In vitro insulin refolding: Characterization of the intermediates and the putative folding pathway

The in vitro refolding process of the double-chain insulin was studied based on the investigation of in vitro single-chain insulin refolding. Six major folding intermediates, named P1A, P2B, P3A, P4B, P5B, and P6B, were captured during the folding process. The refolding experiments indicate that all of these intermediates are on-pathway. Based on these intermediates and the formation of hypothetic transients, we propose a two-stage folding pathway of insulin. (1) At the early stage of the folding process, the reduced A chain and B chain individually formed the intermediates: two A chain intermediates (P1A and P3A), and four B chain intermediates (P2B, P4B, P5B, and P6B). (2) In the subsequent folding process, transient I was formed from P3A through thiol/disulfide exchange reaction; then, transients II and III, each containing two native disulfides, were formed through the recognition and interaction of transient I with P4B or P6B and the thiol group's oxidation reaction mainly using GSSG as oxidative reagent; finally, transients II and III, through thiol/mixture disulfide exchange reaction, formed the third native disulfide of insulin to complete the folding.     

<Emphasis Type="Italic">In vitro</Emphasis> insulin refolding: Characterization of the intermediates and the putative folding pathway

The in vitro refolding process of the double-chain insulin was studied based on the investigation of in vitro single-chain insulin refolding. Six major folding intermediates, named P1A, P2B, P3A, P4B, P5B, and P6B, were captured during the folding process. The refolding experiments indicate that all of these intermediates are on-pathway. Based on these intermediates and the formation of hypothetic transients, we propose a two-stage folding pathway of insulin. (1) At the early stage of the folding process, the reduced A chain and B chain individually formed the intermediates: two A chain intermediates (P1A and P3A), and four B chain intermediates (P2B, P4B, P5B, and P6B). (2) In the subsequent folding process, transient I was formed from P3A through thiol/disulfide exchange reaction; then, transients II and III, each containing two native disulfides, were formed through the recognition and interaction of transient I with P4B or P6B and the thiol group’s oxidation reaction mainly using GSSG as oxidative reagent; finally, transients II and III, through thiol/mixture disulfide exchange reaction, formed the third native disulfide of insulin to complete the folding.     

In vitro refolding of human proinsulin. Kinetic intermediates,putative disulfide-forming pathway folding initiation site,and potential role of C-peptide in folding process

Human insulin is a double-chain peptide that is synthesized in vivo as a single-chain human proinsulin (HPI). We have investigated the disulfide-forming pathway of a single-chain porcine insulin precursor (PIP). Here we further studied the folding pathway of HPI in vitro. While the oxidized refolding process of HPI was quenched, four obvious intermediates (namely P1, P2, P3, and P4, respectively) with three disulfide bridges were isolated and characterized. Contrary to the folding pathway of PIP, no intermediates with one- or two-disulfide bonds could be captured under different refolding conditions. CD analysis showed that P1, P2, and P3 retained partially structural conformations, whereas P4 contained little secondary structure. Based on the time-dependent distribution, disulfide pair analysis, and disulfide-reshuffling process of the intermediates, we have proposed that the folding pathway of HPI is significantly different from that of PIP. These differences reveal that the C-peptide not only facilitates the folding of HPI but also governs its kinetic folding pathway of HPI. Detailed analysis of the molecular folding process reveals that there are some similar folding mechanisms between PIP and HPI. These similarities imply that the initiation site for the folding of PIP/HPI may reside in the central alpha-helix of the B-chain. The formation of disulfide A20-B19 may guide the transfer of the folding information from the B-chain template to the unstructured A-chain. Furthermore, the implications of this in vitro refolding study on the in vivo folding process of HPI have been discussed.     

