You are on page 1of 10

Biophysical Journal

Volume 80

June 2001

2751–2760

2751

Hydration and Protein Folding in Water and in Reverse Micelles: Compressibility and Volume Changes
D. Valdez,* J -Y. Le Hue ´ rou,* M. Gindre,* W. Urbach,† and M. Waks*
*Laboratoire d’Imagerie Parame ´ trique, UMR 7623 CNRS Universite ´ Pierre et Marie Curie,75270 Paris cedex 06, France; †Laboratoire de Physique Statistique, UMR 8550 CNRS Ecole Normale Supe ´ rieure, 75231 Paris cedex 05, France

ABSTRACT The partial specific volume and adiabatic compressibility of proteins reflect the hydration properties of the solvent-exposed protein surface, as well as changes in conformational states. Reverse micelles, or water-in-oil microemulsions, are protein-sized, optically-clear microassemblies in which hydration can be experimentally controlled. We explore, by densimetry and ultrasound velocimetry, three basic proteins: cytochrome c, lysozyme, and myelin basic protein in reverse micelles made of sodium bis (2-ethylhexyl) sulfosuccinate, water, and isooctane and in aqueous solvents. For comparison, we use ␤-lactoglobulin (pI ϭ 5.1) as a reference protein. We examine the partial specific volume and adiabatic compressibility of the proteins at increasing levels of micellar hydration. For the lowest water content compatible with complete solubilization, all proteins display their highest compressibility values, independent of their amino acid sequence and charge. These values lie within the range of empirical intrinsic protein compressibility estimates. In addition, we obtain volumetric data for the transition of myelin basic protein from its initially unfolded state in water free of denaturants, to a folded, compact conformation within the water-controlled microenvironment of reverse micelles. These results disclose yet another aspect of the protein structural properties observed in membrane-mimetic molecular assemblies.

INTRODUCTION Although the interiors of folded proteins, or protein domains, are compact and well-packed (Klapper, 1971; Richards, 1977; Richards and Lim, 1994), the packing of aminoacid residues is not perfect (i.e., proteins are compressible). In addition to imperfect packing, a significant fraction of protein compressibility originates from the finite rigidity of the intermolecular interaction potentials. Gekko and Nogushi (1979) pointed out that, in addition to imperfect packing, the presence of internal cavities (Rashin et al., 1986) contributes positively to protein adiabatic (isoentropic) compressibility as well as volume changes. When such voids are large enough to accommodate water molecules, internal water can then act as a structure stabilizer (Takano et al., 1997) by maintaining good hydrogen bonds between the domains and filling sites of imperfect packing. As a universal solvent, water interacting at the protein surface contributes negatively to protein volume and compressibility. Difficulties in interpreting the precise changes in protein volume and compressibility induced by hydration at protein-water interfaces have led to empirical estimates of the intrinsic protein volume, that is, the volume that cannot be penetrated by the solvent (Paci and Velikson, 1997). There is no consensus on the exact value of the compressibility of the solvent-inaccessible protein core, or intrinsic protein compressibility, due to the diverse approaches used by different investigators (Ghekko and Noguchi, 1979; Gavish et al., 1983; Kharakoz and Sarvazyan, 1993; Chalikian et al., 1996). Recent developments in acoustic techniques (difference ultrasound velocimetry) have rendered possible high precision compressibility measurements of small volumes of solutes, in particular biopolymers (Sarvazyan, 1991). Thus, the experimental study of protein volume and adiabatic compressibility variation as a function of increasing levels of hydration should become possible and offer unique insights into these observations. Experiments designed to address the latter issue cannot be performed in a traditional solvent such as water in the bulk state, a medium in which investigations have previously been carried out. We have therefore used a nonconventional solvent designated as microemulsions, or reverse micelles, in this work. Reverse micelles can be described as water microdroplets dispersed in water-immiscible apolar solvents (i.e., oils) and stabilized by a monolayer of surfactant acting as an interface between oil and water. The surfactant non-polar tails protrude into the oil, while the polar headgroups are in direct contact with the central water core. An important feature of the anionic surfactant sodium bis (2-ethylhexyl) sulfosuccinate (AOT) is the tight binding by one polar headgroup of about 7 to 10 water molecules. There is a wealth of information available concerning reverse micelles, such as their structure, their phase behavior, and the physical properties of the surfactant micellar bound and free waters (Luisi and Magid, 1986; Oldfield, 1994; Amararene et al., 2000). The size of the droplets, which are in thermal equilibrium, depends only on the water concentration. As water is added, their radius (range, 1.5 to 10 nm) increases as a function of the water:surfactant molar ratio (W0). (Amararene et al., 2000). The water content of the system can

Received for publication 28 June 2000 and in final form 19 March 2001. Address reprint requests to Dr. Marcel Waks, LIP, UMR 7923 CNRS, Universite Pierre et Marie Curie, 15, Rue de L’Ecole de Medecine, Paris, France 75006. Tel: 33144414976; Fax: 33146335673; E-mail: marcel. waks@lip.bhdc.jussieu.f. © 2001 by the Biophysical Society 0006-3495/01/06/2751/10 $2.00

