Volume, expansivity and isothermal compressibility changes associated with temperature and pressure unfolding of staphylococcal nuclease1

doi:10.1006/jmbi.2001.4517 available online at http://www.idealibrary.com on

J. Mol. Biol. (2001) 307, 1091±1102

Volume, Expansivity and Isothermal Compressibility Changes Associated with Temperature and Pressure Unfolding of Staphylococcal Nuclease Heiko Seemann1, Roland Winter1* and Catherine A. Royer2 1

University of Dortmund Department of Chemistry Physical Chemistry I, OttoHahn-Str. 6, D-44227 Dortmund, Germany 2

Centre de Biochimie Structurale, Faculte de Pharmacie, 15 ave. Charles Flahault, F-34060, Montpellier France

We have characterized the temperature- and pressure-induced unfolding of staphylococcal nuclease (Snase) using high precision densitometric measurements. The changes in the apparent speci®c volume, expansion coef®cient and isothermal compressibility were determined by these measurements. To our knowledge, these are the ®rst measurements of the volume and isothermal compressibility changes of a protein undergoing pressure-induced unfolding. In order to aid in interpreting the temperature and pressure dependence of the apparent speci®c volume of Snase, we have also carried out differential scanning calorimetry under the solution conditions which are used for the volumetric studies. We have seen that large compensating volume and compressibility effects accompany the temperature and pressure-induced protein unfolding. Measurements of the apparent speci®c volume and thermal expansion coef®cient of Snase at ambient pressure indicate the formation of a pretransitional, molten globule type of intermediate structure about 10  C below the actual unfolding temperature of the protein. Compared to the folded state, the apparent speci®c volume of the unfolded protein is about 0.3-0.5 % smaller. In addition, we investigated the pressure dependence of the apparent speci®c volume of Snase at a number of different temperatures. At 45  C we calculate a decrease in apparent speci®c volume due to pressure-induced unfolding of ÿ3.3 10ÿ3 cm3 gÿ1 or ÿ55 cm3 molÿ1. The threefold increase in compressibility between 40 and 70 MPa re¯ects a transition to a partially unfolded state, which is consistent with our results obtained for the radius of gyration of the pressuredenatured state of Snase. At the lower temperature of 35  C, a signi®cant increase in compressibility around 30 MPa is indicative of the formation of a pressure-induced molten globule-like intermediate. Changes in the apparent volume, expansion coef®cient and isothermal compressibility are discussed in terms of instrinsic, hydrational and thermal contributions accompanying the unfolding transition. # 2001 Academic Press

*Corresponding author

Keywords: staphylococcal nuclease; unfolding; speci®c volume; expansion coef®cient; isothermal compressibility

Introduction A detailed understanding of the structural, dynamic and thermodynamic properties of the unfolded states of proteins has been a major topic of research in the ®eld of protein chemistry and Abbreviations used: Snase, staphylococcal nuclease; DSC, differential scanning calorimetry; FT-IR, Fourier transform infrared. E-mail address of corresponding author: [email protected] 0022-2836/01/041091±12 $35.00/0

biophysics in recent years. Partial volume and its temperature and pressure derivatives are basic thermodynamic parameters determining protein stability and unfolding. A large body of empirical data on the partial speci®c volume and adiabatic compressibility of folded proteins and their lowmolecular mass constituents has been accumulated (for a review, see1,2). However, only very few volumetric studies of the denatured state and on the unfolding process have been undertaken so far.2-17 Here, we present data obtained from a volumetric study of the temperature- and pressure-induced # 2001 Academic Press

1092

Unfolding of Snase

unfolding of staphylococcal nuclease (Snase). This protein has been widely used as a model in conformational and protein folding studies .18-20 Snase is a small globular protein of molecular mass 16.8 kDa which contains 149 amino acid residues and no disul®de bonds. The backbone of Snase forms a ®ve-strand b-barrel packed against three approximately parallel a-helices.21 The fact that proteins can be unfolded by pressure implies by the general thermodynamic relationship:   d ln K
V  ˆÿ RT dp T

…1†

(K is the equilibrium constant of the unfolding reaction, at standard state  ) that at constant temperature an overall negative volume change should be involved in the unfolding process. This implies that the volume change
V  accompanying the pressure-induced unfolding of a protein must be negative at and above the unfolding pressure (for a review, see, e.g., 22-25). On the contrary, at ambient pressure, the volume change may be either negative or positive, depending on the sign and the magnitude of the change in protein compressibility associated with unfolding.4,9 Several proteins studied to date exhibit a decrease in speci®c volume upon unfolding. The underlying molecular basis for this decrease in system volume upon unfolding has been attributed to the combined effects of electrostriction, hydrophobic hydration and the elimination of packing de®ciencies upon unfolding.6,9,22,26-28 Often, the volume changes observed are very close to zero, however, and in some cases also small positive values have been found.29 The characteristic differences between the native and denatured structures are probably more sensitively manifested themselves in the temperature and pressure derivatives of the speci®c volume. The partial isothermal compressibility is an important physical quantity directly related to the conformational state of a protein molecule through the volume ¯uctuations, and thus sensitively re¯ects the changes in structural characteristics of the protein upon unfolding. Thus far, only the partial adiabatic compressibility kS. of the solute has been measured using sound velocity measurements.30-33,12 The partial adiabatic compressibilities: kS ˆ ÿ

  1 dV  V  dp S

…2†

of the solutions and solvent were calculated from the sound velocity, u, and the density, d, of the solution and solvent, respectively, according to kS ˆ 1/(d u2). Vo is the partial speci®c volume of the solute (protein) in this case. Only very limited data have been reported on the adiabatic compressibility change associated with protein unfolding.3,4,6,33,34 Generally, small or negative

values were found for acid, base and guanidinium hydrochloride-induced unfolding reactions. Important aspects relating to volumetric properties of proteins are the hydration-volume and hydration-compressibility relationships. The major part of the change in volume and compressibility upon un/refolding of proteins is probably due to hydration processes. For these reasons, applying temperature or pressure changes may be more appropriate methods for such studies than using strong chemical denaturants or pH. Here, we present data of a volumetric study of protein conformational transitions and characterize and compare the thermodynamic parameters derived for the temperature- and pressure-induced unfolding of Snase. To our knowledge, these are the ®rst measurements of the volume and isothermal compressibility changes of a protein undergoing pressure-induced unfolding.

