Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2001 Apr;10(4):741–752. doi: 10.1110/ps.18501

Roles of dimerization in folding and stability of ketosteroid isomerase from Pseudomonas putida biotype B

Do-Hyung Kim 1,2,4, Gyu Hyun Nam 1, Do Soo Jang 1, Sunggoo Yun 2,3, Gildon Choi 1,2, Hee Cheon Lee 2,3, Kwan Yong Choi 1,2
PMCID: PMC2373975  PMID: 11274465

Abstract

Equilibrium and kinetic analyses have been performed to elucidate the roles of dimerization in folding and stability of KSI from Pseudomonas putida biotype B. Folding was reversible in secondary and tertiary structures as well as in activity. Equilibrium unfolding transition, as monitored by fluorescence and ellipticity measurements, could be modeled by a two-state mechanism without thermodynamically stable intermediates. Consistent with the two-state model, one dimensional (1D) NMR spectra and gel-filtration chromatography analysis did not show any evidence for a folded monomeric intermediate. Interestingly enough, Cys 81 located at the dimeric interface was modified by DTNB before unfolding. This inconsistent result might be explained by increased dynamic motion of the interface residues in the presence of urea to expose Cys 81 more frequently without the dimer dissociation. The refolding process, as monitored by fluorescence change, could best be described by five kinetic phases, in which the second phase was a bimolecular step. Because <30% of the total fluorescence change occurred during the first step, most of the native tertiary structure may be driven to form by the bimolecular step. During the refolding process, negative ellipticity at 225 nm increased very fast within 80 msec to account for >80% of the total amplitude. This result suggests that the protein folds into a monomer containing most of the α-helical structures before dimerization. Monitoring the enzyme activity during the refolding process could estimate the activity of the monomer that is not fully active. Together, these results stress the importance of dimerization in the formation and maintenance of the functional native tertiary structure.

Keywords: Ketosteroid isomerase, folding, dimerization

Δ5–3-KSI is an enzyme that catalyzes the conversion of a variety of 3-oxo-Δ5-steroids to their conjugated Δ4-isomers by an intramolecular proton transfer (Scheme 1 ▶) (Batzold et al. 1976). KSI is one of the most proficient enzymes, enhancing the catalytic rate by a factor of 11 orders of magnitude compared with the corresponding nonenzymatic reaction (Pollack et al. 1989). This enzyme has been the subject of intensive studies as a prototype for understanding the enzyme mechanism of the allylic rearrangement (Wang et al. 1963; Hawkinson et al. 1991, 1994; Xue et al. 1991). Recent determinations of X-ray crystal structures as well as NMR solution structures of the enzymes from Pseudomonas putida biotype B and Comamonas testosteroni have contributed significantly to understanding the enzyme mechanism for efficient catalysis (Kim et al. 1997a; Wu et al. 1997; Cho et al. 1998; Massiah et al. 1998).

Scheme 1.

Scheme 1.

Open in a new tab

KSI catalyzed isomerization reaction. The reaction is known to be stereospecific; the β proton at C4 of the substrate, 5-androstene-3,17-dione, is transferred to the catalytic base of the enzyme to generate the reaction intermediate in the middle, and then the same proton is transferred to the β side of C6 to generate the product, 4-androstene-3,17-dione.

Recently, C. testosteroni KSI was subject to an analysis of its folding mechanism (Kim et al. 2000b). As a small dimeric protein, KSI provides a good model system for folding mechanism studies of dimeric proteins. It is highly expressed in Escherichia coli and conveniently purified by affinity chromatography (Kuliopulos et al. 1987a; Kim et al. 1994). KSI is a homodimeric enzyme consisting of 125 or 131 amino acids for C. testosteroni KSI (TI) or P. putida KSI (PI), respectively (Benson et al. 1971; Kim et al. 1994). The two enzymes are 34% identical in the amino acid sequences. Their crystal structures revealed that they are folded into six-stranded β-sheets and three α-helices in each monomer (Fig. 1 ▶) (Kim et al. 1997a; Cho et al. 1998). One of the most noticeable features of the structures is that each monomer forms a deep apolar active-site cavity made of the residues from the same monomer. Most of the active-site cavity is lined with hydrophobic amino acids and includes aromatic residues that play important roles in steroid binding (Kim et al. 1999). The catalytic residues are located at the bottom of the cavity forming a hydrogen-bond network (Kim et al. 2000a). Another interesting structural feature of KSI is that the monomers interact with each other over a narrow and long patch of β-sheet of each monomer (Kim et al. 1997a; Massiah et al. 1998). The dimeric interface is well defined and is formed between the convex surfaces of the β-sheets. Many neutral or apolar amino acids are found within 3.8 Å from the partner monomer, suggesting that hydrophobic interaction is important for dimerization. As well, such residues as Val 74, Arg 75, Ala 76, His 78, Val 101, Ser 121, Gln 122, and Asn 124 in PI are also involved in the intermolecular interaction by forming hydrogen bonds.

Fig. 1.

Fig. 1.

Open in a new tab

Ribbon diagram of Pseudomonas putida KSI (Kim et al. 1997a) seen in a direction of twofold symmetry axis. The side chains responsible for fluorescence signal are displayed and labeled. Residues 1–131 and 201–331 refer to the two subunits, respectively. The program Molscript (Kraulis 1991) was used to draw the figure.

Dimerization is commonly found even when dimers are made of identical monomers containing all of the functional groups. Contribution of dimerization to conformational stability can be analyzed by equilibrium unfolding studies, which allow for identification of monomeric intermediates for some dimeric proteins (Jaenicke 1987; Park and Bedouelle 1998). Dimerization also plays an important role in folding of dimeric proteins. Dimeric protein folding proceeds via two steps of monomeric folding and association, which are either concerted or sequential. For P22 Arc repressor, monomeric folding occurs in a concerted manner with dimerization (Milla et al. 1995; Walburger et al. 1996). In cases of proteins including E. coli Trp repressor (Gloss and Matthews 1998), triosephosphate isomerase (Rietveld and Ferreira 1998), and human glutathione transferase A1–1 (Wallace and Dirr 1999), dimerization induces monomeric intermediates to fold into the complete tertiary structure. Recent equilibrium and kinetic analysis of TI folding revealed that a monomeric folding intermediate is generated transiently during the refolding process but thermodynamically stable intermediates were not detected (Kim et al. 2000b). The folding mechanisms of PI and TI should commonly involve the formation of the apolar active-site cavity as well as the dimeric interactions, in both of which many hydrophobic residues participate. Comparison of their folding mechanisms may allow for characterization of specific recognition at the dimeric interface essential for folding and may provide a valuable insight into the roles of long-range or quaternary interactions in the folding mechanism.

