| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Biophys J, February 1998, p. 918-930, Vol. 74, No. 2
*Institut Curie, The interaction of dodecyl maltoside with lipids was
investigated through the studies of solubilization and reconstitution processes. The solubilization of large unilamellar liposomes was analyzed through changes in turbidity and cryo-transmission electron microscopy. Solubilization was well described by the three-stage model
previously reported for other detergents, and the critical detergent/phospholipid ratios at which lamellar-to-micellar transition occurred (Rsat = 1 mol/mol) and finished
(Rsol = 1.6 mol/mol) were determined. The
vesicle-micelle transition was further observed in the vitrified
hydrated state by cryo-transmission electron microscopy. A striking
feature of the solubilization process by dodecyl maltoside was the
discovery of a new phase consisting of a very viscous "gel-like"
sample. It is shown that this equilibrium cohesive phase is composed of
long filamentous thread-like micelles, over microns in length. Similar
structures were observed upon solubilization of sonicated liposomes,
multilamellar liposomes, or biological Ca2+ ATPase
membranes. This "gel-like" phase was also visualized during the
process of liposome reconstitution after detergent removal from
lipid-dodecyl maltoside micelles. The rate of detergent removal, controlled through the use of SM2 Bio-Beads, was demonstrated to
drastically influence the morphology of reconstituted liposomes with a
propensity for multilamellar liposome formation upon slow transition
through the "gel-like" phase. Finally, on the basis of these
observations, the mechanisms of dodecyl maltoside-mediated reconstitution of bacteriorhodopsin were analyzed, and optimal conditions for reconstitution were defined.
Structural and functional studies of membrane
proteins have made important advances during the past decade. However,
in many instances, these studies are still limited because of the lack of reproducible methods for the solubilization, reconstitution and
crystallization steps (for reviews, see Helenius and Simons, 1975 Dodecyl maltoside (DOM) is a nonionic detergent with a low critical
micellar concentration (cmc) and is characterized by an intermediate
length of the hydrophobic moiety and a bulky hydrophilic sugar
headgroup. For the last ten years, DOM has gained widespread use in the
solubilization of diverse functionally active membrane proteins (Suarez
et al., 1984 In the present work, the transitional changes induced by the
interaction of DOM on phosphatidyl choline/phosphatidic acid liposomes
were studied by means of light-scattering and cryo-transmission electron microscopy (cryo-TEM). The turbidity data reported made it
possible to accurately define the different steps of the solubilization process and to quantify the mixed bilayer-mixed micelle
interconversion. The results were related to the three-stage model
describing the interaction of detergents with membranes (Lichtenberg,
1985 We have also investigated the mechanisms of DOM-mediated liposome
reconstitution upon detergent removal from lipid-DOM micellar solutions. The optimal conditions for removing this low-cmc detergent by hydrophobic adsorption onto SM2 Bio-Beads were determined. The rate
of detergent removal was shown to critically affect the morphology of
resulting liposomes, demonstrating the role of kinetic factors in
liposome reconstitution. In particular, our data indicated that a slow
transition through the "gel-like phase" led to the formation of
multilamellar liposomes instead of homogeneous unilamellar liposomes
upon rapid transition. Finally, on the basis of these observations, we
have studied the process of DOM-mediated membrane protein
reconstitution according to a general method developed in our
laboratory (Rigaud et al., 1995 Materials
Egg phosphatidyl choline (EPC) and egg phosphatidic acid (EPA)
of the highest purity were purchased from Avanti. The nonionic surfactant n-dodecyl- Methods
Liposome preparation
ABSTRACT
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
INTRODUCTION
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
;
Silvius, 1992; Rigaud et al., 1995; Kühlbrandt, 1992; Dolder et
al., 1996; Garavito et al., 1995). This bottleneck is mainly related to
the amphiphilic character of membrane proteins, which require the use
of detergents as a means of disintegrating the structure of native
membranes in the initial step of their purification. Therefore, a
comprehensive survey of the physicochemical properties and the use of
detergents is still needed.
; Kragh-Hansen et al., 1993; Brandolin et al., 1993;
Pourcher et al., 1995; Knol et al., 1996; Buchanan and Walker, 1996).
