Protein Engineering, Vol. 12, No. 10, 841-850,
October 1999
© 1999 Oxford University Press
Contribution of amino acid substitutions at two different interior positions to the conformational stability of human lysozyme
1 Institute for Protein Research, Osaka University, Yamadaoka, Suita, Osaka 565-0871 and 2 Graduate School of Pharmaceutical Sciences, Osaka University, Yamadaoka, Suita, Osaka 565-0871, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
To elucidate correlative relationships between structural change and thermodynamic stability in proteins, a series of mutant human lysozymes modified at two buried positions (Ile56 and Ile59) were examined. Their thermodynamic parameters of denaturation and crystal structures were studied by calorimetry and X-ray crystallography. The mutants at positions 56 and 59 exhibited different responses to a series of amino acid substitutions. The changes in stability due to substitutions showed a linear correlation with changes in hydrophobicity of substituted residues, having different slopes at each mutation site. However, the stability of each mutant was found to be represented by a unique equation involving physical properties calculated from mutant structures. By fitting present and previous stability data for mutant human lysozymes substituted at various positions to the equation, the magnitudes of the hydrophobicity of a carbon atom and the hydrophobicity of nitrogen and neutral oxygen atoms were found to be 0.178 and –0.013 kJ/mol.Å2, respectively. It was also found that the contribution of a hydrogen bond with a length of 3.0 Å to protein stability was 5.1 kJ/mol and the entropy loss of newly introduction of a water molecules was 7.8 kJ/mol.
Keywords: calorimetry/human lysozyme/mutant protein/protein stability/X-ray structural analysis
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The conformational stability of most proteins that results from a combination of effects, i.e. hydrophobicity, hydrogen bonding, conformational entropy and other causes, is low, only 40 kJ/mol (Privalov and Gill, 1988
). The magnitude of each of these factors (interactions) is not so small but they compensate for each other (Dill, 1990). Studies of mutant proteins should be useful for estimating the effect of various factors on protein stability, but the same kinds of substitutions have given different results. For example, in the case of Ile 
Val mutants, the denaturation Gibbs energy changes (
G) range from –1.5 to –5.0 kJ/mol for five mutant human lysozymes (Takano et al., 1995), from –2.1 to –7.5 kJ/mol for five mutant staphyloccocal nucleases (Shortle et al., 1990) and from –0.8 to –7.5 kJ/mol for eight mutant barnases (Serrano et al., 1992). This shows that the effect of a single amino acid substitution on protein stability depends on the location of the mutation site and its environment within the protein structure. Moreover, a substitution may affect not only the mutation site but also other parts of the protein far from the site, although the structural changes are not large (Takano et al., 1997a).
The first correction to be made in reconciling different values is to compare G values at the same solvent accessibility of the residue (Pace, 1992), because the G values for hydrophobic mutants increase with the extent of buried non-polar surface area (Yutani et al., 1987; Matsumura et al., 1988a). Although this can be partially corrected, in the case of Ile Val mutant barnases, the G values still range from –2.1 to –7.5 kJ/mol after the correction (Pace et al., 1996). Eriksson et al. (1992) have shown that the range of G values observed for Leu Ala mutant phage T4 lysozymes results from the mutants having different size cavities. The same has been observed for other T4 lysozyme mutants (Xu et al., 1998) and for Ile Ala barnase mutants (Buckle et al., 1996). For the Ile Val and Val Ala mutations within the hydrophobic core of barnase and human lysozyme, however, such a correlation between the size of cavity created by the substitution and G value has not been observed (Buckle et al., 1993; Takano et al., 1995, 1997a). In the case of five Ile Val and nine Val Ala mutant human lysozymes, a correlation between G values and changes in hydrophobic surface area exposed by denaturation has been found, if the effect of the secondary structure propensity is taken into account (Takano et al., 1995, 1997a). In order to reconcile a number of conflicting reports concerning the contribution of different factors to protein stability, it is highly desirable that data on structure and stability changes are increased by carrying out a more systematic study of mutant proteins with predetermined substitutions.
Human lysozyme (130 residues) is a good model for study because it is possible to obtain qualitative thermodynamic parameters from differential scanning calorimetry (DSC) measurements of the heat-denaturation process and high resolution three-dimensional structures of the mutant proteins. Studies on the structure and stability of 21 hydrophobic mutant human lysozymes (5 Ile Val, Ala; 2 Ile Gly; 9 Val Ala) have been reported by Takano et al. (1995, 1997a,b) to clarify the contribution of the hydrophobic effect on conformational stability. Recently, the contribution of the hydrophobic effect and hydrogen bonds on the stability of human lysozyme, from 14 mutants [5 Ile Val; 9 Val Ala mutants lacking an S–S bond between Cys77 and Cys95 (Takano et al., 1998)] and 12 mutants [6 Tyr Phe (Yamagata et al., 1998) and 6 Ser Ala (Takano et al., 1999a)], respectively, have also been examined.