In this paper, we report the results of both equilibrium and kinetic studies of PI folding. The folding reversibility was confirmed by comparing fluorescence, CD, onedimensional (1D) NMR spectra, and activity of the refolded protein with those of the native protein. Urea-induced equilibrium unfolding was investigated by fluorescence and ellipticity measurements. The unfolding experimental results were consistent with a two-state mechanism involving only the native dimer and the unfolded monomer. Size-exclusion chromatography and 1D NMR spectral analyses were performed but did not show any evidence for a monomeric intermediate at equilibrium. Analysis of chemical modification of a dimeric interface cysteine residue monitored the stability of the interface interaction. Refolding kinetics was analyzed by fluorescence and ellipticity measurements during the refolding process. Refolding kinetics identified a bimolecular refolding step and a monomeric intermediate containing most of the α-helical structures. Significant functional roles of dimerization were also deduced from measurements of enzyme activity during the refolding process.

Results

Reversibility of folding

Reversibility of PI folding was investigated by comparing fluorescence and CD spectra of the refolded protein with those of the native protein after successively diluting the unfolded protein in 8 M urea to lower urea concentrations in a reducing condition. The unfolded protein exhibited a maximum emission band at 355 nm almost matching that of free tryptophan. The fluorescence intensity decreased in the range of 327–400 nm and increased in the range of 300–327 nm, and the maximum wavelength was blue-shifted as the urea concentration was lowered (Fig. 2A ▶). The fluorescence spectrum of the refolded protein was almost identical to that of the native protein with the maximum emission at 340 nm (Fig. 2B ▶). These spectral patterns are expected for PI containing two tryptophan residues, Trp 92 and Trp 120, located at the protein surface and the hydrophobic active site, respectively (Fig. 1 ▶). The good agreement between the fluorescence spectra of the native and the refolded proteins indicates that the overall tertiary structure was recovered after refolding. CD spectra of the refolded and native proteins were also identical in the range of 205–300 nm with the very similar intensities at 212 and 222-nm bands. A marginal difference might originate from the presence of urea in the refolding buffer. These results suggest that the secondary and tertiary structures were recovered after refolding.

Fig. 2.

Fig. 2.

Open in a new tab

Reversibility of PI folding in secondary and tertiary structures. Fluorescence (A) and CD (B) spectra of native, unfolded, and refolded PI are displayed for comparison. The fluorescence spectra of the proteins were obtained after excitation at 285 nm. The protein concentration was 15 μM. The unfolded protein was prepared by incubating the protein at 7 M urea longer than 48 h. The refolded protein was prepared by diluting the unfolded protein to 0.2 M urea and incubating longer than 48 h. The experiments were performed at 25°C in the buffer containing 20 mM potassium phosphate at pH 7.0, 0.5 mM EDTA, and 1 mM DTT.

The folding reversibility was also confirmed by 1D NMR spectral analysis of the native and refolded enzymes. As shown in Fig. 3 ▶, the NMR spectrum of the refolded protein was matched to that of the native protein in the overall spectral range. This result strongly supports that the native and refolded proteins are almost the same in the three-dimensional structures. Consistent with this result, enzyme activity of the refolded protein was recovered up to over 95% of that of the native protein (data not shown).

Fig. 3.

Fig. 3.

Open in a new tab

1D NMR spectra of PI at different urea concentrations. The protein was incubated in the presence of urea longer than 48 h for equilibration. The representative spectra obtained at 0, 3, 5.5, and 8 M urea are displayed. The refolded protein was prepared by dialyzing the denatured protein with the refolding buffer containing 20 mM potassium phosphate, 0.5 mM EDTA, and 3 mM DTT. The protein concentration was 500 μM for all of the cases. The spectra were obtained according to the protocol as described in Materials and Methods. The peaks labeled a, b, and c represent some of the side chain protons of Leu 70, Val 88, and Tyr 16/Tyr 57, respectively. DTT peaks are labeled d.

Equilibrium studies

Equilibrium unfolding of PI was investigated by monitoring the fluorescence intensity at 320 nm, the wavelength at the maximum fluorescence, and the CD ellipticity at 222 nm as a function of urea concentration at 25°C. The data were compared with one another by normalizing the results to the apparent fraction of unfolded protein, FU (Fig. 4A ▶). The fluorescence and ellipticity intensities as well as the maximum wavelength did not change significantly up to and including 4 M urea, whereas a drastic change was observed between 4 and 6 M urea. The transition curves were very similar for the three different kinds of measurements. This result is consistent with a two-state transition between native dimers and unfolded monomers without any thermodynamically stable intermediates.

Fig. 4.

Fig. 4.

Open in a new tab

Equilibrium unfolding transition of PI. (A) The equilibrium unfolding transitions monitored by fluorescence and CD measurements are displayed. PI was incubated in the buffer containing 20 mM potassium phosphate at pH 7.0, 0.5 mM EDTA, 1 mM DTT, and urea at different concentrations. The fluorescence intensity at 320 nm (υ) and the maximum wavelength (×) were obtained after excitation at 285 nm, and the CD ellipticity was measured at 222 nm (O). The protein concentration was 15 μM. (B) Protein concentration dependence of the equilibrium unfolding transition. The equilibrium unfolding transition was analyzed by the fluorescence measurement at three different protein concentrations of 1, 5, and 25 μM. The Y-axis represents the fraction of unfolded protein. The data points were fitted to equation 2 to obtain the transition curve.

When the equilibrium unfolding experiment was performed at different protein concentrations of 1, 5, and 25 μM, the midpoints of unfolding transitions were about 4.6, 5.0, and 5.4 M urea, respectively (Fig. 4B ▶). This trend is consistent with the two-state model describing the protein concentration dependence of bimolecular reactions (Bowie and Sauer 1989). When the spectroscopic data were fitted to a two-state model, the free energy difference, ΔGUH2O, and the m value were determined to be ∼24.0 ± 2.4 kcal•mol−1 and 3.30 ± 0.43 kcal•mol−1•M−1, respectively. The m value is comparable to 3.0 kcal•mol−1•M−1 that can be calculated from surface area change on unfolding (Myers et al. 1995).