the protein concentration. We have selected a method based on time-of-flight determination. while ␤-lactoglobulin cannot be completely solubilized at a working concentration (ϳ3 mg/ml) below W0 ϭ 10. heart muscle cytochrome c (C 7752) and bovine milk ␤-lactoglobulin (L 3908). can accommodate proteins within their water cores.2752 Valdez et al. cytochrome c remains partially insoluble at W0 values above 22. the sing-around. which we do not observe. for example. was obtained from Nikko Chemicals (Japan). For comparison. Waks. 1985. carefully dried in vacuo.6 due to its high affinity for membrane-bound water (Nicot and Waks. the time-of-flight measurement.04 ml). turbid samples detected by optical density at 350 nm) were ruled out. The precision obtained on density is better than 10Ϫ5 g⅐mlϪ1 for a single measurement. a precise solubility curve (concentration versus W0) was established by optical density at 280 nm. The stability of the apparatus is crucial: no evaporation of solvent occurs during the measurement procedure. significantly different from bulk water. For this purpose.. Compressibility of nano inclusions in complex fluids by ultrasound velocity measurements. 1985). MATERIALS AND METHODS Materials AOT purchased from Sigma (SIGMA ULTRA) was 99% pure and used after sufficient desiccation in vacuum over phosphorus pentoxide (SICAPENT from Merck). a high precision provided by recent instrumental and computational advances (Le Hue ´ rou et al.. compatible with total protein solubilization.. Therefore. 1986). and they do not denature during the solubilization process (Nicot and Waks. tetraethylene glycol monododecyl ether (C12E4). and circular dichroism) are available. Since micellar solutions are optically clear. a peripheral membrane protein from the nervous system. In all the proteins examined. Obviously. which offers. Experiments under conditions in which protein would be only partially solubilized and remain suspended in the oil (i. . resonance. In the aqueous core of reverse micelles. During all the experiments the greatest care has been taken to avoid evaporation of the organic solvent. 1995). Submitted for publication). MBP folds into a well-ordered.. 1 and 2.00 Ϯ 0. we have found limits of solubilization for each protein studied as a function of W0 in relation to their respective size and surface charges. After several minutes of gentle shaking and a few seconds of sonication. The nonionic surfactant. lysozyme and myelin basic protein. we have selected three basic proteins of known structure in aqueous solution and in micellar solutions: cytochrome c. while decane Ͼ 99% pure came from Sigma. 1975). the same experimental strategy has allowed us to obtain interesting volumetric data on MBP. was purchased from Merck. The proteins were from Sigma: hen egg lysozyme (L 6876). since extinction coefficients are identical in water and in reverse micelles (Luisi and Magid. For example. however. By systematically applying this procedure. appears to mirror different aspects of complex cellular conditions. as they would not have been optically transparent. they were Biophysical Journal 80(6) 2751–2760 Ultrasound velocity measurements Several methods of ultrasound velocity determination have been described to date. the dried. The controlled micellar microenvironment. We have been unable to detect traces of protein in the organic solvent by fluorescence spectrophotometry.15 M C12E4 in decane) with the appropriate amount of water required to achieve the desired water:surfactant molar ratio. we observe a decrease in compressibility as a function of hydration. Sample preparation All the samples were prepared by first weighing. and consequently a systematic drift of the density meter. and pulse overlap methods (Braezaele et al. thus be precisely varied and experimentally controlled. After dialysis against cold water and lyophillization. Moreover. lyophillized proteins on a Model 1712 Sartorius balance. with the exception of myelin basic protein (MBP). consequently. The conformational properties of the above proteins in reverse micelles have been reviewed. the results averaged. in 20 ml volumetric flasks (class A Ϯ 0. We have undertaken this work to explore the effect of hydration on the specific volume and compressibility of proteins sequestered in the micellar aqueous core. The samples were always kept in ground glassware and transferred without delay to the densitometer and the air-tight ultrasound measurement cells (see below) by the means of glass syringes. Due to their positive charge they experience a strong electrostatic field in AOT reverse micelles. and serves as a model of the myelin interlamellar space (Nicot and Waks. This fact is clearly illustrated in Figs.g. solubility and conformation can be determined with high precision. 1985). Water-free isooctane.1). As in water. Note that proteins are not soluble in isooctane. At the lowest water concentration.e. and judged to be Ͼ 99% pure from gas chromatography. adiabatic compressibility is more sensitive in reverse micelles to protein hydration than volume. MBP displays a solubility maximum at W0 ϭ 5. compact periodic structure (Nicot et al.. 1985). the protein displays a high degree of conformational flexibility with little periodic structure (Liebes et al. We have taken advantage of this unique structural property to characterize both the aqueous unfolded and the micellar folded conformational states of MBP. with a precision of Ϯ 0. MBP was prepared and purified from bovine brain. Water used in this study was of MILLIQ purity (pH around 6). fluorescence. in addition to its simplicity. optically clear solutions were ready for use. Each set of measurements was carried out at least five times. evaporation would result in increasing protein concentration and density values. Pro Analysi grade.. we have also carried out measurements with ␤-lactoglobulin (pI ϭ 5.03 mg. in which MBP is located in myelin.01°C using a vibrating tube Anton Paar DMA 58 digital density meter standardized by water and dry air. we report large compressibility values in the range of those of a protein interior. Volumetric experiments were always carried out well below the solubility limit. 1986). absorption. Reverse micelles are protein-sized and. a water immiscible solvent. In water free of denaturing agents.1M AOT in isooctane or 0. as has been previously described (Nicot et al.. all current spectroscopic measurement techniques (e. 1981). In contrast to the other proteins. Volumetric measurements The densities of micellar solutions and protein-containing micellar solutions were determined at 25. The solvents used to make up the volume at 20°C were solutions of the surfactants (0. For all proteins used. and the value then used to calculate the protein apparent specific volume ␸v (see below).