In addition, this detergent has been used in reconstitution studies,
and some reports have recently appeared in the literature dealing with
the formation of active proteoliposomes (Groth and Walker, 1996; Knol
et al., 1996) and 2D crystals (Rigaud et al., 1997). Surprisingly,
despite all of these studies that employed DOM, very little information
is available that quantitatively characterizes the solubilization and
reconstitution processes using this detergent (Kragh-Hansen et al.,
1993; De la Maza and Parra, 1997). Thus the main scope of the present
investigation was to investigate the phase behavior of mixed
DOM-phospholipid systems with the aim of providing a basis for
developing rational, reproducible, and efficient reconstitution schemes
that are useful for further functional and structural studies of
membrane proteins.
; Silvius, 1992; Rigaud et al., 1995). Then, using cryo-TEM, we
have directly visualized the structures formed during the
lamellar-to-micellar transition. It is worth noting that our data
reveal an unexpected structural change during the solubilization
process and report a new equilibrium "gel-like" phase specific to
DOM-lipid interaction. It is shown that this cohesive viscous phase is
composed of very long overlapping thread-like micelles, over microns in
length, which has never been reported for other detergents.
), and using bacteriorhodopsin (BR) as a
prototypic membrane protein. The optimal conditions for the use of this
glycosylated detergent in reconstitution experiments are defined and
integrated into the general model proposed for other detergents.
MATERIALS AND METHODS
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
-D-maltoside (DOM) was
obtained from Sigma, and its radioactive derivative,
1-[14C]dodecylmaltoside, was from CEA Saclay (France).
SM2 Bio-Beads (20-50 mesh; BioRad) were extensively washed with
methanol and water before use (Holloway, 1973). Polycarbonate filters
were purchased from Nucleopore Corporation. All other reagents were of
analytical grade.
). Small unilamellar vesicles were prepared by sonicating the large vesicles for 30 min under argon at 4°C. Multilamellar vesicles were
prepared by rehydration of a dry lipid film. The liposomes were
prepared with 9:1 molar mixtures of EPC and EPA as lipids, in a buffer
composed of 100 mM K2SO4 and 10 mM
KH2PO4 (pH 7.2).
Liposome solubilization
Liposomes were adjusted to the desired lipid concentration (from 1 to 10 mM). Solubilization of liposomes was carried out by adding increasing amounts of DOM to aliquoted vesicle suspensions, under constant stirring. After equilibration at 20°C, the turbidities of the different phospholipid-detergent suspensions were measured at 400-500 nm with a Unicam UV2 spectrophotometer.
Cryo-transmission electron microscopy
A few microliters of the lipid-DOM mixtures were applied to a
holey carbon film on a copper grid that was held by tweezers mounted on
a shaft above a liquid ethane bath cooled by liquid nitrogen (Leica EM
CPC). The sample was blotted with filter paper and immediatly plunged
into the liquid ethane. The vitreous specimens were transferred under
liquid nitrogen to the cryo-TEM cold stage (model 626 Gatan), which was
inserted into the Philipps CM120 electron microscope and maintained at
170°C throughout specimen observation. Specimens were imaged at 120 kV by low dose techniques, and micrographs were recorded on Kodak SO163
films with a magnification of 45,000× and a 1-µm defocus.
Membrane protein reconstitution
Purple membrane was isolated from Halobacterium
halobium strain S9, according to the method of Oesterhelt and
Stoeckenius (1974). Monomers of BR in detergents were prepared as
described previously (Seigneuret et al., 1991). BR-containing
proteoliposomes were reconstituted by a step-by-step procedure,
according to the method previously reported (Rigaud et al., 1988,
1995). Liposomes were first treated with different amounts of DOM to
reach the desired stage in the solubilization process. In a second
step, solubilized proteins were added to the equilibrated
detergent-phospholipid mixtures. Finally, the detergent was removed by
direct addition of SM2 Bio-Beads (Levy et al., 1990a,b; Rigaud et al.,
1997).
