Skip Navigation


PEDS Advance Access originally published online on May 23, 2006
Protein Engineering Design and Selection 2006 19(8):355-358; doi:10.1093/protein/gzl019
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
19/8/355    most recent
gzl019v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Wunderlich, M.
Right arrow Articles by Schmid, F. X.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wunderlich, M.
Right arrow Articles by Schmid, F. X.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

The correlation between protein stability and dipole moment: a critical test

Michael Wunderlich and Franz X. Schmid1

Laboratorium für Biochemie und Bayreuther Zentrum für Molekulare Biowissenschaften, Universität Bayreuth D-95440 Bayreuth, Germany

1To whom correspondence should be addressed. E-mail: fx.schmid{at}uni-bayreuth.de


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Improving the stability of proteins is a major aim in basic and applied protein science. Querol and coworkers calculated changes in the quasi-electric dipole moment of a protein and used it as a simple criterion to predict stabilizing charge mutations. They employed this method to propose for the bacterial cold shock protein Bc-Csp a number of charge mutations that should have a strong influence on stability. We produced eight variants of Bc-Csp with such mutations and measured their stabilities experimentally. However, we could not find a correlation between the stability and the quasi dipole moment of these variants. Possibly, the quasi dipole moment reflects only a secondary aspect of the changes that are caused by charge mutations in a protein.

Keywords: cold shock protein/electric dipole moment/electrostatic interactions/protein stabilization


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Robust proteins are important in almost all areas of biochemistry and biotechnology. Accordingly, much experimental and theoretical work concentrates on understanding and improving protein stability. Comparisons between proteins from thermophilic organisms (‘thermophilic’ proteins) and their homologs from mesophilic organisms (‘mesophilic' proteins) suggested that the number and the distribution of the charged residues are important determinants for the thermostability of a protein (Jaenicke and Böhm, 1998Go; Karshikoff and Ladenstein, 2001Go). Mutational analyses of several small proteins showed that charge replacements at the protein surface can indeed strongly change protein stability (Yano and Poulos, 2003Go).

Several theoretical approaches have been used to analyze and optimize the contributions of charged residues to protein stability (Kundrotas and Karshikoff, 2002Go; Marshall et al., 2002Go; Luisi et al., 2003Go). Mozo-Villarias et al. (2003)Go proposed a particularly simple and straightforward electrostatic criterion for predicting the changes in protein stability caused by the mutation of charged residues. Their criterion is the quasi dipole moment P between the centers of the negative and the positive charge of the protein. P is thus a measure for the asymmetry in the spatial distribution of the charges. A decrease in the P-value (e.g. upon mutation) should lead to an increase in stability, an increase in P to a decrease. P is an attractive stability parameter because it can easily be calculated for protein mutants.

Mozo-Villarias et al. (2003)Go calculated P-values for homologous pairs of mesophilic and thermophilic proteins as well as for wild-type proteins and the corresponding charge mutants. Overall they found an inverse correlation between this dipole moment and stability. In addition, they used the P criterion to predict the effects of charge changes on the stability of the cold shock proteins from the mesophile Bacillus subtilis (Bs-CspB) and from the thermophile Bacillus caldolyticus (Bc-Csp).

The bacterial Csp molecules are small monomeric proteins. They consist of 65–70 residues, which form a five-stranded antiparallel ß-barrel structure (Figure 1) (Schindelin et al., 1993Go). The mesophilic Bs-CspB and the thermophilic Bc-Csp have often been used as models for analyzing the contributions of charged residues to protein stability in both experimental and theoretical studies (Perl et al., 2000Go; Martin et al., 2001Go; Perl and Schmid, 2001Go; Sanchez-Ruiz and Makhatadze, 2001Go; Dominy et al., 2002Go; Martin et al., 2002Go; Torrez et al., 2003Go; Zhou and Dong, 2003Go; Garofoli et al., 2004Go; Wunderlich et al., 2005Go). Mozo-Villarias et al. (2003)Go noted a strong correlation between the stability and the quasi dipole moment P for Bs-CspB and Bc-Csp. The decrease in the P-value from 9.7 D for the mesophilic Bs-CspB to 6.6 for the thermophilic Bc-Csp is accompanied by an increase in the midpoint of thermal unfolding from 53.8 to 76.9°C (Table I) and a corresponding increase in stability of 16.5 kJ/mol.


