Isothermal titration calorimetric and spectroscopic studies on (alcohol + salt) induced partially folded state of α-lactalbumin and its binding with 8-anilino-1-naphthalenesulfonic acid

Available online at www.sciencedirect.com

J. Chem. Thermodynamics 40 (2008) 1141–1151 www.elsevier.com/locate/jct

Isothermal titration calorimetric and spectroscopic studies on (alcohol + salt) induced partially folded state of a-lactalbumin and its binding with 8-anilino-1-naphthalenesulfonic acid Ruchika Sharma, Nand Kishore * Department of Chemistry, Indian Institute of Technology Bombay, Powai, Mumbai 400 076, India Received 29 December 2007; received in revised form 17 February 2008; accepted 19 February 2008 Available online 26 February 2008

Abstract Interaction of 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) and isopropanol in the presence of equimolar quantities of guanidine thiocyanate (GndSCN) with bovine a-lactalbumin (a-LA) has been investigated by using a combination of isothermal titration calorimetry, circular dichroism, fluorescence, and ultra-violet spectroscopies at in 20  103 mol  dm3 phosphate buffer pH 7.0. All the thermal unfolding transitions, in the presence of both the (alcohol + salt) mixtures were found to be reversible as judged by the same values of absorbance observed at different temperatures during cooling after the completion of thermal unfolding. In the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture and 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture, a-lactalbumin was observed to be in the partially folded state with significant loss of native tertiary structure. Intrinsic fluorescence results, acrylamide and potassium iodide quenching, 8-anilino-1-naphthalenesulfonic acid (ANS) binding, and energy transfer results also corroborate the presence of partially folded states of a-lactalbumin. Apart from the generation of the partially folded states, it was also observed that destabilizing action of GndSCN is reduced in the presence of isopropanol compared to that in HFIP. Isothermal titration calorimetry has been used to characterize the energetics of ANS binding to the partially folded states of the protein. ITC results indicate that ANS binds to these partially folded states at pH 7.0 due to the presence of two sequentially binding sites on the protein under the solvent conditions employed. For example, ANS binds to the 0.25 mol  dm3 (HFIP + GndSCN) induced partially folded state with affinity constants K1 = (858 ± 220), K2 = (1.12 ± 0.25)  103; enthalpies of binding DH1 = (4.4 ± 1.0) kJ  mol1, DH2 = (2.1 ± 0.2) kJ  mol1; and entropies of binding DS1 = 70 J  K1  mol1 and DS2 = 65 J  K1  mol1, respectively at these two sequential binding sites. In light of the fluorescence results, possible binding sites where ANS can bind to the protein have also been suggested. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: a-Lactalbumin; Partially folded states; Isothermal titration calorimetry; Spectroscopy; (Salt + alcohol) mixture; 8-anilino-1-naphthalenesulfonic acid

1. Introduction In order to understand the conformational behaviour of a protein, defining the details of the structures of intermediate species that populate protein folding pathway constiAbbreviations: HFIP, 1,1,1,3,3,3-hexafluoroisopropanol; GndSCN, guanidine thiocyanate; a-LA, a-lactalbumin; ANS, 8-Anilino-1-naphthalenesulfonic acid; MG, molten globule. * Corresponding author. Tel.: +91 22 2576 7157; fax: +91 22 2576 7152. E-mail address: [email protected] (N. Kishore). 0021-9614/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jct.2008.02.009

tutes an important challenge in the understanding of the mechanism of protein folding. Significant work has been done on the detection and characterization of these folding intermediates [1–3]. It has been suggested [4–8] that a close similarity exists between the compact species formed during the early stages of refolding a protein and the partially folded states observed at equilibrium under partial denaturing conditions. Thus, insights into the transient kinetic partially folded states, which are otherwise difficult to trap because of their transient nature, can be gained through

1142

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

detailed structural studies of the stable equilibrium partially folded state. Alcohols are known to denature the native state of proteins by weakening the hydrophobic interactions and enhance local polar interactions such as hydrogen bonds. They also stabilize the a-helical conformation in proteins by minimizing the exposure of peptide backbone [9–12]. Fluoro-substituted alcohols have been shown to exhibit structure stabilizing effects on proteins and peptides [13– 15]. In recent years, use of 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) has been reported in the generation of intermediate conformations and several other applications in biochemical studies [16–21]. The biological implication of HFIP has been reported in the investigation of prion diseases and Alzhiemer’s amyloid peptides [22,23]. It has been used to induce refolding in case of Cobra neurotoxin in the presence of sodium dodecyl sulphate [20]. HFIP has effectively formed aggregates in cecropin AD [19] and molten globule (MG) states in myoglobin [18]. The regulation of milk lactose biosynthesis is highly dependent on the action of a specifier protein, a-lactalbumin (a-LA) which is a well characterized protein with a molar mass of 14.2 kDa and 123 amino acid residues making up the two domains [24]. In this work, the effectiveness of isopropanol and its fluorosubstituted homologue, HFIP, to generate the partially folded state in a-LA has been examined in the presence of guanidine thiocyanate (GndSCN). The GndSCN is a strong denaturant while alcohols are capable of producing local secondary structures, and here we have made an attempt to find whether an partially folded state at equilibrium can be generated under milder conditions when they are used together. A combination of isothermal titration calorimetry, UV–Visible, circular dichroism and fluorescence spectroscopy has been used to characterize these partially folded states both qualitatively and quantitatively.

reported pH is that of the dialysate, determined using a standard Control Dynamics pH meter at room temperature.

