Liposomal formulation of retinoids designed for enzyme triggered release

Article abstract of DOI:10.1021/jm100190c

The design of retinoid phospholipid prodrugs is described based on molecular dynamics simulations and cytotoxicity studies of synthetic retinoid esters. The prodrugs are degradable by secretory phospholipase A₂ IIA and have potential in liposomal drug delivery targeting tumors. We have synthesized four different retinoid phospholipid prodrugs and shown that they form particles in the liposome size region with average diameters of 94-118 nm. Upon subjection to phospholipase A₂, the lipid prodrugs were hydrolyzed, releasing cytotoxic retinoids and lysolipids. The formulated lipid prodrugs displayed IC₅₀ values in the range of 3-19 μM toward HT-29 and Colo205 colon cancer cells in the presence of phospholipase A ₂, while no significant cell death was observed in the absence of the enzyme.

Full text of DOI:10.1021/jm100190c

3782 J. Med. Chem. 2010, 53, 3782–3792
DOI: 10.1021/jm100190c
Liposomal Formulation of Retinoids Designed for Enzyme Triggered Release
Palle J. Pedersen,† Sidsel K. Adolph,| Arun K. Subramanian,† Ahmad Arouri,‡ Thomas L. Andresen,§ Ole G. Mouritsen,‡
†
Robert Madsen, Mogens W. Madsen, Gunther H. Peters, and Mads H. Clausen*
|
†
,†
€
†Department of Chemistry, Technical University of Denmark, Kemitorvet, Building 201 and 207, DK-2800 Kgs. Lyngby, Denmark,
‡Department of Physics and Chemistry, MEMPHYS-Center for Biomembrane Physics, University of Southern Denmark, Campusvej 55,
DK-5230 Odense M, Denmark, §Department of Micro- and Nanotechnology, Technical University of Denmark, DK-4000 Roskilde, Denmark, and
|LiPlasome Pharma A/S, Technical University of Denmark, Diplomvej 378, DK-2800 Kgs. Lyngby, Denmark
Received February 12, 2010
The design of retinoid phospholipid prodrugs is described based on molecular dynamics simulations and
cytotoxicity studies of synthetic retinoid esters. The prodrugs are degradable by secretory phospholipase
A2 IIA and have potential in liposomal drug delivery targeting tumors. We have synthesized four
different retinoid phospholipid prodrugs and shown that they form particles in the liposome size region
with average diameters of 94-118 nm. Upon subjection to phospholipase A2, the lipid prodrugs were
hydrolyzed, releasing cytotoxic retinoids and lysolipids. The formulated lipid prodrugs displayed IC50
values in the range of 3-19 μM toward HT-29 and Colo205 colon cancer cells in the presence of
phospholipase A2, while no significant cell death was observed in the absence of the enzyme.
Introduction
Retinoids,1 such as all-trans retinoic acid (ATRA,a Figure 1),
are known for their broad and diverse biological functions,
and various strategies have been explored to make retinoids
applicable as drugs in the treatment of diseases.2 One of the
biological functions of ATRA is its anticancer activity toward
a broad range of cancer types, like breast, prostate, and colon
cancer.3 For example, orally administrated ATRA is used
clinically in the treatment of leukemia.4 However, the oral
administration route is restricted by a relative low bioavail-
ability4c and a fast clearance from the bloodstream,5 and thus
Figure 1. Structure of the cytotoxic compounds ATRA and 2 and the
ATRA phospholipid prodrug (1) that is not hydrolyzed by sPLA2.7b.
alternative ways of administrating ATRA would be benefi-
cial. Intravenous administration is hampered by the low water
solubility of ATRA, but this can be circumvented by formu-
lating ATRA in liposome based drug delivery systems.6 Un-
fortunately, the liposomal formulation strategies have often
been plagued by problems with leakage, which have led to an
uncontrolled release of ATRA from the carrier system. Re-
cently, we have introduced new liposomal drug delivery
systems7 that address the formulation challenges of relatively
hydrophobic drugs. The drug delivery systems consist of
secretory phospholipase A2 (sPLA2) IIA sensitive phospholi-
pids and are constructed of drug-lipid prodrugs, in which the
lipophilic anticancer drugs are covalently attached to the
phospholipids. Despite successful incorporation of the well-
known anticancer drug chlorambucil in the sn-2 position,7a it
was later experienced that the desired delivery of free ATRA
was not feasible due to lack of sPLA2 activity on substrates
like 1 (Figure 1).7b ATRA is a rigid molecule and contains a
methyl substituent in close proximity to the carboxylic acid
moiety. These features contrast with naturally occurring fatty
acids, which are predominantly saturated and flexible mole-
cules without branching. Knowing that some esters and amides
of ATRA retain their activity,8 we decided to incorporate an
aliphatic C6-linker between ATRA and the lipid backbone
(Figure 2). To gain further insight into the substrate specificity
of the sPLA2 enzyme, hydrolysis of the lipid prodrugs was
investigated by molecular dynamics (MD) simulation, which
has earlier been demonstrated as a valuable tool for assessing
sPLA2 activity toward a range of substrates.9
*To whom correspondence should be addressed. Phone: þ45
45252131. Fax: þ45 45933968. E-mail: mhc@kemi.dtu.dk.
a Abbrevations: ATRA, all-trans retinoic acid; sPLA2, secretory
phospholipase A2; MD, molecular dynamics; RAR, retinoic acid recep-
tor; RXR, retinoic X receptor; rmsd, root-mean-square deviation; S,
substrate; PMB, p-methoxybenzyl; DDQ, 2,3-dichloro-5,6-dicyano-1,4-
benzoquinone; DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; DIAD, diiso-
propyl azodicarboxylate; NBS, N-bromosuccinimide; DLS, dynamic
light scattering; DPPG, 1,2-dipalmitoyl-sn-glycero-3-phosphoglycerol;
DSPG, 1,2-distearoyl-sn-glycero-3-phosphoglycerol; 1-O-DSPG, 1-O-
stearyl-2-stearoyl-sn-glycero-3-phosphoglycerol; POPG, 1-palmitoyl-2-
oleoyl-sn-glycero-3-phosphoglycerol; NBD-DPPE, 1,2-dipalmitoyl-sn-
glycero-3-phosphoethanolamine-N-[7-nitro-2-1,3-benzoxadiazol-4-yl];
MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight;
DHB, 2,5-dihydroxybenzoic acid; MTT, 3-(4,5-dimethylthiazolyl)-2,5-
diphenyltetrazolium bromide.
One major drawback in applying ATRA as a drug is its
nonselective activation of retinoic acid receptor (RAR) sub-
types (RARs, R, β, γ), retinoic X receptor subtypes (RXRs, R,
r
pubs.acs.org/jmc
Published on Web 04/20/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3783
Figure 2. The sPLA2 degradablephospholipidprodrugs3-6, withanaliphaticC6-linker incorporatedbetween the drug, R0COOH (ATRAor2),
and the lipid backbone.
β, γ), and/or the subtype RAR isoforms, R1, R2, β1-β5, γ1,
and γ2.3a,10 The lack of selectivity is believed to be responsible
for the severe side effects that have been observed during
chronic administration of ATRA,10 and therefore there is
much interest in discovering more selective RAR agonists.11
Targeting of RARβ2 has been found to have a suppressive
effect on human tumors,12 and recently 4-(4-octylphenyl)-
benzoic acid (2, Figure 1) was identified as a selective RARβ2
agonist.13 The lipophilicity of 2 makes it is a good candidate
for incorporation into a liposomal drug delivery system
through covalent attachment to the lipid backbone. However,
since Bonsen et al. have demonstrated that benzoic acid is not
released from the sn-2-position by sPLA2,14 direct attachment
Figure 3. Representation of the active site in sPLA2 with hydrogen
of 2 to the phospholipids was ruled out. Instead, we decided to
investigate how attachment of an aliphatic C6-linker would
bonds and ionic interactions indicated with dashed bonds. Key
protein residues are drawn in blue, the substrate in red, and the
influence the cytotoxicity of 2 and the sPLA2 activity toward
the corresponding prodrug.
calcium ion and the water molecule both in brown. The two gray
circles indicate the H-S region, where the overlap in dark gray
represents the water count region. Labels are shown in black, and
atom types indicated in parentheses refer to the Protein Data Bank
nomenclature.
The retinoid-lipid prodrugs are designed to include the
phosphatidylglycerol headgroup because human sPLA2 IIA
has strong affinity for negatively charged phospholipids.15
The lipid backbone of the prodrugs contains either an sn-1-
ester or an sn-1-etherfunctionality. UponactivationofsPLA2,
the prodrugs will be converted into the free drug and lysolipids
(Figure 2) and while sn-1-ether lysolipids have good metabolic
stability and are cytotoxic against many cell lines,16 sn-1-ester
lysolipids are rapidly metabolized (by e.g., lysophospholi-
pases) and have been dismissed as suitable lysolipid drug
candidates.16a Information about the cytotoxicityofsn-1-ester
lysolipids is limited and covers only lysophosphatidylcholine
lipids,17 but these lipids have shown to be slightly better
substrates for sPLA2 than sn-1-ether lipids.18 Therefore, we
found it interesting to synthesize both types of prodrugs in
order to study their relative hydrolysis rate and cytotoxicity.
time evolution of the root-mean-square deviation (rmsd) of
the CR atoms with respect to the protein structure obtained
after minimization. Stable rmsd data were obtained after 1-2ns
for the sPLA2-DPPG, sPLA2-3 and sPLA2-5 complexes
˚
reaching a plateau of ∼1.4 A (Supporting Information). rmsd
data for the sPLA2-1 complex were not stable, increasing to
˚
∼2.5 A after ∼6.5 ns (Supporting Information).
