Protein Engineering, Vol. 14, No. 12, 967-974,
December 2001
© 2001 Oxford University Press
Energetics of three-state unfolding of a protein: canine milk lysozyme
1 Research Fellow of the Japan Society for the Promotion of Science (JSPS). 4 Division of Biological Sciences, Graduate School of Science,Hokkaido University, Kita-ku, Sapporo 060-0810, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Thermodynamics of thermal transitions of a calcium-binding lysozyme, canine milk lysozyme (CML), was studied using differential scanning calorimetry and compared with those for homologous proteins, human
-lactalbumin (-hLA) and equine milk lysozyme (EML). The results showed that CML and EML exhibit two clear heat absorption peaks in the absence of calcium ions (apo-form), although the cooperative thermal transition of -hLA is apparently absent in this form. The first peak represents the unfolding transition from the native to an unfolding intermediate state (N–I transition) and the second peak represents that from the intermediate to the thermally unfolded state (I–U transition). We interpret that the cooperative thermal transition, which is observed between the intermediate and the thermally unfolded states of CML and EML, comes from the native-like packing interaction in their intermediate states. Furthermore, to examine the role of the stabilization mechanism of CML intermediate, we constructed four variant CMLs (H21G, I56L, A93S and V109K), in which the residues of CML are substituted for those of EML, and also investigated their thermal stability. Especially the His21 and Val109 of CML play a role in stabilization of the intermediate state and their contributions to the unfolding free energy are estimated to be 2.0 and 1.8 kJ/mol, respectively. From the results of the mutational analysis, a few differences in the local helical interactions within the -domain are found to be predominant in stabilizing the intermediate state.
Keywords: canine milk lysozyme/cooperativity/differential scanning calorimetry/thermodynamics/three-state transition
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The lysozyme–
-lactalbumin superfamily is one of the most investigated protein superfamilies with biophysical and biochemical methodologies and has been especially utilized as useful model proteins for studying the folding mechanism (Kuwajima, 1989
, 1996). These two proteins have similar amino acid sequences (Brew et al., 1970) and very similar three-dimensional structures characterized by the presence of a deep cleft that divides the molecule into two domains (the - and ß-domains) (Struart et al., 1986; Acharya et al., 1991; Tsuge et al., 1992; Grobler et al., 1994; Pike et al., 1996). Nevertheless, the apparent equilibrium unfolding behaviors of these two proteins are known to be very different to each other. Thanks to the progress in folding studies, however, it was found that certain calcium-binding lysozymes (canine and equine milk lysozymes), in addition to -lactalbumin, exhibit a stable unfolded intermediate state at equilibrium (Morozova et al., 1991, 1995, 1997, 1999; Van Dael et al., 1993; Griko et al., 1995; Kikuchi et al., 1998; Mizuguchi et al., 1998; Koshiba et al., 1999,2000; Kobashigawa et al., 2000), although most of the conventional lysozymes usually unfold in a cooperative two-state manner without the accumulation of the intermediate state.
The intermediate states of canine (CML) and equine milk lysozymes (EML) are much more native-like than that of -lactalbumin, and they are thought to be stabilized by non-specific interactions and by specific side-chain packing interactions (Morozova et al., 1995, 1997,1999; Mizuguchi et al., 1998; Kobashigawa et al., 2000; Koshiba et al., 2000). Furthermore, our investigations have shown that the intermediate state of CML is extraordinarily stable compared with that of EML (Kobashigawa et al., 2000; Koshiba et al., 2000), so it has been thought that CML is one of the proteins that shows the most stable intermediate state in the members of the lysozyme–-lactalbumin superfamily, which has been studied previously (Koshiba et al., 2000). However, the detailed molecular mechanism of the stabilization of the intermediate state of CML has not yet been fully elucidated.