2. Materials and methods

The thermodynamic parameters obtained for the reversible thermal denaturation of the protein were used to calculate the value of DC23, the change in preferential solvation parameter of component 2 (protein) by component 3 ((alcohol + salt) mixture) using the equation [27]   oT DH ox1=2  3 ; DC23 ¼ CD3  CN 3 ¼ ð2Þ lg a3 2 RT 1=2 o ox 3

Bovine a-lactalbumin (type I) (mass fraction purity 0.85), 1,1,1,3,3,3-hexafluoroisopropanol (mass fraction purity 0.99), ANS (mass fraction purity 0.99), potassium iodide (mass fraction purity 0.996), acrylamide (mass fraction purity 0.994), sodium thiosulphate(Na2S2O3) (mass fraction purity >0.995), guanidine thiocyanate (mass fraction purity 0.99) and potassium phosphate (mass fraction purity 0.99) were purchased from Sigma Chemical Company and used without further purification. Isopropanol was procured from SRL and was used as such. The base NaOH (mass fraction purity 0.99) was of reagent grade. Entries in the parentheses represent purity of the compound as listed by the vendor. The water used for preparing the solutions was double distilled followed by deionization using a Cole-Parmer research mixed-bed ion exchange column. The stock solutions of a-LA for all the experiments were prepared by extensive dialysis of the protein at T = 277 K against 20  103 mol  dm3 potassium phosphate buffer, at pH 7.0. The

2.1. UV–Visible experiments The concentration of a-lactalbumin was determined using an extinction coefficient corresponding to an absorbance of A1% 280 ¼ 20:1 at pH 7.0 [25] in phosphate buffer. This value is assumed to be reliable as it has been thoroughly reported in the literature to calculate the concentration of a-lactalbumin. A Shimadzu UV265 spectrophotometer was used for the concentration determination and thermal denaturation scans. The protein concentration in the thermal denaturation experiments was kept at 7.0  103 mol  dm3 and the absorbance at different temperatures was measured at a fixed wavelength of 293 nm. The temperature around the cuvettes was controlled with a Cole-Parmer Polystat temperature controller unit to within 0.1 K. The results in terms of transition temperature (T1/2) and enthalpy of thermal unfolding (DH) were analyzed by EXAM programme of Kirchhoff [26] and the other thermodynamic parameters accompanying the thermal transitions were evaluated. Each experiment was repeated at least twice to check the reproducibility. The reversibility of the thermal denaturation was checked by heating the sample to just above the transition temperature, cooling it down gradually and measuring the absorbance. The Gibbs free energy (DG) of thermal denaturation was calculated at different temperatures for denaturation of the protein in these solvents by using the values of heat capacity of denaturation (DCp) obtained from slope of the plot of DH against T1/2 in the following relationship,     T 1=2  T T 1=2  DG ¼ DH  DC p ðT 1=2  T Þ þ T DC p ln : T 1=2 T ð1Þ

where CD3 and CN3 represent the preferential solvation of the protein by the (alcohol + salt) mixture in the denatured and native states, respectively. Here, x is the mole fraction and ‘a’ is the activity of alcohol. The highest mole fraction of alcohol used in the calorimetric experiments is 0.0174; therefore, the activity was taken to be equal to the mole fraction. 2.2. Fluorescence experiments The fluorescence experiments were done on a Perkin– Elmer LS-55 spectrofluorimeter at T = 298 K with a quartz

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

cell of 1-cm path length. The protein concentration in all the experiments was kept between 3.5  106 and 5  106 mol  dm3. The excitation and emission slit widths were fixed at 5 nm. The excitation wavelength was kept at 295 nm to selectively excite the tryptophan molecules. For the ANS binding and energy transfer experiments, the ANS concentration was kept at 12  105 mol  dm3 which was determined using the 1 extinction coefficient of ANS as E1%  dm3  350 ¼ 5000 mol 1 cm [28]. For ANS binding and energy transfer experiments, the excitation wavelengths were fixed at 365 and 295 nm, respectively. Quenching experiments were performed with acrylamide and KI where the concentration of the quencher was varied from (0.05 to 0.30) mol  dm3. The background spectra containing same amount of additive in buffer was subtracted from all the plots. For quenching experiments with KI, 1.0  105 mol  dm3 Na2S2O3 solution was added to prevent the formation of I 3.

K 2 ¼ ½ML2 =½ML½L;

1143

ð4Þ

where L denotes ligand and M denotes the macromolecule. The heat content Q after any ith injection is then expressed as Q ¼ M t V o ½F 1 DH 1 þ F 2 ðDH 1 þ DH 2 Þ þ    þ F n ðDH 1 þ DH 2 þ DH 3 þ    þ DH n Þ;

ð5Þ

where Fi is the fraction of total macromolecule having i bound ligands and Vo is the active cell volume. The pertinent calculated heat effect (DQ) for the ith injection is   dV i ðQðiÞ þ Qði  1ÞÞ DQ ¼ QðiÞ þ  Qði  1Þ; ð6Þ 2 Vo which is then used in the Marquardt minimization algorithm to obtain best fitting values for the thermodynamic parameters.

2.3. Circular dichroism experiments

3. Results and discussion

The CD experiments were performed on a Jasco-810 CD spectropolarimeter at T = 298 K. The protein concentration and path length of the cell used were 5  106 mol  dm3 and 0.1 cm for far UV CD and 20  106 mol  dm3 and 1 cm for near UV CD, respectively. The spectropolarimeter was purged with N2 prior to the experiment. Each CD plot was an average of three accumulated plots scanned at 500 nm  min1. The plots were baseline corrected. The molar ellipticity was calculated from the observed ellipticity, h, as 100  h/c  l, where c is the concentration of the protein solution in mol  dm3 and l is the path length of the cell in centimetres.

3.1. Interaction of a-lactalbumin with (HFIP + GndSCN) and (isopropanol + GndSCN) mixtures 3.1.1. UV–Visible experiments The representative thermal denaturation profiles of a-lactalbumin in the absence and presence of varying concentrations of equimolar (HFIP + GndSCN) and (isopropanol + GndSCN) mixtures at pH 7.0, as studied by the change in absorbance at 293 nm as a function of temperature, are shown in figure 1 as fraction of native protein (fN). The value of fN was calculated using the following equation:

2.4. Isothermal calorimetry experiments fN ¼ The ITC measurements were carried out at T = 298.15 K on an isothermal titration calorimeter (VPITC from Micro Cal, Northampton, MA). Before loading, the solutions were thoroughly degassed by stirring under vacuum in a Thermovac unit supplied with the instrument. The reference cell was filled with the degassed buffer and the sample cell contained protein solution at a concentration of 1.4  104 mol  dm3. Titrations were carried out using a 250 ldm3 syringe filled with the solution of interest which was added sequentially in 5 ldm3 aliquots (for a total of 50 injection, 20 s duration each) at 4 min intervals. The heats of dilution were determined in parallel experiments for the ligand, protein and buffer mixing, and were subtracted from the main binding profile prior to the curve fitting. Origin 5.0 software was used to fit the thermodynamic parameters to the measured heats of reaction. The data on binding of ANS to a-lactalbumin fitted best to two sequential binding sites model. For sequential binding, the binding constants are defined relative to the progress of saturation, such as K 1 ¼ ½ML=½M½L;

