National Academy of Sciences of RA ÐÐ ¶ÇïáõÃÛáõÝÝ»ñÇ ²½·³ÛÇÝ ²Ï³¹»Ùdz Национальная Академия наук РА

Electronic Journal of ¾É»ÏïñáݳÛÇÝ ³Ùë³·Çñ Электронный журнал

NATURAL SCIENCES ´Ü²Î²Ü ¶ÆîàôÂÚàôÜܺð ЕСТЕСТВЕННЫЕ НАУКИ

2(15), 2010
Bioinformatics Բիոինֆորմատիկա Бионформатика

INTERAC TION OF L-A LAN INE AND L-VA LIN E AMINOACID MOLECU LES
W I TH PHO SPH O LI PI D B ILA Y E R .
MOLECU LAR DYNAMICS SIMU LA TION

S. A. Aramyan 1, A. H. Poghosyan 2,
Cor. Member of NAS RA A. A. Shahinyan 1
1
Institute of Applied Problems of Physics of National Academy of Sciences of Armenia,
Hr. Nersisian 25, 0014, Yerevan, Republic of Armenia; Tel.: (+374 10) 245 896; e-mail: sona_aramyan@yahoo.com;
2
International Scientific and Educational Center of National Academy of Sciences of Armenia,
Marshal Baghramian Ave. 24d, 0019 Yerevan, Republic of Armenia; Tel.: (+374 10) 583 691

ABSTRACT
It is well known the importance of investigations of interactions between proteins and phospholipid bilayers. We decided to give a
deeper sight to this problem by the determination of ways of interactions between aminoacids and phospholipid bilayers using
computer simulation methods. By means of simulation, remarkable progress is being made in the understanding of protein folding,
lipid bilayer behavior and peptide-lipid interactions. There were used the molecules of L-alanine and L-valine and phospholipids
bilayer consisting of 128 Dipalmitoylphosphatidylcholine (dppc) molecules and 3655 water molecules, which corresponds to fully
hydrated state of dppc bilayer. For the computer experiment the GROMACS simulation package and GROMOS force fields were
used. There were done two main simulations of systems containing L-ala and bilayer and correspondingly L-val and bilayer. Both
simulations were done with the duration of 15 ns and under the same conditions with same parameters.
We have got that areas per lipid and bilayer thicknesses of aminoacid neighboring part of bilayers are decreasing, correspondingly
from 0.62 nm2 to 0.4 nm2 and from 3.7 nm to 3.5 nm. Aminoacid centers of masses are penetrating into the membrane from thaier
starting point by approximately 0.4 nm. And at last we have determined the possible aminoacid orientation in yhe bilayer. Thus
aminoacids are orientating by their nitrogen atom down to lipid’s negative phosphorus atoms and by their carboxyl carbon atom
upwards to lipid’s positively charged nitrogen atoms.

Key words: MD simulations, DPPC, aminoacid

INTRODUCTION
During last decades the accumulation of biological information is continuing, besides this during last years it is
accumulating in forms of electronic versions, such as texts, images and so on. Because of the presence of this huge
amount of information is rising a problem of systematization and adequate usage of this amount of experimental data.
Bioinformatical methods give an appropriate salvation to these problems. Besides systematization, during last decade, it
gives an opportunity to make models and investigate complex biological and chemical molecules and processes.
Many experimental studies on hydrated phospholipid bilayers using x-ray diffraction have yielded important
information about bilayer structure [1]. However, due to the fluid disorder present in such systems, these experiments
provide only one-dimensional information about membrane structure in the direction normal to the of the membrane.
Molecular dynamics (MD) simulations, on the other hand, inherently provide three-dimensional structural and dynamic
information about bilayers [2]. MD simulation is a bioinformatical method to investigate molecular systems of biologi-
cal relevance. The method is well suited to investigations of biomembranes’ dynamical structure, because it is characte-
rized by atomic resolution and makes possible the observations of the processes on the time scale of 10-8-10-7 s with a
time resolution of a few femtoseconds. To carry out MD simulations of a molecular system, relevant atom-atom inter-
action have to be defined. Of all inter-atomic interactions that are commonly included in the potential function of the
system the electrostatic ones have the largest contribution at longer distances. A correct treatment of long-ranged
electrostatic interactions is essential for obtaining meaningful simulation results. For this propose there are several
methods of short-rang and long-range interaction treatments [3].
The aim of this work was to determine the ways of interaction of aminoacids with phospholipid bilayers using computer
simulation methods. Particularly for these simulations the method of molecular dynamics was used. It was interesting to
find out in details the ways of aminoacid-phospholipids bilayer interactions in atomic scale. There were used molecules
of L-alanine and L-valine with the bilayer consisting of 128 dipalmitoilphosphatidilcholine (dppc) and 3655 water
molecules. In this experiment there are elucidated many nuances of aminoacid behavior and bilayer property changes
during aminoacid addition to bilayer.