In this study, we investigated the thermodynamics of unfolding transitions of CML by differential scanning calorimetry (DSC) measurement and compared the values obtained with those obtained for homologous proteins, human -lactalbumin (-hLA) and EML. We first showed that the intermediate state of CML and EML is a thermodynamic state, whereas that of -hLA is thermodynamically indistinguishable from the unfolded state (Pfeil, 1998). We also examined the stabilization mechanism of the extraordinarily stable intermediate state of CML and compared it with the results for EML and variant CMLs (see Figure 1). These substitutions of CML (His21, Ile56, Ala93 and Val109) were designed for the examination of the role of the stabilization mechanism of CML intermediate due to the helical interaction in the -domain, with consideration of the NMR results for the thermal unfolding (Koshiba et al., 2000). From the results of the present study, it is concluded that the native-like packing interaction in the intermediate state induces a remarkably positive heat absorption peak with thermal unfolding as seen in the cases of the two calcium-binding lysozymes and the local helical interactions play a critical role in the stabilization of the intermediate state.
|
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Materials
All enzymes used for genetic engineering were purchased from Takara Shuzo (Kyoto, Japan) or Toyobo (Osaka, Japan). Oligonucleotide DNA primers were synthesized by Kurabo (Osaka, Japan). All other reagents were of biochemical research grade.
Site-directed mutagenesis of CML
Plasmid pHK7CML, which contains T7 transcription and translation signals and DNA encoding the chemically synthesized canine milk lysozyme sequence, was used as a starting material for mutagenesis (Koshiba et al., 1999). To obtain the variant CMLs (H21G, I56L, A93S and V109K), His21, Ile56, Ala93 and Val109 in CML were replaced by glycine, leucine, serine and lysine, respectively. Site-directed mutagenesis was performed as described previously (Koshiba et al., 1998). The mutations were confirmed by DNA sequencing.
Expression and refolding of wild-type and variant CMLs
Overexpression of wild-type and variant CMLs in Escherichia coli, as inclusion bodies and refolding procedures using E.coli thioredoxin, were performed as described previously (Koshiba et al., 1998, 1999).
Preparation of apo- and holo-proteins
Apo- and holo-proteins were prepared as described previously (Koshiba et al., 1999, 2000).
Differential scanning calorimetry (DSC)
DSC measurements of CML were carried out with an MC-2 microcalorimeter (MicroCal, Northampton, MA) at a scan rate of 1.0 K/min. Sample solutions for DSC measurements between pH 2.0 and 4.0 were prepared by dissolving the proteins in 50 mM glycine hydrochloride buffer; solutions between pH 4.0 and 4.5 were prepared using 50 mM sodium acetate buffer. The protein concentrations used were 1.2–1.8 mg/ml. The pH of the sample solution was confirmed before and after each measurement. To extend the temperature range, all DSC measurements were performed under an excess pressure of 1.5 atm.
Structure analysis
The atomic coordinates of the proteins analyzed here were obtained from the Protein Data Bank [CML (Koshiba et al., 2000), -hLA (Acharya et al., 1991) and EML (Tsuge et al., 1992)]. The hydrogen exchange rate data of CML, -hLA and EML at pH 2.0 (5 and 25°C) were obtained from previous reports (Morozova et al., 1995, 1997; Schulman et al., 1995; Kobashigawa et al., 2000). Molecular graphics were produced using WebLab Viewer (Molecular Simulations) and MOLSCRIPT (Kraulis, 1991).
Estimation of protein concentration
The protein concentrations were estimated spectrophotometrically using an extinction coefficient at 280 nm of E1%1 cm = 23.2 for wild-type and its variant of CML (Kikuchi et al., 1998).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Thermal unfolding of CML in the absence of calcium ions
An important question regarding the intermediate state is whether it is a thermodynamic state. To address this question, DSC measurements of the CML were carried out at acidic pH between 2.0 and 4.5 in the absence of calcium ions (apo-form; Figure 2) and compared with those for -hLA (Mizuguchi et al., 2000) and EML (Van Dael et al., 1993; Griko et al., 1995) reported previously. In this pH region, the thermal unfolding of CML was reversible. This was confirmed by reheating the protein solution in the calorimeter cell immediately after cooling from the first scan. Repeated thermal scans of the samples exhibited similar transitions (data not shown). The cooperative thermal transition is apparently absent in the molten globule state (at acidic pH; pH 2.2) of -hLA (Mizuguchi et al., 2000), suggesting that this state is thermodynamically indistinguishable from the unfolded state. This could mean that there are no energy barriers between the molten globule and unfolded states, as previously reported for bovine -LA (Yutani et al., 1992; Griko et al., 1994; Pfeil, 1998).
