Human serum contains factors that promote oxidative folding of disulphide proteins. We demonstrate this here using hirudin as a model. Hirudin is a leech-derived thrombin-specific inhibitor containing 65 amino acids and three disulphide bonds. Oxidative folding of hirudin in human serum is shown to involve an initial phase of rapid disulphide formation (oxidation) to form the scrambled isomers as intermediates. This is followed by the stage of slow disulphide shuffling of scrambled isomers to attain the native hirudin. The kinetics of regenerating the native hirudin depend on the concentrations of both hirudin and human serum. Quantitative regeneration of native hirudin in undiluted human serum can be completed within 48 h, without any redox supplement. These results cannot be adequately explained by the existing oxidized thiol agents in human serum or the macromolecular crowding effect, and therefore indicate that human serum may contain yet to be identified potent oxidase(s) for assisting protein folding.
- disulphide oxidase activity
- folding kinetics
- human serum
- oxidative folding of hirudin
- protein disulphideisomerase
Nascent disulphide proteins require proper folding and formation of native disulphide bonds in order to attain their stable native structures. This folding process, which involves oxidation and isomerization of disulphide bonds, is generally referred to as ‘oxidative folding’. In vivo oxidative folding occurs in the periplasmic space of prokaryotic cells and is assisted by a cascade of sophisticated electron transfer reactions catalysed by a specific class of thiol-disulphide oxidoreductases, including DsbA–DsbD [1–4]. DsbA is a primary catalyst of disulphide bond formation in bacteria. DsbC is a protein disulphide bond isomerase. DsbB and DsbD are two enzymes responsible for maintaining DsbA and DsbC in oxidized and reduced states respectively [1–3]. In eukaryotic cells, oxidative folding of disulphide proteins takes place at the ER (endoplasmic reticulum) and is ensured and facilitated mainly by PDI (protein disulphide-isomerase) and its related proteins [5–10]. PDI exists in both oxidized and reduced states and may exhibit oxidase or isomerase activity in promoting disulphide formation and isomerization of oxidative folding [9,10]. In addition, mechanisms of the unfolded protein response exist in the ER for dealing with the accumulation of unfolded and misfolded proteins [11,12].
In vitro, oxidative folding of fully reduced proteins can also be achieved under diverse conditions, with or without PDI and Dsb proteins. The process allows tracking and identification of folding intermediates and has been utilized to elucidate the common mechanism of protein folding. This technique, first applied to the study of BPTI (bovine pancreatic trypsin inhibitor) [13–15] and RNase A [16,17], was subsequently employed by many different laboratories to characterize the folding mechanisms of numerous disulphide proteins [18–30]. The results obtained so far have revealed a broad diversity and versatility of the mechanism of oxidative folding, which is displayed by both folding pathways and folding kinetics. Even among small proteins with comparable sizes and the same number of disulphide bonds, their folding pathways may vary significantly. These differences are mainly manifested by (i) the extent of heterogeneity of folding intermediates, (ii) the presence (or absence) of predominant intermediates containing native disulphide bonds and (iii) the level of accumulation of fully oxidized scrambled isomers as folding intermediates. The folding pathways of BPTI [13–15] and hirudin [18,19] (both three disulphides) represent two models at the opposite ends of such diversity. Studies have shown that the underlying cause of such diversity is in part dependent on the extent of stability of protein subdomains [31,32].
Aside from the diversity of folding pathways, the kinetics of in vitro oxidative folding are highly adaptable. They are generally affected by redox supplements in a two-stage mechanism . Oxidation agents (GSSG, Cys-Cys or oxidized PDI) act as oxidases. Their presence accelerates disulphide formation and the flow of partially oxidized (e.g. 1- and 2-disulphide) intermediates [19,20,22]. Reduced forms of thiol agents (GSH, Cys or reduced form of PDI) function as disulphide isomerases. Their presence promotes disulphide shuffling and the conversion of fully oxidized isomers into the native structure. Efficient in vitro oxidative folding of disulphide proteins therefore requires the presence of optimized concentrations of oxidase and isomerase, a mechanism similar to that taking place in vivo [1–3]. This is illustrated in the case of hirudin folding. Kinetically, in vitro oxidative folding of hirudin can be completed within seconds or days [18,19], depending on the conditions of folding. For instance, oxidative folding of hirudin can be accomplished in a buffer including only catalytic amount of thiol (e.g. 0.2 mM GSH). However, under these conditions, both disulphide formation and shuffling are slow and oxidative folding of hirudin is completed within approx. 24 h. In the buffer containing an optimized redox potential (a mixture of Cys, Cys-Cys and PDI), which promotes both disulphide formation and shuffling, quantitative recovery of native hirudin can be achieved within 15 s, a near 5000-fold acceleration of in vitro folding . Several small disulphide proteins [20–23] investigated in our laboratory exhibit properties of folding kinetics similar to those of hirudin.