in a temperature controlled room. 1994. a short electrical pulse is applied to a piezoelectric transducer.Protein Compressibility in Reverse Micelles 2753 We can thus relate the micellar protein solution compressibility (␤) to both the micellar (␤1) and protein (␤p) compressibilities: Principle In ultrasound velocity measurements using the time-of-flight technique. We then describe each of the constituent components as a function of the relative volumes of the constituent phases in terms of several parameters. such as evaporation. dried. converting the electrical wave into an acoustical one which then propagates through the sample.. Experiments are carried out at 25. As described in the Materials and Methods section. and precipitation. Cytochrome c This small (13 kDa) basic (pI ϭ 10. 1998). For example. u1 ⅐ cp RESULTS The reverse micellar system investigated in this work has been previously characterized in a protein-free state by difference ultrasound velocimetry (Amararene et al. This difficulty is circumvented by the use of a set of two cells of identical acoustic path enclosed in a single metal thermostated block (Sarvazyan. and velocity measurements are carried out. for example. u and u1 for each value of the water:surfactant molar ratio (W0). Depending on the microenvironment. Thus. Each of the constituent phases is described by parameters corresponding to the pure phase. considering the various caveats discussed in the Materials and Methods section.5) protein has been studied in detail. and the second phase consists of micellar inclusions randomly embedded in the continuous phase. using Laplace’s equation: ␤ϭ 1 ␳u2 (2) The adiabatic compressibility of the reference micellar solution becomes: ␤1 ϭ 1 ␳ 1u 12 (3) where u1 is the velocity of the reference micellar solution In this work we make use of the effective medium theory (Ye et al. We take one phase to be the continuum fluid. ␤1. Table 1 summarizes our results: the experimental errors are of the same order of magnitude as those reported in the literature.. The time interval elapsed between the emitted and received signals combined with knowledge of the precise distance between the two transducers allows determination of the ultrasound velocity. the existence of protein-free micelles does not interfere with our measurements since the values obtained for such micelles will cancel out by difference. our data characterize solely the protein. The introduction of proteins into the micellar aqueous core adds an additional complexity to the experiments. (␳) the density of the protein micellar solution. the measuring cell is drained. including its interactions. 2000). 1982).e. achieving a considerable improvement in sensitivity.00 Ϯ 0. The setup allows the sequential determination of the ultrasonic velocity difference between the reference and the measuring cell at identical temperature. The micelle occupancy by the protein implies a fraction of protein-free micelles. great care has been taken to control and avoid all possible sources of error. and the crystal structure of ferricytochrome c is known at high resolution (Bushnell et al. Cortese et al. because the relevant acoustic wavelength is always orders of magnitude larger than the micelle size. at W0 ϭ 5 the ratio of free to filled micelles is ϳ5 for a concentration of 3 mg per ml. and ␾p are obtained from ␳. the protein-containing micellar solution). The experimental measurements of the protein micellar solution density (␳) and the sound velocity (u) allows the determination of the solution adiabatic compressibility.. Membrane-bound cytochrome c has also been thoroughly studied (Pinheiro and Watts. ␤ ϭ ␤ 1͑ 1 Ϫ ␾ p͒ ϩ ␤ p␾ p (4) ␤. The difference technique uses as a reference an identical micellar solution devoid of protein. protein insolubility. (5) can be written as a function of the protein apparent specific adiabatic compressibility (␸k) ␤p ϭ where ͓u͔ ϭ ␤1 ␸k 1 2␸v Ϫ Ϫ 2͓u͔ ϭ ␸v ␳1 ␸v ͩ ͪ (6) u Ϫ u1 . the memBiophysical Journal 80(6) 2751–2760 Protein volume and compressibility in reverse micelles The protein apparent specific volume is given by: ␸v ϭ 1 ␳ Ϫ ␳1 Ϫ ␳1 c p␳ 1 (1) This equation applies to both simple aqueous solutions and reverse micelles where (␳1) is the density of the reference micellar solution. Nevertheless. Stringent precautions are required to avoid errors in protein concentration that would lead to irreproducible results. . cytochrome c appears to adopt a number of conformational states. and refilled with the liquid under investigation (i.. The protein compressibility is obtained by: ␤p ϭ ␤1 ϩ ␤ Ϫ ␤1 c p␸ v (5) Measurement system One of the principal limitations of the custom-built system resides in possible temperature drifts. A second transducer converts the received wave into an electrical signal.01°C. Submitted for publication). 1990). In a second step. the two acoustic cells are filled with the protein-free micellar reference solution (Ϸ3 ml) using a glass syringe. This rules out the presence of more than one protein molecule per micelle or the existence of protein aggregates. The experimental setup has been described in detail elsewhere (Le Hue ´ rou et al. The final precision in ultrasound velocity determination is better than 10Ϫ5. 1991) to account for the behavior of reverse micelles.. The cells are made airtight with Teflon o-rings to avoid any liquid leakage or solvent evaporation.. which is compared to the excitation signal. ␳1. First. cp. Eq. and cp the protein concentration. rinsed.

.. as hydration decreases further from W0 ϭ 11 to 8. 1.730 Ϯ 0.8 and 19. In AOT reverse micelles.. On the far right. since its absolute value is still 0. The experimental conditions are the same as in Fig. lysozyme (Ⅺ) and ␤-lactoglobulin (Œ) as a function of W0. we still find a value of 0. identical spectral characteristics have been reported by Brochette et al.1 M in isooctane). 1997). Pinheiro et al. brane anionic environment perturbs the native structure of cytochrome c. The authors have interpreted this result by the opening of the cytochrome c heme crevice under the intense electrostatic field generated by the membrane anionic polar headgroups (Pinheiro et al. Each experiment was carried out at least five times and the data averaged.753 ml ⅐ gϪ1. FIGURE 2 Plot of the adiabatic compressibility ␤ of cytochrome c (F). In highly hydrated reverse micelles (W0 ϭ 17. in the same range as Chalikian et al. When the water amount decreases to W0 ϭ 11. 1996. at the lowest water content (W0 ϭ 8).2) the micellar microenvironment does not seem to affect the specific volume much.2754 Valdez et al. we find a compressibility ␤ ϭ 3 Ϯ 0. 1995). 1).2 W0 range. 1997). However.2 ϫ 10Ϫ11 PaϪ1 ⅐ ml ⅐ gϪ1). In such states the absence of the characteristic absorbance band at 695 nm is paralleled with the disruption of Met-80 ligation to the iron atom (de Jongh et al. Biophysical Journal 80(6) 2751–2760 .2 to 19. Experimental errors are shown in Table 1. the volume increases abruptly to 0. In reverse micelles.5 ϫ 10Ϫ11 PaϪ1 (␸k ϭ 2. We have carried out experiments in water and reverse micelles in the 8.725 ml ⅐ gϪ1. the native protein displays a partial specific volume of 0.726 Ϯ 0. (1996). (1988) at high water content. Compressibility For the native protein in water. Volume In water. It is remarkable that such a volume jump is observed for a rather small change in W0 (Fig.. lysozyme (Ⅺ) and ␤Ϫlactoglobulin (Œ) as a function of W0.003 ml ⅐ gϪ1.. FIGURE 1 Plot of the apparent specific volume ␸v of cytochrome c (F). in agreement with the literature (Chalikian et al. 1995). values obtained in pure water are shown. The volume scale and the data points on the far right of the figure correspond to values obtained in pure water. the water:surfactant molar ratio (AOT is 0. leading to molten globule-like states (Bychkova et al.002 ml ⅐ gϪ1. we obtain a high value for adiabatic compressibility.