ð3Þ

AN  AT ; AN  AD

ð7Þ

where AN, AD and AT are the values of the absorbance of the protein in the native state, denatured state, and at any general temperature T. The thermal denaturation reaction of the protein occurs in the manner as represented by: Native a-LAðN Þ ¼ Thermally denatured a-LA ðDÞ:

ð8Þ

The corresponding thermodynamic parameters accompanying the transitions are reported in table 1. The most obvious effect observed in the presence of (alcohol + salt) mixtures was the lowering of the thermal transition temperature and the enthalpy of the protein unfolding. The thermal unfolding in presence of both the additive mixtures was found to be reversible and so we could calculate the associated thermodynamic parameters. Beyond 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture, no clear thermal transitions were observed. At higher concentrations of the (isopropanol + GndSCN) equimolar mixture, the transition curves were observed to be close to each other as also revealed by the entropy and enthalpy of the transitions (table 1). In the presence of the (HFIP + GndSCN) equimolar mixture, the heat capacity

1144

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

II

1.0

fraction of native protein

fraction of native protein

I

0.8 0.6 0.4 0.2 0.0

1.0 0.8 0.6 0.4 0.2 0.0

280 290 300 310 320 330 340 350 360

280 290 300 310 320 330 340 350 360

T/K

T/K

FIGURE 1. Plot of fraction of native protein against temperature to show the denaturation profiles of a-LA at pH 7.0 in the presence of the equimolar (I) (HFIP + GndSCN) mixture: 0 (j), 0.05 (d), 0.1 (N), 0.2 (.), 0.25 (), 0.3 (), 0.4 (*) and 0.5 (+) mol  dm3, (II) (isopropanol and GdSCN) mixture: 0 (j), 0.1 (d), 0.3 (N), 0.5 (.), 0.6 (), 0.85 (+) and 1.0 () mol  dm3.

TABLE 1 Transition temperature (T1/2), enthalpy (DH), and entropy (DS) of the thermal unfolding of 7  106 mol  dm3 a-lactalbumin at pH 7.0 in the presence of the (HFIP + GndSCN) equimolar mixture and (isopropanol (IP) + GndSCN) equimolar mixture according to equation (8)a [HFIP + GndSCN]/ (103 mol  dm3)

T1/2/ K

DH/ (kJ  mol1)

DS/ (kJ  K1  mol1)

[IP+GndSCN]/ (103mol  dm3)

T1/2/ K

DH/ (kJ  mol1)

DS/(kJ  K1  mol1)

0 50 75 100 150 200 250

336.6 324.6 321.8 316.8 313.1 306.1 301.8

285 273 241 221 217 195 190

0.85 0.84 0.75 0.70 0.69 0.64 0.62

0 100 300 500 600 750 850

336.6 336.7 329.3 324.0 318.0 310.7 303.8

285 282 263 249 244 235 200

0.85 0.84 0.79 0.77 0.76 0.76 0.66

a Each experiment was repeated at least twice to check the reproducibility. When error in sample preparation, reproducibility, and sample impurities are incorporated, the errors in the values of T1/2 and DH are ±0.2 K and 5%, respectively.

of denaturation of a-lactalbumin (DCp) from a plot of DH against T1/2 was found to be (2.2 ± 0.2) kJ  K1  mol1 and that in the presence of the (isopropanol + GndSCN) equimolar mixture was found to be (2.3 ± 0.2) kJ  K1  mol1. Exposure of the non-polar groups that are previously buried in the native structure to the surrounding solvent molecules around them results in large and positive values of DCp for protein denaturation. However, it is also known that a negative contribution from the exposure of the polar groups is expected to offset partially the positive contribution (DCpol < 0; DCapolar > 0) [29,30]. In the absence of any additive, the value of DCp for a-lactalbumin is known to be (4.6 ± 0.3) kJ  K1  mol1 [16]. Lowering the value of DCp for a-lactalbumin under these experimental conditions indicates enhanced hydrophobic interactions between the non-polar regions of the unfolded protein and the alkyl groups of isopropanol and –CF3 groups of HFIP, respectively, leading to reduction of the extent of reordering of water molecules surrounding the hydrophobic regions of the protein. The DCp value obtained in the presence of the (isopropanol + GndSCN) equimolar mixture were comparable to that obtained in the presence of the (HFIP + GndSCN) equimolar mixture which shows a

similar extent of structural reorganization under these conditions. The values of denaturational change in the preferential solvation parameter (DC23) accompanying the thermal unfolding of a-lactalbumin in the presence of the (HFIP + GndSCN) equimolar mixture and (isopropanol + GndSCN) equimolar mixture were calculated [31] with respect to the mole fraction of the alcohol (table 2). The positive value of DC23 and its increase with the increasing concentrations of both the mixtures indicate (figure 2) that the co-solute have greater affinity towards the protein surface exposed upon denaturation than that in the native state, thus shifting the Native (N) D Denatured (D) equilibrium to the right, leading to decrease in the denaturation temperature of a-lactalbumin. The plot of denaturational change in the standard Gibbs free energy calculated by using equation (1) against temperature at different concentrations of the (HFIP + GndSCN) equimolar mixture (table 3) and the (isopropanol + GndSCN) equimolar mixture (table 4) shows that there is a continuous decrease in the value of DG with the increasing concentration of the (alcohol + salt) mixture (figure 3), suggesting that destabilization is favoured at higher concentrations of these two mixtures. The temperature at the

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151 TABLE 2 Change in the preferential solvation parameter (DC23) due to the thermal unfolding of 7  106 mol  dm3a-LA at pH 7.0 in the presence of equimolar (HFIP + GndSCN) and equimolar (isopropanol (IP) + GndSCN) mixtures Mole fraction of HFIP

DC23

Mole fraction of isopropanol

DC23

0.0009 0.0014 0.0018 0.0028 0.0038 0.0048

1.8 2.4 3.0 4.6 5.9 7.4

0.0009 0.0014 0.0018 0.0028 0.0037 0.0046

0.6 0.8 1.1 1.7 2.2 2.8

8 7 6

Δ Γ 23

5 4 3 2 1 0 0.001

0.002

0.003

0.004

0.005

mole fraction of alcohol FIGURE 2. Plot of the change in the preferential solvation parameters (DC23) against mole fraction alcohol accompanying denaturation of 7.0  106 mol  dm3 a-lactalbumin at pH 7.0 in the presence of equimolar (HFIP + GndSCN) mixture (j) and (isopropanol + GndSCN) mixture (d) as a function of mole fraction of alcohol.