Figure 1
View larger version (30K):
[in this window]
[in a new window]
  		Fig. 1. Ribbon drawing of the tertiary structure of wild-type Bc-Csp. The side-chains of the residues that were investigated are shown in the ball-and-stick representation. The mutations predicted to be stabilizing and destabilizing are marked by arrows pointing up and down, respectively. The Figure was prepared using the PDB file 1c9o (Schindelin et al., 1993Go) and MOLMOL (Koradi et al., 1996Go).

 

View this table:
[in this window]
[in a new window]
  		Table I. Stability data for wild-type Bs-CspB, wild-type Bc-Csp and the variants of Bc-Csp with increased or decreased P-values

 
For both cold shock proteins Mozo-Villarias et al. (2003)Go calculated the changes in the P-values that occur when, at every position of the protein chain, the wild-type residue was mutated to a negatively or to a positively charged amino acid. Charged residues were mutated to an oppositely charged or an uncharged residue.

In the present study we investigated experimentally whether the quasi dipole moment P is a good predictor of protein stability, i.e. whether there is an inverse correlation between protein stability and the quasi dipole moment P. For Bc-Csp we produced the four variants that showed the strongest decrease in the calculated P-value as well as the four variants with the strongest increase in P and determined their thermodynamic stability. Although they span a very wide range of P-values (between 2.0 and 12.5 D), we could not find a significant correlation between P and the conformational stability. These results for Bc-Csp raise doubts whether the quasi dipole moment is a robust predictor for changes in protein stability upon mutation of charged residues.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Acknowledgements
 References
 
Protein expression and purification

The gene for Bc-Csp was amplified by PCR with primers including restriction sites for NdeI at the N terminus and BamHI at the C-terminus. It was cloned into the plasmid pET11a (Novagen, Madison, Wisconsin, USA). The mutations were introduced by amplifying the whole plasmid by PCR with primers coding for the mutations, blunt end ligation of the PCR product and transformation into Escherichia coli XL1Blue. The mutant form of the plasmid was isolated and transformed into E. coli BL21(DE3)pLysS (Stratagene, La Jolla, USA). The proteins were expressed and purified as described previously (Schindelin et al., 1992Go; Mueller et al., 2000Go), with minor modifications.

Stability measurements by CD spectroscopy and differential scanning calorimetry

Thermal unfolding transitions were measured in 0.1 M Na cacodylate/HCl, pH 7.0, at protein concentrations of 4.0 µM in the presence of 0 M or 2.0 M NaCl. The CD signal at 222.6 nm at 1 cm path length was used to monitor the transitions at heating rates of 60°C/h. Transitions were evaluated using a nonlinear least-squares fit according to a two-state model (Schindler et al., 1995Go) with a fixed heat capacity change {Delta}Cp of 4 kJ/mol/K as described previously (Perl et al., 2000Go). Thermal unfolding was reversible and identical transitions were obtained upon a second heating, provided that the protein was not kept in the thermally unfolded state for more than 10 min. The Gibbs free energies of unfolding ({Delta}GD) were calculated from the stability curves (Figure 2) at 75°C. This temperature is within the transition region for all variants of Bc-Csp and therefore the equilibrium constant could be determined with high precision.


Figure 2
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 2. Thermal unfolding transitions measured by circular dichroism at 222.6 nm of wild-type BcBs-CspB (filled triangle) and Bc-Csp (filled circle) and the variants of Bc-Csp E36K (1), E50K (2), E12K (3), E21K (4), N55K (5) and R56E (6). All transitions were measured in 0.1 M sodium cacodylate-HCl (pH 7.0) The fractions of native protein as obtained after two-state analysis of the data are shown as a function of temperature. The corresponding thermodynamic parameters are given in Table I.

