Subscriber access provided by UNIV DE CORDOBA
The Journal of Physical Chemistry B is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Effect of the Barometric Phase Transition of a DMPA Bilayer on the Lipid/
Water Interface. An Atomistic Description by Molecular Dynamics Simulation
J. J. Giner Casares, L. Camacho, M. T. Martn Romero, and J. J. Lpez Cascales
J. Phys. Chem. B, 2007, 111 (49), 13726-13733 • DOI: 10.1021/jp075948v
Downloaded from http://pubs.acs.org on February 9, 2009
More About This Article
Additional resources and features associated with this article are available within the HTML version:
Access to high resolution figures
Links to articles and content related to this article
Copyright permission to reproduce figures and/or text from this article
Effect of the Barometric Phase Transition of a DMPA Bilayer on the Lipid/Water Interface.
An Atomistic Description by Molecular Dynamics Simulation
J. J. Giner Casares,†L. Camacho,†M. T. Martı ´n Romero,†and J. J. Lo ´pez Cascales*,‡
Departmento Quı ´mica Fı ´sica y Termodina ´mica Aplicada, Ed. Marie Curie, Campus de Rabanales, UniVersidad
de Co ´rdoba, 14014 Co ´rdoba, Spain, and UniVersidad Polite ´cnica de Cartagena, Centro de Electroquı ´mica y
Materiales Inteligentes (CEMI), Aulario II, Campus de Alfonso XIII, 30203 Cartagena, Murcia, Spain
ReceiVed: July 27, 2007; In Final Form: September 24, 2007
Understanding the structure and dynamics of phospholipid bilayers is of fundamental relevance in biophysics,
biochemistry, and chemical physics. Lipid Langmuir monolayers are used as a model of lipid bilayers, because
they are much more easily studied experimentally, although some authors question the validity of this model.
With the aim of throwing light on this debate, we used molecular dynamics simulations to obtain an atomistic
description of a membrane of dimyristoylphosphatidic acid under different surface pressures. Our results
show that at low surface pressure the interdigitation between opposite lipids (that is, back-to-back interactions)
controls the system structure. In this setting and due to the absence of this effect in the Langmuir monolayers,
the behavior between these two systems differs considerably. However, when the surface pressure increases
the lipid interdigitation diminishes and so monolayer and bilayer behavior converges. In this work, four
computer simulations were carried out, subjecting the phospholipids to lateral pressures ranging from 0.17 to
40 mN/m. The phospholipids were studied in their charged state because this approach is closer to the
experimental situation. Special attention was paid to validating our simulation results by comparison with
available experimental data, therebeing in general excellent agreement between experimental and simulation
data. In addition, the properties of the lipid/solution interface associated with the lipid barometric phase transition
This work was focused on the venerable and still prominent
topic of lipid bilayer structure, which plays a crucial role
controlling the diffusion of nutrients, ions, and water between
the inner and outer cell together, while providing a suitable
environment to other molecules embedded in the membrane,
such as membrane proteins or cholesterol, among others.1,2
Understanding the atomic interactions that control the behavior
of this complicated system is of unquestionable importance. One
of the aspects that has attracted increasing interest during recent
decades is the polymorphic phase transition of lipid bilayers
associated with temperature due to the importance of this process
from a biological viewpoint. In this respect, a temperature
denoted transition temperature (Tm) defines the temperature at
which lipids change from a gel (P′?) to a liquid crystalline state
(LR) at a constant external pressure.3,4
However, polymorphic phase transition may also be associ-
ated with different states induced by an external pressures or
barometric transition.5,6To understand the atomic interactions
that govern these phase transitions induced by an external
pressure, Langmuir air-water lipid monolayers have been
widely studied,7mainly because they enable the surface pressure
on the lipids that form the monolayer to be studied, and also
because a monolayer is basically half of a symmetric lipid
bilayer. For this purpose, different experimental techniques have
been used, such as scattering techniques, X-ray and neutron
reflectivity measurements, infrared reflection-absorption spec-
trometry (IR), NMR studies, and the ellipsometric technique.7-13
In the case of monolayers, different barometric phases have been
denoted, such as G (gaseous), LE (liquid expanded), LC (liquid
condensed), and SC (solid condensed).7
Improvements made in simulation algorithm and the increases
in the computing power attained during recent years have led
to molecular dynamics simulation14emerging as a powerful and
precise technique that provides atomic insight into the dynamic
Keeping this in mind with the aim of obtaining an atomic insight
into the interactions that control the phase transition of a lipid
bilayer, a computer simulation of a lipid bilayer of dimyris-
toylphosphatidic acid (DMPA-) was carried out. The main
reasons why we chose this type of lipid instead of DPPC or
DPPE were the following: first, because charged lipids at
physiological conditions play a crucial role in the structure and
functions of cellular membranes introducing an asymmetry
between the inner and outer side of the membrane, which can
constitute up to 20% of the total lipids, and second, because
DMPA-is a lipid widely used in the study of lipid monolay-
ers.20,21,9,11,13,22In this setting, considering that electrostatic
interactions control the lipid-lipid interactions in charged
lipids,16it is expected that different lipid packing associated to
their barometric phase transitions can modify noticeably the
Thus, the main goal of this work was, in a first instance, to
validate the force field of DMPA-, because as far as we know,
this is the first time that this type of lipid has been simulated.