intersection of the plot with the dotted line is the transition temperature for thermal denaturation of the protein in presence of the respective concentrations of the mixtures. In order to check the validity of group additivity in the thermodynamic parameters following relationships were considered D1 ¼ T 1=2 ðmixtureÞ  T 1=2 ðbufferÞ; D2 ¼ T 1=2 ðalcoholÞ  T 1=2 ðbufferÞ;

ð9Þ ð10Þ

D3 ¼ T 1=2 ðGndSCNÞ  T 1=2 ðbufferÞ:

ð11Þ

At a particular concentration where D2 + D3 = D1, the group additivity would hold, implying that the two additives exert their individual denaturing influence when used as a mixture. It was observed that for the 0.2 mol  dm3 (HFIP + GndSCN) equimolar mixture group additivity holds whereas for the same concentration of the (isopropanol + GndSCN) equimolar mixture, it does not. This can be explained on the basis of the fact that alcohols lower the dielectric constant of an aqueous solution [32]. When this happens, at lower concentration of alcohol, the salt would form an even stronger ion pair, and would not denature the protein in its full capacity. It was observed that T1/2 of the protein in the presence of GndSCN or isopropanol alone is lower than that in the presence of the (isopropanol + GndSCN) equimolar mixture suggest-

1145

TABLE 3 Standard Gibbs free energy of thermal denaturation (DG) of 7.0  106 mol  dm3 a-LA at pH 7.0 at different temperatures, in the presence of different concentrations of the equimolar (HFIP + GndSCN) mixture DG/(kJ  mol1)

T/K

0

0.05

0.1

0.2

0.25

3

273 278 283 288 293 298 303 308 313 318 323 328 333 338 343 348 353 358 363 368 373

39.71 37.68 35.45 33.03 30.42 27.61 24.63 21.46 18.12 14.60 10.90 7.04 3.01 1.19 5.55 10.07 14.75 19.58 24.57 29.71 35.00

[HFIP + GndSCN]/(mol  dm ) 33.85 23.56 17.00 31.45 21.61 14.97 28.85 19.46 12.75 26.06 17.12 10.33 23.08 14.58 7.72 19.90 11.86 4.92 16.55 8.96 1.94 13.01 5.87 1.22 9.29 2.60 4.56 5.40 0.84 8.08 1.34 4.46 11.77 2.90 8.24 15.63 7.30 12.20 19.66 11.87 16.32 23.86 16.60 20.60 28.21 21.49 25.04 32.73 26.54 29.64 37.41 31.75 34.39 42.24 37.11 39.30 47.22 42.62 44.36 52.36 48.28 49.58 57.64

15.00 12.86 10.52 7.98 5.26 2.34 0.76 4.04 7.50 11.14 14.95 18.93 23.07 27.39 31.86 36.50 41.29 46.24 51.34 56.59 62.00

TABLE 4 Standard Gibbs free energy of thermal denaturation (DG) of 7.0  106 mol  dm3 a-LA at pH 7.0 at different temperatures, in the presence of different concentrations of the equimolar (isopropanol + GndSCN) mixture T/K

273 278 283 288 293 298 303 308 313 318 323 328 333 338 343 348 353 358 363 368 373

DG/(kJ  mol1) 0

0.1

39.71 37.68 35.45 33.03 30.42 27.62 24.63 21.46 18.12 14.60 10.90 7.04 3.00 1.19 5.55 10.07 14.75 19.58 24.57 29.71 35.00

[Isopropanol + GndSCN]/(mol  dm3) 38.88 33.20 29.44 26.83 23.03 37.04 31.26 27.46 24.65 20.63 34.99 29.11 25.27 22.25 18.023 32.74 26.76 22.89 19.66 15.21 30.29 24.21 20.30 16.86 12.20 27.64 21.46 17.51 13.87 9.00 24.80 18.52 14.54 10.68 5.60 21.77 15.39 11.37 7.30 2.01 18.55 12.07 8.02 3.74 1.76 15.15 8.57 4.48 0 5.72 11.56 4.89 0.76 3.93 9.86 7.80 1.03 3.13 8.03 14.17 3.86 3.00 7.20 12.31 18.66 0.25 7.21 11.44 16.76 23.33 4.53 11.59 15.86 21.38 28.16 8.98 16.13 20.44 26.17 33.17 13.60 20.84 25.19 31.13 38.33 18.38 25.72 30.10 36.25 43.67 23.32 30.75 35.17 41.53 49.16 28.42 35.94 40.39 46.96 54.81 33.68 41.29 45.78 52.56 60.61

0.3

0.5

0.6

0.75

0.85 16.56 14.39 12.02 9.44 6.66 3.69 0.52 2.83 6.37 10.10 14.01 18.09 22.35 26.78 31.39 36.16 41.09 46.19 51.45 56.87 62.44

ing that in a mixture protein is slightly stabilized. Whereas this kind of stabilization was not observed in the presence of the (HFIP + GndSCN) equimolar mixture which can be

1146

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

B

45

45

30

30

15

15

ΔG0/ (kJ.mol-1)

ΔG0/ (kJ.mol-1)

A

0 -15 -30 -45

0 -15 -30 -45

-60

-60

-75

-75

-90 260

280

300

320

340

360

-90 260

380

280

300

320

T/K

340

360

380

T/K

FIGURE 3. Plot of the standard Gibbs free energy (DG) of thermal denaturation against temperature for 7.0  106 mol  dm3 a-LA at pH 7.0 in the presence of the equimolar (A) (HFIP + GndSCN) mixture: 0 (j), 0.05 (d), 0.1 (N), 0.2 (.), 0.25 () mol  dm3 (B) (isopropanol + GndSCN) mixture: 0 (j), 0.1 (d), 0.3 (N), 0.5 (.), 0.6 (), 0.75 (+) and 0.85 () mol  dm3.

attributed to the very strong denaturing influence of HFIP due to the presence of two –CF3 groups which enhance the hydrophobicity of the alcohol manifolds.

1.0 mol  dm3 equimolar mixtures, suggesting that protein is also in the partially folded conformation at these concentrations of the additives.