 
For wild-type Bc-Csp and the variants H29E and E46K the unfolding was followed by differential scanning calorimetry with a VP-DSC instrument (MicroCal, Northampton, MA, USA) at protein concentrations between 20 and 150 µM in 0.1 M sodium cacodylate/HCl (pH 7.0) and 0 M or 2.0 M NaCl with a scan rate of 1.5 K/min (cell volume 0.523 ml). The measured excess molar heat capacity Formula (T) was analyzed by a Levenberg–Marquardt nonlinear least-squares method according to a non-two-state model after correction for a progressive baseline (assumption of zero {Delta}Cp). For the analysis the origin software provided by MicroCal was used.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Acknowledgements
 References
 
The mutants that caused the strongest changes in P were constructed for the thermophilic protein Bc-Csp (TM = 76.9°C). For Bc-Csp a wide temperature range is available to detect and characterize the effects of the expected strong variations in the TM values. For the replacements we chose Glu when a negatively charged side chain was required and Lys to change to a positively charged residue.

The E12K, E36K, E50K and N55K mutations led to the strongest increases in the calculated quasi dipole moment and are thus predicted to be strongly destabilizing. Their P-values range between 10.1 and 12.5 D (Mozo-Villarias et al., 2003Go) and are thus higher than the P-value of the mesophilic protein Bs-CspB (Table I). The E21K, H29E, E46K and R56E variants at the other end of the scale showed the lowest P-values, between 4.9 and 2.0 D, and thus they are expected to be strongly stabilized. All mutated residues are at the surface of the protein and well exposed to the solvent (Figure 1).

The protein variants were overexpressed and purified by anion exchange chromatography, hydrophobic interaction chromatography and gel filtration. Six of the eight variants eluted as monomers from the gel filtration column at the same elution volume as wild-type Bc-Csp. The elution volumes of the variants H29E and E46K were slightly smaller. The thermal transitions were followed by circular dichroism at 222.6 nm. The normalized transitions are shown in Figure 2. The CD change accompanying the unfolding of the variants H29E and E46K was small and therefore their thermal unfolding transitions were followed by differential scanning calorimetry.

Four variants of the thermophilic protein Bc-Csp showed P-values between 10.1 and 12.5 D, which are thus higher than the P-value of the mesophilic protein Bs-CspB (9.7 D). Nevertheless, their stabilities were almost unchanged, and their TM values ranged between 73.5 and 77.7°C (Table I), which is very close to the TM of wild-type Bc-Csp (76.9°C), but ~20° higher than the TM of Bs-CspB. In particular, the N55K mutation increased the conformational stability of Bc-Csp (Table I) although it raised the P-value from 6.6 to 10.1 D.

The four variants with the lowest P-values are predicted to be strongly stabilized, but their TM values were also almost unchanged. They all ranged between 73.5 and 80°C, close to the TM of the wild-type protein (Table I). Incidentally, the highest TM (80.0°C) was found for the variant with the lowest P-value (2.0). The variants H29E and E46K are probably dimeric. This is indicated by their calorimetric unfolding transitions. They are concentration-dependent, and the van't Hoff enthalpies of unfolding, {Delta}HD, are 443 and 477 kJ/mol, respectively, which is twice the {Delta}HD value of the monomeric variants (Table I). This suggests that for these two variants the unfolding of the monomers is coupled with dimer dissociation. For wild-type Bc-Csp the TM values and the {Delta}HD values from the CD and the calorimetric transitions were identical (Table I), supporting our conclusion that the variants H29E and E46K are dimeric. The stabilities of these two variants are thus overestimated, because they gain additional stability from dimer formation.

Figure 3a suggests that the thermal stability as expressed by the TM value is not correlated with the quasi dipole moment as expressed by the P-value. It should be noted that the P-values of the single mutants cover a broad range from 2.0 to 12.5 D. The changes in the Gibbs free energy {Delta}{Delta}GD also do not correlate with the P-values (Figure 3b). A correlation is seen only for the parent proteins, the thermophilic Bc-Csp and the mesophilic Bs-CspB. Here the increase in P from 6.6 to 9.7 D leads to a 16.5 kJ/mol loss in stability (Figure 3b).


Figure 3
View larger version (7K):
[in this window]
[in a new window]
  		Fig. 3. Correlation between the calculated quasi-electric dipole moment P and the experimentally determined (a) melting points (TM), (b) total changes of the Gibbs free energy of unfolding relative to the wild-type protein {Delta}{Delta}GD and (c) the coulombic part of the changes in Gibbs free energy measured as the difference between the values at 0 M NaCl and 2 M NaCl of wild-type Bs-CspB (filled square), wild-type Bc-Csp (filled circle), and the investigated variants, as measured by circular dichroism (open circles) or by differential scanning calorimetry (inverted triangle). The data are taken from Table I.