727 Ϯ 0. We carried out measurements on lysozyme in water and in reverse micelles in the W0 ϭ 8 to 40 range.726 Ϯ 0. 1997).5 kDa. MBP. close to the value obtained in bulk water. 1 and 2). an enzyme of 14. At W0 ϭ 8. At the same time.726 7 Ϫ1 0. ␤ ϭ 37 ϫ 10Ϫ11 PaϪ1 (␸k ϭ 28 Ϯ 2 ϫ 10Ϫ11 PaϪ1 ⅐ ml ⅐ gϪ1). 1972).gϪ1.003 ml ⅐ g .0.735 0.2 11 17. of the same order as cytochrome c) and decreases progressively until W0 ϭ 40. We have taken advantage of the this unusual conformational property to investigate the volume and compressibility changes of the protein first in water.725 0.1. we have obtained a value for ␤ of 6. then upon subunit dissociation at pH 2. Its structure has been solved at 1.732 0. ␤ ϭ 37 ϫ 10Ϫ11 PaϪ1 (i.731 0. at the highest hydration level measured (W0 ϭ 40).2.776 0. its value of 0.728 0. Like other basic proteins. *In reverse micelles of 0.752 0.. and finally in AOT reverse micelles between W0 values of 11 and 50 (Figs.753 0. It dissociates into two monomers at pH 2. the compressibility ␤ drops to 4. 1993. we have obtained a value of 0. is also well documented (Imoto et al.4.8 Å resolution (Brownlow et al.731 ␸V ␤ 0. Lysozyme The structure of lysozyme (pI ϭ 10). where it reaches 18 ϫ 10Ϫ11 PaϪ1. In AOT reverse micelles we observe a progressive decrease of the adiabatic compressibility as a function of increasing hydration. 1996). at the lowest W0 measured (8.743 ml. the compressibility decreases to 21 Ϯ 2 ϫ 10Ϫ11 PaϪ1 (Fig. In this work. implying the religation of Met-80 by heme (spectrum not shown).726 ␤ 38 25 22 21 ␸V 0.754 0.8.0. its conformation may be affected by the electrostatic field generated by the anionic surfactant polar headgroups and the protein positive charges. lysozyme.727 0. By increasing hydration to W0 ϭ 11.. (1981). Its behavior in reverse micelles has been described by Grandi et al. reaching a value of 0. finally. The specific volume then increases progressively as the amount of water present decreases.754 0. 19.8 (Fig. In AOT reverse micelles. in nonionic micelles of C12E4 (Merdas et al.0 ϫ 10Ϫ11 PaϪ1 at W0 ϭ 22..6 8.002 ml ⅐ gϪ1 for the apparent specific volume. and ␤-lactoglobulin as a function of W0 Cytochrome c W0* 5.. It is a predominantly ␤-sheet protein.4 30 40 45 50 H2O Lysozyme MBP ␤-Lactoglobulin ␤ 35 36 33 33 0. 1996).15 M tetraethylene glycol monododecyl ether (C12E4) in decane W0 ϭ 22.003 ml ⅐ gϪ1 is similar to that measured in water and remains constant down to W0 ϭ 17.732 0. Volume In water. where electrostatic interactions are offset. retaining the same x-ray structure as at neutral pH (Kuwata et al.725 Ϫ1 Apparent volumes are expressed in ml ⅐ g . 2). Priev et al. ␤-Lactoglobulin The major whey protein of the milk of ruminants and other mammals has an isoelectric point of 5. 1999).751 ml ⅐ gϪ1 range. the partial specific volume found in literature for ␤-lactoglobulin is quite variable in the 0. It always remains higher than in bulk water (Fig.2 22. the protein exists as a dimer of about 35 kDa.2). At neutral pH. † In reverse micelles of 0.754 0.8 ϫ 10Ϫ11 PaϪ1. e.732 0.727 ␤ 37 32 29 25 19 18 ␸V 0.743 0.734 – 0.Protein Compressibility in Reverse Micelles 2755 TABLE 1 Apparent volumes ␸V and adiabatic compressibilities ␤ of cytochrome c. in good agreement with the literature (Sasahara et al. the maximum error is Ϯ 2 ϫ 10Ϫ11 PaϪ1.1 M AOT in isooctane.2 (the protein upper solubility limit). and. 1). the maximum error is Ϯ 0. Compressibility In water.8 19. 1999).750 39 33 27 20 13 10 9 ␸V 0.765 0.750 Biophysical Journal 80(6) 2751–2760 . Volume In bulk water we find a value of 0. consisting of a ␤Ϫbarrel made of eight antiparallel ␤-strands.730 3 Ϫ1 0. 17. In contrast. compressibility values for cytochrome c and MBP are: ␤ ϭ 4 ϫ 10Ϫ11 PaϪ1 and 2 ϫ 10Ϫ11 PaϪ1. respectively. Adiabatic compressibilities are expressed in 10Ϫ11 PaϪ1... probably due to different experimental conditions. in agreement with published data (Kharakoz and Sarvazyan. 2). similar to cytochrome c.727 0. we observe the full restoration of the visible absorbance band at 695 nm.