Positions 56 and 59 of human lysozyme, which are located between two ß-strands and in a ß-strand, respectively, are completely buried in the interior of the protein molecule. Previous studies have demonstrated dramatically different responses to Ile Val or Ala substitutions at both sites (Takano et al., 1995, 1997b). To estimate the effect of the structural changes on the stability in the different environments, this paper focuses on a series of amino acid modifications at positions 56 and 59 (Ile Gly, Ala, Val, Leu, Met, Phe, Ser, Thr or Tyr). The thermodynamic parameters upon denaturation and crystal structures for these mutant proteins were determined by DSC and high-resolution X-ray crystallography, respectively. The present data were combined with similar data from previous studies of mutant human lysozymes (Takano et al., 1995, 1997a, b, 1998; Funahashi et al., 1996; Yamagata et al., 1998) to estimate the contribution of the hydrophobic effect of carbon atoms and neutral oxygen/nitrogen atoms to protein stability. Furthermore, using these parameters, the contributions of introducing a water molecule and hydrogen bond could be estimated.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Mutant proteins
Mutagenesis, expression and purification of a series of mutant human lysozymes at Ile56 and Ile59 were performed as described (Takano et al., 1995). Mutant proteins (15 mutants) with Ala, Val, Leu, Met, Phe or Thr substitutions at positions 56 and 59 and Gly, Ser or Tyr substitutions at position 59 were prepared, but other mutant proteins could not be obtained owing to the extremely low yield in the yeast expression system. DNA sequence analysis was carried out using an automated DNA sequencer at the Research Center for Protein Engineering, Institute for Protein Research, Osaka University. Protein concentration was determined spectrophotometrically using E1%1 cm = 25.65 at 280 nm for human lysozyme (Parry et al., 1969) and its mutants, except for the Tyr-substituted mutant. The concentration of the Tyr mutant protein at position 59 was determined spectrophotometrically using E1%1 cm = 26.59 at 280 nm with a correction for the increase in the molar absorption coefficient of Tyr (Wetlaufer, 1962).
X-ray crystallography
Mutant human lysozymes were crystallized, diffraction data collected and the structures refined (Brunger, 1992) as described previously (Takano et al., 1995, 1997a), except for data collection for I59S and structure determination for I56M and I56F. Crystals of most mutants belong to the same crystal form as the wild-type proteins (P212121, a = 56.7, b = 61.1, c = 33.8 Å; type I) (Takano et al., 1995, 1997a,b, 1998, 1999a,b; Funahashi et al., 1996; Yamagata et al., 1998). However, the crystals of I56M and I56F (P212121, a = 64.7, b = 110.3, c = 43.6 Å; type II) differed from that of the wild-type protein. The two kinds of crystal forms corresponded to those of the wild-type human lysozyme (a = 57.1, b = 61.0, c = 33.0 Å and a = 65.3, b = 110.5, c = 43.7 Å) as reported by Osserman (1969).
For I59S, the crystal was small. The data set was collected using synchrotron radiation at the Photon Factory (Tsukuba) on beam line 18B (wavelength 1.0 Å) with a Weissenberg camera (Sakabe, 1991). The data were processed with DENZO (Otwinowski, 1990).
The structure of I56M was solved by the molecular replacement technique using the program AMoRe (Navaza, 1994) with the wild-type structure as a search model. The refinement of the I56F structure was carried out using the model of I56M. The mutants that crystallized non-isomorphously exhibited a pronounced increase in the r.m.s. coordinate deviation for the structure considered as a whole. The r.m.s. deviations (0.3–0.4 Å) for the C
atoms among three kinds of molecules in type I and type II crystals were larger than those (0.1–0.15 Å) for C atoms among crystallographically identical molecules. The crystallographic data of mutant proteins are given in Table I.
|
The coordinates of the mutant human lysozymes, I56F, I56L, I56M, I59F, I59L, I59M, I59S, I59T and I59Y have been deposited in the Brookhaven Protein Data Bank, accession numbers 2MEA to 2MEI.
DSC measurements
Calorimetric measurements and data analyses were carried out as described (Takano et al., 1995). The scan rate was 1.0 K/min. Sample solutions for DSC measurements were prepared by dissolving the lysozyme in 0.05 M glycine buffer between pH 2.4 and 3.2. Under these solvent conditions, heat denaturation of human lysozyme was reversible. The lysozyme concentrations were 0.7–1.5 mg/ml. Data analysis was performed using Origin software (MicroCal, Northampton, MA). The thermodynamic parameters for denaturation as a function of temperature were calculated using the following equations:
 |
 |
 |
Cp does not depend on temperature (Privalov and Khechinashivili, 1974). |
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
X-ray structures of the mutant human lysozymes
All mutant structures determined were essentially identical with the wild-type structure. However, in the case of hydrophobic mutant human lysozymes (Ile Val and Val Ala), the substitutions affect not only the mutation site but also other parts of the protein far from the site, although the structural changes are not large (Takano et al., 1995, 1997a,b). Therefore, when investigating the relationship between changes in thermodynamic parameters and molecular structures caused by mutations, subtle changes due to rearrangements of the overall structure should also be considered.