NMR spectrum of PI at different urea conditions

To characterize the structural feature of PI during unfolding, we observed 1D NMR spectra of PI at different urea concentrations. The NMR spectra at 1–3 M urea were slightly different at some peak positions from those of the native state (Fig. 3 ▶). These small changes in chemical shift and peak intensities by low concentration of denaturant commonly occur in proteins without a large conformational transition (Lumb and Dobson 1992). The overall spectra did not change significantly between 1 and 3 M urea. This suggests that the structures at these urea concentrations might be very similar. During the unfolding transition, noticeable changes were observed at positions of −0.5, 0.3, 0.8, 1.2, 1.5, 2.9, and 3.1 ppm in the aliphatic region. The peak intensity increased gradually near the positions of 0.8, 1.2, 1.5, 2.9, and 3.1 ppm and decreased near the positions of −0.5 and 0.3 ppm as the urea concentration increased. In TI, the peaks between −0.5 and 0.5 ppm in the upfield region had been assigned to the side chain protons of Val-11, -29, -65, -71, -74, -84, -109, -110, and Leu-23, -67 (Kuliopulos et al. 1987b). Among them, Val 65, Val 84, and Leu 67 are located in the hydrophobic active site. Because Leu 67 and Val 84 are conserved between TI and PI, the corresponding peaks in the PI spectra would be expected to represent the side chain protons of the conserved leucine and valine residues at positions 70 and 88 in the active site. Thus, the decrease of these peak intensities would reflect the unfolding of the active site. In the aromatic region, the peaks near 6.5 ppm may be assigned to some of the side chain protons of Tyr 16 and Tyr 57, as referenced to the 1D NMR spectrum of TI (Kuliopulos et al. 1987b). The decrease of these peak intensities might thus also reflect the unfolding process of the active-site structure. The spectrum of PI at intermediate urea concentrations appeared to be a mixture of those obtained at 3 M and 8 M urea. We could not identify any distinct peaks at 5.5 M urea relative to the spectra at 3 M and 8 M urea, supporting the two-state mechanism for PI equilibrium unfolding.

Gel-filtration chromatography of PI at different urea concentrations

Gel-filtration chromatography was performed to analyze the hydrodynamic properties of PI at different urea conditions. Figure 5 ▶ shows the elution profiles of PI after gel-filtration chromatography at different urea concentrations. The unfolded monomer at 7 M urea was eluted at earlier retention time than the native dimer at 0 M urea, probably because of its extended conformation. In contrast to TI, the peaks representing different oligomeric states could not be resolved at the urea concentrations of 4–6 M in the unfolding transition region (Kim et al. 2000b). The resolution was not improved, even when a different gel-filtration column, Superdex 75 HR column (Amersham Pharmacia Biotech), was used or different flow rates were tried. There was no evidence for monomeric intermediates that might be eluted at a later time than the native dimer. When the elution time was analyzed as a function of urea concentration, a transition occurred near 5 M urea (Fig. 5B ▶).

Fig. 5.

Fig. 5.

Open in a new tab

Size-exclusion chromatography analysis of PI. (A) Elution profiles are displayed for PI at different urea concentrations. The protein was incubated in the presence of urea at the indicated concentrations longer than 48 h. The incubated protein was then loaded onto the column equilibrated with the buffer containing 20 mM potassium phosphate, 0.5 mM EDTA, 10 mM β-mercaptoethanol at pH 7, and urea at the respective concentration. The flow rate was 0.4 mL/min and the protein peak was monitored by the absorbance at 280 nm. (B) Dependence of the peak elution time on the urea concentration. The elution time obtained at different urea concentrations was plotted against the urea concentration.

Chemical modification of the interface residue Cys 81

To investigate the stability of the dimeric interaction, we monitored the chemical modification of Cys 81 at the dimeric interface at different urea concentrations. For this experiment, Cys 69 and Cys 97 were replaced by serines to leave Cys 81 alone as a cysteine residue in each monomer. The chemical reagent DTNB, which specifically modifies cysteine residues, will react with Cys 81 if the residue is exposed to solvent. PI-C81 did not react with DTNB, even after a prolonged incubation longer than several hours (data not shown), indicating that Cys 81 is completely buried in the dimeric interface. PI-C81 was incubated at different urea concentrations longer than 48 h, and then DTNB was treated to monitor the exposure of Cys 81. Just after the addition of DTNB, the extent of modification was analyzed by measuring the absorbance change at 412 nm. The absorbance changed little during the time of measurement, suggesting that the DTNB modification might not affect the refolding or assembly processes significantly (data not shown). The absorbance originates from nitrothiobenzoate released on the reaction of the cysteine residue with DTNB. The absorbance was little changed in the presence of 0–2 M urea, whereas a transition occurred between 2 and 5 M urea with a midpoint of about 3.4 M (Fig. 6A ▶). After the transition region, the absorbance value remained at about 0.19 in the 5.5–8 M urea range, which reflects the complete modification of Cys 81.

Fig. 6.

Fig. 6.

Open in a new tab

Chemical modification and equilibrium unfolding analysis of PI-C81. (A) The transition curves are compared between the chemical modification of the dimeric interface Cys 81 (×) and the equilibrium unfolding of PI-C81 (O). PI-C81 at 15 μM was incubated at different urea concentrations longer than 48 h. Just after the addition of DTNB, the absorbance at 412 nm was measured and plotted against urea. The equilibrium unfolding of PI-C81 was also analyzed by monitoring the protein fluorescence intensity at 320 nm after excitation at 285 nm. For comparison, the equilibrium unfolding transition curve for the wild-type PI is also displayed (ν). The Y-axis represents the fraction of modified or unfolded protein in the total protein. (B) Dependence of the equilibrium unfolding transition of PI-C81 on the protein concentration. The unfolding transition was analyzed by the fluorescence measurements at two different protein concentrations of 1 and 15 μM. The data points were fitted to equation 2 to obtain the transition curve.

Because the double mutation replacing Cys 69 and Cys 97 by serines could affect the conformational stability of PI, the equilibrium unfolding was analyzed for PI-C81 by monitoring the fluorescence at 320 nm at different urea concentrations. The mutant protein was unfolded earlier than the wild-type protein with a transition midpoint of ∼4.7 M (Fig. 6A ▶), suggesting that the double mutation moderately affects the conformational stability. However, the unfolding transition occurred at significantly higher urea concentrations than the chemical modification transition curve, suggesting that PI-C81, and probably PI, might expose the interface residues to solvent before the complete unfolding. The equilibrium unfolding transition of PI-C81 occurred at a higher urea concentration as the protein concentration increased, suggesting that PI-C81, like PI, also follows the two-state model (Fig. 6B ▶). When the spectroscopic data were fitted to the two-state model, ΔGUH2O and m values were determined to be 22.5 ± 1.3 kcal•mol−1 and 3.43 ± 0.15 kcal•mol−1•M−1, respectively. Thus, the cysteine mutation may not significantly affect the conformational stability of PI-C81.