3.1.2. Circular dichroism Near-UV CD spectra for a-lactalbumin in the presence of both the (alcohol + salt) mixtures (figure 4) show the characteristic negative maxima at 272 nm, which follows a concentration dependent decrease with increase in the concentrations of the mixtures. The total loss of tertiary structure was observed at and above the 0.3 mol  dm3 (HFIP + GndSCN) equimolar mixture (figure 4A). Significant loss of tertiary structure was observed in the presence of 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture, indicating a partially folded state of a-lactalbumin under these conditions. Near-UV CD plots for a-lactalbumin in the presence of the (isopropanol + GndSCN) equimolar mixture (figure 4B), show an appreciable loss of tertiary structure in the presence of 0.85 mol  dm3 and

3.1.3. Fluorescence experiments Intrinsic fluorescence experiments were done to monitor change in the tryptophan environment in the presence of the two sets of additives. In the absence of any additive, the fluorescence spectra of the protein give the characteristic fluorescence intensity maxima (kmax) at 331 nm, which is consistent with that reported in the literature [33]. In the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture, shift in the value of kmax from (331 to 345) nm was observed. For the 8.5 mol  dm3 urea denatured protein, kmax at 354 nm corresponding to a fully denatured protein was obtained. The intrinsic fluorescence plot in the presence of the (isopropanol + GndSCN) equimolar mixture, first shows a lowering of intensity, a red shift in the kmax from (331 to 345) nm for 0.85 mol  dm3

B

1

0 0

[Θ ](104deg · cm2 · dmol-1)

[Θ ] (104deg · cm2 · dmol-1)

A

-1 -2 -3 -4

-1 -2 -3 -4 -5

-5 260

280

300

320

λ /nm

340

360

260

280

300

320

340

360

λ /nm

FIGURE 4. Near-UV CD spectra of 20  106 mol  dm3 a-lactalbumin at pH 7.0, in the presence of the equimolar (A) (HFIP + GndSCN) mixture: 0 (j), 0.05 (d), 0.1 (N), 0.15 (.), 0.2 (), 0.25 (+), 0.3 (), 0.4 (*), 0.5 () mol  dm3 (B) (isopropanol + GndSCN) mixture: 0 (j), 0.1 (d), 0.3 (N), 0.6 (.), 0.85 (), 1.00 (+) mol  dm3.

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

and a further shift to 352 nm for 1.0 mol  dm3 of the same additives, which suggest that at 1.0 mol  dm3 of the mixture, the protein is near denaturation. Characteristically, partially folded states are expected to have kmax lying between that of the native and fully denatured state [24]. Our experimental results also show that at pH 7.0, for concentrations at and above those of the 0.2 mol  dm3 (HFIP + GndSCN) equimolar mixture and 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture, the value of kmax is in between 331 nm (corresponding to the native protein) and 354 nm (corresponding to fully denatured protein). 3.1.3.1. ANS binding and energy transfer. ANS is a wellknown hydrophobic fluorophore which has been extensively used for characterizing the folding intermediates of proteins. The ANS binding and energy transfer experiments were done in order to get further information as to which concentration of the (alcohol + salt) mixtures led to partially folded conformations of a-lactalbumin. When ANS binds to exposed hydrophobic residues in proteins there is energy transfer between the tryptophans of the protein and the bound ANS. The tryptophan residues lose their fluorescence intensity and the ANS fluorescence gains in intensity. In the presence of a partially folded protein with exposed hydrophobic surfaces, the fluorescence of ANS is enhanced and the emission maximum is blue-shifted from 510 nm, corresponding to free ANS to 470 nm corresponding to protein bound ANS [34]. From ANS binding and energy transfer studies, it was observed that there is minimal binding of ANS with the protein in the native state which gradually increases on addition of the additive mixtures. The ANS binding and energy transfer fluorescence intensity was maximum in the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture and 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture suggesting that a-lactalbumin could possibly be in a stable, partially folded conformation under these conditions. These observations are consistent with the well established fact that ANS binds more to a partially folded state compared to native and denatured states of the protein [34].

1147

3.1.3.2. Acrylamide quenching. To investigate further the change in conformation of the protein in the presence of the additive mixtures, experiments with acrylamide as the quencher were performed. For dynamic and collisional quenching, the process is governed by the following equation [33], F 0 =F ¼ 1 þ K SV ½Q;

ð12Þ

where F0 is the fluorescence intensity in absence of the quencher Q, F is the value in the presence of the quencher at concentration [Q] and KSV is the Stern–Volmer quenching constant. The values of KSV were obtained from the slope of the plot of (F0/F  1) against [Q] (table 5, figure 5). The extent of quenching and hence the value of KSV depends on the degree to which the quencher achieves the encounter distance of the fluorophore. Acrylamide can penetrate both hydrophobic as well as hydrophilic regions of the protein and this is indicated by the linear plot of (F0/F  1), against [Q] suggesting that all four tryptophan residues of a-lactalbumin are equally accessible to the quencher. The value of the Stern–Volmer quenching constant, KSV for the native protein was found to be 1.97 mol1  dm3, whereas for 8.5 mol  dm3 urea denatured protein it was 7.16 mol1  dm3 which is almost 3.6 times higher than that of the native protein. These findings suggest that in the native state the tryptophans are shielded from the quencher to the highest degree and in the denatured state at the least amount. In the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture the value of KSV is 6.23 mol1  dm3, which is in between that of the native and denatured state and thus, suggestive of an intermediate conformation of a-lactalbumin. In the presence of the 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture, the value of KSV is 4.88 mol1  dm3, again in between that for the native and the denatured protein. 3.1.3.3. KI quenching. The quenching experiments were also done in the presence of KI, which being an ionic compound can encounter and hence quench only those tryptophans, which are exposed to the solvent [35]. The a-lactalbumin has four tryptophan residues at positions,

TABLE 5 The values of relative change in fluorescence intensity [(F0/F)  1] upon addition of acrylamide in 5.0  106 mol  dm3a-LA in the absence and presence of the 0.25 mol  dm3 equimolar (HFIP + GndSCN) mixture, 0.85 mol  dm3 equimolar (isopropanol (IP) + GndSCN) mixture and 8.5 mol  dm3 urea [Acrylamide]/ (mol  dm3) 0.025 0.05 0.075 0.1 0.15 0.2 0.3 0.4 0.5

(F0/F)  1 a-LA 0.089

0.361 0.532 0.751 1.009

a-LA + 0.25 mol  dm3(HFIP + GndSCN)  0.138 0.303 0.415 0.579 0.896 1.126 1.909

a-LA+0.85 mol  dm3(IP + GndSCN)

a-LA+ 8.5 mol  dm3 urea

0.256

0.105 0.256 0.482

0.456 0.706 0.915 1.415 1.962

1.346 2.072

1148

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

B 2.2 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0

2.0 1.8 1.6

(F0/F-1)

(F0 /F-1)

A

1.4 1.2 1.0 0.8 0.6 0.4 0.2

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.0 0.0

-3

0.1

0.2

0.3

0.4

0.5

[Acrylamide]/(mol · dm-3)

[Acrylamide]/(mol · dm )

FIGURE 5. Plot of (F0/F  1) against the concentration of acrylamide, for 5.0  106 mol  dm3 a-lactalbumin, (A) in the absence (j) and the presence of the 0.25 mol  dm3 equimolar (HFIP + GndSCN) mixture (d) and 8.5 mol  dm3 urea (N), (B) in absence (j) and presence of 0.85 (N) mol  dm3 equimolar (isopropanol + GndSCN) mixture and 8.5 (.) mol  dm3 urea.

