Similarities and differences of serotonin and its precursors in their interactions with model membranes studied by molecular dynamics simulation

Article abstract of DOI:10.1016/j.molstruc.2013.04.011

In this work, we report a molecular dynamics (MD) simulations study of relevant biological molecules as serotonin (neutral and protonated) and its precursors, tryptophan and 5-hydroxy-tryptophan, in a fully hydrated bilayer of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidyl-choline (POPC). The simulations were carried out at the fluid lamellar phase of POPC at constant pressure and temperature conditions. Two guest molecules of each type were initially placed at the water phase. We have analyzed, the main localization, preferential orientation and specific interactions of the guest molecules within the bilayer. During the simulation run, the four molecules were preferentially found at the water-lipid interphase. We found that the interactions that stabilized the systems are essentially hydrogen bonds, salt bridges and cation-π. None of the guest molecules have access to the hydrophobic region of the bilayer. Besides, zwitterionic molecules have access to the water phase, while protonated serotonin is anchored in the interphase. Even taking into account that these simulations were done using a model membrane, our results suggest that the studied molecules could not cross the blood brain barrier by diffusion. These results are in good agreement with works that show that serotonin and Trp do not cross the BBB by simple diffusion.

Full text of DOI:10.1016/j.molstruc.2013.04.011

Journal of Molecular Structure 1045 (2013) 124–130
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Structure
journal homepage: www.elsevier.com/locate/molstruc
Similarities and differences of serotonin and its precursors in their
interactions with model membranes studied by molecular dynamics
simulation
Irene Wood , M. Florencia Martini a,b, Mónica Pickholz a,b,
a
⇑
a
Department of Pharmaceutical Technology, University of Buenos Aires, Buenos Aires CP 1113, Argentina
Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Av. Rivadavia 1917, C1033AAJ-Buenos Aires, Argentina
b
h i g h l i g h t s
ꢀ
ꢀ
ꢀ
ꢀ
We simulate four NPT molecular dynamics simulations of serotonin and related molecules on lipid bilayers.
Main interactions were studied for all simulation systems.
Trp, 5-hydroxy-tryptophan and serotonin show similar localization on lipid bilayer-water interphase.
No crossing events were observed through hydrophobic lipid region.
a r t i c l e i n f o
a b s t r a c t
Article history:
In this work, we report a molecular dynamics (MD) simulations study of relevant biological molecules
as serotonin (neutral and protonated) and its precursors, tryptophan and 5-hydroxy-tryptophan, in a
fully hydrated bilayer of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidyl-choline (POPC). The simula-
tions were carried out at the fluid lamellar phase of POPC at constant pressure and temperature condi-
tions. Two guest molecules of each type were initially placed at the water phase. We have analyzed,
the main localization, preferential orientation and specific interactions of the guest molecules within
the bilayer. During the simulation run, the four molecules were preferentially found at the water–
lipid interphase. We found that the interactions that stabilized the systems are essentially hydrogen
Received 22 November 2012
Received in revised form 4 April 2013
Accepted 5 April 2013
Available online 18 April 2013
Keywords:
Serotonin
L-tryptophan
bonds, salt bridges and cation-p. None of the guest molecules have access to the hydrophobic region
5
-Hydroxytryptophan
of the bilayer. Besides, zwitterionic molecules have access to the water phase, while protonated serotonin
is anchored in the interphase. Even taking into account that these simulations were done using a model
membrane, our results suggest that the studied molecules could not cross the blood brain barrier by dif-
fusion. These results are in good agreement with works that show that serotonin and Trp do not cross the
BBB by simple diffusion.
Molecular dynamics
Lipid bilayer
Ó 2013 Elsevier B.V. All rights reserved.
1
. Introduction
HTP) by tryptophan hydroxylase and a second step where the 5-
HTP is decarboxylated by the aromatic
into 5-HT [2] (see Fig. 1).
L-amino acid decarboxylase
Serotonin (5-hydroxy-tryptamine, 5-HT) is a monoamine neu-
rotransmitter and hormone implicated in several physiological
processes by activating multiple receptors. Abnormalities of the
serotonergic system have been implicated in many psychiatric dis-
orders including anxiety, depression, psychosis, migraine, and oth-
ers [1]. Serotonin is physiologically synthesized by a two-step
reaction: a first and limiting step where the essential amino acid
During several years 5-HTP was extensively proposed for the
treatment of migraine and others syndromes affecting serotonergic
system [3,4]. This serotonin precursor could be extracted from
Griffonia simplicifolia seeds [3] and orally administered to enhance
serotonin transmission or improve several symptoms. The admin-
istration of 5-HTP, instead of Trp, is a good and more effective
choice because it by-passes the first limiting step of 5-HT biosyn-
thesis and it is entirely converted to 5-HT [5]. However, the clinical
use of 5-HTP still under discussion because its contraindications
and side effects.
