HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Seefeldt, M. B. Articles by Randolph, T. W. Search for Related Content PubMed PubMed Citation Articles by Seefeldt, M. B. Articles by Randolph, T. W. Social Bookmarking                   What's this?

High-pressure refolding of bikunin: Efficacy and thermodynamics

¹ Department of Chemical and Biological Engineering, Center for Pharmaceutical Biotechnology, University of Colorado, Boulder, Colorado 80309-0424, USA
² Bayer Corp., Berkeley, California 94710, USA
³ Department of Pharmaceutical Sciences, Center for Pharmaceutical Biotechnology, University of Colorado Health Sciences Center, Denver, Colorado 80262, USA

Reprint requests to: Theodore W. Randolph, Department of Chemical and Biological Engineering, Center for Pharmaceutical Biotechnology, ECCH 111, University of Colorado, Boulder, CO 80309-0424, USA; e-mail: randolph{at}pressure3.colorado.edu; fax: (303) 492-4341.

(RECEIVED May 27, 2004; FINAL REVISION July 6, 2004; ACCEPTED July 6, 2004)

	Abstract

TOP Abstract Introduction Results Discussion Materials and methods References

 
Bikunin is a glycosylated protein that aggregates extensively during mammalian cell culture, resulting in loss of activity, loss of native secondary structure, and the formation of nonnative disulfide bonds. We investigated the use of high hydrostatic pressure (1000–3000 bar) for the refolding of bikunin aggregates. The refolding yield obtained with pressure-modulated refolding at 2000 bar was 70 (±5%) by reverse-phase chromatography (RP-HPLC), significantly higher than the value of 55 (±6%) (RP-HPLC) obtained with traditional guanidine HCl "dilution-refolding." In addition, we determined the thermodynamics of pressure-modulated refolding. The change in volume for the transition of aggregate to monomer V_refolding was calculated to be –28 (±5) mL/mole. Refolding was accompanied by a loss of hydrophobic exposure, resulting in a positive contribution to the V_refolding. These findings suggest that the disruption of electro-static interactions or the differences in size of solvent-free cavities between the aggregate and the monomer are the prevailing contributions to the negative V_refolding.

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.04891204.

	Introduction

TOP Abstract Introduction Results Discussion Materials and methods References

 
Production of recombinant proteins through bacterial fermentation is often unsuccessful if the desired protein requires glycosylation. In this case, mammalian cell lines are often used for production (Chu and Robinson 2001). Expressed proteins are typically properly folded and glycosylated. However, slower growth rates, complex growth media, and lower production levels increase production costs. Furthermore, inefficient processing of polypeptide chains within the endoplasmic reticulum often results in the formation of high–molecular weight aggregates containing nonnative disulfide bonds (Braakman et al. 1992; Schroder and Friedl 1997; Schroder et al. 2002; Soukharev et al. 2002). Even when correctly expressed, sustained exposure to fermentation temperatures would be expected to cause aggregation of secreted proteins. Numerous studies of purified proteins demonstrate that proteins readily aggregate under analogous solution conditions (Kim et al. 2000; Chi et al. 2003a). Protein aggregation during mammalian cell culture is often overlooked but can result in low production titers.

High hydrostatic pressures have been shown to unfold native proteins, typically at pressures above 4000 bar (Hawley 1971; Chatani et al. 2002; Royer 2002). Moderate pressures of approximately 2000 bar have been shown to dissociate native oligomers (Gross and Jaenicke 1994; Balny 2002). Aggregates behave like multimers in the sense that they can be dissociated with moderate pressure (St. John et al. 1999). Meanwhile, the native state remains favored at these same pressures, enabling disaggregated proteins to maintain some residual structure, which could increase re-folding efficiency (Kunugi and Tanaka 2002). Previous research has shown that moderate pressures successfully dissociate and refold nonnative aggregates of human growth hormone and lysozyme, while favoring the native state (St. John et al. 2001, 2002). However, the protein aggregates previously studied have been from nonglycosylated proteins produced through prokaryotic expression (St. John et al. 1999, 2001; Ferrao-Gonzales et al. 2000; Balny 2002).

The purpose of this study was to determine whether high hydrostatic pressure could be useful in refolding a glycosylated protein containing multiple disulfide bonds from aggregates produced during mammalian cell culture. We studied the effect of high hydrostatic pressure on the refolding of recombinant human placental bikunin from soluble, disulfide cross-linked, nonnative aggregates and compared the results to those obtained using traditional chaotrope-based refolding methods. We investigated the effects of processing conditions (pressure, temperature, pH, ionic strength, depressurization rate, chaotrope concentrations, and refolding time) to determine their effect on aggregate dissociation and native protein formation. Disaggregation and refolding were measured with size-exclusion chromatography (SEC-HPLC), reverse phase chromatography (RP-HPLC), and kallikrein activity assays to determine native monomer recovery from aggregates. To determine structural properties of monomers and aggregates, circular dichroism (CD), infrared spectroscopy (IR), and second-derivative ultraviolet spectroscopy (2D-UV) were used. The free energy of refolding could be calculated by obtaining equilibrium compositions of aggregates and monomers through the use of gas-phase electrophoretic-mobility macromolecular analysis (GEMMA). Finally, differential scanning calorimetry (DSC) was used to obtain the thermodynamic parameters governing thermal denaturation of native bikunin for comparison to refolding thermodynamics.

	Results

TOP Abstract Introduction Results Discussion Materials and methods References

 
Definitions
Throughout the remainder of this article, the term "refolding" will be used to describe the recovery of native monomer from aggregate. "Refolding" in this context describes the overall process of aggregate dissociation, native disulfide bond formation, and protein folding from a perturbed conformational state. In contrast, we have defined "denaturation" as the unfolding of native monomeric bikunin to the denatured conformation.

Bikunin monomer and aggregate structural characterization: IR, UV CD, and 2D-UV spectroscopy
Bikunin aggregates are covalently crosslinked through non-native disulfide bonds and require both dissociation and disulfide modification to impart refolding. Native, monomeric bikunin has a molecular weight of 19–24 kDa (as a function of two possible glycosylation sites) and contains six disulfide bonds. Because high-resolution structural analyses (nuclear magnetic resonance [NMR], X-ray diffraction) have not been previously reported for placental bikunin, we used IR and far-UV CD spectroscopic studies to estimate the extent of secondary structure in monomers and aggregates.