A prerequisite for successful hydrolysis is that a stable
Michaelis-Menten complex is obtained and that a water
molecule acting as a nucleophile can enter the catalytic cleft.
To monitor the stability of the Michaelis-Menten complex,
we have chosen distances according to their importance in
the calcium-dependent enzymatic reaction.19 The catalytic
mechanism has been identified by X-ray crystallography,
revealing that the catalytic device of sPLA2 is essentially
characterized by an aspartic acid-histidine dyad, a calcium-
binding site, and a water molecule acting as the nucleophile.
A schematic representation of the catalytic mechanism is
shown in Figure 3, indicating that the calcium ion (cofactor)
is coordinated to the D48 carboxylate groups (atoms: OD1
and OD2), the G29 carbonyl (O), the carbonyl (O22) in the
substrate (S), and the phosphate (O4) in S. Atom types given
in parentheses refer to the Protein Data Bank nomenclature.
Furthermore, the catalytic residue, H47, is stabilized by D91,
Molecular Dynamics Simulations
Molecular dynamics (MD) simulations were performed to
7b
understand why 1 is not hydrolyzed by sPLA2 and which
structural modification(s) of the prodrug would regain enzy-
matic hydrolysis. Simulations were carried out for sPLA2-
DPPG, sPLA2-1, and as discussed below also for sPLA2-3
and sPLA2-5. Simulations for each complex were repeated
five times to estimate the statistical uncertainties in the
calculated quantities. The stability of the simulations of the
different sPLA2 complexes were checked by computing the

3784 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Pedersen et al.
Figure 4. (A) ATRA induced structural perturbation of the Michaelis-Menten complex. Images taken at the beginning (top) and at the end
(bottom) of an sPLA2-1 simulation revealed that ATRA induced conformational changes (as evident from the G29(O)-Ca2þ-S(O22) distances
indicated by the triangles) leading to the distortion of the Michaelis-Menten complex. (B) 2D scatter plots of distances between selected residue
pairs important for stabilizing the Michaelis-Menten complex. (C) Histograms of relative water counts as a function of distance from both
H47ND1 and SC21 (H-S region; Figure 3).
which forms hydrogen bonds with Y51 and Y66. The water
molecule acting as the nucleophile enters the region between
H47ND1 and SC21. This region will be referred to as the H-S
region (Figure 3).
We have checked the stability of the Michaelis-Menten
complexes by monitoring the following distances during the
were similar to those extracted from sPLA2-DPPG simula-
tions (Figure 4). Contrarily, the relative water counts ex-
tracted from sPLA2-1 were very low when compared to the
other complexes. Because complexes of DPPG, 3, and 5 show
similar relative water counts and their Michaelis-Menten
complexes are stable, these results predict that 3 and 5 will be
hydrolyzed by sPLA2.
simulations: Ca2þ-G29O, Ca2þ-D48OD1, Ca2þ-SO22, SO22
-
,
G29HN, H47HE2-D91OD1, H47ND1-SC21, D91OD2-Y66HH
and D91OD1-Y51HH. The time evolution of these distances is
provided in the Supporting Information, and 2D scatter plots
are shown in Figure 4.
Synthesis of Retinoid Esters
The linker moietytert-butyl6-hydroxyhexanoate (11)20 was
coupled with ATRA by a Mitsunobu reaction (Scheme 1),21
which we previously have demonstrated as a reliable method
for forming esters of ATRA.7b The following tert-butyl
deprotection led to degradation of the ATRA skeleton under
For prodrug1, nostable Michaelis-Menten complex could
be observed, and hence in accordance with experimental data,
nohydrolysiscanoccur. The rigid nature of the ATRAester at
the sn-2-position of the prodrug causes distortion of several
distances involved in the Michaelis-Menten complex (Figure 4).
For instance, distances Ca2þ-SO22 and G29HN-SO22 are signi-
ficantly larger for sPLA2-1than sPLA2-DPPG(Figure4).The
distortion is also highlighted by the images taken at the begin-
ning and upon completion of an sPLA2-1 simulation, respec-
tively (Figure 4). Introduction of a C6-linker between the
phospholipid backbone and 1 was sufficient to confer stability
of the Michaelis-Menten complex in sPLA2-3 and sPLA2-5.
Although a stable Michaelis-Menten complex could be
obtained by introducing a C6-linker, the question remained if
water can reach the H-S region (Figure 3) in sPLA2 com-
plexed with 3 or 5. To quantify the probability that water
molecules enter the H-S region, we counted the number of
22
acidic conditions (5% TFA or ZnBr2 in CH2Cl2 or HF in
23
MeCN), but using 2,6-lutidine and TMSOTf in CH2Cl2
afforded the carboxylic acid 9 in a good yield (Scheme 1). The
ATRA ester 9 was isolated as yellow crystals and was stable
when stored at -20 °C under aninertatmosphere, whereas the
oily tert-butyl ester of 9 decomposes upon less than one month
of storage. This highlights the importance of storing ATRA-
analogues as solids in order to avoid decomposition. The
corresponding linker-molecule of 224 was synthesized by a
Steglich coupling25 with 11 followed by a deprotection of the
tert-butyl ester with TFA in CH2Cl2 (Scheme 1). In contrast to
the ATRA-analogues, we never observed any decomposition
of the synthetic derivatives of 2 neither as oils nor solids. The
cytotoxicity of ATRA, 2, 9, and 10 were evaluated in three cell
lines, and the retinoid esters had activities comparable to the
acids (vide infra).
˚
˚
water molecules within certain distances (d = 3 A - 6 A;
Δd = 0.5 A) from both H47ND1 and SC21, and the counts were
˚
normalized by dividing the respective water counts by the
˚
water count determined at 6 A (Figure 4). Averages and
standard deviations of normalized water counts were calcu-
lated from the five simulations of each complex. The relative
water counts extracted from sPLA2-3 and -5 simulations
Synthesis of Lipid Prodrugs
The cytotoxicity of the esters 9 and 10 and the results from
the MD simulations prompted us to proceed with the synthesis

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3785
Scheme 1. Synthesis of the Retinoid Estersa
a Reagents: (a) (i) diisopropyl azodicarboxylate (DIAD), PPh3, THF; (ii) TMSOTf, 2,6-lutidine, CH2Cl2; (b) (i) dicyclohexylcarbodiimide (DCC),
DMAP, CH2Cl2; (ii) TFA, triisopropylsilane, CH2Cl2.
Scheme 2. Synthesis of the Lipid Precursors 15 and 18a
a Reagents: (a) (i) tetrazole, CH2Cl2, MeCN; (ii) tBuOOH; (iii) DDQ, H2O, CH2Cl2; (iv) 14, DCC, DMAP, CH2Cl2; (v) DDQ, H2O, CH2Cl2; (b) (i) 14,
EDCI, DMAP, CH2Cl2; (ii) NBS, DMSO, THF, H2O; (c) (i) 13, tetrazole, CH2Cl2, MeCN; (ii) tBuOOH; (iii) DDQ, H2O, CH2Cl2.
Scheme 3. Synthesis of the sn-1-Ether and sn-1-Ester prodrugs 3, 4, 5, and 6a
a Reagents: (a) (i) 15, ATRA, DIAD, PPh3, THF or 18, ATRA, DCC, DMAP, Et3N, Et2O; (ii) DBU, CH2Cl2; (iii) HF, H2O, CH2Cl2, MeCN; (b) (i) 2,
DCC, DMAP, CH2Cl2; (ii) DBU, CH2Cl2; (iii) HF, H2O, CH2Cl2, MeCN.
of the corresponding sn-1-ether and sn-1-ester lipid prodrugs
3-6. The backbone of the sn-1-ether prodrugs was con-
structed by a tetrazole mediated coupling of alcohol 12 and
phosphoramidite 13, followed by an oxidation of the phos-
phite to the phosphate with tBuOOH (Scheme 2).7a Deprotec-
tion of the p-methoxybenzyl (PMB) group in moist CH2Cl2
with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) af-
forded the secondary alcohol, which was reacted with 6-(4-
methoxybenzyloxy)-1-hexanoic acid (14) to give an ester. The
newly introduced PMB-group was then removed, affording
alcohol 15 in an overall yield of 51% (Scheme 2). Secondary
alcohol 16, synthesized from 2,3-O-isopropylidene-sn-glycerol
according to the procedures by Gaffney et al26 and Burgos
et al.,27 served as the starting material for the sn-1-ester pro-
drugs (Scheme 2). Diester 17 was obtained by coupling of 14
and 16, and the following TBDMS deprotection was best
achieved using N-bromosuccinimide (NBS) in a mixture of
DMSO, THF, and H2O (Scheme 2),27,28 while alternative
conditions like Bu4NF and imidazole in THF29 or HF in
MeCN led to a significant degree of acyl migration. Mosher
ester analysis30 of 17 showed that the enantiomeric purity was
>95%. The phosphate headgroup was attached in the same
way as for the sn-1-ether lipids, and PMB deprotection gave
the primary alcohol 18 (Scheme 2). Phospholipids 15 and 18
were converted to the desired prodrugs of ATRA and 2
(Scheme 3). The ATRA moiety was introduced either by the
Mitsunobu reaction or the Steglich coupling, and the final
prodrugs 3 and 4 were obtained after removal of the cyano-
ethyl group with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)
andTBDMS-deprotectionmediatedbyHFinMeCN/CH2Cl2/
H2O. Prodrugs 5 and 6 were accessed by coupling between 2
and the primary alcohols 15 and 18, after which the remaining
chemistry was similar to the synthesis of 3 and 4 (Scheme 3). 31P
NMR of the synthesized prodrugs showed one signal resonat-
ing between -1 ppm and 2 ppm, demonstrating high diaster-
eomeric and regioisomeric purity (>95%).