Despite our understanding of the mechanisms of in vivo and in vitro oxidative folding, little is known about the mechanism of maintenance of disulphide proteins in the extracellular space, or whether redox systems also exist outside the cells (e.g. in human serum). It is plausible that such components may be present in serum for the purpose of preserving the oxidized state of circulating proteins or repairing partially reduced and accidentally denatured proteins. This subject is relevant as native disulphide proteins may possibly undergo isomerization (scrambling) or partial reduction after they are exported into the extracellular milieu; both chemical reactions can lead to inactivation or complication of their biological functions. Studies of in vitro reductive unfolding have shown that the stability of individual disulphide bonds of native proteins varies substantially [31,33,34]. Some of them may be reduced under very mild conditions. For example, the exceedingly stable BPTI  comprises one native disulphide bond (Cys14–Cys38) that can be rapidly reduced at micromolar concentration of dithiothreitol . More significantly, disulphide bonds of native proteins have the propensity to isomerize at elevated temperature or under denaturing conditions if adequate thiol compounds are present [36,37], which is likely to be the case in serum [38,39]. Indeed, the conformational stability of disulphide proteins in biological systems may not be directly deduced from in vitro studies, in which analysis is typically conducted in the absence of free thiols and native disulphide bonds are kept intact. For instance, the conformational stability of bovine pancreatic phospholipase A2, Δ(ΔGH2O), is reduced by 6.2 kcal/mol (1 cal=4.184 J) when disulphide isomerization is allowed in the presence of thiol catalyst .
We hypothesize that human serum may include biological apparatus for the purpose of maintaining the stability and intactness of circulating disulphide proteins. In order to validate this hypothesis, we have conducted in the present study oxidative folding of reduced hirudin in human serum. Hirudin was selected because the pathway of its in vitro oxidative folding has been investigated [18,19]. Hirudin is a thrombin-specific inhibitor isolated from the leech Hirudo medicinalis . It is the most potent thrombin inhibitor known, both natural and synthetic. Hirudin comprises 65 amino acids and is stabilized by three native disulphide bonds. It consists of two functional domains : a compact N-terminal domain (residues 1–49) that binds to the catalytic site of thrombin and a disordered acidic C-terminal tail (residues 50–65) that interacts with the fibrinogen recognition site (anion binding exosite) of the enzyme [43–45]. Both recombinant hirudin and its C-terminal peptide-derived compounds are clinically effective anticoagulants.
Recombinant HV1 (hirudin variant 1; CGP 39393) was obtained from Novartis AG (Basel, Switzerland). The protein is more than 98% pure as judged by HPLC and N-terminal sequence analysis. Human serum was kindly provided by Dr Ken Fujise of the University of Texas at Houston. Three batches of human serum were used throughout the present study. Dithiothreitol, GSH, GSSG, 2-mercaptoethanol and guanidinium chloride were purchased from Sigma with a purity of >99%. BSA and dextran 70 were also obtained from Sigma.
Preparation of fully reduced and denatured hirudin
Native hirudin (5.0 mg/ml) was reduced and denatured in the Tris/HCl buffer (0.1 M, pH 8.4) containing dithiothreitol (20 mM) and guanidinium chloride (6 M). The reaction was carried out at 22 °C for 90 min. Reduced and denatured hirudin was then separated from the denaturant and reducing agent by passing the sample through a NAP-5 column (Amersham Biosciences, Piscataway, NJ, U.S.A.) equilibrated with 0.1% TFA (trifluoroacetic acid). Purified hirudin sample was freeze-dried (1 mg/vial) and stored at −20 °C.
