of conformers need to be analyzed (Fig. 1.2). For a protein with only 3 disulfide bonds up to 74 different disulfide-bonded intermediates can populate along the folding pathway. As the number of disulfides in a protein increases, the complexity of the theoretical intermediate ensemble grows exponentially. Therefore, highly effective fractionation methods are required to analyze intermediates of oxidative folding reactions. The quenching method might condition the subsequent separation procedure. In this way, acid trapped intermediates should be kept at low pH, which precludes the use of many chromatographic and electrophoretic methods. For acid-trapped intermediates, reversed phase high-performance liquid chromatography (RP-HPLC) at low
8
S. Ventura and R.J.Y. Chang
pH has become the most useful analysis technique. Although the ensemble of acid-trapped intermediates only differ in the number and the identity of their disulfide bonds, these covalent links endorse them with differential compactness and hydrophobicity, allowing their separation and isolation by RP-HPLC. When covalent chemical quenchers are used, the preferred separation method would depend on the nature of the cysteine-blocking reagent, ion-exchange chromatography can be used for molecules that alter the overall protein charge and SDS-PAGE can be used for large blocking reagents. Importantly, even when using relative small groups for derivatization, like vinyl-pyridine, the corresponding increase in molecular weight can be monitored by mass spectrometry (MS). Because the increase in molecular mass is proportional to the number of blocked cysteines, and accordingly to the number of remaining reactive thiols, MS analysis thus allows us to follow the kinetics of disulfide formation during oxidative folding. To determine the specific disulfide pairing of selected, purified, and derivatized folding intermediates, they are digested with endopeptidases in order to obtain peptide mass fingerprints under reducing and nonreducing conditions that would allow identification of the preformed covalent bonds in the isolated intermediate.
1.3.2 Reductive Unfolding Experiments In reductive unfolding experiments, native disulfide-containing proteins are treated with different concentrations of a reducing agent and the disruption of their disulfide bonds is monitored. In many cases, the disulfide bonds are totally or partially protected by the native protein structure and cannot be easily reduced, requiring first a total or local unfolding event that exposes the bond to the reducing agent. Therefore, the stability of a disulfide bond in front of reduction depends not only on its chemical environment and degree of exposition but also on the dynamic properties of the protein. The thermodynamic value of the structural transition leading to the exposition of a previously buried disulfide bond can be calculated from its rate of reduction at different reducing reagent concentrations (Welker et al. 2001a). In proteins with multiple disulfide bonds the reduction of the first bond renders a natively bonded des-species. Depending on its structure and stability, this species can be metastable and display a native conformation that is only lost after additional reduction of one or more disulfide bonds. Such des-species might also be highly unstable with the rest of the disulfides being exposed to solvent and rapidly reduced in a coordinated manner. Reductive unfolding pathways are sensitive to pH of the solution. Reshuffling reactions are favored when reductive unfolding is performed at basic or neutral pH and mostly avoided when the reaction takes places at acidic pH, since the new free cysteines are protonated and therefore not reactive. Reduced dithiothreitol (DTTred) is usually used for reductive unfolding experiments at basic to neutral pH, whereas tris(2-carboxyethyl)phosphine (TCEP) is the preferred reducing agent at acidic pH.
1 Protein Oxidative Folding
9
1.3.3 Oxidative Folding Experiments In oxidative folding experiments, the protein is first denatured by using high concentrations of chaotropic agents, typically urea or guanidinium chloride, and placed under strong reducing conditions to completely reduce the disulfide bonds that still link covalently the polypeptide chain in the denatured state. This denatured and reduced state emulates the conformation of the protein molecules as they emerge from the ribosome after their synthesis. Denaturing and reducing agents are simultaneously removed and the proteins are allowed to refold. The sequential regeneration of their disulfide bonds is then monitored as described above. The kinetics of oxidative folding reactions are highly sensitive to the experimental conditions, specifically to the ionic strength, pH, and redox potential of the solution. As a general trend, the presence of salts accelerates oxidative folding, likely by competing with nonspecific ionic intramolecular interactions (Arolas et al. 2004). Because the oxidative reaction depends on the protonation state of the protein-free cysteines, the pH of the solution has a strong influence on the velocity and efficiency of the folding process. Most works employed pH 8.5 as the default choice to maximize the folding and recovery of native disulfide proteins. Oxidative folding experiments can be performed in the absence or presence of redox agents. When no external redox agents are added, the molecular oxygen acts as an oxidant. However, under these conditions, most oxidative folding reactions become extremely slow and inefficient. Accordingly, both oxidative and reducing agents are commonly used in oxidative folding experiments (Chatrenet and Chang 1993). Oxidative agents such as oxidized glutathione (GSSG), Cys–Cys, or oxidized protein disulfide isomerase (PDI) act as oxidases, accelerating the formation of disulfide-bonds, which often promotes the fast accumulation of fully oxidized isomers or scrambled species as major folding intermediates. Reducing agents such as reduced glutathione (GSH), Cys, 2-mercaptoethanol, or reduced PDI act as reductases. They catalyze and promote disulfide shuffling of already formed disulfide bonds by allowing the folding intermediates to fold toward a new energy minimum in the way to attain the native disulfide connectivity. Therefore, the optimal rates of in vitro oxidative folding are achieved at specific concentrations of oxidative and reducing agents and usually correspond to redox potentials in which the prevalence of thermodynamic and kinetic traps is minimized (Kibria and Lees 2008). Unfortunately, these concentrations should be empirically and individually adjusted for each particular disulfide-rich protein model under study. Despite different redox conditions might result in extremely different refolding rates and efficiencies for a given protein, it is important to note that, as a rule, the overall folding pathway of the protein remains essentially unaffected. This is best demonstrated in the case of oxidative folding of hirudin (Chang 1994). In other words, the conformational preferences, the identity, and the progression of intermediates that populate the folding reaction of a disulfide-rich protein are likely intrinsic properties depending on its specific primary sequence and mostly independent of the solution conditions, which affect mainly the folding kinetics.
10
S. Ventura and R.J.Y. Chang
Although, in principle, during in vitro oxidative folding experiments one would like to approximate physiological conditions by reproducing the ionic strength, pH, and redox potential of biologically relevant setting, still little is known on the specific environment an unfolded and reduced disulfide-rich protein faces inside the cell or extracellularly. Importantly, natural redox agents such as ascorbate, tocopherol, and vitamin K are emerging as efficient catalyzers of oxidative folding reactions both in vivo and in vitro at physiological concentrations (Saaranen et al. 2010; Margittai et al. 2009), which suggest links between oxidative protein folding and redox cellular metabolism.
1.4 Diversity of Oxidative Folding Pathways Beginning with the characterization of the oxidative folding pathway of bovine trypsin inhibitor (BPTI) by Creighton in the mid-1970s (Creighton 1974), the folding and unfolding reactions of a large number of disulfide-containing proteins have been studied (Arolas et al. 2006; Mamathambika and Bardwell 2008). Surprisingly, these studies illustrated a high degree of diversity of folding pathways, especially for small disulfide-rich proteins (Arolas et al. 2006). The diversity of folding mechanisms has precluded transforming this vast amount of detailed experimental information into a unified folding theory that might allow us to somehow predict the folding pathway from the primary sequence. More significantly, even with proteins which share sequence homology and similar 3D structure, it is impossible to forecast whether they would fold by similar pathway. This is best illustrated by the cases of tick anticoagulant peptide (TAP) and BPTI, two proteins that share very similar conformation and identical disulfide connectivity but exhibit strikingly different oxidative folding pathways (Weissman and Kim 1991; Chang 1996; Chang and Li 2005). An oxidative folding pathway is characterized by three main features: (a) the level of the heterogeneity of the folding intermediates that populate along the folding pathway, (b) the prevalence of intermediate species displaying native disulfide bonds and native-like structures, and (c) the accumulation along the folding reaction of fully oxidized scrambled isomers containing at least two non-native disulfides (Chang 2004, 2008). The two extreme mechanisms of oxidative folding are illustrated by (a) proteins that fold through a selected number of intermediates containing exclusively native disulfide bonds and (b) proteins fold via a highly heterogeneous population of intermediates containing mostly non-native disulfides, including fully oxidized scramble isomers. For many proteins, folding pathways share common characteristics of both extreme models.
1.4.1 Folding Through Native Intermediates: The Case of Leech-Derived Trypsin Inhibitor BTPI represents the best characterized model of small disulfide-rich protein which folds through a limited number of intermediates that acquire predominantly native
1 Protein Oxidative Folding
11
Fig. 1.3 Oxidative folding of LDTI. (a) Ribbon plot of the 3D structure of LDTI (3 disulfide bonds and 46 residues). The Protein Data Bank ID is 1LDT and the figure was prepared with PyMOL. The native disulfide bonds are indicated. (b) Scheme of the oxidative folding pathway of LDTI. R and N correspond to the fully reduced/unfolded and native states, respectively. The disulfide pairings of major folding intermediates are indicated. 1S is an ensemble of intermediates containing one disulfide bond. Species in which all the formed disulfide bonds are native are shown in gray and the rest in black
disulfide connectivity and native-like local structures. The folding pathway of BPTI has been widely discussed and reviewed (Creighton 1979, 1990; Creighton and Goldenberg 1984; Weissman and Kim 1991, 1992; Dadlez 1997). Therefore, in this chapter we have chosen to exemplify the folding behavior of this type of proteins, which we will denote as BPTI-like, by describing the folding pathway of leech-derived trypsin inhibitor (LDTI), characterized recently from the kinetic and structural point of view by one of our groups. LDTI is Kazal-type protease inhibitor isolated from the medicinal leech Hirudo medicinalis (Sommerhoff et al. 1994; Auerswald et al. 1994; Arolas et al. 2006). It binds tightly to human tryptase, a trypsin-like serine proteinase involved in allergic and inflammatory diseases. LDTI consists of 46 residues and contains three disulfide bonds (Cys4–Cys29, Cys6–Cys25, and Cys14–Cys40) (Fig. 1.3). LDTI folds into a defined conformation comprising a short central a-helix and a small triple-stranded antiparallel b-sheet (Di Marco and Priestle 1997; Stubbs et al. 1997; Muhlhahn et al. 1994). The first two disulfide bonds stabilize the a-helix by connecting
12
S. Ventura and R.J.Y. Chang
this secondary structure element to the N terminus of the inhibitor forming a CSH motif that is also shared by other bioactive peptides. The third disulfide stabilizes the small b-sheet by linking the b1-strand and the end of the b3-strand. The folding of LDTI from the initially denatured and reduced state to the native conformation proceeds through a sequential oxidation of its cysteine residues (Fig. 1.3). Up to 74 different folding intermediates can theoretically populate along the oxidative folding pathway of LDTI. Nevertheless, similar to the case of BPTI, only five of these disulfide isomers accumulate significantly (Arolas et al. 2008a). The initial oxidation of one of the LDTI disulfides renders an ensemble of 1S intermediates that rapidly reaches equilibrium. The distribution of disulfide bonds in this population does not obey to loop entropy contributions, but rather a preferential accumulation of the native bond linking b1- and b3-strand is observed. Structural studies of this 1S intermediate (PantojaUceda et al. 2009) indicate that, very likely, the triple-stranded antiparallel b-sheet of the inhibitor is already formed in this form. Secondary structure preferences might promote the establishment of transient interactions between the b-sheet elements and as a result Cys14 and Cys40 would come close to each other preferentially, relative to the other free Cys residues, allowing the formation of Cys14–Cys40 native disulfide bond. The new disulfide bond would significantly stabilize the preformed b-sheet and in turn this stable secondary structure would preferentially protect the disulfide from subsequent reduction reactions thus facilitating in kinetic terms the accumulation of this species as a major folding intermediate in the 1S ensemble. The formation of the native bond in the b-sheet of LDTI illustrates how, in BPTI-like proteins, conformational folding and native disulfide bond formation are tightly connected processes during oxidative folding. Only three 2S folding species accumulate in the oxidative folding of LDTI. Two of them, (Cys4–Cys29, Cys14–Cys40) and (Cys6–Cys25, Cys14–Cys40), originate directly from the further oxidation of the predominant 1S folding intermediate (Cys14–Cys40). The rest of 1S intermediates rearrange their disulfide bonds and oxidize their free thiols to form (Cys4–Cys29, Cys6–Cys25), a 2S intermediate which lacks the disulfide bond connecting the b-sheet. Importantly, only native disulfide bonds are present in the three LDTI 2S intermediates. This implies that, in principle, the three 2S have the potential to become productive species, since the direct oxidation of their remaining two free cysteines would form a natively bound, stable, and functional inhibitor in all cases. However, the experimental data clearly indicate that only (Cys6–Cys25, Cys14–Cys40) is the productive intermediate which is able to reach native LDTI. This suggests that the other two native-like 2S intermediates need to reshuffle their disulfide bonds to render the productive 2S species prior to their conversion to the native structure. In (Cys4–Cys29, Cys6–Cys25) the formation of the third disulfide bond is kinetically impeded whereas in (Cys4– Cys29, Cys14–Cys40) the free Cys25 is highly reactive, promoting the reshuffling of the preformed disulfide bonds instead of bonding to Cys6, which is in fact far away in this particular conformer. In summary, LDTI fold via BPTI-like mechanism in which the folding pathway comprises only four natively bonded intermediates without significant accumulation of scrambled forms. In BPTI-like proteins, the low complexity of the conformational space that the polypeptide chain has to explore to attain the native conformation makes the folding reaction fast and highly efficient.
1 Protein Oxidative Folding
13
Fig. 1.4 Oxidative folding of hirudin. (a) Ribbon plot of the 3D structure of hirudin (3 disulfide bonds and 65 residues). The Protein Data Bank ID is 4HTC and the figure was prepared with PyMOL. The native disulfide bonds are indicated. (b) Scheme of the oxidative folding pathway of hirudin. R and N correspond to the fully reduced/unfolded and native states, respectively. 1S and 2S correspond to ensembles of intermediates containing one and two disulfide bonds, respectively. 3S indicates scrambled species in which all the cysteines are oxidized but they contain at least two non-native disulfide bonds. Species in which all the formed disulfide bonds are native are shown in gray and the rest in black
1.4.2 Folding Through Scrambled Isomers: The Case of Hirudin Similar to LDTI, hirudin is a protein isolated from the medicinal leech H. medicinalis (Chang 1983). Hirudin is a 65-residue long thrombin inhibitor that folds into two structurally and functionally distinguishable domains (Fig. 1.4). It consists of an N-terminal globular domain formed by a four-stranded b-sheet and stabilized by the presence of three disulfide bonds (Cys6–Cys14, Cys16–Cys28, and Cys22–Cys39). It also comprises an acidic C-terminal domain devoid of any regular secondary structure (Fig. 1.4) (Folkers et al. 1989; Rydel et al. 1990; Grutter et al. 1990). In contrast to LDTI, the oxidative folding reaction of hirudin is characterized by the absence of a predominant folding route, without any preferential accumulation of folding intermediates (Fig. 1.4) (Chatrenet and Chang 1992, 1993; Chang 1994). Apparently, the folding of hirudin from its initially denatured and reduced state is independent of any conformational constraint and responds to a “trial-and-error” process in which all the free cysteines become similarly involved in complex disulfide reshuffling reactions that finally result in the formation of the native state. Accordingly, the metastable intermediates that accumulate in the folding reaction of this protease inhibitor correspond mainly to 3S-scrambled isomers, that is, hirudin species with at least two non-native disulfide bonds. The folding of this kind of
14
S. Ventura and R.J.Y. Chang
proteins, which we denote as hirudin-like proteins, comprises two different steps: an initial stage characterized by nonspecific packing into collapsed structures and a final stage in which these conformations consolidate to form the native functional protein. The first step consists of the sequential oxidation of hirudin-free cysteines to render equilibrated ensembles of 1S, 2S, and scrambled 3S intermediates. All these populations display a high degree of conformational heterogeneity without any apparent bias in the intermediate distribution resulting from protein structural propensities or noncovalent contacts. The theoretical number of 1S and 2S species for a three disulfide-containing protein is 60, from these, at least 30 have been shown to accumulate in the folding reaction, and 11 of the 14 theoretical 3S scrambled isomers have been identified (Chang et al. 1995). This sums ~40 folding intermediates. Therefore, the number of transient species that accumulate in the oxidative folding of hirudin is one order of magnitude higher than those populating the folding reaction of LDTI, strongly suggesting that the forces driving the packing-stage in hirudin have a nonspecific nature and probably respond to the collapse of previously exposed hydrophobic side chains. This lack of selectivity in the side chain and backbone interactions results in an almost unbiased formation of scrambled forms containing most of the possible disulfide bond combinations at this stage. In these species, all the six cysteine residues are involved in disulfide bonds and accordingly no free thiolates exist that can attack the preformed disulfides and promote their reshuffling toward new energy minima in the conformational space. This is the reason why scrambled isomers act as major kinetic traps, their conversion into the native structure being the rate-limiting step in the pathway. Accordingly, the consolidation step is strongly accelerated by the presence of thiol catalysts that allow newly formed scrambled isomers to initiate the required disulfide rearrangements in order to attain the native connectivity. In the past, it has long been debated if scrambled isomers constitute dead-end species or are instead on-pathway intermediates. It is now generally accepted, that they are essential and productive species in the folding pathways of hirudin-like proteins. In hirudin-like proteins, the formation of the native species from scrambled forms requires the establishment of nativelike noncovalent interactions. However, in contrast to BPTI or LDTI, there is apparently no sequential or even a preferential pathway for the formation of such contacts, in such a way that the consolidation stage follows again a “trial-and-error” process which renders the overall folding of hirudin-like proteins slow and inefficient.
1.4.3 Mixed Oxidative Folding Pathways: The Case of Epidermal Growth Factor Folding pathways of many proteins exhibit both BPTI-like and hirudin-like mechanisms. One of these cases is the folding reaction of epidermal growth factor (EGF). EGF constitutes the founding member of a family of diverse proteins having very similar structure but different function. EGF is a 53-residue protein consisting of
1 Protein Oxidative Folding
15
Fig. 1.5 Oxidative folding of EGF. (a) Ribbon plot of the 3D structure of EGF (3 disulfide bonds and 53 residues). The figure was prepared with PyMOL. The native disulfide bonds are indicated. (b) Scheme of the oxidative folding pathway of EGF. R and N correspond to the fully reduced/ unfolded and native states, respectively. The disulfide pairings of the major folding intermediates is indicated. 1S is an ensemble of intermediates containing one disulfide bond. 3S indicate scrambled species in which all the cysteines are oxidize but they contain at least two non-native disulfide bonds. Species in which all the formed disulfide bonds are native are shown in gray and the rest in black
three different loops stabilized by three intramolecular disulfide bonds (Fig. 1.5) (Montelione et al. 1992; Ogiso et al. 2002). The N-terminal A loop is linked by the Cys6−Cys20 disulfide bond. The B loop forms a b-hairpin constrained by a disulfide bond between Cys14 and Cys31. Finally, the C-terminal C loop includes the third Cys33−Cys42 disulfide linkage. The folding of reduced and denatured EGF to the native conformation involves initially the formation of a heterogeneous and transient population of 1S intermediates (Wu et al. 1998; Chang et al. 2001). This ensemble transforms rapidly into a single, predominant 2S disulfide species, which represents up to 85% of the total protein along the folding pathway. This predominant 2S intermediate comprises two native disulfide bonds (Cys14−Cys31 and Cys33−Cys42), indicating that a single oxidative reaction between the remaining two free cysteines (Cys6 and Cys20) in this conformer would allow formation of native EGF. Surprisingly, the two free thiols of Cys6 and Cys20 cannot react directly. Instead, kinetic analysis indicates that substantial unfolding and disulfide rearrangement through the formation of scrambled isomers is an obligatory step to reach the
16
S. Ventura and R.J.Y. Chang
native conformation. The reorganization of these totally oxidized species into the native form constitutes the major rate-limiting step in the folding of EGF, significantly slowing the overall folding reaction (Fig. 1.5). Therefore, the folding of EGF resembles that of hirudin in the initial accumulation of a complex ensemble of intermediates and at the final rate-limiting step involving the conversion of scrambled isomers into the functional structure. However, like in the case of LDTI, the major folding intermediate corresponds to a reduced native-like bonded species. The case of EGF illustrates how the search for a functional conformation might result in a complicated folding scenario in which already folded native-like structural elements are not preferentially retained, as will be expected, but lost during the progress of the reaction. This suggests that the folding pathway is shaped mainly on the basis of final thermodynamic requirements.
1.5 Structural Characterization of Folding Intermediates For small proteins without disulfide bonds, the topology of the native structure usually suffices to determine, in general terms, the folding rate and the location of the rate-limiting step in the pathway (transition state), independently of the proteinspecific sequence (Plaxco et al. 1998; Guerois and Serrano 2000). The divergent folding behavior of BPTI and TAP clearly illustrates that this is not the case for small disulfide-containing proteins, because conformational folding is necessarily linked to disulfide-bond formation. In these proteins, the formation of native disulfide bonds can promote and accelerate the arrangement of local or global secondary or tertiary structural elements and conversely strong structural propensities might force the preferential formation of native disulfide bonds. However, it also occurs that the rather unspecific hydrophobic collapse that initiates the folding process of many proteins lead to the formation of wrongly paired disulfide bonds which, by restraining the polypeptide chain would prevent the subsequent formation of native-like structural elements. In turn, the conformational properties of both natively and non-natively connected folding intermediates affect the reactivity, accessibility, and proximity of their free cysteines and disulfide bonds, making very difficult to predict the flow of folding species along the pathway. Because in disulfide-rich proteins, neither the sequence nor the native structure appear to be informative by themselves about the conformational space these proteins have to explore to fold into the native conformation, we can only try to understand it by exploring at the atomic level the specific structural properties of the metastable folding intermediates that populate along the folding course. Chemically inert analogs for structural studies have been usually obtained by derivatization of the free thiols present in the intermediate (Kortemme et al. 1996). However, the packing of these disulfide intermediates is usually affected by the size and charge of the blocking reagent. Mutation of the intermediate free cysteines by protein engineering to alanines or serines generates inert analogs, which are expected to closely mimic the conformational properties of folding intermediates (Laity et al. 1997;
1 Protein Oxidative Folding
17
Fig. 1.6 Solution structures of genuine folding intermediates of oxidative folding. Ribbon plots of the 3D average structures of major folding intermediates of (a) LCI, (b) MCoTI-II, (c, d) TCI, and (e–g) LDTI, respectively. The Protein Data Bank ID codes are as follows: 1ZFL, 2PO8, 2K2Y, 2K2Z, 2KMP, 2KMQ, and 2KMR, respectively. The figure was prepared with PyMOL
Shimotakahara et al. 1997). However, the hydrophobic character of alanine might further stabilize the analog relative to the intermediate when the original thiols are buried in the structure and destabilize entropically the protein when they free cysteines that are originally exposed (Arolas et al. 2009). The opposite effect is expected for cysteines to serines substitutions. To avoid these problems, efforts are now being focused on solving the structure of genuine folding intermediates isolated directly from the oxidative folding or reductive unfolding reactions under acid conditions, in which isolated and unmodified intermediates are stable yet thiol/disulfide exchange is almost unnoticeable. Because these intermediates display significant conformational flexibility in comparison to the native state, they are not amenable to crystallographic and X-ray studies and their structures should be addressed in solution by NMR spectrometry of the previously 15N and/or 13C labeled species. Seven of such structures have been solved in the last 5 years. Protease inhibitors are in many cases small-disulfide inhibitors and therefore have become important models to study oxidative folding reactions. Accordingly, all genuine oxidative folding intermediates structurally solved to date correspond to conformers of protease inhibitors: LDTI (Pantoja-Uceda et al. 2009), Leech carboxypeptidase inhibitor (LCI) (Arolas et al. 2005), Momordica cochinchinensis trypsin inhibitor II (MCoTI-II) (Cemazar et al. 2008) and tick carboxypeptidase inhibitor (TCI) (Arolas et al. 2008b). It is important to note that, although the number of free cysteines in these species varies from two to six, in all cases, the formed disulfide bonds display native connectivity (Fig. 1.6). However, it remains a challenging task in the characterization of structural properties of non-natively bound intermediates and specifically scrambled species. This is crucial since it would allow us to decipher at the molecular level why they are such strong kinetic and, in some cases, thermodynamic
18
S. Ventura and R.J.Y. Chang
traps in the folding pathway. This is also important because the accumulation of scrambled isomers has been associated with pathogenic phenotypes in genetic diseases like hypercolesterolemia (Arias-Moreno et al. 2008). The structural characterization of the above-mentioned set of folding intermediates (Fig. 1.6) has allowed us to decipher some molecular clues to understand why a particular productive intermediate is able to convert to the native state through direct oxidation of their free thiols and why other species are nonproductive, meaning that they have to rearrange the connectivity of preformed disulfide bonds to render a productive folding intermediate in order to convert to the native species. Nonproductive folding intermediates, usually posses a native-like conformation in which both the free thiols and the preformed disulfide bonds are buried in the structure and have a low accessibility to the solvent, in such a way that they become almost insensitive to external thiols (Arolas et al. 2004). The protection of the free cysteines in a stable structure impedes its direct oxidation and the progress of the folding reaction. Accordingly, structured nonproductive folding intermediates act as strong kinetic traps. As stated above, the lack of one or more covalent disulfide bonds render these conformers flexible, relative to the native state. On one hand, this flexibility stabilizes the intermediate from the entropic point of view, but on the other hand it promotes structural fluctuations that allow it to escape from this particular energy minimum (Arolas et al. 2005). These intermediates usually correspond to what Scheraga and coworkers defined as “disulfide-insecure” species (Narayan et al. 2003b; Creighton 1988b), because, due to their similar burial, conformational changes that expose the free thiols also expose at the same time the disulfide bonds, which are attacked by the reactive cysteines promoting their reshuffling instead of the formation of the missing native disulfide bonds. In productive folding intermediates, the already formed native disulfide bonds are preferentially protected in regular secondary structural elements or located at the hydrophobic core, in both cases with reduced accessibility or exchange with the solvent. This is in contrast to their free thiols, which are either already accessible to solvent or, more frequently, located in flexible and exposed protein regions. They are “disulfide-secure” species (Welker et al. 2001b; Narayan et al. 2003b), in the sense that the rigidity of the native-like intermediate scaffold protects the native disulfide bonds from conformational fluctuations. This would allow the remaining and solvent accessible free cysteines to undergo direct oxidation to form the native protein. A long pursued goal in the field of oxidative folding has been the characterization of the structural properties of the complete ensemble of species that populate the folding pathway of a disulfide-rich protein. This objective has been recently attained by one of our groups for the LDTI protein model. The obtained results confirm that conformational folding and the assembly of native disulfide bonds need to occur coordinately in a productive folding reaction (Pantoja-Uceda et al. 2009). Importantly, although structural propensities define essentially the number and conformation of the different folding intermediates in the pathway, this effect is kinetically controlled by the reactivity and accessibility of thiols and disulfide bonds in these molecules. Therefore, the interplay of structural and kinetic constraints appears as a recurrent factor governing the folding of disulfide-rich proteins.
1 Protein Oxidative Folding
19
1.6 Concluding Remarks The vast amounts of data obtained from the study of the oxidative folding reactions of an important number of evolutionarily unrelated disulfide-rich protein models constitute the ground work for our future efforts to understand how the linear information contained in the protein sequence is precisely translated into the 3D conformation. A first step in this direction would be the understanding of the sequential and structural determinants that determine the heterogeneity of a protein folding pathway. In this respect, the resolution of an increasing number of genuine folding intermediates at atomic level is shedding light on the structural, thermodynamic, and kinetic factors shaping the complex interplay between conformational folding and disulfide formation. This suggests that it is likely that in the near future we can arrive at the construction of a clear and uniform description of the mechanism of oxidative protein folding.
References Abkevich VI, Shakhnovich EI (2000) What can disulfide bonds tell us about protein energetics, function and folding: simulations and bioinformatics analysis. J Mol Biol 300:975–985 Anfinsen CB (1972) The formation and stabilization of protein structure. Biochem J 128:737–749 Arias-Moreno X, Arolas JL, Aviles FX, Sancho J, Ventura S (2008) Scrambled isomers as key intermediates in the oxidative folding of ligand binding module 5 of the low density lipoprotein receptor. J Biol Chem 283:13627–13637 Arolas JL, Bronsoms S, Lorenzo J, Aviles FX, Chang JY, Ventura S (2004) Role of kinetic intermediates in the folding of leech carboxypeptidase inhibitor. J Biol Chem 279:37261–37270 Arolas JL, D’silva L, Popowicz GM, Aviles FX, Holak TA, Ventura S (2005) NMR structural characterization and computational predictions of the major intermediate in oxidative folding of leech carboxypeptidase inhibitor. Structure 13:1193–1202 Arolas JL, Aviles FX, Chang JY, Ventura S (2006) Folding of small disulfide-rich proteins: clarifying the puzzle. Trends Biochem Sci 31:292–301 Arolas JL, Bronsoms S, Aviles FX, Ventura S, Sommerhoff CP (2008a) Oxidative folding of leechderived tryptase inhibitor via native disulfide-bonded intermediates. Antioxid Redox Signal 10:77–85 Arolas JL, Pantoja-Uceda D, Ventura S, Blanco FJ, Aviles FX (2008b) The NMR structures of the major intermediates of the two-domain tick carboxypeptidase inhibitor reveal symmetry in its folding and unfolding pathways. J Biol Chem 283:27110–27120 Arolas JL, Castillo V, Bronsoms S, Aviles FX, Ventura S (2009) Designing out disulfide bonds of leech carboxypeptidase inhibitor: implications for its folding, stability and function. J Mol Biol 392:529–546 Auerswald EA, Morenweiser R, Sommerhoff CP, Piechottka GP, Eckerskorn C, Gurtler LG, Fritz H (1994) Recombinant leech-derived tryptase inhibitor: construction, production, protein chemical characterization and inhibition of HIV-1 replication. Biol Chem Hoppe Seyler 375:695–703 Baldwin RL (1994) Finding intermediates in protein folding. Bioessays 16:207–210 Bryngelson JD, Onuchic JN, Socci ND, Wolynes PG (1995) Funnels, pathways, and the energy landscape of protein folding: a synthesis. Proteins 21:167–195 Camacho CJ, Thirumalai D (1995) Modeling the role of disulfide bonds in protein folding: entropic barriers and pathways. Proteins 22:27–40
20
S. Ventura and R.J.Y. Chang
Cemazar M, Joshi A, Daly NL, Mark AE, Craik DJ (2008) The structure of a two-disulfide intermediate assists in elucidating the oxidative folding pathway of a cyclic cystine knot protein. Structure 16:842–851 Chang JY (1983) The functional domain of hirudin, a thrombin-specific inhibitor. FEBS Lett 164:307–313 Chang JY (1994) Controlling the speed of hirudin folding. Biochem J 300:643–650 Chang JY (1996) The disulfide folding pathway of tick anticoagulant peptide (TAP), a Kunitz-type inhibitor structurally homologous to BPTI. Biochemistry 35:11702–11709 Chang JY (2004) Evidence for the underlying cause of diversity of the disulfide folding pathway. Biochemistry 43:4522–4529 Chang JY (2008) Diversity of folding pathways and folding models of disulfide proteins. Antioxid Redox Signal 10:171–177 Chang JY, Li L (2005) Divergent folding pathways of two homologous proteins, BPTI and tick anticoagulant peptide: compartmentalization of folding intermediates and identification of kinetic traps. Arch Biochem Biophys 437:85–95 Chang JY, Schindler P, Chatrenet B (1995) The disulfide structures of scrambled hirudins. J Biol Chem 270:11992–11997 Chang JY, Li L, Lai PH (2001) A major kinetic trap for the oxidative folding of human epidermal growth factor. J Biol Chem 276:4845–4852 Chatrenet B, Chang JY (1992) The folding of hirudin adopts a mechanism of trial and error. J Biol Chem 267:3038–3043 Chatrenet B, Chang JY (1993) The disulfide folding pathway of hirudin elucidated by stop/go folding experiments. J Biol Chem 268:20988–20996 Creighton TE (1974) Renaturation of the reduced bovine pancreatic trypsin inhibitor. J Mol Biol 87:563–577 Creighton TE (1979) Electrophoretic analysis of the unfolding of proteins by urea. J Mol Biol 129:235–264 Creighton TE (1986) Disulfide bonds as probes of protein folding pathways. Methods Enzymol 131:83–106 Creighton TE (1988a) The protein folding problem. Science 240(267):344 Creighton TE (1988b) Toward a better understanding of protein folding pathways. Proc Natl Acad Sci USA 85:5082–5086 Creighton TE (1990) Protein folding. Biochem J 270:1–16 Creighton TE (1997) Protein folding coupled to disulphide bond formation. Biol Chem 378:731–744 Creighton TE, Goldenberg DP (1984) Kinetic role of a meta-stable native-like two-disulphide species in the folding transition of bovine pancreatic trypsin inhibitor. J Mol Biol 179:497–526 Dadlez M (1997) Hydrophobic interactions accelerate early stages of the folding of BPTI. Biochemistry 36:2788–2797 Daggett V, Fersht A (2003) The present view of the mechanism of protein folding. Nat Rev Mol Cell Biol 4:497–502 Darby NJ, Creighton TE (1993) Dissecting the disulphide-coupled folding pathway of bovine pancreatic trypsin inhibitor. Forming the first disulphide bonds in analogues of the reduced protein. J Mol Biol 232:873–896 Darby N, Creighton TE (1997) Probing protein folding and stability using disulfide bonds. Mol Biotechnol 7:57–77 Di Marco S, Priestle JP (1997) Structure of the complex of leech-derived tryptase inhibitor (LDTI) with trypsin and modeling of the LDTI-tryptase system. Structure 5:1465–1474 Dill KA, Ozkan SB, Weikl TR, Chodera JD, Voelz VA (2007) The protein folding problem: when will it be solved? Curr Opin Struct Biol 17:342–346 Dobson CM, Evans PA (1988) Protein structure. Trapping folding intermediates. Nature 335:666–667 Eisenberg D, Marcotte EM, Xenarios I, Yeates TO (2000) Protein function in the post-genomic era. Nature 405:823–826 Fetrow JS, Giammona A, Kolinski A, Skolnick J (2002) The protein folding problem: a biophysical enigma. Curr Pharm Biotechnol 3:329–347
1 Protein Oxidative Folding
21
Folkers PJ, Clore GM, Driscoll PC, Dodt J, Kohler S, Gronenborn AM (1989) Solution structure of recombinant hirudin and the Lys-47-Glu mutant: a nuclear magnetic resonance and hybrid distance geometry-dynamical simulated annealing study. Biochemistry 28:2601–2617 Gruebele M (1999) The fast protein folding problem. Annu Rev Phys Chem 50:485–516 Grutter MG, Priestle JP, Rahuel J, Grossenbacher H, Bode W, Hofsteenge J, Stone SR (1990) Crystal structure of the thrombin-hirudin complex: a novel mode of serine protease inhibition. EMBO J 9:2361–2365 Guerois R, Serrano L (2000) The SH3-fold family: experimental evidence and prediction of variations in the folding pathways. J Mol Biol 304:967–982 Hogg PJ (2003) Disulfide bonds as switches for protein function. Trends Biochem Sci 28:210–214 Hogg PJ (2009) Contribution of allosteric disulfide bonds to regulation of hemostasis. J Thromb Haemost 7(Suppl 1):13–16 Honig B, Ray A, Levinthal C (1976) Conformational flexibility and protein folding: rigid structural fragments connected by flexible joints in subtilisin BPN. Proc Natl Acad Sci USA 73:1974–1978 Karplus M, Weaver DL (1994) Protein folding dynamics: the diffusion-collision model and experimental data. Protein Sci 3:650–668 Kibria FM, Lees WJ (2008) Balancing conformational and oxidative kinetic traps during the folding of bovine pancreatic trypsin inhibitor (BPTI) with glutathione and glutathione disulfide. J Am Chem Soc 130:796–797 Kim PS, Baldwin RL (1982) Specific intermediates in the folding reactions of small proteins and the mechanism of protein folding. Annu Rev Biochem 51:459–489 Kortemme T, Hollecker M, Kemmink J, Creighton TE (1996) Comparison of the (30–51, 14–38) two-disulphide folding intermediates of the homologous proteins dendrotoxin K and bovine pancreatic trypsin inhibitor by two-dimensional 1H nuclear magnetic resonance. J Mol Biol 257:188–198 Laity JH, Lester CC, Shimotakahara S, Zimmerman DE, Montelione GT, Scheraga HA (1997) Structural characterization of an analog of the major rate-determining disulfide folding intermediate of bovine pancreatic ribonuclease A. Biochemistry 36:12683–12699 Mamathambika BS, Bardwell JC (2008) Disulfide-linked protein folding pathways. Annu Rev Cell Dev Biol 24:211–235 Margittai E, Csala M, Mandl J, Banhegyi G (2009) Participation of low molecular weight electron carriers in oxidative protein folding. Int J Mol Sci 10:1346–1359 Montelione GT, Wuthrich K, Burgess AW, Nice EC, Wagner G, Gibson KD, Scheraga HA (1992) Solution structure of murine epidermal growth factor determined by NMR spectroscopy and refined by energy minimization with restraints. Biochemistry 31:236–249 Muhlhahn P, Czisch M, Morenweiser R, Habermann B, Engh RA, Sommerhoff CP, Auerswald EA, Holak TA (1994) Structure of leech derived tryptase inhibitor (LDTI-C) in solution. FEBS Lett 355:290–296 Narayan M, Welker E, Wedemeyer WJ, Scheraga HA (2000) Oxidative folding of proteins. Acc Chem Res 33:805–812 Narayan M, Welker E, Scheraga HA (2003a) Native conformational tendencies in unfolded polypeptides: development of a novel method to assess native conformational tendencies in the reduced forms of multiple disulfide-bonded proteins. J Am Chem Soc 125:2036–2037 Narayan M, Welker E, Wanjalla C, Xu G, Scheraga HA (2003b) Shifting the competition between the intramolecular reshuffling reaction and the direct oxidation reaction during the oxidative folding of kinetically trapped disulfide-insecure intermediates. Biochemistry 42:10783–10789 Nishimura C, Lietzow MA, Dyson HJ, Wright PE (2005) Sequence determinants of a protein folding pathway. J Mol Biol 351:383–392 Ogiso H, Ishitani R, Nureki O, Fukai S, Yamanaka M, Kim JH, Saito K, Sakamoto A, Inoue M, Shirouzu M, Yokoyama S (2002) Crystal structure of the complex of human epidermal growth factor and receptor extracellular domains. Cell 110:775–787 Pantoja-Uceda D, Arolas JL, Aviles FX, Santoro J, Ventura S, Sommerhoff CP (2009) Deciphering the structural basis that guides the oxidative folding of leech-derived tryptase inhibitor. J Biol Chem 284:35612–35620
22
S. Ventura and R.J.Y. Chang
Plaxco KW, Simons KT, Baker D (1998) Contact order, transition state placement and the refolding rates of single domain proteins. J Mol Biol 277:985–994 Poland DC, Scheraga HA (1965) Comparison of theories of the helix-coil transition in polypeptides. J Chem Phys 43:2071–2074 Ptitsyn OB (1991) How does protein synthesis give rise to the 3D-structure? FEBS Lett 285:176–181 Rydel TJ, Ravichandran KG, Tulinsky A, Bode W, Huber R, Roitsch C, Fenton JW II (1990) The structure of a complex of recombinant hirudin and human alpha-thrombin. Science 249:277–280 Saaranen MJ, Karala AR, Lappi AK, Ruddock LW (2010) The role of dehydroascorbate in disulfide bond formation. Antioxid Redox Signal 12:15–25 Sanchez R, Pieper U, Melo F, Eswar N, Marti-Renom MA, Madhusudhan MS, Mirkovic N, Sali A (2000) Protein structure modeling for structural genomics. Nat Struct Biol 7(Suppl):986–990 Scheraga HA, Konishi Y, Rothwarf DM, Mui PW (1987) Toward an understanding of the folding of ribonuclease A. Proc Natl Acad Sci USA 84:5740–5744 Shimotakahara S, Rios CB, Laity JH, Zimmerman DE, Scheraga HA, Montelione GT (1997) NMR structural analysis of an analog of an intermediate formed in the rate-determining step of one pathway in the oxidative folding of bovine pancreatic ribonuclease A: automated analysis of 1H, 13C, and 15N resonance assignments for wild-type and [C65S, C72S] mutant forms. Biochemistry 36:6915–6929 Sommerhoff CP, Sollner C, Mentele R, Piechottka GP, Auerswald EA, Fritz H (1994) A Kazaltype inhibitor of human mast cell tryptase: isolation from the medical leech Hirudo medicinalis, characterization, and sequence analysis. Biol Chem Hoppe Seyler 375:685–694 Stubbs MT, Morenweiser R, Sturzebecher J, Bauer M, Bode W, Huber R, Piechottka GP, Matschiner G, Sommerhoff CP, Fritz H, Auerswald EA (1997) The three-dimensional structure of recombinant leech-derived tryptase inhibitor in complex with trypsin. Implications for the structure of human mast cell tryptase and its inhibition. J Biol Chem 272:19931–19937 Ventura S (2008) Oxidative protein folding: from the test tube to in vivo insights. Antioxid Redox Signal 10:51–53 Wedemeyer WJ, Welker E, Narayan M, Scheraga HA (2000a) Disulfide bonds and protein folding. Biochemistry 39:7032 Wedemeyer WJ, Welker E, Narayan M, Scheraga HA (2000b) Disulfide bonds and protein folding. Biochemistry 39:4207–4216 Wedemeyer WJ, Xu X, Welker E, Scheraga HA (2002) Conformational propensities of protein folding intermediates: distribution of species in the 1S, 2S, and 3S ensembles of the [C40A, C95A] mutant of bovine pancreatic ribonuclease A. Biochemistry 41:1483–1491 Weissman JS (1995) All roads lead to Rome? The multiple pathways of protein folding. Chem Biol 2:255–260 Weissman JS, Kim PS (1991) Reexamination of the folding of BPTI: predominance of native intermediates. Science 253:1386–1393 Weissman JS, Kim PS (1992) Kinetic role of nonnative species in the folding of bovine pancreatic trypsin inhibitor. Proc Natl Acad Sci USA 89:9900–9904 Welker E, Narayan M, Wedemeyer WJ, Scheraga HA (2001a) Structural determinants of oxidative folding in proteins. Proc Natl Acad Sci 98:2312–2316 Welker E, Wedemeyer WJ, Narayan M, Scheraga HA (2001b) Coupling of conformational folding and disulfide-bond reactions in oxidative folding of proteins. Biochemistry 40:9059–9064 Welker E, Hathaway L, Scheraga HA (2004) A new method for rapid characterization of the folding pathways of multidisulfide-containing proteins. J Am Chem Soc 126:3720–3721 Wouters MA, Fan SW, Haworth NL (2010) Disulfides as redox switches: from molecular mechanisms to functional significance. Antioxid Redox Signal 12:53–91 Wu J, Yang Y, Watson JT (1998) Trapping of intermediates during the refolding of recombinant human epidermal growth factor (hEGF) by cyanylation, and subsequent structural elucidation by mass spectrometry. Protein Sci 7:1017–1028 Zavodszky M, Chen CW, Huang JK, Zolkiewski M, Wen L, Krishnamoorthi R (2001) Disulfide bond effects on protein stability: designed variants of Cucurbita maxima trypsin inhibitor-V. Protein Sci 10:149–160
Chapter 2
The Case of Oxidative Folding of Ribonuclease A: Factors Impacting Fold Maturation of ER-Processed Proteins Mahesh Narayan
Abstract Proteins that are membrane-bound or secreted outside the cell often possess disulfide bonds. The maturation of such proteins is termed oxidative protein folding and takes place within the endoplasmic reticulum (ER) of eukaryotic cells prior to their export. The events that comprise oxidative folding are complex. A variety of intrinsic and extrinsic factors finely orchestrate the formation of a biologically viable molecule almost from the time of its genesis in the ribosome. This chapter recapitulates our current understanding of factors impacting fold maturation of ER-processed proteins which emerges from a large number of in vitro regeneration experiments performed in a number of laboratories on several small single-domain disulfide bondcontaining proteins. The focus here is on advances made in our understanding of oxidative folding through studies on the folding pathways of bovine pancreatic ribonuclease A which were primarily undertaken by the Scheraga laboratory over 30 years. In addition, laboratory techniques and biophysical and biochemical manipulations which facilitate the study of oxidative protein folding are also discussed. Keywords Fold maturation • Thiol–disulfide exchange • Endoplasmic reticulum • Misfolding • Disulfide bond • Oxidative folding • Conformational folding • Trafficking
Abbreviations 1S 2S 3S
One-disulfide intermediates Two-disulfide intermediates Three-disulfide intermediates
M. Narayan (*) Department of Chemistry, The University of Texas at El Paso, 500 W. Univ. Ave., El Paso, TX 79968, USA e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_2, © Springer Science+Business Media, LLC 2011
23
24
M. Narayan
AEMTS 2-Aminoethyl methanethiosulfonate [(NH2) C2H5SSO2CH3] BPTI Bovine pancreatic trypsin inhibitor des [40, 95] RNase A Ribonuclease A lacking the disulfide bond between the cysteine residues denoted in the brackets des [65–72] RNase A Ribonuclease A lacking the disulfide bond between the cysteine residues denoted in the brackets DTT ox Oxidized dithiothreitol DTT red Reduced dithiothreitol EDTA Ethylenediaminetetraacetic acid GSH Reduced glutathione GSSG Oxidized glutathione HEPES 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid HPLC High-performance liquid chromatography MS Mass spectrometry RNase A Bovine pancreatic ribonuclease A Tris Tris(hydroxymethyl)aminomethane
2.1 Introduction The eukaryotic cell contains a specialized compartment, the endoplasmic reticulum, which is charged with the task of ensuring the maturation of those proteins that reside within membranes of cells, are secreted into specialized compartments, or secreted outside the cell (Tu and Weissman 2004). Such proteins normally contain disulfide bonds formed by oxidation of cysteine residues (Narayan et al. 2000). The acquisition of the biologically active, tertiary structure of such proteins is a complex event termed as oxidative folding (or oxidative regeneration) (Narayan et al. 2000; Arolas et al. 2006; Mamathambika and Bardwell 2008; Wedemeyer et al. 2000; Welker et al. 2001a, b; Chang 2004; Chang 2008; Woycechowsky and Raines 2000). It involves the acquisition of the native set of disulfide bonds from the fully reduced nascent polypeptide chain and a conformational folding reaction which preserves the native disulfide bonds within stable tertiary structure (Welker et al. 2001b; Wedemeyer et al. 2002). The endoplasmic reticulum in contrast to the cytosol has an environment that is oxidizing in nature and facilitates the formation of disulfide bonds. The chief redox couple in the ER is glutathione with the ratio of GSSG to GSH varying between 1:1 and 1:3, in contrast to 1:30 for the cytosol (Hwang et al. 1992). In addition to glutathione, there are a variety of chaperones and cofactors that facilitate the maturation of ER-processed substrate proteins prior to their secretion. Principal among these is the ER-resident oxidoreductase chaperone, protein disulfide isomerase (PDI) (Gilbert 1998; Wilkinson and Gilbert 2004; Hawkins and Freedman 1991; Fewell et al. 2001; Shin and Scheraga 2004; Tian et al. 2006; Xiao et al. 2004), which is similar in function to the prokaryotic DSB oxidoreductases (Hiniker and Bardwell 2003; Gleiter and Bardwell 2008). Lastly, molecular oxygen also plays a pivotal role in the ER by serving as the terminal electron acceptor via Ero1 (Romisch 2004, 2005).
2 Fold Maturation in the ER
25
The following sections describe the process by which the nascent polypeptide oxidatively regenerates in the ER. A variety of biochemical and biophysical studies and methodological inroads, referenced herein, have facilitated our current understanding of this process. These studies have focused on what can now be termed as “model” or “classical” disulfide bond-containing proteins. The list of proteins include, but is not limited to, bovine pancreatic ribonuclease A (RNase A), onconase, hirudin, hen-egg white lysozyme, tick anticoagulant peptide, leech carboxypeptidase, bovine pancreatic trypsin inhibitor (BPTI), etc. (Weissman and Kim 1995; Arolas et al. 2004, 2006; Narayan et al. 2000; Chang and Li 2005; Salamanca and Chang 2005; Chang et al. 2006; Lin and Chang 2007; Ardelt et al. 2008; Gahl et al. 2008; Mamathambika and Bardwell 2008). In this chapter, we have chosen to describe the folding pathway of RNase A as the paradigm for oxidative folding of ER-processed proteins (Narayan et al. 2000). This is especially appropriate considering that the protein adopts several different pathways by which it regenerates, possessing features that are representative of facets prevalent in the oxidative folding landscape of other disulfide bond-containing proteins. The chapter is divided according to the stages observed in the oxidative folding pathway of RNase A which generally applies to other proteins as well. In each step of the folding process factors that impact the regeneration rate and regeneration pathway are discussed. Extensive work performed in the Scheraga laboratory in unraveling the regeneration pathway of RNase A is widely referenced. It is also becoming increasingly clear that despite the presence of a finely tuned mechanism complete with an orchestra of folding adjuvants and highly evolved folding pathways, the traffic that enters the ER does not always exit via the secretory pathway (Romisch 2004, 2005; Pal et al. 2010a); the hazards are many and misfolding is common. Such misfolding, if terminal, becomes victim to the subcellular housekeeping machinery and exits the ER via the ERAD (ER associated degradation) pathway (Romisch 2004). Recent work in our lab is aimed at understanding how (mis)folding events within the ER can impact accumulation of cellular debris which is the bedrock of several neurodegenerative disorders (Wang and Narayan 2008; Pal et al. 2010a).
2.2 General Terms and Conditions Pertaining to the Oxidative Folding of Bovine Pancreatic Ribonuclease A Bovine pancreatic RNase A is a 124-amino acid endonuclease that cleaves single-stranded RNA (Crook et al. 1960) (Fig. 2.1). This activity was leveraged to determine refolding rate and yield during the early years when folding studies on this protein were initiated (Konishi et al. 1982). It is a member of the ribonuclease superfamily which includes onconase, RNase H, and RNase T1 among others (Bientema 1998). RNase A has four disulfide bonds at positions: Cys26– Cys84, Cys58–Cys110, Cys40–Cys95, and Cys65–Cys72. The tertiary structure is a two-layer a + b protein folded in half, with a deep cleft for binding RNA substrate.
26
M. Narayan
Fig. 2.1 Bovine pancreatic ribonuclease A (ribbon diagram)
2.2.1 Terminology Before proceeding toward description of its folding pathway, we describe terminology that the reader should become familiarized with: R, refers to the fully reduced RNase A. Thus, R would constitute unfolded RNase A possessing eight cysteines; des [40, 95], ribonuclease A lacking the disulfide bond between the cysteine residues denoted in the brackets; des [65, 72], ribonuclease A lacking the disulfide bond between the cysteine residues denoted in the brackets; 1S, one-disulfide-containing intermediates; 2S, two-disulfide-containing intermediates; 3S, three-disulfide-containing intermediates; 4S, four-disulfide-containing intermediates (also known as scrambled or dead-end species); nS, an isomeric ensemble of species containing n number of disulfide bonds; N, native protein.
2.2.2 Folding Conditions Folding conditions at which RNase A regeneration was studied in the Scheraga laboratory generally involved pH values at which chemical reactions required for successful disulfide bond formation could be easily captured. A folding buffer at
2 Fold Maturation in the ER
27
pH 8 and composed of 100 mM Tris–HCl and 1 mM EDTA was mostly employed. The choice of redox reagent, though initially GSH/GSSG, was dithiothreitol (DTT red)/trans-4,5-dihydroxy-1,2-dithiane (DTT ox). DTT ox/red was selected over glutathione for two reasons. (1) DTT ox is a weak oxidizing agent. Therefore, kinetic control over the folding is possible over a wide range of DTT ox concentrations. Furthermore, (2) DTT red is an unusually strong reducing agent, owing to its high conformational propensity to form a six-membered ring with an internal disulfide bond (Rothwarf and Scheraga 1992). Thus, intermediate mixed-disulfide states between protein thiols and DTT is unstable and poorly populated because the second thiol of DTT has a high propensity to close the ring, forming oxidized DTT and leaving behind a reduced disulfide bond (Rothwarf and Scheraga 1992, 1993a). This facilitates analytical techniques such as HPLC fractionation and mass analysis of intermediates with differing numbers of disulfide bonds (Rothwarf and Scheraga 1993a, b, c, d). A noteworthy disadvantage of DTT red is that it being a strong reducing agent, the regeneration rate of the reaction at higher protein concentrations becomes sensitive to the reducing equivalents (i.e., DTT red) generated during the course of the reaction. This can be overcome by maintaining low concentrations of the protein under study (Rothwarf and Scheraga 1993a).
2.2.3 Analytical Techniques and Methods Used to Study Regeneration Rates and Pathways Characterization of fold regeneration requires a good analytical handle on kinetics of formation and disappearance of all intermediates and the formation of the native species. This requirement in turn necessitates the separation of intermediates from one another, ability to isolate them, and requires their structural characterization. The Scheraga laboratory has pioneered the use of 2-aminoethyl methanethiosulfonate (AEMTS; Bruice and Kenyon 1982), a potent thiol-blocking agent, to facilitate analytical study of the regeneration process (Rothwarf and Scheraga 1993b). AEMTS attaches to free thiols and imparts a charge of +1 per blocked thiol. Blocking the thiol is necessary to determine the reaction progress. However, use of AEMTS facilitates fractionation of intermediates by charge. On a cation-exchange column, the AEMTS-blocked R with a positive charge of +8 relative to N is most strongly retained relative to other intermediates (Fig. 2.2). The order of elution of the intermediates is N, 3S, 2S, 1S, and R which are well separated on a strong cationexchange column (Xu et al. 1996). Furthermore, their separation into isomer ensembles upon AEMTS-blocking and fractionation (see Sect. 2.3.1) facilitates characterization of their disulfide bond connectivity, mass , and kinetic study (order and rate of appearance/disappearance) (Rothwarf and Scheraga 1993a; Xu et al. 1996; Volles et al. 1999; Xu et al. 2005) (Fig. 2.2). A reduction-pulse is a convenient tool to convert unstructured intermediates to the fully reduced protein (R) by the application of a small amount of reducing
28
M. Narayan equilibrium constants rate-determining step pathway rate-constants
N
intermediates DTTred GdnSCN
reduction
data analysis
N R
DTTox at 25 OC, pH 8.0
regeneration
fractionation R
3S
N
ion-exchange R HPLC N
R AEMTS-block quenching
1S
2S
4S
3S
1S
2S
4S
Fig. 2.2 Typical regeneration schematic of a multidisulfide protein (e.g., RNase A). RNase A is first fully reduced and unfolded by application of strongly denaturing and reducing conditions. The fully reduced protein is then introduced into refolding conditions and aliquots are periodically withdrawn, blocked with AEMTS, and fractionated using cation-exchange chromatography. Kinetic analysis of the folding rate is performed by integrating the peak areas under the chromatogram as previously described (Rothwarf and Scheraga 1993b)
equivalents (1–5 mM DTT red) for a short period of time (1–2 min) (Rothwarf et al. 1998a; Narayan et al. 2000; Narayan et al. 2008). This simplifies the regeneration chromatogram and facilitates the collection of structured intermediates for further characterization (Welker et al. 1999). It has also been successfully applied to characterize and contrast the regeneration pathways of proteins within and across families and to expedite analysis of the regeneration rates of proteins (Welker et al. 2004; Wang and Narayan 2008).
2.3 The Folding Pathways of RNase A RNase A folds via multiple (parallel) pathways as shown in Fig. 2.3 (Rothwarf et al. 1998a, b). In order to understand the mechanism(s) by which the native fold is generated, the folding of this protein from its fully reduced state I can be divided into three stages: initial stage of oxidative folding; the rate-determining step in regeneration; and post-folding steps (Narayan et al. 2000). The folding pathway of most disulfide-proteins can also be divided into the above stages with minor differences
2 Fold Maturation in the ER
29
R↔1S ↔ 2S ↔ 3S ↔ 4S 3S*
Des [40-95] Des [65-72]
N Fig. 2.3 Regeneration scheme of RNase A resulting in the formation of the native protein (N) from its unstructured precursors (R, 1S, 2S, 3S, and 4S). The rate-determining step involves the formation of two structured 3S* species (des [40–95] and des [65–72]) from its unstructured 3S ensemble isomers. The two 3S* species then form N by the oxidation of the remaining cysteines
observed among proteins (Narayan et al. 2000). Various factors impact the folding of this protein (and by extension others) during each step of the folding process.
2.3.1 Initial Stages of Oxidative Folding The initial stages of the oxidative regeneration involve sequential oxidation of the cysteines of the fully reduced polypeptide (R) by an extrinsic oxidizing agent (Rothwarf and Scheraga 1993a, b, c, d). In this stage, the rates of formation of disulfide bonds from cysteines are governed by pH of the folding medium and choice of the oxidizing agent (redox reagent). Typical oxidative folding conditions include an appropriate choice of pH (between 7 and 9; this is important because the thiolate is the active form of the sulfhydryl capable of being involved in thiol– disulfide exchange reactions) (Rothwarf and Scheraga 1993a). The pKas of free thiols generally vary around 8.3. At a pH of 8.3, 50% of the cysteines would be in the active, deprotonated form and the formation of disulfide bonds would have a certain rate. However, if the pH is lowered by one unit, there would be a tenfold increase in protonation of thiolates to form thiols resulting in a decreased oxidative folding rate. Nevertheless, pH serves as a valuable handle to study intermediates in oxidative folding because control over the reaction rate can be achieved by tuning the pH of the folding buffer. Of the large number of available redox reagents (Singh et al. 1995; Gough et al. 2006), the Scheraga laboratory has resorted to the use of DTT ox after initially working with glutathione because it does not form mixed disulfides with protein thiols (Rothwarf and Scheraga 1992). DTT ox is also a relatively weak oxidizing agent and thus incapable of rapidly oxidizing R to a fully oxidized state. Again, the use of a relatively weak oxidizing agent, compared to GSSG, provides greater control over the folding rate. Under folding conditions (pH 8, 100 mM DTT ox, 1 mM EDTA), as regeneration progresses, a percentage of R becomes oxidized to 1S; a fraction of the latter species then becomes oxidized to 2S and so on, until R, 1S, 2S, 3S, and 4S are populated
30
M. Narayan
(Rothwarf and Scheraga 1993a, b, c, d). The ratio of these populations depends upon pH, choice of oxidizing agent, and sampling time (after initiation of regeneration). As discussed, a strong oxidizing agent would favor the populations of species with larger number of disulfide bonds (i.e., 4S and 3S over 2S and 1S). The reduction of DTT ox by the protein thiolates produces DTT red, a strong reducing agent. This can result in 1S → 4S becoming reduced. The tendency to become reduced is further promoted by the fact that in RNase A and many other proteins, R → nS is not structured and thus susceptible to reduction (Narayan et al. 2000). The potency of reducing equivalents generated during the course of the regeneration process therefore depends upon the nature of the redox system and also on the protein concentration and numbers of protein cysteines. High protein concentration (i.e., high concentrations of R) coupled with a large number of free cysteines can arrest regeneration from progressing to completion because of the high concentration of reducing equivalents generated. 2.3.1.1 Nature of the nS Species of RNase A The 1S → 4S species are ensemble mixtures of such species (Xu et al. 1996; Volles et al. 1999). In case of a four-disulfide-bond-containing protein such as RNase A, the 1S ensemble consists of 28 isomers that are in equilibrium with one another. Each isomer interconverts to another via intramolecular thiol–disulfide exchange reactions between existing disulfide bonds and free cysteines (Xu et al. 1996). Similarly, 210 species in the 2S ensemble, 420 in the 3S ensemble, and 104 isomers in the 4S ensemble are predicted to exist (Rothwarf and Scheraga 1993a). Nevertheless, the isomers are not present in equal concentrations and the factors affecting their distribution are discussed next. 2.3.1.2 “Structure” and Disulfide-Connectivity During the Initial Stages of Regeneration A variety of structural probes and application of enzyme activity assays have revealed that little or no structure exists in the R → 4S ensembles of RNase A (Saito et al. 2001; Narayan et al. 2003). Nevertheless, these intermediates are not “random coils” but instead can be described as “statistical coils” (Matheson and Scheraga 1978). Fully reduced RNase A was examined thoroughly by a variety of techniques and found to be “locally ordered.” Fluorescence labeling of select residues and distance measurements revealed a native-like trace of the backbone with select side-chain distances not very different from the native protein (Navon et al. 2001a, b). Nevertheless, when “native tendency” of R was measured, the value was 0.6% suggesting that cysteines destined to become paired in the native structure were not spatially aligned in R (Narayan et al. 2003). (Note: In the oxidative folding pathways of certain proteins, predominant occurrence of native disulfide bonds has been observed.) A valuable tool has been developed to assess the “native tendency” of the fully reduced forms of ER-processed proteins (Narayan et al. 2003). The tool permits
2 Fold Maturation in the ER
31
estimation of the impact of chain conformation, sans disulfides, on the tendency to form native disulfides without resorting to corrective thiol–disulfide reactions. The “native tendency factor” therefore is a report on the “native-like” orientation of the chain in its fully reduced form which suggests that the amino acid residues are involved in promoting native interactions in the initial stages of the folding process. Native tendency values can range from 0 to 100%. The acquisition of the first disulfide bond impacts the chain significantly. Chain entropy is lost upon loop (disulfide bond) formation (Flory 1953; Xu et al. 1996). An examination of the population distribution of one disulfide connectivities in the 1S ensemble of RNase A indicates that the distribution of disulfide bonds in the 1S ensemble are mostly entropically ordained (Xu et al. 1996), i.e., loops that result in the least loss of chain entropy are those that are favored to the greatest extent. Therefore, proximal cysteines tend to form disulfide bonds to a greater extent than cysteines that are further apart in primary sequence. A similar trend was observed upon mapping the 2S ensemble of RNase A (Volles et al. 1999). Nevertheless, in both populations, deviations were found from loop entropy correlations. These deviations were attributed to enthalpic interactions that might bring sequentially distant residues in close special proximity and reduce the loss in entropy that might be encountered by the chain in the absence of such interactions (Xu et al. 1996; Volles et al. 1999). Such enthalpic interactions can include hydrophobic interactions, electrostatic interactions, etc. which would likely manifest themselves in a polar environment prior to the formation of a disulfide bond. The loss in chain entropy upon formation of a disulfide bound would thus be reduced compared to conditions in which such enthalpic interactions were absent. Notably, the 65–72 disulfide bond of RNase A is favored well-above the entropically ordained value and fourfold over the entropically equi-probable 58–65 disulfide bond (Xu et al. 1996). Disulfide bond distribution of a synthetic 58–72 peptide fragment of RNase A had revealed a similar fourfold bias of the 65–72 disulfide over the 58–65 disulfide bond (Altmann and Scheraga 1990). These results suggested that the enthalpic interactions favoring the formation of the 65–72 disulfide bond over other reside within the 58–72 peptide segment of RNase A. Structural, computational, and mutational studies of this peptide and native RNase A suggested the presence of a type III b-turn encompassing Cys65Lys66-Asn67-Gly68 which reduced the entropy loss associated with the formation of 65–72 disulfide bond, thus favoring it over 58–65.
2.3.2 The Rate-Determining Steps in the Oxidative Folding of RNase A The unstructured ensembles of RNase A initially populate the folding landscape of the protein. Gradually, there is the appearance of enzymatic activity. Chromatographic separations, mass spectrometric studies, disulfide mapping, and structural studies reveal that the appearance of the native protein (N) is preceded by the appearance of two structured (native-like) three disulfide bond-containing isomers from their unstructured 3S isomers (Rothwarf et al. 1998a, b). These intermediates were
32
M. Narayan
Fig. 2.4 Schematic of the rate-determining step in the regeneration of RNase A. The 3S ensemble consists of 416 non-native disulfide bond-containing isomers and four native disulfide bond-containing isomers in equilibrium with each other. The chemical exchange between these species is shown where a 3Snonnatbnd species reshuffles to form an intermediate with three native disulfide bonds (3Snatbnd). The physical conformational folding of 3Snatbnd to 3S* results in the protection of native disulfide bonds within tertiary structure and prevention of further isomerization reactions; conformational folding is pivotal to regeneration because it separates a structured intermediate from the pool of unstructured intermediates. The overall 3S → 3S* step is rate determining because the formation of 3Snatbnd is stochastic and because the formation of 3S* in RNase A is proline-isomerization limited and hence slow
determined to be des [40–95] and des [65–72] (see Sect. 2.2.1) and referred to as 3S* species (Laity et al. 1997; Shimotakahara et al. 1997). Fitting of the regeneration data as well as isolated 3S → 3S* experiments demonstrated that the reshuffling and conformational folding of the 3S ensemble to form the two 3S* species was the ratedetermining step in RNase A regeneration (at pH 8, 100 mM DTT ox, 100 mM Tris– HCl, 1 mM EDTA) (Rothwarf et al. 1998b). Note that the rate-determining step is dependent upon the folding conditions. It is possible to conduct regeneration experiments under a highly oxidizing environment, which would rapidly oxidize R to the scrambled (dead-end) 4S ensemble. Under such conditions, the rate-determining step is likely to be the reduction of 4S back to 3S which is necessary before the 3S → 3S* reaction can take place. Similarly, for the purpose of argument, under a reducing environment, the rate-determining step would be the oxidation of R to 1S and so on. Under physiological conditions, the rate-determining step might also be influenced by the presence of cofactors and chaperones. Nevertheless, in vitro folding studies have permitted us to infer important facets of the regeneration process and the rate-determining step of the regeneration pathway. In RNase A and several other proteins, the rate-determining step involves a competition between chemical and physical reactions (Welker et al. 2001a, b; Wang and Narayan 2008; Pal et al. 2010b) (Fig. 2.4). The chemical reactions are thiol–disulfide
2 Fold Maturation in the ER
33
exchange reactions that are necessary to populate an intermediate with the native set, or native subset, of disulfide bonds. Such an intermediate can fold in a physical conformational folding reaction and protect its newly formed native disulfide bonds from attack by solvent thiolates (redox agents) or intramolecular thiolates. The formation of the native protein or native-like structured intermediate first requires the formation of a critical number of native disulfide bonds. In RNase A, this number is three (of the four possible disulfide bonds). However, of the 420 possible three disulfide bond-containing intermediates in an eight-cysteine protein such as RNase A, only four species possessing three native disulfide bonds exist. Thus, statistically, less than 1% (4/420) of all species in the 3S ensemble possesses native disulfide bonds. The process of forming the native-like 3S* intermediates in RNase A from its unstructured 3S ensemble isomers is depicted in Fig. 2.4 (Wang and Narayan 2008). The process of forming 3S* from its unstructured 3S precursors constitutes the ratedetermining step in RNase A folding (Rothwarf et al. 1998b). The four native-disulfide-bond-containing intermediates (3Snative) must first form from the pool of non-native-disulfide-bond-containing isomers through intramolecular thiol–disulfide exchange reactions (Fig. 2.4). The 3Snative intermediate is at the cross-roads of the rate-determining step. It has two possible fates. It can physically (conformationally) fold to form the structured (native-like) 3S* species. Or it can be attacked by the thiolates and become back-reshuffled and revert to the pool of 3Snonnative species (Saito et al. 2001). The forward reaction is fruitful to the regeneration process because it removes the molecule from the “unstructured pool” of intermediates that constitute the folding landscape. This is because the newly formed native disulfide (as those in the 3Snative intermediate) disulfides become buried within the stable tertiary structure of the native-like intermediate and are protected from back-reshuffling reactions or attack by extrinsic redox reagents. Furthermore, the forward reaction results in a structure that has the remainder of the cysteines aligned to form disulfide bonds (Gonzalez et al. 2010). 2.3.2.1 Factors Influencing the Rate-Determining Step in the Regeneration of RNase A There are several factors impacting the rates of the chemical and physical halves of the rate-determining step. At pH 8, in RNase A ~30% of the 3Snative species back reshuffles to form 3Snonnative (Welker et al. 2001a; Saito et al. 2001). This is because conformational folding in RNase A is slowed by proline isomerization (Houry et al. 1994; Iwaoka et al. 1999). In RNase A, isomerization of three of the four prolines contribute significantly to the conformational folding rate (Schmid et al. 1993; Schmid 2001). Therefore, RNase A conformational folding is multiphasic with discrete populations and lifetimes attributable to the dispositions of the prolines in the unfolded states of the molecule (Houry et al. 1994). If conformational folding is not initiated from the equilibrium unfolded state, then the conformational folding phases will have a different set of lifetimes. For example, if native RNase A (or 3S*) is
34
M. Narayan
introduced into unfolding conditions for a very brief period (<1 s) before reintroduction into folding conditions, then conformational folding occurs on the millisecond timescale (Saito, et al. 2001; Gonzalez et al. 2010). This is because the prolines do not have sufficient dwell-time under unfolded conditions to isomerize to the equilibrium unfolded distribution. Under such fast conformational folding scenarios, it is likely that reshuffling reactions cannot compete with the conformational folding reaction (Gonzalez et al. 2010). Conversely, if 3Snative is allowed to refold at pH 7, a higher fraction of 3S* species is observed to be formed (Saito et al. 2001; Wang and Narayan 2008; Pal et al. 2010a). This is not because the conformational folding rate increases as a function of lowered pH but because of the impact that pH has on the reshuffling reaction. Reduction of pH by one unit results in the protonation of cysteines (protonation of 3Snative species to thiols) and loss in ability to attack disulfide bonds. As a result, at pH 7, the thiol–disulfide isomerization (3Snative to 3Snon-native step) is slowed relative to the rate at pH 8. As a result, the physical conformational folding of 3Snative is able to outcompete the chemical thiol–disulfide exchange reactions and the yield of 3S* species increases. If 3Snative is introduced into folding conditions at pH 9, the yield of 3S* decreases from 30% to about 8%. Again, this is because pH increase from 8 to 9 increases the reshuffling rate which can easily outcompete proline- isomerization-limited conformational folding of 3Snative to 3S*. Other factors that can impact the rate-determining step in oxidative folding reactions include temperature, chaotropes, kosmotropes, and catalysts (Low et al. 2000, 2002). In a protein whose conformational folding rate is proline isomerization limited, the presence of a catalyst such as peptidylprolyl isomerase is likely to increase the conformational folding rate and tilt the competition between the physical and chemical processes to favor folding (Wang and Narayan 2008). Our laboratory has recently investigated the impact of the chief endoplasmic reticulum oxidoreductase, protein disulfide isomerase (PDI) on the rate-determining step of RNase A, and on the conformational folding of select disulfide-bond containing proteins (Wang and Narayan 2008; Pal et al. 2010b). Our results indicate that though PDI accelerates 11-fold the formation of 3S* from its unstructured isomers (3Snonnative), its isomerase activity limits its potential in catalyzing the ratedetermining step in oxidative protein folding (Fig. 2.5). This is because the presence of PDI in the 3Snative to 3S* step inhibits the folding reaction by facilitating the backreshuffling of 3Snative to 3S* at a rate that outcompetes the conformational folding rate (Wang and Narayan 2008; Pal et al. 2010b). An eightfold decrease in the formation of 3S* from 3Snative is observed in the presence of PDI compared to control (Pal et al. 2010b). These results suggest that maximum potential catalytic impact of PDI in the rate-determining step of RNase A folding could reach 11 × 8 = 88-fold over control levels except for the self-limiting behavior of the oxidoreductase at the structure forming step (physical half) of the oxidative folding reaction. Similarly, the conformational folding of other proteins such as a-lactalbumin and a three-disulfide mutant of RNase A were investigated in the presence of PDI (Pal et al. 2010b). In a-lactalbumin, the conformational folding is Ca2+-dependent (Veprintsev et al. 1997) and is in competition with external reducing agents (such
2 Fold Maturation in the ER
35 Chemical Reshuffling rxn
Physical Conf. Folding rxn
↔
3SNonnatbnd
3Snatbnd
3S*nativelikestruct.
~10.5x w/PDI ~7x w/PDI Fig. 2.5 The presence of PDI catalyzes the rate-determining step (3S ® 3S*) ~11-fold over uncatalyzed rates. However, it retards the conformational folding component of the rate-determining step by a factor of 7. This is because its oxidoreductase activity back reshuffles 3Snatbnd (to its 3Snonnatbnd isomers) before it can conformationally fold to 3S*
as GSH). In the three-disulfide mutant of RNase A conformational folding of 3Snative to 3S* (N) state competes with extrinsic thiols. The inclusion of PDI in the fold maturation step of both proteins results in a PDI-dependent diminution of the rate of native protein formation (Pal et al. 2010b). These observations suggest that oxidative regeneration in the ER appears to be orchestrated by a complex number of features some of which may act in concert with each other. In vitro studies performed to-date provide an initial inroad into advancing our understanding of this complex phenomenon.
2.3.3 Postfolding Steps in the Regeneration of RNase A In case of RNase A and many other proteins in which a structured (native-like) intermediate is formed prior to the formation of the native molecule, a final step (or series of steps) involving the oxidation of any remaining thiols is required to complete the regeneration process (Wedemeyer et al. 2000; Narayan et al. 2000). These postrate-determining steps are generally rapid and well directed. This is because the formation of native-like structure essentially orients remaining thiols in correct pairs, thereby averting any non-native disulfide bond formation. In RNase A, the final step in the regeneration process involves the oxidation of des [40–95] and des [65–72] to N (Wedemeyer et al. 2000). The formation of the fourth and final disulfide bond stabilizes the protein (increases Tm by ~15°C) and restores full enzymatic activity (Shimotakahara et al. 1997; Laity et al. 1997).
36
M. Narayan
In other proteins which may require cofactors for structure stabilization and biological activity, the binding of the cofactor may constitute the final stage in oxidative folding (Narayan et al. 2000). Considering that poststructure forming regeneration steps are rapid, it becomes transparent that proteins that are able to acquire native-like structure early during the regeneration process would be able to avert non-native disulfide bond formation and progress rapidly toward the native fold. An example of such a protein is onconase which is a homolog of RNase A and possesses four disulfide bonds (Leland and Raines 2001; Ardelt et al. 2008). A comparison of the regeneration rates of both proteins reveals that the rate of formation of native onconase is approximately 2-fold faster than RNase A at 25°C (Welker et al. 2004, 2007). Inspection of the regeneration pathway of onconase reveals the formation of a stable two disulfide bond-containing intermediate (2S*). Furthermore, the three disulfide ensemble comprises of two major structured intermediates and near lack of non-native (unstructured) isomers (Gahl et al. 2008; Gahl and Scheraga 2009). The formation of the stable two disulfide intermediate likely results in the orientation of the remaining cysteines such that they form native disulfide bonds. It thereby serves to limit the formation of non-native 3S isomers and enhances the folding rate of the protein. Note: Considering that there is an inherent competition between aggregation and folding, it is in the best interest of the organism to select those proteins that fold rapidly and avoid becoming aggregated.
2.3.4 Other Pathways (Minor Pathways, Parallel Pathways, and Off Pathway Intermediates in RNase A Regeneration) The above described pathway (scheme 1) represents the major mechanisms by which RNase A regenerates. However, the protein is capable of regenerating through other mechanisms which represent minor pathways when the major mechanisms are in effect (Iwaoka et al. 1998; Xu and Scheraga 1998).
2.3.4.1 Minor Regeneration Pathways C65AC72A and C40AC95A are two three disulfide bond-containing mutant analogs of des [65–72] and des [40–95] RNase A which have been extensively characterized in the Scheraga laboratory (Shimotakahara et al. 1997; Laity et al. 1997). Both proteins possess native-like structure and activities that mimic their wild-type analogs. Regeneration studies of both mutants reveal that they are able to acquire native structure via a “minor” pathway involving the oxidation-coupled conformational folding step of a two native-disulfide-bond-containing intermediate to the native protein (2Snative → 3S*) (Xu and Scheraga 1998; Iwaoka et al. 1998). These minor pathways constitute ~5% each of the flux through all possible pathways that the protein can regenerate. However, as far as each mutant is concerned, the flux would be 100% through the “minor” pathway because there is no other pathway possible.
2 Fold Maturation in the ER
37
2.3.4.2 Off-Pathway Intermediates (Kinetically Trapped Intermediates) When the regeneration of RNase A is carried out at 15°C, two new intermediates are found to appear (Welker et al. 1999). Importantly, both intermediates remain populated within the regeneration after all other intermediates have been converted to the native protein, i.e., they are kinetically trapped (Welker et al. 1999, 2001a; Narayan et al. 2001). Furthermore, while they survive mildly reducing conditions (i.e., an applied reduction-pulse) at 15°C they are not stably populated when the regeneration is performed at 25°C or higher. These observations suggest that they are structured and possess a thermal transition between 15 and 25°C. Mapping of the disulfide connectivities in both intermediates revealed their identity as des [58–110] and des [26–84]. It also helped illuminate why both species are kinetically trapped. Both pairs of cysteines in the two intermediates are buried within the tertiary structure of the proteins (Wlodawer et al. 1982). Thus, oxidizing them into disulfide bonds is difficult and the species become kinetically trapped. Local unfolding is required to expose the buried cysteines for oxidation to disulfide bond. However, given their location, the magnitude of the unfolding event is equal to a “global” unfolding reaction. Such an unfolding response also results in the exposure of other native disulfides in each species, making them susceptible to intramolecular thiol–disulfide attack (by the formerly buried cysteines) (Welker et al. 2001a; Narayan et al. 2001). This intramolecular reshuffling reaction is in competition with the oxidation of the cysteines by the external oxidizing agent (DTTox). However, the intramolecular attack on the existing disulfides by the cysteines is generally more effective than the oxidation of the cysteines to disulfides. This is because of the higher “effective” concentration of those cysteines relative to the existing disulfides compared to the concentration and nature of the external oxidizing agent. The unfolding events depend upon the temperature of the environment. Even at 15°C, fluctuations in des [58–110] and des [26–84] eventually result in their disappearance from the regeneration landscape via reshuffling to form unstructured isomers (3S). These 3S isomers then revert to the thermodynamically favored des species. However, the formation of the new des species is not stoichiometric because the 3S ensemble gets distributed among all four des species, two of which become N. Through such iterations the kinetically trapped intermediates are finally consumed (Welker et al. 2001a; Narayan et al. 2001). While it may be argued that in the kinetically trapped intermediates in RNase A were generated using nonphysiological conditions, it is important to note that these studies promote our understanding of when a species becomes kinetically trapped. Such kinetically trapped intermediates have been previously observed in BPTI, a-lactalbumin, human epidermal growth factor, etc. (Weissman and Kim 1995; Chang et al. 2001; Salamanca and Chang 2005). The presence of multiple (parallel) pathways (mechanisms) by which a protein can fold is extremely important for cell viability (Narayan et al. 2000). The milieu and certain stress conditions within the cell may reduce the flux through a particular
38
M. Narayan
pathway or in extreme instances may completely shut it down. This is because intermediates populating that particular pathway may be susceptible to stress or environmental changes and become destabilized, resulting in attrition of the regenerative flux through that oxidative folding arm. In such a scenario, the presence of alternate pathway with separate intermediates having physicochemical properties resistant to the emergent stress conditions would result in enhanced relative regeneration flux through that pathway. It is recognizable that the overall regeneration rate under conditions of insult would be reduced, relative to homeostasis conditions. Neverthe less, the presence of multiple pathways would ensure that the protein can regenerate through an arm rather than through an all-or-none phenomenon.
2.4 Conclusions The work done by the Scheraga laboratory on RNase A, in conjunction with work done on other model proteins in other laboratories referenced above, has advanced our understanding of how disulfide bond-containing proteins mature within the oxidative environs of the ER. A general picture has emerged from the above biophysical studies which suggests that the nascent polypeptide is subjected to two opposing forces before exiting the ER; an inherent, evolutionarily tailored, tendency to fold which is in fierce competition with a tendency to aggregate. Several factors with the polypeptide contribute to the latter phenomena and include the placement of cysteines along the primary sequence vis-a-vis native disulfide bonds, presence of prolines, the inability to distinguish between native and non-native disulfide bonds, among others. The body of work contributed to date is the stepping stone for furthering our understanding of traffic within the ER and factors contributing to misfolding under conditions of stress and homeostasis. By applying the principles of fold formation to these scenarios, we may be able to understand the biochemical and biophysical foundation of misfolding-related disorders.