In the case of mutants at position 56, the cavities created by substitutions remained empty, whereas new water molecules were found in cavities created at I59S and I59T. Figure 1 shows the structures in the vicinity of residue 59 in the wild-type and mutant proteins. An additional water molecule held by two hydrogen bonds in the cavity of I59V (Figure 1d) (Takano et al., 1995) and additional water molecules in the cavities of I59G and I59A, resulting in the formation of a hydrogen bonding network (Figure 1b and c), are observed (Takano et al., 1997b). For I59T, the hydroxyl group of the introduced Thr59 formed a hydrogen bond with a newly introduced water molecule (Figure 1f). For I59S, the hydroxyl group of the introduced Ser59 participated in the hydrogen bonding network, including additional water molecules (Figure 1e). When the hydroxyl groups are introduced into the cavity, all of the mutant proteins examined [I56T (Funahashi et al., 1996), I59S, I59T and I59Y] formed hydrogen bonds with buried water molecules or polar groups. In the case of I56T (Funahashi et al., 1996), the substitution residue (Thr) forms a hydrogen bond with an original water molecules in the wild-type protein. This indicates that polar groups in the interior of a protein molecule are prone to form hydrogen bonds.
|
Stability of the mutant human lysozymes
Changes in the conformational stability of human lysozyme due to substitutions were measured by DSC at pH values between 2.4 and 3.2. In the acidic pH region, a high reversibility of thermal denaturation of mutant and wild-type proteins was observed. Typical excess heat capacity profiles of wild-type protein and I59S are shown in Figure 2. All examined proteins gave similar profiles. The denaturation temperature (Td), the calorimetric enthalpies (Hcal), the van't Hoff enthalpies (HvH) and the heat capacity changes (Cp) were obtained directly from an analysis of these curves (Table II). To minimize errors arising from estimating the experimental results, the thermodynamic parameters of denaturation of the mutant proteins were compared at the denaturation temperature of wild-type protein at pH 2.7 (64.9°C). As shown in Table III, all mutant proteins were destabilized relative to the wild-type protein. Identical substitution at different positions and different substitutions at the same position resulted in different degrees of destabilization.
|
|
|
The calorimetric enthalpy changes ranged from 409 to 480 kJ/mol (Table III
). This range was slightly high when compared with the range 440–484 kJ/mol for Val Ala mutants (Takano et al., 1997a). |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Linear hydrophobic correlation has different slopes for mutant human lysozymes at positions 56 and 59
Takano et al. (1997b) have found differences in environment between positions 56 and 59, which are completely buried in the interior of a protein, based on the calculation of cavity volume for I56A and I59A. A large cavity has been detected in the model structure created by deleting methylene groups of Ile59 based on the coordinate of the wild-type and in the real mutant structure (I59A) with a probe radius of 1.6 Å, which is larger than that typically assumed for the radius of a water molecule (1.4 Å). However, no cavities have been detected in either the model or real structures of I56A with a probe radius of 1.6 Å. These results indicate that the space in the vicinity of the side chains around position 56 is tightly packed, more so than that of position 59. Although none of the position 56 mutants had any water molecule newly introduced into the cavities created by the substitutions, some position 59 mutants (I59G, I59A, I59V, I59S, I59T) had newly introduced water molecules. Differences between the positions 56 and 59 mutants were also observed in their stabilities.
A linear correlation between stability changes in the mutant proteins and hydrophobicities of the substituted residues, except for mutant proteins with aromatic amino acids (Phe, Tyr or Trp), have been reported in studies on tryptophan synthase subunit (Yutani et al., 1987), T4 lysozyme (Matsumura et al., 1988a), kanamycin nucleotidyl transferase (Matsumura et al., 1988b), barnase (Kellis et al., 1989), gene V protein (Sandberg and Terwilliger, 1991) and RNase HI (Akasako et al., 1997). However, the slopes of the linear correlation differed from each other, indicating that the contribution of amino acid residues to the stability differs depending on the location of the site. Figure 3 shows the correlation between unfolding Gibbs energy changes (G) of mutant human lysozymes substituted at positions 56 and 59 and differences in transfer Gibbs energy (Gtr) from ethanol to water of the substituted residues (Tanford, 1962). For mutant human lysozymes, a linear correlation was also found at each position except for I56F, I59F and I59Y (Figure 3), but their slopes differed from each other (2.5 versus 1.6, respectively). These results indicate that the effect of substituted residues on protein stability varies depending on the environment of the mutation site (Eriksson et al., 1993), but the linearity at each position is still maintained except for bulky mutant residues.