Results Thermal unfolding The thermal unfolding of Snase was measured by means of a differential scanning micro-calorimeter. Runs were carried out at protein concentrations ranging from 0.25 to 5 % by weight and in 10 mM bis-Tris buffer at pH 7.0. No signi®cant changes were observed either in the onset temperature or in the value of the enthalpy change between different concentrations. In Figure 1 is shown the scan obtained at pH 7 and at 0.25 wt% Snase. The protein begins to unfold (onset temperature at 43 ( 0.5)  C and the temperature of the unfolding midpoint, Tm, was found to be 52 ( 0.1)  C. We observed a decrease of the midpoint of the transition with increasing concentration, which may be due to some protein aggregation. The enthalpy change amounts 164 ( 5) kJ/mol. We ®nd an increase in Cp of 8.3 ( 1.1) kJ molÿ1 Kÿ1 between the native and unfolded state at 65  C. Figure 2(a) shows the temperature dependence of the apparent speci®c volume, V*, of Snase at ambient pressure. The V* values increase with increasing temperature between 15 and 35  C, in a gradual, non-linear fashion, with a slope that decreases with increasing temperature. In the pretransition and transition range (between 37.5  C and 50  C) the apparent speci®c volume exhibits complex behavior. First we observe an increase in slope between 37.5 and 42.5  C, whereupon the value of V* (not the slope) actually decreases at temperatures up to 47.5  C. Above that temperature and in the post-transition region a gradual linear increase of V* is observed. Multiple repetitions of the temperature runs revealed that these features were entirely reproducible. In Figure 2(b), the actual apparent thermal expansion coef®cient, a, is plotted as a function of temperature. Since a corresponds to (1/V*) (dV* /dT), the volume normalized derivative of

1093

Unfolding of Snase

Figure 1. DSC trace of a 0.25 wt% solution of Snase at pH 7 in bis-Tris buffer (scan rate: 1 deg. C/minute).

the temperature dependence of the speci®c volume, the complexity observed in Figure 2(a) in the transition region is even more apparent than in the plot of V* versus. T (Figure 2(a)). Between 15 and 35  C the value of a decreases from a value near (8.9  1.6)  10ÿ4 Kÿ1 to a value near (2.6  0.5)  10ÿ4 Kÿ1 at 35  C. At approximately 37.5  C, just prior to the onset of thermal unfolding as detected in the DSC scan, a signi®cant increase in the thermal expansion coef®cient is observed. This increase is maximized at 40  C and then a large decrease to negative values of a is observed up to 45  C. This temperature is still below the midpoint in unfolding as detected by DSC. Above 45  C, a increases to a relatively constant level in the post-transition temperature range. If we assume that the Snase remains in the folded state between 15 and 35  C, then the slope of the plot of the apparent speci®c volume versus. temperature (dV*/dT) (Figure 2(a)) and a ((1/V* )(dV*/dT)) (Figure 2(b)) corresponds to the thermal expansivity of the folded state. And if we assume that since the thermal unfolding transition is complete around 55  C then we can assume that above that temperature, the slope of the plot of the apparent speci®c volume versus temperature (Figure 2(a)) or a (Figure 2(b)) correspond to the unfolded state. The thermal expansivity of the unfolded state appears to be fairly constant (V* versus temperature is linear between 55 and 65  C), whereas that for the folded state decreases with increasing temperature. The non-linearity of the low temperature section of the pro®le may arise from the energetic constraints on the volume ¯uctuations (dV*/V*) imposed by the structure of the folded state. The increase of the slope of the speci®c volume versus temperature plot (Figure 2(a)) observed in

the pre-transition zone (37.5-42.5  C) indicates the formation (at least fractional) of an intermediate that is somewhat expanded. Likewise, the large increase in the apparent thermal expansivity in this range (Figure 2(b)) also indicates the population of an intermediate which expands more readily with increasing temperature than does the folded state. The decrease in the value of the apparent speci®c volume (negative expansivity value) between 42.5 and 47.5  C is directly due to the decrease in volume upon unfolding at these temperatures. Pressure denaturation Figure 3 shows data for the apparent speci®c volume of Snase as a function of pressure V*(p) between atmospheric pressure and 70 (or 100) MPa for selected temperatures between 25 and 55  C. The data for T ˆ 25  C cover the purely native, folded state of the protein and the data shown for T ˆ 55  C are within the denatured, unfolded state, while the plots obtained at the intermediate temperatures include contributions from the folded, intermediate and unfolded states in varying proportions. In the folded state at 25  C the value of kT, as obtained from the V*(p) data shown in Figure 3, decreases from (11  2)  10ÿ11 Paÿ1 at ambient pressure to (5.4  2.9)  10ÿ11 Paÿ1 (at 70 MPa) with increasing pressure over this range, indicating that the internal constraints of the folded state limit the effect of pressure. As shown in Figure 4, the behavior of the pressure dependence of kT at 35  C is quite complex. For example, the value of kT at atmospheric pressure and up to 20 MPa is tenfold lower than that observed at 25  C, (1.0  0.5)10ÿ11 Paÿ1. Above 20 MPa, the value of kT undergoes an eightfold

1094

Unfolding of Snase

dependence of the apparent speci®c volume) we calculate the decrease in apparent speci®c volume due to the unfolding transition, itself, at 100 MPa as ÿ3.3  10ÿ3 cm3 gÿ1 or ÿ55 cm3 molÿ1.