in the nonionic (C12E4) reverses micelles devoid of electrostatic interactions (although other interactions remain operative). devoid of disulfide bridges. Volume We have found for the partial specific volume of MBP in water a value of 0. However. This unusual behavior. the compressibility obtained is rather high: 36 Ϯ 2 ϫ 10Ϫ11 PaϪ1.754 Ϯ 0.. as the dimer protein dissociates into two subunits. This fact is probably due to its relatively low isoelectric point (pI ϭ 5.0 ϫ 10Ϫ11 PaϪ1 (␸k ϭ 1. Compressibility We obtain for the adiabatic compressibility of MBP in pure water a value of ␤ ϭ Ϫ1. we obtain a difference of only 61 ml/mole of protein. 1993. 1978) evidencing extensive solvation. similar to its apparent specific volume (Fig. At the highest water content measured (W0 ϭ 22. Furthermore.. For example. It progressively decreases as a function of hydration. At the highest W0 value (50). Moreover. the volume increases to 0.5 kDa) peripheral membrane basic (pI ϭ 10.6) the volume remains almost constant (Fig. will be discussed below.000 Å3 to 47. At W0 ϭ 11. after correction for carboxyl and histidine protonation (Foygel et al.4). 1975). The uncorrected difference corresponds to a value of 370 ml/mole of protein.770 ml ⅐ gϪ1. MBP. MBP MBP is essential for the formation of the myelin sheath. The micellar cavity thus Biophysical Journal 80(6) 2751–2760 DISCUSSION We present in this work novel experimental results obtained by densitometry and sound velocimetry for proteins in both aqueous solvents and protein-sized reverse micelles.01 M HCl at pH 2. its value increases slightly from 0. There is no information available concerning the actual value of the apparent partial specific volume of refolded MBP in its native. its three-dimensional structure in native myelin is not known. similar to the aqueous solution value. at the lowest W0 value measured (5.2. ␤ increases to 12 ϫ 10Ϫ11 PaϪ1 for the monomeric protein. from 22.4.300 Å3 (Martenson. in the same manner as for the other proteins studied. it remains unchanged at W0 values of 30 and 22. and the curve can be extrapolated to the value obtained in aqueous solvents. upon dimer dissociation. Thus.731 Ϯ 0. its compressibility is the highest measured in this study: ␤ ϭ 39 ϫ 10Ϫ11 PaϪ1.4. In 0.6) in AOT reverse micelles. 3). it exposes to the solvent a surface of 568 Å2 per monomer (Brownlon et al. It is interesting that at W0 ϭ 22. in agreement with the literature (Gekko and Hesawaga.1) and to its size.5 ϫ 10Ϫ11 PaϪ1 ⅐ ml ⅐ gϪ1).8.002 ml ⅐ gϪ1. indeed a very modest variation taking into account experimental errors. in good agreement with Gekko and Hasegawa (1986). 1997) and its volume increases to 0.0 ϫ 10Ϫ11 PaϪ1 ⅐ ml ⅐ gϪ1). Priev et al.. membranebound form in myelin. In dilute aqueous solutions. a number of parameters might contribute to the behavior of proteins in such a complex system. At W0 ϭ 11. ordered conformation..6) protein has been completely sequenced. Note that in 0. we find ␤ ϭ 2. We have shown previously that in AOT reverse micelles MBP refolds from an unordered structure to a stable. a small increase in volume parallels a larger increase in compressibility. 2) is obviously different from the other basic proteins studied. specific to MBP.725 in water to 0. in addition to hydration. where micelle-assisted folding can proceed in the absence of aggregation (Nicot and Waks. The latter value being in the same range as that found for native cytochrome c in water (see Table 1).2756 Valdez et al. The molecular volume of the protein doubles upon hydration.5 ϫ 10Ϫ11 PaϪ1 (␸k ϭ Ϫ1. It remains constant regardless of the amount of water present up to W0 ϭ 22.8 ϫ 10Ϫ11 PaϪ1 for the dimer. 8. 3). Our aim was to correlate the volumetric measurements with increasing levels of controlled hydration. In contrast. displays a high degree of conformational flexibility (Liebes et al. close to experimental error. Although the small (18. Note that in AOT reverse micelles.0. 1993).003 ϫ 10Ϫ3 ml ⅐ gϪ1. We are fully aware that. although the curve profile (Fig. Turning now to reverse micelles at the highest water content studied in this work (W0 ϭ 50). We determined the volume of the refolded protein in the aqueous core of AOT reverse micelles as a function of hydration. the protein. 1995). we find a volume of 0. It is clear that in reverse micelles the volume of MBP is independent of the amount of water present. ␤-lactoglobulin can thus serve as a reference for comparison with the basic proteins.003 ml ⅐ gϪ1. and 5. we obtain a value of ␤ ϭ 8. 1996).720 ml ⅐ gϪ1 reported by Liebes et al. in both the refolded and unfolded conformation without the use of chemical denaturing agents. the protein compressibility decreases to 10 ϫ 10Ϫ11 PaϪ1. ml ⅐ gϪ1. much less prone to proteolysis than the flexible aqueous form (Nicot et al. the strong electrostatic field present in AOT reverse micelles results in . Kharakoz and Sarvazyan.776 ml ⅐ gϪ1. Compressibility In water. in agreement with the value of 0. (1975) from the amino acid composition. provides MBP with a chaperonin-like microenvironment.01 M HCl at pH 2.0. 1986. at all other W0 values measured (17.. ␤Ϫlactoglobulin cannot be solubilized at lower W0 values. At the same time this strategy has provided useful volumetric data for a protein. 1995).725 Ϯ 0.4.

