Isomerization of all-(E)-retinoic acid mediated by carbodiimide activation synthesis of ATRA ether lipid conjugates

Article abstract of DOI:10.1002/ejoc.200901128

Treatment of the lysolipid 1-O-hexadecyl-sn-phosphatidylcholine with all-(E)-retinoic acid, DCC and DMAP resulted in poor acylation and caused (Z)Z(E) isomerization of the α-β double bond. In the presence of a proton, source, the carbodiimide-activated all-(E)-retinoic acid undergoes fast isomerization to give a final mixture of (13E)/(13Z) isomers in a 3:1 ratio. Similar treatment of (13Z)-retinoic acid leads to the same isomer ratio. The isomerization was circumvented successfully by using a Mitsunobu reaction, which provided an efficient synthesis of all-(.E)-retinoic acid sn-2-conjugated to phosphatidylcholine and phosphatidylglycerol etherlipids.

Full text of DOI:10.1002/ejoc.200901128

FULL PAPER
DOI: 10.1002/ejoc.200901128
Isomerization of all-(E)-Retinoic Acid Mediated by Carbodiimide Activation –
Synthesis of ATRA Ether Lipid Conjugates
Mikkel S. Christensen,[a] Palle J. Pedersen,[a] Thomas L. Andresen,[b] Robert Madsen,[a] and
Mads H. Clausen*[a]
Keywords: Phospholipids / Retinoic acid / Acylation / Mitsunobu reaction / Isomerization
Treatment of the lysolipid 1-O-hexadecyl-sn-phosphatidyl-
choline with all-(E)-retinoic acid, DCC and DMAP resulted
in poor acylation and caused (Z)/(E) isomerization of the α–β
double bond. In the presence of a proton source, the carbodi-
imide-activated all-(E)-retinoic acid undergoes fast isomer-
ization to give a final mixture of (13E)/(13Z) isomers in a 3:1
ratio. Similar treatment of (13Z)-retinoic acid leads to the
same isomer ratio. The isomerization was circumvented suc-
cessfully by using a Mitsunobu reaction, which provided an
efficient synthesis of all-(E)-retinoic acid sn-2-conjugated to
phosphatidylcholine and phosphatidylglycerol etherlipids.
to two signals with the proton of the major isomer resonat-
ing at δ = 5.8 ppm and the minor at δ = 5.6 ppm, data that
is consistent with the isomers being esters of all-(E)-retinoic
acid (2) and (13Z)-retinoic acid (3), respectively
(Scheme 1).[6]
Introduction
Liposomes are efficient drug-delivery systems for on-
cology, and a number of such systems are on the market or
in clinical development.[1] We have recently disclosed novel
types of drug-delivery systems, where the active drug is co-
valently linked to the lipid constituting the carrier system
and liberated in situ by secretory phospholipase A2
(sPLA2).[2,3] In extension of these findings, we were inter-
ested in synthesizing sn-1 ether lipids with all-(E)-retinoic
acid (ATRA) in the sn-2 position. ATRA is known to have
cytotoxic properties towards a range of cancer cells and as
such constitutes an interesting compound for cancer treat-
ment.[4] Furthermore, ATRA is a highly lipophilic com-
pound sensitive to acid and light that could potentially ben-
efit from a liposomal formulation and the protection of-
fered by localization to the liposome membrane.
Based on our experience with prodrugs with the antican-
cer agent chlorambucil in the sn-2 position, we have investi-
gated the synthesis of (E)-4 from lysolipid 1[3] and ATRA
(2) under standard acylation conditions.[5] However, when
1 was treated with excess 2, dicyclohexylcarbodiimide
(DCC) and 4-(dimethylamino)pyridine (DMAP) in dichlo-
romethane the conversion was poor and most retinoic acid
formed either the anhydride or the N-acylurea byproduct.
Furthermore, the isolated ester 4 was a mixture of two in-
Scheme 1. DCC/DMAP-mediated ester coupling with concurrent
isomerization.
1
separable isomers. In the H NMR spectrum, H14 gave rise
[a] Department of Chemistry, Technical University of Denmark,
Kemitorvet, Building 201, 2800 Kgs. Lyngby, Denmark
Fax: +45-4593-3968
Results and Discussion
To examine the acylation with ATRA and its concomi-
tant isomerization we chose hexadecanol for acylation ex-
periments to simplify spectral characterization and purifica-
tion (Scheme 2 and Table 1). As displayed in Table 1 (En-
try 1) our first choice of activation reagent was 3-[3-(di-
E-mail: mhc@kemi.dtu.dk
[b] Department of Micro- and Nanotechnology, Technical Univer-
sity of Denmark,
4000 Roskilde, Denmark
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.200901128.
Eur. J. Org. Chem. 2010, 719–724
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
719

M. H. Clausen et al.
FULL PAPER
methylamino)propyl]-1-ethylcarbodiimide (EDC) hydro- the exact same (E)/(Z) = 3:1 ratio (see Supporting Infor-
chloride hoping to minimize the formation of N-acylu- mation). Because this ratio is established within 2 h in the
rea.[7,8] Indeed, the isolated yield was an excellent 98%, but presence of the carbodiimide and HCl, it is evident that the
isomerization was prominent with a (13E)/(13Z) isomer ra- isomerization process takes place only after formation of
tio of 3:1. Employing DCC instead (Entry 2) led to a lower the activated intermediate, whereafter equilibrium is
yield of 80% but notably without the concomitant isomer- reached quickly.
ization. Addition of pyridine hydrochloride (pyr·HCl) as a
proton source (Entry 3) did not improve the yield but re-
stored a (13E)/(13Z) isomer ratio of 3:1. At this point, it
appeared that HCl was necessary for isomerization to occur,
and another experiment employing tetrabutylammonium
chloride (TBACl) as additive alongside DCC and DMAP
(Entry 4) showed the chloride anion to have no significant
effect on isomerization or yield. Simple treatment of reti-
noic acid with pyr·HCl in CD2Cl2 did not cause significant
isomerization within 2 h (data not shown), which was well
within the time frame of isomerization in the coupling ex-
periments (vide infra). However, as indicated in Entry 5,
employing EDC hydrochloride alone without DMAP did
not cause any significant isomerization either. It is thus evi-
dent that both a proton source and DMAP are involved in
the isomerization step.
Scheme 3. Sequential addition of coupling reagents in NMR ex-
periment.
Naturally occurring retinoic acids exist in biological sys-
tems as mixtures of mainly all-(E)-retinoic acid, (13Z)-reti-
noic acid and (9Z)-retinoic acid and are all known to be
efficacious in certain cancer treatments.[4] They have been
shown to be interconverted enzymatically by glutathione S-
methyltransferase,[10] and they may also be interconverted
by treatment with thiol reagents.[11] Mechanistically, the
isomerization process probably involves a 1,4-addition fol-
lowed by single-bond rotation and elimination.[10] This
mechanism has also been suggested in more general cases
of (E)/(Z) isomerization of α,β-unsaturated carboxylic acid
Scheme 2. Reagents and conditions for acylation experiments.
In order to obtain a better understanding of the isomer- derivatives when their mixed anhydrides were treated with
1
ization process, the reaction was monitored by H NMR DMAP in esterification reactions[12] or a Yamaguchi lac-
spectroscopy. All-(E)-retinoic acid (2) was dissolved in tonization.[13] However, to the best of our knowledge, this is
CD2Cl2 in an NMR tube, and reagents were added sequen- the first clear documentation of the importance of a proton
tially (Scheme 3, Figure 1 and Supporting Information).[9] source in addition to DMAP as nucleophile, and these find-
DMAP addition did not cause any observable isomerization ings may improve the scope of this conventional method for
within 1 h, and EDC hydrochloride was added. A (13E)/ ester formation.
(13Z) isomer ratio of ca. 3:1 for the activated intermediate
However, even with the reactive cetyl alcohol the yield
was reached within 2 h, and this remained unchanged for for the carbodiimide coupling under conditions where mini-
24 h (not shown). Thus, equilibrium had been reached, and mum isomerization occurs (Table 1, Entries 2, 4 and 5) was
quenching this mixture with [D4]MeOH gave rise to two considered inadequate. Employing a Mitsunobu reaction[14]
new signals conserving the 3:1 isomer ratio. Applying the to obtain cetyl esters of both all-(E)- and (13Z)-retinoic
same methodology with (13Z)-retinoic acid (3) resulted in acid (Table 1, Entries 6 and 7) completely suppressed iso-
Table 1. Cetyl esters of retinoic acid.
Entry
Retinoic acid
Conditions[a]
(13E)/(13Z)[b]
Yield[c]
1
2
3
4
5
6
7
2
2
2
2
2
2
3
EDC·HCl (1.5 equiv.), DMAP (1.5 equiv.)
DCC (1.5 equiv.), DMAP (1.5 equiv.)
DCC (1.5 equiv.), DMAP (1.5 equiv.), pyr·HCl (1.5 equiv.)
DCC (1.5 equiv.), DMAP (1.5 equiv.), TBACl (1.5 equiv.)
EDC·HCl (2.0 equiv.)
DIAD (1.5 equiv.), Ph3P (1.5 equiv.)
DIAD (1.5 equiv.), Ph3P (1.5 equiv.)
3:1
1:0
3:1
16:1
16:1
1:0
98%
80%
69%
73%
53%
99%
81%
0:1
[a] Reactions performed with C16H33OH (2.0 equiv.) in CH2Cl2 at 20 °C (Entries 1–5) or THF at 0 °C (Entries 6–7). [b] Determined by
1H NMR spectroscopy. [c] Isolated yield after 24 h (48 h for Entry 5).
720
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 719–724

Synthesis of ATRA Ether Lipid Conjugates
Figure 1. Successive NMR spectra showing H14 signals acquired
during activation of all-(E)-retinoic acid by EDC·HCl followed by
quenching with [D4]MeOH in CD2Cl2 (solvent residue at δ =
5.3 ppm, indicated for reference of scale).
Scheme 4. Reagents and conditions: (a) 2, Ph3P, DIAD, THF; (b)
aq. HF, MeCN (50%, 2 steps); (c) i: POCl3, CH2Cl2, ii: choline
tosylate, pyridine, iii: H2O (40%, 3 steps); (d) i: 9, 1H-tetrazole,
MeCN, CH2Cl2, ii: tBuOOH, iii: DBU (89%, 3 steps); (e) aq. HF,
MeCN (82%).
merization for both esters. For this reason and due to the
difficulties of obtaining acylation precursors that were com-
pletely free of residual protons, we opted for this reaction
for the preparation of the desired phospholipids (E)-4 and
8 containing all-(E)-retinoic acid in the sn-2 position
(Scheme 4). Thus, the known[15] alcohol 5, obtained in 82%
yield from -1,2-isopropylidineglycerol in three steps, was
esterified with all-(E)-retinoic acid by a Mitsunobu reac-
tion. Coelution of excess diisopropyl azodicarboxylate
(DIAD) with the product precluded isolation on a larger
scale, and the ester was therefore deprotected immediately.
Desilylation with TBAF in THF was hampered by acyl mi-
gration from the sn-2 to the sn-3 position. In contrast, the
TBDMS group was removed cleanly by using aq. HF in
acetonitrile and afforded 6 in a 50% yield over the two
steps. The phosphatidylcholine headgroup was installed by
treatment of 6 with excess phosphoryl trichloride followed
by choline tosylate and then water to give 4 in a 40% yield.
Alternatively, the phosphorylglycerol headgroup could be
introduced to give 7 in 89% yield over three steps by stan-
dard amidite chemistry employing the TBDMS protected
phosphoramidite 9.[3] The amidite was activated with 1H-
To confirm complete inversion in the Mitsunobu reac-
tion, a sample of the ester was reductively cleaved with di-
isobutylaluminum hydride and subsequently converted into
the para-methoxybenzoate for analytical convenience to
give (+)-10 (Scheme 5). Comparison with the enantiomer
(–)-10 by chiral HPLC (chiralcel OD-H column) showed
clean inversion with retention times of 6.0 min for (–)-10
and 6.5 min for (+)-10 (see Supporting Information for de-
tails).
Scheme 5. Reagents and conditions: (a) 2, Ph3P, DIAD, THF, 0 Ǟ
20 °C; (b) DIBAL-H, CH2Cl2, –78 °C; (c) PMBzCl, DMAP, Et3N,
CH2Cl2, 0 Ǟ 20 °C.
To asses whether the phospholipid 8 was a substrate for
tetrazole, and subsequent phosphorus oxidation was secretory phospholipase A2, the lipid was suspended in a
achieved with careful sequential addition of tBuOOH in stirred emulsion of 10 m HEPES buffer (pH = 7.5,
decane to avoid oxidation of the olefins.[16] The TBDMS 0.03 m CaCl2) and Et2O (5:2) and treated with snake
groups were removed in 82% yield to give 8 again by em- (Naja mossambica mossambica) venom phospholi-
ploying aq. HF in acetonitrile, conditions tolerated by the pase A2.[18] Evidently, neither TLC nor MALDI-TOF MS
sensitive unsaturated ester. Compared to the more conven- of the ethereal phase showed hydrolytic cleavage after 2 d
tional acetonide protection of the glycerol headgroup,[17] of incubation, whereas the natural substrate dipalm-
the milder deprotection conditions required for the silyl itoylphosphatidylglycerol (DPPG) was cleaved quantita-
ethers make them very well suited for acid-sensitive lipids.
tively under the same conditions in less than 2 h. Likewise,
Eur. J. Org. Chem. 2010, 719–724
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
721

M. H. Clausen et al.
FULL PAPER
dissolved in THF (10 mL) and cooled to 0 °C. Diisopropyl azod-
icarboxylate (0.20 mL, 1.0 mmol) was added slowly over 10 min.
The mixture was stirred for 2 h before another portion of Ph3P, 2
and DIAD was added. The mixture was stirred for 3 h, concen-
trated in vacuo, and the residue was chromatographed (CH2Cl2/
hexane, 1:1) to give an oil, which was suspended in MeCN (9.0 mL)
and cooled to 0 °C before 40% aq. HF (1.0 mL) was added drop-
wise. Cooling was then removed and the mixture stirred for 9 h
before it was poured into satd. aq. NaHCO3 (50 mL) and extracted
with CH2Cl2 (3ϫ20 mL). The organic phase was dried (Na2SO4)
liposomes formulated from 8 with an average diameter of
96 nm as determined by dynamic light scattering (DLS)
were not degraded by sPLA2, neither from Naja mossam-
bica mossambica nor Agkistrodon piscivorus piscivorus (see
Supporting Information). Similarly, phosphatidylcholine li-
pid (E)-4 was not hydrolyzed by sPLA2 (data not shown).
Since it has been reported[19] that retinoids can inhibit
PLA2 activity, we decided to investigate the hydrolysis of
DPPG in the presence of ATRA. Our experiments ruled
out that the lack of hydrolysis was due to inhibition by and concentrated in vacuo before flash chromatography on silica
(EtOAc/heptane, 1:9 Ǟ 1:1) gave 6 as a gum (0.25 g, 50%). [α]2D0
=
ATRA, since DPPG was completely hydrolyzed by PLA2
within 24 h in the presence of both 0.1 and 1 equiv. of
ATRA (see Supporting Information).
+12.2 (c = 2.4, CHCl ). IR (neat): ν = 3446, 2924, 2853, 1710, 1609,
˜
3
1584, 1458, 1358, 1239, 1152, 1050, 966 cm–1. 1H NMR (300 MHz,
CDCl3): δ = 7.01 (dd, 3JH,H = 11.4, 15.0 Hz, 1 H, 11-H), 6.30–6.11
(m, 4 H, 12-H, 10-H, 8-H, 7-H), 5.82 (s, 1 H, 14-H), 5.04 (q, 3JH,H
= 4.7 Hz, 1 H, sn-2-H), 3.84 (m, 2 H, sn-3-H), 3.65 (m, 2 H, sn-1-
H), 3.47 (m, 2 H, 1Ј-H), 2.37 (d, 3JH,H = 1.0 Hz, 3 H, 13-Me), 2.00
(m, 5 H, 9-Me, 3-H), 1.71 (s, 3 H, 2-Me), 1.63–1.44 (m, 6 H, 4-H,
2Ј-H, 5-H), 1.24 (br., 26 H, 3Ј-H to 15Ј-H), 1.02 (br., 6 H, 2 6-Me),
Conclusions
We have shown the importance of a proton source and a
nucleophile such as DMAP in the isomerization of all-(E)-
retinoic acid to its (13Z) isomer when activated with car-
bodiimide reagents. To the best of our knowledge, this rep-
resents a new aspect of this isomerization and may have
consequences for acylations with α,β-unsaturated acids in
general. The enzyme experiments show that ATRA–lipid
conjugates are not hydrolyzed by PLA2. Comparison to the
natural substrates and our previous investigation of chlor-
ambucil–lipid conjugates suggest that the rigid and steri-
cally demanding structure of ATRA is responsible for the
inertness of these compounds to PLA2 degradation.
3
0.87 (t, JH,H = 6.6 Hz, 3 H, 16Ј-H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 166.77, 153.94, 139.91, 137.62, 137.18, 134.87, 131.39,
130.07, 129.40, 128.82, 117.88, 72.27, 71.88, 70.12, 63.20, 39.54,
34.22, 33.08, 31.90, 29.68–29.61 (8 C), 29.52, 29.44, 29.35, 28.93 (2
C), 26.02, 22.68, 21.73, 19.17, 14.12, 13.94, 12.90 ppm. HRMS
(ESI): calcd. for C39H66O4Na [M + Na]+ 621.4854; found 621.4848.
1-Hexadecyl-2-[all-(E)-retinoyl]-sn-glycero-3-phosphorylcholine [(E)-
4]: A solution of 6 (166 mg, 0.28 mmol) and Et3N (50 µL,
0.36 mmol) in CH2Cl2 (4 mL) was dropwise added to a solution of
POCl3 (32 µL, 0.35 mmol) in CH2Cl2 (1.5 mL) at 0 °C over 15 min.
The solution was stirred at room temperature for 30 min, after
which pyridine (180 µL, 2.23 mmol) and choline tosylate (153 mg,
0.55 mmol) were added. The solution was stirred at room tempera-
ture for 16.5 h. Water (0.2 mL) was added, and the mixture was
stirred for 40 min. Continuous concentration with ethanol/toluene
(1:1, 20 mL) gave a yellow residue, which was redissolved in THF/
H2O (9:1) and slowly passed through an MB-3 column. The solvent
was removed by continuous concentration with ethanol/toluene
(1:1, 20 mL). The crude product was purified by flash chromatog-
raphy on silica gel (CH2Cl2/MeOH, 4:1 Ǟ CH2Cl2/MeOH/H2O,
15:10:1) to give (E)-4 as a yellow oil (84 mg, 40%). [α]2D0 = –7.7 (c
Experimental Section
General: All commercial reagents were used as supplied. Unless
otherwise stated reactions were performed under nitrogen and an-
hydrous solvents used. Flash-column chromatography was carried
out by using Merck silica gel 60 (particle size 0.040–0.063 mm).
Mixtures were stirred by using a magnetic stir-bar and the reactions
monitored by TLC on aluminum plates, pre-coated with silica gel
60 and spots visualized by dipping in an ethanolic solution of phos-
phomolybdic acid (12 g/250 mL) followed by heating. Optical rota-
tions were measured with a Perkin–Elmer 241 polarimeter, and IR
spectra were recorded with a Bruker Alpha FT-IR spectrometer.
NMR spectral characterization was carried out with either a Varian
Mercury 300 MHz instrument or a Varian Unity Inova 500 MHz
instrument as indicated in the text. High-resolution mass spectra
were recorded with an Ionspec Ultima Fourier Transform mass
spectrometer at the Department of Physics and Chemistry, Univer-
sity of Southern Denmark.
= 0.6, CHCl ). IR (neat): ν = 2923, 2853, 2490, 1706, 1458, 1238,
˜
3
1152, 1085, 967 cm–1. 1H NMR (500 MHz, CDCl3/CD3OD, 4:1): δ
3
3
= 7.03 (dd, JH,H = 11.5, 15.0 Hz, 1 H, 11-H), 6.30 (d, JH,H
=
3
16.0 Hz, 1 H, 7-H), 6.29 (d, JH,H = 15.0 Hz, 1 H, 12-H), 6.15 (d,
3JH,H = 16.0 Hz, 1 H, 8-H), 6.15 (d, JH,H = 11.5 Hz, 1 H, 10-H),
3
5.79 (s, 1 H, 14-H), 5.20–5.16 (m, 1 H, sn-2-H), 4.23 (br., 2 H,
CH2CH2NMe3), 4.09–4.02 (m, 2 H, sn-3-H), 3.67–3.62 (m, 2 H,
sn-1-H), 3.58 (br., 2 H, CH2CH2NMe3), 3.49–3.42 (m, 2 H, 1Ј-H),
3.20 (s, 9 H, NMe3), 2.34 (s, 3 H, 13-Me), 2.05–2.00 (m, 5 H, 3-
H), 1.72 (s, 3 H, 2-Me), 1.66–1.60 (m, 2 H, 4-H), 1.57–1.51 (m, 2
H, 3Ј-H), 1.49–1.47 (m, 2 H, 5-H) 1.26 (br., 26 H, 3Ј-H to 15Ј-H),
1.03 (s, 6 H, 2 6-Me), 0.88 (t, 3JH,H = 6.9 Hz, 3 H, 16Ј-H) ppm. 13
C
NMR (75 MHz, CDCl3/CD3OD, 4:1): δ = 166.49, 153.48, 139.75,
137.30, 136.87, 134.36, 131.27, 129.74, 129.00, 128.63, 117.55,
71.37, 70.90 (d, JC,P = 9.0 Hz), 68.89, 66.15 (d, JC,P = 9.0 Hz),
63.90 (d, JC,P = 5.3 Hz), 58.53 (d, JC,P = 5.3 Hz), 53.79, 53.75,
53.70, 39.19, 33.86, 32.70, 31.55, 29.33–29.29 (8 C), 29.14, 29.13,
28.99, 28.47 (2 C), 25.64, 22.30, 21.24, 18.81, 13.60, 13.39,
12.41 ppm. HRMS (ESI): calcd. for C44H78NO7P [M + Na]+
786.5408; found 786.5422.
Example for Entry 2 (Table 1): Acid 2 (100 mg, 0.33 mmol), DMAP
(61 mg, 0.50 mmol) and cetyl alcohol (161 mg, 0.66 mmol) were
dissolved in CH2Cl2 (3 mL), and DCC (1.0  in CH2Cl2, 0.50 mL,
0.50 mmol) was added dropwise. The mixture was stirred for 24 h
before flash chromatography on silica (CH2Cl2/hexane, 2:3) af-
1
forded the product. H NMR spectra of the products in Entries 1,
6 and 7 are enclosed in the Supporting Information.
1-Hexadecyl-2-[all-(E)-retinoyl]-sn-glycerol (6): Acid
2
(0.30 g,
1-Hexadecyl-2-[all-(E)-retinoyl]-sn-glycero-3-phosphoryl-1Ј-[2Ј,3Ј-bis-
1.0 mmol), Ph3P (0.44 g, 2.0 mmol) and 5 (0.36 g, 0.84 mmol) were (tert-butyldimethyl)-sn-glycerol] (7): Compound 6 (131 mg,
722
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 719–724

Synthesis of ATRA Ether Lipid Conjugates
0.22 mmol) and amidite 9 (171 mg, 0.33 mmol) were dissolved in
CH2Cl2 (5 mL), and molecular sieves (3-Å, 250 mg) were added.
After 30 min, 1H-tetrazole in MeCN (0.50 mL, 0.45 , 0.23 mmol)
was added, and the mixture was stirred for 1 h. tBuOOH in decane
(60 µL, 5.5 , 0.33 mmol) was added, and, after 1 h, another por-
tion was added (30 µL, 0.17 mmol). The mixture was then stirred
for 30 min before addition of DBU (90 µL, 0.60 mmol). After
30 min, the mixture was concentrated in vacuo. Flash chromatog-
raphy on silica gel (EtOAc, then CH2Cl2/MeOH, 4:1) gave 7
(193 mg, 89%) as a gum. [α]2D0 = –68 (c = 5.2, CHCl3). IR (neat):
wise addition of DIAD (0.20 mL, 1.0 mmol) over 30 min. The mix-
ture was stirred for 24 h, concentrated in vacuo and purified by
flash chromatography on silica gel (CH2Cl2/heptane, 1:1) to yield
the ester (127 mg). 1H NMR (300 MHz, CDCl3): δ = 6.98 (dd,
3JH,H = 11.4, 15.0 Hz, 1 H, 11-H), 6.30–6.09 (m, 4 H, 12-H, 10-H,
3
8-H, 7-H), 5.79 (s, 1 H, 14-H), 5.04 (q, JH,H = 5 Hz, 1 H, sn-2-
H), 3.80–3.70 (m, 2 H, sn-3-H), 3.63–3.54 (m, 2 H, sn-1-H), 3.50–
3.37 (m, 2 H, 1Ј-H), 2.16 (s, 3 H, 13-Me), 2.03–1.99 (m, 5 H, 9-
Me, 3-H), 1.70 (s, 3 H, 2-Me), 1.64–1.44 (m, 6 H, 4-H, 2Ј-H, 5-H),
1.24 (br., 26 H, 3Ј-H to 15Ј-H), 1.01 (s, 6 H, 2 6-Me), 0.87 (s, 12
H, SiCMe3, 16Ј-H), 0.04 (s, 6 H, SiMe2) ppm. 13C NMR (75 MHz,
ν = 2926, 2855, 1709, 1610, 1582, 1463, 1360, 1250, 1152,
˜
1105 cm–1. 1H NMR (500 MHz, CDCl3/CD3OD, 3:1): δ = 6.85 (dd, CDCl3): δ = 166.47, 152.97, 139.53, 137.64, 137.24, 135.14, 130.92,
3JH,H = 11.4, 15.0 Hz, 1 H, 11-H), 6.13–5.95 (m, 4 H, 12-H, 10-H, 129.95, 129.48, 128.59, 118.60, 72.36, 71.56, 68.89, 61.70, 39.55,
3
8-H, 7-H), 5.59 (s, 1 H, 14-H), 4.99 (q, JH,H = 5 Hz, 1 H, sn-2- 34.22, 33.07, 31.91, 30.91, 29.68–29.58 (8 C), 29.47, 29.35, 28.92 (2
H), 3.80 (m, 2 H, sn-3-H), 3.65 (m, 2 H, sn-1Ј-H), 3.58 (m, 1 H,
sn-2Ј-H), 3.46–3.36 (m, 4 H, sn-1-H, sn-3Ј-H), 3.26 (m, 2 H, 1Ј-H), 12.87, –5.41, –5.43 ppm. The ester (127 mg, 0.18 mmol) was then
2.16 (s, 3 H, 13-Me), 1.86–1.83 (m, 5 H, 9-Me, 3-H), 1.53 (s, 3 H, dissolved in CH2Cl2 (5 mL) and cooled to –78 °C before DIBAL-
C), 26.06, 25.79 (3 C), 22.68, 21.72, 19.18, 18.22, 14.11, 13.86,
2-Me), 1.46–1.28 (m, 6 H, 4-H, 2Ј-H, 5-H), 1.07 (br., 26 H, 3Ј-H H in hexane (1.0 , 0.25 mL, 0.25 mmol) was added slowly over a
to 15Ј-H), 0.85 (s, 6 H, 2 6-Me), 0.71 (s, 21 H, 2 SiCMe3, 16Ј-H), 5 min period and the mixture then stirred for 1 h. Then, MeOH
–0.08 (s, 6 H, SiMe2), –0.12 (s, 6 H, SiMe2) ppm. 13C NMR
(75 MHz, CDCl3/CD3OD, 3:1): δ = 166.39, 153.61, 139.63, 137.30,
136.91, 134.43, 131.21, 129.64, 129.06, 128.53, 117.46, 72.37 (d,
JC,P = 9.6 Hz), 71.36, 70.72 (d, JC,P = 8.7 Hz), 68.79, 66.47 (d, JC,P
(0.2 mL) was added, the mixture stirred for 15 min and satd. aq.
Rochelle’s salt solution (20 mL) added before the mixture was left
to warm up for 16 h. The mixture was diluted with water (10 mL)
and extracted with CH2Cl2 (2ϫ10 mL), dried (Na2SO4) and con-
= 5.7 Hz), 64.64, 63.72 (d, JC,P = 4.9 Hz), 39.19, 33.84, 32.69, centrated in vacuo. The crude mixture was then dissolved in
31.56, 29.33–29.16 (10 C), 29.00, 28.48 (2 C), 25.65, 25.46 (3 C), CH2Cl2 (2 mL) containing Et3N (0.2 mL) and cooled to 0 °C.
25.36 (3 C), 22.30, 21.25, 18.81, 17.94, 17.73, 13.62, 13.40, 12.40, PMBzCl (100 µL, 1.30 mmol) was added followed by a spatula tip
–5.17, –5.24, –5.89, –5.94 ppm. 31P NMR (202 MHz, CDCl3/
of DMAP. The mixture was stirred for 24 h before water (0.1 mL)
CD3OD, 3:1): δ = –2.05 ppm. HRMS (ESI): calcd. for C54H100O9- was added. The mixture was then stirred for 2 h, dried (Na2SO4)
PSi2Na2 [M + Na]+ 1025.6434; found 1025.6443.
and purified by flash chromatography on silica gel (CH2Cl2/hep-
tane, 1:1) to yield (+)-10 (56 mg, 55%). [α]2D0 = +4.8 (c = 0.8,
CHCl3). HPLC (chiralcel OD-H column, n-hexane/iPrOH, 400:1,
1-Hexadecyl-2-[all-(E)-retinoyl]-sn-glycero-3-phosphoryl-1Ј-sn-gly-
cerol (8): Compound 7 (57 mg, 58 µmol) was dissolved in MeCN
(1.8 mL), cooled to 0 °C, and 40% aq. HF (0.2 mL) was added.
Cooling was removed, and the mixture was stirred vigorously for
5 h before it was poured into saturated aq. NaHCO3 (5 mL). The
mixture was diluted with brine (10 mL) and extracted with CH2Cl2
(5 ϫ 10 mL) and then EtOAc (10 mL). The combined organic
phases were dried (Na2SO4), concentrated in vacuo, and the residue
1.00 mLmin–1): t = 6.5 min. IR (neat): ν = 2924, 2854, 1715, 1607,
˜
r
1511, 1463, 1255, 1167, 1100, 1033, 836, 770 cm–1
.
1H NMR
(300 MHz, CDCl3): δ = 8.03–7.98 (m, 2 Ar-H), 6.93–6.88 (m, 2 Ar-
3
H), 5.21 (q, JH,H = 5 Hz, 1 H, sn-2-H), 3.85–3.84 (m, 5 H, sn-3-
H, OMe), 3.69–3.68 (m, 2 H, sn-1-H), 3.52–3.40 (m, 2 H, 1Ј-H),
1.58–1.49 [m, 2 H, overlap with residual water at δ = 1.56 (2Ј-H)
ppm], 1.25 (br., 26 H, 3Ј-H to 15Ј-H), 0.87 (s, 12 H, SiCMe3, 16Ј-
H), 0.04 (s, 3 H, SiMe), 0.03 (s, 3 H, SiMe) ppm. 13C NMR
(75 MHz, CDCl3): δ = 165.79, 163.27, 131.71 (2 C), 122.80, 113.47
(2 C), 73.51, 71.60, 68.95, 61.75, 55.38, 31.91, 29.69–29.61 (9 C),
29.46, 29.35, 26.08, 25.78 (3 C), 22.68, 18.21, 14.12, –5.42,
–5.44 ppm. HRMS (ESI): calcd. for C33H60O5SiNa [M + Na]+
587.4102; found 587.4076.
was purified by flash chromatography on silica gel (CH2Cl2
Ǟ
CH2Cl2/MeOH/H2O, 65:25:4) to afford 8 (36 mg, 82%) after azeo-
tropic removal of residual water with toluene at 40–50 °C in vacuo.