Discussion

Figure 2. (a) Temperature dependence of the apparent speci®c volume V* of a 4 wt% solution of Snase at pH 7 (in bis-Tris buffer) and p ˆ 1 bar, and (b) temperature dependence of the apparent speci®c expansion coef®cient a of Snase at pH 7 (in bis-Tris buffer) calculated from the data shown in the upper curve. The dotted line indicates linear extrapolation of the low temperature region (see the text).

increase to (8.2  1.0)  10ÿ11 Paÿ1 at 35 MPa, and then decreases again gradually with pressure to near (3.1  1.0)  10ÿ11 Paÿ1 at 70 MPa. At 45  C, Snase undergoes its unfolding transition at pressures between 70 and 110 MPa.20 For this reason we sought to extend the pressure range of the volumetric measurements at this temperature well beyond the 70 MPa limit indicated for the apparatus. In Figure 5 is replotted the pressure dependence of V* at 45  C up to 100 MPa, and in Figure 6, is shown the pressure dependence of the apparent isothermal compressibility. We see that indeed the apparent speci®c volume decreases signi®cantly between 60 and 100 MPa, corresponding to the expected decrease in volume upon unfolding. At 45  C, kT decreases with pressure up to 30 MPa, then increases by approximately a factor of 2-3, and then remains essentially constant above the transition temperature. Snase is more than 80 % unfolded at 100 MPa and 45  C.20 Extrapolating the low pressure slope to 100 MPa (which approximates the contributions of the compressibilities of the folded and unfolded states to the pressure

In order to aid in interpreting the temperature and pressure dependence of the apparent speci®c volume of Snase, we have carried out differential scanning calorimetry under the solution conditions which are used for the volumetric studies. According to our DSC results, the thermal unfolding of staphylococcal nuclease occurs with a Tm of 52  C, an enthalpy change of 164 kJ molÿ1 and an increase in heat capacity of 8.3 kJ molÿ1 Kÿ1 at 65  C. The pro®le we obtain is quite similar to those reported by Privalov and co-workers,35 although at somewhat lower pH. The lack of any additional salt in our buffer may suppress the pHdependent increase in the enthalpy change and transition temperature observed previously.35 As previously discussed by Chalikian et al.,4-6,9 changes in partial speci®c volume, V , of a protein can be considered to be dissected into essentially three different contributions: (1) the intrinsic volume, Vintr, which originates from the van der Waals volume of the constituent atoms plus the volume of intrinsic voids within the water-inaccessible protein interior, (2) a hydrational term, dVhydr, or interaction volume, describing the solvent volume associated with the hydration of solvent-accessible protein atomic groups, i.e. from solute-solvent interactions around the charged (electrostriction), polar (hydrogen-bonding), and non-polar (hydrophobic hydration) atomic groups on the protein surface, and (3) the thermal volume, Vtherm, which results from thermally induced mutual molecular vibrations and re-orientations. A similar equation for the apparent speci®c volumetric contributions Vi* is assumed. This leads to:    ‡ dVhydr ‡ Vtherm V  ˆ Vintr

…3†

Since the constitutive atomic volume of a protein does not change signi®cantly with temperature, the observed temperature dependence of the apparent speci®c volume can be ascribed essentially to the modi®cation of the internal voids (
V*intr ˆ
V*void), to hydration changes on the protein surface and to changes in the thermal volume:    ‡
dVhydr ‡
Vtherm V  ˆ
Vvoid

…4†

The apparent speci®c expansion coef®cient of the native protein, a, of about (5.2  0.5)  10ÿ4 Kÿ1 at 25  C is on the order of that which is generally observed for native proteins.8,11,29 We have found that the apparent coef®cient of thermal expansion of Snase decreases with increasing temperature in the range of temperatures (15-35  C) over which the protein remains in the folded state

Unfolding of Snase

1095

Figure 3. Pressure dependence of the apparent speci®c volume V*(p) of Snase at pH 7 (4 wt%, in bis-Tris buffer) for four selected temperatures, 25, 35, 45 and 55  C. For clarity the V*(p) curves are offset against each other.

(as assessed by the DSC measurements). The apparent speci®c volume of the protein increases with increasing temperature, mostly due to thermal volume and perhaps increases in the internal void volume, but the more the temperature is increased, the smaller its effect. This is likely due to the constraints on the expansion of voids and thermal volume imposed on the protein by the interactions in the folded structure. Such constraints on the temperature dependence of tryptophan dynamics in the interior of proteins have been reported (36 and references therein). The decrease in a with temperature in the native state of the protein resembles the behavior of a for bulk water with decreasing temperature, and might be expected for a hydrated hydrophobic particle,

whose structural and dynamic properties resemble those of bulk water at much lower temperatures. In the pre-transition range of temperatures, the apparent speci®c volume and the apparent thermal expansivity exhibit a signi®cant and reproducible increase. Such behavior indicates the population (at least fractional population) of an expanded form of the protein which has fewer internal constraints (softer structure). In addition to thermal volume, an increase in internal void volume is likely for this intermediate. A noticeable result is that the apparent speci®c volume and expansion coef®cient in the transition region seem to deviate from the transition curves expected for a simple two-state unfolding model. Snase has been shown by circular dichroism, NMR and ¯uorescence to

Figure 4. Pressure dependence of the apparent isothermal compressibility of Snase at 35  C (4 wt%, in bis-Tris buffer). The data for three different measurements with different sample preparations are shown (different symbols).

1096

Unfolding of Snase

Figure 5. Pressure dependence of the apparent speci®c volume of Snase at 45  C (4 wt%, in bis-Tris buffer). The dotted lines indicate linear extrapolations of the low and high pressure regions (see the text).

populate intermediates during kinetic refolding experiments at atmospheric pressure.37-41 Mutants of Snase also clearly exhibit intermediate states in equilibrium unfolding experiments as assessed by circular dichroism and calorimetry.35 These equilibrium intermediates arise from either a stabilization of the beta-barrel subdomain or a destabilization of the alpha-helical sub-domain by the mutation such that at intermediate denaturant concentrations or pH values, a form (or forms) of the protein exists in which the beta-barrel domain remains structured, while the alpha-helical subdomain is disrupted, while the kinetic intermediates implicate at least in part cis/trans prolyl peptide bond isomerizations. Although intermediates have

not been observed for the wild-type protein at equilibrium using the above methods previously, it may be that the sensitivity of the volumetric measurements allow for its detection. Above the transition temperature (between 55 and 65  C) in the unfolded state, the constant increase in volume with temperature likely arises from a dominant contribution from the thermal volume term. In the range of temperatures over which the transition from the folded to unfolded form occurs, the apparent speci®c volume and the apparent thermal expansivity both decrease signi®cantly. The decrease in volume due to the unfolding reaction between 42.5 and 47.5  C is offset to some degree by the increase in apparent speci®c volume

Figure 6. Pressure dependence of the apparent isothermal compressibility kT of Snase at 45  C (4 wt%, in bis-Tris buffer). The data for three different measurements with different sample preparations are shown (different symbols).