Oxidative folding of hirudin in buffers
Control folding experiments in buffers were carried out at 37 °C in the Tris/HCl buffer (0.1 M, pH 8.4) containing 0.15 M NaCl, sodium phosphate buffer (20 mM, pH 7.4) containing 0.15 M NaCl, or sodium phosphate buffer (20 mM, pH 7.4) containing both NaCl (0.1 M) and GSH/GSSG (1 mM/0.5 mM). The protein concentration was 0.5 mg/ml. Folding was quenched after various time periods by mixing aliquots of folding sample with 2 vol. of 4% (v/v) TFA in water. Trapped folding intermediates were analysed by HPLC or stored at −20 °C.
Oxidative folding of hirudin in a buffer containing mixed macromolecular crowding agents
Oxidative folding of reduced hirudin was also carried out in the Tris/HCl buffer (0.1 M, pH 8.4) containing mixed crowding agents (10 g/l BSA and 90 g/l dextran). The folding was performed in the absence and presence of GSH/GSSG (1 mM/0.5 mM). The protein concentration was 0.5 mg/ml. Folding was quenched in a time-course manner by mixing aliquots of folding sample with 2 vol. of 4% TFA in water, and directly analysed by HPLC.
Oxidative folding of hirudin in human serum
Three major sets of folding experiments with reduced hirudin in human serum were performed. All experiments were carried out at 37 °C. (i) Folding experiments were performed in solutions that comprised 100, 50, 30 and 10% (v/v) human serum, diluted with sodium phosphate buffer (20 mM, pH 7.4; 0.15 M NaCl). The hirudin concentration was 0.5 mg/ml. (ii) Folding experiments were carried in 100% human serum with serially reduced concentrations of hirudin (1, 0.75, 0.5, 0.25 and 0.1 mg/ml). (iii) Folding experiments were carried out in 95% human serum (diluted with 5% PBS) containing selected concentrations of GSH or GSSG (0.1, 0.3 and 1 mM), and a hirudin concentration of 0.5 mg/ml. Folding samples were quenched after various time periods by mixing aliquots of folding sample with 2 vol. of 4% TFA in water, and further diluted with another 2 vol. of 0.1% TFA. The samples were centrifuged at 16000 g for 10 min. The supernatant was filtered by passing it through a 0.2 μm PVDF syringe filter (4 mm diameter). Then the filtrate was analysed by HPLC or stored at −20 °C.
Analysis of the folding intermediates by reversed-phase HPLC
The reversed-phase HPLC analysis was performed on an Agilent 1100 Series HPLC system, using the following conditions. Solvent A for HPLC was water containing 0.088% TFA. Solvent B was acetonitrile/water (9:1, v/v) containing 0.084% TFA. The gradient was 14–27% solvent B linear in 10 min and 27–39% solvent B linear from 10 to 30 min. The flow rate was 0.5 ml/min. The column was Zorbax 300SB C-18 for peptides and proteins (4.6 mm×150 mm, 3.5 μm). Column temperature was 23 °C.
Analysis of the folding intermediates by MS
The reaction mixtures or HPLC-fractionated intermediates (freeze-dried, 20 μg) were derivatized with 50 μl of vinyl pyridine (0.13 M) in Tris/HCl buffer (0.1 M, pH 8.4) at 23 °C for 45 min. Vinyl pyridine-derivatized samples were further desalted by passing them through a NAP-5 column equilibrated with 0.1% TFA, freeze-dried and then subjected to MALDI–TOF-MS (matrix-assisted laser-desorption ionization–time-of-flight MS; Voyager-DE™ STR Biospectrum™ workstation; PerSeptive Biosystems, Foster City, CA, U.S.A.).
Anti-thrombin assay: anti-amidolytic assay
Anti-amidolytic activity of hirudin was assessed by its ability to inhibit α-thrombin from digesting chromozym TH (a p-nitroaniline-based substrate). The assay was performed at 23 °C in 1 ml of Tris/HCl buffer [67 mM, pH 8.0; containing 133 mM NaCl and 0.13% poly(ethylene glycol) 6000]. The digestion of chromazym TH was monitored at 405 nm for a period of 2 min. Concentrations were 0.2 mM for chromozym substrate, 1 nM for α-thrombin and 0.5 nM for control hirudin (the activity was taken as 100%).