extensive interactions between the basic proteins and the micellar inner wall. (1995). This effect leads to the expulsion of electrostricted water and the concentration of counterions.764 ml ⅐ gϪ1). as well as dehydration of the uncharged surface. The compressibility ␤ of the protein in water is slightly negative and out of scale (see Table 1). all these proteins display apparent volumes similar to those found in bulk water. display the highest compressibility values (35 to 39 ϫ 10Ϫ11 PaϪ1). full protein hydration within micelles is ruled out. It is also clear that protein melting and aggregation can be excluded. The difference may be due to strong interactions occurring between a protein in close contact with a number of non-solvated surfactant polar headgroups. 1992). As discussed in the preceding paragraph. At the lowest W0 value compatible with full protein solubilization. which remain poorly hydrated. and ␤-lactoglobulin as a function of W0 (Fig. Nevertheless. Compressibility All the proteins examined at the lowest water content (in the 5.e.Protein Compressibility in Reverse Micelles 2757 FIGURE 3 Plot of the adiabatic compressibility ␤ (left {) and apparent specific volume ␸v (right ࡗ) of myelin basic protein as a function of the water content of the system. based on x-ray coordinate data for 12 globular proteins (Richards. Under these conditions (i. can make hydration data difficult to interpret. in the 8 to 11 W0 range. lysozyme. The ion solvation and solvent reorganization in the vicinity of the surfactant-solute interface must also be mentioned. They are independent of their amino acid sequence. 1977). this result may be difficult to interpret. isoelectric point. W0. contributing to increase both protein volume and compressibility. Under these experimental conditions. 1) reveals a very similar profile for the three proteins. recall that each polar AOT headgroup binds ϳ7 to 10 water molecules very tightly. which then has to compete for it. the absence of free water). and conformation. The value found is thus probably the sum of two parameters: one due to the interior of the barely-hydrated protein and a second one originating from interactions with the dry polar Biophysical Journal 80(6) 2751–2760 . do we experimentally measure intrinsic protein compressibility? The value found in this report is somewhat higher than the empirical estimates proposed by Kharakoz and Sarvazyan (1993) or Chalikian et al. 2000). the apparent specific volumes reach their highest value for all the proteins examined. At high water content.776 ml ⅐ gϪ1) obtained at low W0 are of the same order of magnitude as the partial specific volume of an average globular protein interior (0. The reduced space available within the micelle may induce a perturbation of the chemical potential of macromolecules (Minton. 1977). The anomalous volume and compressibility values of micellar water in both free and bound states. This result clearly indicates that the complex micellar system does not affect volumetric parameters of these proteins.6 to 10 W0 range). The arrow on the far right indicates the volume of MBP in water. It is remarkable that the results obtained fall within the same order of magnitude for all the proteins investigated. the compressibility of the protein interior would not be expected to vary greatly because the interior packing density of normal globular proteins falls within a narrow range (Richards. the amount of water present is not sufficient to solvate both the surfactant polar headgroups and the protein. leaving very little water available for the protein. at W0 values around 10 or less. Because of the compensating mechanisms operative in volume measurements.. including ␤-lactoglobulin. In fact. Volume The comparison of the variation in apparent specific volume of cytochrome c.. as compared to bulk water (Amararene et al. molecular weight. We discuss these below.743 to 0. It is interesting that the volume values (0.

which is also observed by time-resolved fluorescence experiments (Nicot et al. leading to a reduced amount of electrostricted water and protein structural perturbations.006 Ϯ 0. even in 6 M guanidinium hydrochloride.. and has been interpreted by Chalikian and Breslauer (1996) on the basis of the thermal volume concept. on the other hand. Yet. 1999) constitutes another example of the growing importance of an in depth understanding of the correct protein packing and folding. we observe differences between the two basic proteins. Pinheiro et al. 1985). The unfolded form of MBP displays a small negative value of adiabatic compressibility in pure water. Vidugiris and Royer. headgroups interacting with sites on the protein surface. It is also in agreement with recent studies concluding that the unfolded peptide state. is not modeled by a random-coil polypeptide chain (Dill and Shortle. Such an interpretation is confirmed by the low values of ␤ found in nonionic reverse micelles in the range of aqueous values. 1998. with probable expulsion of interfacial water. seem to level off at much higher values than those obtained in aqueous solvents. it has been recently reported that ␤Ϫinterferon administered to multiple sclerosis patients may induce lethal complications (Du- . absent in bulk aqueous solutions. Further investigation of the volumetric properties that characterize the correct folding of proteins obtained from genetic engineering should also improve our understanding of the structural properties of peptides used as therapeutic agents. Our results show that the transition between the two forms of MBP are accompanied by a very small volume variation (0. can be attributed to electrostatic interactions between the negatively charged surfactant polar headgroups and the positively charged proteins. This effect results from the tight and extensive contacts of the protein adsorbed on the micellar wall as observed on lipid membranes (Mueller et al. in parallel with the recovered heme ligation of cytochrome c and native conformation. due to an increase in the number and/or size of protein voids. 2000). linked to neurodegeneration (Aguzzi. 1998). 1974). 1985) in the 3 to 30 W0 range. 1995). We believe that information obtained in this work may prove important for the construction of de novo stable proteins. To avoid these drawbacks. Sasahara et al. 2). The volumetric properties of MBP in reverse micelles thus reflect a different aspect of the protein specific structural features. 1991). The recent emergence of misfolding related diseases.. 40% ␤-sheet) is unaffected by the water amount present (Nicot et al. 1996. that the final MBP micellar conformation (20% ␣-helix. probably due to the extent of their structural alteration in AOT reverse micelles. 1986).. may exist within the electrostatic field of reverse micelles. Lysozyme and cytochrome c compressibilities.. independent of the water concentration (Fig.. In contrast to all other proteins studied in this report. since it is known that a residual ␤-pleated sheet and hydrophobic cluster persists in the water-unfolded form of MBP (Martenson. However. which is related to volume changes for protein unfolding involving changes in the solvent-accessible surface area. which appears to indicate a substantial structural disruption with the possibility of a molten globule-like state (Bychkova et al. This effect. however. 1999).2758 Valdez et al. 1974). we observe. 1998. and can be extrapolated to its value in pure water. One of the proposed mechanisms suggests a preferential binding of the denaturants to the polypeptide chains (Lee and Timasheff. in which correct chain packing is crucial (Harbury et al. as expected. the myelin protein displays constant apparent volume and compressibility values. 2). 1996. we cannot rule out that volume effects. Cohen.003 ml ⅐ gϪ1).. the result is not inconsistent with the above empirical estimation. For example. a progressive decrease of adiabatic compressibility (Fig. on the other hand. The folding transition of MBP The apparent specific volume and compressibility of denatured. there is very little or no Biophysical Journal 80(6) 2751–2760 volume difference between the aqueous unfolded conformation. whereas cytochrome c has no disulfides. but less than the value estimated by using the additivity scheme proposed by Kharakoz (1997) for fully extended oligopeptides and polypeptides. This result has been referred to as the protein volume paradox. the exact nature of the molecular interaction between the peptide chain and the denaturing agents is not well understood. 1997). Although our data show a very modest change in MBP volume. However. obvious differences between the basic proteins and ␤-lactoglobulin. the relative conformational perturbation of lysozyme is obviously less extensive. it seems difficult to estimate their respective contributions. The protein thus displays a form of a frozen structure in reverse micelles (Waks and Beychok. unfolded proteins and peptides have been previously investigated by exposing the solutes to chemical additives (urea or guanidinium hydrochloride) and by extrapolating the denaturant concentration at infinite dilution (Tamura and Gekko.. Although cytochrome c exhibits a sharp compressibility transition between W0 values 8 and 11 (Fig. As is indeed reported in literature (Chalikian and Breslauer. The compressibility values obtained in this work between the two extreme W0 values (10 and 50) reflect indeed the actual hydration status of ␤-lactoglobulin in reverse micelles and validate our experimental strategy. The compressibility profile of the latter declines smoothly. which is not observed in ␤-lactoglobulin. and the compact refolded state of proteins. There are. When the amount of water increases. for example. we have carried out experiments with unfolded MBP in water free of denaturing agents. 3) and thus of hydration. Its four disulfide bonds act as structural stabilizers. as well as in reverse micelles where MBP refolds. We know. At this point.

