Tài liệu Báo cáo Y học: Biphasic reductive unfolding of ribonuclease A is temperature dependent pdf

Biphasic reductive unfolding of ribonuclease A is temperature dependent Yong-Bin Yan 1,2 , Ri-Qing Zhang 1,2 and Hai-Meng Zhou 3 1 NMR Laboratory, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China; 2 State Key Laboratory of Biomembrane and Membrane Biotechnology, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China; 3 Protein Science Laboratory of the Ministry of Education, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China The kinetics of the reversible thermal unfolding, irreversible thermal unfolding, and reductive unfolding processes of bovine pancreatic ribonuclease A (RNase A) were investi- gated in NaCl/P i solutions. Image parameters including Shannon entropy, Hamming distance, mutual information and correlation coefficient were used in the analysis of the CD and 1D NMR spectra. The irreversible thermal unfolding transition of RNase A was not a cooperative process, pretransitional structure changes occur before the main thermal denaturation. Different dithiothreitol (dithio- threitol red ) concentration dependencies were observed between 303 and 313 K during denaturation induced by a small amount of reductive reagent. The protein selectively follows a major unfolding kinetics pathway with the selec- tivity can be altered by temperature and reductive reagent concentration. Two possible explanations of the selectivity mechanism were discussed. Keywords: image analysis; proton nuclear magnetic reson- ance; reductive unfolding; thermal unfolding; unfolding kinetics. Dynamic analysis of the unfolding and refolding pathways and identification of the specific conformational changes which form the individual intermediates involved in the rate-limited pathway(s) can distinguish one pathway from another and play fundamental importance for protein folding [1,2]. It is usually of considerable interest to estimate the conformational changes of both the whole protein tertiary structure and of specific sites observed by spectro- scopic techniques in different redox systems and solvent conditions. Protein unfolding is highly pertinent to protein folding [3] and is more controllable for more comprehensive study by slowing the unfolding process carried out at physiological pH and temperature [4]. Such studies in turn provide new insights into the functional properties and mechanisms of proteins, which will lead to a more detailed and more complete description of biological functions [5]. Bovine pancreatic ribonuclease A (RNase A; EC 3.1.27.5) contains 124 residues with four native disulfide bonds (Cys26–Cys84, Cys40–Cys95, Cys58–Cys110, and Cys65–Cys72). RNase A has played a crucial role as a model system in studies of protein structure, folding, stability, and chemistry [6]. It folds and unfolds through multiple pathways, with the rate-limiting steps in the well- populated pathways involving the formation of distinct transition intermediates [1–3,7,8]. The complexity of the multiple pathways means that different mechanisms may occur with different types of redox systems and different solvent conditions [9,10]. Thus a comprehensive investiga- tion under different conditions using different methods is essential to elucidate the protein folding and unfolding processes. Much effort has been devoted in recent years to understanding the mechanism and the main factors that control protein folding, and to developing approaches that allow researchers to investigate the multifarious aspects of protein folding and unfolding. While the determination of the protein structural transitions that occur in the pathways is at the heart of studies on unfolding and refolding processes [10], dynamic analysis of the different processes is necessary to evaluate the different pathways. These studies will also clarify the key factor(s) controlling the processes and clarify the unfolding and refolding mechanism associ- ated with the redox properties and solvent conditions. Therefore, we have used 1 H NMR spectra to investigate the unfolding dynamics of RNase A during denaturation by different concentrations of reductive reagent dithiothreitol (dithiothreitol red ) at different temperatures. Similar studies were carried out by Rothwarf and Scheraga [10] in which the temperature dependence of RNase A regeneration was studied with dithiothreitol ox /dithiothreitol red and GSSG/ GSH systems. Their results suggested that the regeneration process with the two types of redox reagents proceeded through different pathways with significantly different temperature dependencies. Here we present a temperature Correspondence to Y B. Yan, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing 100084, China. E-mail: ybyan@mail.tsinghua.edu.cn or H M. Zhou, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing 100084, China. E-mail: zhm-dbs@mail.tinghua.edu.cn Abbreviations: C, Correlation coefficient; des-[40–95], RNase A lacking the 40–95 disulfide bond; des-[65–72], RNase A lacking the 65–72 disulfide bond; DSS, 2,2-dimethyl-2-sila-pentanesulfonate; FID, free induction decay; GSH, reduced glutathione; GSSG, oxidized glutathione; H, Shannon entropy; HD, Hamming distance; MI, Shannon mutual information; RNase A, Bovine pancreatic ribonuclease A. (Received 23 June 2002, revised 3 September 2002, accepted 11 September 2002) Eur. J. Biochem. 