265
8 Molecular Dynamics Simulations of Lipid Bilayers with Incorporated Peptides
the simulations restricted by several ns. At the same time they received similar
results with shorter—around 5 ns and longer—around 10 ns simulations. However,
it should be noted that characteristic time of relaxation of phospholipid dipole moments following membrane disturbance by voltage jump lies between micro- to
mili-second scale. Longer time probably corresponds to the collective movement
of lipid clusters [93]. Incorporation of the short peptide influences this relaxation
time significantly [92]. We can therefore expect that the relaxation time of the short
peptides, like L 24 , should be comparable or even larger than that for phospholipids.
Therefore, in order to receive equilibrium state of the peptide in a membrane, the
simulations should last in order of microseconds. Currently it is possible by means
of the coarse-graining (CG) models, which however lack the details of atomic resolution. In addition due to less degree of freedom the CG systems move rapidly than
atomic models. However, even simple CG models allow to obtain important information on the features of peptide-lipid interactions. This explains growing interest
to this method in recent five years, which include also combination of full atomic
and CG approach in modeling the peptide/lipid interactions (see Polyansky et al.
for recent review [30]). However as stated by Monticelli et al. [95], more shorter
simulations can provide better sampling of conformation space than longer one.
Despite the large number of limitations, MD represents a useful approach for the
study of fast conformational movements of peptides and phospholipids in a membrane, though we cannot be sure whether the model system reached equilibrium
or not. But results obtained by MD are consistent with experiments, in respect of
inducing hydrophobic mismatch and disordering effect of peptide on the membrane
in the gel state.
8.7 Conclusion
Our results confirmed the tendency of Lys-flanked peptides to compensate the positive mismatch between peptide and membrane hydrophobic core by tilting. Some
of the peptides, however, produce superhelical double-twisted structure. This only
occurs in the membrane in the gel phase, where only a small hydrophobic mismatch exists. The peptide also alters certain properties of the surrounding lipids
such as membrane ordering, the amount of dihedral angles in trans conformation
and the number of transitions between trans and gauche conformation. It is likely
that these effects should provide some preferable structural state of the peptides in
a membrane. The lipid structural state around the peptide is probably between gel
and liquid-crystalline state. This effect depends on peptide amino acid composition.
Amino acids with large side chains branched at C
β
(Ile, Val) produce helix, which
has more side chains fluctuates than that of a poly-Leu helix. This holds also for
small side chains (Ala).
Acknowledgments The access to the METACentrum and CERIT computing facilities provided
under the research intent MSM6383917201 is highly appreciated. This work was also supported by
the Slovak Research and Development Agency (Projects No. APVV—0410-10 and LPP-0341-09).
8 Molecular Dynamics Simulations of Lipid Bilayers with Incorporated Peptides
the simulations restricted by several ns. At the same time they received similar
results with shorter—around 5 ns and longer—around 10 ns simulations. However,
it should be noted that characteristic time of relaxation of phospholipid dipole moments following membrane disturbance by voltage jump lies between micro- to
mili-second scale. Longer time probably corresponds to the collective movement
of lipid clusters [93]. Incorporation of the short peptide influences this relaxation
time significantly [92]. We can therefore expect that the relaxation time of the short
peptides, like L 24 , should be comparable or even larger than that for phospholipids.
Therefore, in order to receive equilibrium state of the peptide in a membrane, the
simulations should last in order of microseconds. Currently it is possible by means
of the coarse-graining (CG) models, which however lack the details of atomic resolution. In addition due to less degree of freedom the CG systems move rapidly than
atomic models. However, even simple CG models allow to obtain important information on the features of peptide-lipid interactions. This explains growing interest
to this method in recent five years, which include also combination of full atomic
and CG approach in modeling the peptide/lipid interactions (see Polyansky et al.
for recent review [30]). However as stated by Monticelli et al. [95], more shorter
simulations can provide better sampling of conformation space than longer one.
Despite the large number of limitations, MD represents a useful approach for the
study of fast conformational movements of peptides and phospholipids in a membrane, though we cannot be sure whether the model system reached equilibrium
or not. But results obtained by MD are consistent with experiments, in respect of
inducing hydrophobic mismatch and disordering effect of peptide on the membrane
in the gel state.
8.7 Conclusion
Our results confirmed the tendency of Lys-flanked peptides to compensate the positive mismatch between peptide and membrane hydrophobic core by tilting. Some
of the peptides, however, produce superhelical double-twisted structure. This only
occurs in the membrane in the gel phase, where only a small hydrophobic mismatch exists. The peptide also alters certain properties of the surrounding lipids
such as membrane ordering, the amount of dihedral angles in trans conformation
and the number of transitions between trans and gauche conformation. It is likely
that these effects should provide some preferable structural state of the peptides in
a membrane. The lipid structural state around the peptide is probably between gel
and liquid-crystalline state. This effect depends on peptide amino acid composition.
Amino acids with large side chains branched at C
β
(Ile, Val) produce helix, which
has more side chains fluctuates than that of a poly-Leu helix. This holds also for
small side chains (Ala).
Acknowledgments The access to the METACentrum and CERIT computing facilities provided
under the research intent MSM6383917201 is highly appreciated. This work was also supported by
the Slovak Research and Development Agency (Projects No. APVV—0410-10 and LPP-0341-09).