L-tryptophan (Trp) is converted into 5-hydroxytryptophan (5-
⇑
Corresponding author. Address: Cátedra de Farmacotecnia II, Facultad de
Farmacia y Bioquímica, Universidad de Buenos Aires, Junín 956 2° Piso (1113),
Capital Federal, Argentina. Tel.: +54 011 4964 8273.
The blood–brain barrier (BBB) acts like an interphase that re-
stricts and regulates exchange of substances – ions, nutrients, ami-
E-mail address: mpickholz@gmail.com (M. Pickholz).
0
022-2860/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molstruc.2013.04.011

I. Wood et al. / Journal of Molecular Structure 1045 (2013) 124–130
125
Fig. 1. Structures of four guest molecules studied: (a) L-Tryptophan (Trp), (b) 5-hydroxy-tryptophan (5-HTP), (c) neutral serotonin (n5-HT), (d) charged serotonin (p5-HT).
The four guest molecules studied share a 3 substituted indolic moiety. Just Trp and 5-HTP are present as zwitterions because they bear a carboxylate group at 2-aminoethyl
þ
þ
substituent, in addition to the NH3 group. The molecules n5-HT and p5-HT bear a NH
2
and NH3 group respectively. Trp is the unique molecule which is not bearing OH group
at position 5 of indole.
phobic nature of the indolic ring. Furthermore, Johansson and
Lindahl [16] studied the solvation of several amino acids in trans-
membrane segments of membrane proteins by MDs. They found
that Trp, like tyrosine (Tyr), acts as membrane helix anchors.
Physicochemical studies of Trp analogues and small peptides
containing Trp suggest that this behavior arises from enhanced
stability due to distinct interfacial interactions such as hydrogen
bonds, dipolar and cation-p interactions [17].
In this work, we study by molecular dynamics (MD) simulations
the interactions of serotonin, 5-HTP and Trp with model mem-
branes of 1-palmitoil-2-oleoil-sn-glicero-3-phosphatidyl-choline
(
POPC). Model membranes of POPC were chosen since phosphati-
dyl choline lipid types are the most abundant in mammal endothe-
lial cell membranes [18] such as those of BBB. Furthermore, model
bilayers of POPC are used to study permeation and diffusion of sev-
eral molecules and drugs as simplified model membranes of the
BBB [19]. Trp and 5-HTP are amino acid, so they are found at zwit-
terionic form. Serotonin is essentially found protonated at physio-
logical pH [20]. Nevertheless, in this study we consider both
ionization states: protonated (p5-HT) and neutral (n5-HT). We fo-
cus our attention on the localization, orientation and specific inter-
actions of the guest molecules within the bilayer. Since, they are
very similar molecules, with different physiological properties,
we aim to understand the main similarities and differences be-
tween them in their interaction with model membranes.
Fig. 2. Snapshot of initial configuration for 5-HTP in water region, surrounding a
POPC bilayer, visualized by VMD 1.8.7 software package [48].
no acids, peptides and neurotransmitters – between central ner-
vous system (CNS) and circulating blood. It is well known that
both, 5-HT and Trp enter to the CNS through the BBB, during a
process facilitated by transporters. 5-HT is transported by seroto-
nin transporter SERT, located both at luminal (blood) and ablumi-
nal (brain) faces of brain capillary endothelial cells membrane [6].
The SERT carrier is involved in serotonin transport both from and
to CNS. Furthermore, the large transporter LAT-1 is responsible
for the transport of most neutral amino acids through BBB [7]. In
this way Trp should compete with other amino acid, such as phen-
ylalanine and tyrosine, for binding to LAT-1 [8,9]. However, in our
knowledge, there is not experimental data supporting that the
intermediate 5-HTP cross BBB by a specific carrier. Some authors
such Birdsall [5] suggest that it could cross BBB by passive diffu-
sion. In relation to this fact, recent studies demonstrate that the
intravenous administration of 5-HTP increase levels of 5-HT in
brain extracellular fluid and then in whole blood [10]. This indi-
cates that 5-HT could cross from brain to blood through BBB, and
that 5-HTP could cross from blood to brain through an unknown
mechanism.
2. Methodology
Here, we study the interaction between Trp, n5-HT, p5-HT or
5-HTP (Fig. 1) and a bilayer of POPC. All simulated systems consist
in a lipid bilayer containing 150 POPC molecules (75 in each leaflet)
fully hydrated with 4200 water molecules and two identical mol-
ecules (5-HT and precursors), under periodic boundary conditions
(PBC). The initial structures were built using Packmol package [21].
The guest molecules were originally placed into the water phase
(as shown in Fig. 2). In order to assure the electroneutrality, we
have added chloride counter ions when protonated molecules are
present.
Simulations were performed using the NAMD2 program [22]
with the PARAM27 CHARMM parameter set [23] and the water
molecules were described by the TIP3P model [24]. The amino acid
Trp is already parameterized into CHARMM force field [25].
