Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Proteins. Author manuscript; available in PMC 2011 May 1.
Published in final edited form as:
PMCID: PMC2853881

Thermodynamics of the Trp-cage Miniprotein Unfolding in Urea


The thermodynamic properties of unfolding of the Trp-cage mini protein in the presence of various concentrations of urea have been characterized using temperature-induced unfolding monitored by far-UV circular dichroism spectroscopy. Analysis of the data using a two-state model allowed the calculation of the Gibbs energy of unfolding at 25°C as a function of urea concentration. This in turn was analyzed by the linear extrapolation model that yielded the dependence of Gibbs energy on urea concentration, i.e. the m-value for Trp-cage unfolding. The m-value obtained from the experimental data, as well as the experimental heat capacity change upon unfolding, were correlated with the structural parameters derived from the three dimensional structure of Trp-cage. It is shown that the m-value can be predicted well using a transfer model, while the heat capacity changes are in very good agreement with the empirical models based on model compounds studies. These results provide direct evidence that Trp-cage, despite its small size, is an excellent model for studies of protein unfolding and provide thermodynamic data that can be used to compare with atomistic computer simulations.


The design of the Trp-cage mini protein (only 20 amino acid residues) by Andersen et al 1 provided the protein folding field with a convenient model to examine the thermodynamic and folding properties of a small protein with all the structural attributes of larger globular proteins 25, i.e. a well defined secondary and tertiary structure, and a well developed hydrophobic core (Figure 1A). The small size of the protein also made it an attractive model for computer simulations in atomistic details (see e.g. references 613). In this paper we provide experimental data on the effects of urea on the thermodynamic properties of the Trp-cage mini protein. These results have two important consequences. One, they provide the experimental data for the effects of urea on the smallest known protein and, thus, will allow us to test empirical predictions for the effects of urea on globular proteins, which were developed based on larger proteins. Second, they provide the thermodynamic information that can be used to calibrate and validate the computational modeling of the effects of urea on protein stability.

Figure 1
A) Ribbon diagram of the Trp-cage miniprotein representing the average of all NMR conformers of 1L2Y, as computed by PyMol 37. Several core amino acids are displayed as ball and stick, including Trp6 (light gray) and the amino acids that effectively cage ...

Materials and Methods

Sample Preparation

The Trp-cage peptide (amino acid sequence: 2HN-NLYIQWLKDGGPSSGRPPPS-COOH) was synthesized using standard Fmoc chemistry (Penn State College of Medicine Macromolecular Core Facility). The peptide was purified as previously described on a C-18 column 2 and purity was confirmed with MALDI-TOF. Circular Dichroism (CD) experiments were performed in one of two buffers: the first containing 5 mM sodium phosphate and 5 mM sodium citrate, pH 7.0, and the second containing 1 mM sodium phosphate, 1 mM sodium citrate, and 1 mM borate, pH 7.0. For urea denaturation, an 8 M urea stock solution was also made in the above buffers. Both buffers gave similar melting profiles. All peptide samples were dialyzed into MilliQ water and diluted to 10 μM in the above buffers, containing the following urea concentrations: 0 M, 0.5 M, 0.75 M, 1.0 M, 1.5 M, 2.0 M, 2.5 M, 3.0 M, 3.5M, 4.0 M, 6.0 M. Concentrations of urea were confirmed both before and after CD experiments by comparing measured values with tabulated data from the solution refractive index 14.

