263
8 Molecular Dynamics Simulations of Lipid Bilayers with Incorporated Peptides
8.6 Hydrophobic Mismatch
It is possible to calculate peptide’s effective thickness from its tilt angle and its
length. Effective thickness means length of projection of the peptide to normal of
the membrane. This parameter should be equal to length of membrane hydrophobic
core to minimize system energy. The effective lengths (perpendicular to membrane
surface) of hydrophobic parts of peptides and both lipid shells of the membrane
are compared in Table 8.1. When comparing these values, it is important to keep in
mind that Lys side chains can flip in or out. This resulted in shortening or prolonging ofeffective length of peptide (max. to 0.4–0.5 nm). Thus, it is not necessary to
keep the same effective length of peptide and 1
st
shell of lipids to satisfy hydrophobic mismatch.
The thickness of hydrophobic part of unmodified membranes has a value of
3.6 nm for DPPC/gel, 3.2 nm for DMPC/gel, 3.0 nm for DPPC/LC and 2.6–2.7 nm
for DMPC/LC. In all cases the thickness of the second shell (compared with the first
shell) are closer to the unmodified membrane: the average thickness is 3.58 nm for
DPPC/gel, 2.93 nm for DPPC/LC, 3.12 nm for DMPC/gel and 2.62 nm for DMPC/
LC. Nearly in all cases of simulations, which resulted in the peptide tilt, the membrane is by 0.3 nm thicker than the effective length of the peptide. In those cases the
peptide tilt is bigger (according to the simulation results) so its effective length is
smaller. But the average membrane thickness does not contain direct information on
the orientation of individual lipid chains. Lipid chains can still be longer even in the
LC state (higher order parameters, more trans conformations of dihedral angles),
because they can tilt like the peptide.
The simulations of membrane in a gel state (V 24 /DPPC, P 24 /DPPC, LA 12 /DMPC,
and LA 12 /DPPC) suggest that the membrane affects the peptides conformation. The
whole helix is twisted into superhelical structure—helix composed from helical
chain (see Fig. 8.10). The whole structure resembles single chain from coiled coil
conformation. This structure is produced only in the gel membrane phase, where
there is only small hydrophobic mismatch. The difference of thickness between
peptide and membrane in LC state are too large to solve the situation similarly like
for gel state.
The I 24 /DPPC/LC simulations suggest that the average peptide tilt is very small
and didn’t solve the mismatch. Visual observation shows that at the beginning the
peptide tilted (up to approx. 20°), but it didn’t stay in this conformation, rather went
back nearly into its starting position. After short period of time it tilted again in
random direction, but again returned back. During the 40 ns of simulation, whole
tilting and returning process is repeated 5 times. The reason of this behavior remains
unknown.
Because some membrane phenomena are quite rare, we ran some (I 24 &DMPC/
LC, I 24 &DPPC/gel, I 24 &DPPC/LC, LA 12 &DMPC/gel, L 24 &DMPC/gel, V 24 &DPPC/
gel, V 24 /DPPC/LC) simulations to 100 ns. In all cases the conformation didn’t differ
much from end of original simulations. E.g. tilt in I 24 &DMPC/LC decreased by 2°,
I 24 &DPPC/LC finally stabilized at average angle 5.15° (effective thickness: 3.2 nm,
8 Molecular Dynamics Simulations of Lipid Bilayers with Incorporated Peptides
8.6 Hydrophobic Mismatch
It is possible to calculate peptide’s effective thickness from its tilt angle and its
length. Effective thickness means length of projection of the peptide to normal of
the membrane. This parameter should be equal to length of membrane hydrophobic
core to minimize system energy. The effective lengths (perpendicular to membrane
surface) of hydrophobic parts of peptides and both lipid shells of the membrane
are compared in Table 8.1. When comparing these values, it is important to keep in
mind that Lys side chains can flip in or out. This resulted in shortening or prolonging ofeffective length of peptide (max. to 0.4–0.5 nm). Thus, it is not necessary to
keep the same effective length of peptide and 1
st
shell of lipids to satisfy hydrophobic mismatch.
The thickness of hydrophobic part of unmodified membranes has a value of
3.6 nm for DPPC/gel, 3.2 nm for DMPC/gel, 3.0 nm for DPPC/LC and 2.6–2.7 nm
for DMPC/LC. In all cases the thickness of the second shell (compared with the first
shell) are closer to the unmodified membrane: the average thickness is 3.58 nm for
DPPC/gel, 2.93 nm for DPPC/LC, 3.12 nm for DMPC/gel and 2.62 nm for DMPC/
LC. Nearly in all cases of simulations, which resulted in the peptide tilt, the membrane is by 0.3 nm thicker than the effective length of the peptide. In those cases the
peptide tilt is bigger (according to the simulation results) so its effective length is
smaller. But the average membrane thickness does not contain direct information on
the orientation of individual lipid chains. Lipid chains can still be longer even in the
LC state (higher order parameters, more trans conformations of dihedral angles),
because they can tilt like the peptide.
The simulations of membrane in a gel state (V 24 /DPPC, P 24 /DPPC, LA 12 /DMPC,
and LA 12 /DPPC) suggest that the membrane affects the peptides conformation. The
whole helix is twisted into superhelical structure—helix composed from helical
chain (see Fig. 8.10). The whole structure resembles single chain from coiled coil
conformation. This structure is produced only in the gel membrane phase, where
there is only small hydrophobic mismatch. The difference of thickness between
peptide and membrane in LC state are too large to solve the situation similarly like
for gel state.
The I 24 /DPPC/LC simulations suggest that the average peptide tilt is very small
and didn’t solve the mismatch. Visual observation shows that at the beginning the
peptide tilted (up to approx. 20°), but it didn’t stay in this conformation, rather went
back nearly into its starting position. After short period of time it tilted again in
random direction, but again returned back. During the 40 ns of simulation, whole
tilting and returning process is repeated 5 times. The reason of this behavior remains
unknown.
Because some membrane phenomena are quite rare, we ran some (I 24 &DMPC/
LC, I 24 &DPPC/gel, I 24 &DPPC/LC, LA 12 &DMPC/gel, L 24 &DMPC/gel, V 24 &DPPC/
gel, V 24 /DPPC/LC) simulations to 100 ns. In all cases the conformation didn’t differ
much from end of original simulations. E.g. tilt in I 24 &DMPC/LC decreased by 2°,
I 24 &DPPC/LC finally stabilized at average angle 5.15° (effective thickness: 3.2 nm,