Structure and heterogeneity of the one- and two-disulfide folding intermediates of tick anticoagulant peptide

Tick anticoagulant peptide (TAP) is a factor Xa-specific inhibitor and is structurally homologous to bovine pancreatic trypsin inhibitor (BPTI). The fully reduced TAP refolds spontaneously to form the native structure under a wide variation of redox buffers. The folding intermediates of TAP consist of at least 22 fractions of one-disulfide, two-disulfide, and three-disulfide scrambled isomers. Three species of well-populated one- and two-disulfide intermediates were isolated and structurally characterized. The predominant one-disulfide species contains TAP-(Cys33—Cys55). Two major two-disulfide isomers were TAP-(Cys33—Cys55, Cys15—Cys39) and TAP-(Cys33—Cys55, Cys5—Cys39). Both Cys33—Cys55 and Cys15—Cys39 are native disulfides of TAP. These three species are structural counterparts of BPTI-(Cys30—Cys51), BPTI-(Cys30—Cys51, Cys14—Cys38), and BPTI-(Cys30—Cys51,Cys5—Cys38), which have been shown to be the major intermediates of BPTI folding. In addition, time-course-trapped folding intermediates of TAP, consisting of about 47% one-disulfide species and 30% two-disulfide species, were collectively digested with thermolysin, and fragmented peptides were analyzed by Edman sequencing and mass spectrometry in order to characterize the disulfide-containing peptides. Among the 15 possible single-disulfide pairings of TAP, 10 (2 native and 8 nonnative) were found as structural components of its one- and two-disulfide folding intermediates. The results demonstrate that the major folding intermediates of TAP bear structural homology to those of BPTI. However, the folding pathway of TAP differs from that of BPTI by (a) a higher degree of heterogeneity of one- and two-disulfide intermediates and (b) the presence of three-disulfide scrambled isomers as folding intermediates. Mechanism(s) that may account for these diversities are proposed and discussed.     

Oxidative folding of bovine pancreatic ribonuclease A: insight into the overall catalysis of the refolding pathway by phosphate

The effects of the strong stabilizing anion, phosphate, on the oxidative folding of bovine pancreatic ribonuclease A were examined. Phosphate was found to catalyze several steps involved in the oxidative folding process at pH 8.0 and 25°C, resulting in an increase in the rate of pre-equilibration of unstructured species on the folding pathway. In the presence of 400 mM phosphate, the overall increase in the rate of regeneration of native protein was caused primarily by the increased formation and stabilization of tertiary structure in the nativelike intermediates, des-[40-95] and des-[65-72], involved in the rate-determining step. Based on the regeneration of native protein and the stability of Cys Ala substituted mutant analogs of the des-species, (C40A, C95A) and (C65A, C72A), it is suggested that the primary role of phosphate is to catalyze the overall regeneration of native protein through nonspecific electrostatic and hydrogen-bonding effects on the protein and solvent.     

A study of intermediates involved in the folding pathway for recombinant human macrophage colony-stimulating factor (M-CSF): evidence for two distinct folding pathways.

The folding pathway for a 150-amino acid recombinant form of the dimeric cytokine human macrophage colony-stimulating factor (M-CSF) has been studied. All 14 cysteine residues in the biologically active homodimer are involved in disulfide linkages. The structural characteristics of folding intermediates blocked with iodoacetamide reveal a rapid formation of a small amount of a non-native dimeric intermediate species followed by a slow progression via both monomeric and dimeric intermediates to the native dimer. The transition from monomer to fully folded dimer is complete within 25 h at room temperature at pH 9.0. The blocked intermediates are stable under conditions of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and thus represent various dimeric and folded monomeric species of the protein with different numbers of disulfide bridges. Peptide mapping and electrospray ionization mass spectrometry revealed that a folded monomeric species of M-CSF contained three of the four native disulfide bridges, and this folded monomer also showed some biological activity in a cell-based assay. The results presented here strongly suggest that M-CSF can fold via two different pathways, one involving monomeric intermediates and another involving only dimeric intermediates.     

Recombinant human insulin: X simplification of folding of the biotechnological precursor of recombinant human insulin

The folding of biotechnological precursor of the human insulin precursor was carried out from solubilized inclusion bodies without a preliminary oxidizing or reducing its Cys residues. The inclusion bodies were dissolved in 8 M urea with the addition of 10 mM 2-mercaptoethanol. Hydrophobic cell components were removed from the solution by passing through a neutral weakly hydrophobic sorbent, the solution was five times diluted and refolded upon addition of 0.3 mM cystine for initiation of disulfide rearrangement. The presence of nucleic acids and cell protein impurities does not affect the folding efficiency. The resulting precursor of folded human insulin was purified by metal-chelate affinity chromatography and converted into insulin by two-stage enzymatic cleavage.     

