Novel zwitterionic reverse micelles for encapsulation of proteins in low-viscosity media

Article abstract of DOI:10.1002/chem.200501422

Large proteins remain inaccessible to structural NMR studies because of their unfavorable relaxation properties. Their solubilization in the aqueous core of reverse micelles, in a low-viscosity medium, represents a promising approach, provided that their

Full text of DOI:10.1002/chem.200501422

DOI: 10.1002/chem.200501422
Novel Zwitterionic Reverse Micelles for Encapsulation of Proteins in
Low-Viscosity Media
Sylvain Doussin,[a] Nicolas Birlirakis,*[b] Dominique Georgin,[c] FrØdØric Taran,[c] and
Patrick Berthault*[a]
Abstract: Large proteins remain inac-
cessible to structural NMR studies be-
cause of their unfavorable relaxation
properties. Their solubilization in the
aqueous core of reverse micelles, in a
unfolding, due to strong electrostatic
interactions between the polar head
groups and the protein charges. To
design reverse micelles in which these
interactions are weakened, a new zwit-
terionic surfactant molecule was syn-
thesized and studied by high-resolution
NMR spectroscopy, for which cyto-
chrome C and 15N-labeled ubiquitin
were used as guest candidates. At dif-
ferent ionization states, both proteins
are encapsulated in the absence of salts
or other additives, in a folded confor-
mation similar to the native one.
low-viscosity medium, represents
a
promising approach, provided that
their native tertiary structure is main-
tained. However, the use of classical
ionic surfactants may lead to protein
Keywords: micelles · NMR spec-
AHCTREUNG
Introduction
and 15N-labeled macromolecules, have been used to extend
the size limit of the molecules studied by solution-state
NMR spectroscopy.
The study of large biomolecular systems by high-resolution
NMR spectroscopy is impeded by their slow rotational
motion, which leads to a very efficient transverse relaxation
and, thus, to line broadening and shortening of signal life-
time. Relaxation-interference methods based on chemical-
shift anisotropy-dipole–dipole cross-correlation in transverse
relaxation optimized spectroscopy (TROSY)-type experi-
ments,[1] combined with the extensive deuteration of 13C-
In 1998, A. J. Wand and co-workers proposed a quite in-
novative approach for the structural study of large proteins
(>30 kDa).[2] This involves the substitution of water as bulk
solvent by a low-viscosity medium, which would accelerate
the soluteꢀs Brownian motion and, therefore, lengthen its T2
relaxation time. To maintain the natural environment of the
protein for the study of its native three-dimensional struc-
ture, it was proposed to encapsulate it with some hydration
layers into reverse micelles.
[a] S. Doussin, Dr. P. Berthault
Although this ambitious approach was validated for ubiq-
uitin (8 kDa) in reverse micelles of sodium bis(2-ethylhex-
yl)sulfosuccinate (AOT), it has not yet been applied to the
structural studies of large proteins. Indeed, the problems en-
countered are numerous, for example: 1) some surfactants
lack the swelling capacity necessary to encapsulate large
molecular objects, 2) certain micellar systems suffer from
time-instability, 3) the proteins may be encapsulated, but in
a non-native state. Such a conformational transition of pro-
teins encapsulated in AOT micelles could be explained by
the presence of strong electrostatic interactions between the
surfactant polar head groups and the charged residues of
the guest molecule. Therefore, the attenuation of these in-
teractions through the design of new surfactants may facili-
tate protein structural studies in reverse micelles. Here, a
comparison of AOT, CTAB (cetyltrimethylammonium bro-
Laboratoire Structure et Dynamique par RØsonance MagnØtique
DSM/DRECAM/Service de Chimie MolØculaire
URA CEA/CNRS 331, CEA/Saclay, 91191 Gif sur Yvette (France)
Fax : (+33)169-089-806
E-mail: pberthault@cea.fr
[b] Dr. N. Birlirakis
Laboratoire de Chimie et Biologie Structurales
Institut de Chimie des Substances Naturelles
CNRS, 91190 Gif sur Yvette (France)
Fax : (+33)169-823-784
E-mail: nicolas.birlirakis@icsn.cnrs-gif.fr
[c] D. Georgin, Dr. F. Taran
Laboratoire de Marquage au 14C
DSV/DBJC/Service de Marquage MolØculaire et
Chimie Bioorganique
CEA/Saclay, 91191 Gif sur Yvette (France)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
4170
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4170 – 4175

FULL PAPER
mide), and a newly synthesized zwitterionic surfactant, 2-
ammonioethyl 2,3-bis(3-ethylheptanoyloxy)propyl phos-
phate, is presented. Unlabeled cytochrome C was chosen as
a guest candidate, and H NMR spectroscopy was used to
confirm not only that the protein was encapsulated, but also
that it had maintained its folded structure.
