the molecular processes occurring beyond the reachable time scale by all-atom
simulations. Here, we focus on a comparison of the respective diffusion coefficients
of the vesicles simulated.
Interestingly, the lipid diffusion in the outer leaflet of each vesicle is typically
enhanced over that in a flat bilayer, but the lipid mobility in the inner leaflet is
suppressed. Moreover, the diffusion coefficient of lipid molecules in the outer
leaflet is almost constant for any size vesicle, whereas that in the inner leaflet
shows a strong size dependency. The outer leaflet of the inner bilayer of the MLV
shows slightly lower mobility, due to the effect of the outer bilayer, although a
similar effect on the inner leaflet of the opposing bilayer is not evident.
As for the ULV, we should see the convergence of D, as the size of vesicle
increases, to the diffusion coefficient of a flat DMPC bilayer. This is not apparently
observed in Table 1, suggesting a possible effect on the estimation of D due to the
thermal fluctuations or undulation of membranes. The net rotation of the vesicle
(each leaflet independently) may also need to be subtracted, although we did not
observe a clear motion of this type in the time scale simulated.
7 Perspectives and Challenges
We have briefly reviewed a few applications of our CG model for MD simulations of
amphiphilic polymers to predict their self-assembly into supramolecular structures
with surprising detail. We speculate that, in the near future, the assembly into
complex supramolecular structures will be studied as easily as chemical reactions
between small molecules have been modeled successfully in the past few decades by
electronic structure methods. However, several potential issues may arise: first and
foremost, the accuracy of the CG models used will be put to the test, because higher
computing power will allow the study of larger and more complex systems.
Another important challenge to CG–MD simulations of self-assembly will be the
ability to properly describe the dynamic equilibrium between chemical species in
mixed systems. To remain on the main subject of this brief review, biological
Table 1 Calculated lateral self-diffusion coefficient (D) of a DMPC molecule in flat and vesicular
membranes
Geometry
Number of DMPC molecules
D (Â 10
À6 cm
2
/s)
Inner leaflet
Outer leaflet
Flat bilayer
8,142
1.41
ULV
1,512
0.75
1.53
2,500
0.92
1.52
3,500
1.00
1.52
5,000
1.05
1.55
MLV
1,512
0.72
1.38
5,000
1.02
1.56
DMPC 1,2-dimyristoyl-sn-glycero-3-phosphocholine, ULV unilamellar vesicle, MLV
multilamellar vesicle
104
G. Fiorin et al.
simulations. Here, we focus on a comparison of the respective diffusion coefficients
of the vesicles simulated.
Interestingly, the lipid diffusion in the outer leaflet of each vesicle is typically
enhanced over that in a flat bilayer, but the lipid mobility in the inner leaflet is
suppressed. Moreover, the diffusion coefficient of lipid molecules in the outer
leaflet is almost constant for any size vesicle, whereas that in the inner leaflet
shows a strong size dependency. The outer leaflet of the inner bilayer of the MLV
shows slightly lower mobility, due to the effect of the outer bilayer, although a
similar effect on the inner leaflet of the opposing bilayer is not evident.
As for the ULV, we should see the convergence of D, as the size of vesicle
increases, to the diffusion coefficient of a flat DMPC bilayer. This is not apparently
observed in Table 1, suggesting a possible effect on the estimation of D due to the
thermal fluctuations or undulation of membranes. The net rotation of the vesicle
(each leaflet independently) may also need to be subtracted, although we did not
observe a clear motion of this type in the time scale simulated.
7 Perspectives and Challenges
We have briefly reviewed a few applications of our CG model for MD simulations of
amphiphilic polymers to predict their self-assembly into supramolecular structures
with surprising detail. We speculate that, in the near future, the assembly into
complex supramolecular structures will be studied as easily as chemical reactions
between small molecules have been modeled successfully in the past few decades by
electronic structure methods. However, several potential issues may arise: first and
foremost, the accuracy of the CG models used will be put to the test, because higher
computing power will allow the study of larger and more complex systems.
Another important challenge to CG–MD simulations of self-assembly will be the
ability to properly describe the dynamic equilibrium between chemical species in
mixed systems. To remain on the main subject of this brief review, biological
Table 1 Calculated lateral self-diffusion coefficient (D) of a DMPC molecule in flat and vesicular
membranes
Geometry
Number of DMPC molecules
D (Â 10
À6 cm
2
/s)
Inner leaflet
Outer leaflet
Flat bilayer
8,142
1.41
ULV
1,512
0.75
1.53
2,500
0.92
1.52
3,500
1.00
1.52
5,000
1.05
1.55
MLV
1,512
0.72
1.38
5,000
1.02
1.56
DMPC 1,2-dimyristoyl-sn-glycero-3-phosphocholine, ULV unilamellar vesicle, MLV
multilamellar vesicle
104
G. Fiorin et al.