1097

Unfolding of Snase

over this same range due to the thermal expansion of the unfolded, and perhaps to a small extent, folded, state. Since the thermal unfolding is apparently not entirely two-state, and an expanded intermediate is populated in the pre-transition zone, the small decrease of V* above 42.5  C might be due to loss of the void volume in the partially expanded intermediate. Similar to the apparent speci®c volume, the apparent expansion coef®cient is a composite quantity, comprised of several contributions that may differ in sign. Upon unfolding, a redistribution of bound and free water occurs concomitant with the corresponding changes in the relative contributions. Assuming negligibly small expansion of the van der Waals volume of the constitutive atoms of the protein, a for any state of the protein can be described as the temperature derivatives of the void volume, V*void, the volume change due to the hydration contribution, dV*hydr, and the volume change due to thermal excitation of vibrational and reorientational motions, V*therm: 

  ddVhydr dVtherm dV  dVvoid ˆ ‡ ‡ dT dT dT dT

…5†

In the folded state, the ®rst term is probably small but positive due to a decrease in the local atomic packing density with increasing temperature. In the unfolded state this ®rst term will be essentially negligible. The second term is positive for both states, although a larger positive dV*therm/dT value is expected for the unfolded state compared to the folded state due to the loss upon unfolding of the energetic constraints present in the folded state. The third term, ddV*hydr/dT, should become more positive (or less negative) with increasing temperatures due to a decrease in H-bonding interactions, and this effect should be larger in the denatured state since the hydration is greater. The hydration around polar and non-polar groups is more sensitive to temperature variation than electrostricted hydration. Therefore, the increases in these two types of hydration should contribute to make ddV*hydr/dT more positive. Thus, both the thermal and hydration term would positively contribute to an increased expansion coef®cient of the thermally denatured protein. Assuming that the essentially constant value of the thermal expansion coef®cient for the fully unfolded form measured above 55  C is valid over the entire temperature range investigated, one can use a linear extrapolation of the high temperature (55-65  C, unfolded form) slope of the V* versus T plot back to 15  C (dotted lines in Figure 2(a)) to obtain an estimate of the apparent speci®c volume of the unfolded state over the entire range of temperatures tested. For other proteins, the V* values in the pre- and post-tansitional branches of the V*(T) curves often appeared to be linearly dependent on temperature as found by other groups (42,29 and references therein). Between 15 and 35  C, the estimated value

of the apparent speci®c volume of the unfolded state is smaller than the observed value of the folded state. The values of these negative volume changes for unfolding below 35  C range from ÿ0.002 to ÿ0.004 cm3gÿ1 (ÿ40 to ÿ60 cm3 molÿ1), which corresponds to about 0.3-0.5 % of the speci®c volume of the protein. If we assume that the apparent thermal expansivity of the folded state either remains at the value determined at 35  C or even decreases to zero (Figure 2(b), broken line), then at some point, as temperature increases above 35  C, the apparent speci®c volume of the unfolded state would become larger than that of the folded state, and at these higher temperatures, the volume change upon unfolding would then become positive. We have previously proposed 20,43,44 a difference in coef®cients of thermal expansion between the folded and unfolded states as the basis for the pressure-induced folding of proteins at high temperatures, and the decrease in the absolute value of the volume change upon unfolding with increasing temperature.20,43 These small values for the volume change upon unfolding, which can be negative or positive, depending upon the temperature in the case of Snase and on the protein in a more general sense, must therefore result from very large compensations between the change in V*intr, dV*hydr, and V*therm.6,9 Disruption of the secondary structure is accompanied by replacement of peptide-peptide hydrogen bonds by peptide-water hydrogen bonds. The resulting increase in hydration around the polar groups should contribute to a decrease in the apparent speci®c volume of the denatured protein since the volume changes due to polar hydration are negative.12 On average, the dV*hydr contribution to V* of a globular protein is found to be signi®cant, it may be as large as about 8 % for Snase.5,17 The contribution of electrostricted hydration to the observed volume changes should be small since most ionic groups are already located on the protein surface in the folded state. The contribution of the polar hydration term upon unfolding should be negative and signi®cant. The sign and magnitude of the changes in volume upon hydration of hydrophobic residues, however, is still subject to debate. Although transfer studies of liquid hydrocarbons show a decrease in volume upon transfer to water,45,46 the applicability of such studies to protein folding is compromised by the fact that the packing density of hydrophobic residues in the protein interior is much greater than the density of liquid hydrocarbons.47,48 Moreover, the structure of water around exposed hydrophobic groups may be quite different than bulk water. In fact, transfer studies of water to liquid hydrocarbons indicate that the speci®c volume of water may increase in hydrophobic media.46 The change in the internal void volume on thermal unfolding depends upon the degree to which the structure is disrupted and the access of solvent to any residual structure. Our recent temperature SAXS analysis of Snase19 showed that the radius of

1098

Unfolding of Snase

Ê in the native gyration, Rg, increases from 17 A Ê state to 43 A in the thermally unfolded state, which is only slightly smaller than the value of a hypothetical random-coil conformation (about Ê ). These results suggest that very little internal 46 A void volume remains in the thermally unfolded protein. Since the electrostriction and hydrogen bonding contribution to the hydration term, and the cavity loss contribution are negative, the increase in thermal volume and perhaps a positive contribution from hydrophobic hydration represent the missing positive contribution to the volume change accompanying thermally induced protein unfolding. We have also investigated the pressure dependence of the speci®c volume of Snase at a number of different temperatures. At 25  C the protein remains in the folded state up to 60 MPa26 and the derivative of the decrease in apparent speci®c volume yields the apparent isothermal compressibility of the folded state at that temperature. At 55  C the protein is essentially unfolded and the derivative of the plot of speci®c volume versus pressure yields the apparent isothermal compressibility of the unfolded state at that temperature. We note that although these values are not constant with pressure they are fairly similar, between (11  2)  10ÿ11 Paÿ11 and (7.5  5)  10ÿ11 Paÿ1 at 25 and 55  C for ambient pressure, respectively. For T ˆ 60  C a similar apparent isothermal compressibility value of (10  5)  10ÿ11 Paÿ1 has been obtained. The kT value at 25  C is very consistent with the values of the adiabatic compressibility reported previously. The two thermodynamic quantities are related as follows: 2