Biophysical Characterization
The lipid prodrugs (3-6) were formulated as liposomes by
extrusionin HEPES buffer using the dry lipid film technique,31
yielding clear solutions. The particle size of the formulated

3786 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Pedersen et al.
Table 1. Measurement of Particle Size with DLS
Differential scanning calorimetry scans (15-65 °C) of the
lipid solutions of 3 and 5 displayed no gel-to-liquid crystalline
or any other thermotropic phase transition in the tested
temperature range (data not shown), indicating that the lipid
bilayers are in a fluid state. This was not surprising taking into
account the bulky and stiff nature of the substituents in the
sn-2-position, which hampers well-ordered chain packing.
The ability of sPLA2 to hydrolyze the formulated lipids at
37 °C was investigated with matrix-assisted laser desorption/
ionization time-of-flight (MALDI-TOF) MS. MALDI-TOF
MS is a fast and sensitive technique which has been shown to
be a reliable tool for detection of lipids.34 In Figure 6, the
obtained MS data for the sn-1-ether prodrug 3 subjected to
purified sPLA2 from snake (Agkistrodon piscivorus piscivorus)
venom is shown. Gratifyingly, and as evident from the spectra
(Figure 6), the prodrug (M þ Hþ) is consumed by the enzyme
and the desired constituents, lysolipid 7 (M þ Hþ) and ATRA
ester 9 (M þ Naþ), are released. The MALDI-TOF MS
analysis of the sn-1-ether prodrugs 3 and 5 and the sn-1-ester
prodrugs 4 and 6 revealed that all of the prodrugs were
degraded by the enzyme (Table 2) and the desired molecules
were released (see Supporting Information). The observation
that the sn-1-ether prodrugs were consumed by sPLA2 to the
same extent as the sn-1-ester prodrugs illustrate and confirm
that sn-1-ether phospholipids are excellent substrates for
sPLA2. On the basis of the MALDI-TOF MS analysis of
the sPLA2 activity on the various prodrugs, we can conclude
that the incorporation of an aliphatic C6-linker between
ATRA and the lipid backbone resolved the issue of sPLA2
activity on ATRA-lipid prodrugs. The experimental observa-
tions are in agreement with the outcome from the MD
simulations, illustrating the power and potential of using
MD simulations in the design of sPLA2 degradable lipid
prodrugs. To rule out that the degradation of the prodrugs
was caused by nonenzymatic hydrolysis, a sample of each
prodrug was subjected to the reaction conditions in the
absence of sPLA2. As evident from the spectra in Figure 6
(and in Supporting Information), no degradation of the
prodrugs was observed within 48 h and none of the release
products were detectable. Finally, we compared the enzyme
activity of the purified sPLA2 from snake (Agkistrodon pisci-
vorus piscivorus) venom with the commercially available
sPLA2 from snake (Naja mossambica mossambica) venom,
and as can be seen in Table 2, the two enzymes perform
equally well on all substrates.
particle size
prodrug
diameter (nm)
polydispersity index
3
4
5
6
94
97
0.08
0.13
0.16
0.05
118
118
Figure 5. Time-dependent fluorescence from NBD labeled lipids
quenched by dithionite.
lipids was measured by dynamic light scattering (DLS). The
DLS analysis revealed that all of the lipids were able to form
particles with a diameter around 100 nm (Table 1 and Sup-
porting Information) and with a low polydispersity, indicating
the formation of unilamellar vesicles. The formation of a uni-
lamellar bilayer is further supported by the fluorometric
procedure described by McIntyre et al.,32 which was used to
study the permeability. Vesicles formed from prodrug 5 was
compared to vesicles of 1,2-dipalmitoyl-sn-glycero-3-phos-
phoglycerol (DPPG) and 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphoglycerol (POPG). The lipids were formulated with
1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-[7-ni-
tro-2-1,3-benzoxadiazol-4-yl] (NBD-DPPE, 0.15mol %). The
fluorescence of the NBD group was monitored over time
before and after the addition of dithionite (S2O4-2), which
functions as a quenching agent. At the end of the experiment,
the detergent Triton X-100 was added to solubilize the lipid
vesicles and cause 100% quenching (see Figure 5). As evident
from the figure, the decrease in the fluorescence of the NBD-
labeled gel-state DPPG occurs in two distinct steps, an
immediate and fast one completed after 20 min followed by
much slower decay of fluorescence. We anticipate that the
quick drop in fluorescence signal corresponds to quenching of
the dye in the outer leaflet and is dependent on the fraction of
lipids exposed to the surrounding, i.e., vesicle size and lamel-
larity. The slower second step is due to quenching of the dye in
the inner leaflet, which is controlled by membrane permeabil-
ity to dithionite. In the case of the fluid POPG vesicles, the
quenching of both leaflets occurs simultaneously, resulting
effectively in a single exponential-like decay in the NBD
fluorescence. A comparable behavior for gel-state and fluid
membranes has been reported previously.32,33 The quenching
of the NBD-labeled vesicles for prodrug 5 resembles that of
DPPG, demonstrating that 5 forms a bilayer membrane,
however seemingly more permeable to dithionite than the
vesicles of DPPG.
Fawzy et al.35 and Hope et al.36 have demonstrated two
decades ago that ATRA is an inhibitor for sPLA2, reporting
IC50 values of <50 and 10 μM respectively. However, we have
later shown that DPPG is fully hydrolyzed by sPLA2 in the
presenceof 0.1 and1.0equiv of ATRA,7b proving that the lack
of hydrolysis for the ATRA-lipid prodrug 1 is not a conse-
quence of inhibition by ATRA. These observations are
further supported by the present work, in which it was found
that 9 (IC50 = 8 μM, Ki = 6 μM, see Supporting Information)
inhibits sPLA2 at the same level as ATRA (IC50 = 15 μM,
Ki = 11 μM), and whereas the ATRA-lipid prodrug 1 is not
consumed by the enzyme, the prodrugs 3 and 4 are fully
degraded. This demonstrates that under the conditions used,
the released ATRA ester 9 does not inhibit sPLA2, and the
difference in degradation of the prodrugs therefore solely
relies on the ability of sPLA2 to hydrolyze the different
prodrugs. Additionally, Cunningham et al. have reported Ki
values <100 nM37 for potent sPLA2 inhibitors, illustrating
that ATRA and 9 are weak inhibitors.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3787
Figure 6. MALDI-TOF MS monitoring of snake (Agkistrodon piscivorus piscivorus) venom sPLA2 activity on the sn-1-ether prodrug 3. The
spectra demonstrate that the prodrug 3 (top left) is consumed and that the lysolipid 7 (top right) and the ATRA ester 9 (bottom) are released.
Table 2. Measurement of sPLA2 Activity on the Prodrugs by MALDI-
TOF MS
Table 3. IC50 (μM) Values for the Retinoids in Three Cancer Cell Linesa
MT-3
IC50 (μM)
HT-29
IC50 (μM)
Colo205
IC50 (μM)
a
hydrolysis by sPLA2
compd
Agkistrodon piscivorus
piscivorus
Naja mossambica
mossambica
ATRA
9
30 ( 4
17 ( 1
51 ( 3
14 ( 1
4.3 ( 0.2
3.6 ( 1.8
>200
37 ( 1
17 ( 2
>200
prodrug
2
10
3
4
5
6
þ
þ
nd
þ
>200
127 ( 18
b
b
þ
þ
a Cytotoxicity was measured using the MTT assay as cell viability 48 h
after incubation with the indicated substances for 24 h and shown by
mean ( SD (n = 3).
þ
þ
a Determined by MALDI-TOF MS after 48 h incubation at 37 °C
with purified sPLA2 from snake (Agkistrodon piscivorus piscivorus or
Naja mossambica mossambica) venom. b After 48 h, a small signal for
prodrug 5 remains; nd = not determined.
Table 4. IC50 (μM) Values for the Prodrugs 3-6, the Lysolipids 7 and 8,
DSPG, and 1-O-DSPG in HT-29 and Colo205 Cancer Cell Linesa
b
HT-29
IC50 (μM)
HT-29 þ sPLA2
IC50 (μM)
Colo205
IC50 (μM)
compd
Cytotoxicity
3
4
5
6
7
8
>200
>200
>200
>200
11 ( 6
7 ( 1
>200
>200
>200
c
7 ( 2
6 ( 1
3 ( 1
8 ( 1
nd
16 ( 4
12 ( 3
7 ( 6
19 ( 4
25 ( 2
22 ( 3
54 ( 7
33 ( 4
>200
c
The cytotoxicity of ATRA, 2, 9, and 10 were evaluated in
MT-3 breast carcinoma, HT-29 colon carcinoma, and
Colo205 colon adenocarcinoma cell lines (Table 3). Interest-
ingly, the retinoid esters were either more active than ATRA
and 2 or equal in potency in the tested cell lines. A possible
explanation for the enhancedactivityofthesemolecules is that
the derivatization increases the lipophilicity of the drugs,
augmenting transport over the cell membrane. RARβ2 ago-
nists 2 and 10 showed very little activity against HT-29 cells,
and presumably this is because growth inhibition in this cell
line is induced by RARR agonists38 and 2 has a low affinity for
that receptor.13 The result is also a strong indication that the
cytotoxicity of 10 in MT-3 and Colo205 originate from
RARβ2 activation.
nd
DSPG
1-O-DSPG
25 ( 11
9 ( 2
nd
C
sPLA2
17H35COOH
c
a Cytotoxicity was measured using the MTT assay as cell viability 48 h
after incubation with the indicated substances for 24 h and shown by
mean ( SD (n g 3); nd = not determined. b Snake (Agkistrodon piscivorus
piscivorus) venom sPLA2 was added to a final concentration of 5 nM. c No
change in cell viability was observed after 24 h.