|
) and Ile59 (•) mutants, plotted against difference between isoleucine and individual amino acid of transfer Gibbs energy from ethanol to water (Tanford, 1962). Open symbols (
and 
) represent mutant proteins with aromatic amino acids.The slope of the correlation for the mutant proteins at position 56 was steeper than that at position 59. In the case of mutant proteins at position 59, the hydrogen bond(s) with newly introduced water(s) and/or hydroxyl groups of substituted residues might suppress the decrease in stability due to a decrease in hydrophobicity. It has been reported that a water molecule in a cavity created in the interior of a protein molecule contributes favorably to stability (Takano et al., 1997b
). On the other hand, at position 56 the increase in unfilled cavity volume created by substitutions might participate in the decrease in stability in addition to the decrease in hydrophobicity. Hence these correlations showed different slopes depending on their mutation sites, but the stability of a protein would basically originate from a combination of various kinds of forces involved in each unique three-dimensional structure. To estimate synthetically the protein stability data at different positions, we attempted to relate protein stabilities to physical properties observed from mutant structures, using the present and previous information on structure and stability of mutant human lysozymes substituted in different environments (Takano et al., 1995, 1997a,b, 1998, 1999a; Funahashi et al., 1996; Yamagata et al., 1998). Magnitude of the contribution of the hydrophobic effect to protein stability
The hydrophobic effect is one of the most important stabilizing forces of a folded structure (Kauzmann, 1959). As described before, studies using a series of single amino acid substitutions (Yutani et al., 1987; Matsumura et al., 1988a) have shown that changes in the transfer Gibbs energies of residues substituted in the interior of proteins correlate with the changes in the stability of proteins (G). On the other hand, experimental studies on model compounds have shown that the transfer Gibbs energies of small hydrocarbon side chains or small polar groups from water to hydrophobic solvents relate linearly to the accessible surface area (ASA) (Chothia, 1974, 1976; Richards, 1977; Spolar et al., 1992; Makhatadze and Privalov, 1993, 1995; Privalov and Makhatadze, 1993; Oobatake and Ooi, 1993). Then, the change in G due to hydrophobic effect between the wild-type and mutant proteins (GHP) can be expressed as follows:
|
|
ASAnon-polar and ASApolar represent the difference in ASA of non-polar and polar atoms of all residues in a protein, respectively, upon denaturation between the wild-type and mutant proteins.
Mutant human lysozymes, for which it can be assumed that other factors, except for the hydrophobic effect, do not contribute to changes in protein stability due to substitutions, such as no introduction of new water molecules or hydrogen bonds, judged from changes in X-ray structures of mutant proteins, were chosen as type A (I56A, I56V, I56L, I56M, I59L, I59M). For estimating the contributions from factors stabilizing the conformation of a protein, the data for structure and stability of a series of mutant human lysozymes, three Ile Val (Takano et al., 1995), three Val Ala (Takano et al., 1997a), two Ile Ala (Takano et al., 1997b), six 3ss (3 Ile Val; 3 Val Ala mutants lacking an S–S bond between Cys77 and Cys95) (Takano et al., 1998), a Tyr Phe (Yamagata et al., 1998), a Ser Ala (Takano et al., 1999a), a Thr Val and a Thr Ala (Takano et al., 1999c) mutants, were used as type A. For type A mutants (total 24 mutants in addition to the present data), G was calculated as follows:
|
|
Gconf is considered as only the contribution due to changes in conformational entropy of the substituted residue. Pickett and Sternberg (1993) defined conformational entropy for amino acid side chains as the Boltzmann sampling over all states, that is, S = –R
pilnpi; pi has been estimated from the observed distribution of exposed side chain rotamers in 50 non-homologous protein crystal structures (Pickett and Sternberg, 1993; Doig and Sternberg, 1995). The entropy change for protein folding can be written as S = Sd + Sconf + Sother, where Sd and Sconf denote the desolvation entropy and the conformational entropy, respectively (Vajda et al., 1994). These two terms contribute more than 95% to the total entropy of unfolding (Murphy and Freire, 1992), with the remaining small term, Sother, due to some vibrational effects and others (Oobatake and Ooi, 1993; Xie and Freire, 1994). In this study, Sd is contained in GHP and Sother is neglected. Then, it was assumed that Gconf = –TSconf and that Sconf corresponds to the difference in conformational entropy between residues as reported by Doig and Sternberg (1995).
Gconf (at 65°C) and ASA for each mutant protein in Equation 5 were evaluated from the value reported by Doig and Sternberg (1995) and calculated as the difference between the surface area in the native state, obtained from the X-ray structure of each mutant protein, and that in the unfolded state, modeled as an extended three-residue peptide including the mutation site in the middle (Oobatake and Ooi, 1993; Takano et al., 1997a), respectively. For the calculation of ASA, C/S atoms in residues assigned to ASAnon-polar and N/O atoms to ASApolar were used. A least-squares fit of the GHP value to the (G – Gconf) value in Equation 5 for type A mutants gave = 0.178 and ß = –0.013 kJ mol–1 Å–2 (Figure 4a). The correlation coefficient (R) and standard deviation (SD) were 0.85 and 2.28, respectively. Contributions of these factors to the stability of type A mutants are shown in Tables IV and V.
|
|
|
Eisenberg and McLachlan (1986) have estimated the contribution of each protein atom to the solvation Gibbs energy as the product of the accessibility of the atom to the solvent and its atomic solvation parameter (ASP). This scale is based on the experimentally determined values for the hydrophobicity of the buried residues of Fauchere and Pliska (1983). Eisenberg and McLachlan (1986) proposed the following equation for estimating the contribution of solvation to the Gibbs energy of protein folding:
 |
i and (ASA)i are an atom type (C, N/O, O–, N+ or S), ASP value of atom i and difference in ASA values of atom i between folded and denatured states, respectively. Coefficients of and ß in this paper are equivalent to C and N/O, respectively. Table VI shows ASP values obtained by Eisenberg and McLachlan (1986), Wesson and Eisenberg (1992), Pickett and Sternberg (1993) and Vajda et al. (1994), as compared with the present results ( and ß). Sharp et al. (1991) revised experimental hydrophobicity values derived from transfer experiments, considering the different volumes of the solute and the solvent. Pickett and Sternberg (1993) recalculated ASA values using the revised hydrophobic values (Sharp et al., 1991).
