TICB-382; No of Pages 9
Review
TRENDS in Cell Biology
Vol.xxx No.x
Special issue: Membrane Dynamics
Biophysical properties of lipids and
dynamic membranes
P.A. Janmey1 and P.K.J. Kinnunen2
1
Institute for Medicine and Engineering, Departments of Physiology, Physics, Bioengineering, University of Pennsylvania,
Philadelphia, PA 19104, USA
2
Helsinki Biophysics & Biomembrane Group, Department of Medical Chemistry, Institute of Biomedicine, University of Helsinki,
FIN-00014 Helsinki, Finland
The lipid bilayer is a 3D assembly with a rich variety of
physical features that modulate cell signaling and protein function. Lateral and transverse forces within the
membrane are significant and change rapidly as the
membrane is bent or stretched and as new constituents
are added, removed or chemically modified. Recent studies have revealed how differences in structure between
the two leaflets of the bilayer and between different
areas of the bilayer can interact together with membrane
deformation to alter the activities of transmembrane
channels and peripheral membrane binding proteins.
Here, we highlight some recent reports that the physical
properties of the membrane can help control the
function of transmembrane proteins and the motordependent elongation of internal organelles, such as
the endoplasmic reticulum.
Introduction
The lipid bilayer of a cell membrane might seem to be a
passive film that blocks flow of water and solutes and in
which the truly regulatory elements – proteins – are
inserted. But the variety of lipids and their controlled spatial
organization, which define the biophysical properties of the
membrane, have an active role in cell function. For example,
the length and degree of saturation of the lipid acyl chains
determine the thickness and ordering of the hydrophobic
region of the membrane; the electrostatic charge of anionic
lipids mediates interactions with cationic regions of membrane-associated proteins; and specific interactions with
selected lipids, such as polyphosphoinositides, are important for spatial organization of their protein ligands.
The chemical compositions of the two leaflets of the lipid
bilayer are complex and very different from each other. For
example, nearly all anionic lipids in eukaryotic cells face
the cytoplasm, whereas most lipids with large glycosylated
headgroups are exposed to the extracellular environment.
Differences in bilayer asymmetry between eukaryotic and
prokaryotic membranes are essential for the activity of
endogenous antimicrobial factors that rupture bacterial
membranes but are harmless to eukaryotic cells [1]. The
chemical composition of the bilayer affects its mechanical
Corresponding authors: Janmey, P.A. (janmey@mail.med.upenn.edu);
Kinnunen, P.K.J. (paavo.kinnunen@helsinki.fi).
Available online xxxxxx.
www.sciencedirect.com
properties and, conversely, application of forces to the
membrane can alter its chemical composition. Some essential aspects of bilayer structure are summarized in Box 1.
It has long been recognized that mechanical forces can
have physiologically relevant effects on cells [2,3], and
recent studies begin to suggest how the lipid bilayer acts
in concert with transmembrane and peripherally bound
proteins to detect and respond to forces. A brief summary of
the mechanical properties of lipid bilayers is shown in
Boxes 2,3.
Here, we highlight a few areas of cell biology in which
the physical properties of membranes, and not only chemically specific lipid–protein interactions, have recently
been identified as essential for proper cell function and
intracellular signaling. These examples include the role of
membrane tension in gating of mechanically sensitive
channels, stress-activation of enzyme activity, and how
the biophysical properties of the membrane can effect
membrane bending and stretching.
Transverse lipid asymmetry
The bilayer in a typical eukaryotic cell has a thickness of
5 nm and a continuous surface area of hundreds of square
microns, containing hundreds of different lipid types and
>108 individual molecules. The types of lipids in the inner
and outer leaflets are very different (see Box 1). The transbilayer asymmetry is produced primarily in the transGolgi network (TGN), but also at the plasma membrane
by several proteins that require ATP hydrolysis [4–6]. ATP
binding cassette (ABC) transporters seem to move (flop)
phospholipids from the inner to the outer leaflet. Aminophospholipid translocases that control movement and
retrieval (flipping) of lipids to the inner leaflet include
P-type ATPases. Several candidate flippases have been
identified in yeast, with various specificities for different
lipids [7]. One such ATPase, Drs2p, a transmembrane
protein implicated in protein transport from the TGN,
localizes to the TGN and specifically translocates phosphatidylserine (PS) but not other lipids. A third class of
proteins, scramblases, dissipate the transbilayer asymmetry [8]. These proteins, which are even less definitively
characterized than are aminophospholipid translocases,
are generally found to be activated by increased intracellular Ca2+ levels but do not require ATP for activity.
0962-8924/$ – see front matter ß 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tcb.2006.08.009
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
2
TRENDS in Cell Biology Vol.xxx No.x
Box 1. Lipid asymmetry in eukaryotic cell membranes
Hundreds of different lipid species are present in the plasma
membrane [5,6,37]. Many of the rare but important signaling
phospholipids, such as polyphosphoinositides, seem to be exclusively generated or delivered to the cytoplasmic face (inner leaflet) of
the plasma membrane and to specific classes of internal membranes.
By contrast, cholesterol accounts for a large fraction of both the inner
and outer leaflets but seems to be more abundant in the outer leaflet
(Figure I). This asymmetry is not strictly conserved, and different cell
types, organelles and cells at different states of activity are likely to
change the lipid distribution.
Within the lipid bilayer, disorder is introduced by differences in
chain length and saturation of the hydrophobic chains in the
membrane interior and the lateral distribution of different lipids
within each leaflet, which can alter the biophysical properties of
the membrane. Whether lipids within each leaflet are randomly
distributed or organized into domains is a crucial, unresolved and
contentious issue in membrane biochemistry, with many implications
for cell signaling. Proteins also constitute 50% of the cross-sectional
area of the membrane, and peripheral proteins interact with both
extracellularly and cytoplasmically directed lipids.
One consequence of trans-bilayer asymmetry is sequestration of
acidic phospholipids away from the external face of the membrane,
a feature that distinguishes eukaryotic lipid outer membranes from
prokaryotic ones, which are highly anionic. Other consequences
are likely to arise from differences in membrane bending and
stretching moduli (see Box 2) of different lipid compositions and
from differences in permeability to water and other small
molecules.
Figure I. The lipids found in the plasma membrane. Abbreviations: PI, phosphatidylinositol; PIP, phosphatidylinositol phosphate; PIP2, PtdIns(4,5)P2;
PIP3, phosphatidylinositol (3,4,5)-trisphosphate.
Loss of trans-bilayer asymmetry to expose PS on the
outer surface is often a sign of injury and leads to activation
of blood coagulation or recognition by phagocytes of cells
undergoing apoptosis. Intriguingly, cancer cells and vascular endothelial cells in tumors also expose PS, causing
increased coagulation and thrombosis in cancer patients
[8,9]. Acidic phospholipids on the outer membrane can also
provide an environment that is sufficiently different from
surfaces composed of lipids with no net charge and cholesterol to trigger the formation of amyloid-type fibers by
several proapoptotic, cytotoxic and antimicrobial proteins
and peptides [10–12].
Whereas creating and maintaining trans-bilayer asymmetry requires ATP and essentially depends on specific
proteins, the mechanism of scrambling the lipids is less
clear and might in some contexts occur by physical rather
than specific biochemical actions in the membrane. For
example, whereas the first identified scramblase proteins
require intracellular calcium increases for activity, some
scrambling activities might function in concert with
cholesterol, and other scrambling mechanisms can be
triggered by factors that promote lateral sequestration
of inner leaflet polyphosphoinositides. For example
polyamines [13] and a phosphoinositide-specific peptide
based on the phosphatidylinositol (4,5)-bisphosphate
(PtdIns(4,5)P2)-regulatory site of gelsolin [14] both
strongly promote exposure of PS at the outer leaflet without increasing cytosolic Ca2+ levels. A purely physical
mechanism for lipid scrambling has recently been proposed [15], based on changes in lymphocyte PS exposure
during changes in cell volume leading to changes in membrane lipid packing. In this mechanism, shown in Figure 1,
imposition of membrane curvature dilates one leaflet while
compressing the other depending on whether the curvature is concave or convex. Compression of inner leaflets
rich in PS apposed to dilated outer leaflets lowers the
activation energy for spontaneous translocation (flopping)
of PS and phosphatidylethanolamine (PE) to the outer
leaflet, and compensatory flipping transitions of PC occur
at regions of opposite curvature. This example illustrates
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
TRENDS in Cell Biology
Box 2. Forces controlling membrane shape
Membrane tension
The cell membrane tends to maintain a specific lipid packing density
and therefore an optimal surface pressure on the order of 30 mN m 1.
Increasing the lipid spacing by osmotic swelling, for example, is
strongly resisted, and leads to rupture when the membrane is strained
slightly above its optimal packing. Compression within the plane of
the membrane would also be resisted, but the membrane buckles out
of plane before significant compression occurs.
Spontaneous curvature and bending stiffness
The default shape for most membrane constituents is not flat.
Instead, each lipid shape that deviates from a cylinder contributes a
spontaneous curvature to the membrane.
Molecules that have an overall inverted conical shape, such as
detergent molecules, lysophospholipids and polyphosphoinositides, form structures with a positive curvature, such as micelles
(Figure Ia). Cylindrical-shaped lipid molecules, such as phosphatidylcholine and sphingomyelin, preferentially form flat bilayer
structures (Figure Ib). Lipid molecules that have an overall conical
shape, such as diacylglycerol and PE, with a small hydrophilic crosssection, form structures with a negative curvature, such as the
inverted hexagonal phase of tubes with headgroups inside and
hydrophobic tails outside (Figure Ic). The local shape of a membrane
depends on which lipids are present and on how they are spatially
distributed. Insertion or removal of lipids into the inner or outer
leaflet leads to area mismatches that also alter curvature.
Membranes resist bending because changing local curvature
alters both the headgroup spacing and the entropy of the hydrophobic chains. Bending stiffness is characterized by two bending
moduli quantifying stiffness in the two orthogonal radii of curvature
possible for a planar membrane. For an initially flat membrane in the
x–y plane, one bending direction can be visualized in the z direction
along the x-axis and the other in the z direction along the y-axis. The
bending stiffness is strongly dependent on the nature of the lipids
and their spatial distribution.
Figure I. Structures formed by different lipids: (a) inverted conical lipids, such
as detergent molecules, lysophospholipids and polyphosphoinositides; (b)
cylindrical-shaped lipids, such as phosphatidylcholine and sphingomyelin; (c)
conical lipids, such as diacylglycerol and PE.
how physical features of the membrane, in this case the
increase in free energy of the membrane as one leaflet is
stretched while the other is compressed, lead to selective
movement of lipid molecules from one leaflet to the other to
equalize lateral tensions without necessarily activating
any specific protein-based lipid translocation complex.
Lateral lipid asymmetry
The lipid bilayer is also heterogeneous laterally, with
various descriptions of this asymmetry having been put
forward as evidence of rafts or other domains [16]. Formation of specialized domains in the inner leaflet of the
Vol.xxx No.x
3
plasma membrane are hypothesized to be important in
recruitment of signaling complexes to sites where transmembrane receptors are activated, and in establishing
polarity necessary for directed cell locomotion. In contrast
to the broad consensus and clear evidence for the general
features of transmembrane asymmetry, the nature and
even existence of physiologically relevant lateral membrane domains is still controversial [17,18]. This lack of
consensus is surprising, because lateral segregation of
cholesterol-induced microdomains in sphingomyelin
bilayers and other synthetic mixed lipid systems [19–25]
was demonstrated soon after development of the fluid
mosaic model of lipid membranes [26], and models of
domains based on studies on cellular membranes were
reported not long after [21,27–30].
In part the difficulty in demonstrating that domains exist
in cell membranes, as opposed to model membranes in which
the evidence is clear, is that domains in cells are too small to
visualize by existing methods, and manipulations that render them large enough to visualize are open to the criticism
that the manipulation itself caused them to form. Unlike
trans-bilayer lipid movements, which are slow in the
absence of perturbations and therefore relatively easy to
maintain, lateral movements of lipids within a leaflet are
very rapid, and domains can form and disappear on a
millisecond timescale, allowing measurements by spectroscopic methods but perhaps not direct visualization.
Furthermore, stabilization of domains could be due as much
to protein-lipid binding as to lipid–lipid interactions [31].
However, the insights gained from model systems that
show, for example, the often dominant effect of cholesterol
or lipids with long chain fatty acids on mixing or segregation
of lipids in mixed systems might suggest strategies by which
to detect or manipulate lipids domains in vivo. Current
problems in studying lipid domains in cell membranes, with
emphasis on the technical challenges that limit visualization of these small domains, the special role of cholesterol
[32,33] and other lipids in domain formation, and in the
conceptual challenges to relate equilibrium phase diagrams
of pure systems to small, transient domains in the cell, have
been discussed in several recent reviews [31,34–36].
Lateral membrane pressures and the regulation of
integral membrane proteins
The conformation of amphiphilic molecules is a compromise of free energies of their hydrophobic and hydrophilic
parts. In a lipid bilayer this compromise results in neither
the hydrophobic nor hydrophilic part of the phospholipid
being in the lowest energy configuration that it would take
if it were not tied to its chemically incompatible partner.
The hydrophilic headgroups at the surface of the membrane are crowded together more tightly than they would
be if free in solution. This frustration is evident when a
headgroup such as Ins(1,4,5)P3 is liberated from the membrane by a phospholipase and diffuses into the cell interior
to activate its cytoplasmic targets. The small diacylglycerol
remaining in the membrane after PtdIns(4,5)P2 hydrolysis
can alter the membrane structure, because its small headgroup renders it unstable in a flat bilayer; removal of
PtdIns(4,5)P2 from the plasma membrane has recently
been shown to be necessary for proper secretory vesicle
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
4
TRENDS in Cell Biology Vol.xxx No.x
Box 3. Forces within the lipid bilayer
Transverse forces
Membrane thickness is determined by the hydrophobic length of the
lipid. The optimal membrane thickness depends on the chain length,
the degree of saturation and the angle of tilt within the membrane.
The transition from a thicker to a thinner membrane generates
packing disorders that increase elastic energy. Transmembrane
proteins also have a specified length of hydrophobic contour that
can differ from the optimal hydrophobic thickness of the bilayer. This
hydrophobic mismatch can lead to stretching or compression of lipids
and proteins within the membrane (Figure I) or to tilting of transbilayer helices to decrease the hydrophobic height. Insertion of
different lipids in an isolated domain and insertion of proteins can
both affect the thickness of the membrane.
Figure I. Schematic representation of the hydrophobic mismatch between a
membrane protein of hydrophobic length dP in a lipid bilayer in which the
unperturbed hydrophobic thickness dL is smaller (top) or larger (bottom) than dP.
The influence of the protein extends over a certain distance from the protein
surface and progressively vanishes, so that the bilayer recovers its unperturbed
thickness dL. Reproduced with permission from Refs. [68,69].
Line tension
When a heterogeneous population of membrane lipids separate into
domains, the border between domains results in lipid packing that
is different from that inside and outside the domain, resulting from
such effects as the differences in height between the domains. The
deformation of molecules at the domain boundary that occurs mainly
to prevent exposure of hydrophobic regions to water costs energy,
fusion, perhaps because of the membrane perturbation
that follows release of the large PtdIns(4,5)P2 headgroup
[37]. By contrast, the hydrophobic acyl chains of phospholipids are generally stretched out more than they would be
without their hydrophilic anchors. The end to end distance
of, for example, a 16 carbon chain in a dipalmitoyl phosphatidyl choline bilayer is much longer than the end to end
distance of hexadecane in bulk, and the loss of entropy that
comes from straightening out the chain results in a significant lateral pressure within the lipid bilayer that varies
with the depth into the bilayer (Box 1). Structural and
theoretical work have provided quantitative estimates for
how much different regions of the phospholipid acyl chains
deviate from a random configuration [38], and this deviation results in a lateral pressure gradient throughout the
lipid bilayer [39] that can affect membrane curvature or
the structure of transmembrane proteins.
Even though the bilayer as a whole might be stable,
each part of it is highly stressed. In general the
and this energy per length of the boundary is called the line tension.
The magnitude of the line tension, which is generally not measurable
directly, contributes to the parameters that determine domain size
and stability.
Lateral pressures
A pressure due to loss of chain entropy within the hydrophobic
domain creates compressive forces within the bilayer, the magnitude
of which depends on the distance into the center of the bilayer, the
nature of the hydrophobic chains (e.g. saturated, unsaturated, single
chains or sterols) and the membrane curvature.
A compression force acts at the hydrophilic interface to crowd the
headgroup close enough to minimize exposure of the hydrophobic
chains to water. These lateral forces are present even if no external
force is applied to the membrane. Because such forces – resulting
from, for example, osmotic stress, membrane bending, or pulling on
transmembrane proteins – deform the membrane, the lateral forces
are also affected and therefore the structure of proteins inserted in the
bilayer can change. See Figure II.
Figure II. (a) The forces that act within the bilayer. Black lines represent the
hydrophobic chains and blue dots the hydrophilic headgroup. (b) The
corresponding lateral pressure, p(z), at different distances (z) across the bilayer
thickness. Strong tensions at the interfaces are balanced by positive pressures
through the interior, which are greatest near the interfaces. When the areas
under the curves add to zero, the membrane is globally at rest. The red
arrows show how a mismatch in the thickness of a transmembrane protein and
the lipid bilayer acyl chains moves the regions of high pressure up or down
along the z-axis, and the blue arrows show how bending the membrane alters
the pressure gradient within the bilayer. Adapted from [39,69].
hydrophobic–hydrophilic interface exerts interfacial
tension, pulling the molecules together; this is due to the
hydrophobic effect minimizing contacts of the hydrocarbon
parts with the aqueous phase, which is balanced by the
steric repulsion between the headgroups and entropic repulsion between the acyl chains in the monolayer leaflets,
exerting lateral pressure that tends to compress proteins
embedded within. As the forces acting on the system are
confined to very narrow zones of only few Ångstroms within
the bilayer, the prevailing pressures can be sufficient to
influence transmembrane protein structure. These lateral
stresses, which depend sensitively on lipid composition,
curvature, pH, divalent cations, drugs and binding to proteins, are increasingly considered in models for how transmembrane channels and other proteins, especially those
that respond to force, can alter their configurations when
they are stimulated [40]. Similar conformational changes
can also be produced by transverse forces that result from
hydrophobic mismatch [41], as illustrated in Box 3.
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
TRENDS in Cell Biology
Figure 1. Physical model for transmembrane lipid scrambling. In a flat membrane,
lipid asymmetry is built into the inner and outer leaflets by the work of flippases
and other enzymes. In the oversimplified view here one the outer leaflet is all
phosphatidylcholine (PC, blue) and the inner leaflet all phosphatidylserine (PS,
pink). When the membrane is bent in a concave shape, the area per lipid of the
outer leaflet increases while the area per lipid in the inner leaflet decreases. The
energy imbalance of lipids too crowded together (red) on one side and too far
apart on the other (light blue) to satisfy the optimal packing constraints of the lipids
lowers the activation energy for trans-bilayer flipping and lipids in the crowded
leaflet move to occupy space in the dilated leaflet. In a typical protrusion there are
regions of positive and negative curvature on both leaflets, so PS will flip to the
outside and PC will flip to the inside, leading to lipid scrambling.
An example of how interfacial forces contribute to channel function is provided by two recent theoretical models
that consider the changes in free energy as the spatial
relationship between a membrane channel and the membrane in which it is embedded changes. As shown in
Figure 2, a typical transmembrane channel, for example
a mechanosensitive ion channel such as a TRP channel,
has an asymmetric profile within the hydrophobic part of
the lipid bilayer, characterized by the angle u and a
hydrophobic height W that is less than the bilayer
Vol.xxx No.x
5
hydrophobic thickness. In the resting (closed) state, the
protein is in mechanical equilibrium with both lateral and
transverse forces within the lipid bilayer.
These membrane forces can change in several ways. For
example, if the bilayer is dilated as in a patch clamp
exerting suction pressure or perhaps in a cell undergoing
hypo-osmotic swelling, the lateral pressure within the
bilayer decreases, with a resulting increase in the lateral
cross-section of the protein, leading to channel opening.
Two recent studies suggest alternative means to activate
the channel [42,43]. If the membrane bends, or its lipid
composition changes in one or the other leaflet so as to alter
lateral pressures or hydrophobic height, the pressure profiles will also change and the protein conformation
responds. If the channel were to tilt within the membrane,
perhaps because of the imposition of force on the proteins
but not directly on the lipid, the spatial relation between
protein and membrane would also change and potentially
activate the channel. Similar scenarios can be envisioned
for other classes of transmembrane proteins to provide
mechanisms that alter transmembrane protein function
even in the absence of direct activation by a specific chemical ligand. The theoretical models are consistent with
experimental data [42,43], and future work will probably
enable unambiguous distinction between physical and
chemical activation mechanisms. In particular, quantitative measurements of the forces exerted on transmembrane proteins due to shear flow, osmotic stress or cellcell interactions need to be compared with the forces that
are estimated from modeling studies to be required to
produce the requisite changes in protein structure.
The possibility that lateral pressures in the membrane
owing to lipid packing can alter protein function and affect
cellular signals, including the potency of anesthetic agents
[44], is reinforced by recent findings that some peptides
hypothesized to alter ion channel activity by binding the
eukaryotic cell channel protein are also potent antibacterial peptides [45]. It is possible that these small peptides
have specific and distinct protein ligands on prokaryotic
and eukaryotic cells, and the finding that stereoisomers of
steroids [46,47] can have distinct effects on ion channels
Figure 2. Cross section of a mechanosensitive transmembrane complex such as a channel in its closed configuration. The geometry of the protein is described by three
parameters: the radius R, of the folded polypeptide, the thickness W of its hydrophobic domain, and the angle u that the hydrophobic domain boundary makes with a line
perpendicular to the membrane surface. The hydrophobic mismatch, 2U, is the difference between the hydrophobic protein thickness, W, and the bilayer equilibrium
thickness, 2a. Changes in the lateral bilayer forces in each leaflet can alter the polypeptide structure embedded within, and tilting of the protein [42] or moving it up or down
[40] will subject it to a different force profile within the bilayer that can alter the folding of the polypeptide. Adapted from Ref. [43].
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
6
TRENDS in Cell Biology Vol.xxx No.x
suggests some degree of chemical specificity. However,
unique protein ligands for many amphipathic compounds
that affect cell function have not yet been identified, and an
effect elicited by amphipathic molecules on membrane
biophysics is also plausible as an explanation for their
biological activities.
Elasticity of the membrane
The cell membrane resists deformation, and the magnitude of this resistance to forces applied in various directions is characterized by several elastic constants that
characterize resistance to different geometries of deformation: shear, bending and stretching, discussed above. These
physical properties depend on the chemical composition of
the bilayer and on the lateral and transverse asymmetries,
discussed below.
Shear deformation and viscous flow
Shear deformations within the plane of the fluid bilayer
present in eukaryotic cells meet no elastic resistance
because the lipids and the transmembrane proteins can
flow past each other. An underlying protein mesh, such as
the spectrin–actin network, endows the membrane with
resistance to shear, and the composite of 2D protein network and lipid bilayer together determine the remarkable
viscoelastic properties of erythrocytes and other cells [48].
The lack of resistance to shear places limits on how forces
can be applied, for example by motor proteins at the
membrane surface. For example, myosin 1 [49,50], kinesin
[51,52] and other motor proteins have specific binding sites
for phosphoinositides or other acidic lipids residing in the
inner leaflet, suggesting that these lipids anchor them to
the plasma membrane or the surface of a vesicle. Such an
anchor could suffice to transport a vesicle within the
cytoplasm as the motor walks along its track of actin or
microtubules, but it is less clear whether this mechanism
can be used to displace the plasma membrane with respect
to the cytoskeleton. Without a shear elastic modulus, a
resistance to static deformation and not only to flow, the
lipid part of the membrane might allow movement of
the cytoskeleton if the motor moved rapidly enough for
the viscous resistance to be significant, but a slow movement would result in passive flow of the lipid to which the
motor is anchored, with no relative motion of the cytoskeleton. This scenario is changed if the anchoring lipid is
bound within a larger structure or sequestered within a
rigid domain. Studies of the types of movement generated
by membrane-localized motors with and without linkage to
transmembrane proteins therefore have the potential to
determine whether domains of increased mechanical stability can form within the lipid bilayer or whether controlled movements of the membrane require a protein
lattice to stabilize the membrane.
Membrane bending
Even without proteins, lipid bilayer membranes resist
stretching and bending with elastic constants that are
physiologically relevant. On the other hand, many membrane phospholipids, especially those in the inner leaflet,
introduce a spontaneous curvature because they prefer to
pack into curved but not flat bilayers. In addition to the
bilayer sheets that form the plasma membrane and
surround internal organelles, membrane lipids can form
a wide variety of structures, many of which are found in
biology [53]. A striking feature of the chemical composition
of cell membranes is that many if not most of their lipid
constituents are, by themselves, unable to form planar
bilayer membranes. Phosphatidylcholine (PC) and PS
are the common constituents of the outer and inner leaflets, respectively of eukaryotic cell plasma membrane and
do form flat or gently curved planar membranes in vitro,
but PE, cholesterol and other abundant cellular lipids, and
important rarer lipids such as phosphoinositides, diacylglycerol, ceramides and lyosphospholipids, cannot form
bilayers except when mixed with other lipids (see Box
2). The presence of these lipids in planar membranes
destabilizes them, and indeed this destabilization seems
to be essential for the biological function of membranes and
for their ability to undergo vesicle budding, fusion and
other shape transformations. Therefore, local accumulation of these lipids in specialized domains will have
mechanical as well as biochemical consequences. Localization of acidic lipids into fluid domains, for example, can
trigger phospholipase A2 activity [54], with resulting
destabilization of membrane structure as lysophospholipids accumulate.
An increment in the internal pressure of the bilayer
hydrocarbon region by lipids such as diacylglycerol or PE
with unsaturated chains increases the tendency for the
membrane to curl, while remaining lamellar. Such a state
is called ‘frustrated’ as these lipids increase the tendency of
the membrane to adopt a negative curvature while the
lamellar state remains relatively flat to accommodate
other contributions to its total free energy. One of the
key enzymes in cellular signaling cascades, protein kinase
C, a peripheral membrane protein, can be activated by this
membrane stress [55] because the lipid packing around its
substrate is altered to allow the kinase to access the site of
phosphorylation. As another example, a novel type of a
peripheral lipid–protein interaction called ‘extended-lipid
anchorage’, has been described for cytochrome c, in which
high internal pressure in the membrane hydrocarbon
region promotes movement of the acyl chain to the membrane surface and further into a hydrophobic cavity inside
this protein, thus establishing a hydrophobic lipid–protein
interaction in the absence of intercalation of the protein
into the bilayer [56]. These studies demonstrate that the
activity of membrane proteins can be regulated (i) by direct
lipid-protein interactions, with specific lipids acting as
allosteric effectors, and (ii) by lipids influencing the physical state of the membrane. Obviously, these two mechanisms are not mutually exclusive for a given lipid.
Active remodeling of lipid bilayers
Enzymes acting on membrane lipids can have pronounced
consequences not only in causing changes in the lateral
distribution of lipids but also in producing changes in the
3D organization of membranes. For example, removal of
the phosphocholine headgroup of sphingomyelin by sphingomyelinase (SMase) yields ceramide, a lipid with very
different physicochemical properties that can alter spontaneous curvature and lateral packing. Although
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
TRENDS in Cell Biology
Vol.xxx No.x
7
sphingomyelin in the absence of cholesterol is miscible in
phosphatidylcholine, ceramide has a profound tendency for
segregation into microdomains, driven by intermolecular
hydrogen bonding [57]. This results in tight packing and
reduced trans-to-gauche isomerization of the hydrocarbon
chains, which changes acyl chain packing and increases
bending rigidity. As ceramide has a tendency to promote
the formation of the inverted hexagonal phase (Figure Ic in
Box 2), the domains enriched in ceramide form projections
with high curvature. Experiments using microinjection of
SMase on the surface of giant vesicles composed of phosphatidylcholine and sphingomyelin have demonstrated the
formation of smaller vesicles emerging from the larger
vesicle, consistent with this model. More specifically formation of either endocytotic vesicles into the internal
cavity of giant liposomes or shedding of vesicles from the
outer surface of the substrate liposome following the action
of SMase on the external or internal leaflet of the giant
liposome was seen [58]. In this model system, the stresses
induced on the initially relatively flat membrane due to the
hydrolytic activity of SMase are sufficient to get either
inward or outward budding of new vesicles depending on
which side of the vesicle the enzyme is delivered. These
results show how reorganization of cellular membranes
can be driven without ATP, simply by inducing enzymatically a phase transition of the membrane lipid composite in
one leaflet that causes structural changes to the bilayer.
Membrane stretching
Lipid bilayers strongly resist stretching because increasing
the average distance between head groups increases exposure of the hydrophobic domain to water. Some membranes, such as the plasma membrane of leukocytes,
have much greater surface area than needed to enclose
the cell volume and so deformation of the membrane does
not lead to bilayer stretching. Deformation of cellular
membranes, such as the formation of tubular invaginations from the plasma membrane, can be controlled by
changes in lipid composition and by the binding of specific
proteins, such as those containing BAR domains, which are
thought to bind the membrane to produce large tubular
structures [59]. Some organelle membranes might be near
the limit at which further deformation is elastically
resisted, and this resistance can influence changes in
shape. For example, the membrane of the endoplasmic
reticulum is pulled into tubes by motors that run along
microtubules [60]. This process has recently been reproduced in vitro with a minimal set of proteins and purified
lipids [61,62]. As shown in Figure 3, thin tubes can be
pulled out of a large vesicular reservoir that can mimic the
tubes pulled out of the ER. An important finding of this
study is that the force needed to pull out a pure lipid tube is
on the order of 50 pN and therefore requires coordinated
pulling by multiple motors. The resistance of the membrane to deformation can be sufficiently large to stall the
motors, stop tube elongation, and in some cases lead to
elastic recoil of the tube. Therefore, tube extension in vivo
could be initiated by decreasing membrane tension as well
as by activating motors, and extension might be stopped
when membrane tension reaches a high value or when the
motor is inactivated. Changes in the lipid composition of
Figure 3. Shape changes in a membrane vesicle pulled by molecular motors. (a) A
confocal side-view image of a fluorescently labeled lipid vesicle shows a long
membrane tube pulled out of the vesicle by kinesin motor proteins that bind to the
membrane by biotin–avidin links and translocate along microtubules that are
firmly attached (but not visualized) to the bottom surface. (b) A schematic
representation of the geometry and attachment sites at the tip of the tube. (c) A
diagram of the hypothetical arrangement of motors at the tip of the tube. Note the
accumulation of multiple motors at the tip of the protrusion that occurs as motors
slow down when they reach the end of the tube. Adapted from Ref. [62].
the ER, the binding of peripheral membrane proteins that
alter surface tensions, or changes in internal pressure can
all impact the rate and direction of tube extension, even
without a change in the number of bound or activated
motor proteins. Conversely, the mechanical work of pulling
out a tube might change the lipid composition at distinct
regions within the tube compared with that in the parent
vesicle if the membrane contains a mix of lipids that prefer
curved surfaces that occur at the tip and the base of the
extended tube [63].
Tube formation and fission into vesicles in vivo can use a
range of proteins that alter or sense changes in membrane
mechanical properties. For example, proteins containing
BAR domains, which are crescent shaped structures that
interact electrostatically with curved membranes, are
implicated in the formation of membrane tubes [59]. Membrane tube formation is also facilitated by proteins homologous to Schizosaccharomyces pombe Cdc15 (PCH
proteins), which bind to PS and PtdIns(4,5)P2 to deform
the membrane [64]. Additional changes to membrane
structure occur when twisting motions within membrane
tubules coupled with the GTPase activity of dynamin
increase membrane tension, leading to fission [65]. Some
BAR domain proteins also bind dynamin, suggesting that
the effects of these two classes of proteins on membrane
curvature and stability are coordinated in vivo [59].
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
8
TRENDS in Cell Biology Vol.xxx No.x
Apart from the lateral pressure profile, it is important
also to take into account the so-called equilibrium lateral
pressure pe, that is, the pressure at which the tension due
to the hydrophobic effect and the counteracting steric
repulsion between the headgroups and acyl chains of vicinal lipids are at balance. Although the measurement of
this pressure in biomembranes remains a challenge, theoretical studies and various types of experimental
approaches have estimated the value of pe at 33–
35 mN m 1. The best examples of proteins controlled by
lateral pressure are the phospholipases A2. Specific types
of phospholipase A2 from different species have maximal
enzymatic activities in vitro at specific values of lipid
lateral packing that can be tightly controlled in lipid
monolayers using Langmuir balances. Extrapolating from
this approach, it is evident that equilibrium lateral pressure and thus phospholipase A2 activity can be controlled
by stretching of membranes by osmotic pressure gradients.
More specifically, a pressure of 33–35 mN m 1 is too high
for some enzymes, preventing proper insertion of the protein in the lipid bilayer. However, subjecting the membrane to tension (for example as a result of osmotic swelling
that decreases lateral pressure) lowers pe, enabling phospholipase A2 to bind and orient itself in the surface in the
manner required for the expression of its catalytic activity
[66]. Importantly, this example demonstrates a fundamental role for the physical state of the membrane in controlling the activity of a membrane protein: a mechanical force
is directly converted into a biochemical signal, and the
membrane is thus acting as an osmotic response element
[67].
Conclusion
The biophysical features of the cell membrane are increasingly recognized to be important control elements in cell
signaling and membrane protein function. We have discussed here a few examples where physical effects can be
as important as specific biochemical reactions in the function of the cell membrane. However, a separation between
physical and chemical events in the membrane is subtle.
Nearly any chemical change in the membrane caused by
lipid hydrolysis, trafficking or sequestration in the membrane has a physical consequence, manifested, for example, as a change in pressure or curvature. Likewise,
mechanical work done on the membrane to bend or expand
it will cause redistribution of the hundreds of distinct lipid
species that form the bilayer. It seems likely that physical
and chemical features have evolved together to form the
complexity of interactions responsible for cell function.
References
1 Sato, H. and Feix, J.B. (2006) Peptide-membrane interactions and
mechanisms of membrane destruction by amphipathic alphahelical antimicrobial peptides. Biochim Biophys Acta 45, 9997–
10007
2 Vogel, V. and Sheetz, M. (2006) Local force and geometry sensing
regulate cell functions. Nat. Rev. Mol. Cell Biol. 7, 265–275
3 Janmey, P.A. and Weitz, D.A. (2004) Dealing with mechanics:
mechanisms of force transduction in cells. Trends Biochem. Sci. 29,
364–370
4 Devaux, P.F. et al. (2006) Proteins involved in lipid translocation in
eukaryotic cells. Chem. Phys. Lipids 141, 119–132
5 Balasubramanian, K. and Schroit, A.J. (2003) Aminophospholipid
asymmetry: A matter of life and death. Annu. Rev. Physiol. 65, 701–
734
6 Boon, J.M. and Smith, B.D. (2002) Chemical control of phospholipid
distribution across bilayer membranes. Med. Res. Rev. 22, 251–281
7 Natarajan, P. et al. (2004) Drs2p-coupled aminophospholipid
translocase activity in yeast Golgi membranes and relationship to in
vivo function. Proc. Natl. Acad. Sci. U. S. A. 101, 10614–10619
8 Zwaal, R.F. et al. (2005) Surface exposure of phosphatidylserine in
pathological cells. Cell. Mol. Life Sci. 62, 971–988
9 Utsugi, T. et al. (1991) Elevated expression of phosphatidylserine in the
outer membrane leaflet of human tumor cells and recognition by
activated human blood monocytes. Cancer Res. 51, 3062–3066
10 Segrest, J.P. et al. (1976) Amyloid A: amphipathic helixes and lipid
binding. Biochemistry 15, 3187–3191
11 Zhao, H. et al. (2004) Formation of amyloid fibers triggered by
phosphatidylserine-containing membranes. Biochemistry 43, 10302–
10307
12 Zhao, H. et al. (2005) Binding of endostatin to phosphatidylserinecontaining membranes and formation of amyloid-like fibers.
Biochemistry 44, 2857–2863
13 Bucki, R. et al. (2000) Phosphatidylinositol 4,5-bisphosphate domain
inducers promote phospholipid transverse redistribution in biological
membranes. Biochemistry 39, 5838–5844
14 Bucki, R. et al. (2001) Involvement of phosphatidylinositol 4,5bisphosphate in phosphatidylserine exposure in platelets: use of a
permeant phosphoinositide-binding peptide. Biochemistry 40,
15752–15761
15 Elliott, J.I. et al. (2006) Phosphatidylserine exposure in B lymphocytes:
a role for lipid packing. Blood DOI: 10.1182/blood-2005-11-012328
(www.bloodjournal.org)
16 Simons, K. and Vaz, W.L. (2004) Model systems, lipid rafts, and cell
membranes. Annu. Rev. Biophys. Biomol. Struct. 33, 269–295
17 Nichols, B. (2005) Cell biology: without a raft. Nature 436, 638–639
18 Munro, S. (2003) Lipid rafts: elusive or illusive? Cell 115, 377–388
19 Suurkuusk, J. et al. (1976) A calorimetric and fluorescent probe study of
the gel-liquid crystalline phase transition in small, single-lamellar
dipalmitoylphosphatidylcholine vesicles. Biochemistry 15, 1393–1401
20 Galla, H.J. and Sackmann, E. (1975) Chemically induced phase
separation in mixed vesicles containing phosphatidic acid. An
optical study. J. Am. Chem. Soc. 97, 4114–4120
21 Stier, A. and Sackmann, E. (1973) Spin labels as enzyme substrates.
Heterogeneous lipid distribution in liver microsomal membranes.
Biochim. Biophys. Acta 311, 400–408
22 Shimshick, E.J. and McConnell, H.M. (1973) Lateral phase separations
in binary mixtures of cholesterol and phospholipids. Biochem. Biophys.
Res. Commun. 53, 446–451
23 Shimshick, E.J. et al. (1973) Lateral phase separations in membranes.
J. Supramol. Struct. 1, 285–294
24 Trauble, H. and Sackmann, E. (1972) Studies of the crystalline-liquid
crystalline phase transition of lipid model membranes. 3. Structure of a
steroid-lecithin system below and above the lipid-phase transition. J.
Am. Chem. Soc. 94, 4499–4510
25 Sackmann, E. and Trauble, H. (1972) Studies of the crystalline-liquid
crystalline phase transition of lipid model membranes. I. Use of spin
labels and optical probes as indicators of the phase transition. J. Am.
Chem. Soc. 94, 4482–4491
26 Singer, S.J. and Nicolson, G.L. (1972) The fluid mosaic model of the
structure of cell membranes. Science 175, 720–731
27 Klausner, R.D. et al. (1979) Ionophore A23187 disrupts membrane
structure by modifying protein-lipid interactions. Nature 281,
82–83
28 Karnovsky, M.J. et al. (1982) The concept of lipid domains in
membranes. J. Cell Biol. 94, 1–6
29 Karnovsky, M.J. et al. (1982) Lipid domains in membranes. Ann. N. Y.
Acad. Sci. 401, 61–75
30 Klausner, R.D. et al. (1980) Lipid domains in membranes. Evidence
derived from structural perturbations induced by free fatty acids and
lifetime heterogeneity analysis. J. Biol. Chem. 255, 1286–1295
31 Hancock, J.F. (2006) Lipid rafts: contentious only from simplistic
standpoints. Nat. Rev. Mol. Cell Biol. 7, 456–462
32 Mouritsen, O.G. and Zuckermann, M.J. (2004) What’s so special about
cholesterol? Lipids 39, 1101–1113
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009
TICB-382; No of Pages 9
Review
TRENDS in Cell Biology
33 Miao, L. et al. (2002) From lanosterol to cholesterol: structural
evolution and differential effects on lipid bilayers. Biophys. J. 82,
1429–1444
34 London, E. (2005) How principles of domain formation in model
membranes may explain ambiguities concerning lipid raft formation
in cells. Biochim. Biophys. Acta 1746, 203–220
35 Kenworthy, A.K. (2005) Where do we go from here? Meeting Report on
the Biophysical Society Discussion on ‘Probing Membrane
Microdomains’, October 28–31, 2004, Asilomar, CA, USA. Traffic 6,
518–523
36 Mukherjee, S. and Maxfield, F.R. (2004) Membrane domains. Annu.
Rev. Cell Dev. Biol. 20, 839–866
37 Hammond, G.R. et al. (2006) Elimination of plasma membrane
phosphatidylinositol (4,5)-bisphosphate is required for exocytosis
from mast cells. J. Cell Sci. 119, 2084–2094
38 Sonne, J. et al. (2005) Methodological problems in pressure profile
calculations for lipid bilayers. J. Chem. Phys. 122, 124903
39 Cantor, R. (1997) Lateral pressures in cell membranes: a mechanism
for modulation of protein function. J. Phys. Chem. B 101, 1723–1725
40 Kung, C. (2005) A possible unifying principle for mechanosensation.
Nature 436, 647–654
41 Sackmann, E. (1984) Physical basis for trigger processes and
membrane structures. In Biological Membranes (Chapman, D.,
ed.), pp. 105–143, Academic Press
42 Turner, M.S. and Sens, P. (2004) Gating-by-tilt of mechanically
sensitive membrane channels. Phys. Rev. Lett. 93, 118103
43 Wiggins, P. and Phillips, R. (2004) Analytic models for
mechanotransduction: Gating a mechanosensitive channel. Proc.
Natl. Acad. Sci. U. S. A. 101, 4071–4076
44 Mohr, J.T. et al. (2005) Anesthetic potency of two novel synthetic
polyhydric alkanols longer than the n-alkanol cutoff: evidence for a
bilayer-mediated mechanism of anesthesia? J. Med. Chem. 48, 4172–
4176
45 Jung, H.J. et al. (2006) Lipid membrane interaction and antimicrobial
activity of GsMTx-4, an inhibitor of mechanosensitive channel.
Biochem. Biophys. Res. Commun. 340, 633–638
46 Romanenko, V.G. et al. (2002) Modulation of endothelial inwardrectifier K+ current by optical isomers of cholesterol. Biophys. J. 83,
3211–3222
47 Li, W. et al. (2006) Enantiomers of neuroactive steroids support a
specific interaction with the GABA-C receptor as the mechanism of
steroid action. Mol Pharmacol 69, 1779–1782
48 Discher, D.E. (2000) New insights into erythrocyte membrane
organization and microelasticity. Curr. Opin. Hematol. 7, 117–122
49 Tang, N. et al. (2002) Dynamics of myo1c (myosin-ibeta) lipid binding
and dissociation. J. Biol. Chem. 277, 42763–42768
50 Takeda, T. and Chang, F. (2005) Role of fission yeast myosin I in
organization of sterol-rich membrane domains. Curr. Biol. 15, 1331–
1336
51 Klopfenstein, D.R. and Vale, R.D. (2004) The lipid binding pleckstrin
homology domain in UNC-104 kinesin is necessary for synaptic vesicle
transport in Caenorhabditis elegans. Mol. Biol. Cell 15, 3729–3739
Vol.xxx No.x
9
52 Klopfenstein, D.R. et al. (2002) Role of phosphatidylinositol(4,5)
bisphosphate organization in membrane transport by the Unc104
kinesin motor. Cell 109, 347–358
53 Snedden, J. and Templer, R. (1995) Polymorphism of lipid-water
systems. In Structure and dynamics of membranes (Lipowsky, R.
and Sackman, E., eds), pp. 97–160, Elsevier
54 Leidy, C. et al. (2006) Domain-induced activation of human
phospholipase A2 type IIA: local versus global lipid composition.
Biophys. J. 90, 3165–3175
55 Drobnies, A.E. et al. (2002) CTP:phosphocholine cytidylyltransferase
and protein kinase C recognize different physical features of
membranes: differential responses to an oxidized phosphatidylcholine.
Biochim. Biophys. Acta 1564, 82–90
56 Tuominen, E.K. et al. (2002) Phospholipid-cytochrome c interaction:
evidence for the extended lipid anchorage. J. Biol. Chem. 277, 8822–
8826
57 Holopainen, J.M. et al. (1998) Sphingomyelinase induces lipid
microdomain
formation
in
a
fluid
phosphatidylcholine/
sphingomyelin membrane. Biochemistry 37, 17562–17570
58 Holopainen, J. et al. (2000) Vectorial budding of vesicles by
asymmetrical enzymatic formation of ceramide in giant liposomes.
Biophys. J. 78, 830–838
59 Itoh, T. et al. (2005) Dynamin and the actin cytoskeleton cooperatively
regulate plasma membrane invagination by BAR and F-BAR proteins.
Dev. Cell 9, 791–804
60 Waterman-Storer, C.M. and Salmon, E.D. (1998) Endoplasmic
reticulum membrane tubules are distributed by microtubules
in living cells using three distinct mechanisms. Curr. Biol. 8, 798–
806
61 Roux, A. et al. (2002) A minimal system allowing tubulation with
molecular motors pulling on giant liposomes. Proc. Natl. Acad. Sci.
U. S. A. 99, 5394–5399
62 Leduc, C. et al. (2004) Cooperative extraction of membrane nanotubes
by molecular motors. Proc. Natl. Acad. Sci. U. S. A. 101, 17096–
17101
63 Roux, A. et al. (2005) Role of curvature and phase transition in lipid
sorting and fission of membrane tubules. EMBO J. 24, 1537–1545
64 Tsujita, K. et al. (2006) Coordination between the actin cytoskeleton
and membrane deformation by a novel membrane tubulation domain of
PCH proteins is involved in endocytosis. J. Cell Biol. 172, 269–279
65 Roux, A. et al. (2006) GTP-dependent twisting of dynamin implicates
constriction and tension in membrane fission. Nature 441, 528–531
66 Lehtonen, J.Y. and Kinnunen, P.K. (1995) Phospholipase A2 as a
mechanosensor. Biophys. J. 68, 1888–1894
67 Kinnunen, P.K. (2000) Lipid bilayers as osmotic response elements.
Cell. Physiol. Biochem. 10, 243–250
68 Dumas, F. et al. (1999) Is the protein/lipid hydrophobic matching
principle relevant to membrane organization and functions? FEBS
Lett. 458, 271–277
69 Jensen, M.O. and Mouritsen, O.G. (2004) Lipids do influence protein
function-the hydrophobic matching hypothesis revisited. Biochim.
Biophys. Acta 1666, 205–226
www.sciencedirect.com
Please cite this article as: P.A. Janmey, P.K.J. Kinnunen, Biophysical properties of lipids and dynamic membranes, TRENDS in Cell Biology (2006), doi:10.1016/j.tcb.2006.08.009