References Altmann KH, Scheraga HA (1990) Local-structure in ribonuclease-A – effect of amino-acid substitutions on the preferential formation of the native disulfide loop in synthetic peptides corresponding to residues cys-58 cys-72 of bovine pancreatic ribonuclease-A. J Am Chem Soc 112:4926–4931 Ardelt W, Shogen K, Darzynkiewicz Z (2008) Onconase and amphinase, the antitumor ribonucleases from Rana pipiens oocytes. Curr Pharm Biotechnol 9:215–225 Arolas JL, Bronsoms S, Lorenzo J, Aviles FX, Chang JY, Ventura S (2004) Role of kinetic intermediates in the folding of carboxypeptidase inhibitor. J Biol Chem 279:37261–37270 Arolas JL, Aviles FX, Chang JY, Ventura S (2006) Folding of small disulfide-rich proteins: clarifying the puzzle. Trends Biochem Sci 31:292–301 Bientema JJ (1998) Introduction: the ribonuclease A superfamily. Cell Mol Life Sci 54:763–765
2 Fold Maturation in the ER
39
Bruice TW, Kenyon GL (1982) Novel alkyl alkanethiolsulfonate sulfhydryl reagents. Modification of derivatives of L-cysteine. J Protein Chem 1:47–58 Chang JY (2004) Evidence for the underlying cause of diversity of the disulfide folding pathway. Biochemistry 43:4522–4529 Chang JY (2008) Diversity of folding pathways and folding models of disulfide proteins. Antioxid Redox Signal 10:171–177 Chang JY, Li L (2005) Divergent folding pathways of two homologous proteins, BPTI and tick anticoagulant peptide: compartmentalization of folding intermediates and identification of kinetic traps. Arch Biochem Biophys 437:85–95 Chang JY, Li L, Lai PH (2001) A major kinetic trap for the oxidative folding of human epidermal growth factor. J Biol Chem 276:4845–4852 Chang JY, Lu BY, Lai PH (2006) Oxidative folding of hirudin in human serum. Biochem J 394: 249–2457 Crook EM, Mathias AP, Rabin BR (1960) Spectrophotometric assay of bovine pancreatic ribonuclease by the use of cytidine 2¢:3¢-phosphate. Biochem J 4:234–238 Fewell SW, Travers KJ, Weissman JS et al (2001) The action of molecular chaperones in the early secretory pathway. Annu Rev Genet 35:149–191 Flory PJ (1953) Principles of polymer chemistry. Cornell University Press, Ithaca, NY Gahl RF, Scheraga HA (2009) Oxidative folding pathway of onconase, a ribonuclease homologue: insight into oxidative folding mechanisms from a study of two homologues. Biochemistry 48:2740–2751 Gahl RF, Narayan M, Xu G et al (2008) Dissimilarity in the oxidative folding of onconase and ribonuclease A, two structural homologues. Protein Eng Des Sel 21:223–231 Gilbert HF (1998) Protein disulfide isomerase. Methods Enzymol 290:26–50 Gleiter S, Bardwell JC (2008) Disulfide bond isomerization in prokaryotes. Biochim Biophys Acta 1783(4):530–534. Epub 2008 Feb 23. Review Gonzalez V, Pal R, Narayan M (2010) The oxidoreductase behavior of protein disulfide isomerase impedes fold maturation of endoplasmic reticulum-processed proteins in the pivotal structure-coupled step of oxidative folding: implications for subcellular protein trafficking. Biochemistry 49(29):6282–6289 Gough JD, Barrett EJ, Silva Y et al (2006) Ortho- and meta-substituted aromatic thiols are efficient redox buffers that increase the folding rate of a disulfide-containing protein. J Biotechnol 125:39–47 Hawkins HC, Freedman RB (1991) The reactivities and ionization properties of the active-site dithiol groups of mammalian protein disulphide-isomerase. Biochem J 275:335–339 Hiniker A, Bardwell JC (2003) Disulfide bond isomerization in prokaryotes. Biochemistry 42(5):1179–1185. Review Houry WA, Rothwarf DM, Scheraga HA (1994) A very fast phase in the refolding of disulfideintact ribonuclease A: implications for the refolding and unfolding pathways. Biochemistry 33:2516–2530 Hwang C, Sinskey AJ, Lodish HF (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257:1496–1502 Iwaoka M, Juminaga D, Scheraga HA (1998) Regeneration of three-disulfide mutants of bovine pancreatic ribonuclease A missing the 65–72 disulfide bond: characterization of a minor folding pathway of ribonuclease A and kinetic roles of Cys65 and Cys72. Biochemistry 37:4490–4501 Iwaoka M, Wedemeyer WJ, Scheraga HA (1999) Conformational unfolding studies of three-disulfide mutants of bovine pancreatic ribonuclease A and the coupling of proline isomerization to disulfide redox reactions. Biochemistry 38:2805–2815 Konishi Y, Ooi T, Scheraga HA (1982) Regeneration of ribonuclease A from the reduced protein. Rate-limiting steps. Biochemistry 21:4734–4740 Laity JH, Lester CC, Shimotakahara S et al (1997) Structural characterization of an analog of the major rate-determining disulfide folding intermediate of bovine pancreatic ribonuclease A. Biochemistry 36:12683–12699
40
M. Narayan
Leland PA, Raines RT (2001) Cancer chemotherapy-ribonucleases to the rescue. Chem Biol 8:405–413 Lin CC, Chang JY (2007) Pathway of oxidative folding of bovine alpha-interferon: predominance of native disulfide-bonded folding intermediates. Biochemistry 46:3925–3932 Low LK, Shin HC, Scheraga HA (2000) Acceleration of oxidative folding of bovine pancreatic ribonuclease A by anion-induced stabilization and formation of structured native-like intermediates. FEBS Lett 472:67–72 Low LK, Shin HC, Scheraga HA (2002) Oxidative folding of bovine pancreatic ribonuclease A: insight into the overall catalysis of the refolding pathway by phosphate. J Protein Chem 21:19–27 Mamathambika BS, Bardwell JC (2008) Disulfide-linked protein folding pathways. Annu Rev Cell Dev Biol 24:211–235 Matheson R Jr, Scheraga HA (1978) A method for predicting nucleation sites for protein folding based on hydrophobic contacts. Macromolecules 11:819–829 Narayan M, Welker E, Wedemeyer WJ et al (2000) Oxidative folding of proteins. Acc Chem Res 33:737–820 Narayan M, Welker E, Scheraga HA (2001) Development of a novel method to study the ratedetermining step during protein regeneration: application to the oxidative folding of RNase A at low temperature reveals BPTI-like kinetic traps. J Am Chem Soc 123:2909–2910 Narayan M, Welker E, Scheraga HA (2003) Native conformational tendencies in unfolded polypeptides: development of a novel method to assess native conformational tendencies in the reduced forms of multiple disulfide-bonded proteins. J Am Chem Soc 123:2909–2910 Narayan M, Welker E, Zhai H et al (2008) Detecting native folds in mixtures of proteins that contain disulfide bonds. Nat Biotechnol 26:427–429 Navon A, Ittah V, Gussakovsky EE et al (2001a) Local and long-range interactions in the thermal unfolding transition of bovine pancreatic ribonuclease A. Biochemistry 40:93–104 Navon A, Ittah V, Landsman P et al (2001b) Distributions of intramolecular distances in the reduced and denatured states of bovine pancreatic ribonuclease A. Folding initiation structures in the C-terminal portions of the reduced protein. Biochemistry 40:105–118 Pal R, Cristan EA, Schnittker K et al (2010a) Rescue of ER oxidoreductase function through polyphenolic phytochemical intervention: implications for subcellular traffic and neurodegenerative disorders. Biochem Biophys Res Commun 392:567–571 Pal R, Gonzalez V, Narayan M (2010b) Reshuffling activity of protein disulfide isomerase reduces refolding yield in the structure-forming step of the oxidative protein folding reaction. Chem Lett 39:263–265 Romisch K (2004) A cure for traffic jams: small molecule chaperones in the endoplasmic reticulum. Traffic 5:815–820 Romisch K (2005) Endoplasmic reticulum-associated degradation. Annu Rev Cell Dev Biol 21:435–456 Rothwarf DM, Scheraga HA (1992) Equilibrium and kinetic constants for the thiol–disulfide interchange reaction between glutathione and dithiothreitol. Proc Natl Acad Sci USA 89:7944–7948 Rothwarf DM, Scheraga HA (1993a) Regeneration of bovine pancreatic ribonuclease A. 1. Steady-state distribution. Biochemistry 32:2671–2679 Rothwarf DM, Scheraga HA (1993b) Regeneration of bovine pancreatic ribonuclease A. 2. Kinetics of regeneration. Biochemistry 32:2680–2689 Rothwarf DM, Scheraga HA (1993c) Regeneration of bovine pancreatic ribonuclease A. 3. Dependence on the nature of the redox reagent. Biochemistry 32:2690–2697, Erratum in: Biochemistry 32:7064 Rothwarf DM, Scheraga HA (1993d) Regeneration of bovine pancreatic ribonuclease A. 4. Temperature dependence of the regeneration rate. Biochemistry 32:2698–2703 Rothwarf DM, Li YJ, Scheraga HA (1998a) Regeneration of bovine pancreatic ribonuclease A: detailed kinetic analysis of two independent folding pathways. Biochemistry 37:3767–3776 Rothwarf DM, Li YJ, Scheraga HA (1998b) Regeneration of bovine pancreatic ribonuclease A: identification of two nativelike three-disulfide intermediates involved in separate pathways. Biochemistry 37:3760–3766
2 Fold Maturation in the ER
41
Saito K, Welker E, Scheraga HA (2001) Folding of a disulfide-bonded protein species with free thiol(s): competition between conformational folding and disulfide reshuffling in an intermediate of bovine pancreatic ribonuclease A. Biochemistry 40:15002–15008 Salamanca S, Chang JY (2005) Unfolding and refolding pathways of a major kinetic trap in the oxidative folding of alpha-lactalbumin. Biochemistry 44:744–750 Schmid FX (2001) Prolyl isomerases. Adv Protein Chem 59:243–282 Schmid FX, Mayr LM, Schönbrunner ER et al (1993) Prolyl isomerases: role in protein folding. Adv Protein Chem 44:25–66 Shimotakahara S, Rios CB, Laity JH et al (1997) NMR structural analysis of an analog of an intermediate formed in the rate-determining step of one pathway in the oxidative folding of bovine pancreatic ribonuclease A: automated analysis of 1H, 13C, and 15N resonance assignments for wild-type and [C65S, C72S] mutant forms. Biochemistry 36:6915–6929 Shin HC, Scheraga HA (2004) Catalysis of the oxidative folding of bovine pancreatic ribonuclease A by protein disulfide isomerase. J Mol Biol 300:995–1003 Singh R, Lamoureux GV, Lees WJ et al (1995) Reagents for rapid reduction of disulfide bonds. Methods Enzymol 251:167–173 Tian G, Xiang S, Noiva R et al (2006) The crystal structure of yeast protein disulfide isomerase suggests cooperativity between its active sites. Cell 124:61–73. Erratum in: (2006). Cell 124:1085–1088 Tu BP, Weissman JS (2004) Oxidative protein folding in eukaryotes: mechanisms and consequences. J Cell Biol 164:341–346 Veprintsev DB, Permyakov SE, Permyakov EA et al (1997) Cooperative thermal transitions of bovine and human apo-alpha-lactalbumins: evidence for a new intermediate state. FEBS Lett 412:625–628 Volles MJ, Xu X, Scheraga HA (1999) Distribution of disulfide bonds in the two-disulfide intermediates in the regeneration of bovine pancreatic ribonuclease A: further insights into the folding process. Biochemistry 38:7284–7293 Wang YH, Narayan M (2008) pH dependence of the isomerase activity of protein disulfide isomerase: insights into its functional relevance. Protein J 27:181–185 Wedemeyer WJ, Welker E, Narayan M et al (2000) Disulfide bonds and protein folding. Biochemistry 39:4207–4216 Wedemeyer WJ, Welker E, Scheraga HA (2002) Proline cis-trans isomerization and protein folding. Biochemistry 41:14637–14644 Weissman JS, Kim PS (1993) Efficient catalysis of disulphide bond rearrangements by protein disulphide isomerase. Nature 365:185–188 Weissman JS, Kim PS (1995) A kinetic explanation for the rearrangement pathway of BPTI folding. Nat Struct Biol 2:1123–1130 Welker E, Narayan M, Volles MJ et al (1999) Two new structured intermediates in the oxidative folding of RNase A. FEBS Lett 460:477–479 Welker E, Narayan M, Wedemeyer WJ et al (2001a) Structural determinants of oxidative folding in proteins. Proc Natl Acad Sci USA 98:2312–2316 Welker E, Wedemeyer WJ, Scheraga HA et al (2001b) Coupling of conformational folding and disulfide-bond reactions in oxidative folding of proteins. Biochemistry 40:9059–9064 Welker E, Hathaway L, Scheraga HA (2004) A new method for rapid characterization of the folding pathways of multidisulfide-containing proteins. J Am Chem Soc 126:3720–3721 Welker E, Hathaway L, Xu G et al (2007) Oxidative folding and N-terminal cyclization of onconase. Biochemistry 46:5485–5493 Wilkinson B, Gilbert HF (2004) Protein disulfide isomerase. Biochim Biophys Acta 1699:35–44 Wlodawer A, Bott R, Sjölin L (1982) The refined crystal structure of ribonuclease A at 2.0 A resolution. J Biol Chem 257(3):1325–1332 Woycechowsky KJ, Raines RT (2000) Native disulfide bond formation in proteins. Curr Opin Chem Biol 4:533–539 Xiao R, Wilkinson B, Solovyov A et al (2004) The contributions of protein disulfide isomerase and its homologues to oxidative protein folding in the yeast endoplasmic reticulum. J Biol Chem 279:49780–49786
42
M. Narayan
Xu X, Scheraga HA (1998) Kinetic folding pathway of a three-disulfide mutant of bovine pancreatic ribonuclease A missing the [40–95] disulfide bond. Biochemistry 37:7561–7571 Xu X, Rothwarf DM, Scheraga HA (1996) Nonrandom distribution of the one-disulfide intermediates in the regeneration of ribonuclease A. Biochemistry 35:6406–6417 Xu G, Narayan M, Scheraga HA (2005) The oxidative folding rate of bovine pancreatic ribonuclease is enhanced by a covalently attached oligosaccharide. Biochemistry 44:9817–9823
Chapter 3
Cystine Knot Folding in Cyclotides Norelle L. Daly, Christian W. Gruber, Ulf Göransson, and David J. Craik
Abstract Cyclotides are naturally occurring plant-based proteins of approximately 30 amino acids in size that contain a head-to-tail cyclized backbone and a cystine knot motif formed by their three conserved disulfide bonds. Their exceptional stability and unique topology make them valuable frameworks in drug design or protein engineering applications. To facilitate such applications and to explore structure–activity relationships of cyclotides it is useful to be able to chemically synthesize them, a process that is readily achieved via solid phase peptide synthesis followed by oxidative folding. This chapter describes what is known about the oxidative folding of cyclotides, both in chemical folding buffers and assisted by a protein disulfide isomerase enzyme isolated from a cyclotide-producing plant. Formation of the cystine knot motif is readily achieved, despite its apparent topological complexity. Keywords Cyclic peptide • Cycloviolacin O2 • Cystine knot • Folding • Kalata B1 • Plant proteins • Protein disulfide isomerase • Structure
3.1 Introduction 3.1.1 Background and Structure Cyclotides are a family of plant-based proteins approximately 30 amino acids in size, containing six conserved cysteine residues that form a knotted structural motif (Craik et al. 1999). Their main characteristic feature is a head-to-tail cyclized peptide backbone, i.e., these proteins have no N or C terminus; instead they have a
D.J. Craik (*) Institute for Molecular Bioscience, Division of Chemistry and Structural Biology, The University of Queensland, QLD 4072, Brisbane, Australia e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_3, © Springer Science+Business Media, LLC 2011
43
44
N.L. Daly et al.
Fig. 3.1 Three-dimensional structure of kalata B1 (PDB ID code 1NB1). The disulfide bonds are shown in ball-and-stick format and numbered with Roman numerals. The backbone loops are numbered 1–6 and the b-strands are shown as arrows. The diagram was made with MOLMOL (Koradi et al. 1996)
continuous cycle of peptide bonds (Trabi and Craik 2002; Craik 2006). This circular backbone and cystine knot engenders cyclotides with exceptional stability (Colgrave and Craik 2004). The cystine knot motif of cyclotides (Craik 2001) is formed by the threading of a disulfide bond (CysIII–CysVI) through a ring created by two other disulfide bonds (CysI–CysIIV, and CysII–CysV) and their connecting backbone segments as illustrated in Fig. 3.1. Cystine knot motifs are present in a wide variety of proteins, ranging from toxins to growth factors, and occur in a wide range of organisms, including insects, plants, and animals (Isaacs 1995; Craik et al. 2001). However, only in the cyclotides the cystine knot is combined with a head-to-tail cyclized backbone, resulting in the cyclic cystine knot (CCK) motif. As in other cystine knots, a b-sheet structure is intimately associated with the cystine knot in cyclotides. As might be expected, the CCK motif makes the cyclotides exceptionally stable (Colgrave and Craik 2004). The internal disulfide ring of the cystine knot of cyclotides comprises eight amino acids, resulting in a very tight threaded structure. In this chapter, we describe studies that have investigated how cyclotides fold to this topologically intriguing structure. First, we provide a background on the sequences and structures of cyclotides, their discovery, and their distribution in the plant kingdom and their applications. Individual plants produce suites of up to 100 cyclotides, and it has been estimated that the cyclotide family might comprise up to 50,000 members (Gruber et al. 2008). Cyclotides are expressed in virtually all tissues of the plants in which they occur,
3 Cyclotide Folding
45
Table 3.1 Sequences of selected cyclotides Peptide Sequence Kalata B1 -GLPVCGET---CVGGT-CNTPG---CTCSWPV-CTRNKalata B2 -GLPVCGET---CFGGT-CNTPG---CSCTWPI-CTRDKalata B6 -GLPTCGET---CFGGT-CNTPG---CSCSSWPICTRNKalata B7 -GLPVCGET---CTLGT-CYTQG---CTCSWPI-CKRNCycloviolacin O1 -GIP-CAES---CVYIP-CTVTALLGCSCSNRV-CYNCycloviolacin O2 -GIP-CGES---CVWIP-CISSAIG-CSCKSKV-CYRNMCoTI-I -GGV-CPKILQRCRRDSDCPGA----CICRGNGYCGSGSDMCoTI-II -GGV-CPKILKKCRRDSDCPGA----CICRGNGYCGSGSDCysteine residues are shown in bold. The sequences shown are for a selection of cyclotides mentioned in the text. A full list of sequences is available on CyBase (http://www/cybase.org.au) (Wang et al. 2008)
including leaves, stems, flowers, and roots (Trabi et al. 2004). Table 3.1 shows some representative cyclotide sequences, including the prototypic cyclotide, kalata B1, which was the first to be structurally characterized (Sletten and Gran 1973; Saether et al. 1995). As is clear from the table, most cyclotides have a similar spacing of their six conserved cysteine residues, but there is variability in the amino acid content of the backbone loops between successive cysteine residues in the sequence. The sequences of all known cyclotides are contained in CyBase (http://www.cybase. org.au) (Mulvenna et al. 2006; Wang et al. 2008). The CCK motif dictates that cyclotides are rigid and compact proteins, but their six loops contain the amino acids that are exposed on the molecular surface and hence responsible for their biomolecular interactions and biological activities. Most cyclotides contain a number of hydrophobic residues that tend to cluster in a patch on the surface, giving cyclotides the potential to bind to membranes (Kamimori et al. 2005; Svangård et al. 2007; Huang et al. 2009). This surface-exposed hydrophobic patch also tends to result in cyclotides being rather late-eluting on reverse phase (RP)-HPLC and has implications for the conditions used to oxidatively fold cyclotides, as is further described in Sect. 3.2. Cyclotides fall into two major subfamilies: Möbius and bracelet, based on the presence or absence, respectively, of a cis peptide bond preceding a proline residue in loop 5 of the sequence (Craik et al. 1999). Most peptide bonds in proteins have a trans (w = 180°) arrangement, but the bond proceeding proline residues, particularly when a bulky amino acid is nearby, is often in a cis (w = 0°) configuration. Conceptually, this places a 180° twist in the otherwise trans backbone, leading us to introduce the term Möbius (a Möbius strip is a topological entity formed by joining the ends of a ribbon after making a 180° twist in it) to refer to this subclass of cyclotides. In the case of kalata B1, a Trp–Pro bond has the cis linkage (Rosengren et al. 2003). A third subclass of cyclotides is referred to as the trypsin inhibitor cyclotides and presently comprises only two members, namely MCoTI-I and MCoTI-II (Hernandez et al. 2000). The names reflect their discovery in the vine plant Momordica cochinchinensis and the fact that they are trypsin inhibitors, which are present in the seeds of these plants. From the sequence shown in Table 3.1, it is
46
N.L. Daly et al.
clear that MCoTI-II, the more extensively studied of the two peptides, is significantly different compared to other cyclotides, but we include it in the cyclotide family because it contains a CCK motif (Felizmenio-Quimio et al. 2001; Heitz et al. 2001) and is a plant-derived head-to-tail cyclized peptide. Since MCoTI-I and MCoTI-II have high homology to a class of proteins that are part of the knottin family of proteins, they are also referred to as cyclic knottins (Chiche et al. 2004).
3.1.2 Distribution Cyclotides have so far been found mainly in plants from the Rubiaceae and Violaceae families, with the two trypsin inhibitor cyclotides found in the Cucurbitaceae family (Craik et al. 2004). The Rubiaceae is one of the largest plant families known, and comprises more than 13,000 species, including coffee, and is thus also known as the coffee family. In comparison, the Violaceae, the violet family, is a much smaller family, comprising only about 900 species. Interestingly, every member of the Violaceae examined so far has been found to contain cyclotides (Göransson et al. 1999; Burman et al. 2010); whereas only about 5–10% of Rubiaceae plants examined contain cyclotides (Gruber et al. 2008). A single Cucurbit plant, M. cochinchinensis, has been reported to contain cyclotides (Hernandez et al. 2000; Felizmenio-Quimio et al. 2001). The plant in which cyclotides were first discovered is a member of the Rubiaceae family, Oldenlandia affinis, a common weed that grows in the tropical region of Africa. The uterotonic properties of an extract of this plant, used as a medicinal tea by women in the Democratic Republic of Congo was first reported by Gran (1970), who characterized the amino acid composition of kalata B1 in the early 1970s (Gran 1973; Sletten and Gran 1973). The name of the peptide reflects the name of the medicinal tea used for the acceleration of childbirth, “kalata–kalata.” The CCK motif of the peptide was not elucidated until 1995 (Saether et al. 1995), and the structure that was determined using NMR spectroscopy is shown in Fig. 3.1.
3.1.3 Biosynthesis Unlike many earlier known cyclic peptides, which are typically smaller than 12 amino acids in size and are nonribosomally synthesized, cyclotides are ribosomally synthesized and are excised from precursor proteins by a process that is only now beginning to be understood. Figure 3.2 shows a schematic representation of the precursor protein for kalata B1 along with the precursors for a few other cyclotides, including kalata B2, B3, and B6. In each case, the precursor protein contains an endoplasmic reticulum (ER) signal sequence, a pro-domain, a region corresponding to the mature cyclotide domain, and a small C-terminal tail of generally hydro phobic residues (Jennings et al. 2001). Some precursors contain multiple copies
3 Cyclotide Folding
47
Fig. 3.2 Schematic representation of the precursor proteins of cyclotides. Representations of the precursor proteins for kalata B1, kalata B2, and kalata B6 and B7 are given. These precursors have been isolated from Oldenlandia affinis (Jennings et al. 2001). The proteins have an endoplasmic reticulum signal sequence (ER), a pro-region (PRO), a conserved repeated fragment (NTR) and either one or multiple copies of the mature peptides. Each precursor has a single C-terminal hydrophobic tail. The cleavage sites for excision occur in loop 6 of the mature peptide and are indicated with arrows on the structure
(one, two, or three) of the mature domain and in all such cases, a small section of the pro-domain just upstream of the mature peptide region is also repeated and is referred to as the N-terminal repeat (NTR). Although the processing events that result in excision of the mature peptide sequence from the precursor protein, followed by cyclization, are not fully understood, asparaginyl endopeptidase (AEP) enzymes have been implicated in the processing (Saska et al. 2007; Gillon et al. 2008; Saska and Craik 2008). This hypothesis has been suggested based on the ubiquitous presence of an asparagine or an aspartic acid residue at the C-terminal end of the mature peptide domain within the linear precursor sequence. Although AEPs are implicated in the cyclization process, much less is known about the folding process that results in cyclotides forming a cystine knot within a cyclic peptide backbone. It is assumed that the oxidative folding occurs before cyclization, and may be facilitated by biological chaperones (Jennings et al. 2001; Gruber et al. 2007a). Indeed, Gruber et al. (2007b) recently reported the discovery of a protein disulfide isomerase (PDI) from O. affinis leaves (OaPDI) and showed that it accelerates folding in vitro (Gruber et al. 2007a), as described in more detail in Sect. 3.3 of this chapter. Recently, PDI analogues were also reported from a Violaceae species, indicating that similar cyclotide-oxidizing systems have been developed in other cyclotide-expressing plants (Burman et al. 2010).
3.1.4 Applications Cyclotides have a wide range of biological activities, and indeed it was these activities that were first responsible for their discovery, since some of these activities had potential pharmaceutical applications. The first pharmaceutically relevant discovery was related to the observation of the uterotonic effect of kalata B1 in the medicinal tea from O. affinis (Gran 1970). Subsequent studies of the bark from a
48
N.L. Daly et al.
Tanzanian tree conducted at the National Cancer Institute in the USA found anti-HIV activity for the cyclotides circulin A and circulin B (Gustafson et al. 1994), and a group of researchers at Merck Pharmaceuticals found an inhibitory effect of cyclopsychotride A on neurotensin binding to membrane preparations (Witherup et al. 1994). This peptide was extracted from a South American tree. In the Violaceae plant family, cyclotides were first discovered because of their hemolytic properties (Schöpke et al. 1993): the hemolytic and cytotoxic effects of cyclotides have now been the focus of extensive studies (Lindholm et al. 2002; Barry et al. 2003; Herrmann et al. 2006; Svangard et al. 2007; Simonsen et al. 2008). None of these pharmaceutically relevant activities are particularly useful to a plant, and it appears that the natural function of cyclotides is as insecticidal agents, based on the results of feeding trials in which Helicoverpa punctigera caterpillar larvae were fed diets containing cyclotides in comparable quantities to the cyclotide concentration in leaf tissue (Jennings et al. 2001). The cyclotide-fed caterpillars had marked growth retardation or mortality compared with controls that consumed cyclotide-free diets. Since that initial report, insecticidal activity has been confirmed in a number of other cyclotides, including kalata B2 (Jennings et al. 2005; Gruber et al. 2007a) and kalata B7 (A. Elliott et al., unpublished data). The structure of the latter peptide was recently reported (Shenkarev et al. 2008). Furthermore, various cyclotides have a range of other pesticidal activities that have potential applications in agriculture, including anthelmintic (Colgrave et al. 2008a, b, 2009) and molluscicidal activities (Plan et al. 2008). Although the natural function of MCoTI-II and MCoTI-II is not known, their trypsin inhibitory activity (Hernandez et al. 2000) might protect seeds from degradation in the guts of foraging animals. In addition to the potential of cyclotides as drugs themselves, based on their various pharmaceutically relevant activities, they also have been proposed as templates in protein engineering applications (Hernandez et al. 2000; Craik et al. 2006, 2007; Leta Aboye et al. 2008). Because of their stability and small size, they make excellent mini-protein scaffolds, onto which foreign bioactive peptide sequences can be grafted to stabilize them. For example, we recently showed that it is possible to graft anti-angiogenic activity on the kalata B1 scaffold (Gunasekera et al. 2008) and Thongyoo et al. (2008) re-engineered the MCoTI-II scaffold to introduce inhibitory activity against a protease associated with the foot and mouth disease virus, and against two serine proteases involved in inflammatory diseases (Thongyoo et al. 2009). To capitalize on the potential of cyclotides as protein engineering tools, it is essential to have efficient means to make them, and so considerable efforts from our laboratories over the last few years have been directed toward methods for the synthesis and folding of cyclotides. Section 3.2 of this chapter describes in vitro folding studies, and Sect. 3.3 describes folding studies using biological catalysts. Another interesting approach that shows great promise but is outside the scope of this chapter is the use of inteins to generate cyclotides (Kimura et al. 2006). It has recently been shown that this approach can in principle be used to make libraries of cyclotides inside bacterial cells (Camarero et al. 2007).
3 Cyclotide Folding
49
3.2 Chemical Folding Studies The in vitro synthesis of cyclotides typically involves peptide chain assembly using solid phase peptide synthesis (SPPS) followed by cyclization and folding. At around 30 amino acids in size, cyclotides are well suited to SPPS and can be assembled in a few days using manual synthesis or overnight using an automated peptide synthesizer. Initial studies were aimed at synthesizing native forms of cyclotides, whereas more recent studies have focused either on the synthesis of variants to explore structure–activity relationships (Simonsen et al. 2008) or on grafted cyclotides for protein engineering applications (Gunasekera et al. 2008; Thongyoo et al. 2008, 2009). In all of these studies, thioester-based chemistry has been found to be the most efficient means of chemically producing the cyclic product (Tam and Lu 1997; Daly et al. 1999; Tam et al. 1999a, b). Folding to form the cystine knot motif of cyclotides has been achieved using selective protection of cysteines during synthesis or alternatively, the use of undirected disulfide formation followed by separation of the correctly folded product by RP-HPLC from the crude folding mixture. The former approach was initially applied to the prototypic kalata B1, the circulins A and B and cyclopsychotride A, but yields were low (Tam et al. 1999b). Although protection of one pair of cysteines cuts the number of disulfide isomers from 15 to three, the three isomers were found in similar amounts (Tam et al. 1999b). In the undirected oxidative folding approach however, we were able to form the cystine knot motif in kalata B1 in a one-step reaction. The breakthrough here involved adding 50% isopropanol to the usual bicarbonate buffer system to facilitate the formation of the native isomer (Daly et al. 1999). It appears that the hydrophobic nature of the isopropanol-containing buffer stabilizes the development of the surface-exposed hydrophobic patch present in the native structure. (We noted earlier in this chapter that one of the characteristics of cyclotides is their surface-exposed patch of hydrophobic residues). In principle it is possible to form a CCK motif either by first cyclizing the peptide backbone and then by oxidizing the cysteine residues to form the disulfide bonds, or first forming the disulfide bonds and then the cyclic backbone. Interestingly, in the case of kalata B1 formation of the cyclic backbone prior to disulfide bond formation results in higher yields of correctly folded peptide compared with oxidation before cyclization (Daly et al. 1999). Although very important for in vitro studies, this findingmight not have biological significance because in vivo the formation of the disulfide bonds is likely to occur before cyclization, as described in Sect. 3.3. Following on from the initial studies aimed simply at synthesizing cyclotides, the mechanism by which the cyclotides form the cystine knot connectivity has been explored in recent studies. Analysis of the oxidative folding of a sample of fully reduced kalata B1 using RP-HPLC, mass spectrometry, and NMR spectroscopy revealed that a species, which contains two native disulfide bonds (CysII–CysV, CysIII–CysVI) and two reduced cysteine residues (CysI and CysIV), is the major intermediate present during the folding process in the isopropanol-containing buffer system (Daly et al. 2003). NMR studies suggested that a native-like fold is present
50
N.L. Daly et al.
in this intermediate and this suggestion was confirmed by calculation of the three-dimensional structure of a two-disulfide form of kalata B1 where CysI and CysIV were replaced with alanine residues. Intriguingly, although the two-disulfide intermediate has a native-like structure, it is not the direct precursor of the native peptide during in vitro folding, but rather appears to be a “trap” that requires the presence of shuffling reagents such as glutathione to allow formation of the native connectivity. The reason for the accumulation of this intermediate appears to be related to its stable native-like structure. Furthermore, this intermediate is highly abundant during folding in buffers containing isopropanol, but it is of minor abundance in buffers without isopropanol, demonstrating that the native-like structure is stabilized by the presence of organic solvent. Analysis of the corresponding reductive unfolding process for native kalata B1 did not detect this two-disulfide intermediate, and indeed the only intermediates observed were present at very low concentrations and were dependent on the conditions used for the reductive unfolding reaction (Daly et al. 2003; Göransson and Craik 2003). When the unfolding studies were repeated in the presence of the reducing agent tris(2-carboxyethyl)phosphine intermediates could be isolated and it appears that the first disulfide bond to break in the reduction process is CysII–CysV, suggesting that des[CysII–CysV]kalata B1 is the direct precursor of the native peptide during oxidative refolding (Daly et al. 2003; Göransson and Craik 2003). Figure 3.3 summarizes intermediates that have been characterized on the folding/unfolding pathway, and demonstrates the impact of choice of folding buffer.
Fig. 3.3 Snapshot of the folding of the Möbius cyclotide kalata B1. (a) Summary of intermediates that have been characterized on the folding (the top series) and the unfolding pathways (Daly et al. 2003; Göransson and Craik 2003; U. Göransson et al., unpublished results). The two-disulfide species occurring on the unfolding pathway, des[CysII–CysV]kalata B1, has been suggested to be the direct precursor to the native peptide (N), instead of the intermediate that has been identified on the folding pathway, des[CysI–CysIV]kalata B1. The latter intermediate is the major two-disulfide species in the standard folding buffer system containing isopropanol, shown in the upper HPLC-trace in (b). The folding pattern and kinetics are highly affected by the presence of isopropanol: the HPLC traces are snapshots of the folding progress in the same buffer but with and without isopropanol (after 30 min, 2/1 mM reduced and oxidized glutathione, 0.1 M ammonium bicarbonate). Notably, the folding pathway involves a large number of intermediates as judged by the number of chromatographic peaks. Peaks containing characterized species are marked
3 Cyclotide Folding
51
Fig. 3.3 continued
The folding/unfolding studies described so far relate to kalata B1, a member of the Möbius subfamily of cyclotides. Given the differences in sequences of the two other subfamilies of cyclotides it was of interest to also examine the oxidative folding of prototypic members from these subfamilies. The oxidative folding of MCoTI-II, from the trypsin inhibitor subfamily of cyclotides, produces a topologically similar two-disulfide intermediate, i.e., des[CysI–CysIV], to that observed during the folding of kalata B1 (Cĕmažar et al. 2006). Interestingly, this intermediate is the direct precursor to the native peptide, in sharp contrast to the case for the analogous intermediate for kalata B1. This difference is also reflected in differences in reductive unfolding as the des[CysI–CysIV] intermediate accumulates during reductive unfolding of MCoTI-II (Cĕmažar et al. 2006). The oxidative folding of MCoTI-II is similar to that of a linear trypsin inhibitor, EETI-II (Le-Nguyen et al. 1993), which has significantly higher sequence homology to MCoTI-II compared with kalata B1, albeit without the cyclic backbone. A nativelike intermediate is observed during the oxidative refolding of EETI-II that is the immediate precursor to the native. The differences in the size of the cystine knot
52
N.L. Daly et al.
Fig. 3.4 Dissecting the folding of the bracelet cyclotide cycloviolacin O1. Analysis of the folding of a series of hybrids of kalata B1 and cycloviolacin O1 indicated that two changes could dramatically increase the yields of the correctly folded product under standard folding conditions (Gunasekera et al. 2009). The two changes are highlighted on the sequence and structure of cycloviolacin O1; an amino acid substitution was made in loop 2 where the Ile is highlighted and an amino acid addition was made to loop 6 at the site indicated by an arrow
ring have been suggested to account for these differences in folding (Cĕmažar et al. 2006). The disulfide ring of the trypsin inhibitor cyclotides, at 11 amino acids is larger than that of kalata B1 (eight amino acids) and other cyclotides. Larger rings contribute to independence of disulfide bonds and may favor an on-pathway twodisulfide (CysII–CysV, CysIII–CysVI) intermediate. The bracelet subfamily also appears to have some differences in its in vitro folding relative to the other two subfamilies. Members of this subfamily have so far proved to be the most difficult to fold in significant yields. Indeed, this has been a significant limitation on the use of bracelet cyclotides as protein engineering frameworks, but promising results on their folding have been published recently. After extensive folding trials, a buffer containing optimized concentrations of dimethyl sulfoxide, detergent (Brij 35), and redox agents was found to promote folding of the bracelet cyclotide cycloviolacin O2 to yields of more than 50% (Leta Aboye et al. 2008). Furthermore, we dissected factors that contributed to the folding of the bracelet cyclotide cycloviolacin O1 by making a series of synthetic hybrids between this peptide and the Möbius cyclotide kalata B1 (Gunasekera et al. 2009). It was found that two minor changes, a replacement of an Ile for a Gly in loop 2 and insertion of a Thr in loop 6, resulted in high yields of the native isomer in vitro. The locations of these residues are shown on the structure of cycloviolacin O1 in Fig. 3.4. In contrast to kalata B1 and MCoTI-II, the major intermediate accumulating during the folding of bracelet cyclotides is a non-native three-disulfide bond species.
3 Cyclotide Folding
53
Fig. 3.5 Comparison of the folding pathways of the cyclotide subfamilies. Schematic representations of the reduced (left), major intermediates (middle), and native structures (right) are given. A three-disulfide bond intermediate with non-native connectivities accumulates during the folding of cycloviolacin O2 (bracelet cyclotide), which slowly converts to the native product (Leta Aboye et al. 2008). A two-disulfide intermediate with native-connectivities accumulates during the folding of MCoTI-II (trypsin inhibitor cyclotide), which is the immediate precursor to the native form (Ceˇmažar et al. 2006). A topologically equivalent intermediate accumulates during the folding of kalata B1, but it requires shuffling reagents to form the native product (Daly et al. 2003)
This intermediate has been identified both in the folding of the naturally occurring cycloviolacin O2 and a designed synthetic hybrid of cycloviolacin O1 and kalata B1, and contains disulfide bonds between neighboring cysteines, i.e., CysI–CysII, CysIII–CysIV, and CysV–CysVI (Leta Aboye et al. 2008; Gunasekera et al. 2009). Furthermore, the kinetics of the folding process distinguishes bracelet cyclotides from kalata B1 and MCoTI-II, whereas the folding of the latter is characterized by the occurrence of one- and two-disulfide species, cycloviolacin O2 collapses within minutes into the misfolded three-disulfide species (T. Leta Aboye et al., unpublished results). The identity of one- and two-disulfide species preceding this collapse is not yet known. It is clear though that non-native one-disulfide species containing bonds between neighboring cysteines form as primary events in the folding of kalata B1 (U. Göransson et al., unpublished results). A summary of the major folding intermediates for the three subfamilies of cyclotides is given in Fig. 3.5. Overall, it appears that formation of the CCK fold in vitro occurs through a variety of pathways. More generally, these differences reflect the complexity of the oxidative folding of a wide range of cystine-rich
54
N.L. Daly et al.
peptides (Chang et al. 2000). The small and compact nature of cyclotides makes them excellent model systems for studies of the oxidative folding of proteins in general. Overall, the in vitro studies are of course valuable for the production of cyclotides, but detailed analyses of the folding of cyclotides in vivo are also important to provide a more coherent picture of folding pathways in a biological setting.
3.3 Biological Folding Studies The discussion so far has focused on in vitro folding studies on CCK proteins; however, it is important to note that there have been recent developments concerning the folding of cyclotides using biological auxiliaries. We recently isolated and biochemically characterized a PDI from OaPDI and demonstrated that it facilitates the production of cyclotides. PDI is a generic term for oxidoreductase enzymes of the thioredoxin superfamily (Ellgaard et al. 1999). PDI has a major role in the oxidative folding of polypeptides in the endoplasmic reticulum (ER) of eukaryotic cells and functions as an ER chaperone (Wilkinson and Gilbert 2004). The exact mechanism of PDI action in cells is not clear but it is believed to bind polypeptides through hydrophobic interactions, and carry out the formation (oxidation), breakage (reduction), and/or shuffling (isomerization) of disulfide bonds in substrate molecules via a dithiol–disulfide exchange between its active-site CXXC motif and the substrate polypeptide (Gruber et al. 2006).
3.3.1 Isolation and Biochemical Characterization of a Novel Plant PDI We isolated a cDNA clone encoding the novel OaPDI and showed that the corresponding protein is expressed in the cyclotide-producing plant. After establishing a method for the expression and purification of the recombinant PDI, we compared enzyme activities of OaPDI with that of human PDI (hPDI) proteins and related the results to unique primary sequence features of the plant PDI (Gruber et al. 2007b). For instance, OaPDI exhibited lower isomerase activity, i.e., 70% of that of hPDI. This was explained by an amino acid substitution at a position following the conserved thioredoxin active-site motif (CXXC), which plays a critical role in oxidoreductase activity. Another distinctive feature of the new OaPDI was the high isoelectric point (pI 6.6) of the C-terminal extension domain, which is very acidic (low pI) in PDI proteins than other organisms (Koivunen et al. 1999; Tian et al. 2004, 2006). We provided the first activity data for a PDI containing a nonacidic C-terminal extension domain and concluded that this domain does not have to be acidic for a PDI to be a functional chaperone. Other sequence features of the OaPDI are similar to known PDIs, including its domain structure and redox potential. Although the biological behavior of
3 Cyclotide Folding
55
d ithiol–disulfide oxidoreductases cannot be deduced solely from biochemical enzyme characteristics, the redox potential is often a valuable indicator for in vivo function of a novel oxidoreductase. Hence, by analogy, and with the known function of other PDI proteins, OaPDI was thought to have a role in the folding of cellular proteins, particularly the cyclotides, which are major compounds in O. affinis. In a more recent study, we reported a detailed analysis of the oligomerization behavior of OaPDI using various biochemical and biophysical techniques, including SDS-PAGE, size-exclusion chromatography, NMR spectroscopy, surface plasmon resonance, and atomic force microscopy (Gruber et al. 2009). This baseline study might be valuable for future in vivo studies and could be of interest for the in vitro use of this novel PDI in a variety of biotechnological applications. More generally, the biological and physiological role of PDI oligomerization remains unknown. Rat PDI dimer formation reportedly has no catalytic redox function, but it is instead considered important for storage and detoxification of metal ions (Solovyov and Gilbert 2004). Similarly, crystals of yeast PDI did not indicate dimer formation, but showed the presence of tetramers that are probably disulfide-linked. To date, these tetramers are not considered to have a physiological role (Tian et al. 2006). Our observations showed that the formation of monomers or dimers/oligomers depends on the exposure of the proteins to physicochemical conditions. The ability to switch between different oligomer states might be a mechanism by which increased surface area is generated for binding of PDI substrate molecules. These larger polypeptide-binding clefts might be important for the oxidative folding of larger nascent proteins that would not otherwise be folded; hence, this mechanism would constitute a very important cellular function.
3.3.2 Enzymatic Folding of Cyclotides In Vitro After a biochemical characterization of the OaPDI, we examined its role in enzymatic (oxidative) folding of kalata B1 and a synthetic linear analog (Gruber et al. 2007b). The recombinant OaPDI enzyme significantly increased the yield of correctly disulfide-bonded species in vitro and can shuffle non-native disulfide bonds into their correct connectivity. This isomerization step is necessary to produce fully functional insecticidal cyclotides. The results of these studies provided valuable insight into the mechanism of PDI. When folding was catalyzed by OaPDI, the yield of correctly folded product increased by ~10- and ~26-fold for linear and cyclic kalata B1, respectively, as compared with the controls (see Fig. 3.6a). The PDI-folded cyclotides were tested in larval feeding trials and showed identical insecticidal activity to native kalata B1, confirming the identity of naturally occurring and enzymatically in vitro folded peptides (see Fig. 3.6b). To obtain more information about the role of OaPDI in the oxidative folding of cyclotides, we examined the major intermediate species on folding pathways in the presence and absence of PDI. PDI-assisted folding leads to oxidized peptides that either have folded into their native disulfide connectivities or appear as energetically
56
N.L. Daly et al.
Fig. 3.6 Enzymatic folding of kalata B1 using OaPDI. (a) HPLC traces of cyclic and linear kalata B1 folded for 24 h at pH 7.5 with or without the addition of OaPDI. The enzyme significantly increases the yield of correctly folded kalata peptide (*) after 24 h under physiological conditions. (b) The OaPDI-produced peptides were fully functional as insecticidal molecules compared with plant-extracted kalata B1, as analyzed in feeding trials by measuring the size and weight of Helicoverpa punctigera larvae, the amount of frass produced, and the amount of peptide diet left over. (c) Folding intermediates of both cyclic and linear kalata B1 without OaPDI [dashed rectangle in (a)] consist of three-disulfide misfolded and two-disulfide partially folded peptides. Regardless of the presence or absence of the enzyme, the peptides undergo thiol oxidation to form disulfide bonds, but shuffling of non-native disulfide bonds into their native connectivity is greatly compromised without OaPDI. Hence, OaPDI appears to have an important role in the isomerization of disulfide bonds during the folding of cystine knot peptides. (d) Structure of the Oak1 precursor protein, highlighting the N-terminal repeat (NTR) region amino acid sequence. Various deletion constructs (+4, +8, +12, and +17) were analyzed for their propensity of binding to OaPDI to study the influence of size and hydrophobicity to the strength of biomolecular interaction
3 Cyclotide Folding
57
trapped non-native three-disulfide species. Without PDI, the folding still proceeds to the initial stage of non-specific disulfide bond formation by oxidation (“packing”), to the point where mostly oxidized and partially oxidized (3SS and 2SS) peptides are present. However, it appears that they are trapped and cannot effectively undergo disulfide reshuffling (“consolidation”) to the native species, resulting in a lower yield of correctly oxidized peptides. In short, oxidation proceeds without the enzyme, but the shuffling of misfolded, or partially folded, intermediates into their native disulfide connectivity is significantly reduced (see Fig. 3.6c). This reinforces the notion that isomerization is a major function of the new plant PDI.
3.3.3 Enzymatic Folding of Cyclotides In Planta Besides the biochemical description and in vitro folding activity of the plant PDI, it is important to study the mechanism of action of PDI-assisted folding of cyclotides in vivo. The current model for the production of cyclotides in planta involves oxidation of the disulfide bonds in the precursor molecule within the ER, and processing and cyclization further downstream in the secretory pathway. As noted in Sect. 3.1 of this chapter, an AEP enzyme (Saska et al. 2007; Gillon et al. 2008) recognizes a conserved Asn (or Asp) residue at the C-terminal end of the mature cyclotide sequence and is implicated in the cyclization of the mature peptide. This implies that oxidative folding occurs before cyclization in the ER. We have shown that OaPDI and the Oak1 cyclotide precursor co-express on a transcriptional level, although, at the time the experiments were conducted, it was not possible to show co-localization of OaPDI and Oak1 on a subcellular level using antibody experiments because the Oak1 protein could not be detected by the antibody (C. Gruber et al., unpublished observations). This is not surprising though, because the antibody to Oak1 is very weak (Jennings et al. 2001; Saska et al. 2007). The reason for this lack of sensitivity was thought to be rapid processing of the Oak1 precursor in vivo. To overcome these limitations, surface plasmon resonance was used to demonstrate a strong biomolecular interaction of OaPDI and Oak1. It has been reported previously that the presence of the peptide pro-region in cone snail toxin precursors increases the enzymatic folding rate of those conotoxins using PDI (Buczek et al. 2004). The NTR region of cyclotide precursors appears similar to the pro-region of conotoxin precursors and it is probable that there are sequence or structural elements in the N-pro or NTR region of Oak1 that could have a similar role. Unlike the conotoxin precursors, the NTR sequences of various cyclotide precursors do not share much commonality within different plant species; however, it has been suggested that the NTRs form a conserved structural motif (Dutton et al. 2004). To analyze the molecular mechanisms of this interaction, a number of Oak1 deletion proteins were analyzed by surface plasmon resonance experiments. The protein substrates contained the mature cyclotide domain + C-terminal tail and a varying number of residues of the NTR region (see Fig. 3.6d). By increasing the length of the NTR region from four to eight residues, the interaction with OaPDI
58
N.L. Daly et al.
significantly increased (C. Gruber et al., unpublished observations) and hence this region might be important for interaction of Oak1 and OaPDI. This average resonance response decreases slightly by adding another four (+12 NTR) or nine (+17 NTR) residues of the NTR region, but the interaction response is still significantly greater compared with the (+4) NTR. The (+8) NTR construct contains an N-terminal exposed Phe residue (see Fig. 3.6d), which, due to its hydrophobicity, may increase the strength of binding to the hydrophobic surface of OaPDI. Increasing the length of the NTR might result in the Phe residue becoming buried; hence, it would not be available to promote this stronger interaction. The increase in length may promote a better interaction of the elongated substrates with the hydrophobic area of OaPDI. This is in agreement with recent findings that an increase in the length of the NTR region results in an increased tendency for a disordered structure and hence a more hydrophobic surface (S. Gunasekera et al., unpublished results). Although we are yet to provide direct evidence of a protein–protein interaction of OaPDI and Oak1 in vivo, we have shown that they co-express on a transcriptional level and show a strong bio-molecular interaction, which together suggest an in vivo interaction of the two proteins. Furthermore, we suggest a novel role for the NTR region of cyclotide precursors, i.e., it may be important for interaction and substrate recognition with the folding enzyme OaPDI.
3.4 Outlook In conclusion, in this chapter we have described studies from our laboratories in which we have examined the oxidative folding of cyclotides using a range of approaches, from in vitro chemical studies to biological studies with the aid of a newly discovered PDI from O. affinis. These findings are potentially important for biotechnological applications of cyclotides, and more generally for a range of disulfide-rich proteins. One limiting factor for the overproduction of proteins is the formation of the correct disulfide connectivity, and there is a need for effective folding systems and catalysts to form disulfide-rich peptides. OaPDI might become an important tool for the production of cyclotides and other proteins in vitro and in vivo. Cyclotides have a range of biological activities that make them attractive candidates for drug development, agricultural applications, and as novel molecular scaffolds (Craik 2006). To increase the rate and yield of the folding process, OaPDI could be used as an additive in the final refolding step for synthetically made cyclotides or those made in transgenic plants by OaPDI co-expression. From a more fundamental perspective, the studies described here have assisted us in (literally and metaphorically) unraveling a knotty problem in protein folding. Acknowledgments Work in DC’s laboratory on cyclotides is supported by grants from the Australian Research Council (ARC) and the National Health and Medical Research Council (NHMRC). NLD is a Queensland Smart State Fellow and DJC is an NHMRC Principal Research Fellow. UG is supported by grants from the Swedish Research Council (VR) and the Swedish Foundation for Strategic Research (SSF).
3 Cyclotide Folding
59
References Barry DG, Daly NL, Clark RJ et al (2003) Linearization of a naturally occurring circular protein maintains structure but eliminates hemolytic activity. Biochemistry 42:6688–6695 Buczek O, Olivera BM, Bulaj G (2004) Propeptide does not act as an intramolecular chaperone but facilitates protein disulfide isomerase-assisted folding of a conotoxin precursor. Biochemistry 43:1093–1101 Burman R, Gruber CW, Rizzardi K et al (2010) Cyclotide proteins and precursors from the genus Gloeospermum: filling a blank spot in the cyclotide map of Violaceae. Phytochemistry 71: 13–20 Camarero JA, Kimura RH, Woo Y-H et al (2007) Biosynthesis of a fully functional cyclotide inside living bacterial cells. Chembiochem 8:1363–1366 Ceˇmažar M, Daly NL, Haggblad S et al (2006) Knots in rings: the circular knotted protein Momordica cochinchinensis trypsin inhibitor-II folds via a stable two-disulfide intermediate. J Biol Chem 281:8224–8232 Chang JY, Li L, Bulychev A (2000) The underlying mechanism for the diversity of disulfide folding pathways. J Biol Chem 275:8287–8289 Chiche L, Heitz A, Gelly JC et al (2004) Squash inhibitors: from structural motifs to macrocyclic knottins. Curr Protein Pept Sci 5:341–349 Colgrave ML, Craik DJ (2004) Thermal, chemical, and enzymatic stability of the cyclotide kalata B1: the importance of the cyclic cystine knot. Biochemistry 43:5965–5975 Colgrave ML, Kotze AC, Huang YH et al (2008a) Cyclotides: natural, circular plant peptides that possess significant activity against gastrointestinal nematode parasites of sheep. Biochemistry 47:5581–5589 Colgrave ML, Kotze AC, Ireland DC et al (2008b) The anthelmintic activity of the cyclotides: natural variants with enhanced activity. Chembiochem 9:1939–1945 Colgrave ML, Kotze AC, Kopp S et al (2009) Anthelmintic activity of cyclotides: in vitro studies with canine and human hookworms. Acta Trop 109:163–166 Craik DJ (2001) Plant cyclotides: circular, knotted peptide toxins. Toxicon 39:1809–1813 Craik DJ (2006) Seamless proteins tie up their loose ends. Science 311:1563–1564 Craik DJ, Daly NL, Bond T et al (1999) Plant cyclotides: a unique family of cyclic and knotted proteins that defines the cyclic cystine knot structural motif. J Mol Biol 294:1327–1336 Craik DJ, Daly NL, Waine C (2001) The cystine knot motif in toxins and implications for drug design. Toxicon 39:43–60 Craik DJ, Daly NL, Mulvenna J et al (2004) Discovery, structure and biological activities of the cyclotides. Curr Protein Pept Sci 5:297–315 Craik DJ, Cĕmažar M, Daly NL (2006) The cyclotides and related macrocyclic peptides as scaffolds in drug design. Curr Opin Drug Discov Devel 9:251–260 Craik DJ, Clark RJ, Daly NL (2007) Potential therapeutic applications of the cyclotides and related cystine knot mini-proteins. Expert Opin Investig Drugs 16:595–604 Daly NL, Love S, Alewood PF et al (1999) Chemical synthesis and folding pathways of large cyclic polypeptides: studies of the cystine knot polypeptide kalata B1. Biochemistry 38:10606–10614 Daly NL, Clark RJ, Craik DJ (2003) Disulfide folding pathways of cystine knot proteins. Tying the knot within the circular backbone of the cyclotides. J Biol Chem 278:6314–6322 Dutton JL, Renda RF, Waine C et al (2004) Conserved structural and sequence elements implicated in the processing of gene-encoded circular proteins. J Biol Chem 279:46858–46867 Ellgaard L, Molinari M, Helenius A (1999) Setting the standards: quality control in the secretory pathway. Science 286:1882–1888 Felizmenio-Quimio ME, Daly NL, Craik DJ (2001) Circular proteins in plants: solution structure of a novel macrocyclic trypsin inhibitor from Momordica cochinchinensis. J Biol Chem 276:22875–22882 Gillon AD, Saska I, Jennings CV et al (2008) Biosynthesis of circular proteins in plants. Plant J 53:505–515
60
N.L. Daly et al.
Göransson U, Craik DJ (2003) Disulfide mapping of the cyclotide kalata B1. Chemical proof of the cystic cystine knot motif. J Biol Chem 278:48188–48196 Göransson U, Luijendijk T, Johansson S et al (1999) Seven novel macrocyclic polypeptides from Viola arvensis. J Nat Prod 62:283–286 Gran L (1970) An oxytocic principle found in Oldenlandia affinis DC. Medd Nor Farm Selsk 12:173–180 Gran L (1973) On the effect of a polypeptide isolated from “Kalata–Kalata” (Oldenlandia affinis DC) on the oestrogen dominated uterus. Acta Pharmacol Toxicol 33:400–408 Gruber CW, Cemazar M, Heras B et al (2006) Protein disulfide isomerase: the structure of oxidative folding. Trends Biochem Sci 31:455–464 Gruber CW, Cemazar M, Anderson MA et al (2007a) Insecticidal plant cyclotides and related cystine knot toxins. Toxicon 49:561–575 Gruber CW, Cemazar M, Clark RJ et al (2007b) A novel plant protein-disulfide isomerase involved in the oxidative folding of cystine knot defense proteins. J Biol Chem 282:20435–20446 Gruber CW, Elliott AG, Ireland DC et al (2008) Distribution and evolution of circular miniproteins in flowering plants. Plant Cell 20:2471–2483 Gruber CW, Cemazar M, Mechler A et al (2009) Biochemical and biophysical characterization of a novel plant protein disulfide isomerase. Biopolymers 92:35–43 Gunasekera S, Foley FM, Clark RJ et al (2008) Engineering stabilized vascular endothelial growth factor-A antagonists: synthesis, structural characterization, and bioactivity of grafted analogues of cyclotides. J Med Chem 51:7697–7704 Gunasekera S, Daly NL, Clark RJ et al (2009) Dissecting the oxidative folding of circular cystine knot miniproteins. Antioxid Redox Signal 11:971–980 Gustafson KR, Sowder RCI, Henderson LE et al (1994) Circulins A and B: Novel HIV-inhibitory macrocyclic peptides from the tropical tree Chassalia parvifolia. J Am Chem Soc 116:9337–9338 Heitz A, Hernandez JF, Gagnon J et al (2001) Solution structure of the squash trypsin inhibitor MCoTI-II. A new family for cyclic knottins. Biochemistry 40:7973–7983 Hernandez JF, Gagnon J, Chiche L et al (2000) Squash trypsin inhibitors from Momordica cochinchinensis exhibit an atypical macrocyclic structure. Biochemistry 39:5722–5730 Herrmann A, Svangard E, Claeson P et al (2006) Key role of glutamic acid for the cytotoxic activity of the cyclotide cycloviolacin O2. Cell Mol Life Sci 63:235–245 Huang YH, Colgrave ML, Daly NL et al (2009) The biological activity of the prototypic cyclotide kalata B1 is modulated by the formation of multimeric pores. J Biol Chem 284:20699–20707 Isaacs NW (1995) Cystine knots. Curr Opin Struct Biol 5:391–395 Jennings C, West J, Waine C et al (2001) Biosynthesis and insecticidal properties of plant cyclotides: the cyclic knotted proteins from Oldenlandia affinis. Proc Natl Acad Sci USA 98:10614–10619 Jennings CV, Rosengren KJ, Daly NL et al (2005) Isolation, solution structure, and insecticidal activity of kalata B2, a circular protein with a twist: do Mobius strips exist in nature? Biochemistry 44:851–860 Kamimori H, Hall K, Craik DJ et al (2005) Studies on the membrane interactions of the cyclotides kalata B1 and kalata B6 on model membrane systems by surface plasmon resonance. Anal Biochem 337:149–153 Kimura RH, Tran A-T, Camarero JA (2006) Biosynthesis of the cyclotide kalata B1 by using protein splicing. Angew Chem Int Ed Engl 118:987–990 Koivunen P, Pirneskoski A, Karvonen P et al (1999) The acidic C-terminal domain of protein disulfide isomerase is not critical for the enzyme subunit function or for the chaperone or disulfide isomerase activities of the polypeptide. EMBO J 18:65–74 Koradi R, Billeter M, Wüthrich K (1996) MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph 14:29–32 Le-Nguyen D, Heitz A, Chiche L et al (1993) Characterization and 2D NMR study of the stable [9–21, 15–27] 2 disulfide intermediate in the folding of the 3 disulfide trypsin inhibitor EETI II. Protein Sci 2:165–174 Leta Aboye T, Clark RJ, Craik DJ et al (2008) Ultra-stable peptide scaffolds for protein engineeringsynthesis and folding of the circular cystine knotted cyclotide cycloviolacin O2. Chembiochem 9:103–113
3 Cyclotide Folding
61
Lindholm P, Goransson U, Johansson S et al (2002) Cyclotides: a novel type of cytotoxic agents. Mol Cancer Ther 1:365–369 Mulvenna JP, Wang C, Craik DJ (2006) CyBase: a database of cyclic protein sequence and structure. Nucleic Acids Res 34:D192–D194 Plan MR, Saska I, Cagauan AG et al (2008) Backbone cyclised peptides from plants show molluscicidal activity against the rice pest Pomacea canaliculata (golden apple snail). J Agric Food Chem 56:5237–5241 Rosengren KJ, Daly NL, Plan MR et al (2003) Twists, knots, and rings in proteins. Structural definition of the cyclotide framework. J Biol Chem 278:8606–8616 Saether O, Craik DJ, Campbell ID et al (1995) Elucidation of the primary and three-dimensional structure of the uterotonic polypeptide kalata B1. Biochemistry 34:4147–4158 Saska I, Craik DJ (2008) Protease-catalysed protein splicing: a new post-translational modification? Trends Biochem Sci 33:363–368 Saska I, Gillon AD, Hatsugai N et al (2007) An asparaginyl endopeptidase mediates in vivo protein backbone cyclisation. J Biol Chem 282:29721–29728 Schöpke T, Hasan Agha MI, Kraft R et al (1993) Hämolytisch aktive komponenten aus Viola tricolor L. und Viola arvensis Murray. Sci Pharm 61:145–153 Shenkarev ZO, Nadezhdin KD, Lyukmanova EN et al (2008) Divalent cation coordination and mode of membrane interaction in cyclotides: NMR spatial structure of ternary complex Kalata B7/Mn2+/DPC micelle. J Inorg Biochem 102:1246–1256 Simonsen SM, Sando L, Rosengren KJ et al (2008) Alanine scanning mutagenesis of the prototypic cyclotide reveals a cluster of residues essential for bioactivity. J Biol Chem 283:9805–9813 Sletten K, Gran L (1973) Some molecular properties of kalatapeptide B-1: a uterotonic polypeptide isolated from Oldenlandia affinis DC. Medd Nor Farm Selsk 7–8:69–82 Solovyov A, Gilbert HF (2004) Zinc-dependent dimerization of the folding catalyst, protein disulfide isomerase. Protein Sci 13:1902–1907 Svangard E, Burman R, Gunasekera S et al (2007) Mechanism of action of cytotoxic cyclotides: cycloviolacin O2 disrupts lipid membranes. J Nat Prod 70:643–647 Tam JP, Lu Y-A (1997) Synthesis of large cyclic cystine-knot peptide by orthogonal coupling strategy using unprotected peptide precursors. Tetrahedron Lett 38:5599–5602 Tam JP, Lu Y-A, Yu Q (1999a) Thia zip reaction for synthesis of large cyclic peptides: mechanisms and applications. J Am Chem Soc 121:4316–4324 Tam JP, Lu YA, Yang JL et al (1999b) An unusual structural motif of antimicrobial peptides containing end-to-end macrocycle and cystine-knot disulfides. Proc Natl Acad Sci USA 96:8913–8918 Thongyoo P, Roque-Rosell N, Leatherbarrow RJ et al (2008) Chemical and biomimetic total syntheses of natural and engineered MCoTI cyclotides. Org Biomol Chem 6:1462–1470 Thongyoo P, Bonomelli C, Leatherbarrow RJ et al (2009) Potent inhibitors of beta-tryptase and human leukocyte elastase based on the MCoTI-II scaffold. J Med Chem 52:6197–6200 Tian R, Li SJ, Wang DL et al (2004) The acidic C-terminal domain stabilizes the chaperone function of protein disulfide isomerase. J Biol Chem 279:48830–48835 Tian G, Xiang S, Noiva R et al (2006) The crystal structure of yeast protein disulfide isomerase suggests cooperativity between its active sites. Cell 124:61–73 Trabi M, Craik DJ (2002) Circular proteins – no end in sight. Trends Biochem Sci 27:132–138 Trabi M, Svangard E, Herrmann A et al (2004) Variations in cyclotide expression in Viola species. J Nat Prod 67:806–810 Wang CK, Kaas Q, Chiche L et al (2008) CyBase: a database of cyclic protein sequences and structures, with applications in protein discovery and engineering. Nucleic Acids Res 36:D206–D210 Wilkinson B, Gilbert HF (2004) Protein disulfide isomerase. Biochim Biophys Acta 1699:35–44 Witherup KM, Bogusky MJ, Anderson PS et al (1994) Cyclopsychotride A, A biologically active, 31-residue cyclic peptide isolated from Psychotria longipes. J Nat Prod 57:1619–1625
Chapter 4
In Vitro Folding of Single/Double Chain Insulins and Related Proteins You-Min Feng
Abstract Insulin is not only a very important protein hormone in clinic but also a perfect model molecule in protein science. Almost each important achievement in the field of protein science was established from insulin study, such as protein crystal, protein sequence, protein synthesis, protein expression, and protein engineering since its discovery by Banting and Best in 1921. However, concerning in vitro folding of insulin it was not well characterized until the early 2000, probably the major difficult being that insulin consists of two polypeptide chains linked by disulfides. Because insulin is synthesized in vivo as a single chain peptide and folds well, recently the study on the in vitro folding of insulin has been investigated from single chain insulin to double chain insulin. This chapter briefly summarizes the current studies on the in vitro folding of single/double chain insulins and related proteins including their folding process, intermediates, and putative folding pathways. Besides, the common folding behavior of single/double chain insulins and related proteins and the in vivo folding process of nascent proinsulin, as suggestion or speculation, are discussed. Keywords Insulin • Proinsulin • Porcine insulin precursor • Amphioxus insulinlike peptide • Insulin-like growth factor-1 • Disulfide-containing protein • Folding • Folding process • Folding intermediate • Putative folding pathway
Y.-M. Feng (*) Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yue-Yang Road, Shanghai 200031, China e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_4, © Springer Science+Business Media, LLC 2011
63
64
Y.-M. Feng
4.1 Introduction Insulin is a very important protein hormone in clinic and a perfect model protein in the field of protein science. Therefore, insulin has been broadly and deeply studied both theoretically and practically since its discovery in 1921 by Banting and Best (1921). In the protein science field, almost each important achievement was established from insulin study, such as protein purification, protein crystal, protein sequence, chemical synthesis of protein, gene expression of protein, protein engineering, and so on. Several achievements of insulin study are the milestones not only in medical science but also in protein science. However, the folding process, intermediates, and folding pathway of insulin are not well characterized until the early 2000, which might be mainly due to the difficulties caused by the double chain nature of insulin. Insulin consists of two polypeptide chains (designated as A chain and B chain, respectively) linked by three disulfides: two inter-AB chain disulfides, one intra-A chain disulfide. Although insulin consists of two polypeptide chains, it is synthesized in vivo as a properly folded single chain polypeptide (Steiner 1967). Because insulin is synthesized as a single chain precursor (proinsulin) and folds correctly in vivo, the investigation of in vitro folding of insulin started with single chain insulins including a recombinant porcine insulin precursor (PIP), human proinsulin (HPI), and a recombinant amphioxus insulin-like peptide (rAILP) to investigate their folding process, intermediates, and putative folding pathway. This review briefly summarizes the current progresses in the studies on in vitro folding of single/ double chain insulins and related proteins. Single/double chain insulin was simultaneously used to study their in vitro folding, which may not only help to reveal the relationship between single chain and double chain insulins in the folding processes but may also help to learn the folding mechanism of proteins with two polypeptide chains linked by disulfides.
4.2 In Vitro Folding of Single Chain/Double Chain Insulins and Related Proteins For understanding the in vitro folding of insulin we have firstly investigated the in vitro folding processes of single chain insulins and related proteins and then investigated the double chain insulin. Figure 4.1 shows the primary structures of insulin, single chain insulins, and related proteins.
4.2.1 In Vitro Folding of Single Chain Insulins and Related Proteins In this part, we mainly summarize the progress in the studies on in vitro folding of single chain insulins and related proteins, including PIP and HPI, rAILP, and insulinlike growth factor-1 (IGF-1) (Fig. 4.1).
4 Folding of Single/Double Chain Insulin
65
Fig. 4.1 Sequence of insulin, PIP, HPI, AILP, rAILP, and IGF-1. C, connect peptide (C-peptide), which links B chain C terminus to A chain N terminus. The length of C-peptide in insulin, PIP, HPI, AILP, rAILP, and IGF-1 is different: for porcine insulin precursor the “C-peptide” is dipeptide, Ala-Lys; for human proinsulin is 35 amino acids; for AILP is 31 amino acids; for rAILP is tripeptide, Ala-Ala-Ly; for IGF-1 is 12 amino acids; for insulin we may say that the “C-peptide” is zero or infinite length
4.2.1.1 Single Chain Insulins Porcine Insulin Precursor PIP is recombinant single chain insulin precursor, in which the B30Ala and A1Gly of porcine insulin are linked by a dipeptide, Ala-Lys (Fig. 4.1). The recombinant PIP can be expressed and secreted in a correctly folded and soluble form from transformed yeast Saccharomyces cerevisiae cells and can be converted into human insulin by means of in vitro transpeptidation. This variant has been used in clinic (Markussen et al. 1987; Zhang et al. 1996). The in vitro folding reaction of PIP was carried out by adding the fully reduced/ denatured PIP into the refolding buffer (50 mM Tris–HCl, 0.5 M l-Arg, 5 mM EDTA, 5 mM GSH, and 0.5 mM GSSG, pH 9.5) at 16°C with final protein concentration of 0.1 mg/ml (Qiao et al. 2001). Refolding yield can reach 85%. The folding intermediates were captured at different time points. The temporal distribution of intermediates is shown in Fig. 4.2a and three obvious intermediates were captured: one intermediate (designated 1SSPIP) contains one disulfide, possibly A6–A11; two intermediates (designated 2SSPIPa and 2SSPIPb) contain two disulfides, disulfide A20–B19 and another interchain disulfide (A7–B7 or A6–B7) as shown in Table 4.1. Figure 4.2b shows the native pH 8.3 PAGE patterns of the intermediates. CD spectra analysis indicated that the one-disulfide intermediate contains little ordered secondary
66
Y.-M. Feng
Fig. 4.2 Temporal distribution and PAGE pattern of the intermediates as well as the putative refolding pathway of PIP. (a) Temporal distribution of the intermediates; Peak I, 2SSPIPa; Peak II, 2SSPIPb, Peak III, 1SSPIP. (b) Native pH 8.3 PAGE pattern: lane 3, purified 1SSPIP; lane 8, purified 2SSPIPa; lane 9, purified 2SSPIPb. (c) The putative refolding pathway of PIP: intermediates are represented by their disulfide linkage pattern. Intermediates with an overstriking rectangle represent the intermediates captured in this study (Qiao et al. 2001)
structures, while the two two-disulfide intermediates retain significant amount of helical structures (Qiao et al. 2001). Based on the time-dependent formation and distribution of the captured intermediates, two different putative folding pathways were proposed as shown in Fig. 4.2c. The first folding pathway involves the rapid formation of the intra-A chain disulfide, followed by the slower formation of one interchain disulfide to complete refolding. The second folding pathway starts with the formation of the disulfide A20–B19, followed immediately by another interchain disulfide formation to complete refolding (Qiao et al. 2001). In the folding process of PIP, only three intermediates, [A6–A11]PIP, [A20–B19, A7–B7]PIP, and [A20–B19, A6–B7] or [A20–B19, A11–B7]PIP, were trapped (Qiao
4 Folding of Single/Double Chain Insulin
67
Table 4.1 The disulfide linkage of the captured intermediates during the refolding of PIP, HPI, and rAILP Protein Intermediate SS number Disulfidea PIP 1SSPIP 1SS A6–A11 2SSPIPa 2SS A20–B19, A7–B7 2SSPIP 2SS A20–B19, A6–B7 A20–B19, A11–B7 HPI
P2
3SS
P3
3SS
A6–B7, A7–A11, A20–B19 A6–A7, A11–B7, A20–B19
A20–B7, A6–A7, A11–B19 A20–B7, A6–A11, A7–B19 A20–B7, A7–A11, A6–B19 P4 3SS B7–B19, A6–A7, A11–A20 B7–B19, A6–A11, A7–A20 B7–B19, A6–A20, A7–A11 rAILP P1 2SS A20–B19, A6–B7 P2 2SS A20–B19, A7–B7 P3 2SS A20–B19, A6–A11 P4 1SS A20–B19 a Due to some technical limitations, the exact elucidation of the disulfide linkage is not possible, so the possible linkages for the intermediates are all listed
et al. 2001). Our previous work has shown that [A6–A11]PIP cannot be secreted at all from transformed yeast cells (Guo and Feng 2001). The formation of the intermediate [A6–A11]PIP in the folding process probably results from the proximity of the two Cys residues in sequence and the sequence of A6–A11 together with the disulfide A6–A11 form a 20-membered cyclic structure (Katsoyannis and Tometsko 1966). The other one-disulfide intermediate [A20–B19]PIP should be critical during the folding, although it cannot be trapped in the folding process, which may indicate that once formed the remaining folding process is completed in a fast way. For further understanding the folding process of PIP, five model peptides of possible intermediates with one or two disulfide were prepared by the means of protein engineering and named according to their remaining disulfide(s) as follows: [A20–B19] PIP (Yan et al. 2003), [A20–B19, A7–B7]PIP, [A20–B19, A6–B7]PIP, [A20–B19, A6–A11]PIP, and [A20–B19, A7–A11] (Jia et al. 2003). The model peptide [A20– B19]PIP can be secreted from transformed yeast cells and adopts a partially folded conformation. In vitro, the fully reduced [A20–B19]PIP can quickly and efficiently form the disulfide A20–B19. Therefore, it is possible to deduce that this one-disulfide intermediate is just like a template to usher the later folding process of PIP (Yan et al. 2003). The model peptide [A20–B19]PIP can be converted into (desB30) [A20–B19]insulin by digestion with endoproteinase Lys-C. The receptor binding activity of (desB30)[A20–B19]insulin is too low to be quantified. However, (desB30)[A20–B19, A6–A11]insulin, which was obtained from [A20–B19, A6–A11]PIP by the same means, retains a receptor binding activity of about 0.1% of that in native insulin (Guo and Feng 2001).
68
Y.-M. Feng
The other four two-disulfide model peptides all adopt partially folded structures. In redox buffer, the disulfides of the model peptides are more easily reduced than those of the wild-type PIP. During in vitro refolding, the reduced model peptides share similar relative folding rates but different folding yields, which indicate that the folding intermediates corresponding to the present model peptides adopt partially folded conformations and can be formed during PIP refolding, while the chance of forming the intermediate with disulfide connectivity [A20–B19, A7–A11] is much lower (Jia et al. 2003). Human Proinsulin HPI and PIP differ in the length of their connecting peptide (C-peptide), which in HPI consists of 35 residues, while in PIP is only two residues (Fig. 4.1). The refolding reaction of fully reduced/denatured HPI was carried out in the refolding buffer (100 mM Tris–HCl, 5 mM EDTA, 1 mM GSH, and 5 mM GSSG, pH 10.0) at 16°C with final protein concentration of 0.1 mg/ml (Qiao et al. 2003). By comparing the folding conditions used in HPI with that of PIP we found that l-Arg was not absolutely required as it is the predominance of GSSG in the redox buffer for the refolding of HPI, whereas the refolding of PIP requires l-Arg and GSH (Qiao et al. 2001), which suggest that the C-peptide might have significant effect on the folding kinetics. The temporal distribution of the refolding intermediates is shown in Fig. 4.3. Four disulfide intermediates (designated P1, P2, P3, and P4) were captured and characterized. The molecular mass of the four intermediates determined by electrospray ionization-mass spectrometry (ESI-MS) indicated that all of them are 9,388, that is identical to that of native HPI, which demonstrate that the intermediate P1, P2, P3, and P4 contain three intact disulfides (Table 4.1). All of the possible disulfide linkages of intermediates P2, P3, and P4 are shown in Fig. 4.4a. The CD spectra analysis indicated that P4 was more similar to the fully reduced/denatured HPI, whereas P1 and P3 were much closer to HPI, which suggested that P4 has less folded conformation than P1 and P3, two species with interchain disulfides containing a partially structured conformation. However, all the kinetic intermediates can finally refold into native HPI through disulfide rearrangement in the refolding buffer. The P2 intermediate that contains the disulfide A20–B19 is a crucial intermediate during HPI refolding because the other three intermediates need to rearrange their disulfides to form P2 during their refolding to native HPI (Qiao et al. 2003). Based on the intermediates properties, a putative in vitro disulfide-forming pathway of HPI is proposed and shown in Fig. 4.4b. During the refolding process, the fully reduced/denatured PHI can pair its free thiol groups randomly and rapidly to form in turn one-disulfide, two-disulfides until consequently native HPI, or the three-disulfides intermediates such as P1, P2, P3, and P4. One-disulfide and twodisulfide intermediates exist too short-life to be captured. P4 without interchain disulfide could convert directly or indirectly by way of P1 or P3 into P2 by disulfide rearrangement. P1 and P3 could interconvert each other, and both of them could convert into P2 before they finally refold into native HPI.
4 Folding of Single/Double Chain Insulin
69
Fig. 4.3 Temporal distribution of the intermediates of HPI trapped by trifluoroacetic acid (a) and IAA (b) (Qiao et al. 2003)
70 Fig. 4.4 Schematic representation of the possible disulfide linkages of the intermediates and putative refolding pathway of HPI. (a) Native pH 8.3 PAGE pattern of the purified intermediates: Lanes 1 and 7, native HPI and denatured HPI. Lanes 2, 3, 4, and 6, the purified intermediates P1, P2, P3, and P4, respectively. Lane 5, a mixture of P3 and P4. The additional minor bands indicated by an asterisk are deamidated products of each intermediate caused by acid during the purification; (b) Possible disulfide linkages of the HPI folding intermediates; (c) The putative folding pathway of HPI in vitro: I, II and III represent the intermediates population with one, two and three disulfides, respectively. R and N are fully reduced and native HPI, respectively. The double arrow indicates the equilibrium between each intermediates species and the equilibrium has a greater tendency for the end with the bigger arrow (Qiao et al. 2003)
Y.-M. Feng
4 Folding of Single/Double Chain Insulin
71
The quick formation of the three disulfides can protect the free thiols in the polypeptide chain from reaction during the folding process. This would prevent the protein from aggregation caused by disulfide cross-linking between polypeptide chains. Thereafter, the native disulfide bonds are regenerated by intramolecular disulfide reshuffling (Qiao et al. 2003).
4.2.1.2 Single Chain Insulin Related Proteins Amphioxus Insulin-Like Peptide Amphioxus insulin-like peptide (AILP) belongs to the insulin superfamily and is recognized as the common ancestor of insulin and IGF-1. The AILP is deduced from the cDNA sequence and displays structural characteristics of both mammalian insulin and IGF-1 (Chan et al. 1990). A recombinant single chain AILP (rAILP) was constructed, in which the deduced B- and A-domains were linked together by a tripeptide, Ala-Ala-Lys (Fig. 4.1). The rAILP was successfully expressed in yeast cells and adopted an insulin-like fold (Shen et al. 2001). The in vitro refolding reaction of fully reduced/denatured rAILP was carried out in a refolding buffer consisting of 50 mM Tris–HCl, 1 mM EDTA, and 1 mM GSSG, pH 9.5, at 16°C with final protein concentration of 1 mg/ml (Chen et al. 2004a). The temporal distribution of the refolding intermediates during the folding process is shown in Fig. 4.5a. The Fig. 4.5b shows the magnified HPLC profile of fully reduced/denatured rAILP refolding reaction stopped at the time of 15 min. There were six intermediates can be captured and designated as P1, P2, P3, P4, P5, and P6, the native pH 8.3 PAGE analysis of the six intermediates is shown in Fig. 4.5c. The disulfide linkages of the intermediates are listed in Table 4.1. CD spectra analysis indicates that the purified intermediate P1, P2, P3, P4, P5, and P6 remaining a-helix content is 22.2, 55, 52.5, 23.4, 13.4, and 1.8%, respectively, from that in the native rAILP. During the refolding process, the fully reduced/denatured rAILP acquired its three disulfides through the 1SS, 2SS, and 3SS stages (Chen et al. 2004a). The schematic flow chart of the refolding pathway of rAILP is shown in Fig. 4.6. In the refolding process of rAILP, A20–B19 disulfide is formed first (intermediate P4), followed by the formation of the second disulfide whose pairings have much more flexibility (intermediates P2, P1 or P3). The non-native disulfides can be converted to native disulfides through disulfide reshuffling (Chen et al. 2004a). Insulin-Like Growth Factor-1 IGF-1 is a 70-residue single chain globular protein composed of B-, C-, A- and D-domains from the N terminus to the C terminus and linked with three disulfides 6–48, 18–61, and 47–52 corresponding to the disulfide A7–B7, A20–B19, and
72
Y.-M. Feng
Fig. 4.5 Refolding of rAILP. (a) Temporal distribution of refolding intermediates; (b) the magnified HPLC profile of rAILP refolding at 15 min shows refolding intermediate P1–P6; (c) native pH 8.3 PAGE pattern of the purified intermediates (Chen et al. 2004a)
A6–A11 in insulin (Humbel 1990). IGF-1 and insulin belong to the same family, insulin superfamily, both share high sequence homology, the common ancestor molecule, AILP (Chan et al. 1990), and a common structural motif (Guo et al. 2008). However, their folding behaviors are significantly different: insulin folds into one unique thermodynamically controlled structure, while IGF-1 folds into two thermodynamically controlled disulfide isomers (native and swap). The native form contains disulfide 6–48, 18–61, and 47–52 with a yield 60%; the swap form contains disulfide 6–47, 18–61, and 48–52 (corresponding to the disulfide A6–B7, A20–B19, and A7–A11 in insulin) with a yield 40% (Miller et al. 1993). The native form adopts an insulin-like structure mainly including three a-helical segments (8–18, a-I; 42–49, a-II; and 54–61, a-III) in its A- and B-domains, while the a-II is unfolded in swap form. The conformation of C- and D-domains is highly flexible. So, the insulin-like structure of IGF-1 is mainly encoded by its A- and B-domains (Miller et al. 1993).
4 Folding of Single/Double Chain Insulin
73
Fig. 4.6 Schematic flow chart of major oxidative refolding and reductive unfolding pathway of rAILP. R fully reduce rAILP, N native rAILP, S swap rAILP; intermediates P1–P4 were identified during the oxidative refolding process, U1–U4 were identified during the reductive unfolding process (Chen et al. 2004a)
Recently, not only the sequence determinants of the different folding behaviors of insulin and IGF-1 have been addressed by studies on the folding behaviors of IGF-1/PIP hybrids (Guo et al. 2002; Chen et al. 2004b; Huang et al. 2007) but also the origin of their different folding mechanism has been revealed through investigation of rAILP refolding (Wang et al. 2003). The refolding of IGF-1 was generally performed in refolding buffer: 0.1 M Tris– HCl (pH 7.8) containing 1 mM GSSG and 10 mM GSH, 0.2 M KCl, and 1 mM EDTA with protein concentration of 0.1 mg/ml (Yang et al. 1999). The folding reaction was quenched at different time points by acidification to capture the folding intermediates. During the in vitro refolding of IGF-1, four main intermediates were captured, including one one-disulfide intermediate: [18–61]IGF-1, three two-disulfide intermediates: [18–61, 6–48]IGF-1, [18–61, 6–48]IGF-1, [18–61, 6–52]IGF-1 (Miller et al. 1993; Rosenfeld et al. 1997; Milner et al. 1999; Yang et al. 1999). Based on the detected intermediates, a putative folding pathway for IGF-1 has been proposed. During the folding process, the one-disulfide intermediate [18–61]IGF-1 formed first followed by the two-disulfide intermediate [18–61, 6–48]IGF-1 in
74
Y.-M. Feng
which both disulfides are native, to further complete the folding of native form [6–48, 18–61, 47–52]IGF-1; a disulfide rearrangement between two-disulfide intermediates is likely the most favorable pathway for the folding of the swap form [6–47, 18–61, 48–52]IGF-1 (Miller et al. 1993; Rosenfeld et al. 1997; Milner et al. 1999; Yang et al. 1999). Since the disulfide 18–61 (corresponding to the disulfide A20–B19 in insulin) is present in all of the captured intermediates, it should be the first disulfide to be formed during the refolding process. However, formation of the second disulfide looks random. Among the two-disulfide intermediates, the intermediate [18–61, 6–48]IGF-1 and [18–61, 6–47]IGF-1 seems the immediate precursor of the native and swap forms, respectively. Other two-disulfide intermediates need to undergo disulfide rearrangement to form the above two intermediates before they can fold into the native and swap IGF-1 species.
4.2.2 Double Chain Insulin and Related Proteins 4.2.2.1 Double Chain Insulin Mature insulin is a double chain disulfide-containing protein hormone, which consists of two polypeptide chains: A chain (21 residues) and B chain (30 residues) linked by three disulfides. The resynthesis of insulin from separated native A chain and B chain is actually the in vitro folding of insulin (also named insulin chain combination), which was reported in the early stage of 1960s (Dixon and Wardlaw 1960; Du et al. 1961). Since then, Tsou and coworkers have further investigated the mechanism of the resynthesis of insulin and demonstrated that the insulin A and B chains contain sufficient structural information to form the native molecule (Wei et al. 1992; Qian and Tsou 1987; Tang and Tsou 1990; Tang et al. 1988; Wang and Tsou 1991). Insulin molecule contains three a-helical structures locate at both Nand C terminals of A chain (A2–A8 and A13–A19, designated as a-II, a-III) and B chain center part (B9–B19, designated as a-I) and three disulfides (one intrachain disulfide, A6–A11; two interchain disulfides, A7–B7 and A20–B19) (The Peking Insulin Structure Research Group 1974; Baker et al. 1988). In the studies of insulin chain combination (in vitro refolding), Weiss and co-workers (Hua et al. 2002) have investigated the effect of a-II and a-III on insulin biological activity and chain combination. The results showed that loss of a-II strongly decreases the insulin activity but has no effect on insulin chain combination; while when a-III was disturbed by substitution of the residues located in a-III insulin chain combination was seriously blocked. Thus, the contribution of the two a-helices in A chain to insulin activity and folding are different. However, the in vitro folding process, intermediates, and folding pathway of insulin were not well characterized. Based on the studies of in vitro refolding of single chain insulins and related proteins (Qiao et al. 2001, 2003; Chen et al. 2004a), Feng and co-workers (Tang et al. 2007) have further
4 Folding of Single/Double Chain Insulin
75
investigated the in vitro folding process and intermediates of insulin and a putative in vitro folding pathway of insulin has been proposed. The folding process, intermediates, and putative folding pathway of insulin are summarized as follows. Temporal Distribution of Intermediates in the Refolding Process of Insulin The refolding reaction was initiated by directly adding the fully reduced insulin (frIns) into the refolding buffer (0.1 M glycine, pH 10.6, 8 mM oxidized glutathione (GSSG), 2.7 mM dithiothreitol (DTT)) making frIns concentration was 2 mg/ml and carried out at 0°C. The folding process of insulin and temporal distribution of the intermediates are shown in Fig. 4.7a. The reverse-phase HPLC (rpHPLC) profile indicates that there are six obvious intermediates as well as the refolded insulin (Ins) and remained frIns can be observed at 30 min (Fig. 4.7b), which are designated in order as P1A, P2B, P3A, Ins, P4B, P5B, P6B, and frIns, respectively. The peak frIns is a mixture of fully reduced A chain and B chains, which could not be separated under those conditions. Isolation, Purification, and Characterization of the Intermediates For preparation of the six intermediates, the refolding reaction was stopped at 5 min by adding the solvent A (a solvent for rpHPLC analysis which is a solution of 0.15% trifluoroacetic acid (TFA) and 10% acetonitrile) and the reaction mixture was separated by rpHPLC (Fig. 4.7b). The peaks were collected manually and lyophilized, respectively. The P1A, P2B, P3A, P4B, P5B, and P6B were further purified by rpHPLC. The native pH 8.3 PAGE patterns of the purified intermediates are shown in Fig. 4.7c. The gel was stained by Coomassie Brilliant blue R250, the stain of the P1A was too weak to be observed, but when it was reduced and modified with iodoacetic acid sodium salt (IAA), native pH 8.3 PAGE indicated that P1A corresponds to the A chain (data not shown). The molecular mass of the purified intermediates was determined by ESI-MS and listed in Table 4.2, which indicates that the P1A is A chain derivative with two intraA chain disulfides; P2B is B chain derivative contains two –GS groups; P3A is other A chain derivative with one disulfide and two –SH groups; P4B and P6B both are B chain derivatives and each contains one –SH group and one –GS group; P5B is also B chain derivative with one intra-B chain disulfide. The analysis of ESI-MS (Table 4.2) and rpHPLC (Fig. 4.7a) indicated that the refolded insulin was identical to that of native porcine insulin (N. Ins). Interestingly, there was no detectable aggregated A chain or B chain as well as double chain intermediate to be observed under the present condition. Although the disulfide linkage of P1A could not be determined, according to the reports of Chang et al. (Chang 1997; Jiang and Chang 2005) it may contain two scrambled disulfides with A6–A7, A11–A20 disulfide linkage.
76
Y.-M. Feng
a
b
P1A
0 min
P5B P6B
1 min
P4B P2B Ins
frIns
5 min
30 min
P3A
30 min
1 hr
c P1A P2B
P3A P4B
P5B P6B frIns
4 hr
16 hr
N. Ins Elution time
Fig. 4.7 Temporal distribution and native pH 8.3 PAGE pattern of the folding intermediates of insulin. (a) Temporal distribution of the major intermediates observed at different time points during the refolding process of insulin. N. Ins, represents native porcine insulin. (b) Show the major intermediates P1A, P2B, P3A, P4B, P5B, and P6B as well as Ins (refolded insulin) at 30 min. (c) Native pH 8.3 PAGE pattern of the IAA modified intermediates and frIns (Tang et al. 2007)
To further elucidate the disulfide composition, the endoproteinase V8 (V8) was used to digest P3A and P4B intermediates. The peptide fragments were purified by rpHPLC and characterized. Their molecular masses as well as disulfide connectivity of P3A and P4B are listed in Table 4.3. The results demonstrate that in P3A the disulfide formed is A20Cys with one Cys of A6, A7, or A11; so the two –SH groups should locate at two of the three Cys residues of A6, A7, and A11. In P4B, B7Cys has one –SH group, B19Cys has one –GS group. P4B and P6B are isomers but display different rpHPLC behavior, it is reasonable to deduce that in P6B the –SH group should be located at B19Cys, whereas the –GS group would be located at B7Cys.
4 Folding of Single/Double Chain Insulin
77
Table 4.2 Molecular mass of the refolding intermediates and Ins Sample Molecular mass measured Structure of intermediate a P1A 2,379.2 (2,379.7) A chain + 2 intrachain disulfides P2B 4,008.8 (4,009.9) B chain + 2 –GS groups P3A 2,381.1 (2,381.7) A chain + 1 intrachain disulfide and 2 –SH groups Ins 5,777.4 (5,777.5) Refolded porcine insulin P4B 3,703.5 (3,704.9) B chain + 1 –GS group and 1 –SH group P5B 3,396.7 (3,397.9) B chain + 1 intrachain disulfide P6B 3,703.7 (3,704.9) B chain + 1 –GS group and 1 –SH group a The figures in the brackets are theoretical data
Table 4.3 Molecular mass of V8 digestion fragments of P3A and P4B and their disulfide linkage Sample Molecular mass measured Disulfide linkage of P3A and P4B The intermediate P3A P3A-1 2,117.0 (2,117.2)a A5–17 + A18–21 + 1 disulfide and 2 –SH groups P3A-2 2,099.9 (2,099.2) A5–21 + 1 disulfide and 2 –SH groups The intermediate P4B P4B-1 1,540.7 (1,540.7) P4B-2 1,172.5 (1,172.0) P4B-3 1,086.4 (1,086.3) a The figures in the brackets are theoretical data
B1–13 + 1 –SH group B14–21 + 1 –GS group B22–30
ime-Dependent Distribution of the Intermediates T and Refolded Insulin The relationships between frdIns, intermediates, and Ins are shown in Fig. 4.8. Figure 4.8a shows the relationship between the frIns and the formation of Ins during the refolding process. The data indicate that when frIns decreased to an almost steady level, Ins started to form and then increased quickly, suggesting that Ins is formed from intermediates and not directly from frdIns. The refolding yield is increased along with the refolding time, the yield was about 30% at 16 h, and reached about 50% at 160 h, as calculated from Fig. 4.8a. Figure 4.8b shows the relationships between frdIns, Ins, and A chain intermediate P1A and P3A. The P3A was generally kept in a low level during the folding process and it reached its maximum level at 5 min but the percentage of P3A peak area was only about 5% of total area of the peaks (frIns, Ins, and all intermediates). Then P3A decreased and kept in a low level during the subsequent process. While the P1A increased quickly and dropped down slowly in the subsequent process. The different distribution change of P1A and P3A suggests that P3A may be an active A chain intermediate which directly leads the formation of the Ins; while P1A may be rather stable but it can also be slowly converted into active A chain intermediate, such as P3A to form Ins.