Mutational analysis of the BPTI folding pathway: I. Effects of aromatic-->leucine substitutions on the distribution of folding intermediates.

The roles of aromatic residues in determining the folding pathway of bovine pancreatic trypsin inhibitor (BPTI) were analyzed mutationally by examining the distribution of disulfide-bonded intermediates that accumulated during the refolding of protein variants in which tyrosine or phenylalanine residues were individually replaced with leucine. The eight substitutions examined all caused significant changes in the intermediate distribution. In some cases, the major effect was to decrease the accumulation of intermediates containing two of the three disulfides found in the native protein, without affecting the distribution of earlier intermediates. Other substitutions, however, led to much more random distributions of the intermediates containing only one disulfide. These results indicate that the individual residues making up the hydrophobic core of the native protein make clearly distinguishable contributions to conformation and stability early in folding: The early distribution of intermediates does not appear to be determined by a general hydrophobic collapse. The effects of the substitutions were generally consistent with the structures of the major intermediates determined by NMR studies of analogs, confirming that the distribution of disulfide-bonded species is determined by stabilizing interactions within the ordered regions of the intermediates. The plasticity of the BPTI folding pathway implied by these results can be described using conformational funnels to illustrate the degree to which conformational entropy is lost at different stages in the folding of the wild-type and mutant proteins.     

Simulation and experiment conspire to reveal cryptic intermediates and a slide from the nucleation-condensation to framework mechanism of folding

There is a change from three-state to two-state kinetics of folding across the homeodomain superfamily of proteins as the mechanism slides from framework to nucleation-condensation. The tendency for framework folding in this family correlates with inherent helical propensity. The cellular myeloblastis protein (c-Myb) falls in the mechanistic transition region. An earlier, preliminary report of protein engineering experiments and molecular dynamics simulations (MD) showed that the folding mechanism for this protein has aspects of both the nucleation-condensation and framework models. In the more in-depth analysis of the MD trajectories presented here, we find that folding may be attributed to both of these mechanisms in different regions of the protein. The folding of the loop, middle helix, and turn is best described by nucleation-condensation, whereas folding of the N and C-terminal helices may be described by the framework model. Experimentally, c-Myb folds by apparent two-state kinetics, but the MD simulations predict that the kinetics hide a high-energy intermediate. We stabilized this hypothetical folding intermediate by deleting a residue (P174) in the loop between its second and third helices, and the mutant intermediate is long-lived in the simulations. Equilibrium and kinetic experiments demonstrate that folding of the DeltaP174 mutant is indeed three-state. The presence and shape of the intermediate observed in the simulations were confirmed by small angle X-ray scattering experiments.     

Protein folding, misfolding, and refolding of therapeutic proteins

Substantial progress has been made towards understanding the folding mechanisms of proteins in vitro and in vivo even though the general rules governing such folding events remain unknown. This paper reviews current folding models along with experimental approaches used to elucidate the folding pathways. Protein misfolding is discussed in relation to disease states, such as amyloidosis, and the recent findings on the mechanism of converting normally soluble proteins into amyloid fibrils through the formation of intermediates provide an insight into understanding the pathogenesis of amyloid formation and possible clues for the development of therapeutic treatments. Finally, some commonly adopted refolding strategies developed over the past decade are summarized.     

Folding of the yeast prion protein Ure2: kinetic evidence for folding and unfolding intermediates.

The Saccharomyces cerevisiae non-Mendelian factor [URE3] propagates by a prion-like mechanism, involving aggregation of the chromosomally encoded protein Ure2. The N-terminal prion domain (PrD) of Ure2 is required for prion activity in vivo and amyloid formation in vitro. However, the molecular mechanism of the prion-like activity remains obscure. Here we measure the kinetics of folding of Ure2 and two N-terminal variants that lack all or part of the PrD. The kinetic folding behaviour of the three proteins is identical, indicating that the PrD does not change the stability, rates of folding or folding pathway of Ure2. Both unfolding and refolding kinetics are multiphasic. An intermediate is populated during unfolding at high denaturant concentrations resulting in the appearance of an unfolding burst phase and roll-over in the denaturant dependence of the unfolding rate constants. During refolding the appearance of a burst phase indicates formation of an intermediate during the dead-time of stopped-flow mixing. A further fast phase shows second-order kinetics, indicating formation of a dimeric intermediate. Regain of native-like fluorescence displays a distinct lag due to population of this on-pathway dimeric intermediate. Double-jump experiments indicate that isomerisation of Pro166, which is cis in the native state, occurs late in refolding after regain of native-like fluorescence. During protein refolding there is kinetic partitioning between productive folding via the dimeric intermediate and a non-productive side reaction via an aggregation prone monomeric intermediate. In the light of this and other studies, schemes for folding, aggregation and prion formation are proposed.     