The cytotoxicity of the prodrugs 3-6 was investigated in
HT-29 colon carcinoma and Colo205 colon adenocarcinoma
cells. HT-29 colon carcinoma cells do not secrete sPLA2,
which allowed us to test the activity in the presence and
absence of sPLA2. As evident from Table 4 and the dose-
response curve for 3 and 5 (Figure 7), none of the prodrugs
were able to induce significant cell death in the absence of
sPLA2, whereas upon sPLA2 addition, all of the prodrugs
displayed IC50 values below 10 μM in HT-29 cells and complete
cell death was obtained when higher concentrations were
applied (see Figure 7). Evidently, the cytotoxicity is induced
by sPLA2 triggered breakdown of the prodrugs into 9 or 10 and
the lysolipids. Interestingly, we observed that the sn-1-ester
prodrug 6 in the presence of sPLA2 displayed almost the same
cytotoxicity toward HT-29 cells as the corresponding sn-1-ether
prodrug 5, and taking the low activity of 10 in HT-29 cells

3788 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Pedersen et al.
free lysolipids but similar to 9, indicating that the majority of the
cytotoxicity can be ascribed to the released ATRA ester 9. For
prodrugs 5 and 6, the IC50 values indicate a cumulative effect
because these prodrugs are more potent than both of the released
cytotoxic compounds by themselves, demonstrating the advan-
tage of the prodrug formulation.
Conclusion
In the present study we have successfully synthesized lipid
prodrugs of retinoids and demonstrated that they can be
formulated as liposomes that are degraded by sPLA2. On
the basis of MD simulations, we incorporated a C6-linker
between the glycerol backbone of the lipids and the retinoids
to overcome the lack of enzymatic activity stemming from the
rigid nature of ATRA and 2. The resulting prodrugs 3-6 were
hydrolyzed completely by sPLA2 within 24 h and were
demonstrated to be cytotoxic to cancer cells in the presence
of the enzyme.
The liposomal formulation is very well suited for active
substances with low water solubility, and the novel prodrugs
present a solution to formulation of retinoids for applications
in cancer therapy. Furthermore, our findings underpin the
overall strategy of covalently linking cytotoxic compounds to
lipids and widen the scope of the approach because even very
sterically hindered substrates can now be effectively incorpo-
ratedintoliposomesandreleasedthroughanenzyme-triggered
degradation.
Figure 7. Dose-response curves for the treatment of HT-29 cells
with the sn-1-ether prodrugs 3 (b), 3 þ sPLA2 (9), 5 ([), and 5 þ
sPLA2 (2).
Experimental Section
General. Starting materials, reagents, and solvents were pur-
chased from Sigma-Aldrich and used without further purifica-
tion. POPG, DPPG, DSPG, 1-stearoyl-2-hydroxy-sn-glycero-3-
phospho-(10-rac-glycerol) (8), and NBD-DPPE were purchased
from Avanti Polar Lipids (Alabaster, AL). The purified snake
venom sPLA2 from Agkistrodon piscivorus piscivorus was do-
nated by Dr. R. L. Biltonen (University of Virginia), and sPLA2
from Naja mossambica mossambica was purchased from Sigma-
˚
Aldrich. CH2Cl2 was dried over 4 A molecular sieves, and
Figure 8. Dose-response curves for the treatment of Colo205 cells
with sn-1-ester prodrugs 4 (9) and 6 (2).
THF was dried over sodium/benzophenone and distilled before
use. Evaporation of solvents was done under reduced pressure
(in vacuo). TLC was performed on Merck aluminum sheets
precoated with silica gel 60 F254. Compounds were visualized by
charring after dipping in a solution of Ce(IV) (6.25 g of (NH4)6-
Mo7O24 and 1.5 g of Ce(SO4)2 in 250 mL of 10% aq H2SO4) or
an ethanolic solution of phosphomolybdenic acid (48 g/L).
Flash column chromatography was performed using Matrex
(Table 3) into account, these results indicate that sn-1-ester
lysolipids (8) contribute with the same degree of potency as sn-
1-ether lysolipids (7). These findings were verified when the free
lysolipids 7 (see Supporting Information) and 8 were tested
against HT-29 cells. Both lysolipids displayed IC50 values close
to 10 μM, and similar results were obtained for DSPG and 1-O-
stearyl-2-stearoyl-sn-glycero-3-phosphoglycerol (1-O-DSPG, see
Supporting Information) in the presence of sPLA2 (Table 4). We
conclude that even though sn-1-ester lysolipids generally are
rapidly metabolized, under these in vitro conditions, there is no
significant metabolism of the lipid backbone and therefore we
observe an equal potency of the two lysolipids. With the control
experiments in hand, it is possible to determine the origin of the
cytotoxicity of the prodrugs in HT-29 cells, and as evident from
Tables 3 and 4, the majority of the activity arises from the
lysolipids in prodrug 5 and 6 while for the prodrugs 3 and 4 there
appear to be an equal contribution from the lysolipids
(7 and 8) and the ATRA ester 9. Colo205 cells express sPLA2,
and encouragingly, the four prodrugs induce cell death with IC50
values below 20 μM (Table 4) and complete cell death was
obtained when higher concentrations were applied (see Figure 8),
indicating that the secreted sPLA2 in Colo205 cells provides the
desired hydrolysis and release of the anticancer agents. Addi-
tionally, the prodrugs 3 and 4 displayed IC50 values below the
˚
60 A silica gel. The purity of all tested compounds was found to
be >95% by HPLC (see Supporting Information). HPLC was
performed on a Waters Alliance HPLC equipped with a DAD,
using a LiChrospher Si 60 column and eluting with water/
isopropanol/hexane mixtures.39 NMR spectra were recorded using
a Bruker AC 200 MHz spectrometer, a Varian Mercury 300 MHz
spectrometer, or a Varian Unity Inova 500 MHz spectrometer.
Chemical shifts were measured in ppm and coupling constants in
Hz, and the field is indicated in each case. IR analysis was carried out
on a Bruker Alpha FT-IR spectrometer, and optical rotations were
measured with a Perkin-Elmer 341 polarimeter. HRMS was re-
corded on an Ionspec Ultima Fourier transform mass spectrometer.
Molecular Dynamics Simulations. The crystal structures of
bee venom (Apis Mellifera) phospholipase A2 complexed with
the transition-state analogue, L-1-O-octyl-2-heptylphosphonyl-
sn-glycero-3-phosphoethanolamine (diC8(2Ph)PE),19 resolved
˚
to 2.0 A, and human phospholipase A2 IIA complexed with 6-
phenyl-4(R)-(7-phenyl-heptanoylamino)-hexanoic acid,40 resol-
ved to 2.1 A, were obtained from the Protein Data Bank41 (entry
˚
codes: 1poc and 1kqu, respectively). The initial modeling step

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3789
involved placing diC8(2Ph)PE into the binding cleft of human
phospholipase A2 IIA, as described previously.18 A previously
assembled sPLA2-substrate complex18 was used as the template
to place lipid prodrugs into the active site. The optimized lipid
prodrugs were manually overlapped with the pre-existing sub-
strate, which was subsequently deleted. NAMD software42 with
the Charmm27 all-atom parameter set and TIP3 water model
was used for all simulations.43 The lipid prodrugs were first
energy-minimized for 1000 steps before using them in simula-
tions. Missing force field parameters for the prodrug molecules
were taken from similar atom types in the CHARMM27 force
field. The sPLA2-lipid complexes were solvated using the
program SOLVATE.44 Eighteen water molecules were ran-
domly replaced with chloride ions to neutralize the systems.
Ultimately, the system contained ∼4900 water molecules in a
(12 μL, 71 μM)) venom sPLA2. Sampling was done after 0, 2, 24,
and 48 h by collecting 100 μL of the reaction mixture and rapidly
mixing it with a solution of CHCl3/MeOH/H2O/AcOH 4:8:1:1
(0.5 mL) in order to stop the reaction. The mixture was washed
with water (0.5 mL), and the organic phase (80 μL) was isolated
by extraction and then concentrated in vacuo. The extract was
mixed with 9 μL of 2,5-dihydroxybenzoic acid (DHB) matrix
(0.5 M DHB, 2 mM CF3COONa, 1 mg/mL DPPG in MeOH),
and 0.5 μL of this mixture was used for the MS analysis.
Cytotoxicity. Colon cancer HT-29 cells were cultured in
McCoy’s 5A medium in the presence of 10% fetal calf serum
and 1% Pen-Strep (InVitrogen). Breast cancer MT-3 and colon
cancer Colo205 cells were cultured in RPMI 1640 supplemented
with 10% fetal calf serum and 1% Pen-Strep in a humidified
atmosphere containing 5% CO2. Colo205 cells secrete sPLA2-
IIa, however as the cytotoxicity assay was made with low cell
density and for a short incubation period, inadequate concen-
trations were reached in the medium, which is why conditioned
medium was used. Conditioned medium was made as follows:
Colo205 cells were grown to confluency for 72 h, at which point
the cells were pelleted and the medium, which contained approx-
imately 100 ng/mL sPLA2 IIA, was collected and used for the
assay. Cells were plated in 96-well plates at a density of 1 ꢀ 104
cells per well for HT-29 and MT-3 cells and 2 ꢀ 104 cells per well
for Colo205 cells, 24 h prior to addition of the tested compound.