78
Y.-M. Feng
Fig. 4.8 Time-dependent distribution of the intermediates and refolded insulin during the refolding. (a) The relationship between frIns and Ins; (b) the relationship between frIns, Ins, and A chain intermediates; (c) the relationship between frIns, Ins, and the B chain intermediate P2B and P4B; (d) the relationship between frIns, Ins, and the B chain intermediate P5B and P6B. The percentage of intermediates, frIns, and Ins wase calculated from the corresponding peak area integration on rpHPLC (Tang et al. 2007)
Figure 4.8c, d shows the relationships between frIns, Ins, and B chain intermediates P2B, P4B, P5B, and P6B. For the intermediate P2B, both its increase and decrease were slow in the folding process. P4B appeared early and remain at high level during the early stage of the process, then it quickly decreased until “invisible” along with Ins appearance and increase (Fig. 4.8c). Therefore, it is deduced that P4B may be the most active B chain intermediate and directly forms Ins during the refolding process. The intermediate P5B is a quite stable B chain intermediate, however, it can be converted into other active B chain intermediates. The role of P6B in the formation of Ins may by similar to that of P4B but displaying a lower reactivity (Fig. 4.8d). Refolding of the Intermediates into Insulin Refolding of B Chain Intermediates with Fully Reduced A Chain
Figure 4.9a shows the reaction processes of refolding of the B chain intermediates (P2B, P4B, P5B, or P6B) with fully reduced A chain. The rpHPLC profiles indicate that when the reaction proceeded for 5 min, the peak Ins can be obviously observed in the process of P4B; the peak of Ins in the refolding process of P6B is smaller than that
4 Folding of Single/Double Chain Insulin
79
Fig. 4.9 Formation of refolded insulin during the refolding of the intermediates with corresponding counterpart. (a) Formation of the Ins during the refolding of P2B, P4B, P5B, or P6B with fully reduced A chain, respectively, in which P1, P2, P3, P4, P5, and P6 in the column P4B represent the intermediate P1A, P2B, P3A, P4B, P5B, and P6B, respectively; (b) formation of Ins during the refolding of A chain intermediate P1A or P3A with the B chain intermediate P4B, respectively (Tang et al. 2007)
of P4B; while almost no Ins speak can be observed in the refolding process of P2B and P5B. The results demonstrate that the P4B is more active than P6B in the formation of Ins during the refolding. The P2B and P5B are not active and may not directly form Ins in the refolding process. However, all B chain intermediates can fold into Ins with fully reduced A chain when the reaction proceeded for 16 h (Fig. 4.9a), which suggest that the four B chain intermediates are all on-pathway species. The results also suggest that the four intermediates can convert into each other during the refolding process. efolding of A Chain Intermediates (P1A and P3A) R with the B Chain Intermediate P4B The experiment of refolding of the B chain intermediates with fully reduced A chain has indicated that the intermediate P4B is most active in the B chain intermediates (Fig. 4.9a), so it was used as the counterpart in the refolding of P1A and P3A. Figure 4.9b shows that the Ins formed from P3A and P4B appears earlier and with higher yield than that from P1A and P4B, which indicates that the A chain intermediate P3A is more active than P1A to form Ins during the refolding. When the reaction of refolding proceeded for 16 h, both P3A and P4B almost completely disappear (Fig. 4.9b right column), while P1A still remains high level (Fig. 4.9b left column), which suggests that the intermediate P1A, like the B chain intermediate P2B and P5B, is more stable, but it can also fold into Ins with its counterpart (Fig. 4.9b left column).
80
Y.-M. Feng
Fig. 4.10 Schematic representation of the disulfide, mixture disulfide, and thiol group in the folding intermediates (Tang et al. 2007)
The Putative In Vitro Folding Pathway of Insulin he Intermediates Captured During Insulin Refolding T Are All On-Pathway The refolding of insulin was performed in a refolding system consisting in 8 mM GSSG and 2.7 mM DTT in glycine buffer, which is favorable for the disulfide formation by oxidation of thiol groups mainly using GSSG as oxidative reagent and the disulfide/thiol exchange reaction. The results show that in early stages of the refolding, the fully reduced insulin (i.e., reduced A chain and B chain) formed six major intermediates. As shown in Fig. 4.10, the intermediate P1A contains two intrachain disulfides, P3A contains one intrachain disulfide and two thiol groups, P2B contains two mixture disulfides, P5B contains one intrachain disulfide, P4B and P6B both contain one mixed disulfide and one thiol group but located in different positions. The refolding experiment of the intermediates indicates that
4 Folding of Single/Double Chain Insulin
81
Fig. 4.11 Schematic representation of the formation of transient intermediate I, II, and III and their folding to insulin (Tang et al. 2007)
among the intermediates, P1A and P5B are very stable, P3A and P4B are very active, but all of the intermediates are on-pathway since they all can finally fold into Ins with their counterpart (Fig. 4.9). However, there is no double-chain intermediate captured in this work (Tang et al. 2007). Tsou and co-worker (Wei et al. 1992) also reported that double-chain intermediate could not been observed in their work.
ormation of the Hypothetic Double-Chain Intermediates F and the Putative Folding Pathway of Insulin The basis of the formation of transient intermediates and its role in insulin in vitro folding is as follows: (1) the six intermediates captured in this work are all onpathway; (2) the separated A and B chains can recognize and interact each other (Wei et al. 1992); (3) the reactions of thiol/disulfide exchange and thiol groups oxidation to form disulfide play a central role in the disulfide-coupled protein folding (Creighton 1981, 1997; Gilbert 1990; Wedemeyer et al. 2000); (4) Xu et al. (1996) observed that during the protein folding, the disulfide pairing is not random. So Tang et al. (2007) hypothesize that in the subsequent folding process of insulin, the A and B chain intermediates should form double-chain intermediate as transient intermediate (Fig. 4.11) to finally complete the folding. First, the nonnative disulfide A20-(A6, A7, or A11) in P3A (Fig. 4.10) should be broken through thiol/disulfide exchange to form native disulfide A6–A11 and at the same time let A7Cys and A20Cys residues contain free –SH group, by which a new A chain intermediate, designated as transient I (TI), which contains native disulfide A6–A11 and two free –SH groups, is formed (Fig. 4.11a). Scheraga and co-workers (Song and Scheraga 2000) have indicated that in the formation of native structure by thiol–disulfide
82
Y.-M. Feng
exchange reactions, the protein can serve as its own redox reagent. Therefore, the transient I will further direct the folding process by quickly reacting with B chain intermediates either P4B or P6B through recognition and interaction with each other and thiol group oxidation by the GSSG present in the refolding system to form second native disulfide, by which two double-chain transient intermediates, named as transient II (TII, Fig. 4.11b) and transient III (TIII, Fig. 4.11c) were formed. Transient I reacts with P4B is faster than with P6B, since the disulfide A6–A11 together with the sequence of A6–A11 form a 20-membered cyclic structure, which make the –SH group in A7Cys of TI be more reactive (Katsoyannis and Tometsko 1966). During the folding, the formation of native structure in transient intermediates can increase the effective concentrations of the reactive thiol and disulfide groups, thus favoring the forward reaction (Welker et al. 2001). The transient II (Fig. 4.11b) should be very fast to fold into insulin and complete the folding, because of that transient II contains two native disulfide A6–A11 and A7–B7 (Fig. 4.11b), which makes it much more stable and leads the thiol/disulfide exchange reaction occurs between mixture disulfide B19Cys-GS and –SH group in A20Cys become intra-molecule reaction to easily form the third native disulfide A20–B19. The transient III contains two native disulfide A6–A11 and A20–B19 (Fig. 4.11c), which is also much more stable and leads the thiol/disulfide exchange of the mixture disulfide B7Cys-GS and –SH group in A7Cys be an intramolecular reaction to easily form the third native disulfide A7–B7 and quickly fold into insulin to complete the refolding (Fig. 4.11c). Transient I can react with both P4B and P6B. This may explain why transient I was not obtained in this work since the amount of transient I is too small to be captured. The same reason may also explain why transient II and transient III were not obtained. In both cases, their fold into insulin would be too fast to be captured as stable intermediates. Based on the intermediates and the formation of hypothetic transient forms, a two-stage folding pathway of insulin has been proposed by Tang et al. (2007) as shown in Fig. 4.12. (1) At the early stage of the folding process, the reduced A chain and B chain individually form the intermediates: two A chain intermediates (P1A and P3A), four B chain intermediates (P2B, P4B, P5B, and P6B), of which are all on-pathway. (2) In the subsequent folding process, the transient I was formed from P3A through disulfide/thiol exchange reaction; then the transient II and III, both contain two native disulfides, were formed through recognition and interaction of transient I with P4B or P6B and thiol groups oxidation reaction using GSSG as main oxidative reagent; finally through intramolecular mixture disulfide/thiol exchange reaction in transient II and III molecules to easily form the third native disulfide to complete the folding. Although intermediates P1A, P2B, and P5B are on-pathway and can finally fold into insulin, P1A must convert into P3A; P2B and P5B must convert into P4B or P6B before folding into insulin. The disulfides in P1A and P5B are intrachain non-native disulfides (Table 4.2 and Fig. 4.11), which had also been observed in the folding intermediates of HPI (Qiao et al. 2003) and PIP (Qiao et al. 2001) refolding processes. The similarity between the intermediates observed in single chain insulin, HPI and PIP, and
4 Folding of Single/Double Chain Insulin
83
Fig. 4.12 Schematic representation of the putative in vitro folding pathway of insulin (Tang et al. 2007)
double chain insulin may not only suggest that the double chain and single chain insulins share similar folding pathway, but also indicate that the formation of intrachain disulfides in folding early stage may play an important role in preventing the formation of intermolecule disulfides, which may lead to subsequent protein aggregation. 4.2.2.2 Double Chain Insulin-Related Proteins Two double chain insulin-related proteins, double chain IGF-1 and double chain rAILP, are available. They were obtained from recombinant single chain mini-IGF-1 and rAILP, respectively, after removing their connecting peptides using endoproteinase Lys-C (Guo et al. 2002; Shen et al. 2001). Although the in vitro refolding of double chain IGF-1 and double chain rAILP is not completely characterized yet, the analysis of circular dichroism and receptor-binding has shown that both double chain proteins have identical disulfide pairing to that of the double chain insulin (Guo et al. 2002; Shen et al. 2001). The investigation of the in vitro folding of double chain IGF-1 and rAILP, and the comparison with that of their single chain forms will allow to gain more information about the folding mechanism of double chain proteins linked by disulfide(s).
4.3 Concluding Remarks 4.3.1 The Common Folding Behaviors of the Single Chain Insulin and Related Protein Based on the studies of single chain insulins and related proteins in vitro folding, it is suggested that the single chain insulins and related proteins share some common folding behaviors as follows.
84
Y.-M. Feng
1. The disulfide A20–B19 is the most important and formed at the refolding initial stage. In the refolding pathway of PIP, both of the 2SS intermediates contain the native disulfide A20–B19 (Qiao et al. 2001). The most important intermediate P2 of HPI also contains the disulfide A20–B19 (Qiao et al. 2003). During refolding of rAILP, an essential 1SS intermediate containing A20–B19 has been identified (Chen et al. 2004a). The disulfide 18–61 (corresponding to the disulfide A20–B19 in insulin) is present in all of the captured intermediates in folding process of IGF-1 (Miller et al. 1993; Rosenfeld et al. 1997; Milner et al. 1999; Yang et al. 1999). For PIP, the peptide model with only disulfide A20–B19, [A20–B19]PIP, still retained partially folded native conformation, and could refold efficiently in vitro (Yan et al. 2003). All of the above observations strongly support that the disulfide A20–B19 plays a critical and conservative role in the initial stage of folding of the single chain insulins and related proteins. 2. The second disulfide pairing is flexible and can be rearranged from non-native to native. In all of the refolding pathways mentioned above (Qiao et al. 2001, 2003; Chen et al. 2004a; Yang et al. 1999), the formation of the second disulfide, based on 1SS intermediate, which contains disulfide A20–B19, seems random. All of the identified intermediates during the oxidative refolding process of PIP, HPI, and rAILP are listed in Table 4.1. The non-native disulfide in the 2SS species can be finally converted into native one (or swap disulfide in rAILP and IGF-1) by disulfide rearrangement, which is an intramolecular reaction and can minimize the chance of aggregation. To get more details of pairing property of the second disulfide, four model peptides of possible folding intermediates with two disulfides, [A20–B19, A7–B7]PIP, [A20–B19, A6–B7]PIP, [A20–B19, A6–A11]PIP, and [A20–B19, A7–A11]PIP, were prepared and their properties were analyzed, which indicated that the four model peptides all adopt partially folded structure with moderate conformational differences (Jia et al. 2003). 3. The a-I, located in the B-chain/domain center (B9Ser–B19Cys) which is close to the disulfide A20–B19, is probably the folding initiation site. The a-I is the longest helical structure in the single chain insulins and related proteins, and it is also the most robustly ordered structure. When the other two a-helical segments (a-II and a-III) located at the N- and C terminal of A-chain/domsin were disturbed by disulfide deletion or disulfide miss-pair, the a-I still remained intact. At the initial stage of refolding, the a-I might be rapidly formed because of its high helical propensity, and subsequently stabilized by formation of the disulfide A20–B19 (Yan et al. 2003; Jia et al. 2003; Guo and Feng 2001; Guo et al. 2004). Dupradeau et al. have reported that a B-chain mutant, [B7Ser, B19Ser]B-chain, can form an a-helix between B9–B19 and a b-turn between B20–B23, which indicates that even B-chain alone can partially fold into structures similar to that of a-I in the insulin crystal structure (Dupradeau et al. 2002).
4 Folding of Single/Double Chain Insulin
85
4.3.2 The Folding Information of Single/Double Chain Insulins and Related Proteins is Mainly Encoded in the Insulin Structural Motif The investigations of in vitro folding of single/double chain insulins and related proteins mentioned above showed that either single chain insulins and related proteins or double chain insulin can finally fold into identical or very similar insulin three-dimensional structures, although following different folding pathways, which suggested that the folding information of the single/double chain insulins and related proteins is encoded in a similar sequential context. Single single/double chain insulins and the related protein IGF-1 and AILP belong to insulin superfamily. All of the superfamily members share a common characteristic amino acid sequence, CGxxxxxxxxxxCnCCxxxCxxxxxxxxC, which is named as insulin structural motif (Guo et al. 2008). Independently, if the mature proteins of the superfamily members are double chain forms (such as insulin and relaxin) or single chain forms (such as IGF-1), they are all synthesized in vivo as single chain precursors and contain the insulin structural motif. The structural motif consists of six fully conserved Cys residues located at A6, A7, A11, A20, B7, and B19, and a Gly residue located at B8, and three constant regions: 10× residues, 3× residues, and 8× residues between the residue B8 and B19, A7 and A11, A11 and A20, respectively, as well as a variable connect peptide “n.” The six absolutely conserved Cys residues form three disulfides in the folded proteins: two interchain/domain disulfides (A7–B7 and A20–B19) and one intra-A chain/domain disulfide (A6–A11). The three disulfides are crucial for the native structure and biological function of the superfamily members. The a-I and a-III located at the constant residue number 10× and 8×, respectively. As mentioned above, a-I is the longest helical segment and most robustly ordered structure in the single/double chain insulins and related proteins and play an important role in the folding. The report by Hua et al. (2002) indicated that loss of a-II does not effect on insulin chain combination; while a-III was disturbed that causes insulin chain combination seriously blocked. Therefore, we suggest that the folding information of single/double chain insulins and related proteins is mainly encoded in the insulin structural motif.
4.3.3 Possible Role of C-Peptide: Implications for In Vivo Folding of Insulin Inspection of the C-peptide of single chain/double chain insulins and related proteins reveals that its length is quiet variable (Fig. 4.1): for HPI the C-peptide consists of 35 amino acids (Steiner 1967), for PIP the “C-peptide” is the dipeptide, Ala-Lys (Markussen et al. 1987; Zhang et al. 1996); for AILP is 31 amino acids (Chan et al. 1990); for rAILP is the tripeptide, Ala-Ala-Lys (Shen et al. 2001);
86
Y.-M. Feng
Fig. 4.13 Schematic representation of in vivo folding processes of nascent proinsulin. (I) The B chain was synthesized first and formed an intra-B chain disulfide, B7–B19, to prevent from aggregation at very early stage of proinsulin biosynthesis. (II) Formation of intra-A disulfides, like scramble disulfide, as the intermediates captured the in vitro folding process of HPI and insulin. (III) Formation and accumulation of secondary structure, especially the central a-helix and b-turn of B chain, which made the B chain act as a template for subsequent folding of the A chain, and initiation of disulfide rearrangement through thiol/disulfide exchange. (IV) Formation of the transient state (molten globule) and rearrangement of the conformation, which led complete the folding. (V) The well folded proinsulin
for IGF-1 is 12 amino acid (Humbel 1990); for insulin we may say that the “C-peptide” is zero or infinite length. Besides, Derewenda et al. have reported a mini-proinsulin, in which B29 is directly linked with A1 while having identical crystal structure to that of insulin. The C-peptide length of the mini-proinsulin is negative one amino acid (Derewenda et al. 1991). The above reports demonstrated that although the “C-peptide” is quiet different, single/double chain insulins and related proteins can fold into their native structure. Therefore, it is suggested that in the folding process, C-peptide seems to act just as a flexible tie. On the one hand, C-peptide provides a flexible environment for A and B chains to fold freely; on the other hand, the A and B chains are tied together by C-peptide, which makes the folding highly efficiency. From the above described data on the in vitro folding of single chain insulins in redox conditions similar to those in the endoplasmic reticulum (Hwang et al. 1992; Frand et al. 2000), we may deduce that the in vivo folding process of nascent proinsulin might proceed as follows: (1) we know well the order of the gene and in vivo synthesis of proinsulin is from the B-chain to C-peptide to A-chain, so the B chain is synthesized first and forms an intra-B chain disulfide, B7–B19, like intermediate P4 captured in the refolding process of HPI (Qiao et al. 2003), thus preventing aggregation by formation of intermolecule disulfide(s); (2) when the A chain is synthesized, two intra-A chain disulfides A6–A7 and A11–A20, like the scrambled disulfides (Jiang and Chang 2005), observed in the refolding process of HPI (Qiao et al. 2003) and insulin (Tang et al. 2007), are formed before disulfide rearrangement to prevent again the formation of intermolecule
4 Folding of Single/Double Chain Insulin
87
disulfide(s); (3) the central a-helix (a-I) is formed, which may lead the intra-B chain disulfide B7–B19 to be broken and release two –SH groups, meanwhile the b-turn (B20–B23) is formed, making the B chain C-terminal bend and approximates the A chain N-terminal region, leading the –SH groups of B7Cys and B19Cys close to the intra-A chain non-native disulfides A6–A7 and A11–A20; (4) through thiol/disulfide exchange and oxidation the native disulfides A20–B19, A7–B7 and A6–A11 are formed to complete the in vivo folding of proinsulin. The schematic disulfide-forming pathway and folding processes of nascent proinsulin are shown in Fig. 4.13.
References Baker EN, Blundell TL, Cutfield JF et al (1988) The structure of 2Zn pig insulin crystals at 1.5Å resolution. Philos Trans R Soc Lond B Biol Sci 319:369–456 Banting FG, Best CH (1921) The internal secretion of the pancreas. J Lab Clin Med 7:251–266 Chan SJ, Cao QP, Steiner DF (1990) Evolution of the insulin superfamily: cloning of a hybrid insulin/ insulin-like growth factor cDNA from amphioxus. Proc Natl Acad Sci USA 87:9319–9323 Chang JY (1997) A two-stage mechanism for the reductive unfolding of disulfide-containing proteins. J Biol Chem 272:69–75 Chen Y, Jin R, Dong HY et al (2004a) In vitro refolding/unfolding pathways of amphioxus insulinlike peptide: implications for folding behavior of insulin family proteins. J Biol Chem 279:55224–55233 Chen Y, You Y, Jin R et al (2004b) Sequences of B-chain/domain 1-10/1-9 of insulin and insulinlike growth factor 1 determine their different folding behavior. Biochemistry 43:9225–9233 Creighton TE (1981) Accessibilities and reactivities of cysteine thiols during refolding of reduced bovine pancreatic trypsin inhibitor. J Mol Biol 151:211–213 Creighton TE (1997) Protein folding coupled to disulfide bond formation. J Biol Chem 378:731–744 Derewenda U, Derewenda Z, Dodson EJ et al (1991) X-ray analysis of the single chain B29-A1 peptide-linked insulin molecule. A completely inactive analogue. J Mol Biol 220:425–433 Dixon GH, Wardlaw AC (1960) Regeneration of insulin activity from the separated and inactive A and B chains. Nature 188:721–724 Du YC, Zhang YS, Lu ZX et al (1961) Resynthesis of insulin from its glycyl and phenylalanyl chains. Sci Sin 10:84–104 Dupradeau FY, Richard T, Le Flem G et al (2002) A new B-chain mutant of insulin: comparison with the insulin crystal structure and role of sulfonate group in the B-chain structure. J Pept Res 60:56–64 Frand AR, Cuozzo JW, Kaiser CA (2000) Pathways for protein disulfide bond formation. Trends Cell Biol 10:203–210 Gilbert HF (1990) Molecular and cellular aspects of thiol-disulfide exchange. Adv Enzymol Relat Areas Mol Biol 63:69–172 Guo ZY, Feng YM (2001) Effects of cysteine to serine substitutions in the two inter-chain disulfide bonds of insulin. Biol Chem 382:443–448 Guo ZY, Jia XY, Feng YM (2004) Replacement of the interchain disulfide bridge-forming amino acids A7 and B7 by glutamate impairs the structure and activity of insulin. Biol Chem 385:1171–1175 Guo ZY, Qiao ZS, Feng YM (2008) The in vitro oxidative folding of the insulin superfamily. Antioxid Redox Signal 10:127–139 Guo ZY, Shen L, Feng YM (2002) The different folding behavior of insulin and insulin-like growth factor 1 is mainly controlled by their B-chain/domain. Biochemistry 41:1556–1567
88
Y.-M. Feng
Hua QX, Chu YC, Jia W et al (2002) Mechanism of insulin chain combination. Asymmetric roles of A-chain a-helices in disulfide pairing. J Biol Chem 277:43443–43453 Huang QL, Zhao J, Tang YH et al (2007) The sequence determinant causing different folding behaviors of insulin and insulin-like growth factor-1. Biochemistry 46:218–224 Humbel RE (1990) Insulin-like growth factor I and II. Eur J Biochem 190:445–462 Hwang C, Sinskey AJ, Lodish HF (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257:1496–1502 Jia XY, Guo ZY, Wang Y et al (2003) Peptide models of four possible insulin folding intermediates with two disulfides. Protein Sci 12:2412–2419 Jiang C, Chang JY (2005) Unfolding and breakdown of insulin in the presence of endogenous thiols. FEBS Lett 579:3927–3931 Katsoyannis PG, Tometsko A (1966) Insulin synthesis by recombination of A and B chains: a highly efficient method. Proc Natl Acad Sci USA 55:1554–1561 Markussen J, Hougaard P, Ribel U et al (1987) Soluble, prolonged-acting insulin derivatives. I. Degree of protraction and crystallizability of insulins substituted in the termini of the B-chain. Protein Eng 1:205–213 Miller JA, Narhi LO, Hua QX et al (1993) Oxidative refolding of insulin-like growth factor 1 yields two products of similar thermodynamic stability: a bifurcating protein-folding pathway. Biochemistry 32:5203–5213 Milner SJ, Carver JA, Ballard FJ et al (1999) Probing the disulfide folding pathway of insulin-like growth factor-I. Biotechnol Bioeng 62:693–703 The Peking Insulin Structure Research Group (1974) Studies on the insulin crystal structure: the molecule at 1.8Å resolution. Sci Sin 17:752–778 Qian YQ, Tsou CL (1987) Resynthesis of insulin from its A and B chains in the presence of denaturants. Biochem Biophys Res Commun 146:437–442 Qiao ZS, Gou ZY, Feng YM (2001) Putative disulfide-forming pathway of porcine insulin precursor during its refolding in vitro. Biochemistry 40:2662–2668 Qiao ZS, Min CY, Hua QX et al (2003) In vitro refolding of human proinsulin. Kinetic intermediates, putative disulfide-formation pathway, folding initiation site, and potential role of C-peptide in folding process. J Biol Chem 278:17800–17809 Rosenfeld RD, Miller JA, Narhi LO et al (1997) Putative folding pathway of insulin-like growth factor-1. Arch Biochem Biophys 342:298–305 Shen L, Guo ZY, Chen Y et al (2001) Expression, purification, characterization of amphioxus insulin-like peptide and preparation of polyclonal antibody to it. Sheng Wu Hua Xue Yu Sheng Wu Wu Li Xue Bao (Shanghai) 33:629–633 Song MC, Scheraga HA (2000) Formation of native structure by intermolecular thiol-disulfide exchange reaction without oxidants in the folding of bovine pancreatic ribonuclease A. FEBS Lett 471:177–181 Steiner DF (1967) Evidence for a precursor in the biosynthesis of insulin. Trans N Y Acad Sci 30(1):60–68 Tang JG, Tsou CL (1990) The insulin A and B chains contain structural information for the formation of the native molecule. Studies with protein disulphide-isomerase. Biochem J 268:429–435 Tang JG, Wang CC, Tsou CL (1988) Formation of native insulin from the scrambled molecule by protein disulphide-isomerase. Biochem J 255:451–455 Tang Y, Wang S, Chen Y et al (2007) In vitro insulin refolding: characterization of the intermediates and the putative folding pathway. Sci China C Life Sci 50:717–725 Wang CC, Tsou CL (1991) The insulin A and B chains contain sufficient structural information to form the native molecule. Trends Biochem Sci 16:279–281 Wang S, Guo ZY, Shen L et al (2003) Refolding of amphioxus insulin-like peptide: implications of a bifurcating evolution of the different folding behavior of insulin and insulin-like growth factor 1. Biochemistry 42:9687–9693 Wedemeyer WJ, Welker E, Narayan M et al (2000) Disulfide bonds and protein folding. Biochemistry 39:4207–4216
4 Folding of Single/Double Chain Insulin
89
Wei J, Xie L, Lin YZ et al (1992) The pairing of the separated A and B chains of insulin and its derivatives. FTIR studies. Biochim Biophys Acta 1120:69–74 Welker E, Wedemeyer WJ, Narayan M et al (2001) Coupling of conformational folding and disulfide-bond reactions in oxidative folding of proteins. Biochemistry 40:9059–9064 Xu X, Rothwart DM, Scheraga HA (1996) Nonrandom distribution of the one-disulfide-intermediates in the regeneration of ribonuclease A. Biochemistry 35:6406–6417 Yan H, Guo ZY, Gong XW et al (2003) A peptide model of insulin folding intermediate with on disulfide. Protein Sci 12:768–775 Yang Y, Wu J, Watson JT (1999) Probing the folding pathways of long R(3) insulin-like growth factor-I (LR(3)IGF-I) and IGF-I via capture and identification of disulfide intermediates by cyanylation methodology and mass spectrometry. J Biol Chem 274:37598–37604 Zhang Y, Hu H, Cai R et al (1996) Secretory expression of a single-chain insulin precursor in yeast and its conversion into human insulin. Sci China C Life Sci 39:225–233
Chapter 5
Unfolding and Refolding of Disulfide Proteins Using the Method Disulfide Scrambling Rowen J.Y. Chang
Abstract Unfolding and refolding of disulfide proteins can be investigated by the method of disulfide scrambling which is based on the reversible conversion between the native (N) and scrambled isomers (X). The method of disulfide scrambling presents a number of unique features in elucidation of pathways of protein unfolding and refolding. (a) It allows trapping and isolation of diverse intermediates (unfolding and folding) for further structural analysis. (b) It demonstrates that protein denaturation and unfolding can be quantified independently. Denaturation is calculated by the conversion of native (N) to non-native X-isomers. Unfolding is measured by the progressive unfolding of X-isomers. (c) It shows that folding experiment can be initiated with a structurally defined X-isomer possessing the highest free energy among all unfolded X-isomers. (d) It reveals that the energy landscape of conformational heterogeneity (unfolding and refolding) can be illustrated by a diamond-shaped model. At two extreme ends of the energy landscape, the conformational heterogeneity is reduced to minimum. Keywords Method of disulfide scrambling • Isomers of unfolded protein • Unfolding by disulfide scrambling • Folding by disulfide scrambling • Scrambled proteins
Abbreviations a-LA HPLC BPTI TAP
a-Lactalbumin High-pressure liquid chromatography Bovine pancreatic trypsin inhibitor Tick anticoagulant peptide
R.J.Y. Chang (*) Research Center for Protein Chemistry, Brown Foundation Institute of Molecular Medicine, 1825 Pressler Street, Houston, TX 77030, USA Department of Biochemistry and Molecular Biology, The University of Texas, Houston, TX 77030, USA e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_5, © Springer Science+Business Media, LLC 2011
91
92
R.J.Y. Chang
5.1 Introduction Folding mechanism and pathway of disulfide protein can be investigated by various techniques. The most commonly applied method is the conventional “disulfide intact” approach, in which disulfide bonds of native proteins remain intact throughout the process of unfolding and refolding. In this approach, native proteins are first unfolded by denaturant, extreme temperature, or pH. Folding experiment is then initiated by removal (or dilution) of denaturant, temperature jump, or pH adjustment. The mechanism of protein refolding is then monitored by the restoration of selective physicochemical signals that distinguishes the native and unfolded states. The most commonly used signals are spectra of fluorescence, circular dichroism, infrared, and NMR coupled with amide proton exchange (Englander and Mayne 1992; Shortle 1996; Dyson and Wright 1996, 2004). Folding mechanisms of many native disulfide proteins have been studied by this conventional “disulfide intact” approach. Among them, hen egg lysozyme represents one of the most extensively investigated models (Matagne and Dobson 1998). Lysozyme was shown to fold via multiple pathways with detectable intermediates (Radford et al. 1992). Specifically, GdmCl (6 M) unfolded lysozyme was shown to refold via kinetically partitioned pathways (Kiefhaber 1995). About 80% of the unfolded lysozyme molecules refold on a slow pathway with well-populated intermediates. The remaining 20% of denatured lysozyme refold on a fast track without detectable intermediate. The major limitation of this conventional “disulfide intact” approach is that it does not permit, in most cases, isolation of folding intermediates. To date, the majority of our knowledge about protein folding has been acquired by this approach. A second distinct method for analyzing the folding mechanism of disulfide protein is “disulfide oxidation” (commonly known as protein oxidative folding) (Creighton 1986; Scheraga et al. 2001; Arolas et al. 2006). In this approach, a protein is first fully reduced and denatured by reducing agent (e.g., dithiothreitol) and denaturant (e.g., 6 M GdmCl). The folding experiment is initiated by exclusion of reducing agent and denaturant and is allowed to proceed via disulfide oxidation (formation) in a buffer containing redox agent. A fully reduced 3-disulfide protein may undergo 15 one-disulfide, 45 two-disulfide, and 14 three-disulfide isomers as possible folding intermediates. These disulfide isomers are typically trapped by sample acidification and isolated by HPLC. The disulfide folding pathway is then characterized by the heterogeneity, structure, and kinetic property of 74 possible disulfide isomers that accumulate along pathway leading to the native protein. This technique, pioneered by Creighton (Creighton 1986, 1992), has been applied to the extensive elucidation of folding pathways of numerous disulfide containing proteins, including among many others (see Chap. 1), bovine pancreatic trypsin inhibitor (BPTI) (Creighton 1990; Goldenberg 1992; Weissman and Kim 1991), ribonuclease A (Rothwarf et al. 1998; Welker et al. 2001), hirudin (Chatrenet and Chang 1993; Chang 1994). A third method for studying the mechanism of protein folding is designated as “disulfide scrambling” (Chang 1999a; Chang and Li 2001; Chang 2002). The method of disulfide scrambling can be considered as a hybrid of “disulfide intact”
5 Protein Folding via Disulfide Scrambling
93
YNRL C I K PRDWI DECDSNEGGERAYFRNGKGGCDSFWI CPEDHTGADYYSSYRDCFNAC I
5
15
33
39
55 59
Native protein
YNRLCIKPRDWI DECDN S EGGR E AYFRNGKGGD C SFWI CPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCIKPRDWI DECDN S EGGR E AYFRNGKGGD C SFWI CPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCIKPRDWI DECDN S EGGR E AYFRNGKGGD C SFWI CPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
Unfolding
Refolding
Denaturant+ thiol catalyst
Thiol catalyst
YNRLCIKPRDWIDECDN S EGGR E AYFRNGKGGDF C S WICPEDHTGADYYSSYRDF C NACI
5
15
33 39
55 59
YNRLCIKPRDWIDECDN S EGGR E AYFRNGKGGDF C S WICPEDHTGADYYSSYRDF C NACI
5
15
33 39
55 59
YNRLCIKPRDWIDECDN S EGGR E AYFRNGKGGDF C S WICPEDHTGADYYSSYRDF C NACI
5
15
33 39
55 59
YNRLCIKPRDWIDECDN S EGGR E AYFRNGKGGD C SFWICPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCIKPRDWIDECDN S EGGR E AYFRNGKGGD C SFWICPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCIKPRDWIDECDN S EGGR E AYFRNGKGGD C SFWICPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCI KPRDWI DECDSNEGGR E AYFRNGKGGCDSFWI CPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCI KPRDWI DECDSNEGGR E AYFRNGKGGCDSFWI CPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCI KPRDWI DECDSNEGGR E AYFRNGKGGCDSFWI CPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCI KPRDWI DECDN S EGGR E AYFRNGKGGD C SFWICPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCI KPRDWI DECDN S EGGR E AYFRNGKGGD C SFWICPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
YNRLCI KPRDWI DECDN S EGGR E AYFRNGKGGD C SFWICPEDHTGADYYSSYRDCFNACI
5
15
33 39
55 59
14 unfolded X-isomers Fig. 5.1 The method of disulfide scrambling for reversible unfolding and refolding between the native protein and unfolded X-isomers. A 3-disulfide protein is presented here as an example. For a 4-disulfide protein, unfolding by disulfide scrambling may generate a maximum of 104 X-isomers
and “disulfide oxidation” techniques. In this method, a native disulfide protein is allowed to unfold in the alkaline buffer (~pH 8.0) containing denaturant and thiol catalyst (e.g. 0.2 mM b-mercaptoethanol or 1 mM cysteine). Under these conditions, the native protein unfolds by shuffling its native disulfide bonds and converts to a mixture of fully oxidized scrambled isomers (X-isomers). For instance, an unfolded 3-disulfide protein may comprise 14 possible X-isomers (Fig. 5.1). All X-isomers can be isolated by HPLC and allowed to refold to form the native protein via disulfide scrambling. A diagram illustrating the reversible conversion between the native protein and unfolded X-isomers is presented in Fig. 5.1. The method of disulfide scrambling provides unique advantages in analyzing unfolding and folding of disulfide proteins. For the process of unfolding of native protein (1) it permits isolation and structural characterization of diverse X-isomers of unfolded proteins and (2) it permits distinction between protein denaturation and protein unfolding. For the process of refolding of X-isomers (1) folding can be initiated with a structurally defined X-isomer that possesses the highest free energy among all unfolded X-isomers and (2) there are defined numbers of folding intermediates which adopt varied extent of unfolding. A 3-disulfide and 4-disulfide proteins have 13 and 103 X-isomers that may serve as folding intermediates between the starting X-isomer and the ending N-protein. Each intermediate can be isolated and thoroughly characterized for its structural and kinetic property. This permits identification of on-pathway and off-pathway intermediates, kinetic traps, energy compartments, etc. These unique features will be illustrated in the following sections.
94
R.J.Y. Chang
5.2 Disulfide Scrambling Represents a Major Activity in the Pathway of Oxidative Folding of Disulfide Proteins Development of the method of disulfide scrambling has been originated from practicing the method of disulfide oxidation (oxidative folding). The pathway and mechanism of protein oxidative folding exhibit enormous diversity (see Chap. 1), which is most evident among widely investigated 3-disulfide proteins. In the well-established BPTI model, folding intermediates were shown to consist of only selected number of 1-disulfide and 2-disulfide isomers (Creighton 1990; Goldenberg 1992; Weissman and Kim 1991). However, we have found that folding intermediates of many 3-disulfide proteins comprise 3-disulfide X-isomers in addition to 1- and 2-disulfide isomers. This is best exemplified by the case of hirudin (Chatrenet and Chang 1993; Chang 1994) (Fig. 5.2). Oxidative folding of fully reduced hirudin (R) is characterized by a sequential flow of intermediates via heterogeneous 1-disulfide isomers (I), 2-disulfide isomers (II), and 3-disulfide X-isomers (III or X) to reach the native structure. The final step of folding, conversion of 3-disulfide X-isomers to the native hirudin, requires the presence of thiol catalyst. When folding was carried out in the buffer alone (Fig. 5.2a),
Fig. 5.2 (a) Oxidative folding of hirudin in the buffer without redox agent. (b) Oxidative folding of hirudin in the buffer containing GSH (0.5 mM). Folding intermediates were trapped by sample acidification and analyzed by reversed phase HPLC. “R” and “N” stand for fully reduced and native hirudin, respectively. Folding intermediates comprise heterogeneous 1-disulfide isomers (I), 2-disulfide isomers (II), and eleven 3-disulfide isomers (III/X) (Chang 1994)
5 Protein Folding via Disulfide Scrambling
95
disulfide formation was promoted by air oxidation and conversion of fully oxidized X-isomer to native hirudin was catalyzed by thiols of residual cysteines of partially oxidized intermediates (I + II). As folding progresses, thiols of cysteines are gradually depleted and X-isomers thus become trapped, unable to convert to the native hirudin. The inclusion of thiol catalyst (e.g. 0.5 mM GSH) in the folding buffer facilitates quantitative recovery of native hirudin (Fig. 5.2b). Numerous disulfide proteins fold via mechanism similar to that of hirudin. They include potato carboxypeptidase inhibitor (Chang et al. 1994), leech carboxypeptidase inhibitor (LCI) (Salamanca et al. 2003), tick anticoagulant peptide (Chang 1996), Ascaris carboxypeptidase inhibitor (Arolas et al. 2009), human proinsulin (Qiao et al. 2003), LA5 module of LDL receptor (Arias-Moreno et al. 2008), cardiotoxin-III (Chang et al. 2006), and Amaranthus alpha-amylase inhibitor (Cemazar et al. 2003, 2004). For these proteins, rapid accumulation of fully oxidized X-isomers occurs when folding was performed in the presence of oxidizing agents (GSSG or cystine) (Chang 1994) and conversion of X-isomers to form the native protein by disulfide scrambling become a rate-limiting step of folding. Extensive studies of the final stage of protein oxidative folding in our laboratory (see 24 h sample of Fig. 5.2a) led to the development of the method of disulfide scrambling for analyzing reversible conversion between N-protein and X-isomers (Chang 1997, 1999a, 2002; Chang and Li 2001).
5.3 Unfolding of Disulfide Proteins via Disulfide Scrambling Unfolding is an essential step to generate starting material for folding experiment. This is typically carried out by incubating a native disulfide protein in a mild alkaline buffer (pH 8.0–8.5) containing a thiol catalyst and increasing concentrations of a selected denaturant. The denaturant can be urea, GdmCl, GdmSCN, or organic solvent. The thiol catalyst can be cysteine (~0.1–2.0 mM), GSH (~0.2–2.0 mM), or b-mercaptoethanol (~0.02–0.2 mM). Under these conditions, the native protein (N-protein) unfolds by shuffling its disulfide bonds and converts to a mixture of fully oxidized X-isomers (Fig. 5.1). At 22°C, the unfolding reaction usually reaches equilibrium within 0.5–2 h. N-Protein and X-isomers are fractionated and quantified by HPLC. The results can be used to distinguish and quantify the extent of denaturation and the extent of unfolding separately. (1) The extent of denaturation is determined by the recovery of X-isomers as % of total protein, i.e., X/(N + X). This is based on the definition that any structure that is not native is considered as denatured. (2) The extent of unfolding is determined by the recovery of the most unfolded X-isomer (one with the highest associated free energy) as % of total protein, i.e., Xa/(N + X) in which Xa is the most unfolded X-isomer. This is based on the notion that denatured X-isomers may adopt varied extent of unfolding. The unfolding of tick anticoagulant peptide (TAP) and BPTI illustrates the potential of the method to discriminate between denatured and unfolded states. TAP and BPTI are both kunitz-type protease inhibitors. They have almost identical size (58 a.a. vs. 60 a.a.). They also share modest sequence homology, the same disulfide
96
R.J.Y. Chang
Fig. 5.3 (a) Denaturation and unfolding of TAP using the method of disulfide scrambling. Native TAP was incubated in the Tris–HCl buffer (0.1 M, pH 8.4) containing 0.2 mM of b-mercaptoethanol and increasing concentrations (1–6 M) of urea, GdmCl, and GdmSCN. The reactions were carried out at 23°C for 16 h to allow the reaction to reach the equilibrium. Unfolded samples were acidified and analyzed directly by HPLC (Chang 1999a). Seven major X-TAP isomers (a–g) were identified as unfolding intermediates. Their disulfide structures are given in Fig. 5.4. (b) Denaturation curves of TAP. The denaturation curves are determined by the fractions of TAP converted to X-TAP. (c) Denaturation and unfolding of BPTI by increasing concentrations of GdmSCN. Seven fractions of X-BPTI isomers were identified. Among them, the most predominant X-BPTI-a adopts beads-form disulfide configuration. (d) Denaturation curves of BPTI. Note that BPTI remains intact even at 8 M urea
pattern (3 disulfide bonds), similar content of secondary structures, and nearly superimposable 3D structures (Antuch et al. 1994; Lim-Wilby et al. 1995). These two proteins were allowed to denature and unfold in the presence of increasing concentrations of urea, GdmCl, and GdmSCN (Chang 1999a; Chang and Ballatore 2000). The results are elaborated and summarized in the following (Fig. 5.3): (1) Unfolded TAP and BPTI each comprise about seven X-isomers (50% of possible
5 Protein Folding via Disulfide Scrambling
97
Fig. 5.4 Folding of X-TAP-a and X-BPTI-a using the method of disulfide scrambling (left) X-TAP-a and X-BPTI-a represent two most extensively unfolded isomers of denatured TAP (Chang 1999a) and BPTI (Chang and Ballatore 2000), respectively. They were allowed to refold at 23°C in the Tris–HCl buffer (pH 8.4) containing cysteine, 0.5 mM for X-TAP-a, and 0.1 mM for X-BPTI-a, to form their native conformation. Intermediates of folding were trapped by acidification and analyzed by HPLC using the conditions described in (Chang and Li 2005). “N” indicates the elution position of the native protein. (Right) Disulfide structures of isomers of 10 X-TAP and 8 X-BPTI isomers
isomers). The disulfide structures of seven X-TAP isomers (a–g) (Fig. 5.3a) and the most predominant X-BPTI isomer (X-BPTI-a) (Fig. 5.3b) are presented in Fig. 5.4. (2) The denaturation curves of TAP and BPTI, calculated from X/(N + X), showed that midpoint denaturation of [GdmCl]1/2 at 4.2 and 7.5 M, respectively (Fig. 5.3b and d). It is interesting to notice that BPTI remains totally intact at 8 M urea, a usual and remarkable stability shared only with few other proteins. (3) The structure (composition) of unfolded X-isomers continues to evolve as the strength of denaturant increases. There is only one X-isomer each in both unfolded TAP and BPTI whose recovery (as % of total protein) rises constantly as the strength of denaturant increases. These two X-isomers, represented by X-TAP-a and X-BPTI-a, adopt beads-form disulfide pattern with three disulfide bonds bridged by consecutive cysteines, a configuration consistent with the extensively unfolded state. They are considered as the X-isomer with highest free energy and their recoveries can be used to plot the unfolding curves (Chang 2009). Isolated X-TAP-a and X-BPTI-a are also ideal isomers as starting materials of folding experiments using the method of disulfide scrambling (see following section).
98
R.J.Y. Chang
In addition to TAP and BPTI, the technique of disulfide scrambling has been applied to study denaturation/unfolding of numerous disulfide proteins. Many of them exhibit unfolding properties similar to that of TAP. These proteins include insulin-like growth factor (Chang et al. 1999), lysozyme (Chang and Li 2002), a-lactalbumin (Chang and Li 2001), potato carboxypeptidase inhibitor (Chang et al. 2000), LCI (Salamanca et al. 2002), lipid transfer proteins (Lin et al. 2004), complement protein C3a (Chang et al. 2008), phospolipase A2 (Singh and Chang 2004), cardiotoxin (Chang et al. 1998), and human proinsulin (Min et al. 2004). It is important to mention that conformational stability of disulfide proteins measured by the “disulfide scrambling” technique is typically lower than that obtained by the conventional “disulfide intact” method using CD or fluorescence as signals (Pace 1986). This should not be surprising because intact disulfides increase the conformational stability mainly by constraining the unfolded conformations of the protein and thereby decreasing their conformational entropy. With disulfide scrambling method, the native disulfide framework was teased to break and reform according to the conformational changes induced by the denaturant. Once the rigid native-framework of disulfides was rendered flexible, the conformational stability provided by the noncovalent interactions which maintain the secondary and tertiary structure will be somehow weakened. For example, the GdmCl concentrations required to denature 50% of bovine pancreatic phospholipase A2 (7 disulfide protein) were found to be 6.4 M by disulfide intact method and 2.5 M by disulfide scrambling method (Singh and Chang 2004). Similar phenomenon was also observed in the cases of other disulfide containing proteins such as ribonuclease A, a-lactalbumin, and human anapylatoxin C3a. Ribonuclease A (123 amino acids with four disulfide bonds) needed 7.5 M urea to denature 50% of the native form when disulfide intact protein was denatured (Pace et al. 1990), whereas by the method of disulfide scrambling 6 M urea was enough to denature 50% of the native protein (Chang 1999b). In the case of calcium bound a-lactalbumin (122 amino acids with four disulfide bonds), 3.75 M GdmCl was needed to denature 50% of the native form by disulfide intact method (Mizuguchi et al. 1999), whereas by disulfide scrambling 2.25 M GdmCl was enough to denature 50% of the native protein (Chang and Li 2001). In the case of C3a, the relative concentration of [GdmCl]1/2 was 4.6/3.4 M (disulfide intact/disulfide scrambling). Interestingly, when the stability of native C3a was measured by CD in the presence of b-mercaptoethanol (0.2 mM), [GdmCl]1/2 was reduced from 4.6 to 3.2 M, comparable to 3.4 M obtained by disulfide scrambling method (Chang et al. 2008). Therefore, the difference in the susceptibility of the protein to denaturants when denaturation is done by the conventional method (disulfide-intact) and by disulfidescrambling reflects the extent of dependence of the respective protein on the rigid framework of its native disulfide bonds to maintain its native structure.
5.4 Folding of Disulfide Proteins via Disulfide Scrambling Using the method of disulfide scrambling, folding experiment can be initiated (1) with a mixture of heterogeneous X-isomers produced by strong denaturant; (2) with
5 Protein Folding via Disulfide Scrambling
99
any isolated X-isomer; or (3) with an isolated X-isomer representing the most unfolded state. The experiment is typically carried out by incubating isolated X-isomer(s) of an unfolded protein in a mild alkaline buffer (pH 8.0–8.5) containing a thiol catalyst. The thiol catalyst can be cysteine (~0.1–2.0 mM), GSH (~0.2– 2.0 mM), or b-mercaptoethanol (~0.02–0.2 mM). Under these conditions, X-isomers refold by shuffling its non-native disulfide bonds, undergo intermediates (also X-isomers) and form the N-protein (Fig. 5.1). Experiments are typically conducted at 22°C. Folding intermediates are trapped by sample acidification and analyzed by HPLC. This folding technique, developed mainly in our laboratory, has been applied to study the folding behaviors of a number of disulfide proteins. The results provide distinctive information regarding the structural and kinetic properties of folding intermediates, including identification of kinetic traps, energy compartments, onpathway/off-pathway intermediates, and the mechanism of early-phase of protein folding. These will be elaborated case-by-case in the following sections.
5.4.1 Unfolding and Refolding of TAP and BPTI Unfolding of TAP and BPTI has been described in the last section. For both proteins, the most unfolded X-isomers were identified as X-TAP-a and X-BPTI-a, respectively (Fig. 5.3). These two isomers were isolated and allowed to refold in parallel in the presence of thiol catalyst to form the native structure. The HPLC profiles of acid trapped folding intermediates and the disulfide structures of isolated folding intermediates are presented in Fig. 5.4. Despite their structural similarities, X-TAP-a and X-BPTI-a fold via very different pathways and mechanisms (Chang and Li 2005). These differences can be summarized as follows. (1) Folding intermediate of TAP is more heterogeneous than that of BPTI. There are about seven well-identified folding intermediates for TAP versus only two predominant intermediates for BPTI. (2) During the early stage of folding, both X-TAP-a and X-BPTI-a reshuffle disulfide bonds at the C-terminal region and convert rapidly to homologous predominant intermediates, X-TAP-d and X-BPTI-d, suggesting that folding activity of both proteins begins at the C-terminal region. However, this is the lone similarity of folding mechanism between TAP and BPTI. (3) One of the major intermediates X-BPTI-b detected in BPTI folding contains Cys30-Cys51, a native disulfide bond which has also been identified as a major 1-disulfide intermediate in the oxidative folding of fully reduced BPTI (Creighton 1990; Weissman and Kim 1991). Its counterpart X-TAP-b appears only as a minor intermediate in TAP folding. (4) Three isomers of X-TAP (X-TAP-f, X-TAP-g, and X-TAP-k) sharing a stable non-native disulfide bond Cys15-Cys33 are shown to act as kinetic traps of TAP folding. Their counterparts are conspicuously absent in the BPTI folding. (5) Most interestingly, stop/go folding experiments have shown that folding intermediates of TAP are largely interconvertible and exist in dynamic equilibrium (Fig. 5.5a), whereas folding intermediates of BPTI are found to be energetically compartmentalized (Fig. 5.5b). A striking demonstration is the energy booth accommodating isomers X-BPTI-e, n, and m (Chang and Li 2005). X-BPTI-e, X-BPTI-n,
100
R.J.Y. Chang
Fig. 5.5 Stop–Go folding experiments of purified folding intermediates of X-TAP (A) and X-BPTI (B). Four folding intermediates of X-TAP-a (d–g) and five folding intermediates of X-BPTI-a (isomers b, d, n, e, i) were purified by HPLC and freeze-dried. To carry on the folding, the samples were dissolved in Tris–HCl buffer (pH 8.4) containing cysteine (1 mM for X-TAP and 0.1 M for X-BPTI). The protein concentration was 0.5 mg/ml. Folding was performed at 23°C for 10 min, trapped by acidification, and analyzed by HPLC. “N” indicates the elution position of the native TAP and BPTI. Numbers given at the left-hand side of each chromatogram indicate the recovery of native TAP and BPTI
and X-BPTI-m equilibrate rapidly among each other, but these three isomers are unable to cross energy barriers and convert to X-BPTI-b, X-BPTI-d, or X-BPTI-i. Structurally, X-BPTI-e, n, and m share Cys5-Cys30 (Fig. 5.4), implying that stability provided by this non-native long-range disulfide has locked these three X-BPTI intermediates in the same energy compartment. A detailed structural characterization of these intermediates (e.g., by NMR) will be needed in order to unravel the noncovalent interactions that reinforce the stability of this non-native disulfide bond.
5.4.2 Unfolding and Refolding of Lysozyme Hen egg lysozyme is one of the most widely and extensively investigated models for understanding the mechanism of protein folding using the conventional “disulfide
5 Protein Folding via Disulfide Scrambling
101
unfolded lysozyme was shown to fold through kinetically partitioned pathways (Kiefhaber 1995; Wildegger and Kiefhaber 1997; Matagne et al. 1997). About 75–85% of the unfolded lysozyme molecules refold on a slow track (t = 420 ms) via well-populated, partially folded intermediates. The remaining 15–25% molecules refold by a fast track (t = 50 ms) via a two-state process with little detectable intermediates. The origin of this kinetic partitioning has been suggested to arise either from the structural heterogeneity of the unfolded state or from the burst-phase collapsed state which resulted in two distinct populations of folding molecules (Wildegger and Kiefhaber 1997; Matagne and Dobson 1998). Our studies using the method of disulfide scrambling clearly demonstrate that two populations of unfolded lysozyme molecules, with molar ratio of 80:20 (X-lysozyme-a/X-lysozyme-b), exist in equilibrium at the unfolded state. These two unfolded isomers also exhibit folding kinetic and folding properties strikingly similar to that observed in the kinetically partitioned pathways (Kiefhaber 1995). These data, taken together, strongly suggest that the structural heterogeneity of the unfolded state, not the conformers in the collapsed state, account for the kinetic partitioning of slow and fast tracks in lysozyme folding. However, it remains possible that the structures of 6 M GdmCl denatured proteins generated by the “disulfide intact” and “disulfide scrambling” methods may differ from each other.
5.4.3 Unfolding and Refolding of a-Lactalbumin a-Lactalbumin (aLA) and lysozyme share sequence homology (~40%), identical disulfide pattern, and similar 3D conformation. It is another popular model that has been extensively investigated by the conventional “disulfide intact” method. Specifically an intermediate of the “molten globule” state has been identified along the pathways of unfolding and refolding of aLA in the presence of four intact native disulfide bonds (Kuwajima 1989; Ptitsyn 1995; Redfield et al. 1999; Luo and Baldwin 1999). The structure of aLA molten globule is characterized by a high degree of native-like secondary structure, a fluctuated tertiary fold, a disordered b-sheet region, and a structured a-helical domain (Kuwajima 1996; Ptitsyn 1995; Wu et al. 1995; Wu and Kim 1998). Our laboratory has analyzed reversible unfolding and refolding of aLA in great details using the method of “disulfide scrambling.” aLA (calcium free) exhibits midpoint denaturation of [GdmCl]1/2 and [Urea]1/2 at 1.1 and 3.4 M, respectively, significantly lower than that of 2.8 and 7.4 M for lysozyme (Chang and Li 2001). At low concentration of GdmCl (1–3 M), denatured aLA comprises about 40 fractions of well identified X-aLA isomers, including partially unfolded X-aLA-b and X-aLA-c (Fig. 5.7). At high concentration of GdmCl (6–8 M), denatured aLA consists of two extensively unfolded isomers, X-aLA-d and X-aLA-a, that constitute about 45% of the total denatured aLA (Chang and Li 2001). Based on the unfolding curve, X-aLA-d and X-aLA-a were identified as the first and second most unfolded X-aLA isomers and were used as starting materials for refolding (Chang 2002). The
102
R.J.Y. Chang
intact” approach (Radford et al. 1992; Matagne and Dobson 1998). It comprises 4 disulfide bonds and 129 amino acids. A denatured 4-disulfide protein, like lysozyme, may adopt 104 possible X-isomers. Denaturation and unfolding of lysozyme via disulfide scrambling have been investigated (Chang and Li 2002). It exhibits midpoint denaturation of [GdmCl]1/2 and [Urea]1/2 at 2.8 and 7.4 M, respectively. The structural heterogeneity of unfolded lysozyme is dependent on the denaturing conditions. Heat-denatured lysozyme consists of about 18 major and minor fractions of X-lysozyme isomers, whereas GdmCl-denatured lysozyme comprises two predominant isomers, X-lysozyme-a and X-lysozyme-b, that constitute more than 90% of unfolded isomers (Fig. 5.6). X-lysozyme-a contains four non-native disulfide bonds arranged in a beads-form configuration (Cys6-Cys30, Cys64-Cys76, Cys80Cys94, and Cys115-Cys127). X-lysozyme-b comprises two non-native disulfide bonds, but retains two native disulfide bonds at the b-sheet domain (Cys6-Cys30, Cys64Cys80, Cys76-Cys94, and Cys115-Cys127) (Fig. 5.6). At the unfolded state, X-lysozyme-a and X-lysozyme-b exist in equilibrium in the presence of 6 M GdmCl, with molar ratio of 80:20. They were isolated and allowed to refold in parallel via disulfide scrambling (Chang et al. 2009). The results show that the folding of X-lysozyme-a is about eight to ten times slower than that of X-lysozyme-b (measured by the recovery of native lysozyme). The population of folding intermediates of X-lysozyme-a is also more heterogeneous than that of X-lysozyme-b. In addition, folding of X-lysozyme-a undergoes X-lysozyme-b as intermediate, whereas X-lysozyme-b folds directly to form native lysozyme, indicating that X-lysozyme-a possesses a higher free energy (Fig. 5.6). These findings may not be surprising as X-lysozyme-b already contains two native disulfide bonds and can be considered as an on-pathway folding intermediate. Results of this simple experiment have significant implications in the elucidation of the early-stage folding mechanism. Experiment of protein folding is typically initiated with 6 M GdmCl denatured proteins, which are generally considered as fully unfolded. However, studies conducted by various laboratories have shown that many 6 M GdmCl denatured proteins are structurally heterogeneous and still retain native-like residual structures. This has been corroborated by our data obtained with the disulfide scrambling technique. The fact that many starting materials of folding experiments still contain residual structures raises the question as to whether the intermediates detected at the dead-time of folding reactions are results of the collapsed state or the unfolded state (Chang 2009). In another word, are these nascent intermediates derived from fully unfolded state via dead-time burst-phase collapse? Or they correspond instead to residual structures of the unfolded state already existing at the starting point of folding experiments? To answer this question will require the detailed characterization of the conformational properties of the unfolded state. If these locally ordered structures are already detected at the unfolded state, then they will constitute unfolding intermediates instead of the usually assumed early nascent folding intermediates. These questions may be answered to some extent by comparing the results of lysozyme folding elucidated by the “disulfide intact” and “disulfide scrambling” methods. Using the conventional “disulfide intact” approach, GdmCl (6 M) denatured/
N
6
30
64
76 80
115 127
94
a a
N
a
N
keq b
b
3M GdmCl
h 75 C 5 min
6M GdmCl
6
30
64 76 80
a
94
115 127
b
o
6
30
64 7680
94
115 127
b
a 1.5 min N N
5 min
N h
N b
b
10 min N
h
N
c 20 min c
h
45 min
Fig. 5.6 Unfolding and refolding of lysozyme. Heterogeneity of denatured lysozyme is dependent on the denaturing conditions (Chang and Li 2002). Samples denatured by 3 M GdmCl, 6 M GdmCl, and 75°C (5 min) are shown here. GdmCl denatured lysozyme comprises two major fractions of X-isomers, X-lysozyme-a, and X-lysozyme-b. They were purified by HPLC and allowed to refold in Tris–HCl buffer (0.1 M, pH 8.4) containing b-mercaptoethanol (0.1 mM). Folding intermediates were trapped by acidification and analyzed by HPLC. The disulfide structures of N-lysozyme, X-lysozyme-a, and X-lysozyme-b are indicated. Folding kinetics of X-lysozyme-a is eightfold slower than that of X-lysozyme-b. The population of folding intermediates of X-lysozyme-a is also more heterogeneous than that of X-lysozyme-b
104
R.J.Y. Chang
C
CTTFHTSGYDTQAIVQNNDSTEYGLFQINNKIWCKDDQNPHSSNICNISCDKFLDDDLTDDIMCVKKILDKVGINYWLAHKALCSEKLDQWLCEK
EQLTK EVFRELKDLKGYGGVSLPEWV
6
28
61
73 77
91
111 120
C
C
C
CC
C
C
6
28
61
73 77
91
111 120
Refolding
Unfolding
X-αLA-d
C
EQLTK EVFRELKDLKGYGGVSLPEWV TTFHTSGYDTQAIVQNNDSTEYGLFQINNKIW KDDQNPHSSNI NIS DKFLDDDLTDDIM VKKILDKVGINYWLAHKAL SEKLDQWL EK
X-αLA-a
C
C
C
CC
C
C
C
EQLTK EVFRELKDLKGYGGVSLPEWV TTFHTSGYDTQAIVQNNDSTEYGLFQINNKIW KDDQNPHSSNI NIS DKFLDDDLTDDIM VKKILDKVGINYWLAHKAL SEKLDQWL EK
6 C
C
C
111 120
X-αLA-b
CNISCDKFLDDDLTDDIMCVKKILDKVGINYWLAHKALCSEKLDQWLCEK
EQLTK EVFRELKDLKGYGGVSLPEWV TTFHTSGYDTQAIVQNNDSTEYGLFQINNKIW KDDQNPHSSNI
6
28
61
73 77
91
111 120
C
CTTFHTSGYDTQAIVQNNDSTEYGLFQINNKIWCKDDQNPHSSNICNISCDKFLDDDLTDDIMCVKKILDKVGINYWLAHKALCSEKLDQWLCEK
EQLTK EVFRELKDLKGYGGVSLPEWV
6
X-αLA-h
28
61
73 77
91
111 120
X-αLA-k CEVFRELKDLKGYGGVSLPEWVCTTFHTSGYDTQAIVQNNDSTEYGLFQINNKIWCKDDQNPHSSNICNISCDKFLDDDLTDDIMCVKKILDKVGINYWLAHKALCSEKLDQWLCEK
EQLTK
6
111 120
X-αLA-c
C
C
C
CC
C
C
6
28
61
73 77
91
111 120
C
EQLTK EVFRELKDLKGYGGVSLPEWV TTFHTSGYDTQAIVQNNDSTEYGLFQINNKIW KDDQNPHSSNI NIS DKFLDDDLTDDIM VKKILDKVGINYWLAHKAL SEKLDQWL EK
N-αLA Fig. 5.7 Reversible unfolding and refolding of a-lactlbumin (aLA). X-aLA-d represents the highest free energy among unfolded X-aLA isomers (Chang and Li 2001) and was used as the starting material of folding experiment (Chang 2002). The major folding pathway of X-aLA-d depicts two predominant on-pathway intermediates (X-aLA-b and X-aLA-c) and two major offpathway intermediates (X-aLA-h and X-aLA-k)
results, summarized in Fig. 5.7 depict two on-pathway intermediates (X-aLA-b and X-aLA-c) and two major off-pathway intermediates (X-aLA-h and X-aLA-k). The folding kinetics of on-pathway intermediates (measured by the rate of recovery of N-aLA) are three- to fourfold faster than that of off-pathway intermediates. In addition, X-aLA-b and X-aLA-c fold to form N-aLA without the accumulation of any significant intermediate. In contrast, conversion of X-aLA-h and X-aLA-k to N-aLA undergoes heterogeneous intermediates, including X-aLA-b and X-aLA-c (Chang 2002). Specifically, isomer X-aLA-c, which comprises a structured a-helical domain and an unfolded b-sheet domain (Chang et al. 2001), is structurally equivalent to the molten globule of aLA observed with “disulfide intact” approach (Kuwajima 1996; Ptitsyn 1995; Wu et al. 1995; Wu and Kim 1998).
5 Protein Folding via Disulfide Scrambling
105
5.5 Concluding Remarks One of the most challenging issues in protein folding is the lack of structural information of the starting material of folding experiment. In vitro folding is typically initiated with proteins denatured with denaturant, heat, or extreme pH. But even with strong denaturant (e.g., 6 M GdmCl), many proteins still retain residual structures and thus comprise heterogeneous conformational isomers (Neri et al. 1992; Shortle and Abeygunawardana 1993; Gorovits et al. 1995; Shortle and Ackerman 2001). This has led to the proposal of widely accepted model of “folding funnel” (Onuchic et al. 1995; Dill and Chan 1997; Socci et al. 1998), which depicts that protein folding undergoes funnel-shaped energy landscapes and starts with heterogeneous conformations. The method of “disulfide scrambling” provides crucial information that should further advance understanding and instigate debate of protein folding landscapes: (1) The structural heterogeneity of a denatured protein can be quantified, albeit by rough grouping of X-isomers. (2) The conformational heterogeneity of a denatured protein is actually characterized by a bell-shaped curve against the increasing concentration of the denaturant and the ascending of free energy. At the two extreme ends of energy landscape, conformational heterogeneity of the protein is reduced to minimum. (3) Folding experiments do not necessarily have to start with heterogeneous conformational isomers. It can be initiated with a single X-isomer representing the highest free energy among all possible unfolded X-isomers. In this case, folding energy landscape can be fittingly illustrated by a diamond-shaped model instead of funnel-shaped model (Chang 2009).
References Antuch W, Güntert P, Billeter M, Hawthorne T, Grossenbacher H, Wüthrich K (1994) NMR solution structure of recombinant tick anticoagulant protein (rTAP), a factor Xa inhibitor from the tick Ornithodoros moubata. FEBS Lett 352:251–257 Arolas JL, Aviles FX, Chang J-Y, Ventura S (2006) Folding of small disulfide-rich proteins: clarifying the puzzle. Trends Biochem Sci 31:292–301 Arias-Moreno X, Arolas JL, Aviles FX, Sancho J, Ventura S (2008) Scrambled isomers as key intermediates in the oxidative folding of liogand binding module 5 of the low density lipoprotein receptor. J Biol Chem 283:13627–13637 Arolas JL, Sanglas L, Lorenzo J, Bronsoms S, Aviles FX (2009) Insights into the two-domain architecture of the metallocarboxypeptidase inhibitor from the Ascaris parasite inferred from the mechanism of its oxidative folding. Biochemistry 48:8225–8232 Cemazar M, Zahariev S, Lopez JJ, Carugo O, Jones JA, Hore PJ, Pongor S (2003) Oxidative folding intermediates with nonnative disulfide bridges between adjacent cysteine residues. Proc Natl Acad Sci USA 100:5754–5759 Cemazar M, Zahariev S, Pongor S, Hore PJ (2004) Oxidative folding of Amaranthus alpha-amylase inhibitor: disulfide bond formation and conformational folding. J Biol Chem 279:16697–16705 Chang J-Y (1994) Controlling the speed of hirudin folding. Biochem J 300:643–650 Chang J-Y, Canals F, Schindler P, Querol E, Aviles FX (1994) The disulfide folding pathway of potato carboxypeptidase inhibitor. J Biol Chem 269:22087–22094
106
R.J.Y. Chang
Chang J-Y (1996) The disulfide folding pathway of tick anticoagulant peptide, a kunitz-type inhibitor structurally homologous to BPTI. Biochemistry 35:11702–11709 Chang J-Y (1997) A two-stage mechanism for the reductive unfolding of disulfide-containing proteins. J Biol Chem 272:69–75 Chang J-Y, Kumar TKS, Yu C (1998) Unfolding and refolding of cardiotoxin III elucidated by reversible conversion of the native and scrambled species. Biochemistry 37:36745–6751 Chang J-Y (1999a) Denatured state of tick anticoagulant peptide. J Biol Chem 274:123–128 Chang J-Y (1999b) Quantitative analysis of the native and scrambled ribonuclease A. Anal Biochem 268:147–150 Chang J-Y, Maerki W, Lai PH (1999) Analysis of the extent of unfolding of denatured insulin-like growth factor. Protein Sci 8:1463–1468 Chang J-Y, Li L, Canals F, Aviles FX (2000) The unfolding pathway and conformational stability of potato carboxypeptidase inhibitor. J Biol Chem 275:14205–14211 Chang J-Y, Ballatore A (2000) The structure of denatured bovine pancreatic trypsin inhibitor (BPTI). FEBS Lett 473:183–187 Chang J-Y, Bulychev A, Li L (2001) A stabilized molten globule protein. FEBS Lett 487:298–300 Chang J-Y, Li L (2001) The structure of denatured alpha-lactalbumin elucidated by the technique of disulfide scrambling. J Biol Chem 276:9705–9712 Chang J-Y (2002) The folding pathway of alpha-lactalbumin elucidated by the technique of disulfide scrambling. J Biol Chem 277:120–126 Chang J-Y, Li L (2002) The unfolding mechanism and the disulfide structures of denatured lysozyme. FEBS Lett 511:73–78 Chang J-Y, Li L (2005) Divergent folding pathways of two homologous proteins, BPTI and tick anticoagulant peptide. Arch Biochem Biophys 437:85–95 Chang J-Y, Lu B-Y, Lin C-J, Yu C (2006) Fully oxidized scrambled isomers are essential and predominant folding intermediates of cardiotoxin-III. FEBS Lett 580:656–660 Chang J-Y, Lin C-J, Salamanca S, Pangburn MK, Wetsel RA (2008) Denaturation and Unfolding of Human Anaphylatoxin C3a: An unusually low covalent stability of its native disulfide bonds. Arch Biochem Biophys 480:104–110 Chang J-Y (2009) Structural heterogeneity of 6M GdmCl denatured proteins: Implications for the mechanism of protein folding. Biochemistry 48:9340–9346 Chang J-Y, Bao-Yuan-Lu, Li Li (2009) Fast and slow tracks in lysozyme folding elucidated by the technique of disulfide scrambling. Protein J 28:300–304 Chatrenet B, Chang J-Y (1993) The disulfide folding pathway of hirudin elucidated by stop/go folding experiments. J Biol Chem 268:20988–20996 Creighton TE (1986) Disulfide bonds as probes of protein folding pathways. Methods Enzymol 131:83–106 Creighton TE (1990) Protein folding. Biochem J 270:1–16 Creighton TE (1992) The disulfide folding pathway of BPTI. Science 256:111–114 Dyson HJ, Wright PE (1996) Insights into protein folding from NMR. Annu Rev Phys Chem 47:369–395 Dyson HJ, Wright PE (2004) Unfolded proteins and protein folding studied by NMR. Chem Rev 104:3607–3622 Dill KA, Chan HS (1997) From Levinthal to pathways to funnels. Nat Struct Biol 4:10–18 Englander SW, Mayne L (1992) Protein folding studied using hydrogen-exchange labeling and two-dimensional NMR. Annu Rev Biophys Biomol Struct 21:243–265 Goldenberg DP (1992) Native and non-native intermediates in the BPTI folding pathway. Trends Biochem Sci 17:257–261 Gorovits BM, Seale JW, Horowitz PM (1995) Residual structure in urea-denatured chaperonin GroEL. Biochemistry 34:13928–13933 Kiefhaber T (1995) Kinetic traps in lysozyme folding. Proc Natl Acad Sci USA 92:9029–9033 Kuwajima K (1989) The molten globule state as a clue for understanding the folding and cooperativity of globular-protein structure. Proteins 6:87–103 Kuwajima K (1996) The molten globule state of alpha-lactalbumin. FASEB J 10:102–109
5 Protein Folding via Disulfide Scrambling
107
Lim-Wilby MSL, Hallenga K, De Maeyer M, Lasters I, Vlasuk GP, Brunck TK (1995) NMR structure determination of tick anticoagulant peptide. Protein Sci 4:178–186 Lin C-H, Li L, Lyu P-C, Chang J-Y (2004) Distinct unfolding and refolding pathways of lipid transfer proteins LTP-1 and LTP-2. Protein J 23:553–566 Luo Y, Baldwin RL (1999) The 28–111 disulfide bond constrains the alpha-lactalbumin molten globule and weakens its cooperativity of folding. Proc Natl Acad Sci USA 96:11283–11287 Matagne A, Dobson CM (1998) The folding process of hen lysozyme: a perspective from the “new view”. Cell Mol Life Sci 54:363–371 Matagne A, Radford SE, Dobson CM (1997) Fast and slow tracks in lysozyme folding: insight into the role of domains in the folding process. J Mol Biol 267:1068–1074 Min CY, Qiao ZS, Feng YM (2004) Unfolding of human proinsulin, intermediates and possible role of its C-peptide in folding/unfolding. Eur J Biochem 271:1737–1747 Mizuguchi M, Masaki K, Nitta K (1999) The molten globule state of a chimera of human alphalactalbumin and equine lysozyme. J Mol Biol 292:1137–1148 Neri D, Billeter M, Wider G, Wuthrich K (1992) NMR determination of residual structure in a urea-denatured protein, the 434-repressor. Science 257:1559–1563 Onuchic JN, Wolynes PG, Luthey-Schulten Z, Socci ND (1995) Toward an outline of the topography of a realistic protein-folding funnel. Proc Natl Acad Sci USA 92:3626–3630 Pace CN (1986) Determination and analysis of urea and GdmCl denaturation curves. Methods Enzymol 131:266–280 Pace CN, Laurents DV, Thomson JA (1990) pH dependence of the urea and guanidine hydrochloride denaturation of ribonuclease A and ribonuclease T1. Biochemistry 29:2564–2572 Ptitsyn OB (1995) Molten globule and protein folding. Adv Protein Chem 47:83–229 Qiao ZS, Min CY, Hua QX, Weiss MA, Feng YM (2003) Kinetic intermediates, putative disulfide forming pathway, folding initiation site and potential role of C-terminal peptide in folding process. J Biol Chem 278:17800–17809 Radford SE, Dobson CM, Evans PA (1992) The folding of hen lysozyme involves partially structured intermediates and multiple pathways. Nature 358:302–307 Redfield C, Schulman BA, Milhollen MA, Kim PS, Dobson CM (1999) Alpha-lactalbumin forms a compact molten globule in the absence of disulfide bonds. Nat Struct Biol 6:948–952 Rothwarf D, Li Y-J, Scheraga HA (1998) Regeneration of bovine pancreatic ribonuclease A: identification of two native-like three-disulfide intermediates involved in separate pathways. Biochemistry 37:3760–3766 Salamanca S, Li L, Vendrell J, Aviles FX, Chang J-Y (2003) Major kinetic traps for the oxidative folding of leech carboxypeptidase inhibitor. Biochemistry 42:6754–6761 Salamanca S, Villegas V, Vendrell J, Aviles FX, Chang J-Y (2002) The unfolding pathway of leech carboxypeptidase inhibitor. J Biol Chem 277:17538–17543 Scheraga HA, Wedemeyer WJ, Welker E (2001) Bovine pancreatic ribonuclease A: oxidative and conformational folding studies. Methods Enzymol 341:189–221 Shortle DR (1996) Structural analysis of non-native states of proteins by NMR methods. Curr Opin Struct Biol 6:24–30 Shortle D, Abeygunawardana C (1993) NMR analysis of the residual structure in the denatured state of a mutant of staphylococcal nuclease. Structure 1:121–134 Shortle D, Ackerman MS (2001) Persistence of native-like topology in a denatured protein in 8M urea. Science 293:487–489 Singh RR, Chang J-Y (2004) Investigating conformational stability of bovine pancreatic phospholipase A2. A novel concept in evaluating the contribution of “native-framework” of disulfides to the global conformational stability of proteins. Biochem J 377:685–692 Socci ND, Onuchic JN, Wolynes PG (1998) Protein folding mechanisms and the multidimensional folding funnel. Proteins 32:136–158 Weissman JS, Kim PS (1991) Re-examination of the folding of BPTI: predominance of native intermediates. Science 253:1386–1393 Welker E, Narayan M, Wedemeyer WJ, Scheraga HA (2001) Structural determinants of oxidative folding in proteins. Proc Natl Acad Sci 98:2312–2316
108
R.J.Y. Chang
Wildegger G, Kiefhaber T (1997) Three-state model for lysozyme folding: triangular folding mechanism with an energetically trapped intermediate. J Mol Biol 270:294–304 Wu LC, Peng ZY, Kim PS (1995) Bipartite structure of the alpha-lactalbumin molten globule. Nat Struct Biol 2:281–286 Wu LC, Kim PS (1998) A specific hydrophobic core in the alpha-lactalbumin molten globule. J Mol Biol 280:175–182
Chapter 6
Oxidative Protein Folding with Small Molecules Watson J. Lees
Abstract The in vitro oxidative folding of disulfide containing proteins has traditionally been enhanced by the addition of redox active small molecules, usually aliphatic thiols such as glutathione. Although enhanced, in many cases folding is still kinetically slow and/or low yielding. In response to the need for improved folding, additional types of redox active small molecules have recently been examined, such as bis(cysteine) containing peptides, small molecule dithiols, aromatic thiols, and selenols. In general, these molecules have demonstrated the ability to improve either or both protein folding rates and yields compared to the standard reagent glutathione, although only a limited number of proteins have been examined so far. Some promising results have also been obtained with in vivo folding. Herein, the evidence in support of each type of molecule is presented and discussed, and some general conclusions are presented. Ultimately, the field is likely to undergo both expansion and consolidation as more examples are investigated. Keywords Protein folding • Oxidation • Thiols • Selenols
Abbreviations 1S A protein folding intermediate with one disulfide bond 2S A protein folding intermediate with two disulfide bonds 3S A protein folding intermediate with three disulfide bonds 4S A protein folding intermediate with four disulfide bonds ALP Alkaline phosphatase ArSH Small molecule aromatic thiol BMC Trans-1,2-bis(2-mercaptoacetamido)cyclohexane or Vertase P W.J. Lees (*) Department of Chemistry and Biochemistry, Florida International University, 11200 SW 8th Street, Miami, FL 33199, USA e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_6, © Springer Science+Business Media, LLC 2011
109
W.J. Lees
110
BME BPTI CGC CMC CTAB CXXC Cya Cys des DMSO Dsb DsbA DsbB DTAB DTT E0¢
Beta-mercaptoethanol or 2-mercaptoethanol Bovine pancreatic trypsin inhibitor A tripeptide containing cysteine-glycine-cysteine Critical micellular concentration Cetyltrimethylammonium bromide Peptide sequence containing two cysteines separated by two amino acids Cystamine Cysteine A protein folding intermediate missing one native disulfide bond Dimethylsulfoxide Disulfide bond formation protein Disulfide bond formation protein A Disulfide bond formation protein B Dodecyltrimethylammonium bromide Dithiothreitol Reduction potential at pH 7
En ER Grx GSH GSeH GSeSeG GSSG h [H+] HPLC Ka Keq M MAP mg min mL mM mV N=N PDI pH pKa R RNase RSH RS−
Ethylenediamine Endoplasmic reticulum Glutaredoxin Reduced glutathione Reduced selenoglutathione Oxidized selenoglutathione or selenoglutathione diselenide Oxidized glutathione or glutathione disulfide Hour H+ ion concentration High-performance liquid chromatography Acid dissociation constant Equilibrium constant Molar Multiple antigen peptide Milligram Minutes Milliliter Millimolar Millivolts Nitrogen to nitrogen double bond Protein disulfide isomerase −log[H+] −log[Ka] Reduced protein Ribonuclease Small molecule thiol Small molecule thiolate
6 Oxidative Protein Folding with Small Molecules
RSSR s Sc SN2 SL Trr Trx
111
Small molecule disulfide Seconds Central sulfur of disulfide Second order nucleophilic substitution Leaving group sulfur of disulfide Thioredoxin reductase Thioredoxin
6.1 Introduction In many proteins, especially extracellular ones, protein disulfide bonds are essential for activity. Disulfide bonds lower the entropy of the unfolded protein and therefore can contribute significantly to overall protein stability and stabilize particular secondary and tertiary structures required for the native fold. A protein disulfide bond is formed by the oxidation of two cysteine thiols, which requires an oxidizing agent. However, not all protein disulfide bonds formed during protein folding are found in the native protein and therefore these incorrect disulfide bonds need to be rearranged to the correct disulfide bonds. Sometimes, disulfide bonds even need to be reduced back again to the dithiol. The simultaneous process of forming, rearranging, and breaking disulfide bonds until ultimately native protein is obtained is called oxidative protein folding (Scheme 6.1) (Arolas et al. 2006). The oxidative folding of proteins in vitro is essential to the biotechnology industry and biochemical research (Sahdev et al. 2008; Mainathambika and Bardwell 2008). Many disulfide-containing proteins, such as insulin, human growth hormone, and erythropoietin, are used as drugs to improve the lives of millions. The most important of these proteins, insulin, is produced with the use of in vitro protein folding. E. coli can rapidly and efficiently produce proteins. However, E. coli is not so efficient at folding proteins, especially disulfide containing ones (Fahnert et al. 2004). Poorly folded proteins aggregate and precipitate out inside the bacterium; the aggregates are known as inclusion bodies. The formation of inclusion bodies protects the proteins from cellular proteases, allows for much higher levels of expression, and can simplify protein purification. However, in order to obtain native protein, the inclusion bodies need to be resolubilized using denaturants, such as urea and guanidine hydrochloride chloride, and the resulting soluble protein needs to be refolded in vitro. In vitro protein folding is considered the most difficult and slowest step in the production of native protein. If the protein of interest contains disulfide bonds, then oxidative protein folding is undertaken. In vivo, oxidative protein folding usually takes place in the endoplasmic reticulum (ER) of eukaryotes or the periplasmic space of Gram-negative bacteria. Inside the ER and periplasmic space, oxidative protein folding is aided by the presence of chaperones such as protein disulfide isomerase (PDI) and the Dsb series of proteins, respectively (Hatahet and Ruddock 2009; Ito and Inaba 2008). The inhibition of PDI is proposed to be a contributing factor in a number of human diseases (Nakamura
W.J. Lees
112
Disulfide Bond Formation
Small Molecule Disulfide R
R
S
S
R S
R
-RSH
S
S + S
S
~S
S
SH
Mixed Disulfide
Reduced Protein
Oxidized Protein
Disulfide Bond Rearrangement S S
S S
~S
S
“Correct” Disulfide Bond
“Incorrect” Disulfide Bond
Disulifde Bond Breaking R
S R
S
~
S
~ R
S
S
+H+
S
R
R
S
S S
+
SH
SH
Mixed Disulfides can also rearrange R
R
S S
S
S
S
S S
S
S
S
Scheme 6.1 Disulfide bond formation, rearrangement, and breaking using a small molecule disulfide (RSSR) and a small molecule thiol (RSH)
and Lipton 2008). In Gram-negative bacteria such as E. coli the periplasmic space is accessible to small molecules that may enhance oxidative protein folding, as the Dsb series of proteins can become overwhelmed during the overexpression of disulfide containing proteins.
6 Oxidative Protein Folding with Small Molecules
113
In this chapter, small molecules that promote oxidative protein folding by directly affecting protein disulfide bond formation, breaking, and rearrangement will be examined. These molecules may or may not be true catalysts, in the strict sense, as in some cases they directly provide the oxidizing equivalents to form the disulfide bonds. First, the reactions underlying protein folding will be presented. The properties of PDI will then be considered, as most small molecule catalysts have been designed to mimic PDI. The small molecules themselves are described in the last three sections: (1) thiol-based catalysts; (2) selenium-based catalysts; (3) reagents that oxidize protein thiols to disulfides but do not catalyze disulfide bond rearrangement. The focus is on small molecules developed in the last two decades (Kersteen and Raines 2003; Lees 2008).
6.2 Underlying Reaction In order to design small molecules that catalyze the formation of disulfide bonds, it is necessary to examine the underlying reaction (Szajewski and Whitesides 1980). We will focus on the thiol–disulfide interchange reaction but parallel reactions exist when the sulfur is replaced with selenium (Scheme 6.2) (Steinmann et al. 2010). The reaction involves an SN2 like nucleophilic attack of a negatively charged thiolate on a disulfide bond. The disulfide bond contains two sulfur atoms, one of which is the site of nucleophilic attack (the central sulfur, SC). The second sulfur becomes part of the leaving group (SL). Although only the anionic thiolate is reactive and not the neutral thiol, nucleophiles are compared based upon the pKa value of their conjugate acid, the neutral thiol (Szajewski and Whitesides 1980). In general, thiols with pKa values close to the solution pH make the best nucleophiles. Thiols with pKa values higher than the solution pH exist mainly in the thiol form and not in the reactive thiolate form. Thiols with pKa values lower than the solution pH are less reactive, as the thiolate is more stable, as indicated by the lower pKa value. Leaving groups are also ranked by the thiol pKa value of their conjugate acids. The lower the thiol pKa value, the better the leaving group. Selenolates are 100–1000 fold more nucleophilic than thiolates but are similar or slightly worse as leaving groups, once pKa values have been taken into account. Selenium in diselenides is 10,000-fold better as the site of nucleophilic attack (equivalent to SC in Scheme 6.2) than sulfur in disulfides (Steinmann et al. 2010). Selenols have lower pKa values (pKa = 5–6) than thiols (pKa = 8–9). All these factors will play a role in oxidative protein folding with small molecule catalysts, however, additional factors such as protein disulfide/ thiol accessibility and protein conformation will also be major factors (Arolas et al. 2006; Narayan et al. 2000; Wedemeyer et al. 2000). R'
Scheme 6.2 The thiol– disulfide interchange reaction
SN R
+
SC
R' SL
SN
R"
R
SC
+
SL R"
W.J. Lees
114 Protein Thiolate as Nucleophile
I
S S S
III
R
S HS S
S
R S +
S S ∼S
II
SH S ∼S
IV
S
RSSR
S S
R
S HS S
R S + R S S S S
R S S
S
HS S
Small Molecule Thiolate as Nucleophile
V
R S +
HS
R
HS
R S + R
S S
R S VII + R
S HS S
S S S
S HS S
HS
HS
VI
R S
R S R VIII +
S HS S
HS
RSSR
R S
+ RSSR
RSSR + HS HS S
Scheme 6.3 The eight types of thiol–disulfide interchange reactions that occur during protein folding where RS− is the small molecule thiolate
Eight general thiol–disulfide interchange reactions take place during oxidative protein folding using small molecule disulfides or diselenides (Scheme 6.3). The eight reactions are the result of having two general types of nucleophiles in solution and four general types of electrophiles. Four reactions involve the protein thiolate acting as a nucleophile and four reactions involve the small molecule thiolate or selenolate acting as a nucleophile. Enhancing the nucleophilicity of the small molecule would increase the rate of four of these reactions (V–VIII), and enhancing the leaving group ability of the small molecule would increase the rate of four of these reactions (II, III, VI, VII). Two of these reactions have no net change if only one small molecule thiolate is used (VI and VII), which is most often the case. Most importantly, seven of these eight reactions involve the small molecule thiol or disulfide and only one, the intramolecular rearrangement of protein disulfide bonds (I), does not. Therefore, enhancing the nucleophilicity and/or leaving group ability of the small molecule catalyst is likely to improve the rate of protein folding, assuming all else stays the same.
6 Oxidative Protein Folding with Small Molecules Scheme 6.4 Folding of reduced RNase A, R, via des intermediates containing all but the indicated native disulfide bond as proposed by Scheraga et al. (Narayan et al. 2000). 1S, 2S, 3S, and 4S refer to the number of disulfide bonds in the intermediates
R
115
1S
3S
2S
Des[40-95]
4S
Des[65-72]
Native
Equilibria between protein folding intermediates and the small molecule reagents used for protein folding have a significant effect upon folding. If the small molecule is strongly reducing, such as reduced DTT, the formation of disulfide bonds becomes difficult as the equilibrium is shifted toward reduced protein and away from native protein. If the environment is too oxidizing, disulfide bonds are formed but rearrangement becomes more difficult, as there are only a few free protein thiolates. As a consequence of these effects, the addition of more small molecule catalyst, which would increase the reaction rates of many of the eight underlying reactions, does not always lead to faster folding (Lyles and Gilbert 1991). With glutathione (GSH) and glutathione disulfide (GSSG) many proteins have optimal conditions where the addition of more or less of one or both reagents does not improve the folding rate. In the case of RNase A, this is proposed to be because there is a pre-equilibrium mixture of intermediates containing one (1S), two (2S), three (3S), or four disulfide bonds (4S), which is shifted away from the optimal concentration of 3S intermediates as the ratio of GSSG to GSH is altered (Scheme 6.4) (Narayan et al. 2000). Therefore, varying the ratio of reduced to oxidized reagent used during protein folding can have a profound effect upon folding rates. In summary, although the oxidative folding of disulfide-containing proteins is a complex process, the design of small molecule catalysts for protein folding needs to take into account the underlying reactions. The small molecule catalyst should increase reaction rates and create the potential to vary the equilibrium composition of the intermediates.