Thermal Unfolding of Trp-cage Monitored Using Far-UV Circular Dichroism (CD)

CD measurements were performed on a Jasco J-715 spectropolarimeter. All measurements were carried out either in a 1 cm rectangular quartz cell with temperature controlled by a six-position Peltier cell changer or in a 1 mm cylindrical cell with the temperature controlled by an external, circulating water bath. Melting profiles were monitored at 222 nm or 225 nm (to reduce the contribution of high concentrations of urea to the CD signal) and from 4 to 100°C, in 1°C increments. The actual temperature inside the cell was measured using an Omega HH24 internal temperature probe. Reversibility of thermal unfolding was confirmed by comparing far-UV CD spectra at 4°C both before and after thermal melting. Measured ellipticity values, Θ, were converted into mean residue ellipticity, [Θ], using:


where MR is the average mass per amino acid, l is the optical length of the cell in centimeters, and c is the concentration of the peptide in mg/ml. All experiments were performed a minimum of two times, with the average values presented.

Calculation of the Accessible Surface Area (ASA)

The accessible surface area was calculated using all 20 Trp-cage NMR structures (PDB code: 1L2Y 1, using NACCESS V2.1.1 15), and the average ASA are reported. The unfolded state was modeled in Swiss PDB Viewer v3.7b2 as an extended beta sheet (ϕ=− 120°, ψ=120°) and also processed using NACCESS. The difference between the accessible surface areas, ΔASA, was calculated using:


where ASAUnfolded and ASAfolded are the surface areas of the unfolded and native states, respectively.

Data Analysis

All CD unfolding data was fitted to a two-state unfolding model, where it is assumed that only the native (N) and unfolded (U) states are populated at any given temperature. For such a two-state model, the temperature dependence of enthalpy, ΔH, and entropy, ΔS, can be defined as:




where the changes in the heat capacity upon unfolding, ΔCp, are assumed to be temperature independent. The temperature-dependence of Gibbs free energy can be derived via Gibbs-Helmholtz relationship:


Direct DSC analysis of Trp-cage unfolding in aqueous solution (i.e. in the absence of urea) led to the following thermodynamic parameters: ΔH(Tm)=56±2 kJ·mol−1, Tm=43.9±0.8 °C and ΔCp=0.3±0.1 kJ·mol−1·K−1 2. During the fit to a two-state model, ΔH(Tm) and ΔCp were allowed to vary within the experimental uncertainties.

For urea concentrations greater than 3.5 M, the fraction of the native Trp-cage molecules was less than 50% even at the lowest experimental temperatures (4°C). Consequently, Tm was no longer a valid reference temperature and Ts, the temperature at which entropy is equal to zero, was used 16. For this case, the temperature dependence of ΔH and ΔS can be defined as:




where again the changes in the heat capacity upon unfolding, ΔCp, are assumed to be temperature independent. The temperature-dependence of Gibbs free energy can then be defined as:


It has been shown that both enthalpy and entropy of unfolding are dependent on urea concentration 17. In addition, the heat capacity of unfolding, ΔCp, has also been shown to be dependent on urea concentration 17,18. To account for these effects, ΔCp was expressed as the sum of two terms:


where the first term, ΔCp, refers to the heat capacity of unfolding in the absence of urea and the second term, δCp, is the increase in the heat capacity of unfolding due to the addition of urea. The value of δCp=0.013±0.006 kJ·mol−1·K−1 per mole of urea was obtained from the average dependence of the heat capacity on urea concentration for several proteins 17, and was constrained within the experimental uncertainty during the fit.

Using nonlinear regression analysis software (NLREG), the CD unfolding data obtained under low urea concentrations was fit to equations 3, 4, 5 and 9, holding the previously obtained values for ΔH(Tm), Tm, and ΔCp, constant for 0 M urea. The CD data obtained at concentrations of urea of 3.5 M and higher was fit to equations 6, 7, 8, and 9. The CD signal for the native state cannot be determined experimentally and therefore in the fit, the temperature dependence of the CD signal for the native state was assumed to be the same for all melting curves within a corresponding data set. Similarly, the CD signal for the unfolded state was assumed to be the same for all melting curves.