Refolding kinetics

Refolding kinetics of PI was studied by analyzing fluorescence or ellipticity change during the refolding process. Unfolded proteins at 7 M urea were diluted to induce refolding at 0.64 M urea. Figure 7 ▶ exhibits the kinetic trace monitored by fluorescence change. The trace started with an abrupt increase of fluorescence and subsequently decreased at moderate rates. The fluorescence intensity was then increased very slowly until 1500 sec to reach equilibrium. This kinetic trace could be best described by five kinetic phases. Among the five kinetic phases, the second phase was dependent on the protein concentration. The representative traces at different protein concentrations are displayed in Figure 8A ▶. The plot of the rate constant against the protein concentration exhibited a linear relationship at the protein concentrations below about 1 μM, giving a second-order rate constant of 8 × 105 M−1•s−1, but it was deviated from linearity at higher concentrations (Fig. 8B ▶). Even though it was inevitable that the data at lower protein concentrations are less sensitive, the obtained values were best fitted to the data than any others. The first phase was a unimolecular process with the relaxation time, τ, of 26 msec. At 5 μM protein, the relative amplitudes of the first and second phases accounted for about 27% and 28%, respectively. This indicates that the tertiary structure forms partly at the first step and the subsequent bimolecular and unimolecular steps lead to the formation of most native tertiary structures. Following the bimolecular step, the fluorescence signal continued to decrease until about 40 sec and then increased slowly until 1500 sec. This slow stage could be best described by three kinetic phases with τ's of 1.7sec, 12.5 sec, and 770 sec, respectively. They accounted for about 6%, 7%, and 32%, respectively, of the total amplitude change at 5 μM protein.

Fig. 7.

Fig. 7.

Open in a new tab

PI refolding kinetics monitored by fluorescence measurement. The refolding was induced by diluting the denatured protein dissolved in 7 M urea to 0.64 M urea. The representative trace obtained for 5 μM PI is displayed. The fluorescence intensity passed through the 305-nm cutoff filter was monitored after the excitation at 285 nm. Five traces were accumulated to obtain the final data. The trace was best fitted by five exponential functions. Their relaxation times were 0.026, 0.26, 1.7, 12.5, and 770 sec, respectively, in the subsequent order of the kinetic phases.

Fig. 8.

Fig. 8.

Open in a new tab

Dependence of the second refolding phase on the protein concentration. (A) Representative traces of PI refolding monitored by fluorescence change are displayed at the final protein concentrations of 0.4, 1, 5, and 15 μM. The Y-axis represents the normalized fluorescence change with taking the fluorescence maximum and minimum as 1 and 0, respectively. (B) Dependence of the rate constant of the second phase on the protein concentration. The second-order rate constant was determined to be ∼8 × 105 M−1sec−1 in the range of low protein concentrations.

Because PI contains eight proline residues per each monomer and one of them, Pro40, forms a trans X-Pro peptide bond, proline isomerization reactions would be expected to make the refolding process complex to analyze. To determine whether any of the refolding phases is affected by proline isomerizations, we investigated the refolding process in the presence of Cyclophilin A, a peptidyl-prolyl isomerase. Among the five kinetic phases of the PI refolding process, the earlier two phases were little affected by Cyclophilin A, but the last two phases were accelerated over the range of 0–3 μM Cyclophilin A (Fig. 9 ▶). The rate constant of the fourth phase increased by about 1.6-fold at 3 μM Cyclophilin A, and it was more highly dependent on the prolyl isomerase than was the fifth phase. The amplitudes of the fourth and fifth phases increased 5%–14% relative to those obtained in the absence of the prolyl isomerase. The rate constant and amplitude of the third phase was marginally decreased by the prolyl isomerase. This implies that prolyl isomerizations might change the population of denatured proteins with different prolyl peptide bonds.

Fig. 9.

Fig. 9.

Open in a new tab

Catalytic effect of Cyclophilin A on the fourth and fifth refolding phases. The refolding trace of PI was obtained by monitoring the fluorescence change in the presence of Cyclophilin A. The PI concentration was 3 μM, and two different concentrations, 1 and 3 μM, of Cyclophilin A were tried. The rate constants for the fourth and fifth phases were plotted against Cyclophilin A concentration. The ratio of k/k0 was calculated from the refolding rates in the presence (k) and the absence (k0) of Cyclophilin A.

When the refolding kinetics was monitored by ellipticity change at 225 nm, the CD signal abruptly changed during the dead time to make it hard to analyze at a very early stage. More than 80% ellipticity change occurred within 80 msec (Fig. 10 ▶). This suggests that α-helical structures may form very fast before the dimerization step. Thus, PI folds as a monomer containing most of the native α-helical structures at a very early stage. It is not clear whether β-strands at the dimeric interface form at the early stage. Because the three α-helices of PI are not involved in the dimeric interaction, the ellipticity change at 225 nm would be expected to reflect these α-helical structures rather than the β-strands at the interface.

Fig. 10.

Fig. 10.

Open in a new tab

PI refolding kinetics monitored by ellipticity change at 225 nm. The ellipticity change was observed until 6000 sec after a manual mixing. The unfolded protein was induced to refold by diluting the urea concentration from 7 to 0.64 M. The protein concentration was 5 μM. The Y-axis represents the mean residue molar ellipticity. The inset represents the kinetic trace in the range of 0–4 sec obtained by use of the Bio-Logic stopped-flow system. Ten traces were accumulated to obtain the final data. The initial ellipticity value (ν) for the unfolded protein is indicated.

Activity of monomer

To investigate the role of dimerization in the enzymatic function of PI, the enzyme activity of PI monomer was estimated indirectly by monitoring the activity simultaneously during the refolding process. Activity measurements were started 2 sec after dilution of the unfolded enzyme into the assay buffer. The slope of the absorbance change at 248 nm gradually increased to reach a maximum value until a time point. The slope did not change any more after this time point, suggesting that the activity reached its maximum. The time required to reach the half of the maximum slope of activity was analyzed against the protein concentration. If the monomer were fully active, the reactivation would follow a first-order process independent of the protein concentration. The half-time was independent of the protein concentration at concentrations above 10 nM, whereas it became dependent on the protein concentration at lower concentrations of the protein (Fig. 11 ▶). This result suggests that the dimeric interaction affects the PI activity, and consequently the PI monomer is not fully active. Interestingly, the half-time was several times lower for PI than for TI, implying that the dimerization step is relatively favorable to occur in PI. This is consistent with the higher second-order rate constant for the bimolecular refolding step of PI, which was determined to be about 15-fold higher than that of TI (Kim et al. 2000b).

Fig. 11.

Fig. 11.