Oxidative folding of hirudin in buffers: control folding experiments
In order to understand the mechanism of hirudin folding in human serum, it is essential to perform control folding experiments in buffers. The mechanism of in vitro oxidative folding of hirudin is influenced by two major parameters: the pH, which ranges between 7 and 9 [13,15], and the redox buffer , which contains various concentrations of GSH/GSSG (0.1–2 mM), Cys/Cys-Cys or reduced/oxidized dithiothreitol. Higher pH facilitates disulphide oxidation (formation) and increases the efficiency of hirudin folding. Redox agents facilitate hirudin folding and the flow of folding intermediates in a two-stage mechanism. GSSG (or Cys-Cys) accelerates the rate of disulphide formation, whereas the presence of GSH (or Cys, 2-mercaptoethanol or PDI) promotes the disulphide shuffling . Three selected control folding experiments of hirudin in buffers are described follow.
At pH 8.4, oxidative folding of hirudin takes place regardless of the absence or presence of redox agent . However, quantitative recovery of the native hirudin at pH 8.4 still needs the presence of redox agents or at least an additional thiol catalyst (e.g. GSH, Cys or 2-mercaptoethanol). When oxidative folding of hirudin was carried out in the buffer (pH 8.4) alone (Figure 1A), the recovery of native hirudin was approx. 50±10%. The remaining protein was consistently trapped as fully oxidized scrambled isomers, unable to shuffle their non-native disulphide bonds to reach the native structure. Under these folding conditions, air oxidation serves to promote disulphide oxidation and the formation of 1-, 2- and 3-disulphide intermediates. The folding kinetics are slow. During the early stage of folding, free cysteine residues of 1- and 2-disulphide intermediates function as thiol catalyst and catalyse the conversion of 3-disulphide intermediates to form the native hirudin. As the folding progresses, free cysteine residues begin to be depleted because more cysteine residues become involved in disulphide pairing and less are available as thiol catalyst for disulphide shuffling; therefore scrambled 3-disulphide isomers of hirudin (∼50% of the total protein, depending on the protein concentration) become trapped and accumulate as intermediates, unable to be converted into the native hirudin due to the absence of thiol catalyst. At this stage, addition of a trace amount of thiol agent will facilitate the process of disulphide shuffling and permit complete conversion of the scrambled hirudins into the native structure .
At pH 7.4, which mimics physiological conditions, redox agents are required for successful folding of hirudin. In the buffer alone (PBS solution, pH 7.4), oxidative folding of hirudin hardly occurs. This is demonstrated in Figure 1(B). Under these conditions, less than 1% of the native hirudin was formed after 24 h of sample incubation at 37 °C. In the presence of GSH/GSSG (1 mM/0.5 mM) at pH 7.4, folding of hirudin was complete within 12 h (Figure 1C) and the profile of folding intermediates was indistinguishable from that when folding was performed at pH 8.4.
To further elucidate the mechanism of hirudin folding, intermediates of the three control folding experiments were isolated by HPLC, derivatized with vinyl pyridine and characterized by MALDI–TOF-MS (method outlined in Figure 2) in order to determine the concentrations of disulphide species present in each time-course trapped sample. The data obtained from the analysis of samples of Figures 1(A)–1(C) are shown in Figures 3(A)–3(C) respectively. All three folding experiments are characterized by a sequential flow from 0-disulphide (R) through 1-disulphide and 2-disulphide to 3-disulphide isomers of hirudin. When folding of hirudin was performed in the buffer alone (pH 8.4) (Figure 3A), the end-product (24 h folding) comprised approx. 50% of native hirudin and 50% of 3-disulphide scrambled isomers. In the buffer of pH 7.4 alone (Figure 3B), most of the hirudin sample was recovered as 1- and 2-disulphide species after 24 h of sample incubation at 37 °C. In the presence of GSH/GSSH at pH 7.4 (Figure 3C), the formation of 3-disulphide intermediates was followed by the quantitative recovery of native hirudin.
Oxidative folding of hirudin in human serum: folding pathway and characterization of folding intermediates
Oxidative folding of hirudin was shown to occur in human serum (pH 7.4) without supplement. An initial standard folding experiment was carried out in an undiluted human serum with a hirudin concentration of 0.5 mg/ml. Fully reduced and denatured hirudin was prepared in freeze-dried form and mixed directly with 100% human serum to initiate the folding. Folding intermediates were then quenched at different time points by acidification, separated from other serum proteins and analysed by HPLC. The results show that the folding proceeds through intermediates with unvarying HPLC profiles (Figure 4, top panel) and the production of native hirudin exhibits close to first-order kinetics. A 70% recovery of native hirudin was achieved after 24 h of folding at 37 °C.