|
There is considerable difference in ASP values reported for transfer Gibbs energies from vacuum to water (v/w) and from octanol to water (o/w) (Table VI
). The contribution of polar surfaces (N/O), derived using the v/w transfer Gibbs energy, is larger than that derived using the o/w transfer Gibbs energy, resulting in the possibility that the contribution of hydration to protein stability might be overestimated when using the v/w transfer Gibbs energy. The average value (–0.001) of N/O using the o/w transfer Gibbs energy in Table VI was comparable to the present result (ß = –0.013), suggesting that the hydration of nitrogen and neutral oxygen (N/O) scarcely contributes to protein stability. On the other hand, the contribution of the non-polar surface (C) is positive, but the values are different from each other. The present result ( = 0.178) is the highest among the values listed in Table VI. A reason for this may be that Equation 5 contains the contribution due to entropic effects (Gconf). If Equation 5 is used without the term Gconf for the fitting, the estimate for is 0.131 (R = 0.83). In a previous paper, the linear correlation between G obtained experimentally and ASAHP (the changes in surface area of hydrophobic residues exposed upon denaturation) was found to have a slope of 0.12 kJ mol–1 Å–2 for hydrophobic mutant human lysozymes (Takano et al., 1997b). Effects of hydrogen bonding and buried water molecule on protein stability
The mutant human lysozymes, except for type A, showed several structural changes, where new water molecules were introduced and/or new hydrogen bonds formed. The contribution of these structural changes to protein stability could be estimated using the parameters ( and ß) of the hydrophobic effect obtained in the present study and the information on stability–structure data for these mutants.
In order to evaluate the effect of forming or removing a hydrogen bond (GHB), the contribution of a hydrogen bond to protein stability was considered as 
(kJ Å/mol); is normalized by the distance, rHB, because the magnitude of the electrostatic interaction between two heavy atoms that are forming a hydrogen bond with each other is proportional to the reciprocal of the distance between the two atoms:
 |
(kJ/mol):
 |
NH2O is the change in the number of water molecules due to substitution. In the case of I59G, I59A, I59V, I59S, I59T and I106A, NH2O was 3, 3, 1, 2, 1 and 1, respectively. This term does not include the effect of hydrogen bonds formed by water molecules.
Mutant human lysozymes, for which there are contributions to GHB or GH2O from the formation and removal of hydrogen bonds or introduction and removal of water molecules judged from X-ray structures of mutant proteins, were chosen as type B (I56T, I59G, I59A, I59V, I59S, I59T). Similarly, additional mutant human lysozymes, an Ile Ala mutant (Takano et al., 1997b) and five Tyr Phe mutants (Yamagata et al., 1998), were chosen as type B. For type B mutants (total 12 mutants), G could be described by the equation
|
|
GHB + GH2O value to the G –Gconf –GHP value in Equation 9 gave = 15.53 kJ Å/mol and = –7.79 kJ/mol (Figure 4b). The correlation coefficient (R) and standard deviation (SD) were 0.97 and 1.37, respectively. Contributions of these factors to the stability of type B mutants are shown in Tables IV and V.
This result suggests that, if a hydrogen bond with a length of 3.0 Å is removed by substitution, the mutant protein should be destabilized by 5.1 kJ/mol. Recently, GHB for the intramolecular hydrogen bond has been estimated to be 7.5 kJ/mol, using a mutant (Tyr Phe) human lysozyme most appropriate for the evaluation (Yamagata et al., 1998). This value is larger than the present estimate, in which all hydrogen bonds are taken into account, including those formed by water molecules. GHB values for hydrogen bonds with a water molecule and with inter-residues might be different, although they were not distinguishable in the present study. These values are comparable to the values estimated by other investigators. Myers et al. (1997), for example, concluded that hydrogen bonds stabilize proteins and the average net stabilization is 4.2– 8.4 kJ/mol (1–2 kcal/mol).
The value of means that when one water molecule is newly introduced into the interior of a mutant protein, the protein is destabilized by 7.8 kJ/mol due to the entropic effect. Dunitz (1994) estimated that the decrease in entropy caused by the transfer of a water molecule from the solvent to the interior of a protein corresponds to a maximum Gibbs energy cost of 10 kJ/mol at 65°C. However, if the introduced water molecule forms two hydrogen bonds in the interior of a protein, the entropic loss might be compensated by GHB. Further, it has been reported that a water molecule in a cavity created in the interior of a protein contributes favorably to its stability (Takano et al., 1997b).