Open in a new tab

Dependence of half-time for the activity recovery from denatured states on the protein concentration. The time required to reach the maximal slope of the reactivation rate was taken as the half-time. The half-time was plotted against the protein concentration (ν). The activity was assayed by using 5-AND as a substrate by monitoring the absorbance change at 248 nm. The refolding buffer was 34 mM potassium phosphate at pH 7.0, 1 mM EDTA, and 1 mM DTT. For comparison, the half-times were also displayed for TI (O) (Kim et al. 2000b).

The apparent rate constant for the bimolecular step could be estimated from the half times, which gave the values of 0.04–0.4 sec−1 in the 1–12 nM range of the protein concentration. These values are considerably higher than the rate constants of the bimolecular step, 10−3 to 10−2 sec−1, estimated from the second-order rate constant at nanomolar protein concentrations. This difference might be due to the influence of urea on the refolding rate. The refolding rate would be higher at the lower urea concentration, 0.033 M, used for the activity measurement, than at 0.64 M, used for the fluorescence measurement. Alternatively, the used substrate concentration, 116 μM, would not be high enough to maintain the uniform reaction condition during the reactivation process. Because the Km value for the used substrate is about 50 μM (Kim et al. 1999), the substrate-saturated enzyme concentration may decrease gradually as the reaction proceeds. The enzyme activity of PI derived from the maximal slope of activity during the refolding process was substantially lower than its known specific activity. This might be because of incomplete folding of PI or the influence of urea on the enzyme activity. Enzyme activity of PI was affected significantly by urea and decreased exponentially as urea concentration increased, suggesting that the active site is less stable than the tertiary structure as a whole (data not shown).

Discussion

The results of equilibrium unfolding studies were apparently consistent with a two-state model in which only native dimers and unfolded monomers can exist at equilibrium. The 1D NMR spectra of PI at different urea concentrations did not show any distinct peaks not found in those of the native and unfolded states (Fig. 3 ▶). The gel-filtration chromatography analysis also suggested no monomeric folded structure at equilibrium (Fig. 5 ▶). These results suggest that the unfolding and dissociation are concerted processes and the dimerization is essential for maintaining the native tertiary interactions. Nevertheless, the interface interaction appeared to be relatively less stable than the active site at Trp 120 and α-helices, as revealed by the chemical modification experiment (Fig. 6 ▶). The exposure of the interface Cys 81 earlier than the structural unfolding might be caused by a partial disruption of the dimeric interaction by urea. This inconsistent result might be explained by increased dynamic motion of the interface interactions in the presence of urea. The chemical modification may be able to occur even in the native dimer if dynamic motions increased enough to expose the cysteine residue more frequently without the dimer dissociation.

The PI refolding process consists of at least five kinetic phases. The fluorescence change during the refolding process may reflect the different kinetic behaviors of Trp 92 and Trp 120. Trp 92 located on the protein surface will give rise to lower fluorescence emission if Trp 120 is located near to it during refolding (Fig. 1 ▶). Trp 120 is located at the hydrophobic active-site pocket and its movement into the active site will raise the fluorescence intensity during refolding. The abrupt increase of fluorescence at the early stage might thus reflect the internalization of Trp 120 into the active-site pocket during the first refolding step. There are also three tyrosine residues in the active site, among which Tyr 16 (Tyr 14 in TI) is known to exhibit high fluorescence emission in TI (Wu et al. 1994). Interestingly, fluorescence intensity increased by about 1.5 fold when Tyr 16 of PI was replaced by phenylalanine (data not shown), implying that Tyr 16 somehow affects the fluorescence emission of Trp 120 to decrease the intensity. Thus, the decrease in fluorescence intensity during refolding can be related to this interaction between Tyr 16 and Trp 120. The solvent accessible areas were around 300 Å for both Trp 92 and Trp 120, suggesting that the nearby amino acids might be more important than the solvent accessibility for the fluorescence change during refolding.

The bimolecular refolding step became less dependent on the protein concentration at higher protein concentrations (Fig. 8 ▶). This implies that the rate-limiting step for this phase changed to a slower monomeric folding step from the monomer–monomer association step at high protein concentrations. Because PI contains eight proline residues and Pro 40 forms a trans-peptide bond, prolyl isomerization reactions could be considered as one of plausible rate-limiting steps. However, the bimolecular step was not affected by Cyclophilin A, suggesting that the rate-limiting step might not be directly related to prolyl isomerizations. The fluorescence and ellipticity changes during the refolding process revealed that the monomeric intermediate might form most α-helical structures and partial tertiary structure before dimerization. Because the dimerization step may require a properly folded monomer, the monomeric intermediate should take a conformation with exposed dimeric interface. Thus, the rate-limiting step may be related to formation of the monomer containing a properly structured dimeric interface.

It is apparent that the equilibrium unfolding transition of PI occurs at higher urea concentrations than that of TI (Kim et al. 2000b). The ΔGUH2O value was about 2 kcal•mol−1 higher for PI than for TI. Although this difference was marginal, it could be related to the stability of dimeric interaction. The dimeric interaction is very similar between these two enzymes, but some differences could be identified in their crystal structures (Kim et al. 1997a; Cho et al. 1998). Relative to TI, PI was found to have more bound water molecules and hydrogen bonds at the dimeric interface. PI also has more residues located within 3.8 Å from the partner monomer than does TI. The stable dimeric interaction might also be related to the preferential monomer–monomer association in PI relative to TI during the refolding process. The folding mechanism of PI is similar to that of TI because the monomeric folding intermediate occurs transiently before the dimerization, but the dimerization rate constant, 8 × 105 M−1sec−1, of PI was determined to be about 15-fold higher than that of TI (Kim et al. 2000b). It would be expected that the monomeric folding intermediate of PI might have a preferential binding moiety on the dimeric interface compared with that of TI.

In conclusion, the analysis of equilibrium states of PI at different urea concentrations revealed that the dimerization is essential for maintaining the native tertiary interactions. A folding intermediate is generated transiently before dimerization during the refolding process. This monomeric folding intermediate may contain most of the α-helical structures and partial native tertiary structures. The dimerization step contributes to substantial conformational changes to form the functional native tertiary structure. We could identify the difference in both conformational stability and refolding kinetics between PI and TI. This difference might be related to the structural features of the dimeric interaction. Detailed analysis of the folding kinetics combined with mutational studies will help elucidate the roles of dimeric interactions in the folding mechanism. These studies will provide ideas on how quaternary interactions can define the kinetics of oligomeric protein folding.