On the other hand, the parameters for 5-HT and 5-HTP were
mostly taken from the CHARMM Trp, because of their structural
similarities, and estimated from quantum chemical calculations.
In this way, we carried out quantum chemical calculation, using
Gaussian03W package [26], in order to obtain the optimized geom-
etry of these molecules. The optimization was done at the density
functional theory (DFT) [27] level using the B3LYP [28] functional
Despite its low occurrence in nature [11], the amino acid Trp
was extensively studied. Trp residues in membrane proteins are re-
lated with several functions including protein organization,
anchoring and orientation within the bilayer. Besides, it is common
to find Trp at the water–lipid interphase in membrane proteins
ꢁ
ꢁ
[
12,13]. There are several works that addressed the interaction of
molecules such indole with lipid bilayers through MD simulations
13,14]. Comparing with other amino acids analogues studied by
and 6-311G [29] basis set. In this way we obtained the equilib-
rium bond lengths and angles of the structures of interest. The par-
ꢁ
[
tial atomic charges were obtained from a single point HF/6-31G
simulation [15], molecules such indole, have interfacial preference
due to a combination of weak electrostatic factors and the hydro-
[30] calculation within the Merz–Singh–Kollman protocol [31].
The calculations were performed using the polarized continuum

1
26
I. Wood et al. / Journal of Molecular Structure 1045 (2013) 124–130
model (PCM) [32] with solvent dielectric constant of 80, corre-
sponding to liquid water at environment conditions. PCM was
the chosen methodology in order to obtain the zwitterionic charge
separation of the amino acids.
Classical atomistic MDs were performed at the NPT conditions.
The temperature was maintained constant at 310 K, through a
Langevin thermostat by applying friction and random force [22].
The Langevin thermostat was used without coupling to hydrogens
and with a damping coefficient of 1/ps. Besides, pressure was
maintained at 1 atm by using a Langevin barostat [33] with a pis-
ton period of 5 ps and a damping time of 5 ps. The simulation
box was fully flexible. A multiple-time step algorithm, RESPA
In order to facilitate the comparison, we have included the total
guest molecule (in turquoise) and the phosphates POPC EDPs (in
dotted black). A representative scheme of the preferential orienta-
tion of the molecules is shown in the onset of Fig. 4A, B and D. Since
no preferential orientation was found for the n5-HT (Fig. 4C) case -
the maximum of the different groups distributions are located at
the same position in z axis – the onset scheme show a random cho-
sen orientation of the molecule. Besides, the area per lipid of the
2
bilayer slightly increases in presence of the n5-HT (61.5 Å ) respect
2
to the other guest molecules ꢂ60 Å . This effect can be explained
by the less restricted rotations – and consequently higher effective
volume – of the n5-HT compared to the other molecules. The area
per lipid of a plain bilayer using the same model and simulation
[
34], was used with the shortest time step of 2 fs. The short-range
2
forces were computed using a cut-off of 10 Å and the long-range
forces were taken into account by means of the particle mesh
Ewald (PME) technique [35].
The initial structures were energy-minimized for 2000 steps
using the Steepest Descent algorithm to remove bad contacts.
The simulation consisted of an equilibration period of about
conditions was reported as 58 Å [36].
On the other hand, for Trp (A), 5-HTP (B) and p5-HT (D) we can
see that the indole moiety enters deeper in the bilayer than the lat-
eral chain. Besides, we can see a superposition between the pyrrole
ring and the phosphate groups of the POPC lipids. In order to elu-
cidate if this superposition could be related to one specific interac-
tion, we have calculated the radial distribution functions between
the H atom of indolic NH (guest molecules) and O phosphate group
of POPC for the four guest molecules. The results are shown in
1
0 ns, within each of the interesting molecules migrated to their
preferential localization relatively to the membrane, followed by
5
0 ns production run.
3
. Results and discussion
Here, we summarized the main results of the analysis of our
simulations, performed over the 50 ns of the simulations run of
the 4 systems.
3.1. Localization of guest molecules in the bilayer
In order to have an idea of the overall organization of the sys-
tems under study, we have calculated the electron density profile
EDP), by time averaging the net charges per 0.1 Å thick slabs, as
(
function of the z coordinate (normal to the lipid bilayer). The
z = 0 corresponds to the bilayer center. In Fig. 3 we show the EDPs
of the different systems under study (from top to bottom): (a) Trp,
(
b) 5-HTP, (c) n5-HT and (d) p5-HT. Besides, the EDPs correspond-
ing to the different molecule types are shown in the following col-
ors: POPC in black, water in blue and guest molecules in red. As we
can see from the water EDP there are three well defined regions: a
bulk aqueous phase, an intermediate complex water–lipid inter-
phase in which the polar headgroups are hydrated, and a poor hy-
drated hydrophobic region. The POPC electron density profiles
reveal a distribution characteristic of a lipid bilayer organization.
Superimposing the POPC distributions of the four studied cases (re-
sults not shown) we can see that the bilayer organization was not
significantly affected.