For BR proteoliposomes reconstituted in a medium containing 100 mM
K2SO4, 10 mM PIPES (pH 7.2), light-induced
transmembrane pH gradients were measured as changes in the fluorescence
intensity of 9-aminoacridine (Cladera et al., 1996). Fluorescence was
monitored with a Perkin-Elmer LS50B fluorimeter, using 400 and 460 nm
for excitation and emission, respectively. Illumination was performed with a 250-W xenon lamp through a flexible glass fiber guide equipped with a low wavelength cutoff at 500 nm and a heat filter.
| |
RESULTS |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
|
|---|
Solubilization of liposomes by dodecyl maltoside
Turbidity measurements
Previous studies demonstrated that changes in optical density of liposome suspensions upon the addition of detergent constituted a convenient technique for a survey of the bilayer solubilization by detergents (Lichtenberg, 1985; Paternostre et al., 1988; Silvius, 1992;
Rigaud et al., 1995). Accordingly, the solubilizing perturbation produced by incremental addition of DOM to EPC/EPA liposomes was monitored by this technique.
However, because detergent was externally added, we have first
determined the time needed for full equilibration. Control kinetic
studies of the interaction of DOM with liposomes indicated that up to
~2-3 h was needed to achieve a constant light-scattering level when
DOM was added to preformed liposomes (data not shown). In the
following, to ensure complete detergent equilibration, the interaction
of DOM with liposomes was studied through the changes in the OD of
detergent-lipid suspensions equilibrated 24 h before measurement
(see also De La Maza and Parra, 1997).
Fig. 1 A shows the
solubilization curves of liposomal suspensions with increasing DOM
concentrations (lipid concentration in the range 2.5-10 mM). The shape
of each curve is similar for all lipid concentrations, which makes it
possible to identify several points that represent a unique assembly of
macromolecular structures. Previous studies indicated that each curve
can be interpreted according to a three-stage model (Lichtenberg, 1985; Rigaud et al., 1995). Stage I involves partitioning of detergent monomers between the aqueous medium and the lipid bilayer, a process recently analyzed in detail by De La Maza and Parra (1997). During this
stage, optical density shows a slight decrease followed by a slight
increase. These changes in OD are difficult to interpret because they
can be related to many parameters, such as changes in the average
diameter, in the number of particles, in viscosity, and/or in
refractive index. At the end of this stage (denoted by white
arrows), liposomes are saturated with detergent. During stage II,
detergent addition promotes gradual liposome solubilization and
lipid-detergent micelles begin to form, inducing a large and significant decrease in OD. Stage III is characterized by complete solubilization of lipid into small mixed micelles, and the suspension becomes optically transparent (denoted by black arrows).
|
![D<SUB><UP>total</UP></SUB>=D<SUB><UP>water</UP></SUB>+R<SUB><UP>eff</UP></SUB> · [<UP>Lipid</UP>]](/content/vol74/issue2/fulltext/918/img001.gif)
|
; Knol et al., 1996), which could be
related to different lipid compositions and/or different modes of
liposome preparations.
It is noteworthy that one of the most striking features not visible in
the turbidity curves was a very drastic change in the consistency of
the suspensions for Reff's around 1.4-1.5
mol/mol (i.e., near the end of the lamellar-to-micellar
transformation). This change, consisting of an increase in the
viscosity of the suspension, was clearly observed for phospholipid
concentrations above 2 mM, and thickening progressed with phospholipid
concentration to give a so-called gel-like sample: at a phospholipid
concentration of 8 mM, when the sample was turned upside down, it
remained stacked at the bottom of the tube. The new optically
transparent phase was stable at room temperature for weeks, appearing
to be isotropic between cross-polarizers. It should also be stressed
that stirring enhanced the gel consistency and that during stirring,
the interface moved inward and climbed at the center: this flow
behavior is reminiscent of the so-called Weissenberg effect
(Weissenberg, 1947) that is observed in semidilute polymer solutions
and which implies the presence of highly entangled objects.
Cryo-transmission electron microscopy
We have correlated the different stages observed by optical density measurements with long-lived structures observed in the vitrified hydrated state using cryo-TEM. This technique, which avoids the artifacts of staining and drying procedures, permits the observation of undistorted samples. As depicted in Fig. 2 a, the EPC/EPA liposomes formed by reverse-phase evaporation, followed by filtration through 0.2-µm Nucleopore filters, appeared by cryo-TEM to be spherical shells with delineated 4-5-nm lipid bilayer walls. No multilamellar vesicles were observed, and liposomes had a monomodal distribution centered around 120 nm (Fig. 3 A), in agreement with previous studies by freeze-fracture electron microscopy (Gulick-Krzywicki et al., 1987).
|
|
-D-glucopyranoside, OG), where long
flexible cylindrical structures, 200-500 nm in length and similar in
thickness to that of a phospholipid bilayer, were observed (Vinson et
al., 1989). Whatever the final interpretation, it should be noted that
the stringlike structures in Fig. 2 d appear, in our case,
as contrasted as liposome bilayers.