The retinoids (ATRA, 2, 9, 10, and stearic acid) were solubilized
in DMSO and water (final DMSO concentration e0.5%).
Liposomes were diluted in PBS, and initial lipid concentrations
in the liposome solutions were determined by phosphorus
analysis.50 After 24 h of incubation, the substances were removed
and the cells were washed and incubated in complete medium for
another 48 h. Cytotoxic activity was assessed using a standard
3-(4,5-dimethylthiazolyl)-2,5-diphenyltetrazolium bromide (MTT)
assay (Cell Proliferation Kit I, Roche, Germany).51 Cell viability is
expressed as percentage reduction of incorporated MTT.
1-O-Octadecyl-2-(6-(all-trans-retinoyloxy)-hexanoyl)-sn-glycero-
3-phospho-(S)-glycerol (3).Alcohol 15 (13 mg, 0.014 mmol), ATRA
(6 mg, 0.019 mmol), and PPh3 (9 mg, 0.043 mmol) were dissolved
in THF (1.0 mL). DIAD (11 μL, 0.043 mmol) was added, and the
reaction mixture was stirred at 20 °C for 1 h and then concen-
trated in vacuo and purified by flash column chromatography
(heptane/EtOAc 2:1 then heptane/EtOAc 1:2) to give 13 mg that
was dissolved in CH2Cl2 (1.0 mL) and DBU (2 μL, 0.012 mmol)
was added. The reaction mixture was stirred for 30 min at 20 °C
and then purified directly by flash column chromatography
(heptane/EtOAc 1:1 then CH2Cl2/MeOH 9:1) to afford 8 mg
that was dissolved in MeCN (0.9 mL) and CH2Cl2 (0.3 mL) and
cooled to 0 °C. Aqueous HF (40%, 30 μL) was added dropwise,
and the reaction mixture was allowed to reach 20 °C. After 3.5 h,
the reaction was quenched by dropwise addition of MeOSiMe3
(0.3 mL) and the mixture was stirred for 30 min, after which
NaHCO3 (6 mg, 0.07 mmol) was added and the mixture was
concentrated in vacuo and purified by flash column chromato-
graphy (CH2Cl2/MeOH 10:1 then CH2Cl2/MeOH 4:1) to afford
4 mg (31% over 3 step) of 3 as a yellow amorphous solid. Rf =
0.44 (CH2Cl2/MeOH 4:1). 1H NMR (500 MHz, CDCl3/CD3OD
4:1): δ 7.02 (dd, J = 15.0, 11.4 Hz, 1H), 6.32-6.27 (m, 2H), 6.15
(d, J = 11.4 Hz, 1H), 6.14 (d, J = 16.2 Hz, 1H), 5.77 (s, 1H),
5.18-5.13 (m, 1H), 4.11 (t, J = 6.6 Hz, 2H), 4.03-3.94 (m, 2H),
3.94-3.90 (m, 2H), 3.82-3.77 (m, 1H), 3.62 (t, J = 5.2 Hz, 2H),
3.60-3.57 (m, 2H), 3.49-3.39 (m, 2H), 2.39-2.33 (m, 5H),
2.05-2.00 (m, 5H), 1.72 (s, 3H), 1.71-1.60 (m, 6H), 1.57-1.51
(m, 2H), 1.50-1.45 (m, 2H), 1.45-1.40 (m, 2H), 1.31-1.23 (m,
30H), 1.03 (s, 6H), 0.88 (t, J = 6.9 Hz, 3H). 13C NMR (50 MHz,
CDCl3/CD3OD 4:1): δ 174.0, 168.0, 153.7, 140.0, 138.0, 137.7,
135.3, 131.7, 130.3, 129.8, 129.2, 118.5, 72.2 (2C), 71.2, 69.4, 66.7,
64.5, 64.1, 62.7, 40.0, 34.5 (2C), 33.4, 32.3, 30.1, 29.7 (12C), 29.2
(2C), 28.7, 26.4, 25.9, 24.9, 23.0, 22.0, 19.6, 14.3, 14.1, 13.3. IR
(neat) 3325, 2923, 2853, 1733, 1709, 1609, 1584, 1457, 1259, 1064,
798 cm-1; m/z (M þ Hþ) 917.58.
3
˚
simulation cell of dimensions 52.7 ꢀ 51.7 ꢀ 67.3 A . An initial
energy minimization of the complex was carried out for 1000
steps, and the system was further minimized for 100, 200, 300,
400, and 500 steps to generate different starting structures.
Periodic boundary conditions were applied in x, y, and z
directions. Initial MD simulations were conducted for ∼100
ps in which each system was slowly heated to 300 K. Simulations
were carried out for 10 ns in the NPT ensemble, i.e., at constant
number of atoms (N), pressure (P), and temperature (T). A time
step of 1 fs and an isotropic pressure of 1 atm using the Langevin
piston method45 were employed. Electrostatic forces were cal-
culated using the particle mesh Ewald method with a uniform
˚
46
˚
grid spacing of ∼1 A. A 12 A cutoff was used for terminating
the van der Waals interactions in combination with a switching
˚
function starting at 10 A. Analysis of the trajectories were per-
formed using the Visualization Molecular Dynamics software
suite.47
Liposome Preparation and Particle Size Determination. The
lipid prodrugs 3-6 were dissolved in CHCl3 in a glass tube
and dried under vacuum for 15 h to form a thin film. The lipid
prodrugs (2 mM) were solubilized by addition of aqueous buffer
(0.15 M KCl, 30 μM CaCl2, 10 μM EDTA, 10 mM HEPES, pH
7.5) and vortexed periodically over 1 h at 20 °C. Subsequently,
the solutions were extruded through a 100 nm polycarbonate
cutoff membrane (20-30 repetitions) using a Hamilton syringe
extruder (Avanti Polar Lipids, Birmingham, AL). The particle
size distribution of the formulated lipids was measured by
DLS. The DLS measurements were obtained using a Zetasizer
nano particle analyzer (ZS ZEN3600, Malvern Instrument,
Westborough, MA).
Permeability Assay. The permeability of the lipid vesicles was
examined using the procedure described by McIntyre et al.32
The lipid vesicles were labeled with NBD-DPPE (0.15 mol %)
and prepared in an aqueous buffer (10 mM HEPES, 1 M sucrose,
150 mM NaCl, pH 7.4) by sonication for 30 min at 45 °C and then
extrusion through a 100 nm filter at 45 °C. Sucrose was added
to minimize the osmolarity difference between the external solu-
tion and the vesicle interior that will arise after the addition of
dithionite.48 The experiments were performed in duplicates or
more using FLUOstar Omega microplate reader (BMG LABTECH)
in a 96-well microplate at 20 °C (150 μL). The lipid samples
(100 μM) were allowed to equilibrate at 20 °C before the
addition of sodium dithionite (Na2S2O4) solution (33 mM),
which was used as a quenching agent. Because of the high
instability of dithionite,49 the dithionite solution was freshly
prepared prior to use. The NBD fluorescence was observed at
520 nm (ex 485 nm). After 85 min, Triton X-100 was added to
the medium (0.3 wt %).
sPLA2 Activity Measurements Monitored by MALDI-TOF
MS. The formulated lipid prodrugs (0.40 mL, 2 mM) were
diluted in an aqueous buffer (2.1 mL, 0.15 M KCl, 30 μM CaCl2,
10 μM EDTA, 10 mM HEPES, pH 7.5), and the mixture was
stirred at 37 °C in a container protected from light. The catalytic
reaction was initiated by addition of snake (Agkistrodon pisci-
vorus piscivorus (20 μL, 42 μM) or Naja mossambica mossambica

3790 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Pedersen et al.
1-Octadecanoyl-2-(6-(all-trans-retinoyloxy)-hexanoyl)-sn-glycero-
3-phospho-(S)-glycerol (4). Alcohol 18 (47 mg, 0.052 mmol), ATRA
(31 mg, 0.104 mmol), DMAP (13 mg, 0.104 mmol), and Et3N
(72μL, 0.52 mmol) were dissolved inEt2O (1.2 mL). DCC (27mg,
0.13 mmol) was added, and the reaction mixture was stirred at
20 °C for 22 h and then concentrated in vacuo and purified by
flash column chromatography (toluene then toluene/EtOAc 5:1)
to give 32 mg that was dissolved in CH2Cl2 (2.2 mL) and DBU
(4.2 μL, 0.028 mmol) was added. The reaction mixture was stirred
for 30 min at 20 °C and then purified directly by flash column
chromatography (heptane/EtOAc 1:1 then CH2Cl2/MeOH 9:1)
toafford 21 mgthat was dissolved inMeCN(1.5 mL) and CH2Cl2
(0.5 mL) and cooled to 0 °C. Aqueous HF (40%, 90 μL) was
added dropwise, and the reaction mixture was allowed to reach
20 °C. After 3.5 h, the reaction was quenched by dropwise
addition of MeOSiMe3 (0.3 mL) and the mixture was stirred
for 30 min, after which NaHCO3 (6 mg, 0.07 mmol) was added
and the mixture was concentrated in vacuo and purified by flash
column chromatography (CH2Cl2/MeOH 10:1 then CH2Cl2/
MeOH 4:1 then CH2Cl2/MeOH/H2O 65:25:1) to afford 15 mg
(31% over 3 step) of 4 as a yellow amorphous solid. Rf = 0.56
(CH2Cl2/MeOH 5:1). 1H NMR (500 MHz, CDCl3/CD3OD 4:1):
δ 7.02 (dd, J = 15.0, 11.4 Hz, 1H), 6.32-6.27 (m, 2H), 6.15 (d,
J = 11.4 Hz, 1H), 6.15 (d, J = 16.2 Hz, 1H), 5.77 (s, 1H),
5.25-5.20 (m, 1H), 4.39 (dd, J = 12.0, 3.3 Hz, 1H), 4.16 (dd, J =
12.0, 6.6 Hz, 1H), 4.11 (t, J = 6.6 Hz, 2H), 3.99-3.91 (m, 4H),
3.83-3.78 (m, 1H), 3.63 (t, J = 5.1 Hz, 2H), 2.36 (t, J = 7.5 Hz,
2H), 2.34 (s, 3H), 2.31 (t, J = 7.6 Hz, 2H), 2.05-2.00 (m, 5H),
1.71 (s, 3H), 1.71-1.65 (m, 4H), 1.65-1.57 (m, 4H), 1.50-1.45
(m, 2H), 1.45-1.41 (m, 2H), 1.34-1.21 (m, 28H), 1.03 (s, 6H),
0.88 (t, J = 7.0 Hz, 3H). 13C NMR (50 MHz, CDCl3/CD3OD
4:1): δ 174.2, 173.5, 168.0, 153.6, 140.1, 138.0, 137.6, 135.3, 131.6,
130.3, 129.7, 129.1, 118.5, 71.2, 70.7, 66.8, 64.0 (2C), 62.6 (2C),
39.9, 34.3 (3C), 33.4, 32.2, 29.9 (12C), 29.2 (2C), 28.7, 25.8, 25.1,
24.8, 23.0, 21.9, 19.5, 14.3, 14.1, 13.1. 31P NMR (202 MHz,
CDCl3/CD3OD 4:1): δ -0.08. IR (neat): 3390, 2924, 2853,
1734, 1709, 1661, 1458, 1260, 1237, 1153, 1050, 805 cm-1; m/z (M
þ Hþ) 931.57.