|
However, CML exhibits a cooperative thermal unfolding with two remarkably positive heat absorption peaks (Figure 2
). This indicates that the first peak represents the unfolding transition from the native to an unfolding intermediate state and the second peak represents the unfolding transition from the intermediate to the thermally unfolded state. Although similar behaviors (two positive heat absorption peaks) have been observed in apo-EML (Van Dael et al., 1993; Griko et al., 1995), the transition from the intermediate to the thermally unfolded state of apo-CML occurred at a higher temperature and 
HI–U was larger than that of apo-EML. Among the -LA and lysozyme family of proteins studied thus far (Yutani et al., 1992; Van Dael et al., 1993; Griko et al., 1994, 1995; Pfeil, 1998; Griko, 2000; Koshiba et al., 2000), CML exhibits the most stable intermediate state with the largest HI–U. Thermal unfolding of CML in the presence of calcium ions
The temperature dependence of the heat capacity of CML in 50 mM sodium acetate buffer (pH 4.5) in the presence of 10 mM CaCl2 and in the absence of CaCl2 is shown in Figure 3. An increase in calcium concentration (1, 2, 5 and 10 mM) leads to a shift of the first heat absorption peak to higher temperatures (data not shown). On the other hand, the second peak does not show this calcium dependence, which indicates that the shift of the first peak reflects the stabilization of the native state by calcium binding.
|
Under conditions such as pH 4.5 and the presence of 10 mM CaCl2,
-hLA and EML are known to show a single heat absorption peak caused by a cooperative transition from the native to the thermally unfolded state (Van Dael et al., 1993; Griko et al., 1995; Mizuguchi et al., 2000). In the case of CML, however, two clear heat absorption peaks were observed in the holo-form (Figure 3), unlike the case of -hLA or EML. Therefore, the comparison of the heat absorption peaks in the apo- and holo-forms of CML also indicates that the extraordinary stability of the intermediate state of this protein has led to the appearance of the second heat absorption peak in the holo-protein. Analysis of the excess heat capacity function of CML
To understand fully the stability of a protein, it is necessary to elucidate the thermodynamic parameters of the thermal unfolding. In order to estimate the thermodynamic parameters such as melting temperature (Tm), enthalpy change (H), entropy change (S) and Gibbs free energy change (G) of a three-state unfolding protein, the heat capacity changes of the protein in a thermodynamic state must first be assumed.
The heat capacity of a macromolecule under conformational equilibrium, Cp,obs, is given by the following equation (Griko et al., 1995):
 | (1) |
Cpi is the relative heat capacity of the ith state compared with that of the native state. The second term on the right-hand side of Equation 1 represents the fluctuation of conformation of the macromolecule (Griko et al., 1995) and the third term accounts for the heat capacities of conformers other than the native state.
In this study, the thermal unfolding profile is that of a three-state reversible unfolding (N, I and U). Therefore, Equation 1 can be rewritten as follows:
 | (2) |
CpN–U of hen egg white lysozyme (HEL) are temperature dependent and are represented as a linear and a quadratic function of temperature, respectively (Griko et al., 1994, 1995; Gomez et al., 1995; Xie et al., 1995), as follows:
 | (3) |
 | (4) |
). Given that the heat capacities of structurally similar proteins are similar (Gomez et al., 1995; Griko et al., 1995), we can assume for the purpose of our study that the CpN and CpU of CML can be approximated as those of HEL. The heat capacity change of each conformational transition of CML is assumed as
 | (5) |
 | (6) |
represents the fraction of heat capacity change of the N–I transition relative to that of the N–U transition and is assumed to be independent of temperature.
The enthalpy and entropy changes, from the native to an unfolded intermediate state [HN–I(T), SN–I(T)] and from the intermediate to the thermally unfolded state [HI–U(T), SI–U(T)], can be written as
 | (7) |
 | (8) |
 | (9) |
 | (10) |
 | (11) |
 | (12) |
 | (13) |
HN–I(Tm1), HI–U(Tm2), Tm1, Tm2, and CpN–U(T). The constants HN–I(Tm1), HI–U(Tm2), Tm1, Tm2, and A were determined for each set of experimental data using the non-linear least-squares method. Determination of the thermodynamic parameters of wild-type and variant CMLs
First, the parameter was treated as an unknown constant to be determined by the non-linear least-squares method for six data sets of wild-type apo-CML at different pHs from 2.74 to 4.50 (Figure 2). The mean of the resulting values for was 0.579 with a standard deviation of 0.018. Then, the thermodynamic parameters were recalculated with a fixed of 0.579, because the heat capacity change is usually not very dependent on pH. This value of was also assumed for holo- and variant CMLs.