To this end, the properties of the bilayer were studied and
* To whom correspondence should be addressed. E-mail: javier.lopez@
†Universidad de Co ´rdoba.
‡Universidad Polite ´cnica de Cartagena, Centro de Electroquı ´mica y
Materiales Inteligentes (CEMI).
J. Phys. Chem. B 2007, 111, 13726-13733
10.1021/jp075948v CCC: $37.00© 2007 American Chemical Society
Published on Web 11/16/2007
compared with available experimental data. After checking and
validating the force field for their liquid crystalline state (LR),
the system was subjected to high-surface pressures, bringing
the bilayer close to liquid condensed state (LC), with the goal
of describing its barometric behavior. Thus, changes in the lipid
structure associated with pressure are described and how these
variations affect the lipid/solvent interface properties is dis-
Assuming that lipid bilayers and monolayers are in fact
different physical systems, our hypothesis is that these systems
behave differently, and we propose an explanation in terms of
lipid-lipid interactions. What we present here is a systematic
and detailed study that supports our hypothesis. The obtained
results suggest that the simulated system behaves like a bilayer
in its liquid crystalline state at low surface pressure but behaves
more like a monolayer in its liquid condensed state when
subjected to high pressure.
For low surface pressures, the most important lipid interaction
is related with lipid tail interdigitation (back-to-back interac-
tions). However, because this kind of interaction is impossible
to achieve in a lipid monolayer, monolayers do not reproduce
bilayer behavior. However, at higher values of surface pressure,
the interdigitation associated with lipid tails vanishes, and the
most important effects are related with lateral interaction
between neighboring lipids. Thus, at sufficiently high pressures,
lipid bilayers and monolayers show the same behavior. In this
way, as has been seen for DPPC bilayers,19lipid bilayers at
high surface pressures reproduce air-water monolayer behavior.
Finally, a detailed study of the lipid/water interface was carried
out for different lipid packing of the membrane.
2. Method and Model
Periodical boundary conditions were considered along the
three-dimensional space, X, Y, and Z. The computer simulation
box was generated by placing two leaflets of 144 DMPA-(i.e.,
288 lipds in total), pointing the lipid heads toward the two water
layers placed on both sides of the computing box. Figure 1
depicts two snapshots of the system at two different surface
pressures. To maintain the electroneutrality of the system, 288
water molecules were substituted by 288 sodium ions (Na+),
so that the whole system was constituted by 288 DMPA-, 288
Na+, and 9780 SPCE water molecules,23which amounted 41436
atoms in total.
Gromacs 3.3. package24,25was used to run the MD simula-
tions, and the generated trajectories were analyzed with a
software code developed by ourselves. A dipalmitoylphosphati-
dylserine (DPPS-) modified force field16was used to carry out
these simulations. The charge distribution of a lipid DMPA-
molecule was determined by CNDO semiempirical method
implemented in HyperChem.26Thus the charge borne by each
atom of the system is depicted in Table 1, corresponding to the
atomic numeration shown in Figure 2. Long range interactions
were modeled by the Lennard-Jones potential with a cutoff of
0.8 nm, and the electrostatic interactions by the Ewald algo-
rithm.27,28All the bonds of the system were constrained by
SHAKE.29The time step used in all the simulations was 2 fs.