Furthermore, guest molecules are essentially found at water–li-
pid interphase, with access to the water phase. In some cases, this
access leads to both molecules to be in the same monolayer, and
this is the reason for the observed asymmetric distributions. The
5
-HTP molecule shows higher access to the water phase than the
other molecules. By the other hand, the p5-HT, even bearing a
net charge, shows a more compacted distribution than the other
molecules at the interphase.
In order to find more details on the water–lipid interphase
localization of the guest molecules, we have investigated the orien-
tation of separated groups of the guest molecules, as well as their
main specific interactions with both, POPC lipids and water mole-
cules. In this way, as a first step, we performed a detailed EDP of
the different groups of the guest molecule EDP, shown in
Fig. 4A–D. The selected groups are: the OH lateral group in purple
Fig. 3. Electron density profiles for all systems: Trp-POPC, a; 5-HTP-POPC, b; n5-
HT-POPC, c; p5-HT-POPC, d. POPC electron density is depicted in black, water in
blue and guest molecules in red (15-fold to easiest visualization). z = 0 correspond
to bilayer center. (For interpretation of the references to colour in this figure legend,
the reader is referred to the web version of this article.)
(
10,000-fold for visualization purposes), the amino lateral group in
black, and benzene and pyrrole rings in red and green, respectively.

I. Wood et al. / Journal of Molecular Structure 1045 (2013) 124–130
127
Fig. 4. Detailed electron density profiles for all systems and possible guest molecule orientation: Trp, A; 5-HTP, B; n5-HT, C; p5-HT, D. Total electron density for any guest
molecule is depicted in turquoise, amino lateral group in black, benzene ring in red, pyrrole ring in green, OH in purple (10,000-fold to easiest visualization) and Phosphate of
the POPC in black dotted line and scaled to serve as reference location. Guest molecules orientations are schematized in the same color scale than detailed EDP for the
functional groups. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
ated molecule (p5-HT), as expected, followed by the zwitterionic
ones (Trp and 5-HTP). Furthermore, in order to see the stability
of SB for each case, we have calculated the minimum distance
between lateral amino groups of guest molecules and phosphate
of POPC during time. The results are shown in the onset of Fig. 6.
3
.3. Cation-
p
Interactions
Cation-
p
interaction is a noncovalent molecular interaction
between the face of an electron-rich
p system (e.g. benzene, ethyl-
Fig. 5. Radial distribution functions for all systems. The atom pair considered is
non-ester Op (phosphate group, POPC) and H (indolic NH group, guest molecule).
Trp is depicted in black, 5-HTP in red, n5-HT in green and p5-HT in blue. (For
interpretation of the references to colour in this figure legend, the reader is referred
to the web version of this article.)
Fig. 5. We can see a well defined peak ꢂ1.7 Å in all the cases, char-
acteristic of hydrogen bond kind of interactions.
3.2. Salt bridge Interactions
Salt bridge (SB) is a kind of non-covalent electrostatic interac-
tion established between charged pairs [37–40]. In membrane pro-
teins, that kind of interaction is involved in transmembrane
segments binding and orientation processes [41,42].
In our case, the salt bridges between the center of mass of lat-
eral amino group (guest molecule) and center of mass of phosphate
group (POPC) was investigated through the radial distribution
function analysis. In Fig. 6 we show the gSB(r) for the three studied
cases. From this analysis, we found that charged and zwitterionic
guest molecules form SB, as evidenced by the first peak at ꢂ4 Å.
However, the intensity of these interaction is higher for the proton-
Fig. 6. Radial distribution functions for the three charged systems. The atom pair
considered is lateral amino group (guest molecules) and phosphate groups (POPC).
Trp is depicted in black, 5-HTP in red and p5-HT in blue. Onset: minimum distance
between lateral amino groups of guest molecules and phosphate of POPC during
time. Trp, A; 5-HTP, B; p5-HT, C. (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.)

1
28
I. Wood et al. / Journal of Molecular Structure 1045 (2013) 124–130
Table 1
Average number of hydrogen bonds per molecule between lateral amino, NH indolic
and OH (guest molecules) and non-ester oxygen atoms of phosphate and carbonyl
groups, Op and Oc (POPC). The statistical errors are also shown.
