Biology Reference
In-Depth Information
aims to study changes in structural parameters such as area per lipid or membrane
thickness (see the succeeding text). In contrast to pressure coupling, the approach
used for controlling the temperature is frequently more standard. Nevertheless,
the choice of the temperature will obviously impact the timescale of the simulation
needed to reach convergence. Since both lipid composition and temperature will de-
termine the phase behavior of our systems, we need to know in advance what lipid is
holding the highest chain-melting transition temperature (
T
m
) in the mixture we are
simulating (
Koynova & Caffrey, 1998; Maulik & Shipley, 1996
).
Although frequently overlooked, the hydration level of lipid bilayers is another
important parameter, particularly if one attempts to study structural parameters such
as order or hydrogen-bonding networks (
Ho, Slater, & Stubbs, 1995
). Water
ordering at membrane interfaces plays an important role in different biological
events (
Kasson, Lindahl, & Pande, 2011; Robinson, Besley, O'Shea, & Hirst,
2011
), so using adequate hydration levels seems a more physiological approach
than just coarsely hydrating. Again, thanks to the last CHARMM force field update
(
Klauda et al., 2010
), the hydration of lipids can be realistically modeled.
Experimental hydration levels are available for different bilayers, although these
values frequently correspond to phosphatidylcholine bilayers (
Nagle & Tristram-
Nagle, 2000; Tristram-Nagle, Petrache, & Nagle, 1998
). Yet, the number of water
molecules per lipid (
nw
), a common experimental value that represents hydration,
is known to range approximately between 20 and 32.5 (
Kasson et al., 2011;
Robinson et al., 2011; Steinbauer, Mehnert, & Beyer, 2003
). Often, 30 water
molecules per phospholipid (
nw
¼
30) is taken as an adequate hydration level for
membrane systems.
Building and simulating pure membranes (i.e., one-component mixtures) is al-
ways a much simpler task when compared to membrane-protein systems. Despite
building a starting structure of a pure membrane has not always been easy, today,
we can even do this with a molecular visualization software such as VMD (visual
molecular dynamics) (
Humphrey, Dalke, & Schulten, 1996
). The last releases of this
tool offer a plug-in to build pure membranes already solvated. In contrast to earlier
years where scientist could only simulate pure membranes, today, we have access to
different automated tools that can build for us starting structures of multicomponent
membranes. Out of these set of tools, one web-based graphical interface, the
CHARMM-GUI (
http://www.charmm-gui.org/
)
(
Jo, Kim, Iyer, & Im, 2008
), is be-
coming very popular in both pure membrane and membrane-protein simulations
using the CHARMM force field. In particular, this tool offers a specific module
input/membrane
)
intended to help building pure
and mixed membranes (
Jo, Lim, Klauda, & Im, 2009
) or membrane-protein systems
(
Jo, Kim, & Im, 2007
) (see
Section 4.2.2
). This module is very useful to generate
input structures ready for MD simulations and provide a flexible and straightforward
set of options for this purpose.
The large number of lipid types available in the CHARMM-GUI membrane
builder, 77 to date, allows building membrane systems as heterogeneous as needed.
Four straightforward steps guide the user during the process. Firstly, the type and size
¼