269, 5314–5322 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03251.x dependence study based on the same reductive reagent (dithiothreitol red ) but different reagent concentrations. In addition to discussing the temperature effect, the thermal denaturation transition of RNase A and the effect of the reductive reagent concentration, which have not been previously associated with protein folding and unfolding, will also be discussed. A new approach of image analysis, which is established by us recently [4,11], was used to analyze the unfolding dynamics and also was introduced into the analysis of the thermal transition study by CD and 1 HNMR. MATERIALS AND METHODS Sample preparation Highly purified lyophilized ribonuclease A (type XII-A, lot 110K7665) from bovine pancreas (RNase A) was pur- chased from Sigma Chemical Co. (St. Louis, MO, USA) and used without further purification. Dithiothreitol red was also a product of Sigma Chemical Co. All other reagents were of the highest grade commercially available. Samples of the native RNase A were dissolved in 0.5 mL 100 m M NaCl/P i or triple-distilled water, pH 8.0. The protein concentration was 14 mgÆmL )1 (for NMR measurements) or 0.3 mgÆmL )1 (for CD measurements). The NMR samples contained 10% D 2 O to provide a signal for the lock and the final pH was 7.6 which was measured with no corrections for isotope effects. NMR Spectroscopy All 1 H NMR experiments were carried out on a Bruker AM 500 superconductor spectrometer (Bruker, Fa ¨ llanden, Switzerland) at Tsinghua University. The carrier frequency wasfixedatthecenteroftheH 2 O/HOD resonance frequency. The chemical shifts of the spectra were referenced to the most upfield resonance which in turn had been calibrated against 2,2-dimethyl-2-sila-pentanesulfonate (DSS). The 90° pulse width was 6.5 ls, the sweep width was 8333 Hz and each FID had 400 scans or 240 scans with 16 K data points. Two dummy scans were performed for each FID with a recycle delay of 1.8 s or 1 s. Solvent suppression was carried out by presaturation at all times except during acquisition. For a given sample, the unfold- ing experiments were carried out over a contiguous block of time (about 10–72 h) without removing the sample from the spectrometer. The Fourier transform of the FID signal was obtained without additional modification. The phase correction parameters were the same as for the first spectrum. The irreversible thermal transition measurements was carried out by increasing the temperature in one degree steps from 303 to 333 K. Approximately 5 min was allowed for thermal equilibration at each measured temperature. The spectra were collected at 30 min intervals (including thermal equilibration), with 400 scans per spectrum to get a better signal to noise ratio. The experimental temperature for the reductive unfolding was maintained at 303, 308 and 313 K. The protein stability was examined by maintaining the sample without dithio- threitol red at 313 K for 48 h and no difference was found between the original and end spectra (data not shown). RNase A was unfolded with a dithiothreitol red concentra- tion of 10–100 m M (10–100-fold molar excess of the protein, see below). Spectra were collected every 20 min or 8 min after dithiothreitol red was added. Zero corresponds to the dithiothreitol red injection time into the cold NMR tube. To allow for temperature equilibrium and operational delay, the first spectra were obtained at 10 min. RNase A was also reduced with a 100-fold molar excess of dithiothreitol red in NaCl/P i at 313 K for comparison of the reductive denatur- ation endpoint (a complete unfolding was obtained in 4 h with 100-fold molar excess of dithiothreitol red )[1].No aggregation was observed along the experiments. The data analysis was carried out by spectral image analysis method established recently [4,11]. The spectrum parameters (Shannon entropy, H; Mutual Shannon Infor- mation, MI and Correlation coefficient, C) that describe the nature of each image (spectrum) and the correlation between different images were calculated using MATLAB software ( MATLAB 5.2, The MathWorks, Inc., Natick, MA, USA) by programs developed in-house. The spectral window for the reductive unfolding analysis was 6.3– 10.0 p.p.m. The thermal denaturation curves were repre- sented directly by these parameters, while the Hamming distance was calculated instead of the mutual information. The reductive unfolding kinetics was analyzed by a linear expansion least-squares algorithm and graphically using a least mean squares fit procedure. The rate constant errors were defined as the standard deviation. Thermal denaturation monitored by CD spectropolarimetry CD measurements were performed on a Jasco J-715 spectropolarimeter equipped with a thermoelectrically con- trolled cell holder. CD spectra of RNase A in NaCl/P i and in triple-distilled