See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/10622745

Advanced Flicker Spectroscopy of Fluid Membranes

Article in Physical Review Letters · August 2003

DOI: 10.1103/PhysRevLett.91.048301 · Source: PubMed

CITATIONS READS
76 176

6 authors, including:

Hans-Günther Döbereiner Gerhard Gompper

University of Bremen Forschungszentrum Jülich
107 PUBLICATIONS 5,497 CITATIONS 596 PUBLICATIONS 26,998 CITATIONS

SEE PROFILE SEE PROFILE

Daniel M Kroll Karin A Riske

North Dakota State University Federal University of São Paulo
142 PUBLICATIONS 7,230 CITATIONS 124 PUBLICATIONS 4,390 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Gerhard Gompper on 19 May 2014.

The user has requested enhancement of the downloaded file.

 PHYSICA L R EVIEW LET T ERS week ending
VOLUME 91, N UMBER 4 25 JULY 2003

Advanced Flicker Spectroscopy of Fluid Membranes

Hans-Günther Döbereiner,1,* Gerhard Gompper,2 Christopher K. Haluska,1 Daniel M. Kroll,3
Peter G. Petrov,1,† and Karin A. Riske1
1
Max-Planck-Institut für Kolloid- und Grenzflächenforschung, Am Mühlenberg 1, 14476 Golm, Germany
2
Institut für Festkörperforschung, Forschungszentrum Jülich, 52425 Jülich, Germany
3
Department of Medicinal Chemistry and Minnesota Supercomputer Institute, University of Minnesota,
308 Harvard Street SE, Minneapolis, Minnesota 55455, USA
(Received 26 June 2002; published 23 July 2003)
The bending elasticity of a fluid membrane is characterized by its modulus and spontaneous
curvature. We present a new method, advanced flicker spectroscopy of giant nonspherical vesicles,
which makes it possible to simultaneously measure both parameters for the first time. Our analysis is
based on the generation of a large set of reference data from Monte Carlo simulations of randomly
triangulated surfaces. As an example of the potential of the procedure, we monitor thermal trajectories
of vesicle shapes and discuss the elastic response of zwitterionic membranes to transmembrane pH
gradients. Our technique makes it possible to easily characterize membrane curvature as a function of
environmental conditions.

DOI: 10.1103/PhysRevLett.91.048301 PACS numbers: 82.70.Uv, 05.40.–a, 87.16.Dg

The bending elasticity of amphiphilic interfaces is a This is somewhat surprising since spontaneous curvature
fundamental concept introduced more than 30 years ago plays the key role in determining the morphology of
by Helfrich [1] to describe the morphology of biomem- biomembranes [2], lipid vesicles [3], and polymersomes
branes. Subsequently, it has been extensively applied to [11]. Spontaneous curvature is crucial for maintaining the
model the elastic behavior of soft interfaces [2 – 4]. The spatial organization of, and traffic between, cellular or-
principal parameters characterizing the elastic energy of ganelles and the plasma membrane; e.g., there is growing
an amphiphilic interface are the bending modulus
, evidence for a coupling between cell signaling and endo-
which sets the energy scale, and the effective spontaneous cytosis [12]. Further, the functional state of certain in-
curvature C
0 , which describes the preferred curvature of tegral membrane proteins [13,14] and membrane fusion
the interface. In this Letter we describe a new procedure competence [15] is thought to be controlled by monolayer
which makes it possible, for the first time, to simultane- spontaneous curvature. More generally, control of inter-
ously determine both the bending modulus and the spon- facial curvature is required to tune the structure of mate-
taneous curvature of a membrane in a single experiment. rials on the nanoscale [16]. It is therefore important to
There exist several techniques for measuring the bend- develop a simple, straightforward procedure for the di-
ing modulus [5–7]. The method most widely used for rect determination of the spontaneous curvature and
membranes of higher bending rigidity is the flicker spec- bending modulus. The technique we have developed —
troscopy of giant, quasispherical vesicles [5,7]. Although flicker spectroscopy of nonspherical vesicles — avoids the
this technique can be used to determine precise values of shortcomings of traditional spectroscopy techniques by