Table I summarizes the thermodynamic parameters of wild-type CML obtained by the above methods with the fixed at several pHs. The temperature dependence of HN–I(Tm1) and HI–U(Tm2) is shown in Figure 4. The data for HN–I(Tm1) and HI–U(Tm2) were taken from Table I. These plots were well fitted to each enthalpy function as follows:
 | (14) |
 | (15) |
|
|
Table II
shows the thermodynamic parameters of the wild-type, H21G, I56L, A93S and V109K variants of CML and of -hLA and EML in the absence of calcium ions at pH 4.5 (see also Figure 5). The thermal stability of the intermediate state of all variant CMLs decreased relative to that of wild-type CML. In fact, the thermal stability of the intermediate state of the H21G and V109K variants was about 3–4 °C less stable than that of wild-type CML. Thus, the data indicated that the four substituted residues in CML played an important role in creating the extraordinary stability of the intermediate state.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Cooperative folding, which is specific to globular proteins, results from tight and unique packing of amino acid residues in the native structure (Privalov, 1979
; Privalov and Gill, 1988; Shakhnovich and Finkelstein, 1989). Therefore, the thermal transition of the ‘classical’ molten globule states is expected to be a weakly or non-cooperative process, since they do not possess a unique tertiary structure (Kuwajima et al., 1976; Kuwajima, 1989, 1996; Ptitsyn, 1992, 1995). However, our DSC results clearly show that the unfolding intermediate state of apo- and holo-CML can be considered as a highly cooperative intermediate as seen in the case of apo-EML (Griko et al., 1995), whereas that of the molten globule state of -hLA can be considered as a non-cooperative intermediate (Mizuguchi et al., 2000). In order to explain the cooperative thermal transition of the intermediate state of CML, we compared the hydrogen exchange properties (protection factors) of the intermediate state (Kobashigawa et al., 2000) with those of the sequentially and structurally homologous proteins -hLA and EML (Morozova et al., 1995, 1997; Schulman et al., 1995).
There are several important features of the amide hydrogen exchange protection exhibited in the - and ß-domains of CML, -hLA and EML (Figure 6). In the case of -hLA, the protected protons were detected only within the -domain and none of the amide protons within the ß-domain were significantly protected from solvent exchange in the molten globule state (Schulman et al., 1995) (Figure 6, top). This indicates that the molten globule state of -hLA has a bipartite structure with a disordered ß-domain and an -domain containing substantial secondary structure and a native-like tertiary topology (Peng and Kim, 1994; Peng et al., 1995; Schulman et al., 1995; Wu et al., 1995). Therefore, the cooperative thermal transition of -hLA is apparently absent in this state because of the compact but dynamic structure (less ordered structure). In contrast, the number of protected amide protons in the intermediate state of CML (64 amide protons) and EML (54 amide protons) were much larger than that of -hLA (34 amide protons) (Morozova et al., 1995,1997; Kobashigawa et al., 2000). The protected protons of CML and EML in this state were observed not only within the -domain but also within the ß-domain (Figure 6; middle and bottom). It is suggesting that highly persistent structures exist in the - and ß-domain of CML and EML in the intermediate states and especially, the ß-domain of the two proteins are more ordered than that of -hLA. Thus, it can be proposed that the cooperative thermal transitions of the intermediate states, which have been observed in CML and EML, come from a highly ordered structure of the ß-domain in the intermediate states. This is one of the main reasons why the cooperative thermal transitions are observed in the intermediate states of CML and EML and not in that of the molten globule state of -hLA.
|
As discussed above, it has been shown that CML and EML show the presence of native-like tertiary folding in the intermediate state, which, in these two proteins, is a thermodynamic state. The unfolding intermediate state of CML, in particular, has been shown to be extraordinarily stable in comparison with that of EML, 9.0 kJ/mol more stable at pH 4.5 and 66.4°C in the absence of calcium ions (Table III
). However, what causes these differences in the stability of the intermediate state between CML and EML? One way to investigate the differences is to compare thoroughly the physicochemical properties of the unfolding intermediate states of CML and EML and also to consider the amino acid substitutions of these two proteins. Several considerations provide good evidence to explain the stabilization mechanism of the intermediate state of CML.