Simulations were carried out on a HPC160 parallel computer
using 16 processors with a performance of 1.5 ns of trajectory
length per hour.
In this regard, after setting up the system, the whole system
was subjected to a steepest descent energy minimization process
to remove undesired overlaps between neighboring atoms.
Once the system reached a certain energy minimum, four
simulations were carried out at the same normal pressure of 1
atm (z-axis) and each one to different surface pressures (in the
X-Y plane of the membrane) at 1, 6, 60, and 225 atm (which
correspond to 0.17, 1, 10, and 40 mN/m, respectively). Because
Figure 1. Snapshots of the DMPA-bilayer at two different surface
pressures: (a) 0.17mN/m and (b) 40mN/m.
Figure 2. Atom numeration used in this work for a molecule of
dimyristoylphosphatidic acid, DMPA-.
TABLE 1: Charge Density for Each Atom of Lipid DMPA-
Molecule, Na+Ions, and SPCE Water Molecules, As
Obtained by CNDO Method
atom number atom type
DMPA Bilayer on the Lipid/Water Interface
J. Phys. Chem. B, Vol. 111, No. 49, 2007 13727
a thermodynamic normal temperature and pressure (NTP) state
has been simulated, the system was coupled to an external
temperature and pressure bath with temperature coupling
constants, τT, of 0.1 ps and a pressure coupling constant, τP, of
1 ps. In both cases, the Berendsen’s algorithm was used.30
To avoid undesired simulation artifacts associated to the
starting conformation of the system, the first simulation at 0.17
mN/m was carriead out at 500 K during 500 ps. The end
conformation of this simulation was the starting point at 0.17
mN/m. Next, the system was cooled down to 350 K, and the
above-mentioned four simulations were carried out for 20 ns
where the end conformation of each trajectory was the starting
point of the next simulation at higher surface pressures. In the
four cases, the temperature of 350 K is above the DMPA-phase
transition temperature Tmof 328 K in a range of pH from 4 to
9.31The first 2 ns of trajectory length were discarded for analysis
purposes because this was the time required to reach the
equilibrium in the four cases.
3. Results and Discussion
3.1. Lipid Area. Figure 3 depicts the area evolution of lipids
as a function of time for several surface pressures. From these
curves, we observe how for a trajectory length above 2 ns that
the system reaches a steady state in the four cases. Thus, the
dimension of the equilibrated computing boxes were: (8.82,
8.68, 7.92), (8.76, 8.72, 8.04), (8.38, 8.26, 8.75), and (7.61, 7.41,
10.27), for the x-, y-, and z-axis of the computing box in nm at
0.17, 1, 10, and 40 mN/m surface pressure, respectively.
The area per DMPA-molecule obtained from our simulation
was compared with experimental data. Note that at low surface
pressure, 0.17 mN/m (or 1 atm), the simulated area was 0.54 (
obtained by Ziegler et al.,32of 0.58 nm2for its liquid crystalline
state LR. At the same surface pressure, for DMPA-air-water
monolayers, a surface area of 0.8 nm2was measured from the
isotherm surface pressure-lipid area (π-A isotherm) by Lozano
et al.33When surface pressure is increased to 40 mN/m, the
surface area per DMPA-molecule obtained by our simulations
was 0.40 ( 0.07 nm2, which agrees perfectly with the area of
0.4 nm2obtained for monolayer π-A isotherm in its LC state.33
These results are depicted in Figure 4.
The effect of the interdigitation between opposite lipid leaflets
can be observed from the density of methylene groups of lipid
tails of the two lipid leaflets, as shown in Figure 5. Thus, we
observe how an increase in the surface pressure diminishes the
methylene overlap of opposite leaflets, that is, a diminishing in
the lipid-lipid interdigitation.
2, which agrees well with the experimental data
3.2. Lipid-Lipid Lateral Interactions and Lipid Hydra-
tion. The radial distribution function g(r) provides valuable
atomic information concerning neighboring molecules. Thus,
the radial distribution function g(r) is defined as
where N(r) is the number of atoms in a spherical shell at distance
r and thickness δr from a reference atom, and F is the number
density in the computing box.