In IR studies (Fig. 1), the ratios of peaks at 1692–1678 cm^–1 (-turn), 1664 cm^–1 (loop), 1656 cm^–1 (-helix), and 1645–1620 cm^–1 (-sheet) were compared to estimate the secondary structure of the two forms (Dong et al. 1990). The monomer contained 24% -turn and 26% -sheet, and the remainder was a combination of -helix and random structures, indicating that monomer has a relatively disordered secondary structure, consistent with X-ray structural data available for urinary bikunin homologs (Xu et al. 1998). On aggregation, a significant decrease in -helix and loop was observed, with concomitant formation of intermolecular -sheet. Far-UV CD spectra (Fig. 2A) mirror the IR results, with decreased -helix signal at 222 nm observed in spectra for the aggregates.

View larger version (19K): [in this window] [in a new window]   Figure 1. Fourier transform infrared spectra of bikunin monomer and aggregate. The solid line represents the monomer spectra, which has higher levels of native random coil and -helix (1662 nm). The dashed line represents the aggregate, which contains higher levels of sheet.

View larger version (25K): [in this window] [in a new window]   Figure 2. (A) Far-CD scan of monomer and aggregate. A decrease in -helix is seen on aggregation. (B) Near-CD scan of monomer and aggregate. Secondary structure decreases on aggregation and exposes hydrophobic residues to the protein surface. A loss of specific aromatic solvent environments is observed on aggregation.

Bikunin monomer thermodynamic characterization: Differential scanning calorimetry
Table 1 shows values of G_denaturation, Cp_denaturation, and the T_m obtained with differential scanning calorimetry (DSC). The G_denaturation value of 9 kJ/mole for bikunin is relatively small compared to "typical" values of 40 kJ/mole seen for many proteins (Voet 1995). However, this value is not surprising because the T_m of bikunin at pH 5.8 (53°C) is relatively low compared to that of many globular proteins, and the protein contains only moderate levels of secondary structure.

View this table: [in this window] [in a new window]   Table 1. DSC table for the denaturation of native bikunin providing Gdenaturation, Tm, and Cpdenaturation

The structure of monomeric bikunin under pressure: 2D-UV spectroscopy
Structural changes of bikunin monomer as a function of pressure were analyzed using 2D-UV spectroscopy, shown in Figure 3. For 2D-UV, the "r-ratio" was used to monitor the solvent environment of tyrosine residues (Ragone et al. 1984). For bikunin, the r-ratio increases slightly with pressure, from a value of 2.16 at 1 bar to 2.2 at 3000 bar. In contrast, during thermal or guanidine-induced denaturation studies, the r-ratio decreased (1.35 at 75°C and 1.84 in 6 M guanidine HCl) as the temperature and chaotrope concentration increased. The increase of the r-ratio with pressure implies that the native structure is stabilized by application of pressures up to 3000 bar. This conclusion is further supported by the red shift near 288 nm that occurs during pressurization (Fig. 3), indicative of tyrosine burial.

View larger version (28K): [in this window] [in a new window]   Figure 3. 2D-UV spectra of pressure denaturation of bikunin. The red shift occurring near 288 nm is representative of hydrophobic burial. Note the strong differences in pressure spectra compared to the guanidine perturbed and temperature denatured spectra. In comparison, the native state is still favored at pressures up to 3000 bar.

Yields from dilution refolding were not identical by SEC-HPLC, RP-HPLC, or activity assays. Yields determined by activity assays (52% [±16%]) and RP-HPLC (55% [±6%]) are lower than those determined by SEC-HPLC (80% [±2%]) demonstrating the presence of monomer-sized species that are nonnative.

Pressure-modulated refolding of bikunin aggregates
To determine the effectiveness of pressure-modulated re-folding, bikunin at a concentration of 0.5 mg/mL was placed into an optimized refolding buffer (50 mM Tris at pH 8.0, 4 mM GSSG, 2 mM DTT, 157 mM NaCl), incubated at 2000 bar for 24 h at 25°C, and slowly depressurized (see Materials and Methods). The optimized pressure refolding conditions resulted in a 70% (±5%) refolding yield by RP-HPLC. This yield was higher than those obtained from traditional dilution refolding methods. When the initial aggregate concentration was reduced to 0.06 mg/mL, refolding yields greater than 96% were obtained.

After pressure-modulated refolding, yields determined by SEC-HPLC (77% [±3%]), RP-HPLC (70% [±5%]), and activity assays (79% [±17%]) were identical, suggesting that a high percentage of dissociated aggregates are refolded to the native state. This result is in contrast to dilution refolding, where monomer content by SEC-HPLC was higher than native monomer content as determined by RP-HPLC and activity measurements.

Optimization of pH conditions
Bikunin contains six disulfide bonds, and proper disulfide bond formation is essential for refolding. Correct pH is critical because it controls the effectiveness of glutathione disulfide shuffling (Clark et al. 1998). Disulfide exchange occurs through the nucleophilic substitution of a thiol anion into a disulfide (Gilbert 1995). As a result, basic pHs are required to facilitate the removal of the thiol hydrogen atom (pK_a = 8.5; Gilbert 1995). During refolding at 2000 bar, a minimal pH of 8.0 is needed to maximize refolding yield, whereas pHs lower than 8.0 decreased the refolding yield substantially, reaching a baseline value of approximately 10% at pH 7.0 (Fig. 4).

View larger version (15K): [in this window] [in a new window]   Figure 4. The effect of pH on refolding yield, determined by SEC-HPLC and RP-HPLC. pHs greater than 8.0 are needed to facilitate glutathione reactivity and disulfide shuffling.