For example, in the case of AOT micelles in alkanes, it
has been proposed that the protein is located in the interfa-
cial region of the reverse micelles, due to strong electrostatic
attractions between cytochrome C and the negatively charg-
ed surfactant layer of AOT.[6,7] This local environment leads
to a conformational transition of cytochrome C from its
folded structure to an amphiphilic entity.[8] Hence, the
NOESY spectra of cytochrome C in AOT micelles dissolved
in pentane, shown in Figure 1c and d, exhibit a very peculiar
behavior. Whereas the spectrum in AOT/pentane at pH 9.3
reflects a folded protein (without indication of its three-di-
mensional structure), the spectrum at pH 6.4 is more indica-
tive of a molten globule. A reduced amide-proton region,
similar to that of Figure 1d, was also observed by the group
of Wand.[9] By comparing the experiments made in aqueous
solution (Figure 1a and b), the rates of proton exchange
with the solvent cannot be responsible for this spectral deg-
radation. The variation in protein charge between the two
pH values (net charge +5.3 and +9.8, for pH 9.3 and 6.4, re-
spectively), leading to different electrostatic strengths, re-
mains the only explanation for the appearance of the
NOESY maps displayed in Figure 1c and d.
To confirm that the sulfosuccinate function of AOT is ac-
tually responsible for this effect, an AOT analogue with the
same head group, but a different hydrophobic tail, providing
higher solubility in apolar solvents,[10] was synthesized. The
NMR spectra of cytochrome C in micelles of this compound
at two pH values seem very similar to those displayed in
Figure 1c and d (see Supporting Information), indicating
that the nature of the head group is probably the cause of
the observed effects. This also suggests that the influence of
the surfactant tail on the protein conformational state is
rather minor.
1
Results and Discussion
The choice of commercially available cytochrome C for our
experiments was dictated by reasons of simplicity, as the so-
lution becomes red upon solubilization of the protein. More-
over, cytochrome C, being a basic protein (pI=10.2) of 104
amino acids, is strongly positively charged at neutral pH.
Due to the presence of several hydrophilic residues at its ex-
ternal surface, cytochrome C remains highly soluble in
water, even at pH conditions close to its isoelectric point.
For unlabeled proteins, the NOESY experiment can give
a rapid estimation of their conformational state. The pres-
ence of numerous cross-peaks between amide protons com-
bined to an expanded amide region indicate a folded pro-
tein. This is the case of the oxidized paramagnetic ferrocyto-
chrome C, at basic or neutral pH, as shown in Figure 1a and
b. The increase in spectral quality achieved by lowering the
pH from 9.3 to 6.4 reflects the decrease in proton-exchange
rates between water and protein amide groups.[3]
For the design of our new surfactant, some characteristics
were desired. Firstly, it should be able to form reverse mi-
celles in apolar solvents and to encapsulate proteins of dif-
ferent isoelectric points at the pH values used for NMR
studies in aqueous solution. Secondly, the use of additives,
such as linear-chain alcohols, should be avoided if possible.
Indeed, with respect to the final objective of the study of
large proteins, the addition of such cosolvents in significant
proportions could become a drawback as they increase the
viscosity of the medium. Finally, the micellar system should
not require a buffer or high salt content.
Clearly, the design of a new surfactant capable of encap-
sulating various protein types in organic solvents is some-
what empirical. In dealing with the shape of the aggregates,
some basic rules exist, such as consideration of the packing
parameter, the hydrophilic–lipophilic balance, or use of the
flexible surface model.[4] However, no accurate prediction of
the objects formed can be made easily, which is even more
the case when a protein is present, even at low concentra-
tion.