Analysis of the molecular mass of folding intermediates (method described in Figure 2) revealed a rapid formation of scrambled 3-disulphide isomers during the process of folding. The 3 min trapped intermediates consisted of approx. 25% 2-disulphide isomers and 75% 3-disulphide isomers (Figure 3D). After 10 min of folding, the only detectable intermediates were fully oxidized 3-disulphide scrambled species (Figure 3D), which is consistent with the static HPLC profile of folding intermediates (Figure 4, top panel). The results demonstrate that oxidative folding of hirudin in human serum undergoes an initial stage of rapid disulphide oxidation to form the scrambled species as intermediates. This is followed by a rate-limiting step of disulphide shuffling that converts scrambled 3-disulphide isomers to form the native hirudin. The hirudin refolded in human serum exhibits anti-amidolytic activity indistinguishable from that of native hirudin.
Oxidative folding of hirudin in human serum: dependence of folding kinetics on the concentration of serum
The efficiency of folding is dependent on the concentration of human serum. This was demonstrated in a systematic analysis of folding experiments with serially diluted serum. At a selected concentration of hirudin (0.5 mg/ml), dilution of human serum with PBS was shown to reduce the efficiency of hirudin folding. The recoveries of native hirudin following 24 h of folding were found to be 70% in 100% serum, 48% in 50% serum, 37% in 30% serum and 22% in 10% serum (Figure 4). Dilution of human serum also affects the kinetics of flow of folding intermediates. For instance, in 100% serum (Figure 4, top panel), the pattern of folding intermediates remained practically the same after 10 min of sample incubation, due to the rapid oxidation of disulphide bonds and the immediate accumulation of 3-disulphide intermediates. In the 10-fold diluted serum (Figure 4, bottom panel), a considerably slower formation of 3-disulphide intermediates is evident and the patterns of folding intermediates bear a resemblance to those observed in the control folding experiments (Figures 1A and 1C).
Oxidative folding of hirudin in human serum: dependence of folding kinetics on the concentration of hirudin
In addition to the dependence on the concentration of serum, the efficiency of hirudin folding also depends upon the concentration of hirudin itself. We have conducted oxidative folding of hirudin in 100% human serum using different concentrations of hirudin (0.1–1.0 mg/ml). Folding intermediates were similarly analysed by HPLC and the results are presented in Figure 5. The results demonstrate a positive correlation between the recovery of native hirudin and the protein concentration. A near quantitative recovery of native hirudin can be achieved within 24 h at the hirudin concentration of 1 mg/ml. Specifically, the yield of native hirudin displayed a precipitous decrease as the concentration of hirudin was lowered from 0.5 to 0.25 mg/ml. At 0.1 mg/ml, the recovery of native hirudin after 24 h of folding was reduced to 23%.
The HPLC data (Figure 5) reveal a rapid accumulation and predominance of 3-disulphide isomers as folding intermediates, irrespective of the hirudin concentration. This is noticeable from the HPLC profiles of all 10 min trapped intermediates (Figure 5). It becomes apparent that at low concentrations of hirudin, a decreased activity of disulphide shuffling of scrambled isomers accounts for the diminished recovery of the native hirudin.
As a control, the effect of hirudin concentration (0.1–1 mg/ml) on its folding kinetics was also investigated in the Tris/HCl buffer in the absence and presence of GSH/GSSG (1 mM/0.5 mM). The results reveal an inverse, albeit trivial, correlation between hirudin concentration and its folding kinetics. Low concentrations of hirudin have a positive effect in promoting folding. In the buffer alone (Figure 6, open symbols), the decrease in protein concentration enhances the rate of disulphide oxidation, the disappearance of reduced hirudin and the accumulation of partially and fully oxidized (scrambled) hirudin. In the buffer containing redox agents (Figure 6, filled symbols), the decrease in protein concentration also slightly increases the rate of hirudin folding, notably the recovery of native hirudin.
Oxidative folding of hirudin in human serum: dependence of folding kinetics on supplementation with redox agents
Inclusion of redox supplements in human serum can accelerate the kinetics of hirudin folding. We have investigated the independent influence of varying concentrations of GSH and GSSG on hirudin folding. These experiments were carried out in 95% human serum with a hirudin concentration of 0.5 mg/ml. Folding intermediates of each experiment were trapped at different time points and analysed by HPLC (Figure 7). The results show that GSH displays a more significant effect in accelerating the recovery of hirudin. The inclusion of 0.3 and 1 mM GSH increase the initial rate of recovery of native hirudin by a factor of 2- and 4-fold respectively. These results are not unexpected since the rate-limiting step of hirudin folding in serum is the conversion of scrambled isomers into the native structure, which required a thiol reagent as catalyst.