Near the end of stage II (Reff 
1.5), the new
transitional "gel-like" state was detected, and vitrified specimens
of these samples were observed by cryo-TEM (Fig. 2 e). The
most striking features are very long filaments that almost completely
fill the field of view. In some images, these filaments appear to
emerge from the few liposomes still present (see, for example, Fig.
4, a and b). They
reach over a few microns in length and overlap. Such extended
structures, never reported for other detergent-lipid mixed systems
(Vinson et al., 1989; Walter et al., 1991; Edwards et al., 1993),
appear similar, however, to the very long threadlike micelles described
by Danino et al. (1997) in highly diluted micellar solutions of
phosphoglucolipids.
|
; Walter
et al., 1991).
It is worth noting that the "gel-like" phase has also been
encountered during the lamellar-to-micellar transition of different preparations of liposomes. Fig. 4, a and b,
reports some representative micrographs, observed in the "gel-like"
phase (Reff 1.5) for multilamellar and
sonicated liposomes, respectively. In both cases, many long
filamentous structures, characteristic of DOM-lipid mixtures, are
visualized.
For comparison, we have also analyzed the process of solubilization by
DOM of biological membranes, namely sarcoplasmic reticulum (SR)
vesicles from skeletal muscle. One important finding was the absence of
the "gel-like" phase reported during the solubilization of
liposomes by DOM. Nevertheless, analysis by cryo-TEM of the structures
formed near the end of the lamellar-to-micellar transition of SR
vesicles indicated the presence of long filamentous structures (Fig. 4
c). However, it has to be stressed that these structures, although rather similar to those observed in the "gel-like" phase of DOM-liposome mixtures, were much less numerous and progressively broke down, decreasing in length, before full solubilization was obtained.
Liposome and proteoliposome reconstitution
Liposome reconstitution
Bilayer formation upon detergent removal from mixed detergent-phospholipid micelles has been demonstrated to be the symmetrical opposite of bilayer solubilization (Levy et al., 1990b).
However, although three steps in the reconstitution process have been
observed at detergent/lipid ratios similar to those observed during
solubilization, there is no information about the intermediate
structures formed during reconstitution, which may be very different in
the two symmetrical processes. In addition, kinetic factors have been shown to be important in determining the size distribution of reconstituted liposomes (Levy et al., 1990a,b, 1992). Because these
aspects are important for further reconstitution experiments, we have
studied in more detail the morphology of liposomes formed after
detergent removal from DOM-lipid micellar solutions.
Because of the very low cmc of DOM, dialysis or dilution as means of
detergent removal is relatively inefficient. We have thus adapted to
this detergent the method previously developed for the study of
liposome formation from mixed micellar solutions containing
Triton X-100 or C12E8. This method relies on
hydrophobic adsorption of detergents onto polystyrene beads. Previous
studies indicated that this batch procedure was well suited for almost complete removal of different detergents and for controlling the rate
of detergent removal by simply controlling the amount of beads
(Levy et al., 1990a,b, 1992; Rigaud et al., 1997).
The standard procedure was the following. Solutions of mixed
lipid-detergent micelles were obtained by adding 8 mM 14C
DOM to a suspension containing 5 mM lipid. After 24 h of
incubation to ensure complete solubilization, SM2 Bio-Beads were added
directly at different bead/detergent ratios, and detergent removal was analyzed as a function of time. Fig. 5
shows time courses of DOM removal after additions of different amounts
of beads. DOM removal was essentially complete in ~30 min in the
presence of an excess of beads (open triangles in Fig. 5) or
extended to many hours by successive additions of small portions of
beads at different time intervals (filled triangles in Fig.
5).
|
|
Proteoliposome reconstitution
We have studied in detail the reconstitution of BR, a prototypic membrane protein which, after incorporation into closed proteoliposomes, is able to generate a light-induced transmembrane pH gradient (Oesterhelt et al., 1992).