Multiple parallel-pathway folding of proline-free Staphylococcal nuclease

When a protein exhibits complex kinetics of refolding, we often ascribe the complexity to slow isomerization events in the denatured protein, such as cis/trans isomerization of peptidyl prolyl bonds. Does the complex folding kinetics arise only from this well-known reason? Here, we have investigated the refolding of a proline-free variant of staphylococcal nuclease by stopped-flow, double-jump techniques, to examine the folding reactions without the slow prolyl isomerizations. As a result, the protein folds into the native state along at least two accessible parallel pathways, starting from a macroscopically single denatured-state ensemble. The presence of intermediates on the individual folding pathways has revealed the existence of multiple parallel pathways, and is characterized by multi-exponential folding kinetics with a lag phase. Therefore, a "single" amino acid sequence can fold along the multiple parallel pathways. This observation in staphylococcal nuclease suggests that the multiple folding may be more general than we have expected, because the multiple parallel-pathway folding cannot be excluded from proteins that show simpler kinetics.     

Structure of stable protein folding intermediates by equilibrium phi-analysis: the apoflavodoxin thermal intermediate

Protein intermediates in equilibrium with native states may play important roles in protein dynamics but, in cases, can initiate harmful aggregation events. Investigating equilibrium protein intermediates is thus important for understanding protein behaviour (useful or pernicious) but it is hampered by difficulties in gathering structural information. We show here that the phi-analysis techniques developed to investigate transition states of protein folding can be extended to determine low-resolution three-dimensional structures of protein equilibrium intermediates. The analysis proposed is based solely on equilibrium data and is illustrated by determination of the structure of the apoflavodoxin thermal unfolding intermediate. In this conformation, a large part of the protein remains close to natively folded, but a 40 residue region is clearly unfolded. This structure is fully consistent with the NMR data gathered on an apoflavodoxin mutant designed specifically to stabilise the intermediate. The structure shows that the folded region of the intermediate is much larger than the proton slow-exchange core at 25 degrees C. It also reveals that the unfolded region is made of elements whose packing surface is more polar than average. In addition, it constitutes a useful guide to rationally stabilise the native state relative to the intermediate state, a far from trivial task.     

Sequence determinants of a protein folding pathway

Local hydrophobic collapse of the polypeptide chain and transient long-range interactions in unfolded states of apomyoglobin appear to occur in regions of the amino acid sequence which, upon folding, bury an above-average area of hydrophobic surface. To explore the role of these interactions in protein folding, we prepared and characterized apomyoglobins with compensating point mutations designed to change the average buried surface area in local regions of the sequence, while conserving as much as possible the constitution of the hydrophobic core. The behavior of the mutants in quench-flow experiments to determine the folding pathway was exactly as predicted by the changes in the buried surface area parameter calculated from the amino acid sequence. In addition, spin label experiments with acid-unfolded mutant apomyoglobin showed that the transient long-range contacts that occur in the wild-type protein are abolished in the mutant, while new contacts are observed between areas that now have above-average buried surface area. We conclude that specific groupings of amino acid side-chains, which can be predicted from the sequence, are responsible for early hydrophobic interactions in the first phase of folding in apomyoglobin, and that these early interactions determine the subsequent course of the folding process.     

Mechanism of protein folding: I. General considerations and refolding of myoglobin

To explain the rapidity of the process of protein folding, we cite two aspects of hydrophobic interaction: its long-range nature and the specificity of pairing after the formation of secondary structures. These two factors, when incorporated with the growth-type mechanism, can determine the folding pathway of proteins. This mechanism is applied to myoglobin. Appropriate introduction of side chains of amino acid residues and the heme group attached to His 93 yield a refolded tertiary structure that is in good agreement with the native structure.     