Oxidative folding of hirudin in a buffer containing mixed macromolecular crowding agents
These folding conditions are designed to mimic the intracellular milieu [46,47] in which the concentration of macromolecules is estimated to be in the range 80–400 g/l [48,49]. Reduced hirudin was allowed to refold in the Tris/HCl buffer in the absence and presence of mixed BSA/dextran (100 g/l). The results, shown in Figure 8, demonstrate that the presence of mixed BSA/dextran had no effect on the kinetics of hirudin folding, as measured by the decrease in fully reduced hirudin, the recovery of native hirudin (Figure 9) and the patterns of intermediates that appeared along the pathway of folding. Unlike the folding performed in the human serum, in which quantitative formation of 3-disulphide hirudin occurred within 10 min (Figure 4), complete oxidation of reduced hirudin in the presence of macromolecular crowding agents did not happen until 20–24 h of folding (Figure 8B).
These two experiments were repeated in the presence of GSH/GSSG (1 mM/0.5 mM). The inclusion of redox agents significantly accelerated the process of hirudin folding (Figure 9). Again, however, no effect of macromolecular crowding agents was seen.
Technical challenges of performing protein oxidative folding in serum
In performing the oxidative folding of hirudin in human serum, we have anticipated a challenging task of analysing folding intermediates and refolded hirudin by HPLC. Human serum contains at least 3700 species of major proteins with molecular masses ranging from a few thousand to hundreds of thousands of daltons . Many of them were initially expected to interfere with the analysis of hirudin isomers. This concern turned out to be unwarranted. As shown by HPLC data presented here, hirudin isomers appear to be essentially free of interference from serum proteins. Three causes may chiefly account for the tidiness of hirudin analysis by HPLC. (i) The step of acid quenching allows precipitation and removal of a large fraction of serum proteins. (ii) Hirudin is an acidic and highly soluble protein; both properties favour its rapid and early elution from the reversed-phase HPLC. (iii) Numerous serum proteins may indeed co-elute with hirudin isomers, but their concentrations are far lower than that of hirudin.
Concerns also existed as to whether reduced hirudin would form heterodimers or heteropolymers with serum proteins via disulphide formation, or whether a denatured protein, like reduced hirudin, would be prone to proteolytic fragmentation in serum during prolonged incubation. If these reactions do occur, the yield of active hirudin would be drastically diminished. To answer this question, we have compared the total hirudin recovery of two folding experiments (both in triplicate) conducted under parallel methods, one in PBS (in the presence of GSH/GSSG) and the other in 100% human serum. The results reveal that recoveries of hirudin in these two experiments are nearly indistinguishable (<±3–5%). These results demonstrate that potential formation of heterodimers and enzymatic cleavage of denatured hirudin in human serum should not be issues of concern. However, this may not be the case for other disulphide proteins.
Mechanism of oxidative folding of hirudin in human serum
Oxidative folding of hirudin cannot proceed in PBS in the absence of redox agents (Figure 1). However, it occurs in human serum without supplementation with redox agents and with a close to quantitative yield of the native structure (Figure 5). The folding undergoes an initial stage of rapid disulphide formation (oxidation), which leads to the accumulation of fully oxidized scrambled isomers as folding intermediates. In 100% serum at 37 °C, oxidation of disulphide bonds of hirudin is completed within approx. 6–7 min. This is followed by a second stage of disulphide shuffling that converts scrambled isomers into the native hirudin. The process of disulphide shuffling is slow and may take up to 48 h depending on the concentration of hirudin.
The kinetics of disulphide oxidation are influenced by the concentration of serum. Successive dilution of human serum progressively decreases the rate of disulphide oxidation and the accumulation of scrambled hirudin (Figure 4). This is likely to be due to the dilution of serum components responsible for promoting the oxidation of reduced hirudin. The kinetics of disulphide shuffling, in contrast, are mainly affected by the concentration of hirudin (Figure 5). Higher concentrations of hirudin result in a better yield of native hirudin. Analysis of HPLC data (Figure 5) reveals that at higher concentrations of hirudin (e.g. 1 mg/ml), intermediates appear to comprise minor fractions of partially oxidized hirudin which may serve as thiol catalyst and promote the conversion of fully oxidized intermediates to form the native hirudin. This property, however, has not been observed in the hirudin folding conducted in the buffer solution (Figure 6). The underlying cause of this discrepancy is not immediately clear to us.
