| 1 |
\documentclass[11pt]{article} |
| 2 |
|
| 3 |
\usepackage{graphicx} |
| 4 |
\usepackage{color} |
| 5 |
\usepackage{floatflt} |
| 6 |
\usepackage{amsmath} |
| 7 |
\usepackage{amssymb} |
| 8 |
\usepackage[ref]{overcite} |
| 9 |
|
| 10 |
|
| 11 |
|
| 12 |
\pagestyle{plain} |
| 13 |
\pagenumbering{arabic} |
| 14 |
\oddsidemargin 0.0cm \evensidemargin 0.0cm |
| 15 |
\topmargin -21pt \headsep 10pt |
| 16 |
\textheight 9.0in \textwidth 6.5in |
| 17 |
\brokenpenalty=10000 |
| 18 |
\renewcommand{\baselinestretch}{1.2} |
| 19 |
\renewcommand\citemid{\ } % no comma in optional reference note |
| 20 |
|
| 21 |
|
| 22 |
\begin{document} |
| 23 |
|
| 24 |
|
| 25 |
\title{A Mesoscale Model for Phospholipid Simulations} |
| 26 |
|
| 27 |
\author{Matthew A. Meineke\\ |
| 28 |
Department of Chemistry and Biochemistry\\ |
| 29 |
University of Notre Dame\\ |
| 30 |
Notre Dame, Indiana 46556} |
| 31 |
|
| 32 |
\date{\today} |
| 33 |
\maketitle |
| 34 |
|
| 35 |
\section{Background and Research Goals} |
| 36 |
|
| 37 |
Simulations of phospholipid bilayers are, by necessity, quite |
| 38 |
complex. The lipid molecules are large molecules containing many |
| 39 |
atoms,and the head group of the lipid will typically contain charge |
| 40 |
separated ions which set up a large dipole within the molecule. Adding |
| 41 |
to the complexity are the number of water molecules needed to properly |
| 42 |
solvate the lipid bilayer. Because of these factors, many current |
| 43 |
simulations are limited in both length and time scale due to to the |
| 44 |
sheer number of calculations performed at every time step and the |
| 45 |
lifetime of the researcher. A typical |
| 46 |
simulation\cite{saiz02,lindahl00,venable00,Marrink01} will have around |
| 47 |
64 phospholipids forming a bilayer approximately 40~$\mbox{\AA}$ by |
| 48 |
50~$\mbox{\AA}$ with roughly 25 waters for every lipid. This means |
| 49 |
there are on the order of 8,000 atoms needed to simulate these systems |
| 50 |
and the trajectories in turn are integrated for times up to 10 ns. |
| 51 |
|
| 52 |
These limitations make it difficult to study certain biologically |
| 53 |
interesting phenomena that don't fit within the short time and length |
| 54 |
scale requirements. One such phenomena is the existence of the ripple |
| 55 |
phase ($P_{\beta'}$) of the bilayer between the gel phase |
| 56 |
($L_{\beta'}$) and the fluid phase ($L_{\alpha}$). The $P_{\beta'}$ |
| 57 |
phase has been shown to have a ripple period of |
| 58 |
100-200~$\mbox{\AA}$.\cite{katsaras00,sengupta00} A simulation of this |
| 59 |
length scale would require approximately 1,300 lipid molecules and |
| 60 |
roughly 25 waters for every lipid to fully solvate the bilayer. With |
| 61 |
the large number of atoms involved in a simulation of this magnitude, |
| 62 |
steps \emph{must} be taken to simplify the system to the point where |
| 63 |
these numbers are reasonable. |
| 64 |
|
| 65 |
Another system of interest would be drug molecule diffusion through |
| 66 |
the membrane. Due to the fluid like properties of a lipid membrane, |
| 67 |
not all diffusion takes place at membrane channels. It is of interest |
| 68 |