Folding barrier in horse cytochrome c: support for a classical folding pathway

Native-state structures and conformations of ferrocytochrome c, nitrosylcytochrome c, and carbonmonoxycytochrome c are very similar. They are, however, immensely different from each other in terms of thermodynamic stability. The dramatic destabilization of ferrocytochrome c to the extent of 12 kcal mol(-1) produces no effect on the folding rate, and this is so in spite of the fact that all three test-tube variants fold in an apparent two-state manner. For all three proteins the folding barrier is early in time, sizable in energy, and is of the same magnitude (approximately 6.5 kcal mol(-1)). These results raise some challenges to the "new view" of protein folding. An early transition state, the search for which consumes most of the observed folding time, is suggested.     

Redox conditions for stimulation of in vitro folding and assembly of the glycoprotein hormone chorionic gonadotropin

The formation of native disulfide bonds during in vitro protein folding can be limiting in obtaining biologically active proteins. Thus, optimization of redox conditions can be critical in maximizing the yield of renatured, recombinant proteins. We have employed a folding model, that of the beta subunit of human chorionic gonadotropin (hCG- beta), to investigate in vitro oxidation conditions that facilitate the folding of this protein, and have compared the in vitro rates obtained with the rate of folding that has been observed in intact cells. Two steps in the folding pathway of hCG-beta were investigated: the rate-limiting events in the folding of this protein, and the assembly of hCG-beta with, hCG-alpha. The rates of these folding events were determined with and without protein disulfide isomerase (PDI) using two different types of redox reagents: cysteamine and its oxidized equivalent, cystamine, and reduced and oxidized glutathione. Rates of the rate-limiting folding events were twofold faster in cysteamine/cystamine redox buffers than in glutathione buffers in the absence of PDI. Optimal conditions for hCG-beta folding were attained in a 2 mM glutathione buffer, pH 7.4, that contained 1 mg/mL PDI and in 10muM cysteamine/cystamine, pH 8.7, without PDI. Under these conditions, the half-time of the ratelimiting folding event was 16 to 20 min and approached the rate observed in intact cells (4 to 5 min). Moreover, folding of the beta subunit under these conditions yields a functional protein, based on its ability to assemble with the alpha subunit. The rates of assembly of hCG-beta with hCG-alpha in the cysteamine/cystamine or glutathione/PDI redox buffers were comparable (t(1/2/sb> = 9 to 12 min)). These studies show that rates of folding and assembly events that involve disulfide bond formation can be optimized by a simple buffer system composed of cysteamine and cystamine. (c) 1994 John Wiley & Sons, Inc.     

Acyl cystamine: small-molecular foldase mimics accelerating oxidative refolding of disulfide-containing proteins

Based on the structural characteristic of Protein disulfide isomerases and DsbA that have hydrophobic regions around the active sites, hydrophobic alkyl tails are linked to cystamine to create new small molecular foldase mimics, acyl cystamine. Both the oxidizing power and oxidation specificity of cystamine are enhanced by n-octanoyl or n-hexanoyl tail. N-octanoyl and n-hexanoyl cystamine are very effective to facilitate oxidative protein refolding at strong reducing environments. In the presence of 0.42 mM DTT, the activity recovery of lysozyme is over 90% by 90-min refolding with 0.1 mM n-octanoyl cystamine and 0.1 mM cystamine as oxidant, while almost no activity is recovered with 0.2 mM GSSG by 160-min refolding. For the refolding of 0.2 mg/mL lysozyme, with 0.6 mM n-hexanoyl cystamine and 1.12 mM residual DTT as redox agents, the activity recovery reaches as high as 93% after refolding for only 20 min. For ribonuclease A (RNase A) refolding, with 0.4 mM n-hexanoyl cystamine and 1.30 mM DTT, the recovery of activity reaches as high as 90% within 3 h. Thus, with n-octanoyl or n-hexanoyl cystamine as the oxidants, the necessity to remove excess DTT in the reduced and denatured protein solutions can be greatly alleviated. With a moderate hydrophobicity, n-hexanoyl cystamine is promising for application in oxidative protein refolding at an extensive concentration range. It is observed that in the oxidative refolding of 0.2 mg/mL lysozyme and RNase A, only about half of n-hexanoyl cystamine is needed when compared to cystamine to achieve the same kinetic effect.