, it has the disadvantage that analytical results required utilizing extensive Monte Carlo simulations of dynami-
for extracting
from experimental data are only available cally triangulated vesicles to generate data for a wide
in the spherical limit. In this limit, the membrane is under range of reduced volumes and spontaneous curvatures,
a lateral tension which dominates the long-wavelength which can then be used to extract the elastic parameters of
part of the spectrum. To determine
it is therefore the membrane from flicker spectroscopy data (see Fig. 1).
necessary to measure the spectrum up to high mode
number, which in turn requires very large vesicles. In
addition, since the volume-to-area ratio changes with
temperature, a vesicle which is quasispherical at one
temperature is not quasispherical at another, so that the
bending modulus of a single vesicle cannot be determined
as a function of temperature. Finally, membrane fluctua-
tions are insensitive to the effective spontaneous curva-
ture in this limit, so that C
0 cannot be determined. FIG. 1. Experimental and theoretical vesicle shapes.
Although values for the spontaneous curvature of vari- (A) Phase contrast micrograph (v
0:828). The scale bar
ous lipid molecules forming an inverted hexagonal phase corresponds to 5 m, (B) simulation snapshot. The parameters
[8] have been inferred, there is little experimental litera- in the simulations are v
0:825,
=kB T
25,
0:90, c
0

ture dealing with bilayer spontaneous curvature [9,10]. 0:28, and g
0:37.

048301-1 0031-9007=03=91(4)=048301(4)$20.00  2003 The American Physical Society 048301-1

 PHYSICA L R EVIEW LET T ERS week ending
VOLUME 91, N UMBER 4 25 JULY 2003

The fluctuation spectra contain information on both the spherical harmonics. The Fourier expansion of the two-
bending modulus and the spontaneous curvature. Using dimensional contour in a plane through the center of
this technique it is therefore possible to determine both mass, parallel to the substrate, is easily obtained from
elastic constants simultaneously from one experiment this fit. Most of our data are obtained for triangulated
and to study their dependence on environmental condi- surfaces of 407 vertices; the shapes are fitted with 196
tions, such as temperature and solvent pH. spherical harmonics, which implies some smoothing of
Experimentally, thermal shape fluctuations of giant the simulated shapes on the scale of the lattice constant.
fluid vesicles are well characterized using light micros- The simulated vesicles are therefore analyzed in the same
copy [5,10,17]. In these experiments, fluctuating prolate way as real vesicles in experiment. We characterize the
vesicles are stabilized by gravity — due to a small density mean vesicle shape by the first two amplitudes ha2 i and
difference of the solvent inside and outside the vesicle — ha4 i, which is known to be a very good approximation
on the bottom of a temperature-controlled microcham- [17]. With increasing c
0 , a transition from an oblate to a
ber. The focal plane of a phase contrast microscope is prolate shape is observed, which leads to a pronounced
adjusted to include the long axis of the vesicle, and shape increase of ha2 i and ha4 i. The oblate-to-prolate transition
contours are obtained by real time video image analysis; is reflected in a sharp peak of the fluctuations in a2 ; see
for details see Ref. [17]. The raw data therefore consist of a Fig. 2. Similarly, the budding transition of the prolate
time sequence of closed 2D loops. Choosing a coordinate shapes, which occurs at higher spontaneous curvature, is
system in which the x coordinate lies along the long axis signaled by a strong increase of the mean-square ampli-
of the vesicle, the contours are then represented
P in po- tude ha23 i as the budding spinodal is approached (data
lar
P coordinates r; ’ as r’
r0 1  a
n n cosn’  not shown).
n b n sinn’, where the angle ’ is measured from the The fitting of the experimental spectra to the
positive x axis. The time-dependent amplitudes fan ; bn g Monte Carlo data sets proceeds as follows. We employ
encode the full experimental information. The mean val- the average amplitudes ha2 i, ha4 i, and the mean-square
ues han i describe the mean vesicle shape; for the oriented fluctuation amplitudes ha22 i, ha23 i, ha24 i, and ha25 i.
contours used here, hbn i
0. The mean-square ampli- The Monte Carlo data of all these quantities are first
tudes ha2n i han  han i2 i measure the thermal fluc- interpolated linearly as a function of c
0 , then again
tuations of the vesicles about their mean shape. linearly as a function of the reduced volume v. A least-
Our Monte Carlo simulations are based on a model of square fit to the experimental data then determines
, c
0 ,
dynamically triangulated surfaces [18]. The bilayer mem- and v simultaneously for a given vesicle.
brane is modeled by a single triangulated network of The analysis of 14 prolate stearoyl-oleoyl-
spherical topology; see Fig. 1(B). In contrast to previous phosphatidylcholine (SOPC) vesicles in sucrose/glucose
simulations [19], we use the full Hamiltonian of the area- solution (100 mM) yielded results for the bending modu-
difference-elasticity model [20] and include the contribu- lus which clustered around two values,