3
 METHODS
The molecular dynamics simulations were carried out using GROMACS simulation package (version 3.2.1, single
precision) [4]. Simulations were performed on Armcluster, which was founded in Institute of Informatics and
Automatization Problems of NAS RA by ISTC A-823 project. The VMD package [5] was used for visualization and
molecular graphics.
The interactions between, L-valine/bilayer, L-alanine/bilayer were examined. There were investigated two systems
consisting of L-alanine, and L-valine molecules and phospholipid bilayers (128 DPPC and 3655 water molecules, which
corresponds to fully hydrated state of phospholipid bilayer) (Fig. 1). For the system preparation water was removed
from bilayer and then the aminoacid molecule was added right onto the bilayer surface. After aminoacid addition water
was palced back onto its starting configuration.

Fig. 1. Starting (first column) and final (second column) structures of bilayers correspondingly with L-alanine and L-valine
(water molecules are not shown for figure simplicity).

The starting configurations of DPPC bilayers were taken from Tieleman’s group website
http://moose.bio.ucalgary.ca/Downloads. Force fields and topologies made available by the Tieleman group at
http://moose.bio.ucalgary.ca/Downloads were also used. For DPPC were used parameters from Berger et al. [6].
The flexiable simple point charge (SPC) water model was used for solvent [7].
All simulations were made at same conditions with same parameters. The integration time step was chosen as 2 fs. All
bonds were constrained to fixed length during position updates via the SHAKE algorithm [8]. The long-ranged
Coulomb interactions were cut off to 2.5 nm and the short-range van der Waals interactions were cut off to 1.4 nm.
Simulations were done under NPT ensemble.To keep temperature stable around the room temperature (303 K) Nose-
Hoover coupling algorithm was used [9, 10].
For keeping pressure stable Parrinello-Rahman pressure coupling algorithm and isotropic coupling type were used [11,
12].
At first there were done energy minimizations on both systems. After getting the appropriate geometrical configuration,
molecular dynamics simulations were done on systems with duration of 15 ns each.
Several parameters were measured and investigated for understanding the bilayer parameter changes and aminoacid
behavior in the membrane. Data processing and analyzing were done using already existing GROMACS analyzing tools
[4, 13].

4
 RESULTS
Bilayer properties
For the description of bilayer properties there were mea-
sured some parameters such as area per lipid and bilayer
thickness. In most of cases measurements were done on
the nearest to aminoacid part of bilayer. On Fig. 2 you
can see changes of area per lipid (on the aminoacid
neighboring part of bilayer) during simulation time.
From the real physical experiment this value was received
as 0.62 nm2 [14]. As you can see this value is decreasing
up to approximately 0.4 nm2. This means that lipids are
getting closer to each other after aminoacid addition. On
Fig. 3 are presented the changes of bilayer thickness mea- Fig. 2. Area per lipid on the aminoacid neighboring part of
bilayers with L-alanine and L-valine.
sured as a distance between aminoacid neighboring lipid
headgroup phosphorus of upper and lower layers.
This value from real experiment was 3.7 nm but you can
see that thickness is decreasing too during simulation to
~3.5 nm [14].
There were also measured total energies of bilayers (Fig.
4). From this value we can see whether system is stable or
not. As you can see after 1.5 ns system stabilizes and
there are little changes in energy after this time.

