Preparation of fluorescence quenched libraries containing interchain disulphide bonds for studies of protein disulphide isomerases

Protein disulphide isomerase is an enzyme that catalyses disulphide redox reactions in proteins. In this paper, fluorogenic and interchain disulphide bond containing peptide libraries and suitable substrates, useful in the study of protein disulphide isomerase, are described. In order to establish the chemistry required for the generation of a split-synthesis library, two substrates containing an interchain disulphide bond, a fluoroescent probe and a quencher were synthesized. The library consists of a Cys residue flanked by randomized amino acid residues at both sides and the fluoroescent Abz group at the amino terminal. All the 20 natural amino acids except Cys were employed. The library was linked to PEGA‒beads via methionine so that the peptides could be selectively removed from the resin by cleavage with CNBr. A disulphide bridge was formed between the bead‒linked library and a peptide containing the quenching chromophore (Tyr(NO₂)) and Cys(pNpys) activated for reaction with a second thiol. The formation and cleavage of the interchain disulphide bonds in the library were monitored under a fluoroescence microscope. Substrates to investigate the properties of protein disulphide isomerase in solution were also synthesized. © 1998 European Peptide Society and John Wiley & Sons, Ltd.     

Purification of the STB enterotoxin of Escherichia coli and the role of selected amino acids on its secretion, stability and toxicity

The methanol-insoluble heat-stable enterotoxin of Escherichia coli (STB) was purified and characterized by automated Edman degradation and tryptic peptide analysis. The amino-terminal residue, Ser-24, confirmed that the first 23 amino acids inferred from the gene sequence were removed during translocation through the E. coli inner membrane. Tryptic peptide analysis coupled with automated Edman degradation revealed that disulphide bonds are formed between residues Cys-33 and Cys-71 and between Cys-44 and Cys-59. Oligonucleotide-directed mutagenesis performed on the STB gene demonstrated that disulphide bond formation does not precede translocation of the polypeptide through the inner membrane and that disulphide bridge formation is a periplasmic event; apparently, elimination of either of two disulphides of STB renders the molecule susceptible to periplasmic proteolysis. In addition, a loop defined by the Cys-44-Cys-59 bond contains at least two amino acids (Arg-52 and Asp-53) required for STB toxic activity.     

Topological analysis of the hepatitis B virus core particle by cysteine-cysteine cross-linking.

The nucleocapsid, or core particle, of hepatitis B virus is formed by 180 subunits of the core protein, which contains Cys at positions 48, 61, 107 and 183, the latter constituting the C terminus. Upon adventitious oxidation, some or all of these cysteine residues participate in the formation of disulphide bridges, leading to polymerization of the subunits within the particle. To utilize the cysteine residues as topological probes, we reduced the number of possible intersubunit crosslinks by replacing these residues individually, or in all combinations, by serine. A corresponding set of variants was constructed within the context of an assembly-competent core protein variant that lacks the highly basic C-terminal region. Analysis, by polyacrylamide gel electrophoresis under non-reducing conditions, of the oxidative crosslinking products formed by the wild-type and mutant proteins expressed in Escherichia coli, revealed a clear distinction between the three N-proximal, and the C-terminal Cys: N-proximal Cys formed intermolecular disulphide bonds only with other N-proximal cysteine residues, leading to dimerization. Cys48 and Cys61, in contrast to Cys107, could be crosslinked to the homologous cysteine residues in a second subunit, and are therefore located at the dimer interface. Cys 183 predominantly formed disulphide bonds with Cys183 in subunits other than those crosslinked by the N-proximal cysteine residues. Hence, the polymers generated by oxidation of the wild-type protein are S-S-linked dimeric N-terminal domains interconnected via Cys183/Cys183 disulphide bonds. The intermolecular crosslinks between the N-proximal cysteine residues were apparently the same in the C-terminally truncated and in the full-length proteins, corroborating the model in which the N-terminal domain and the C terminus of the HBV core protein form two distinct and structurally independent entities. The strong tendency of the N-terminal domain for dimeric interactions suggests that core protein dimers are the major intermediates in hepatitis B virus nucleocapsid assembly.     