26, 60,104, and 118 [36]. Out of these four tryptophans, the one at position 60 is the only tryptophan residue exposed to the solvent in the native state and contributes only 7% to the total fluorescence of a-lactalbumin, hence in the native state the quenching of the protein fluorescence by KI is almost nil. The tryptophan residues at positions 104 and 118 are part of the 310 helix and the one at position 26 is part of the a-helix [37]. These three residues are essential for the formation of partially folded state of a-lactalbumin. In the presence of the 0.25 mol  dm3 (HFIP + GdSCN) equimolar mixture and 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture, the protein showed considerable amount of quenching with KI. On plotting (F0/F  1) against [KI], nonlinear plots were obtained (table 6; figures 6 and 7) indicative of fluorophores interacting with KI in different environments. Hence to interpret the result, the modified Stern–Volmer equation [36] was used

F 0 =ðF 0  F Þ ¼ 1=fa þ 1=ðfa  K a  ½QÞ;

ð13Þ

where fa is the fraction of fluorophore accessible to the quencher and Ka is the association constant of the quencher to the fluorophore. Making use of the modified Stern– Volmer equation [F0/(F0  F)] was plotted against 1/[KI] and a linear plot was obtained (figures 6 and 7). In the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture, the fraction of tryptophan accessible to the quencher was found to be 0.66, which means that 66% of the tryptophans were accessible to the quencher. The fraction of fluorophore accessible to the quencher in the presence of the 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture was found to be 0.50 which means that out of four tryptophans of a-lactalbumin, 50% are exposed to the solvent. In the presence of 8.5 mol  dm3 urea, all four tryptophan residues are subjected to a similar degree of the fluorescence quenching by

TABLE 6 The values of relative change in fluorescence intensity [(F0/F)  1] upon addition of [KI] in 5.0  106 mol  dm3a-LA in the absence and presence of the 0.25 mol  dm3 equimolar (HFIP + GndSCN) mixture, 0.85 mol  dm3 equimolar (isopropanol (IP) + GndSCN) mixture and 8.5 mol  dm3 urea [KI]/(mol  dm3)

(F0/F)  1 a-LA

0.013 0.025 0.040 0.050 0.060 0.075 0.080 0.100 0.120 0.140 0.150 0.160 0.180 0.200

0.006

a-LA+0.25 mol  dm3(HFIP + GndSCN)

a-LA+0.85 mol  dm3(IP + GndSCN)

a-LA+8.5 mol  dm3 urea

0.066 0.079 0.176  0.230

0.021 0.047 0.070

0.213

0.003 0.003

0.276 0.313 0.364 0.386

0.015

0.020

0.474  0.520

0.275 0.075  0.087 0.079 0.096 0.109  0.083 0.0747 0.096

0.337 0.389

0.513

0.636

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

I

18

II

0.7

1149

16

0.6 14

F0 /(F0-F)

(F0 /F-1)

0.5 0.4 0.3

12 10 8

0.2

6

0.1

4

0.0

2

0.00

0.05

0.10

0.15

0

0.20

10

20

30

40

50

60

1/[KI]/ M-1

[KI]/ M

FIGURE 6. (I) Plot of (F0/F  1) against KI concentration for 5.0  106 mol dm3 a-lactalbumin in the absence (j), in the presence of the 0.25 mol  dm3 equimolar (HFIP + GndSCN) mixture (d) and 8.5 mol  dm3 urea (N). (II) The plot of F0/(F0  F) against the reciprocal of KI concentration under the same conditions.

I

0.6

II 18

0.5 16 14

F0 /(F0-F)

(F0 /F-1)

0.4 0.3 0.2

12 10 8

0.1

6 0.0 4 0.00

0.05

0.10

0.15

0.20

[KI]/(mol · dm-3)

4

6

8

10

12

14

16

18

1/[KI]/(mol-1 · dm3)

FIGURE 7. (I) Plot of (F0/F  1) against KI concentration for 5.0  106 mol  dm3 a-lactalbumin in the absence (j), in the presence of the 0.85 (.) mol  dm3 equimolar (isopropanol + GndSCN) mixture and 8.5 mol  dm3 urea (d). (II) The plot of F0/(F0  F) against the reciprocal of KI concentration under the same conditions.

KI, i.e. all of them are exposed to the solvent in the urea denatured protein. 3.2. Binding interaction of ANS with a-lactalbumin in its native and partially folded conformations: isothermal titration calorimetry Isothermal titration calorimetric thermograms were obtained by titrating 7.0  103 mol  dm3 ANS with 0.14  103 mol  dm3 a-lactalbumin in the native and partially folded states of the protein. Integration of the area of cell feed back by subtracting the dilution heats of both the ligand and protein gives the differential curves, which provides the amount of heat generated per injection as a function of the molar ratio of ANS to protein. The calorimetric titration profile for the binding of ANS with a-lactalbumin in the native state at pH 7.0 fitted well to a single binding site model with an average of 4.4 mol of ANS binding per mole of the protein. According to equation (14) with n = 4.4

a-LA þ nANS ! a-LA  nANS: ð14Þ The values of association constant, enthalpy, and entropy of binding are (7.1 ± 1.3)  103 mol1  dm3, (836 ± 150) J  mol1and 59 J  K1  mol1, respectively. The thermodynamic parameters indicate that the binding of ANS to the native state of a-lactalbumin is weak and is accompanied by a slight exothermic effect indicating absence of significant hydrophobic interactions between the ligand and the protein. The hydrophobic interactions are accompanied by endothermic heat effects [38]. The calorimetric titrations of ANS with a-lactalbumin at pH 7.0 in the presence of the 0.25 mol  dm3 (HFIP + GndSCN) mixture and 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture were carried out. Here also, a model of two sequential binding sites fits adequately to the calorimetric data and the binding may be represented as, a-LA þ ANS ¼ ½a-LA  ANS; ½a-LA  ANS þ ANS ¼ ½a-LAðANSÞ2 :