Materials and methods

Reagents and experimental conditions

Ultrapure urea was purchased from Sigma. Urea solution was prepared freshly for every experiment. The Superose 12 gel-filtration column was obtained from Amersham Pharmacia Biotech. Cyclophilin A was obtained from Roche Molecular Biochemicals. 5-AND was purchased from Steraloids. All other chemicals were reagent grade and obtained from Sigma. The temperature was maintained at 25°C for all experiments. Protein concentration was determined by using the difference spectrum change at 295 nm as described previously (Kim et al. 1997b). All of the protein concentrations are reported for monomer unless otherwise specified.

Protein source and purification

The wild-type PI was overexpressed in E. coli strain BL21(DE3) containing the plasmid pKK-KSI and purified to homogeneity according to the methods described previously (Kim et al. 1994). The purity of the protein was confirmed by identifying a single band on SDS-polyacrylamide gels stained with Coomassie blue. Specific activity of the wild-type PI was determined to be approximately 50,000 μmol/min/mg by the assay method described previously (Kim et al., 1994).

Fluorescence and CD spectra

The fluorescence spectra of the native, denatured, and refolded enzymes were obtained by use of a fluorescence spectrophotometer (Shimadzu Model RF-5000). The emission spectrum was monitored in the range from 300 to 420 nm after excitation at 285 nm. The bandwidths were 2 nm for both the excitation and emission wavelengths. The step resolution and the integration time were 1 nm and 1 sec, respectively. The denatured enzyme was obtained by incubating the enzyme in a buffer solution containing 20 mM potassium phosphate at pH 7.0, 0.5 mM EDTA, 1 mM DTT, and 7 M urea. The refolded enzyme was prepared by diluting the denatured enzyme into 0.2 M urea. The final protein concentration was 15 μM. The CD spectra of the native, denatured, and refolded proteins were also obtained by use of a spectropolarimeter (Jasco 715). The proteins were prepared in the same way as described earlier. A cuvette with 0.2-cm path length was used for all of the CD spectral measurements. The CD spectra were obtained with the scan speed of 20 nm/min and the bandwidth of 2 nm. Scans were collected at 0.2-nm intervals with a response time of 0.25 sec and accumulated three times. All of the CD spectra were corrected by subtracting the spectrum of the solution containing the used buffer.

Equilibrium unfolding

Fifteen micromolar PI was preincubated in the phosphate buffer containing 20 mM potassium phosphate at pH 7.0, 0.5 mM EDTA, 1 mM DTT, and urea at 0–8 M for longer than 48 h. The intramolecular fluorescence of KSI was measured by use of a fluorescence spectrophotometer (Shimadzu model RF-5000) equipped with a thermostatically controlled cell holder. The excitation wavelength was 285 nm, and we measured the emission fluorescence at 320 nm and the wavelength at the maximum fluorescence. The CD ellipticity was also obtained for PI at different urea concentrations with a spectropolarimeter (Jasco J-715). A cuvette with 0.2-cm path length was used for all of the CD spectral measurements. The ellipticity at 222 nm was recorded and analyzed.

The changes in the optical properties of the protein were compared by normalizing each transition curve to the apparent fraction of the unfolded form, FU, which was obtained by the following equation:

graphic file with name M1.gif 1

where Y is the observed fluorescence intensity or molar ellipticity at a given urea concentration, and YF and YU are the observed values for the native and unfolded forms, respectively, at the same denaturant concentration. A linear dependence of Y on the denaturant concentration was observed in both the native and unfolded baseline regions for all spectroscopic methods. Linear extrapolations from these baselines were made to obtain estimates of YF and YU in the transition region. To obtain ΔGUH2O and m values, we fitted the data of a urea denaturation curve to equation 2 (Kim et al. 2000b) by a nonlinear least-squares analysis by using the Kaleidagraph version 2.6 program (Abelbeck Software).

graphic file with name M2.gif 2

NMR experiment

The sample of 500 μM PI was prepared in the phosphate buffer containing 90% H2O and 10% D2O (v/v). The pH was readjusted to 7.0 after dissolving urea at respective concentration. NMR spectra were collected on a spectrometer (Bruker DRX500) equipped with a triple resonance, pulse-field gradient probe with actively shielded Z-axis gradients, and a gradient amplifier unit. Signals from urea were suppressed via a low-power presaturation method. WATERGATE sequence (Piotto et al. 1992) was used to suppress the water signal. The observed 1H chemical shifts were determined relative to that of sodium 2,2-dimethyl-2-silapentane-5-sulfonate (DSS), which is reported to be insensitive to changes in temperature. The experiment was performed at 300 K with 40 scans acquired for each spectrum and the relaxation delay was 1 sec. The spectral width of 8012 Hz was used in 16,384 data points. The data were processed on a workstation (Silicon Graphics IndyPC) by using software program XWIN-NMR (Bruker).

Gel-filtration chromatography of PI at different urea concentration

PI (50 μM) was incubated at room temperature longer than 48 h in various concentrations of urea at 0, 3, 3.5, 4, 4.5, 5, 5.5, 6, and 7 M, respectively. The incubated proteins were then loaded onto a Superose 12 gel-filtration column equilibrated previously with a buffer containing 20 mM potassium phosphate at pH 7.0, 0.5 mM EDTA, and 10 mM β-mercaptoethanol, and urea at the respective concentrations. The column was eluted with the buffer at the flow rate of 0.4 mL/min. The absorption peak was monitored at 280 nm with a UV-absorbance detector (Amersham Pharmacia Biotech, Uvicord II).

Site-directed mutagenesis

Double-point mutation (KSI-C81) replacing Cys 69 and Cys 97 by Ser simultaneously was generated by site-directed mutagenesis by using a uracil-containing, single-stranded DNA as a template (Kunkel, 1985); E. coli RZ1032 (ung -dut -) was transformed with the pSK(−) containing the C67S mutant DNA (Kim et al. 1997b), and the single-stranded DNA complementary to the coding strand of the KSI gene was prepared from the transformant and used as a template. The oligonucleotide of 5′-TGG AAC GGC CAG CCC AGC GCA CTG GAT GTC-3′ was synthesized and used as a primer for the mutagenesis, in which the underlined nucleotides represent the changed nucleotides by the point mutation. The entire PI gene in pSK(−) was sequenced by the dideoxynucleotide chain termination method to confirm the mutagenesis without changing other sequence. The mutated gene in pSK(−) was isolated by digestion with EcoRI and HindIII, and then subcloned into pKK223–3 (Amersham Pharmacia Biotech), an expression vector, to express the protein in E. coli.