6.3 Design of Small Molecules for Protein Folding The design of small molecule folding catalysts has been driven in large part by the desire to mimic nature. The best known in vivo protein folding catalyst, protein disulfide isomerase, PDI, has served as the paradigm (Kozlov et al. 2010). PDI has two cysteines in each active site separated by two amino acids (CXXC motif); one of the cysteine thiols is solvent exposed while the other is buried. The solvent exposed thiol has enhanced reactivity with exogenous disulfides and has a relatively low thiol pKa value, pKa = 6.7 as most protein thiols have pKa values around 9
W.J. Lees
116 Dithiols with the escape mechanism
S HS
HS
S
S3 S4
S1 S2
S3 S4
S3
S2
S
S4
S S3H S4
Escape
S1 S2
S1 S2
HS S1
S
HS S1
S3
S4
S
S2
Scheme 6.5 Folding of proteins with monothiols and dithiols illustrating the escape mechanism
(Hawkins and Freedman 1991). In addition, the oxidized form of PDI, which is a 14-membered cyclic disulfide, is relatively unstable as indicated by the redox potential, E0¢ = -180 mV (Lundstroem and Holmgren 1993). The lower the redox potential, the more stable the disulfide. The difference in redox potential corresponds to the equilibrium constant between a thiol and a disulfide: 0.0296 × (log (Keq)) = change in redox potential, presuming standard temperature and pressure, and a pH of 7. The equilibrium constant between reduced PDI and oxidized glutathione is 3 mM (Lundstroem and Holmgren 1993). Small molecule catalysts of protein folding have thus been design to have reactive thiols with low pKa values, two thiol groups, and high reduction potentials.
6 Oxidative Protein Folding with Small Molecules
117
Small molecule catalysts containing two thiol groups with low pKa values, reactive thiols, and high reduction potentials have many potential advantages. A low thiol pKa value will enhance the leaving group ability of the corresponding thiolate. A thiol pKa value close to solution pH will enhance the nucleophilicity of the corresponding thiolate. Both of these factors will enhance the rate of the underlying thiol disulfide interchange reactions. Having two thiol groups in each active site will limit the lifetime of the mixed disulfide intermediates that occur during folding (Scheme 6.5). With monothiols, mixed disulfides between the protein and the monothiol can become kinetic traps, but with dithiols the mixed disulfide intermediates have a limited lifetime because the second thiol can attack the mixed disulfide intramolecularly, removing the dithiol. This is called the escape mechanism, and the second thiol is said to act as an “intramolecular clock.” (Walker and Gilbert 1997) A high reduction potential allows PDI to exist in the endoplasmic reticulum (ER) as a nearly 1:1 mixture of oxidized and reduced protein. The oxidized form of PDI can help form protein disulfide bonds via oxidation while the reduced form of PDI can help rearrange protein disulfide bonds. Since the reduction potential found inside the ER is close to optimal for the folding of many disulfide containing proteins (it sets up the equilibria between the different folding intermediates), any small molecule dithiol should have a redox potential similar to PDI’s. With small molecule dithiols, the redox potential is based upon the stability of the corresponding cyclic disulfide and the thiol pKa value. The lower the thiol pKa value, the higher the redox potential because the equilibrium mixture is shifted toward the thiol due to deprotonation, especially if the thiol pKa value is lower than the solution pH. Biochemical redox potentials are referenced to pH 7 by definition. With monothiols, the redox potential is mainly determined by the thiol pKa value, as most noncyclic disulfides have similar stabilities in the absence of pKa effects, barring steric hindrances (Lees and Whitesides 1993).
6.4 Folding Strategies The oxidative folding of disulfide-containing protein involves both the formation of disulfide bonds and the intramolecular rearrangement of disulfide bonds to form native protein (Scheme 6.1). If the formation of disulfide bonds occurs without rearrangement, fully oxidized protein is obtained but very little native protein is produced, as most of the protein molecules contain non-native disulfide bonds. In the case of RNase A with four disulfide bonds, rapid oxidation of the reduced protein produces about 0.6% native protein (Narayan et al. 2003a). If disulfide bond formation does not occur, there are no disulfide bonds to rearrange. In many cases, both steps are performed simultaneously, but it is also common to perform them separately. For example, RNase A may be fully air oxidized under denaturing conditions to form scrambled RNase A, which is RNase A with an approximately random distribution of its four disulfide bonds. The scrambled RNase A is subsequently folded to native RNase A in the presence of a small molecule catalyst or enzyme that promotes the rearrangement of disulfide bonds. Native bovine pancreatic trypsin
W.J. Lees
118
inhibitor (BPTI) and tick anticoagulent peptide have also been prepared by a similar two-step procedure, providing additional information about the protein folding pathway (Chang and Li 2005).
6.5 Catalysts that Form and Rearrange Disulfide Bonds In order for small molecules to act as oxidative protein folding catalysts, they need to both form and break protein disulfide bonds in the same reaction mixture. Therefore, molecules used to catalyze oxidative protein folding have so far been limited to disulfides or diselenides and their corresponding thiols or selenols. Disulfides and diselenides have the advantage that they can oxidize protein thiols to protein disulfides, while the corresponding thiols and selenols are able to rearrange incorrect protein Traditional Reagents NH2 NH2
H N
HO2C O
HO2C
O N H SH
H N
Ox CO2H
Red
O
HO
SH SH
DTT HO
Ox
HO
N H
Red
HO
Ox
BME HO
SH
Cystamine H2N
SH
S
HO
Red
HO
Ox
H2N
Red
S
S S
S S
H2N O H2N
O Cysteine
H2N
Ox
S
OH SH
OH
S
Red
OH
H2N O
CO2H
S
O
NH2
N H S
GSSG HO2C
GSH
O
H N O
CO2H
6 Oxidative Protein Folding with Small Molecules
119
disulfide bonds to the correct native ones. During protein folding, disulfide bonds are continually formed and broken until the thermodynamically favorable native protein is obtained. The catalyst section is broken into two parts. The first part looks at disulfidebased catalysts and the second part looks at diselenide-based catalysts.
6.5.1 Thiol Based Catalysts 6.5.1.1 Glutathione Glutathione has historically been used as a protein folding catalyst because it occurs where disulfide-containing proteins are folded in vivo in eukaryotes. Glutathione is found in the ER at a total concentration of about 10 mM. Although, the ratio of reduced glutathione to oxidized glutathione to protein mixed disulfides with glutathione is a little uncertain, the ratio of glutathione entities in the reduced form to glutathione entities in the oxidized form is between 3:1 and 1:1 with more recent results tending toward 1:1 (Hwang et al. 1992; Bass et al. 2004). Given its importance in vivo, glutathione is often used as the starting point for protein folding experiments and as the standard to which other results are compared. This is especially true for the model proteins traditionally used for examining protein folding catalysts, such as RNase A, lysozyme, and BPTI. With the possible exception of RNase A, however, no standard set of conditions is used for comparison, making it difficult to compare protein-folding catalysts directly. The relative concentrations of reduced and oxidized glutathione can have an effect upon folding rates and yields. Thus, nonstandard conditions may lead to poor folding results with glutathione. For RNase A the typical conditions are 0.2 mM GSSG and 1 mM GSH at pH 7–8 and 20–30°C (Lyles and Gilbert 1991). When the concentration of GSH or GSSG changes, the protein folding rate decreases. Surprisingly, the yield and rate reported for typical conditions can be quite different. We have found that the source and/or preparation of the reduced RNase A or the scrambled RNase A can make a difference in the results for typical conditions. With lysozyme at pH 7 and 25°C we found the best conditions to be 7 mM GSH and 2 mM GSSG, similar to results obtained by others (Gurbhele-Tupkar et al. 2008; Hevehan and Clark 1997), while for BPTI they were 2 mM GSH and 2 mM GSSG or 5 mM GSH and 5 mM GSSG at pH 7.4 and 25°C (Kibria and Lees 2008). The best conditions tend to have a relatively broad optimum (Fig. 6.1).
6.5.1.2 Dithiothreitol The main advantage of DTT over glutathione is that it does not form stable mixed disulfides with proteins. Thus, interpretation of protein folding pathways becomes much simpler as mixed disulfides do not accumulate. The main disadvantage of DTT with respect to glutathione is that it is strongly reducing, as it forms a stable six-membered cyclic disulfide. Therefore, most oxidative folding reactions start with oxidized DTT in relatively high concentrations and little or no reduced DTT;
W.J. Lees
120 Fig. 6.1 Folding of reduced BPTI in the presence of various concentrations of GSH and GSSG at pH 7.4 (Kibria and Lees 2008). The yield of native protein after 24 h of folding is indicated
%Native protein after 24 h
90
60
30
0
[GSSG] 0.125 mM 0.5 mM 2 mM 5 mM
0
5
20
15 10 [GSH] (mM)
25
otherwise, protein disulfide bonds are not formed. Furthermore, protein folding rates with DTT tend to be slower than those obtained with glutathione. Thus, DTT is mainly used to fold proteins for mechanistic studies (Narayan et al. 2000). 6.5.1.3 Beta-Mercaptoethanol, Cysteine, Cystamine The main advantage of these compounds over glutathione is that they are less expensive and therefore attractive for larger scale preparations of proteins. In addition, in some cases they can provide improved folding rates and yields. Cystamine has a net positive charge at pH 7, unlike glutathione, which has a net negative charge. Cystamine has a lower thiol pKa value than glutathione. Due to the charge difference, the interactions of glutathione and cystamine with proteins may be slightly different.
SH H N
Ac-Trp-NH
NH
O N H
O SH Ac-Ala-NH
Thiol based Reagents
N
PDI-[34-41]
H N
O
H N O
H N
O O Thioredoxin reductase-[134-141] SH H N
Ac-Trp-NH O
Lys-Ala-Leu-NH2
SH
O
SH
HO N H
Glutaredoxin-[10-17]
O
O O
[His37]-Thioredoxin-[31-38]
Ac-Gly-NH
O
Asp-Gly-PheNH2
O
H N
N H
O
SH
H N
N
N
OH
Val-Arg-Ala-NH2
O
Linear peptides with CXXC motif
O
SH
H N
Lys-His-Ile-NH2 S
O Lys-His-Ile-NH2 SH
Ox Red
N O
S N H
O
NH-Trp-Ac
6 Oxidative Protein Folding with Small Molecules Pro Cyclic peptide with CXXC motif
Tyr
Gly
Cys
Cys
SH Cys
121
Pro
Gly
Cys
Cys
SH
Gly
Lys
Trp
Trp
Lys
SH
Lys-Cys-Ala-Thr-Cys-Asp-Lys-Lys
NH
NH
O N=N
N= N
Photoactive Peptide with CXXC motif
Lys
S
S Cis Isomer
Lys-Cys-Ala-Thr-Cys-Asp-Lys-Lys O
Cys S
Red
PDI peptide
Thioredoxin peptide
SH
S Cys
Cys HS
Trp
Val
Glutaredoxin peptide
His
Ox
SH
HS
HS
Gly
His
Ox hν
hν
hν
Red
S
SH
SH
Lys-Cys-Ala-Thr-Cys-Asp-Lys-Lys
Trans Isomer
NH
O
S
Lys-Cys-Ala-Thr-Cys-Asp-Lys-Lys N
O
N
NH
=
=
N
N
CO2H SH H2N
O
SH
O
H N
N H CGC
O HN
Ox OH
O
Red
S S
HN NH2
O
O O
H N
BMC
N H
O
SH
Ox
SH
Red
HN
Gemini
()
n
S S
HN O
ArsH R = CH2CO2H, SO3H, COOH, R CH2PO3H2, CH2N+(CH3)3
+ N
S
Ox S
R
SH Red
Ox SH Red
( )n
+ N
S S
hν
+ N
( )n
R
122
W.J. Lees
6.5.1.4 Linear, Cyclic, and Photoactive Peptides with the CXXC Motif In order to mimic the active site of PDI, which has a CXXC motif, linear peptides with a similar motif were synthesized and tested for their ability to oxidatively fold reduced RNase A. Then more constrained cyclic peptides and photoactive peptides with a CXXC motif were prepared to better represent the active site of PDI. Constraining the peptide should lead to a less stable and more reactive disulfide bond. In an initial report, an eight amino acid peptide, containing the CXXC active site of PDI, was attached to multiple antigen peptide (MAP) resin (Ookura et al. 1995). Although promising results were obtained with this approach it was not taken further. Later, linear peptides with eight amino acids each that contained the CXXC active sites of thioredoxin (Trx, His inserted at position 37), thioredoxin reductase (Trr), glutaredoxin (Grx), and PDI were prepared (Moroder et al. 1996). In the presence of stoichiometric amounts of oxidized peptide or GSSG (relative to disulfide bonds), reduced RNase A was oxidized to 60% native protein in about 5 days. The peptide derived from thioredoxin was slightly faster than GSSG. Doubling the stoichiometry led to faster folding. The thioredoxin peptide now gave double the rate of GSSG. Higher yields were also obtained with the peptides, giving 75–90% versus 70% for GSSG in about a day. Addition of the peptide to the standard conditions of 0.2 mM GSSG and 1 mM GSH or the use of 0.1 mM oxidized peptide did not improve folding relative to the standard conditions (90% in 8 h). The thioredoxin peptide had the lowest reduction potential of all the peptides and was therefore probably the most reactive disulfide, due to strain. However, the thioredoxin peptide did not improve upon the standard conditions. To induce more strain into the disulfide bond, cyclic peptide dithiols containing the CXXC motif and six amino acids were prepared (Cabrele et al. 2002a). Again the amino sequence was based upon the active sites of thioredoxin, thioredoxin reductase, glutaredoxin, and PDI. The corresponding equilibrium constants with cystine were 4.2, 230, 33, and 0.8 mM, respectively, indicating that the least stable disulfide was that derived from the PDI sequence. The first thiol pKa values were all in the range of 7.3–7.9, lower than that of glutathione (8.7). Folding results with RNase A demonstrated that the less stable the cyclic disulfide was, the better the peptide performed in terms of yield and rate. Using a 1:40:8:0.8 molar ratio of reduced RNase A:GSH:GSSG:peptide (0.025, 1, 0.2, and 0.02 mM) the folding yield could be increased to 95% (PDI peptide) from the 83% obtained with the standard conditions (1 mM GSH, 0.2 mM GSSG). The rate also went up by 20%. All peptides showed improvement. By using a 1:20:4 ratio of RNase A:GSH:oxidized peptide the folding rate could be doubled over standard conditions, but the yield declined to 66%. Direct addition of stoichiometric amounts of peptide to reduced RNase A gave the same rate and yield as the standard conditions. Photoactive cyclic peptides are able to adjust the stability of their disulfide bond with light (Cabrele et al. 2002b; Cattani-Scholz et al. 2002). A peptide containing eight amino acids derived from the active site of thioredoxin reductase was connected at both ends with a photoreactive azobenzene group. The azobenzene moiety
6 Oxidative Protein Folding with Small Molecules
123
has an N=N double bond that can be isomerized between the trans and cis isomers with light. The conformation of the peptide also changes upon isomerization, as does the stability of the disulfide bond in the peptide. The cis isomer contains the least stable disulfide bond as it has an equilibrium constant with glutathione disulfide of 0.8 mM compared with 49 mM for the trans isomer. Neither isomer significantly enhanced the folding rate of RNase A relative to standard conditions (1 mM GSH and 0.2 mM GSSG). However, in the presence of a 1:20:2:2 ratio of reagents (RNase A (0.024 mM):GSH:GSSG:either conformation of oxidized peptide) the folding yield was approximately 100%, while in the presence of the standard conditions it was only 77%, although, the folding rate decreased by a factor of 2–3. Replacing more or less of the GSSG with either isomer of the cyclic peptide only decreased the yield. Ultimately, the addition of cyclic peptides was able to significantly increase the protein-folding yield of RNase A but only had a limited effect upon folding rates.
6.5.1.5 CXC Motif The CXC motif allows for the formation of a strained 11-membered cyclic disulfide upon oxidation (Woycechowsky and Raines 2003). The CGC sequence was selected, as it corresponds to the active site of thioredoxin without an intervening proline. Scrambled RNase A, which is RNase A with an approximately random distribution of disulfide bonds, was folded in the presence of 1 mM GSH or 0.5 mM CGC or 0.5 mM BMC (vide infra). Since standard conditions were not used, it is difficult to make direct comparisons, but in 2 h and in the presence of CGC at 30°C and pH 7.6, scrambled RNase A was folded in 60% yield. The same group published an earlier paper using standard conditions (1 mM GSH and 0.2 mM GSSG) in which the scrambled RNase A folded in about 3 h in about a 40% yield (Woycechowsky et al. 1999). Thus, CGC appears to have improved the yield significantly and also the folding rate somewhat.
6.5.1.6 Trans-1,2-bis(2-mercaptoacetamido)cyclohexane (BMC or Vertase P) Small molecules that do not mimic the active site of PDI but have some of the important physical properties of PDI have also been designed. BMC is a small molecule dithiol that forms a relatively unstable cyclic disulfide upon oxidation as indicated by its redox potential of −0.24 V, which is higher than that of glutathione (−0.25 V). In addition, it has a slightly lower thiol pKa value (8.3) than glutathione (8.7). Initially, it was shown that the addition of 1 mM of BMC to a folding mixture containing 1 mM GSH and 0.5 mM GSSG would enhance the folding yield of scrambled RNase A at pH 7.6 and 30°C from about 40% to over 90% (Woycechowsky et al. 1999). The addition of the monothiol analog only raised the yield to 60%. The addition of BMC to the growth media of S. cerevisiae resulted in a threefold
124
W.J. Lees
increase in the amount of a secreted acid phosphatase (Woycechowsky et al. 1999). The addition of 0.05 mM BMC to a mixture of reduced RNase A and 50 mM oxidized DTT doubled the folding rate (Fink et al. 2008). BMC was also shown in a factorial screen to be more efficient at folding disulfide-containing protein than GSSG/GSH in terms of yield, and it was recommended that groups folding disulfide containing proteins should consider using BMC (Willis et al. 2005). The in vitro folding yield of proinsulin could be doubled from 20 to 40% by the addition of 0.01 mM BMC to a folding mixture containing 2 mM GSSG and 1 mM GSH, although the folding rate did not increase significantly (Winter et al. 2002). Higher concentrations of BMC decreased the yield. It was speculated that BMC was affecting protein aggregation during the early stages of folding. In vivo, a 60% improvement in yield was noted when BMC (0.02–0.05 mM) was added to the growth media of E. coli, which was secreting proinsulin into its periplasmic space (Winter et al. 2002). Again higher concentrations were disadvantageous. Both in vivo and in vitro the addition of BMC significantly improved protein-folding yields, even at low concentrations.
6.5.1.7 Aromatic Thiols (ArSH) Aromatic thiols have lower thiol pKa values than most aliphatic thiols, but the corresponding thiolate is more nucleophilic once the thiol pKa value has been taken into account (DeCollo and Lees 2001). The lower thiol pKa will make the corresponding thiolate a better leaving group but will not decrease the nucleophilicity as aromatic thiolates are inherently more nucleophilic than their aliphatic counterparts. Therefore, aromatic thiols should enhance the rates of most of the thiol–disulfide interchange reactions that occur during protein folding. Folding rates of RNase A have been enhanced by a factor of up to 23 at pH 6, up to 12 at pH 7, and up to 8 at pH 7.7 relative to GSH under standard conditions (Gough et al. 2002, 2003). Approximately 2 mM of thiol in the neutral SH form in solution was optimal. Also, the best aromatic thiol had a thiol pKa value about 1–2 units below the pH of the solution. In addition, the leaving group ability of the thiol was important. Even the catalytic activity of PDI could be increased by the use of aromatic thiols instead of glutathione (Gough and Lees 2005). Folding of reduced lysozyme was also accelerated with aromatic thiols relative to GSH. At high lysozyme concentration (1 mg/mL) folding rates at pH 8 were increased sevenfold and folding yields were increased from 70 to 96% relative to glutathione (Madar et al. 2009). At pH 7 the effects were even more dramatic, with a tenfold rate acceleration, and the yield increased from 48 to 89%. Overall the use of aromatic thiols can significantly increase both the rate and yield of protein folding even at high protein concentrations.
6.5.1.8 Gemini Surfactants Gemini surfactants offer the advantage of being both disulfides and surfactants that will bind to proteins (Potempa et al. 2010). Quaternary ammonium salt surfactants
6 Oxidative Protein Folding with Small Molecules
125
linked by a disulfide bond (gemini surfactant) were used to oxidatively fold lysozyme (0.2–0.3 mg/mL) at pH 8.5 in the presence of 0.03 M guanidinium chloride. The surfactant was initially added to the reduced lysozyme for 15 min (capture phase) and then methyl-b-cyclodextrin was added to strip the detergent away from the lysozyme (release phase). A number of surfactants of varying length (n = 7–14) were examined. The highest yield, 80%, was obtained when the carbon chain was 11 carbons long (n = 9). For this compound, the yield increased with increasing concentration in a sigmoidal fashion until about 1 mM, where it leveled off. The corresponding critical micellular concentration (CMC) was 0.55 mM. For comparison, a 40% yield was obtained with the 16-carbon long gemini detergent (n = 14), and a 60% yield was obtained using a combination of cetyltrimethylammonium bromide (CTAB, a detergent with a 16-carbon chain and no disulfide bond) and glutathione/glutathione disulfide in a similar manner. Using only glutathione and glutathione disulfide without a detergent gave a yield of 18%. Chains shorter than 11 appeared to cause protein aggregation during the first 15 min (capture phase) as measured by turbidity. However, chains longer than 11 carbons appeared to cause faster aggregation during the release phase after the cyclodextrin was added, as measured by kinetic models. Both RNase A and alkaline phosphatase (ALP) were also successfully folded using a similar strategy (Potempa et al. 2010). RNase A (2 mg/mL) was folded with 9.4 mM gemini disulfide (n = 10) and 18.9 mM corresponding thiol in 71% yield. The control of 9.4 mM GSSG and 18.9 mM GSH provided a 51% yield. ALP (0.18 mg/mL) was folded in the presence of 4.65 mM gemini disulfide (n = 10) and 23.2 mM corresponding thiol in 18% yield. The same yield was obtained with 4.6 mM GSSG and 23.2 mM GSH or 27.9 mM dodecyltrimethylammonium bromide (DTAB, a detergent with a 12-carbon chain), 4.6 mM GSSG and 23.2 mM GSH. Ultimately, the use of disulfide-containing surfactants in a capture release methodology is generally applicable to protein folding, although direct comparisons with other methods are difficult as nonstandard conditions were used as controls. Interestingly, the effect of gemini disulfide chains on folding appears to be different from that of their nondisulfide analogs.
Selenol Based Reagents H2N
SeH
Ox
H2N
Red
H2N
SeIenocystamine
Se Se
NH2 NH2
H N
HO2C O
GSeH
O N H SeH
Ox CO2H
Red
H N
HO2C O
CO2H
H N
CO2H
Se
O
NH2
N H Se
GSeSeG HO2C
O
N H
O
126
W.J. Lees
6.5.2 Selenol Based Catalysts 6.5.2.1 Selenoglutathione (GSeH) and Selenocystamine Diselenides and selenates offer advantages over disulfides and thiolates as they are more electrophilic and more nucleophilic, respectively (Steinmann et al. 2010). In addition, selenols have lower pKa values than thiols, and are more rapidly oxidized by oxygen or air to the diselenide, than thiols to disulfides. GSeH has a reduction potential of −0.41 V and a selenol pKa of 5.3 (Beld et al. 2007). Initially, glutathione diselenide (GSeSeG) was used to fold reduced RNase A and BPTI. Reduced RNase A at early time points folded twice as fast with 0.02 mM GSeSeG as with 0.02 mM GSSG, and the final yield of native RNase A was the same with both reagents. Next, the folding of BPTI was followed by HPLC (Beld et al. 2007). HPLC-based assays, in general, provide more information as the appearance and disappearance of intermediates can be followed, not just the appearance of native protein. With BPTI the folding kinetics with 0.15 mM of both reagents was quite different under aerobic conditions. Initially, GSeSeG produced more native BPTI (first 8 min), then GSSG surpassed it (next 2 h) and finally GSeSeG again produced more native protein. With GSeSeG the protein folded completely within 4 h. GSSG yielded only about 70% native protein after 4 h. GSeSeG appeared to be faster than GSSG at converting reduced BPTI to intermediates with one disulfide bond, slower at converting intermediates with one disulfide bond to intermediates with two disulfide bonds, and faster at converting intermediates with two disulfide bonds to native protein with three disulfide bonds. Under anaerobic conditions, GSeSeG produced the equivalent amount or more of native protein at all time points relative to GSSG, although folding was not complete within 4 h. The biggest differences occurred at longer times. Under aerobic conditions, we speculate that the results might be due to the fact that GSeSeG is faster at forming protein disulfide bonds directly than GSSG but not quite as efficient at rearranging them. Under aerobic conditions, GSeH, which is important for catalyzing rearrangements, will be oxidized to GSeSeG by air. Conversion from reduced protein to intermediates with one disulfide bond and from intermediates with two disulfide bonds to native protein may only involve direct oxidation. Under anaerobic conditions where the concentration of GSeH will be higher (no air oxidation) the conversion of intermediates with one disulfide bond to intermediates with two disulfide bonds is faster. However, because the GSeSeG concentrations is lower (not formed by air oxidation) the conversion of intermediates with two disulfide bonds to native protein is slower and the yield of native protein after 4 h is lower. The folding of reduced RNase A with GSeSeG was next examined under a variety of conditions (Beld et al. 2008). In the presence of air, GSeSeG was shown to act catalytically as an oxidant. Under the standard conditions, 1 mM GSH or GSeH and 0.2 mM GSSG or GSeSeG at pH 8, folding was two- to threefold faster with GSeSeG and the yield of native protein was the same (80% native protein was obtained in about 2 h with GSeSeG and about 6 h with GSSG). More importantly, reduced RNase
6 Oxidative Protein Folding with Small Molecules
127
A could be folded by substoichiometric concentrations of GSeSeG but not GSSG, because of the rapid oxidation of GSeH to GSeSeG by air. With a 10:1 ratio of protein thiols to GSeSeG at pH 8, reduced RNase A was oxidized to 80% native protein in 24 h. With GSSG only about 10% native protein was produced. At pH 5, the effect was even more dramatic: after 50 h 50% native protein was obtained with GSeSeG and essentially no native protein was obtained with GSSG. Selenocystamine was also examined at pH 5 and 7 but provided lower yields than GSeSeG. Interestingly, the addition of PDI also improved folding in the presence of diselenides. The ability of diselenides to catalyze the air oxidation of thiols was used to significantly improve protein folding in vivo (Beld et al. 2010). Selenols are rapidly oxidized by oxygen to diselenides but thiols are only very slowly oxidized by oxygen to disulfides. Gram-negative bacteria oxidize and fold disulfide-containing proteins in their periplasmic space, which is accessible to small molecule additives. In E. coli the enzyme DsbA, with a CXXC motif in its active site, directly oxidizes protein dithiols to protein disulfides. In DsbA-deficient cells, the formation of disulfide bonds in periplasmic proteins is retarded. Using three different in vivo assays, Hilvert et al. were able to demonstrate that the addition of selenocystamine could restore much if not all of the lost activity of DsbA at concentrations of 1–10 mM. Selenocystamine was the most effective of several diselenides tested, and disulfides had little or no effect at much higher concentrations. The authors propose that the positive charge on selenocystamine might lead to its preferential accumulation in the periplasmic space, which is lined with negatively charged phospholipids. In DsbAdeficient E. coli the expression of ALP, with two structural disulfide bonds required to attain its active conformation, could be restored to the same level either by the addition of 1 mM selenocystamine or incorporation of a plasmid encoding DsbA. The motility of a DsbA-deficient E. coli could be restored by the addition of 1 mM selenocystamine. One of the flagellar proteins required for motility contains an essential disulfide bond. Even in E. coli deficient in both DsbA and DsbB, motility could be restored by the addition of selenocystamine. Based on the synergistic effects of selenocystamine in vitro with PDI it is expected that the addition of selenocystamine to cells expressing normal levels of DsbA will also be synergistic, leading to the use of selenocystamine in the biotechnological production of proteins.
6.6 Rapid Formation of Disulfide Bonds Without Rearrangement The oxidation of thiols to disulfides can be accomplished using a number of reagents, but in addition to their advantages, all have disadvantages such as incompatibility with other amino acids, side reactions, slow reaction rates, or nonaqueous reaction conditions. Examples include oxygen, iodine, DMSO, potassium ferricyanide, Ellman’s reagent, and thallium trifluoroacetate (Annis et al. 1998; Annis et al. 1997). Recently, two mild reagents that rapidly and selectively oxidize protein thiols to disulfides in aqueous solution have been applied to protein folding.
W.J. Lees
128 O Selenoxide HO
Se
Se Red OH
HO
OH
6.6.1 Selenoxides Selenoxides can be used to rapidly oxidize protein thiols to protein disulfides. Once the formation of the protein disulfide bonds has taken place, the protein disulfide bonds can rearrange via intra- and intermolecular protein reactions to form more stable intermediates, provided residual thiols remain on the protein (Iwaoka et al. 2008). For example, within 60 s the addition of 3 equivalents of selenoxide to reduced RNase A at pH 8 resulted in a mixture of RNase A with an average of 3 disulfide bonds each (native RNase A contains 4 disulfide bonds). The rearrangement of disulfide bonds to form more stable intermediates was then followed. Other reagents that rapidly and irreversibly oxidize thiols to disulfides should have similar properties. Selenoxides also have the potential to be used as catalysts for the formation of protein disulfide bonds (Kumakura et al. 2010). In the presence of hydrogen peroxide, selenides can catalyze the formation of protein disulfides at a wide range of pH values, especially at low pH. The slow step in the reaction is the oxidation of the selenide to a selenoxide by hydrogen peroxide. RNase A (0.02 mM) in the presence of 1 equivalent of selenide and 100 equivalents of hydrogen peroxide will be completely oxidized in a couple of hours at pH 4 and 25°C.
6.6.2 Platinum-Based Compounds The rapid oxidation of thiols to disulfides can be accomplished selectively and efficiently using organometallic reagents. Using a slight excess of trans-[Pt(en)2Cl2]2+ a series of dithiol containing peptides was oxidized to the corresponding disulfides quantitatively at pH 4 in less than 2 h (Shi and Rabenstein 2000). The ability of [Pt(en)2Cl2]2+ to rapidly oxidize protein thiols to disulfides was used to develop a method to determine the native conformational tendencies of unfolded polypeptides (Narayan et al. 2003a). In less than 5 min at pH 4.7, fully reduced RNase A was completely oxidized with a twofold (or greater) excess of [Pt(en)2Cl2]2+ (on a per disulfide basis). However, the yield of native protein was only 0.6%, approximately what would be expected from a random distribution of the 104 possible four-disulfide bond containing isomers. The reagent was also used to directly oxidize stable folding intermediates missing one native disulfide bond (des intermediates) to native protein rapidly. This is potentially a slow process, as it may otherwise require the intramolecular rearrangement of disulfide bonds (Narayan et al. 2003b).
6 Oxidative Protein Folding with Small Molecules
129
6.6.3 Applications Rapid oxidizing reagents offer the possibility of enhancing the in vitro folding of disulfide containing proteins. Two possible scenarios are the addition of reagent at the beginning or the end of the folding process (Lu and Liu 2008). At the beginning, substoichiometric or stoichiometric amounts of reagent could be employed to rapidly form disulfide bonds and generate a mixture of folding intermediates potentially less prone to aggregation. Other less oxidizing reagents would then be added to rearrange the newly formed disulfide bonds to native disulfide bonds. If added at the end of the protein folding process, these reagents would rapidly and directly oxidize the des species that tend to accumulate as kinetic traps at the end of many protein folding pathways, e.g., RNase A, lysozyme, BPTI, to native protein. Under normal folding conditions, many des species would rearrange intramolecularly to other des species before being oxidized to native protein. An additional advantage of such reagents is that they do not form doubly mixed disulfides, which could hinder the direct oxidation pathway of des species. Scheraga et al. have demonstrated such an application with RNase A (Narayan et al. 2003b).
6.7 Conclusion An expanding number of redox active small molecules are being applied to oxidative protein folding problems both in vitro and in vivo. Improvements over traditional reagents such as glutathione started with peptide dithiols and then moved on to small molecule dithiols such as BMC. In general, dithiols improve oxidative protein folding yields both in vivo and in vitro but not the corresponding folding rates. The improvement in yields was proposed to be due to the prevention of aggregation early in the folding process, possibly due to the destabilizing of kinetic traps via the “intramolecular clock” or escape mechanism. Aromatic thiols, which significantly increase protein-folding rates, especially at neutral or lower pH values, as well as yields, were then introduced and have been shown to be generally applicable to in vitro protein folding. The greater nucleophilicity of aromatic thiolates and the low thiol pKa values of their corresponding thiols enhance reaction rates. Gemini disulfides offer the promising advantage of acting as chaperones as well as disulfide reagents. However, their potential is still unclear. Unlike disulfides, diselenides can catalyze both the air oxidation of thiols, otherwise a slow process, and disulfide bond rearrangements. The use of substoichiometric amounts of diselenides to catalyze the air oxidation of protein thiols has been elegantly employed both in vitro and in vivo. Diselenides, however, only slightly improve protein folding rates compared to disulfides during in vitro protein folding with RNase A, although more impressive results were obtained with BPTI, where the conversion of intermediates with two disulfide bonds to native protein with three disulfide bonds is accelerated.
130
W.J. Lees
The in vitro application of the current generation of small molecule redox reagents will continue to increase, and we expect that the tuning of redox potentials during protein folding will become more important (Lu and Liu 2008). Early on in protein folding, disulfide bond rearrangement is generally the slow step but as kinetic traps are formed, in many cases missing one native disulfide bond (a des species), direct oxidation becomes the rate-determining step. The in vivo application of small molecule redox reagents to the production of proteins in bacteria and yeast has already shown significant promise in raising yields and will become more generally applicable when more examples appear in the literature. As the importance of oxidative protein folding in human health becomes more evident, the use of the current generation of small molecule redox reagents, or the next, as drugs may become a reality.
References Annis I, Hargittai B, Barany G (1997) Disulfide bond formation in peptides. Methods Enzymol 289:198–221 Annis I, Chen L, Barany G (1998) Novel solid-phase reagents for facile formation of intramolecular disulfide bridges in peptides under mild conditions. J Am Chem Soc 120:7226–7238 Arolas JL, Aviles FX, Chang JY, Ventura S (2006) Folding of small disulfide-rich proteins: clarifying the puzzle. Trends Biochem Sci 31:292–301 Bass R, Ruddock LW, Klappa P, Freedman RB (2004) A major fraction of endoplasmic reticulumlocated glutathione is present as mixed disulfides with protein. J Biol Chem 279:5257–5262 Beld J, Woycechowsky KJ, Hilvert D (2007) Selenoglutathione: efficient oxidative protein folding by a diselenide. Biochemistry 46:5382–5390 Beld J, Woycechowsky KJ, Hilvert D (2008) Catalysis of oxidative protein folding by smallmolecule diselenides. Biochemistry 47:6985–6987 Beld J, Woycechowsky KJ, Hilvert D (2010) Small-molecule diselenides catalyze oxidative protein folding in vivo. ACS Chem Biol 5:177–182 Cabrele C, Cattani-Scholz A, Renner C, Behrendt R, Oesterhelt D, Moroder L (2002) Photomodulation of the redox and folding adjuvant properties of bis(cysteinyl) peptides. Eur. J. Org. Chem. 2144–2150 Cabrele C, Fiori S, Pegoraro S, Moroder L (2002) Redox-active cyclic bis(cysteinyl)peptides as catalysts for in vitro oxidative protein folding. Chem Biol 9:731–740 Cattani-Scholz A, Renner C, Cabrele C, Behrendt R, Oesterhelt D, Moroder L (2002) Photoresponsive cyclic bis(cysteinyl)peptides as catalysts of oxidative protein folding. Angew Chem Int Ed Engl 41:289–292 Chang JY, Li L (2005) Divergent folding pathways of two homologous proteins, BPTI and tick anticoagulant peptide: compartmentalization of folding intermediates and identification of kinetic traps. Arch Biochem Biophys 437:85–95 DeCollo TV, Lees WJ (2001) Effects of aromatic thiols on thiol-disulfide interchange reactions that occur during protein folding. J Org Chem 66:4244–4249 Fahnert B, Lilie H, Neubauer P (2004) Inclusion bodies: formation and utilisation. Adv Biochem Eng Biotechnol 89:93–142 Fink M, Nieves P, Chang S, Narayan M (2008) Non-redox-active small-molecules can accelerate oxidative protein folding by novel mechanisms. Biophys Chem 132:104–109 Gough JD, Lees WJ (2005) Increased catalytic activity of protein disulfide isomerase using aromatic thiol based redox buffers. Bioorg Med Chem Lett 15:777–781
6 Oxidative Protein Folding with Small Molecules
131
Gough JD, Williams RH Jr, Donofrio AE, Lees WJ (2002) Folding disulfide-containing proteins faster with an aromatic thiol. J Am Chem Soc 124:3885–3892 Gough JD, Gargano JM, Donofrio AE, Lees WJ (2003) Aromatic thiol pKa effects on the folding rate of a disulfide containing protein. Biochemistry 42:11787–11797 Gurbhele-Tupkar MC, Perez LR, Silva Y, Lees WJ (2008) Rate enhancement of the oxidative folding of lysozyme by the use of aromatic thiol containing redox buffers. Bioorg Med Chem 16:2579–2590 Hatahet F, Ruddock LW (2009) Protein disulfide isomerase: a critical evaluation of its function in disulfide bond formation. Antioxid Redox Signal 11:2807–2850 Hawkins HC, Freedman RB (1991) The reactivities and ionization properties of the active-site dithiol groups of mammalian protein disulfide-isomerase. Biochem J 275:335–339 Hevehan DL, Clark ED (1997) Oxidative renaturation of lysozyme at high concentrations. Biotechnol Bioeng 54:221–230 Hwang C, Sinskey AJ, Lodish HF (1992) Oxidized redox state of glutathione in the endoplasmicreticulum. Science 257:1496–1502 Ito K, Inaba K (2008) The disulfide bond formation (Dsb) system. Curr Opin Struct Biol 18:450–458 Iwaoka M, Kumakura F, Yoneda M, Nakahara T, Henmi K et al (2008) Direct observation of conformational folding coupled with disulphide rearrangement by using a water-soluble selenoxide reagent – a case of oxidative regeneration of ribonuclease a under weakly basic conditions. J Biochem 144:121–130 Kersteen EA, Raines RT (2003) Catalysis of protein folding by protein disulfide isomerase and small-molecule mimics. Antioxid Redox Signal 5:413–424 Kibria FM, Lees WJ (2008) Balancing conformational and oxidative kinetic traps during the folding of bovine pancreatic trypsin inhibitor (BPTI) with glutathione and glutathione disulfide. J Am Chem Soc 130:796–797 Kozlov G, Maattanen P, Thomas DY, Gehring K (2010) A structural overview of the PDI family of proteins. FEBS J 277:3924–3936 Kumakura F, Mishra B, Priyadarsini KI, Iwaoka M (2010) A water-soluble cyclic selenide with enhanced glutathione peroxidase-like catalytic activities. Eur. J. Org. Chem. 440–445 Lees WJ (2008) Small-molecule catalysts of oxidative protein folding. Curr Opin Chem Biol 12:740–745 Lees WJ, Whitesides GM (1993) Equilibrium-constants for thiol disulfide interchange reactions – a coherent, corrected set. J Org Chem 58:642–647 Lu D, Liu Z (2008) Dynamic redox environment-intensified disulfide bond shuffling for protein refolding in vitro: molecular simulation and experimental validation. J Phys Chem B 112:15127–15133 Lundstroem J, Holmgren A (1993) Determination of the reduction-oxidation potential of the thioredoxin-like domains of protein disulfide-isomerase from the equilibrium with glutathione and thioredoxin. Biochemistry 32:6649–6655 Lyles MM, Gilbert HF (1991) Catalysis of the oxidative folding of ribonuclease A by protein disulfide isomerase: dependence of the rate on the composition of the redox buffer. Biochemistry 30:613–619 Madar DJ, Patel AS, Lees WJ (2009) Comparison of the oxidative folding of lysozyme at a high protein concentration using aromatic thiols versus glutathione. J Biotechnol 142:214–219 Mainathambika BS, Bardwell JC (2008) Disulfide-linked protein folding pathways. Annu Rev Cell Dev Biol 24:211–235 Moroder L, Besse D, Musiol H-J, Rudolph-Boehner S, Sideler F (1996) Oxidative folding of cystinerich peptides versus regioselective cysteine pairing strategies. Biopolymers 40:207–234 Nakamura T, Lipton SA (2008) Emerging roles of S-nitrosylation in protein misfolding and neurodegenerative diseases. Antioxid Redox Signal 10:87–101 Narayan M, Welker E, Wedemeyer WJ, Scheraga HA (2000) Oxidative folding of proteins. Acc Chem Res 33:805–812
132
W.J. Lees
Narayan M, Welker E, Scheraga HA (2003a) Native conformational tendencies in unfolded polypeptides: development of a novel method to assess native conformational tendencies in the reduced forms of multiple disulfide-bonded proteins. J Am Chem Soc 125:2036–2037 Narayan M, Welker E, Wanjalla C, Xu GQ, Scheraga HA (2003b) Shifting the competition between the intramolecular reshuffling reaction and the direct oxidation reaction during the oxidative folding of kinetically trapped disulfide-insecure intermediates. Biochemistry 42: 10783–10789 Ookura T, Kainuma K, Kim H-J, Otaka A, Fujii N, Kawamura Y (1995) Active site peptides with CXXC motif on MAP-resin can mimic protein disulfide isomerase activity. Biochem Biophys Res Commun 213:746–751 Potempa M, Hafner M, Frech C (2010) Mechanism of gemini disulfide detergent mediated oxidative refolding of lysozyme in a new artificial chaperone system. Protein J 29:457–465 Sahdev S, Khattar SK, Saini KS (2008) Production of active eukaryotic proteins through bacterial expression systems: a review of the existing biotechnology strategies. Mol Cell Biochem 307:249–264 Shi T, Rabenstein DL (2000) Discovery of a highly selective and efficient reagent for formation of intramolecular disulfide bonds in peptides. J Am Chem Soc 122:6809–6815 Steinmann D, Nauser T, Koppenol WH (2010) Selenium and sulfur in exchange reactions: a comparative study. J Org Chem 75:6696–6699 Szajewski RP, Whitesides GM (1980) Rate constants and equilibrium-constants for thiol-disulfide interchange reactions involving oxidized glutathione. J Am Chem Soc 102:2011–2026 Walker KW, Gilbert HF (1997) Scanning and escape during protein-disulfide isomerase-assisted protein folding. J Biol Chem 272:8845–8848 Wedemeyer WJ, Welker E, Narayan M, Scheraga HA (2000) Disulfide bonds and protein folding. Biochemistry 39:4207–4216 Willis MS, Hogan JK, Prabhakar P, Liu X, Tsai K et al (2005) Investigation of protein refolding using a fractional factorial screen: a study of reagent effects and interactions. Protein Sci 14:1818–1826 Winter J, Lilie H, Rudolph R (2002) Recombinant expression and in vitro folding of proinsulin are stimulated by the synthetic dithiol Vectrase-P. FEMS Microbiol Lett 213:225–230 Woycechowsky KJ, Raines RT (2003) The CXC motif: a functional mimic of protein disulfide isomerase. Biochemistry 42:5387–5394 Woycechowsky KJ, Wittrup KD, Raines RT (1999) A small-molecule catalyst of protein folding in vitro and in vivo. Chem Biol 6:871–879
Chapter 7
Protein Disulfide Isomerase and the Catalysis of Oxidative Protein Folding Hiram F. Gilbert
Abstract For proteins that are processed in the eukaryotic endoplasmic reticulum and destined for the cell surface, the correct formation of protein disulfides is critical to their folding and function. Protein disulfide isomerase (PDI) is a resident of the endoplasmic reticulum that catalyzes disulfide formation and also provides mechanisms to correct folding mistakes. The protein consists of multiple thioredoxin domains with one or more active sites, generally in the sequence CGHC. The active site cycles between an oxidized (disulfide) state which can catalyze disulfide formation and a reduced (dithiol) state that can break disulfides that are incorrectly paired. During the folding of disulfide-containing proteins, the folding process must allow the proper cysteines to closely approach each other while allowing steric access to oxidants. Early in folding, disulfide formation is error-prone and a mechanism is required to break incorrect disulfides. In addition to catalyzing disulfide formation, PDI provides this critical proof-reading function as well. In this chapter, we consider the structural organization of PDIs, their chemical and catalytic properties, the factors that may affect specificity, and their role in the eukaryotic cell. Keywords Protein disulfide isomerase • Disulfide formation • Disulfide isomerization • Protein folding • Oxidation • Reduction
7.1 Oxidative Protein Folding In proteins that contain disulfide bonds, the relatively rapid conformational rearrangements that constitute folding, must be coupled with the formation of the correct disulfides. In the cell, the chemical process of disulfide formation is facilitated
H.F. Gilbert (*) Verna and Marrs McLean Department of Biochemistry and Molecular Biology, Baylor College of Medicine, Houston, TX 77030, USA e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_7, © Springer Science+Business Media, LLC 2011
133
134
H.F. Gilbert
by a folding assistant and catalyst, protein disulfide isomerase (PDI), which catalyzes a variety of thiol/disulfide exchange reactions to ensure that proper disulfides are formed (Gilbert 1990). The process, termed oxidative protein folding, begins with a protein in which all of the cysteines are reduced and ends, hopefully, with a folded protein in which the correct pairs of cysteines have been connected with a disulfide bond. In the cell, the reduced protein is co-translationally inserted into the lumen of the endoplasmic reticulum (ER) where it eventually acquires its correct disulfide bonding pattern before it exits the ER. A great deal of what is known about the details of the mechanisms of oxidative folding comes from studies of oxidative folding in vitro in the presence and absence of folding catalysts such as PDI. Before turning to a discussion of catalysis of correct disulfide formation, it is instructive to review key points of oxidative folding in the absence of catalysis. Disulfide bonds are covalent modifications that contribute to protein stability. Unlike the noncovalent forces that stabilize the folded state, disulfide formation is a chemical process that results in the formation of a covalent bond between two cysteine residues. The reaction is a net oxidation (loss of electrons) in which an electron acceptor (oxidant) becomes reduced as the two cysteine residues loose two electrons and form a covalent bond. In the absence of a reducing agent, the disulfide bond and the stability it imparts to the folded protein are irreversible. However, under the appropriate conditions, disulfide formation can become reversible in the presence of a reducing agent (Chang 2008). In the laboratory, molecular oxygen often serves as the oxidant; however, this process is irreversible in the absence of a reducing agent. Because there are often mistakes in disulfide pairing, efficient folding to the native disulfides often requires conditions where incorrect disulfides can be reduced. Conditions that support both oxidation and reduction are provided by a redox buffer, a mixture of a low molecular weight thiol and its corresponding disulfide. The most abundant low molecular weight thiol in cells is glutathione and a redox buffer consisting of glutathione (GSH) and its disulfide (GSSG) is often used to provide an oxidant (GSSG) while also providing a reductant (GSH) that can support the reduction of any incorrect disulfides through a reversible process of thiol–disulfide exchange (Fig. 7.1). In all of these reaction, the reactive species is the thiolate anion (RS−), so that the reaction rate increases with increasing pH up to the pKa of the thiol. During oxidative protein folding, the chemical process of disulfide formation is coupled to the noncovalent interactions that stabilize structure (and vice-versa). The thermodynamic cycle of Fig. 7.2 shows that the contribution of a disulfide to the stability of a given structure depends on the stability of the disulfide that forms (Gilbert 1990; Lin and Kim 1989). The converse is also true; stable noncovalent interactions will also stabilize a disulfide. This linkage between noncovalent interactions that direct protein folding and the stability of the resulting disulfide has consequences for the mechanism of oxidative protein folding since the most thermodynamically stable disulfides are those that are compatible with the fold that produces the most stable noncovalent interactions. Like other protein folding mechanisms, oxidative folding often proceeds through a complex mixture of intermediates in which portions of the structure have formed and others have not. In the case of oxidative folding, the connection of specific
7 Protein Disulfide Isomerase
135
Fig. 7.1 Reversible disulfide bond formation through thiol/disulfide exchange reactions with a glutathione redox buffer. Disulfide formation in a redox buffer is a reversible process. The position of the equilibrium will depend on the concentrations of GSSG and GSH2 and the overall equilibrium constant for the reaction, Kox. Kox may be viewed as a thermodynamic effective molarity of the two cysteines that form the disulfide. The larger Kox, the more stable the disulfide
Fig. 7.2 Relationship between the stability of a protein disulfide and its contribution to protein stability. The ratio of the disulfide bond stability in the native, folded state to the disulfide bond stability in the unfolded state is equivalent to the ratio of the equilibrium stabilities of the disulfide state to the reduced state. The harder a disulfide is to reduce thermodynamically the more it contributes to the stability of the protein
cysteines can serve as a long-lived covalent reporter of portions of the structure that are sufficiently close at different times during folding to enable the covalent linking of two cysteines (Creighton 1986). This approach has been very productive in helping to elucidate the pathway(s) by which disulfides form during oxidative folding. Chang has provided a recent comprehensive review of the variety of oxidative folding pathways displayed by various disulfide-containing proteins (Chang 2008). The distinguishing feature of the diverse routes to the native state is often the number of disulfide intermediates that form during the process. Some proteins, for example RNase A (Rothwarf et al. 1998) and hirudin (Chang 1994), accumulate a large number of disulfide isomers containing one, two, or three disulfides. During the early
136
H.F. Gilbert
stages of folding of these proteins, there is insufficient information in noncovalent interactions alone to specify correct disulfide pairing. However, in the cases of BPTI (Weissman and Kim 1991) and a-interferon (Lin and Chang 2007), there are a limited number of disulfide isomers with the major disulfide isomers having native disulfide pairings. In these cases, there appears to be sufficiently well-defined structure guided by noncovalent reactions to guide the formation of correct disulfides. Intramolecular disulfide formation in an entirely unfolded protein in denaturant is largely driven by the proximity of the two cysteine residues in the sequence. Oxidation of the unfolded protein under denaturing conditions produces a large, almost random assortment of disulfides (Galat et al. 1981). Under conditions that promote folding, more distant cysteines may be brought close enough together by noncovalent interactions to promote specific disulfide formation. In the case of RNase A (four disulfides), which has been studied in great detail, early disulfide formation populates a large number of intermediates with one, two, or three disulfides. However, in the rate-determining step, the large ensemble of 3S intermediates is converted to one of two native-like three-disulfide intermediates, missing the disulfide (65–72) or (40–95). The steps converting the various 3S intermediates to these native-like three disulfide intermediates involve rearrangement of disulfide bonds and proline isomerization (Saito et al. 2001). Recently, Gahl and Scheraga have compared the oxidative folding pathways of two structurally related proteins, RNase A and onconase. These two proteins have marginal (30%) sequence identity (Gahl and Scheraga 2009) but arrive at very similar folds. In onconase, the folding is more efficient due to an increased stability of specific one and two-disulfide intermediates which limits the appearance of the large ensemble of 3S intermediates observed in RNase A refolding. Thus, the increased contribution of noncovalent interactions to stability and conformational folding of folding intermediates can stabilize the formation of specific disulfides and direct the folding pathway. Limiting the number of incorrect disulfides that are formed may make disulfide formation more efficient; however, given the opportunity to break incorrectly paired disulfides, the protein will eventually form the most stable set of disulfides. In the case of bovine pancreatic trypsin inhibitor (BPTI, three disulfides), the oxidative folding pathway also involves multiple pathways for reaching the native disulfide arrangement (Weissman and Kim 1991; Creighton 1992; Goldenberg 1992). Unlike RNase A, the major intermediates that are observed are predominately those with native disulfides. The first disulfides to form are one of the two disulfide located in the hydrophobic core of the molecule. When one of these disulfides form, a second disulfide [14–38] forms rapidly producing a two-disulfide intermediate that has the remaining two cysteines trapped in the protected core of the protein and in the reduced state. After a very slow intramolecular rearrangement of the two disulfides to produce both of the core disulfides and reduce the [14–38] disulfide, the protein is rapidly oxidized to the native state (Creighton and Goldenberg 1984). The diversity of folding “pathways” for disulfide containing proteins suggests two consistent problems that arise during the initial stages of disulfide formation. The insufficiency of secondary structure and other noncovalent interactions to bring two correct cysteines into proximity results in a large number of folding intermediates,
7 Protein Disulfide Isomerase
137
often with non-native cysteine pairings. In these cases, the subsequent evolution of structure and native-cysteine pairings results from an interplay between noncovalent and disulfide-dependent stabilization of the native fold. In other cases, the correct disulfides are formed but in the wrong temporal order. Formation of a nativedisulfide(s) traps cysteines in a relatively stable structure (a “kinetic trap”) where one of the native disulfides must be broken before the native state can be formed. In both cases, the resolution of the problem requires a mechanism, either intramolecular or intermolecular to break incorrect disulfides whether they are structurally or temporally mis-paired. In the cell, catalysts like PDI not only catalyze disulfide formation but also provide an efficient mechanism to break these incorrect disulfides.
7.2 Protein Disulfide Isomerase Structure The PDI family is comprised of a large number of functionally related proteins, generally found in the eukaryotic endoplasmic reticulum (ER) (Ellgaard and Ruddock 2005). There is a related set of bacterial proteins involved in disulfide formation and isomerization in the periplasmic space and the reader is referred to an excellent, recent review of disulfide formation in prokaryotes (Nakamoto and Bardwell 2004). In one of the simplest eukaryotes, Saccharomyces cerevisiae, there are five genes that encode PDI-like proteins in the endoplasmic reticulum (Norgaard et al. 2001). Pdi1p is a classic PDI that consists of four tandem thioredoxin domains (Fig. 7.3) in which the N- and C-terminal thioredoxin domains (a and a¢) each contain an active site sequence CGHC (Edman et al. 1985). The two internal thioredoxin folds (b and b¢) are structural, noncatalytic domains, thought to provide for interactions with PDI substrates (Pirneskoski et al. 2004). The other family member in the yeast ER with a classical four-domain structure is Eug1p (Norgaard et al. 2001); however, the two active sites have sequences, CLHS and CIHS, that do not have the second more C-terminal active site cysteine. Eug1p is not a very effective catalyst of either disulfide formation or isomerization, but mutation of its active CXXS active sites to CXXC motifs dramatically increases its isomerase activity (Norgaard and Winther 2001). The in vivo function of Eug1p is not clear. Mpd1p and Mpd2p both have two thioredoxin domains, one catalytic and one structural. They are both normally expressed at very low levels; however, when overexpressed, both can rescue the lethal deletion of the Pdi1p (Norgaard et al. 2001). Mpd1p has recently been shown to interact with yeast calnexin (Kimura et al. 2005) and a single-cysteine mutant of Mpd2 (CQHA) has been found in association with the
Fig. 7.3 General domain architecture for protein disulfide isomerase. The similarity in sequence is greatest between the a and a¢ domains and the b and b¢ domains; however, each of the four domains has a thioredoxin fold. The c region is a negatively charged helical tail with an ER retention signal
138
H.F. Gilbert
yeast disulfide oxidase, Ero1p (Frand and Kaiser 1999) (see below) suggesting that these two proteins may provide some oxidase activity to help complement Pdi1p function, at least when the Pdi1p function is compromised. The final yeast PDI family member is Eps1p, which is a transmembrane protein with one thioredoxin domain. Its function is largely unknown. In higher eukaryotes, a more diverse set of PDI homologues is present that includes approximately 20 proteins. Some have a classical PDI arrangement of four thioredoxin domains (two catalytic and two structural domains) along with the classical CGHC active site (PDI, EPp57, PDIp) while others have one (ERp44, TMX, hAG, ERp18), two (P5), three (ERp72, Erp46, PDIr), or five (ERdj5) catalytic domains with variable numbers of noncatalytic domains ranging from one to two. In addition, there is considerable variation in the active site sequences among some of the members (Ellgaard and Ruddock 2005). The physiological function(s) of most of these homologues is not known, but what is known suggests that the differences in domain architecture and active site sequences affect the interactions of these proteins with substrates and other proteins. For example, Erp57 plays a critical role in the oxidative folding of glycosylated substrates and interacts with both calnexin and calreticulin through its b¢-like structural domain (Russell et al. 2004), although the calnexin/calreticulin interaction does not seem to affect the ER folding of MHC Class I molecules (Zhang et al. 2009). The pancreas-specific isozyme, PDIp interacts specifically with a number of different substrate peptides derived from proteins found only in the pancreas (Ruddock et al. 2000). Hatahet and Ruddock (2009) have recently provided an excellent and comprehensive review of the known functional properties of the significant PDI-homologues. Although there a number of structures of individual domains, full-length PDI proved relatively resistant to crystallization. Recently, crystal structures of a complete PDI molecule been reported. The two structures available reveal an interesting conformational flexibility. In the initial structure (Tian et al. 2006), the two catalytic domains and their active sites face each other over the structural b–b¢ domains forming a “twisted-U” conformation similar to that observed for the DsbC isomerase from Escherichia coli (McCarthy et al. 2000) with the two active sites of the a and a¢ domain facing each other over the base of the “U” formed by the b and b¢ domains. However a later, lower resolution structure obtained by crystallization at a higher temperature showed that the domain architecture now resembled a “boat” (Tian et al. 2008) (Fig. 7.4). Interestingly, in this conformation, the two active sites of the a and a¢ domains no longer face each other. Mutations that were designed to crosslink (through disulfides) the a and a¢ domains to the b¢ domain in the “U” conformation and restrict the conformational mobility of the two active site domains. Mutations introducing a stable disulfide to restrict the mobility of the a domain resulted in a 60–70% decrease in catalytic activity when assayed by the ability to catalyze the oxidative folding of RNase A. However, restricting the mobility of the a¢ domain resulted in only a 24% decrease in activity. While these results might suggest that the two catalytic domains act cooperatively in folding through flexible movements relative to each other, inactivation of either the a or a¢ domain by mutating the active site cysteines to alanine has shown in the mammalian protein that the activities of the two active sites are at best additive (Lyles and Gilbert 1994).
7 Protein Disulfide Isomerase
139
Fig. 7.4 Crystal structures of two alternative conformations of yeast PDI. In the “twisted U” conformation, the two active sites point toward a shared central scaffold created by the b and b¢ domains. In the boat conformation, the two active sites point in opposite directions. The PDB coordinates for 2B5E (Tian et al. 2006) and 3BOA (Tian et al. 2008) were used along with the Swiss-PdbViewer (http://www.expasy.ch/spdbv/mainpaige.html) to render the structures
Another structural study has suggested conformational mobility of the a¢ domain as an important contribution to the catalytic cycle. In a construct consisting of the b¢-domain with the relatively long (19 amino acids) linker between the b¢ and a¢ domains attached (the x-region) showed that the x-region peptide occupies a distinct site on the b¢ domain that had previously been associated with substrate binding (Nguyen et al. 2008). This led to the suggestion that the linker between b¢ and a¢ might normally occupy the protein–peptide binding site on the b¢ domain and be displaced upon substrate binding. This would provide an interesting mechanism that could link substrate binding to the b¢ domain to the release and subsequent increased mobility of the a¢ domain through displacement of the bound x-peptide linker by substrate. Mutations introduced into the full length PDI molecule that inhibit the binding of substrate were found to increase the amount of conformer where the x-peptide occupied the b¢ binding site.
7.3 Catalysis of Disulfide Formation Disulfide formation is a chemical oxidation requiring an oxidizing agent to remove electrons from two thiols and form a disulfide bond. During oxidative folding, the disulfides that form initially will depend on the kinetic reactivity of individual cysteine residues with the oxidant which, in turn, will be governed by the intrinsic chemical reactivity and pKa of the nucleophilic cysteine residue, the reactivity of the
140
H.F. Gilbert
Fig. 7.5 Kinetic partitioning of the PDI-S-S-Protein mixed disulfide intermediate formed after the attack of the nucleophilic cysteine of PDI on a substrate disulfide determines how the isomerization reaction occurs. If the substrate disulfide is stable then the effective molarity of the newly formed substrate cysteine will be high and disulfide will simply reform. This and steric accessibility make stable disulfides poor PDI substrates. If the intermediate can rearrange in an intramolecular reaction in which a rearranged cysteine expels reduced PDI from the complex faster than the resolving cysteine can displace it, the catalyzed isomerization will involve an intramolecular protein rearrangement. Otherwise, the resolving cysteine displaces the substrate, preventing PDI from becoming stuck in a slowly rearranging complex and reducing the substrate disulfide
oxidant, steric accessibility of the cysteine to the oxidant, and the proximity of a second cysteine (Fig. 7.5) (Jensen et al. 2009). The folding assistant and catalyst, PDI, catalyzes the initial formation of disulfides through the high chemical reactivity of its active site disulfide but it has little effect on which disulfides are formed, i.e., PDI adds no new “information” that helps specify which disulfides to connect. The active site disulfide forms of PDI are very good oxidants (poor reductants) thermodynamically. The equilibrium constant (Kox) for forming the active site disulfide using glutathione disulfide as an oxidant is between 1 and 3 mM (−180 mV) (Lundstrom and Holmgren 1993; Darby and Creighton 1995). By comparison, the equilibrium constant for oxidizing the two active site thiols of thioredoxin is 1.4 × 103 M (−270 mV) (Aslund et al. 1997). Thus, on a thermodynamic basis, the active site disulfide of PDI is a better oxidant than the thioredoxin active site by a factor of 103. While both proteins have a very similar overall structure, the active site environments modulate the stability of the active site disulfide considerably. The stability of thioredoxin-like active sites (CXXC) depends on the sequence between the two cysteine residues and nearby charged residues that can modulate the pKa (Grauschopf et al. 1995). Other structural factors that can affect the stability of the
7 Protein Disulfide Isomerase
141
active site disulfide include other disulfides. In the a-domain of yeast PDI, the presence of a second structural disulfide is used to destabilize the active site disulfide, making it a better oxidant thermodynamically by 18-fold (Wilkinson et al. 2005). The thermodynamic instability of the active site disulfide of oxidized PDI is also reflected in its kinetic reactivity. Walker and Gilbert (1995) used a b-lactamase model in which folding of reduced, denatured b-lactamase was used to set up a competition between folding and burying its two cysteines and forming a disulfide through oxidation by GSSG or PDI. In this system, PDI was kinetically a better oxidant than GSSG by approximately 500-fold. Studies with the individual catalytic domains of human PDI expressed and purified as individual proteins show that the a and a¢ domains still catalyze oxidation as well as they do when they are present in full-length PDI and are individually about 500-fold more reactive than low molecular weight disulfides (Darby and Creighton 1995). Thus, there is little specificity in disulfide formation catalyzed by PDI, and its ability to rapidly oxidize protein and nonprotein thiols does not depend significantly on the presence of the b or b¢ domains, which provide for interactions between PDI and its larger protein substrates. When exposed to reduced, denatured RNase (Shin and Scheraga 2000) or BPTI (Weissman and Kim 1993), in the presence of PDI, the initial disulfides that form are essentially the same as those formed in the absence of PDI. Thus, PDI does not significantly affect the overall mechanism of oxidative folding, including which protein cysteines are paired, it simply accelerates the thiol–disulfide exchange reactions that would occur in its absence.
7.4 Catalysis of Disulfide Isomerization Perhaps the most defining chemistry catalyzed by PDI during oxidative protein folding lies in its ability to promote the isomerization of protein disulfides. The error-prone nature of disulfide oxidation and the inability of PDI to initially oxidize the cysteines in their correct spatial or temporal pairing, make it necessary to have some mechanism for correcting these incorrect disulfides – a proofreading activity. The disulfide isomerase reaction of PDI, which rearranges the disulfide connectivity without a net change in the disulfide redox state of the substrate protein, provides this mechanism. When provided with a substrate of scrambled RNase (or other proteins), PDI rapidly catalyzes the reshuffling of the disulfides to yield the native disulfide arrangement. It was, in fact, this ability that led Anfinsen to the original discovery and naming of PDI (Fuchs et al. 1967). PDI catalysis of protein disulfide formation is achieved largely by the high chemical reactivity of its active site. The more N-terminal active site cysteines of the reduced PDI active sites are also hyper-reactive with disulfide substrates, principally due to a lower pKa (Darby and Creighton 1995), but this reactivity alone is insufficient to promote disulfide isomerization – the noncatalytic b and b¢ domains are essential to PDI isomerase activity. Domain deletion mutants show that the only species that have significant isomerase activity (13–20% of wild type isomerase activity) contains at
142
H.F. Gilbert
least the b¢ and a¢-domains (Darby et al. 1998; Xiao et al. 2001). The activity of the b¢–a¢ construct in catalyzing isomerization is essentially identical to the activity observed in the full-length PDI molecule with the a¢ domain active site cysteines mutated to serine (Xiao et al. 2001). The b¢ domain of PDI is involved in the interaction with unfolded protein and peptide substrates. Small peptides and proteins inhibit PDI-catalyzed isomerization suggesting a binding site for peptides (Morjana and Gilbert 1991). The b¢ domain is labeled by peptides with a radioactive photoaffinity label, but other sites may contribute to the interactions with larger substrates (Klappa et al. 1998). Mutations in the b¢ domain also result in decreased binding of peptides as well as decreased isomerization activity, and the b¢-x-peptide linker crystal structure reveals a site on the b¢ domain that may be opened by a conformation change to reveal a peptide binding site during the catalytic cycle (Nguyen et al. 2008). Despite the presence of one or more peptide–protein binding sites to interact with substrates, PDI displays little sequence preference in binding peptides. A variety of peptides inhibits PDIcatalyzed isomerization with little sequence specificity and in direct proportion to their length (Morjana and Gilbert 1991). Using a kinetic assay to determine which peptide-disulfides react fastest with a reduced PDI active site, Winther and colleagues (Westphal et al. 1998) found a weak consensus sequence of “small/helix breaker-cysteine-X-hydrophobic/basic-hydrophobic.” Since there are few simple sequence cues that reveal which disulfides form during folding, it is not surprising that PDI has developed interaction sites that show little specificity. This enables PDI to act on a large number of substrate disulfides. The rather weak binding to its substrates (in the high mM to mM range) may be compensated for by the very high concentration of PDI in the endoplasmic reticulum. The initiation of PDI catalyzed protein disulfide isomerization requires a reduced active site cysteine (Fig. 7.5). In each active site, the cysteine of the CGHC motif that is closer to the protein’s N terminus is the nucleophilic cysteine. It has a low pKa and is solvent exposed (Kortemme et al. 1996). The more C-terminal cysteine residue of the active site is buried and normally unreactive with substrate – its reactions are confined to reacting with the nucleophilic cysteine within the same active site (Darby and Creighton 1995). After the initial formation of a mixed disulfide between the protein disulfide substrate and the reduced PDI active site, the more C-terminal cysteine acts as a “resolving” nucleophile and determines the kinetic fate of the PDI-S-S-Protein intermediate. The resolving cysteine serves as a “clock” to help control the maximum lifetime of the intermediate (Walker and Gilbert 1997). The simplest fate of the PDI-S-S-Protein intermediate is to reverse the reaction and reform the original substrate disulfide. This will not have an effect on how or if isomerization occurs, but if reversal is rapid, the substrate disulfide will be reduced slowly and resistant to isomerization. If a pathway exists in the intermediate for an intramolecular rearrangement of the substrates disulfides while bound to PDI, the rearrangement must result in a final displacement of the PDI by attack of the newly released protein sulfhydryl group to regenerate the reduced active site and release the rearranged substrate. If this intramolecular isomerization pathway is too slow, the “resolving” cysteine at the active site will displace the substrate and form an
7 Protein Disulfide Isomerase
143
oxidized PDI active site and a reduced protein. This will not only allow PDI to escape from a slowly rearranging substrate but would also allow the protein an additional attempt to form a correct disulfide. Given the variety of oxidative folding mechanisms (Chang 2008), it is likely that the various folding intermediates will have a variety of intramolecular rearrangement and oxidation/reduction pathways available. At least for two proteins, RNase A and BPTI, the reduction and subsequent reoxidation are a major mechanism for disulfide isomerization. In the case of the isomerization of scrambled RNase A, mutation of PDI’s resolving cysteine to serine results in a significant loss of isomerase activity, an unexpected result if an intramolecular isomerization pathway is important in the rate-limiting step (Schwaller et al. 2003). In addition, the formation of native RNase from scrambled RNase requires some oxidized PDI, again suggesting that a significant fraction of the substrate is released from the active site in the reduced form and must be reoxidized to reach the native state. With the [5–55:14–38] kinetically trapped intermediate in BPTI refolding, addition of reduced PDI in the absence of any redox buffer results in the rapid reduction of the [14–38] disulfide to generate oxidized PDI and reduced BPTI. The oxidized PDI then donates its disulfide to help form alternative two disulfide intermediates until all of the two disulfide intermediates disappear and a mixture of native (threedisulfide) and one-disulfide species are formed (Wilkinson and Gilbert unpublished). In both these cases, substrate reduction involving displacement of the PDI-S-S-Protein by the resolving cysteine is a rapid pathway and suggests that PDI catalyzes isomerization by providing the substrate with multiple attempts to reach the correct arrangement of disulfide. If a folding attempt results in an incorrect disulfide leading to a kinetic trap, PDI can simply reduce the offending disulfide and allow the reduced protein to try again.
7.5 Oxidative Protein Folding in the Cell In the endoplasmic reticulum, the pathways that introduce disulfides into folding proteins require a source of oxidizing equivalents and a mechanism to deliver them to proteins that are co-translationally inserted into the ER (Fig. 7.6). The Kaiser (Frand and Kaiser 1998) and Wiseman (Pollard et al. 1998) laboratories simultaneously reported the discovery of the source of oxidizing equivalents in the yeast ER, a flavin-dependent oxidase, Ero1p. In higher eukaryotes, there are two Ero1 isozymes, both of which are induced by the unfolded protein response (Cabibbo et al. 2000; Pagani et al. 2000). Ero1-La is expressed in most cell types while Ero1-Lb is expressed largely in the stomach and pancreas (Dias-Gunasekara et al. 2005). With the individual a and a¢ domains of PDI, Ero1 oxidizes each of them equally effectively. However, in the full-length molecule, the a¢ domain is a better substrate for Ero1p than the a-domain, and the presence of substrate further inhibits the oxidation of the a-domain. This suggests that there may be different functions for the two catalytic PDI domains in the cell – the a¢ domain serves as the major oxidase to insert
144
H.F. Gilbert
Fig. 7.6 Transfer of electrons during disulfide formation in the endoplasmic reticulum. The immediate acceptor of electrons from substrate cysteines is performed by PDI. The electrons are subsequently transferred to the active site of Ero1, where they move from one Ero1 disulfide to another through an intramolecular transfer and are then deposited onto an oxidized flavin which is normally oxidized by oxygen yielding hydrogen peroxide. In the absence of oxygen other substrates may serve as the terminal electron acceptor
new disulfides into folding proteins while the a-domain functions to reduce or isomerize incorrect disulfides (Kulp et al. 2006). When a mutant PDI missing the resolving cysteine (nearer the C-terminus) is introduced into the yeast ER, PDI is found associated with Ero1p through a protein–protein disulfide, suggesting that the transfer of electrons can occur directly between Ero1p and PDI (Frand and Kaiser 1999). The structure of Ero1p reveals an active site with two pairs of disulfides, C100–C105 and C352–C355 (Gross et al. 2004). The disulfide (C100–C105) accepts electrons directly from the PDI a and a¢ active site dithiols to oxidize the PDI active sites. Subsequently, the Ero1p C352–C355 disulfide accepts the electrons from the C100–C105 dithiol and then transfers them to the flavin cofactor (Sevier and Kaiser 2006). In the presence of oxygen, the flavin then transfers the electrons to oxygen to form hydrogen peroxide. However, in the absence of oxygen, disulfides can still be formed using either stoichiometric flavin or a variety of other electron acceptors that can accept electrons from the flavin (Gross et al. 2006). In vitro, efficient oxidative folding requires a redox environment that can support disulfide formation and reduce incorrect disulfides. The oxidative folding of RNase A occurs fastest with a GSH concentration of 1 mM and a GSSG concentration of 0.2 mM (Lyles and Gilbert 1991). The relatively high concentration of GSH compared to GSSG is needed to maintain a sufficient concentration of PDI in the reduced state to break incorrect disulfides. Hwang et al. (1992) used the retrieval of a cysteine-containing glycoprotein marker that visited the ER/Golgi after being taken up by the cell to show that a glutathione mixed disulfide was found in conjunction with ER/Golgi modifications to the carbohydrate. From the ratio of reduced glutathionylated peptide, it was concluded that the GSH/GSSG ratio of the ER was approximately 5–10, near the redox optimum determined for PDI. Using a gel-shift
7 Protein Disulfide Isomerase
145
assay that results in a 15-kDa shift in apparent molecular mass on SDS-PAGE, Wilkinson et al. observed the redox state of the PDI active sites in the yeast ER and found that the PDI active sites were approximately half-reduced, consistent with the need for catalyzing oxidation and reduction (Xiao et al. 2004). The thiol/disulfide redox state of the ER is maintained by a complex interplay between the insertion of reduced protein substrates into the ER, oxidation of Erop1, and the transfer of oxidizing and reducing equivalents to PDI substrates. A conditional allele of Ero1p can be complemented by the deletion of the GSH1 gene, suggesting that low GSH levels can rescue a defect in Ero1p oxidation and may normally serve to keep the redox state of Ero1p at the appropriate level (Cuozzo and Kaiser 1999). Sitia and coworkers have shown in permeabialized mammalian cells, that correct disulfide formation in the ER can be restored by adding GSH to the medium and that there is a link between the activity/amount of Ero1p present and the amount of GSH needed to maintain proper ER oxidative folding (Molteni et al. 2004). The availability of fluorescent disulfide mutants of GFP as probes of the ER thiol disulfide redox state confirms that the ER is considerably more oxidizing than the cytosol; however, measuring the exact redox state will require probes that are more difficult to oxidize than those available (Bjornberg et al. 2006). Nevertheless, these probes have shown changes in ER redox state in response to induction of the unfolded protein response or the addition of exogeneous oxidants and reductants that suggest the importance of mechanisms to maintain a redox environment in the ER that is conducive to folding (Merksamer et al. 2008; Dooley et al. 2004). There have been only a few detailed mechanistic studies in cells to see if the overall mechanism of disulfide formation is comparable to what is observed in vitro. Ruddon and coworkers used in vivo pulse-chase labeling followed by detailed peptide mapping by HPLC to show that the oxidative folding pathway of human chorionic gonadotropic in the cell is populated by the same intermediates observed in vitro (Huth et al. 1993). Disulfide formation can begin while the nascent peptide chain is still attached to the ribosome (Bergman and Kuehl 1979) or disulfide formation may be delayed. The LDL receptor (30-disulfides) folds in the ER through early intermediates characterized by dithiothreitol (DTT)-sensitive disulfides that become DTT resistant as the LDL receptor folds into its native state. Although, the LDL receptor is organized in discrete multiple domains with only intradomain disulfides, the initial folding intermediates connect very distal parts of the structure indicating that disulfides do not form in order as the protein is translocated into the ER (Jansens et al. 2002). Thus, it would appear that the variety of oxidative folding mechanisms observed in the catalyzed and uncatalyzed oxidative folding of proteins in vitro is likely to occur in the cell as well. Even with the complement of folding catalysts and chaperones of the eukaryotic endoplasmic reticulum, the system adds little new “information” to direct folding along pathways not initially specified in the primary sequence of the substrate protein. In vitro, the distinguishing activity of PDI is its disulfide isomerase activity. Other thioredoxin homologues and the individual catalytic a and a¢ domains themselves can put disulfides into proteins; however, all organisms from S. cerevisiae to man have one or more multidomain PDI molecules with a conserved four-domain
146
H.F. Gilbert
structure and isomerase activity. It has recently been possible to cleanly separate the isomerase and oxidase activities of PDI to determine which of these two activities makes PDI an essential gene in yeast. Surprisingly, expressing just the a or a¢ domains from the normal PDI1 promoter is sufficient to rescue the lethal deletion of the PDI1 gene in yeast (Solovyov et al. 2004) despite the fact that the a or a¢ domains have <5% of the normal isomerase activity of yeast Pdi1p. Titration of the expression Pdi1p and the a¢-domain using the GAL1-10 promoter revealed that wild-type growth in rich medium requires at least 60% of the oxidase activity of PDI but less than 6% of its isomerase activity (Solovyov et al. 2004). In yeast, PDI’s isomerase activity is not its essential activity in the cell despite the fact that some proteins do require isomerization to exit the ER correctly folded. This raises the possibility that there are mechanisms available in yeast, and presumably higher eukaryotes, to compensate for a defect in disulfide isomerization pathways in the ER. Xiao et al. (2004) found that even when all of the ER homologues of PDI have been deleted from yeast, the expression of the a¢ domain under the control of the PDI1 promoter is sufficient to rescue the lethal PDI1 deletion. Although the growth rate of this multiple deletion mutant was near wild-type, disulfide isomerization of yeast carboxypeptidase Y (CPY) was significantly decreased compared to wild-type isomerization rates. It is possible that in yeast proteins that are required for growth and survival, including nutrient transporters, have evolved not to have disulfides that form rapidly and correctly in the absence of PDI. However, other mechanisms may exist that compensate for a slower exit of disulfide-containing proteins from the ER. A recent screen for genes that become essential when Pdi1p is replaced by the isomerase deficient a¢-domain suggests (Kim et al. 2009) detecting a large number of genes that are involved in regulating vesicle trafficking, including endocytosis and secretion pathways. Mechanisms may exist that would coordinate production, secretion, and degradation of proteins to help compensate for defects in folding in the ER. For example, a slow exit of a disulfide-containing protein from the ER would tend to decrease the steady-state levels of disulfide-containing proteins at the yeast plasma membrane; however, a corresponding decrease in the endocytosis and degradation of plasma membrane proteins would help compensate for this defect and maintain near normal levels of disulfide-containing proteins.
References Aslund F, Berndt KD, Holmgren A (1997) Redox potentials of glutaredoxins and other thiol- disulfide oxidoreductase of the thioredoxin superfamily determined by direct protein-protein redox equilibria. J Biol Chem 272:20780–30786 Bergman LW, Kuehl WM (1979) Formation of an intrachain disulfide bond on nascent immunoglobulin light chains. J Biol Chem 254:8869–8876 Bjornberg OI, Ostergaard H, Winther JR (2006) Measuring intracellular redox conditions using GFP-based sensors. Antioxid Redox Signal 8:354–361 Cabibbo A, Pagani M, Fabbri M, Rocchi M, Farmery MR, Bulleid NJ, Sitia R (2000) ERO1-L, a human protein that favors disulfide bond formation in the endoplasmic reticulum. J Biol Chem 275:4827–4833
7 Protein Disulfide Isomerase
147
Chang J-Y (1994) Controlling the speed of hirudin folding. Biochem J 300:643–650 Chang J-Y (2008) Diversity of folding pathways and folding models of disulfide proteins. Antioxid Redox Signal 10:171–177 Creighton TE, Goldenberg DP (1984) Kinetic role of a meta-stable native-like two-disulphide species in the folding transition of bovine pancreatic trypsin inhibitor. J Mol Biol 179:497–526 Creighton TE (1986) Disulfide bonds as probes of protein folding pathways. Meth Enzymol 131:83–106 Creighton TE (1992) The disulfide folding pathway of BPTI. Science 256:111–114 Cuozzo JW, Kaiser CA (1999) Competition between glutathione and protein thiols for disulphidebond formation. Nat Cell Biol 1:130–135 Darby NJ, Creighton TE (1995) Characterization of the active site cysteine residues of the thioredoxinlike domains of protein disulfide isomerase. Biochemistry 34:16770–16780 Darby NJ, Penka E, Vincentelli R (1998) The multi-domain structure of protein disulfide isomerase is essential for high catalytic efficiency. J Mol Biol 276:239–247 Dias-Gunasekara S, Gubbens J, van Lith M, Dunne C, Williams JA, Kataky R, Scoones D, Lapthorn A, Bulleid NJ, Benham AM (2005) Tissue-specific expression and dimerization of the endoplasmic reticulum oxidoreductase Ero1beta. J Biol Chem 280:33066–33075 Dooley CT, Dore TM, Hanson GT, Jackson WC, Remington SJ, Tsien RY (2004) Imaging dynamic redox changes in mammalian cells with green fluorescent protein indicators. J Biol Chem 279:22284–22293 Edman JC, Ellis L, Blancher RW, Roth RA, Rutter WJ (1985) Sequence of protein disulphide isomerase and implications of its relationship to thioredoxin. Nature 317:267–270 Ellgaard L, Ruddock LW (2005) The human protein disulphide isomerase family: substrate interactions and functional properties. EMBO Rep 6:28–32 Frand AR, Kaiser CA (1998) The ERO1 gene of yeast is required for oxidation of protein dithiols in the endoplasmic reticulum. Mol Cell 1:161–170 Frand AR, Kaiser CA (1999) Ero1p oxidizes protein disulfide isomerase in a pathway for disulfide bond formation in the endoplasmic reticulum. Mol Cell 4:469–477 Fuchs S, De Lorenzo F, Anfinsen CB (1967) Studies on the mechanism of the enzymic catalysis of disulfide interchange in proteins. J Biol Chem 242:398–402 Gahl RF, Scheraga HA (2009) Oxidative folding pathway of onconase, a ribonuclease homologue: insight into oxidative folding mechanisms from a study of two homologues. Biochemistry 48:2740–2751 Galat A, Creighton TE, Lord RC, Blout ER (1981) Circular dichroism, Raman spectroscopy, and gel filtration of trapped folding intermediates of ribonuclease. BiocheCircular dichroism, Raman spectroscopy, and gel filtration of trapped folding intermediates of ribonuclease. Biochemistry 3:594–601 Gilbert HF (1990) Molecular and cellular aspects of thiol/disulfide exchange. Adv Enzymol 63:69–172 Goldenberg DP (1992) Native and non-native intermediates in the BPTI folding pathway. Trends Biochem Sci 17:257–261 Grauschopf U, Winther JR, Korber P, Zander T, Dallinger P, Bardwell JC (1995) Why is DsbA such an oxidizing disulfide catalyst? Cell 83:947–955 Gross E, Sevier CS, Heldman N, Vitu E, Bentzur M, Kaiser CA, Thorpe C, Fass D (2006) Generating disulfides enzymatically: reaction products and electron acceptors of the endoplasmic reticulum thiol oxidase Ero1p. Proc Natl Acad Sci USA 103:299–304 Gross E, Kastner DB, Kaiser CA, Fass D (2004) Structure of Ero1p, source of disulfide bonds for oxidative protein folding in the cell. Cell 117:601–610 Hatahet F, Ruddock LW (2009) Protein disulfide isomerase: a critical evaluation of its function in disulfide bond formation. Antioxid Redox Signal 11:2807–2850 Huth JR, Perini F, Lockridge O, Bedows E, Ruddon RW (1993) Protein folding and assembly in vitro parallel intracellular folding and assembly. Catalysis of folding and assembly of the human chorionic gonadotropin alpha beta dimer by protein disulfide isomerase. J Biol Chem 268:16472–16482
148
H.F. Gilbert
Hwang C, Sinskey AJ, Lodish HF (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257:1496–1502 Jansens A, van Duijn E, Braakman I (2002) Coordinated nonvectorial folding in a newly synthesized multidomain protein. Science 298:2401–2403 Jensen KS, Hansen RE, Winther JR (2009) Kinetic and thermodynamic aspects of cellular thioldisulfide redox regulation. Antioxid Redox Signal 11:1047–1058 Kim JH, Zhao Y, Pan X, He X, Gilbert HF (2009) The unfolded protein response is necessary but not sufficient to compensate for defects in disulfide isomerization. J Biol Chem 284:10400–10408 Kimura T, Hosoda Y, Sato Y, Kitamura Y, Ikeda T, Horibe T, Kikuchi M (2005) Interactions among yeast protein-disulfide isomerase proteins and endoplasmic reticulum chaperone proteins influence their activities. J Biol Chem 280:31438–31441 Klappa P, Ruddock LW, Darby NJ, Freedman RB (1998) The b¢ domain provides the principal peptide-binding site of protein disulfide isomerase but all domains contribute to binding of misfolded proteins. EMBO J 17:927–935 Kortemme T, Darby NJ, Creighton TE (1996) Electrostatic interactions in the active site of the N-terminal thioredoxin like domain of protein disulfide isomerase. Biochemistry 35:14503–14511 Kulp MS, Frickel E-M, Ellgaard L, Weissman JS (2006) Domain architecture of protein-disulfide isomerase facilitates its dual role as an oxidase and an isomerase in Ero1p-mediated disulfide formation. J Biol Chem 281:876–884 Lin TY, Kim PS (1989) Urea dependence of thiol-disulfide equilibria in thioredoxin: confirmation of the linkage relationship and a sensitive assay for structure. Biochemistry 13:5282–5287 Lin CJ, Chang J-Y (2007) Pathway of oxidative folding of bovine alpha-interferon: predominance of native disulfide bonded folding intermediates. Biochemistry 46:3925–3932 Lundstrom J, Holmgren A (1993) Determination of the reduction-oxidation potential of the thioredoxin-like domains of protein disulfide-isomerase from the equilibrium with glutathione and thioredoxin. Biochemistry 32:6649–6655 Lyles MM, Gilbert HF (1991) Catalysis of the oxidative folding of ribonuclease A by protein disulfide isomerase: Dependence of the rate on the composition of the redox buffer. Biochemistry 30:613–619 Lyles MM, Gilbert HF (1994) Mutations in the thioredoxin sites of protein disulfide isomerase reveal functional non-equivalence of the N- and C-terminal domains. J Biol Chem 269:30946–30952 McCarthy AA, Haebel PW, Törrönen A, Rybin V, Baker EN, Metcalf P (2000) Crystal structure of the protein disulfide bond isomerase, DsbC, from Escherichia coli. Nat Struct Biol 7:196–199 Merksamer PI, Trusina A, Papa FR (2008) Real-time redox measurements during endoplasmic reticulum stress reveal interlinked protein folding functions. Cell 135:933–947 Molteni SN, Fassio A, Ciriolo MR, Filomeni G, Pasqualetto E, Fagioli C, Sitia R (2004) Glutathione limits Ero1-dependent oxidation in the endoplasmic reticulum. J Biol Chem 279:32667–32673 Morjana NA, Gilbert HF (1991) Effect of protein and peptide inhibitors on the activity of protein disulfide isomerase. Biochemistry 30:4985–4990 Nakamoto H, Bardwell JC (2004) Catalysis of disulfide bond formation and isomerization in the Escherichia coli periplasm. Biochim Biophys Acta 1783:520–529 Nguyen VD, Wallis K, Howard MJ, Happalainen AM, Salo KEH, Saaranen MJ, Sidhu A, Wierenga RK, Freedman RB, Ruddock LW, Williamson RA (2008) Alternative conformations of the x region of human protein disulphide isomerase modulate exposure of the substrate binding b¢ domain. J Mol Biol 383:1144–1155 Norgaard P, Winther JR (2001) Mutation of yeast Eug1p CXXS active sites to CXXC results in a dramatic increase in protein disulphide isomerase activity. Biochem J 358:269–274 Norgaard P, Westphal V, Tachibana C, Alsoe L, Holst B, Winther JR (2001) Functional differences in yeast protein disulfide isomerases. J Cell Biol 152:553–562 Pagani M, Fabbri M, Benedetti C, Fassio A, Pilati S, Bulleid NJ, Cabibbo A, Sitia R (2000) Endoplasmic reticulum oxidoreductin 1-lbeta (ERO1-Lbeta), a human gene induced in the course of the unfolded protein response. J Biol Chem 275:23685–23692 Pirneskoski A, Klappa P, Lobell M, Williamson RA, Byrne L, Alanen HI, Salo KE, Kivirikko KI, Freedman RB, Ruddock LW (2004) Molecular characterization of the principal substrate
7 Protein Disulfide Isomerase
149
binding site of the ubiquitous folding catalyst protein disulfide isomerase. J Biol Chem 279:10374–10381 Pollard MG, Travers KJ, Weissman JS (1998) ERO1p: a novel and ubiquitous protein with an essential role in oxidative protein folding in the endoplasmic reticulum. Mol Cell 1:171–182 Rothwarf DM, Li Y-J, Scheraga HA (1998) Regeneration of bovine pancreatic ribonuclease a: detailed kinetic analysis of two independent folding pathways. Biochemistry 37:3767–3776 Ruddock LW, Freedman RB, Klappa P (2000) Specificity in substrate binding by protein folding catalysts: tyrosine and tryptophan residues are the recognition motifs for the binding of peptides to the pancrease-specific protein disulfide isomerase PDIp. Protein Sci 9:758–764 Russell SJ, Ruddock LW, Salo KE, Oliver JD, Roebuck OP, Llewellyn DH, Roderick HL, Koivunen P, Myllyharju J, High S (2004) The primary substrate binding site in the b0 domain of ERp57 is adapted for endoplasmic reticulum lectin association. J Biol Chem 279:18861–18869 Saito K, Welker E, Scheraga HA (2001) Folding of a disulfide-bonded protein species with free thiol(s): competition between conformational folding and disulfide reshuffling in an intermediate of bovine pancreatic ribonuclease A. Biochemistry 40:15002–15008 Schwaller MF, Wilkinson B, Gilbert HF (2003) Reduction/reoxidation cycles contribute to catalysis of disulfide isomerization by protein disulfide isomerase. J Biol Chem 278:7154–7159 Sevier CS, Kaiser CA (2006) Disulfide transfer between two conserved cysteine pairs imparts selectivity to protein oxidation by Ero1. Mol Biol Cell 17:2256–2266 Shin HC, Scheraga HA (2000) Catalysis of the oxidative folding of bovine pancreatic ribonuclease A by protein disulfide isomerase. J Mol Biol 300:995–1003 Solovyov A, Xiao R, Gilbert HF (2004) Sulfhydryl oxidation, not disulfide isomerization, is the principal function of protein disulfide isomerase in yeast Saccharomyces cerevisiae. J Biol Chem 279:34095–34100 Tian G, Kober F-X, Lewandrowski U, Sickmann A, Lennarz WJ, Schindelin H (2008) The catalytic activity of protein disulfide isomerase requires a conformationally flexible molecule. J Biol Chem 283:33630–33640 Tian G, Xiang S, Noiva R, Lennarz WJ, Schindelin H (2006) The crystal structure of yeast protein disulfide isomerase suggests cooperativity between its active sites. Cell 124:61–73 Walker KW, Gilbert HF (1997) Scanning and escape scanning and escape during protein disulfide isomerase-assisted protein folding. J Biol Chem 272:8845–8848 Walker KW, Gilbert HF (1995) Oxidation of kinetically trapped thiols by protein disulfide isomerase. Biochemistry 34:13642–13650 Weissman JS, Kim PS (1991) Re-examination of the folding of BPTI: predominance of native intermediates. Science 253:1386–1393 Weissman JS, Kim PS (1993) Efficient catalysis of disulfphide bond rearrangements by protein disulphide isomerase. Nature 365:185–188 Westphal V, Spetzler JC, Meldal M, Christensen U, Winther JR (1998) Kinetic analysis of the mechanism and specificity of protein-disulfide isomerase using fluorescence-quenched peptides. J Biol Chem 273:24992–24999 Wilkinson B, Xiao R, Gilbert HF (2005) A structural disulfide of yeast protein disulfide isomerase destabilizes the active site disulfide of the N-terminal thioredoxin domain. J Biol Chem 280:11483–11487 Xiao R, Solovyov A, Gilbert HF, Holmgren A, Lundström-Ljung J (2001) Combinations of protein-disulfide isomerase domains show that there is little correlation between isomerase activity and wild-type growth. J Biol Chem 276:27975–27980 Xiao R, Wilkinson B, Solovyov A, Lundstron-Ljung J, Winther JR, Holmgren A, Gilbert HF (2004) Protein disulfide isomerase is an oxidase and isomerase in the S. cerevisiae endoplasmic reticulum. J Biol Chem 279:49780–49786 Zhang Y, Kozlov G, Pocanschi CL, Brockmeier U, Ireland BS, Maattanen P, Howe C, Elliott T, Gehring K, Williams DB (2009) ERp57 does not require interactions with calnexin and calreticulin to promote assembly of class I histocompatibility molecules, and it enhances peptide loading independently of its redox activity. J Biol Chem 284:10160–10173
Chapter 8
Allosteric Disulfide Bonds Jason W.H. Wong and Philip J. Hogg
Abstract Protein disulfide bonds link cysteine residues in the polypeptide chain. The bonds contribute, sometimes crucially, to protein stability and function and are strongly conserved through the evolution of species. By analyzing the conservation of all structurally validated disulfide bonds across 29 completely sequenced eukaryotic genomes, we found that disulfide-bonded cysteines are even more conserved than tryptophan – the most conserved amino acid. Moreover, the rate of acquisition of disulfide bonds shows a strong positive correlation with organism complexity, which probably reflects the requirement for more sophistication in protein function in complex species. The majority of disulfide bonds perform a structural role by stabilizing the mature protein. Some disulfide bonds perform a functional role in the mature protein and can be divided into catalytic or allosteric disulfides. Catalytic disulfides/dithiols transfer electrons between proteins, while the allosteric bonds control the function of the protein in which they reside when they break and/or form. There are currently a dozen or so examples of allosteric disulfide bonds. The features of these bonds and their involvement in the respective proteins’ function are discussed. A common aspect of 11 of the 12 allosteric bonds discussed herein is that they link b-strands or b-loops. Keywords Allosteric disulfide bonds • Functional disulfide bonds • Disulfide bond evolution • CD4 • gp120 • b2-glycoprotein I
Abbreviations b2GPI b2-glycoprotein I Csk Carboxyl-terminal Src kinase ERp5 Endoplasmic reticulum protein 5 P.J. Hogg (*) Lowy Cancer Research Centre, Prince of Wales Clinical School, University of New South Wales, 2052 Sydney, Australia e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_8, © Springer Science+Business Media, LLC 2011
151
J.W.H. Wong and P.J. Hogg
152
GFP HIV MHCI NK PDI TG2 VWC VWF
Green fluorescent protein Human immunodeficiency virus Major histocompatibility complex class I Natural killer Protein disulfide isomerase Transglutaminase 2 von Willebrand factor type C domains von Willebrand factor
8.1 Introduction Cystine residues are two cysteine amino acids linked by a disulfide bond. The evolution of cystine resides in eukaryotic proteins will be considered initially to set the scene for discussion of the different types of disulfide bonds. Cystine residues or protein disulfide bonds perform either a structural or functional role and the functional disulfides can be classified as either catalytic or allosteric. The catalytic bonds are found at the active sites of enzymes that mediate thiol/disulfide exchange in other proteins, the oxidoreductases (Berndt et al. 2008), while the allosteric bonds control the function of the protein in which they reside by mediating a change when they are reduced or oxidized (Hogg 2003; Schmidt et al. 2006) (Fig. 8.1). The functional
Fig. 8.1 Definition of the allosteric disulfide bond. Allosteric disulfide bonds control the function of the protein in which they reside by mediating a change when they are reduced or oxidized. The redox state of the allosteric disulfides are controlled by catalytic disulfides of the oxidoreductases
8 Allosteric Disulfide Bonds
153
disulfides are mechanistically linked in that the redox state of the allosteric disulfides is controlled by catalytic disulfides. This chapter focuses on the allosteric disulfides.