The reconstitution procedure was derived from the general method
previously described for different membrane proteins using other
detergents (Rigaud et al., 1995; Cladera et al., 1997). The procedure
was carried out in three steps. Liposomes were first treated with
different amounts of detergent to reach the desired stage in the
solubilization process. After 3 h of incubation, previously
solubilized membrane proteins were added and incorporation was studied
at each step of the lamellar-to-micellar transition. The third step in
the reconstitution is to remove the detergent by using SM2 Bio-Beads.
In view of the results reported above, we have studied the effects of
the rate of detergent removal upon the efficiency of the reconstitution
of BR. Starting from totally solubilized samples, much higher (about
twofold) light-induced pH gradients were measured after fast detergent
removal than after slow detergent removal (traces a and
b in Fig. 7 A).
These data can be explained simply on the basis of the electron
micrographs reported in Fig. 6. Indeed, as for the reconstitution of
pure liposomes, slow detergent removal leads to the formation of
multilamellar BR-containing proteoliposomes (Fig. 6 c),
which are expected to be much less efficient than the homogeneous
unilamellar, 110-120-nm-diameter proteoliposomes produced upon fast
detergent removal (Figs. 6 b and Fig. 3 B).
|
; Pitard et al., 1996). Another piece of
evidence for impermeability relies on the effects of valinomycin on the
kinetics of the light-induced pH gradients across reconstituted
BR-proteoliposomes. As illustrated in Fig. 7 A, the rates of
light-induced internal acidification by BR-proteoliposomes are strongly
accelerated in the presence of this ionophore. Such an acceleration can
be related to the collapse of a light-induced electrical transmembrane
potential, generated in the absence of valinomycin, and reported to
retroinhibit the H+ pumping by BR (Seigneuret and Rigaud,
1986). In the presence of valinomycin, this electrical potential is
overcome by compensatory K+ movements, and a large pH
gradient can develop rapidly because of the high, uninhibited rate of
proton pumping.
| |
DISCUSSION |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
|
|---|
The main goal of this study was a detailed investigation of the phase behavior of mixed phospholipid-detergent systems to increase our understanding of solubilization and reconstitution processes. We have extended the scope of previous investigations on different classes of detergents by including dodecyl maltoside, a glycosylated nonionic detergent with a low critical micellar concentration, which has gained widespread interest in membrane protein biochemistry during the last decade.
Solubilization process
The interaction of dodecyl maltoside with EPC/EPA vesicles was
studied at different lipid/detergent ratios by correlating the
macroscopic changes observed by bulk sample turbidity measurements with
structures observed using cryo-TEM. From turbidity data, the
solubilization process by DOM was described in line with the three-stage model: detergent incorporation, lamellar-to-micellar transition, and total solubilization (Lichtenberg, 1985; Rigaud et al.,
1995). Turbidity measurements as a function of lipid concentration allowed quantitative determination of the critical effective
DOM-to-lipid ratios at which the lamellar-to-micellar transition
started (Rsat = 1 mol/mol) and finished
(Rsol = 1.6 mol/mol). These values are similar
to those reported by De La Maza and Para (1997) and can be compared
with those measured under strictly similar conditions using other
detergents. For example, Triton X-100, C12E8,
octylglucoside, and cholate saturated liposomes at molar ratios of
0.64, 0.7, 1.3, and 0.3, respectively, whereas total solubilization
occurred at respective ratios of 2.5, 2.2, 3, and 0.9 (Paternostre et
al., 1988; Rigaud et al., 1995). Such a comparison indicates that 1) The lamellar-to-micellar transition is much shorter with DOM as compared to other detergents. There is a factor of only 1.6 between Rsol and Rsat for DOM,
compared to a factor of ~3 for other detergents. 2) At
subsolubilizing levels, DOM has a higher capacity for saturating EPC/EPA liposomes than other low-cmc detergents, and up to 1 mol DOM/mol lipid can be incorporated at saturation. 3) At solubilizing levels, DOM has a slightly higher ability to solubilize bilayers than
the other low-cmc nonionic detergents and a much higher (about twofold)
solubilization efficiency than octylglucoside. It is difficult,
however, to rationalize all of these comparative data, which have to
take into account interfering physicochemical properties of detergents
such as shape, size, geometry, amphiphilicity, cmc, packing defects,
and fusogenic ability.