For the choice of the polar head, some considerations
from previous works were applied. The article of Shioi et al.
on protein extraction[5] suggests that ionic polar heads favor
the formation of reverse micelles in organic solvents, the
transfer of the protein into the aqueous core being governed
by electrostatic interactions. However, if the protein is
strongly charged, these interactions are also likely to occur
inside the formed micelles between these charges and the
anionic or cationic polar heads.
The encapsulation of 15N-labeled cytochrome C in cationic
CTAB reverse micelles was reported recently.[11] In such an
environment, the native conformation of the protein is pre-
served, because the electrostatic interactions between the
protein and the CTAB head groups are now repulsive.[12]
However, in the search for a unique surfactant system capa-
ble of encapsulating proteins of different isoelectric points,
this cannot represent our final solution. We suspect that cat-
ionic micelles would unfold acidic proteins if the pH of the
aqueous core is largely inferior to the protein isoelectric
point. However, Wand et al. have successfully encapsulated
a mutant of flavodoxin (pI~4.2) in a folded state in CTAB
micelles. Nevertheless, the formation of CTAB micelles in
short-chain alkanes is quite difficult and requires the addi-
tion of cosolvents, such as n-hexanol at a proportion of 8–
10% of the total volume, thereby penalizing the viscosity of
the solution.[13] In the case of flavodoxin in CTAB/n-hexanol
micelles, the presence of the alcohol may dilute the charge
of the micelle, and may explain the folded form of the pro-
tein. In the absence of cosolvent, we were unable to solubi-
lize cytochrome C in CTAB/pentane with a water-to-surfac-
tant molar ratio W of 20, even after considerable sonication.
After addition of n-hexanol at 10% v/v, the solution turns
slightly red, however, a proportion of the protein remains
Chem. Eur. J. 2006, 12, 4170 – 4175
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4171

P. Berthault, N. Birlirakis et al.
Figure 1. Partial contour plots of two-dimensional NOESY spectra of cytochrome C at 298 K (mixing time 150 ms). Each sample contained around 1 mg
of protein, except for (f), in which it was estimated as 0.3 mg. (a) and (b): Spectra recorded in H2O/D2O of 95:5; (a) pH 9.3, (b) pH 6.4. (c) and (d): Spec-
tra recorded in the AOT/
N
N
system; (e) pH 9.3, W=8, (f) pH 6.4, W=12.0.
precipitated (see Supporting Information for the recorded
NOESY spectrum). This surfactant is, therefore, not suitable
for our purpose.
Mixtures of anionic, cationic, and even nonionic surfac-
tants have been proposed recently by the group of Wand to
encapsulate proteins in low-viscosity solvents, such as liquid
ethane.[13,14] These surfactant mixtures appear to increase
the stability of the micelles formed, as they facilitate the en-
capsulation of proteins in low-density solvents. The pres-
sures required to stabilize the microemulsions can be de-
4172
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4170 – 4175

Zwitterionic Reverse Micelles
FULL PAPER
creased if an appropriate cosolvent is used. Thus, a real
breakthrough is achieved by the lengthening of the trans-
verse relaxation time. However, in view of the work of sev-
eral teams who use catanionic reverse micelles for the syn-
thesis of hierarchical nanostructures, the risk of a charge re-
distribution/segregation along the inner surface of the mi-
celle cannot be neglected. The cationic-to-anionic surfactant
ratio influences the structure of the aggregates, and this
value is taken into account in the modification of the size
and shape of the nanocrystals growing in these aggregates
(see, for example, reference [15]). Moreover, adaptation of
the micellar charges according to the surface-charge distri-
bution of the protein is possible.
protein due to electrostatic interactions. Clearly, the risk of
redistribution of micellar charge according to the protein
surface potential, as can be the case for a mixture of surfac-
tants (see below), is avoided. This surfactant may act itself
as a buffer for pH levels between its two pK values: ~5 for
the phosphate group and ~8 for the ammonium function.[17]
Based on geometric considerations and basic predictions,
we have kept an AOT-like hydrophobic tail to create the
curvature required for an inverse micelle. The new zwitter-
ionic surfactant, 2-ammonioethyl 2,3-bis(3-ethylheptanoyl-
A
a synthetic pathway that allows for a future perdeuteration
(Figure 2).