to study certain molecules that may incorporate themselves directly |
| 69 |
into the membrane. These molecules may then have an appreciable |
| 70 |
waiting time (on the order of nanoseconds) within the |
| 71 |
bilayer. Simulation of such a long time scale again requires |
| 72 |
simplification of the system in order to lower the number of |
| 73 |
calculations needed at each time step. |
| 74 |
|
| 75 |
|
| 76 |
\section{Methodology} |
| 77 |
|
| 78 |
\subsection{Length and Time Scale Simplifications} |
| 79 |
|
| 80 |
The length scale simplifications are aimed at increasing the number of |
| 81 |
molecules that can be simulated without drastically increasing the |
| 82 |
computational cost of the system. This is done by a combination of |
| 83 |
substituting less expensive interactions for expensive ones and |
| 84 |
decreasing the number of interaction sites per molecule. Namely, |
| 85 |
point charge distributions are replaced with dipoles, and unified atoms are |
| 86 |
used in place of water, phospholipid head groups, and alkyl groups. |
| 87 |
|
| 88 |
The replacement of charge distributions with dipoles allows us to |
| 89 |
replace an interaction that has a relatively long range, |
| 90 |
$(\frac{1}{r})$, for the coulomb potential, with that of a relatively |
| 91 |
short range, $(\frac{1}{r^{3}})$, for dipole - dipole |
| 92 |
potentials. Combined with a computational simplification algorithm |
| 93 |
such as a Verlet neighbor list,\cite{allen87:csl} this should give |
| 94 |
computational scaling by $N$. This is in comparison to the Ewald |
| 95 |
sum\cite{leach01:mm} needed to compute the coulomb interactions, which |
| 96 |
scales at best by $N \ln N$. |
| 97 |
|
| 98 |
The second step taken to simplify the number of calculations is to |
| 99 |
incorporate unified models for groups of atoms. In the case of water, |
| 100 |
we use the soft sticky dipole (SSD) model developed by |
| 101 |
Ichiye\cite{Liu96} (Section~\ref{sec:ssdModel}). For the phospholipids, a |
| 102 |
unified head atom with a dipole will replace the atoms in the head |
| 103 |
group, while unified $\text{CH}_2$ and $\text{CH}_3$ atoms will |
| 104 |
replace the alkyl units in the tails (Section~\ref{sec:lipidModel}). |
| 105 |
|
| 106 |
The time scale simplifications are introduced so that we can take |
| 107 |
longer time steps. By increasing the size of the time steps taken by |
| 108 |
the simulation, we are able to integrate the simulation trajectory |
| 109 |
with fewer calculations. However, care must be taken that any |
| 110 |
simplifications used, still conserve the total energy of the |
| 111 |
simulation. In practice, this means taking steps small enough to |
| 112 |
resolve all motion in the system without accidently moving an object |
| 113 |
too far along a repulsive energy surface before it feels the effect of |
| 114 |
the surface. |
| 115 |
|
| 116 |
In our simulation we have chosen to constrain all bonds to be of fixed |
| 117 |
length. This means the bonds are no longer allowed to vibrate about |
| 118 |
their equilibrium positions. Bond vibrations are typically the fastest |
| 119 |
periodic motion in a dynamics simulation. By taking this action, we |
| 120 |