Results and Discussion

Experimental determination of the urea m-value for Trp-cage unfolding

Figure 1B shows the temperature-induced unfolding profiles of Trp-cage for a representative set of urea concentration as monitored using circular dichroism (CD). As expected, increasing concentrations of urea have a destabilizing effect on the Trp-cage mini-protein. The absolute ellipticity values at 4°C decrease with increasing urea, which suggests a decrease in the fraction of the native Trp-cage population. In addition, the transition profiles for higher urea concentrations are shifted towards lower temperatures, suggesting reduced thermal stability. Previously, we have shown that Trp-cage unfolding in the absence of urea closely follows a two-state unfolding model 2. Therefore, it was assumed that the experimental data in the presence of urea can be equally well described by a two-state process. For 0 M urea, a Tm of 41±2 °C was obtained, which is within the error of previously reported values 13.

Using the parameters obtained from the fit to a two-state model, the values of the change in Gibbs free energy at 25°C were calculated using equation 5 or 8 (see Table 1). The dependence of ΔG(25°C) on urea concentration appears to be a linear function of denaturant (see Figure 2). According to the linear extrapolation model 19,20, the slope of this dependence corresponds to the m-value and the intercept with the y-axis corresponds to the Gibbs energy of Trp-cage unfolding in 0 M urea:

Figure 2
The plot of ΔG(25°C) values versus the concentration of urea. Symbols represent the average of multiple experiments at that urea concentration. The solid line represents the fit to a linear extrapolation model (Equation 10), with m-value ...
Table 1
Thermodynamic parameters of unfolding of the Trp-cage mini protein in the presence of different concentrations of urea

Linear regression yields an m-value of 1.36±0.08 kJ mol−1 M−1 and ΔG(H2O)=2.8±0.3 kJ·mol−1.

Comparison of the m-value for Trp-cage with predictions by empirical methods

It is known that there is a strong correlation between m-values and the total exposed surface area upon unfolding, ΔASAtotal 21,22. However, this relationship has not been examined for very small proteins. The Trp-cage mini-protein, consisting of only 20 residues, is therefore a prime candidate for examining this correlation.

According to NACCESS calculations (see Materials and Methods for details), unfolding of Trp-cage results in a change in total accessible surface area, ΔASAtotal, of 990±50 Å2. This value corresponds well with the predicted ASA of 953 Å2, using the relationship obtained by Myers et al. 21:


Such a good agreement suggests that the total changes in the accessible surface area of Trp-cage follow a similar trend as those of larger proteins.

According to Myers et al. 21, there is a linear relationship between ΔASAtotal and the m-value:


Using this relationship and the ΔASAtotal of 990±50 Å2, the predicted m-value is 2.0±0.2 kJ mol−1 M−1,21. It should be noted, however, that equation 12 is strictly empirical; therefore, an m-value of 0 can only be obtained if ΔASAtotal is negative. If the regression line is forced to pass through the origin, it may be possible to more accurately predict the m-value for proteins with small ΔASAtotal values, such a Trp-cage. Using the same 45 proteins as Myers et al. 21, this new correlation can be written as


Using this equation, the predicted m-value for Trp-cage is 0.54±0.03 kJ·mol−1·M−1, which is lower than the experimental m-value of 1.36±0.08 kJ·mol−1 ·M−1 for Trp-cage. Since the true m-value lies somewhere between the original regression line and one constrained by thermodynamic principles, a better prediction may be possible if the data is fit using a polynomial regression:


Using this equation, the predicted m-value is 1.3±0.03 kJ·mol−1·M−1, which corresponds well to the experimental value. Although this approach is inconsistent with the interpretation of m-value being proportional to the accessible surface area exposed upon unfolding, it is strictly an empirical correlation, similar to equation 12.