Optimization of ionic strength
Ionic strength was altered while the remaining refolding conditions were held constant (2000 bar, 25°C, 4 mM GSSG, 2 mM DTT, 50 mM Tris at pH 8.5 [to prevent ionic strength effects from adversely decreasing pH] for 24 h with slow depressurization). Refolding yield decreased from 79.5% (±0.7%) to 70% (±0.5%) (SEC-HPLC) and from 79% (±3%) to 66.2% (±0.7%) (RP-HPLC) in the 50 mM Tris buffer lacking NaCl (Fig. 5).

View larger version (17K): [in this window] [in a new window]   Figure 5. The effect of ionic strength on refolding yield, determined by SEC-HPLC and RP-HPLC. Decreases in ionic strength lower refolding yields.

View larger version (17K): [in this window] [in a new window]   Figure 6. The effect of pressure on refolding yield, determined by SEC-HPLC and RP-HPLC. A critical pressure of 1500 bar is required to maximize refolding, whereas further increases in pressure do not increase yield.

View larger version (14K): [in this window] [in a new window]   Figure 7. SEC and RP analysis of the effect of temperature on refolding yield. Lower temperatures resulted in an increase of refolding yield. Increases in temperature resulted in large decreases in refolding yield.

Determination of G_refolding, Cp_refolding, H_refolding, and S_refolding

View larger version (14K): [in this window] [in a new window]   Figure 8. The Grefolding as a function of temperature. The data were fit through a nonlinear regression employing equation 1 and is represented by the solid line. Through this regression, Cprefolding and Srefolding could be calculated. Cprefolding is small compared to the Cpdenaturation because of the linear nature of the regression.

(1)

where the subscript ₀ refers to the reference state, 25°C and 2000 bar. H_refolding at 298K was obtained from

(2)

A summary of the thermodynamic refolding parameters obtained is found in Table 3.

Effect of pressure on refolding yield: Determination of G_refolding and V_refolding
The pressure dependence of refolding is described by the change in volume on refolding (V_refolding), which is defined as the difference in partial molar volume between the aggregated and native state.

The pressure was varied from 0 to 1500 bar, and the V_refolding determined by applying the equation

(3)

V_refolding was –28 (±5) mL/mole, as shown in Figure 9.

View larger version (13K): [in this window] [in a new window]   Figure 9. The natural log of the equilibrium constant, Krefolding, as a function of pressure. A linear regression of pressures up to 1500 bar was used to obtain the Vrefolding of –28 ± 5 mL/mole.

Effect of ionic strength on refolding yield: Determination of G_refolding

Comparison of refolding and denaturation thermodynamics
The reference concentration required for G_refolding prevents direct comparison to G_denaturation. However, the differential parameters (V_refolding, Cp_refolding, H_refolding, and S_refolding) can be compared directly to the denaturation parameters. Three key points can be drawn: First, the V_refolding is negative and significant (–28 [±5] mL/mole) and is in sharp contrast to pressure’s negligible effect on the native state; second, Cp_refolding is small and negative, differing from the typical large and positive Cp_denaturation; last, enthalpic forces favor the dissociation and refolding of protein aggregates. The enthalpic contribution is slightly larger and opposite in sign compared to the entropic contribution.

	Discussion

TOP Abstract Introduction Results Discussion Materials and methods References

 
Improved yields of pressure refolding over chaotrope refolding
Chaotrope refolding requires complete protein denaturation for aggregate dissociation. Dilution refolding results in relatively high yield of native bikunin (55% based on RP-HPLC), but there is a significant fraction of nonnative monomer formed (comparison of SEC-HPLC to RP-HPLC activity). In contrast, pressures of 2000 bar improved re-folding over traditional chaotrope methods. We speculate that higher yields are obtained using high-pressure because, in contrast to chaotrope refolding, disaggregation occurs under conditions that favor native state conformation, and the protein is never fully unfolded. V_refolding was negative (–28 [±5] mL/mole). However, native monomer structure was unaffected by pressures up to 3000 bar at 25°C (Fig. 3).

Electrostatic shielding improves refolding yield
Electrostatic interactions are important within catalytic sites and can either increase or decrease protein stability (Takano et al. 2000; Kumar and Nussinov 2002). Research has also shown that electrostatic interactions are important for colloidal stability (Chi et al. 2003a,b). As a typical example, recombinant GCSF is stable at pH 3.5 but aggregates readily at pH 7.0 in 10 mM PBS buffer and 140 mM NaCl (Chi et al. 2003a). The net positive charge of GCSF at pH 3.5 resulted in electrostatic repulsion of protein species and prevented aggregation. In the second case, the addition of NaCl provided shielding and minimized the strength and distance of electrostatic repulsion, enabling GCSF to "see" itself and aggregate (Israelachvili 1992; Chi et al. 2003a).

An opposite trend is seen with bikunin. In this case, the addition of 157 mM NaCl improved the refolding yield. Two potential explanations arise from this result. First, bikunin could have an asymmetric distribution of surface charge. If positive and negative ions were not distributed equally on the protein surface, the molecule would act like a dipole and enable electrostatic attraction. Because bikunin is glycosylated, this scenario would not be surprising because of the significant negative charge of acids within the carbohydrate chains (Voet 1995). A second possibility is that specific salt bridges may be present within the aggregate or monomer that strongly influence the protein’s stability. The addition of NaCl could disrupt stabilizing salt bridges within the aggregate or, conversely, disrupt destabilizing salt bridges within the native state. Recent research on the strength of surface salt bridges in human lysozyme supports this hypothesis (Takano et al. 2000). The bond strength of salt bridges with greater than 50% solvent accessibility fell to essentially zero on the addition of 200 mM KCl (Takano et al. 2000). Detailed structural knowledge of both the monomer and the aggregate (which is unavailable) would be required to further explain the ionic strength effects seen.

Refolding buries hydrophobic residues
The large positive heat capacity change on protein denaturation is a result of the hydration of apolar protein surfaces (Privalov 1990). Hydrophobic hydration results in a positive increase in the heat capacity, whereas hydration of polar groups decreases the heat capacity by approximately 30% of the hydrophobic contribution (Privalov 1997). Hydration effects on the heat capacity change of denaturation are generally non–protein specific, and consequently, they can be estimated from the length of the primary sequence with reasonable accuracy (Edelhoch and Osborne 1976). Using a correlation of 49 J/mole-°C-residue, the estimated Cp_denaturation for bikunin was 8530 J/mole-K, compared to the experimental value obtained by DSC (8000 [±2000] J/mole-K) (Table 1) confirming typical hydration/protein characteristics.