35  3kB T
tion of a gravitational force acting on the vesicle interior and

75  7kB T, corresponding to single and double
[21]. The discretization of the mean curvature is described
in Ref. [18]. The ratio of volume V and area A of the
vesicles is characterized by the reduced volume v

RV =RA 3 , where RA
A=41=2 and RV
3V=41=3 0.006
are effective vesicle radii. The influence of gravity is v=0.82
measured via the dimensionless parameter g
<∆a22>
g0 R4A =
, where g0 is the local acceleration and  is
0.004
the excess mass density of the interior vesicle solution.
The membrane asymmetry is measured by the dimen-
sionless parameter c
0
C
0 RA . The effective spontaneous v=0.91
curvature C
0 has two contributions [17]: the first is due to 0.002
solution and/or bilayer asymmetry, the second — which
v=0.96
is weighted by the ratio of the membrane stretching
modulus to the bending rigidity ( ’ 1:4 for phosphocho- 0
line bilayers [20]) — is due to the area difference in the -8 -6 -4 -2 0 2 4 6 8
c0
inner and outer monolayers. These two contributions can-
not be independently determined from a single vesicle FIG. 2. Simulated mean-square amplitude ha22 i of shape
fluctuation spectrum [17]. fluctuations as a function of the effective spontaneous cur-
The Monte Carlo simulations follow the procedure vature c
0 . Note the peak at the prolate-to-oblate transition.
described previously [18]. In order to extract the experi- Three different values of the reduced volume are shown, as
mentally relevant information, we first fit the simulated indicated. The other parameters are
=kB T
25,
0:90,
three-dimensional shape data using an expansion in and g
0:37.

048301-2 048301-2
 PHYSICA L R EVIEW LET T ERS week ending
VOLUME 91, N UMBER 4 25 JULY 2003

bilayer membranes, respectively. The measured bending therefore present only on the outside of the vesicle.
modulus is in good agreement with the value

32kB T Illuminating the sample leads to an increase of pH in
obtained from fluctuation spectroscopy of quasispherical the external solution due to photoaquation of the iron
vesicles [7]. It is important to point out that we can complex, as well as a subsequent curvature change of
analyze vesicles as small as 3–4 m, which is much the vesicle membrane [24]. The magnitude of the pH
smaller than those used in Ref. [7]. change depends on the illumination intensity. We have
In order to test our ability to determine the sponta- calibrated the pH change in the external vesicle solution
neous curvature, we explored thermal shape trajectories using the fluorescent probe 4-methylumbelliferone.
of SOPC vesicles in parameter space v; c
0 T. A suit- Figure 4 contains normalized histograms of the ellip-
able vesicle was selected and driven via temperature tical mode a2 of a typical vesicle as a function of external
across two shape transitions. We followed the vesicle pH. At low pH, the vesicle is in a bistable state with a
from an oblate shape at low temperatures (T < 23:0  C), predominantly oblate shape. It shows pronounced prolate-
through an elongated prolate form at higher temperatures, oblate fluctuations due to the proximity of the weak first-
to the budding transition (T
44:0  C). We obtain