Op
Oc
þ
Trp
NH lateral
1.118(4)
0.354(2)
0.107(2)
0.077(2)
3
NH indole
þ
5-HTP
NH lateral
0.803(3)
0.352(2)
0.329(2)
0.023(1)
0.006(1)
0.100(2)
3
NH indole
OH substituent
n5-HT
p5-HT
NH
NH indole
OH substituent
2
lateral
0.125(2)
0.349(2)
0.367(2)
0.072(2)
0.028(1)
0.098(2)
þ
NH lateral
NH indole
OH substituent
1.568(4)
0.507(2)
0.122(1)
0.043(1)
0.050(1)
0.168(2)
3
Fig. 7. Schematization of distance (r) requirements for cation-
candidates from the center of the six-member ring of indole. The rigid indole ring
is considered as a plane, in which n is the normal of this plane. Distance out-of-
p
interaction
the four cases under study are shown in Fig. 8. We can see a well-
defined peak at ꢂ4.0 Å for all guest molecules.
plane h and in-plane d determine the angular distance dC6-
cation donor for cation- interaction.
p requirement to find a
3.4. Hydrogen bonds
p
We adopted the following geometric criteria for the hydrogen
bond (HB): the distance between the acceptor and the hydrogen
is shorter than or equal to 2.0 Å, and the angle between donor-
hydrogen-acceptor should be smaller than 30°.
ene, acetylene) and an adjacent cation. An important implication of
the cation- interaction and the related ‘‘polar- ’’ interactions is
p
p
that phenylalanine (Phe), tyrosine (Tyr), and Trp should not be
considered simply ‘‘hydrophobic’’ amino acids. They are in fact
distinct from the conventional hydrophobic residues such as valine
The number of HB made by a selected pair was determined by
counting the number of pairs that satisfy these geometric criteria
and by averaging over time (50 ns).
(
Val), leucine (Leu), and isoleucine (Ile). Because of the dual
Taken into account that POPC just can act as acceptor through
the non-ester oxygen atoms of phosphate (Op) and carbonyl (Oc)
groups – oxygen atoms that are just bonded to one heteroatom –
we have calculated number of HB between them and the 3 donor
groups (amino lateral, NH indole and OH lateral) of guest mole-
cules. In Table 1, we summarized the average number of HB be-
tween these groups.
nature of benzene hydrophobic yet quadrupolar aromatic amino
acids are expected to play a unique role in protein structure and
function [43].
In particular, the indole ring is an electron rich aromatic system
[
44]. It has hydrogen bond donor character, due to its indolic NH
group, and ability to form cation- interactions, due to its elec-
p
p
tron system. Since aromatic Trp, and related molecules like 5-
HTP and 5-HT, are electrically characterized by its quadrupolar
Amino lateral: For the charged and zwitterionic guest molecules
we found a strong interaction between the hydrogens of this group
with Op oxygens. The stronger interaction was found for the p5-HT
(with an average of 1.56 HBs per guest molecule). This strong inter-
action could be one of the factors of the anchoring of this molecule,
as discussed above. Furthermore, for n5-HT, as expected since this
group is neutral and has just two hydrogens atoms, the number of
HB is one order of magnitude lower (0.13 HB per molecule). By the
other hand, the HB formation with the Oc is considerable low for
the four molecules.
moment, acts like a
p-donor interacting with the positively
charged amino group of POPC choline. This kind of interaction
was studied from the radial distribution function between the cen-
troid of benzene ring of indole, of the guest molecules, and charged
nitrogen (N) atom of choline, in POPC (as shown in Fig. 7) such in
p
the work of Petersen et al. [17]. The calculated gc- (r) functions for
NH indole: The ability of indolic structure to form a hydrogen
bond could explain differences of behavior between molecules, like
Trp, with other hydrophobic and aromatic amino acid like
L-Phe
[
17]. In Phe the benzene group lacks a hydrogen bond donor such
indolic NH. Our results show a comparable HB number (ꢂ0.36 HB)
for the Trp, n5-HT and 5-HTP and grows up to ꢂ0.5 HB for the p5-
HT (anchoring). By the other hand, the HB interaction with the Oc
is quite low for the four guest molecules.
OH lateral: This group is not present in Trp. By the other hand,
the other molecules are doing HB ꢂ30% of the time (5-HTP and
n5-HT) and ꢂ10% of the time (p5-HT) with the Op and 10–15%
with the Oc.
By the other hand, the four guest molecules have the ability to
form hydrogen bonds with water, since they bear donor and
acceptor hydrogen bond groups in their structures. In Table 2,
we show the results of the number of HB formed, per guest mol-
ecule, with water. As water molecules can act both as donor and
acceptor we have separated in Table 2 the HB made by the Ow
(acceptor) and the total number of HB. It is important to point
Fig. 8. Radial distribution functions for all systems. The atom pair considered is
centroid of benzene ring of indole (guest molecule) and charged N atom of choline
groups (POPC). Trp is depicted in black, 5-HTP in red, n5-HT in green and p5-HT in
blue. (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)

I. Wood et al. / Journal of Molecular Structure 1045 (2013) 124–130
129
Table 2
Average number of HB per molecule between lateral amino, carboxylate, NH indolic
and OH (guest molecules) with water oxygen atoms and total water, and are
expressed per molecule. The statistical errors are also shown.