In contrast, refolding of bikunin aggregates is associated with a small (–700 [±900] J/mole-K) (Table 3) change in heat capacity. We hypothesize that this is the result of a hydrophobic burial effect. The refolding reaction is accompanied by a reduction of exposed hydrophobic surface area and a corresponding decrease in hydration effects. This argument is supported by CD and UV spectroscopy, which demonstrated that tyrosine groups were more solvent exposed in the aggregate than in the monomer (Figs. 2, 3). Furthermore, bis-ANS studies showed that fluorescence intensity increased by 10% in the presence of aggregates compared with solutions containing monomer (data not shown). Last, RP-HPLC elution times for monomer were shorter than those for aggregates.

V_denaturation and V_refolding:Cavities and hydration effects
Pressure favors molecular transitions that decrease total system volume (a negative V). The partial molar volume of protein can be described by the equation V_protein = V_atoms + V_cavities + V_hydration, where V_atoms is the volume of all atoms, V_cavities is the volume of void cavities within the protein, and V_hydration is the system volume change associated with solvation (Gekko and Hasegawa 1986; Chalikian 2003). V_atoms remains constant in a system, consequently the measurement of V_refolding represents the sum of V_{hydration, refolding} or the V_{cavities, refolding} (Gekko and Hasegawa 1986; Chalikian 2003).

V_cavities are the result of imperfect packing within the dense interior core of the protein. On denaturation, some of these cavities can be eliminated, contributing toward a negative V_denaturation(Gekko and Hasegawa 1989). In addition, the penetration of water molecules into cavities on pressurization disrupts native structure (Akasaka and Li 2001; Kamatari et al. 2001; Kuwata et al. 2001; Akasaka 2003; Williamson et al. 2003).

Hydrophobic, electrostatic and hydrogen bonding all contribute to V_hydration. Hydrophobic effects are disfavored on application of pressure (V_{hydration-hydrophobic}). For example, simulations of a water droplet enclosed in a spherical hydrophobic interface provides insight into pressure’s effect on the hydrophobic effect (Wallqvist et al. 2001). The simulation showed that water density near a hydrophobic surface is similar to the density gradient at the water-vapor interface, with gaslike densities of water found at distances up 4 Å away from the hydrophobic surface. On pressurization near 1000 bar, wetting occurs, demonstrating the diminishing of the hydrophobic effect (Wallqvist et al. 2001).

Pressure also disfavors salt bridges through the property of electrostriction (V_{hydration-electrostriction}) (Mozhaev et al. 1996). On average, deprotonation of protein carboxylic acids decreases the total volume by 14 mL/mole, whereas protonation of amine groups decreases volume by 20 mL/ mole (Van Eldik et al. 1989).

In contrast, pressure has been shown to have negligible or slightly stabilizing effects on hydrogen bonds (V_{hydration-hydrogen bonds}). The formation of hydrogen bonded helices from poly (L-lysine) and poly (A+U) random coils exhibits a V of -1 and +1 mL/mole, respectively (Gross and Jaenicke 1994). Application of 2000 bar pressures shortened N-H hydrogen bonds with solvent water and carboxyl groups in basic pancreatic trypsin inhibitor (Li et al. 1998).

Bikunin aggregate has more hydrophobic exposure than the native state, which would be expected to result in a positive contribution of V_{hydration-hydrophobic} to the total V_refolding. Thus, because our measured V_refolding is negative, V_cavities and V_{hydration-electrostriction} must be negative and large enough in magnitude to overcome the positive V_hydration of hydrophobic burial. There is some evidence that this trend is specific not only for bikunin but also applies to other nonnative insoluble aggregates. Bis-ANS binding studies of nonnative carbonic anhydrase aggregate have also shown higher levels of hydrophobic exposure (Kundu and Guptasarma 2002).

The magnitude and sign of the V_{hydration-electrostriction} cannot be obtained, but general observations can be made. In the presence of 200 mM KCl, salt bridges with greater than 50% solvent accessible surface area are effectively shielded and do not contribute to protein stability (Takano et al. 2000). The refolding buffer used with bikunin contained 50 mM Tris and 157 mM NaCl; thus, we would expect pressure to influence only short-range solvent inaccessible surface salt-bridges. In these cases, pressures of 1500 bar would decrease salt bridge bond strength by approximately 3 kJ/mole. In comparison, salt bridges with 10% and 30% solvent accessibility in human lysozyme had bond strengths of 6 kJ/mole and 3.5 kJ/mole, respectively (Takano et al. 2000). Salt bridges buried within the protein core would be inaccessible to water without significant structural perturbation, which was not seen.

The relative packing between the aggregate and the monomer could also be an important parameter in determining whether pressure will be an effective refolding tool. Recent NMR studies of protein unfolding have drawn an analogous conclusion (Akasaka and Li 2001; Kamatari et al. 2001). Specifically, the presence of cavities within the protein core is the most important parameter in determining its stability toward pressure and provides the sites of pressure-modulated structural changes. Again, the magnitude and sign of the V_cavities cannot be discerned in this case.

The effect of temperature on refolding yield
Refolding yields are increased at lower temperatures in the range 0°–50°C. This effect does not increase the stability of the native monomer (calculated from equation 1); rather, the entropic cost of refolding is decreased and favors aggregate dissociation. As temperature is decreases from 25° to 0°C, H_refolding decreased by 16.5 kJ/mole, whereas (TS) decreased by 14.6 kJ/mole, resulting in a net decrease in G_refolding of 1.9 kJ/mole. Interestingly, this result was not seen in the case of recombinant human growth hormone, where increased temperatures improved refolding by disrupting strong nonnative hydrogen bonds (St. John et al. 2001). Addition of nondenaturing levels of guanidine HCl increases refolding yields of human growth hormone from aggregates, presumably by disrupting hydrogen bonds by preferentially binding to the protein (St. John et al. 2001). In contrast, addition of guanidine HCl had no effect on the refolding yield of bikunin.