kT ˆ kS ‡ a TV  =Cp

…6†

(a is the partial expansion coef®cient). At 25  C a value of (2.9  1.2)  10ÿ11 Paÿ1 has been reported for the native state value of kS of Snase.17 Taking Vo ˆ 0.763 cm3 gÿ1,17 a ˆ (5.2  0.5)  10ÿ4 Kÿ1, and Cp ˆ 1.25 J gÿ1 Kÿ1 49 we calculate a value of (7.9  2.1) 10ÿ11 Paÿ1 for kT, which is well within the range of values reported here for the apparent isothermal compressibility (kTˆ(11  2)  10ÿ11 Paÿ1). We assume that the apparent isothermal compressibility coef®cient, kT, also consists of the different contributions mentioned above:       1 dV  1 dVvoid 1 dVtherm ˆ ÿ  ‡ ÿ  kT ˆ ÿ  V dp V dp V dp !  1 ddVhydr …7† ‡ ÿ  dp V The intrinsic contribution, kT,intr, which consists mainly of the void contribution kT,void, is positive. The thermal contribution kT,therm, is positive as well and the hydration contribution, dkT,hyr, is

negative. The intrinsic contribution, kT,void, re¯ects the compressible atomic contacts due to imperfect packing in the solvent-inaccessible interior of proteins. The thermal contribution, kT,therm, re¯ects the damping of molecular vibrational and reorientational motions under pressure, and the hydration contribution, kT,hydr, depends on the difference in compressibility between water in the hydration shell and in bulk water. The hydration of protein surfaces is known to signi®cantly decrease the compressibility of the solution at room temperature because of the compact packing of water molecules in the hydration shell.12,13,15 The higher the total solvent-accessible protein surface area, the more negative the hydration contribution. The negative value of dV*/dp and hence the apparent isothermal compressibility coef®cient slightly decreases with increasing pressure in the folded state (25  C). This is likely to arise from a decreasing value of all three contributions in equation (7), that of the void volume probably being of greatest importance. The apparent isothermal compressibility kT slightly increases with pressure at 55  C, from (7.5  5)  10ÿ11 to (15  4)  10ÿ11 Paÿ1. This might be due to a further unfolding of residual folded protein with increasing pressure at these temperatures. At such high temperatures, the hydration contribution to equation (7), although negative, is likely to be much smaller than hydration effects on speci®c volume at lower temperatures. Given the value of the radius of gyration of Snase measured at 55  C and ambient pressure,20 which is quite close to that expected for a random coil conformation, the contribution of compression of void volumes in the unfolded form of the protein to the pressure dependence of the compressibility is expected to be negligible at this temperature. However, if a small fraction of folded protein persists, which is indeed the case (see below), then the thermal ¯uctuations may increase slightly as this residual folded protein undergoes pressureinduced unfolding, exposing its side chains. This increase in apparent isothermal compressibility with pressure at high temperatures must necessarily be suppressed at higher pressures (unfortunately beyond the range of the high pressure densitometer used here). As expected, this effect is less pronounced for the higher temperature of 60  C, where kT ˆ (10  5)  10ÿ11 Paÿ1 at 0.1 MPa and kT ˆ (12  3)  10ÿ11 Paÿ1 at 70 MPa. The plots of the pressure dependence of kT at temperatures between 35 and 55  C, in addition to the intrinsic isothermal compressibilities of the native, intermediate(s) and unfolded states, also contain large direct contributions from the conformational transitions between the folded state, the intermediate(s) and the unfolded state, which occur over this range of temperature and pressure. At 35  C in particular the apparent isothermal compressibility at atmospheric pressure is tenfold lower than that at 25  C, which is probably essentially due to a negative hydration contribution to

1099

Unfolding of Snase

kT, and then increases by a factor of about 8 upon application of pressures between 25 and 35 MPa (see Figure 4). We recall that large changes in the thermal coef®cient of expansion at ambient pressure are observed just after this temperature as well. The increase in compressibility around 30 MPa at 35  C would be consistent with apparent volume decreases involved in the formation of a molten globule-like intermediate,9 and not from a change in compressibility, per se. At 45  C, the two to threefold increase in compressibility between 40 and 70 MPa re¯ects the change in slope in V*(p) which occurs with the onset of the actual unfolding transition, and thus again is not compressibility per se, but re¯ects for the most part, the actual decrease in volume upon unfolding. The observed increase of the compressibility value would be consistent with a transition to a partially unfolded state with a considerable amount of void volume. This is in agreement with results obtained for the radius of gyration of the pressure-denatured state of Snase. Previous studies by our group demonstrated that increasing pressure in the range of 1 bar (native state) to 3.5 kbar (denatured state) results in an increase in the value Ê , an of the radius of gyration from 17 to 35 A Rg-value which is indeed considerably less than the one expected for the fully unfolded protein Ê ). Also the corresponding FT-IR data (Rg  46 A showed that the pressure-denatured state of Snase still contains a considerable amount of secondary structure elements.19,20 At pressures above 80 MPa, kT starts to decrease again which is probably due to an increased hydration contribution. As pointed out in Results, extrapolation to 100 MPa of the low pressure slope in Figure 5 allows for the estimation of this change in volume due to the unfolding transition at about ÿ55 cm3 molÿ1. This value is reasonably consistent with that obtained from ®ts of high pressure ¯uorescence and FT-IR pro®les versus temperature at 40  C ( ÿ52 cm3 molÿ1 20). One could argue that the complex behavior of the pressure dependence of the apparent isothermal compressibility in the densitometric measurements, which require higher concentrations than the calorimetry, might be caused by some aggregation occuring. A transient aggregation in the pathway of folding/unfolding has indeed been described for simple monomeric proteins,50 and might explain, at least partially, the complex pressure dependence of kT for example at 35  C by assuming, an intermediate dimeric state, especially in the light of recent reports that pressure can dissociate or prevent aggregation.51,52 At least at concentrations up to about 2 wt% Snase, where SAXS measurements were performed, no indications of such aggregation effects were observed, however. Also in the FT-IR data, at least within the sensitivity of the method, no indications of aggregation were observed at concentrations of 5 wt% Snase.19,20 It is also noteworthy that the value of the apparent isothermal compressibilities of Snase at