|
First, Gly21 of EML is replaced by His21 in CML. In the native structure of CML, His21 is located at the turn region between the A- and B-helices and is close to the C-helix, as shown in Figure 7A
(also see Figure 1). The distance between the N
2 atom of His21 and the O atom (C=O) of Val99 is 3.13 Å, indicating the formation of a hydrogen bond between these two atoms in the native structure. As a result, the A- and B-helices become much closer to the C-helix and hydrophobicity around this region is increased. In fact, hydrogen exchange rates in this region of CML have exhibited uniform behavior. Eleven aromatic and aliphatic residues within this region (Arg10, Lys11, Leu12, Lys13 and Me-15 of the A-helix; Phe20 and Tyr23 of the A–B loop; and Cys30, Met31, Ala32 and Glu33 of the B-helix) exhibit very high protection factors (P > 100) in the intermediate state (Kobashigawa et al., 2000). In this condition (at pH 2.0 and 25°C; see also Figure 2), the most protected residues of the A- and B-helices are located in the interface of these helices with the C-helix. However, protection factors of EML in this region were much lower (10 < P < 1) than those of CML in the intermediate state (Figure 7B). The DSC measurements of H21G variant show that this amino acid replacement, which we discussed above, actually decreases the stability of the unfolding intermediate state of CML and the value for H21G variant is 2.0 kJ/mol lower than that of the wild-type CML (Figure 5 and Table III). Therefore, it can be estimated that the replacement of a histidine residue by a glycine residue would destabilize the protein at most by 2.0 kJ/mol in the G of the I–U transition and by 7.7 kJ/mol in the H of the I–U transition.
|
Second, residue 109 is located at the N-terminus of the D-helix (Figure 1
). This residue is valine and lysine in CML and EML, respectively (Figure 8A and B). The helix contact between the B- and D-helices is revealed clearly in the hydrogen exchange data for the intermediate states of CML and EML. Seven aromatic and aliphatic residues around this region (Cys30, Met31, Ala32 and Glu33 of the B-helix; Trp108 of the C–D loop; Val109 of the D-helix; and His114 of the loop) exhibit the highest protection factors (P > 100) at pH 2.0 and 25°C in CML. The most protected residues of the B-helix are also located in the interface of this helix with the D-helix. Val109 is particularly strongly immobilized in the intermediate state (P = 330) and seems to stabilize the D-helix. However, the residue at this site is lysine in EML and such high protection is not observed in the case of the intermediate state of EML. This suggests that the hydrogen exchange kinetics of Lys109 occur so rapidly that the protection factor cannot be determined. This difference in the exchange of Val109 and Lys109 must be a major reason for the destabilization of the D-helix of EML. In fact, the thermal stability of V109K variant is 1.8 kJ/mol lower than that of the wild-type CML (Table III). In the case of the V109K variant, two positively charged side chains (Lys109 and -113) are thought to destabilize the native structure due to electrostatic repulsion. Between the thermodynamic parameters for the I–U transition of CML and V109K variant, the major term to destabilize the intermediate state is the entropy term (TS = 5.8 kJ/mol) (Table III). The enthalpy terms seem to compensate only partly for the entropic destabilization of V109K variant. It appears that the destabilization of the D-helix is caused by the increase in chain entropy in the intermediate state, which is caused by the existence of the lysine residue at the N-terminus of the D-helix. This may indicate that the helical contact between the B- and D-helices of EML is much weaker than that of CML.