water were measured in the far-UV range (195–245 nm) in 2 mm pathlength quartz cells. For the thermal denaturation, the measurements were typically made at two degree intervals in the temperature range from 293 to 353 K, at a heating rate of 1 KÆmin )1 . Spectra were scanned twice at a rate of 100 nmÆmin )1 , a resolution of 0.5 nm, and a bandwidth of 1 nm. The thermal denaturation curves were obtained by measuring the ellipticity at 222 nm and analyzed with a nonlinear least-squares algorithm. The image parameters (H, HD, C) were also calculated for the image analysis of the CD spectra. The calculating routines were the same as those for the NMR spectra. Curve fits was obtained by the standard method using the Marqurdt-Levenberg routine [12] as provided in the ORIGIN 6.0 software (Microcal Inc., Northampton, MA, USA). RESULTS The thermal stability of proteins, especially enzymes, has long been a practical concern, because this is usually the factor that most limits their usefulness. Moreover, tempera- ture and the relevant free energy change DG, usually influence the behavior of protein folding and unfolding. Both the protein folding and unfolding processes are accelerated by increased temperature. Protein conformation ensembles are undoubtedly affected by temperature, but can temperature affect the folding and unfolding mechanism of Ó FEBS 2002 Temperature dependence of protein unfolding kinetics (Eur. J. Biochem. 269) 5315 proteins due to the different conformation ensembles at different temperatures. This research investigates the rela- tionship between the unfolding mechanism and the con- formational change at different temperatures. Thermal denaturation measured by far-UV CD The changes in the secondary structure of RNase A as a function of temperature were followed by CD measure- ments. Figure 1A shows the reversible thermal transition curves of RNase A in triple-distilled water and in NaCl/P i measured by far-UV CD spectra at 222 nm. To compare the results from NMR spectra, the spectral image param- eters, Shannon entropy (H), correlation coefficient (C)anda new parameter, Hamming distance (HD), were introduced into the CD spectra analysis (mutual information was not presented for large errors). Figure 1B,D show the thermal transition curves described by the Shannon entropy and the correlation coefficient. HD is defined as HD ¼ 1 N X abs h ki À h kj ÀÁ ð1Þ where h k means the ellipticity at k nm, N means the number of data points used in the calculation, i and j are the CD spectra at temperatures i and j. Here, we used i ¼ 1to calculate the relative ellipticity change referenced to the first CD spectra. The relative ellipticity change rate can be calculated by j ¼ i + 1 (data not shown). It can be seen in Fig. 1C that the Hamming distance exactly reflects the ellipticity change but with minimal errors relative to the traditional method. All these parameters present a well- defined two-state process for RNase A thermal denatura- tion, which is consistent with the Ômultiprobe principleÕ [13]. The thermodynamic properties obtained from the curves in Fig. 1 and given in Table 1 are quite consist with previous studies [14]. The data clearly indicates that RNase A in NaCl/P i andinH 2 O have a similar transition properties but those in NaCl/P i are more stable. A small amount (5–10%) of irreversible denaturation was observed reproducibly in both solutions. Irreversible thermal denaturation measured by NMR The extent of irreversible denaturation of proteins has been determined previously to depend on the protein concentra- tion and the period of time that the protein solution is incubated at the higher temperatures during the heating and cooling process [15]. Therefore, the denaturation and renaturation times were usually chosen which minimize the thermal equilibration at each temperature. However, a temperature appearing in the predenaturation part of a thermodynamic curve does not mean stabilization for relatively long-term investigations at this temperature. Thus, a thermal denaturation study of a high concentration protein used in reductive unfolding studies will be useful to determine the initial conformational ensembles of the investigation. Here we present a study of irreversible unfolding using image analysis of the NMR spectra to Fig. 1. Thermal denaturation profiles of RNase A in H 2 O(s) and in NaCl/P i (h) measured by far-UV CD spectra presented by (A) ellipticity at 222 nm; (B) Shannon entropy; (C) Hamming distance; (D) correlation coefficient. Table 1. Thermodynamic properties of RNase A. The errors were calculated as the standard deviations of the curve fits. T m (K) DH 0 (T m ) (kJÆmol )1 ) DS 0 (T m ) (eu) CD in H 2 O h 222 337.9 ± 0.7 419 ± 33 1.24 ± 0.09 HD 337.7 ± 0.2 364 ± 11 1.08 ± 0.03 H 338.6 ± 0.9 389 ± 48 1.15 ± 0.15 C 342.2 ± 0.4 421 ± 18 1.23 ± 0.05 CD in NaCl/P i h 222 340.1 ± 0.6 426 ± 37 1.25 ± 0.1 HD 340.6 ± 0.2 388 ± 12 1.14 ± 0.03 H 341.0 ± 0.9 535 ± 76 1.5 ± 0.2 C 345.6 ± 0.8 466 ± 17 1.35 ± 0.05 NMR in NaCl/P i CH3(Val63) 318.84 ± 0.09 HD 321 H (upfield) 321.5 MI 320 C 320 5316 Y B. Yan et al. (Eur. J. Biochem. 