 
Potential changes in stability that are linked with the variation of the quasi dipole moment should originate from coulombic interactions. Such interactions are sensitive to screening by salt, and therefore, as in our previous work (Perl et al., 2000Go; Martin et al., 2001Go, 2002Go; Wunderlich et al., 2005Go), we measured the stabilities of the Bc-Csp variants at 0 M and at 2 M NaCl. The difference between the two values (Table I) reflects the coulombic contributions to the stability difference. It provides a lower estimate because in proteins not all coulombic interactions can be screened by 2 M NaCl (Lee et al., 2002Go; Luisi et al., 2003Go; Spencer et al., 2005Go). The screenable part of the coulombic contributions is also not correlated with the P-value (Figure 3c). It should also be noted that the expected inverse correlation between stability and net charge is poor as well. Intriguingly, the variant with the highest net charge (R56E) shows the highest thermal stability.


   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Acknowledgements
 References
 
Mozo-Villarias et al. (2003)Go proposed the quasi-electric dipole moment P as a simple electrostatic criterion to predict the influence of charge mutations on the stability of a protein. Mutations that increase P should be destabilizing and those that decrease P should be stabilizing. We used the cold shock protein Bc-Csp and tested this prediction by examining the effects of those charge mutations that led to the largest increases or the largest decreases in P. For this set of mutated proteins we could not find a correlation between the stability and the quasi dipole moment. All the mutations had negligible effects on the thermal stability of Bc-Csp, and the TM values clustered in a narrow range of ±3°C around the TM of the wild-type protein (Figure 3a), although the P-values varied in a wide range, between 2.0 and 12.5 D. Even at the qualitative level the P-value is a poor predictor. Three mutations that were predicted to be strongly stabilizing are in fact slightly destabilizing. These data suggest that the strong difference between the stabilities of the thermophilic Bc-Csp and the mesophilic Bs-CspB ({Delta}TM = 23.1 degrees) is not caused by the increase in P from 6.6 to 9.7 D. Rather, it originates predominantly from the local repulsion between Glu3 and Glu66 in Bs-CspB, which is absent in Bc-Csp (Perl et al., 2000Go). The importance of charge interactions for the stability of cold shock proteins has been shown in several experimental studies (Martin et al., 2001Go; Perl and Schmid, 2001Go; Sanchez-Ruiz and Makhatadze, 2001Go; Makhatadze et al., 2004Go; Wunderlich et al., 2005Go).

The positions with the strongest changes in P upon charge variation are all located at the protein periphery, mostly in protruding loops, remote from the center of the molecule (Figure 1). Therefore the quasi dipole moment changes strongly when they are mutated. The residues 12, 36, 55 and 56 are more than 60% exposed to the solvent, which are the highest values found in Bc-Csp. This probably explains why the mutations to other charged residues had only minor effects on the stability. These residues are mobile and are not involved in local charge clusters, and the dielectric constant in their (aqueous) environment is high. Therefore they contribute only marginally to the electrostatic interactions in the protein (Kumar and Nussinov, 2002Go; Lazaridis and Karplus, 2003Go). Residues 55 and 56 (Figure 1) provide a good example. They are located at the tip of a long exposed ß hairpin. The introduction of a positive charge by the N55K mutation thus increased the quasi dipole moment to 10.1 D, the positive-to-negative change by the R56E mutation decreased it to 2.0 D. Experimentally, both mutations are slightly stabilizing.

The approach of using quasi dipole moments for predicting stabilizing mutations is appealing. Its predictive power is probably limited, because those charged residues that are farthest away from the center of the protein dominate the calculated P-value and because the local electrostatic environment and changes in desolvation are not accounted for. For the cold shock proteins such calculations (Sanchez-Ruiz and Makhatadze, 2001Go; Dominy et al., 2002Go) led to a better coincidence with the experimental stability data. Overall, we could not find a significant correlation between the conformational stability and the quasi dipole moment P. It thus appears to be an inadequate criterion to predict stabilizing charge mutations.