8.2 Evolution of Disulfide Bonds in Eukaryotes 8.2.1 Increasing Frequency of Cysteine in Modern Proteomes It is generally assumed that the 20 common amino acids of present-day proteins did not arise simultaneously. Evidence based on sequence analysis (Brooks et al. 2002), aminoacyl-tRNA synthetase structural analysis (Klipcan and Safro 2004), and experimental simulations (Miller 1953, 1987) consistently indicate that simple amino acids such as glycine, alanine, and valine are likely to have been predominately present in early evolution. On the other hand, more complex amino acids including cysteine, tyrosine, and phenylalanine have been incorporated more recently as part of the genetic code. It has been shown that cysteine composition in proteins has more than doubled from 0.4 to 0.84% in a set of 65 highly conserved proteins across 26 species from all three domains of life (Brooks et al. 2004). In addition to the increased usage of cysteines in the proteome of modern organisms, whether the expansion of cysteines continues in current protein evolution has also been examined. Jordan et al. (2005) compared the change in amino acid frequencies in triplets of 15 closely related genomes across all three kingdoms. In all sets of genomes analyzed, cysteine content increased in the sister genomes when compared with the outgroup. Notably, cysteine was the most significantly accrued in 11 of the 15 sets. Interestingly, the same trend is observed when a comparable analysis was performed on human single-nucleotide polymorphisms. The finding suggests that cysteines continue to be positively selected in current protein evolution. The neutral theory of molecular evolution stipulates that the vast majority of protein evolutionary changes are caused by random drift of selectively neutral mutations. Based on this theory, the frequency of all amino acids should eventually reach equilibrium. Three amino acids, cysteine, tyrosine, and phenylalanine, were found to have increased since the last universal ancestor, suggesting that these amino acids have yet to reach equilibrium (Brooks et al. 2002). Cysteine is the only amino acid where the observed frequency in modern proteomes is substantially below that of the predicted. The observed frequency of tyrosine was found to be comparable with the predicted frequency, while the observed frequency of phenylalanine exceeded that of the predicted frequency. This indicates that cysteine has been acquired particularly slowly through evolution. The propensity for cysteines to form disulfide bonds may require the amino acid to be acquired simultaneously in pairs in order to have a nondeleterious effect on the protein. At the same time, cysteine is the second most conserved amino acid (after tryptophan) in protein evolution (Jones et al. 1992) suggesting that there is clearly strong positive selection for cysteines. The countering forces of positive selection and purifying selection in the processes of disulfide bond formation may in part explain the slow accrual of cysteines through evolution.
154
J.W.H. Wong and P.J. Hogg
8.2.2 Disulfide Bond Conservation Given that cysteines are the second most conserved amino acid after tryptophan (Gonnet et al. 1992), it is commonly accepted that disulfide bonds are also highly conserved in proteins. In an early comparative analysis, Thornton (1981) performed a comprehensive analysis on 55 known disulfide bonds elucidated from all available protein crystal structures at that time. It was reported that mutations in disulfide bonds only occurred when homology between sequences were greater than 50%, suggesting that mutations only occurred in distantly related proteins. The loss of disulfide bonds was frequently observed when comparing distantly related members of protein superfamilies. For example, within the serine protease superfamily, the number of disulfide bonds varies from six in trypsin to just two in streptogrisin-A in bacteria. In the examples studied by Thornton, based on crystallographic evidence, the presence or absence of a disulfide bond within a protein superfamily does not change the basic structure. For example, extra disulfide bonds in trypsin can be threaded with no alteration of structure in chymotrypsin. A notable feature is that in almost all cases examined by Thornton, the disulfide bond was solvent exposed and therefore expected to have the least effect on the overall structure. Increasing numbers of X-ray protein structures deposited in the Protein Data Bank now makes possible an analysis of all the types of disulfide bonds including buried, surface-exposed, structural, and functional disulfides. We recently analyzed the evolution of disulfide bonds across 29 eukaryotic species with completely sequenced genomes (see Table 8.1 for list of species) (Wong et al. 2010). There were 63,090 disulfide bonds in X-ray crystal structures in the protein databank (Berman et al. 2000) as of May 11, 2009. All disulfide bonds that mapped to identical UniProt residues and bonds with an undefined mapping were culled. Redundant disulfide bonds from the same protein domain or fold were also culled. This left a dataset of 5,181 disulfide bonds. One-to-one orthologs/in-paralogs of eukaryotic proteins containing these disulfide bonds were identified and conservation of specific disulfide bonds was determined using the Needleman–Wunsch algorithm. A disulfide bond was only considered to be conserved if both disulfide cysteine residues were perfectly aligned. The conservation of unpaired cysteine residues (as defined by X-ray crystal structures), tryptophan residues, and all amino acids was also determined and used as controls for relative conservation. The conservation of all nonredundant Saccharomyces cerevisiae, Drosophila melanogaster, Gallus gallus, Bos taurus, Rattus norvegicus, Mus musculus, and Homo sapiens disulfides against orthologous proteins in other metazoa and fungi species was calculated (Fig. 8.2). Half-cystines are significantly more conserved in all species than unpaired cysteines (p < 0.001, Wilcoxon signed-rank test), when conservation is expressed as a function of time since the last common ancestor. This is interesting since most unpaired cysteines also play functional roles in redox and metal binding. In particular, in almost all cases, half-cystines are even more conserved than tryptophan. The only exception being H. sapiens, M. musculus, and R. norvegicus where tryptophan is more conserved than half-cystines for the most
8 Allosteric Disulfide Bonds
155
Table 8.1 The 29 species used in the analysis of disulfide bond conservation in eukaryotes Last common ancestor with Species Common name humans (MYA) Homo sapiens Human 0 Pan troglodytes Chimpanzee 6.3 Macaca mulatta Macaque 30.5 Mus musculus Mouse 91 Rattus norvegicus Rat 91 Canis lupus Dog 97.4 Equus caballus Horse 97.4 Bos taurus Cow 97.4 Sus scrofa Pig 97.4 Monodelphis domestica Opossum 176.1 Ornithorhynchus anatinus Platypus 220.2 Gallus gallus Chicken 324.5 Taeniopygia guttata Finch 324.5 Danio rerio Zebrafish 454.6 Tribolium castaneum Beetle 910 Apis mellifera Bee 910 Anopheles gambiae Mosquito 910 Drosophila melanogaster Fruit fly 910 Drosophila yakuba Fruit fly 910 Drosophila simulans Fruit fly 910 Drosophila pseudoobscura Fruit fly 910 Caenorhabditis elegans Roundworm 910 Caenorhabditis briggsae Roundworm 910 Saccharomyces cerevisiae Baker’s yeast 1,368 Pichia stipitis Brewer’s yeast 1,368 Candida albicans Monilia 1,368 Schizosaccharomyces pombe Fission yeast 1,368 Aspergillus niger Black mold 1,368 Aspergillus nidulans Black mold 1,368 The common names and time of the last common ancestor is shown (Hedges et al. 2006)
distantly related organisms (i.e., fungi). The striking conservation of half-cystines reinforces the fact that disulfide bonds play a significant role in the maintenance of protein structure and function.
8.2.3 Rate of Disulfide Bond Acquisition As discussed earlier, a number of studies on amino acid composition in proteins have indicated that cysteines appear to be increasing in modern proteomes relative to other amino acids (Brooks and Fresco 2002; Jordan et al. 2005). Since roughly 50% of cysteines in proteins form disulfide bonds (based on human UniProt sequence
156
J.W.H. Wong and P.J. Hogg
Fig. 8.2 Conservation of disulfide bonds in eukaryotic protein evolution. Conservation of all nonredundant S. cerevisiae, D. melanogaster, G. gallus, B. taurus, R. norvegicus, M. musculus, and H. sapiens half-cystines, cysteines, tryptophan, and all amino acids against orthologous proteins in other species of Metazoan and Fungi as a function of the time of the last common ancestor (Million Years Ago, MYA) (Hedges et al. 2006). The data points and error bars are the mean and S.D.
annotations), it is reasonable to assume that the frequency of disulfide bonds may also be increasing. To investigate this, the relationship between organismal complexity and the rate of half-cystine acquisition was examined. The number of cell types was used to quantify organismal complexity (Carroll 2001). The rate of acquisition of unpaired cysteines was also measured as control. The rate of acquisition was calculated from the change in the level of half-cystine conservation as a function of the estimated time since the last common ancestor. This measure reflects how rapidly half-cystine or cysteine residues had to have been acquired in order to reach the frequency presently observed in today’s organisms.
8 Allosteric Disulfide Bonds
157
Fig. 8.3 Rate of half-cystine and unpaired cysteine acquisition as a function of cell types. The rate of acquisition is defined as the negative rate of conservation over estimate time of the last common ancestor with each of the species analyzed. The number of cell types is used as a measure of organismal complexity (Carroll 2001). Note that the R2 values should be interpreted cautiously because the data points are phylogenetically nonindependent
The correlation between organismal complexity and the rate of half-cystine acquisition is distinctly positive (Fig. 8.3). This relationship reflects the tendency for ancient disulfide bonds, such as those in yeast, to be retained. Multicellular organisms including humans have acquired new disulfide bonds more rapidly to reach the present-day frequency (see slope of half-cystine conservation in Fig. 8.2). The faster rate of accrual of disulfide bonds in complex organisms is consistent with the greater diversity of proteins in these species. In contrast, the correlation between organismal complexity and unpaired cysteine acquisition is marginally negative (Fig. 8.3), suggesting that unpaired cysteines are not being accrued. The overall rate of acquisition is higher than that for half-cystines which reflects the higher mutability of unpaired cysteine residues.
8.2.4 Disulfide Bonded Cysteines Are Usually Acquired in Pairs Thornton (1981) observed that in all cases where a disulfide bond was not conserved, both half-cystines were simultaneously mutated. This observation has subsequently
158
J.W.H. Wong and P.J. Hogg
Fig. 8.4 The proportion of disulfide bonds in H. sapiens conserved in cysteine pairs as a function of the time of the last common ancestor
been supported by analyses using larger datasets. A more recent study mapped disulfide conservation across five distinct protein families (Kreisberg et al. 1995). It was found that half-cystines were almost always replaced in pairs. In another recent study, Rubinstein and Fiser (2008) analyzed the disulfide bond conservation pattern from structural classification of proteins (Murzin et al. 1995; Hubbard et al. 1997). Multiple sequence alignment of the proteins showed that both halfcystines had been replaced in 81% of all cases. However, manual inspection revealed that conservation or replacement of one of the half-cystines was nearly always due to misalignment or a divergence in domain structure. By removing these incidences from the analysis, it was concluded that pairs of half-cystines were simultaneously replaced in 99% of all cases. The potential reactivity of free thiol groups is the likely reason for rare observations for one of two free cysteines to remain through protein evolution. In fact, this has been used as the basis for disulfide bond prediction within protein superfamilies (Rubinstein and Fiser 2008). We further investigated the proportion of disulfide bonds that are conserved in pairs across all human proteins from our dataset (Fig. 8.4). The proportion of disulfide bonds that are not conserved in pairs is seemingly high compared to that reported by Rubinstein and Fiser (2008). The discrepancy is likely due to the difference in the evolutionary distance between proteins of the two datasets. In the case of Rubinstein and Fiser, comparisons were made between families of proteins, where as in our dataset, comparisons were made within homologs from closely (6.3 MYA) to distantly (1,368 MYA) related species. Not surprisingly, the percentage of disulfide bonds acquired in pairs is more broadly similar for more distantly related species (>910 MYA). The increasing trend of disulfide bonds being conserved in cysteine pairs may suggest that through protein evolution, in some cases at least, disulfide bonds are formed through the acquisition of individual cysteines through time.
8 Allosteric Disulfide Bonds
159
Since disulfide bond patterns in homologous proteins are only inferred based on the amino acid sequence alignment, it is likely that in some cases, the loss of one half-cystine has been offset by the acquisition of another in different species. Reports of this type of disulfide bond reshuffling are rare in the literature, but an example was recently described in ribonuclease 8, where in the formation of a disulfide with Cys50, the lost Cys109 in monkeys has been offset by the gain of Cys93 in apes (Zhang 2007). We have also observed a number of cases of where a single half-cystine was acquired to form disulfide bonds in humans or murines. Examples include fibroblast growth factor 8 (Cys128-Cys145), MHC class I polypeptide-related sequence A (Cys36-Cys41), and T-cell surface glycoprotein CD4 (Cys130-159).
8.3 Disulfide Bond Structures We have hypothesized that the different types of disulfide bonds, structural versus functional, have in common particular configurations. To test this theory, the general features of all the disulfide bonds in X-ray and NMR structures in the Protein Data Bank were examined (Schmidt et al. 2006; Schmidt and Hogg 2007). A disulfide bond has been characterized as consisting of six atoms including the two a-carbons of the cysteine residues, Ca–Cb–Sg–Sg¢–Cb¢–Ca¢ (Fig. 8.5a). These six atoms define five c (chi) angles and each c angle can be either positive or negative. This definition equates to 20 possible disulfide bond configurations (Schmidt et al. 2006). The three fundamental types of disulfide are defined by the angle of the central three bonds: the spirals, hooks, or staples. If the c3 angle is positive the bond is right-handed and left-handed if the c3 angle is negative (Richardson and Richardson 1989). A disulfide is a minus left-handed spiral (−LHSpiral), for example, if the c1, c2, c3, c2¢, and c1¢ angles are negative, positive, negative, positive, and negative, respectively. Examples of three disulfide bond configurations are shown in Fig. 8.5b. The most common disulfides are the spirals. –LHSpiral bonds, for instance, account for nearly 30% of all disulfides in X-ray structures and 20% of the disulfides in NMR structures (Schmidt and Hogg 2007). Nearly all the catalytic disulfides are +/−RHHooks because they share a common CysXXCys motif in a thioredoxinfold domain structure. Of the known allosteric disulfides, the –RHStaple is the most common configuration (Schmidt et al. 2006) (see Sect. 1.4). A particular feature of –RHStaples is the close proximity of the a-carbon atoms of the two cysteine residues; 4.32 Å (95% CI, 4.26–4.37 Å) compared to a mean of 5.63 Å (95% CI, 5.61–5.65 Å) for all disulfides (Schmidt et al. 2006). The a-carbon atom distance is short because of the secondary structures that the cysteines link. About 60% of these bonds link adjacent strands in the same b-sheet. The strands are so close in the b-sheet that they need to pucker to accommodate the disulfide bond, which can strain the bond (Schmidt et al. 2006; Wouters et al. 2004).
160
J.W.H. Wong and P.J. Hogg
Fig. 8.5 Disulfide bond configuration. (a) The six atoms of the disulfide bond define five c angles and each c angle can be either positive or negative. This definition equates to 20 possible disulfide bond configurations. The –RHStaple disulfide bonds are characterized by the close proximity of the a-carbon atoms of the two cysteine residues. (b) Examples of three disulfide bond configurations: the CD4 Cys130-Cys159 –RHStaple (PDB ID 3CD4), the fibronectin Cys45-Cys56 –LHStaple (PDB ID 1O9A), and the b2-glycoprotein I Cys288-Cys326 −/+RHHook (PDB ID 1C1Z)
Analysis of the –RHStaple disulfides in solution NMR structures indicates that they can exist in the –LHStaple configuration, which has an even shorter mean Ca–Ca¢ distance and higher mean torsional energy (Schmidt and Hogg 2007). This finding suggests that the allosteric staple bonds can toggle between right- and lefthanded configurations and it may be that the left-handed configuration is more amenable to reduction. Specific examples of allosteric disulfide bonds are discussed in the remainder of this chapter. The examples have been separated into two groups: clear examples of regulation of mature protein function by allosteric disulfides (Table 8.2) and possible examples of such bonds (Table 8.3). There are strict criteria for inclusion of the examples in these lists. There must be one or more X-ray structure of the disulfide bond in question and the consequence of cleavage or formation of the bond for protein function is known. The catalytic disulfide/dithiol that controls the redox state of the disulfide has been determined in some cases. There are several other examples of possible allosteric disulfides but their current characterization does not meet these criteria (Wouters et al. 2004; Chen and Hogg 2006).
288-326
CD4
b2-glycoprotein I
MICA Angiotensinogen VWC domain
Human
Human
Human Human Eukaryotes
−/+RHHook
−RHStaple
−RHStaple
Thioredoxin
Thioredoxin
Not known
Oxidoreductase PDI/thioredoxin
Matthias et al. (2002), Maekawa et al. (2006), Matthias et al. (2010) Passam et al. (2010b), Ioannou et al. (2010) Kaiser et al. (2007) Zhou et al. (2010) Unpublished
References Barbouche et al. (2003), Gallina et al. (2002), Azimi et al. (2010) Fischer and Montal (2007)
202-259 −/+LHHook ERp5 18-138 −/+RHHook Not known 1-4 +/−RHSpiral Thioredoxin 3-5 −RHStaple Not known Criteria for inclusion in the table include the existence of one or more X-ray structures of the disulfide bond in question and the consequence of cleavage or formation of the bond for protein function is known. The oxidoreductase that controls the redox state of the allosteric disulfide is listed when known
429-453 (type A) 436-445 (type B) 130-159
Botulinum neurotoxins
Bacteria (Clostridium botulinum)
Table 8.2 Current examples of control of protein function by allosteric disulfide bonds Organism Protein Disulfide Cys Configuration Virus (HIV-1) gp120 296-331 −RHStaple
8 Allosteric Disulfide Bonds 161
Chloroplast m-type thioredoxin Not known
DsbL
Oxidoreductase DsbA (E. coli)
References Jacob-Dubuisson et al. (1994), Zav’yalov et al. (1997) Malojcic et al. (2008), Grimshaw et al. (2008) Gopalan et al. (2004)
Mills et al. (2007), Ogawa et al. (2002) Not known Pinkas et al. (2007), Lai et al. (2008) Human Tissue factor 186-209 −RHStaple Not known Chen et al. (2006), Ahamed et al. (2006) With these examples, the respective disulfide bonds have been implicated in control of the mature protein function but have not been proven to be involved. As for the examples in Table 8.2, the criteria for inclusion include the existence of one or more X-ray structures of the disulfide bond in question and the consequence of cleavage or formation of the bond for protein function is known. The oxidoreductase that controls the redox state of the allosteric disulfide is listed when known
Table 8.3 Possible examples of control of protein function by allosteric disulfide bonds Organism Protein Disulfide Cys Configuration Gram-negative bacteria PapD-like 207-212 (PapD) −RHStaple chaperones Bacteria Aryl Sulfotransferase 418-424 −RHStaple (Escherichia coli) Plant AtFKBP13 106-111 −RHStaple (Arabidopsis thaliana) Human C-terminal Src 122-164 +/−RHSpiral Kinase (Csk) +LHHook Human Transglutaminase 2 370-371 −RHStaple
162 J.W.H. Wong and P.J. Hogg
8 Allosteric Disulfide Bonds
163
8.4 Well-Characterized Allosteric Disulfide Bonds 8.4.1 gp120 The human immunodeficiency virus (HIV) is the agent responsible for AIDS. The HIV envelope glycoprotein (env) is translated as a single polypeptide chain (gp160) that is cleaved by host cell subtilisins into two noncovalently bound fragments, the surface glycoprotein subunit (gp120) and the transmembrane (gp41) subunit that is anchored in the viral membrane (Einfeld 1996). env is a trimer of gp41/gp120 heterodimers on the viral surface that is activated by binding to the immune co-receptor, CD4, and a chemokine receptor, CXCR4 or CCR5, on susceptible cells. gp120 dissociates from gp41, which allows the fusion peptide to be inserted into the target membrane. The end result is formation of a six-helix bundled gp41 ectodomain that drives the membrane merger and eventual fusion (Gallo et al. 2003). One or two of the nine disulfide bonds in gp120 are cleaved during fusion of the viral and target cell membranes (Barbouche et al. 2003; Gallina et al. 2002; Markovic et al. 2004; Ryser et al. 1994; Azimi et al. 2010). This event appears to be critical for HIV entry and infection. It is thought that reduction of the gp120 bond(s) facilitates unmasking of the gp41 fusion peptide and its insertion into the target cell membrane. The gp120 disulfides can be reduced by protein disulfide isomerase (PDI), thioredoxin, or glutaredoxin-1 (Barbouche et al. 2003; Gallina et al. 2002; Ou and Silver 2006; Auwerx et al. 2009). This observation implies that the gp120 bonds must have a high redox potential, as PDI is a poor reductant with a standard redox potential of −175 mV (Lundstrom and Holmgren 1993). A high redox potential suggests that the disulfide bonds are primed for reduction. Seven of the nine disulfide bonds are present in the eight crystal structures of gp120 in the protein databank. Five of these bonds can exist in either –RHStaple or –LHStaple configurations in the different structures (Schmidt and Hogg 2007). Thioredoxin cleaves the Cys296-Cys331 bond that constrains the b-loop structure of the V3 domain (Azimi et al. 2010) (Fig. 8.6a), which is the domain that binds chemokine receptor. This was determined using a thioredoxin trapping mutant and mass spectrometry analysis of the thioredoxin-gp120 covalent complex. Cleavage of the V3 Cys296-Cys331 disulfide may facilitate release of gp120 from the receptor and proper positioning of the fusion peptide (Barbouche et al. 2003). Interestingly, a recombinant gp140 trimer was found to be disulfide-bonded through the V3 domain (Billington et al. 2007). This result confirms the redox activity of the Cys296-Cys331 V3 bond in that it can form an intermolecular disulfide bond with other V3 bonds. Binding of purified CD4 to purified gp120 or purified CD4 to gp120 expressed on the surface of 293 cells (Azimi et al. 2010) facilitates cleavage of the gp120 V3 bond by thioredoxin, suggesting that CD4 binding induces a conformational change in gp120 that unmasks the Cys296-Cys331 disulfide bond. This conclusion is supported by the studies of Liu et al. (2008), who investigated the three-dimensional structures of trimeric env displayed on native HIV-1 in the unliganded state and in complex with CD4. They showed that CD4 binding results in a major reorganization of the env trimer, causing an outward rotation and displacement of each gp120
164
J.W.H. Wong and P.J. Hogg
Fig. 8.6 Allosteric disulfides in viral and bacterial proteins. (a) (i) Ribbon structure of gp120 showing the V3 domain loop and the Cys296-Cys331 and Cys385-Cys418 disulfide bonds in space-filling representation. (ii) Close-up of the two –RHStaple disulfide bonds, which are in the same b-sheet. The structure is that of PDB ID 1YYM (Huang et al. 2005). (b) (i) Ribbon structure of the di-chain botulinum neurotoxin type B showing the catalytic, receptor-binding and translocation domains and the Cys436-Cys445 disulfide bond in space-filling representation. (ii) Close-up of the Cys436-Cys445 –RHStaple disulfide bond, which constrains a b-loop structure in the single chain molecule. The structure is that of PDB ID 1EPW (Swaminathan and Eswaramoorthy 2000)
8 Allosteric Disulfide Bonds
165
monomer. This was coupled with a rearrangement of the gp41 trimer leading to closer contact between the viral and target cell membranes. It is not clear at this stage if a second disulfide bond is cleaved in gp120. Only the Cys296-Cys331 gp120 bond is cleaved by thioredoxin (Azimi et al. 2010). If a second bond is cleaved, then the Cys385-Cys418 disulfide is a good candidate (Schmidt and Hogg 2007). The Cys385-Cys418 bond is found as a –RHStaple in two structures (PDB ID’s 1YYM and 1YYL) and as a –LHStaple in one structure (PDB ID 1GC1). In the remaining structures, it is a –LHHook (Schmidt and Hogg 2007). The strain energies of the Cys385-Cys418 bond are particularly high for all bond configurations and the a-carbon separation is short, ranging from 3.7 to 3.9 Å. The Cys385-Cys418 bond is in the same b-barrel as the Cys296-Cys331 disulfide (Fig. 8.6a(ii)). It is possible that the reduction of the Cys296-Cys331 V3 bond leads to accessibility of the reductant to the Cys385-Cys418 disulfide. The cleavage of these two disulfides in one structural motif should allow for a large conformational change in the domain.
8.4.2 Botulinum Neurotoxins Neurotoxins from Clostridium botulinum are the causative agent of botulism. They block synaptic exocytosis in peripheral synapses causing flaccid paralysis (reviewed in (Montal 2010)). The neurotoxins are synthesized as a single polypeptide with a Mr of 150 kDa. The polypeptide is activated through cleaved by bacterial or host proteases into a heteromer consisting of a 100-kDa heavy chain and a 50-kDa light chain. The heavy and light chains are linked by a –RHStaple disulfide bond (Cys429Cys453 in type A neurotoxins and Cys436-Cys445 in type B) (Fig. 8.6b). The toxins consist of three domains: an N-terminal catalytic domain (light chain), an N-terminal half translocation domain (heavy chain), and the receptor-binding domain (C-terminal half). The catalytic domain is an endopeptidase which shares structural similarity to thermolysin, a metalloprotease. The neurotoxins enter sensitive cells by receptor-mediated endocytosis. These receptors involve a specific ganglioside (GT1b) together with other protein components that determine the neurotropism. The two-chain molecule integrates into the endosomal membrane where the –RHStaple is reduced, releasing the light-chain into the cytosol. It is proposed that the process is mediated by the following sequence: pH-induced heavy chain incorporation in the endosomal membrane, coupled light chain unfolding and its conduction through the heavy chain channel, release of the light chain by reduction of the –RHStaple disulfide bond, and finally light chain refolding in the cytosol (Fischer and Montal 2007). Unique components of SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) complex, the synaptic vesicle fusion complex required for membrane fusion, are cleaved by the light-chain endopeptidase. The –RHStaple disulfide linkage is crucial for the chaperone function of the heavy chain and light chain translocation and release (Fischer and Montal 2007). The identity of the oxidoreductase that reduces the –RHStaple disulfide is not known.
166
J.W.H. Wong and P.J. Hogg
Fig. 8.7 Allosteric disulfides in mammalian proteins. (a) (i) Ribbon structure of extracellular domains 1 and 2 of CD4 and the Cys130-Cys159 disulfide bond. The structure is that of PDB ID 3CD4 (Garrett et al. 1993). (ii) Structure of the predicted domain-swapped CD4 dimer, linked via the –RHStaple disulfide bonds (Sanejouand 2004). (b) Ribbon structure of b2-glycoprotein I showing the four complement control modules, the unusual fifth domain, and the Cys288-Cys326 disulfide bond in space-filling representation. The structure is that of PDB ID 1C1Z (Schwarzenbacher et al. 1999). (c) Ribbon structure of the extracellular part of MICA showing the Cys202-Cys259 allosteric disulfide bond in the a3 domain. The unusual Cys36-Cys41 –RHStaple disulfide bond in the a1 domain is also shown. The structure is that of PDB ID 1B3J (Li et al. 1999). (d) Ribbon structure of angiotensinogen showing the Cys18-Cys138 allosteric disulfide bond in space-filling representation. The structure is that of PDB ID 2WXW (Zhou et al. 2010). (e) Ribbon structure of the VWC domain of crossveinless 2 showing the Cys1-Cys4 and Cys3-Cys5 disulfide bonds. The structure is that of PDB ID 3BK3 (Zhang et al. 2008)
8.4.3 CD4 CD4 (cluster of differential 4) is a type I integral membrane glycoprotein consisting of four immunoglobulin-like domains, a trans-membrane domain and a short cytoplasmic region. It is a co-receptor for binding of T cells to antigen presenting cells. CD4 is also the primary receptor for HIV. Disulfide bond Cys130-Cys159 in the second extracellular domain of CD4 is a –RHStaple (Fig. 8.7a(i)). It is reduced on the T cell surface by thioredoxin, which leads to formation of disulfide-linked
8 Allosteric Disulfide Bonds
167
homo-dimers (Matthias et al. 2002; Ou and Silver 2006; Schwertassek et al. 2007). Domain swapping of D2 has been identified as the most likely mechanism for formation of the dimer. The domains likely dimerize through the formation of two disulfide-bonds between Cys130 in one monomer and Cys159 in the other (Maekawa et al. 2006; Sanejouand 2004) (Fig. 8.7a(ii)). Ablating the D2 disulfide-bond by mutagenesis promotes HIV entry and envmediated cell–cell fusion (Matthias et al. 2010) but impairs CD4’s co-receptor function (Maekawa et al. 2006). These observations suggest that monomeric reduced CD4 is preferentially selected for entry by HIV, while dimeric CD4 is the preferred receptor for binding to antigen presenting cells. Disrupting CD4 dimerization increases HIV-1 entry into susceptible cells (Bourgeois et al. 2006), which supports this conclusion. A structural model of the T-cell receptor and domain-swapped CD4 dimer bound to MHC class II and antigen also suggests that the domain-swapped dimer is the immune co-receptor (Maekawa et al. 2006). The unusual nature of the domain 2 bond and its involvement in CD4 function prompted an investigation into all known C2-type immunoglobulin superfamily domains for an equivalent disulfide link. Of the 1,275 domains, CEACAM20 is the only protein in addition to primate and rodent CD4 to contain a similar disulfide bond (Wong et al. 2010). This finding suggests that acquisition of the unusual disulfide bond in CD4 domain 2 is a relatively rare event in terms of immunoglobulin domain evolution.
8.4.4 b2-Glycoprotein I b2-glycoprotein I (b2GPI) is a 50-kDa plasma protein that circulates in blood at the relatively high concentration of ~0.2 mg/mL. Its primary function is not completely understood, but it is thought to play a role in apoptotic cell clearance. It has also been shown to be involved in coagulation through interactions with serine proteases, anionic phospholipid, and cell surface receptors (Miyakis et al. 2004). It has been strongly implicated in antiphospholipid syndrome, an autoimmune condition characterized by pathogenic circulating anti-b2GPI antibodies (Giannakopoulos et al. 2009). The syndrome is associated with enhanced oxidative stress, vascular thrombosis, accelerated atherosclerosis, and recurrent miscarriages. The structure of b2GPI is defined by four complement control protein modules and an unusual fifth domain (Bouma et al. 1999; Schwarzenbacher et al. 1999) (Fig. 8.7b). The fifth domain contains two small helices, a central b-spiral of four antiparallel b-sheets and an extended C-terminal loop region (Fig. 8.7b). The four complement modules each contain two conserved disulfides, while the fifth domain contains three disulfide bonds. The protein contains no lone cysteines. In b2GPI, of all the disulfide bonds Cys288-Cys326 in the fifth domain is particularly solvent accessible. Cys326 is the C-terminal residue and the disulfide bond links it to Cys288 which is within a b-loop (Fig. 8.7b). This bond has a −/+RHHook configuration (Passam et al. 2010b).
168
J.W.H. Wong and P.J. Hogg
The Cys288-Cys326 disulfide bond can be reduced by both PDI and thioredoxin, in platelets in culture or on the surface of endothelial cells (Passam et al. 2010b; Ioannou et al. 2010). As free cysteine thiol(s) in b2GPI can be labeled in human blood, cleavage of the b2GPI Cys288-Cys326 disulfide must occur in vivo (Ioannou et al. 2010). The reduced form of b2GPI confers protection of endothelial cells from oxidative stress. It is believed that S-nitrosylation of the free b2GPI thiols may contribute to this protection. It is possible that the primary physiological function of b2GPI is to protect vascular cells from oxidative damage via the Cys288-Cys326 disulfide bond (Ioannou et al. 2010). A high concentration of b2GPI circulating in plasma is consistent with such a role. Furthermore, it has been shown that reduced b2GPI also binds to VWF in a thiol-dependent manner. Binding of glycoprotein Iba to VWF and platelets to activated VWF (Passam et al. 2010a) is increased. These interactions may contribute to the redox regulation of platelet adhesion and thrombosis (Essex 2008; Furie and Furie 2008).
8.4.5 MICA NKG2D is an activating immune-receptor that is expressed on most natural killer (NK) cells, CD8 ab T cells, gd T cells, and macrophages (Gonzalez et al. 2006). One of its many ligands include the distant major histocompatibility complex class I (MHCI) homolog, MICA (Gonzalez et al. 2006), which is expressed in intestinal epithelium and epithelium-derived tumors (Groh et al. 1996, 1999). Unlike typical MHCs, MICA does not present antigen. It has been shown to be activated by cellular stress, including transformation, noxious conditions, or infection. The NKG2D homodimer–MICA monomer complex has a comparable crystal structure to that of the T-cell receptor–MHCI complex (Li et al. 2001). The MICA and MHCI proteins comprise three extracellular domains (a1, a2, and a3), a transmembrane-spanning domain, and a small cytoplasmic domain (Li et al. 1999). Despite having the same domain structure, their sequence identity is less than 30% (Li et al. 1999). The MHCI family proteins contain two highly conserved disulfide bonds, one each in the a2 (Cys101-Cys164) and a3 (Cys203-Cys259) domains. MICA contains these two disulfide bonds (Cys96-Cys164 and Cys202-Cys259), plus an extra one in the a1 domain (Cys36-Cys41). NKG2D positive cells can mediate effective tumor rejection in the early stages of tumorigenesis. However, late stage tumors can evade tumor immune response by sustained surface expression and shedding of MICA (Kaiser et al. 2007; Gonzalez et al. 2006). Shed MICA induces internalization and degradation of NKG2D, consequently stimulating population expansions of NKG2D+CD4+ T cells. These normally rare cells generally have negative regulatory functions. It has been shown that cleavage of the MICA a3 domain Cys202-Cys259 disulfide bond by endoplasmic reticulum protein 5 (ERp5) is responsible for the shedding of MICA at the cell-surface (Kaiser et al. 2007). This bond links the b-sheets of
8 Allosteric Disulfide Bonds
169
the immunoglobulin-like fold (Fig. 8.7c). The buried nature of the Cys202-Cys259 bond in the crystal structure implies that a major conformational change in the a3 domain is necessary for ERp5 to react with the disulfide (Kaiser et al. 2007). The extra disulfide bond in the MICA a1 domain, Cys36-Cys41, is interesting. This bond is a –RHStaple disulfide that constrains a b-loop structure, features typical of other allosteric bonds (Fig. 8.7c).
8.4.6 Angiotensinogen Angiotensinogen is the plasma protein source of the angiotensin peptides that control vasoconstriction and blood pressure (Fyhrquist and Saijonmaa 2008). Angiotensinogen has been shown to exist in oxidized and reduced forms in blood, defined by the redox state of the Cys18-Cys138 disulfide bond (Fig. 8.7d) (Zhou et al. 2010). The bond is a −/+RHHook that links an a-helix (Cys18) to a hydrogenbonded turn (Cys138). The disulfide has a standard redox potential of −230 mV, which is similar to the redox potentials of –RHStaple bonds (see below). The balance of the oxidized and reduced forms shifts toward the oxidized protein in the maternal circulation in pre-eclampsia, a hypertensive crisis of pregnancy (Zhou et al. 2010). The oxidized angiotensinogen is more efficiently cleaved by renin bound to cell surface pro-renin receptor, resulting in an increased release of angiotensin and rise in blood pressure.
8.4.7 VWC Domain More than 1,000 proteins contain von Willebrand factor type C (VWC) domains. The domain is 60–80 amino acids in length and is defined by a conserved spacing of ten cysteine residues (Zhang et al. 2007). Both intracellular and secreted proteins contain one or more copies of the VWC domain. The intracellular proteins are involved in transcription, DNA repair, protein turnover, membrane transport, and signal transduction, while the secreted proteins are involved in cell adhesion, migration, homing, and pattern formation. Mutations in VWC domains have been linked to pseudorheumatoid dysplasia (Hurvitz et al. 1999), a progressive skeletal muscle disorder, and von Willebrand’s disease (Lester et al. 2007), a defect in hemostasis resulting in excessive bleeding. A common feature of VWC domain-containing proteins is their involvement in multiprotein complexes. A number of proteins have been shown to self-associate or interact with other partners via their VWC domains (Abreu et al. 2002; Choi et al. 2007; Zhang et al. 2007). We have examined the mechanism of self-association of VWC domains from chordin and von Willebrand factor. Both domains form disulfide-linked dimers and higher-order oligomers via the Cys1-Cys4 and Cys3-Cys5 disulfides/dithiols when expressed in mammalian cells (T Ganderton, I Azimi, C Schroeder, JWH Wong
170
J.W.H. Wong and P.J. Hogg
and PJ Hogg, unpublished observations). The Cys1-Cys4 disulfide bond links a β-strand (Cys4) to a loop structure (Cys1) and has a +/–RHSpiral configuration, while the Cys3-Cys5 disulfide is a –RHStaple bond that links adjacent β-strands (Fig. 8.7e). Thioredoxin reduces the Cys1-Cys4 disulfide bond in the chordin VWC domain, suggesting that it may be the physiological reductant (Zhang et al. 2008). For von Willebrand factor, lateral dimers form when the Cys1 thiolate anion of the VWC2 domain attacks the Cys1 sulfur atom of the Cys1-Cys4 disulfide bond of another VWF molecule, while lateral trimers and higher-order oligomers form when the Cys3 thiolate anion of one of the VWF molecules in the dimer attacks the Cys1 sulfur atom of the Cys1-Cys4 disulfide bond of another VWF molecule.
8.5 Possible Allosteric Disulfide Bonds 8.5.1 PapD-Like Chaperones P pili are adhesive fibers made by Gram-negative bacteria. They mediate specific recognition and attachment to host tissues. The pili are assembled from individual subunits by the immunoglobulin-fold PapD-like chaperones in the periplasm. The PapD chaperones play a role in the temporary binding, stabilization, and capping of subunits until they are ready to be assembled into the pilus. The pili subunits lack the seventh b-strand in their immunoglobulin-fold. The chaperone provides this strand allowing the capping of the subunit until it reaches the pilus assembly site. Each subunit within the mature pilus structure supplies the final strand to complete the fold of its adjacent subunit (Sauer et al. 2002). PapD and family members contain a CD4-like C-terminal domain. Within this domain there is a –RHStaple disulfide bond (Cys207-Cys212) linking the last two b-strands (Jacob-Dubuisson et al. 1994; Zav’yalov et al. 1997; Piatek et al. 2005) (Fig. 8.8a). This disulfide bond has been shown to be required for in vivo activity and its formation is mediated by the DsbA/DsbB redox couple in the periplasm. The role of the –RHStaple disulfide bond in chaperone function has been examined in the F1 antigen/Caf1M system of Yersinia pestis (Zav’yalov et al. 1997). This system functions in an analogous manner to the P pilus/PapD system of E. coli (Sauer et al. 2002). The same –RHStaple bond is present in both PapD and Caf1M. Zav’yalov et al. have hypothesized that the –RHStaple bond undergoes a redox cycle which is necessary for the release of the Caf1 subunit on interaction with Caf1A, an outer-membrane usher protein. The Caf1–oxidized Caf1M complex interacts with Caf1A. Caf1M is subsequently released as the Caf1 subunits assemble forming the F1 capsule. Reduction of the –RHStaple disulfide bond by DsbA is possibly the trigger for dissociation of Caf1M. The reduced Caf1M is able to re-enter the cycle once it has been re-oxidized. This mechanism has not been proven, though, hence the assignment of the PapD Cys207-Cys212 bond to the possible allosteric disulfides.
8 Allosteric Disulfide Bonds
171
Fig. 8.8 Structures of potential allosteric disulfides. (a) Ribbon structure of PapD showing the Cys207-Cys212 –RHStaple bond linking the last two b-strands of the C-terminal domain. The structure is that of PDB ID 1N0L (Sauer et al. 2002). (b) Ribbon structure of aryl sulfotransferase showing the Cys418-Cys424 –RHStaple bond in a b-loop of the sixth propeller blade and the catalytic His436 residue. The structure is that of PDB ID 3ELQ (Malojcic et al. 2008). (c) Overlay ribbon structures of the reduced (green) and oxidized (blue) forms of AtFKBP13 showing the Cys106-Cys111 –RHStaple bond in an extended b-loop structure. The loop shifts position when the bond is reduced. The structures are that of PDB ID’s 1Y0O_C (oxidized) and 1U79_C (reduced) (Gopalan et al. 2004). (d) Overlay ribbon structures of the reduced (green) and oxidized (blue) forms of the SH2 domain of Csk showing the Cys122-Cys164 disulfide bond that links a b-loop (Cys164) to coil (Cys122). The backbone of the structure shifts position slightly when the bond is reduced. The structures are that of PDB ID’s 3EAZ_A (oxidized) and 3EAC_A (reduced). (e) Ribbon structure of transglutaminase 2 showing the Cys370-Cys371 –RHStaple bond in a b-loop of the catalytic domain. The structure is that of PDB ID 2Q3Z. (f) Ribbon structure of Tissue Factor showing the Cys186-Cys209 –RHStaple bond that links adjacent b-strands in the membraneproximal C-terminal domain. The structure is that of PDB ID 2HFT (Muller et al. 1996)
8.5.2 Aryl Sulfotransferase Bacterial aryl sulfotransferases are enzymes that catalyze the transfer of sulfuryl groups between phenolic compounds. They reside in the periplasm and mediate host–pathogen interactions and intercell interactions. A –RHStaple disulfide bond (Cys418-Cys424) is essential for activity of E. coli aryl sulfotransferase (Malojcic et al. 2008; Grimshaw et al. 2008). The bond is located close to the surface of the protein in a b-loop of the sixth propeller blade. It is just 12 residues N terminal of the catalytic His436 that undergoes transient sulfurylation in the catalytic reaction (Fig. 8.8b).
172
J.W.H. Wong and P.J. Hogg
It is thought that the bond may be necessary for the maintenance of conformational equilibrium in the protein. Interestingly, the DsbL/Dsbl redox couple appears to have co-evolved with aryl sulfotransferase to control the redox state of the bond in the bacterial periplasm, as the three genes are encoded in a tricistronic operon (Grimshaw et al. 2008). As with the PapD –RHStaple bond, it has not been shown whether the aryl sulfotransferase bond can break and/or form in the mature protein, so it has been also listed as a possible allosteric disulfide.
8.5.3 AtFKBP13 Photosynthesis in plants and eukaryotic algae occurs in the chloroplasts. Oxygen evolution occurs in the lumen of these organelles, known as the thylakoids. Within the thylakoid lumen, there are more than a dozen resident immunophilins. These proteins were originally defined as targets for immunosuppressive drugs and were subsequently shown to have peptidyl-prolyl cis–trans isomerase activity. AtFKBP13 is a representative of the FKBP immunophilin group which is resident in the thylakoid lumen. AtFKBP13 consists of six b-strands which form two a-helices and an antiparallel b-sheet (Gopalan et al. 2004). AtFKBP13 contains a –RHStaple disulfide bond (Cys106-Cys111) that is not conserved in FKBPs from animals or yeast (Fig. 8.8c). The disulfide links a b-loop motif that shifts considerably upon the reduction of the disulfide bond (Fig. 8.8c). Chloroplast m-type thioredoxin has been shown to be able to reduce this disulfide (Gopalan et al. 2004). Upon reduction, peptidyl-prolyl isomerase activity of AtFKBP13 is reduced. It has not been proven that m-type thioredoxin regulates AtFKBP13 function in the chloroplast thylakoids, however, so it has been included under possible allosteric disulfides.
8.5.4 Carboxyl-Terminal Src Kinase Src family kinases are important for many aspects of cellular function including cell division and differentiation, cell attachment and movement, and cell survival and death. Src family tyrosine kinases in the cytoplasm are regulated by the C-terminal Src kinase (Csk) in a negative manner. Csk consists of SH3, SH2, and kinase domains (Ogawa et al. 2002). There is an unusual disulfide bond (Cys122-Cys164) within the SH2 domain that can exist in either a + LHHook or +/−RHSpiral configuration in a 1 Å resolution crystal structure (PDB ID 3EAC). The two central antiparallel b-sheets within the SH2 domain are flanked by two a-helices. The Cys122-Cys164 disulfide bond is solvent exposed and links a b-loop (Cys164) to random coil (Cys122) (Fig. 8.8d). The NMR solution structure of the complete protein shows that upon the reduction of the disulfide bond, residues that extend from the bond across the molecule change to a surface that is in close proximity
8 Allosteric Disulfide Bonds
173
with the kinase domain. Furthermore, molecular dynamics analyses indicate that formation of the disulfide bond affects residues within the active-site cleft of the kinase domain. Crystal structures of the oxidized (PDB ID 1K9A:C) and reduced (PDB ID 1K9A:B) SH2 domain show differing positioning of the central antiparallel b-sheets in the two structures (Fig. 8.8d). There is tenfold increase in kinase activity in the reduced form of the enzyme compared to the oxidized form (Mills et al. 2007). This suggests that enzyme activity may be controlled by redox change of the disulfide bond. This has not been shown, though, so the Csk disulfide has been included as a possible allosteric bond.
8.5.5 Transglutaminase 2 Transglutaminase 2 (TG2), or tissue transglutaminase, functions inside and outside the cell (Griffin et al. 2002; Siegel and Khosla 2007). In the cytoplasm, TG2 functions as a G-protein in regulation of the phospholipase C signal transduction cascade. Extracellularly, TG2 catalyzes transamidation of glutamine to lysine in a Ca2+dependent manner resulting in Ne(g-glutamyl)lysyl isopeptide bonds. TG2 also has a role in cell adhesion, differentiation, signaling, and motility by binding tightly to the extracellular matrix via fibronectin and cell-surface integrins. Human TG2 contains an N-terminal b-sandwich, a catalytic domain and two C-terminal fibronectin type III b-barrels (Pinkas et al. 2007). There is a –RHStaple disulfide bond (Cys370-Cy371) in the catalytic domain and is within an extended b-loop structure (Fig. 8.8e). There is evidence that oxidation of Cys370-Cy371 to form the disulfide bond results in the inactivation of the enzyme. It is believe that the bond interferes with the ability of TG2 to undergo a change in conformation upon binding to GTP (Chung and Folk 1970; Connellan and Folk 1969; Begg et al. 2006). The evidence is circumstantial that the Cys370-Cys371 TG2 bond controls the function of the mature protein, so it has been included as a possible allosteric disulfide. Nevertheless, this cysteine pair is conserved in a number of other members of the transglutaminase family, including TG1, TG4, TG5, and TG7.
8.5.6 Tissue Factor Injury to a blood vessel rapidly initiates events in the vessel wall and in blood that maintain hemostasis (Furie and Furie 2008). Platelets in the circulation are recruited to the site of injury and are incorporated into the developing thrombus. At the same time, blood coagulation stabilizes the developing thrombus through the production of fibrin. Aberrant arterial thrombus formation leads to myocardial infarction and stroke. Venous thrombosis is in fact one of the leading causes of fatality in cancer patients (Furie and Furie 2008). Blood coagulation is triggered by Tissue Factor, a transmembrane cofactor. It resides on the cell-surface in mostly an inactive configuration (cryptic) that must be
174
J.W.H. Wong and P.J. Hogg
activated (de-encrypted) before it triggers coagulation. Cryptic Tissue Factor binds coagulation factor VIIa and cleaves a peptidyl substrate while active Tissue Factor binds factor VIIa and cleaves both a peptidyl substrate and coagulation factor X (Ploplis et al. 1987; Sakai et al. 1989; Le et al. 1992). A characteristic feature of cryptic Tissue Factor is the slow rate of binding of factor VII/VIIa. Binding of VII/VIIa to active Tissue Factor reaches equilibrium within 1 min, while binding to cryptic Tissue Factor takes 1–2 h to reach equilibrium. Activation of factor X leads to a cascade of protease activations culminating in thrombin formation and production of fibrin. The membrane-proximal domain of Tissue Factor contains an unusual disulfide bond (Cys186-Cys209). This bond has been implicated in the de-encryption or activation of the cofactor (Chen et al. 2006; Ahamed et al. 2006; Liang and Hogg 2008; Chen and Hogg 2006). The cofactor contains two fibronectin type III domains in the extracellular part, each with a disulfide bond. The N-terminal domain Cys49-Cys57 disulfide is typical for fibronectin type III domains, while the C-terminal domain Cys186-Cys209 disulfide is a –RHStaple bond that straddles the F and G strands of the antiparallel b-sheet and is exposed to solvent (Chen et al. 2006) (Fig. 8.8g). The –RHStaple bond is hypothesized to be reduced in cryptic Tissue Factor, while oxidation of the disulfide results in de-encryption. Thiol alkylating compounds stop the activation of Tissue Factor while dithiol cross-linkers activate it, suggesting that the disulfide is involved in the activation process (Chen et al. 2006). Also, an intact Cys186-Cys209 bond is required for efficient Tissue Factor coagulant activity (Ahamed et al. 2006). This mechanism of Tissue Factor de-encryption remains controversial, with alternative mechanisms proposed (Pendurthi et al. 2007; Bach 2006). PDI (Ahamed et al. 2006; Cho et al. 2008; Reinhardt et al. 2008), glutathione (Reinhardt et al. 2008), and NO (Ahamed et al. 2006) have been implicated in redox control of the Tissue Factor Cys186-Cys209 disulfide, although the involvement of PDI has been questioned (Persson 2008; Versteeg and Ruf 2007), however. Notably, PDI blockade markedly reduces thrombus formation in vivo (Cho et al. 2008; Reinhardt et al. 2008). Complexes of Tissue Factor-VIIa and Tissue Factor-VIIa–Xa also signal in inflammation, angiogenesis, and tumor progression through the cleavage of protease activated receptor 2 (Chen and Hogg 2006). Interestingly, an intact Cys186-Cys209 disulfide is not required for binary Tissue Factor-VIIa complex signaling (Ahamed et al. 2006). The Cys209Ala Tissue Factor mutant retains signaling activity while the Cys186Ala mutant is inactive in signaling (Ahamed et al. 2006). Evidence for the formation of the Cys186-Cys209 disulfide bond being necessary for Tissue Factor de-encryption is still circumstantial; therefore, it is listed as a possible allosteric disulfide.
8.6 A –RHStaple Bond with a Catalytic Function Dsbs are a class of proteins involved in the mediation of disulfide bond formation and isomerization in the E. coli periplasm. DsbD is located in the inner membrane and is responsible for electron transfer from thioredoxin in the cytoplasm to other
8 Allosteric Disulfide Bonds Table 8.4 Redox potentials of –RHStaple disulfide bonds Standard redox Protein Disulfide potential (mV) CD4 Cys130-Cys159 −241 DsbD Cys103-Cys109 −229 Tissue factor Cys186-Cys209 −278 Green fluorescent Cys149-Cys202 −261 protein
175
Reference Hogg (2009) Collet et al. (2002) Hogg (2009) Ostergaard et al. (2001)
oxidoreductases in the periplasm. DsbD contains a periplasmic N-terminal domain, a central transmembrane domain, and a periplasmic C-terminal domain (Rozhkova and Glockshuber 2008). Each of these domains contains two cysteine residues that transport two electrons. Cytoplasmic thioredoxin donates two electrons to the central transmembrane domain cysteines, which are shuttled to a catalytic +/−RHHook disulfide of the periplasmic C-terminal domain. The electrons are finally shuttled to a –RHStaple disulfide in the periplasmic N-terminal domain. The –RHStaple then donates the e lectrons to periplasmic CcmG, DsbC, and DsbG .
8.7 Redox Potentials of –RHStaple Disulfide Bonds The redox potentials of four –RHStaple bonds are known (Table 8.4). These are the Cys186-Cys209 Tissue Factor bond, the Cys149-Cys202 bond engineered into green fluorescent protein (GFP), the Cys130-Cys159 CD4 bond, and the Cys103Cys109 DsbD bond. The redox potentials of these four –RHStaple bonds are closer to that of the strong protein reductant thioredoxin (Krause et al. 1991), than the more oxidizing PDI (Lundstrom and Holmgren 1993). For reference, the catalytic disulfides of PDI and thioredoxin have redox potentials of −175 and −270 mV, respectively (Berndt et al. 2008).
8.8 Stereochemistry of Disulfide Bond Reduction The bimolecular nucleophilic substitution (SN2) reaction has been shown to be the mechanism by which disulfide bonds are reduced (Fernandes and Ramos 2004; Wiita et al. 2007). In this reaction, the three sulfur atoms involved (the sulfur ion nucleophile of the oxidoreductase and the two sulfur atoms of the disulfide bond) must form a ~180° angle (Fig. 8.9). The reaction is highly directional and proceeds via a transition state. The three sulfur atoms must be appropriately positioned to enable the reaction to proceed. Therefore, allosteric disulfide bonds must have a favorable stereochemistry to be reduced by oxidoreductases.
176
J.W.H. Wong and P.J. Hogg
Fig. 8.9 Stereochemistry of disulfide bond cleavage. Reduction of disulfide bonds occurs via a bimolecular nucleophilic substitution reaction (SN2). In this reaction, the arrangement of incoming atom (nucleophile), target atom (electrophile), and leaving group (outgoing atom) is linear. The proposed details of the SN2 reaction of a reductase (such as thioredoxin) with a disulfide RSSR¢ is shown: (1) deprotonation of reduced reductase by buffer to give I as the initial state; (2) SN2 reaction of I with RSSR¢ with nucleophile (blue), electrophile (red), and leaving group (brown) gives III via transition state II with byproduct R¢SH; (iii) deprotonation of III by buffer to give IV; (iv) intramolecular SN2 reaction of IV with nucleophile (blue), electrophile (red), and leaving group (brown) gives oxidized reductase (VI) via transition state V with byproduct RSH. The transition states II and V are presumed to have the central sulfur atom in a trigonal bipyramidal geometry (when the two nonbonding electron pairs on sulfur are considered) with the S…S…S angle close to or exactly 180°
8.9 Disulfide Bond Analysis Tool To assist with the identification of possible functional disulfide bonds in proteins, a simple interface to obtain structural information about disulfide bonds in X-ray or NMR structures has been generated (www.cancerresearch.unsw.edu.au/CRCWeb.nsf/ page/Disulfide+Bond+Analysis). The tool allows users to rapidly obtain geometric measures, secondary structural information, solvent accessibility values, and the classification of disulfide bonds. Any number of structures obtained by any method (e.g., X-ray diffraction, solution NMR, or even predicted structures) can be analyzed simultaneously. This tool assisted with the identification of the b2GPI allosteric disulfide (Passam et al. 2010a, b; Ioannou et al. 2010).
8.10 Conclusions The strong evolutionary conservation of cystines suggests that disulfide bonds are important components in the determination of protein structure and function. Allosteric disulfide bonds in particular control the function of mature proteins
8 Allosteric Disulfide Bonds
177
when they break and/or form. The redox state of these bonds is controlled by the catalytic disulfides of the oxidoreductases. There are currently seven examples of allosteric disulfide bonds that clearly control the function of the protein in which they reside, and there are several other potential examples at various stages of characterization. Most of the allosteric bonds characterized so far have a –RHStaple configuration and a common feature of 11 of the 12 bonds discussed herein is that they link b-strands or b-loop structures. The current indications are that this means of post-translational control of protein function may prove to be a generally important mechanism in prokaryotes and eukaryotes.
References Abreu JG, Coffinier C, Larraín J, Oelgeschläger M, De Robertis EM (2002) Chordin-like cr domains and the regulation of evolutionarily conserved extracellular signalling systems. Gene 287:39–47 Ahamed J, Versteeg HH, Kerver M, Chen VM, Mueller BM, Hogg PJ, Ruf W (2006) Disulfide isomerization switches tissue factor from coagulation to cell signaling. Proc Natl Acad Sci USA 103(38):13932–13937 Auwerx J, Isacsson O, Soderlund J, Balzarini J, Johansson M, Lundberg M (2009) Human glutaredoxin-1 catalyzes the reduction of hiv-1 gp120 and cd4 disulfides and its inhibition reduces hiv-1 replication. Int J Biochem Cell Biol 41(6):1269–1275 Azimi I, Matthias LJ, Center RJ, Wong JW, Hogg PJ (2010) Disulfide bond that constrains the hiv-1 gp120 v3 domain is cleaved by thioredoxin. J Biol Chem 285(51):40072–40080 Bach RR (2006) Tissue factor encryption. Arterioscler Thromb Vasc Biol 26(3):456–461 Barbouche R, Miquelis R, Jones IM, Fenouillet E (2003) Protein-disulfide isomerase-mediated reduction of two disulfide bonds of hiv envelope glycoprotein 120 occurs post-cxcr4 binding and is required for fusion. J Biol Chem 278(5):3131–3136 Begg GE, Carrington L, Stokes PH, Matthews JM, Wouters MA, Husain A, Lorand L, Iismaa SE, Graham RM (2006) Mechanism of allosteric regulation of transglutaminase 2 by gtp. Proc Natl Acad Sci USA 103(52):19683–19688 Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE (2000) The protein data bank. Nucleic Acids Res 28(1):235–242 Berndt C, Lillig CH, Holmgren A (2008) Thioredoxins and glutaredoxins as facilitators of protein folding. Biochim Biophys Acta 1783(4):641–650 Billington J, Hickling TP, Munro GH, Halai C, Chung R, Dodson GG, Daniels RS (2007) Stability of a receptor-binding active human immunodeficiency virus type 1 recombinant gp140 trimer conferred by intermonomer disulfide bonding of the v3 loop: Differential effects of protein disulfide isomerase on cd4 and coreceptor binding. J Virol 81(9):4604–4614 Bouma B, de Groot PG, van den Elsen JM, Ravelli RB, Schouten A, Simmelink MJ, Derksen RH, Kroon J, Gros P (1999) Adhesion mechanism of human beta(2)-glycoprotein i to phospholipids based on its crystal structure. EMBO J 18(19):5166–5174 Bourgeois R, Mercier J, Paquette-Brooks I, Cohen EA (2006) Association between disruption of cd4 receptor dimerization and increased human immunodeficiency virus type 1 entry. Retrovirology 3:31 Brooks DJ, Fresco JR (2002) Increased frequency of cysteine, tyrosine, and phenylalanine residues since the last universal ancestor. Mol Cell Proteomics 1(2):125–131 Brooks DJ, Fresco JR, Lesk AM, Singh M (2002) Evolution of amino acid frequencies in proteins over deep time: Inferred order of introduction of amino acids into the genetic code. Mol Biol Evol 19(10):1645–1655
178
J.W.H. Wong and P.J. Hogg
Brooks DJ, Fresco JR, Singh M (2004) A novel method for estimating ancestral amino acid composition and its application to proteins of the last universal ancestor. Bioinformatics 20(14): 2251–2257 Carroll SB (2001) Chance and necessity: The evolution of morphological complexity and diversity. Nature 409(6823):1102–1109 Chen VM, Ahamed J, Versteeg HH, Berndt MC, Ruf W, Hogg PJ (2006) Evidence for activation of tissue factor by an allosteric disulfide bond. Biochemistry 45(39):12020–12028 Chen VM, Hogg PJ (2006) Allosteric disulfide bonds in thrombosis and thrombolysis. J Thromb Haemost 4(12):2533–2541 Cho J, Furie BC, Coughlin SR, Furie B (2008) A critical role for extracellular protein disulfide isomerase during thrombus formation in mice. J Clin Invest 118(3):1123–1131 Choi H, Aboulfatova K, Pownall HJ, Cook R, Dong JF (2007) Shear-induced disulfide bond formation regulates adhesion activity of von willebrand factor. J Biol Chem 282(49):35604–35611 Chung SI, Folk JE (1970) Mechanism of the inactivation of guinea pig liver transglutaminase by tetrathionate. J Biol Chem 245(4):681–689 Collet JF, Riemer J, Bader MW, Bardwell JC (2002) Reconstitution of a disulfide isomerization system. J Biol Chem 277(30):26886–26892 Connellan JM, Folk JE (1969) Mechanism of the inactivation of guinea pig liver transglutaminase by 5,5’-dithiobis-(2-nitrobenzoic acid). J Biol Chem 244(12):3173–3181 Einfeld D (1996) Maturation and assembly of retroviral glycoproteins. Curr Top Microbiol Immunol 214:133–176 Essex DW (2008) Redox control of platelet function. Antioxidants & redox signaling 11(5):1191–1225 Fernandes PA, Ramos MJ (2004) Theoretical insights into the mechanism for thiol/disulfide exchange. Chemistry (Weinheim an der Bergstrasse, Germany) 10 (1):257–266 Fischer A, Montal M (2007) Crucial role of the disulfide bridge between botulinum neurotoxin light and heavy chains in protease translocation across membranes. J Biol Chem 282(40): 29604–29611 Furie B, Furie BC (2008) Mechanisms of thrombus formation. N Engl J Med 359(9):938–949 Fyhrquist F, Saijonmaa O (2008) Renin-angiotensin system revisited. J Intern Med 264(3): 224–236 Gallina A, Hanley TM, Mandel R, Trahey M, Broder CC, Viglianti GA, Ryser HJ (2002) Inhibitors of protein-disulfide isomerase prevent cleavage of disulfide bonds in receptor-bound glycoprotein 120 and prevent hiv-1 entry. J Biol Chem 277(52):50579–50588 Gallo SA, Finnegan CM, Viard M, Raviv Y, Dimitrov A, Rawat SS, Puri A, Durell S, Blumenthal R (2003) The hiv env-mediated fusion reaction. Biochim Biophys Acta 1614(1):36–50 Garrett TP, Wang J, Yan Y, Liu J, Harrison SC (1993) Refinement and analysis of the structure of the first two domains of human cd4. J Mol Biol 234(3):763–778 Giannakopoulos B, Passam F, Ioannou Y, Krilis SA (2009) How we diagnose the antiphospholipid syndrome. Blood 113(5):985–994 Gonnet GH, Cohen MA, Benner SA (1992) Exhaustive matching of the entire protein sequence database. Science 256(5062):1443–1445 Gonzalez S, Groh V, Spies T (2006) Immunobiology of human nkg2d and its ligands. Curr Top Microbiol Immunol 298:121–138 Gopalan G, He Z, Balmer Y, Romano P, Gupta R, Heroux A, Buchanan BB, Swaminathan K, Luan S (2004) Structural analysis uncovers a role for redox in regulating fkbp13, an immunophilin of the chloroplast thylakoid lumen. Proc Natl Acad Sci USA 101(38):13945–13950 Griffin M, Casadio R, Bergamini CM (2002) Transglutaminases: Nature’s biological glues. Biochem J 368(Pt 2):377–396 Grimshaw JP, Stirnimann CU, Brozzo MS, Malojcic G, Grutter MG, Capitani G, Glockshuber R (2008) Dsbl and dsbi form a specific dithiol oxidase system for periplasmic arylsulfate sulfotransferase in uropathogenic escherichia coli. J Mol Biol 380(4):667–680 Groh V, Bahram S, Bauer S, Herman A, Beauchamp M, Spies T (1996) Cell stress-regulated human major histocompatibility complex class i gene expressed in gastrointestinal epithelium. Proc Natl Acad Sci USA 93(22):12445–12450
8 Allosteric Disulfide Bonds
179
Groh V, Rhinehart R, Secrist H, Bauer S, Grabstein KH, Spies T (1999) Broad tumor-associated expression and recognition by tumor-derived gamma delta t cells of mica and micb. Proc Natl Acad Sci USA 96(12):6879–6884 Hedges SB, Dudley J, Kumar S (2006) Timetree: A public knowledge-base of divergence times among organisms. Bioinformatics 22(23):2971–2972 Hogg PJ (2003) Disulfide bonds as switches for protein function. Trends Biochem Sci 28(4):210–214 Hogg PJ (2009) Contribution of allosteric disulfide bonds to regulation of hemostasis. J Thromb Haemost 7(Suppl 1):13–16 Huang CC, Stricher F, Martin L, Decker JM, Majeed S, Barthe P, Hendrickson WA, Robinson J, Roumestand C, Sodroski J, Wyatt R, Shaw GM, Vita C, Kwong PD (2005) Scorpion-toxin mimics of cd4 in complex with human immunodeficiency virus gp120 crystal structures, molecular mimicry, and neutralization breadth. Structure 13(5):755–768 Hubbard TJ, Murzin AG, Brenner SE, Chothia C (1997) Scop: A structural classification of proteins database. Nucleic Acids Res 25(1):236–239 Hurvitz JR, Suwairi WM, Van Hul W, El-Shanti H, Superti-Furga A, Roudier J, Holderbaum D, Pauli RM, Herd JK, Van Hul EV, Rezai-Delui H, Legius E, Le Merrer M, Al-Alami J, Bahabri SA, Warman ML (1999) Mutations in the ccn gene family member wisp3 cause progressive pseudorheumatoid dysplasia. Nat Genet 23(1):94–98 Ioannou Y, Zhang J-Y, Passam FH, Rahgozar S, Qi JC, Giannakopoulos B, Yu MQP, Yu DM, Hogg PJ, Krilis SA (2010) Naturally occurring free thiols within ß2-glycoprotein i in vivo: Nitrosylation, redox modification by endothelial cells and regulation of oxidative stress induced cell injury. Blood:in press Jacob-Dubuisson F, Pinkner J, Xu Z, Striker R, Padmanhaban A, Hultgren SJ (1994) Papd chaperone function in pilus biogenesis depends on oxidant and chaperone-like activities of dsba. Proc Natl Acad Sci USA 91(24):11552–11556 Jones DT, Taylor WR, Thornton JM (1992) The rapid generation of mutation data matrices from protein sequences. Comput Appl Biosci 8(3):275–282 Jordan IK, Kondrashov FA, Adzhubei IA, Wolf YI, Koonin EV, Kondrashov AS, Sunyaev S (2005) A universal trend of amino acid gain and loss in protein evolution. Nature 433(7026):633–638 Kaiser BK, Yim D, Chow IT, Gonzalez S, Dai Z, Mann HH, Strong RK, Groh V, Spies T (2007) Disulphide-isomerase-enabled shedding of tumour-associated nkg2d ligands. Nature 447(7143): 482–486 Klipcan L, Safro M (2004) Amino acid biogenesis, evolution of the genetic code and aminoacyltrna synthetases. J Theor Biol 228(3):389–396 Krause G, Lundstrom J, Barea JL, Pueyo de la Cuesta C, Holmgren A (1991) Mimicking the active site of protein disulfide-isomerase by substitution of proline 34 in escherichia coli thioredoxin. J Biol Chem 266(15):9494–9500 Kreisberg R, Buchner V, Arad D (1995) Paired natural cysteine mutation mapping: Aid to constraining models of protein tertiary structure. Protein Sci 4(11):2405–2410 Lai TS, Liu Y, Tucker T, Daniel KR, Sane DC, Toone E, Burke JR, Strittmatter WJ, Greenberg CS (2008) Identification of chemical inhibitors to human tissue transglutaminase by screening existing drug libraries. Chem Biol 15(9):969–978 Le DT, Rapaport SI, Rao LV (1992) Relations between factor viia binding and expression of factor viia/tissue factor catalytic activity on cell surfaces. J Biol Chem 267(22):15447–15454 Lester WA, Guilliatt AM, Enayat MS, Rose P, Hill FG (2007) The r2464c missense mutation in the von willebrand factor gene causes a novel abnormality of multimer electrophoretic mobility and falls into the subgroup of type 2 von willebrand disease “unclassified”. Thromb Haemost 97(1):159–160 Li P, Morris DL, Willcox BE, Steinle A, Spies T, Strong RK (2001) Complex structure of the activating immunoreceptor nkg2d and its mhc class i-like ligand mica. Nat Immunol 2(5):443–451 Li P, Willie ST, Bauer S, Morris DL, Spies T, Strong RK (1999) Crystal structure of the mhc class i homolog mic-a, a gammadelta t cell ligand. Immunity 10(5):577–584 Liang HP, Hogg PJ (2008) Critical importance of the cell system when studying tissue factor de-encryption. Blood 112 (3):912–913; author reply 913
180
J.W.H. Wong and P.J. Hogg
Liu J, Bartesaghi A, Borgnia MJ, Sapiro G, Subramaniam S (2008) Molecular architecture of native hiv-1 gp120 trimers. Nature 455(7209):109–113 Lundstrom J, Holmgren A (1993) Determination of the reduction-oxidation potential of the thioredoxin-like domains of protein disulfide-isomerase from the equilibrium with glutathione and thioredoxin. Biochemistry 32(26):6649–6655 Maekawa A, Schmidt B, de St F, Groth B, Sanejouand YH, Hogg PJ (2006) Evidence for a domainswapped cd4 dimer as the coreceptor for binding to class ii mhc. J Immunol 176(11):6873–6878 Malojcic G, Owen RL, Grimshaw JP, Brozzo MS, Dreher-Teo H, Glockshuber R (2008) A structural and biochemical basis for paps-independent sulfuryl transfer by aryl sulfotransferase from uropathogenic escherichia coli. Proc Natl Acad Sci USA 105(49):19217–19222 Markovic I, Stantchev TS, Fields KH, Tiffany LJ, Tomic M, Weiss CD, Broder CC, Strebel K, Clouse KA (2004) Thiol/disulfide exchange is a prerequisite for cxcr4-tropic hiv-1 envelopemediated t-cell fusion during viral entry. Blood 103(5):1586–1594 Matthias LJ, Azimi I, Tabrett CA, Hogg PJ (2010) Reduced monomeric cd4 is the preferred receptor for hiv. J Biol Chem 285(52):40793–40799 Matthias LJ, Yam PT, Jiang XM, Vandegraaff N, Li P, Poumbourios P, Donoghue N, Hogg PJ (2002) Disulfide exchange in domain 2 of cd4 is required for entry of hiv-1. Nat Immunol 3(8):727–732 Miller SL (1953) A production of amino acids under possible primitive earth conditions. Science 117(3046):528–529 Miller SL (1987) Which organic compounds could have occurred on the prebiotic earth? Cold Spring Harb Symp Quant Biol 52:17–27 Mills JE, Whitford PC, Shaffer J, Onuchic JN, Adams JA, Jennings PA (2007) A novel disulfide bond in the sh2 domain of the c-terminal src kinase controls catalytic activity. J Mol Biol 365(5):1460–1468 Miyakis S, Giannakopoulos B, Krilis SA (2004) Beta 2 glycoprotein i–function in health and disease. Thromb Res 114(5–6):335–346 Montal M (2010) Botulinum neurotoxin: A marvel of protein design. Annu Rev Biochem 79:591–617 Muller YA, Ultsch MH, de Vos AM (1996) The crystal structure of the extracellular domain of human tissue factor refined to 1.7 a resolution. J Mol Biol 256(1):144–159 Murzin AG, Brenner SE, Hubbard T, Chothia C (1995) Scop: A structural classification of proteins database for the investigation of sequences and structures. J Mol Biol 247(4):536–540 Ogawa A, Takayama Y, Sakai H, Chong KT, Takeuchi S, Nakagawa A, Nada S, Okada M, Tsukihara T (2002) Structure of the carboxyl-terminal src kinase, csk. J Biol Chem 277(17): 14351–14354 Ostergaard H, Henriksen A, Hansen FG, Winther JR (2001) Shedding light on disulfide bond formation: Engineering a redox switch in green fluorescent protein. EMBO J 20(21):5853–5862 Ou W, Silver J (2006) Role of protein disulfide isomerase and other thiol-reactive proteins in hiv-1 envelope protein-mediated fusion. Virology 350(2):406–417 Passam FH, Rahgozar S, Qi M, Raftery MJ, Wong JWH, Tanaka K, Ioannou Y, Zhang JY, Gemmell R, Qi JC, Giannakopoulos B, Hughes WE, Hogg PJ, Krilis SA (2010a) Redox control of b2gpivon willebrand factor interaction by thioredoxin-1. J Thromb Haemost 8(8):1754–1762 Passam FH, Rahgozar S, Qi M, Raftery MJ, Wong JWH, Tanaka K, Ioannou Y, Zhang JY, Gemmell R, Qi JC, Hughes WE, Hogg PJ, Krilis SA (2010b) Beta 2 glycoprotein i is a substrate of thiol oxidoreductases. Blood 116(11):1995–1997 Pendurthi UR, Ghosh S, Mandal SK, Rao LV (2007) Tissue factor activation: Is disulfide bond switching a regulatory mechanism? Blood 110(12):3900–3908 Persson E (2008) Protein disulfide isomerase has no stimulatory chaperone effect on factor x activation by factor viia-soluble tissue factor. Thrombosis Res 123(1):171–176 Piatek R, Zalewska B, Kolaj O, Ferens M, Nowicki B, Kur J (2005) Molecular aspects of biogenesis of escherichia coli dr fimbriae: Characterization of drab-drae complexes. Infect Immun 73(1):135–145 Pinkas DM, Strop P, Brunger AT, Khosla C (2007) Transglutaminase 2 undergoes a large conformational change upon activation. PLoS Biol 5(12):e327
8 Allosteric Disulfide Bonds
181
Ploplis VA, Edgington TS, Fair DS (1987) Initiation of the extrinsic pathway of coagulation. Association of factor viia with a cell line expressing tissue factor. J Biol Chem 262(20):9503–9508 Reinhardt C, von Bruhl ML, Manukyan D, Grahl L, Lorenz M, Altmann B, Dlugai S, Hess S, Konrad I, Orschiedt L, Mackman N, Ruddock L, Massberg S, Engelmann B (2008) Protein disulfide isomerase acts as an injury response signal that enhances fibrin generation via tissue factor activation. J Clin Invest 118(3):1110–1122 Richardson J, Richardson D (1989) Prediction of protein structure and the principles of protein conformation. Plenum Press, New York Rozhkova A, Glockshuber R (2008) Thermodynamic aspects of dsbd-mediated electron transport. J Mol Biol 380(5):783–788 Rubinstein R, Fiser A (2008) Predicting disulfide bond connectivity in proteins by correlated mutations analysis. Bioinformatics 24(4):498–504 Ryser HJ, Levy EM, Mandel R, DiSciullo GJ (1994) Inhibition of human immunodeficiency virus infection by agents that interfere with thiol-disulfide interchange upon virus-receptor interaction. Proc Natl Acad Sci USA 91(10):4559–4563 Sakai T, Lund-Hansen T, Paborsky L, Pedersen AH, Kisiel W (1989) Binding of human factors vii and viia to a human bladder carcinoma cell line (j82). Implications for the initiation of the extrinsic pathway of blood coagulation. J Biol Chem 264(17):9980–9988 Sanejouand YH (2004) Domain swapping of cd4 upon dimerization. Proteins 57(1):205–212 Sauer FG, Pinkner JS, Waksman G, Hultgren SJ (2002) Chaperone priming of pilus subunits facilitates a topological transition that drives fiber formation. Cell 111(4):543–551 Schmidt B, Ho L, Hogg PJ (2006) Allosteric disulfide bonds. Biochemistry 45(24):7429–7433 Schmidt B, Hogg PJ (2007) Search for allosteric disulfide bonds in nmr structures. BMC Struct Biol 7:49 Schwarzenbacher R, Zeth K, Diederichs K, Gries A, Kostner GM, Laggner P, Prassl R (1999) Crystal structure of human beta2-glycoprotein i: Implications for phospholipid binding and the antiphospholipid syndrome. EMBO J 18(22):6228–6239 Schwertassek U, Balmer Y, Gutscher M, Weingarten L, Preuss M, Engelhard J, Winkler M, Dick TP (2007) Selective redox regulation of cytokine receptor signaling by extracellular thioredoxin-1. EMBO J 26(13):3086–3097 Siegel M, Khosla C (2007) Transglutaminase 2 inhibitors and their therapeutic role in disease states. Pharmacol Ther 115(2):232–245 Swaminathan S, Eswaramoorthy S (2000) Structural analysis of the catalytic and binding sites of clostridium botulinum neurotoxin b. Nat Struct Biol 7(8):693–699 Thornton JM (1981) Disulphide bridges in globular proteins. J Mol Biol 151(2):261–287 Versteeg HH, Ruf W (2007) Tissue factor coagulant function is enhanced by protein-disulfide isomerase independent of oxidoreductase activity. J Biol Chem 282(35):25416–25424 Wiita AP, Perez-Jimenez R, Walther KA, Grater F, Berne BJ, Holmgren A, Sanchez-Ruiz JM, Fernandez JM (2007) Probing the chemistry of thioredoxin catalysis with force. Nature 450(7166):124–127 Wong JWH, Ho SYW, Hogg PJ (2011) Disulfide bond acquisition through eukaryotic protein evolution. Mol Biol Evol 28(1):327–334 Wouters MA, Lau KK, Hogg PJ (2004) Cross-strand disulphides in cell entry proteins: Poised to act. Bioessays 26(1):73–79 Zav’yalov VP, Chernovskaya TV, Chapman DA, Karlyshev AV, MacIntyre S, Zavialov AV, Vasiliev AM, Denesyuk AI, Zav’yalova GA, Dudich IV, Korpela T, Abramov VM (1997) Influence of the conserved disulphide bond, exposed to the putative binding pocket, on the structure and function of the immunoglobulin-like molecular chaperone caf1m of yersinia pestis. Biochem J 324(Pt 2):571–578 Zhang J (2007) Disulfide-bond reshuffling in the evolution of an ape placental ribonuclease. Mol Biol Evol 24(2):505–512 Zhang JL, Huang Y, Qiu LY, Nickel J, Sebald W (2007) Von willebrand factor type c domaincontaining proteins regulate bone morphogenetic protein signaling through different recognition mechanisms. J Biol Chem 282(27):20002–20014
182
J.W.H. Wong and P.J. Hogg
Zhang JL, Qiu LY, Kotzsch A, Weidauer S, Patterson L, Hammerschmidt M, Sebald W, Mueller TD (2008) Crystal structure analysis reveals how the chordin family member crossveinless 2 blocks bmp-2 receptor binding. Developmental Cell 14(5):739–750 Zhou A, Carrell RW, Murphy MP, Wei Z, Yan Y, Stanley PL, Stein PE, Broughton Pipkin F, Read RJ (2010) A redox switch in angiotensinogen modulates angiotensin release. Nature 468(7320):108–111
Chapter 9
The Problem of Expression of Multidisulfide Bonded Recombinant Proteins in E. coli Silvia A. Arredondo and George Georgiou
Abstract Recombinant proteins currently play an important role in the pharmaceutical industry. Very frequently, proteins of therapeutic value contain complex disulfide bond patterns that are necessary for folding, stability, and/or function. Although the folding of proteins with multiple disulfide bonds in E. coli poses considerable challenges, a number of approaches developed in recent years can now be deployed for the production of such proteins at significant yields. Here, we present a summary of disulfide bond formation in E. coli and the main strategies aimed toward optimization of multidisulfided recombinant protein expression by secretion into the periplasmic space, expression in the cytoplasm of strains engineered to favor the formation of disulfide bonds in that compartment, and finally cell-free synthesis. Keywords Disulfide • Recombinant • Expression • Escherichia coli • Oxidative • Folding • DsbC • DsbA
Abbreviations AP BPTI CHO DTT ER
Alkaline phosphatase Bovine pancreatic trypsin inhibitor Chinese hamster ovary Dithiothreitol Endoplasmic reticulum
G. Georgiou (*) Department of Chemical Engineering, University of Texas, Austin, TX 78712, USA Department of Biomedical Engineering, University of Texas, Austin, TX 78712, USA Department of Molecular Genetics and Microbiology, University of Texas, Austin, TX 78712, USA Institute for Cell and Molecular Biology, University of Texas, Austin, TX 78712, USA e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_9, © Springer Science+Business Media, LLC 2011
183
S.A. Arredondo and G. Georgiou
184
FDA GAPDH GSH GSSG GST l-Arg l-Glu MBP NADPH scFv Sec-pathway SRP
Food and drug administration d-Glyceraldehyde-3-phosphate dehydrogenase Glutathione Glutathione disulfide Glutathione S-transferase l-Arginine l-Glutamine Maltose-binding protein Nicotinamide adenine dinucleotide phosphate reduced Single-chain fragment variable antibody Secretory pathway Signal recognition particle
9.1 Introduction Almost 30 years ago, advances in molecular biology and protein expression led to the beginning of a medical revolution marked by the introduction of the first human recombinant protein used for treatment – recombinant human insulin (Humulin) (Goeddel et al. 1979; Keefer et al. 1981; Sun 1980). Since then, the utilization of recombinant proteins in clinical practice has dramatically increased and now plays a key role in the pharmaceutical industry. As of 2008, more than 130 proteins or peptides had been approved by the FDA and many more candidates are currently in development or clinical trials (Leader et al. 2008). These therapeutics are used in the treatment of a wide range of diseases including cancer, immune disorders, infectious diseases, diabetes, endocrinology disorders, and many others (Aggarwal 2008; Leader et al. 2008; Pavlou and Reichert 2004). The era of protein therapeutics was precipitated by the benefits of recombinant proteins which include faster and less expensive production, reduction of the risk of exposure to animal or human disease agents, and versatility and control for improvement of function or specificity (Leader et al. 2008). The most commonly used production technologies for recombinant protein therapeutics are bacteria, yeast, and mammalian cells. Selecting the best system for the production of a particular protein depends on factors such as post-translational modifications, production yield and volumetric productivity, speed of process development and cell banking, and regulatory issues. The gram-negative bacterium Escherichia coli is very attractive for preparative protein expression because it is well characterized and the process development can be fast, inexpensive, and relatively simple to scale-up. Growth hormones, many cytokines and interferons are, at present, commercially produced in E. coli (Schmidt 2004).However, very frequently, proteins of therapeutic value contain complex disulfide bond patterns that are necessary for folding, stability, and/or function. The need for expressing proteins that contain the authentic (native) disulfide bond pattern with minimal amounts of misfolded proteins (with incorrect, non-native disulfide bonds) poses a considerable challenge for E. coli since the disulfide bond formation
9 Multidisulfide Bonded Proteins in E. coli
185
machinery of the bacterium is equipped to efficiently handle polypeptides containing only a few disulfide bonds (<5). It is estimated that disulfide bonds are required for the folding of approximately one third of human proteins (Hatahet and Ruddock 2009). From a biotechnological point of view, disulfide bond formation is a very significant challenge as many important multidisulfide proteins cannot be expressed efficiently in bacteria and, therefore, have to be produced using more costly expression systems. The optimization of disulfide bond formation for the expression of recombinant proteins in E. coli will certainly lead to higher yields and dramatically lower costs for the production of protein therapeutics and, at the same time, will open possibilities for investigation of new candidates. In this review, we present a summary of disulfide bond formation in E. coli and the main biotechnological efforts aimed toward the optimization of multidisulfided recombinant protein expression.