ð15Þ ð16Þ

1150

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

Existence of a pair of sequential binding site means that binding of ligand to one site will be influenced by whether or not ligands are bound to the other site. The thermodynamic parameters extracted from the model fitting are summarized in table 7 which indicates that the binding of ANS molecules to both the binding sites in the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture is endothermic with a positive entropy change and also that the binding constant K2 for the second site (high affinity site) is higher than the first binding site. For both the sites, binding is entropically driven since the values of both DH and DS are positive. The low values of enthalpy of binding and binding constants suggest that the binding is very weak. Energetic contributions of many individual interactions such as hydrogen bonds, van der Waals, electrostatic, polar, and dipolar interactions, etc. between the two binding molecules are reflected by the enthalpy term. The entropy represents the gain in degrees of freedom of the water molecules that prior to the binding are localized in the surface of the binding molecules and are released to the bulk solvent upon binding. It is usually positive (favourable), since the binding molecules have a greater percentage of non-polar than polar surface area buried upon binding [39]. In the present case also, significantly positive values of the entropy of binding are obtained. The binding of a-lactalbumin at pH 7.0 with ANS in the presence of the 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture also shows best nonlinear fitting to two sequential binding sites giving the best fitting thermodynamic parameters which are summarized in table 7. The values of heat change involved in the binding process are very small, yet they show a definite trend. Again the binding seemed to be governed mainly by the entropic contribution (table 7). It was observed that the first binding site is the high affinity binding site, with K1 higher than K2 (table 7). This type of binding can be best described as that exhibiting negative cooperativity and is marked by appearance of two phases in the binding isotherm, corresponding to saturation of the first site, and then the second site. Comparing the binding constants in the presence of the two different (alcohol + salt) mixtures, we infer that the two partially folded states of a-lactalbumin are different in nat-

TABLE 7 Association constant (K), enthalpy (DH), and entropy (DS) of binding of ANS to a-lactalbumin in the presence of the 0.25 mol  dm3 (HFIP + GndSCN) equimolar mixture and 0.85 mol  dm3 (isopropanol + GndSCN) equimolar mixture at pH 7.0 and T = 298.15 K according to two sequential binding sites model described by equations (15) and (16)

K1/dm3  mol1 DH1/(kJ  mol1) DS1/(J  K1  mol1) K2/dm3  mol1 DH2/(kJ  mol1) DS2/(J  K1  mol1)

0.25 mol  dm3 (HFIP + GndSCN)

0.85 mol  dm3 (isopropanol + GndSCN)

858 ± 220 4.4 ± 1.0 70 (1.120 ± 0.25)  103 2.1 ± 0.2 65

(9.35 ± 3.0)  104 1.3 ± 0.1 100 68.7 ± 7.4 200 ± 60 706

ure as they exhibit a different binding affinity towards ANS. This difference in the binding parameters suggests that these two sets of co-solutes lead to conformational states of protein which are partially folded to different extent. The binding parameters and more importantly the small heat changes for the binding of ANS with the suggested partially folded intermediate states indicate that the nature of binding is very weak. This further corroborates that in the presence of the 0.25 mol  dm3 (HFIP + GndSCN) and 0.85 mol  dm3 (isopropanol + GndSCN) mixtures a-LA exists in partially folded conformations which are different from the conventional MG state where a very strong binding with ANS has been observed [40]. 3.3. Possible binding sites for ANS In principle, binding of ANS to a protein surface is dictated by a variety of non-covalent interactions such as hydrophobic interactions, electrostatic interactions, van der Waals interactions, and hydrogen bond interactions. Amongst these, hydrophobic interactions have been assumed to play a significant role in the binding processes [41,42]. The crystal structure of a-lactalbumin is well established and is known to have a hydrophobic region, aromatic cluster I and a calcium binding site as the rigid and functional part of the enzyme. The hydrophobic region is comprised of residues 53(Phe), 55(I), 95(I), 103(Tyr), and 104(Trp) from the C and D helices of a domain and from the b domain, whereas the aromatic cluster I consists of residues 31(Phe), 32(His), 117(Gln) and 118(Trp). The Phe31 and Gln117 are relatively exposed, whereas His32 and Trp118 are somewhat shielded from the solvent by the surrounding residues [43]. We suggest here that the proposed partially folded states of a-lactalbumin could be binding with ANS through these two regions which may be getting exposed due to the denaturing influence of the additives. There is a reason to believe this since the two tryptophans in these regions (Trp104 and Trp118) have been known to play an important role in formation of the partially folded state and have been shown to be exposed to different levels in the quenching experiments (%Trp exposure being different for each of the proposed partially folded state) indicating that these regions might also be exposed to different levels in presence of different additive mixture and their concentrations. However, importance of the electrostatic interactions in ANS binding has also been reported in the literature [44]. Even though the major force of binding here is the result of hydrophobic interactions (since the binding is endothermic in nature), the electrostatic interactions through the sulphonate group of ANS with the cationic groups (histidine, arginine, lysine chains etc.) donated by the protein, also cannot be ruled out. This is also supported by our fluorescence experiments where a very low intensity of binding is observed. The sulphonate group of ANS could possibly be forming an ion pair with guanidine group of GndSCN, leading to a lesser