A significant feature of the present data was the slowness (2-3 h) of
DOM equilibration times when mixed with EPC/EPA unilamellar liposomes,
as compared to previously reported equilibration times (a few minutes)
observed for other detergents (Paternostre et al., 1988; Ollivon et
al., 1988; Almog et al., 1990). This observation, made on pure
liposomes, corroborates that made during DOM solubilization of
biological membranes (Kragh-Hansen et al., 1993). A possible interpretation would be to consider the bulky glycosylated headgroup of
DOM, which could hinder the binding of the detergent to the interface
region of the liposomes, and its integration into the hydrocarbon
region by rearrangement of detergent and lipid molecules (see
discussion below on protein insertion). Other factors, such as
viscosity and detergent flip-flop, are probably involved, in addition
to detergent-detergent, lipid-detergent interactions and geometrical
constraints. Whatever the final interpretation, one must include time
among the important variables examined in optimizing the solubilization
conditions by DOM of liposomes and biological membranes.
Intermediate structures in the vesicle-micelle transition
To address directly the microstructures of the aggregates formed during the solubilization process, we have performed cryo-transmission electron microscopic studies. Cryo-TEM has several advantages for the present study. First, cryo-TEM, by analyzing vitrified samples, traps the potentially labile structures associated with intermediates in membrane solubilization. Second, entire hydrated aggregates of liposomes can be imaged, as opposed to the single fracture planes through the specimen that are observed with freeze-fracture electron microscopy. Third, no stains or cryoprotectants are used, reducing the number of artifacts.
Between the apparent upper and lower phase boundaries of the
lamellar-to-micellar transition, different structures were observed: open vesicles, large bilayer sheets, and long threadlike micelles. Although relatively similar structures have already been reported for
other detergent-lipid systems, despite the vast difference in
Reff's at each transition, our data on DOM
relate an unexpected new microstructure, corresponding to a macroscopic
"gel-like" phase. Near the end of the lamellar-to-micellar
transition, the DOM-lipid suspensions became very viscous, forming a
gel-like phase. This new phase is composed of filamentous structures
spanning the field of view. When one compares these threadlike micelles to the cylindrical, wormlike or rodlike micelles seen in other surfactant systems or in phospholipid-detergent mixtures (Vinson et
al., 1989; Walter et al., 1991; Edwards et al., 1993; Silvander et al.,
1996), their unusual feature is their long persistence length, over
microns in length. Such long and numerous entangled structures, which
are observed whatever the mode of preparation of liposomes, may cause
the high viscosity and the gel consistency of the DOM-lipid mixtures.
They can also explain the flow behavior of the phase upon stirring,
namely an increase in the gel consistency and a clear Weissenberg
effect. Solubilization of SR vesicles, i.e., vesicles with a high
protein content, demonstrate that long threadlike micelles are still
present, but the presence of protein and/or the lipid heterogeneity in
SR (see Kragh-Hansen et al., 1993) prevents the large increase in
number and length observed for lipid-DOM mixtures, possibly explaining
the absence of the gel-like consistency of the phase.
It is of interest to relate these observations to those reported in the
studies of the liposome-micelle transition by octylglucoside, another
glycosylated detergent. Near the micellar phase boundary, the
appearance of oil-like droplets, accompanied by a dramatic increase in
the turbidity, was observed in octylglucoside-lipid mixtures
(Paternostre et al., 1988; Ollivon et al., 1988; Almog et al., 1990).
Such turbid mixtures separated into two bulk phases; indeed, after a
few hours of incubation at room temperature, a spontaneous macroscopic
phase separation occurred, resulting in a clear, viscous lower phase
and a turbid upper phase (Ollivon et al., 1988; Almog et al., 1990). On
the basis of lipid and detergent composition, the upper phase was
suggested to be composed of lamellar structures, whereas the structure
of the lower phase, which was clear, viscous, and enriched in lipid and
OG has not been determined. Although there is no a priori reason to
expect OG and DOM interactions with phospholipid to be identical (there
are large differences in the hydrophobic chain length and in the
structure of the polar headgroup, leading to large differences in
cmc's and in the Reff's for the transitions),
it is tempting to correlate the macroscopic gel-like phase observed in
the presence of DOM with the viscous phase observed after phase
separation in the presence of OG. Furthermore, because long threadlike
structures similar to those reported here for DOM have been observed
for phosphoglycolipid micelles (Danino et al, 1997), it can be
proposed that surfactants with a glycosylated polar headgroup have the
tendency to give this very viscous phase the viscosity or gel-like
consistency, depending on the structure of the glycosylated headgroup,
the amphiphilicity of the detergent molecule, and/or the lipid and
protein composition of the micelles.