The apparent mechanism of hirudin folding in human serum indicates that a major rate-limiting step for the recovery of native protein is the disulphide shuffling of scrambled 3-disulphide intermediates, which requires a thiol compound as an initiator [18,19]. As expected, addition of GSH to human serum significantly improves the kinetics of recovery of native hirudin (Figure 7). The data imply that a wide range of thiol-containing compounds, such as Cys and PDI, can be used as supplements or cofactors for regulating the kinetics of hirudin folding in human serum.
The HPLC patterns of folding intermediates of hirudin, despite extensive overlap, are impossible to differentiate when folding is performed in human serum (Figure 4) or carried out in buffers (pH 7.4 and 8.4; Figure 1). These results, together with our previous studies of hirudin core domain [18,19], demonstrate two important properties of the folding mechanism of hirudin. (i) The folding kinetics of hirudin, including kinetics of the recovery of native structure and the flow of intermediates, may vary substantially, depending on the redox potential and folding conditions. (ii) The folding pathway of hirudin (the composition of folding intermediates) remains essentially unchanged, regardless of folding conditions.
Human serum may contain oxidase(s) that promote the oxidative folding of hirudin
In comparison with the folding kinetics of reduced hirudin performed in PBS (pH 7.4) and assisted by GSSG, Cys-Cys or oxidized PDI, 10-fold diluted human serum exhibits an oxidase activity comparable with that of 500 μM GSSG, 100 μM Cys-Cys or 10 μM oxidized PDI. This rapid disulphide formation of reduced hirudin in human serum cannot be accounted for by the known serum components. Human serum indeed contains low-molecular-mass thiol compounds [e.g. Cys, HomoCys (homologous with Cys with a longer side chain -CH2-CH2-SH) and glutathione] that exist in reduced, oxidized or protein-bound forms. In their oxidized form, they are able to accelerate the disulphide formation of protein oxidative folding. However, the concentrations of their oxidized forms in human serum were shown to be in the low micromolar range (55 μM for oxidized Cys, 2 μM for oxidized HomoCys and 1.4 μM for GSSG) . These concentrations of oxidized thiol compounds are obviously inadequate to account for the fast disulphide formation observed in human serum.
The rapid disulphide formation of reduced hirudin in human serum also cannot be explained by the molecular crowding effect. It has been reported  that ‘mixed macromolecular crowding accelerates the oxidative refolding of reduced denatured lysozyme’ (title of ). We have evaluated the effect of a mixture of BSA/dextran and failed to show that such macromolecules exhibit any detectable effect on promoting the oxidative folding of hirudin, on either the kinetics of disulphide bond formation or the recovery of native hirudin (Figures 8 and 9). Despite the implications of the title, the above-mentioned published work  actually demonstrates that mixed crowding agents are more effective than single crowding agents. The refolding yield of lysozyme was shown to be equally high in the absence and presence of crowding agents. Our results are therefore not inconsistent with these previous observations.
However, human serum also harbours at least 3700 species of identifiable protein molecules , some of which may potentially exhibit properties of thiol-redoxin proteins, such as DsbA, DsbC or oxidized and reduced forms of PDI [3,9,10]. Thus it is most likely that additional oxidase-like molecules in human serum are accountable for promoting the folding of hirudin. It is plausible that such components may be present in serum for the purposes of maintaining the oxidized state of native proteins or repairing partially reduced and accidentally denatured proteins. The identification and isolation of such prospective oxidase(s) would permit further understanding of the molecular mechanism of the oxidative folding of hirudin and other proteins in serum.
J.-Y. C. acknowledges the generous support from the Robert Welch Foundation. We are also grateful to Dr Ken Fujise for providing the sample of human serum.
Abbreviations: BPTI, bovine pancreatic trypsin inhibitor; ER, endoplasmic reticulum; MALDI–TOF-MS, matrix-assisted laser-desorption ionization–time-of-flight MS; PDI, protein disulphide-isomerase; TFA, trifluoroacetic acid
- The Biochemical Society, London