Contribution of amino acid substitutions at two different positions to the stability of a series of mutant proteins
As described, mutagenesis studies have shown that changes in the transfer Gibbs energies of residues substituted in the interior of proteins correlate with the changes in the stability of proteins, but their slopes differ from each other depending on the location of the site (Yutani et al., 1987; Matsumura et al., 1988a,b; Kellis et al., 1989; Sandberg and Terwilliger, 1991; Akasako et al., 1997). Furthermore, the effects of hydrogen bonds on protein stability also depend on substitution sites (Myers et al., 1997; Yamagata et al., 1998; Takano et al., 1999a). Thus, each site has peculiarities, which are rationalized on a per site basis. Similarly, the mutant proteins substituted at different positions (Ile56 and Ile59) exhibited different responses to the substitutions and showed different correlations between G and Gtr (Figure 4). We then divided G into each component of the stabilizing factors at each atom, considering structural changes of mutant proteins, as described, to estimate synthetically the stabilities of mutant proteins. To test whether the stabilities of the mutants at positions 56 and 59 could be estimated synthetically, their stabilities were calculated using Equation 10. Each parameter obtained above was used.
|
|
, showing the correlation between Gexperiment and Gcalculation, suggests that the stability of mutant human lysozymes substituted at different positions (Ile56 and Ile59) could be estimated synthetically using Equation 10.
|
Other effects on protein stability
Factors contributing to conformational stability were analyzed in connection with the structural changes, using G values of each mutant. The errors in Gexperiment from experiments are mainly caused by experimental errors in temperature (Td). If the experimental error of our calorimeter (DASM4) is maximally ±0.5°C, this would correspond maximally to ±0.7 kJ/mol as an error in G. To confirm the reliability of the fitting of the parameters, attempts were made to obtain coefficients for fitting by various means. First, the four coefficients , ß, and were determined in two steps; in the first step, (0.178) and ß (–0.013) were determined, and in the second step, (15.53) and (–7.79) were determined using the values of and ß. Second, the four coefficients were determined at the same time using Equation 11 for 36 mutant proteins:
|
|
, ß, and were calculated as 0.173, –0.015, 15.33 and –7.84, respectively. The correlation coefficient, R, was 0.92. These values were very close to those obtained from the two steps. This result suggests that Gibbs energy components are additive and their parameters are reliable.
Considering all of these facts, the deviations (2.0 kJ/mol) between Gexperiment and Gcalculation in Figures 4 and 5 are larger than the experimental errors (maximally 0.7 kJ/mol). This suggests that the difference between them corresponds to the contribution of other factors to protein stability, which are not considered in the present calculation, such as structural changes in the denatured structures, changes in steric hindrance, changes in cavity volume, packing density and electrostatic interaction. Table IV shows the contributions of various factors calculated using the values of , ß, and for a series of mutants at positions 56 and 59. The data for type C mutant proteins in Table IV were not used in determining the coefficients because the effect of steric hindrance, due to the substitution (I56F, I59F, I59Y), could not be estimated. The volumes of Phe and Tyr residues are much larger than that of the Ile residue. The difference between Gexperiment and Gcalculation of type C (I56F, I59F and I59Y) was 8~18 kJ/mol and larger than those of the other type of mutant proteins (maximally 4.2 kJ/mol).
There are numerous estimates of contributions of different factors to protein stability, e.g. hydrophobic effect (Shortle et al., 1990; Eriksson et al., 1992; Pace, 1992), hydrogen bonding (Byrne et al., 1995), cavity volume (Matthews, 1996; Xu et al., 1998), packing density (Otzen et al., 1995) and -helix propensity (Blaber et al., 1994). In this work, we have attempted to estimate directly the magnitude of some factors which play major roles in determining protein stability from experimental results without the utilization of data from model compounds. However, other factors remain to be estimated. The present study suggests that we could estimate other factors quantitatively by examining relationships between stability and structural changes using systematic amino acid substitutions that might affect a particular factor.
|
Acknowledgments |
|---|
We thank Takeda Chemical Industries (Osaka) for providing plasmid pGEL125. This work was supported in part by Fellowships from the Japan Society for the Promotion of Science for Young Scientists (J.F. and K.T.), by the Sakabe project of TARA, University of Tsukuba (Y.Y.), by a grant-in-aid for special project research from the Ministry of Education, Science and Culture (Y.Y. and K.Y.) and by the Japan Space Utilization Promotion Center (K.Y.).
|
Notes |
|---|
3 To whom correspondence should be addressed.E-mail: yutani{at}protein.osaka-u.ac.jp
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Akasako,A., Haruki,M., Oobatake,M. and Kanaya,S. (1997) J. Biol. Chem., 272, 18686–18693.
Blaber,M., Zhang,X.-J., Lindstrom,J.D., Pepiot,S.D., Baase,W.A. and Matthews,B.W. (1994) J. Mol. Biol., 235, 600–624.[Web of Science][Medline]
Brunger,A.T. (1992) X-PLOR Manual, Version 3.1. Yale University, New Haven, CT.
Buckle,A.M., Henrick,K. and Fersht,A.R. (1993) J. Mol. Biol., 234, 847–860.[Web of Science][Medline]
Byrne,P.M., Manuel,R.L., Lowe,L.G. and Stites,W.E. (1995) Biochemistry, 34, 13949–13960.[Medline]
Chothia,C. (1974) Nature, 248, 338–339.[Medline]
Chothia,C. (1976) J. Mol. Biol., 105, 1–14.[Web of Science][Medline]
Dill,K.A. (1990) Biochemistry, 29, 7133–7155.[Medline]
Doig,A.J. and Sternberg,M.J.E. (1995) Protein Sci., 4, 2247–2251.[Web of Science][Medline]
Dunitz,J.D. (1994) Science, 264, 670.