25.8, 24.9, 23.0 (2C), 14.3 (2C). 31P NMR (202 MHz, CDCl3/
CD3OD 4:1): δ -1.02. IR (neat): 3320, 2920, 2851, 1717, 1276,
1102, 1069 cm-1; m/z (M þ Hþ) 927.58.
1-Octadecanoyl-2-(6-(40-octyl-4-phenylbenzoyloxy)-hexanoyl)-
sn-glycero-3-phospho-(S)-glycerol (6). The synthesis was performed
as for 5, starting from alcohol 18 (70 mg, 0.077 mmol) and afford-
ing 51 mg (70% over 3 step) of 6 as a colorless amorphous solid.
1
Rf = 0.54 (CH2Cl2/MeOH 5:1). H NMR (500 MHz, CDCl3/
CD3OD 4:1): δ 8.08 (d, J = 8.3 Hz, 2H), 7.67 (d, J = 8.3 Hz, 2H),
7.56 (d, J = 8.1 Hz, 2H), 7.29 (d, J = 8.1 Hz, 2H), 5.27-5.22 (m,
1H), 4.39 (dd, J = 12.0, 3.2 Hz, 1H), 4.34 (d, J = 6.6 Hz, 2H), 4.17
(dd, J = 12.0, 6.6 Hz, 1H), 4.05-4.00 (m, 2H), 3.98-3.91 (m, 2H),
3.85-3.80 (m, 1H), 3.65-3.61 (m, 2H), 2.66 (t, J = 7.7 Hz, 2H),
2.40 (t, J = 7.5 Hz, 2H), 2.31 (t, J = 7.6 Hz, 2H), 1.86-1.78 (m,
2H), 1.76-1.69 (m, 2H), 1.68-1.62 (m, 2H), 1.62-1.56 (m, 2H),
1.55-1.49 (m, 2H), 1.39-1.20 (m, 38H), 0.89 (t, J = 6.7 Hz, 3H),
0.88 (t, J = 6.8 Hz, 3H). 13C NMR (50 MHz, CDCl3/CD3OD 4:1):
δ 174.4, 173.6, 167.4, 146.3, 143.7, 137.6, 130.4 (2C), 129.4 (2C),
129.0, 127.5 (2C), 127.2 (2C), 71.2, 70.8, 67.1, 65.2, 64.4, 62.7 (2C),
36.0, 34.4 (2C), 32.3 (2C), 31.8, 29.9 (15C), 28.8, 25.9, 25.2, 24.9,
23.0 (2C), 14.3 (2C). 31P NMR (202 MHz, CDCl3/CD3OD 4:1): δ
-1.30. IR (neat): 3314, 2921, 2851, 1737, 1467, 1277, 1103 cm-1
;
m/z (M þ Hþ) 941.55.
Acknowledgment. We would like to acknowledge Prof.
Helena Danielson (Department of Biochemistry and Organic
Chemistry, Uppsala University) for her assistance with the
enzyme inhibition assays. The technical help of Lars Duelund
(MEMPHYS, University of Southern Denmark) is much
appreciated. We thank the Danish Council for Strategic
Research (NABIIT Program) for financial support. MEM-
PHYS-Center for Biomembrane Physics is supported by the
Danish National Research Foundation.
Supporting Information Available: Analytical and spectral
data for all synthesized compounds, experimental procedures
for the synthesis of 2, 7, 9, 10, 14, 15, 17, 18, and 1-O-DSPG,
Mosher ester analysis data of 17, DLS analysis, further MAL-
DI-TOF MS data for sPLA2 degradation experiments, experi-
mental procedures for the inhibition experiments and figures
from MD simulations. This material is available free of charge
via the Internet at http://pubs.acs.org.
1-O-Octadecyl-2-(6-(40-octyl-4-phenylbenzoyloxy)-hexanoyl)-sn-
glycero-3-phospho-(S)-glycerol (5). Alcohol 15 (62 mg, 0.068
mmol), carboxylic acid 2 (43mg, 0.14 mmol), and DMAP(25 mg,
0.20 mmol) were dissolved in CH2Cl2 (4.5 mL). DCC (42 mg, 0.20
mmol) was added, and the reaction mixture was stirred at 20 °C
for 18 h. The mixture was concentrated in vacuo and purified by
flash column chromatography (toluene then toluene/EtOAc 5:1)
to give 65 mg that was dissolved in CH2Cl2 (5.0 mL) and DBU
(9 μL, 0.06 mmol) was added. The reaction mixture was stirred
for 35 min at 20 °C and then purified directly by flash column
chromatography (heptane/EtOAc 1:1 then CH2Cl2/MeOH 10:1)
toafford58mg that was dissolved in a mixtureofMeCN(4.2 mL)
and CH2Cl2 (1.4 mL) and cooled to 0 °C. Aqueous HF (40%,
250 μL) was added dropwise, and the reaction mixture was allowed
to reach 20 °C. After 3.5 h, the reaction was quenched by drop-
wise addition of MeOSiMe3 (0.93 mL), and the mixture was
stirred for 30 min, after which NaHCO3 (8 mg, 0.095 mmol) was
added and the mixture was concentrated in vacuo and purified by
flash column chromatography (CH2Cl2/MeOH 10:1 then CH2Cl2/
MeOH 4:1) to afford 24 mg (38% over 3 step) of 5 as a color-
less amorphous solid. Rf = 0.23 (CH2Cl2/MeOH 4:1). 1H NMR
(300 MHz, CDCl3/CD3OD 4:1): δ 8.08 (d, J = 8.3 Hz, 2H), 7.67
(d, J = 8.3 Hz, 2H), 7.55 (d, J = 8.1 Hz, 2H), 7.29 (d, J = 8.1 Hz,
2H), 5.19-5.14 (m, 1H), 4.34 (t, J = 6.6 Hz, 2H), 4.06-3.93 (m,
4H), 3.85-3.80 (m, 1H), 3.66-3.61 (m, 2H), 3.60-3.56 (m, 2H),
3.48-3.38 (m, 2H), 2.66 (t, J = 7.7 Hz, 2H), 2.40 (t, J = 7.5 Hz,
2H), 1.86-1.79 (m, 2H), 1.76-1.70 (m, 2H), 1.69-1.62 (m, 2H),
1.56-1.49 (m, 4H), 1.39-1.19 (m, 40H), 0.89 (t, J = 6.8 Hz, 3H),
0.88 (t, J = 6.9 Hz, 3H). 13C NMR (50 MHz, CDCl3/CD3OD
4:1): δ 173.9, 167.3, 146.1, 143.6, 137.5, 130.3 (2C), 129.3 (2C),
128.9, 127.4 (2C), 127.1 (2C), 72.1 (2C), 71.3, 69.3, 66.7, 65.2,
64.6, 62.6, 35.9, 34.5, 32.1 (2C), 31.8, 30.0, 29.7 (15C), 28.8, 26.3,
References
(1) Barua, A. B.; Furr, H. C. Properties of Retinoids. Mol. Biotechnol.
1998, 10, 167–182.
(2) Rolewski, S. L. Clinical Review: Topical Retinoids. Dermatol.
Nurs. 2003, 15, 447–465.
(3) (a) Okuno, M.; Kojima, S.; Matsushima-Nishiwaki, R.; Tsurumi,
H.; Muto, Y.; Friedman, S. L.; Moriwaki, H. Retinoids in Cancer
Chemoprevention. Curr. Cancer Drug Targets 2004, 4, 285–298.
(b) Freementle, S. J.; Spinella, M. J.; Dmitrovsky, E. Retinoids in cancer
therapy and chemoprevention: promise meets resistance. Oncogene
2003, 22, 7305–7315.
(4) (a) Huang, M. E.; Ye, Y. C.; Chen, S. R.; Chai, J.; Lu, J. X.; Zhoa,
L.; Gu, L. J.; Wang, Z. Y. Use of all-trans retinoic acid in the
treatment of acute promyelocytic leukemia. Blood 1988, 72, 567–
572. (b) Castaigne, S.; Chomienne, C.; Daniel, M. T.; Ballerini, P.;
Berger, R.; Fenaux, P.; Degos, L. All-trans retinoic acid as a differentia-
tion therapy for acute promyelocytic leukemia. I. Clinical results. Blood
1990, 76, 1704–1709. (c) Regazzi, M. B.; Iacona, I.; Gervasutti, C.;
Lazzarino, M.; Toma, S. Clinical Pharmacokinetics of Tritinoin. Clin.
Pharmacokinet. 1997, 32, 382–402.
(5) Muindi, J. R. F.; Frankel, S. R.; Huselton, C.; DeGrazia, F.;
Garland, W. A.; Young, C. W.; Warrell, R. P. Clinical pharmaco-
logy of oral all-trans-retinoic acid in patients with acute promye-
locytic leukemia. Cancer Res. 1992, 52, 2138–2142.
(6) (a) Shimizu, K.; Tamagawa, K.; Takahashi, N.; Takayama, K.;
Maitani, Y. Stability and antitumor effects of all-trans retinoic
acid-loaded liposomes contained sterylglucoside mixture. Int.
J. Pharm. 2003, 258,45–53. (b) Kawakami, S.; Opanasopit, P.; Yokoyama,
M.; Chansri, N.; Yamamoto, T.; Okano, T.; Yamashita, F.; Hasdida, M.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9 3791
Biodistribution Characteristics of All-trans Retinoic Acid Incorporated in
Liposomes and Polymeric Micelles Following Intravenous Administration.
Satoh, A.; Fujita, M.; Yoshida, M.; Shimosegawa, T. Lysophosphati-
dylcholine induces apoptosis in AR42J cells. Pancreas 2001, 22,
75–83. (c) Kogure, K.; Nakashima, S.; Tsuchie, A.; Tokumura, A.;
Fukuzawa, K. Temporary membrane distortion of vascular smooth
muscle cells is responsible for their apoptosis induced by platelet-
activating factor-like oxidized phospholipids and their degradation
product, lysophosphatidylcholine. Chem. Phys. Lipids 2003, 126,
29–38.
€
J. Pharm. Sci. 2005, 94, 2606–2615. (c) Díaz, C.; Vargas, E.; Gatjens-
Boniche, O. Cytotoxic effect induced by retinoic acid loaded into galatosyl-
sphingosine containing liposomes on human hepatoma cell lines. Int.
J. Pharm. 2006, 325, 208–115. (d) Suzuki, S.; Kawakami, S.; Chansri, N.;
Yamashita, F.; Hasdida, M. Inhibition of pulmonary metastasis in mice by
all-trans retinoic acid incorporated in cationic liposomes. J. Controlled
Release 2006, 116, 58–63.
€
(18) Peters, G. H.; Møller, M. S.; Jørgensen, K.; Ronnholm, P.;
(7) (a) Pedersen, P. J.; Christensen, M. S.; Ruysschaert, T.; Linderoth,
L.; Andresen, T. L.; Melander, F.; Mouritsen, O. G.; Madsen, R.;
Clausen, M. H. Synthesis and Biophysical Characterization of
Chlorambucil Anticancer Ether Lipid Prodrugs. J. Med. Chem.
2009, 52, 3408–3415. (b) Christensen, M. S.; Pedersen, P. J.; Andresen,
T. L.; Madsen, R.; Clausen, M. H. Isomerization of all-(E)-Retinoic
Acid Mediated by Carbodiimide Activation;Synthesis of ATRA Ether
Lipid Conjugates. Eur. J. Org. Chem. 2009, 719–724. (c) Linderoth,
L.; Peters, G. H.; Madsen, R.; Andresen, T. L. Drug Delivery by an
Enzyme-Mediated Cyclization of a Lipid Prodrug with Unique Bilayer-
Formation Properties. Angew. Chem., Int. Ed. 2009, 48, 1823–1826.
(8) (a) Newton, D. L.; Henderson, W. R.; Sporn, M. B. Structure-
Activity Relationships of Retinoids in Hamster Tracheal Organ
Culture. Cancer Res. 1980, 40, 3413–3425. (b) Um, S. J.; Kwon, Y. J.;
Han, H. S.; Park, S. H.; Park, M. S.; Rho, Y. S.; Sin, H. S. Synthesis and
Biological Activity of Novel Retinamide and Retinoate Derivatives.
Chem. Pharm. Bull. 2004, 52, 501–506. (c) Gediya, L. K.; Khandelwal,
A.; Patel, J.; Belosay, A.; Sabnis, G.; Mehta, J.; Purushottamachar, P.;
Njar, V. C. O. Design, Synthesis, and Evaluation of Novel Mutual
Prodrugs (Hybrid Drugs) of All-trans-Retinoic Acid and Histone Dea-
cetylase Inhibitors with Enhanced Anticancer Activities in Breast and
Prostate Cancer Cells in Vitro. J. Med. Chem. 2008, 51, 3895–3904.
(d) Ulukaya, E.; Wood, E. J. Fenretinide and its relation to cancer. Cancer
Treat. Rev. 1999, 25, 229–235.
(9) Linderoth, L.; Andresen, T. L.; Jørgensen, K.; Madsen, R.; Peters,
G. H. Molecular Basis of Phospholipase A2 Activity toward
Phospholipids with sn-1 Substitutions. Biophys. J. 2008, 94, 14–26.
(10) Camacho, L. H. Clinical applications of retinoids in cancer medi-
cine. J. Biol. Regul. Homeostatic Agents 2003, 17, 98–114.
(11) (a) Maden, M. Retinoic Acid in the Development, Regeneration
and Maintenance of the Nervous System. Nat. Rev. Neurosci. 2007,
8, 755–765. (b) Swift, C. B.; Hays, J. L.; Petty, J. W. Distinct Functions
of Retinoic Acid Receptor Beta Isoforms: Implications for Targeted
Therapy. Endocr., Metab. Immune Disord.: Drug Targets 2008, 8,
47–50.
Mikkelsen, M.; Andresen, T. L. Secretoty Phospholipase A2
Hydrolysis of Phospholipid Analogues is Dependent on Water
Accessibility to the Active Site. J. Am. Chem. Soc. 2007, 129,
5451–5461.
(19) Scott, D. L.; White, S.; Otwinowski, Z.; Yuan, W.; Gelb, M. H.;
Sigler, P. B. Interfacial Catalysis;The Mechanism of Phospholi-
pase-A2. Science 1990, 250, 1541–1546.
(20) tert-Butyl 6-hydroxyhexanoate (11) was synthesized according to
the litterature. Larock, R. C.; Leach, D. R. Organopalladium
Approaches to Prostaglandins. 3. Synthesis of Bicyclic and Tricyclic
7-Oxaprostaglandin Endoperoxide Analoguesnvia Oxypalladation
of Norbornadiene. J. Org. Chem. 1984, 49, 2144–2148.
(21) Mitsunobu, O. The Use of Diethyl Azodicarboxylate and Triphe-
nylphosphine in Synthesis and Transformation of Natural Pro-
ducts. Synthesis 1981, 1–28.
(22) Kaul, R.; Brouillette, Y.; Sajjadi, Z.; Hansford, K. A.; Lubell,
W. D. Selective tert-Butyl Ester Deprotection in the Presence of
Acid Labile Protection Groups with ZnBr2. J. Org. Chem. 2004, 69,
6131–6133.
(23) Borgulya, J.; Bernauer, K. Transformation of Carboxylic Acid
t-Butyl Esters into the Corresponding Trimethylsilyl Esters or Free
Acids under Nonacidic Conditions. Synthesis 1980, 545–547.
(24) Compound 2 is commercially available (CAS no. 59662-49-6), but
during this work, 2 was synthesized in one step from the corres-
ponding nitrile (CAS no. 52709-84-9) by a NaOH mediated
hydrolysis in 98% yield (see Supporting Information).
(25) Neises, B.; Steglich, W. Simple Method for the Esterification of
Carboxylic Acids. Angew. Chem., Int. Ed. Engl. 1978, 17, 522–524.
(26) Gaffney, P. R. J.; Reese, C. B. Synthesis of naturally occurring
phosphatidylinositol 3,4,5-triphosphate [PtdIns(3,4,5)P3] and its
diastereoisomers. J. Chem. Soc., Perkin Trans. 1 2001, 192–105.
(27) Burgos, C. E.; Ayer, D. E.; Johnson, R. A. A New, Asymmetric
Synthesis of Lipids and Phospolipids. J. Org. Chem. 1987, 52,
4973–4977.
(28) Batten, R. J.; Dixon, A. J.; Taylor, R. J. K. A New Method for
Removing the t-Butyldimethylsilyl Protecting Group. Synthesis
1980, 234–236.
(29) Qin, D.; Byun, H. S.; Bittman, R. Synthesis of Plasmalogen via 2,3-
Bis-O-(40-methoxybenzyl)-sn-glycerol. J. Am. Chem. Soc. 1999,
121, 662–668.
(30) Dale, J. A.; Mosher, H. S. NuclearMagnetic Resonance Enantiomer
Reagents. Configurational Correlations via Nuclear Magnetic
Resonance Chemical Shifts of Diastereomeric Mandelate,
O-Methylmandelate, and R-Methoxy-R-trifluoromethylphenyla-
cetate (MTPA) Esters. J. Am. Chem. Soc. 1973, 95, 512–519.
(31) (a) Lichtenberg, D.; Barenholz, Y. Liposomes;Preparation,
Characterization and Preservation. Methods Biochem. Anal.
1988, 33, 337–462. (b) Tirrell, D. A.; Takigawa, D. Y.; Seki, K.
Interactions of synthetic polymers with cell-membranes and model
membrane systems 7. pH sensitization of phospholipid-vesicles via
complexation with synthetic poly(carboxylic acid)s. Ann. N.Y. Acad.
Sci. 1985, 446, 237–248.
(12) Zhuang, Y.; Faria, T. N.; Chambon, P.; Gudas, L. J. Identification
and characterization of retinoic acid receptor beta2 target genes in
F9 teratocarcinoma cells. Mol. Cancer Res. 2003, 1, 619–630.
(13) (a) Piu, F.; Gauthier, N. K.; Olsson, R.; Currier, E. A.; Lund,
B. W.; Croston, G. E.; Hacksell, U.; Brann, M. R. Identification of
novel subtype selective RAR agonists. Biochem. Pharmacol. 2005,
71, 156–162. (b) Lund, B. W.; Piu, F.; Gauthier, N. K.; Eeg, A.; Currier,
E.; Sherbukhin, V.; Brann, M. R.; Hacksell, U.; Olsson, R. Discovery of
a Potent, Orally Available and Isoform-Selective Retinoic Acid β2
Receptor Agonist. J. Med. Chem. 2005, 48, 7517–7519. (c) Lund, B.
W.; Knapp, A. E.; Piu, F.; Gauthier, N. K.; Begtrup, M.; Hacksell, U.;
Olsson, R. Design, Synthesis, and Structure-Activity Analysis of
Isoform-Selective Retinoic Acid Receptor β Ligands. J. Med. Chem.
2009, 52, 1540–1545.
(14) Bonsen, P. P. M.; de Haas, G. H.; Pieterson, W. A.; van Deenen,
L. L. Studies on Phospholipase A and its Zymogen from Porcine
Pancreas. IV. The Influence of Chemical Modification of the
Lecithin Structure on Substrate Properties. Biochim. Biophys. Acta
1972, 270, 364–382.
(32) McIntyre, J. C.; Sleight, R. G. Fluorescence assay for phospholipid
membrane asymmetry. Biochemistry 1991, 30, 11819–11827.
(33) (a) Huster, D.; Muller, P.; Arnold, K.; Herrmann, A. Dynamics of
membrane penetration of the fluorescent 7-nitrobenz-2-oxa-1,3-
diazol-4-yl (NBD) group attached to an acyl chain of phosphati-
dylcholine. Biophys. J. 2001, 80, 822–831. (b) Langner, M.; Hui, S. W.
Dithionite penetration through phospholipid bilayers as a measure of
defects in lipid molecular packing. Chem. Phys. Lipids 1993, 65, 23–30.
(34) (a) Harvey, D. J. Matrix-Assisted Laser Desorption/Ionization
Mass Spectrometry of Phospholipids. J. Mass Spectrom. 1995, 30,
1333–1346. (b) Schiller, J.; Arnhold, J.; Benard, S.; M€uller, M.; Reichl,
S.; Arnold, K. Lipid Analysis by Matrix-Assisted Laser Desorption and
Ionization Mass Spectrometry: A Methodological Approach. Anal.
(15) (a) Buckland, A. G.; Wilton, D. C. Anionic phospholipids, inter-
facial binding and the regulation of cell functions. Biochim. Bio-
phys. Acta 2000, 1483, 199–216. (b) Canaan, S.; Nielsen, R.;
Ghomashchi, F.; Robinson, B. H.; Gelb, M. H. Unusual mode of binding
of human group IIA secreted phospholipase A2 to anionic interfaces as
studied by continuous wave and time domain electron paramagnetic
resonance spectroscopy. J. Biol. Chem. 2002, 277, 30984–30990.
(c) Houlihan, W. J.; Lohmeyer, M.; Workman, P.; Cheon, S. H. Phos-
pholipid antitumor agents. Med. Res. Rev. 1995, 15, 157–223.
(16) (a) Houlihan, W. J.; Lohmeyer, M.; Workman, P.; Cheon, S. H.
Phospholipid antitumor agents. Med. Res. Rev. 1995, 15, 157–223.
(b) Andresen, T. L.; Jensen, S. S.; Madsen, R.; Jørgensen, K. Synthesis
and Biological Activity of Anticancer Ether Lipids that are Specifically
Released by Phospholipase A2 in Tumor Tissue. J. Med. Chem. 2005,
48, 7305–7314.
ꢀ
Biochem. 1999, 267, 46–56. (c) Petkovic, M.; Muller, J.; Muller, M.;
Schiller, J.; Arnold, K.; Arnhold, J. Application of matrix-assisted laser
desorption/ionization time-of-flight mass spectrometry for monitoring
the digestion of phosphatidylcholine by pancreatic phospholipase A2.
Anal. Biochem. 2002, 308, 61–70.
(17) (a) Hsieh, C. C.; Yen, M. H.; Liu, H. W.; Lau, Y. T. Lysopho-
sphatidylcholine induces apoptotic and non-apoptotic death in
vascular smooth muscle cells: in comparison with oxidized LDL.
Atherosclerosis 2000, 151, 481–491. (b) Masamune, A.; Sakai, Y.;
(35) Fawzy, A. A.; Wishwanath, B. S.; Franson, R. C. Inhibition of
human nonpancreatic phospholipases A2 by retinoids and flavo-
noids. Mechanism of action. Agents Actions 1988, 25, 394–400.

3792 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 9
Pedersen et al.
(36) Hope, W. C.; Patel, B. J.; Fiedler-Nagy, C.; Wittreich, B. H.
Retinoids inhibit Phospholipase A2 in human synovial fluid and
arachidonic acid release from rat peritoneal macrophages. Inflam-
mation 1990, 14, 543–559.
(37) Cunningham, T. J.; Maciejewski, J.; Yao, L. Inhibition of secreted
phospholipase A2 by neuron survival and anti-inflammatory pep-
tide CHEC-9. J. Neuroinflammation 2006, 3, 25.
(38) Nicke, B.; Kaiser, A.; Wiedernmann, B.; Riecken, E. O.; Rosewicz,
S. Retinoic Acid Receptor R Mediates Growth Inhibition by
Retinoids in Human Colon Carcinoma HT29 Cells. Biochem.
Biophys. Res. Commun. 1999, 261, 572–577.
(39) (a) Geurts van Kassel, W. S. M.; Hax, W. M. A.; Demel, R. A.;
de Gier, J. High Performance Liquid Chromatographic Separation
and Direct Ultraviolet Detection of Phospholipids. Biochim. Bio-
phys. Acta 1977, 486, 524–530. (b) Rivnay, B. Combined Analysis of
Phospholipids by High-Performance Liquid Chromatography and
Thin-Layer Chromatography. J. Chromatogr. 1984, 294, 303–315.
(40) Hansford, K. A.; Reid, R. C.; Clark, C. I.; Tyndall, J. D. A.;
Whitehouse, M. W.; Guthrie, T.; McGeary, R. P.; Schafer, K.;
Martin, J. L.; Fairlie, D. P. D-Tyrosine as a chiral precusor to
potent inhibitors of human nonpancreatic secretory phospholipase
A(2) (IIa) with antiinflammatory activity. ChemBioChem 2003, 4,
181–185.
(43) Jorgensen, W. L.; Chandrasekhar, J.; Medura, J. D.; Impey, R. W.;
Klein, M. L. Comparison of simple potential models for simulating
liquid water. J. Chem. Phys. 1983, 79, 926–935.
€
(44) Grubmuller, H. Solvate: a program to create atomic solvent models.
Electronic publication: http://www.mpibpc.gwdg.de/abteilungen/071/
solvate/docu.html; accessed 02/11/2010.
(45) Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks., B. R. Constant
pressure molecular dynamics simulation: the Langevin piston
method. J. Chem. Phys. 1995, 103, 4613–4621.
(46) (a) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald an
Nlog(n) method for Ewald sums in large systems. J. Chem. Phys.
1993, 98, 10089–10092. (b) Darden, T.; York, D.; Pedersen, L. Particle
mesh Ewald: an N-log(N) method for Ewald sums in large systems.
J. Chem. Phys. 1995, 103, 8577–8593.
(47) Humphrey, W.; Dalke, A.; Schulten, K. VMD;Visual Molecular
Dynamics. J. Mol. Graphics 1996, 14, 33–38.
(48) Mills, J. K.; Needham, D. Lysolipid incorporation in dipalmitoyl-
phosphatidylcholine bilayer membranes enhances the ion perme-
ability and drug release rates at the membrane phase transition.
Biochim. Biophys. Acta 2005, 1716, 77–96.
(49) (a) de Carvalho, L. M.; Schwedt, G. Polarographic determination
of dithionite and its decomposition products: kinetic aspects,
stabilizers, and analytical application. Anal. Chim. Acta 2001,
436, 293–300. (b) Stutts, K. J. Liquid chromatographic assay of
dithionite and thiosulfate. Anal. Chem. 1987, 59, 543–544.
(50) Chen, P. S.; Toribara, T. Y.; Warner, H. Microdetermination of
phosphorus. Anal. Chem. 1956, 28, 1756–1758.
(41) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J.; Meyer, E. E.; Brice,
M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi, M.
Protein Data Bank;Computer-Based Archival File for Macro-
molecular Structures. J. Mol. Biol. 1977, 112, 535–542.
(42) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.;
Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable
molecular dynamics with NAMD. J. Comput. Chem. 2005, 26,
1781–1802.
(51) Carmichael, J.; DeGraff, W. G.; Gazdar, A. F.; Minna, J. D.;
Mitchell, J. B. Evaluation of a tetrazolium-based semiautomated
colorimetric assay: assessment of chemosensitivity testing. Cancer
Res. 1987, 47, 936–942.