269) Ó FEBS 2002 investigate the relationship between the protein initial conformational ensembles and the reductive unfolding mechanism for different temperatures. The dynamics of the thermal transition can be analyzed by observing specific resonance’s split phenomena, chemical shift changes and the integral areas. The resonance at 0.4 p.p.m. was resolved as CH 3 of Val63. Its thermody- namic curve is shown in Fig. 2 as a typical two-state process. If the protein shows a cooperative two-state transition during the thermal unfolding process, the thermodynamic curves obtained by different parameters should give similar results. The image parameters (H, MI, C, HD)calculatedto characterize the 1 H NMR spectra of the thermal unfolding process of RNase A are presented in Fig. 3. As described previously [4, 11], Shannon entropy values reflect the ÔentropyÕ of the spectra, which is the sum probability of the peak intensity in a certain region. In other words, the H-values reflect the dispersion properties of the resonance in selected regions. Mutual Shannon information reflects the correlation of the Shannon entropy between the corres- ponding data of two spectra. The MI value increases as the degree of correlation increases. MI reflects how many unchangeable parts occur in the spectra. The correlation coefficient reflects the correlation between two spectra using their covariance. A larger value of C indicates more correlation. In Fig. 3, the significant increase of the values above 323 K in the downfield region was due to the protein aggregation and bad resolution. Three of these parameters have similar curves between the downfield and upfield region if the aggregation parts are ignored. The HD values have nearly equal curves, while the Shannon entropy values have rather different curves between the downfield and upfield regions which may be due to the H-values of the downfield region is more affected by the change of microenvironment. No clear explanation can be given now for the differences obtained from H. One explanation may be that H cannot exactly reflect the main changes during thermal denaturation. A second explanation may be that there are some significant conformational differences that were only be observed by H. However, there were some conformational differences between 303 and 313 K, and the thermal transition curves characterized by C and HD were flat between 310 and 316 K, which suggests a relative thermal stabilization. The transition between 310 and 323 K characterized by MI, C,andHD shows a similar two state transition to Fig. 3. The T m values obtained from the MI, C and HD curves are presented in Table 1. Though the thermal transition is irreversible when the protein is heated to 323 K, it is nearly reversible (% 95%) in the temperature range of 301–319 K (data not shown). Reductive denaturation The reductive denaturation of 1 m M RNase A induced by 10–100 m M dithiothreitol red at 303, 308 and 313 K was investigated by image analysis of the 1D NMR spectra as Fig. 2. Integral area of CH3(Val63) during irreversible thermal dena- turation of RNase A in 0.5 mL 100 m M NaCl/P i buffer. Fig. 3. Time course of Shannon entropy (A), mutual information (B), correlation coefficient (C), and Hamming distance (D) values of 1 H NMR spectra during irreversible thermal denaturation of RNase A. (s) Upfield region and (h) downfield region. Ó FEBS 2002 Temperature dependence of protein unfolding kinetics (Eur. J. Biochem. 269) 5317 performed in a previous paper [4]. As relatively low dithiothreitol red concentrations (only 10–100-fold molar excess of the protein) and relatively high temperatures were used, no intermediates peaks were observed from the 1D NMR spectra or 2D NOESY NMR spectra (data not shown). The image parameters (Shannon entropy, mutual information, and correlation coefficient) were calculated to characterize the unfolding process of RNase A induced by dithiothreitol red . The Hamming distance is not presented because the curve is similar to those of the other parameters. All the results were fit best by a model of two consecutive first-order reactions, with rate constants listed in Table 2. A comparison of the rate constants from MI for different dithiothreitol red concentrations and different temperatures is presented in Figs 4 and 5 (Similar curves can be obtained from other parameters). As can be seen in Fig. 4A,B, the rate constants increased while increasing temperature, as described in a previous paper [4]. The rate constants of the fast process (k f ) have similar increases with temperature at the two dithiothrei- tol red concentrations. It should be noted that the rate constants of the slow process (k s ) increased dramatically when the temperature was increased from 303 to 313 K for a dithiothreitol red concentration of 30 m M , while no such large increase occurred for a dithiothreitol red concentration of 50 m M . Figure 5A,B present the relationships between the rate constants and the dithiothreitol red concentration at different temperatures. The dithiothreitol red concentration affected the fast and slow process unfolding rate