Aminoacid behavior
In Fig. 5 you can see the permeation of aminoacid into the
bilayer. During simulation aminoacids are entering to
Fig. 3. Thickness of aminoacid neighboring part of bilayers
correspondingly with L-alanine and L-valine.

Fig. 5. Aminoacids center of mass distance from neighboring Fig. 4. Total energies of systems with L-alanine and L-valine.
upper layer phosphors.

bilayer from their starting point until phosphorus level (7.5 ns) and then even go deeper then average phosphors. From
figure it is clear that at the end of simulation, starting from 12 ns till end, L-valine penetrates deeper then L-alanine,
which can be explained by L-valines more hydrophobicity compared with L-alanine.
For the determination of aminoacid preferable orientation in bilayer there were measured and compared aminoacid’s
nitrogen and carboxyl group carbon distances from neighboring upper layer phosphors. The results of these measure-
ments you can see on Fig. 6. In both cases carboxyl groups carbon is placed upper in the bilayer than nitrogen.
Also it was calculated the angle between bilayer normal z axis and aminoacid carboxyl carbon-nitrogen vector.

Fig. 6. L-alanine’s and L-valine’s nitrogen and carboxyl group carbon distances from neighboring upper layer phosphors.

5
This angle is blunt during simulation time (Fig. 7). From these data we can present the possible orientation of aminocids
in the bilayer (Fig. 8).

Fig. 7. Angle between z axis and aminoacid’s carboxyl

Fig. 8. Possible orientation of aminoacids in bilayer.
carbon-nitrogen vector.

In Fig. 9 and 10 are presented L-alanine and L-valine center of mass diffusion on x y plane. From these figures we can
see that during 15 ns of simulation aminoacids are moving in the 1.5 nm2 area in the x y plane.

Fig. 9. L-alanine center of mass diffusion on x y plane. Fig. 10. L-valine center of mass diffusion on x y plane.

For full understanding the aminoacid diffusion, aminoacid

center of mass diffusion across z axis was analysed (Fig.
11).
There were compared aminoacid diffusion by z axis and
thickness of aminoacid neighboring part of bilayer (Fig.
12, 13). We have got very interesting relationship between
aminoasid transversal diffusion and bilayer thickness
behavior. As you can see when aminoacid penetrates into
the bilayer, thickness of its neighboring part of bilayer is
decreasing and vice versa.
Fig. 11. Aminoacids center of mass diffusion across z axis.

Fig. 12. L-alanine center of mass diffusion across z axis Fig. 13. L-valine center of mass diffusion across z axis and
and thickness of aminoacid neighboring part of bilayer. thickness of aminoacid neighboring part of bilayer.