|
Finally, we have previously noted the differences between CML and EML in their amino acid sequences (Koshiba et al., 2000
) and these differences should be responsible for the differences in the stability of the intermediate state. Among these substitutions, those at residues 56 (CML
Ile; EMLLeu) and 93 (CMLAla; EMLSer) seem to be important in relation to differences in the stability of the intermediate states for the following reasons. Residue Ile56 is located in the ß-domain of CML (Figure 1). This residue is thought to interact with a hydrophobic cluster within the -domain and thought to play a domain-anchored role (- and ß-domain). However, a substitution from isoleucine to leucine may be formed a new hydrophobic contact between Leu56 and the cluster and, as a result, the thermal stabilities of native and intermediate states are decreased (Figure 5). In fact, the thermal stability of I56L variant is 1.0 kJ/mol lower than that of the wild-type CML (Table III). Consistent with this view, the ß-domain of CML is more ordered than that of EML (Figure 6). Moreover, residue 93 is located in the middle position of the C-helix (Figure 1). Alanine is thought to be more favorable for forming an -helix than serine when the alanine residue is located in the central position of the -helix (Merutka et al., 1990). The DSC measurements of A93S variant show that this amino acid replacement, which we discussed above, actually decreases the stability of the unfolding intermediate state of CML and the value for A93S variant is 1.1 kJ/mol lower than that of the wild-type CML (Figure 5 and Table III). Conclusion
We propose that the cooperative thermal transitions of the intermediate states, which have been observed in CML and EML, come from the highly persistent structures existing not only within the -domain but also within the ß-domain. This is one of the main reasons why the cooperative thermal transitions are observed in the intermediate states of CML and EML and not in that of the molten globule state of -hLA. Also, our results suggest that a few differences in the local helical interactions (A-, B- and C-helix interactions; B- and D-helix interactions) predominantly stabilize the intermediate state.
|
Notes |
|---|
2 Present address: Division of Biology, MC156-29, California Institute of Technology, 1200 E. California Blvd, Pasadena, CA 91125, USA
3 To whom correspondence should be addressed.E-mail: nitta{at}sci.hokudai.ac.jp
|
Acknowledgments |
|---|
We gratefully acknowledge the invaluable advice of Professor Isao Tanaka, Dr Min Yao of the Division of Biological Sciences, Graduate School of Science, Hokkaido University, Dr Mineyuki Mizuguchi of the Faculty of Pharmaceutical Sciences, Toyama Medical and Pharmaceutical University and of Dr Munehito Arai of The University of Tokyo. We also thank Ms Hiroko Yamamoto (Hokkaido University) for producing the variant CMLs. This study was supported by Grants-in-Aid from the Ministry of Agriculture, Forestry and Fishery of Japan and by Fellowships from the Japan Society for the Promotion of Science for Young Scientists (T.K.).
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Acharya,K.R., Ren,J., Struart,D.I., Phillips,D.C. and Fenna,R.E. (1991) J. Mol. Biol., 221, 571–581.[Web of Science][Medline]
Brew,K., Castellino,F.J., Vanaman,T.C. and Hill,R.L. (1970) J. Biol. Chem., 245, 4570–4582.
Gomez,J., Hilser,V., Xie,D. and Freire,E. (1995) Proteins, 22, 404–412.[Web of Science][Medline]
Griko,Y.V. (2000) J. Mol. Biol., 297, 1259–1268.[Web of Science][Medline]
Griko,Y.V., Freire,E. and Privalov,P.L. (1994) Biochemistry, 33, 1889–1899.[Medline]
Griko,Y.V., Freire,E., Privalov,G., Van Deal,H. and Privalov,P.L. (1995) J. Mol. Biol., 252, 447–459.[Web of Science][Medline]
Grobler,J.A., Rao,K.R., Pervaiz,S. and Brew,K. (1994) Arch. Biochem. Biophys., 313, 360–366.[Web of Science][Medline]
Kikuchi,M., Kawano,K. and Nitta,K. 1998. Protein Sci., 7,2150–2155.[Web of Science][Medline]
Kobashigawa,Y., Demura,M., Koshiba,T., Kumaki,Y., Kuwajima,K. and Nitta,K. (2000) Proteins, 40, 579–589.[Web of Science][Medline]
Koshiba,T., Tsumoto,K., Masaki,K., Kawano,K., Nitta,K. and Kumagai,I. (1998) Protein Eng., 11, 683–690.
Koshiba,T., Hayashi,T., Ishido,M., Kumagai,I., Ikura,T., Kawano,K., Nitta,K. and Kuwajima,K. (1999) Protein Eng., 12, 429–435.