Conclusion: Application of pressure refolding and thermodynamics
This study has quantitatively confirmed that high hydrostatic pressure can effectively refold glycosylated protein containing multiple disulfide bonds from aggregates produced during mammalian cell culture. Refolding yields of 70% (±5%) were obtained at 2000 bar, which were significantly higher than yields obtained through chaotrope refolding (55% [±6%]). Chaotrope refolding requires folding from the denatured state, which can lead to increased reaggregation. Pressure is an advantageous refolding tool because of its ability to dissociate aggregates while favoring the native confirmation. This key attribute allows folding to be initiated from a state that is less perturbed than that found in traditional chaotrope refolding processes, resulting in higher yields of active protein than those found after conventional refolding.

Under pressure, bikunin aggregates are preferentially destabilized, shifting equilibrium toward the native state, with a V_refolding of –28 (±5) mL/mole. Refolding is not hydrophobically driven: Numerous experiments verified that the aggregate contained higher levels of hydrophobic exposure than the native state. To enable pressure refolding, the V_cavities and V_{hydration-electrostriction} must be negative and large enough in magnitude to overcome the positive V_{hydration-hydrophobic}.

	Materials and methods

TOP Abstract Introduction Results Discussion Materials and methods References

 
Materials
Recombinant human placental bikunin was produced by fermentation in Chinese hamster ovary cells. During fermentation, a substantial fraction of the bikunin formed soluble aggregates. Samples of monomeric (6.3 mg/mL) and aggregated (7.3 mg/mL and 9.2 mg/mL) bikunin fractions were used as received from Bayer Corp. The protein was stored frozen, thawed before use, and diluted to a concentration of 0.5 mg/mL for refolding studies. All reagents and buffers were obtained from Sigma Corp. The 100-mM solutions of dithiothreitol (DTT) and oxidized glutathione (GSSG) were stored frozen and thawed immediately before use to prevent oxidation. The free thiol content of the DTT stock solution was periodically checked using Ellman’s reagent to ensure that oxidation had not occurred. Sodium citrate, sodium phosphate, 2-amino-2–(hydroxymethyl)-1,3-propanediol (Tris), guanidine HCl, and sodium chloride were used to produce various buffers. For monomer spectroscopy and low pH refolding, a sodium citrate buffer (pH 5.8) was used at a concentration of 22 mM along with 150 mM NaCl. For aggregate spectroscopy, a sodium phosphate buffer was used (pH 7.2) at concentration of 100 mM along with 150 mM NaCl. Tris at a concentration of 50 mM was used to prepare all refolding buffers at various pH (7.0–9.0). Tris was used as a buffer system of choice because of its minimal pH shift at elevated pressures (Byrne and Laurie 1999). Sodium chloride was used at a concentration of up to 157 mM to adjust the ionic strength of the refolding buffers.

Pressure generation
Pressure was generated by using 10-fold hydraulic intensifier equipment driven by high-pressure nitrogen (200–400 bar), with oil as a pressure transmitting fluid (High Pressure Equipment). A 2-L cloverleaf reactor (rated at 2000 bar) was used in conjunction with the hydraulic intensifier (High Pressure Equipment). In addition, manually driven 4000-bar and 6750-bar high-pressure generators were used with water as a pressure transmitting fluid (High Pressure Equipment). Custom-built pressure cells were used with these pressure generators.

Pressure refolding
Aggregated bikunin at a concentration of 0.5 mg/mL in various solutions was placed in heat-sealed 1-mL polypropylene syringes (enabling pressure transfer) along with 4 mM GSSG and 2 mM DTT to provide a redox environment for disulfide shuffling. Samples were slowly pressurized (over 10 min) to minimize system heating and were then held at the desired temperature and pressure. Two different depressurization protocols were used. For slow depressurization, pressures were reduced by 5% of the initial pressure every 30 min until a pressure 20% of the initial pressure was reached. From this point, pressure was reduced by 10% (of the initial pressure) every 30 min until atmospheric pressure was obtained (about 8 h). Rapid depressurization occurred over a few seconds, when the pressurized system was opened to atmosphere. After depressurization, SEC-HPLC or RP-HPLC was conducted while the remaining samples were stored frozen at –20°C until further testing with SEC-HPLC, RP-HPLC, GEMMA, and activity assays. Freeze–thaw studies conducted at Bayer verified that monomer structure was unaffected by multiple freeze–thaw cycles (data not shown).

In addition, no significant changes in refolding yields and chromatograms were seen before and after freezer storage through SEC-HPLC and RP-HPLC. Refolding yield was determined by dividing the monomer mass fraction by the total protein concentration (monomer and aggregate) and then converting the result to a percentage.

Chaotrope-modulated refolding (guanidine HCl)
Aggregated bikunin was incubated at 37°C in the presence of 6M guanidine HCl and 12 mM DTT to reduce disulfide bonds and disassociate the aggregate (Clark et al. 1998). Two refolding protocols were employed, "dialysis refolding" and dilution refolding. Dialysis refolding was conducted by placing the denatured protein (0.5 mg/mL) in a dialysis cassette and submerging it in excess refolding buffer of 50 mM Tris (pH 8.0), 157 mM NaCl, 4 mM GSSG, and 2 mM DTT. The samples were dialyzed overnight. Dilution refolding was conducted by rapidly diluting solutions containing 50 mM Tris (pH 8.0), 157 mM NaCl, 12 mM DTT, and 6 M guanidine HCl to a final solution containing 50 mM Tris (pH 8.0), 157 mM NaCl, 4 mM GSSG, 2 mM DTT, and 1 M guanidine HCl at a protein concentration of 0.4 mg/mL. After dilution, the samples were incubated at 25°C for 8 h (Clark et al. 1998). The samples were then frozen and stored until analysis by SEC-HPLC, RP-HPLC, GEMMA, and activity assays.