In addition to the empirical correlations discussed above, there have been several attempts to predict m-values based on the free energy transfer of the side chains and backbone of a protein. The most notable of these is Tanford’s transfer model, which relies on the assumption that the folding free energy is equivalent to the difference in solvation between the native and denatured states 23. More recently, Bolen et. al have suggested that m-values can be accurately predicted using information from the three-dimensional structure 24. Assuming the relationship between m-value and transfer energies:




the calculated m-value for Trp-cage is 1.69±0.68 kJ mol−1 M−1. It is worth noting that this calculation involves modeling the solvent accessibility of the unfolded state as the average of two extremes25,26. While more elaborate and rigorous methods have been developed since (see e.g. ref. 27), they are in good agreement with approach used here. It is also worth noting that this approach is associated with a significant error. Bolen et al. stress that this is the result of the inability to accurately quantify the ASA of the denatured state, and not the fault of the model itself 24. Another potential error is that the correction for changes in activity coefficients of model compounds is applied only to the transfer energy of the backbone. Regardless of these shortcomings, the calculated m-value for Trp-cage is in good agreement with that obtained experimentally, suggesting the validity of the transfer model. O’Brien et al 28,29 reached a similar conclusion based on computer simulations of urea-induced unfolding of lysozyme.

Comparison of the ΔCp for Trp-cage with predictions by empirical methods

Experimentally determined value of heat capacity of unfolding, ΔCp, for Trp-cage is positive 0.3±0.1 kJ·mol−1·K−1 consistent with the observed positive heat capacity changes for globular protein unfolding 2. The small absolute value of ΔCp for Trp-cage is probably the result of smaller size of the Trp-cage miniprotein, and consequently smaller changes of the ASA exposed upon unfolding. A detailed analysis of the factors contributing to the change in heat capacity of protein unfolding suggests that the exposure of non-polar and polar ASA need to be accounted for separately. It has been shown, using the transfer of model compounds, that the exposure of non-polar ASA contributes positively to ΔCp of unfolding while polar ASA contributes negatively 21,30,31. Different absolute values for these contributions per square angstrom of non-polar and polar exposed surfaces were proposed by Meyers et al. 21, Spolar et al. 31, Murphy and Freire 30, respectively:




Furthermore, the contributions from non-polar and polar surface areas can be subdivided into non-polar aliphatic, non-polar aromatic, polar backbone and polar side chains 32:


Calculations performed using equations 1720 give ΔCp values of 0.19±0.04 kJ·mol−1·K−1, 0.40±0.05 kJ·mol−1·K−1, 0.43±0.07 kJ·mol−1·K−1, and 0.22±0.08 kJ·mol−1·K−1, respectively. These predictions compare well with the ΔCp value obtained by DSC and CD, 0.3±0.1 kJ·mol−1·K−1 2.

Comparison with the computer simulations

One of the attractive features that Trp-cage offers is that its small size allows full-atom computer simulations of unfolding of this protein 411,13,3336. For example, Paschek et al 13 used replica exchange molecular dynamics simulations and obtained a value of ΔCp of 0.2±0.05 kJ mol−1 K−1 for unfolding to Trp-cage, very similar to the one reported here. The same group obtained an m-value of 0.41±0.03 kJ mol−1 M−1 after performing simulations on Trp-cage in the presence of urea (Angel Garcia, personal communication). This value is in a reasonable correspondence with the experimental value obtained here. Availability of the thermodynamic parameter for Trp-cage reported here will further stimulate computational modeling efforts.

Concluding remarks

Overall, the Trp-cage mini-protein is well behaved and shares many similarities to larger-sized proteins, during temperature-induced unfolding in both the presence and absence of urea. Globally fitting individual data sets using a two-state model allowed for reliable estimates of ΔG(25°C) in the presence of different concentrations of urea. Using the linear extrapolation model, the urea m-value for Trp-cage unfolding was determined. The experimental m-value appears to be best predicted by the empirical transfer model 24. The experimental ΔCp can be well predicted by empirical models that account for the changes in the polar and non-polar accessible surface areas 21,32. These experimental thermodynamics data will be essential for validation of atomistic simulations of the effects of urea of protein stability by computational methods.


We thank Angel Garcia for sharing the results of unpublished work. This work was supported in part by a grant from NIH NIGMS (GM054537).