Eisenberg,D. and McLachlan,A.D. (1986) Nature, 319, 199–203.[Medline]
Eriksson,A.E., Baase,W.A., Zhang,X.-J., Heinz,D.W., Blaber,M., Baldwin,E.P. and Matthews,B.W. (1992) Science, 225, 178–183.
Eriksson,A.E., Baase,W.A. and Matthews,B.W. (1993) J. Mol. Biol., 229, 747–769.[Web of Science][Medline]
Fauchere,J.-L. and Pliska,V. (1983) Eur. J. Med. Chem., 18, 369–375.[Web of Science]
Funahashi,J., Takano,K., Ogasahara,K., Yamagata,Y. and Yutani,K. (1996) J. Biochem., 120, 1216–1223.
Kauzmann,W. (1959) Adv. Protein Chem., 14, 1–63.[Web of Science][Medline]
Kellis,J.T., Nyberg,K. and Fersht,A.R. (1989) Biochemistry, 28, 4914–4922.[Medline]
Makhatadze,G.I. and Privalov,P.L. (1993) J. Mol. Biol., 232, 639–659.[Web of Science][Medline]
Makhatadze,G.I. and Privalov,P.L. (1995) Adv. Protein Chem., 47, 307–429.[Web of Science][Medline]
Matsumura,M., Becktel,W.J. and Matthews,B.W. (1988a) Nature, 334, 406–410.[Medline]
Matsumura,M., Yahanda,S., Yutani,K. and Aiba,S. (1988b) Eur. J. Biochem., 171, 715–720.[Web of Science][Medline]
Matthews,B.W. (1996) FASEB J., 10, 35–41.[Abstract]
Murphy,K.P. and Freire,E. (1992) Adv. Protein Chem., 43, 313–361.[Web of Science][Medline]
Myers,J.K., Pace,C.N. and Scholtz,J.M. (1997) Proc. Natl Acad. Sci. USA, 94, 2833–2837.
Navaza,J. (1994) Acta Crystallogr., A50, 157–163.
Oobatake,M. and Ooi,T. (1993) Prog. Biophys. Mol. Biol., 59, 237–284.[Web of Science][Medline]
Osserman,E.F., Cole,S.J., Swan,I.D.A. and Blake,C.C.F. (1969) J. Mol. Biol., 46, 211–212.[Web of Science][Medline]
Otwinowski,Z. (1990) DENZO Data Processing Package. Yale University, New Haven, CT.
Otzen,D.E., Rheinnecker,M. and Fersht,A.R. (1995) Biochemistry, 34, 13051–13058.[Medline]
Pace,C.N. (1992) J. Mol. Biol., 226, 29–35.[Web of Science][Medline]
Pace,C.N., Shirley,B.A., Mcnutt,M. and Gajiwala,K. (1996) FASEB J., 10, 75–83.[Abstract]
Parry,R.M., Chandan,R.C. and Shahani,K.M. (1969) Arch. Biochem. Biophys., 130, 59–65.[Web of Science][Medline]
Pickett,S.D. and Sternberg,M.J. (1993) J. Mol. Biol., 231, 825–839.[Web of Science][Medline]
Privalov,P.L. and Gill,S.J. (1988) Adv. Protein Chem., 39, 191–234.[Web of Science][Medline]
Privalov,P.L. and Khechinashvili,N.N. (1974) J. Mol. Biol., 86, 665–684.[Web of Science][Medline]
Privalov,P.L. and Makhatadze,G.I. (1993) J. Mol. Biol., 232, 660–679.[Web of Science][Medline]
Richards,F.M. (1977) Annu. Rev. Biophys. Bioengng, 6, 151–176.[Web of Science][Medline]
Sakabe,N. (1991) Nucl. Instrum. Methods Phys. Res., A303, 448–463.
Sandberg,W.S. and Terwilliger,T.C. (1991) Proc. Natl Acad. Sci. USA, 88, 1706–1710.
Serrano,L., Kellis,J.T., Cann,P., Matouschek,A. and Fersht,A.R. (1992) J. Mol. Biol., 224, 783–804.[Web of Science][Medline]
Sharp,K.A., Nicholls,A., Friedman,R. and Honig,B. (1991) Biochemistry, 30, 9686–9697.[Medline]
Shortle,D., Stites,W.E. and Meeker,A.K. (1990) Biochemistry, 29, 8033–8041.[Medline]
Spolar,R.S., Livingstone,J.R. and Record,M.T., Jr (1992) Biochemistry, 31, 3947–3955.[Medline]
Takano,K., Ogasahara,K., Kaneda,H., Yamagata,Y., Fujii,S., Kanaya,E., Kikuchi,M., Oobatake,M. and Yutani,K. (1995) J. Mol. Biol., 254, 62–76.[Web of Science][Medline]
Takano,K., Yamagata,Y., Fujii,S. and Yutani,K. (1997a) Biochemistry, 36, 688–698.[Medline]
Takano,K., Funahashi,J., Yamagata,Y., Fujii,S. and Yutani,K. (1997b) J. Mol. Biol., 274, 132–142.[Web of Science][Medline]
Takano,K., Yamagata,Y. and Yutani,K. (1998) J. Mol. Biol., 280, 749–761.[Web of Science][Medline]
Takano,K., Yamagata,Y., Kubota,M., Funahashi,J., Fujii,S. and Yutani,K., (1999a) Biochemistry, 38, 6623–6629.[Medline]
Takano,K., Ota,M., Ogasahara,K., Yamagata,Y., Nishikawa,K. and Yutani,K. (1999b) Protein Engng, 12, 663–672.