Hence, the radial distribution function of the atom numbers
4, 11, and 28 (two carbonyl and a phosphate oxygens) around
the phosphorus of a neighboring lipid DMPA-was calculated.
By integrating the first maximum of the radial distribution
function g(r), the coordination numbers of the above-mentioned
atoms (4, 11, and 28) over a phosphorus atom as a function of
the surfaces pressures were calculated and are depicted in Table
Note that an increase in the surface pressure applied produces
an increase in the coordination numbers N(r) between lipids,
which clearly reflects the importance of lateral interactions at
high values of surface pressures. Indeed, the coordination
Figure 3. Lipid running area as a function of time for different surface
Figure 4. DMPA-surface area for different surface pressures: circles
represent the simulated data, squares represent the experimental data
for a DMPA-bilayers, and triangles represent the experimental data
Figure 5. Lipid tail methylene density F in kg/m3along the z-axis for
surface pressures of 0.17 mN/m (a) and 40 mN/m (b). The origin of
the z-axis was placed in the center of the box.
13728 J. Phys. Chem. B, Vol. 111, No. 49, 2007
Casares et al.
number for atom 4 (phosphate oxygen) around a phosphorus
atom undergoes a dramatic increase of 73% from 0.17 to
40 mN/m. Less pronounced is the increase in coordination of
the carbonyl oxygens around the phosphorus atom, which is
only around 25% under analogous conditions.
The coordination numbers of water oxygens around atoms 4
and 28 of the lipid molecule (oxygen of phosphate and carbonyl
group, respectively) are depicted with their standard error in
Figure 6. From this figure, we observe that there is no significant
variation in the coordination number for pressures ranging from
0.1 to 10 mN/m. For higher pressures (40 mN/m or higher), a
dehydration of the polar lipid heads becomes clear. To quantify
this effect, the hydration number of the lipid head was calculated
from the radial distribution function of water around the heads
of the lipid molecules. It was found that 4.20, 4.19, 4.16, and
4.05 water molecules formed the first hydration shell around
phosphate oxygen for 0.17, 1, 10, and 40 mN/m, respectively.
In regard to the carbonyl oxygen tail, values of 1.12 (for 0.17
mN/m), 1.11 (for 1 mN/m), 1.09 (for 10 mN/m), and 0.85 (for
40 mN/m) were calculated. With these data, lipid hydration
(estimated as the sum of the hydration number of the phosphate
group) was calculated, reaching a value of 6.44 (for 0.1 mN/
m), 6.41 (for 0.17 mN/m), 6.34 (for 10 mN/m), and 5.75 (for
40 mN/m) water molecules. These values agree well with
experimental data provided by Schalke et al.9for DMPA-air-
water monolayers, where the number of water molecules that
coordinates the lipid head ranges from 6.2 to 4.8 for a surface
pressure ranging from 2 to 45 mN/m, respectively.
3.3. Hydrocarbon Tail Structure. The structure of the
hydrocarbon tail can be studied in NMR experiments by
measuring the deuterium order parameters along the lipid
ethylene tails. The order parameter tensor is defined as
where x, y, and z are the local coordinates of the system, θais
the angle made by the molecular axis with the bilayer normal,
and δab is the Kronecker delta. From simulation, the order
parameter -SCDcan be determined using the relation proposed
by Egberts and Berendsen34
where Sxxand Syyare the terms of the order parameter tensor of
Using eq 3, we can calculate the order parameters of a
DMPA-molecule for different surface pressures. These results
in Figure 7 can be compared with experimental data of the
DMPA-bilayer in its liquid crystalline state.20From a
comparison simulation with experimental data, we observe the
good agreement between both results at low pressures. Morover,
from the simulation we observe how the hydrocarbon structure
of the lipid tails is not perturbed at surface pressures up to 1
mN/m. Above this, increasing the surface pressures increases
the order of the lipid tails. This increase is more noticeable for
pressures above 40 mN/m, corresponding to the pressure at
which the barometric transition state of the lipids takes place
from its liquid crystalline state (LR) to gel state (L′?), which is
comparable with the liquid condensed state (LC) in lipid
Furthermore, an angle θ was defined as the angle between
the vector along the lipid chains and the perpendicular direction
to the lipid/water interface. The values of the angle θ obtained
were of θ ) 29.5, 29.7, 22.1, and 9.2° for 0.17, 1, 10, and 40
mN/m, respectively. In this regard, Schale et al.,9measured an
angle θ of 9° at 40 mN/m by X-ray, which agrees well with
our simulation results, for pressures at which bilayer and
monolayer behaviors converge. In this sense, these data confirm
the increasingly packed arrangement found in both monolayer
Figure 6. Coordination number for water molecules. Circles, carbonyl
oxygens (atom 28), and triangles, phosphate oxygen (atom 4), attending
atom numeration of Figure 2.