Liposome formation upon detergent removal
Another important part of this work is related to the study of
liposome reconstitution upon detergent removal from DOM-lipid mixed
micelles. The protocol that was employed, namely adsorption of DOM onto
Bio-Beads, was well suited to investigating the role of kinetics
factors in the vesiculation process. In this context, our cryo-TEM
studies demonstrated that the rate of detergent removal critically
affected the morphology of reconstituted liposomes. Upon slow detergent
removal, a high proportion of multilamellar structures were formed,
whereas upon rapid removal, almost unilamellar liposomes were produced.
Such an observation appears to be specific to DOM-mediated
reconstitutions, because our previous studies with OG, Triton X-100,
C12E8, and different ionic detergents have shown that, when the same strategy of detergent removal was used, whatever the rate of removal, only unilamellar liposomes were produced.
The only significant effect of the rate of detergent removal was a
propensity to form small liposomes upon rapid detergent removal (Levy
et al., 1990a,b, 1992; Cladera et al., 1997).
This specific property of DOM-mediated reconstitutions led us to
tentatively correlate the multilamellar tendency to a slow transition
through the specific gel-like phase. In the light of the models
proposed for vesicle formation by detergent depletion techniques
(Lasic, 1988; Wrigglesworth et al., 1987; Schurtenberger et al., 1984),
it has been proposed that three steps may occur during the overall
process: micellar equilibration (micellar growth), bilayer closure, and
liposome growth (due to residual detergent in the formed liposomes).
Thus a possible interpretation of multilamellar formation upon slow DOM
removal from mixed micelles would be that in the early stage of
detergent removal, mixed micelles would fuse, leading to the formation
of very long threadlike micelles. Upon further detergent removal, these
entangled filamentous micelles would bend and coalesce, leading to
multilamellar bilayer formation.
In this context, we would like also to mention that recent 2D
crystallization trials, using SM2 Bio-Beads to remove DOM, have shown
that the rate of detergent removal drastically influenced the
morphology and shape of melibiose permease crystals (Rigaud et al.,
1997), confirming the importance of kinetic factors in the DOM
micellar-to-lamellar transition.
Membrane protein reconstitution
Using the same strategy of reconstitution with other detergents
and other proteins (Rigaud et al., 1995; Pitard et al., 1996; Cladera
et al., 1996, 1997), we have identified three mechanisms by which
membrane proteins can associate with lipids to give functional proteoliposomes. Depending on the nature of the detergent, proteins can
be either directly incorporated into detergent-saturated liposomes (OG-mediated reconstitutions), transferred from mixed micelles to
detergent-saturated liposomes (Triton X-100-mediated
reconstitutions), or participate in proteoliposome formation during
the micellar-to-lamellar transition (ionic detergents). Our data on
DOM-mediated reconstitution of bacteriorhodopsin are in agreement with
the mechanisms described for OG-mediated reconstitutions of different
classes of membrane proteins, i.e., by direct incorporation of proteins
into detergent-saturated liposomes. In this context, recent studies,
using our step-by-step protocol for reconstitution of the bovine heart
mitochondrial ATP synthase (Groth and Walker, 1996) and of the lactose
transport protein of Streptococcus thermophilus (Knol et
al., 1996), demonstrated that the highest transport activities were
obtained when the liposomes were titrated with saturating amounts of
DOM. Although not analyzed in this paper, an ultimate consequence of
this mechanism of direct protein incorporation into preformed liposomes
is related to the final orientation of the protein: indeed, upon direct
incorporation, protein orientation has been shown to be asymmetrical,
leading to much more efficient biological activities than
reconstitution from micellar solutions in which protein orientation is
generally more random (Rigaud et al., 1995).