Another solution that combines the presence of charged
groups in the polar head and the decrease in micellar net
global charge is to conceive a zwitterionic surfactant. It has
been reported that lecithin, a zwitterionic natural phospholi-
pid (a mixture of phosphatidylcholines with acyl chains of
different lengths and degrees of saturation), self-assembles
in alkanes into long, wormlike objects in the presence of
small amounts of water.[16] Such behavior observed at high
During preparation of the samples of cytochrome C en-
capsulated in micelles of either AOT or 1, a remarkable dif-
ference was observed. Whereas the dissolution of the pro-
tein after addition of AOT micelles is immediate, in the case
of 1, it takes much longer (several hours, even with ultra-
sound). For the same reason as stated previously, this may
be due to the zwitterionic nature of the surfactant polar
head, which could hide the net electrostatic interactions re-
quired to form the reverse micelles.
Figure 1e and f displays the NOESY contour maps of cy-
tochrome C initially prepared at pH 9.3 and 6.4, respectively,
in reverse micelles of surfactant 1. The spectra are similar
(the lower signal-to-noise ratio in spectrum f is due to a
lower protein concentration). As for the aqueous solution
studies, and contrary to the spectrum recorded in AOT at
neutral pH, the amide protons cover a wide spectral region,
which characterizes a folded protein.
surfactant concentrations could arise from the large size of
+
the polar head group (containing a N
A
tive to that of the hydrophobic tail. For this reason, the new
surfactant molecule, 2-ammonioethyl 2,3-bis(3-ethylhepta-
noyloxy)propyl phosphate (compound 1), possesses a small-
er head group (NH3+), namely that of phosphatidylethanol-
amine, the phospholipid located primarily at the external
surface of biological membranes. The zwitterionic head pre-
vents the release of counter ions into the water pool and the
balance between positive and negative fixed ions gives a
neutral character to the internal wall, making the protein–
micelle interactions weaker than those with other ionic sur-
factants. The alternation of fixed opposite charges on the in-
ternal surface of the micelle reduces perturbation on the
Experiments to encapsulate larger proteins are underway
in our laboratory. The use of nonionic surfactants from the
poly(oxyethylene) alkyl ether family (CiEj) to encapsulate
cytochrome C at various pH values was also successful in
our hands, however, the resulting spectra gave broad peaks
Figure 2. Pathway for synthesis of surfactant 1. Details of steps i–vii are given in the Experimental Section.
Chem. Eur. J. 2006, 12, 4170 – 4175
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4173

P. Berthault, N. Birlirakis et al.
and the protein was not fully solubilized. Moreover, the
region of the phase diagram indicating reverse micelles in
short-chain alkanes is very sharp and depends strongly upon
temperature.[18] For these last experiments, the addition of
AOT to C12E4 could have improved the encapsulation of cy-
tochrome C, as electrostatic interactions between the surfac-
tant head groups and the protein are paradoxically impor-
tant for this.[5]
Conclusion
This new zwitterionic surfactant appears to facilitate the en-
capsulation of proteins of low or high isoelectric point in the
absence of any buffer or salt. The formation of reverse mi-
celles occurs without the need for any cosurfactant. For
1
both ubiquitin and cytochrome C, the line widths of the H
or 15N NMR spectra show that small micellar objects are
formed in short-chain alkanes, with W values close to 10.
These reverse micelles are stable over a very long period:
after several months no precipitation or phase separation
was observed. The encapsulation of larger proteins and their
NMR analysis in low-viscosity solvents is now underway in
our laboratory. It is also expected that modification of the
aliphatic tail to increase solubility of this surfactant in
apolar solvents would improve its capacity to encapsulate
proteins in greater quantities, or in solvents of lower viscosi-
ty.
Another critical point in the usefulness of our new surfac-
tant in the approach of Wand is its solubility in low-viscosity
solvents and the time-stability of the micelles containing a
protein. Indeed, in the subcritical region, for a chemical
family, such as n-alkanes, viscosity of a medium is strongly
related to its solvating power through the cohesive energy
densities, as introduced by Hildebrand.[19] Thus, the solubili-
ty of this new surfactant was evaluated in an ethane/pro-
pane/pentane mixture (ca. 45:22:33 v/v). 15N-labeled ubiqui-
tin, a protein of 76 amino acids (pI=6.6) used as a model
1
compound by Wand, was chosen. The H,15N-HSQC spec-
trum displayed in Figure 3 resembles that of ubiquitin in
water, thereby indicating a similar three-dimensional struc-
ture. The spectrum obtained with eight scans per transient,
by using a 500 MHz spectrometer equipped with a classical
inverse probe head, exhibits an excellent signal-to-noise
ratio. This is encouraging for future studies of larger pro-
teins by using triple-resonance techniques. The stability of
the micellar system was confirmed, as the same experiment
repeated one month later showed no alteration in the
HSQC spectrum.