are able to take time steps of 3 fs while still maintaining constant |
| 121 |
energy. This is in contrast to the 1 fs time step typically needed to |
| 122 |
conserve energy when bonds lengths are allowed to oscillate. |
| 123 |
|
| 124 |
\subsection{The Soft Sticky Water Model} |
| 125 |
\label{sec:ssdModel} |
| 126 |
|
| 127 |
%\begin{floatingfigure}{55mm} |
| 128 |
%\includegraphics[width=45mm]{ssd.epsi} |
| 129 |
%\caption{The SSD model with the oxygen and hydrogen atoms drawn in for reference. \vspace{5mm}} |
| 130 |
%\label{fig:ssdModel} |
| 131 |
%\end{floatingfigure} |
| 132 |
|
| 133 |
The water model used in our simulations is a modified soft |
| 134 |
Stockmayer-sphere model.\cite{stevens95} Like the Stockmayer-sphere, the SSD |
| 135 |
model\cite{Liu96} consists of a Lennard-Jones interaction site and a |
| 136 |
dipole both located at the water's center of mass (Figure |
| 137 |
\ref{fig:ssdModel}). However, the SSD model extends this by adding a |
| 138 |
tetrahedral potential to correct for hydrogen bonding. |
| 139 |
|
| 140 |
The SSD water potential is then given by the following equation: |
| 141 |
\begin{equation} |
| 142 |
V_{\text{SSD}} = V_{\text{LJ}}(r_{i\!j}) + V_{\text{dp}}(\mathbf{r}_{i\!j}, |
| 143 |
\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j}) |
| 144 |
+ V_{\text{sp}}(\mathbf{r}_{i\!j},\boldsymbol{\Omega}_{i}, |
| 145 |
\boldsymbol{\Omega}_{j}) |
| 146 |
\label{eq:ssdTotPot} |
| 147 |
\end{equation} |
| 148 |
$V_{\text{LJ}}$ is the Lennard-Jones potential: |
| 149 |
\begin{equation} |
| 150 |
V_{\text{LJ}} = |
| 151 |
4\epsilon_{ij} \biggl[ |
| 152 |
\biggl(\frac{\sigma_{ij}}{r_{ij}}\biggr)^{12} |
| 153 |
- \biggl(\frac{\sigma_{ij}}{r_{ij}}\biggr)^{6} |
| 154 |
\biggr] |
| 155 |
\label{eq:lennardJonesPot} |
| 156 |
\end{equation} |
| 157 |
where $r_{ij}$ is the distance between two $ij$ pairs, $\sigma_{ij}$ |
| 158 |
scales the length of the iteraction, and $\epsilon_{ij}$ scales the |
| 159 |
energy of the potential. For SSD, $\sigma_{\text{SSD}} = 3.051 \mbox{ |
| 160 |
\AA}$ and $\epsilon_{\text{SSD}} = 0.152\text{ kcal/mol}$. |
| 161 |
$V_{\text{dp}}$ is the dipole potential: |
| 162 |
\begin{equation} |
| 163 |
V_{\text{dp}}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i}, |
| 164 |
\boldsymbol{\Omega}_{j}) = \frac{1}{4\pi\epsilon_{0}} \biggl[ |
| 165 |
\frac{\boldsymbol{\mu}_{i} \cdot \boldsymbol{\mu}_{j}}{r^{3}_{ij}} |
| 166 |
- |
| 167 |
\frac{3(\boldsymbol{\mu}_i \cdot \mathbf{r}_{ij}) % |
| 168 |
(\boldsymbol{\mu}_j \cdot \mathbf{r}_{ij}) } |
| 169 |
{r^{5}_{ij}} \biggr] |
| 170 |
\label{eq:dipolePot} |
| 171 |
\end{equation} |
| 172 |
where $\mathbf{r}_{ij}$ is the vector between $i$ and $j$, |
| 173 |
$\boldsymbol{\Omega}$ is the orientation of the species, and |
| 174 |
$\boldsymbol{\mu}$ is the dipole vector. The SSD model specifies a dipole |
| 175 |
magnitude of 2.35~D for water. |
| 176 |
|
| 177 |
The hydrogen bonding of the model is governed by the $V_{\text{sp}}$ |
| 178 |
term of the potential. Its form is as follows: |
| 179 |
\begin{equation} |
| 180 |
V_{\text{sp}}(\mathbf{r}_{i\!j},\boldsymbol{\Omega}_{i}, |
| 181 |
\boldsymbol{\Omega}_{j}) = |
| 182 |
v^{\circ}[s(r_{ij})w_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i}, |