   Footnotes
 
Edited by Dr Ammon Horovitz


   Acknowledgements
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
References
 
The authors thank the members of the laboratory for many discussions. This work was supported by grants from the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Dominy B.N., Perl D., Schmid F.X., Brooks C.L. III. (2002) J. Mol. Biol. 319:541–554.[CrossRef][Web of Science][Medline]

Garofoli S., Falconi M., Desideri A. (2004) J. Biomol. Struct. Dyn. 21:771–780.[Medline]

Jaenicke R. and Böhm G. (1998) Curr. Opin. Struct. Biol. 8:738–748.[CrossRef][Web of Science][Medline]

Karshikoff A. and Ladenstein R. (2001) Trends Biochem. Sci. 26:550–556.[CrossRef][Web of Science][Medline]

Koradi R., Billeter M., Wüthrich K. (1996) J. Mol. Graph. 14:29–32.

Kumar S. and Nussinov R. (2002) Chembiochem. 3:604–617.[CrossRef][Web of Science][Medline]

Kundrotas P.J. and Karshikoff A. (2002) Protein Sci. 11:1681–1686.[CrossRef][Web of Science][Medline]

Lazaridis T. and Karplus M. (2003) Biophys. Chem. 100:367–395.[CrossRef][Web of Science][Medline]

Lee K.K., Fitch C.A., Garcia-Moreno E.B. (2002) Protein Sci. 11:1004–1016.[CrossRef][Web of Science][Medline]

Luisi D.L., Snow C.D., Lin J.J., Hendsch Z.S., Tidor B., Raleigh D.P. (2003) Biochemistry. 42:7050–7060.[CrossRef][Medline]

Makhatadze G.I., Loladze V.V., Gribenko A.V., Lopez M.M. (2004) J. Mol. Biol. 336:929–942.[CrossRef][Web of Science][Medline]

Marshall S.A., Morgan C.S., Mayo S.L. (2002) J. Mol. Biol. 316:189–199.[CrossRef][Web of Science][Medline]

Martin A., Sieber V., Schmid F.X. (2001) J. Mol. Biol. 309:717–726.[CrossRef][Web of Science][Medline]

Martin A., Kather I., Schmid F.X. (2002) J. Mol. Biol. 318:1341–1349.[CrossRef][Medline]

Mozo-Villarias A., Cedano J., Querol E. (2003) Protein Eng. 16:279–286.[Abstract/Free Full Text]

Mueller U., Perl D., Schmid F.X., Heinemann U. (2000) J. Mol. Biol. 297:975–988.[CrossRef][Web of Science][Medline]

Perl D. and Schmid F.X. (2001) J. Mol. Biol. 313:343–357.[CrossRef][Web of Science][Medline]

Perl D., Mueller U., Heinemann U., Schmid F.X. (2000) Nat. Struct. Biol. 7:380–383.[CrossRef][Web of Science][Medline]

Sanchez-Ruiz J.M. and Makhatadze G.I. (2001) Trends Biotechnol. 19:132–135.[CrossRef][Web of Science][Medline]

Schindelin H., Herrler M., Willimsky G., Marahiel M.A., Heinemann U. (1992) Proteins 14:120–124.[CrossRef][Web of Science][Medline]

Schindelin H., Marahiel M.A., Heinemann U. (1993) Nature 364:164–168.[CrossRef][Medline]

Schindler T., Herrler M., Marahiel M.A., Schmid F.X. (1995) Nature Struct. Biol. 2:663–673.[CrossRef][Web of Science][Medline]

Spencer D.S., Xu K., Logan T.M., Zhou H.X. (2005) J. Mol. Biol. 351:219–232.[CrossRef][Web of Science][Medline]

Torrez M., Schultehenrich M., Livesay D.R. (2003) Biophys. J. 85:2845–2853.[Web of Science][Medline]

Wunderlich M., Martin A., Schmid F.X. (2005) J. Mol. Biol. 347:1063–1076.[CrossRef][Web of Science][Medline]

Yano J.K. and Poulos T.L. (2003) Curr. Opin. Biotechnol. 14:360–365.[CrossRef][Web of Science][Medline]

Zhou H.X. and Dong F. (2003) Biophys. J. 84:2216–2222.[Medline]

Received March 13, 2006; revised April 7, 2006; accepted April 12, 2006.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
19/8/355    most recent
gzl019v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Wunderlich, M.
Right arrow Articles by Schmid, F. X.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wunderlich, M.
Right arrow Articles by Schmid, F. X.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?