Bongioanni. Nat. 1996. and P. Hackenbrock. Mueller. Oggero. Protein Sci. A. M.. 1988. K. J.. E. Biochemistry. C. R. Lee. Biotechnol. Edmonds. 1975. M. J. Kahn and Dr. 20: 409 – 474. 25:837– 843. The hydration of globular proteins as derived from volume and compressibility measurements: cross correlating thermodynamic and structural data.. In Ultrasonics. Biophysical Journal 80(6) 2751–2760 relli et al. and A. and J. Volumetric characterizations of the native. Foygel. D. P. 1974. Abagayan. 24:7024 –7032. 1986. R. and L. Imoto. On the nature of protein interior. Limited proteolysis of myelin basic protein in a system mimetic of the myelin interlamellar aqueous spaces. 1999. Waks. editor. Cooper. S. Ferrero. H-J. Biotechnol. P. Vacher. Hasegawa. Watts. J. Marzano. Academic Press. W. Ober. J. Waks. Mol. Kuwata. 1994. E. 92:3505–3511. probably resulting from incomplete folding and/or poor packing. 229:557–566. Dujsekina. Breslauer. 1995. Goto. P. 282: 1462–1467. 1981. T. I. D. Solution structure and dynamics of bovine ␤-lactoglobulin. Gindre. D.. Paci. V. 33:2451–2458. C. D. Neurology. 260:588 – 603. Rev. FEBS Lett.. Biopolymers. Phys.. Shortle.. D.. C. Totrov. Tidor. Biochemistry. The use of gel filtration to follow conformational changes in proteins: conformational flexibility of bovine myelin basic protein. V. and M. 1979. D. N. On volume changes accompanying conformational transitions in biopolymers. C. J. Sci. J. C. Biochemistry. Chalikian. 1993. Denatured states of proteins. and K. 46:1612–1622. G. 63:1090 –1100. A. P. Morais Cabral. Tiktopulo. Med. M. 39: 619 – 626. and P. 1993.. Rev. and M. 1995. Proteins as invited guests of reverse micelles: conformational effects. C. A. Batt. Yewdall. Cohen. E. 293:313–320. Le Hue ´ rou. Waters for the careful reading of the manuscript. Molten globule-like state of cytochrome c. 19. Petit. T. and M. Alber. Solubilization of enzymes and nucleic acids in hydrocarbon micellar solutions. Chatterjee. R. C. W. Durelli. A. C. Luisi. T. Plecs. E. 1997. Amararene for his contribution to the early stages of this work. J. and thank Dr. Louie. J. Nonionic surfactant reverse micelles of C12E4 in dodecane: temperature dependence of size and shape. V. 2000. J. and K. T. 1999. Lipid specificity in the interaction of cytochrome c with anionic phospholipid bilayers revealed by solidstate 31P NMR. H. Hardy. Mol. 1994. Urbach. Vincent. Chem. and W. Timasheff. Gallay. D. We also express our gratitude to Dr. N.. E. Martenson. 1996. 1983. Gindre. Membrane proteins in reverse micelles: myelin basic protein in a membranemimetic environment. M. Kahn. Voglino. Vol. Rizzetto. N. Engin. Brownlow. S. A. 13:267–314. Heyman. 100:15180 –15186. J. 1998. Biochem. M. Mol. Micellar solubilization of biopolymers in organic solvents. Spector. 5:481– 495. and Y. J. Biochemistry. Nicot. 1990. Kahn. Enzymes in water-in-oil microemulsions (reversed micelles): principles and applications. and A. Bayer. Nicot. molten globule and unfolded states of cytochrome c at acidic pH. and G. Biophys.8 Å resolution-still an enigmatic lipocain.. Kim. We thank Dr. 35:6058 – 6063. Protein. Neurochem. Sarvazyan. A. 61:682– 689. 1986. Biochemistry. Structural and kinetic description of cytochrome c unfolding induced by the interaction with lipid vesicles. Policarpov. Smith.. 1981. I. Biol. J. Hoshino. Chem. Biol. Urbach. L. J.. Rev. Chem. C. C. J. Structure. 4:1125–1126.. A. 83:2706 –2714. . T. T. and M.. K. and M. 1986. 1998. Roder. D. Klapper. and P. U. K. 1992. J.. Cantrell. New York. 23:10276 –10285. Biochemistry. REFERENCES Aguzzi. 1996. H. 1998. Proc. A. Phys. R.. Boyer.. Adiabatic compressibility of AOT [sodium bis(2ethylhexyl)sulfosuccinate] reverse micelles: analysis of a simple model based on micellar size and volumetric measurements. M. Acta.. H. and H. J. Engin. Breazaele. Breslauer.A. In The Enzymes. A. and L. Biopolymers. E. Kharakoz. Watts. J. and J. Possible hydrophobic region in myelin basic protein consisting of an orthogonally packed beta sheet. Nicot. 665– 808. D. Compressibility of proteins at 25°C. M. North. Waks. and C. High resolution three dimensional structure of horse heart cytochrome c. T. 1985. P. 1996. Rupley. Phys. 36:13122–13132. Confinement as a determinant of macromolecular structure and reactivity. Harbury. Bovine ␤-lactoglobulin at 1..Protein Compressibility in Reverse Micelles 2759 Gekko. applications. A. 50:570. L. S. New York. Elo ¨ ve. P. Compressibility-structure relationship of globular proteins. 1971. and O. N. Biochim.. S. T. Amararene. Ultrasonic wave velocity and attenuation measurements. Gavish.-Y.S. Genet. F. 250:291–306. Merdas.. V. 2000. A. P. Noguchi. and C. 8:2541–2545. Dill. Natl. B. Chem. T. H. Lipid specificity for membrane mediated partial unfolding of cytochrome c. A. Biol. Liebes. de Jongh. L. Acad. J. Grandi. Vacher. CRC Crit. L. Solution behavior. 1995. On the volume of macromolecules. Biochim.... Biochemistry. and E. Pileni. 60: 1283–1291. Biol. J. and Y. Johnson. Phillips. A. Biochem. S. M. J. K. P. L. J. Sawyer. Magid. Bychkova. Sci. Science. K. W. J. S. E. High-resolution protein design with backbone freedom. Acta. Pinheiro. J. V. Adiabatic compressibility of globular proteins. A. V. E. A. 1997. Forge. 1997. Killian. J. P. Nicot. 13:11–26. Velikson. Protein conformation dictates prion strain. Gindkin. 360:255–260. J. Philips. and D.. L.. R. T. Annu. Klenin. Partial specific volumes and interactions with solvent components of proteins in guanidine hydrochloride. S. G. K. 60:795– 825. S. Cytochrome c in sodium bis 2-(ethylhexyl) sulfosuccinate reverse micelles: structure and reactivity. 41:785–797. and M. Chem. T. Revs. 1995. J. P. 214: 585–595. A. Oldfield. 405:27–39. J. B. Waks. R. significance. S. A. and J. and S.. Mol. Physiol. Brochette. Lysozyme. R. P. B. Zand. F. P. 1997. circular dichroism and 220MHz PMR studies of the bovine myelin basic protein. Biochemistry.. Uversky. G. Biopolymers. Chem. Biophys. and H. Breslauer. Ptitsyn. 4:1426 –1429. Chalikian. V. Protein misfolding and prion diseases. and K. C. 104:4552– 4559. Rev. Martenson. Pinheiro. Multiple conformations of physiological membrane-bound cytochrome c. Hydrational and intrinsic compressibilities of globular proteins... V. Era. Gratton.. L. 37:6402– 6409. C. Denoroy. 80:750 –754. R. Cortese.. B. I.. Adsorption of membraneassociated proteins to lipid bilayers studied with an atomic force microscope: myelin basic protein and cytochrome c. H. 13:257–265. North. Biol. J. Flower. and B. Phys. 1972. Biol.. Valdez. C. Minton. Chalikian. Volume changes of the molten globule transition of horse heart ferricytochrome c: a thermodynamic cycle. J. R. 25:8885– 8893. D. B. Luisi. C. Buff. 1978. Bushnell. C. Gekko. 1991. 1998). Nicot. Neurochem. Academic Press. This work was supported in part by the Centre National de la Recherche Scientifique (Physique et Chimie du Vivant) and a grant from the Simone and Cino Del Duca Fundation. Ritsema. editor. A. Interferon treatment for multiple sclerosis: autoimmune complications may be lethal. Partial volumes and compressibilities of extended polypeptide chains in aqueous solution: additivity scheme and implication of protein unfolding at normal and high pressure. Kharakoz. 25:6563– 6571. Bamberg. T. M. Genet. Waks. and R. 67–135. 1998. E. E. 12:255–327. R. Biophys.