Modification of Cys 81 by DTNB

The availability of Cys 81 in the dimeric interface between the two monomers to the modification by DTNB was analyzed for PI-C81 incubated at different urea conditions as follows: PI-C81 (15 μM) was incubated in the presence of different urea concentrations longer than 48 h. The protein was then reacted with 500 μM DTNB in 20 mM potassium phosphate at pH 7.0, and 1 mM EDTA. The extent of the reaction was monitored by measuring the absorbance at 412 nm (ɛ = 14,150 M−1•cm−1) with a spectrophotometer (Cary 3E) equipped with a temperature-controlled cell holder (Riddles et al., 1983). The equilibrium unfolding transition of PI-C81 was also analyzed according to the procedure described for the wild-type PI.

Folding kinetics

The folding kinetic experiments were performed by using Bio-Logic SFM-4 stopped-flow modules equipped with Xenon/Mercury lamp supplied from Bio-Logic. Fifty microliters of the protein solution in 7 M urea at pH 7.0 was mixed with 500 μL of the buffer containing 20 mM potassium phosphate at pH 7.0. The dead time of the instrument was 8 msec. As a cuvette, FC20 (Bio-Logic) was used for both fluorescence and CD experiments. During the refolding process, the fluorescence intensity passed through the 305-nm cutoff filter (Oriel) was measured after the excitation at 285 nm. The fluorescence change was recorded with sampling times of 1, 10, and 200 msec depending on the time scale. The ellipticity change was monitored at 225 nm during the refolding process by use of the Bio-Logic stopped-flow instrument equipped with a photoelastic modulator (PEM-90, HINDS Instruments, Inc). The CD signal was also monitored after a manual mixing by use of a spectropolarimeter (Jasco 715). To enhance the signal-to-noise ratio, we averaged multiple shots; 4–10 shots were averaged for the fluorescence signals, and >10 shots were averaged for the CD signals. Different protein concentrations were tried to analyze the dimerization step in the folding process. The refolding kinetics was also analyzed in the presence of Cyclophilin A at 1 and 3 μM to investigate which kinetic phases are related to prolyl isomerizations.

The rate constants were obtained by fitting the data to the following equation:

graphic file with name M3.gif 3

where Ft is the signal at time t, F∞ is the signal of the final state, Fi and Fj are the amplitudes of the kinetic phases, and τi and τj are the relaxation times for the refolding. Data fitting was performed by using the Kaleidagraph program.

Half-time for recovery of full activity from denatured state

Enzyme activity was measured during the refolding process by monitoring the absorbance change at 248 nm. Ten microliters of the unfolded enzyme in 7 M urea was added into 3 mL of the assay buffer containing 34 mM potassium phosphate, 2.5 mM EDTA at pH 7.0, 1 mM DTT, and 116 μM 5-AND. The absorbance measurements were started just after the addition of the unfolded protein. The final urea concentration was 0.0233 M. The half-time was taken as the time at which the slope of the absorbance change reached half of the maximum slope. The half-time was determined at different protein concentrations in the range of 0.5–12 nM.

Acknowledgments

This work was supported by a grant from Korea Science and Engineering Foundation, by the Academic Research Fund of Korea's Ministry of Education and Culture, and in part by the Brain Korea 21 project.

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.

Abbreviations

  • KSI, ketosteroid isomerase

  • PI, KSI of Pseudomonas putida biotype B

  • TI, KSI of Comamonas testosteroni

  • DTT, dithiothreitol

  • pSK(−), pBluescript SK(−)

  • EDTA, ethylenediaminetetraacetic acid

  • CD, circular dichroism

  • NMR, nuclear magnetic resonance

  • KD, dissociation constant

  • ΔGUH2O, Gibbs free energy change for unfolding in the absence of urea and at 25°C

  • DTNB, 5,5′-dithio-bis(2-nitrobenzoate)

  • 5-AND, 5-androstene-3,17-dione

Article and publication are at www.proteinscience.org/cgi/doi/10.1110/ps.18501.