Experimental Section
Syntheses
Reagents, solvents, and cytochrome C were purchased from Fluka. 15N-la-
beled ubiquitin was purchased from VLI Research and deuterated sol-
vents from Eurisotop.
Synthesis of the AOT analogue: The AOT analogue was synthesized ac-
cording to S. Nave et al.,[20] overall yield of 66.5%. ESI-MS: m/z=467.2
[M+Na] corresponding to m/z=444.2 [M]. The purity of the product
(white solid) was verified by NMR spectroscopy.
Synthesis of compound 1
Step i: dibenzyl phosphoryl chloride: Diphenyl phosphonate (7.5 g,
28.5 mmol) and chlorosuccinimide (3.75 g, 28.5 mmol) were dissolved in
50 mL of anhydrous toluene and stirred for 2 h at ambient temperature.
The mixture was filtrated in an anhydrous atmosphere and the solvent
was evaporated under reduced pressure. Due to its instability, the crude
product was used in the subsequent reaction without any purification.
Step ii: dibenzyl (2,2-dimethyl-1,3-dioxolan-4-yl)methyl phosphate: Solke-
tal (2.5 g, 18.5 mmol) was dissolved in 45 mL of anhydrous pyridine and
the mixture was cooled in an ice bath. Dibenzyl phosphoryl chloride was
introduced under vigorous stirring and the mixture was left for 2 h at
08C. Unreacted dibenzyl phosphoryl chloride was hydrolyzed by the ad-
dition of aqueous sodium carbonate. After extraction with dichlorome-
thane and purification by flash chromatography using hexane/ethyl ace-
tate (6:4 v/v), 2.1 g of the product was obtained (29%, based on solke-
tal).
Step iii: dibenzyl 2,3-dihydroxypropyl phosphate: Dibenzyl (2,2-dimethyl-
1,3-dioxolan-4-yl)methyl phosphate (2.1 g, 5.35 mmol), previously ob-
tained, was mixed with 1.5 g of Dowex 5W-X8(H+) resin and 40 mL of
methanol. After 4 h, 1.5 g of resin was added and the reaction was left
for 48 h under stirring. The mixture was filtrated and purified by flash
chromatography using ethanol/ethyl acetate (1:9 v/v) to yield 1.03 g of
product (55%).
Step iv: 3-(bis(benzyloxy)phosphoryloxy)propane-1,2-diyl bis(3-ethylhep-
tanoate): Dicyclohexylcarbodiimide (3.352 g) dissolved in 10 mL of di-
chloromethane was added at 08C under stirring to a solution of dibenzyl
2,3-dihydroxypropyl phosphate (1 g, 3.125 mmol), 3-ethyl-heptanoic acid
(1.169 g, 7.40 mmol, obtained in 63% yield by chain elongation from 2-
ethylhexyl bromide), and dimethylaminopyridine (763 mg, 6.25 mmol) in
20 mL dichloromethane. The reaction proceeded for 10 min at 08C and
overnight at ambient temperature. After filtration of the precipitated
urea and evaporation of the solvent, the crude mixture was purified by
Figure 3. 1H,15N-HSQC spectrum of ubiquitin encapsulated in micelles of
surfactant 1 dissolved in an ethane/propane/
A
45:22:33 v/v. Eight scans were recorded over 160 t1 increments.
4174
www.chemeurj.org
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4170 – 4175

Zwitterionic Reverse Micelles
FULL PAPER
flash chromatography using pentane/diethyl ether (2:1 v/v) to yield 1.26 g
of product (63%).