Biophys. Yedgar. F. 1997. S. Induced conformational states in human apohemoglobin on binding haptoglobin 1–1. and B. M. A. and unfolded state. P. 113:15–21. Biol. and S. Chem. M. A. and C. Ye... Honig. Developments of methods of precise ultrasonic measurements in small volumes of liquids. Glycerol decreases the volume and compressibility of protein interior. J. Vidugiris. 1995. and K. Takano. P. 34:1878 –1884. Tamura. Contribution of water molecules in the interior of a protein to the conformational stability. Annu. Rev. Biophys. 20:321–342. Beychok.. Areas. and B. Internal cavities and buried waters in globular proteins. D. Phys. Mol. Rashin. 1991. Waks. Sakurai. and K. J. Y. Annu. 1996. 1998. A. Effect of added heme as a probe of frozen structures. The volume and compressibility changes of lysozyme associated with guanidinium chloride and pressure-assisted unfolding. S. Biophys. Determination of the volume changes for pressure-induced transitions of apomyoglobin between the native. packing and protein structure.. Weitz. 26:4. A. Y. Sound propagation in sodium di-2ethyl-hexylsulfosuccinate micelles and microemulsions. 1977. J. volumes. Biochemistry. Liu. Proteins and peptides in water-restricted environments. Iofin. Biochemistry. 75: 463– 470. molten globule. S.. Richards. Lim. Huang. Yamagata. Royer. 6:151–176. and J. L. and K. 44:8249 – 8263. J. 35: 2061–2066. Bioengin. Gekko. Rev. J. Richards. A. Sarvazyan. Q. 1994. Yutani. M. A. Rev. Biophys. 1986. 1982. 274:132–142. Valdez et al. J.. Sheng. 1986. Gavish. Biophys. A. P. Sasahara. A. K. 1999. Fujii.. F. Compactness of thermally and chemically denatured ribonuclease A as revealed by volume and compressibility. Rev. G.. Biol. An analysis of packing in the protein folding problem. 291:693–701. A. M. Sarvazyan. Waks.. Biophysical Journal 80(6) 2751–2760 . Ultrasonics. 20:151–154.: 423– 498. Biochemistry. 1991. A. and W. Funahashi. Nitta. Proteins. Mol. Ultrasonic velocimetry of biological compounds.2760 Priev. K. M. Almagor. 23:3619 –3625. 1:4 –15. 1974. Biochemistry. M.