constants in a complex way. The k f values increased similarly when the dithiothreitol red concentration was increased from 30 m M to 50 m M at different temperatures. As the dithiothreitol red concentration was increased from 30 m M to 50 m M ,thek s values increased at 303 K, but decreased dramatically at 308 and 313 K. DISCUSSION It has been proposed that the regeneration of RNase A follows different pathways in the GSSG/GSH and dithio- threitol ox /dithiothreitol red systems [9]. Li et al. also found that the reductive unfolding of RNase A took a parallel biphasic pathway for unfolding induced by dithiothreitol red [3]. Recent studies demonstrated that the oxidative folding of RNase A has a multiple pathway mechanism [7,8,16–18]. Although the oxidative regeneration and the reductive cleavage of the disulfide bond were suggested to be kinetically the same [3], des-[65–72] RNase A is thought to be the most highly populated intermediate in the unfolding pathways [19]. However, the unfolding and Table 2. Rate constants for changes in Shannon entropy, mutual information and correlation coefficient observed in the reductive denaturation of 1 m M RNase A in NaCl/P i , at 313 K,308 K and 303 K. Temperature (K) [dithiothreitol red ] (m M ) k H (10 4 min )1 ) a k MI (10 4 min )1 ) a k C (10 4 min )1 ) a k Hf k Hs k MIf k MIs k Cf k Cs 313 50 618 b 4.4 731 8.3 436 7.8 313 30 370 6.8 435 18.2 429 12.2 313 20 335 5.6 279 11.1 211 8.9 313 10 134 1.4 143 4.8 106 4.8 308 100 140 1.5 145 4.0 146 3.8 308 50 149 1.6 183 3.7 118 5.1 308 30 57 0.9 61 5.9 57 4.3 303 100 55 3.0 62 6.9 70 3.0 303 50 56 1.4 52 3.4 40 3.4 303 30 13 0.14 15 0.4 10.8 0.12 a Rate constants of Shannon entropy, mutual information and correlation coefficient, respectively. b The errors were of the order of 20–30% determined by repeating experiments. Fig. 4. Comparison of the fast and slow process rate constants, k f (A) and k s (B), from mutual information as a function of temperature reduced by 30 m M and 50 m M dithiothreitol red . 5318 Y B. Yan et al. (Eur. J. Biochem. 269) Ó FEBS 2002 folding processes, especially at relatively high temperatures [20], of the protein are usually dominated by a major pathway related to a major rate-determining intermediate in similar redox systems. For a single pathway, the biphasic process can be expressed as N À! k 1 I À! k 2 U ð2Þ where N, I, and U mean the native state, transition intermediate, and unfolding state, respectively. The observed image parameter rate constants, k f and k s ,can be expressed by a 1 k 1 [N][dithiothreitol red ], a 2 k 2 [I][dithiothre- itol red ], where a 1 and a 2 are defined as the link constants between the image parameter rates and the actual rates. If the unfolding process of RNase A was dominated by a major pathway, the observed rate constants should increase as the dithiothreitol red concentration increased. Thus the results shown in Fig. 5 cannot be explained by a single or major pathway mechanism although similar redox systems were used. The dithiothreitol red concentration and the temperature have a relatively simple effect on the rate constants of the fast process (Figs 4A and 5A), but have a complex effect on the rate constants of the slow process (Figs 4B and 5B). As there are multiple unfolding pathways with multiple three disulfide intermediates [7,8,16–18], these intermediates seem have different sensitivities to dithiothre- itol red and thus may exhibit different kinetics from NMR spectra. Therefore, a change in dithiothreitol red concentra- tion will change the relative population of these intermedi- ates, which will be reflected by the image analysis of the NMR spectra. These results suggest that the external solvent conditions affect the properties of the unfolding rate-limiting intermediates. Thus different unfolding inter- mediates may dominate under different solvent conditions. A possible mechanism for the results in Figs 4 and 5 is shown in Fig. 6. The protein presents a selective behavior of unfolding process under different conditions. The selectivity of the major unfolding pathway may be altered by the temperature and the dithiothreitol red concentration, though the mechanism for this selectivity is not clear now. The effects of pH and phosphate concentration were also studied, and it was found that ions only affect the rate of protein unfolding but have no effect on the pathways (data not shown). N 313 and N 303 mean the different conformational ensembles at 313 and 303 K (Fig. 6). The free energy change for folding at any temperature DG(T), can be obtained using the modified Gibbs-Helmholtz equation [21], DGðTÞ¼DH m ð1 À T=T m Þ þ DC p ½ðT À T m ÞÀT lnðT=T m Þ ð3Þ Here DC p ¼ 1.15 (0.08 kcalÆK )1 Æmol )1 as recently obtained by Pace et al. [22]. Using the parameters listed in Table 1, the DG(T) change between 313 and 303 K was calculated to be 6.6 ± 0.4 kJÆmol )1 (1.6 ± 0.1 kcalÆmol )1 ). Though there is no other evidence that the conformation of RNase A at 313 K is stabilized by a thermal transition state, the different features of the NMR spectra at 303 and 320 K suggest different molecular chemical