Solid-phase Synthesis of ω-Agatoxin IVA,a P-type Calcium Channel Blocker

ω-Agatoxin IVA, isolated from the venom of funnel web spider Agelenopsis aperta, blocks potently and selectively P-type calcium channels. This toxin, composed of 48 amino acids and containing 8 cysteine residues, was synthesized by the solid-phase procedure. The Cys residues were protected by acetamidomethyl (Acm) groups which were removed by mercuric acetate. During treatment with mercuric acetate, a by-product was detected, involving modification of tryptophan residues by the Acm groups. This side reaction can be completely prevented by addition of an excess of tryptophan in the reaction medium during Acm deprotection. The resulting peptide was submitted to an oxidative refolding, in different conditions, in order to determine the most favourable protocol. After formation of the four disulphide bonds, the toxin was purified by successive preparative HPLC, on two different supports, and fully characterized by analytical HPLC, capillary electrophoresis, amino acid analysis, mass spectrometry and Edman degradation. It was found to block the P-type calcium channel with a similar biological potency as described for the natural product.     

Interactions between cysteine residues as probes of protein conformation: the disulphide bond between Cys-14 and Cys-38 of the pancreatic trypsin inhibitor.

The rate constants for the reversible reduction by dithiothreitol of the disulphide bond linking eysteines 14 and 38 of the native pancreatic trypsin inhibitor have been measured. The results are consistent with this disulphide bond being formed as the last step in refolding of the fully reduced inhibitor.The rates of reduction of several model linear disulphides have been measured under the same conditions. A linear relationship was found between the rate of reduction and the ionization tendency of the thiol group generated.The apparent pK value of the thiol groups of cysteines 14 and 38 of the selectively reduced inhibitor were measured by the pH-dependence of their rate of alkylation with iodoacetamide. The rate of reduction of the disulphide bond between these two residues was very close to that predicted from the model compounds.The kinetics and thermodynamics of disulphide bond formation and breakage are shown to be useful for experimental determination of conformational transitions in proteins and model peptides.     

Conformational restrictions on the pathway of folding and unfolding of the pancreatic trypsin inhibitor

The kinetic roles of the partially folded, intermediate protein species with two disulphide bonds in folding and unfolding of the pancreatic trypsin inhibitor have been investigated further. Formation of a second disulphide bond between Cys5 and Cys55 during refolding of the reduced inhibitor, which would yield the species with the 30–51 and 5–55 disulphide bonds and, possibly, the native-like conformation of the protein, is not significant. Instead, three other second disulphide bonds (5–14, 5–38 and 14–38) are formed approximately 10⁵ times more readily, but each of these two-disulphide species then rearranges intramolecularly to the native-like, two-disulphide intermediate. Therefore, the reduced protein does not simply form sequentially the three disulphide bonds of the native state. Unfolding of the native state takes place by the reverse of this process.The kinetic importance for folding and unfolding of this transition between two-disulphide intermediates under the conditions used here was illustrated experimentally by a modified form of the inhibitor in which the thiols of Cys14 and Cys38 were blocked irreversibly. In the folded conformation, this modified protein is more stable to unfolding than normal, but after unfolding cannot readily regain the native-like conformation, because Cys14 or Cys38 are required to be involved in disulphide bonds during the interconversion of the two-disulphide intermediates.Some conception of the conformational transitions that take place at each stage of the folding transition may be inferred from the relative propensities of the six cysteine residues to make or rearrange disulphide bonds. It is concluded that the inhibitor probably does not refold by sequential adoption of the native conformation by the unfolded polypeptide chain. Instead, it appears that essentially all elements of the native conformation are attained simultaneously in the final stage of folding, within an unstable and flexible, yet relatively compact, form of the entire polypeptide chain produced by weak interactions between groups distant in the primary structure.     