6
 CONCLUSIONS
In conclusion we can say that we have modeled for the first time the above mentioned aspect of interaction between
phospholipid bilayer and aminoacids.
Several parameters, describing the structure of the bilayer, such as area per lipid on the surface of bilayer, thickness of
bilayer, arrangement of aminoacid molecules in the bilayer were investigated.
We have got that areas per lipid for the aminoacid neighboring part of bilayer are decreasing from 0.62 to 0.4 nm2. And
the bilayers thicknesses again for the aminoacid neighboring part of bilayers are decreasing from 3.7 to 3.5 nm.
Aminoacids center of masses are penetrating into the bilayer deeper then lipids average phosphorus for 0.3 nm,
furthermore L-valine is penetrating deeper than L-alanine.
Besides these all we have suggested the possible orientation of aminoacids in the bilayer. For this purpose we have
examined aminoacid’s nitrogen and carboxyl group carbon distances from neighboring upper layer phosphors. From
this data you can see that in both cases carboxyl groups carbon is placed upper in the bilayer than nitrogen. We also
have calculated the angle between bilayer normal z axis and aminoacid carboxyl carbon-nitrogen vector. As this angle
is blunt we can assume the most possible aminoacid orientation in the bilayer.
Next aspect was the aminoacid diffusion on x y plane and by z axis. We have shown that on x y plane during 15 ns
aminoacids’ center of masses are moving in the 1.5 nm2 area and across z axis by ~0.5 nm.
At the end we have shown very interesting relationship between aminoacid diffusion by z axis and thickness of amino-
acid neighboring part of bilayer. As you have already seen when aminoacid penetrates into the bilayer, thickness of its
neighboring part of bilayer is decreasing and vice versa.
So we can say that areas per lipid and bilayer thicknesses for the aminoacid neighboring part of bilayers are decreasing,
aminoacids center of masses are penetrating into the bilayer deeper then lipids average phosphorus for 0.3 nm, further-
more L-valine is penetrating deeper than L-alanine and finally we have suggested the possible orientation of aminoacids
in the bilayer.

REFERENCES
[1] Sun W., Suter S., Tristram-Nagle S. et al., 1994. Order and disorder in fully hydrated unoiiented bilayers of gel phase
dipalmitoylphosphatidylcholine / Phys. Rev. E, v. 49, №5: 4665-4676.
[2] Benz R.W., Castro-Roman F., Tobias D. J., White S.H., 2005. Experimental validation of molecular dynamics simulations of
lipid bilayers: a new approach / Biophysical J., v. 88: 805-817.
[3] Rog T., Murzyn K., Pasenkiewicz-Gierula M., 2003. Molecular dynamics simulations of charged and neutral lipid bilayers:
treatment of electrostatic interactions / Acta Biophisica Polonica, v. 50, №3: 789-798.
[4] Gromacs 3.1.3 or 3.0
[5] Humphrey W., Dalke A., Schulten K., 1996. VMD - Visual Molecular Dynamics / J. Molec. Graphics, 14.1: 33-38.
[6] Berger O., Edholm O., Jahnig F., 1997. Molecular dynamics simulations of a fluid bilayer of dipalmitoylphosphatidylcholine at
full hydration, constant pressure, and constant temperature / Biophys. J., 72: 2002–2013.
[7] Berendsen H.J.C., Postma J.P.M., van Gunsteren W.F., Hermans J., 1981. Interaction models for water in relation to protein
hydration. In Intermolecular Forces / B. Pullman, ed. D. Reidel Publishing Company, Dordrecht, The Netherlands: 331-342.
[8] Ryckaert J.P., Ciccotti G., Berendsen H.J.C., 1977. Numerical integration of the Cartesian equations of motion of a system
with constraints; molecular dynamics of n-alkanes / J. Comp. Phys, 23: 327–341.
[9] Nos´e S., 1984. A molecular dynamics method for simulations in the canonical ensemble / Mol. Phys, 52: 255–268.
[10] Hoover W.G., 1985. Canonical dynamics: equilibrium phase-space distributions / Phys. Rev., A 31: 1695–1697.
[11] Parrinello M., Rahman A., 1981. Polymorphic transitions in single crystals: A new molecular dynamics method / J. Appl. Phys.
52: 7182–7190.
[12] Nos´e S., Klein M.L., 1983. Constant pressure molecular dynamics for molecular systems / Mol. Phys., 50: 1055–1076.
[13] Gromacs manual
[14] Tieleman D.P., Berendsen H.J.C., 1996. Molecular dynamics simulations of a fully hydrated dipalmitoylphosphatidylcholine
bilayer with different macroscopic boundary conditions and parameters / J. Chem. Phys. 105: 4871–4880.

7
Copyright of Electronic Journal of Natural Sciences is the property of Presidium of National Academy of
Sciences, Republic of Armenia and its content may not be copied or emailed to multiple sites or posted to a
listserv without the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.