TABLE 2: Different Values of Coordination Number
around the Phosphorus Atoms (Atom 3) by Oxygen Atoms
(Atoms 4, 11, and 28) from Neighboring Lipids at Different
Value of Surface Pressure
Figure 7. Order parameters (-SCD) along hydrocarbon tails for 0.17,
1, 10, and 40 mN/m. The experimental results20are represented by
Sab)〈3 cos(θa)cos(θb) - δab〉
a, b ) x, y, z
DMPA Bilayer on the Lipid/Water Interface
J. Phys. Chem. B, Vol. 111, No. 49, 2007 13729
and bilayer as pressure increases and agrees well with the surface
area per lipid reported above.
Considering the lipid packaging, DMPA-monolayers are
arranged in hexagonal structure at high surface pressures. In
Figure 8, the center of mass of the ethylene lipid tails is plotted
in the X-Y plane (lipid plane). The hexagonal structure
associated with the gel structure L′?(in bilayers) and liquid
condensed state LC (in monolayers) can be clearly seen for a
pressure of 40 mN/m.
In addition, the lipid head orientation with respect to the
normal to the lipid plane was studied, obtaining values of 49.7,
48.8, 42.3, and 44.7° at 0.17, 1, 10, and 40 mN/m respectively.
These values show how the DMPA-heads in their liquid
crystalline state are much more oriented toward the water layer
than DPPC, where an orientation almost parallel to the surface
of the lipid layer has been reported.10,19This increase in head
orientation can be explained by the greater degree of head
hydration and lipid packaging of them than the DPPC ones. In
this regard, Schalke et al.9reported that the head orientation of
DMPA-in its liquid condensed state LC for a pressure of 40
mN/m was 47°, which agrees with our simulation result of 44.7°.
Unfortunately, no experimental data of DMPA-head orientation
in bilayers has been reported to compare with its liquid
crystalline state LR.
By means of FTIR, Lozano et al.33observed that in the
monolayers of pure DMPA-at high surface pressures the
transition dipole moment of the carbonyl group aligned with
the interface. The angular distribution function of the bonding
vector for this CdO group was calculated for each surface
pressure studied. Thus, from Figure 9 we can observe that for
high pressures, for example, 40 mN/m, the distribution takes a
sharper form with two maximum peaks at 90 and 97°, reflecting
the above-mentioned change in orientation of the carbonyl
moiety. This dense packing of a DMPA-monolayer at high
surface pressure is in a perfect accordance with experimental
3.4. Electrostatic Potential. The electrostatic potential ψ
across the lipids layers was computed from the double integral
of the charge density35
where the origin z of the electrostatic potential ψ(0) is taken at
the middle of the lipid bilayer. The electrostatic potential
computed in this way agrees with the computed charge density
using Poisson’s equation without using a cutoff radius.36Hence,
Figure 10 depicts the electrostatic potential averaged on both
symmetric lipid leaflets.