| 183 |
\boldsymbol{\Omega}_{j}) |
| 184 |
+ |
| 185 |
s'(r_{ij})w^{x}_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i}, |
| 186 |
\boldsymbol{\Omega}_{j})] |
| 187 |
\label{eq:spPot} |
| 188 |
\end{equation} |
| 189 |
Where $v^\circ$ scales strength of the interaction. |
| 190 |
$w_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j})$ |
| 191 |
and |
| 192 |
$w^{x}_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j})$ |
| 193 |
are responsible for the tetrahedral potential and a correction to the |
| 194 |
tetrahedral potential respectively. They are, |
| 195 |
\begin{equation} |
| 196 |
w_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j}) = |
| 197 |
\sin\theta_{ij} \sin 2\theta_{ij} \cos 2\phi_{ij} |
| 198 |
+ \sin \theta_{ji} \sin 2\theta_{ji} \cos 2\phi_{ji} |
| 199 |
\label{eq:spPot2} |
| 200 |
\end{equation} |
| 201 |
and |
| 202 |
\begin{equation} |
| 203 |
\begin{split} |
| 204 |
w^{x}_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j}) = |
| 205 |
&(\cos\theta_{ij}-0.6)^2(\cos\theta_{ij} + 0.8)^2 \\ |
| 206 |
&+ (\cos\theta_{ji}-0.6)^2(\cos\theta_{ji} + 0.8)^2 - 2w^{\circ} |
| 207 |
\end{split} |
| 208 |
\label{eq:spCorrection} |
| 209 |
\end{equation} |
| 210 |
The angles $\theta_{ij}$ and $\phi_{ij}$ are defined by the spherical |
| 211 |
polar coordinates of the position of sphere $j$ in the reference frame |
| 212 |
fixed on sphere $i$ with the z-axis aligned with the dipole moment. |
| 213 |
The correction |
| 214 |
$w^{x}_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j})$ |
| 215 |
is needed because |
| 216 |
$w_{ij}(\mathbf{r}_{ij},\boldsymbol{\Omega}_{i},\boldsymbol{\Omega}_{j})$ |
| 217 |
vanishes when $\theta_{ij}$ is $0^\circ$ or $180^\circ$. |
| 218 |
|
| 219 |
Finally, the sticky potential is scaled by a cutoff function, |
| 220 |
$s(r_{ij})$ that scales smoothly between 0 and 1. It is represented |
| 221 |
by: |
| 222 |
\begin{equation} |
| 223 |
s(r_{ij}) = |
| 224 |
\begin{cases} |
| 225 |
1& \text{if $r_{ij} < r_{L}$}, \\ |
| 226 |
\frac{(r_{U} - r_{ij})^2 (r_{U} + 2r_{ij} |
| 227 |
- 3r_{L})}{(r_{U}-r_{L})^3}& |
| 228 |
\text{if $r_{L} \leq r_{ij} \leq r_{U}$},\\ |
| 229 |
0& \text{if $r_{ij} \geq r_{U}$}. |
| 230 |
\end{cases} |
| 231 |
\label{eq:spCutoff} |
| 232 |
\end{equation} |
| 233 |
|
| 234 |
Despite the apparent complexity of Equation \ref{eq:spPot}, the SSD |
| 235 |
model is still computationaly inexpensive. This is due to Equation |
| 236 |
\ref{eq:spCutoff}. With $r_{L}$ being 2.75~$\mbox\AA}$ and $r_{U}$ |
| 237 |
being equal to either 3.35~$\mbox{\AA}$ for $s(r_{ij})$ or |
| 238 |
4.0~$\mbox{\AA}$ for $s'(r_{ij})$, the sticky potential is only active |
| 239 |
over an extremly short range, and then only with other SSD |
| 240 |
molecules. Therefore, it's predominant interaction is through it's |
| 241 |
point dipole and Lennard-Jones sphere. |
| 242 |
|
| 243 |
\subsection{The Phospholipid Model} |
| 244 |
\label{sec:lipidModel} |
| 245 |
|
| 246 |
\begin{floatingfigure}{90mm} |
| 247 |
\includegraphics[angle=-90,width=80mm]{lipidModel.epsi} |
| 248 |
\caption{A representation of the lipid model. $\phi$ is the torsion angle, $\theta$ is the bend angle, $\mu$ is the dipole moment of the head group, and n is the chain length. \vspace{5mm}} |