1. Neidigh JW, Fesinmeyer RM, Andersen NH. Designing a 20-residue protein. Nat Struct Biol. 2002;9(6):425–430. [PubMed]
2. Streicher WW, Makhatadze GI. Unfolding thermodynamics of Trp-cage, a 20 residue miniprotein, studied by differential scanning calorimetry and circular dichroism spectroscopy. Biochemistry. 2007;46(10):2876–2880. [PubMed]
3. Qiu L, Pabit SA, Roitberg AE, Hagen SJ. Smaller and faster: the 20-residue Trp-cage protein folds in 4 micros. Journal of the American Chemical Society. 2002;124(44):12952–12953. [PubMed]
4. Barua B, Lin JC, Williams VD, Kummler P, Neidigh JW, Andersen NH. The Trp-cage: optimizing the stability of a globular miniprotein. Protein Eng Des Sel. 2008;21(3):171–185. [PMC free article] [PubMed]
5. Bunagan MR, Yang X, Saven JG, Gai F. Ultrafast folding of a computationally designed Trp-cage mutant: Trp2-cage. The journal of physical chemistry. 2006;110(8):3759–3763. [PubMed]
6. Simmerling C, Strockbine B, Roitberg AE. All-atom structure prediction and folding simulations of a stable protein. Journal of the American Chemical Society. 2002;124(38):11258–11259. [PubMed]
7. Snow CD, Zagrovic B, Pande VS. The Trp cage: folding kinetics and unfolded state topology via molecular dynamics simulations. Journal of the American Chemical Society. 2002;124(49):14548–14549. [PubMed]
8. Chowdhury S, Lee MC, Xiong G, Duan Y. Ab initio folding simulation of the Trp-cage mini-protein approaches NMR resolution. J Mol Biol. 2003;327(3):711–717. [PubMed]
9. Pitera JW, Swope W. Understanding folding and design: replica-exchange simulations of “Trp-cage” miniproteins. Proceedings of the National Academy of Sciences of the United States of America. 2003;100(13):7587–7592. [PubMed]
10. Zhou R. Trp-cage: folding free energy landscape in explicit water. Proceedings of the National Academy of Sciences of the United States of America. 2003;100(23):13280–13285. [PubMed]
11. Beck DA, White GW, Daggett V. Exploring the energy landscape of protein folding using replica-exchange and conventional molecular dynamics simulations. J Struct Biol. 2007;157(3):514–523. [PMC free article] [PubMed]
12. Paschek D, Nymeyer H, Garcia AE. Replica exchange simulation of reversible folding/unfolding of the Trp-cage miniprotein in explicit solvent: On the structure and possible role of internal water. Journal of Structural Biology. 2007;157(3):524–533. [PubMed]
13. Paschek D, Hempel S, Garcia AE. Computing the stability diagram of the Trp-cage miniprotein. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(46):17754–17759. [PubMed]
14. Pace CN. Determination and analysis of urea and guanidine hydrochloride denaturation curves. Methods Enzymol. 1986;131:266–280. [PubMed]
15. Hubbard SJ, Thornton JM. ‘NACCESS’, Computer Program. Department of Biochemistry and Molecular Biology, University College London; 1993.
16. Ibarra-Molero B, Makhatadze GI, Sanchez-Ruiz JM. Cold denaturation of ubiquitin. Biochim Biophys Acta. 1999;1429(2):384–390. [PubMed]
17. Makhatadze GI, Privalov PL. Protein interactions with urea and guanidinium chloride. A calorimetric study. J Mol Biol. 1992;226(2):491–505. [PubMed]
18. Zweifel ME, Barrick D. Relationships between the temperature dependence of solvent denaturation and the denaturant dependence of protein stability curves. Biophysical Chemistry. 2002;101:221–237. [PubMed]
19. Bolen DW, Santoro MM. Unfolding free energy changes determined by the linear extrapolation method. 2. Incorporation of delta G degrees N-U values in a thermodynamic cycle. Biochemistry. 1988;27(21):8069–8074. [PubMed]