O–H
2
O
2
H O
þ
Trp
NH lateral
0.658(4)
–
0.338(2)
1.229(4)
4.176(4)
0.312(2)
3
ꢃ
COO lateral
NH indole
þ
5
-HTP
NH lateral
0.602(4)
–
0.356(2)
0.341(2)
1.541(4)
4.105(4)
0.410(2)
0.779(3)
3
ꢃ
COO lateral
NH indole
OH substituent
n5-HT
p5-HT
NH
NH indole
OH substituent
2
lateral
0.352(2)
0.321(2)
0.329(2)
0.913(4)
0.319(2)
0.779(3)
Fig. 9. Water centered trajectories along z axes. In red and black are depicted every
one of both residues of any guest molecule. Trp, A; 5-HTP, B; n5-HT, C; p5-HT, D.
z = 0 correspond to center of water region. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
þ
NH lateral
NH indole
OH substituent
0.552(3)
0.303(2)
0.353(2)
1.045(3)
0.265(2)
0.725(3)
3
mass trajectories for each of the guest molecules over the 50 ns
of the simulation run, and the results are shown in Fig. 9. While
in Figs. 3 and 4 are centered in the middle of the lipid bilayer, in
Fig. 9 z = 0 corresponds to the center of the water phase (taking
into account the periodic boundary conditions). It is important to
point out here that this approximation could be compared against
multilamellar vesicle, since the amount of water was estimated for
a fully hydrated lipid bilayer [45]. The comparison with the cell dif-
fusion condition is somehow not straightforward. During the 50 ns
simulation time we did not find any crossing events of the guest
molecules through lipid hydrophobic core. Crossing events was de-
fined as the guest molecule passes from one lipid monolayer to the
other one through hydrophobic region. 50 ns is enough time to let
small molecules cross the lipid bilayer: for example, local anesthet-
ics (containing 40 atoms) are able to cross a POPC bilayer in a lower
time scale of 20 ns [46].
Furthermore, in three of the four cases we observe crossing
events through the water region. It is important to point out, that
these crossing events occur with a stabilized system. The only sys-
tem not showing these events is the p5-HT. The reason for that was
discussed above and it is related with the stronger interactions that
have these molecules with the interphase, which anchor them in
the interphase region. By the other hand, n5-HT does not show
strong specific interactions, and no preferential orientation. How-
ever, even if n5-HT access to the water phase, this is a fast event
and then re-enter quickly into the bilayer. By the other hand, the
zwitterionic molecules – especially 5-HTP – can remain more time
in water.
out that both, Trp and 5-HTP are zwitterionic and bear an extra
carboxylate group (COO ) that is able to interact with the water
molecules.
ꢃ
As an overall view, we can say that 5-HTP is the guest molecules
that most interact with water through HB (ꢂ6.8). This is due to the
fact that this molecule has greater number of polar groups in its
structure, and consequently more groups to be solvated. These
two molecules, Trp and 5-HTP, have more access to the water
phase, as we can see in the EDP and will discuss below. Our results
ꢃ
show that the COO group strongly interacts with water forming
ꢂ4.1 HB in average for both zwitterionic structures.
We can see that serotonin in both ionization states are the mol-
ecules that have less HB with water (ꢂ1.9). For the n5-HT amino
lateral group we can see, if we compare the donor acceptor type
of HB, that is the only one that act as an acceptor, as well (0.64),
because it is not charged.
The OH substituent has also the capability to act as donor an
acceptor. We can see from Table 2 that the interaction of this group
as donor is higher for the p5-HT (competing with the Op and Oc)
groups. This group is quite polar, and needs to be solvated, acting
as a donor. In this way, the three molecules that bear this group
are essentially 80% of the time forming HB.
By the other hand, the NH of indole ring acts as donor in the
range of 70–85% of the time.
Looking for the performance of the other molecules, the number
of HB formed with water are in good agreement with the amount
of donor and acceptor groups bearing in the structure and ioniza-
tion state. In addition, detailed EDP demonstrates preferential ori-
þ
ꢃ
entation of charged groups (NH3 and COO ) in Trp and 5-HTP and
þ
4. Conclusions
NH in p5-HT. The fact of facing these groups to water explains
3
highest number of HB established with water: 5-HTP > Trp > n5-
HT > p5-HT, in addition to simulation time spends in water region.
Considering the localization of guest molecules into lipid bilay-
ers and taking into account all the detailed specific interactions, it
should be reasonable to analyze if these molecules cross the mem-
brane through the lipid core. In this way, in the next section we
analyze the trajectories of all the studied molecules along the sim-
ulation run time.
In this work, we carried out molecular dynamics simulations of
compounds of physiological and clinical importance as are seroto-
nin and its precursors in a model lipid bilayer, at very low concen-
tration. We aim to recognize similarities and differences of these
molecules regarding their localization and main interactions with
the lipid bilayer.
Our results show that the four guest molecules are preferen-
tially found at the water–lipid interphase, some of them with ac-
cess to the water phase. Furthermore, charged and zwitterionic
molecules show a preferential orientation within the bilayer: the
aromatic ring enters deeper in the lipid region. By the other hand,
the neutral serotonin did not show, in average, any preferential ori-
entation. The interactions that stabilize the systems are mainly
3.5. Trajectories
The EDP analysis gave us a first picture of the distribution of the
guest molecules within the bilayer. In most of cases, we found that
the molecules partitioned between the water phase and the water–
lipid interphase. In order to obtain more detailed information
about this distribution, we have followed the molecules center of
hydrogen bonds, salt bridges and cation-p.