Periplasmic disulphide bond formation is essential for cellulase secretion by the plant pathogen Erwinia chrysanthemi

Secretion to the cell exterior of cellulase EGZ and of at least six pectinases enables the Gram-negative Erwinia chrysanthemi to cause severe plant disease. The C-terminal cellulose-binding domain (CBD) of EGZ was found to contain a disulphide bond which forms, in the periplasm, between residues Cys-325 and Cys-382. Dithiothreitol (DTT)-treatment of native EGZ showed that the disulphide bond was dispensable, both for catalysis and cellulose binding. Adding DTT to E. chrysanthemi cultures led to immediate arrest of secretion of EGZ which accumulated in the periplasm where the CBD was eventually proteolysed. Site-directed mutagenesis that affected Cys residues involved in disulphide bond formation resulted in molecules that were catalytically active and able to bind to cellulose but were no longer secreted, Instead they accumulated in the periplasm. Interestingly, the region around EGZ Cys-325 is conserved in two pectinases secreted by the same pathway as EGZ. We conclude that the conserved Cys, and possibly adjacent residues, bear essential information for EGZ to be secreted and that periplasmic disulphide bond formation is an obligatory step which provides a pre-folded functional form of EGZ with secretion competence.     

Cosolvent-assisted oxidative folding of a bicyclic alpha-conotoxin ImI.

alpha-Conotoxin ImI is a 12-amino acid peptide, found in the venom of the marine snail Conus imperialis. This conotoxin is a selective antagonist of alpha7 nicotinic acetylcholine receptors. To produce biologically active alpha-ImI, disulfide bonds must be formed between Cys2-Cys8 and Cys3-Cys12. Oxidative folding of bicyclic conotoxins, such as alpha-ImI, has been traditionally achieved using two-step oxidation protocols with orthogonal protection on two native pairs of cysteines. In this work, two alternative oxidation protocols were explored: (1) the recently described one-pot oxidation of t-butyl/4-methylbenzyl protected Cys pairs and (2) direct oxidative folding. In contrast to the first method, the latter one resulted in high yields of correctly folded alpha-ImI. The addition of organic cosolvents, such as methanol, ethanol or isopropanol into the folding mixture significantly increased the accumulation of the native peptide. This effect was also observed for another conotoxin, alpha-PnIA. It is suggested that cosolvent-assisted direct oxidation might be of general use for other bicyclic alpha-conotoxins, but efficiency should be assessed on a case-by-case basis.     

The single-disulphide intermediates in the refolding of reduced pancreatic trypsin inhibitor

The intermediate species with one disulphide bond in the renaturation of reduced pancreatic trypsin inhibitor have been trapped, isolated, and the Cys residues involved in the disulphide bonds determined. Approximately half the intermediate species had the disulphide bond between Cys-30 and 51, a disulphide bond also present in the native inhibitor. The next most predominant species, representing one-quarter of the total, had a disulphide bond between Cys-5 and 30, and two more minor species involving Cys-30 and 55 and Cys-5 and 51 were detected; these disulphide bonds are not present in the native inhibitor.The nature of the disulphide bonds present are concluded to reflect primarily the conformational forces acting at this stage of folding, which may be primarily interactions between segments with propensities for secondary structure, either helices or β-sheet. The general importance of such interactions in protein folding is discussed.     

Conotoxin GI: disulfide bridges, synthesis, and preparation of iodinated derivatives

The 13 amino acid toxic peptide from the marine snail Conus geographus, conotoxin GI, blocks the acetylcholine receptor at the neuromuscular junction. In this report, we describe a method for analyzing disulfide bonding in nanomole amounts of small cystine-rich peptides. The procedure involves partial reduction and a double-label alkylation of cysteine residues. Using this method, we show that the natural conotoxin GI has a (2-7, 3-13) disulfide configuration. The structure of conotoxin GI has been confirmed by chemical synthesis. The preparation and purification of molecularly homogeneous, iodinated derivatives of this toxin are also described. All derivatives, including the [diiodohistidine,diiodotyrosine]conotoxin GI, retained at least half of the biological activity of unmodified toxin. Since the tetraiodinated toxin, which is greater than 25% by weight iodine, retains considerable toxicity, unmodified histidine and tyrosine residues in conotoxin GI are not crucial for biological activity.     

Accessibilities and reactivities of cysteine thiols during refolding of reduced bovine pancreatic trypsin inhibitor

The experimental basis of the pathway of refolding of reduced bovine pancreatic trypsin inhibitor that accompanies disulphide bond formation is explained in the light of a recent suggestion that the inability of certain Cys residues to form disulphide bonds could be explained simply by their thiol groups being inaccessible to disulphide reagents. This explanation is not valid, because part of the experimental evidence for inability to form disulphides is that the Cys residues accumulate as mixed-disulphides with the reagent. That these thiol groups are observed to react normally with the reagent, and with iodoacetic acid, is direct positive proof that they were not inaccessible or otherwise unreactive. The experimentally determined refolding pathway accurately reflects the energetics of the protein folding transitions and is consistent with all general observations of the folding transitions of other small proteins, whether or not disulphide bond formation is involved.     