9.2 Biological Considerations In general, folding of a polypeptide in the cell may begin immediately as soon as the N-terminus of the amino acid chain exits the ribosome. Some proteins might be able to reach their three-dimensional final conformations in the necessary biological timeframe without assistance. However, in the very crowded cellular environment, many proteins are prone to nonproductive intermolecular aggregation and/or premature folding that can lead to the formation of insoluble protein deposits (inclusion bodies), degradation of the polypeptide, or toxicity to the cell. To avoid these problems, cells have evolved a complex system of chaperones and folding catalysts. Molecular chaperones facilitate folding by binding to the accessible hydrophobic areas of polypeptides in unfolded, intermediate, or misfolded conformations, thus preventing aggregation. Folding enzyme catalysts, on the other hand, accelerate rate-limiting conformational transitions, namely peptidyl bond isomerization and disulfide bond formation (Hartl and Hayer-Hartl 2002; Kolaj et al. 2009; Young et al. 2004).
9.2.1 Disulfide Bond Formation in Eukaryotes Despite decades of study, eukaryotic oxidative folding is still not fully understood. In eukaryotic cells, the formation of disulfide bonds takes place primarily in the endoplasmic reticulum (ER), although some vaccinia virus proteins have been reported to form disulfides in the cytoplasm (Locker and Griffiths 1999). A high percentage of the proteins exported into the ER contains disulfides; the oxidation and isomerization of these bonds result from the dual catalytic action of protein disulfide isomerase (PDI). To remain in an active state, PDI is recycled by other ER proteins: Ero1-La and Ero1-Lb in mammalian cells, AERO1 and AERO2 in plants, and Ero1p or Erv2p in yeast (Gruber et al. 2006). PDI is by far the best studied, but not the only, oxidoreductase found in the ER. There are 20 PDI-family members in
186
S.A. Arredondo and G. Georgiou
Fig. 9.1 The thioredoxin fold. The placement of the a-helices with respect to the central b-sheet is shown. The asterisks denote residue-insertion points in the structure. Reproduced from Martin (1995) with permission from Elsevier
humans and 5 in yeast, differing in domain organization, tissue localization, and substrate specificity, and many have yet to be characterized in detail. However, they all have at least one thioredoxin domain (Appenzeller-Herzog and Ellgaard 2008; Gruber et al. 2006; Hatahet and Ruddock 2009).The thioredoxin domain is defined by two features: the first is the thioredoxin fold which consists of a four-stranded b-sheet and three flanking a-helices, illustrated in Fig. 9.1 (Martin 1995). The second feature is the catalytic site located in the thioredoxin fold near the N-terminus of one of the a-helices, and defined by the sequence Cys-X-X-Cys, where X represents any amino acid (Aslund and Beckwith 1999; Ren et al. 2009). Among the members of its family, PDI is believed to be the most abundant, constituting about 0.8% of the total cellular protein in yeast and mammalian cells (Ferrari and Soling 1999). PDI is a monomer comprised of five domains (a, a¢, b, b¢ and c) and has two C-G-H-C catalytic sites, located in the a and a¢ domains. The complete structure of mammalian PDI has not yet been reported; however, the crystal structure of yeast PDI (yPDI) has been solved (Fig. 9.2). The enzyme is a “U”-shaped molecule with the residues lining the interior of the U being mostly hydrophobic. It has been suggested that substrate binding interactions occur in this region. The catalytic domains a and a¢ adopt a thioredoxin fold structure; the same fold is also observed in the noncatalytic b and b¢ domains with minor variations (Tian et al. 2006). In general, PDI is thought to have evolved from thioredoxin domain duplication.
9 Multidisulfide Bonded Proteins in E. coli
187
Fig. 9.2 Molecular structure of yeast PDI. Ribbon diagram of yeast PDI showing the a, b, b¢, and a¢ domains and the c-terminal extension. The active site cysteines are shown in space-filling representation. Reproduced from Tian et al. (2006) with permission from Elsevier
Extensive biochemical studies indicate that while all thioredoxin-like domains of PDI are necessary for isomerization activity in complex protein substrates, oxidation requires only the a or a¢ domain and, isomerization of peptides is catalyzed by a combination of b¢ with either a or a¢ (Darby and Creighton 1995; Darby et al. 1998a; Wilkinson and Gilbert 2004; Xiao et al. 2001). Catalysis of isomerization can occur through either of two mechanisms: in the intramolecular pathway, a mixed disulfide is formed between PDI and its substrate; it is then resolved by a cysteine coming from the substrate resulting in a native disulfide. In the reduction/oxidation pathway, cycles of oxidation and reduction leading to a properly folded substrate, take place (Schwaller et al. 2003). PDI is also able to work as a chaperone by inhibiting the aggregation of proteins lacking disulfide bonds; however, the catalytic cysteines are not required for this activity (Hatahet and Ruddock 2009).
9.2.2 Oxidative Folding in E. coli In contrast to oxidative folding in eukaryotes, where a single enzyme, PDI, catalyzes both cysteine oxidation and disulfide bond isomerization, in E. coli two different enzymes perform these processes independently. For disulfide bonds to form, it is essential for the surrounding environment to be oxidative. Accordingly, in E. coli, after synthesis in the cytoplasm, which has a reducing redox potential, disulfide-containing
188
S.A. Arredondo and G. Georgiou
Fig. 9.3 Oxidative folding in the E. coli periplasm. The disulfide bond formation (DsbA/DsbB), and isomerization (DsbC/DsbD) pathways in the periplasmic compartment are illustrated
proteins need to be translocated into the much more oxidizing periplasm. In this compartment, proteins reach their final native conformation with the assistance of the Dsb (disulfide bond) family of oxidoreductases through two distinct pathways: oxidation-DsbA/DsbB, and isomerization-DsbC/DsbG/DsbD (Fig. 9.3). In the oxidation pathway, the oxidase DsbA, catalyzes disulfide bond formation on a polypeptide substrate by donating its own, very unstable, disulfide. Regeneration of the disulfide bond in DsbA is accomplished by the membrane protein DsbB which in turn transfers electrons onto the respiratory chain through ubiquinone or menaquinone, under aerobic or anaerobic conditions, respectively (Inaba and Ito 2008). DsbA was the first protein involved in disulfide bond formation in the bacterial periplasm to be discovered. It is a 21-kDa monomeric enzyme composed of 189 residues (Bardwell et al. 1991; Kamitani et al. 1992). Structurally, DsbA consists of a thioredoxin-fold domain linked by a flexible hinge to a compact a-helical domain (Fig. 9.4). The catalytic site, which as in all thioredoxins contains a CXXC motif, has the sequence Cys30-Pro31-His32-Cys33 and is located at the N-terminus of the first a-helix of the thioredoxin domain (Martin et al. 1993). Among all thiol–disulfide oxidoreductases, DsbA is the second most oxidizing enzyme known, with a redox potential of ~−120 mV. Its disulfide bond is extremely unstable and therefore it is readily transferred to the free cysteines of substrate proteins (Ito and Inaba 2008; Zapun et al. 1993).
9 Multidisulfide Bonded Proteins in E. coli
189
Fig. 9.4 Molecular structure of DsbA. Ribbon diagram of DsbA showing the catalytic site in space-filling representation (Martin et al. 1993)
DsbA is found predominantly in the oxidized state in vivo, in spite of its strong oxidizing potential, because of the recycling action of the inner membrane protein DsbB. In dsbB mutants, DsbA is mostly reduced. DsbB has two pairs of essential cysteines: Cys104-Cys130, involved in the 1:1 stoichiometric redox interaction with DsbA, and Cys41-Cys44 which takes part in transferring electrons to ubiquinone or menaquinone (Bader et al. 2000; Glockshuber 1999; Grauschopf et al. 2003; Guilhot et al. 1995; Inaba et al. 2006). For proteins containing only two disulfide-forming cysteines, oxidative folding is complete after catalysis by DsbA. Conversely, when a protein contains two or more disulfide bonds, the situation can become more complex. Because of its high oxidizing potential and rapid kinetics, DsbA catalyzes the formation of disulfides in an indiscriminate fashion. Moreover, DsbA is able to introduce disulfides even when there is no thermodynamic drive (due to protein folding) to bring two cysteines into close proximity as is the case for the cytoplasmic protein b-galactosidase when it is expressed with a signal peptide which initiates its export into the periplasm (Bardwell et al. 1991). It should be noted though, that DsbA is more likely to catalyze native disulfide bond formation than metal-catalyzed thiol oxidation by copper as evidenced by the finding that copper-mediated disulfide bond formation increases the requirement for isomerization (Hiniker et al. 2005). In addition, DsbA was reported to catalyze the correct folding of RNase I, a native protein containing three consecutive and one nonconsecutive disulfides, both in vivo and in vitro, (Messens et al. 2007). Recent
190
S.A. Arredondo and G. Georgiou
studies by Kadokura and Beckwith indicate that there is a correlation between the accuracy of DsbA and the mode of translocation of the substrate into the periplasm. By following the in vivo, DsbA-mediated, oxidative folding of alkaline phosphatase that requires the formation of two disulfides linking consecutive cysteines, the authors concluded that cotranslational oxidation seems to take place vectorially. The rate of protein translocation across the secretion pore in the inner membrane (formed by the SecY/E complex) is likely to be a significant factor in this process since it is relatively slow when it occurs cotranslationally in an SRP-dependent manner, but more rapid when it occurs via the post-translational secretion pathway (Kadokura and Beckwith 2009). When the action of DsbA results in the incorrect pairing of cysteines, rearrangement of non-native disulfides becomes necessary for the substrate to reach its proper conformation; thus isomerization represents a limiting step in the expression of recombinant proteins in E. coli. In the isomerization pathway, the main bacterial isomerase, DsbC, rearranges incorrect disulfides enabling the pairing of cysteines that form disulfides in the native state. DsbC needs to be in a reduced state in order to perform the catalysis of disulfide bond isomerization; the protein that maintains DsbC in its active state is DsbD, an inner membrane protein that is itself reduced directly by thioredoxin in the cytoplasm (Rietsch et al. 1997). The identification of DsbC by two different groups in 1994 (Missiakas et al. 1994; Shevchik et al. 1994) led to the discovery of the isomerization pathway in E. coli. DsbC is a homodimeric protein that consists of two 23.4 kDa subunits. Each subunit contains four cysteines, two of which form a buried structural disulfide bond while the other two are located in the Cys98-Gly99-Tyr100-Cys101 catalytic site with a redox potential, E0 = −130 mV (Zapun et al. 1995). In contrast to DsbA, the catalytic cysteines of DsbC are found in a reduced state in vivo (Joly and Swartz 1997). The crystal structure of DsbC (Fig. 9.5) reveals that both monomers come together resulting in a V-shaped molecule. Each monomer is formed by two distinct domains: the N-terminal dimerization domain, and the C-terminal catalytic domain where the thioredoxin fold and the CXXC motif are located. These domains are connected by a slightly kinked helical linker. The surface of the cleft formed by the dimerization of the two units is hypothesized to be involved in substrate recognition (McCarthy et al. 2000). The mechanism of isomerization by DsbC has been proposed to occur as follows: first, Cys98 launches a nucleophilic attack onto a non-native disulfide in the substrate protein resulting in an unstable mixed disulfide complex between DsbC and its target. Then, this mixed disulfide is resolved in one of two ways: either (1) through the attack of a second cysteine from the substrate, consequently forming a more stable disulfide and releasing DsbC in a reduced state or, (2) by reacting with Cys101, leaving DsbC oxidized. The ability of DsbC to catalyze isomerization in vitro suggests the possibility of the former resolving mechanism (Collet et al. 2002; Darby et al. 1998b; Maskos et al. 2003). The latter route is supported by recent studies showing that a dedicated reductase, a mutant of TrxP from Bacteroides fragilis, can fully complement a dsbC strain in the efficient in vivo folding of a protein containing three consecutive and one nonconsecutive disulfide bonds. This DsbD-dependent activity
9 Multidisulfide Bonded Proteins in E. coli
191
Fig. 9.5 Molecular structure of DsbC. Ribbon diagram of the DsbC homodimer showing the catalytic and dimerization domains connected by the a-helical linker. The CGYC catalytic site is shown in space-filling representation. Adapted by permission from Macmillan Publishers Ltd: Nature Structural Biology (McCarthy et al. 2000), copyright (2000)
provides direct evidence that proper disulfide formation can result from cycles of reduction by DsbC and re-oxidation by DsbA (Shouldice et al. 2010). DsbC is recycled into the reduced state by DsbD, an inner membrane protein comprised of one transmembrane (b) and two periplasmic domains (a, g), in a process that uses reducing equivalents originating in the cytoplasm, ultimately from NADPH. This electron cascade starts with the reduction of the disulfide in DsbDb by thioredoxin 1. Electrons are then transferred to DsbDg, the C-terminal periplasmic domain, which in turn reduces the disulfide in DsbDa, the N-terminal periplasmic domain. Once reduced, DsbDa is ready for reactivation of DsbC (Cho and Beckwith 2009; Haebel et al. 2002; Katzen and Beckwith 2000; Rietsch et al. 1997; Rozhkova et al. 2004). While the model above is supported by extensive evidence and is most widely accepted, it should be noted that Vertommen et al. have proposed an alternative mechanism based on phenotypic differences of dsbA dsbC cells relative to dsbA, dsbC, and dsbA dsbD mutants. According to this model, DsbC can function independently of DsbD and might be able to catalyze both oxidation and isomerization by cycling from the reduced to the oxidized state upon interaction with the substrate (Vertommen et al. 2008). Besides being a reductase and an isomerization catalyst, DsbC has also been shown to function as a chaperone, assisting the folding of proteins without disulfide bonds, such as d-glyceraldehyde-3-phosphate dehydrogenase (GAPDH); the Cys98-Cys101 catalytic cysteines are not required for this activity (Chen et al. 1999; Liu and Wang 2001).The inner surface of the cleft that forms by the association of the monomers in DsbC accounts for 45% of the solvent-accessible surface area of
192
S.A. Arredondo and G. Georgiou
Fig. 9.6 Ribbon diagram and surface representation of the hydrophobic cleft of DsbC that may account for substrate binding and chaperone activity, observed from the top of the V-shaped molecule. Reprinted by permission from Macmillan Publishers Ltd: Nature Structural Biology (McCarthy et al. 2000), copyright (2000)
the dimer and is lined with 46 uncharged or hydrophobic residues. Having the dimensions 40 × 40 × 25 Å, this cleft is sufficiently large to accommodate small protein domains and it is believed to play an important role in the recognition of unfolded polypeptides (Fig. 9.6, (McCarthy et al. 2000)). Substrate recognition is essential for DsbC to effectively catalyze disulfide bond isomerization and to assist folding as a chaperone, however, little is known about the binding interactions between this enzyme and its substrates. Sun and Wang (2000) determined that the C-terminal domain of DsbC, by itself, is a monomer not able to act as a chaperone or to catalyze isomerization in vitro, and concluded that the N-terminal dimerization domain is indispensable for activity and for substrate binding (Sun and Wang 2000). However, a more recent study employing engineered DsbC fusions containing only one functional catalytic domain concluded that a single active domain is sufficient for isomerase activity as long as the putative substrate recognition area is conserved (Arredondo et al. 2009). It should be noted that similar results have been reported for the eukaryotic isomerase PDI in which the major binding site has been mapped to a small hydrophobic pocket within the b¢ domain (Klappa et al. 1998; Pirneskoski et al. 2004). Moreover, one catalytic domain (a or a¢) in combination with the noncatalytic b¢ domain is sufficient to catalyze simple disulfide isomerization (Wilkinson and Gilbert 2004; Xiao et al. 2001). So far only a few E. coli DsbC substrates have been identified: the penicillin insensitive endopeptidase MepA, ribonuclease RNase I, and acid phosphatase AppA, all of which have nonconsecutive disulfides (Berkmen et al. 2005; Hiniker and Bardwell 2004). Although, information on the substrate specificity of DsbC is not available, we have evidence that the enzyme most likely recognizes conformational, rather than linear, epitopes (Arredondo and Georgiou unpublished data). DsbG, another E. coli periplasmic protein belonging to the thioredoxin superfamily, has also been shown to display disulfide bond isomerization activity. Overexpression of DsbG can complement some of the phenotypes of dsbC cells. However, the isomerization activity of DsbG seems to be limited to a narrower range of substrates. As with DsbC, DsbG has been shown to be an effective molecular chaperone in vitro (Bessette et al. 1999b; Shao et al. 2000). Like DsbC, DsbG is a symmetric V-shaped homodimer comprising of two monomers having
9 Multidisulfide Bonded Proteins in E. coli
193
interacting dimerization domains, each of which is connected via an a-helical linker to a thioredoxin domain. However, the a-helical linkers in DsbG are longer, expanding the volume of the putative substrate-binding region, and there are negatively charged patches present in this same region not found in DsbC (Heras et al. 2004; Yeh et al. 2007). Single point mutations in this region enable DsbG to perform functions that are normally only displayed by DsbC (Hiniker et al. 2007). Very recently, Depuydt et al. discovered that DsbG is an effective reductase of protein cysteines oxidized to sulfenic acid (−SOH). These authors report that the membrane protein YbiS, which requires a reduced single catalytic cysteine for its function, is maintained in its active conformation by DsbG. Given the strongly oxidative environment of the periplasm, partially or fully folded proteins with single or nonpaired cysteines are prone to deleterious oxidation first to sulfenic acid, a process which is reversible, and then to irreversible higher oxidation states, namely to sulfinic and sulfonic acids. DsbG uses reducing equivalents obtained from the cytoplasm through DsbD to catalyze the reduction of protein sulfenic groups. Sulfenylation of periplasmic proteins was reported to occur more extensively in dsbG dsbC indicating that, although to a lesser extent, DsbC is also involved in the protection of single cysteines (Depuydt et al. 2009). It is reasonable to ask what is preventing the cysteine oxidation and disulfide isomerization pathways from establishing a futile electron cycle, or in other words, why is DsbC not oxidized by DsbB, and DsbD by DsbA, since these proteins are all localized within the same subcellular compartment. Rozhkova et al. reported that nonphysiologically relevant reactions between the enzymes of the oxidative pathway (DsbB, DsbA) with those of the reducing pathway (DsbD, DsbC and DsbG) are 103 to 107 fold slower relative to the reaction rates of DsbB with its physiological substrate DsbA and of dsbD with DsbC. Even the DsbB–DsbC reaction, the fastest among the nonproductive thiol–disulfide electron transfer processes, is about 1,000 times slower than the physiologically relevant oxidation of DsbA by DsbB. These kinetic barriers allow the co-existence of both pathways in the periplasmic envelope (Bader et al. 2000; Rozhkova et al. 2004). In vivo studies, biochemical analysis, and structural evidence suggest that the dimerization of DsbC prevents its active sites from being available for oxidation by DsbB. Bader et al. was able to isolate mutants of DsbC capable of catalyzing DsbB-dependent oxidation as a result of these mutations preventing the DsbC monomers to associate (Bader et al. 2001). Binding of DsbC to DsbB is predicted to cause one of the DsbC domains to clash onto the lipid bilayer membrane, hence making it sterically unfavorable (Bader et al. 2001; Inaba et al. 2006). Mutations in an a-helix of DsbB that probably change the orientation of its catalytic cysteines, deletions in the a-helical linker of DsbC that change the opposing orientation of the two catalytic sites of the enzyme, or engineered DsbC variants in which one catalytic domain has been removed compromise the kinetic isolation of DsbC from DsbB/A. In these instances, DsbC is more susceptible to oxidation by DsbB and accumulates in the periplasm as a roughly equimolar mixture of the oxidized and reduced forms (Arredondo et al. 2009; Pan et al. 2008; Segatori et al. 2006).
194
S.A. Arredondo and G. Georgiou
9.3 Expression of Multidisulfided Heterologous Proteins in E. coli Recombinant proteins that fail to fold properly in E. coli are either degraded or accumulate in an aggregated state as inclusion bodies (Bowden et al. 1991). The formation of inclusion bodies can simplify purification but, although some proteins are currently produced at commercial scale in this manner, the refolding process can be expensive or inefficient (Burgess 2009). The refolding of proteins containing disulfides adds another level of complexity. First, a reducing agent such as DTT must be added to disrupt any non-native disulfides. Subsequently, an optimized redox buffer (e.g., GSSG/GSH, Cysteine/Cystine) is required during refolding to enable the formation of native disulfide bonds (Burgess 2009). The drawbacks of protein refolding from inclusion body expression were illustrated by a comparative study on the process economics for the production of human tissue plasminogen activator or tPA (35 cysteines, 17 disulfides). Even though tpA could be expressed as inclusion bodies at a yield of 460 mg/L in E. coli compared to 33.5 mg/L in CHO cells, the downstream processing cost associated with refolding made the E. coli production process highly unfavorable (Datar et al. 1993). Extensive efforts have been dedicated to the development of expression methods for the production of soluble proteins in E. coli. Soluble protein expression can be enhanced, for example, by codon optimization (Angov et al. 2008; Kudla et al. 2009), choice of promoter, e.g., T7, trc, lac, araBAD (Makrides 1996), expression induction at lower temperatures (Qing et al. 2004; Schein 1989), media modification with additives such as l-Arg, l-Glu, GSH, or sucrose (Bowden and Georgiou 1988; Golovanov et al. 2004; Schaffner et al. 2001), fusion tags and proteins, e.g., GST, MBP, DsbA, TrxA, SUMO (Hammarstrom et al. 2002; Malakhov et al. 2004; Waugh 2005; Zhang et al. 1998), and coexpression of molecular chaperones such as GroEL/GroES, DnaK/DnaJ-GrpE, and trigger factor (Hartl and Hayer-Hartl 2002; Hoffmann et al. 2010; Kolaj et al. 2009; Young et al. 2004). Additionally, following the discovery of the oxidation and isomerization pathways in E. coli, the biotechnological potential of manipulating oxidative folding in bacteria for recombinant expression purposes became evident. In early studies, expression of the eukaryotic disulfide isomerase PDI in the bacterial periplasm was shown to be unable to assist recombinant protein folding because PDI was oxidized by DsbB and could not function as an isomerase (Humphreys et al. 1995; Ostermeier et al. 1996; Zhan et al. 1999). Nonetheless, PDI has been shown to positively affect the expression of some proteins such as BPTI and mouse urokinase, possibly because it acted as a slower oxidase relative to DsbA thus leaving enough time for native conformers to form prior to disulfide bond formation (Ostermeier et al. 1996; Stafford and Lund 2000). Qiu et al. (1998) reported a dramatic improvement in the expression of active human tissue plasminogen activator in the E. coli periplasm by co-expressing DsbC. Proper disulfide bond formation in tPA resulted in the production of 180 mg/L of purified active enzyme (Qiu et al. 1998). Coexpression of DsbC
9 Multidisulfide Bonded Proteins in E. coli
195
alone, or sometimes in combination with the other members of the Dsb family, is still the most efficient and widely utilized approach for improving the periplasmic expression of proteins with complex disulfide patterns. Several studies have reported successful production of multidisulfided proteins in E. coli following this approach, including human nerve growth factor, scFv antibodies, and Ragi bifunctional inhibitor with three, two, and five disulfides, respectively (Kurokawa et al. 2001; Maskos et al. 2003; Zhang et al. 2002) (see Table 9.1 for a more extensive list). In certain cases, expression of multidisulfided proteins in the periplasm can be further enhanced by the action of the folding modulators FkpA, SurA, and/or Skp. Although these proteins are not directly involved in disulfide formation or isomerization, they do participate in periplasmic folding as chaperones and/or cis–trans proline isomerases (Missiakas et al. 1996) and therefore their coexpression in combination with DsbA/B/C/D has proven to be beneficial. FkpA and SurA are bacterial peptidyl-prolyl cis–trans isomerases (PPIases) that catalyze the isomerization of the peptide bond preceding a proline residue from the trans to the cis position, often a rate-limiting step in protein folding. However, the ability of these two enzymes to assist in the expression of heterologous proteins may be related more to their chaperone activity rather than to the PPIase activity which is localized in a structurally different domain (Behrens et al. 2001; Hennecke et al. 2005; Ramm and Pluckthun 2000; Saul et al. 2004). The effect of simultaneous overexpression of DsbC, DsbA, FkpA, and SurA from a single plasmid has been reported by Schlapschy et al. to enhance the expression of human plasma retinol-binding protein, the extracellular domain of a dendritic cell membrane receptor (three disulfides each), and of a malaria vaccine candidate Pfs48/45 fragment with ten cysteines (Outchkourov et al. 2008; Schlapschy et al. 2006). Skp is a molecular chaperone whose main function involves binding to early outer-membrane protein folding intermediates to prevent their aggregation in the periplasm. More recently, however, soluble periplasmic substrates of Skp have also been identified (Jarchow et al. 2008; Schafer et al. 1999). The coexpression of Skp has been reported to assist in folding of T-cell receptors, scAbs and scFvs (Hayhurst and Harris 1999; Mavrangelos et al. 2001; Maynard et al. 2005).
9.3.1 Cytoplasmic Expression In contrast to the periplasmic compartment, the cytoplasm of E. coli has a reducing potential. Although cytoplasmic proteins do not have stable disulfides, metabolic enzymes like the Hsp33 chaperone, the OxyR and RsrA transcription factors, ribonuclease reductase, and sulfoxide reductase depend on disulfide oxidation–reduction cycles to perform their respective functions (Bessette et al. 1999a; Ritz and Beckwith 2001). After catalysis, these proteins are left in an oxidized state and reduction is necessary for continued activity. Cytoplasmic thiol reduction originates from
Table 9.1 Representative summary of soluble recombinant proteins that require oxidative folding reported to be secreted into the periplasmic space of E. coli Periplasmic expression Protein SS# Yield Relevant conditions Application/function References Anti-bladder carcinoma 2 Several fold DsbG-fusion, co-exp: DsbC Antibody Zhang et al. (2002) scFv Anti-domoic acid scFv 2 1 mg/L Co-exp: DsbC, DnaKJE Immunosensor Hu et al. (2007) Anti-idiotypic scFv 2 n/a Co-exp: FkpA Antibody Padiolleau-Lefevre et al. (2006) Art v 1 3, 4 5–10-fold Osmotic stress, codon optimized Immunotherapy candidate Gadermaier et al. (2010) a-sarcin 2 7 mg/L Co-exp:TrxA (cytoplasmic) RNase, protein Garcia-Ortega et al. (2000) biosynthesis inhibitor BDNF 3 2 mg/L Co-exp: DsbA, DsbB, DsbC, Nervous system Hoshino et al. (2002) DsbD development BPTI 3 15-fold Glutathione, co-exp: PDI Protease inhibitor Ostermeier et al. (1996) 3 3.6 mg/L Co-exp: DsbA, DsbB, DsbC, Mannose binding Soanes et al. (2008) Lectin-like domain of SCLRC DsbD c-type lectin 5 cys Several fold Co-exp:GroEL/ES, DnaKJ, Aldehyde and ketone Lee et al. (2004) Cyclohexanone monooxygenase GrpE, DsbA, DsbC biocatalyst HIV-binding lectin Schlapschy et al. (2006) DC-SIGN fragment 3 10-fold Co-exp: DsbA, DsbC, FkpA, SurA Hcps 4 n/a Maybe involved Devi et al. (2006) in immune response Growth factor Kurokawa et al. (2001) h-NGF 3 3-fold Co-exp: DsbA, DsbB, DsbC, DsbD nervous system HRP 4 10-fold Co-exp: DsbA, DsbB, Molecular probe Kurokawa et al. (2000) DsbC, DsbD Human G-CSF 2 236 mg/mL Endoxylanase signal seq Neutropenia Jeong and Lee (2001) Human Renal 2 5-fold Dimerization, Cancer chemotherapy O’Dwyer et al. (2009) Dipeptidase (HDP) unpaired-cys removal candidate Huridin 3 60 mg/L l-Asparaginase signal seq Thrombotic diseases Tan et al. (2002) IgG 16 1 g/L Co-exp: DsbA, DsbC Full-length antibody Reilly and Yansura (2010)
196 S.A. Arredondo and G. Georgiou
1 mg/L 150 mg/L
51.5 mg/L 3 mg/L 5 mg/L
n/a 0.5 mg/L
1 1
3
1–3 2
3 10 cys
3 3 5
2, 3 2
Lipase B (PalB)
Lipocalins Nogo-A ectodomain
Pepsinogen Pfs48/45 fragment
Proinsulin Proinsulin RAGI bifunctional inhibitor scAbs T-cell receptor fragments FkpA-fusion, co-exp: FkpA Heat shock response, co-exp: DsbA
Ecotin-fusion MBP-fusion, co-exp: DsbA, DsbC, FkpA, SurA Ecotin-fusion DsbA-fusion, l-arginine, ethanol Reduced glutathione, co-exp:DsbC
OmpA signal seq Endoxylanase signal seq, Co-exp: DsbA Co-exp:DegP; DsbA; FkpA, SurA, DsbA and DsbC Co-exp: DsbA, DsbC, FkpA, SurA
Low temperature DsbA signal seq, codon optimized
Relevant conditions
Antibody Cellular immune response
Diabetes Diabetes Trypsin inhibitor
Biotransformation, biodiesel production Ligand-binding proteins Target for treatment of spinal cord injuries Aspartic protease Vaccine candidate
Obesity Obesity
Cytokine Carboxypeptidase inhibitor
Application/function
tPA 17 0.2 mg/L Co-exp: DsbC Thrombolytic agent The number of disulfide bonds (SS#), and the reported yield and relevant expression conditions are listed
7.6 mg/L 1 mg/L
n/a 0.2 mg/L
7-fold
n/a 470 mg/L
Yield
5 4
SS#
IL-2Ra Leech carboxypeptidase inhibitor Leptins Leptin
Periplasmic expression Protein
Zhang et al. (2003) Wulfing and Pluckthun (1994) Qiu et al. (1998)
Malik et al. (2007) Winter et al. (2001) Maskos et al. (2003)
Malik et al. (2006) Outchkourov et al. (2008)
Breustedt et al. (2006) Zander et al. (2007)
Xu et al. (2008a)
Guisez et al. (1998) Jeong and Lee (2000)
Dracheva et al. (1995) Puertas and Betton (2009)
References 9 Multidisulfide Bonded Proteins in E. coli 197
198
S.A. Arredondo and G. Georgiou
Fig. 9.7 Cytoplasmic disulfide bond reducing systems in E. coli. The thioredoxin and glutaredoxin pathways are illustrated. The genes encoding for each component are shown in parentheses
NADPH and takes place through the thioredoxin/thioredoxin reductase and glutathione/glutaredoxin pathways (Fig. 9.7). Thioredoxins 1 and 2 and thioredoxin reductase, encoded by trxA, trxC, and trxB respectively, constitute the thioredoxin pathway. Both thioredoxins are kept catalytically active by thioredoxin reductase through the flow of electrons from NADPH. Trx1, the protein in which the “thioredoxin fold” structure is based, contains a C-XX-C catalytic motif, similar to that of the Dsb enzymes, and has a low redox potential (−270 mV) (Martin 1995; Vlamis-Gardikas 2008). The glutaredoxin pathway is comprised of three glutaredoxins which also belong to the thioredoxin superfamily, glutathione and glutathione reductase. Glutaredoxins are reduced by glutathione, a small tripeptide that itself is reduced by glutathione reductase via the transfer of electrons from NADPH (Vlamis-Gardikas 2008). E. coli mutants in which both the thioredoxin and the glutaredoxin pathways are impaired require exogenous reductants for normal growth (Prinz et al. 1997). Early studies suggested that prevention of cysteine oxidation in the cytoplasm is dependent on the thioredoxin and glutaredoxin pathways (Derman et al. 1993). Genetic screens using a signal-peptide less version of alkaline phosphatase, a protein that normally resides in the periplasm, led to the isolation of E. coli mutants that allowed disulfide bond formation in the cytoplasm. The mutation was mapped to the trxB gene that encodes for thioredoxin reductase, the enzyme in charge of maintaining
9 Multidisulfide Bonded Proteins in E. coli
199
the thioredoxins in a reduced state. Therefore, it was initially suggested that the accumulation of oxidized thioredoxin, encoded by trxA, could be responsible for oxidative folding. However, there was no significant difference in the alkaline phosphatase activity of a double mutant lacking both thioredoxin and thioredoxin reductase (trxA trxB) (Derman et al. 1993). Study of a comprehensive set of mutants lacking proteins in the thioredoxin and glutaredoxin pathways indicated that both systems contribute to the reduction of disulfides in the cytoplasm and they are able to partially substitute for each other. Strains deficient in thioredoxin and glutathione reductase (trxA gor), and thioredoxin and glutathione (trxA gshA) grew normally and permitted a high level of disulfide bond formation in alkaline phosphatase in the cytoplasm. Inactivation of both pathways, as in the thioredoxin reductase and glutathione reductase (trxB gor) or the thioredoxin reductase and glutathione (trxB gshA) mutants, was found to result in even higher levels of disulfide bond formation; however, these cells are not viable under aerobic conditions unless an external reducing agent is available (Prinz et al. 1997). Suppression mutations that restored normal growth to trxB gor or trxB gshA mutants were isolated and found to map in the ahpC gene encoding one of the two subunits of alkyl hydroperoxidase. The trxB gor suppressor mutations, most notably an allele that occurred at high frequency and was termed ahpC*, resulted in a change of function in alkyl hydroperoxidase which enabled the protein to reduce glutathione (Bessette et al. 1999a; Faulkner et al. 2008; Ritz et al. 2001; Yamamoto et al. 2008). Interestingly, in trxB gor ahpC* or trxB gshA supp mutants disulfide bond formation was shown to be catalyzed by the thioredoxins. Thus, even though thioredoxins normally catalyze disulfide bond reduction, their function is reversed in mutants in which the cytoplasm is oxidizing and trxB is inactivated. The active involvement of thioredoxins 1 and 2 in disulfide bond formation in the cytoplasm was evident by the finding that both trxA trxC and trxA trxB trxC mutants showed little or no alkaline phosphatase activity (Stewart et al. 1998). The demonstration that active alkaline phosphatase can be expressed in the cytoplasm opened the possibility for the expression of even more complex heterologous proteins containing multiple disulfide bonds in E. coli. Encouraging results came early from Derman et al. who were able to express enzymatically active mouse urokinase, containing six disulfide bonds, in the trxB strain (Derman et al. 1993). Bessette et al. reported the successful high level expression of the more challenging targets vtPA and tPA, having 9 and 17 disulfides, respectively, using the trxB gor ahpC* strain. Furthermore, the coexpression of thioredoxin 1 and more oxidizing thioredoxin mutants in this strain increased the level of active vtPA to varying degrees up to 15-fold. While coexpression of signal-peptide less DsbA actually reduced the level of vtPA activity, cytoplasmic coexpression of DsbC resulted in a 20-fold increase. The level of active vtPA expressed in the cytoplasm of the trxB gor ahpC* strain was higher than the observed in periplasmic expression under optimized conditions (Bessette et al. 1999a). Following these studies, there have been numerous reports of successful expression of multidisulfide bonded proteins in the E. coli cytoplasm (see Table 9.2 for a representative list). Recent examples of proteins expressed in a trxB gor background include human midkine (five disulfides),
2.2 mg/L
8 mg/L
4 mg/L 0.5 mg/L
20 mg/L 10 mg/L
0.5 mg/L 3 mg/L
10 mg/L
3
3
2
3 4 4 6 2
3 3
3 5
4 2
1
b-defensins
Biotinylated hsRAGE b-lactoglobulin
BmK86 Brazzein BSPHI Calobin Chemokines
Cox17 Extracellular domain of ISG75 Fibrolase Gaussia Luciferase Gluc Hepcidin HF6478 domain of LEKTI hLH/CG
2 mg/L 27 mg/L n/a n/a 1–15 mg/L
130 mg/L
0.8 mg/L
5
Anti-digoxin Fab
Yield 12-fold
SS # 2
Protein Anti-c-Met scFv
Cytoplasmic expression
trxB gor, TrxA-fusion
trxB gor, TrxA-fusion trxB gor, TrxA-fusion
GST-fusion SUMO-tag trxB gor, TrxA-fusion trxB gor, co-exp: DsbC, TrxA trxB gor, TrxA-SUMO tag, codon optimized trxB gor trxB gor, Glu and Arg buffer Co-exp: DsbC Cold shock induction
trxB gor, co-exp: DsbC
Relevant conditions trxB gor, sucrose, co-exp: DsbC trxB gor, co-exp: GroEL/ES, TF, Skp, DsbC TrxA-fusion, codon optimized trxB gor
Hormone receptor
Thrombotic diseases Bioluminescent reporter protein Antimicrobial, iron metabolism Proteinase inhibitor
Copper chaperone Sleeping sickness diagnostics
Target for blocking diabetic complications Milk allergen, protein folding studies K+ channel research Potential low calorie sweetener Sperm capacitation Thrombotic diseases Cell signalling molecules
Antimicrobials
Antibody
Application/function Antibody cancer therapy
Lobel et al. (2001)
Gagliardo et al. (2008) Lauber et al. (2001)
Zhang et al. (2006) Rathnayaka et al. (2010)
Voronova et al. (2007) Tran et al. (2008)
Mao et al. (2007) Assadi-Porter et al. (2008) Lefebvre et al. (2009) Yuan et al. (2004) Lu et al. (2009)
Kumano-Kuramochi et al. (2008) Ponniah et al. (2010)
Huang et al. (2008)
Levy et al. (2001)
References Heo et al. (2006)
Table 9.2 Representative summary of soluble recombinant proteins that require oxidative folding reported to be expressed in the cytoplasmic compartment of E. coli
200 S.A. Arredondo and G. Georgiou
5
2
3
2
Human midkine
Human muskelin
Lipase B (PalB)
MMP-2 (FNII-2) domain scFv SPARC T-cell receptor tPA Viscotoxin A3
40 mg/L
130 U/L
1.5 mg/L
2.5 mg/L
n/a
trxB gor rare-tRNAs strain, SUMO-tag trxB gor, DsbA-fusion, co-exp:DsbA trxB, TrxA-fusion
Stress-response proteins fusions, codon optimized trxB gor Cancer therapeutic development Cell adhesion, cytoskeleton dynamics Biodiesel production
Neutropenia
Regulation of extracellular matrix 2 0.3 mg/L trxB gor, co-exp: DsbC antibody 7 n/a trxB, co-exp: DnaKJ Wound healing tissue repair 5 3.4 mg/L trxB gor Antigen recognition 17 15-fold trxB gor, co-exp: TrxA, DsbC Thrombolytic agent 3 5.2 mg/L Rare-tRNAs strain, TrxAActive component in medical fusions mistletoe The number of disulfide bonds (SS#), and the reported yield and relevant expression conditions are listed
2
Human G-CSF
Jurado et al. (2002) Schneider et al. (1997) Liddy et al. (2010) Bessette et al. (1999a) Bogomolovas et al. (2009)
Peisley and Gooley (2007)
Xu et al. (2008b)
Kiedzierska et al. (2008)
Yan et al. (2010)
Song et al. (2009)
9 Multidisulfide Bonded Proteins in E. coli 201
202
S.A. Arredondo and G. Georgiou
hepcidin (four disulfides) expressed as a Trx1 fusion, and snake calobin (six disulfides) assisted by coexpression of Trx1 and signal-peptide less DsbC (Gagliardo et al. 2008; Yan et al. 2010; Yuan et al. 2004). Additionally, a cytoplasmic–periplasmic expression method in which the protein is allowed to first fold and form disulfides in the cytoplasm of oxidizing strains and then translocate into the periplasm has been demonstrated (DeLisa et al. 2003; Kim et al. 2005). In bacteria, there are two major pathways for the export of proteins across the cytoplasmic membrane: the twin-arginine or Tat pathway which translocates proteins that are already in their folded, native state and/or have a cofactor; and, the Sec-pathway that transports proteins in unfolded conformations either cotranslationally in a SRP-dependent fashion, as is the case for membrane proteins, or post-translationally. With few exceptions, nonmembrane proteins, including the ones that will require the formation of disulfide bonds, are translocated posttranslationally via the Sec pathway (Driessen and Nouwen 2008; Lee et al. 2006; Natale et al. 2008). A well-studied example is alkaline phosphatase which is naturally exported through the Sec pathway in an unfolded conformation, and once in the periplasm, disulfide bond formation takes place and the protein becomes active (Kadokura and Beckwith 2009). By fusing AP to Tat-specific leader peptides and providing an oxidizing cytoplasm (trxB gor ahpC*), DeLisa et al. were able to properly fold AP in the cytoplasm and then export it to the periplasm. Similar results were obtained with the scFv and Fab antibody fragments of 26–10 antidigoxin having two and five disulfide bonds, respectively. The Fab yield was significantly increased by coexpression of signal-sequenceless DsbC (DeLisa et al. 2003). Furthermore, active vtPA and tPA (9 and 17 disulfides) were obtained in the periplasmic space following analogous procedures; surprisingly, coexpression of DsbC was not required for this activity (Kim et al. 2005).
9.3.2 Cell-Free Protein Synthesis Systems The constant search for simpler, faster, and more controllable protein expression methods has lead to the investigation of cell-free protein synthesis systems. These systems utilize cell extracts from different organisms, e.g., E. coli, wheat germ, rabbit reticulocytes, and Leishmania tarantolae (Mureev et al. 2009; Nakano and Yamane 1998) in combination with all the factors needed for protein synthesis. E. coli extracts are widely used for in vitro protein synthesis even at preparative scales. The bacterial extracts provide ribosomes and all other enzymes and factors required for translation. Amino acids, nucleotides, salts, an energy-regenerating source, exogenous RNA polymerase, and of course the particular DNA template encoding the target heterologous protein have to be supplied separately (Fig. 9.8 (Nakano and Yamane 1998; Sitaraman and Chatterjee 2009; Swartz et al. 2004)). The folding of multidisulfide bonded proteins produced in cell-free systems has also been a challenge since the standard reaction conditions do not provide a favorable environment for oxidative protein folding to take place. However, addition of a glutathione redox buffer, purified oxidoreductase PDI, and molecular chaperones to
9 Multidisulfide Bonded Proteins in E. coli
203
Fig. 9.8 Cell-free protein synthesis. The schematic shows the coupled processes of transcription and translation obtained with E. coli cell extracts. Cells are grown and lysed to prepare the extract followed by the addition of substrates and salts. Protein synthesis is initiated by the addition of the template. Reproduced from Swartz (2001) with permission from Elsevier
the reaction enabled the in vitro synthesis of properly folded, functional single chain antibodies containing two disulfides (Merk et al. 1999; Ryabova et al. 1997). Subsequently, Kim and Swartz reported that while these conditions did not promote the folding of urokinase which contains six disulfides, blocking of free endogenous sulfhydryl groups by pretreating the cell extract with iodoacetamide in combination with an oxidizing glutathione buffer, pH optimization, and addition of the bacterial oxidoreductase DsbC resulted in the expression of up to 40 mg/mL of enzymatically active urokinase (Kim and Swartz 2004). Continued efforts for the optimization of oxidative folding in cell-free systems include manipulation of the redox potential, addition of chaperones like Skp or GroEL/S, in vivo overexpression of chaperones and DsbC (as opposed to the more expensive addition of purified enzymes), inactivation of the cytoplasmic reduction system of the source strain by deleting gor and/or physically removing TrxB, etc. These manipulations have led to the in vitro expression of multidisulfided proteins like vtPA, human erythropoietin, GM-CSF, scFvs, mGM-scFv fusions, MAK33 IgG, and BCXRH1 Fab (9, 2, 2, 2, 4, 16, and 5 disulfides, respectively) at attractive yields (Frey et al. 2008; Goerke and Swartz 2008; Kang et al. 2005; Knapp et al. 2007; Oh et al. 2010; Yang et al. 2005; Yin and Swartz 2004). Although in vitro expression is emerging as an attractive alternative for the production of some disulfided bonded proteins and there has been substantial improvements in terms of reagent cost and reaction standardization (Calhoun and Swartz 2005; Goerke and Swartz 2008), the scalability of the process still remains to be demonstrated.
204
S.A. Arredondo and G. Georgiou
9.3.3 Protein Engineering Approaches for Aiding the Expression of Multidisulfide Proteins Efforts have been made to facilitate the expression of proteins in the periplasm by engineering periplasmic oxidoreductases that are optimized for assisting the formation of disulfide bonds in heterologous proteins. PDI, the eukaryotic counterpart, is known to catalyze isomerization at higher rates than DsbC in vitro (Segatori et al. 2004; Zhao et al. 2003) indicating that DsbC or other enzymes may have the potential to be engineered for improved catalytic proficiency. Mutations in the C-X-X-C dipeptide within the catalytic sites of DsbA and DsbC exert a strong influence in the redox potential of these enzymes (Grauschopf et al. 1995). Bessette et al. carried out saturation mutagenesis of the active site dipeptide of DsbA isolating mutants that had a lower oxidation potential and enhanced the expression of mouse urokinase three- to fourfold. Likewise, saturation mutagenesis and screening of the library of mutants in the DsbC active site led to the isolation of two variants having a CGFC- or CTFC-catalytic motif instead of the CGYC of the w.t. enzyme. These two DsbC variants conferred higher yields of vtPA and mouse urokinase (Bessette et al. 2001). Thioredoxin (PDI-like) and DsbA demonstrate negligible isomerase activity in vitro (Lundstrom et al. 1992; Segatori et al. 2004) and are unable to catalyze in vivo isomerization in the periplasm (Jonda et al. 1999). However, Segatori et al. showed that fusion of these two enzymes to the dimerization domain of DsbC results in proteins that are capable of catalyzing disulfide isomerization. While the DsbC-TrxA and DsbC-DsbA chimeras resulted in significant in vivo disulfide isomerase activity, the oxidized form of these proteins was also shown to serve as an oxidase, catalyzing de novo disulfide bond formation (Segatori et al. 2004). In an independent study, Zhao et al. presented the biochemical analysis of DsbC-PDIa, a fusion of the a domain of PDI with the dimerization domain of DsbC, which exhibited 50% higher in vitro isomerase activity than DsbC (Zhao et al. 2003). The fusion of DsbA to the dimerization domain and linker of FkpA resulted in the creation of artificial disulfide isomerases that were able to efficiently catalyze disulfide bond isomerization of vtPA in vivo. Interestingly, these chimeras catalyzed both disulfide bond formation and isomerization of vtPA in a dsbA mutant. Random mutagenesis of FkpA-DsbA resulted in isolation of variants containing a His to Tyr mutation in the dipeptide of the catalytic site that conferred higher resistance to Cu(II) in dsbA dsbC cells, a phenotype that is dependent on the disulfide isomerization activity in the periplasm (Arredondo et al. 2008). The directed evolution of DsbG has also been reported. In order to find DsbG mutants that closely resemble the isomerase activity of DsbC, Hiniker et al. subjected a DsbG library to selective pressure based on the copper (II) sensitivity of dsbC strains. The authors isolated point mutants that mapped to regions in DsbG believed to be important for substrate binding. Although these variants were not tested for in vivo folding of multidisulfided proteins, they displayed more efficient isomerization in vitro in comparison to wild-type DsbG (Hiniker et al. 2005, 2007).
9 Multidisulfide Bonded Proteins in E. coli
205
An alternative to engineering the host cell machinery to aid disulfide bond formation and oxidative protein folding is to engineer the heterologous polypeptide either by altering its folding kinetics or by eliminating disulfide bonds altogether. The later strategy requires that Cys are mutated to either Ala or Ser and additional mutations are introduced to increase the stability of the folded state of the polypeptide lacking disulfides. For example, Proba et al. using directed evolution generated a functional cysteine-free scFv that can be expressed in the bacterial cytoplasm (Proba et al. 1998). A different approach that also resulted in the isolation of functional scFvs that do not require disulfide bonds at all, relied on the screening of libraries of scFvs tethered to the inner membrane of cells deficient in DsbA, therefore providing a reducing periplasm (Seo et al. 2009). More recently, a scFv mutant was isolated from a library by allowing folding in the oxidizing cytoplasm of trxB gor E. coli followed by detection in the periplasm after translocation via the Tat system, the yield was doubled compared to the parent scFv (Ribnicky et al. 2007). Fisher et al. were able to evolve previously insoluble scFvs into a soluble version by using a method also based on the Tat system; folding, however, took place in the naturally reducing cytoplasm (Fisher and DeLisa 2009). Although these mutations are not necessarily affecting the cysteines in the targets, they do have an impact in the kinetics and thermodynamics of folding that can render disulfide bond formation less problematic or non essential. These techniques, however, are strongly dependent on the availability of robust high-throughput screening methods.
9.4 Conclusion The systematic study of the disulfide bond formation and isomerization pathways in the periplasm and, of the reducing pathways in the cytoplasm of E. coli has provided a firm foundation to advance the engineering of expression systems suitable for the production of multidisulfide bonded recombinant proteins in E. coli. Unfortunately, the current practical approaches for expression of heterologous proteins requiring oxidative folding are still case-dependent and require trial and error. While some proteins are easily expressed in the periplasm by coexpressing the oxidoreductases, others are better suited for the less oxidative environment of strains with an engineered cytoplasm and others can further benefit from a cell-free environment. Nonetheless, despite some initial time investment required to identify the best expression technology, E. coli still represents the easiest and most cost-effective system for the expression of valuable proteins requiring disulfide bond formation.
References Aggarwal S (2008) What’s fueling the biotech engine-2007. Nat Biotechnol 26(11):1227–1233 Angov E, Hillier CJ, Kincaid RL, Lyon JA (2008) Heterologous protein expression is enhanced by harmonizing the codon usage frequencies of the target gene with those of the expression host. PLoS ONE 3(5):e2189
206
S.A. Arredondo and G. Georgiou
Appenzeller-Herzog C, Ellgaard L (2008) The human PDI family: versatility packed into a single fold. Biochim Biophys Acta 1783(4):535–548 Arredondo S, Segatori L, Gilbert HF, Georgiou G (2008) De novo design and evolution of artificial disulfide isomerase enzymes analogous to the bacterial DsbC. J Biol Chem 283(46):31469–31476 Arredondo SA, Chen TF, Riggs AF, Gilbert HF, Georgiou G (2009) Role of dimerization in the catalytic properties of the Escherichia coli disulfide isomerase DsbC. J Biol Chem 284(36):23972–23979 Aslund F, Beckwith J (1999) The thioredoxin superfamily: redundancy, specificity, and gray-area genomics. J Bacteriol 181(5):1375–1379 Assadi-Porter FM, Patry S, Markley JL (2008) Efficient and rapid protein expression and purification of small high disulfide containing sweet protein brazzein in E.coli. Protein Expr Purif 58(2):263–268 Bader MW, Hiniker A, Regeimbal J, Goldstone D, Haebel PW, Riemer J, Metcalf P, Bardwell JC (2001) Turning a disulfide isomerase into an oxidase: DsbC mutants that Imitate DsbA. EMBO J 20(7):1555–1562 Bader MW, Xie T, Yu CA, Bardwell JC (2000) Disulfide bonds are generated by quinone reduction. J Biol Chem 275(34):26082–26088 Bardwell JC, McGovern K, Beckwith J (1991) Identification of a protein required for disulfide bond formation in vivo. Cell 67(3):581–589 Behrens S, Maier R, de Cock H, Schmid FX, Gross CA (2001) The SurA periplasmic PPIase lacking its parvulin domains functions in vivo and has chaperone activity. EMBO J 20(1–2):285–294 Berkmen M, Boyd D, Beckwith J (2005) The nonconsecutive disulfide bond of Escherichia coli phytase (AppA) renders it dependent on the protein-disulfide isomerase, DsbC. J Biol Chem 280(12):11387–11394 Bessette PH, Aslund F, Beckwith J, Georgiou G (1999a) Efficient folding of proteins with multiple disulfide bonds in the Escherichia coli cytoplasm. Proc Natl Acad Sci USA 96(24):13703–13708 Bessette PH, Cotto JJ, Gilbert HF, Georgiou G (1999b) In vivo and in vitro function of the Escherichia coli periplasmic cysteine oxidoreductase DsbG. J Biol Chem 274(12):7784–7792 Bessette PH, Qiu J, Bardwell JC, Swartz JR, Georgiou G (2001) Effect of sequences of the activesite dipeptides of DsbA and DsbC on in vivo folding of multidisulfide proteins in Escherichia coli. J Bacteriol 183(3):980–988 Bogomolovas J, Simon B, Sattler M, Stier G (2009) Screening of fusion partners for high yield expression and purification of bioactive viscotoxins. Protein Expr Purif 64(1):16–23 Bowden GA, Paredes AM, Georgiou G (1991) Structure and morphology of protein inclusion bodies in Escherichia coli. Biotechnology (N Y) 9(8):725–730 Bowden GA, Georgiou G (1988) The effect of sugars on b-lactamase aggregation in E.coli. Biotechnol Prog 4:97–101 Breustedt DA, Schonfeld DL, Skerra A (2006) Comparative ligand-binding analysis of ten human lipocalins. Biochim Biophys Acta 1764(2):161–173 Burgess RR (2009) Refolding solubilized inclusion body proteins. Methods Enzymol 463:259–282 Calhoun KA, Swartz JR (2005) An economical method for cell-free protein synthesis using glucose and nucleoside monophosphates. Biotechnol Prog 21(4):1146–1153 Chen J, Song JL, Zhang S, Wang Y, Cui DF, Wang CC (1999) Chaperone activity of DsbC. J Biol Chem 274(28):19601–19605 Cho SH, Beckwith J (2009) Two snapshots of electron transport across the membrane: insights into the structure and function of DsbD. J Biol Chem 284(17):11416–11424 Collet JF, Riemer J, Bader MW, Bardwell JC (2002) Reconstitution of a disulfide isomerization system. J Biol Chem 277(30):26886–26892 Darby NJ, Creighton TE (1995) Characterization of the active site cysteine residues of the thioredoxinlike domains of protein disulfide isomerase. Biochemistry 34(51):16770–16780 Darby NJ, Penka E, Vincentelli R (1998a) The multi-domain structure of protein disulfide isomerase is essential for high catalytic efficiency. J Mol Biol 276(1):239–247 Darby NJ, Raina S, Creighton TE (1998b) Contributions of substrate binding to the catalytic activity of DsbC. Biochemistry 37(3):783–791
9 Multidisulfide Bonded Proteins in E. coli
207
Datar RV, Cartwright T, Rosen CG (1993) Process economics of animal cell and bacterial fermentations: a case study analysis of tissue plasminogen activator. Biotechnology (N Y) 11(3):349–357 DeLisa MP, Tullman D, Georgiou G (2003) Folding quality control in the export of proteins by the bacterial twin-arginine translocation pathway. Proc Natl Acad Sci USA 100(10):6115–6120 Depuydt M, Leonard SE, Vertommen D, Denoncin K, Morsomme P, Wahni K, Messens J, Carroll KS, Collet JF (2009) A periplasmic reducing system protects single cysteine residues from oxidation. Science 326(5956):1109–1111 Derman AI, Prinz WA, Belin D, Beckwith J (1993) Mutations that allow disulfide bond formation in the cytoplasm of Escherichia coli. Science 262(5140):1744–1747 Devi VS, Sprecher CB, Hunziker P, Mittl PR, Bosshard HR, Jelesarov I (2006) Disulfide formation and stability of a cysteine-rich repeat protein from Helicobacter pylori. Biochemistry 45(6):1599–1607 Dracheva S, Palermo RE, Powers GD, Waugh DS (1995) Expression of soluble human interleukin-2 receptor alpha-chain in Escherichia coli. Protein Expr Purif 6(6):737–747 Driessen AJ, Nouwen N (2008) Protein translocation across the bacterial cytoplasmic membrane. Annu Rev Biochem 77:643–667 Faulkner MJ, Veeravalli K, Gon S, Georgiou G, Beckwith J (2008) Functional plasticity of a peroxidase allows evolution of diverse disulfide-reducing pathways. Proc Natl Acad Sci USA 105(18):6735–6740 Ferrari DM, Soling HD (1999) The protein disulphide-isomerase family: unravelling a string of folds. Biochem J 339(Pt 1):1–10 Fisher AC, DeLisa MP (2009) Efficient isolation of soluble intracellular single-chain antibodies using the twin-arginine translocation machinery. J Mol Biol 385(1):299–311 Frey S, Haslbeck M, Hainzl O, Buchner J (2008) Synthesis and characterization of a functional intact IgG in a prokaryotic cell-free expression system. Biol Chem 389(1):37–45 Gadermaier G, Jahn-Schmid B, Vogel L, Egger M, Himly M, Briza P, Ebner C, Vieths S, Bohle B, Ferreira F (2010) Targeting the cysteine-stabilized fold of Art v 1 for immunotherapy of Artemisia pollen allergy. Mol Immunol 47(6):1292–1298 Gagliardo B, Faye A, Jaouen M, Deschemin JC, Canonne-Hergaux F, Vaulont S, Sari MA (2008) Production of biologically active forms of recombinant hepcidin, the iron-regulatory hormone. FEBS J 275(15):3793–3803 Garcia-Ortega L, Lacadena J, Lacadena V, Masip M, De Antonio C, Martinez-Ruiz A, Martinez Del Pozo A (2000) The solubility of the ribotoxin alpha-sarcin, produced as a recombinant protein in Escherichia coli, is increased in the presence of thioredoxin. Lett Appl Microbiol 30(4):298–302 Glockshuber R (1999) Protein folding. Where do the electrons go? Nature 401(6748):30–31 Goeddel DV, Kleid DG, Bolivar F, Heyneker HL, Yansura DG, Crea R, Hirose T, Kraszewski A, Itakura K, Riggs AD (1979) Expression in Escherichia coli of chemically synthesized genes for human insulin. Proc Natl Acad Sci USA 76(1):106–110 Goerke AR, Swartz JR (2008) Development of cell-free protein synthesis platforms for disulfide bonded proteins. Biotechnol Bioeng 99(2):351–367 Golovanov AP, Hautbergue GM, Wilson SA, Lian LY (2004) A simple method for improving protein solubility and long-term stability. J Am Chem Soc 126(29):8933–8939 Grauschopf U, Fritz A, Glockshuber R (2003) Mechanism of the electron transfer catalyst DsbB from Escherichia coli. EMBO J 22(14):3503–3513 Grauschopf U, Winther JR, Korber P, Zander T, Dallinger P, Bardwell JC (1995) Why is DsbA such an oxidizing disulfide catalyst? Cell 83(6):947–955 Gruber CW, Cemazar M, Heras B, Martin JL, Craik DJ (2006) Protein disulfide isomerase: the structure of oxidative folding. Trends Biochem Sci 31(8):455–464 Guilhot C, Jander G, Martin NL, Beckwith J (1995) Evidence that the pathway of disulfide bond formation in Escherichia coli involves interactions between the cysteines of DsbB and DsbA. Proc Natl Acad Sci USA 92(21):9895–9899 Guisez Y, Fache I, Campfield LA, Smith FJ, Farid A, Plaetinck G, Van der Heyden J, Tavernier J, Fiers W, Burn P, Devos R (1998) Efficient secretion of biologically active recombinant OB
208
S.A. Arredondo and G. Georgiou
protein (leptin) in Escherichia coli, purification from the periplasm and characterization. Protein Expr Purif 12(2):249–258 Haebel PW, Goldstone D, Katzen F, Beckwith J, Metcalf P (2002) The disulfide bond isomerase DsbC is activated by an immunoglobulin-fold thiol oxidoreductase: crystal structure of the DsbC-DsbDalpha complex. EMBO J 21(18):4774–4784 Hammarstrom M, Hellgren N, van Den Berg S, Berglund H, Hard T (2002) Rapid screening for improved solubility of small human proteins produced as fusion proteins in Escherichia coli. Protein Sci 11(2):313–321 Hartl FU, Hayer-Hartl M (2002) Molecular chaperones in the cytosol: from nascent chain to folded protein. Science 295(5561):1852–1858 Hatahet F, Ruddock LW (2009) Protein disulfide isomerase: a critical evaluation of its function in disulfide bond formation. Antioxid Redox Signal 11(11):2807–2850 Hayhurst A, Harris WJ (1999) Escherichia coli skp chaperone coexpression improves solubility and phage display of single-chain antibody fragments. Protein Expr Purif 15(3):336–343 Hennecke G, Nolte J, Volkmer-Engert R, Schneider-Mergener J, Behrens S (2005) The periplasmic chaperone SurA exploits two features characteristic of integral outer membrane proteins for selective substrate recognition. J Biol Chem 280(25):23540–23548 Heo MA, Kim SH, Kim SY, Kim YJ, Chung J, Oh MK, Lee SG (2006) Functional expression of single-chain variable fragment antibody against c-Met in the cytoplasm of Escherichia coli. Protein Expr Purif 47(1):203–209 Heras B, Edeling MA, Schirra HJ, Raina S, Martin JL (2004) Crystal structures of the DsbG disulfide isomerase reveal an unstable disulfide. Proc Natl Acad Sci USA 101(24):8876–8881 Hiniker A, Bardwell JC (2004) In vivo substrate specificity of periplasmic disulfide oxidoreductases. J Biol Chem 279(13):12967–12973 Hiniker A, Collet JF, Bardwell JC (2005) Copper stress causes an in vivo requirement for the Escherichia coli disulfide isomerase DsbC. J Biol Chem 280(40):33785–33791 Hiniker A, Ren G, Heras B, Zheng Y, Laurinec S, Jobson RW, Stuckey JA, Martin JL, Bardwell JC (2007) Laboratory evolution of one disulfide isomerase to resemble another. Proc Natl Acad Sci USA 104(28):11670–11675 Hoffmann A, Bukau B, Kramer G (2010) Structure and function of the molecular chaperone trigger factor. Biochim Biophys Acta 1803(6):650–661 Hoshino K, Eda A, Kurokawa Y, Shimizu N (2002) Production of brain-derived neurotrophic factor in Escherichia coli by coexpression of Dsb proteins. Biosci Biotechnol Biochem 66(2):344–350 Hu X, O’Hara L, White S, Magner E, Kane M, Wall JG (2007) Optimisation of production of a domoic acid-binding scFv antibody fragment in Escherichia coli using molecular chaperones and functional immobilisation on a mesoporous silicate support. Protein Expr Purif 52(1): 194–201 Huang L, Ching CB, Jiang R, Leong SS (2008) Production of bioactive human beta-defensin 5 and 6 in Escherichia coli by soluble fusion expression. Protein Expr Purif 61(2):168–174 Humphreys DP, Weir N, Mountain A, Lund PA (1995) Human protein disulfide isomerase functionally complements a dsbA mutation and enhances the yield of pectate lyase C in Escherichia coli. J Biol Chem 270(47):28210–28215 Inaba K, Ito K (2008) Structure and mechanisms of the DsbB-DsbA disulfide bond generation machine. Biochim Biophys Acta 1783(4):520–529 Inaba K, Murakami S, Suzuki M, Nakagawa A, Yamashita E, Okada K, Ito K (2006) Crystal structure of the DsbB-DsbA complex reveals a mechanism of disulfide bond generation. Cell 127(4):789–801 Ito K, Inaba K (2008) The disulfide bond formation (Dsb) system. Curr Opin Struct Biol 18(4):450–458 Jarchow S, Luck C, Gorg A, Skerra A (2008) Identification of potential substrate proteins for the periplasmic Escherichia coli chaperone skp. Proteomics 8(23–24):4987–4994 Jeong KJ, Lee SY (2000) Secretory production of human leptin in Escherichia coli. Biotechnol Bioeng 67(4):398–407
9 Multidisulfide Bonded Proteins in E. coli
209
Jeong KJ, Lee SY (2001) Secretory production of human granulocyte colony-stimulating factor in Escherichia coli. Protein Expr Purif 23(2):311–318 Joly JC, Swartz JR (1997) In vitro and in vivo redox states of the Escherichia coli periplasmic oxidoreductases DsbA and DsbC. Biochemistry 36(33):10067–10072 Jonda S, Huber-Wunderlich M, Glockshuber R, Mossner E (1999) Complementation of DsbA deficiency with secreted thioredoxin variants reveals the crucial role of an efficient dithiol oxidant for catalyzed protein folding in the bacterial periplasm. EMBO J 18(12): 3271–3281 Jurado P, Ritz D, Beckwith J, de Lorenzo V, Fernandez LA (2002) Production of functional singlechain Fv antibodies in the cytoplasm of Escherichia coli. J Mol Biol 320(1):1–10 Kadokura H, Beckwith J (2009) Detecting folding intermediates of a protein as it passes through the bacterial translocation channel. Cell 138(6):1164–1173 Kamitani S, Akiyama Y, Ito K (1992) Identification and characterization of an Escherichia coli gene required for the formation of correctly folded alkaline phosphatase, a periplasmic enzyme. EMBO J 11(1):57–62 Kang SH, Kim DM, Kim HJ, Jun SY, Lee KY, Kim HJ (2005) Cell-free production of aggregationprone proteins in soluble and active forms. Biotechnol Prog 21(5):1412–1419 Katzen F, Beckwith J (2000) Transmembrane electron transfer by the membrane protein DsbD occurs via a disulfide bond cascade. Cell 103(5):769–779 Keefer LM, Piron MA, De Meyts P (1981) Human insulin prepared by recombinant DNA techniques and native human insulin interact identically with insulin receptors. Proc Natl Acad Sci USA 78(3):1391–1395 Kiedzierska A, Czepczynska H, Smietana K, Otlewski J (2008) Expression, purification and crystallization of cysteine-rich human protein muskelin in Escherichia coli. Protein Expr Purif 60(1):82–88 Kim DM, Swartz JR (2004) Efficient production of a bioactive, multiple disulfide-bonded protein using modified extracts of Escherichia coli. Biotechnol Bioeng 85(2):122–129 Kim JY, Fogarty EA, Lu FJ, Zhu H, Wheelock GD, Henderson LA, DeLisa MP (2005) Twinarginine translocation of active human tissue plasminogen activator in Escherichia coli. Appl Environ Microbiol 71(12):8451–8459 Klappa P, Ruddock LW, Darby NJ, Freedman RB (1998) The b¢ domain provides the principal peptide-binding site of protein disulfide isomerase but all domains contribute to binding of misfolded proteins. EMBO J 17(4):927–935 Knapp KG, Goerke AR, Swartz JR (2007) Cell-free synthesis of proteins that require disulfide bonds using glucose as an energy source. Biotechnol Bioeng 97(4):901–908 Kolaj O, Spada S, Robin S, Wall JG (2009) Use of folding modulators to improve heterologous protein production in Escherichia coli. Microb Cell Fact 8:9 Kudla G, Murray AW, Tollervey D, Plotkin JB (2009) Coding-sequence determinants of gene expression in Escherichia coli. Science 324(5924):255–258 Kumano-Kuramochi M, Xie Q, Sakakibara Y, Niimi S, Sekizawa K, Komba S, Machida S (2008) Expression and characterization of recombinant C-terminal biotinylated extracellular domain of human receptor for advanced glycation end products (hsRAGE) in Escherichia coli. J Biochem 143(2):229–236 Kurokawa Y, Yanagi H, Yura T (2000) Overexpression of protein disulfide isomerase DsbC stabilizes multiple-disulfide-bonded recombinant protein produced and transported to the periplasm in Escherichia coli. Appl Environ Microbiol 66(9):3960–3965 Kurokawa Y, Yanagi H, Yura T (2001) Overproduction of bacterial protein disulfide isomerase (DsbC) and its modulator (DsbD) markedly enhances periplasmic production of human nerve growth factor in Escherichia coli. J Biol Chem 276(17):14393–14399 Lauber T, Marx UC, Schulz A, Kreutzmann P, Rosch P, Hoffmann S (2001) Accurate disulfide formation in Escherichia coli: overexpression and characterization of the first domain (HF6478) of the multiple Kazal-type inhibitor LEKTI. Protein Expr Purif 22(1):108–112 Leader B, Baca QJ, Golan DE (2008) Protein therapeutics: a summary and pharmacological classification. Nat Rev Drug Discov 7(1):21–39
210
S.A. Arredondo and G. Georgiou
Lee DH, Kim MD, Lee WH, Kweon DH, Seo JH (2004) Consortium of fold-catalyzing proteins increases soluble expression of cyclohexanone monooxygenase in recombinant Escherichia coli. Appl Microbiol Biotechnol 63(5):549–552 Lee PA, Tullman-Ercek D, Georgiou G (2006) The bacterial twin-arginine translocation pathway. Annu Rev Microbiol 60:373–395 Lefebvre J, Boileau G, Manjunath P (2009) Recombinant expression and affinity purification of a novel epididymal human sperm-binding protein, BSPH1. Mol Hum Reprod 15(2): 105–114 Levy R, Weiss R, Chen G, Iverson BL, Georgiou G (2001) Production of correctly folded fab antibody fragment in the cytoplasm of escherichia coli trxB gor mutants via the coexpression of molecular chaperones. Protein Expr Purif 23(2):338–347 Liddy N, Molloy PE, Bennett AD, Boulter JM, Jakobsen BK, Li Y (2010) Production of a soluble disulfide bond-linked TCR in the cytoplasm of Escherichia coli trxB gor mutants. Mol Biotechnol 45(2):140–149 Liu X, Wang CC (2001) Disulfide-dependent folding and export of Escherichia coli DsbC. J Biol Chem 276(2):1146–1151 Lobel LI, Pollak S, Klein J, Lustbader JW (2001) High-level bacterial expression of a natively folded, soluble extracellular domain fusion protein of the human luteinizing hormone/chorionic gonadotropin receptor in the cytoplasm of Escherichia coli. Endocr 14(2):205–212 Locker JK, Griffiths G (1999) An unconventional role for cytoplasmic disulfide bonds in vaccinia virus proteins. J Cell Biol 144(2):267–279 Lu Q, Burns MC, McDevitt PJ, Graham TL, Sukman AJ, Fornwald JA, Tang X, Gallagher KT, Hunsberger GE, Foley JJ, Schmidt DB, Kerrigan JJ, Lewis TS, Ames RS, Johanson KO (2009) Optimized procedures for producing biologically active chemokines. Protein Expr Purif 65(2):251–260 Lundstrom J, Krause G, Holmgren A (1992) A Pro to His mutation in active site of thioredoxin increases its disulfide-isomerase activity 10-fold. New refolding systems for reduced or randomly oxidized ribonuclease. J Biol Chem 267(13):9047–9052 Makrides SC (1996) Strategies for achieving high-level expression of genes in Escherichia coli. Microbiol Rev 60(3):512–538 Malakhov MP, Mattern MR, Malakhova OA, Drinker M, Weeks SD, Butt TR (2004) SUMO fusions and SUMO-specific protease for efficient expression and purification of proteins. J Struct Funct Genomics 5(1–2):75–86 Malik A, Jenzsch M, Lubbert A, Rudolph R, Sohling B (2007) Periplasmic production of native human proinsulin as a fusion to E.coli ecotin. Protein Expr Purif 55(1):100–111 Malik A, Rudolph R, Sohling B (2006) A novel fusion protein system for the production of native human pepsinogen in the bacterial periplasm. Protein Expr Purif 47(2):662–671 Mao X, Cao Z, Yin S, Ma Y, Wu Y, Li W (2007) Cloning and characterization of Bmk86, a novel k+-channel blocker from scorpion venom. Biochem Biophys Res Commun 360(4):728–734 Martin JL (1995) Thioredoxin – a fold for all reasons. Structure 3(3):245–250 Martin JL, Bardwell JC, Kuriyan J (1993) Crystal structure of the dsba protein required for disulphide bond formation in vivo. Nature 365(6445):464–468 Maskos K, Huber-Wunderlich M, Glockshuber R (2003) DsbA and DsbC-catalyzed oxidative folding of proteins with complex disulfide bridge patterns in vitro and in vivo. J Mol Biol 325(3):495–513 Mavrangelos C, Thiel M, Adamson PJ, Millard DJ, Nobbs S, Zola H, Nicholson IC (2001) Increased yield and activity of soluble single-chain antibody fragments by combining highlevel expression and the skp periplasmic chaperonin. Protein Expr Purif 23(2):289–295 Maynard J, Adams EJ, Krogsgaard M, Petersson K, Liu CW, Garcia KC (2005) High-level bacterial secretion of single-chain alphabeta T-cell receptors. J Immunol Methods 306(1–2):51–67 McCarthy AA, Haebel PW, Torronen A, Rybin V, Baker EN, Metcalf P (2000) Crystal structure of the protein disulfide bond isomerase, DsbC, from Escherichia coli. Nat Struct Biol 7(3):196–199
9 Multidisulfide Bonded Proteins in E. coli
211
Merk H, Stiege W, Tsumoto K, Kumagai I, Erdmann VA (1999) Cell-free expression of two singlechain monoclonal antibodies against lysozyme: effect of domain arrangement on the expression. J Biochem 125(2):328–333 Messens J, Collet JF, Van Belle K, Brosens E, Loris R, Wyns L (2007) The oxidase DsbA folds a protein with a nonconsecutive disulfide. J Biol Chem 282(43):31302–31307 Missiakas D, Betton JM, Raina S (1996) New components of protein folding in extracytoplasmic compartments of Escherichia coli SurA, FkpA and Skp/OmpH. Mol Microbiol 21(4):871–884 Missiakas D, Georgopoulos C, Raina S (1994) The Escherichia coli dsbC (xprA) gene encodes a periplasmic protein involved in disulfide bond formation. EMBO J 13(8):2013–2020 Mureev S, Kovtun O, Nguyen UT, Alexandrov K (2009) Species-independent translational leaders facilitate cell-free expression. Nat Biotechnol 27(8):747–752 Nakano H, Yamane T (1998) Cell-free protein synthesis systems. Biotechnol Adv 16(2): 367–384 Natale P, Bruser T, Driessen AJ (2008) Sec- and tat-mediated protein secretion across the bacterial cytoplasmic membrane – distinct translocases and mechanisms. Biochim Biophys Acta 1778(9):1735–1756 O’Dwyer R, Razzaque R, Hu X, Hollingshead SK, Wall JG (2009) Engineering of cysteine residues leads to improved production of a human dipeptidase enzyme in E.coli. Appl Biochem Biotechnol 159(1):178–190 Oh IS, Lee JC, Lee MS, Chung JH, Kim DM (2010) Cell-free production of functional antibody fragments. Bioprocess Biosyst Eng 33(1):127–132 Ostermeier M, De Sutter K, Georgiou G (1996) Eukaryotic protein disulfide isomerase complements Escherichia coli dsbA mutants and increases the yield of a heterologous secreted protein with disulfide bonds. J Biol Chem 271(18):10616–10622 Outchkourov NS, Roeffen W, Kaan A, Jansen J, Luty A, Schuiffel D, van Gemert GJ, van de Vegte-Bolmer M, Sauerwein RW, Stunnenberg HG (2008) Correctly folded pfs48/45 protein of Plasmodium falciparum elicits malaria transmission-blocking immunity in mice. Proc Natl Acad Sci USA 105(11):4301–4305 Padiolleau-Lefevre S, Debat H, Phichith D, Thomas D, Friboulet A, Avalle B (2006) Expression of a functional scFv fragment of an anti-idiotypic antibody with a beta-lactam hydrolytic activity. Immunol Lett 103(1):39–44 Pan JL, Sliskovic I, Bardwell JC (2008) Mutants in DsbB that appear to redirect oxidation through the disulfide isomerization pathway. J Mol Biol 377(5):1433–1442 Pavlou AK, Reichert JM (2004) Recombinant protein therapeutics – success rates, market trends and values to 2010. Nat Biotechnol 22(12):1513–1519 Peisley AA, Gooley PR (2007) High-level expression of a soluble and functional fibronectin type II domain from mmp-2 in the Escherichia coli cytoplasm for solution NMR studies. Protein Expr Purif 53(1):124–131 Pirneskoski A, Klappa P, Lobell M, Williamson RA, Byrne L, Alanen HI, Salo KE, Kivirikko KI, Freedman RB, Ruddock LW (2004) Molecular characterization of the principal substrate binding site of the ubiquitous folding catalyst protein disulfide isomerase. J Biol Chem 279(11):10374–10381 Ponniah K, Loo TS, Edwards PJ, Pascal SM, Jameson GB, Norris GE (2010) The production of soluble and correctly folded recombinant bovine beta-lactoglobulin variants a and b in Escherichia coli for NMR studies. Protein Expr Purif 70(2):283–289 Prinz WA, Aslund F, Holmgren A, Beckwith J (1997) The role of the thioredoxin and glutaredoxin pathways in reducing protein disulfide bonds in the Escherichia coli cytoplasm. J Biol Chem 272(25):15661–15667 Proba K, Worn A, Honegger A, Pluckthun A (1998) Antibody scFv fragments without disulfide bonds made by molecular evolution. J Mol Biol 275(2):245–253 Puertas JM, Betton JM (2009) Engineering an efficient secretion of leech carboxypeptidase inhibitor in Escherichia coli. Microb Cell Fact 8:57
212
S.A. Arredondo and G. Georgiou
Qing G, Ma LC, Khorchid A, Swapna GV, Mal TK, Takayama MM, Xia B, Phadtare S, Ke H, Acton T, Montelione GT, Ikura M, Inouye M (2004) Cold-shock induced high-yield protein production in Escherichia coli. Nat Biotechnol 22(7):877–882 Qiu J, Swartz JR, Georgiou G (1998) Expression of active human tissue-type plasminogen activator in Escherichia coli. Appl Environ Microbiol 64(12):4891–4896 Ramm K, Pluckthun A (2000) The periplasmic Escherichia coli peptidylprolyl cis, trans-isomerase FkpA. II. Isomerase-independent chaperone activity in vitro. J Biol Chem 275(22):17106–17113 Rathnayaka T, Tawa M, Sohya S, Yohda M, Kuroda Y (2010) Biophysical characterization of highly active recombinant Gaussia luciferase expressed in Escherichia coli. Biochim Biophys Acta 1804(9):1902–1907 Reilly DE, Yansura DG (2010) Production of monoclonal antibodies in E.coli. In: Shire SJ, Gombot W, Bechtold-Peters K, Andya J (eds) Current trends in monoclonal antibody development and manufacturing, vol XI. Biotechnology: Pharmaceutical aspects. Springer, New York, pp 295–308 Ren G, Stephan D, Xu Z, Zheng Y, Tang D, Harrison RS, Kurz M, Jarrott R, Shouldice SR, Hiniker A, Martin JL, Heras B, Bardwell JC (2009) Properties of the thioredoxin-fold superfamily are modulated by a single amino acid residue. J Biol Chem 284(15):10150–10159 Ribnicky B, Van Blarcom T, Georgiou G (2007) A scFv antibody mutant isolated in a genetic screen for improved export via the twin arginine transporter pathway exhibits faster folding. J Mol Biol 369(3):631–639 Rietsch A, Bessette P, Georgiou G, Beckwith J (1997) Reduction of the periplasmic disulfide bond isomerase, DsbC, occurs by passage of electrons from cytoplasmic thioredoxin. J Bacteriol 179(21):6602–6608 Ritz D, Beckwith J (2001) Roles of thiol-redox pathways in bacteria. Annu Rev Microbiol 55:21–48 Ritz D, Lim J, Reynolds CM, Poole LB, Beckwith J (2001) Conversion of a peroxiredoxin into a disulfide reductase by a triplet repeat expansion. Science 294(5540):158–160 Rozhkova A, Stirnimann CU, Frei P, Grauschopf U, Brunisholz R, Grutter MG, Capitani G, Glockshuber R (2004) Structural basis and kinetics of inter- and intramolecular disulfide exchange in the redox catalyst DsbD. EMBO J 23(8):1709–1719 Ryabova LA, Desplancq D, Spirin AS, Pluckthun A (1997) Functional antibody production using cell-free translation: effects of protein disulfide isomerase and chaperones. Nat Biotechnol 15(1):79–84 Saul FA, Arie JP, Vulliez-le Normand B, Kahn R, Betton JM, Bentley GA (2004) Structural and functional studies of FkpA from escherichia coli, a cis/trans peptidyl-prolyl isomerase with chaperone activity. J Mol Biol 335(2):595–608 Schafer U, Beck K, Muller M (1999) Skp, a molecular chaperone of gram-negative bacteria, is required for the formation of soluble periplasmic intermediates of outer membrane proteins. J Biol Chem 274(35):24567–24574 Schaffner J, Winter J, Rudolph R, Schwarz E (2001) Cosecretion of chaperones and low-molecularsize medium additives increases the yield of recombinant disulfide-bridged proteins. Appl Environ Microbiol 67(9):3994–4000 Schein CH (1989) Production of soluble recombinant proteins in bacteria. Biotechnology (N Y) 7:1141–1149 Schlapschy M, Grimm S, Skerra A (2006) A system for concomitant overexpression of four periplasmic folding catalysts to improve secretory protein production in Escherichia coli. Protein Eng Des Sel 19(8):385–390 Schmidt FR (2004) Recombinant expression systems in the pharmaceutical industry. Appl Microbiol Biotechnol 65(4):363–372 Schneider EL, Thomas JG, Bassuk JA, Sage EH, Baneyx F (1997) Manipulating the aggregation and oxidation of human SPARC in the cytoplasm of Escherichia coli. Nat Biotechnol 15(6):581–585 Schwaller M, Wilkinson B, Gilbert HF (2003) Reduction-reoxidation cycles contribute to catalysis of disulfide isomerization by protein-disulfide isomerase. J Biol Chem 278(9):7154–7159
9 Multidisulfide Bonded Proteins in E. coli
213
Segatori L, Murphy L, Arredondo S, Kadokura H, Gilbert H, Beckwith J, Georgiou G (2006) Conserved role of the linker alpha-helix of the bacterial disulfide isomerase DsbC in the avoidance of misoxidation by DsbB. J Biol Chem 281(8):4911–4919 Segatori L, Paukstelis PJ, Gilbert HF, Georgiou G (2004) Engineered DsbC chimeras catalyze both protein oxidation and disulfide-bond isomerization in Escherichia coli: reconciling two competing pathways. Proc Natl Acad Sci USA 101(27):10018–10023 Seo MJ, Jeong KJ, Leysath CE, Ellington AD, Iverson BL, Georgiou G (2009) Engineering antibody fragments to fold in the absence of disulfide bonds. Protein Sci 18(2):259–267 Shao F, Bader MW, Jakob U, Bardwell JC (2000) DsbG, a protein disulfide isomerase with chaperone activity. J Biol Chem 275(18):13349–13352 Shevchik VE, Condemine G, Robert-Baudouy J (1994) Characterization of DsbC, a periplasmic protein of erwinia chrysanthemi and escherichia coli with disulfide isomerase activity. EMBO J 13(8):2007–2012 Shouldice SR, Cho SH, Boyd D, Heras B, Eser M, Beckwith J, Riggs P, Martin JL, Berkmen M (2010) In vivo oxidative protein folding can be facilitated by oxidation-reduction cycling. Mol Microbiol 75(1):13–28 Sitaraman K, Chatterjee DK (2009) High-throughput protein expression using cell-free system. Methods Mol Biol 498:229–244 Soanes KH, Ewart KV, Mattatall NR (2008) Recombinant production and characterization of the carbohydrate recognition domain from Atlantic salmon C-type lectin receptor C (SCLRC). Protein Expr Purif 59(1):38–46 Song JA, Han KY, Park JS, Seo HS, Ahn KY, Lee J (2009) Human G-CSF synthesis using stressresponsive bacterial proteins. FEMS Microbiol Lett 296(1):60–66 Stafford SJ, Lund PA (2000) Mutagenic studies on human protein disulfide isomerase by complementation of escherichia coli dsbA and dsbC mutants. FEBS Lett 466(2–3):317–322 Stewart EJ, Aslund F, Beckwith J (1998) Disulfide bond formation in the escherichia coli cytoplasm: an in vivo role reversal for the thioredoxins. EMBO J 17(19):5543–5550 Sun M (1980) Insulin wars: new advances may throw market into turbulence. Science 210(4475):1225–1228 Sun XX, Wang CC (2000) The N-terminal sequence (residues 1–65) is essential for dimerization, activities, and peptide binding of Escherichia coli DsbC. J Biol Chem 275(30): 22743–22749 Swartz JR (2001) Advances in escherichia coli production of therapeutic proteins. Curr Opin Biotechnol 12(2):195–201 Swartz JR, Jewett MC, Woodrow KA (2004) Cell-free protein synthesis with prokaryotic combined transcription-translation. Methods Mol Biol 267:169–182 Tan S, Wu W, Liu J, Kong Y, Pu Y, Yuan R (2002) Efficient expression and secretion of recombinant hirudin III in E.coli using the l-asparaginase II signal sequence. Protein Expr Purif 25(3):430–436 Tian G, Xiang S, Noiva R, Lennarz WJ, Schindelin H (2006) The crystal structure of yeast protein disulfide isomerase suggests cooperativity between its active sites. Cell 124(1):61–73 Tran T, Buscher P, Vandenbussche G, Wyns L, Messens J, De Greve H (2008) Heterologous expression, purification and characterisation of the extracellular domain of trypanosome invariant surface glycoprotein ISG75. J Biotechnol 135(3):247–254 Vertommen D, Depuydt M, Pan J, Leverrier P, Knoops L, Szikora JP, Messens J, Bardwell JC, Collet JF (2008) The disulphide isomerase DsbC cooperates with the oxidase DsbA in a DsbDindependent manner. Mol Microbiol 67(2):336–349 Vlamis-Gardikas A (2008) The multiple functions of the thiol-based electron flow pathways of Escherichia coli: eternal concepts revisited. Biochim Biophys Acta 1780(11):1170–1200 Voronova A, Kazantseva J, Tuuling M, Sokolova N, Sillard R, Palumaa P (2007) Cox17, a copper chaperone for cytochrome c oxidase: expression, purification, and formation of mixed disulphide adducts with thiol reagents. Protein Expr Purif 53(1):138–144 Waugh DS (2005) Making the most of affinity tags. Trends Biotechnol 23(6):316–320 Wilkinson B, Gilbert HF (2004) Protein disulfide isomerase. Biochim Biophys Acta 1699(1–2):35–44
214
S.A. Arredondo and G. Georgiou
Winter J, Neubauer P, Glockshuber R, Rudolph R (2001) Increased production of human proinsulin in the periplasmic space of Escherichia coli by fusion to DsbA. J Biotechnol 84(2):175–185 Wulfing C, Pluckthun A (1994) Correctly folded T-cell receptor fragments in the periplasm of Escherichia coli. Influence of folding catalysts. J Mol Biol 242(5):655–669 Xiao R, Solovyov A, Gilbert HF, Holmgren A, Lundstrom-Ljung J (2001) Combinations of protein-disulfide isomerase domains show that there is little correlation between isomerase activity and wild-type growth. J Biol Chem 276(30):27975–27980 Xu Y, Lewis D, Chou CP (2008a) Effect of folding factors in rescuing unstable heterologous lipase b to enhance its overexpression in the periplasm of Escherichia coli. Appl Microbiol Biotechnol 79(6):1035–1044 Xu Y, Yasin A, Tang R, Scharer JM, Moo-Young M, Chou CP (2008b) Heterologous expression of lipase in Escherichia coli is limited by folding and disulfide bond formation. Appl Microbiol Biotechnol 81(1):79–87 Yamamoto Y, Ritz D, Planson AG, Jonsson TJ, Faulkner MJ, Boyd D, Beckwith J, Poole LB (2008) Mutant AhpC peroxiredoxins suppress thiol-disulfide redox deficiencies and acquire deglutathionylating activity. Mol Cell 29(1):36–45 Yan WK, Goette M, Hofmann G, Zaror I, Sim J (2010) High-level soluble expression, purification and characterization of active human midkine from Escherichia coli. Protein Expr Purif 70(2):270–276 Yang J, Kanter G, Voloshin A, Michel-Reydellet N, Velkeen H, Levy R, Swartz JR (2005) Rapid expression of vaccine proteins for B-cell lymphoma in a cell-free system. Biotechnol Bioeng 89(5):503–511 Yeh SM, Koon N, Squire C, Metcalf P (2007) Structures of the dimerization domains of the Escherichia coli disulfide-bond isomerase enzymes DsbC and DsbG. Acta Crystallogr D Biol Crystallogr 63(Pt 4):465–471 Yin G, Swartz JR (2004) Enhancing multiple disulfide bonded protein folding in a cell-free system. Biotechnol Bioeng 86(2):188–195 Young JC, Agashe VR, Siegers K, Hartl FU (2004) Pathways of chaperone-mediated protein folding in the cytosol. Nat Rev Mol Cell Biol 5(10):781–791 Yuan S, Duan H, Liu C, Liu X, Liu T, Tao H, Zhang Z (2004) The role of thioredoxin and disulfide isomerase in the expression of the snake venom thrombin-like enzyme calobin in Escherichia coli bl21 (de3). Protein Expr Purif 38(1):51–60 Zander H, Hettich E, Greiff K, Chatwell L, Skerra A (2007) Biochemical characterization of the recombinant human Nogo-A ectodomain. FEBS J 274(10):2603–2613 Zapun A, Bardwell JC, Creighton TE (1993) The reactive and destabilizing disulfide bond of DsbA, a protein required for protein disulfide bond formation in vivo. Biochemistry 32(19): 5083–5092 Zapun A, Missiakas D, Raina S, Creighton TE (1995) Structural and functional characterization of DsbC, a protein involved in disulfide bond formation in Escherichia coli. Biochemistry 34(15): 5075–5089 Zhan X, Schwaller M, Gilbert HF, Georgiou G (1999) Facilitating the formation of disulfide bonds in the Escherichia coli periplasm via coexpression of yeast protein disulfide isomerase. Biotechnol Prog 15(6):1033–1038 Zhang ST, Shi J, Zhao J, Qi YF, Guo AG (2006) Expression of soluble and functional snake venom fibrinolytic enzyme fibrolase via the co-expression of DsbC in Escherichia coli. Protein Pept Lett 13(6):559–563 Zhang Y, Olsen DR, Nguyen KB, Olson PS, Rhodes ET, Mascarenhas D (1998) Expression of eukaryotic proteins in soluble form in Escherichia coli. Protein Expr Purif 12(2):159–165 Zhang Z, Li ZH, Wang F, Fang M, Yin CC, Zhou ZY, Lin Q, Huang HL (2002) Overexpression of DsbC and DsbG markedly improves soluble and functional expression of single-chain Fv antibodies in Escherichia coli. Protein Expr Purif 26(2):218–228
9 Multidisulfide Bonded Proteins in E. coli
215
Zhang Z, Song LP, Fang M, Wang F, He D, Zhao R, Liu J, Zhou ZY, Yin CC, Lin Q, Huang HL (2003) Production of soluble and functional engineered antibodies in Escherichia coli improved by FkpA. Biotechniques 35(5):1032–1038, 1041–1032 Zhao Z, Peng Y, Hao SF, Zeng ZH, Wang CC (2003) Dimerization by domain hybridization bestows chaperone and isomerase activities. J Biol Chem 278(44):43292–43298
Chapter 10
NMR-Spectroscopic Investigation of Disulfide Dynamics in Unfolded States of Proteins Robert Silvers, Kai Schlepckow, Julia Wirmer-Bartoschek, and Harald Schwalbe
Abstract The amino acid cysteine plays an important role in protein biochemistry. Besides its catalytic role in the active sites of enzymes it allows for the formation of disulfide bonds which are crucial for the function and stability of proteins. The presence of disulfide bonds also has profound influence on structural and dynamical properties of unfolded states of proteins. The formation of intramolecular loop structures limits the sampling of conformational space and consequently imposes specific structural as well as dynamical constraints onto an unstructured polypeptide chain. In this chapter, it is shown that NMR-spectroscopic studies on model proteins including lysozyme, bovine pancreatic trypsin inhibitor and the prion protein contribute valuably to the elucidation of protein folding and misfolding pathways and how these are impacted by disulfide bonds. Furthermore, NMR techniques are discussed that allow the characterization of unfolded states of proteins on an atomic level which is otherwise difficult to access by other techniques. Keywords Nuclear magnetic resonance spectroscopy • Disulfide bonds • Chain dynamics • Oxidative refolding • Misfolding • Unfolded states of proteins
Abbreviations ALS BMRB BPTI CD
Amyotrophic lateral sclerosis Biological magnetic resonance bank Bovine pancreatic trypsin inhibitor Circular dichroism
H. Schwalbe (*) Institute of Organic Chemistry and Chemical Biology, Center for Biomolecular Magnetic Resonance (BMRZ), Goethe University Frankfurt, Max-von-Laue-Straße 7, 60438, Frankfurt/Main, Germany e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_10, © Springer Science+Business Media, LLC 2011