| 249 |
\label{fig:lipidModel} |
| 250 |
\end{floatingfigure} |
| 251 |
|
| 252 |
The lipid molecules in our simulations are unified atom models. Figure |
| 253 |
\ref{fig:lipidModel} shows a template drawing for one of our |
| 254 |
lipids. The Head group of the phospholipid is replaced by a single |
| 255 |
Lennard-Jones sphere with a freely oriented dipole placed at it's |
| 256 |
center. The magnitude of it's dipole moment is 20.6 D. The tail atoms |
| 257 |
are unified $\text{CH}_2$ and $\text{CH}_3$ atoms and are also modeled |
| 258 |
as Lennard-Jones spheres. The total potential for the lipid is |
| 259 |
represented by Equation \ref{eq:lipidModelPot}. |
| 260 |
|
| 261 |
\begin{equation} |
| 262 |
V_{\mbox{lipid}} = \overbrace{% |
| 263 |
V_{\text{bend}}(\theta_{ijk}) +% |
| 264 |
V_{\text{tors.}}(\phi_{ijkl})}^{bonded} |
| 265 |
+ \overbrace{% |
| 266 |
V_{\text{LJ}}(r_{i\!j}) + |
| 267 |
V_{\text{dp}}(r_{i\!j},\Omega_{i},\Omega_{j})% |
| 268 |
}^{non-bonded} |
| 269 |
\label{eq:lipidModelPot} |
| 270 |
\end{equation} |
| 271 |
|
| 272 |
The non-bonded interactions, $V_{\text{LJ}}$ and $V_{\text{dp}}$, are |
| 273 |
the Lennard-Jones and dipole-dipole interactions respectively. For the |
| 274 |
non-bonded potentials, only the bend and the torsional potentials are |
| 275 |
calculated. The bond potential is not calculated, and the bond lengths |
| 276 |
are constrained via RATTLE.\cite{leach01:mm} The bend potential is of |
| 277 |
the form: |
| 278 |
\begin{equation} |
| 279 |
V_{\text{bend}}(\theta_{ijk}) = k_{\theta}\frac{(\theta_{ijk} - \theta_0)^2}{2} |
| 280 |
\label{eq:bendPot} |
| 281 |
\end{equation} |
| 282 |
Where $k_{\theta}$ sets the stiffness of the bend potential, and $\theta_0$ |
| 283 |
sets the equilibrium bend angle. The torsion potential was given by: |
| 284 |
\begin{equation} |
| 285 |
V_{\text{tors.}}(\phi_{ijkl})= c_1[1+\cos\phi_{ijkl}] |
| 286 |
+ c_2 [1 - \cos(2\phi_{ijkl})] + c_3[1 + \cos(3\phi_{ijkl})] |
| 287 |
\label{eq:torsPot} |
| 288 |
\end{equation} |
| 289 |
All parameters for bonded and non-bonded potentials in the tail atoms |
| 290 |
were taken from TraPPE.\cite{Siepmann1998} The bonded interactions for |
| 291 |
the head atom were also taken from TraPPE, however it's dipole moment |
| 292 |
and mass were based on the properties of the phosphatidylcholine head |
| 293 |
group. The Lennard-Jones parameter for the head group was chosen such |
| 294 |
that it was roughly twice the size of a $\text{CH}_3$ atom, and it's |
| 295 |
well depth was set to be approximately equal to that of $\text{CH}_3$. |
| 296 |
|
| 297 |
\section{Initial Simulation: 25 lipids in water} |
| 298 |
\label{sec:5x5} |
| 299 |
|
| 300 |
\subsection{Starting Configuration and Parameters} |
| 301 |
\label{sec:5x5Start} |
| 302 |
|
| 303 |
Our first simulation was an array of 25 single chained lipids in a sea |
| 304 |
of water (Figure \ref{fig:5x5Start}). The total number of water |
| 305 |
molecules was 1386, giving a final of water concentration of 70\% |
| 306 |
wt. The simulation box measured 34.5~$\mbox{\AA}$ x 39.4~$\mbox{\AA}$ |
| 307 |