environments. Such microenvironmental changes, like the effects induced by ions [13,20,22–24], may result in changes of the protein behavior. It has been proposed that the nonzero surface exposure of Cys40 and Cys65 causes the 40–95 and 65–72 disulfides in RNase A to be the first to break [3]. Thus one possible explanation of the mechanism in Fig. 6 might be that the conformational ensembles at 313 K, a relatively high energy state, may result in little difference between the cystine (or the amino-acid residues around cystine) surface exposures. This may also explain the complexity of the rate constant changes between 15 and 25 °C observed by Li Fig. 6. Possible reductive unfolding pathways of RNase A at 303 and 313 K in NaCl/P i ,pH8.0.N 313 and N 303 represent the different con- formational assemblies at 303 and 313 K. I 1 and I 2 represent two dis- tinct unfolding rate-determining intermediates, while U represents the unfolded state of RNase A. Fig. 5. Comparison of the fast and slow process rate constants, k f (A) and k s (B), from mutual information as a function of dithiothreitol red concentration reduced under three temperatures. Ó FEBS 2002 Temperature dependence of protein unfolding kinetics (Eur. J. Biochem. 269) 5319 et al. [3]. The major parts of intermediates I 1 and I 2 are likely to be the three-disulfide-intact species, though they are not characterized here. Thus the selectivity of the breakage of the first disulfide may be altered by the different cystine surface exposures induced by temperature and dithiothre- itol red concentration. Another possible explanation of the selectivity mechan- ism dependence on temperature and reductive reagent concentration in Fig. 6 may be obtained from the studies of the folding intermediate properties. The solution structure and the thermodynamics of the analogs of the major and minor rate-determining three-disulfide folding intermediates have been characterized in recent years [25–27]. The thermodynamic parameters for these analogs are summar- ized in Table 3. All the T m values of these analogs are lower than 313 K (% 40 °C) at pH 4.6. The major intermediate, des-[65–72] RNase A, was more stable than the minor intermediate, [C40A, C95A] RNase A. The thermal unfolding of des-[65–72] RNase A began at about 303 K and ended at about 318 K [25], while thermal unfolding of [C40A, C95A] RNase A began at about 298 K and ended at about 313 K at pH 4.6. Considering the pH effect on T m , the T m values are expected to well above 313 K at neutral pH [28]. The dramatic increase of k s from 303 to 313 K (Fig. 4B) may result from such intermediate thermody- namic properties. The close-to-unfolding state at 313 K of the three-disulfide intermediates may drive the slow process to the unfolded state. The selectivity mechanism may therefore arise from the different thermodynamic properties between the two intermediates. At 313 K, des-[40–95] RNase A was closer to the midpoint of the thermal unfolding transition than des-[65–72] RNase A. Such a difference accompanied by the increased concentration of the reductive reagent may lead the (major) unfolding process along different pathways. The thermodynamic properties may also resulted in the disulfide reshuffling at such a relatively high temperature of 313 K, though few populated intermediates were observed during the unfolding process (see Results). Recently Sogbein et al. [29] reported that pH has a pronounced influence on the kinetic mechanism of myo- globin unfolding. Their results show that myoglobin unfolds through a short-lived intermediates only at acidic pH, with no intermediates observed for basic conditions. Our results suggest that the protein kinetic unfolding mechanism might depend on external solvent conditions such as temperature. Sogbein et al. hypothesized that the observed pH depend- ence of the protein unfolding mechanism could be related to the pH dependence of heme solubility [29]. From the temperature and reductant concentration dependence of the protein unfolding mechanism presented in this paper, we can further hypothesize that the observed difference in the protein unfolding behavior could be related to different initial conformational ensembles for different external solvent conditions. Furthermore, the protein unfolding behavior could be related to different conformational stabilities which could be demonstrated by different free energy change for different external conditions. It should be noted that such the free energy change is not distributed equally among every amino-acid residue of a well-folded protein. Thermal denaturation studies can help elucidate the structural stabilities of proteins. Traditionally such studies used a simple two-state mechanism of thermally induced transitions in small, compact globular proteins, which are thought to act as single stage systems. Many groups have identified the multiple steps involved in thermal unfolding pathways [30,31]. A theoretical model was presented recently to study the stepwise thermal unfolding of globular proteins using the stabilizing or destabilizing characteristics of amino-acid residues