20. Pace CN, Shaw KL. Linear extrapolation method of analyzing solvent denaturation curves. Proteins. 2000;(Suppl 4):1–7. [PubMed]
21. Myers JK, Pace CN, Scholtz JM. Denaturant m values and heat capacity changes: relation to changes in accessible surface areas of protein unfolding. Protein Sci. 1995;4(10):2138–2148. [PubMed]
22. Courtenay ES, Capp MW, Saecker RMMT, Record J. Thermodynamic analysis of interactions between denaturants and protein surface exposed on unfolding: Interpretation of urea and guanidinium chloride m-values and their correlation with changes in accessible surface area (ASA) using preferential interaction coefficients and the local-bulk domain model. 2000;41:72–85. [PubMed]
23. Ziv G, Haran G. Protein Folding, Protein Collapse, and Tanford’s Transfer Model: Lessons from Single-Molecule FRET. Journal of the American Chemical Society. 2009;131(8):2942–2947. [PMC free article] [PubMed]
24. Auton M, Bolen DW. Predicting the energetics of osmolyte-induced protein folding/unfolding. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(42):15065–15068. [PubMed]
25. Creamer TP, Srinivasan R, Rose GD. Modeling unfolded states of peptides and proteins. Biochemistry. 1995;34(50):16245–16250. [PubMed]
26. Creamer TP, Srinivasan R, Rose GD. Modeling unfolded states of proteins and peptides. II. Backbone solvent accessibility. Biochemistry. 1997;36(10):2832–2835. [PubMed]
27. Gong H, Rose GD. Assessing the solvent-dependent surface area of unfolded proteins using an ensemble model. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(9):3321–3326. [PubMed]
28. O’Brien EP, Brooks BR, Thirumalai D. Molecular origin of constant m-values, denatured state collapse, and residue-dependent transition midpoints in globular proteins. Biochemistry. 2009;48(17):3743–3754. [PubMed]
29. O’Brien EP, Ziv G, Haran G, Brooks BR, Thirumalai D. Effects of denaturants and osmolytes on proteins are accurately predicted by the molecular transfer model. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(36):13403–13408. [PubMed]
30. Murphy KP, Freire E. Thermodynamics of structural stability and cooperative folding behavior in proteins. Adv Protein Chem. 1992;43:313–361. [PubMed]
31. Spolar RS, Livingstone JR, Record MT., Jr Use of liquid hydrocarbon and amide transfer data to estimate contributions to thermodynamic functions of protein folding from the removal of nonpolar and polar surface from water. Biochemistry. 1992;31(16):3947–3955. [PubMed]
32. Makhatadze GI, Privalov PL. Energetics of protein structure. Adv Protein Chem. 1995;47:307–425. [PubMed]
33. Ding F, Buldyrev SV, Dokholyan NV. Folding Trp-cage to NMR resolution native structure using a coarse-grained protein model. Biophys J. 2005;88(1):147–155. [PubMed]
34. Gattin Z, Riniker S, Hore PJ, Mok KH, van Gunsteren WF. Temperature and urea induced denaturation of the TRP-cage mini protein TC5b: A simulation study consistent with experimental observations. Protein Sci. 2009;18(10):2090–2099. [PubMed]
35. Kannan S, Zacharias M. Folding simulations of Trp-cage mini protein in explicit solvent using biasing potential replica-exchange molecular dynamics simulations. Proteins. 2009;76(2):448–460. [PubMed]
36. Zagrovic B, Pande V. Solvent viscosity dependence of the folding rate of a small protein: distributed computing study. Journal of computational chemistry. 2003;24(12):1432–1436. [PubMed]
37. DeLano WL. The PyMOL Molecular Graphics System. 2002.