Of note is the dramatic effect of surface pressure on the
potential difference at the interface, which ranges from
-0.08 V at 0.17 mN/m to -0.22 V at 40 mN/m, an effect that
is of great relevance in the electrochemistry for developing
modified electrodes using Langmuir-Blodgett’s method.37
3.5. Solution Properties. Having established the lipid
structure associated to barometric transitions, the way in which
the solution properties are perturbed by the barometric isomor-
phic lipid transition was studied. For that purpose, the computa-
tion box was sliced to 19 slabs parallel to the X-Y plane (lipid
leaflet), so that slab number 10 matched the midle of lipid
3.5.1. Water and Sodium (Na+) Translational Diffusion
Coefficient, Dt. First, the diffusion coefficients of water and Na+
ions were calculated using the treatment described by Cascales
et al.,16which permitted us to estimate their diffusion coefficient
in each zone of the computing box. The results for water are
shown in Figure 11. As we can see, the regions on both sides
of the box are composed of bulk water, reaching a steady value
for the Dt,xy. Note that the diffusion coefficient of bulk water
for SPC at 349 K is 7.5 × 10-5cm2s-1,38which is very close
to our value of 6.5 × 10-5cm2s-1. Also of interest is that
there is no significative variation in the water diffusion
coefficient with respect to pressure, indicating that our conclu-
sions are not affected by any simulation artifact. From the results
depicted in Figure 11, it can be observed that there is almost
no correlation between the diffusion coefficient and surface
pressure at the lipid/water interface, where a minimum value
of around 2.5 × 10-5cm2s-1was measured in the four cases
in the vicinity of the lipid interface regardless of the isomorphic
state of the lipids.
A similar trend was observed for Na+ions, where the values
of 4.5, 4.3, 4.2, and 4.4 × 10-5cm2s-1were measured for the
bulk water and 1.2, 1.1, 1.2, and 1.2 × 10-5cm2s-1in the
vicinity of the lipid surface, respectively. These results are in
good agreement with experimental data for sodium ion in bulk
water39and in the vicinity DPPS-layers,40where values of 3.8
× 10-5cm2s-1have been reported for bulk water at 298 K
and a value of 2.1 (1.1 × 10-5cm2s-1in the vicinity of a
DPPS-bilayer at 350 K.
The first conclusion from the values reported above is that
the barometric phase transition of lipids does not perturb the
profile of sodium diffusion from the bulk water to the vicinity
of the lipid interface.
Figure 8. Lipid packing for one of the two DMPA-leaflets. Center
of mass of each lipid tail is represented by a circle. Left: surface
pressure of 0.17 mN/m. Right: surface pressure of 40 mN/m. Lines
mark the hexagonal packing of this arrangement.
ψ(z) - ψ(0) ) -1
Figure 9. Angular distribution function for the vector bonding of the
CdO moiety (atoms 27 and 28, considering the atom numeration of
Figure 2) with respect to the perpendicular axis to the lipid/water
interface and different surface pressures.
13730 J. Phys. Chem. B, Vol. 111, No. 49, 2007
Casares et al.
3.5.2. Water Rotational Relaxation Time, τ. Using the same
regions as described above, we estimated the rotational diffusion
coefficient of water molecules.16The reorientation of water
molecules is described by the correlation function
from which an apparent mean relaxation time, τapp, can be
Within this procedure, we calculated τappfrom the beginning
of the lipid/water interface to bulk water at each surface pressure
(see Table 3). As an example, the fit of the calculated data to
eq 6 for 0.17 and 40 mN/m in the central region and near the
lipid/water interface is depicted in Figure 12.
The assumption of bulk water in the midle of the aqueous
region and the absence of any simulation artifact due to surface
pressure is again confirmed by this study. The 2.11 ps of our
simulated value is in good agreement with the 1.7 ps38reported
for bulk water at 349 K. The slight difference between the results
Figure 10. Electrostatic potential across the z-axis for each surface pressure. For clarity, each region is deliniated with dotted lines. The zero volts
is placed in the middle of the water layer.
Figure 11. Translational diffusion coefficient Dt,xyof water for different
positions in the computing box along the axis perpendicular to the lipid/
water interface and different surface pressures.
aiexp(- t/τi) (5)
Figure 12. Reorientational correlation function of water dipole at 0.17
mN/m (top) and 40 mN/m (down) surface pressures. Symbols:
simulation results. Line: fitted curve.