in protein crystals [32]. All these results show that in RNase A, the a helix of the shell residues around 16–22 unfolds in the temperature range 303–318 K, while the b sheet segment 106–118 is relatively stable thermally. Our experiments also confirmed that structural changes occur in RNase A before the main thermal denaturation transition (Fig. 2). The pretransition structural changes occurring in a temperature range of Table 3. Thermodynamic properties of the analogs of the rate-determining three-disulfide folding intermediates. pH T m (°C) DH 0 (T m ) (kcalÆmol )1 ) DS 0 (T m ) (eu) C[65–72]S RNase A a 6.4 44 [C65S, C72S] RNase A b 4.6 38.5 ± 0.2 83.7 ± 2 269 ± 8 des-[65–72] RNase A c 4.6 38.4 ± 0.4 80.3 ± 8 258 ± 24 C[40–95]S RNase A a 6.4 42 [C40A, C95A] RNase A d 4.6 33.7 ± 0.2 72.9 ± 3.7 238 ± 12 [C40S, C95S] RNase A d 4.6 33.6 ± 0.2 71.5 ± 1.8 231 ± 6 a Data from Laity et al. [28]. b Data from Shimotakahara et al. [26]. c Data from Talluri et al. [25]. d Data from Laity et al. [27]. Fig. 7. Effect of external solvent conditions on the kinetic mechanism of protein unfolding. N T(a) and N T(b) represent the initial conformational assemblies under different solvent conditions. I 1 and I 2 represent two distinct unfolding intermediates or different energy states of one intermediate. U T(a) and U T(b) represent the different unfolding assem- bles under different solvent conditions. #1 and #2 represent the fast and slow transition states. 5320 Y B. Yan et al. (Eur. J. Biochem. 269) Ó FEBS 2002 303–318 K may result in significant structural changes of RNase A between 303 and 313 K though the tertiary structure is still maintained as was confirmed by CD measurements. The influence of different initial conforma- tional ensembles on the selectivity of the protein unfolding kinetics is shown in Fig. 7. The different pathways show how the external solvent conditions such as temperature affects the kinetic mechanism of protein unfolding. If the external effect is significant, such as the pH effect on myoglobin unfolding, the protein selectively follows a different unfolding kinetics pathway. If the external effect is very small, the protein follows a simpler pathway (where I 1 and I 2 are the same rate-limiting intermediates). In conclusion, the irreversible thermal unfolding trans- ition of RNase A is not a cooperative process, pretransi- tional structure changes occur before the main thermal denaturation. The different initial conformational ensem- bles at 303 and 313 K may lead to the different dependen- cies on the reductive reagent concentration of the biphasic pathway reductive denaturation. The protein selects a preferred one from several major pathways with the selectivity altered by temperature and reductive reagent concentration. The two possible explanations of the selec- tivity mechanism described here need to be clarified by more detailed investigations. ACKNOWLEDGEMENTS This investigation was supported by the National Key Basic Research Special Funds, P. R. China, No. G1999075607, the National Key Science and Technology Item, P. R. China, no. 96-900-09-03, the 985 Funds of Tsinghua University, P. R. China and THSJZ of Tsinghua University, P. R. China. The authors also thank Dr Bo Jiang, Dr Sen Li, Mrs Xue-Chun Luo, Mrs Xiao-Lan Ding at Tsinghua University, P. R. China, and Dr Guang-Zhong Tu at the Institute of Microchem- istry, P. R. China for expert technical assistance. REFERENCES 1. Rothwarf, D.M. & Scheraga, H.A. (1993) Regeneration of bovine pancreatic ribonuclease A. 1. Steady-state distributon. Biochem- istry 32, 2671–2679. 2. Rothwarf, D.M. & Scheraga, H.A. (1993) Regeneration of bovine pancreatic ribonuclease A. 2. Kinetics of regeneration. Biochem- istry 32, 2680–2689. 3. Li, Y.J., Rothwarf, D.M. & Scheraga, H.A. (1995) Mechanism of reductive protein unfolding. Nat. Struct. Biol. 2, 489–494. 4. Yan, Y B., Jiang, B., Zhang, R Q. & Zhou, H M. (2001) Two-phase unfolding pathway of ribonuclease A during dena- turation induced by dithiothreitol. Protein Sci. 10, 321–328. 5. Alexandrescu, A.T., Rathgeb-Szbo, K., Rumpel, K., Jahnke, W., Schulthess,T.&Kammerer,R.(1998) 15 N backbone dynamics of the S-peptide from ribonuclease A in its free and S-protein bound forms: Toward a site-specific analysis of entropy changes upon folding. Protein Sci. 7, 389–402. 6. Raines, R.T. (1998) Ribonuclease A. Chem. Rev. 98, 1045–1065. 7. Rothwarf, D.M., Li, Y J. & Scheraga, H.A. (1998) Regeneration of bovine pancreatic ribonuclease A: identification of two native like three-disulfide intermediates involved in separate pathway. Biochemistry 37, 3760–3766. 8. Rothwarf, D.M., Li, Y J. & Scheraga, H.A. (1998) Regener- ation of bovine pancreatic ribonuclease A: detailed kinetic analysis of two independent folding pathways. Biochemistry 37, 3767–3776. 9. Rothwarf, D.M. & Scheraga, H.A. (1993) Regeneration of bovine pancreatic ribonuclease A. 3. Dependence on the nature of the redox reagent. Biochemistry 32, 2690–2697. 10. Rothwarf, D.M. & Scheraga, H.A. (1993) Regeneration of bovine pancreatic ribonuclease A. 4. Temperature dependence of the regeneration rate. Biochemistry 32, 2698–2703. 