Takano,K., Yamagata,Y., Funahashi,J., Hioki,Y., Kuramitsu,S. and Yutani,K. (1999c) Biochemistry, 38, in press.
Tanford,C. (1962) J. Am. Chem. Soc., 84, 4240–4247.[Web of Science]
Vajda,S., Weng,Z., Rosenfeld,R. and DeLisi,C. (1994) Biochemistry, 33, 13977–13988.[Medline]
Wesson,L. and Eisenberg,D. (1992) Protein Sci., 1, 227–235.[Web of Science][Medline]
Wetlaufer,D.B. (1962) Adv. Protein Chem., 17, 303–390.[Web of Science]
Xie,D. and Freire,E. (1994) Proteins, 19, 291–301.[Web of Science][Medline]
Xu,J., Baase,W.A., Baldwin,E. and Matthews,B.W. (1998) Protein Sci., 7, 158–177.[Web of Science][Medline]
Yamagata,Y., Kubota,M., Sumikawa,Y., Funahashi,J., Takano,K., Fujii,S. and Yutani,K. (1998) Biochemistry, 37, 9355–9362.[Medline]
Yutani,K., Ogasahara,K., Tsujita,T. and Sugino,Y. (1987) Proc. Natl Acad. Sci. USA, 84, 4441–4444.
Received May 25, 1999; revised June 29, 1999; accepted July 12, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Y. Asada, M. Sawano, K. Ogasahara, J. Nakamura, M. Ota, C. Kuroishi, M. Sugahara, K. Yutani, and N. Kunishima Stabilization Mechanism of the Tryptophan Synthase {alpha}-Subunit from Thermus thermophilus HB8: X-Ray Crystallographic Analysis and Calorimetry J. Biochem., October 1, 2005; 138(4): 343 - 353. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Takano, M. Saito, M. Morikawa, and S. Kanaya Mutational and Structural-Based Analyses of the Osmolyte Effect on Protein Stability J. Biochem., June 1, 2004; 135(6): 701 - 708. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Funahashi, Y. Sugita, A. Kitao, and K. Yutani How can free energy component analysis explain the difference in protein stability caused by amino acid substitutions? Effect of three hydrophobic mutations at the 56th residue on the stability of human lysozyme Protein Eng. Des. Sel., September 1, 2003; 16(9): 665 - 671. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Takano, Y. Yamagata, and K. Yutani Buried water molecules contribute to the conformational stability of a protein Protein Eng. Des. Sel., January 1, 2003; 16(1): 5 - 9. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Funahashi, K. Takano, Y. Yamagata, and K. Yutani Positive Contribution of Hydration Structure on the Surface of Human Lysozyme to the Conformational Stability J. Biol. Chem., June 7, 2002; 277(24): 21792 - 21800. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Takano and K. Yutani A new scale for side-chain contribution to protein stability based on the empirical stability analysis of mutant proteins Protein Eng. Des. Sel., August 1, 2001; 14(8): 525 - 528. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Ota, Y. Isogai, and K. Nishikawa Knowledge-based potential defined for a rotamer library to design protein sequences Protein Eng. Des. Sel., August 1, 2001; 14(8): 557 - 564. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Ohmura, T. Ueda, Y. Hashimoto, and T. Imoto Tolerance of point substitution of methionine for isoleucine in hen egg white lysozyme Protein Eng. Des. Sel., June 1, 2001; 14(6): 421 - 425. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Funahashi, K. Takano, and K. Yutani Are the parameters of various stabilization factors estimated from mutant human lysozymes compatible with other proteins? Protein Eng. Des. Sel., February 1, 2001; 14(2): 127 - 134. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Goda, K. Takano, Y. Yamagata, Y. Katakura, and K. Yutani Effect of extra N-terminal residues on the stability and folding of human lysozyme expressed in Pichia pastoris Protein Eng. Des. Sel., April 1, 2000; 13(4): 299 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Yamagata, K. Ogasahara, Y. Hioki, S. J. Lee, A. Nakagawa, H. Nakamura, M. Ishida, S. Kuramitsu, and K. Yutani Entropic Stabilization of the Tryptophan Synthase alpha -Subunit from a Hyperthermophile, Pyrococcus furiosus. X-RAY ANALYSIS AND CALORIMETRY J. Biol. Chem., March 30, 2001; 276(14): 11062 - 11071. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||