Koshiba,T., Yao,M., Kobashigawa,Y., Demura,M., Nakagawa,A., Tanaka,I., Kuwajima,K. and Nitta,K. (2000) Biochemistry, 39, 3248–3257.[Medline]
Kraulis,P.J. (1991) J. Appl. Crystallogr., 24, 946–950.
Kuwajima,K. (1989) Proteins, 6, 87–103.[Web of Science][Medline]
Kuwajima,K. (1996) FASEB J., 10, 102–109.[Abstract]
Kuwajima,K., Nitta,K., Yoneyama,M. and Sugai,S. (1976) J. Mol. Biol., 106, 359–373.[Web of Science][Medline]
Merutka,G., Lipton,W., Shalongo,W., Park,S.H. and Stellwagen,E. (1990) Biochemistry, 29, 7511–7515.[Medline]
Mizuguchi,M., Arai,M., Yue,K., Nitta,K. and Kuwajima,K. (1998) J. Mol. Biol., 283, 262–277.
Mizuguchi,M., Masaki,K., Demura,M. and Nitta,K. (2000) J. Mol. Biol., 298, 985–995.[Web of Science][Medline]
Morozova,L.A., Haezebrouck,P. and Van Cauwelaert,F. (1991) Biophys. Chem., 41, 185–191.[Web of Science][Medline]
Morozova,L.A., Haynie,D.T., Arico-Muendel,C., Van Dael,H. and Dobson,C.M. (1995) Nature Struct. Biol., 2, 871–875.[Web of Science][Medline]
Morozova,L.A., Arico-Muendel,C., Haynie,D.T., Emelyanenko,V.I., Van Dael,H. and Dobson,C.M. (1997) J. Mol. Biol., 268, 903–921.[Web of Science][Medline]
Morozova,L.A., Jones,J.A., Noppe,W. and Dobson,C.M. (1999) J. Mol. Biol., 289, 1055–1073.[Web of Science][Medline]
Peng, Z-Y and Kim,P.S. (1994) Biochemistry, 33, 2136–2141.[Medline]
Peng, Z-Y, Wu,L.C. and Kim,P.S. (1995) Biochemistry, 34, 3248–3252.[Medline]
Pfeil,W. (1998) Proteins, 30, 43–48.[Web of Science][Medline]
Pike,A.C.W., Brew,K. and Acharya,K.R. (1996) Structure, 4, 691–703.[Medline]
Privalov,P.L. (1979) Adv. Protein Chem., 33, 167–241.[Medline]
Privalov,P.L. and Gill,S.L. (1988) Adv. Protein Chem., 39, 191–234.[Web of Science][Medline]
Privalov,P.L. and Makhatadze,G.I. (1992) J. Mol. Biol., 224, 715–723.[Web of Science][Medline]
Ptitsyn,O.B. (1992) In Creighton,T.E. (ed.), Protein Folding. Freeman, San Francisco, pp. 243–300.
Ptitsyn,O.B. (1995) Adv. Protein Chem., 47, 83–229.[Web of Science][Medline]
Schulman,B.A., Redfield,C., Peng, Z.-y., Dobson,C.M. and Kim,P.S. (1995) J. Mol. Biol., 253, 651–657.[Web of Science][Medline]
Shakhnovich,E. and Finkelstein,A.V. (1989) Biopolymers, 28, 1667–1680.[Web of Science][Medline]
Struart,D.I., Acharya,K.R., Walker,N.P.C., Smith,S.G., Lewis,M. and Phillips,D.C. (1986) Nature, 324, 671–675.
Tsuge,H., Ago,H., Noma,M., Nitta,K., Sugai,S. and Miyano,M. (1992) J. Biochem., 111, 141–143.
Van Dael,H., Haezebrouck,P., Morozova,L.A., Arico-Muendel,C. and Dobson,C.M. (1993) Biochemistry, 32, 11886–11894.[Medline]
Wu,L.C., Peng, Z-Y and Kim,P.S. (1995) Nature Struct. Biol., 2, 281–286.[Web of Science][Medline]
Xie,D., Fox,R. and Freire,E. (1995) Protein Sci., 3, 2175–2184.[Web of Science][Medline]
Yutani,K., Ogasahara,K. and Kuwajima,K. (1992) J. Mol. Biol., 228, 347–350.[Web of Science][Medline]
Received April 19, 2001; revised August 2, 2001; accepted September 4, 2001.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||