R. Sharma, N. Kishore / J. Chem. Thermodynamics 40 (2008) 1141–1151

extent of binding with the exposed hydrophobic surface of the protein. 4. Conclusions The strong denaturing influence of the guanidine salt and the local secondary structure inducing ability of alcohol leads to the formation of equilibrium partially folded states of protein which are relatively stable than the other intermediates occurring during the course of protein folding. The thermal unfolding studies of a-lactalbumin in the presence of equimolar (alcohol + GndSCN) mixtures has permitted the determination of quantitative thermodynamic parameters accompanying all the transitions. Circular dichroism and fluorescence spectroscopic results corroborate the thermal unfolding results and indicate protein conformations with poorly defined tertiary structure in the presence of the 0.25 mol  dm3 HFIP and GndSCN mixture and the 0.85 mol  dm3 isopropanol and GndSCN mixture. Our results also correlate well with the hydrophobicity difference between HFIP and isopropanol. Isothermal titration calorimetric measurements have demonstrated that binding of ANS with the two partially folded states is non-specific as the binding constants for the two binding sites are different and seems to be governed mainly by the entropic contribution. The binding isotherms show best nonlinear fitting to two sequential binding sites. Acknowledgements This work was supported by funding from the Department of Science and Technology New Delhi, India. R.S. is supported by a senior research fellowship from the Council of Scientific and Industrial Research, India. References [1] F. Edwin, Y.V. Sharma, M.V. Jagannadham, Biochem. Biophys. Res. Commun. 290 (2002) 1441–1446. [2] Y.O. Kamatari, D. Konno, T.M. Kataoka, K. Akasaka, J. Mol. Biol. 259 (1996) 512–523. [3] A.L. Fink, L.J. Calciano, Y. Goto, T. Kurotsu, D.R. Palleros, Biochemistry 33 (1994) 12504–12511. [4] V. Forge, R.T. Wijesinha, J. Balbach, K. Brew, C.V. Robinson, C. Redfield, C.M. Dobson, J. Mol. Biol. 288 (1999) 673–688. [5] W.A. Eaton, V. Munoz, S.J. Hagen, G.S. Jas, L.J. Lapidus, E.R. Henry, J. Hofrichter, Annu. Rev. Biophys. Biomol. Struct. 29 (2000) 327–359. [6] S. Chakraborty, Z.Y. Peng, J. Mol. Biol. 298 (2000) 1–6. [7] S.J. Demarest, R. Fairman, D.P. Raleigh, J. Mol. Biol. 283 (2000) 279–291. [8] Z.Y. Peng, P.S. Kim, Biochemistry 33 (1994) 2136–2141. [9] J.W. Nelson, N.R. Kallenbach, Proteins 1 (1986) 211–217.

1151

[10] P.D. Thomas, K.A. Dill, Protein Sci. 2 (1993) 2050–2065. [11] K.A. Dill, S. Bromberg, K. Yue, K.M. Fiebig, D.P. Yee, P.D. Thomas, H.S. Chan, Protein Sci. 4 (1995) 561–602. [12] Y. Liu, D.W. Bolen, Biochemistry 34 (1995) 12884–12891. [13] M. Buck, H. Schwalbe, C.M. Dobson, Biochemistry 34 (1995) 13219– 13232. [14] J.W. Nelson, N.R. Kallenbach, Biochemistry 28 (1989) 5256– 5261. [15] T. Banerjee, N. Kishore, J. Phys. Chem. B 47 (2005) 22655. [16] A. Kundu, N. Kishore, Biopolymers 73 (2004) 405–420. [17] J.R. Cort, N.H. Andersen, Biochem. Biophys. Res. Commun. 233 (1997) 687–691. [18] H.S. Mchaourab, J.S. Hyde, J.B. Feix, Biochemistry 32 (1993) 11895– 11902. [19] A. Galat, J.P. Degelaen, C.C. Yang, E.R. Blout, Biochemistry 20 (1981) 7415–7423. [20] I. Sirangelo, F.D. Piaz, C. Malmo, M. Cassilo, L. Birolo, P. Pucci, G. Marino, G. Irace, Biochemistry 42 (2003) 312–319. [21] N. Hirota, K. Mizuno, Y. Goto, Protein Sci. 6 (1997) 416–421. [22] S.W. Snyder, U.S. Ladror, W.S. Wade, G.T. Wang, L.W. Barrett, E.D. Matayoshi, H.J. Huffaker, G.A. Krafft, T.F. Holzman, Biophys. J. 67 (1994) 1216–1228. [23] H. Zhang, K. Kaneko, J.T. Nguyen, T.L. Livshits, M.A. Baldwin, F.E. Cohen, T.L. James, S.B. Prusiner, J. Mol. Biol. 250 (1995) 514– 526. [24] S.J. Demarest, J.A. Boice, R. Fairman, D.P. Raleigh, J. Mol. Biol. 294 (1999) 213–221. [25] M.J. Kronmann, R.E. Andreotti, R.V. Vitols, Biochemistry 3 (1964) 1152–1160. [26] W.H. Kirchoff, EXAM, U.S. Department of Energy, Thermodynamics Division, National Institute of Standards and Technology, Gaithersburg, MD, USA. [27] E.L. Kovrigin, S.A. Potekhin, Biochemistry 36 (1997) 9195–9199. [28] L. Stryer, J. Mol. Biol. 13 (1965) 482–495. [29] R.S. Spolar, J.R. Livingstone, M.T. Record, Biochemistry 31 (1992) 3947–3955. [30] K.P. Murphy, V. Bhakuni, D. Xie, E. Freire, J. Mol. Biol. 227 (1992) 293–306. [31] E.L. Kovrigin, S.A. Potekhin, Biochemistry 36 (1997) 9195–9199. [32] G. Velicelebi, J.M. Sturtevant, Biochemistry 18 (1979) 1180–1186. [33] O. Stern, M. Volmer, Phys. Z. 20 (1919) 183–188. [34] A.B. Rawitch, R.Y. Hwan, Biochem. Biophys. Res. Commun. 91 (1979) 1383–1389. [35] S.S. Lehrer, Biochemistry 10 (1971) 3254–3263. [36] R.M. France, S.H. Grossman, Biochem. Biophys. Res. Commun. 269 (2000) 709–712. [37] E.D. Chrysina, K. Brew, K.R. Acharya, J. Biol. Chem. 275 (2000) 37021–37029. [38] H.A. Scheraga, G. Nemethy, I.Z. Steinberg, J. Biol. Chem. 237 (1962) 2506–2508. [39] A.V. Campoy, E. Friere, Biophys. Chem. 115 (2005) 115–124. [40] S.K. Singh, N. Kishore, Biopolymers 83 (2006) 205–212. [41] K.K. Sharma, H. Kaur, G.S. Kumar, K. Kester, J. Biol. Chem. 273 (1998) 8965–8970. [42] G. Vanderheeren, I. Hanssens, K. Noyelle, H. Van Dael, M. Joniau, Biophys. J. 75 (1998) 2195–2204. [43] C.W.A. Pike, K. Brew, K.R. Acharya, Structure 4 (1996) 691–703. [44] D. Matulis, R. Lovrien, Biophys. J. 74 (1998) 422–429.

JCT 08-2