TABLE 3: Reorientational Relaxation Time (τapp) of Water
for Different Regions of the Computing Box and Surface
DMPA Bilayer on the Lipid/Water Interface
J. Phys. Chem. B, Vol. 111, No. 49, 2007 13731
can be attributed to the fact that in our simulations some water
molecules that were hydrating sodium ions were considered
rather than the pure water in Postma’s work.38In regard to the
rotational relaxation time in the vicinity of the lipid surface, it
increased up to 11.3 (1.1 ps independently of the surface
pressure. In other words, we can assert that lipid barometric
phase transition did not perturb the dynamic properties of water
from bulk to water/lipid interface in the range of pressures
3.5.3. Sodium Ion Hydration. Using the radial distribution
function of eq 1, we are able to calculate the hydration number
of sodium ions in solution. As it is shown in Figure 13, a value
of 5.0 water molecules was estimated in bulk water and 1.0
water molecule near the lipid/water interface. The results for
bulk water agree well with experimental41and simulation42data
in aqueous solution, where values of 4.6 and 3.8 have been
reported at 298 K. At the lipid/water interface, Cascales et al.
reported40a value of 1.2 in the vicinity of a DPPS-bilayer,
which agrees with the value of 1.1 obtained here for all the
simulated surface pressures.
Four DMPA-bilayers were subjected to different lateral
pressures in a range from 0.17 to 40 mN/m. The results
demonstrated that the lipid bilayer and monolayer follow the
same behavior only when surface pressure is sufficiently high.
On the other hand, their behavior differed with low pressures.
Validation of the DMPA-force field was made by compari-
son with experimental data for both bilayer and monolayer. The
experimental data corresponding to hydrocarbon order param-
eters and surface area per lipid molecule showed excellent
agreement with our simulation results at low surface pressures,
which reproduced the DMPA-bilayer behavior. At high surface
pressures, the area per lipid molecule, orientation of the carbonyl
moiety, tail orientation, head orientation, and hexagonal ar-
rangement of the lipid tails reflected monolayer behavior in the
liquid condensed (LC) phase.
In addition, we focused our interest on how this barometric
polymorphic phase transition affects the lipid/solution interface.
Thus, the water reorientational time, the water and sodium
translational diffusion coefficient, and sodium hydration were
studied. From the simulation results, we can assert that
barometric polymorphic lipid transition does not perturb the
lipid/water interface properties in the range of the pressures
Acknowledgment. J.J.L.C. wishes to thank the Spanish
Government (Ministerio de Educacion y Ciencia, MEC) and
Fundacion Seneca for his financial support through projects
BQU-2001-04777 and 00483/PI/04, respectively. L.C., M.T.M.R.,
and J.J.G.C. thank the Spanish CICYT for financial support of
this research in the framework of Project Nos. CTQ2004-03246/
BQU and MAT2004-03849. Also, J.J.G.C. thanks the Ministerio
de Educacion y Ciencia for a Formacion Profesorado Univer-
sitario (FPU) predoctoral fellowship.
References and Notes
(1) Cullis, P.; Hope, M.; de Kruijff, B.; Verkleij, A.; Tilcock, C. In
Phospholipids and cellular regulation; Kuo, J., Ed.; CRC Press: Boca
Raton, FL, 1985; pp 1-60.
(2) Yeagle, P. FASEB J. 1989, 3 (4), 1833-1842.
(3) Nagle, J.; Tristram-Nagle, S. Biochim. Biophys. Acta 2000, 1469
(4) Yeagle, P. L. In The structure of biological membranes; Yeagle,
P., Ed.; CRC Press: Boca Raton, FL, 1992; pp 73-156.
(5) Ichimori, H.; Hata, T.; Matsuki, H.; Kaneshina, S. Biochim. Biophys.
Acta 1998, 1414 (1-2), 165-174.
(6) R. Winter. Curr. Opin. Colloid Interface Sci. 2001, 6 (3), 303-
(7) Mo ¨hwald, H. In Handbook of biological physics; Lipowsky, R.,
Sackmann, E., Eds.; Elsevier Science B. V.: New York, 1995; pp 161-
(8) Hunt, R.; Mitchell, M.; Dluhy, R. J. Mol. Struct. 1989, (214), 93-
(9) Schalke, M.; Kru ¨ger, P.; Weygand, M.; Lo ¨sche, M. Biochim.
Biophys. Acta 2000, 1464, 113-126.
(10) Dluhy, R. Appl. Spectrosc. ReV. 2000, 35 (4), 315-351.
(11) Schalke, M.; Lo ¨sche, M. AdV. Colloid Interface Sci. 2000, 88 (1-
(12) Mendelsohn, R.; Flach, C. Curr. Top. Membr. 2002, (52), 57-88.