References

  1. Batzold, F.H., Benson, A.M., Covey, D.F., Robinson, C.H., and Talalay, P. 1976. The Δ5–3-ketoteroid isomerase reaction: Catalytic mechanism, specificity and inhibition. Adv. Enzyme Regul. 14 243–267. [DOI] [PubMed] [Google Scholar]
  2. Benson, A.M., Jarabak, J., and Talalay, P. 1971. The amino acid sequence of 5–3-ketosteroid isomerase of Pseudomonas testosteroni. J. Biol. Chem. 246 7514–7525. [PubMed] [Google Scholar]
  3. Bowie, JU and Sauer, R.T. 1989. Equilibrium dissociation and unfolding of the arc repressor dimer. Biochemistry 28 7139–7143. [DOI] [PubMed] [Google Scholar]
  4. Cho, H.-S., Choi, G., Choi, K.Y., and Oh, B.-H. 1998. Crystal structure and enzyme mechanism of ketosteroid isomerase from P. testosteroni. Biochemistry 37 8325–8330. [DOI] [PubMed] [Google Scholar]
  5. Gloss, L.M. and Matthews, C.R. 1998. Mechanism of folding of the dimeric core domain of Escherichia coli Trp repressor: A nearly diffusion-limited reaction leads to the formation of an on-pathway dimeric intermediate. Biochemistry 45 15990–15999. [DOI] [PubMed] [Google Scholar]
  6. Hawkinson, D.C., Eames, T.C., and Pollack, R.M. 1991. Energetics of 3-oxo-Δ5-steroid isomerase: Source of the catalytic power of the enzyme. Biochemistry 30 10849–19858. [DOI] [PubMed] [Google Scholar]
  7. Hawkinson, D.C., Pollack, R.M., and Ambulos Jr., N.P. 1994. Evaluation of the internal equilibrium constant for 3-oxo-Δ5-steroid isomerase using the D38E and D38N mutants: The energetic basis for catalysis. Biochemistry 33 12172–12183. [DOI] [PubMed] [Google Scholar]
  8. Jaenicke, R. 1987. Folding and association of proteins. Prog. Biophys. Mol. Biol. 49 117–237. [DOI] [PubMed] [Google Scholar]
  9. Kim, S.W., Kim, C.Y., Benisek, W.F., and Choi, K.Y. 1994. Cloning, nucleotide sequence, and overexpression of the gene coding for Δ5–3-ketosteroid isomerase from Pseudomonas putida biotype B. J. Bacteriol. 176 6672–6676. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Kim, S.W., Cha, S.-S., Cho, H.-S., Kim, J.-S., Ha, N.-C., Cho, M.-J., Joo, S., Kim, K.-K., Choi, K.Y., and Oh, B.-H. 1997a. High-resolution crystal structures of Δ5–3-ketosteroid isomerase: Active site environment and enzyme mechanism. Biochemistry 36 14030–14036. [DOI] [PubMed] [Google Scholar]
  11. Kim, S.W., Joo, S., Choi, G., Choi, H.-S., Oh, B.-H., and Choi, K.Y. 1997b. Mutational analysis of the three cysteines and active site aspartic acid 103 of ketosteroid isomerase from Pseudomonas putida biotype B. J. Bacteriol. 179 7742–7747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kim, D.-H., Nam, G.H., Jang, D.S., Choi, C., Joo, S., Kim, J.-S., Oh, B.-H., and Choi, K.Y. 1999. Roles of active site aromatic residues in catalysis by ketosteroid isomerase from Pseudomonas putida biotype B. Biochemistry 38 13810–13819. [DOI] [PubMed] [Google Scholar]
  13. Kim, D.-H., Jang, D.S., Nam, G.H., Choi, G., Kim, J.-S., Ha, N.-C., Kim, M.-S., Oh, B.-H., and Choi, K.Y. 2000a. Contribution of the hydrogen-bond network involving a tyrosine triad in the active site to the structure and function of a highly proficient ketosteroid isomerase from Pseudomonas putida biotype B. Biochemistry 39 4581–4589. [DOI] [PubMed] [Google Scholar]
  14. Kim, D.-H., Jang, D.S., Nam, G.H., Yun, S., Cho, J.H., Choi, G., Lee, H.C., and Choi, K.Y. 2000b. Equilibrium and kinetic analysis of folding of ketosteroid isomerase from Comamonas testosteroni. Biochemistry 39 13084–13092. [DOI] [PubMed] [Google Scholar]
  15. Kraulis, P.J. 1991. MOLSCRIPT: A program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24 946–950. [Google Scholar]
  16. Kuliopulos, A., Shortle, D., and Talalay, P. 1987a. Isolation and sequencing of the gene encoding Δ5–3-ketosteroid isomerase of Pseudomonas testosteroni: Overexpression of the protein. Proc. Natl. Acad. Sci. 84 8893–8897 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kuliopulos, A., Westbrook, E.M., Talalay, P., and Mildvan, A.S. 1987b. Positioning of a spin-labeled substrate analogue into the structure of Δ5–3-ketosteroid isomerase by combined kinetic, magnetic resonance, and X-ray diffraction methods. Biochemistry 26 3927–3937. [DOI] [PubMed] [Google Scholar]
  18. Kunkel, TA. 1985. Rapid and efficient site-directed mutagenesis without phenotypic selection. Proc. Natl. Acad. Sci. 82 488–492. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lumb, K.J. and Dobson, C.M. 1992. 1H nuclear magnetic resonance studies of the interaction of urea with hen lysozyme. J. Mol. Biol. 227 9–14. [DOI] [PubMed] [Google Scholar]
  20. Massiah, M.A., Abeygunawardana, C., Gittis, A.G., and Mildvan, A.S. 1998. Solution structure of Δ5–3-ketosteroid isomerase complexed with the steroid 19-nortestosterone hemisuccinate. Biochemistry 37 14701–14712. [DOI] [PubMed] [Google Scholar]
  21. Milla, M.E., Brown, B.M., Waldburger, C.D., and Sauer, R.T. 1995. P22 Arc repressor: Transition state properties inferred from mutational effects on the rates of protein unfolding and refolding. Biochemistry 34 13914–13919. [DOI] [PubMed] [Google Scholar]
  22. Myers, J.K., Pace, C.N., and Scholtz, J.M. 1995. Denaturant m values and heat capacity changes: Relation to changes in accessible surface areas of protein unfolding. Protein Sci. 4 2138–2148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Park, Y.C. and Bedouelle, H. 1998. Dimeric tyrosyl-tRNA synthetase from Bacillus stearothermophilus unfolds through a monomeric intermediate. A quantitative analysis under equilibrium conditions. J. Biol. Chem. 29 18052–18059. [DOI] [PubMed] [Google Scholar]
  24. Piotto, M., Saudek, M., and Sklenar, V. 1992. Gradient-tailored excitation for single-quantum NMR spectroscopy of aqueous solutions. J. Biomol. NMR 2 661–665. [DOI] [PubMed] [Google Scholar]
  25. Pollack, R.M., Zeng, B., Mack, J.P.G., and Eldin, S. 1989. Determination of the microscopic rate constants for the base-catalyzed conjugation of 5-androstene-3,17-dione. J. Am. Chem. Soc. 111 6419–6423. [DOI] [PubMed] [Google Scholar]
  26. Rietveld, A.W.M. and Ferreira, S.T. 1998. Kinetics and energetics of subunit dissociation/unfolding of TIM: The importance of oligomerization for conformational persistence and chemical stability of proteins. Biochemistry 37 933–937. [DOI] [PubMed] [Google Scholar]
  27. Riddles, P.W., Blakeley, R.L., and Zerner, B. 1983. Reassessment of Ellman's reagent. Methods Enzymol. 91 49–60. [DOI] [PubMed] [Google Scholar]
  28. Waldburger, C.D., Jonsson, T., and Sauer, R.T. 1996. Barriers to protein folding: Formation of buried polar interactions is a slow step in acquisition of structure. Proc. Natl. Acad. Sci. 93 2629–2634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Wallace, L.A. and Dirr, H.W. 1999. Folding and assembly of dimeric human glutathione transferase A1–1. Biochemistry 38 16686–16694. [DOI] [PubMed] [Google Scholar]
  30. Wang, S.-F., Kawahara, F.S., and Talalay, P. 1963. The mechanism of the Δ5–3-ketosteroid isomerase reaction: Absorption and fluorescence spectra of enzyme-steroid complexes. J. Biol. Chem. 238 576–585. [PubMed] [Google Scholar]
  31. Wu, P., Li, Y.-K., Talalay, P., and Brand, L. 1994. Characterization of the three tyrosine residues of Δ5–3-ketosteroid isomerase by time-resolved fluorescence and circular dichroism. Biochemistry 33 7415–7422. [DOI] [PubMed] [Google Scholar]
  32. Wu, Z.R., Ebrahimian, S., Zawrotny, M.E., Thornburg, L.D., Perez-Alvarado, G.C., Brothers, P., Pollack, R.M., and Summers, M.F. 1997. Solution structure of 3-oxo-Δ5-steroid isomerase. Science 276 415–418. [DOI] [PubMed] [Google Scholar]
  33. Xue, L., Kuliopulos, A., Mildvan, A.S., and Talalay, P. 1991. Catalytci mechanism of an active-site mutant (D38N) of Δ5–3-ketosteroid isomerase. Direct spectroscopic evidence for dienol intermediates. Biochemistry 30 4991–4997. [DOI] [PubMed] [Google Scholar]

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES