Filling Potholes on the Path to Fusion Pores


New and Notable
Filling Potholes on the Path to
Fusion Pores

Barry R. Lentz,* David P. Siegel,†

and Vladimir Malinin‡

*Molecular and Cellular Biophysics,
Department of Biochemistry and
Biophysics, University of North Carolina at
Chapel Hill, North Carolina, †Givaudan Inc,
Cincinnati, Ohio, and ‡Elan
Pharmaceuticals, Princeton, New Jersey
08540 USA

Membrane fusion is at the heart of
eukaryotic cell life. Carefully regu-
lated membrane fusion is required for
cell compartmentalization, for the im-
port of large molecules into the cell,
and for the export both of waste mol-
ecules and of signaling molecules that
carry information to other cells in a
community. The complex cellular ma-
chinery that carries out these functions
is beginning to come into focus
(Brunger, 2001), but a detailed molec-
ular view of how these machines work
on membrane lipid bilayers to accom-
plish a change in topology of cellular
compartments is still lacking. The dif-
ficulty in dissecting this process lies
not just in the complexity of the ma-
chines, but also in the task of defining
experimentally a dynamic process op-
erating in a semi-ordered system such
as the lipid bilayer. Although experi-
mental approaches are available, it has
been popular for some time to treat this
problem theoretically, in terms of sim-
ple models in which complex arrange-
ments of lipid molecules are treated in
terms of the material properties of ar-
rays of lipid molecules as they occur in
macroscopic semicrystalline lipid
phases. Although this is clearly a vast
simplification, it is probably, if prop-
erly parameterized, a useful zeroth or-
der approach to the very difficult prob-
lem of estimating the free energies of

presumed intermediates on the molec-
ular path to fusion.

The widely accepted model for this
process derives from the original pro-
posal that two bilayers brought into
close contact can merge their contact-
ing (cis) monolayers in a torroidal
“stalk” that joins half the lipid compo-
nents of the two original bilayers (Fig.
2 B from Markin and Albanesi (2001)).
The distal (trans) monolayers of this
structure are not merged, which pre-
vents continuity between the trapped
aqueous compartments of the two fus-
ing membranes. This has been termed
the stalk hypothesis (Markin et al.,
1984). Later, it was recognized that the
stalk may exist in two different forms,
an initial stalk and a transmembrane
contact (Siegel, 1999). It was univer-
sally accepted that such structures
must have a free energy substantially
larger than that of the lipid bilayers
from which it was proposed to arise. A
systematic estimate of this energy was
first made in terms of two major con-
tributions: the energy associated with
bending planar monolayers into
toroids, and the interstice (“void”) en-
ergy associated with the junctions
where monolayers are peeled away
from one another (Siegel, 1993). The
latter was, in the initial level of ap-
proximation, estimated as proportional
to the surface area of the interface be-
tween lamellar structures and a hypo-
thetical void that represented the space
unfilled by uniformly packed mono-
layers distorted to match the hypothe-
sized nonlamellar stalk structure (Sie-
gel, 1993). The bending energy was
estimated in terms of the free energy
required to distort a uniform lamellar
monolayer from its spontaneous or “in-
trinsic” curvature to match the as-
sumed stalk shape. The energy of the
void was parameterized based on the
properties of the phase transition from
a lamellar to a nonlamellar phase (hex-
agonal phase) that was also viewed as
composed of bent lamellar structures

and voids. Based on this treatment, the
free energy of the stalk was estimated
to be so large (�200 kT) that there has
been question as to whether such a
structure could be part of the fusion
process, although these calculations
suggested that the stalk was the lowest-
energy intermediate of several pro-
posed fusion mechanisms (Siegel
1993). Thus, the stalk hypothesis was
found to face an “energy crisis.”

Two papers appear in this month’s
issue of the Biophysical Journal (Ko-
zlovsky and Kozlov, 2001; Markin and
Albanesi, 2001) that, along with an-
other recent paper (Kuzmin et al.,
2001), propose solutions to this energy
crisis. Markin and Albanesi (2001)
propose that the assumption of circular
torroidal geometry for the stalk pro-
duced too large a bending energy.
These authors show that it is possible
to relax this assumption and define, in
terms of two geometric parameters, a
stress-free stalk that has the same av-
erage intrinsic curvature as the lipid
mixture of which it is composed. Just
as for the original stalk hypothesis
(Markin et al., 1984; Siegel, 1993), one
can question some of the geometric
assumptions of this calculation. How-
ever, it is a reasonable result that the
bending energy of a geometrically re-
laxed stalk can be reduced signifi-
cantly from the original estimate that
was based on a rigid (circular torroid)
geometry. Even if the bending free en-
ergy of the stalk can be reduced to
zero, the interstice or void free energy
attributable to the intrinsic geometric
mismatch between lamellar and nonla-
mellar lipid structures must persist.
This is variously estimated as several
tens of kT, still a formidable energy
but much less imposing than the earlier
estimates of the stalk energy.

The other two articles both ask
whether it is appropriate to estimate
the interstice energy in terms of the
surface of the interface between a void
and smoothly bent lamellar structures.

Submitted December 14, 2001, and accepted for
publication December 14, 2001.

© 2002 by the Biophysical Society

0006-3495/02/02/555/03 $2.00

555Biophysical Journal Volume 82 February 2002 555–557



Although it is unlikely that a true mac-
roscopic void exists in intermediate
structures leading to fusion, this con-
struct was originally used as a self-
consistent and convenient way to esti-
mate the free energy associated with
the inherent nonlamellar nature of the
stalk (Siegel, 1993). The question is
whether a better method exists for es-
timating this free energy. Kozlovsky
and Kozlov (2001) and Kuzmin et al.
(2001) both argue that this energy is
better estimated in terms of the energy
needed to bend lamellar structures so
sharply that they accommodate the in-
herently nonlamellar nature of the pro-
posed stalk structure. This requires tilt
of individual lipid molecules and must
occur at an energy price, as tilting lipid
molecules relative to the bilayer nor-
mal expose their hydrophobic regions
to water and elongate acyl chains. Sim-
ilar to the void energy, the tilt defor-
mation energy can be estimated from
experimental data on phase transitions
of lipid liquid crystals from lamellar-
to-nonlamellar states.

Kozlovsky and Kozlov (2001) use
the tilt approach to treat the deforma-
tion of monolayers in terms of a com-
bination of splay (monolayer bending
and/or tilt gradient) and uniform tilt of
lipid molecules away from the local
monolayer normal. The “tilt” formal-
ism provides a means to account for
the interstice energy without the as-
sumption of a void. Figure 8 by Mar-
kin and Albanesi illustrates this ap-
proach relative to the traditional void
approach. Kozlovsky and Kozlov de-
scribe the stalk in terms of two geo-
metric parameters (width of the stalk
base, interbilayer distance). This ap-
proach yields a minimum free energy
for the stalk at an intrinsic curvature of
�0.1 nm�1 of 45–50 kT, which is ap-
proximately the same as one would
estimate for the interstice energy based
on the void model, as pointed out by
Markin and Albanesi. Curiously, at in-
trinsic curvatures of � �0.25 nm�1,
the stalk of Kozlovsky and Kozlov be-
comes thermodynamically stable rela-
tive to planar bilayers.

In summary, one essential finding of
both teams Markin/Albanesi and Ko-
zlovsky/Kozlov is that the energy re-
quired to deform the monolayers into a
stalk is substantially lower than calcu-
lated with the approach used previ-
ously. Basically, the source of the
overestimate in previous models was
the use of a simplistic geometric
model. Both sets of authors show that a
more realistically shaped stalk is likely
to have considerably lower bending
energy, which, relative to planar bilay-
ers, can even become negative for bi-
layers of substantial negative intrinsic
curvature. However, it seems that the
interstice free energy still dominates
the stalk free energy and we may still
have to fill the potholes in the pathway
to fusion pores.

There are tough problems that re-
main to model successfully the mech-
anism of fusion. In both these manu-
scripts, the authors calculate only the
lower bound to the true activation en-
ergy for formation of fusion interme-
diates. Present models account only for
the energy of intermediates viewed as
static structures. Experimentally, the
process of converting closely contact-
ing membranes (initial static structure)
to a fusion pore (final static structure)
involves at least three kinetic steps and
two intermediates (Lentz et al., 2000).
The first and last of these steps involve
changes in the topology of membranes
and trapped compartments and thus de-
mand dramatic rearrangements of lipid
and water molecules. These are not
likely to find description by the sort of
continuum, macroscopic theories re-
viewed here, although some attempts
have been made to do so (Kuzmin et
al., 2001; Markin and Albanesi, 2001).
The second step, interconversion of fu-
sion intermediates, probably does not
require changes in topology and may
be amenable to such an approach. Nor
have we really addressed the problem
of how two membranes become dis-
torted and so closely apposed that the
initial conversion to a fusion interme-
diate can occur. Most importantly, ex-
perimental tests of these models need
to be devised. Given the highly local-

ized and dynamic nature of membrane
fusion events, this will be very diffi-
cult. However, the same rules that gov-
ern the relative energies of different-
geometry intermediates must also
apply to the relative free energies of
lipid molecules in lamellar and nonla-
mellar phases. Therefore, it is probably
worth some effort to see which of the
new proposals best describes the rela-
tive stability and formation kinetics of
complex lipid assemblies such as in-
verted cubic phases. Clearly there is
much lipid physical chemistry that re-
mains to be done!

How do these new results help us
understand the process of protein-in-
duced membrane fusion? First, the res-
olution of the energy crisis increases
our confidence that structures such as
stalks do represent low-energy path-
ways to membrane fusion. Second, the
dominant nature of the void or inter-
stice energy suggests that we look to
how parts of fusion proteins that reside
in or interact with membrane bilayers
might lower this free energy. This will
require combinations of experimental
and theoretical approaches. In doing
so, it is appropriate to keep in mind
that the stalk hypothesis is an hypoth-
esis and that the actual mechanism of
biomembrane fusion may be even
more complex than suggested by this
model.

REFERENCES

Brunger, A. T. 2001. Structural insights into the
molecular mechanism of calcium-dependent
vesicle-membrane fusion. Curr. Opin. Struct.
Biol. 11:163–173.

Kozlovsky, Y., and M. M. Kozlov. 2002. Stalk
model of membrane fusion: solution of energy
crisis. Biophys. J. 82:882– 895.

Kuzmin, P. I., J. Zimmerberg, Y. A. Chiz-
madzhev, and F. S. Cohen. 2001. A quantita-
tive model for membrane fusion based on
low-energy intermediates. Proc. Natl. Acad.
Sci. U.S.A. 98:7235–7240.

Lentz, B. R., V. Malinin, M. E. Haque, and K.
Evans. 2000. Protein machines and lipid
assemblies: current views of cell membrane
fusion. Curr. Opin. Struct. Biol. 10:
607– 615.

Markin, V. S., and J. P. Albanesi. 2002. Mem-
brane fusion: stalk model revisited. Biophys.
J. 82:693–712.

556 Lentz et al.

Biophysical Journal 82(2) 555–557



Markin, V. S., M. M. Kozlov, and V. L. Borov-
jagin. 1984. On the theory of membrane fu-
sion. The stalk mechanism. Gen. Physiol. Bio-
phys. 3:361–377.

Siegel, D. P. 1993. Energetics of intermediates
in membrane fusion: comparison of stalk and
inverted micellar intermediate mechanisms.
Biophys. J. 65:2124 –2140.

Siegel, D. P. 1999. The modified stalk mecha-
nism of lamellar/inverted phase transitions
and its implications for membrane fusion.
Biophys. J. 76:291–313.

Fusion Pores 557

Biophysical Journal 82(2) 555–557