SEC-HPLC
SEC-HPLC was used to determine the ratio of monomeric protein compared to aggregates before and after refolding. SEC-HPLC analysis of protein fractions was conducted on a Beckman Gold HPLC system (Beckman Coulter) equipped with a TosoHaas 2000 SW_XL size exclusion column (TosoHaas Biosep), a 2000 SW_XL guard column, and a 0.2–µm prefilter. A mobile phase of 0.5 M KCl at a rate of 0.6 mL/min was used, with an 80 µL sample injection from a Beckman autosampler (Beckman Coulter). Absorbance was monitored at 215 nm. Aggregate and monomer mass fractions were obtained by integrating the area under their respective peaks. Purified aggregate and monomer standards were used to identify peaks on the SEC chromatograms and to determine the extinction coefficients.

RP-HPLC
RP-HPLC was used to determine the percent of native monomer compared to more hydrophobic nonnative monomers and larger aggregates. RP-HPLC analysis of protein fractions was conducted on a Beckman Gold HPLC system (Beckman Coulter) equipped with a Vydac C18 reverse phase column (Grace Vydac), a matching guard column, and a 0.2-µm prefilter. A gradient chromatography method was used at 50°C, with water containing 1% trifluo-roacetic acid (TFA) as the aqueous phase and acetonitrile with 0.5% TFA as the organic phase. The gradient started at 27% ace-tonitrile, increasing in concentration to 29% after 5 min, 33% after 13 min, 49% after 21 min, 95% after 22 min, and back to 27% at 25 min, until the run ended after 30 min. The total flowrate was maintained at 1 mL/min throughout the entire run. Eighty micro-liters of sample was injected with a Beckman autosampler (Beck-man Coulter). Absorbance was monitored at 215 nm, and aggregate and monomer mass fractions were obtained as in SEC-HPLC.

Kinetic assay of in vitro protease inhibition
The in vitro functional activity of bikunin was measured by its inhibition of the protease activity of human plasma kallikrein. Bikunin samples (0–26 nM) were preincubated with 10 nM human plasma kallikrein (Enzyme Research Laboratories) in 50 mM Tris buffer (pH 8.0) containing 0.1 M NaCl and 0.01% Triton X-100 at 37°C for 30 min. After this preincubation, the protease substrate N-benzoyl-Pro-Phe-Arg-p-nitroanilide was added (156 µM final concentration), and protease activity was monitored optically at 405 nm. The functional concentration of bikunin was determined assuming the presence of two active sites within kallikrein.

CD spectroscopy
Far and near UV CD spectra of bikunin monomer and aggregates were collected on an Aviv 62DS instrument (Proterion). Two scans were averaged and corrected for buffer absorbance. For near-UV CD spectroscopy, 0.5 mg/mL protein in a 1-cm quartz cuvette was scanned from 250 to 340 nm. For far-UV CD spectroscopy, 0.1 mg/mL of protein was scanned from 200 to 260 nm in a quartz cuvette with a pathlength of 0.1 cm. The mean residue molar ellipticity (deg-cm²/dmole) was calculated through the equation:

where _obs is the observed ellipticity, MW is the protein molecular weight, residues is the number of residues, c is the protein concentration (in grams per milliliter), and L is the cell path length (in centimeters).

IR spectroscopy
IR spectra were collected on a Bomem MB-104 IR spectrometer (ABB/Bomem Inc.). Solutions of monomer (15 mg/mL) or aggregates (15 mg/mL) were placed in a BioTools liquid sampling cell, equipped with CaF₂ windows (BioTools Inc.). A 6-µ path length was used. Spectral acquisition and analysis were conducted as previously discussed (Dong et al. 1990) with all mathematical operations conducted in GRAMS software (Thermo Electron Corp.).

Hydrophobic binding studies
4,4'-dianilino-1,1'-binaphthyl-5,5'- disulfonic acid (bis-ANS) was used as a fluorescent probe to examine the hydrophobicity of monomer and aggregates. Aggregate and monomer solutions (0.25 mg/mL) were made with sodium citrate buffer (pH 5.8), 150 mM NaCl with 20 molar excess bis-ANS. Fluorescence measurements were made on an FP-112 fluorimeter (SLM-Aminco) at an excitation wavelength of 395 nm. Emission spectra of 1-mL solutions were scanned from 440 to 510 nm. Peak maximums were obtained near 495 nm and were used to relate the relative hydrophobicity of monomer and aggregate. Appropriate blanks were subtracted from each scan.

UV spectroscopy
Standard UV spectra were obtained on a PerkinElmer Lambda 35 spectrometer (PerkinElmer) for bikunin aggregate and monomer at two concentrations of 0.5 and 0.15 mg/mL. High-pressure UV spectra of monomeric bikunin were collected with the spectrometer equipped with a custom high-pressure cell rated to 6500 bar. Monomer (0.2 mg/mL) was placed in the refolding buffer (50 mM Tris at pH 8.0, 157 mM NaCl), and UV spectra were acquired from 320 to 270 nm as a function of pressure (from 0 to 3000 bar), with 100-bar pressure steps. Samples were equilibrated for 3 min after each change in pressure. The pressure cell was equipped with a heat exchanger to eliminate temperature changes during the pressurization cycle. Derivatives of the spectra were calculated using UV Winlab software (PerkinElmer) and also via in-house software using a Savitsky-Golay algorithm coupled with MATLAB (MathWorks Inc.). Both calculation techniques employ 35-point data smoothing. Structural changes were analyzed by determining the 2D-UV r-ratio as a function of pressure (Ragone et al. 1984; Kornblatt et al. 1995). The r-ratio is a useful technique to examine the solvent environment near tyrosine residues. The r-ratio is calculated by dividing the peak-to-peak distance at 287–283 nm in the 2D-UV spectrum by the peak-to-peak distance at 295–290.5 nm (Ragone et al. 1984). In the case of bikunin, decreases in the r-ratio are indicative of increased solvent exposure of tyrosine groups and subsequent denaturation.

DSC
A MicroCal VP-DSC microcalorimeter (MicroCal LLC) was used to evaluate the thermodynamics of monomer denaturation. A thermal scan of monomeric bikunin at a concentration of 1 mg/mL was collected from 15°–95°C, at a scan rate of 90°C/h. For analysis, MicroCal’s ORIGIN software (MicroCal LLC) was used. The buffer background was subtracted from the denaturation scan and then normalized for protein concentration. A two-state model was fit to the resulting heat capacity plot to evaluate T_m, the temperature midpoint of the thermally induced denaturation; the enthalpy change for denaturation, H; and the heat capacity change for denaturation, Cp. The fitting program implements a Levenberg-Marquardt model to minimize the chi-squared value and, consequently, the model error. The T_m, H, and Cp values obtained from the model were used in the Gibbs-Helmholtz equation to obtain the free energy of denaturation, G_denaturation at 25°C. SEC-HPLC was used to determine that aggregation had not occurred during thermal denaturation and that unfolding was reversible.

GEMMA
GEMMA (TSI) is a relatively new tool that is effective for determining the size and distribution of aggregates and other protein mixtures (Bacher et al. 2001). We used GEMMA because it provides better resolution of aggregate species than does either SEC or RP-HPLC. In the GEMMA technique, protein particles are volatilized by electrospray, separated by gas phase electrophoretic-mobility, and counted via a condensation particle counter. The particle size distribution can be converted to a molecular weight via a calibration curve (Bacher et al. 2001). GEMMA was used to obtain the molecular weight distribution of aggregates before and after refolding. Because a volatile buffer is required for GEMMA analysis, samples were prepared by dialyzing 100 µL of protein (0.5 mg/mL) into 20 mM ammonium acetate buffer at pH 7.0 overnight. After dialysis, samples were diluted to 2 µg/mL, a sufficiently low concentration to meet the instrumental requirement of a single protein molecule or aggregate per electrosprayed droplet, and were then sized using a model 3480 electrospray generator, model 3080 classifier, model 3085 DMA, and a model 3025 ultrafine condensation particle counter.

Thermodynamic calculations: Equal K model
Aggregate size distributions obtained with GEMMA analysis were used to calculate equilibrium constants using a previously described model (Martin 1996). The model assumes that the free energy of monomer addition to an aggregate is independent of aggregate size and has previously been used to characterize the aggregation behavior of lysozyme (Price et al. 1999). At the low levels of higher-order aggregates observed experimentally, aggregates of size pentamer or larger could not be accurately resolved, and thus they were combined with the tetramer concentration. This assumption resulted in a small overestimation of the free energy of refolding, G_refolding. However, neglecting these terms would have produced higher error. The molecular weight distribution obtained through GEMMA was converted to concentration by using RP-HPLC data to quantify total aggregate and monomer concentrations. A reference protein concentration was chosen such that G_refolding = 0 corresponds to one-half of the total protein being found in aggregates. It was assumed that the equilibrium distributions of assembly states obtained at elevated pressures were maintained after depressurization. This assumption was tested by changing the depressurization rate; no dependence of the observed assembly state distribution on depressurization rates was observed. The equilibrium constant for refolding from an aggregate of size N + 1 to form an aggregate of size N and a folded monomer unit

was calculated using the equation

where is equal to the mole fraction of monomer and L is equal to the sample concentration divided by the equilibrium constant K_refolding (moles/L), multiplied by the necessary reference constant (26,956) (Martin 1996). The variable G_refolding was obtained through the equation G_refolding = –RT ln K_refolding and refers to the energy change for forming 1 mole of monomer from aggregate.

Statistics
All error bars listed in the figures and text were derived from 95% confidence intervals of samples run in triplicate determined using standard statistical methods (Himmelblau 1970; Taylor 1997).

	Acknowledgments

 
We thank Willy Grothe for the design and manufacture of our high-pressure equipment. We also would like to thank Anasuya Mitra for conducting the kallikrein assay. Dr. Yong Sung Kim provided technical assistance with CD and IR spectroscopy, with analysis provided by Dr. Mark Manning and Dr. Aichun Dong. Dr. Paul Cachia provided technical assistance with the DSC. Michael Stoner developed the software for the calculation of 2D-UV spectra. Dr. David Clough provided insight toward statistical analysis.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.

	References

TOP Abstract Introduction Results Discussion Materials and methods References

 
Akasaka, K. 2003. Exploring the entire conformational space of proteins by high-pressure NMR. Pure Appl. Chem. 75: 927–936.

Akasaka, K. and Li, H. 2001. Low-lying excited states of proteins revealed from nonlinear pressure shifts in H-1 and N-15 NMR. Biochemistry 40: 8665–8671.[CrossRef][Medline]

Bacher, G., Szymanski, W.W., Kaufman, S.L., Zollner, P., Blaas, D., and All-maier, G. 2001. Charge-reduced nano electrospray ionization combined with differential mobility analysis of peptides, proteins, glycoproteins, non-covalent protein complexes and viruses. J. Mass Spectrom. 36: 1038–1052.[CrossRef][Medline]

Balny, C. 2002. High pressure and protein oligomeric dissociation. High Pressure Res. 22: 737–741.[CrossRef]

Braakman, I., Helenius, J., and Helenius, A. 1992. Role of Atp and disulfide bonds during protein folding in the endoplasmicreticulum. Nature 356: 260–262.[CrossRef][Medline]

Byrne, R.H., and Laurie, S.H. 1999. Influence of pressure on chemical equilibria in aqueous systems—with particular reference to seawater. Pure Appl. Chem. 71: 871–890.