order prolate-oblate transition. At elevated pH values, the
35  5kB T for the mean bending modulus. The reduced prolate shape is stable. Note the strong fluctuations even at
volume v and spontaneous curvature c
0 both vary ap- this pH. Two effects are visible. First, the vesicle contour
proximately linearly with temperature. In the prolate becomes more elliptical with increasing pH. Second, the
phase, we obtain a thermal expansion coefficient A
fluctuations in the elliptical mode decrease. As we have
1:9  0:1  103 =K, consistent with the results of seen above (compare Fig. 2), the latter indicates that the
Ref. [22]. In Fig. 3, the mapped c
0 ; v trajectory is shown prolate vesicle is moving away from the prolate-oblate
together with the two (T
0) spinodal lines of the pro- transition; i.e., the spontaneous curvature increases. The
late phase. Linear extrapolation to obtain the crossing apparent elongation of the vesicle is due to competition
points with the two spinodals gives vsbud
0:777 and between bending elastic energy and gravity. The vesicle
vspro=obl
0:830. Comparison with the reduced volumes, becomes progressively less flattened by gravity with in-
where budding (vbud
0:775) and the prolate-oblate tran- creasing spontaneous curvature. This tends to drive the
sition (vpro=obl
0:834) actually occurred, shows a nice vesicle towards the budding transition, where a small
agreement. These results clearly demonstrate that we are satellite is expelled along the polar axis. Indeed, at suffi-
able to extract precise values for the elastic parameters of ciently high pH we observed, in general, budding of the
the membrane using our technique. vesicle, signaled by an increase in pearlike vesicle fluc-
We have also employed the method to measure electro- tuations near the budding instability. For the vesicle
statically induced spontaneous curvature. It has been shown in Fig. 4, we found ha23 i
0:0204 (pH 7.7),
suggested that a change in pH induces membrane curva- 0.0267 (pH 8.1), and 0.0288 (pH 8.5).
ture via association of hydroxyl ions with the trimethyl- The results of the fitting procedure described above are
ammonium group of the phosphatidylcholine molecule shown in Fig. 5. As already anticipated from the raw data,
[23]. In order to quantify such an effect, we have swollen
giant SOPC vesicles in raffinose solution at neutral pH
and incubated them in an equiosmolar solution of glu- p(a 2)
cose and 50 M potassium hexacyanoferrate, which is 8.5
30
_
5
c0
4 20 8.1
3
7.7
2
10
1

−1 0
0.00 0.05 0.10 0.15 0.20 0.25
0.76 0.78 0.80 0.82 0.84 0.86 a2
v
R
FIG. 3. Thermal trajectory of a prolate vesicle in c
0 ; v FIG. 4. Normalized histograms [ pada
1] of the ellip-
parameter space with a radius RA
4:8 m and g
0:21. tical shape fluctuations of a prolate SOPC vesicle (RA

The experimental points correspond to temperatures T
5:2 m) with varying external pH of 7.7 (open circles), 8.1
25:0  C, 29:4  C, 34:2  C, and 38:8  C. The crossing points of (open squares), and 8.5 (open triangles) at constant internal
a linear fit to the trajectory with the upper (budding) and lower pH 7.7. Gaussian fits are shown. With increasing pH gradient
spinodal (prolate-oblate transition) are indicated. For the bend- the vesicle develops a more elongated shape (larger a2 ) with
ing modulus we find

35  5kB T. decreasing fluctuations.

048301-3 048301-3
 PHYSICA L R EVIEW LET T ERS week ending
VOLUME 91, N UMBER 4 25 JULY 2003
_
c0 (0.950,31.9) nors of The Petroleum Research Fund, administered by
5
the ACS.
(0.951,31.4)
4

3
*Present address: Department of Biological Sciences,
2 Columbia University, New York, NY 10027.
(0.956,31.2) †
Permanent address: School of Physics, University of
1 Exeter, EX4 4QL, United Kingdom.
[1] W. Helfrich, Z. Naturforsch. 28c, 693 (1973).
0 v = 0.953 ± 0.003 [2] Handbook of Biological Physics, edited by R. Lipowsky
(0.953,33.3) κ = (32 ± 1) kT and E. Sackmann (Elsevier, Amsterdam, 1995), Vol. I.
−1 [3] U. Seifert, Adv. Phys. 46, 13 (1997).
[4] S. A. Safran, Adv. Phys. 48, 395 (1999).
7.4 7.6 7.8 8.0 8.2 8.4 8.6
[5] H. P. Duwe, J. Käs, and E. Sackmann, J. Phys. I (France)
pH 51, 945 (1990).
[6] E. Evans and W. Rawicz, Phys. Rev. Lett. 64, 2094
FIG. 5. Spontaneous curvature c
0 as a function of external pH (1990).
for g
0:8. Note that the reduced volume and bending modu- [7] P. Méléard, C. Gerbeaud, T. Pott, L. Fernandez-Puente,
lus, which are given in brackets, v;
, remain constant. v, c0 , I. Bivas, M. D. Mitov, J. Dufourcq, and P. Bothorel,
and
are obtained simultaneously via comparison of the data Biophys. J. 72, 2616 (1997).
shown in Fig. 4 to Monte Carlo simulations. [8] N. Fuller and R. P. Rand, Biophys. J. 81, 243 (2001), and
references therein.
[9] W. Harbich, H. J. Deuling, and W. Helfrich, J. Phys.
we find a strong change in spontaneous curvature at a (Paris) 38, 727 (1977).
constant bending modulus