Hydrogen bond are essentially formed between indolic NH,
amino lateral group and OH lateral group are found, with the

1
30
I. Wood et al. / Journal of Molecular Structure 1045 (2013) 124–130
non-ester phosphate oxygen atoms of the POPC and, in less extent,
with the POPC oxygen atoms of the carbonyl groups. Besides, cat-
[9] W.M. Pardridge, Neurochem. Res. 23 (1998) 635–644.
10] Y. Nakatani, I. Sato-Suzuki, N. Tsujino, A. Nakasato, Y. Seki, M. Fumoto, H. Arita,
Eur. J. Neurosci. 27 (2008) 2466–2472.
11] N. Echols, P. Harrison, S. Balasubramaniam, N. Luscombe, P. Bertone, Z. Zhang,
M. Gerstein, Nucleic Acids Res. 30 (2002) 2515–2523.
12] G. Von Heijne, J. Intern. Med. 261 (2007) 543–557.
13] J. MacCallum, W. Bennett, P. Tieleman, Biophys. J. 94 (2008) 3393–3404.
14] A. Grossfield, T. Woolf, Langmuir 18 (2002) 198–210.
[
[
ion-p interactions were found between the aromatic electronic
system of indole and nitrogen atom of choline group. In this way,
the protonated serotonin is anchored through all these strong
interactions in the water–lipid interphase. However, for the zwit-
terionic molecules, these interactions are destabilized by the pres-
ence of carboxylate group that strongly interacts with water, since
the POPC lipids do not have any HB donor group.
[
[
[
[15] K. Norman, H. Nymeyer, Biophys. J. 91 (2006) 2046–2054.
[
[
[
16] A. Johansson, E. Lindahl, Biophys. J. 91 (2006) 4450–4463.
17] F. Petersen, M. Jensen, C. Nielsen, Biophys. J. 89 (2005) 3985–3996.
18] D.P. Selivonchick, B.I. Roots, Lipids 12 (1977) 165–169.
During the simulation time length, we did not observed any
crossing events through hydrophobic region of the bilayer for
any of the molecules. In particular, it is important to remark that
even the more hydrophobic of the molecules (neutral serotonin)
was unable to cross the bilayer. This is in very good agreement
with the existence of a specific carrier (SERT); otherwise, the acid
basic balance could shift to form neutral specie at aqueous medium
and make the whole serotonin to cross BBB.
Besides, our results show a very similar behavior between the
amino acids Trp and 5-HTP. In addition to their structural similar-
ity, these molecules show the same characteristics in their interac-
tions with and preferential orientation within the bilayer (at the
interphase) as well as their access to the water phase. As was al-
ready mentioned in the introduction it was demonstrated that 5-
HTP could cross from blood to brain through an unknown mecha-
nism [10]. Our results suggest that this mechanism is a passive dif-
fusion because 5-HTP (as Trp) was not able to cross the model
membrane during the simulation time. At this point it is important
to remark the existence of a carrier LAT-1 used by Trp to cross the
BBB [7]. This carrier is also used for aminoacids, such as Tyrosine
[19] H. Fischer, R. Gottschlich, A. Seelig, J Membr. Biol. 165 (1998) 201–211.
[
[
20] G. Alagona, C. Ghio, P. Nagy, JCTC 1 (2005) 801–816.
21] L. Martínez, R. Andrade, E. Birgin, J. Martinez, J. Comput. Chem. 30 (2009)
2157–2164.
[22] L. Kalé, R. Skeel, M. Bhandarkar, R. Brunner, A. Gursoy, N. Krawetz, J. Phillips, A.
Shinozaki, K. Varadarajan, K. Schulten, J. Comput. Phys. 151 (1999) 283–312.
[
[
23] S. Feller, A.D. MacKerell Jr., J. Phys. Chem. B 104 (2000) 7510–7515.
24] W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, J. Chem. Phys. 79 (1983) 926–
935.
[25] A. MacKerell, D. Bashford, M. Bellott, R. Dunbrack, J. Evanseck, M. Field, S.
Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K. Kuczera, F.T.K.
Lau, C. Mattos, S. Michnick, T. Ngo, D.T. Nguyen, B. Prodhom, W.E. Reiher, B.
Roux, M. Schlenkrich, J.C. Smith, R. Stote, J. Straub, M. Watanabe, J. Wirkiewicz-
Kuczera, D. Yin, M. Karplus, J. Phys. Chem. B 102 (1998) 3586–3616.
[
26] M. Frisch, G. Trucks, H. Schlegel, G. Scuseria, M. Robb, J. Cheeseman, J.
Montgomery Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J.
Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson,
H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian,
J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O.
Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K.
Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S.
Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K.
Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J.
Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L.
Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M.
Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A.
Pople, Gaussian03, Revision B.05. Pople, Gaussian, Inc., Pittsburgh, PA, 2003.
27] R. Parr, W. Yang, Density Functional Theory of Atoms and Molecules, Oxford
University Press, New York, 1989.
and
L
-dopa, to reach the CNS. [47]. The same subtle difference be-
-dopa – the presence of an OH substituent
tween the tyrosine and
L
[
in the aromatic ring – is the one between 5-HTP and Trp. In this
way, further studies should be carried out in order to reveal if an
active mechanism is the responsible for the transport of 5-HTP
through the BBB.
[28] A. Becke, J. Chem. Phys. 98 (1993) 5648–5652.
[
[
[
[
[
29] A.D. McLean, G.S. Chandler, J. Chem. Phys. 72 (1980) 5639–5648.
30] W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys. 56 (1972) 2257–2261.
31] U. Singh, P. Kollman, J. Comput. Chem. 5 (1984) 129–145.
32] K. Kim, K.D. Jordan, J. Phys. Chem. 9 (1994) 10089–10094.
33] (a) S. Feller, Y. Zhang, R. Pastor, B. Brooks, J. Chem. Phys. 103 (1995) 4613–
Acknowledgements
4
621;
b) G.J. Martyna, D.J. Tobias, M.L. Klein, J. Chem. Phys 101 (1994) 4177–4189.
[34] M. Tuckerman, B. Berne, G. Martyna, J. Chem. Phys. 94 (1991) 6811–6815.
(
This work was supported with funds from Agencia Nacional de
Promoción Científica y Técnica (PICT (2008/310), associated to the
Project PRH 2007 N° 71). M.P. and M.F.M are members of the Re-
search Career from Consejo Nacional de Investigaciones Científicas
y Técnicas (R. Argentina). I.W. is recipient of a fellowship for Ph.D.
students of the Agencia Nacional de Promoción Científica y Técnica
[
35] U. Essmann, L. Perera, M. Berkowitz, T. Darden, H. Lee, L. Pedersen, J. Chem.
Phys. 103 (1995) 8577–8593.
[
36] M. Pickholz, L.F. Fraceto, E. de Paula, Int. J. Quantum Chem. 108 (2008) 2386–
2391.
[
[
[
37] D.J. Barlow, J.M. Thornton, J. Mol. Biol. 168 (1983) 867–885.
38] S. Kumar, R. Nussinov, J. Mol. Biol. 293 (1999) 1241–1255.
39] M.P. Aliste, J.L. MacCallum, D.P. Tieleman, Biochemistry 42 (2003) 8976–8987.
(
R. Argentina).
[40] S. Kumar, R. Nussinov, Biophys. J. 83 (2002) 1595–1612.
[
41] M.F. Lensink, C. Govaerts, J.M. Ruysschaert, J. Biol. Chem. 285 (2010) 10519–
0526.
42] C.M. Dunkin, A. Pokorny, P. Almeida, H.S. Lee, J. Phys. Chem. B 115 (2011)
188–1198.
1
References
[
1
[
1] H. Gijsman, R. Verkes, J. Schouten-Verhagen, R. Schoemaker, J. Van Gerven, R.
De Rijk, G. Van Kempen, Human Psychopharm. 15 (2000) 397–415.
2] E. Hamel, Cephalalgia 27 (2007) 1293–1300.
3] E.H. Turner, J.M. Loftis, A.D. Blackwell, Pharmacol. Ther. 109 (2006) 325–338.
4] D. Camfield, J. Sarris, M. Berk, Prog. Neuro-psychopharm. Biol. Psych. 35 (2011)
[
[
[
43] J. Ma, D. Dougherty, Chem. Rev. 97 (1997) 1303–1324.
44] S. Mecozzi, A. West, D. Dougherty, JACS 118 (1996) 2307–2308.
45] Norbert Ku cˇ erka, Y. Liu, N. Chu, H.I. Petrache, S. Tristram-Nagle, J.F. Nagle,
Biophys. J. 88 (2005) 2626–2637.
[
[
[
[
46] L. Cabeça, M. Pickholz, E. de Paula, A. Marsaioli, J. Phys. Chem. B 113 (2009)
8
87–895.
2365–2370.
[
[
[
[
5] T.C. Birdsall, Altern. Med. Rev. 3 (1998) 271–280.
6] S. Ohtsuki, Pharmac. Bull. 27 (2004) 1489–1496.
7] W. Pardridge, K. Hargreaves, J. Biol. Chem. 263 (1988) 19392–19397.
8] R. Boado, J. Li, M. Nagaya, C. Zhang, W. Pardridge, PNAS 96 (1999) 12079–
[
[
47] P. Gomes, P. Soares-da-Silva, Brain Res. 829 (1999) 143–150.
48] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graphics 14 (1996) 33–38.
1
2084.