Chalikian, T.V. 2003. Volumetric properties of proteins. Annu. Rev. Biophys. Biomol. Struct. 32: 207–235.[CrossRef][Medline]

Chatani, E., Nonomura, K., Hayashi, R., Balny, C., and Lange, R. 2002. Comparison of heat- and pressure-induced unfolding of ribonuclease A: The critical role of Phe46 which appears to belong to a new hydrophobic chain-folding initiation site. Biochemistry 41: 4567–4574.[CrossRef][Medline]

Chi, E.Y., Krishnan, S., Kendrick, B.S., Chang, B.S., Carpenter, J.F., and Randolph, T.W. 2003a. Roles of conformational stability and colloidal stability in the aggregation of recombinant human granulocyte colony-stimulating factor. Protein Sci. 12: 903–913.[Abstract/Free Full Text]

Chi, E.Y., Krishnan, S., Randolph, T.W., and Carpenter, J.F. 2003b. Physical stability of proteins in aqueous solution: Mechanism and driving forces in nonnative protein aggregation. Pharm. Res. 20: 1325–1336.[CrossRef][Medline]

Chu, L. and Robinson, D.K. 2001. Industrial choices for protein production by large-scale cell culture. Curr. Opin. Biotechnol. 12: 180–187.[CrossRef][Medline]

Clark, E.D., Hevehan, D., Szela, S., and Maachupalli-Reddy, J. 1998. Oxidative renaturation of hen egg-white lysozyme. Folding vs. aggregation. Biotechnol. Prog. 14: 47–54.[CrossRef][Medline]

Dong, A., Huang, P., and Caughey, W.S. 1990. Protein secondary structures in water from 2nd-derivative amide-I infrared-spectra. Biochemistry 29: 3303–3308.[CrossRef][Medline]

Edelhoch, H. and Osborne, J.C. 1976. The thermodynamic basis of the stability of proteins, nucleic acids, and membranes. Adv. Prot. Chem. 30: 183–250.[Medline]

Ferrao-Gonzales, A.D., Souto, S.O., Silva, J.L., and Foguel, D. 2000. The preaggregated state of an amyloidogenic protein: Hydrostatic pressure converts native transthyretin into the amyloidogenic state. Proc. Natl. Acad. Sci. 97: 6445–6450.[Abstract/Free Full Text]

Gekko, K. and Hasegawa, Y. 1986. Compressibility structure relationship of globular-proteins. Biochemistry 25: 6563–6571.[CrossRef][Medline]

———. 1989. Effect of temperature on the compressibility of native globular-proteins. J. Phys. Chem. 93: 426–429.[CrossRef]

Gilbert, H.F. 1995. Thiol/disulfide exchange equilibria and disulfide bond stability. In Biothiols, Part A, pp. 8–28. Academic Press, New York.

Gross, M. and Jaenicke, R. 1994. Proteins under pressure—The influence of high hydrostatic pressure on structure, function and assembly of proteins and protein complexes. Eur. J. Biochem. 221: 617–630.[Medline]

Hawley, S.A. 1971. Reversible pressure-temperature denaturation of chymotrypsinogen. Biochemistry 10: 2436–2442.[CrossRef][Medline]

Himmelblau, D.M. 1970. Process analysis by statistical methods. John Wiley & Sons, New York.

Israelachvili, J. 1992. Intermolecular and surface forces. Academic Press, San Diego, CA.

Kamatari, Y.O., Yamada, H., Akasaka, K., Jones, J.A., Dobson, C.M., and Smith, L.J. 2001. Response of native and denatured hen lysozyme to high pressure studied by N-15/H-1 NMR spectroscopy. Eur. J. Biochem. 268: 1782–1793.[Medline]

Kim, Y.S., Wall, J.S., Meyer, J., Murphy, C., Randolph, T.W., Manning, M.C., Solomon, A., and Carpenter, J.F. 2000. Thermodynamic modulation of light chain amyloid fibril formation. J. Biol. Chem. 275: 1570–1574.[Abstract/Free Full Text]

Kornblatt, J.A., Kornblatt, M.J., and Hoa, G.H.B. 1995. 2nd derivative spectroscopy of enolase at high hydrostatic pressure—An approach to the study of macromolecular interactions. Biochemistry 34: 1218–1223.[CrossRef][Medline]

Kumar, S. and Nussinov, R. 2002. Close-range electrostatic interactions in proteins. Chembiochem 3: 604–617.[CrossRef][Medline]

Kundu, B. and Guptasarma, P. 2002. Use of a hydrophobic dye to indirectly probe the structural organization and conformational plasticity of molecules in amorphous aggregates of carbonic anhydrase. Biochem. Biophys. Res. Comm. 293: 572–577.[CrossRef][Medline]

Kunugi, S. and Tanaka, N. 2002. Cold denaturation of proteins under high pressure. Biochim. Biophys. Acta-Protein Struct. Molec. Enzym. 1595: 329–344.

Kuwata, K., Li, H., Yamada, H., Batt, C.A., Goto, Y., and Akasaka, K. 2001. High pressure NMR reveals a variety of fluctuating conformers in -lacto-globulin. J. Mol. Biol. 305: 1073–1083.[CrossRef][Medline]

Li, H., Yamada, H., and Akasaka, K. 1998. Effect of pressure on individual hydrogen bonds in proteins, basic pancreatic trypsin inhibitor. Biochemistry 37: 1167–1173.[CrossRef][Medline]

Martin, R.B. 1996. Comparisons of indefinite self-association models. Chem. Rev. 96: 3043–3064.[CrossRef][Medline]

Mozhaev, V.V., Heremans, K., Frank, J., Masson, P., and Balny, C. 1996. High pressure effects on protein structure and function. Proteins 24: 81–91.[CrossRef][Medline]

Price, W.S., Tsuchiya, F., and Arata, Y. 1999. Lysozyme aggregation and solution properties studied using PGSE NMR diffusion measurements. J. Am. Chem. Soc. 121: 11503–11512.[CrossRef]

Privalov, P.L. 1990. Cold denaturation of proteins. Crit. Rev. Biochem. Mol. Biol. 25: 281–305.[Medline]

———. 1997. Thermodynamics of protein folding. J. Chem. Thermodyn. 29: 447–474.[CrossRef]

Ragone, R., Colonna, G., Balestrieri, C., Servillo, L., and Irace, G. 1984. Determination of tyrosine exposure in proteins by 2nd-derivative spectroscopy. Biochemistry 23: 1871–1875.[CrossRef][Medline]

Royer, C.A. 2002. Revisiting volume changes in pressure-induced protein unfolding. Biochim. Biophys. Acta-Protein Struct. Molec. Enzym. 1595: 201–209.

Schroder, M. and Friedl, P. 1997. Overexpression of recombinant human anti-thrombin III in Chinese hamster ovary cells results in malformation and decreased secretion of recombinant protein. Biotechnol. Bio