atmospheric pressure for the folded state at 25  C is, within the experimental error, similar to that of the unfolded state at 55 or 60  C. This suggests that the change in compressibility upon unfolding (at atmospheric pressure), regardless of the temperature, should be quite small, and even perhaps negative since, in general, compressibility tends to increase with temperature. Values of
kT of unfolding have been estimated for the unfolding of Snase, from ®ts of NMR pressure-induced unfolding pro®les in terms of both a volume change and a compressibility change upon unfolding.53,54 These values (near 1.1  10ÿ11 Paÿ1) are approximately 10 % of those we measure here for the kT of the folded and unfolded states (ca 10  10ÿ11Paÿ1), consistent with our conclusion that the changes in the actual compressibility upon unfolding must be quite small. Moreover, since compressibility in general must decrease with increasing pressure, we would expect that
kT might also decrease with increasing pressure.

Conclusions These results provide the ®rst direct measure of the effect of pressure and temperature on the apparent speci®c volume of a protein, Snase. In particular, the ranges of pressure and temperature over which these observations have been made provide direct measures of these quantities for the transition from the folded to the unfolded state(s). The complex temperature dependence of the apparent speci®c volume (and thermal expansivity) indicate the population, at equilibrium, of an expanded form of the protein. These results also indicate that the unfolding of Snase proceeds with a decrease in speci®c volume at low temperature and an increase in speci®c volume at high temperature. The pressure dependence of the speci®c volume at 35  C, in the pre-transition region, is also consistent with the population of an expanded intermediate. The pressure dependence of the apparent speci®c volume at 45  C and up to 100 MPa has allowed us to estimate the volume change due to pressure-induced unfolding of Snase at this temperature (assuming the similar compressibility for the folded and unfolded states, consistent with the experimental results). The value of ÿ55 cm3 molÿ1 for the volume change upon unfolding is found to be in good agreement with that obtained from ®ts of the pressure unfolding pro®les obtained under similar conditions. In general these studies have provided novel insight into the changes in hydration which accompany conformational transitions in proteins.

Materials and Methods Protein preparation Recombinant staphylococcal nuclease (Snase) with the sequence of nuclease A from the V8 strain of Staphylococcus aureus was obtained using the l expression system

1100

Unfolding of Snase

in the Escherichia coli strain Arl9 as described by Shortle & Lin.55 The cells were grown as described by Shortle et al. 56 except that SB rather than MOPS media was employed. The protein puri®cation was carried out as described by Shortle & Meeker 57 with modi®cations described by Frye et al.18 Sample concentration was 4 wt% Snase in 10 mM bis-Tris buffer (pH 7.0). The temperature dependence of the pH of the buffer solution was measured to be only ÿ0.01 pH units/deg. C, and thus negligible over the temperature range studied. The pressure dependence of the pH of bis-Tris is also negligible due to a lack of charge change, and thus electrostriction upon deprotonation.

Densitometry We conducted density measurements as a function of temperature and pressure. The densities of sample solutions and solvent were measured with a combination of the high precision oscillating-tube density meter DMA 58 and DMA 512 P (Anton Paar, Graz). The DMA 512 P is designed to measure densities of liquids and gases under high pressure. The pressure range for continuous operation is 0.1 to 70 MPa, the temperature range is ÿ10 to ‡150  C. The oscillating-tube of the pressure cell system DMA 512 P is built of Hastelloy and has an inner volume of about 0.7 cm3. A total volume of 2.3 ml of the sample solution was needed to ®ll the measuring cell. Temperature was controlled by a LAUDA thermostat within 0.1 deg. C. The accuracy of the pressure-dependent density measurements is about 1  10ÿ3 g cmÿ3. The densities of protein solutions were measured in the temperature range of 25 to 60  C in steps of 5 deg. C at pressures ranging from 0.1 to 70 (in one case 100) MPa in steps of 1 MPa using a Heise manometer with an accuracy of 0.1 MPa. For performing highly accurate temperature-dependent density measurements we used the DMA 58, which contains an oscillating tube built of glass. The temperature range for operation of the apparatus is ÿ10 to ‡70  C. Temperature was controlled with a built-in precision temperature sensor with an accuracy of 0.01  C. About 1.2 cm3 of the sample solution were needed to ®ll the measuring cell. The accuracy of density measurements as a function of temperature is very high, about 5 10ÿ5 g cmÿ3. The density of protein solutions was measured at ambient pressure from 15 to 65  C in steps of 2.5  C. The instrument constants were determined by calibration measurements with solvents of known density for each temperature and pressure. The solvents bidestilled water,58 pentane,59 dichlormethane,60 air and cyclohexane61 were used. At least three measurements of three independent sample preparations were performed to yield a reasonable estimate of the absolute error bars of the data. The apparent speci®c volume V* was calculated from the solution (d) and solvent (d0) densities using: V ˆ

  m0 1 1 1 ‡ ÿ d mprot d d0

…8†

where mprot is the mass of the solute (protein) and m0 is the mass of the solvent. The absolute error of V* for measurements of different sample preparations is about 1.5 %, the relative error is much smaller, however, and given as error bars in the Figures.

Acknowledgments This work was supported by the Deutsche Forschungsgemeinschaft (DFG) and the Fonds der Chemischen Industrie (R.W.), and the Institut National de la Sante et la Recherche MeÂdicale (C.R.).