(13) Pedrosa, J.; Perez, M.; Prieto, I.; Martin-Romero, M.; Mo ¨bius, D.;
Camacho, L. Phys. Chem. Chem. Phys. 2002, 4 (11), 2329-2336.
(14) van Gunsteren, W. F.; Berendsen, H. J. C. Angew. Chem., Int. Ed.
Engl. 1990, 29 (9), 992-1023.
(15) Alper, H. E.; Bassolino-Klimas, D.; Stouch, T. R. J. Chem. Phys.
1993, 99 (7), 5554-5559.
(16) Cascales, J.; Berendsen, H.; Garcia de la Torre, J. J. Phys. Chem.
1996, 100 (21), 8621-8627.
(17) Tieleman, D.; Marrink, S.; Berendsen, H. Biochim. Biophys. Acta
1997, 1331 (3), 235-270.
(18) Kaznessis, Y.; Kim, S.; Larson, R. Biophys. J. 2002, 82 (4), 1731-
(19) Cascales, J. L.; Otero, T.; Romero, A. F.; Camacho, L. Langmuir
2006, 22 (13), 5818-5824.
(20) Pott, T.; Maillet, J.; Dufourc, E. Biophys. J. 1995, 69 (5), 1897-
(21) Wu, F.; Gericke, A.; Flach, C.; Mealy, T.; Seaton, B.; Mendelsohn,
R. Biophys. J. 1998, 74, 3273-3281.
(22) Mendelson, R.; Flach, C. Curr. Top. Membr. 2002, 52, 57-78.
(23) Berendsen, H.; Grigera, J.; Straatsma, T. J. Phys. Chem. 1987, 91
(24) Lindahl, E.; Hess, B.; van der Spoel, D. J. Mol. Model. 2001, 7
(25) Berendsen, H.; van der Spoel, D.; van Drunen, R. Comp. Phys.
Comm. 1995, 91 (1-3), 43-56.
(26) Hyperchem 7.5. HyperCube, Inc.
(27) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98 (12),
(28) Essmann, U.; Perea, L.; Berkowitz, M.; Darden, T.; Lee, H.;
Pedersen, L. J. Chem. Phys. 1995, 103 (19), 8577-8593.
(29) van Gunsteren, W.; Berendsen, H. Mol. Phys. 1977, 34 (5), 1311-
(30) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola,
A.; Haak, J. R. J. Chem. Phys. 1984, 8 (8), 3684-3690.
(31) Blume, A.; Eibl, H. Biochim. Biophys. Acta 1979, 558(1), 476-
(32) Ziegler, W.; Blume, A. Spectrochim. Acta, Part A 1995, 51, 1763-
(33) Lozano, P.; Fernandez, A.; Ruiz, J.; Camacho, L.; Martin, M.;
Mun ˜oz, E. J. Phys. Chem. B 2002, 106 (25), 6507-6514.
surrounding the Na+ions for different surface pressures and regions
of the computing box.
Values of coordination number for water molecules
13732 J. Phys. Chem. B, Vol. 111, No. 49, 2007
Casares et al.
(34) Egberts, E.; Berendsen, H. J. Chem. Phys. 1988, 89 (6), 3718-
(35) van Buuren, A.; Berendsen, H. Langmuir 1994, 10 (6), 1703-1713.
(36) Pandit, S.; Bostick, D.; Berkowitz, M. Biophys. J. 2003, 84 (6),
(37) Goldenberg, L. J. Electroanal. Chem. 1994, 379 (1-2), 3-19.
(38) Postma, J. A molecular dynamics study of water. Ph.D. Thesis,
Rijkuniverseit Groningen, The Netherlands, 1985.
(39) Lide, D., Ed. Handbook of Chemistry and Physics; CRC Press:
Boca Raton, FL, 2002.
(40) Cascales, J. L.; de la Torre, J. G. Biochim. Biophys. Acta 1997,
1330 (2), 145-156.
(41) Bockris, J.; Reddy, A. Modern Electrochemistry 1. Ionics; Plenum
Press: New York, 1998.
(42) Impey, R.; Madden, P.; McDonald, I. J. Phys. Chem. 1983, 87 (25),
DMPA Bilayer on the Lipid/Water Interface
J. Phys. Chem. B, Vol. 111, No. 49, 2007 13733