Disulfide structures of highly bridged peptides: a new strategy for analysis.

A new approach is described for analyzing disulfide linkage patterns in peptides containing tightly clustered cystines. Such peptides are very difficult to analyze with traditional strategies, which require that the peptide chain be split between close or adjacent Cys residues. The water-soluble tris-(2-carboxyethyl)-phosphine (TCEP) reduced disulfides at pH 3, and partially reduced peptides were purified by high performance liquid chromatography with minimal thiol-disulfide exchange. Alkylation of free thiols, followed by sequencer analysis, provided explicit assignment of disulfides that had been reduced. Thiol-disulfide exchange occurred during alkylation of some peptides, but correct deductions were still possible. Alkylation competed best with exchange when peptide solution was added with rapid mixing to 2.2 M iodoacetamide. Variants were developed in which up to three alkylating agents were used to label different pairs of thiols, allowing a full assignment in one sequencer analysis. Model peptides used included insulin (three bridges, intra- and interchain disulfides; -Cys.Cys- pair), endothelin and apamin (two disulfides; -Cys.x.Cys- pair), conotoxin GI and isomers (two disulfides; -Cys.Cys- pair), and bacterial enterotoxin (three bridges within 13 residues; two -Cys.Cys- pairs). With insulin, all intermediates in the reduction pathway were identified; with conotoxin GI, analysis was carried out successfully for all three disulfide isomers. In addition to these known structures, the method has been applied successfully to the analysis of several previously unsolved structures of similar complexity. Rates of reduction of disulfide bonds varied widely, but most peptides did not show a strongly preferred route for reduction.     

Intermediates in the refolding of reduced ribonuclease A.

The intermediates with one, two, three or four disulphide bonds which accumulate during unfolding of native ribonuclease and refolding of the reduced protein have been trapped by rapid alkylation with iodoacetate and separated by ionexchange chromatography. They have been characterized to varying extents by their enzymic activity, electrophoretic mobility through polyacrylamide gels, disulphide bonds between cysteine residues, the environments of the six tyrosine residues as indicated by ultraviolet absorption and fluorescence spectra, interaction with antibodies directed against either the trapped unfolded reduced protein or the native folded protein, and for the disruption by urea of any stable conformation producing a change in molecular shape.Correctly refolded ribonuclease was indistinguishable from the original native protein, but virtually all the intermediates with up to four disulphide bonds formed directly from the reduced protein were enzymically inactive and unfolded by these criteria. Unfolding of native ribonuclease was an all-or-none transition to the fully reduced protein, with no accumulation of disulphide intermediates. The intermediates in refolding are separated from the fully folded state by the highest energy barrier in the folding transition; they may be considered rapidly interconvertible, relatively unstable microstates of the unfolded protein. The measured elements of the final conformation are not acquired during formation of the first three disulphide bonds, but appear simultaneously with formation of the fourth native disulphide bond.These observations with ribonuclease are qualitatively similar to those made previously in greater detail with pancreatic trypsin inhibitor and suggest a possible general pattern for the kinetic process of protein unfolding and refolding.     

Initial disulfide formation steps in the folding of an omega-conotoxin

To determine whether the native disulfides of omega-conotoxins are preferentially stabilized early in the folding of these small proteins, the rates and equilibria for disulfide formation were measured for three analogues of omega-conotoxin MVIIA. In each analogue, one of the three pairs of disulfide-bonded Cys residues was replaced with Ala residues, leaving four Cys residues that can form six intermediates with one disulfide and three species with two disulfides. For each analogue, all of the disulfide-bonded species were identified, and the equilibrium constants for forming the individual species via exchange with oxidized and reduced glutathione were measured. These equilibrium constants represent effective concentrations of the Cys thiols and ranged from 0.01 to 0.4 M in the fully reduced protein. There was little or no preference for forming the native disulfides, and the equilibria for forming the first and second disulfides decreased only slightly upon the addition of 8 M urea. The data for the four-Cys analogues, together with equilibrium data for the six-Cys form, were also used to estimate effective concentrations for forming a third disulfide once two native disulfides are present. These effective concentrations were approximately 100 and 10 M in the presence of 0 and 8 M urea, respectively. The results indicate that there is little or no preferential formation of native interactions in the folding of these molecules until two disulfides have formed, after which there is a high degree of cooperativity among the native interactions.     