References 1. Sarvazyan, A. P. (1991). Ultrasonic velocimetry of biological compounds. Annu. Rev. Biophys Biophys. Chem. 20, 321-342. 2. Chalikian, T. V., Sarvazyan, A. P., Funck, T. & Breslauer, K. J. (1994). Partial molar volumes, expansibilities and compressibilities of oligoglycines in aqueous solutions at 18-55  C. Biopolymers, 34, 541553. 3. Chalikian, T. V., Gindikin, V. S. & Breslauer, K. J. (1995). Volumetric characterizations of the native, molten globule and unfolded states of cytochrome c at acidic pH. J. Mol. Biol. 250, 291-306. 4. Chalikian, T. V. & Breslauer, K. J. (1996). Compressibility as a means to detect and characterize globular protein states. Proc. Natl Acad. Sci USA, 93, 10121014. 5. Chalikian, T. V., Totrov, M., Abagyan, R. & Breslauer, K. J. (1996). The hydration of globular proteins as derived from volume and compressibility measurements: cross correlating thermodynamic and structural data. J. Mol. Biol. 260, 588-603. 6. Chalikian, T. V. & Breslauer, K. J. (1996). On volume changes accompanying conformational transitions of biopolymers. Biopolymers, 39, 619-626. 7. Chalikian, T. V., Gindikin, V. S. & Breslauer, K. J. (1996). Spectroscopic and volumetric investigation of cytochrome c unfolding at alkaline pH: characterization of the base-induced unfolded state at 25  C. FASEB J. 10, 291-306. 8. Chalikian, T. V., VoÈlker, J. ,Ana® D. & Breslauer, K. J. (1997). The native and heat-induced denatured states of a-chymotrypsinogen A: thermodynamic and spectroscopic studies. J. Mol. Biol. 274, 237-252. 9. Chalikian, T. V. & Breslauer, K. J. (1998). Thermodynamic analysis of biomolecules: a volumetric approach. Curr. Opin. Struct. Biol. 8, 657-664. 10. Gekko, K. & Hasegawa, Y. (1986). Compressibilitystructure relationship of globular proteins. Biochemistry, 25, 6563-6571. 11. Gekko, K. & Hasegawa, Y. (1989). Effect of temperature on the compressibilty of native globular proteins. J. Phys. Chem. 93, 426-429. 12. Gekko, K. & Noguchi, H. (1979). Compressibility of globular proteins in water at 25  C. J. Phys. Chem. 83, 2706-2714. 13. Kharakoz, D. P. (1991). Volumetric properties of proteins and their analogs in diluted water solutions. 2. Partial adiabatic compressibilities of amino acids at 15-70  C. J. Phys. Chem. 95, 5634-5642. 14. Kharakoz, D. P. (1997). Partial volumes and compressibilities of extended polypeptide chains in aqueous solution: additivity scheme and implication of protein unfolding at normal and high pressure. Biochemistry, 36, 10276-10285. 15. Kharakoz, D. P. & Sarvazyan, A. P. (1993). Hydrational and intrinsic compressibilities of globular proteins. Biopolymers, 33, 11-26.

Unfolding of Snase 16. Sasahara, K., Sakurai, M. & Nitta, K. (1999). The volume and compressibility changes of lysozyme associated with guanidinium chloride and pressureassisted unfolding. J. Mol. Biol. 291, 693-701. 17. Fil®l, R. & Chalikian, T. V. (2000). Volumetric and spectroscopic characterizations of the native and acid-induced denatured states of staphylococcal nuclease. J. Mol. Biol. 299, 827-842. 18. Frye, K. J., Perman, C. S. & Royer, C. A. (1996). Testing the correlation between
A and
V of protein unfolding using m-value mutants of staphylococcal nuclease. Biochemistry, 35, 10234-10239. 19. Panick, G., Malessa, R., Winter, R., Rapp, G., Frye, K. J. & Royer, C. A. (1998). Structural characterization of the pressure-denatured state and unfolding/refolding kinetics of staphylococcal nuclease by synchrotron small angle X-ray scattering and Fourier-transform infrared spectroscopy. J. Mol. Biol. 275, 389-402. 20. Panick, G., Vidugiris, G. J. A., Malessa, R., Rapp, G., Winter, R. & Royer, C. A. (1999). Exploring the temperature phase diagram of staphylococcal nuclease. Biochemistry, 38, 4157-4164. 21. Hynes, T. R. & Fox, R. O. (1991). The crystal strucÊ ture of staphylococcal nuclease re®ned at 1. 7 A resolution. Proteins: Struct. Funct. Genet. 10, 92-105. 22. Mozhaev, V. V., Heremans, K., Frank, J., Masson, P. & Balny, C. (1996). High pressure effects on protein structure. Proteins: Struct. Funct. Genet. 24, 81-91. 23. Markley, J. L., Northrop, D. B. & Royer, C. A. (1996). High-pressure Effects in Molecular Biophysics and Enzymology, Oxford University Press, Oxford. 24. Heremans, K. & Smeller, L. (1997). Pressure versus temperature behaviour of proteins. Eur. J. Solid State Inorg. Chem. 34, 745-758. 25. Winter, R. & Jonas, J. (1999). High Pressure Molecular Science. Kluwer, Academic Publishers, The Netherlands. 26. Vidugiris, G. J. A., Markley, J. L. & Royer, C. A. (1995). Evidence for a molten globule-like transition state in protein folding from determination of activation volumes. Biochemistry, 34, 4909-4912. 27. Vidiguris, G. J. A. & Royer, C. A. (1998). Determination of the volume changes for pressure-induced transitions of apomyoglobin between the native, molten globule and unfolded states. Biophys. J. 75, 473-470. 28. Weber, G. & Drickamer, H. G. (1983). The effect of high pressure upon proteins and other biomolecules. Quart. Rev. Biophys. 16, 89-112. 29. Hinz, H.-J., Vogl, T. & Meyer, R. (1994). An alternative interpretation of the heat capacity changes associated with protein unfolding. Biophys. Chem. 52, 275-285. 30. Gavish, B., Gratton, E. & Hardy, C. J. (1983). Adiabatic compressibility of globular proteins. Proc. Natl Acad. Sci. USA, 80, 750-754. 31. NoÈlting, B., Jiang, M. & Sligar, S. G. (1993). The acidic molten globule state of a-lactalbumin probed by sound velocity. J. Am. Chem. Soc. 115, 9879-9882. 32. NoÈlting, B. & Sligar, S. G. (1993). Adiabatic compressibilty of molten globules. Biochemistry, 32, 1231912323. 33. Tamura, Y. & Gekko, K. (1995). Compactness of thermally and chemically denatured ribonuclease A as revealed by volume and compressibility. Biochemistry, 34, 1878-1884.