NMR spectroscopy: The conformational state of cytochrome C in various
environments was assessed by conducting NOESY experiments. In the
absence of 15N-enriched protein, the results of these 1H NMR experi-
ments gave an insight into both the encapsulation and the folding state of
the protein. However, due to the presence in the NMR spectra of the in-
tense surfactant peaks (concentration~100 mm), a dedicated pulse se-
quence was required for the spectra shown in Figure 1c–f. It was essen-
tially the classical NOESY pulse scheme followed by a frequency-selec-
tive spin echo (Gradient–Soft 1808–Gradient) prior to acquisition. The
soft pulse selectively inverted the 6–11 ppm region and enabled observa-
tion of through-space polarization transfer involving the amide/aromatic
Step v: 2,3-bis(3-ethylheptanoyloxy)propyl phosphate: An amount of 5%
palladium on charcoal (125 mg) was added to a solution of 3-(bis(ben-
zyloxy)phosphoryloxy)propane-1,2-diyl bis(3-ethylheptanoate) (1.26 g) in
methanol, and the mixture was stirred overnight under a hydrogen at-
mosphere. After Celite filtration and solvent evaporation, the crude mix-
ture was purified by flash chromatography using dichloromethane/metha-
nol (9:1 v/v) to yield 0.522 g of product (58%).
Step vi: 2-tert-butoxycarbonylamminoethyl-2,3-bis(3-ethylheptanoyloxy)-
propyl phosphate: Triisopropylbenzene sulfonyl chloride (200 mg) dis-
solved in 4 mL of anhydrous pyridine was added to a solution of 2,3-
bis(3-ethylheptanoyloxy)propyl phosphate (100 mg, 0.22 mmol) and N-
Boc-ethanolamine (69 ml, 0.44 mmol, BOC=tert-butyloxycarbonyl) in
6 mL of anhydrous pyridine. After 4 h of stirring, the reaction was
quenched by the addition of water. Following solvent evaporation under
reduced pressure, diethyl ether was added and the precipitated triisopro-
pylbenzene sulfonic acid was filtered out. Purification by flash chroma-
tography using dichloromethane/methanol (95:5 v/v) gave 134 mg of
product (99%).
1
protons. As the H spectrum covers a range from À25 to +25 ppm, clear-
ly some amide protons are excluded from the plot. These NMR experi-
ments were performed by using a Bruker DRX800 spectrometer (18.7 T)
equipped with a 5-mm z-gradient HCN cryoprobe.
To characterize the conformational state of 15N-labeled ubiquitin in mi-
celles of 1, sensitivity-enhanced 1H,15N-HSQC spectroscopy was per-
formed at 293 K by using a Bruker DRX500 spectrometer (11.7 T) equip-
ped with a 5-mm xyz-gradient broadband inverse probehead.
Step vii: 2-ammonioethyl-2,3-bis(3-ethylheptanoyloxy)propyl phosphate:
Deprotection of the amino function was carried out by using a solution
of 1 mL of trifluoroacetic acid in 10 mL of dichloromethane. After
30 min, the solvent was evaporated under reduced pressure and the prod-
uct was purified by flash chromatography using dichloromethane/metha-
nol (95:5 v/v) to yield 107 mg of final product (95%).
[1] K. Pervushin, R. Riek, G. Wider, K. Wꢂthrich, Proc. Natl. Acad.
Sci. USA 1997, 94, 12366–12371.
[2] A. J. Wand, M. R. Ehrhardt, P. F. Flynn, Proc. Natl. Acad. Sci. USA
1998, 95, 15299–15302.
[3] K. Wꢂthrich, NMR of Proteins and Nucleic Acids, Wiley-Inter-
science, New York, 1986.
[4] E. F. Marques, O. Regev, A. Khan, B. Lindman, Adv. Colloid Inter-
face Sci. 2003, 100–102, 83–104.
ESI-MS: m/z=518.3 [M+Na] corresponding to m/z=495.3 [M]. The
purity of the product (white solid) was verified by NMR spectroscopy.
Sample preparation: The samples of cytochrome in reverse micelles were
prepared by the solid–liquid transfer method.[21] The protein was dis-
solved in water, the pH was adjusted by the addition of HCl, and the pro-
tein was lyophilized. It was then added to a solution containing the sur-
factant, deionized water, and [D12]pentane (98% enriched). At this step,
the amount of water added was that required for W to reach a value of
10. The final W values, estimated from the 1H NMR spectra, are indicat-
ed in the legend of Figure 1c–f. The AOT concentrations were 70 mm in
500 mL. For spectra c and d, 1 mg of protein was used. The concentrations
of 1 were 113 mm in the case of spectrum e (final W value of 8), and
37 mm for spectrum f, due to the limited quantities of the new surfactant.