32  1kB T. Indeed, the [10] H.-G. Döbereiner, O. Selchow, and R. Lipowsky, Eur.
bending modulus of SOPC (see the measurement above) Biophys. J. 28, 174 (1999).
should not change considerably since electrostatic contri- [11] B. M. Discher, Y.-Y. Won, D. S. Ege, J. C-M. Lee, F. S.
Bates, D. E. Discher, and D. A. Hammer, Science 284,
butions to the elastic modulus are only on the order of
1143 (1999).
1kB T [25]. Note that the reduced volume of the vesicle, [12] A. Petrelli, G. F. Gilestro, S. Lanzardo, P. M. Comoglio,
v
0:953  0:003, is also found to be constant, as it N. Migone, and S. Giordano, Nature (London) 416, 187
should be. This is a nontrivial result of the fit consider- (2002).
ing the large changes in apparent area of the vesicle with [13] S. L. Keller, S. M. Bezrukov, S. M. Gruner, M.W. Tate,
pH. The large increase in the spontaneous curvature can I. Vodyanoy, and V. A. Parsegian, Biophys. J. 65, 23
be understood by considering the balance of the electro- (1993).
static free energy and the intrinsic bending energy of the [14] G. S. Attard, R. H. Templer, W. S. Smith, A. N. Hunt, and
membrane. S. Jackowski, Proc. Natl. Acad. Sci. U.S.A. 97, 9032
In summary, we have developed a general technique (2000).
which makes it possible to determine simultaneously [15] B. R. Lentz, D. P. Siegel, and V. Malinin, Biophys. J. 82,
555 (2002).
both the bending modulus and the spontaneous curvature
[16] R. Strey, Colloid Polym. Sci. 272, 1005 (1994).
of a fluid membrane. The general idea is to use a single [17] H.-G. Döbereiner, E. Evans, M. Kraus, U. Seifert, and
nonspherical vesicle as a morphological probe of the M. Wortis, Phys. Rev. E 55, 4458 (1997).
(elastic) interactions of its membrane with the surround- [18] G. Gompper and D. M. Kroll, J. Phys. Condens. Matter 9,
ing fluid. A large number of phenomena can be monitored 8795 (1997).
by analysis of the vesicle fluctuation spectrum. In addi- [19] G. Gompper and D. M. Kroll, Phys. Rev. Lett. 73, 2139
tion to characterizing the static curvature response to en- (1994).
vironmental conditions, the dynamics of (bio-)chemical [20] L. Miao, U. Seifert, M. Wortis, and H.-G. Döbereiner,
reactions at interfaces can be monitored. This has impor- Phys. Rev. E 49, 5389 (1994).
tant consequences for future studies dealing with biologi- [21] M. Kraus, U. Seifert, and R. Lipowsky, Europhys. Lett.
cal or material aspects of interfacial elasticity. 32, 431 (1995).
[22] D. Needham, T. J. McIntosh, and E. Evans, Biochemistry
H. G. D. wants to thank U. Seifert and W. Fenzl for
27, 4668 (1988).
enjoyable and helpful discussions. H. G. D. is grateful [23] J. B. Lee, P. G. Petrov, and H.-G. Döbereiner, Langmuir
to R. Lipowsky for the opportunity to do this work at 15, 8543 (1999).
his department and thanks him and the Deutsche [24] P. G. Petrov, J. Lee, and H.-G. Döbereiner, Europhys. Lett.
Forschungsgemeinschaft for financial support. D. M. K. 48, 435 (1999).
acknowledges support from the National Science [25] T. Chou, M.V. Jaric, and E. D. Siggia, Biophys. J. 72,
Foundation under Grant No. DMR-0083219, and the do- 2042 (1997).

048301-4 048301-4

View publication stats