Assignment of disulphide bonds in synthetic endothelin-1 isomers by fast atom bombardment mass spectrometry.

Endothelin-1 (ET-1), a 21-residue vasoconstrictor peptide possessing four cysteinyl residues at positions 1, 3, 11 and 15, was synthesized by random oxidation of a tetrahydro-ET-1. On reverse-phase high-performance liquid chromatography, crude product was shown to be a mixture of two disulphide isomers. A method was developed to determine the disulphide structure of the isomers. The method consisted of (a) limited digestion with chymotrypsin, (b) cleavage with cyanogen bromide and (c) manual Edman degradation. Through this procedure, each isomer afforded specific fragments containing a single disulphide bond, which were identified by fast atom bombardment mass spectrometry. Isomer 1, the minor component, afforded a fragment containing Cys 3 and Cys 15, and isomer 2, the major component, afforded fragments containing Cys 3 and Cys 11. Since little disulphide exchange was observed, it could be concluded clearly that the disulphide bond pairs in isomer 1 were Cys 1-Cys 11 and Cys 3-Cys 15, while those in isomer 2 were Cys 1-Cys 15 and Cys 3-Cys 11 (the same as natural ET-1). The procedure was successfully applied to two synthetic analogues, [Gly18]-ET-1 and [Pro16]-ET-1.     

The Role of zinc in the disulphide stress-regulated anti-sigma factor RsrA from Streptomyces coelicolor

The regulation of disulphide stress in actinomycetes such as Streptomyces coelicolor is known to involve the zinc-containing anti-sigma factor RsrA that binds and inactivates the redox-regulated sigma factor sigmaR. However, it is not known how RsrA senses disulphide stress nor what role the metal ion plays. Using in vitro assays, we show that while zinc is not required for sigmaR binding it is required for functional anti-sigma factor activity, and that it plays a critical role in modulating the reactivity of RsrA cysteine thiol groups towards oxidation. Apo-RsrA is easily oxidised and, while the Zn-bound form is relatively resistant, the metal ion is readily expelled when the protein is treated with strong oxidants such as diamide. We also show, using a combination of proteolysis and mass spectrometry, that the first critical disulphide to form in RsrA involves Cys11 and one of either Cys41 or Cys44, all previously implicated in metal binding. Circular dichroism spectroscopy was used to follow structural changes during oxidation of RsrA, which indicated that concomitant with formation of this critical disulphide bond is a major restructuring of the protein where its alpha-helical content increases. Our data demonstrate that RsrA can only bind sigmaR in the reduced state and that this state is stabilised by zinc. Redox stress induces disulphide bond formation amongst zinc-ligating residues, expelling the metal ion and stabilising a structure incapable of binding the sigma factor.     

(14-38, 30-51) double-disulphide intermediate in folding of bovine pancreatic trypsin inhibitor: a two-dimensional 1H nuclear magnetic resonance study

An analogue of the BPTI folding intermediate that contains only the disulphide bonds between Cys14 and Cys38 and between Cys30 and Cys51 has been prepared in Escherichia coli by protein engineering methods. The other two Cys residues of native BPTI (at positions 5 and 55) have been replaced by Ser. Essentially complete proton resonance assignments of the analogue were obtained by employing two-dimensional 1H nuclear magnetic resonance techniques. The intermediate has a more extended conformation in the N-terminal (residues 1 to 7) region and there are other differences in the C-terminal (residues 55 to 58) region. The remainder of the protein is substantially identical to native BPTI. The conformational properties of the analogue can explain several aspects of the kinetic role that the normal (14-38, 30-51) intermediate plays in the folding of BPTI.     