217
R. Silvers et al.
218
cGMP CPMG CSA CSI DNA DOSY ER FAD GPI GSH GSSG hetNOE HEWL HMBC hPrP HSQC IUP LACS MHC mPrP NAG NAM NMR NOE NOESY PDI PG-SLED PPS PrP RNA ROESY ROS SSP ThT TOCSY
Cyclic guanosine monophosphate Carr–Purcell–Meiboom–Gill Chemical shift anisotropy Chemical shift index Desoxyribonucleic acid Diffusion-ordered spectroscopy Endoplasmic reticulum Flavin adenine dinucleotide Glycosylphosphatidylinositol Reduced glutathione Oxidized glutathione Heteronuclear nuclear Overhauser enhancement Hen egg white lysozyme Heteronuclear multiple bond coherence Human prion protein Heteronuclear single quantum coherence Intrinsically unstructured protein Linear analysis of chemical shifts Major histocompatibility complex Murine prion protein N-Acetylglucosamine N-Acetylmuramic acid Nuclear magnetic resonance Nuclear Overhauser enhancement Nuclear Overhauser enhancement spectroscopy Protein disulfide isomerase Pulse gradient stimulated echo longitudinal encode-decode Pro-peptide of subtilisin Prion protein Ribonuclear acid Rotational Overhauser enhancement spectroscopy Reactive oxygen species Secondary structure propensity Thioflavin T Total correlation spectroscopy
10.1 Introduction Cysteine is one of the 20 common proteinogenic a-amino acids and occupies an exceptional position among these amino acids, since its thiol group allows a wide variety of functionalities. The thiol group of cysteine residues for instance has a pKa of about 8.4 whereas the alcohol group of tyrosine has a pKa of about 10.5. Hence, thiolate ions can form more easily at physiological pH than alcoholates, making thiolates potent nucleophiles even at physiological conditions. Furthermore, thiolates
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
219
being soft bases have a high metal affinity (especially to mercury). Another characteristic property of thiols is their ability to perform redox chemistry. Hence, cysteine residues can play a major role in a wide range of cellular tasks: 1. Cysteine residues chelate iron in iron–sulfur clusters. Cysteine residues are known to bind iron ions of different oxidation states as 2Fe–2S, 3Fe–4S, 4Fe–4S, and 8Fe–8S clusters (Johnson et al. 2005). These clusters are often found in metalloproteins such as ferredoxins that mediate electron transport by redox chemical mechanisms. Furthermore, iron–sulfur clusters participate in substrate binding and activation, iron/sulfur storage, regulation of gene expression, and enzyme activity. NMR spectroscopic studies of iron–sulfur proteins have been performed to identify interactions between different nuclei and between nuclei and unpaired electrons in the iron–sulfur clusters of rubredoxins, ferredoxins, and high-potential iron proteins (Cheng and Markley 1995; Ciurli et al. 1996). One of the main questions in these studies is how electrons are stored and processed in iron–sulfur clusters as well as how protein residues in the vicinity of these clusters regulate intra- and intermolecular electron transfer. 2. Cysteine residues chelate zinc ions in zinc fingers. Zinc fingers are a class of small protein domains that contain zinc and enable the domain to function as an interaction module capable of binding DNA, RNA, other proteins, and small molecules (Krishna et al. 2003). The best characterized structural motif is the Cys2His2 zinc finger, where a zinc atom is complexed by two cysteine and two histidine residues arranged in a domain with a bba fold with a consensus pattern X2–C–X2,4–C–X12–H–X3,4,5–H, where X stands for any amino acid and the spacing between the two cysteine and histidine residues is variable (Pabo et al. 2001). With the aid of NMR spectroscopy, the structure and dynamics (Isernia et al. 2003; Kochoyan et al. 1991; Qian and Weiss 1992; Sauve et al. 2008) as well as chemical shift perturbation mapping (Schmiedeskamp et al. 1997) of zinc fingers when interacting with target DNA provided key insight into the binding modes of DNA to zinc fingers. 3. Cysteine residues as catalytic residues in cysteine proteases. Cysteine proteases constitute a class of proteases that contain cysteine as a catalytically active residue. These proteases are involved in a wide range of cellular processes such as apoptosis, MHC class II immune response, prohormone processing, and extracellular matrix remodeling (Chapman et al. 1997). Several NMR solution structures (Dubin et al. 2003; Salmon et al. 2006) of cysteine protease inhibitors have been published that play a vital role in the defense against pathogens. 4. Cysteine residues form disulfide bonds to facilitate folding and stabilize protein structure. Cysteine residues can be oxidized to form intra- and intermolecular disulfide bonds. The stabilizing effect of intramolecular disulfide bonds is highlighted by the fact that most disulfide-containing proteins need their disulfide bonds to function and will unfold at least partially upon disulfide bond reduction even in the absence of denaturants. Intermolecular disulfide bonds are often found in fibers such as keratins, where the amount of disulfide bonds (i.e., the degree of cross-linking) is correlated with the stiffness of the fiber. Aberrant formation of intermolecular disulfide bonds can also lead to misfolding and
220
R. Silvers et al.
associated diseases such as amyotrophic lateral sclerosis (ALS), a neurodegenerative disease where misfolded proteins accumulate in the ER lumen. A dysfunction of the Cu, Zn superoxide dismutase (SOD1) has been identified in familial ALS and investigated by NMR spectroscopy (Banci et al. 2007). 5. Cysteine residues are important for metal homeostasis. The yeast copper transport domain Ccc2a is involved in the intracellular copper transport across vesicular membranes. In humans, dysregulation of the intracellular copper transport is associated with diseases such as Wilson’s and Menkens’ disease. Characteristic for Ccc2a is, as is for other proteins of this class, a GMXCXXC motif that is involved in the complexation of copper by two cysteine residues as shown by the NMR solution structure of Ccc2a (Banci et al. 2001). 6. Cysteine residues form disulfide bonds in ROS response. Reactive oxygen species (ROS) such as superoxide (O2−) and hydroxyl radicals are formed during cellular respiration. As one cellular response, superoxide dismutase enzymes are recruited to catalyze the disproportionation of superoxide to hydrogen peroxide and molecular oxygen. One enzyme of this class, the Cu, Zn superoxide dismutase (SOD1), is located in the cytosol, although it was found to contain a disulfide bond. It was shown, that the copper chaperone CCS (Cu-CCS) can posttranslationally regulate SOD1 activity by oxidizing cysteine residues to a kinetically stable disulfide bond even in the presence of excess reductants (Brown et al. 2004; Furukawa et al. 2004). The structure of SOD1 in the absence of zinc and copper with reduced disulfide bond was solved by NMR in 2006 by Banci et al. (2006). 7. Cysteine residues form disulfide bonds as redox sensor. Although disulfide bonds in the cell’s cytosol were thought to be absent due to the reducing conditions, examples for disulfide bond formation have been found. For example, the Ia isoform of guanosine 3¢,5¢-monophosphate (cGMP)-dependent protein kinase (PGKIa) forms an intermolecular disulfide bond in vitro and in cells upon exogenous oxidation stress (Burgoyne et al. 2007). The disulfide bond formation triggers kinase activation and increases the substrate affinity of the kinase. 8. Cysteine residues and redox homeostasis. Being a crucial integral part of glutathione, cysteine acts as the main key molecule in redox homeostasis. The cellular equilibrium between reduced (GSH) and oxidized glutathione (GSSG) dictates the redox state of the cellular compartments and allows maintaining reducing and oxidizing conditions. Oxidative protein folding, i.e., coupling of protein folding with the formation of disulfide bonds has been studied extensively and can be grouped into eukaryotic and prokaryotic systems.
10.1.1 Oxidative Folding in Eukaryotes and Prokaryotes Oxidative folding processes in eukaryotes are almost exclusively restricted to cell compartments such as the endoplasmic reticulum (ER), mitochondria, and chloroplasts. Only these compartments provide the necessary oxidative potential
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
221
Fig. 10.1 Disulfide bond formation (a, b) and isomerization (c) in the lumen of the endoplasmic reticulum (ER) of Saccharomyces cerevisiae. Protein disulfide isomerase (PDI) acts as a central carrier of oxidizing equivalents, it oxidizes thiol groups of unfolded or misfolded substrates to disulfide bonds. Two pathways exist for the reoxidation of PDI: In the first, PDI is reoxidized by the membrane-associated Ero1p (a). In a second pathway, the membrane-associated protein Erv2p (b) transfers oxidizing equivalents to PDI which gets reoxidized in the process. (c) PDI also can reshuffle disulfide bonds by thiol–disulfide exchange while remaining reduced. Figure modified after Mamathambika and Bardwell (2008)
and auxiliary machinery for efficient oxidative folding. The oxidative folding machinery of Saccharomyces cerevisiae is the best studied among eukaryotes. This machinery located at the endoplasmic reticulum (ER) consists mainly of two proteins: Ero1 (ER oxidoreductin 1) and PDI (protein disulfide isomerase) (Frand and Kaiser 1998; Goldberger et al. 1963; Pollard et al. 1998; Sevier et al. 2001; Tu et al. 2000). Ero1 is a luminal glycoprotein that is tightly associated with the ER membrane (Fig. 10.1) (Kersteen and Raines 2003; Sevier and Kaiser 2002). It is essential for the introduction of oxidation equivalents into the ER lumen. PDI is a luminal protein that oxidizes substrate proteins by direct thiol– disulfide exchange reactions. It is also involved in reshuffling disulfide bonds of misfolded proteins in a similar fashion. PDI is reoxidized by Ero1 by direct thiol– disulfide exchange. Mutational analysis of Ero1 revealed four cysteine residues essential for function. The four residues form two disulfide bonds, Cys100–Cys105 and Cys352–Cys355. The disruption of any of these disulfide bonds by removal of any of the four cysteine residues leads to the accumulation of misfolded proteins in the ER lumen. Similar proteins (Ero1a and Ero1b) have been found in human cells exhibiting similar functionality. A third ER protein, named Erv2, has been identified that carries FAD as noncovalent cofactor and can be reoxidized by molecular oxygen as electron acceptor. Erv2 can, like Ero1, reoxidize PDI.
222
R. Silvers et al.
Fig. 10.2 Disulfide bond formation and isomerization in the periplasm of Escherichia coli. (a) DsbA transfers oxidizing equivalents to the substrate by thiol–disulfide exchange. DsbA is reoxidized by the membrane protein DsbB that transfers electrons either to ubiquinone (aerobic conditions) or to menaquinone (anaerobic conditions). These electrons flow to molecular oxygen via the electron transport chain. (b) The disulfide bonds of a misfolded substrate can also be reshuffled by DsbC via thiol–disulfide exchange, in which case DsbC remains reduced. (c) DsbC can also reduce the disulfide bonds of misfolded substrates. In this case, DsbC need to be rereduced by DsbD that is re-reduced subsequently by the thioredoxin (TrxA) system. Figure modified after Mamathambika and Bardwell (2008)
In prokaryotes, oxidative folding occurs in the periplasm of gram-negative bacteria. For the formation of new disulfide bonds, two proteins, DsbA and DsbB, act as electron carrier and oxidize the substrate thiols by direct thiol–disulfide exchange (Fig. 10.2) (Bardwell et al. 1991; Pan and Bardwell 2006; Zapun et al. 1993). DsbA directly interacts with the substrate by a thiol–disulfide exchange mechanism and is reoxidized by the inner-membrane protein DsbB. The electrons from DsbB are passed to ubiquinone under aerobic conditions or menaquinone under anaerobic conditions (Bader et al. 1999, 2000; Kobayashi et al. 1997; Xie et al. 2002). These electrons are subsequently transported to molecular oxygen or anaerobic electron acceptors through an electron transport chain. Two additional proteins have been identified, DsbC and DsbD, that are involved in the reshuffling of misfolded substrates. The reduced form of DsbC interacts directly with misfolded substrates via direct thiol–disulfide exchange reactions. In the case where the disulfide bond of a substrate is reduced and another disulfide bond is formed, the disulfide bond of DsbC remains reduced thus active. Sometimes, however, the disulfide bond of the substrate is only
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
223
reduced and submitted to another oxidation by DsbA. In this case, the disulfide bond of DsbC remains oxidized and inactive until DsbC is re-reduced by DsbD through direct thiol–disulfide exchange, whereas DsbD itself is subsequently rereduced by the cytosolic thioredoxin (TrxA) system. DsbD consists of N-terminal (nDsbD) and C-terminal (cDsbD) periplasmic domains that are connected by a transmembrane domain. The electrons are transported from cDsbD to nDsbD by direct thiol–disulfide exchange before re-reducing DsbC. Mavridou et al. (2007, 2009) investigated the active site properties of the oxidized and reduced cDsbD in extensive NMR studies. Although cysteine residues are involved in a large variety of different functions, we want to focus in this contribution on the effect of cysteine residues on protein folding and stability and what nuclear magnetic resonance (NMR) studies can contribute to this area of research.
10.1.2 Theoretical Considerations Oxidative folding of proteins involves the formation of disulfide bonds and concomitant folding of the polypeptide chain to promote bond formation of sequentially distant disulfide bonds. Under physiological conditions, the final disulfide-bonded form often forms the most stable state, but during kinetic oxidative folding, native as well as non-native states with incorrect disulfide bonds can form. Therefore, intermediates with a different number of disulfide bonds do exist. A fundamental aspect of these intermediates is their transient structure and substantial dynamics that lead to conformational averaging that can be best characterized by NMR spectroscopy at an atomic level.
10.1.3 Structure and Conformational Stability of Disulfide Bonds Disulfides show rotation around the S–S bond with its dihedral angle c 3 . To investigate the height of the rotational barrier for this rotation, Ohsaku and Allinger (1988) investigated the molecular structure and conformational stability of ethyl methyl disulfide, a model for cystin residues. The ab initio self-consistent molecular ‡ orbital calculations resulted in rotational barrier heights of about DGtrans = 4kcal/mol ‡ and DGcis = 12kcal/mol (Fig. 10.3). The most stable conformations are given at c 3 = ±90° . Most disulfide bonds in protein structures obtained from the protein data bank (http://www.pdb.org) (Berman et al. 2000) exhibit similar values. Consequently, during protein folding, rotation around the disulfide bond will become restricted stepwise as the native structure forms until the rotation barrier is too high to be passed. For instance, when a disulfide bond is introduced in a mostly unstructured protein, the rotation around the disulfide bond is restricted by local interactions (i.e., hydrophobic, Coulomb interactions) in the vicinity of the disulfide bond
R. Silvers et al.
224
Fig. 10.3 Schematic representation of the relative energy of the cystin model ethyl methyl disulfide in rotation about the S–S bond. The results of the ab initio self-consistent molecular orbital calculations are shown in black. Hypothetical cases of restricted disulfide bond rotation due to hydrophobic or Coulomb interactions are shown in gray. In case of a folded protein, the rotational barriers becomes very large as insinuated by dashed gray lines
(Fig. 10.3, gray line). During further protein folding, the rotational barrier is restricted until rotation around the disulfide bond very rarely occurs (Fig. 10.3, gray dashed line).
10.1.4 Disulfide Bonds and Thermodynamics One of the main questions is the contribution of the disulfide bond on protein stability: is it of entropic or enthalpic nature? The oxidation of cysteines significantly limits the conformational space of an unfolded protein thus lowering entropy. This effect is correlated with the loop length, i.e., the number of residues that are enclosed to a circle by a disulfide bond. In 1956, Flory based the increase of free energy of the random-coil unfolded state upon the introduction of a single disulfide bond on the decrease of its conformational entropy. In terms of thermodynamics, the equilibrium between the unfolded (U) and native (N) state is shifted to N by the introduction of disulfide bonds. The decrease of entropy can be estimated by the probability that both ends of the protein chain occupy the same space, i.e., the same volume element us . Mathematically, this can be described by
DS = - R ln[3 / (2pl 2 n)3/ 2 ]us
(10.1)
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
225
Fig. 10.4 Schematic representation of the mode of operation of loop permutation analysis. The wild-type protein (left) is circularly permutated (right). Figure adapted from Zhang et al. (1994)
where R is the universal gas constant ( R = 8.314JK -1mol -1 ), n is the number of residues in the loop and l is the average segment length, i.e., the distance between two amino acid residues (l = 3.8 Å). The minimal distance of two thiol groups is about 4.8 Å, resulting in a volume element u s of 57.9 Å3, i.e., the volume of a sphere with a diameter of 4.8 Å. Equation (10.1) then reduces to DS = -2.1 - 1.5 R ln(n) (10.2) Therefore, the decrease in configurational entropy is predicted to be proportional to ln(n) [DS ~ −ln(n)]. To further evaluate the validity of these calculations, several experiments were conducted. The transition between the folded native and the denatured state of a protein can be thermodynamically characterized using the relation DG = DH − TDS. If an existing disulfide bond is removed or a new one is introduced (by mutation for example), one can determine the change of DG, which is DDG = D(DH − TDS). Furthermore, if the effect of this disulfide bond is purely entropic, DDH is zero and DDG can be directly correlated with TDDS. A key obstacle in thermodynamic studies of the effect of disulfide bonds on entropy and enthalpy is to distinguish between the effect of the disulfide bond on the folded protein structure and the effect on the unfolded protein chain. An interesting approach to dissect these contributions of disulfide bonds to entropy, enthalpy, and native-state effects is the loop permutation analysis. In loop permutation analysis, circular permutations are introduced to the protein in order to create several different loop lengths without disrupting the thermodynamic stability of the folded state (Fig. 10.4). Measuring the thermal stability by thermal denaturation of the reduced and oxidized forms of these mutants allows the determination of DDG, the stability relative to the cysteine free forms. The difference in DDG of the oxidized and reduced forms then gives the disulfide contribution.
10.1.5 Disulfide Bonds and the Kinetics of Protein Folding How do cysteine residues find their partner? The number of possible connectivities between cysteine residues can be calculated by simple means. The number of possible disulfide bonds (DSB) is given by
DSB =
n! (n / 2)!2 n / 2
(10.3)
226
R. Silvers et al.
Fig. 10.5 Schematic representation of the hypothetical oxidative folding pathway of hen egg white lysozyme without scrambled intermediates
where n is the number of cysteine residues within the molecule. For a protein with eight cysteine residues such as ribonuclease A or lysozyme, 105 possible disulfide bonds could be formed. Comparable to the Levinthal paradox, it seems impossible for the protein to successively try every possibility. Furthermore, even in vitro lysozyme refolding experiments without proper redox machinery (PDI) do not yield 105 different species. If one assumes that, in the case of lysozyme, only native disulfide bonds form during refolding, 15 stable intermediates can form, including the folded native form (Fig. 10.5). The meticulous investigation of these intermediates by NMR spectroscopy can yield a detailed view of oxidative folding of lysozyme at an atomic level.
10.2 NMR Techniques NMR spectroscopy can be used to determine structures of folded states of proteins. NMR has contributed approximately 15% of the total number of protein structures deposited in the protein data bank (Berman et al. 2000). Since NMR can also be applied under a variety of different solvent conditions and heteronuclear multidimensional techniques provide sufficient resolution to assign most resonances of non-native states of proteins, NMR is probably the most powerful method to characterize non-native states of proteins in solution with atomic resolution. The introduction of one or more disulfide bonds to an unfolded protein usually yields the modulation of several parameters that can be determined by NMR spectroscopy. For instance, the change in hydrodynamic radius between the reduced and oxidized forms of unfolded proteins can be evaluated by Diffusion-ordered spectroscopy (DOSY). Chemical Shifts strongly correlate with secondary structure
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
227
propensity and can reveal populations of residual secondary structure for each amino acid in unstructured proteins. Since it is known that disulfide bonds stabilize secondary structures, chemical shifts can be used to characterize the stabilizing effect of introducing disulfide bonds. The modulation of transient structure and dynamics of backbone and sidechains upon disulfide bond oxidation or reduction can be traced by several NMR methods that characterize transient conformational structure (J Coupling Constants) and dynamics on different timescales (Relaxation Rates and Residual Dipolar Couplings). In this section, we focus on DOSY, chemical shifts, and relaxation rates as NMR methods to show the impact of disulfide bonds on unfolded states of proteins.
10.2.1 General Characteristics of Unfolded Proteins Investigated by NMR A prerequisite for any detailed characterization of structure and dynamics of unfolded states of proteins at residue-specific or even atomic resolution is the assignment of NMR of the amino acid residues. Due to molecular averaging in unfolded ensembles, the chemical shift dispersion observed in 1H-NMR experiments is very low so that a detailed characterization is not feasible. Through the introduction of isotope labeling and heteronuclear NMR spectroscopy it was possible to gain more detailed information due to the introduction of another spectral dimension (15N) in 1H,15N heteronuclear correlation spectroscopy (1H,15N-HSQC). In other words, the low chemical shift dispersion in 1H can be compensated by a second spectral dimension (15N), allowing a detailed analysis even in unfolded proteins (Fig. 10.6, inset).
10.2.2 Identification of Disulfide Bonds by NMR The identification of disulfide bonds in proteins is an important task in protein biochemistry. To date, the exact number of disulfide bonds and their distribution in proteomes is unclear, and predictions from databases proved to be complex due to biased database structures. X-ray crystallography is one possibility to pinpoint the redox state of a cysteine residue in a given protein, provided the protein state can be crystallized. Others include chemical, biochemical, and bioinformatics methods that are often breathtakingly extensive. However, these techniques, especially X-ray crystallography, fail to identify the redox state of cysteine residues in intrinsically unstructured proteins (IUPs) or non-native states of proteins. NMR spectroscopy, by contrast, can be applied to non-native states of proteins, ranging from unfolded, entirely unstructured states to states with different degrees of nonrandom structure including molten globule states (Dyson and Wright 2004). From an NMR point of view, two distinct techniques can be used for the identification of disulfide bonds.
228
R. Silvers et al.
Fig. 10.6 Two-dimensional 1H, 15N-HSQC experiments of hen egg white lysozyme in water pH 3.8 (left) and 8 M urea pH 2.0 (right) at 293 K and 600 MHz 1H spectrometer frequency. The left spectrum shows a high chemical shift dispersion for a folded protein whereas the right spectrum displays only low chemical shift dispersion typical for an unfolded state of proteins
10.2.2.1 Identification by 13C NMR Chemical Shifts Based on analysis of the BioMagResBank (BMRB) and Sheffield databases as well as published data, a database of cysteine 13Ca and 13Cb chemical shifts could be compiled (Sharma and Rajarathnam 2000). These chemical shifts allow the discrimination between cysteine residues in the reduced and oxidized forms. Chemical shifts in either the reduced or oxidized form occupy distinct regions in the 13Ca/13Cb chemical shift map depending on secondary structure elements (Fig. 10.7). 10.2.2.2 Identification by Diselenide Proxies Using 77Se NMR Spectroscopy Since the sulfur isotope 32S is NMR inactive due to its nuclear spin (I = 0) and known stable isotopes are also unfavorable for NMR spectroscopy, 32S has to be replaced by an NMR active atom with similar chemical properties (Mobli et al. 2009). Hence, cysteine residues can be substituted by selenocysteine, often referred to as the 21st amino acid. The natural abundance of 77Se is around 7.6% with a nuclear spin of 1/2. The use of a 2D 1H–77Se heteronuclear multiple bond correlation (HMBC) experiment correlating the Hb protons with the 77Se signal allows the unequivocal assignment of diselenide bond connectivity.
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
229
Fig. 10.7 13Ca and 13Cb chemical shift dependence on both redox state and secondary structure. Data have been extracted from and figure has been modified after Sharma and Rajarathnam (2000)
10.2.3 Chemical Shift Analysis Chemical shifts depend on the chemical environment of the observed nucleus. Generally, non-native states of proteins show lower chemical shift dispersion than proteins in their native state. Due to conformational averaging, NMR active nuclei (1H, 15N, 13C) resonate close to or at their random coil chemical shifts. For an ideal random coil polypeptide chain, NMR spectra look similar to the spectra of the mixture of the amino acids (McDonald and Phillips 1969). This finding supports the concept that nonlocal interactions are absent in the random coil state. However, the local amino acid sequence partially influences random coil chemical shifts of 15N, 1HN, and 13CO resonances (Schwarzinger et al. 2001a). Chemical shifts strongly correlate with secondary structures (Markley et al. 1967; Nakamura and Jardetzky 1968). Determination of secondary structure elements in proteins by using chemical shift index (CSI) calculation is therefore a common method both in solution and in solid-state NMR spectroscopy (Groß and Kalbitzer 1988; Pastore and Saudek 1990; Spera and Bax 1991; Szilágyi 1995; Williamson 1990; Wishart et al. 1991a, b, 1992). The experimental chemical shifts of nuclei from amino acid residues are compared to random coil values to identify secondary structures for a given amino acid. Random coil chemical shifts under several experimental conditions have been determined independently by different research groups (Bienkiewicz and Lumb 1999; Braun et al. 1994; Bundi and
230
R. Silvers et al.
Fig. 10.8 Schematic representation of time scales accessible to NMR spectroscopy and of associated protein dynamics. NMR spectroscopy allows studying dynamics that span more than 12 orders of magnitudes in time
Wüthrich 1979; Glushka et al. 1989; Merutka et al. 1995; Plaxco et al. 1997; Richarz and Wüthrich 1978; Schwarzinger et al. 2000; Thanabal et al. 1994; Wishart et al. 1995a). The effects of secondary structures (helix or sheet) on the observed chemical shift have led to the introduction of secondary chemical shifts Dd (Dd = dobserved − drandom coil). These secondary chemical shifts are mainly influenced by noncovalent interactions, e.g., change of secondary structure, hydrogen bonding, or aromatic stacking. For unfolded or partially folded proteins, secondary chemical shifts are indicative for the presence and population of residual structures. Different methods have been developed to determine residual structures in unfolded or partially folded states of proteins and peptides by using chemical shift data (Pastore and Saudek 1990; Wishart et al. 1991a, 1992; Schwarzinger et al. 2000; Cornilescu et al. 1999; Szilágyi and Jardetzky 1989; Wang and Jardetzky 2002; Wishart and Sykes 1994a). These methods include the Dd method (Reily et al. 1992), the CSI method (Wishart et al. 1992; Wishart and Sykes 1994b), the probability-based method (Wang and Jardetzky 2002), the LACS method (Wang et al. 2005), and the SSP method (Marsh et al. 2006a).
10.2.4 Relaxation Techniques Biomacromolecules can display motions on different time scales ranging from fast motions such as libration, vibration, and side-chain rotation (~fs–ps) to folding processes that can take seconds or even minutes (Fig. 10.8). Relaxation techniques have proven to be a useful means for probing conformational dynamics in biomacromolecules. Generally, in analyzing relaxation rates of nuclei by NMR spectroscopy one can distinguish between two time regimes, the fast (fs–ns, faster than the overall rotational correlation time) and slow (ms–s, of the order of the differences in chemical shifts between different conformational states of the protein) time limit.
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
231
10.2.4.1 Relaxation Processes Sensitive to Dynamics on the Fast Time Scale Heteronuclear spin relaxation rate constants and heteronuclear NOEs are influenced mostly by conformational dynamics on the fast time scale (subnanosecond) (Wagner 1993). These NMR parameters are modulated by the overall rotational tumbling and local fluctuations. For isotropically tumbling molecules, the rotational tumbling is influenced by molecule size (r), temperature (T ), and solvent viscosity ( h ) according to the Einstein–Stokes–Debye relation DR = kbT / (8phr 3 ) (Brilliantov et al. 1991). It is often expressed by the total correlation time t c , which is given by DR = 1 / t c . Local fluctuations, on the other hand, are influenced by chain stiffness, local, and long-range interactions. These local fluctuations are often probed as fluctuations of the NH bond vector of the peptide bond for each residue. Heteronuclear relaxation rates for a spin S (typically 15N) in an IS spin system depend on two relaxation mechanisms – the dipole–dipole interaction of spin S with a spin I (typically 1H) and the chemical shift anisotropy (CSA) of spin S. Thus, the spin–lattice or longitudinal relaxation rate constant (R1) and spin–spin or transverse relaxation rate constant (R2) can be expressed by (10.4) and (10.5), respectively:
R1 = R2 =
d2 [J (w I - w S ) + 3J (w S ) + 6 J (w I + w S )]+ c2 J (w S ) 4
d2 [4 J (0) + J (w I - w S ) + 3J (w S ) + 6J (w I + w S ) + 6J (w I )] 8 c2 + [4 J (0) + 3J (w S )] 6
(10.4)
(10.5)
The heteronuclear nuclear Overhauser enhancement (hetNOE) is given by (10.6):
NOE = 1 +
d2 g I [6 J (w I + w S ) - J (w I - w S )] 4 R1 g S
(10.6)
In (10.4)–(10.6), w S and w I are the Larmor frequencies of spin S and spin I, respectively. The spectral density function at frequency w is J (w ) ; d = m0 hg Sg I / (8rIS3 p 2 ) and c = w S Ds S / 3 , where m 0 is the permeability of the vacuum, h is Planck’s constant, g S and g I are the gyromagnetic ratios of nuclei S and I, respectively, rIS is the distance between nuclei S and I and Ds S is the CSA of spin S. In 1982, Lipari and Szabo (1982a, b) described a “model-free” formalism, which is used to fit the spectral density function J (w ) to experimentally obtained relaxation data. Assuming isotropic diffusion, J(w) is given by
J (w ) =
2 é S 2t c (1 - S 2 )t 2 ù + ê ú 2 2 5 ë1 + w t c 1 + w 2t 2 û
(10.7)
where the order parameter S² is a measure of the spatial restriction of the NH bond vector motion (Fig. 10.9), tc is the overall rotational correlation time of the protein
232
R. Silvers et al.
Fig. 10.9 Schematic representation of model-free parameters for isotropic overall tumbling with the order parameter S2, the effective correlation time te, and the total correlation time tc. Figure modified after Palmer (2004)
Fig. 10.10 (a) Simulation of heteronuclear 15N R2 (solid line) and R1 (dashed line) relaxation rates and of the hetNOE (gray line) in dependence of the global correlation time tc according to (10.4)– (10.6). For the simulation, it was assumed that there is no internal motion and that the molecule tumbles isotropically [see (10.10)]. (b) Dependence of heteronuclear 15N R2 rates on the global correlation time tc under different degrees of internal motion assuming an internal correlation time of 200 ps [see (10.7)]. Order parameters S2 of 1, 0.8, and 0.6 were used (black, gray and light gray lines, respectively).
molecule and 1/t = 1/tc + 1/te with te defining the time-scale of the internal motion of the NH bond vector. Figure 10.10a shows the dependence of R2 and R1 relaxation rates on the total correlation time t c assuming no internal motions (S² = 1, i.e., fully rigid molecule) and overall isotropic tumbling. The simulation of transverse 15N relaxation rates (Fig. 10.10b) in the presence of internal motion shows that the relaxation rate decreases with increasing internal motion (i.e., decreasing order parameter). However, the use of the Lipari–Szabo approach requires that global (tc) and internal (te) motions are not correlated to other motions and can be treated separately. Typically, this requires a separation in time scales of the underlying motions of at least one order of magnitude. The prerequisite of separable global and internal motions is therefore questionable for unfolded states of proteins.
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
233
Hence, other approaches for the analysis of dynamics in unfolded states of proteins have been developed. Alexandrescu and Shortle (1994) introduced a local model-free approach where a global correlation time is fitted per residue. A continuous distribution of correlation times can be used, if extensive averaging over a range of frequencies in the nanosecond regime can be assumed. Buevich and Baum used the Cole–Cole distribution to describe the relaxation behavior of the unfolded propeptide of subtilisin PPS (1999). Yet another improved approach was proposed by Ochsenbein et al. (2002) who used a Lorentzian distribution of correlation times to describe the dynamics of the ensemble of conformers. ∞
J (w ) = ò F (t )J (w , t )dt
(10.8)
D ì for0 t t max 2 ïK 2 F (t ) = í D + (t - t 0 ) ï 0 fort t max î
(10.9)
0
J (w , t ) =
2 t 5 1 + (wt )2
(10.10)
where the Lorentzian distribution is characterized by its center t 0 and width D and is extending from zero to an upper value t max . K is the normalization constant, so ¥ that ò F (t )dt = 1 is satisfied: 0
t -t0 t 1 = arctg max + arctg 0 t K D D
(10.11)
In another approach, Schwalbe et al. (1997) fitted the sequence dependence of the relaxation rate to a simple two-parameter model, called the segmental motion model. The first parameter is the common intrinsic relaxation rate for a given residue and is influenced by the viscosity of the solution, while the second parameter is based on the distance dependence of the influence of neighboring residues on the relaxation rate for a given residue. This approach, despite lacking a detailed timescale analysis, can predict the sequence dependence of relaxation rates in polypeptides with or without disulfide bonds, since the introduction of disulfide bonds leads from an unbranched polypeptide chain to a branched polypeptide chain.
10.2.4.2 Analysis and Interpretation of Relaxation Data in Unfolded States of Proteins R2 relaxation rate constants can be modeled by the segmental motion model (Schwalbe et al. 1997; Klein-Seetharaman et al. 2002). In the case of an unfolded protein without disulfide bonds, two contributions to the experimental relaxation
R. Silvers et al.
234
Fig. 10.11 Simulation of the segmental motion model. (a) An unbranched polypeptide with 129 amino acids was simulated: The baseline, given by (10.12), is shown in gray and the simulation assuming the presence of six hydrophobic clusters (10.11) is shown in black. (b) A branched polypeptide with 129 amino acids and a [30–115] disulfide bond was simulated: The baseline, given by (10.16), is shown in gray and the simulation assuming the presence of six hydrophobic clusters (10.15) and one disulfide bond [30–115] is shown in black
rate can be fitted: the contribution of the random coil polypeptide chain R RC and the contribution of hydrophobic cluster R Cluster . A simulation of (10.12) is shown in Fig. 10.11a. with
R exp (i ) = R RC (i ) + R Cluster (i )
(10.12)
N
R RC (i ) = Rrc å e - ( i - j / l 0 )
(10.13)
j =1
R Cluster (i ) = Rcluster
åe
cluster
- (0.5( i - c cluster )/ lcluster )2
(10.14)
where N is the number of residues, Rrc is a measure for the intrinsic relaxation rate and is determined by the viscosity and temperature of the solution. l0 is the persistence length of the chain and is a measure for the influence of the chain. Deviations from random coil behavior due to hydrophobic cluster can be modeled by (10.13). Rcluster, ccluster, and lcluster are modeling variables and are measures of the amplitude, the position, and the persistence length of the cluster, respectively. In the case of an unfolded protein with disulfide bonds, experimentally obtained relaxation rates can be dissected into maximal three contributions to the experimental relaxation rate [see (10.14)]. The contribution of the random coil polypeptide chain R RC,branched , the contribution of hydrophobic clusters R Cluster , and the contribution of disulfide bonds R DSB to the experimental relaxation rate R exp can be described by (10.15)–(10.17). The simulation of (10.15) is shown in Fig. 10.11b.
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
235
R exp (i ) = R RC,branched (i ) + R Cluster (i ) + R DSB (i )
(10.15)
with N
R RC,branched (i ) = Rrc å e
- ( dmij / l1 )
(10.16)
j =1
N
R dsb (i ) = Rdsb å e - ( i - j / ldsb )
(10.17)
j =1
where N is the number of residues, Rrc is a measure for the intrinsic relaxation rate and is determined by the viscosity and temperature of the solution. l1 is the persistence length of the chain and is a measure for the influence of the chain. Deviations from random coil behavior due to hydrophobic cluster can be modeled by (10.13). The topological distance matrix dmij is a N × N-matrix and contains the information of the shortest path between residue i and j as number of covalent bonds. According to this model, peptide bonds and disulfide bonds are treated equally as an “effective” covalent bond. Rdsb and ldsb are the modeling variables and are measures of the amplitude and the persistence length of the disulfide bond, respectively. 10.2.4.3 Relaxation Processes Sensitive to Dynamics on the Slow Time Scale Another dynamic process influencing the line width of NMR signals is chemical exchange which occurs in the slow time limit in NMR spectroscopy. Chemical exchange between n sites is generally characterized by kex,1
kex,2
kex,n -1
A1 A 2 , A 2 A 3 , , A n -1 A n
To simplify matters, we concentrate on two-site chemical exchange between two states. Typically, a two-site chemical exchange is characterized NMR spectroscopically by the two populations pA1 and pA2 , the exchange rate kex and the difference in chemical shift between the two states Dd . In NMR spectroscopy, three different types of chemical exchange exist: slow, intermediate, and fast exchange (Fig. 10.12). In the case of slow exchange, kex is slow compared to the chemical shift difference Dd resulting in two NMRs that are well separated. If kex faster than the chemical shift difference Dd (fast exchange), there will be only one NMR visible, since both states cannot be resolved on the NMR time scale. Intermediate exchange (kex ~ Dd ) is characterized by a broad resonance of low intensity. There are generally two popular ways to investigate exchange processes by NMR: During R1r (Davis et al. 1994; Deverell et al. 1970) relaxation experiments the magnetization is spin-locked in the rotating frame by application of a radio frequency (rf) field. Another technique is the evolution during spin-echoes by applying CPMG spin echo sequences (Carr and Purcell 1954; Meiboom and Gill 1958). CPMG experiments have also been shown to be useful for the investigation of low populated states (so-called invisible states).
R. Silvers et al.
236
Fig. 10.12 Schematic representation showing the dependence of the appearance of NMR signals on different exchange modes
10.2.5 Diffusion Ordered Spectroscopy The hydrodynamic radius (Rh) is a measure for the dimensions of a spherical protein. As mentioned earlier, one or more disulfide bonds can modulate the compactness of a protein, making Rh a valuable tool for the characterization of disulfide-bonded, unfolded proteins. Rh is inversely correlated to the diffusion constant D and is given by the Stokes–Einstein equation: Rh =
k BT 6phD
(10.18)
with kB = Boltzmann constant, T = temperature, and h = viscosity of the solution. To measure the protein’s hydrodynamic radius by NMR spectroscopy, an internal standard of known properties is used to gauge the measurement. For proteins, dioxane with an Rh of 2.12 Å (Wilkins et al. 1999) is used and the diffusion constant of the protein Dprot and of dioxane Ddiox are measured in the same sample and experiment. Then the protein’s hydrodynamic radius is
Rhprot =
Ddiox diox Ddiox Rh = 2.12 Å Dprot Dprot
(10.19)
These diffusion constants are typically obtained by pulse field gradient (PFG) NMR experiments. The experiments contain a pulse gradient stimulated echo longitudinal encode-decode (PG-SLED) sequence varying the strength of bipolar gradient pulses for diffusion (Wu et al. 1995). This so-called diffusion ordered spectroscopy (DOSY) experiment results for lysozyme and dioxane in a pattern as shown in Fig. 10.13. The signal decay can be fitted by 2
S ( g ) = Ae - Dg
(10.20)
where A is the amplitude factor, D is the diffusion constant, and g the gradient strength. Hence, these signals decay with increasing strength of gradients; the decay
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
237
Fig. 10.13 Example for a DOSY experiment. Signal intensities in a 1H 1D NMR spectrum vary as a function of gradient strength (g.s.). (a) lysozyme mutant 0SS-W62G, (b) dioxane
rate is higher for the small molecule dioxane (B) than for the protein (A). In the case of unfolded states of proteins, the hydrodynamic radius is an averaged value over the ensemble of conformations in contrast to natively folded proteins. Thus, changes in this distribution of the ensemble caused by mutations or cross-linking are translated in the measured Rh.
10.3 Examples In the following, we describe exemplary proteins for which unfolded states with varying degree of disulfide bonds have been characterized by NMR spectroscopy.
10.3.1 Hen Egg White Lysozyme Lysozymes, or muramidases, are enzymes that catalyze the hydrolytic cleavage of a b(1 → 4) glycosidic linkage between N-acetyl muraminic acid (NAM) and N-acetyl glycosamine (NAG) by a mechanism called the Phillips mechanism, after its discoverer. Lysozymes can be found in cells or extracellular spaces as well as different kinds of secretions, i.e., tears, sudor, nasal secretion, saliva, cerumen, albumen of birds, and even latex of different plants. The function of lysozyme and its first description was published by Alexander Fleming in (1922). Hen egg white lysozyme (HEWL) has been isolated from the albumen of chicken. HEWL consists of 129 amino acids and has a molecular weight of 14.3 kDa. The structure of HEWL was the first solved structure of an enzyme and it was solved by Phillips and co-workers in 1965 using X-ray crystallography at a resolution
238
R. Silvers et al.
Fig. 10.14 Ribbon diagram representation of hen egg white lysozyme (HEWL) based on its NMR solution structure (HTTP://www.pdb.org, PDB ID: 1E8L) (Berman et al. 2000; Schwalbe et al. 2001). Disulfide bonds are indicated: [6–127], [30–115], [64–80], and [76–94]. Helices are shown in dark gray, beta sheets in medium gray, and coils in light gray. The structure was rendered with PyMOL
of 2.0 Å (Blake et al. 1965). The native form of HEWL consists of two domains denoted as a-domain (residues 1–35 and 85–129) and b-domain (residues 36–84). The a-domain consists of four a-helices and a 310-helix, whereas the b-domain is built up by a triple-stranded b-sheet (anti-parallel), an a-helix and a loop region (Fig. 10.14). HEWL has eight cysteine residues forming four disulfide bonds, which are Cys6–Cys127, Cys30–Cys115, Cys64–Cys80, and Cys76–Cys94. Oxidative refolding of HEWL proceeds via scrambled two-disulfide species (2SS). The 2SS intermediates fold rapidly into three intermediates each possessing three native disulfide bonds (des[6–127], des[76–94], and des[64–80]; Fig. 10.15) (van den Berg et al. 1999). One disulfide bond ([30–115]) is present in all three intermediates. Formation of the last disulfide bond is slow for des[76–94] since the disulfide connecting the a-domain with the b-domain has to be formed, while it is fast for the other two intermediates, which form the fourth disulfide bond within the a- and b-domains, respectively. 10.3.1.1 Hen Egg White Lysozyme Without Disulfide Bonds (0SS-HEWL) 0SS-HEWL (also HEWL-SMe), which contains no disulfide bonds, is unfolded even in the absence of denaturants as shown by CD measurements and the low chemical shift dispersion in the NMR-spectrum (Klein-Seetharaman et al. 2002; Collins et al. 2005). Using CD measurements and NMR chemical shift deviations, an average 13.4% residual a-helical secondary structure was identified in the protein clustering around Gly22, Val29 and Cys30–SMe, Trp62/Trp63, Trp108/Trp111 and Trp123. The location of these regions corresponds to four out of six hydrophobic regions predicted using the scale of Abraham and Leo (1987).
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
239
Fig. 10.15 Schematic representation of the oxidative refolding pathway of hen egg white lysozyme from its reduced unfolded state (Ured). The positions of cysteine residues in the amino acid sequence are indicated. Upon initiation of refolding, scrambled intermediates form first which contain two disulfide bonds (2SS). These species then fold rapidly into three intermediates each possessing three native disulfide bonds (missing disulfide bond designated by, e.g., des-[6–127]). All of these intermediates form the native state albeit at different rates
Analysis of heteronuclear 15N transverse relaxation rates (R2) identified higher rigidity in these six hydrophobic regions located around Ala9, Trp28, Trp62/Trp63, Leu83, Trp108/Trp111, and Trp123 (Fig. 10.16). Native-like and non-native like longrange interactions between these clusters were found by a combination of R2 relaxation rate measurements with nonconservative single point mutations A9G, W62G, W62Y, W111G, and W123G: Single point mutations change relaxation rates distant from their mutation site, with the strongest effects observed in W62G. Hydrophobic long-range interactions in the unfolded state of the protein as shown here might (a) prevent the protein from aggregation by burying the hydrophobic patches and (b) guide the earliest steps in protein folding. Van Berg et al. showed that three out of the four disulfide bonds of native HEWL are formed very rapidly during protein folding leading to the formation of three different disulfide intermediates. The Cys30–Cys115 disulfide bond is the only disulfide bond always formed during the fast folding step. The results on HEWL-SME shown here suggest that the buildup of a core by clusters 2, 5, and 6 enables the fast formation of this disulfide bond. The impact of these long-range interactions on native disulfide bond formation has recently been investigated by Mishima et al. (2007). They studied the disulfide exchange equilibria of all four single disulfide mutants of lysozyme possessing only one out of four disulfide bonds and either an additional W111G or a W123G mutation compared to their wild-type single disulfide pendants. The mutant
240
R. Silvers et al.
Fig. 10.16 Transverse relaxation rates (R2) in wild-type HEWL-SMe and the A9G, W62G, W62Y, W111G, and W123G mutations of HEWL-SMe. The baselines (light gray) are fitted by (10.13), Gaussian fits are given by black curves for wild-type HEWL-SMe, W111G, and W123G mutations and are modeled by (10.12). Dark gray curves for A9G, W62G, W62Y, W111G, and W123G mutations mark the Gaussian fit curve of wild-type HEWL-SMe (data taken from Wirmer et al. 2004)
W111G significantly reduced the extent of disulfide bond formation of 1SS6–127 and 1SS30–115, whereas disulfide bond formation remained unchanged for the other mutants 1SS64–80 and 1SS76–94. Introduction of an additional W123G mutation on the other hand only diminishes the capability of 1SS6–127 to form disulfide bonds. It is well known that the W111G mutation disrupts the hydrophobic clusters 5 and 6 in HEWL-SME, whereas W123G only disrupts cluster 6. Apparently, the loss of hydrophobic clusters 5 and 6 leads to unfavorable long-range dynamics
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
241
Fig. 10.17 Transverse relaxation rates (R2) in wild-type 4SS-HEWL (a) and 4SSW62G-HEWL (b). The baselines (light gray) are fitted by (10.16), disulfide bonds of 4SS-HEWL are fitted by (10.15) and plotted by a black curve (a) and dark gray curve (b) (data taken from Collins et al. 2005)
for disulfide formation of Cys30–Cys115 and Cys6–Cys127, respectively. Similar studies were used to investigate the effect of a W62G mutation on the disulfide exchange equilibrium of these four disulfide mutants (Ohkuri et al. 2005). These results strongly suggest that the long-range interactions between several hydrophobic clusters within lysozyme modulate disulfide bond formation, hence controlling early protein folding events. 10.3.1.2 Hen Egg White Lysozyme with all Disulfide Bonds (4SS-HEWL) Fully oxidized 4SS-HEWL is native even under acidic conditions. Therefore, 8 M urea at pH 2 is required to unfold the protein as monitored by the low chemical shift dispersion. 15N labeled unfolded HEWL was assigned using TOCSY and NOESY spectra and over 900 NOEs including 130 (i,i + 2) NOEs and 23 (i,i + 3) NOEs could be identified by the analysis of NOESY spectra (Schwalbe et al. 1997). These NOEs were compared to predictions derived from a random coil polypeptide model and it was shown that they are indeed expected in a random coil. Residual secondary structure as investigated by chemical shift perturbations and 3J(HN,Ha) coupling constants is found in the same regions as in unfolded 0SS-HEWL in water. In contrast, the relaxation rate profile is different with the position of maxima in the rates being dominated by the positions of cysteine residues. Furthermore, R2 rates range from 4.4 to 26.9 s−1 indicating a severe rigidification as compared to 0SS-HEWL (Fig. 10.17). The loss in flexibility is due to the interplay of hydrophobic long-range interactions and the disulfide bonds, which could be shown by the investigation of a W62G-4SS mutant in which hydrophobic long-range interactions are destroyed and R2 rates drop to the values expected for a branched polypeptide chain (KleinSeetharaman et al. 2002).
242
R. Silvers et al.
10.3.1.3 Hen Egg White Lysozyme with Three Disulfide Bonds (3SS-HEWL) Des[6–127] HEWL has the same native structure as native lysozyme as shown by Dobson et al. (1991) using chemical shift and NOE data. However, compared to native lysozyme, thermal as well as pH stability are drastically reduced due to the absence of the disulfide bond. The variant is unfolded at pH 2 at room temperature (in contrast to 4SS-HEWL which is native at pH 2) as shown by the small chemical shift dispersion and sharp lines in the NMR spectra as well as by far-UV and nearUV spectra of the protein. Yokota et al. investigated the structure of the other three 3SS variants using NOE and hydrogen exchange measurements (2004). The three variants are native-like with different degrees of unstructured parts. Des[76–94] is quite similar to the wildtype but for the segment from residue 74 to residue 94. Des[64–80] is disordered in between residues 62 and 79 with a network of hydrogen bonds within the b-domain being disrupted. In des[30–115] helix D is disrupted, disturbing the interface to the b-domain as well. 10.3.1.4 Hen Egg White Lysozyme with Two Disulfide Bonds (2SS-HEWL) Studies on lysozyme mutants containing only two disulfide bonds have been conducted by Schwalbe et al. using heteronuclear 15N transverse and rotating frame longitudinal relaxation rates (R2 and R1r, respectively). Their work is based on two lysozyme mutants. 2SSa, which contains the native disulfide bonds in the a-domain and 2SSb in which only the two native disulfide bonds of the b-domain are present. 2SSa is in a native-like conformation in water at pH 3.8 with the a-domain being intact and the b-domain being largely unstructured, with exception of the b1 and b2 strands that form an antiparallel b-sheet as shown by NOE measurements (Noda et al. 2002). 15N R2 relaxation rates are relatively uniform for the folded region of the protein with an average rate of 10.1 s−1, which is of the order expected for a folded protein of this size. Parts of the b-domain could not be assigned. However, relaxation rates of unassigned peaks are around 5 s−1, resembling an R2 value expected for a denatured protein. 2SSa can be denatured by the addition of 8 M urea at pH 2. In this state, the protein is unfolded with R2 relaxation rates ranging from 3.2 to 11.4 s−1 (Fig. 10.18), being significantly lower than in 4SS-HEWL. R1r rates of the mutant are only very slightly lower, indicating a tiny contribution of exchange to the R2 rates. The distribution of R2 rates in 2SSa is quite similar to that of 0SS-HEWL in the b-domain while the R2 rate distribution within the a-domain is similar to the observed R1r rates of 4SS-HEWL in urea, with exception of the region around clusters 5 and 6, where contributions of the hydrophobic clusters around W108/W111 and W123 are clearly seen as well. The correlation of R2 rates of 2SSa with the R1r values of 4SS-HEWL within the a-domain of the protein suggests, that while in the 4SS-HEWL rotational isomerization around the disulfide bond leads to ms–ms timescale motions the
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
243
Fig. 10.18 Transverse relaxation rates (R2) in lysozyme mutants 2SSa-HEWL (a) and 2SSbHEWL (b). 0SS-HEWL is shown in gray, clusters and disulfide bonds are fitted by (10.15) and plotted by a black curve (data taken from Collins et al. 2005)
energy barriers for these motions are diminished by the reduction of the two disulfides within the b-domain, suggesting cooperativity between the two domains. Similar observations can be deduced for the 2SSb mutant which is unfolded both, in water and in urea at pH 2 (Fig. 10.18). In this mutant the R2 rate distribution within the b-domain is similar to the R2 data of 4SS-HEWL while the R2 rates of the a-domain are similar to 0SS-HEWL. R1r rates for the b-domain residues are significantly lower than R2 values, indicating exchange around the disulfide bonds.
10.3.2 Bovine Pancreatic Trypsin Inhibitor One of the best studied proteins in the field of unfolded states of proteins, disulfide bonds, and protein folding is the small globular bovine pancreatic trypsin inhibitor (BPTI). During oxidative refolding of the protein, non-native as well as native disulfides are populated. These intermediates have all been studied by the use of NMRspectroscopy, which will be reviewed below. BPTI is part of the serine protease family of inhibitors. Before maturation, BPTI with its signal peptide and propeptides consists of 100 amino acid residues, in its mature form; however, BPTI has 58 amino acid residues in length. The native BPTI comprises an antiparallel b-sheet and two short a-helices (Fig. 10.19). Three disulfide bonds stabilize the native fold, designated as [5–55], [14–38], and [30–51]. The oxidized refolding of the protein was one of the first pathways studied and remains probably one of the best characterized pathways until today. Depending on the pH value, different pathways are observable (Creighton 1990; Weissman and Kim 1991). In studies at near neutral pH (pH 7.3), which is well below the pKa of the cysteine residues only native disulfide bonds are populated (Weissman and Kim 1991). At pH 8.7 which is at the pKa of the thiols and thus a more suitable pH
244
R. Silvers et al.
Fig. 10.19 Ribbon diagram representation of bovine pancreatic trypsin inhibitor (BPTI) based on its crystal structure (http://www.pdb.org, PDB ID: 1BPI) (Berman et al. 2000; Parkin et al. 1996). Three disulfide bonds are indicated: [5–55], [14–38], and [30–51]. Helices are shown in dark gray, beta sheets in medium gray, and coils in light gray. The structure was rendered with PyMOL
to observe folding, also non-native disulfide species are found (Creighton 1990). This pathway is shown in Fig. 10.20. The first step of refolding is the formation of a number of native and non-native one-disulfide species in rapid equilibrium by intramolecular thiol–disulfide exchange. Three 2SS intermediates are formed from the [30–51] intermediate, two of which contain a non-native second disulphide bond ([30–51, 5–38] and [30–51, 5–14]) and one contains two native disulfide bonds [30–51, 14–38]. Disulfide shuffling results in the native-like 2SS intermediate [30–51, 5–55]N which rapidly converts into the native state. Unfolding by reduction reverses the pathway. Investigation of the reduced state, single-disulfide variants and two-disulfide variants of BPTI were performed by mutating cysteine residues such that no, one, or two disulphide bridges can be formed. 10.3.2.1 BPTI Without Disulfide Bridges Fully reduced BPTI at pH 4.5 is unfolded as shown by the low chemical shift dispersion in the NMR spectra (Kemmink and Creighton 1993; Pan et al. 1995). The region between residue 18 and 24 constituting the b1 sheet in the native conformation has a preference toward extended structure while no b-strand propensities are observed in the b2 region (29–35) as shown by NOE measurements. Native-like side-chain interactions are observed between Y35 and G37 and between Y23 and A25. Particularly, the Y23 A25 interaction proposes the formation of a native-like turn which facilitates folding. Relaxation measurements indicate that residues in the core of the protein are more rigid than in the rest of the protein (Barbar et al. 2001). Furthermore, the protein is slightly more collapsed than expected for a perfect random coil as shown by pulsed field gradient NMR diffusion measurements (Pan et al. 1997).
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
245
Fig. 10.20 Oxidative refolding of BPTI from its reduced unfolded state. The positions of cysteine residues in the unfolded polypeptide chain are indicated by the side-chain thiol groups (SH). Upon initiation of refolding, a number of one-disulfide intermediates is formed with the partially folded intermediate containing the native [30–51] disulfide being the predominating one. From this intermediate, three two-disulfide intermediates form possessing native as well as non-native disulfides. Disulfide reshuffling then leads to the formation of the native-like two-disulfide intermediate designated [30–51, 5–55] which rapidly converts to the native state. The plus sign connecting intermediates [30–51, 5–38] and [30–51, 5–14] illustrates their comparable kinetic roles. Figure modified after Creighton et al. (1996)
10.3.2.2 BPTI with One Disulfide Bridge The [5–55] single disulphide variant has been investigated by proton 1D (Darby et al. 1991) and 2D (van Mierlo et al. 1991a) NMR spectroscopy. 2D NMR spectra of the protein reveal that the native as well as non-native states of the protein are populated at 3°C to around 50% each. The reversible equilibrium between the two states was directly proven by saturation transfer measurements and by temperature-dependent NOESY and ROESY spectra showing weak exchange peaks. The native conformation of the single-disulfide variant is very similar to the native conformation of BPTI as shown by chemical shifts values and NOE contacts, explaining the rapid formation of the [14–38] disulfide in the [5–55] intermediate during protein folding. The [14–38] variant of BPTI consists of a native-like conformation in equilibrium with an unfolded conformation at pH 4.5, with two sets of signals but for G37
246
R. Silvers et al.
where three peaks are observed (Barbar et al. 1995). The native antiparallel b-sheet constituted of the b1 and the b2 strand is intact in the native-like conformation and stabilized by a significant number of long-range tertiary interactions as shown by long-range NOEs. Interestingly, the stable structure of the variant is not in the vicinity of the [14–38] disulfide bond as proved further by relaxation measurements (Barbar et al. 1998). Lowering of the pH value (Ferrer et al. 1995) shifts the equilibrium toward the denatured state of the protein such that at pH 2.5 the protein is predominantly unfolded. The pH denatured state (Barbar et al. 2001) is unfolded in terms of persistent residual secondary structure, however, the residues that fold into the central b-sheet are more ordered than the rest of the protein, as shown by relaxation measurements. Unfolding at pH 5 by the addition of urea is a non-two-state process, and the residues in the central core unfold later than the other residues (Barbar et al. 2001). The most important 1SS species of BPTI is the [30–51] disulfide variant which resembles the major 1SS folding intermediate. Detailed NMR investigations (Staley and Kim 1994; van Mierlo et al. 1993) revealed that the intermediate is partially folded with the first 15 residues being unfolded. Furthermore, weak extra peaks arising from an unfolded conformation in equilibrium with the folded state are observed. By comparison of the protein HSQC spectrum to those of two peptides (1–15) and (12–58) Staley and Kim could show that the two regions of the protein do not interact with each other. Further investigations by R1, R2, and heteronuclear NOE relaxation measurements revealed that besides the 15 N-terminal residues, residues 37–41 are also very flexible. This flexibility explains the formation of the second disulfide in the intermediate between two of the three cysteines Cys5, Cys14, and C38 including the two non-native disulfide bonds [5–14] and [5–38]. 10.3.2.3 BPTI with Two Disulfide Bridges The non-native two-disulfide intermediates [30–51, 5–14] and [30–51, 5–38] which are populated during productive folding of BPTI have been investigated by van Mierlo et al. (1994). Very similar to the [30–51] variant, both species are partially folded, with a minor unfolded conformation observed in the spectra. Significant differences among the three variants (the two 2SS variants and the 1SS variant) are located in the unstructured part close to the non-native disulfide bond. However, the non-native disulfides neither disrupt the native-like core of the protein, nor do they introduce non-native structure in their vicinity. This explains their important role in the folding of the protein: Breaking and rearrangement of these disulfides to form the native [5–55] disulfide are easily possible since no structure is induced by the disulfide bond. The two 2SS variants [30–51, 14–38] and [5–55, 14–38] are non-productive in folding since thiols are not accessible in these variants: they are folded as shown by NMR (van Mierlo et al. 1991b) and X-ray crystallography (Eigenbrot et al. 1990). The productive native 2SS variant [30–51, 5–55] is also very stable, has the native conformation and the same dynamical properties as the native wt protein as shown by R1 and R2 15N relaxation measurements (Beeser et al. 1998). The [14–38] disulfide bond can be easily formed in the intermediate, since the cysteines are well accessible.
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
247
Fig. 10.21 Ribbon diagram representation of human prion protein (hPrP) based on its NMR solution structure (http://www.pdb.org, PDB ID: 1QLX) (Berman et al. 2000; Zahn et al. 2000). The disulfide bond is indicated by [179–214]. Helices are shown in dark gray, beta sheets in medium gray, and coils in light gray. The structure was rendered with PyMOL
10.3.3 Human and Murine Prion Protein As another example, we wish to discuss the prion protein (PrP) in the following. PrP is known to be the causative agent of transmissible spongiform encephalopathies (TSEs), commonly known as prion diseases, which are accompanied by the accumulation of the pathological so-called scrapie form (PrPSc) of the endogenous cellular form (PrPC) in the brain and nervous system. The native structure of human PrP consists of a largely unstructured N-terminal domain (residue 23–120) (Donne et al. 1997; Riek et al. 1996) and a structured C-terminal domain (residue 121–230) that contains three a-helices, a short antiparallel b-sheet and a disulfide bond between cysteine residues 179 and 214, which connect helices 2 and 3 (Fig. 10.21) (Zahn et al. 2000). As is evident from the three-dimensional structure, PrPC has a large content of a-helices (~40%), whereas the fibrillar state of PrPSc is predicted to be mainly composed of a b-sheet superstructure (Caughey et al. 1991; Pan et al. 1993). Consequently, the conversion of cellular PrPC to the pathological PrPSc must include large-scale rearrangements of tertiary structure, a task that cannot be pursued without at least partially unfolding of the tertiary structure. Hence it can be postulated that an intermediate state between the normal, a-helical and abnormal, b-sheet form of PrP exists which has to be partially if not fully unstructured. PrP aggregation studies revealed that, under mildly denaturing conditions favoring the population of partially unstructured states, the presence of the disulfide bond has a dramatic influence on PrP fibrillation. While the disulfide-bonded protein rapidly aggregates upon dissolution in an appropriate aggregation buffer PrP does not show any aggregation if the disulfide is absent under otherwise identical experimental conditions (Gerum et al. 2009). Important insight into the determinants of PrP fibrillation has been gained from the investigation of the urea-denatured state of the C-terminal domain of human PrP
248
R. Silvers et al.
(hPrP) which can be regarded as a model for partially unstructured intermediate states being traversed on the path from PrPC to PrPSc (Gerum et al. 2009). Specifically, both the oxidized (hPrPox) and reduced forms (hPrPred) were studied in order to delineate the influence of the disulfide bond on the structural and dynamical properties of the unfolded state. The analysis of 13Ca and 13Cb secondary chemical shifts (Marsh et al. 2006b; Schwarzinger et al. 2001b; Wishart et al. 1995b) revealed clear structural preferences of the denatured state which are different from the secondary structure in the structured, native state of PrP (Fig. 10.22). Both the urea-denatured states of PrPox and PrPred show almost no residual a-helicity. However, three regions (I–III) with elevated b-sheet propensities can be deduced: region I (Ser135–Gly142) and region II (Tyr157 to Pro165) show b-sheet propensities of around 10%, region III (Ile182–Val189) of around 7%. Apparently, the disulfide bond has negligible influence on the conformational averaging on the level of single amino acid residues. Moreover, the conformational dynamics of the ureadenatured states of hPrPox and hPrPred have been investigated. This was done by measurement of transverse (R2) relaxation rates (Fig. 10.23a, b). Transverse relaxation rates for hPrPox and hPrPred show characteristic sequence patterns which can be analyzed using the segmental motion model (Schwalbe et al. 1997). Relaxation rates predicted by the segmental motion model for a perfect random coil without local or long-range interactions (light gray lines) do not match the experimental data (black lines) for both PrPox and PrPred, indicating the presence of local and/or longrange interactions within PrPox and PrPred. The R2 relaxation rate profiles for hPrPox and hPrPred are very different. The largest deviations occur around the two cysteine residues Cys179 and Cys214 that form the disulfide bond. Elevated relaxation rates in this part of the sequence originate from formation of the intramolecular loop restricting polypeptide chain flexibility significantly. Moreover, rotating frame longitudinal relaxation rates (R1r) do not deviate from transverse relaxation rates indicating the absence of conformational exchange on the micro- to millisecond time-scale (Fig. 10.23a, b). Because of the absence of slow conformational exchange dynamics and in light of the disulfide-dependent PrP fibrillation properties it can be concluded that it is the very pronounced difference in R2 rates accounting for the dramatically different behavior of PrP in aggregation assays. In other words, differences in fast polypeptide chain dynamics on the picoto nanosecond time-scale brought about by the disulfide bond likely account for this striking observation. This might be a general property of PrP as prion proteins from other species also contain a single disulfide bond at corresponding locations in their three-dimensional structures.(López García et al. 2000; Christen et al. 2009; Gossert et al. 2005) In addition, the structural and dynamical characteristics of the oxidized and reduced urea-denatured states of the murine prion protein (mPrP) are very similar as compared to hPrP (Fig. 10.23c, d) (Gerum et al. 2010). In line with this observation, mPrP rapidly fibrillizes when the disulfide bond is present whereas its absence prevents fibrillation entirely. Interestingly, most of the single point mutations which are associated with human prion diseases (Collinge 2001; Kovacs et al. 2002) coincide with the region whose
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
249
Fig. 10.22 Secondary structure propensities (SSP) (Marsh et al. 2006b) and normalized secondary chemical shifts for (a) hPrP ox and (b) hPrP red as a function of residue number recorded for the denatured states in 8 M urea, pH 2.0 and 25°C. Deviations in chemical shifts were calculated by subtracting experimentally determined chemical shifts from sequence-corrected random coil values (Schwarzinger et al. 2001b; Wishart et al. 1995b). Significant Ca, Cb, and CO (grouped into three categories: 0.35 ppm < |Dd| < 0.7 ppm, 0.7 ppm < |Dd| < 1.05 ppm and |Dd| > 1.05 ppm), and significant Ha and HN (grouped into three categories: 0.1 ppm < |Dd| < 0.2 ppm, 0.2 ppm < |Dd| < 0.3 ppm and |Dd| > 0.3 ppm) secondary chemical shifts are indicated by vertical bars. Black bars indicate b-structural preferences, white bars indicate a-helical structure. Regions of increased b-sheet propensities are labeled I–III. The native secondary structure elements and the sequence are indicated on top of each panel. High b-sheet propensities in the C-terminal region calculated by the SSP formalism are due to large chemical shift deviations of the last residue from random coil values, rather than to real structural preferences
250
R. Silvers et al.
Fig. 10.23 Heteronuclear 15N R2 (filled circles) and R1r (empty circles) relaxation rates for the oxidized and reduced denatured states (8 M urea pH 2 and 25°C) of the human (a and b) and murine (c and d) prion proteins. Dark gray lines show fits of the experimental data according to (10.15) (hPrPox and mPrPox) and (10.12) (hPrP red and mPrPred) whereas light gray lines indicate the relaxation rate profiles predicted for ideal branched (a and c) and unbranched (b and d) random coil polypeptide chains. Regions with b-sheet structural propensities as inferred from analysis of secondary chemical shifts (cf. Fig. 10.22) and the location of the single disulfide bond are given on top of each panel
structure and dynamics are modulated by the presence of the disulfide bond. This further underlines the potentially crucial role of the disulfide bond in PrP fibrillation. The prion protein is thus a nice example for the impact of disulfide bonds on the misfolding properties of a protein that is linked to human protein misfolding diseases. Acknowledgments RS is supported by the Stiftung Polytechnische Gesellschaft Frankfurt am Main (Germnay). KS is funded by the Deutsche Forschungsgemeinschaft (Normalverfahren: Charakterisierung struktureller und dynamischer Eigenschaften entfalteter Zustände von Mutanten des Prionenproteins). HS is member of the DFG-funded cluster of excellence: macromolecular complexes. BMRZ is funded by the state of Hesse, Germany.