x 39.4~$\mbox{\AA}$ with periodic boundary conditions invoked. The |
| 308 |
system was simulated in the micro-canonical (NVE) ensemble with an |
| 309 |
average temperature of 300~K. |
| 310 |
|
| 311 |
\subsection{Results} |
| 312 |
\label{sec:5x5Results} |
| 313 |
|
| 314 |
Figure \ref{fig:5x5Final} shows a snapshot of the system at |
| 315 |
3.6~ns. Here it is exciting to note that the system has spontaneously |
| 316 |
self assembled into a bilayer. Discussion of the length scales of the |
| 317 |
bilayer will follow in this section. However, it is interesting to |
| 318 |
note several qualitative properties of the system revealed by this |
| 319 |
snapshot. First, is how the tail chains are tilted to the bilayer |
| 320 |
normal. This is usually indicative of the gel ($L_{\beta'}$) |
| 321 |
phase. Likely, the box size is too small for the bilayer to relax to |
| 322 |
the fluid ($P_{\alpha}$ phase. Showing the need for a constant |
| 323 |
pressure simulation versus the constant volume of the current |
| 324 |
ensemble. |
| 325 |
|
| 326 |
The trajectory of the simulation was analyzed using the pair-wise |
| 327 |
radial distribution function, $g(r)$, which has the form: |
| 328 |
\begin{equation} |
| 329 |
g(r) = \frac{V}{N_{\text{pairs}}}\langle \sum_{i} \sum_{j \neq i} |
| 330 |
\delta(|\mathbf{r} - \mathbf{r}_{ij}|) \rangle |
| 331 |
\label{eq:gofr} |
| 332 |
\end{equation} |
| 333 |
Equation \ref{eq:gofr} gives us information about the spacing of two |
| 334 |
species as a function of radius. Essentially, if the observer were |
| 335 |
located at atom $i$ and were looking out in all directions, $g(r)$ |
| 336 |
shows the relative density of atom $j$ at any given radius, $r$, |
| 337 |
normalized by the expected density of atom $j$ at $r$. In a |
| 338 |
homogeneously mixed fluid, $g(r)$ will approach 1 at large $r$, as a |
| 339 |
fluid contains no long range structure to contribute peaks in the |
| 340 |
number density. |
| 341 |
|
| 342 |
For the species containing dipoles, a second pair wise distribution |
| 343 |
function was used, $g_{\gamma}(r)$. It is of the form: |
| 344 |
\begin{equation} |
| 345 |
g_{\gamma}(r) = foobar |
| 346 |
\label{eq:gammaofr} |
| 347 |
\end{equation} |
| 348 |
Where $\gamma_{ij}$ is the angle between the dipole of atom $j$ with |
| 349 |
respect to the dipole of atom $i$. This correlation will vary between |
| 350 |
$+1$ and $-1$ when the two dipoles are perfectly aligned and |
| 351 |
anti-aligned respectively. This then gives us information about how |
| 352 |
directional species are aligned with each other as a function of |
| 353 |
distance. |
| 354 |
|
| 355 |
Figure \ref{fig:5x5HHCorr} shows the two self correlation functions |
| 356 |
for the Head groups of the lipids. The first peak at 4.03~$\mbox{\AA}$ is the |
| 357 |
nearest neighbor separation of the heads of two lipids. |
| 358 |
|
| 359 |
|
| 360 |
|
| 361 |
|
| 362 |
\section{Second Simulation: 50 randomly oriented lipids in water} |
| 363 |
|
| 364 |
the second simulation |
| 365 |
|
| 366 |
\section{Future Directions} |
| 367 |
|
| 368 |
|
| 369 |
\pagebreak |
| 370 |
\bibliographystyle{achemso} |
| 371 |
\bibliography{canidacy_paper} \end{document} |