1101 34. Hayashi, T., Sakurai, M. & Nitta, K. (1996). The volume and compressibility changes associated with protein denaturation. Chem. Letters, 723-734. 35. Carra, J. H., Anderson, E. A. & Privalov, P. L. (1994). Three-state thermodynamic analysis of the denaturation of staphylococcal nuclease mutants. Biochemistry, 33, 10842-10850. 36. Scarlata, S. F. & Royer, C. A. (1986). Ligand-induced asymmetry as observed through ¯uorophore rotations and free energy couplings: application to neurophysin. Biochemistry, 25, 4925-4929. 37. Chen, H. M., Markin, V. S. & Tsong, T. Y. (1992). pH-induced folding/unfolding of staphylococcal nuclease: determination of kinetic parameters by the sequential-jump method. Biochemistry, 31, 1483-1491. 38. Chen, H. M., Markin, V. S. & Tsong, T. Y. (1992). Kinetic evidence of microscopic states in protein folding. Biochemistry, 31, 12369-12375. 39. Nakano, T., Antonino, L. C., Fox, R. O. & Fink, A. L. (1993). Effect of proline mutations on the stability and kinetics of folding of staphylococcal nuclease. Biochemistry, 32, 2534-2541. 40. Walkenhorst, W. F., Green, S. M. & Roder, H. (1997). Kinetic evidence for folding and unfolding intermediates in staphylococcal nuclease. Biochemistry, 36, 5795-5805. 41. Ikura, T., Tsurupa, G. P. & Kiwajima, K. (1997). Kinetic folding and cis/trans prolyl isomerization of staphylococcal nuclease: a study by stopped-¯ow absorption, stopped ¯ow circular dichroism and molecular dynamics simulations. Biochemistry, 36, 6529-6538. 42. Holcomb, D. N. & van Holde, K. E. (1962). Ultracentrifugal and viscometric studies of the reversible thermal denaturation of ribonuclease. J. Phys. Chem. 66, 1999-2006. 43. Desai, G., Panick, G., Zein, M., Winter, R. & Royer, C. A. (1999). Pressure-jump studies of the folding/ unfolding of trp repressor. J. Mol. Biol. 288, 461-475. 44. Royer, C. A. (1999). Pressure denaturation of proteins. In High Pressure Molecular Science (Winter, R. & Jonas, J., eds), 473, Kluwer, Academic Publishers, The Netherlands. 45. Masterton, W. L. (1954). Partial molal volumes of hydrocarbons in water solution. J. Chem. Phys. 22, 1830-1833. 46. Masterton, W. L. & Seiler, H. K. (1968). Apparent and partial molar volumes of water in organic solvents. J. Phys. Chem. 72, 4257-4262. 47. Richards, F. M. (1977). Areas, volumes, packing, and protein structure. Annu. Rev. Biophys. Bioeng. 6, 151176. 48. Harpaz, Y., Gerstein, M. & Chothia, C. (1994). Volume changes on protein folding. Structure, 2, 641-649. 49. Griko, Y. V., Privalov, P. L., Sturtevant, J. M. & Venyaminov, S. (1988). Cold denaturation of staphylococcal nuclease. Proc. Natl Acad. Sci. U.S.A. 85, 3343-3347. 50. Silow, M., Tan, Y.-J., Fersht, A. R. & Oliveberg, M. (1999). Formation of short-lived protein aggregates directly from the coil in two-state folding. Biochemistry, 38, 13006-13012. 51. St. John, R. J., Carpenter, J. F. & Randolph, T. W. (1999). High pressure fosters protein folding from aggregates at high concentrations. Proc. Natl Acad. Sci. USA, 9, 13029-13033. 52. Foguel, D., Robinson, C. R., de Sousa, P. C., Silva, J. L. & Robinson, A. S. (1999). Hydrostatic pressure

1102

53.

54. 55.

56.

Unfolding of Snase

rescues native protein from aggregates. Biotechnol. Bioeng. 63, 552-558. Lassalle, M. W., Yamada, H. & Akasaka, K. (2000). The pressure-temperature free energy-landscape of staphylococcal nuclease monitored by 1H NMR. J. Mol. Biol. 298, 293-302. Prehoda, K. E., Moorberry, E. S. & Markley, J. L. (1998). Pressure denaturation of proteins: evaluation of compressibility effects. Biochemistry, 37, 5785-5790. Shortle, D. & Lin, B. (1985). Genetic analysis of staphylococcal nuclease: identi®cation of three intragenic global suppressors of nuclease-minus mutations. Genetics, 110, 539-555. Shortle, D., Meeker, A. K. & Gerring, S. L. (1989). Effects of denaturants at low concentrations on the reversible denaturation of staphylococcal nuclease. Arch. Biochem. Biophys. 272, 102-113.

57. Shortle, D. & Meeker, A. K. (1986). Guanidine hydrochloride denaturation studies of mutant forms of staphylococcal nuclease. J. Cell. Biochem. 30, 281289. 58. Saul, A. & Wagner, W. (1989). A fundamental equation for water covering the range from the melting line to 1273 K at pressures up to 25000 MPa. J. Phys. Chem. Ref. Data, 18, 1537-1564. 59. Kratzke, H., MuÈller, S., Bohn, M. & Kohlen, R. (1985). Thermodynamic properties of saturated and compressed liquid n-pentane. J. Chem. Thermodynam. 17, 283-294. 60. Demiriz, A. M. (1986). pVT-messungen an dichlormethan. PhD thesis, UniversitaÈt Bochum, VDIVerlag, DuÈsseldorf. 61. ALLPROPS, (1993). The Center of Applied Thermodynamic Studies, College of Engineering, University of Idaho, Moscow, Idaho, Version 4.0.

Edited by C. R. Mathews (Received 5 October 2000; received in revised form 30 January 2001; accepted 1 February 2001)