11. Yan,Y B.,Luo,X C.,Zhou,H M.&Zhang,R Q.(2002)Two- state kinetics characterized by image analysis of nuclear magnetic resonance spectra. Chinese Sci. Bull. 47, 389–393. 12. Jackson, S.E. & Fersht, A.R. (1991) Folding of chymotrypsin inhibitor 2. 1. Evidence for a two-state transition. Biochemistry 30, 10428–10435. 13. Poklar, N., Petrovc ˇ ic ˇ , N., Oblak, M. & Vesnaver, G. (1999) Thermodynamic stability of ribonuclease A in alkylurea solutions and preferential solvation changes accompanying its thermal denaturation: a calorimetric and spectroscopic study. Protein Sci. 8, 832–840. 14. Reinsta ¨ dler, D., Fabian, H., Backmann, J. & Naumann, D. (1996) Refolding of thermally and urea-denatured ribonuclease A mon- itored by time-resolved FTIR spectroscopy. Biochemistry 35, 15822–15830. 15. Pace, C.N., Shirley, B.A. & Thomson, J.A. (1989) Measuring the conformational stability of a protein. In Protein Structure, a Practical Approach (Creighton, T.E., ed.), pp. 311–330. IRL Press, New York. 16. Iwaoka, M., Juminaga, D. & Scheraga, H.A. (1998) Regeneration of three-disulfide mutants of bovine pancreatic ribonuclease A missing the 65–72 disulfide bond: characterization of a minor folding pathway of ribonuclease A and kinetic roles of Cys65 and Cys72. Biochemistry 37, 4490–4501. 17. Xu, X.B. & Scheraga, H.A. (1998) Kinetic folding pathway of a 3-disulfide mutant of bovine pancreatic ribonuclease A missing the (40–95)-disulfide bond. Biochemistry 37, 7561–7571. 18. Welker, E., Narayan, M., Volles, M.J. & Scheraga, H.A. (1999) Two new structured intermediates in the oxidative folding of RNase A. FEBS Lett. 460, 477–479. 19. Yamamoto, K., Mizutani, Y. & Kitagawa, T. (2000) Nanosecond temperature jump and time-resolved Raman study of thermal unfolding of ribonuclease A. Biophys. J. 79, 485–489. 20. Pace, C.N., Hebert, E.J., Shaw, K.L., Schell, D., Both, V., Krajcikova, D., Sevcik, J., Wilson, K.S., Dauter, Z., Hartley, R.W. & Grimsley, G.R. (1998) Conformational stability and thermodynamics of folding of ribonucleases Sa, Sa2, Sa3. J. Mol. Biol. 279, 271–286. 21. Pace, C.N., Grimsley, G.R., Thomas, S.T. & Makhatadze, G.I. (1999) Heat capacity change for ribonuclease A folding. Protein Sci. 8, 1500–1504. 22. Collins, K.D. (1997) Charge density-dependent strength of hydration and biological structure. Biophys. J. 72, 65–76. 23. Kaushik, J.K. & Bhat, R. (1999) A mechanistic analysis of the increase in the thermal stability of proteins in aqueous carboxylic acid salt solutions. Protein Sci. 8, 222–233. 24. Low, L.K., Shin, H C., Narayan, M., Wedemeyer, W.J. & Scheraga, H.A. (2000) Acceleration of oxidative folding of bovine pancreatic ribonuclease A by anion-induced stabilization and formation of structured native-like intermediates. FEBS Lett. 472, 67–72. 25. Talluri, S., Rothwarf, D.M. & Scheraga, H.A. (1994) Structural characterization of a three-disulfide intermediate of robonuclease A involved in both folding and unfolding pathways. Biochemistry 33, 10437–10449. 26. Shimotakahara, S., Rios, C.B., Laity, J.H., Zimmerman, D.E., Scheraga,H.A.&Montelione,G.T.(1997)NMRstructural analysis of an analog of an intermediate formed in the rate- determining step of one pathway in the oxidative folding of bovine pancreatic ribonuclease A: automated analysis of 1 H, 13 C, and 15 N Ó FEBS 2002 Temperature dependence of protein unfolding kinetics (Eur. J. Biochem. 269) 5321 resonance assignments for wild-type and [C65S,C72S] mutant forms. Biochemistry 36, 6951–6929. 27. Laity, J.H., Lester, C.C., Shimotakahara, S., Zimmerman, D.E., Montelione, G.T. & Scheraga, H.A. (1997) Structural characteri- zation of an analog of the major rate-determining disulfide folding intermediate of bovine pancreatic ribonuclease A. Biochemistry 36, 12683–12699. 28. Laity, J.H., Shimotakahara, S. & Scheraga, H.A. (1993) Expression of wild-type and mutant bovine pancreatic ribonuc- lease A in Escherichia coli. Proc. Natl Acad. Sci. USA 90, 615–619. 29. Sogbein, O.O., Simmons, D.A. & Konermann, L. (2000) Effects of pH on the kinetic reaction mechanism of myoglobin unfolding studied by time-resolved electrospray ionization mass spectrome- try. J. Am. Soc. Mass Spec. 11, 312–319. 30. Matheson, R.R. & Scheraga, H.A. (1979) Steps in the pathway of the thermal unfolding of ribonuclease A. A nonspecific photo- chemical surface-labeling study. Biochemistry 18, 2437–2445. 31. Stelea, S.D., Pancoska, P., Benight, A.S. & Keiderling, T.A. (2001) Thermal unfolding of ribonuclease A in phosphate at neutral pH: Deviations from the two-state model. Protein Sci. 10, 970–978. 32. Muthusamy, R., M.M Gromiha. & P.K Ponnuswamy. (2000) On the thermal unfolding character of globular proteins. J. Protein Chem. 19, 1–8. 5322 Y B. Yan et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . folding and unfolding, will also be discussed. A new approach of image analysis, which is established by us recently [4,11], was used to analyze the unfolding. three-disulfide mutants of bovine pancreatic ribonuclease A missing the 65–72 disulfide bond: characterization of a minor folding pathway of ribonuclease A and