In the latter case, the final W value was 12, as measured from the
1H NMR spectrum. The amounts of cytochrome C were 1 and 0.3 mg for
spectra e and f, leading to concentrations of 0.15 and 0.05 mm, respective-
ly.
[5] A. Shioi, M. Harada, H. Takahashi, M. Adachi, Langmuir 1997, 13,
609–616.
[6] M. Adachi, M. Harada, J. Phys Chem. 1993, 97, 3631–3640.
[7] G. Cassin, Y. Duda, M. Holovko, J. P. Badiali, M. P. Pileni, J. Chem.
Phys. 1997, 107, 2683–2693.
[8] M. Adachi, M. Harada, J. Colloid Interface Sci. 1994, 165, 229–235.
[9] R. W. Peterson, M. S. Pometun, Z. Shi, A. J. Wand, Protein Sci.
2005, 14, 1–3.
[10] J. Eastoe, A. Paul, S. Nave, D. Steytler, B. Robinson, E. Rumsey, M.
Thorpe, R. K. Heenan, J. Am. Chem. Soc. 2001, 123, 988–989.
[11] B. G. Lefebvre, W. Liu, R. W. Peterson, K. G. Valentine, A. J. Wand,
J. Magn. Reson. 2005, 175, 158–162.
[12] K. Vos, C. Laane, S. R. Weijers, A. van Hoek, C. Veeger, A. J. W. G.
Visser, Eur. J. Biochem. 1987, 169, 259–268.
[13] Z. Shi, R. W. Peterson, A. J. Wand, Langmuir 2005, 21, 10632–
10637.
[14] R. W. Peterson, B. G. Lefebvre, A. J. Wand, J. Am. Chem. Soc. 2005,
127, 10176–10177.
[15] H. Shi, L. Qi, J. Ma, H. Cheng, B. Zhu, Adv. Mater. 2003, 15, 1647–
1651.
For the ubiquitin sample, the micellar solution was prepared by the injec-
tion method.[21] The lyophilized protein (1.02 mg) was hydrated by 5.2 mL
of deionized H2O, added by using a Hamilton syringe placed directly into
the NMR tube. A solution containing 17.9 mg of the zwitterionic surfac-
tant dissolved in 500 mL of pentane was added to the hydrated protein.
Manual shaking allowed homogenization of the solution, and a fast
1H,15N-HSQC analysis was performed to ensure that the protein was en-
capsulated in a folded state. An evaporation method[11] was then used to
obtain the final solvent mixture. All of these steps were performed in an
[16] R. Angelico, A. Ceglie, U. Olsson, G. Palazzo, Langmuir 2000, 16,
2124–2132.
[17] D. C. Steytler, D. L. Sargeant, G. E. Welsh, B. H. Robinson Eur. J.
Biochem. 1996, 12, 5312–5318.
[18] A. Merdas, M. Gindre, R. Ober, C. Nicot, W. Urbach, M. Waks, J.
Phys. Chem. 1996, 100, 15180–15186.
[19] J. H. Hildebrand, R. L. Scott, Regular Solutions, Prentice-Hall, New
Jersey, 1962.
NMR sapphire tube, purchased from the laboratory of A. Merbach.[22]
A
dedicated pressure line with a manual pump and pressure gauges was
constructed and used to fill the NMR tube with propane, then ethane.
The solvent volume was determined by measurement of the height of the
liquid in the NMR tube. The propane/ethane ratio was estimated by rela-
tive integration on the 1H NMR spectrum, and the W value was mea-
[20] S. Nave, J. Eastoe, J. Penfold, Langmuir 2000, 16, 8733–8740.
[21] N. Klyachko, A. V. Levashov, Curr. Opin. Colloid Interface Sci.
2003, 8, 179–186.
sured on the final sample at pH 6.2. No precipitation appeared at the
bottom of the tube, however, a small viscous droplet could be observed
some minutes after shaking. The maximum concentration of 15N-enriched
ubiquitin observed in Figure 3 is, thus, 0.25 mm, if one considers that all
of the protein is solubilized.
[22] A. Cusanelli, U. Frey, D. T. Richens, A. E. Merbach, J. Am. Chem.
Soc. 1996, 118, 5265–5271.
Received: November 16, 2005
Published online: March 16, 2006
Chem. Eur. J. 2006, 12, 4170 – 4175
ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4175