Common mechanisms described for OG- and DOM-mediated reconstitution may
be related to the only common property of these two detergents, i.e., a
glycosylated polar headgroup. In this connection, recent NMR studies on
the interactions of OG with lipid bilayers suggested an alignment of
detergent and lipid molecules where the sugar moiety of the detergent
penetrated approximately up to the level of the glycerol backbone, with
the headgroup region tightly packed and the conformation of the
P-N+ dipole of the lipid almost unchanged (Wenk et al.,
1997). The large cross-sectional area of the OG headgroup, as well as
the short C8 chain of the detergent produced packing
defects in the central part of the membrane. Such specific defects
induced by glycosylated detergents on lipid bilayers may be clues in
driving the direct incorporation of membrane proteins into
detergent-saturated liposomes.
As a last remark, this work establishes the potential of SM2 Bio-Beads for removing DOM. Besides providing a convenient way to control the rate of detergent removal and thus the homogeneity and unilamellarity of liposome or proteoliposome preparations, another advantage of Bio-Beads is to allow almost complete detergent removal, leading to relatively impermeant proteoliposomes. Dialysis would have been as efficient in removing all of the detergent, but would have required a very long time for total removal, which might be drastic when dealing with unstable membrane proteins. Furthermore, related to our results on the effects of detergent removal rate, dialysis would result in the formation of multilamellar structures. Thus the most important benefit in using Bio-Beads is to produce unilamellar liposomes and proteoliposomes with a low ionic permeability, which is obviously crucial to the study of transport membrane proteins.
| |
CONCLUSION |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
|
|---|
The solubilization of biological membranes by detergents as well
as the formation of proteoliposomes from lipid-protein-detergent mixtures have found wide application in membrane research. Examples are
the preparation of lipid vesicles of defined size and composition, the
reconstitution of membrane proteins in functional forms, and more
recently, 2D crystallization of membrane proteins. However, it is
clear that a basic understanding of the lipid-detergent interactions is an indispensable prerequisite for efficiently seeking the optimal conditions for solubilization and
reconstitution of a given membrane protein. The present paper on the
use of dodecylmaltoside is a piece of this wide systematic work needed
to optimize the incorporation of membrane proteins for which the use of
this detergent is convenient. Studies on 2D crystallization mediated by
this detergent are currently in progress in our laboratory (Rigaud et
al., 1997).
| |
ACKNOWLEDGMENTS |
|---|
We are grateful to J. J. Lacapère and G. Mosser for helpful discussions and experimental help during the course of this work. We also thank C. Safinya (Santa Barbara) for stimulating discussions at the beginning of this work.
Part of this work was supported by a grant from the EEC (PL962119).
| |
FOOTNOTES |
|---|
Received for publication 7 August 1997 and in final form 10 November 1997.
Address reprint requests to Dr. Jean-Louis Rigaud, Institut Curie, Section de Recherche, UMR-CNRS168 and LCR-CEA 8, 11 rue Pierre et Marie Curie, 75231 Paris Cedex, France. Tel.: 33-1-42-34-6781; Fax: 33-1-40-51-0636; E-mail: rigaud{at}curie.fr.
| |
REFERENCES |
|---|
|
Top
Abstract
Introduction
Materials & Methods
Results
Discussion
Conclusion
References
|
|---|
Biophys J, February 1998, p. 918-930, Vol. 74, No. 2
© 1998 by the Biophysical Society 0006-3495/98/02/918/13 $2.00
This article has been cited by other articles:
 |
|
 |
|
  S. S. Funari, B. Nuscher, G. Rapp, and K. Beyer Detergent-phospholipid mixed micelles with a crystalline phospholipid core PNAS, July 31, 2001; 98(16): 8938 - 8943. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Garavito and S. Ferguson-Miller Detergents as Tools in Membrane Biochemistry J. Biol. Chem., August 24, 2001; 276(35): 32403 - 32406. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||
), 5 (
), 7.5 (
), and 10 mM (
) phospholipid and treated with
different amounts of DOM. Optical densities were measured at 450 nm
after 24 h of equilibration at 20°C. White and black arrows
indicate onset and total solubilization, respectively.
(Inset) Percentage changes in turbidity of multilamellar
(
), sonicated (
), and reverse-phase (
) liposomes upon DOM
addition. Total lipid concentration in all samples: 5 mM (i.e., 4 mg/ml). (B) Lipid-DOM phase diagram. Total DOM
concentrations corresponding to the onset (
, Absorbance of
the lipid-BR-DOM mixtures before detergent removal. White and black
arrows correspond to onset and total solubilization, respectively.