References Abraham DJ, Leo AJ (1987) Extension of the fragment method to calculate amino acid zwitterion and side chain partition coefficients. Proteins 2(2):130–152 Alexandrescu AT, Shortle D (1994) Backbone dynamics of a highly disordered 131 residue fragment of staphylococcal nuclease. J Mol Biol 242(4):527–546
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
251
Bader M et al (1999) Oxidative protein folding is driven by the electron transport system. Cell 98(2):217–227 Bader MW et al (2000) Disulfide bonds are generated by quinone reduction. J Biol Chem 275(34):26082–26088 Banci L et al (2001) Solution structure of the yeast copper transporter domain Ccc2a in the Apo and Cu(I)-loaded states. J Biol Chem 276(11):8415–8426 Banci L et al (2006) Human SOD1 before harboring the catalytic metal: solution structure of copper-depleted, disulfide-reduced form. J Biol Chem 281(4):2333–2337 Banci L et al (2007) Metalation of the amyotrophic lateral sclerosis mutant glycine 37 to arginine superoxide dismutase (SOD1) apoprotein restores its structural and dynamical properties in solution to those of metalated wild-type SOD1. Biochemistry 46(35):9953–9962 Barbar E, Barany G, Woodward C (1995) Dynamic structure of a highly ordered beta-sheet molten globule: multiple conformations with a stable core. Biochemistry 34(36):11423–11434 Barbar E et al (1998) Dynamics of the conformational ensemble of partially folded bovine pancreatic trypsin inhibitor. Biochemistry 37(21):7822–7833 Barbar E et al (2001) NMR-detected order in core residues of denatured bovine pancreatic trypsin inhibitor. Biochemistry 40(32):9734–9742 Bardwell JC, McGovern K, Beckwith J (1991) Identification of a protein required for disulfide bond formation in vivo. Cell 67(3):581–589 Beeser SA, Oas TG, Goldenberg DP (1998) Determinants of backbone dynamics in native BPTI: cooperative influence of the 14–38 disulfide and the Tyr35 side-chain. J Mol Biol 284(5):1581–1596 Berman HM et al (2000) The protein data bank. Nucleic Acids Res 28(1):235–242 Bienkiewicz EA, Lumb KJ (1999) Random-coil chemical shifts of phosphorylated amino acids. J Biomol NMR 15:203–206 Blake CC et al (1965) Structure of hen egg-white lysozyme. A three-dimensional Fourier synthesis at 2 Angstrom resolution. Nature 206(986):757–761 Braun D, Wider G, Wüthrich K (1994) Sequence-corrected 15N “random coil” chemical shifts. J Am Chem Soc 116:8466–8469 Brilliantov NV, Denisov VP, Krapivsky PL (1991) Generalized Stokes–Einstein–Debye relation for charged Brownian particles in solution. Physica A 175(2):293–304 Brown NM et al (2004) Oxygen and the copper chaperone CCS regulate posttranslational activation of Cu, Zn superoxide dismutase. Proc Natl Acad Sci USA 101(15):5518–5523 Buevich AV, Baum J (1999) Dynamics of unfolded proteins: incorporation of distributions of correlation times in the model free analysis of NMR relaxation data. J Am Chem Soc 121(37):8671–8672 Bundi A, Wüthrich K (1979) 1H-NMR parameters of the common amino acid residues measured in aqueous solutions of the linear tetrapeptides H-Gly-Gly-X-L-Ala-OH. Biopolymers 18:285–297 Burgoyne JR et al (2007) Cysteine redox sensor in PKGIa enables oxidant-induced activation. Science 317(5843):1393–1397 Carr HY, Purcell EM (1954) Effects of diffusion on free precession in nuclear magnetic resonance experiments. Phys Rev 94(3):630 Caughey BW et al (1991) Secondary structure analysis of the scrapie-associated protein PrP 27–30 in water by infrared spectroscopy. Biochemistry 30(31):7672–7680 Chapman HA, Riese RJ, Shi GP (1997) Emerging roles for cysteine proteases in human biology. Annu Rev Physiol 59:63–88 Cheng H, Markley JL (1995) NMR spectroscopic studies of paramagnetic proteins: iron-sulfur proteins. Annu Rev Biophys Biomol Struct 24:209–237 Christen B et al (2009) Prion protein NMR structure from tammar wallaby (Macropus eugenii) shows that the beta2-alpha2 loop is modulated by long-range sequence effects. J Mol Biol 389(5):833–845 Ciurli S et al (1996) 1H NMR of high-potential iron–sulfur protein from the purple non-sulfurbacterium Rhodoferax fermentans. Eur J Biochem 236(2):405–411 Collinge J (2001) Prion diseases of humans and animals: their causes and molecular basis. Annu Rev Neurosci 24:519–550
252
R. Silvers et al.
Collins ES et al (2005) Characterisation of disulfide-bond dynamics in non-native states of lysozyme and its disulfide deletion mutants by NMR. Chembiochem 6(9):1619–1627 Cornilescu G, Delaglio F, Bax A (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J Biomol NMR 13:289–302 Creighton TE (1990) Protein folding. Biochem J 270(1):1–16 Creighton TE, Darby NJ, Kemmink J (1996) The roles of partly folded intermediates in protein folding. FASEB J 10(1):110–118 Darby NJ, van Mierlo CP, Creighton TE (1991) The 5–55 single-disulphide intermediate in folding of bovine pancreatic trypsin inhibitor. FEBS Lett 279(1):61–64 Davis DG, Perlman ME, London RE (1994) Direct measurements of the dissociation-rate constant for inhibitor-enzyme complexes via the T1 rho and T2 (CPMG) methods. J Magn Reson B 104(3):266–275 Deverell C, Morgan RE, Strange JH (1970) Studies of chemical exchange by nuclear magnetic relaxation in the rotating frame. Mol Phys 18(4):553–559 Donne DG et al (1997) Structure of the recombinant full-length hamster prion protein PrP (29–231): the N terminus is highly flexible. Proc Natl Acad Sci USA 94(25):13452–13457 Dubin G et al (2003) A novel class of cysteine protease inhibitors: solution structure of staphostatin A from Staphylococcus aureus. Biochemistry 42(46):13449–13456 Dyson HJ, Wright PE (2004) Unfolded proteins and protein folding studied by NMR. Chem Rev 104(8):3607–3622 Eigenbrot C, Randal M, Kossiakoff AA (1990) Structural effects induced by removal of a disulfidebridge: the X-ray structure of the C30A/C51A mutant of basic pancreatic trypsin inhibitor at 1.6 A. Protein Eng 3(7):591–598 Ferrer M, Barany G, Woodward C (1995) Partially folded, molten globule and molten coil states of bovine pancreatic trypsin inhibitor. Nat Struct Biol 2(3):211–217 Fleming A (1922) On a remarkable bacteriologic element found in tissues and secretions. Proc R Soc B Boil Sci 93:306–317 Frand AR, Kaiser CA (1998) The ERO1 gene of yeast is required for oxidation of protein dithiols in the endoplasmic reticulum. Mol Cell 1(2):161–170 Furukawa Y, Torres AS, O’Halloran TV (2004) Oxygen-induced maturation of SOD1: a key role for disulfide formation by the copper chaperone CCS. EMBO J 23(14):2872–2881 Gerum C et al (2009) Unfolded-state structure and dynamics influence the fibril formation of human prion protein. Angew Chem Int Ed Engl 48(50):9452–9456 Gerum C, Schlepckow K, Schwalbe H (2010) The unfolded state of the murine prion protein and properties of single-point mutants related to human prion diseases. J Mol Biol 401(1):7–12 Glushka J et al (1989) Nitrogen-15 chemical shifts of backbone amides in bovine pancreatic trypsin inhibitor and apamin. J Am Chem Soc 111:7716–7722 Goldberger RF, Epstein CJ, Anfinsen CB (1963) Acceleration of reactivation of reduced bovine pancreatic ribonuclease by a microsomal system from rat liver. J Biol Chem 238:628–635 Gossert AD et al (2005) Prion protein NMR structures of elk and of mouse/elk hybrids. Proc Natl Acad Sci USA 102(3):646–650 Groß K-H, Kalbitzer HR (1988) Distribution of chemical shifts in 1H nuclear magnetic resonance spectra of proteins. J Magn Reson 76:87–99 Isernia C et al (2003) NMR structure of the single QALGGH zinc finger domain from the Arabidopsis thaliana SUPERMAN protein. Chembiochem 4(2–3):171–180 Johnson DC et al (2005) Structure, function, and formation of biological iron–sulfur clusters. Annu Rev Biochem 74:247–281 Kemmink J, Creighton TE (1993) Local conformations of peptides representing the entire sequence of bovine pancreatic trypsin inhibitor and their roles in folding. J Mol Biol 234(3): 861–878 Kersteen EA, Raines RT (2003) Catalysis of protein folding by protein disulfide isomerase and small-molecule mimics. Antioxid Redox Signal 5(4):413–424 Klein-Seetharaman J et al (2002) Long-range interactions within a nonnative protein. Science 295(5560):1719–1722
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
253
Kobayashi T et al (1997) Respiratory chain is required to maintain oxidized states of the DsbA– DsbB disulfide bond formation system in aerobically growing Escherichia coli cells. Proc Natl Acad Sci USA 94(22):11857–11862 Kochoyan M, Keutmann HT, Weiss MA (1991) Architectural rules of the zinc-finger motif: comparative two-dimensional NMR studies of native and “aromatic-swap” domains define a “weakly polar switch”. Proc Natl Acad Sci USA 88(19):8455–8459 Kovacs GG et al (2002) Mutations of the prion protein gene phenotypic spectrum. J Neurol 249(11):1567–1582 Krishna SS, Majumdar I, Grishin NV (2003) Structural classification of zinc fingers: survey and summary. Nucleic Acids Res 31(2):532–550 Lipari G, Szabo A (1982a) Model-free approach to the interpretation of nuclear magnetic resonance relaxation in macromolecules. 1. Theory and range of validity. J Am Chem Soc 104(17):4546–4559 Lipari G, Szabo A (1982b) Model-free approach to the interpretation of nuclear magnetic resonance relaxation in macromolecules. 2. Analysis of experimental results. J Am Chem Soc 104(17):4559–4570 López García F et al (2000) NMR structure of the bovine prion protein. Proc Natl Acad Sci USA 97(15):8334–8339 Mamathambika BS, Bardwell JC (2008) Disulfide-linked protein folding pathways. Annu Rev Cell Dev Biol 24:211–235 Markley JL, Meadows DH, Jardetzky O (1967) Nuclear magnetic resonance studies of helix-coil transitions in polyamino acids. J Mol Biol 27:25–40 Marsh JA et al (2006) Sensitivity of secondary structure propensities to sequence differences between a- and g-synuclein: implications for fibrillation. Protein Sci 15:2795–2804 Mavridou DA et al (2007) Active-site properties of the oxidized and reduced C-terminal domain of DsbD obtained by NMR spectroscopy. J Mol Biol 370(4):643–658 Mavridou DA et al (2009) Control of periplasmic interdomain thiol:disulfide exchange in the transmembrane oxidoreductase DsbD. J Biol Chem 284(5):3219–3226 McDonald CC, Phillips WD (1969) Proton magnetic resonance spectra of proteins in random-coil configurations. J Am Chem Soc 91:1513–1521 Meiboom S, Gill D (1958) Modified spin-echo method for measuring nuclear relaxation times. Rev Sci Instrum 29(8):688–691 Merutka G, Dyson HJ, Wright PE (1995) “Random coil” 1H chemical shifts obtained as a function of temperature and trifluoroethanol concentration for the peptide series GGXGG. J Biomol NMR 5:14–24 Mishima T et al (2007) Analyses of native disulfide bond formations in the early stage of the folding process in mutant lysozymes where the long-range interactions in the denatured state were modulated. Biosci Biotechnol Biochem 71(8):2072–2074 Mobli M et al (2009) Direct visualization of disulfide bonds through diselenide proxies using 77Se NMR spectroscopy. Angew Chem Int Ed Engl 48(49):9312–9314 Nakamura A, Jardetzky O (1968) Systematic analysis of chemical shifts in the nuclear magnetic resonance spectra of peptide chains. II. Oligoglycines. Biochemistry 7:1226–1230 Noda Y et al (2002) NMR structural study of two-disulfide variant of hen lysozyme: 2SS[6–127, 30–115] – a disulfide intermediate with a partly unfolded structure. Biochemistry 41(7): 2130–2139 Ochsenbein F et al (2002) Dynamical characterization of residual and non-native structures in a partially folded protein by (15)N NMR relaxation using a model based on a distribution of correlation times. Protein Sci 11(4):957–964 Ohkuri T et al (2005) Effect of the structure of the denatured state of lysozyme on the aggregation reaction at the early stages of folding from the reduced form. J Mol Biol 347(1):159–168 Ohsaku M, Allinger NL (1988) Molecular structure and conformational stability of ethyl methyl disulfide: a model of cystine. J Phys Chem 92(16):4591–4594 Pabo CO, Peisach E, Grant RA (2001) Design and selection of novel Cys2His2 zinc finger proteins. Annu Rev Biochem 70:313–340
254
R. Silvers et al.
Palmer AG 3rd (2004) NMR characterization of the dynamics of biomacromolecules. Chem Rev 104(8):3623–3640 Pan JL, Bardwell JC (2006) The origami of thioredoxin-like folds. Protein Sci 15(10):2217–2227 Pan KM et al (1993) Conversion of alpha-helices into beta-sheets features in the formation of the scrapie prion proteins. Proc Natl Acad Sci USA 90(23):10962–10966 Pan H et al (1995) Extensive nonrandom structure in reduced and unfolded bovine pancreatic trypsin inhibitor. Biochemistry 34(43):13974–13981 Pan H, Barany G, Woodward C (1997) Reduced BPTI is collapsed. A pulsed field gradient NMR study of unfolded and partially folded bovine pancreatic trypsin inhibitor. Protein Sci 6(9):1985–1992 Parkin S, Rupp B, Hope H (1996) Structure of bovine pancreatic trypsin inhibitor at 125 K definition of carboxyl-terminal residues Gly57 and Ala58. Acta Crystallogr D Biol Crystallogr 52(Pt 1):18–29 Pastore A, Saudek V (1990) The relationship between chemical shift and secondary structure in proteins. J Magn Reson 90:165–176 Plaxco KW et al (1997) The effects of guanidine hydrochloride on the “random coil” conformations and NMR chemical shifts of the peptide series GGXGG. J Biomol NMR 10:221–230 Pollard MG, Travers KJ, Weissman JS (1998) Ero1p: a novel and ubiquitous protein with an essential role in oxidative protein folding in the endoplasmic reticulum. Mol Cell 1(2):171–182 Qian X, Weiss MA (1992) Two-dimensional NMR studies of the zinc finger motif: solution structures and dynamics of mutant ZFY domains containing aromatic substitutions in the hydrophobic core. Biochemistry 31(33):7463–7476 Radford SE et al (1991) A three-disulphide derivative of hen lysozyme. Structure, dynamics and stability. Biochem J 273(Pt 1):211–217 Reily MD, Thanabal V, Omecinsky DO (1992) Structure-induced carbon-13 chemical shifts: a sensitive measure of transient localized secondary structure in peptides. J Am Chem Soc 114:6251–6252 Richarz R, Wüthrich K (1978) Carbon-13 NMR chemical shifts of the common amino acid residues measured in aqueous solutions of the linear tetrapeptides H-Gly-Gly-X-L-Ala-OH. Biopolymers 17:2133–2141 Riek R et al (1996) NMR structure of the mouse prion protein domain PrP(121–321). Nature 382(6587):180–182 Salmon D et al (2006) Solution structure and backbone dynamics of the Trypanosoma cruzi cysteine protease inhibitor chagasin. J Mol Biol 357(5):1511–1521 Sauve S et al (2008) NMR assignment of the N-terminal TRAF-like RING zinc finger domain of human FLN29. Biomol NMR Assign 2(1):33–36 Schmiedeskamp M, Rajagopal P, Klevit RE (1997) NMR chemical shift perturbation mapping of DNA binding by a zinc-finger domain from the yeast transcription factor ADR1. Protein Sci 6(9):1835–1848 Schwalbe H et al (1997) Structural and dynamical properties of a denatured protein. Heteronuclear 3D NMR experiments and theoretical simulations of lysozyme in 8 M urea. Biochemistry 36(29):8977–8991 Schwalbe H et al (2001) A refined solution structure of hen lysozyme determined using residual dipolar coupling data. Protein Sci 10(4):677–688 Schwarzinger S et al (2000) Random coil chemical shifts in acidic 8 M urea: implementation of random coil shift data in NMRView. J Biomol NMR 18:43–48 Schwarzinger S et al (2001) Sequence-dependent correction of random coil NMR chemical shifts. J Am Chem Soc 123:2970–2978 Sevier CS, Kaiser CA (2002) Formation and transfer of disulphide bonds in living cells. Nat Rev Mol Cell Biol 3(11):836–847 Sevier CS et al (2001) A flavoprotein oxidase defines a new endoplasmic reticulum pathway for biosynthetic disulphide bond formation. Nat Cell Biol 3(10):874–882 Sharma D, Rajarathnam K (2000) 13C NMR chemical shifts can predict disulfide bond formation. J Biomol NMR 18(2):165–171
10 Disulfide Dynamics in Unfolded States of Proteins by NMR
255
Spera S, Bax A (1991) Empirical correlation between protein backbone conformation and Ca and Cb 13C nuclear magnetic resonance chemical shifts. J Am Chem Soc 113:5490–5492 Staley JP, Kim PS (1994) Formation of a native-like subdomain in a partially folded intermediate of bovine pancreatic trypsin inhibitor. Protein Sci 3(10):1822–1832 Szilágyi L (1995) Chemical shifts in proteins come of age. Prog Nucl Magn Reson Spectrosc 27:325–430 Szilágyi L, Jardetzky O (1989) a-Proton chemical shifts and secondary structure in proteins. J Magn Reson 83:441–449 Thanabal V et al (1994) The 13C chemical shifts of amino acids in aqueous solution containing organic solvents: application to the secondary structure characterization of peptides in aqueous trifluoroethanol solution. J Biomol NMR 4:47–59 Tu BP et al (2000) Biochemical basis of oxidative protein folding in the endoplasmic reticulum. Science 290(5496):1571–1574 van den Berg B et al (1999) Characterisation of the dominant oxidative folding intermediate of hen lysozyme. J Mol Biol 290(3):781–796 van Mierlo CP et al (1991a) Two-dimensional 1H nuclear magnetic resonance study of the (5–55) single-disulphide folding intermediate of bovine pancreatic trypsin inhibitor. J Mol Biol 222(2):373–390 van Mierlo CP et al (1991b) (14–38, 30–51) double-disulphide intermediate in folding of bovine pancreatic trypsin inhibitor: a two-dimensional 1H nuclear magnetic resonance study. J Mol Biol 222(2):353–371 van Mierlo CP et al (1993) Partially folded conformation of the (30–51) intermediate in the disulphide folding pathway of bovine pancreatic trypsin inhibitor. 1H and 15N resonance assignments and determination of backbone dynamics from 15N relaxation measurements. J Mol Biol 229(4):1125–1146 van Mierlo CP et al (1994) 1H NMR analysis of the partly-folded non-native two-disulphide intermediates (30–51,5–14) and (30–51,5–38) in the folding pathway of bovine pancreatic trypsin inhibitor. J Mol Biol 235(3):1044–1061 Wagner G (1993) NMR relaxation and protein mobility. Curr Opin Struct Biol 3(5):748–754 Wang Y, Jardetzky O (2002) Probability-based protein secondary structure identification using combined NMR chemical-shift data. Protein Sci 11:852–861 Wang L et al (2005) Linear analysis of carbon-13 chemical shift differences and its application to the detection and correction of errors in referencing and spin system identifications. J Biomol NMR 32:13–22 Weissman JS, Kim PS (1991) Reexamination of the folding of BPTI: predominance of native intermediates. Science 253(5026):1386–1393 Wilkins DK et al (1999) Hydrodynamic radii of native and denatured proteins measured by pulse field gradient NMR techniques. Biochemistry 38(50):16424–16431 Williamson MP (1990) Secondary-structure dependent chemical shifts in proteins. Biopolymers 29:1423–1431 Wirmer J et al (2004) Modulation of compactness and long-range interactions of unfolded lysozyme by single point mutations. Angew Chem Int Ed Engl 43(43):5780–5785 Wishart DS, Sykes BD (1994a) Chemical shifts as a tool for structure determination. Meth Enzymol 239:363–392 Wishart DS, Sykes BD (1994b) The 13C chemical-shift index: a simple method for the identification of protein secondary structure using 13C chemical-shift data. J Biomol NMR 4:171–180 Wishart DS, Sykes BD, Richards FM (1991a) Relationship between nuclear magnetic resonance chemical shift and protein secondary structure. J Mol Biol 222:311–333 Wishart DS, Sykes BD, Richards FM (1991b) Simple techniques for the quantification of protein secondary structure by 1H NMR spectroscopy. FEBS Lett 293:72–80 Wishart DS, Sykes BD, Richards FM (1992) The chemical shift index: a fast and simple method for the assignment of protein secondary structure through NMR spectroscopy. Biochemistry 31:1647–1651
256
R. Silvers et al.
Wishart DS et al (1995) 1H, 13C and 15N random coil NMR chemical shifts of the common amino acids. I. Investigation of nearest-neighbor effects. J Biomol NMR 5:67–81 Wu DH, Chen AD, Johnson CS (1995) An improved diffusion-ordered spectroscopy experiment incorporating bipolar-gradient pulses. J Magn Reson Ser A 115(2):260–264 Xie T et al (2002) Identification of the ubiquinone-binding domain in the disulfide catalyst disulfide bond protein B. J Biol Chem 277(3):1649–1652 Yokota A et al (2004) NMR characterization of three-disulfide variants of lysozyme, C64A/C80A, C76A/C94A, and C30A/C115A – a marginally stable state in folded proteins. Biochemistry 43(21):6663–6669 Zahn R et al (2000) NMR solution structure of the human prion protein. Proc Natl Acad Sci USA 97(1):145–150 Zapun A, Bardwell JC, Creighton TE (1993) The reactive and destabilizing disulfide bond of DsbA, a protein required for protein disulfide bond formation in vivo. Biochemistry 32(19):5083–5092 Zhang T, Bertelsen E, Alber T (1994) Entropic effects of disulphide bonds on protein stability. Nat Struct Biol 1(7):434–438
Chapter 11
A Half-Century of Oxidative Folding and Protein Disulphide Formation Robert B. Freedman
Abstract The preceding chapters in this volume are a dramatic demonstration of the extent and sophistication of our current understanding of the complex process of protein oxidative folding – a process of much chemical, biotechnological and biomedical interest, which has now been extensively characterised both in vitro and in cellulo. Of course there are many aspects about which we are still ignorant and many questions to which we require answers. But by comparison with the state of knowledge previously – even 20 years ago – we must recognise that this is now a mature field with an established body of knowledge and some reliable techniques and experimental approaches. In this chapter, the half-century from 1961 to 2010 is divided into decades and the key developments within each decade are assessed. The preceding chapters in this volume are a dramatic demonstration of the extent and sophistication of our current understanding of the complex process of protein oxidative folding – a process of much chemical, biotechnological and biomedical interest, which has now been extensively characterised both in vitro and in cellulo. These chapters highlight how the covalent processes of disulphide formation and isomerisation are coupled to the non-covalent process of conformational folding, how diverse the detailed pathways involved are for different proteins and how these processes are catalysed and mediated in eukaryotic cells and in Escherichia coli. Cumulatively, the chapters emphasise the impact of recent technical developments – in separation technology, in mass spectrometry and in high resolution structural techniques – and the extent to which the wealth of knowledge now available and the
R.B. Freedman (*) School of Life Sciences, University of Warwick, Coventry, CV4 7AL, UK e-mail:
[email protected] R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6_11, © Springer Science+Business Media, LLC 2011
257
258
R.B. Freedman
range of proteins now characterised provide a base of understanding in this field which is both wide and deep. Of course there are many aspects about which we are still ignorant and many questions to which we require answers. We lack a real picture of the dynamics and flexibility of intermediates along oxidative folding pathways, and we have no real picture of the dynamic interactions between such intermediates and cellular catalysts of oxidative folding in the bacterial periplasm and the eukaryotic endoplasmic reticulum. We know very little about the newly discovered systems for protein disulphide-bond formation in mitochondria (Riemer et al. 2009). Our efforts to express high-value recombinant disulphide-bonded proteins for pharmaceutical applications are still highly empirical; we cannot predict in advance which expression system or which in vitro folding system is preferred on grounds of yield and cost. But by comparison with the state of knowledge previously – even 20 years ago – we must recognise that this is now a mature field with an established body of knowledge and some reliable techniques and experimental approaches. This maturity is fully documented in the previous chapters in this volume. This chapter, by contrast, aims to sketch out the route by which we have reached this position and to highlight some of the key landmarks along that route. Such a historical perspective is timely because the field celebrates its 50th birthday in 2011. Although there are no absolute beginnings in science, there is no doubt that the key successful work which first established this field came into the public domain in 1960 and 1961 (White 1960, 1961; Anfinsen and Haber 1961; Anfinsen et al. 1961; Bello et al. 1961). In this chapter, the half-century from 1961 to 2010 is divided into decades and the key developments within each decade are assessed.
11.1 Oxidative Protein Folding as a Test of the Reversibility of Denaturation (1961–1970) The work that is often termed “the Anfinsen experiment” arose from a programme of study at NIH from approximately 1957 on the reduction, reoxidation and reactivation of bovine ribonuclease. The technical context of the work was the development of tools for protein chemistry and protein sequencing, specifically the search for chemical reagents that would reliably reduce protein disulphides to generate separate un-cross-linked polypeptide chains, without causing any other covalent change. The conceptual context was the debate about the relationship between a protein’s primary, secondary and tertiary structure and its biological activity, and specifically how the tertiary fold and biological activity were specified (at a time before the nature of the genetic code had been established). Thus, White (1961) introduced his classic paper with the statement “Although there is an abundance of evidence in support of the template hypothesis of protein biosynthesis …. there appears to be no satisfactory experimental basis for deciding whether the role of the template includes the coiling and folding of the protein chain to produce the secondary and tertiary structure … or is restricted to the formation of the amino acid sequence”.
11 A Half-Century in Review
259
Early results from this programme (cited in White 1960) showed some regain of ribonuclease activity after air oxidation of reduced denatured ribonuclease, but the yields were <20% and were dependent on the details of the reduction protocol. Refinements of reagents and protocols led to yields of active material after oxidative refolding of up to 80% (White 1960) or 94% (Anfinsen and Haber 1961). The latter authors made the prophetic observations that protein precipitation competed with reoxidation, and that the yield of soluble product and of active product were both strongly and inversely dependent on the concentration of protein in the reoxidation process. With these improved protocols, it was possible both to isolate the reduced/ reoxidised material for physical characterisation and to monitor kinetics. White (1961) established that the active product of reduction and reoxidation was identical to native ribonuclease by UV spectroscopy, optical rotatory dispersion (ORD) and viscosity and in chromatographic and immunological behaviour, and generated peptide mapping data which implied formation of the “native” four disulphide bonds, while Bello et al. (1961) crystallised the reduced, reoxidised active material and confirmed that its lattice constants were identical to that of native ribonuclease. Anfinsen et al. (1961) showed that during air oxidation, the loss of free protein –SH groups and change in secondary structure (monitored by ORD) preceded the appearance of enzymatic activity, implying the existence of partly oxidised, partly folded inactive states. Previous work had established that reduced denatured ribonuclease approximated a random coil polymer and that “random” reoxidation would result in 0.95% recovery of activity, assuming that activity required fully and correctly disulphide-bonded material; so regeneration of almost 100% activity was very striking. And in a prescient footnote Anfinsen et al. (1961) noted that “Experimental work is now in progress to determine the nature of the pairing of half-cystine residues at various times during the early stages of reoxidation; preliminary results indicate that pairing is quite random and that ‘incorrectly’ formed bonds are present”. Within the following few years, the significance of these results was explored and comparable data were obtained for a range of other proteins. A highly cited review (Epstein et al. 1963) summarised data reporting recoveries of activity of >50% from oxidative refolding studies on lysozyme, taka-amylase, alkaline phosphatase, pepsinogen and human serum albumin, and reported emerging generalisations about optimal conditions (preferably, alkaline pH and low protein concentrations). This work had impact far wider than on the protein chemistry community undertaking it, because it appeared as the most fundamental demonstration of “self-assembly” in biology, in that a “biological” functionality such as enzyme activity was clearly specified at the chemical level of protein primary structure. Given that this work was contemporary with the establishment of the genetic code, this initial work on oxidative folding and disulphide formation formed a key element in the consensus on the core principles of molecular biology. Indeed this was clear from the title (“The Genetic Control of Tertiary Protein Structure: Studies with Model Systems”) of the review of this work by Anfinsen and colleagues (Epstein et al. 1963). This review stated the key principle established by this body of work as follows: “The following concepts now appear to be well-established. Folding of newly-synthesized proteins in vivo is a rapid and highly efficient process…. All genetic information for proper folding is contained within the amino acid sequence.” (ibid p. 447).
260
R.B. Freedman
However, the acceptance of this general principle of self-assembly did not mean that all proteins could – in practice – be oxidatively refolded in good yield from reduced polypeptides. Proteins comprising several polypeptides with inter-chain disulphides provided a particular challenge. Insulin was a key example and a major focus of interest at the time; groups both in China and the West had succeeded in regenerating insulin activity by reoxidation of reduced A and B chains, but the yields were in the range 5–10%, close to what might be expected at random (see Epstein et al. (1963) for original references). Low yields were also observed for the recovery of specific antigen binding after reoxidation of reduced antibodies or Fab fragments, but in this case the expected yield from “random” recombination of Cys residues was so low that any regain of activity was remarkable; within a few years, several groups had described oxidative refolding of antibodies or Fab fragments and reported yields as high as 20–30% (of recovered soluble protein) (Haber 1964; Whitney and Tanford 1965; Freedman and Sela 1966). The majority of the work in these early years used “air oxidation,” oxidation by dissolved O2 species, probably catalysed by trace levels of transition metal cations (Yutani et al. 1968). This was poorly reproducible, required unphysiologically high pH and in general the rates observed were very low, with half-times of hours rather than minutes. This prompted two independent developments both of which had long-term significance – search for enzymic catalysts of oxidative folding, and search for improved chemical oxidants to drive protein disulphide formation. The low rates for recovery of ribonuclease activity in the classic experiments prompted search for potential enzymic catalysts, both at NIH and elsewhere, and the first descriptions of such activity were published in early 1963 (Goldberger et al. 1963; Venetianer and Straub 1963); these papers reported detection of “ribonuclease reactivation” activity in microsomal fractions from liver and pancreas tissue. Almost immediately a partial purification of the enzyme was reported and it was shown that its key activity was not net oxidation but isomerisation of protein disulphides; in the presence of dehydroascorbate as oxidant, the partially purified liver extract catalysed the reactivation of reduced ribonuclease but did not affect the much more rapid rate of loss of free thiol groups (Givol et al. 1964). Further purification enabled work on the mechanism of action of the purified “disulphide interchange enzyme” showing that it contained an essential thiol group in the sequence Cys(Gly,His) which was required to be in the reduced –SH form for activity in disulphide isomerisation (de Lorenzo et al. 1966; Fuchs et al. 1967). From a current perspective, it is interesting that the local dithiol/disulphide motif was not detected and the presence of two such redox-active dithiol/disulphide groups was missed because the enzyme’s molecular weight was estimated as ca. 42 kDa and hence the quantitation of reactive Cys residues per polypeptide was underestimated. Givol et al. (1964) described the ability of their partially purified enzyme preparation to convert “scrambled ribonuclease” to native active ribonuclease and the authors concluded that their result “…favours the hypothesis that the enzyme is a general and non-specific catalyst for disulfide interchange in proteins containing disulfide bonds.” Later the same group supported this conclusion by studies of the action of “disulphide interchange enzyme” on other proteins – chymotrypsin, chymotrypsinogen
11 A Half-Century in Review
261
and insulin – and from its action on insulin concluded that insulin must be derived, like chymotrypsin, by proteolysis of a precursor (Givol et al. 1965). (This was prior to the characterisation of proinsulin in 1967; later it was shown that – in contrast to insulin – proinsulin can be reduced/denatured and oxidatively refolded in relatively high yield (Steiner and Clark 1968)). The data all supported the view that the native disulphides of the immediate biosynthetically produced protein defined a thermodynamically preferred state, but that later proteolysis could trap these as metastable disulphides and this conclusion was further supported by work on RNase S the product of subtilisin treatment of ribonuclease (Kato and Anfinsen 1969). (For a fuller review of early work on PDI see Freedman and Hillson 1980). The other response to the inadequate rates observed in early work was to seek alternative chemical oxidants for oxidative refolding of reduced proteins, and the key work here was a careful study of the use of thiol/disulphide mixtures published at the end of the decade (Saxena and Wetlaufer 1970). This paper studied the oxidative refolding of reduced lysozyme using cysteine/cystine, cysteamine/cystamine and, especially, GSH/GSSG as thiol/disulphide redox agents, showing great improvements in rate and yield over those obtained with air oxidation, i.e. halftimes for reactivation were <5 min, provided pH was >7.5. Concentrations and ratios of GSH and GSSG were varied and a mix of 5 mM GSH + 0.5 mM GSSG was defined as optimal. The process was not inhibited by EDTA, confirming that oxidation in these conditions did not depend on trace metal cations. The long-term influence of this paper has been substantial; it established thiol/disulphide buffers as reproducible agents for protein disulphide regeneration and isomerisation, and confirmed that both reducing and oxidising components were required to facilitate the thiol/disulphide interchanges involved in all key steps. But the paper also claimed, more tendentiously, that since this purely chemical system – using GSH/ GSSG concentrations close to the total cellular concentrations of these species – generated oxidative refolding rates that were comparable to those for biosynthetic oxidative folding, then no enzymic catalysis (of the kind described in the previous paragraph) was required. Intriguingly, a footnote to the paper shows that a referee for the paper asked the authors to reconsider this conclusion in light of the fact that disulphide formation occurred in a compartmentalised situation such as the cisternae of the endoplasmic reticulum, but the authors response was that they “regard the foregoing comments about possible compartmentalization of the intracellular oxidoreduction systems as … unproven.”
11.2 Characterisation of Oxidative Refolding Pathways: Confusion over Enzymic Involvement (1971–1980) Several papers studying protein oxidative refolding in vitro during the 1960s, made attempts to establish the relationships between what we would now term conformational folding, overall thiol oxidation and regain of enzymic activity. In general, they observed relatively rapid thiol oxidation and relatively slow regain of activity,
262
R.B. Freedman
but the kinetics of conformational folding were less clearly established, due to the limitations of the methods available. Early in the 1970s, Scheraga’s group revisited this issue in relation to the oxidative refolding of ribonuclease, using a GSH/GSSG redox system, and employing absorbance and fluorescence spectroscopy to monitor conformational folding, with parallel determinations of enzyme activity and free thiol groups (Hantgan et al. 1974). Under their conditions, the protein thiol titre dropped from 8 –SH/mol to 2.7 –SH/mol within 1 min and to 1.2 –SH/mol within 5 min, while further oxidation was very slow, as was regain of activity, which occurred with a half-time of 80–100 min. Conformational changes determined by spectroscopy involved a “rapid” phase with rate constant of the order of 0.1–0.2 min−1, (considerably slower than the fast rate of protein –SH oxidation) and a slow phase with rate constant of 5–10 × 10−3 min−1 (a rate comparable to that for regain of activity). This implied a complex relationship between folding and disulphide formation and these authors found comparable complexity when attempting to identify chemically the disulphide bonds formed at intermediate stages of refolding; the results indicated the initial formation of many non-native disulphide pairings leading to a model for oxidative refolding involving formation of protein-SSG mixed disulphides “…followed by rapid formation of a large number of different inactive species with incorrectly paired cysteine residues…” which “…then undergo a conformational change and glutathione-catalysed disulfide exchange reactions which ultimately lead to the native conformation”. This was not a revolutionary conclusion; the paper provided more solid and comprehensive experimental evidence for a model that had been around for a decade. The key advance in the field in this period came from the introduction of a new model protein, bovine pancreatic trypsin inhibitor (BPTI), and from the introduction of quenching by iodoacetate to convert free Cys residues to the negatively charged carboxymethyl-Cys form (Creighton 1974a, b). Compared to the four disulphides of the previous preferred model proteins (ribonuclease and lysozyme), BPTI contains only three disulphides, which simplified the disulphide mapping challenge; and the use of iodoacetate enabled electrophoretic or ion-exchange separation of pools of quenched intermediates containing one or two disulphides, from fully reduced and fully re-oxidised BPTI. In a series of striking papers, Creighton exploited this system to provide the most detailed picture to date of an oxidative refolding pathway (Creighton 1975) and, more controversially, to demonstrate that general issues in protein folding could be illuminated by studies of folding associated with disulphide formation. The ability to resolve oxidative refolding intermediates enabled Creighton to show that the “one-disulphide” intermediate pool was dominated by molecules containing the Cys30–Cys51 “native” disulphide (Creighton 1974b), whereas the “two-disulphide” intermediate pool contained four distinct species, all of which contained this Cys30–Cys51 disulphide. However, while two of these species contained a second “native” disulphide, the other two contained a “non-native” disulphide (Creighton 1975); furthermore, only one of the species with two native disulphides (des-[14–38]-BPTI or N(SH)2) could convert directly to active BPTI with three “native” disulphides. Hence Creighton concluded “…that
11 A Half-Century in Review
263
there is a unique and obligatory pathway for refolding. This pathway may be approached rapidly from all disordered states normally present, and some divergence from it is possible, but all molecules must pass through certain obligatory intermediate states.” Further kinetic characterisation of the pathway (summarised in Creighton (1979)) confirmed the summary picture
R « I « II « N (SH )2 « N
and showed that the rate-limiting step of the pathway in both directions (reductive unfolding and oxidative refolding) was the interconversion at the two-disulphide level between N(SH)2 and the mix of species within II. At this stage, conformational characterisation of the intermediates was limited (e.g. Kosen et al. 1980), but studies of the effects of denaturants allowed the derivation of an energy profile showing that all intermediates were metastable at equilibrium and implying that only N(SH)2 was significantly stabilised by native folding. The impact of this work was immediate (see, e.g. Baldwin (1978)) and it provided a model and set a standard for subsequent work. During this period, comparable work confirmed the large number and complexity of the disulphide intermediates in the ribonuclease oxidative pathway (Creighton 1980; Konishi et al. 1981), and showed the difficulties in characterising the pathway for lysozyme (e.g. Acharya and Taniuchi 1977). Work on PDI during the 1970s progressed slowly (see, e.g. Freedman and Hawkins 1977; Freedman and Hillson 1980). Some work attempted to define more precisely its sub-cellular location and in vivo specificity, without achieving definitive results, and this left the way clear for proposals that other enzymes might be responsible for cellular protein disulphide formation (Ziegler and Poulsen 1977; Janolino et al. 1978). There was also a frustrating controversy about the relationship between PDI – defined by its activity in reactivation of reduced or scrambled ribonuclease – and less specific thiol:disulphide oxidoreductase activities – defined, e.g. by catalysis of GSH:insulin oxidoreduction (see, e.g. Chandler and Varandani 1975). The situation was resolved by quantitative analyses of purifications (Hillson and Freedman 1980) and we now understand that while PDI has both activities, many protein families including thioredoxins and glutaredoxins have thiol:disulphide oxidoreductase activities but limited capability to catalyse the conformationally linked protein disulphide isomerisation steps in oxidative folding pathways. At the end of the decade, the ability of purified PDI to catalyse the well-defined oxidative folding pathway of BPTI was explored for the first time and this work extended the picture of PDI’s action. It showed that, with PDI at sub-stoichiometric levels relative to BPTI, the intermediates in the BPTI oxidative folding pathway were not altered but that the rate of every step involving both conformational change and a change in disulphide status was significantly increased. Hence “The protein-disulphide isomerase appears to be a true catalyst of protein unfolding and refolding involving disulphide bond breakage, formation or rearrangement.” (Creighton et al. 1980).
264
R.B. Freedman
11.3 Confirmation of the Cellular Role of PDI and the Impact of Recombinant DNA Technology (1981–1990) Evidence for PDI’s involvement in cellular folding of disulphide-containing proteins was limited through the 1970s, with few convincing studies of the enzyme’s tissue distribution, no data on its developmental properties and rather sketchy understanding of the sub-cellular location of disulphide formation (Freedman and Hawkins 1977). But these deficiencies were repaired in the following decade by studies on PDI activity through chick embryo development, through wheat seed development and in a series of lymphoid cells, establishing clear correlations between PDI activity in a cell and its rate of production of disulphide-bonded extracellular proteins (see Freedman 1984 for primary references). Furthermore, the rapid burst of work on cellfree translation/translocation systems following the establishment of the “signal hypothesis” in the late 1970s clarified the timing and location of cellular disulphide formation, showing that it occurred within the ER lumen as a co-translational or very early post-translational step (Freedman and Hillson 1980). By the mid-1980s, previous work on the latency and detergent sensitivity of PDI activity in microsomal membrane systems had been rationalised by clear evidence that PDI was essentially a free soluble component trapped within microsomal vesicles, and hence a long-term resident of the ER lumen in cellulo (Lambert and Freedman 1985). The application of the newly developed methodology for two-dimensional polyacrylamide gel electrophoresis established that PDI was one of the most abundant of a small number of conserved, acidic proteins found as luminal components in the ER of all secretory cells (Mills et al. 1983; Kaderbhai and Austen 1984). From the mid-1980s, the cellular role of PDI was clearly established, and was further confirmed by evidence that PDI could be cross-linked to nascent secretory proteins in living cells (Roth and Pierce 1987), and that depletion of PDI from cellfree microsomal translocation systems deprived them of the ability to catalyse rapid co- and post-translational formation of protein disulphides (Bulleid and Freedman 1988). The late 1980s corresponded to the period in which the general concept of “molecular chaperones” as key agents in cellular protein folding and assembly became firmly established and this contributed to PDI finally being recognised as an essential chaperone and catalyst of the folding of secretory proteins in the ER lumen. Remarkably, considering its early discovery, the role of PDI had become obscured; although Anfinsen and his group had clearly understood that a cellular catalyst was required, the majority of bioscientists had misinterpreted the message of Anfinsen’s work as being that “proteins know how to fold themselves up” and inferred from this that no cellular assistance was required. This misunderstanding was challenged – and work on protein oxidative refolding was given a boost – by results from the nascent biotechnology industry of the period, which was strongly focussed on the production of high-value recombinant mammalian proteins, following their expression in E. coli. To the surprise of the molecular biologists who established this industry, the secretory mammalian proteins that were their initial targets (such as insulin, interferons and digestive enzymes) generated inclusion bodies in the bacterial cytoplasm and so it became essential for
11 A Half-Century in Review
265
the industry to learn how to generate active proteins from this source by reduction/ denaturation and oxidative refolding (see, e.g. Marston et al. 1984; Marston 1986; Kohno et al. 1990). Experience of expressing recombinant proteins in bacteria also prompted study of how and where disulphide bonds were generated in authentic bacterial proteins, which focussed attention on the periplasm as a more oxidative compartment than the cytoplasm (Talmadge and Gilbert 1982; Pollitt and Zalkin 1983). A further important outcome of the early biotechnology industry was interest in exploiting novel protein disulphide bonds to stabilise valuable proteins, drawing on structural data and the new capability of site-directed mutagenesis to engineer non-natural disulphides into appropriate sites; this approach generated some useful outcomes, among a number of surprises (Wetzel 1987). Another consequence of the challenges of expressing native mammalian proteins in bacteria was acute interest in cloning and sequencing mammalian PDI itself for potential application in refolding protocols. Rat PDI cDNA was first cloned and sequenced by Rutter and colleagues (Edman et al. 1985) with results that were immediately informative and interesting; the inferred sequence of rat PDI showed the presence of a standard secretory signal sequence, plus a pattern of internal sequence repetitions that implied the presence of four domains in an a–b–b¢–a¢ pattern and also revealed an extended thioredoxin-like sequence motif within both the a and a¢ domains, containing the dithiol sequence –WCGHCK–. Not immediately noted was another significant sequence feature; the C terminus of rat PDI had the sequence –KDEL, and the discovery of this sequence at the C terminus of another abundant luminal protein, BiP, soon led to its recognition as an “ER-retention” signal (Munro and Pelham 1987). This was part of a wider realisation that the ER lumen was a key sub-cellular compartment in its own right, with its own activities and set of permanent resident proteins, rather than a passive passage to the cell exterior, as had been pictured previously. By the end of the decade cDNAs had been sequenced that encoded proteins closely related in sequence to PDI (Mazzarella et al. 1990), and the potential existence of a PDI family was apparent. Furthermore, the gene for human PDI had been identified and cloned, its intron/exon pattern established (Tasanen et al. 1988) and a further very unexpected discovery had been made, namely that PDI was multifunctional and existed both as free PDI and as a component of other ER luminal activities. It was found as a subunit of the procollagen-modifying enzyme prolyl-4-hydroxylase (Pihlajaniemi et al. 1987) and, later, of the microsomal triglyceride transfer protein (Wetterau et al. 1990).
11.4 Improved Characterisation of Oxidative Folding Pathways, Intermediates and Catalysts (1991–2000) In the early 1990s, the pathway of oxidative folding of BPTI briefly became highly controversial. New methodologies – the use of acidification rather than alkylation for trapping intermediates, and hplc rather than ion-exchange chromatography for
266
R.B. Freedman
resolving them – led to revised quantitation of the previously identified disulphide intermediates in the pathway (Weissman and Kim 1991). The kinetic pathway outlined above was modified slightly, with the addition of a further – very slowly re-arranging – two disulphide intermediate (des-[30–51]-BPTI or N*), and with a significant quantitative revision in the composition of the pool of two-disulphide intermediates (II) that require disulphide isomerisation to generate the productive two-disulphide intermediate N(SH)2. But this important result did not revise the key features of the previously established BPTI oxidative refolding process. As summarised by Goldenberg (1992), in a contemporary review that is a model of balance and clarity, the new findings confirmed the non-randomness of the pathway (in that only a small fraction of potential intermediates are detectable), confirmed what those intermediates are and confirmed that the process is not simply the sequential acquisition of native disulphides but occurs via disulphide bond rearrangements. This rebirth of interest in oxidative folding pathways was partly stimulated by structural studies on oxidative folding intermediates, or on analogues of such intermediates with Cys residues substituted by Ala or Ser to exclude the possibility of forming specific disulphides. In the case of BPTI, this work established that the 30–51 one-disulphide intermediate was partially folded, with considerable nativelike local structure (van Mierlo et al. 1992, 1993) that intermediates corresponding to components of the two-disulphide pool II (30–51, 5–14 and 30–51, 5–38), were partly folded but had not acquired any additional structure compared to the 30–51 species (van Mierlo et al. 1994), whereas the final intermediate des-[14–38]-BPTI, formed in the rate-determining step of folding, is essentially native-like, in that its variation from the structure of native BPTI is no greater than the variability of native BPTI in three different crystal forms (Eigenbrot et al. 1990). Hence the restricted pathway of oxidative folding of BPTI is partly the result of stabilisation associated with the acquisition of native structure along with the native disulphide 30–51, and partly due to the fact that this early formation of structure limits the potential for facile disulphide formation and exchange at later stages (Goldenberg 1992). Work on other model proteins proceeded along similar lines so that, before long, structural data were available for key disulphide-bonded intermediates in the oxidative folding pathways of a-lactalbumin (Ewbank and Creighton 1993), ribonuclease (Talluri et al. 1994; Laity et al. 1997) and lysozyme (van den Berg et al. 1999a) and the now established methodologies were being extended to determine oxidative folding pathways of less familiar proteins (e.g. Chang et al. 1994). By the end of the decade – driven by advances in mass spectrometry, nuclear magnetic resonance and calorimetry – chemical identification of disulphides, characterisation of intermediates and studies on their stability had all advanced to the point where consensus models of pathways could be presented, with kinetic, thermodynamic and structural detail (Narayan et al. 2000; Wedemeyer et al. 2000). This deeper structural understanding of oxidative folding pathways allowed a re-examination of how PDI operates as catalyst of oxidative folding. Creighton et al. (1993) showed that in conditions mimicking those in the ER, PDI catalysed the
11 A Half-Century in Review
267
oxidative folding of mature BPTI, and its cellular precursor pro-BPTI, with halftimes as low as 2 min, comparable to those observed in co- and post-translational folding in microsomes; the key impact of PDI on BPTI and pro-BPTI folding was to accelerate the conformationally restricted disulphide isomerisations at the twodisulphide stage. Similarly, Weissman and Kim (1993) characterised the action of PDI on isolated two-disulphide intermediates and showed that it accelerated their conversion to the native oxidised material by a factor of 3,000–6,000. Subsequently, similar data were generated for other model proteins. In the case of lysozyme, PDI improved both yields and rates of oxidative refolding, mainly by facilitating the conversion to the native state of the conformationally constrained species des-[76– 94]-lysozyme (van den Berg et al. 1999b). In the case of ribonuclease, the effect of PDI is to accelerate all the disulphide formation steps, including the rate-limiting rearrangement steps in which the native-like intermediates des-[65–72]-ribonuclease and des-[40–95]-ribonuclease are formed from the pool of unstructured threedisulphide intermediates (Shin and Scheraga 2000). In every case, PDI was found not to change the overall pathway or folding mechanism but to accelerate all the steps in which disulphide bonds are formed or isomerised, especially where this requires significant conformational change in the substrate protein. In the absence, at this stage, of a high resolution structure of PDI, it was not simple to develop a mechanistic picture of how PDI achieved this remarkable activity (Freedman et al. 1994). It was clear that PDI comprised a series of domains a–b–b¢–a¢ (Edman et al. 1985), and when the N-terminal a domain was found to be structurally homologous to bacterial thioredoxin (Kemmink et al. 1996), this was no surprise. But when the b domain showed the same overall fold (Kemmink et al. 1997), this was entirely unexpected. The isolated a and a¢ domains were studied and found to be active in simple thiol:disulphide oxidoreductions but they did not reproduce the full activity of PDI (Darby and Creighton 1995). In fact the multi-domain structure of PDI was shown to be essential for high activity in its distinctive catalytic action in rearranging disulphide bonds within partly folded protein substrates, such as the BPTI folding intermediates; analysis of the activities of various constructs implicated the b¢ domain as especially significant (Darby et al. 1998). In parallel, it became clear that PDI had a specific binding site for peptides and unfolded proteins (Klappa et al. 1997) and that the b¢ domain provided the core of this site (Klappa et al. 1998). This finding rationalised the discovery that PDI had non-catalytic chaperone activity towards proteins lacking disulphide bonds (Wang and Tsou 1993), the earlier finding of PDI as a component of stable functional complexes (see above) and the observation that PDI could demonstrate active-site-independent chaperone activity in vivo (McLaughlin and Bulleid 1998). It pointed towards a model of PDI action in which the domains operated synergically – some domains catalysing disulphide interchange reactions in protein substrates while other domains stabilised those substrates in partly unfolded conformations, so coupling disulphide rearrangement and conformational change (Freedman et al. 2002).
268
R.B. Freedman
11.5 Impact of Genetic and Genomic Approaches: The PDI Family and the Thioredoxin Superfamily (1991–2000) Up to the 1980s, discussion about oxidative folding in vivo had mainly been confined to the situation in mammalian tissues or other higher organisms and so genetic approaches had not been available. This changed abruptly in 1991. In that year, five groups – more or less simultaneously – cloned PDI from the yeast Saccharomyces cerevisiae and showed that it was an essential gene in that organism (see below), and the first steps were taken in characterising the system for formation of protein disulphide bonds in bacterial proteins, with the cloning and characterisation of the dsbA gene (Bardwell et al. 1991). The dsbA gene is essential for formation of disulphides in periplasmic proteins and encodes a protein with an obvious redox-active dithiol/disulphide but little other sequence homology to PDI or thioredoxin. The DsbA protein was rapidly crystallised (Martin et al. 1993), revealing a protein with a thioredoxin-fold domain with the active-site at the same position as that in thioredoxin, plus a novel helical domain insert. This discovery provided an intriguing series of proteins with related structures and related redox-active dithiol/disulphide active sites for comparative structure/function and mutagenesis studies – the strongly reducing thioredoxin (−CGPC-), the strongly oxidising DsbA (−CPHC-) and PDI (−CGHC-) with intermediate redox properties (Martin 1995). Throughout the decade, the roles of particular residues around the redox-active site and elsewhere in these proteins were explored vigorously to rationalise the structural and electrostatic basis for their reactivities, pKa values and redox properties (Chivers et al. 1997; Huber-Wunderlich and Glockshuber 1998). The fact that DsbA contains a redox-active dithiol/disulphide active site within a thioredoxin-like domain initially led to the inference that DsbA was the bacterial analogue of PDI (Bardwell and Beckwith 1993) but detailed study of its catalytic activities in vitro showed that DsbA was very effective as an oxidant but very ineffective in catalysing disulphide isomerisations (Zapun and Creighton 1994). This issue was soon resolved by genetic and physiological studies which led to the identification of the dsbB, dsbC, dsbD, etc. genes and characterisation of the complete electron-transfer pathway responsible for disulphide formation and isomerisation in the bacterial periplasm (Bardwell 1994; Rietsch et al. 1996; Rietsch and Beckwith 1998; Bader et al. 2000). DsbC was identified as the major disulphide-isomerase in this system (Bardwell 1994; Zapun et al. 1995) leading to a model in which different proteins are employed to form and then to isomerise disulphide bonds in bacterial secretory proteins. That PDI was by this stage a high-profile research area was demonstrated by the fact that five independent groups reported the cloning of yeast PDI within a single year (LaMantia et al. 1991; Tachikawa et al. 1991; Gunther et al. 1991; Farquhar et al. 1991; Scherens et al. 1991). The PDI1 gene is essential for yeast viability and encodes a protein with a similar overall sequence and organisation to mammalian PDI, but with only 30–32% sequence identity. Subsequent work identified the EUG1
11 A Half-Century in Review
269
gene which encodes an interesting homologue, with active site sequences –CLHS– and –CIHS–, whose overexpression allows yeast to survive and grow in the absence of a functional PDI1 gene (Tachibana and Stevens 1992). Later, the complete sequencing of the S. cerevisiae genome showed that this organism contains a family of five genes encoding proteins which resemble PDI in being ER-located and having thioredoxin-like active site sequences; of these only the PDI1 gene is essential, and the functions and roles of the other family members have not been simple to establish (Norgaard et al. 2001). Parallel genomic work revealed that mammalian species encode a large PDIfamily of related ER-located proteins containing thioredoxin-like sequences and domains, and that the members of this family do not closely resemble the members of the yeast PDI family (Ferrari and Soling 1999). Recently, the functions of some of these mammalian PDI family members are becoming clearer (Ellgaard and Ruddock 2005), but within the 1990s, the only mammalian PDI homologue to be well characterised in terms of function was ERp57; this close structural homologue of PDI was shown to have PDI activity and strong affinity for the ER lectins calnexin and calreticulin and thus to act as a disulphide isomerase towards newly synthesised glycoproteins bound by these lectins (Oliver et al. 1997; Elliott et al. 1997; Zapun et al. 1998).
11.6 The Cellular Environment for Protein Oxidative Folding (1991–2000) Exploration of protein maturation in the secretory pathway of mammalian cells had, by 1990, characterised the numerous modifications and other steps involved in biosynthesis and export of complex multi-domain and multi-chain proteins such as immunoglobulins and viral surface proteins and had identified control steps and agents that blocked the pathway (Rose and Doms 1988). The formation of disulphide bonds had not been manipulated extensively, but in 1992 it was shown that simple addition and removal of DTT could reversibly perturb this process in intact cells (Braakman et al. 1992). The biosynthesis of influenza haemagglutinin (HAO) was shown to pass through various disulphide-bonded partly folded intermediates within the ER, before trimerisation and passage further through the secretory pathway; addition of >1 mM DTT blocked co-translational disulphide formation in HAO and reduced pre-existing disulphides, but this effect was rather rapidly reversed on removal of DTT and replacement with standard cell culture media. This established that the redox environment within the ER of intact cells could be modulated by the addition of exogenous reagents, but also emphasised that an endogenous system within cells maintained a redox system within the ER appropriate for protein disulphide formation. The nature of this system was not immediately apparent, although the abundance and significance of glutathione within the ER were already suspected. Early work
270
R.B. Freedman
on the GSH:GSSG ratio in sub-cellular fractions had shown that the ratio in microsomes derived from the ER was far more oxidising than that in other fractions, especially the cytosol (Isaacs and Binkley 1977) and it was known that addition of GSSG would enable co- or post-translational disulphide formation in proteins translated in vitro in the presence of microsomal vesicles (Scheele and Jacoby 1982; Bulleid and Freedman 1988). But the ratio (and hence effective redox potential, which is a function of [GSH]2/[GSSG]) had not been determined in intact cells. Using a complex and sophisticated method based on an indicator peptide, Hwang et al. (1992) determined that this ratio was in the range 1:1 to 3:1, hence establishing considerably more oxidising conditions in the ER than those in the cytosol, where the ratio is 30:1 to 100:1. This result was a major advance but did not in itself answer key questions about the immediate and ultimate sources of oxidising equivalents for the formation of protein disulphide bonds in the ER. GSSG might or might not be the immediate source, but in either case, was it generated in the ER or transported into the ER? And what was the ultimate source of oxidising equivalents? Yeast genetics provided the next key advance with the discovery of an essential gene ERO1 required for oxidative folding in the ER (Frand and Kaiser 1998; Pollard et al. 1998); mutations in this gene led to the accumulation of misfolded reduced proteins in the ER, and made cells strikingly sensitive to DTT, whereas over-expression of the gene conferred DTT-resistance. Initial physiological studies showed that the gene product Ero1p acts upstream to oxidise PDI in an electrontransfer pathway (analogous to how DsbB acts on DsbA in the bacterial disulphide oxidation pathway), but did not identify the source of oxidation. However, by the end of the decade, some biochemical analysis had been undertaken – establishing that Ero1p is a flavoprotein oxidase, apparently capable of transferring electrons from PDI to molecular oxygen (Tu et al. 2000) – and a mammalian homologue of yeast Ero1p had been identified (Cabibbo et al. 2000). Analogy with the bacterial periplasmic system was a powerful influence on this line of work, and those working on the yeast system tended to present PDI simply as analogous to DsbA, an electron-transfer component in a linear pathway, without regard to PDI’s crucial ability to isomerise disulphides in structured folding intermediates.
11.7 The Near Past and the Future (2001+) The explosion of work in this field in the past decade is comprehensively reviewed in earlier chapters. Particularly, impressive has been the determination of disulphide regeneration pathways for many proteins with challenging disulphide topologies (see Arolas et al. 2006, and Chaps. 1–4 of this volume.). In relation to my particular interest, PDI and the Dsb proteins (see Chaps. 7 and 9), the most striking advances in the decade have been the sudden burst of high-resolution structural data (Gruber et al. 2006; Kozlov et al. 2010) including the definition of the major ligand binding site
11 A Half-Century in Review
271
in PDI. But the key (and unexpected) insight into PDI mechanism was discovery of the role of an Arg side-chain, remote from the active site, in modulating the pKa of the active site thiol groups, by a change of its position at different stages of the catalytic cycle (Lappi et al. 2004; Karala et al. 2010). More detail is emerging about the complex disulphide-linked folding pathways undergone by proteins within the ER (Land et al. 2003) and new data suggest that the redox pathways contributing to disulphide formation in the ER are much more complex and network-like, rather than a simple linear flow of electrons from molecular oxygen via Ero1 and PDI to substrate proteins (Karala et al. 2009; Tavender et al. 2010). But this is a personal selection of highlights from a glittering list!
References Acharya AS, Taniuchi H (1977) Formation of the four isomers of hen egg white lysozyme containing three native disulfide bonds and one open disulfide bond. Proc Natl Acad Sci USA 74: 2362–2366 Anfinsen CB, Haber E (1961) Studies on reduction and reformation of protein disulfide bonds. J Biol Chem 236:1361–1363 Anfinsen CB, Haber E, Sela M, White FH Jr (1961) The kinetics of formation of native ribonuclease during oxidation of the reduced chain. Proc Natl Acad Sci USA 47:1309–1314 Arolas JL, Aviles FX, Chang J-Y, Ventura S (2006) Folding of small disulfide-rich proteins: clarifying the puzzle. Trends Biochem Sci 31:292–301 Bader MW, Xie T, Yu C-A, Bardwell JCA (2000) Disulfide bonds are generated by quinone reduction. J Biol Chem 275:26082–26088 Baldwin RL (1978) The pathway of protein folding. Trends Biochem Sci 3:66–68 Bardwell JCA (1994) Building bridges: disulphide formation in the cell. Mol Microbiol 14:199–205 Bardwell JCA, Beckwith J (1993) The bonds that tie: catalyzed disulfide bond formation. Cell 74:769–771 Bardwell JCA, McGovern K, Beckwith J (1991) Identification of a protein required for disulfide bond formation in vivo. Cell 67:581–589 Bello J, Harker D, de Jarnette E (1961) Note on an x-ray diffraction investigation of reducedreoxidized ribonuclease. J Biol Chem 236:1358–1360 Braakman I, Helenius J, Helenius A (1992) Manipulating disulfide bond formation and protein folding in the endoplasmic reticulum. EMBO J 11:1717–1722 Bulleid NJ, Freedman RB (1988) Defective co-translational formation of disulphide bonds in protein disulphide isomerase-deficient microsomes. Nature 335:649–651 Cabibbo A, Pagani M, Fabbri M, Rocchi M, Farmery MR, Bulleid NJ, Sitia R (2000) ERO1-L, a human protein that favors disulfide bond formation in the endoplasmic reticulum. J Biol Chem 275:4827–4833 Chandler ML, Varandani PT (1975) Kinetic analysis of the mechanism of insulin degradation by glutathione-insulin transhydrogenase (thiol:protein-disulfide oxidoreductase). Biochemistry 14:2107–2115 Chang JY, Canals F, Schindler P, Querol E, Aviles FX (1994) The disulfide folding pathway of potato carboxypeptidase inhibitor. J Biol Chem 269:22087–22094 Chivers PT, Prehoda KE, Raines RT (1997) The CXXC motif: a rheostat in the active site. Biochemistry 36:4061–4066 Creighton TE (1974a) Intermediates in the refolding of reduced pancreatic trypsin inhibitor. J Mol Biol 87:579–602
272
R.B. Freedman
Creighton TE (1974b) The single-disulphide intermediates in the refolding of reduced pancreatic trypsin inhibitor. J Mol Biol 87:603–624 Creighton TE (1975) The two-disulphide intermediates and the folding pathway of reduced pancreatic trypsin inhibitor. J Mol Biol 95:167–199 Creighton TE (1979) Experimental studies of protein folding and unfolding. Prog Biophys Mol Biol 33:231–297 Creighton TE (1980) A three-disulphide intermediate in refolding of reduced ribonuclease A with a folded conformation. FEBS Lett 118:283–288 Creighton TE, Hillson DA, Freedman RB (1980) Catalysis by protein-disulphide isomerase of the unfolding and refolding of proteins with disulphide bonds. J Mol Biol 142:43–62 Creighton TE, Bagley CJ, Cooper L, Darby NJ, Freedman RB, Kemmink J, Sheikh A (1993) On the biosynthesis of bovine pancreatic trypsin inhibitor (BPTI): structure, processing, folding and disulphide bond formation of the precursor in vitro and in microsomes. J Mol Biol 232:1176–1196 Darby NJ, Creighton TE (1995) Functional properties of the thioredoxin-like domains of protein disulphide isomerase. Biochemistry 34:11725–11735 Darby NJ, Penka E, Vincentelli R (1998) The multi-domain structure of protein disulfide isomerase is essential for high catalytic activity. J Mol Biol 276:239–247 De Lorenzo F, Fuchs S, Anfinsen CB (1966) Characterization of a peptide fragment containing the essential half-cystine residue of a microsomal disulfide interchange enzyme. Biochemistry 5:3961–3965 Edman JC, Ellis L, Blacher RW, Roth RA, Rutter WJ (1985) Sequence of protein disulphide isomerase and implications of its relationship to thioredoxin. Nature 317:267–270 Eigenbrot C, Randal M, Kossiakoff AA (1990) Structural effects induced by removal of a disulfidebridge: the X-ray structure of the C30A/C51A mutant of basic pancreatic trypsin inhibitor at 1.6 Å. Protein Eng 3:591–598 Ellgaard L, Ruddock LW (2005) The human protein disulphide isomerise family: substrate interactions and functional properties. EMBO Rep 6:28–32 Elliott JG, Oliver JD, High S (1997) The thiol-dependent reductase ERp57 interacts specifically with N-glycosylated integral membrane proteins. J Biol Chem 272:13849–13855 Epstein CJ, Goldberger RF, Anfinsen CB (1963) The genetic control of tertiary protein structure: studies with model systems. Cold Spring Harb Symp Quant Biol 28:439–449 Ewbank JJ, Creighton TE (1993) Structural characterization of the disulfide folding intermediates of bovine a-lactalbumin. Biochemistry 32:3694–3707 Farquhar R, Honey N, Murant SJ, Bossier P, Schultz L, Montgomery D, Ellis RW, Freedman RB, Tuite MF (1991) Protein disulfide isomerise is essential for viability in Saccharomyces cerevisiae. Gene 108:81–89 Ferrari DM, Soling H-D (1999) The protein disulphide-isomerase family: unravelling a string of folds. Biochem J 339:1–10 Frand AR, Kaiser CA (1998) The ERO1 gene of yeast is required for oxidation of protein dithiols in the endoplasmic reticulum. Mol Cell 1:161–170 Freedman MH, Sela M (1966) Recovery of specific activity upon reoxidation of completely reduced polyalanyl rabbit antibody. J Biol Chem 241:5225–5232 Freedman RB (1984) Native disulphide bond formation in protein biosynthesis: evidence for the role of protein disulphide isomerase. Trends Biochem Sci 9:438–441 Freedman RB, Hawkins HC (1977) Enzyme-catalysed disulphide interchange and protein biosynthesis. Biochem Soc Trans 5:348–357 Freedman RB, Hillson DA (1980) Formation of disulphide bonds. In: Freedman RB, Hawkins HC (eds) The enzymology of post-translational modification of proteins, vol 1. Academic, New York, pp 157–212 Freedman RB, Hirst TR, Tuite MF (1994) Protein disulphide isomerase: building bridges in protein folding. Trends Biochem Sci 19:331–336 Freedman RB, Klappa P, Ruddock LW (2002) Protein disulfide isomerases exploit synergy between catalytic and specific ligand-binding domains. EMBO Rep 3:146–150
11 A Half-Century in Review
273
Fuchs S, de Lorenzo F, Anfinsen CB (1967) Studies on the mechanism of the enzymic catalysis of disulfide interchange in proteins. J Biol Chem 242:398–402 Givol D, Goldberger RF, Anfinsen CB (1964) Oxidation and disulfide interchange in the reactivation of reduced ribonuclease. J Biol Chem 239:PC3114–PC3116 Givol D, de Lorenzo F, Goldberger RF, Anfinsen CB (1965) Disulfide interchange and the threedimensional structure of proteins. Proc Natl Acad Sci USA 53:676–684 Goldberger RF, Epstein CJ, Anfinsen CB (1963) Acceleration of reactivation of reduced bovine pancreatic ribonuclease by a microsomal system from rat liver. J Biol Chem 238:628–635 Goldenberg DP (1992) Native and non-native intermediates in the BPTI folding pathway. Trends Biochem Sci 17:257–261 Gruber CW, Cemazar M, Heras B, Martin JL, Craik DJ (2006) Protein disulfide isomerase: the structure of oxidative folding. Trends Biochem Sci 31:455–463 Gunther R, Brauer C, Janetzky B, Forster H-H, Ehbrecht I-M, Lehle L, Kuntzel H (1991) The Saccharomyces cerevisiae TRG1 gene is essential for growth and encodes a lumenal endoplasmic reticulum glycoprotein involved in the maturation of vacuolar carboxypeptidase. J Biol Chem 266:24557–24563 Haber E (1964) Recovery of antigenic specificity after denaturation and complete reduction of disulfides in a papain fragment of antibody. Proc Natl Acad Sci USA 5:1099–1106 Hantgan RR, Hammes GG, Scheraga HA (1974) Pathways of folding of reduced bovine pancreatic ribonuclease. Biochemistry 13:3421–3431 Hillson DA, Freedman RB (1980) Resolution of protein disulphide-isomerase and glutathioneinsulin transhydrogenase activities by covalent chromatography: properties of the purified protein disulphide-isomerase. Biochem J 191:373–388 Huber-Wunderlich M, Glockshuber R (1998) A single dipeptide sequence modulates the redox properties of a whole enzyme family. Fold Des 3:161–171 Hwang C, Sinskey AJ, Lodish HF (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257:1496–1502 Isaacs J, Binkley F (1977) Glutathione-dependent control of protein disulfide-sulfhydryl content by sub-cellular fractions of hepatic tissue. Biochim Biophys Acta 498:29–38 Janolino VG, Sliwkowski MX, Swaisgood HE, Horton HR (1978) Catalytic effect of sulfhydryl oxidase on the formation of three-dimensional structure in chymotrypsinogen A. Arch Biochem Biophys 191:269–277 Kaderbhai MA, Austen BM (1984) Dog pancreatic microsomal membrane polypeptides analysed by two-dimensional gel electrophoresis. Biochem J 217:145–157 Karala AR, Lappi AK, Saaranen MJ, Ruddock LW (2009) Efficient peroxide-mediated oxidative folding of a protein at physiological pH and implications for oxidative folding in the endoplasmic reticulum. Antioxid Redox Signal 11:963–970 Karala AR, Lappi AK, Ruddock LW (2010) Modulation of an active site cysteine pKa allows PDI to act as a catalyst of both disulfide bond formation and isomerization. J Mol Biol 396:883–892 Kato I, Anfinsen CB (1969) On the stabilization of ribonuclease S-protein by ribonuclease S-peptide. J Biol Chem 244:1004–1007 Kemmink J, Darby NJ, Dijkstra K, Nilges M, Creighton TE (1996) Structure determination of the N-terminal thioredoxin-like domain of protein disulphide isomerase using multi-dimensional 13 C/15N NMR spectroscopy. Biochemistry 35:7684–7691 Kemmink J, Darby NJ, Dijkstra K, Nilges M, Creighton TE (1997) The folding catalyst protein disulphide isomerase is constructed of active and inactive thioredoxin modules. Curr Biol 7:239–245 Klappa P, Hawkins HC, Freedman RB (1997) Interactions between protein disulphide isomerase and peptides. Eur J Biochem 248:37–42 Klappa P, Ruddock LW, Darby NJ, Freedman RB (1998) The b¢ domain provides the principal peptide-binding site of protein disulphide isomerase but all domains contribute to binding of misfolded proteins. EMBO J 17:927–935 Kohno T, Carmichael DF, Sommer A, Thompson RC (1990) Refolding of recombinant proteins. Methods Enzymol 185:187–195
274
R.B. Freedman
Konishi Y, Ooi T, Scheraga HA (1981) Regeneration of ribonuclease A from the reduced protein: isolation and identification of intermediates and equilibrium treatment. Biochemistry 20:3944–3955 Kosen PA, Creighton TE, Blout ER (1980) Ultraviolet difference spectroscopy of intermediates trapped in unfolding and refolding of bovine pancreatic trypsin inhibitor. Biochemistry 19:4936–4944 Kozlov G, Maattanen P, Thomas DY, Gehring K (2010) A structural overview of the PDI family of proteins. FEBS J 277:3924–3936 Laity JH, Lester CC, Shimotakahara S, Zimmerman DE, Montelione E, Scheraga HA (1997) Structural characterization of an analog of the major rate-determining disulfide folding intermediate of bovine pancreatic ribonuclease A. Biochemistry 36:12683–12699 LaMantia M-L, Miura T, Tachikawa H, Kaplan HA, Lennarz WJ, Mizunaga T (1991) Glycosylation site binding protein and protein disulfide isomerise are identical and essential for cell viability in yeast. Proc Natl Acad Sci USA 88:4453–4457 Lambert N, Freedman RB (1985) The latency of rat liver microsomal protein disulphide-isomerase. Biochem J 228:635–645 Land A, Zonneveld D, Braakman I (2003) Folding of HIV-1 envelope glycoprotein involves extensive isomerisation of disulfide bonds and conformation-dependent leader peptide cleavage. FASEB J 17:1058–1067 Lappi AK, Lensink MF, Alanen HI, Salo KEH, Lobell M, Juffer AH, Ruddock LW (2004) A conserved arginine plays a role in the catalytic cycle of the protein disulphide isomerases. J Mol Biol 335:283–295 Marston FAO (1986) The purification of eukaryotic polypeptides synthesized in Escherichia coli. Biochem J 240:1–12 Marston FAO, Lowe PA, Doel MT, Schoemaker JM, White S, Angal S (1984) Purification of calf prochymosin (prorennin) synthesised in Escherichia coli. Bio/technology 2:800–804 Martin JL (1995) Thioredoxin – a fold for all reasons. Structure 3:245–250 Martin JL, Bardwell JCA, Kuriyan J (1993) Crystal structure of the DsbA protein required for disulphide bond formation in vivo. Nature 365:464–468 Mazzarella RA, Srinivasan M, Haugejorden SM, Green M (1990) ERp72, an abundant luminal endoplasmic reticulum protein, contains three copies of the active site sequences of protein disulfide isomerase. J Biol Chem 265:1094–1101 McLaughlin SH, Bulleid NJ (1998) Thiol-independent interaction of protein disulphide isomerase with type X collagen during intracellular folding and assembly. Biochem J 331:793–800 Mills EC, Lambert N, Freedman RB (1983) Identification of protein disulphide-isomerase as a major acidic polypeptide in rat liver microsomal membranes. Biochem J 213:245–248 Munro S, Pelham HRB (1987) A C-terminal signal prevents secretion of luminal ER proteins. Cell 48:899–907 Narayan M, Welker E, Wedemeyer WJ, Scheraga HA (2000) Oxidative folding of proteins. Acc Chem Res 33:805–812 Norgaard P, Westphal V, Tachibana C, Alsoe L, Holst B, Winther JR (2001) Functional differences in yeast protein disulfide isomerases. J Cell Biol 152:553–562 Oliver JD, van der Wal FJ, Bulleid NJ, High S (1997) Interaction of the thiol-dependent reductase ERp57 with nascent glycoproteins. Science 275:86–88 Pihlajaniemi T, Helaakoski T, Tasanen K, Myllyla R, Huhtala M-L, Koivu J, Kivirikko KI (1987) Molecular cloning of the b-subunit of prolyl-4-hydroxylase: this subunit and protein disulphide isomerase are products of the same gene. EMBO J 6:643–649 Pollard MG, Travers KJ, Weissman JS (1998) Ero1p: a novel and ubiquitous protein with an essential role in oxidative protein folding in the endoplasmic reticulum. Mol Cell 1:171–182 Pollitt S, Zalkin H (1983) Role of primary structure and disulfide bond formation in b-lactamase secretion. J Bacteriol 153:27–32 Riemer J, Bulleid N, Herrmann JM (2009) Disulfide formation in the ER and mitochondria: two solutions to a common process. Science 324:1284–1287 Rietsch A, Beckwith J (1998) The genetics of disulfide bond metabolism. Annu Rev Genet 32:163–184
11 A Half-Century in Review
275
Rietsch A, Belin D, Martin N, Beckwith J (1996) An in vivo pathway for disulfide bond isomerisation in Escherichia coli. Proc Natl Acad Sci USA 93:13048–13053 Rose JK, Doms RW (1988) Regulation of protein export from the endoplasmic reticulum. Annu Rev Cell Biol 4:257–288 Roth RA, Pierce SB (1987) In vivo cross-linking of protein disulfide isomerase to immunoglobulins. Biochemistry 26:4179–4182 Saxena VP, Wetlaufer DB (1970) Formation of three-dimensional structure in proteins I. Rapid nonenzymic reactivation of reduced lysozyme. Biochemistry 9:5015–5023 Scheele G, Jacoby R (1982) Conformational changes associated with proteolytic processing of presecretory proteins allow glutathione-catalysed formation of native disulfide bonds. J Biol Chem 257:12277–12282 Scherens B, Dubois E, Messenguy F (1991) Determination of the sequence of the yeast YCL313 gene localized on chromosome III: homology with the protein disulfide isomerase (PDI) gene product of other organisms. Yeast 7:185–193 Shin HC, Scheraga HA (2000) Catalysis of the oxidative folding of bovine pancreatic ribonuclease A by protein disulfide isomerase. J Mol Biol 300:995–1003 Steiner DF, Clark JL (1968) The spontaneous reoxidation of reduced beef and rat proinsulins. Proc Natl Acad Sci USA 60:622–629 Tachibana C, Stevens TH (1992) The yeast EUG1 gene encodes an endoplasmic reticulum protein that is functionally related to protein disulfide isomerase. Mol Cell Biol 12:4601–4611 Tachikawa H, Miura T, Katakura Y, Mizunaga T (1991) Molecular structure of a yeast gene PDI1 encoding protein disulfide isomerase that is essential for cell growth. J Biochem (Tokyo) 110:306–313 Talluri S, Rothwarf DA, Scheraga HA (1994) Structural characterization of a three-disulfide intermediate of ribonuclease A involved in both the folding and unfolding pathways. Biochemistry 33:10437–10449 Talmadge K, Gilbert W (1982) Cellular location affects protein stability in Escherichia coli. Proc Natl Acad Sci USA 79:1830–1833 Tasanen K, Parkkonen T, Chow LT, Kivirikko KI, Pihlajaniemi T (1988) Characterization of the human gene for a polypeptide that acts both as the beta subunit of prolyl 4-hydroxylase and as protein disulfide isomerase. J Biol Chem 263:16218–16224 Tavender TJ, Springate JJ, Bulleid NJ (2010) Recycling of peroxiredoxin IV provides a novel pathway for disulphide formation in the endoplasmic reticulum. EMBO J 29:4185–4197. doi:10.1038/emboj.2010.273 Tu BP, Ho-Schleyer SC, Travers KJ, Weissman JS (2000) Biochemical basis of oxidative protein folding in the endoplasmic reticulum. Science 290:1571–1574 van den Berg B, Chung EW, Robinson CV, Dobson CM (1999a) Characterisation of the dominant oxidative folding intermediate of hen lysozyme. J Mol Biol 290:781–796 van den Berg B, Chung EW, Robinson CV, Mateo PL, Dobson CM (1999b) The oxidative refolding of hen lysozyme and its catalysis by protein disulfide isomerase. EMBO J 18:4794–4803 van Mierlo CPM, Darby NJ, Creighton TE (1992) The partially folded conformation of the Cys30Cys51 intermediate in the disulfide folding pathway of bovine pancreatic trypsin inhibitor. Proc Natl Acad Sci USA 89:6775–6779 van Mierlo CPM, Darby NJ, Keeler J, Neuhaus D, Creighton TE (1993) Partially folded conformation of the (30–51) intermediate in the disulphide folding pathway of bovine pancreatic trypsin inhibitor: 1H and 15N resonance assignments and determination of backbone dynamics from 15N relaxation measurements. J Mol Biol 229:1125–1149 van Mierlo CPM, Kemmink J, Neuhaus D, Darby NJ, Creighton TE (1994) 1H NMR analysis of the partly-folded non-native two-disulphide intermediates (30–51, 5–14) and (30–51, 5–38) in the folding pathway of bovine pancreatic trypsin inhibitor. J Mol Biol 235:1044–1061 Venetianer P, Straub FB (1963) The enzymic reactivation of reduced ribonuclease. Biochim Biophys Acta 67:166–168 Wang C-C, Tsou C-L (1993) Protein disulfide isomerase is both an enzyme and a chaperone. FASEB J 7:1515–1517
276
R.B. Freedman
Wedemeyer WJ, Welker E, Narayan M, Scheraga HA (2000) Disulfide bonds and protein folding. Biochemistry 39:4207–4216 Weissman JS, Kim PS (1991) Reexamination of the folding of BPTI: predominance of native intermediates. Science 253:1386–1393 Weissman JS, Kim PS (1993) Efficient catalysis of disulphide bond rearrangements by protein disulphide isomerase. Nature 365:185–188 Wetterau JR, Combs KA, Spinner SN, Joiner BJ (1990) Protein disulfide isomerase as a component of the microsomal triglyceride transfer protein complex. J Biol Chem 265:9800–9807 Wetzel R (1987) Harnessing disulfide bonds using protein engineering. Trends Biochem Sci 12:478–482 White FH Jr (1960) Regeneration of enzymatic activity by air-oxidation of reduced ribonuclease with observations on thiolation during reduction with thioglycolate. J Biol Chem 235:383–389 White FH Jr (1961) Regeneration of native secondary and tertiary structures by air oxidation of reduced ribonuclease. J Biol Chem 236:1353–1360 Whitney PL, Tanford C (1965) Recovery of specific activity after complete unfolding and reduction of an antibody fragment. Proc Natl Acad Sci USA 53:524–532 Yutani K, Yutani A, Imanishi A, Isemura T (1968) The mechanism of refolding of the reduced random coil form of lysozyme. J Biochem (Tokyo) 64:449–455 Zapun A, Creighton TE (1994) Effects of DsbA on the disulfide folding of bovine pancreatic trypsin inhibitor and a-lactalbumin. Biochemistry 33:5202–5211 Zapun A, Missiakas D, Raina S, Creighton TE (1995) Structural and functional characterization of DsbC, a protein involved in disulfide bond formation in Escherichia coli. Biochemistry 34:5075–5089 Zapun A, Darby NJ, Tessier DC, Michalak M, Bergeron JJM, Thomas DY (1998) Enhanced catalysis of ribonuclease B folding by the interaction of calnexin or calreticulin with ERp57. J Biol Chem 273:6009–6012 Ziegler DM, Poulsen LL (1977) Protein disulfide bond synthesis: a possible intracellular mechanism. Trends Biochem Sci 2:79–81
Index
A AEMTS. See 2-Aminoethyl methanethiosulfonate AEP. See Asparaginyl endopeptidase Alkaline phosphatase (ALP), 125 Allosteric disulfide bonds analysis, 176 angiotensinogen, 169 aryl sulfotransferase, 171–172 AtFKBP13, 172 b2-glycoprotein I, 167–168 botulinum neurotoxins, 165–166 carboxyl-terminal Src kinase, 172–173 CD4, 166–167 definition, 152 eukaryotes conservation, 154–155 cysteine pairs, 157–158 modern proteomes, 153 rate of, 155–157 reshuffling, 159 gp120, 163–165 MICA, 168–169 PapD-like chaperones, 170–171 RHStaple bond catalytic function, 174–175 redox potentials, 175 stereochemistry, 175–176 structures protein function, 160–162 RHStaples, 159–160 X-ray and NMR structures, 159 tissue factor, 173–174 transglutaminase 2, 173 VWC domain, 169–170
2-Aminoethyl methanethiosulfonate (AEMTS), 27 Amphioxus insulin-like peptide, 71 Angiotensinogen, 169 Aromatic thiols (ArSH), 124 ArSH. See Aromatic thiols Aryl sulfotransferase, 171–172 Asparaginyl endopeptidase (AEP), 47 AtFKBP13, 172 B b2-glycoprotein I (b2GPI), 167–168 Botulinum neurotoxins, 165–166 Bovine pancreatic ribonuclease A folding conditions, 26–27 regeneration rates and pathways, 27–28 terminology, 26 Bovine pancreatic trypsin inhibitor (BPTI), 6, 25, 92, 117–118, 265 disulfide shuffling, 244 native and non-native disulphide, 262 with one disulfide bridge, 245–246 oxidative refolding, 244, 245 PDI, 266–267 ribbon diagram, 243, 244 serine protease, 243 with two disulfide bridge, 246 two-disulphide intermediates, 266 without disulfide bridges, 244 C Carboxyl-terminal Src kinase, 172–173 Cluster of differential 4 (CD4), 166–167 CMC. See Critical micellular concentration
R.J.Y. Chang and S. Ventura (eds.), Folding of Disulfide Proteins, Protein Reviews 14, DOI 10.1007/978-1-4419-7273-6, © Springer Science+Business Media, LLC 2011
277
278 Cole–Cole distribution, 233 Coupling conformational folding and disulfide formation disulfide bonds polypeptide chain, 4 reduction and oxidation, 6 thiol–disulfide exchange reaction, 4, 5 folding intermediates disulfide-bond formation, 16 genuine folding intermediates, 17 native disulfide bonds, 18 quenching method, 7–8 trapping method, 6 oxidative folding pathways epidermal growth factor, 14–16 hirudin, 13–14 leech-derived trypsin inhibitor, 10–13 protein folding, 2–3 structural characterization, 16–18 C-peptide, 85–87 Critical micellular concentration (CMC), 125 Cyclic peptide backbone, 47 Cyclotides applications, 47–48 biological folding novel plant PDI, 54–55 in planta, enzymatic folding, 57–58 in vitro, enzymatic folding, 55–57 biosynthesis, 46–47 chemical folding bracelet cyclotide, 52 cycloviolacin structure, 52–53 cystine-rich peptides, 53–54 folding/unfolding pathway, 50–51 Möbius, 51 native-like structure, 50 oxidative folding, 51–52 prototypic kalata B1, 49 distribution, 46 structure cystine knot, 44 Möbius and bracelet, 44–46 peptide bonds, 43–44 Cysteine folding facilitation and protein structure, 219–220 iron–sulfur cluster, 219 metal homeostasis, 220 proteases, 219 proteinogenic a-amino acid, 218 redox sensor and homeostasis, 220 ROS response, 220 zinc finger, 219
Index Cytoplasmic expression active alkaline phosphatase, demonstration, 199 cysteine oxidation, 198–199 cytoplasmic–periplasmic expression method, 202 E. coli, disulfide bond reducing systems, 198 glutaredoxin pathway, 198 levels, disulfide bond formation, 199 soluble recombinant proteins, 195–197 soluble recombinant proteins, oxidative folding, 199–202 thioredoxin fold structure, 198 D Diffusion-ordered spectroscopy (DOSY), 226 Disulfide bond formation, Escherichia coli eukaryotes cytoplasm formation, 185 isomerization, 187 ribbon diagram, yeast PDI, 187 thioredoxin fold, 186 yeast PDI crystal structure, 186 gram-negative bacterium, 184–185 human proteins folding, 185 multidisulfided heterologous proteins biotechnological potential, 194 cell-free protein synthesis systems, 202–203 cytoplasmic expression (see Cytoplasmic expression) human tissue plasminogen, 194–195 molecular chaperone, 195 PPIases, 195 protein engineering, 204–205 proteins refolding, 194 recombinant proteins, 194 oxidative folding cysteines, 189 deleterious oxidation, 193 DsbA molecular structure, 188–189 DsbC crystal structure, 190, 191 hydrophobic cleft, surface representation, 192 isomerization-DsbC/DsbG/DsbD, 188 mechanism, isomerization, 190 metal-catalyzed thiol oxidation, 189–190 oxidation-DsbA/AsbB, 188 recycled DsbC, 191 substrate recognition, 192 thiol–disulfide electron transfer processes, 193 V-shaped homodimer, 192–193
Index Disulfide bond structures protein function, 160–162 RHStaples, 159–160 X-ray and NMR structures, 159 Disulfide formation, catalysis oxidative folding, 139–140 thermodynamic instability, 141 thioredoxin, 140 Disulfide isomerization, PDI catalysis b¢ domain, 142 BPTI refolding, 143 C-terminal cysteine, 142 RNase A, 143 Disulfide scrambling folding a-lactalbumin, 101–104 lysozyme, 100–101 mechanism, 92 TAP and BPTI, 99–100 native proteins, 92 oxidative folding, 94–95 unfolding denaturation curves, 97–98 N-protein and X-isomers, 95 TAP and BPTI, 95–96 X-isomers, 93 Dithiothreitol (DTT), 75 Dithiothreitol (DTTred), 8 DOSY. See Diffusion-ordered spectroscopy Double chain insulin C-peptide, 85–87 folding mechanism, 64 IGF-1 and rAILP, 83 intermediates B chain, 78–79 characterization, 75–77 isolation, 75–77 P1A and P3A, 79–80 purification, 75–77 temporal distribution, 75 time-dependent distribution, 77–78 polypeptide chains, 74 putative in vitro folding pathway hypothetic double-chain intermediates, 81–83 intermediates, 80–81 structural motif, 85 E Electrospray ionization-mass spectrometry (ESI-MS), 68 Endoplasmic reticulum (ER), 46, 54, 117
279 Enzymatic folding in planta, 57–58 in vitro, 55–57 Epidermal growth factor (EGF) Cys6 and Cys20, 15 predominant 2S intermediate, 15 scrambled isomers, 16 ER-processed proteins, fold maturation bovine pancreatic ribonuclease A folding conditions, 26–27 regeneration rates and pathways, 27–28 terminology, 26 RNase A minor regeneration pathways, 36 off-pathway intermediates (kinetically trapped intermediates), 37–38 oxidative folding, 29–36 Eukaryotes conservation, 154–155 cysteine pairs, 157–158 cytoplasm formation, 185 isomerization, 187 modern proteomes, 153 and prokaryotes cytosolic thioredoxin system, 223 E. coli, 222 Ero1 mutational analysis, 221 oxidative potential and auxiliary machinery, 220–221 Saccharomyces cerevisiae, 221 thiol–disulfide exchange mechanism, 222 rate of, 155–157 reshuffling, 159 ribbon diagram, yeast PDI, 187 thioredoxin fold, 186 yeast PDI crystal structure, 186 G Gemini surfactants, 124–125 Glutaredoxin (Grx), 122 Glutathione, 119 gp120, 163–165 Green fluorescent protein (GFP), 175 Grx. See Glutaredoxin H Helicoverpa punctigera, 48 Hen egg white lysozyme (HEWL) albumen, 237 with all disulfide bond (4SS), 241 glycosidic linkage, 237 oxidative refolding pathway, 238, 239
280 Hen egg white lysozyme (HEWL) (cont.) ribbon diagram, 238 with three disulfide bond (3SS), 242 with two disulfide bond (2SS), 242–243 without disulfide bonds (0SS) protein clustering, 238 transverse relaxation rates, 239, 240 W62G mutation, 241 W111G/W123G bond, 239–240 Heteronuclear nuclear Overhauser enhancement (hetNOE), 231 Hirudin, 13–14 Human immunodeficiency virus (HIV), 163 Human proinsulin (HPI) native disulfide bonds, 71 putative in vitro disulfide-forming pathway, 68, 70 refolding intermediates, 68, 69 I Insulin-like growth factor-1 (IGF-1) a-helical segments, 72 rAILP, 73 refolding, 73–74 Insulin structural motif, 85 Iodoacetic acid sodium salt (IAA), 75 K Kalata B1, 49 Kinetically trapped intermediates, 37–38 L a-Lactalbumin (aLA) disulfide intact method, 101–102 disulfide scrambling, 101 on/off-pathway intermediates, 104 structural heterogeneity, 102 Leech carboxypeptidase inhibitor (LCI), 17 Leech-derived trypsin inhibitor (LDTI) BTPI, 10–11 native disulfide bonds, 11, 12 b-sheet elements, 12 2S intermediates, 12 Lipari–Szabo approach, 232 Lysozyme, 100–101, 103 M Mass spectrometry (MS), 8 MICA, 168–169 Minor regeneration pathways, 36
Index Möbius, 51 Momordica cochinchinensis trypsin inhibitor II (MCoTI-II), 17 Multiple antigen peptide (MAP), 122 N Native disulfide bonds, 18 NMR-spectroscopic investigation BPTI (see Bovine pancreatic trypsin inhibitor) chemical shift analysis, 229–230 cysteine, 218–220 diffusion ordered spectroscopy, 236–237 disulfide bonds 13 C NMR chemical shifts, 228, 229 diselenide proxies, 77Se NMR spectroscopy, 228 protein folding kinetics, 225–226 structure and conformational stability, 223–224 thermodynamics, 224–225 X-ray crystallography, 227 DOSY, 226 eukaryotes and prokaryotes cytosolic thioredoxin system, 223 E. coli, 222 Ero1 mutational analysis, 221 oxidative potential and auxiliary machinery, 220–221 Saccharomyces cerevisiae, 221 thiol–disulfide exchange mechanism, 222 HEWL (see Hen egg white lysozyme) human and murine prion protein heteronuclear 15NR2 and R1r relaxation rates, 248, 250 misfolding properties, 249 ribbon diagram, 247 secondary structure propensities, 248, 249 TSEs, 247 urea-denatured states, 248 relaxation techniques biomacromolecules, 230 Cole–Cole distribution, 233 heteronuclear spin relaxation rate constants, 231 hetNOE, 231 hydrophobic cluster, 234 intrinsic relaxation rate, 235 isotropic diffusion, 231, 232 Lipari–Szabo approach, 232 Lorentzian distribution, 233 random coil polypeptide chain, 234
Index segmental motion model, 233–234 slow time scale, 235–236 total correlation time, 232 unfolded proteins, characteristics, 227, 228 N-terminal domain, 174 N-terminal repeat (NTR), 47, 56 O Off-pathway intermediates, 37–38 Oldenlandia affinis, 46 Optical rotatory dispersion (ORD), 259 Oxidative folding, 94–95. See also Oxidative folding pathways; Oxidative protein folding nS species, 30 postfolding steps, 35–36 rate-determining steps a-lactalbumin, 34 PDI, 34–35 prolines, 33–34 reshuffling reaction, 34 3Snative intermediate, 33 structure and disulfide-connectivity, 30–31 unstructured 3S isomers, 31–32 Oxidative folding pathways epidermal growth factor Cys6 and Cys20, 15 predominant 2S intermediate, 15 scrambled isomers, 16 hirudin, 13–14 leech-derived trypsin inhibitor BTPI, 10–11 native disulfide bonds, 11, 12 b-sheet elements, 12 2S intermediates, 12 Oxidative protein folding cysteines, 134 disulfide bonds, 134 ER, 134, 143 Ero1p, 143–144 folding strategies, 117–118 GSH/GSSG, 144–145 LDL receptor, 145 native-cysteine, 137 noncovalent interactions, 136 onconase, 136 protein disulfide bond, 111 protein folding mechanisms, 134–135 rapid formation, disulfide bonds applications, 129 platinum-based compounds, 128 selenoxides, 128 RNase A, 136, 144
281 selenol based catalysts, 126–127 small molecules redox potential, 116 thiol groups, 117 in vivo protein folding catalyst, 115–116 thiol based catalysts aromatic thiols (ArSH), 124 beta-mercaptoethanol, 120–121 BMC/vertase P, 123–124 CXC motif, 123 CXXC motif, 122–123 cystamine, 120–121 cysteine, 120–121 dithiothreitol, 119–120 gemini surfactants, 124–125 glutathione, 119 thiol/disulfide redox state, 145 underlying reaction disulfide bond, 113 nucleophile, 114 protein folding intermediates, 115 thiol-disulfide interchange reaction, 113, 114 yeast proteins, 146 P PapD-like chaperones, 170–171 Peptidyl-prolyl cis–trans isomerases (PPIases), 195 Porcine insulin precursor (PIP) model peptides, 67–68 time-dependent formation and distribution, 66 in vitro folding reaction, 65–66 Prion protein (PrP) heteronuclear 15NR2 and R1r relaxation rates, 248, 250 misfolding properties, 249 ribbon diagram, 247 secondary structure propensities, 248, 249 TSEs, 247 urea-denatured states, 248 Protein disulfide isomerase catalysis disulfide formation, 139–141 disulfide isomerization, 141–143 oxidative protein folding cysteines, 134 disulfide bonds, 134 ER, 134, 143 Ero1p, 143–144 GSH/GSSG, 144–145 LDL receptor, 145
282 Protein disulfide isomerase (cont.) native-cysteine, 137 noncovalent interactions, 136 onconase, 136 protein folding mechanisms, 134–135 RNase A, 136, 144 thiol/disulfide redox state, 145 yeast proteins, 146 structure b and b¢ domains, 139 a and a¢ domains, 138–139 eukaryotes, 138 internal thiore-doxin folds, 137 N- and C-terminal thioredoxin domains, 137 twisted U conformation, 138, 139 Protein disulfide isomerase (PDI), 9, 47 Protein disulphide formation denaturation reversibility “air oxidation,” 261 ORD, 259 polypeptides, 260 reoxidation process, 259 “ribonuclease reactivation,” 260 RNase S, 261 “scrambled ribonuclease,” 260 “the Anfinsen experiment,” 258 thiol/disulphide mixtures, 261 Dsb proteins, 270 high-resolution structural data, 270 non-covalent process, 257 oxidative refolding pathways BPTI, 262, 265–267 Cys30-Cys51 disulphide, 262 enzymic activity, 261 GSH/GSSG redox system, 262 kinetic characterisation, 263 purified PDI, 263, 266–267 PDI Dsb proteins, 270 Ero1, 271 recombinant DNA technology, 264–265 thioredoxin superfamily, 268–269 protein oxidative folding, 269–270 recombinant disulphide-bonded proteins, 258 Protien disulfide isomerase (PDI), 34 PrP. See Prion protein R Recombinant amphioxus insulin-like peptide (rAILP), 64 Reversed phase high-performance liquid chromatography (RP-HPLC), 7
Index RHStaple Bond catalytic function, 174–175 redox potentials, 175 RNase A minor regeneration pathways, 36 off-pathway intermediates (kinetically trapped intermediates), 37–38 oxidative folding initial stages, 29–31 postfolding steps, 35–36 rate-determining steps, 31–35 S Saccharomyces cerevisiae cells, 65 Selenoglutathione (GSeH) and selenocystamine, 126–127 Selenoxides, 128 Single chain insulins folding mechanism, 64 structural motif, 85 in vitro folding amphioxus insulin-like peptide, 71 behaviors of, 83–84 C-peptide, 85–87 human proinsulin, 68–71 insulin-like growth factor-1, 71–74 porcine insulin precursor, 65–68 Solid phase peptide synthesis (SPPS), 49 Stereochemistry, 175–176 T TAP. See Tick anticoagulant peptide TCEP. See Tris(2-carboxyethyl)phosphine TFA. See Trifluoroacetic acid Thiol based catalysts aromatic thiols (ArSH), 124 beta-mercaptoethanol, 120–121 BMC/vertase P, 123–124 CXC motif, 123 CXXC motif, 122–123 cystamine, 120–121 cysteine, 120–121 dithiothreitol, 119–120 gemini surfactants, 124–125 glutathione, 119 Thioredoxin reductase (Trr), 122 Tick anticoagulant peptide (TAP), 10, 95 Trans-1,2-bis(2-mercaptoacetamido) cyclohexane, 123–124 Transglutaminase 2, 173 Transmissible spongiform encephalopathies (TSEs), 247 Trapping method, 6
Index Trifluoroacetic acid (TFA), 75 Tris(2-carboxyethyl)phosphine (TCEP), 8 U Unfolding and refolding a-lactalbumin disulfide intact method, 101–102 disulfide scrambling, 101
283 on/off-pathway intermediates, 104 structural heterogeneity, 102 lysozyme, 100–101, 103 TAP and BPTI, 99–100 X X-isomers, 93