Evidence for intramolecular disulfide bond shuffling in the folding of mutant human lysozyme

Our previous results using the Saccharomyces cerevisiae secretion system suggest that intramolecular exchange of disulfide bonds occurs in the folding pathway of human lysozyme in vivo (Taniyama, Y., Yamamoto, Y., Kuroki, R., and Kikuchi, M. (1990) J. Biol. Chem. 265, 7570-7575). Here we report on the results of introducing an artificial disulfide bond in mutants with 2 cysteine residues substituting for Ala83 and Asp91. The mutant (C83/91) protein was not detected in the culture medium of the yeast, probably because of incorrect folding. Thereupon, 2 cysteine residues Cys77 and Cys95 were replaced with Ala in the mutant C83/91, because a native disulfide bond Cys77-Cys95 was found not necessary for correct folding in vivo (Taniyama, Y., Yamamoto, Y., Nakao, M., Kikuchi, M., and Ikehara, M. (1988) Biochem. Biophys. Res. Commun. 152, 962-967). The resultant mutant (AC83/91) was secreted as two proteins (AC83/91-a and AC83/91-b) with different specific activities. Amino acid and peptide mapping analyses showed that two glutathiones appeared to be attached to the thiol groups of the cysteine residues introduced into AC83/91-a and that four disulfide bonds including an artificial disulfide bond existed in the AC83/91-b molecule. The presence of cysteine residues modified with glutathione may indicate that the non-native disulfide bond Cys83-Cys91 is not so easily formed as a native disulfide bond. These results suggest that the introduction of Cys83 and Cys91 may act to suppress the process of native disulfide bond formation through disulfide bond interchange in the folding of human lysozyme.     

Effect of cysteine residues on the activity of arginyl-tRNA synthetase from Escherichia coli.

Arginyl-tRNA synthetase (ArgRS) from Escherichia coli (E. coli) contains four cysteine residues. In this study, the role of cysteine residues in the enzyme has been investigated by chemical modification and site-directed mutagenesis. Titration of sulfhydryl groups in ArgRS by 5, 5'-dithiobis(2-nitro benzoic acid) (DTNB) suggested that a disulfide bond was not formed in the enzyme and that, in the native condition, two DTNB-sensitive cysteine residues were located on the surface of ArgRS, while the other two were buried inside. Chemical modification of the native enzyme by iodoacetamide (IAA) affected only one DTNB-sensitive cysteine residue and resulted in 50% loss of enzyme activity, while modification by N-ethylmeimide (NEM) affected two DTNB-sensitive residues and caused a complete loss of activity. These results, when combined with substrate protection experiments, suggested that at least the two cysteine residues located on the surface of the molecule were directly involved in substrates binding and catalysis. However, changing Cys to Ala only resulted in slight loss of enzymatic activity and substrate binding, suggesting that these four cysteine residues in E. coli ArgRS were not essential to the enzymatic activity. Moreover, modifications of the mutant enzymes indicated that the two DTNB- and NEM-sensitive residues were Cys(320) and Cys(537) and the IAA-sensitive was Cys(320). Our study suggested that inactivation of E. coli ArgRS by sulfhydryl reagents is a result of steric hindrance in the enzyme.     

Factors governing selective formation of specific disulfides in synthetic variants of alpha-conotoxin

alpha-Conotoxin GI is a snail toxin protein consisting of 13 amino acids cross-linked by 2 intramolecular disulfide bridges. This toxin is an antagonist of acetylcholine receptors. The native sequence has been synthesized, along with nine additional variants in which non-cysteine residues are replaced by alanine or the cysteine positions are altered. Each reduced peptide has been oxidized by reaction with oxygen or glutathione both in a folding buffer and in 6 M guanidine hydrochloride. Purified products of oxidation have been characterized with respect to molecular weights and the positions of disulfides. The four cysteines in conotoxin can form two intramolecular disulfides in three different combinations. Relative yields of each of the three isomers have been determined, thereby permitting evaluation of the roles of non-cysteine residues and cysteine placements in the folding of conotoxin. Cysteine positions dominate factors directing formation of the nativelike isomer in a manner that may be predicted from equilibrium constants for loop formation in model peptides containing two cysteines. Alanine substitutions at several positions which are conserved in naturally occurring conotoxins affect the discrimination between the two most favored disulfide arrangements. Substitutions at three nonconserved positions have no structural effect on isomer yields. It therefore is possible to vary these latter three positions in a manner which might help to generate a functional binding surface which is complementary to receptors in the specific prey of a particular species of snail, without affecting the toxin's folding.