[α]2D0 = –20 (c = 0.5, CHCl ). IR (neat): ν = 2924, 2853, 1709, 1611,
˜
3
1585, 1466, 1388, 1359, 1239, 1118, 1052 cm–1
.
1H NMR
3
(500 MHz, CDCl3/CD3OD, 3:1): δ = 6.84 (dd, JH,H = 11.4,
15.0 Hz, 1 H, 11-H), 6.11–5.93 (m, 4 H, 12-H, 10-H, 8-H, 7-H),
3
5.60 (s, 1 H, 14-H), 4.99 (q, JH,H = 5 Hz, 1 H, sn-2-H), 3.82 (br., 1-(tert-Butyldimethylsilyl)-3-hexadecyl-2-(4-methoxybenzoyl)-sn-
2 H, sn-3-H), 3.72 (br., 2 H, sn-1Ј-H), 3.59 (br., 1 H, sn-2Ј-H), 3.43
(m, 4 H, sn-1-H, sn-3Ј-H), 3.26 (m, 2 H, 1Ј-H), 2.15 (s, 3 H, 13-
Me), 1.83–1.82 (m, 5 H, 9-Me, 3-H), 1.52 (br., 3 H, 2-Me), 1.44
glycerol (–)-10: Compound 5 (75 mg, 0.11 mmol) and Et3N (70 µL,
0.73 mmol) were dissolved in CH2Cl2 (2 mL) and cooled to 0 °C.
Then, PMBzCl (100 µL, 1.30 mmol) was added followed by a spat-
(m, 2 H, 4-H), 1.33 (m, 2 H, 2Ј-H), 1.29 (m, 2 H, 5-H), 1.05 (br., ula tip of DMAP. The mixture was stirred for 24 h before water
26 H, 3Ј-H to 15Ј-H), 0.84 (s, 6 H, 2 6-Me), 0.68 (t, 3JH,H = 6.8 Hz,
3 H, 16Ј-H) ppm. 13C NMR (75 MHz, CDCl3/CD3OD, 3:1): δ =
(0.1 mL) was added, then stirred for another 2 h, dried (Na2SO4)
and purified by flash chromatography on silica gel (CH2Cl2/hep-
166.51, 153.85, 139.81, 137.32, 136.89, 134.39, 131.40, 129.78, tane, 1:1) to yield compound (–)-10 (38 mg, 61%). [α]2D0 = –5.2 (c
129.07, 128.68, 117.40, 71.46, 70.78 (d, JC,P = 8.3 Hz), 70.56 (d,
JC,P = 4.5 Hz), 68.80, 66.10 (d, JC,P = 3.8 Hz), 64.05, 61.87 (d, JC,P
= 1.5 Hz), 39.24, 33.90, 32.75, 31.60, 29.39–29.22 (10 C), 29.04,
28.54 (2 C), 25.70, 22.35, 21.32, 18.85, 13.67, 13.47, 12.48 ppm.
31P NMR (202 MHz, CDCl3/CD3OD, 3:1): δ = –0.79 ppm. HRMS
(ESI): calcd. for C42H72O9PNa2 [M + Na]+ 797.4704; found
797.4731.
= 1.1, CHCl3). HPLC (chiracel OD-H, n-hexane/iPrOH, 400:1,
1.00 mLmin–1): tr = 6.0 min. IR, 1H and 13C NMR spectra are
identical to those of (+)-10. HRMS (ESI): calcd. for C33H60O5SiNa
[M + Na]+ 587.4102; found 587.4086.
Supporting Information (see footnote on the first page of this arti-
1
cle): H, 13C and 31P NMR spectra of compound 6, 7 and 8 and
1H, 13C NMR spectra of 4 and 10 together with NMR assignment
tables for 4 and 8 with the applied atom numbering; HPLC chro-
matograms of compounds (+)-10 and (–)-10, NMR experiment
data, DLS measurements, liposome formulation and MALDI re-
sults for hydrolysis experiments.
3-(tert-Butyldimethylsilyl)-1-hexadecyl-2-(4-methoxybenzoyl)-sn-
glycerol [(+)-10]: Compound 5 (168 mg, 0.24 mmol) and Ph3P
(0.23 g, 0.88 mmol) were dissolved in THF (10 mL) and cooled to
0 °C. Then 2 was added (0.18 g, 0.60 mmol) followed by portion-
Eur. J. Org. Chem. 2010, 719–724
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
723

M. H. Clausen et al.
FULL PAPER
[8] E. P. Boden, G. E. Keck, J. Org. Chem. 1985, 50, 2394–2395.
[9] For these experiments 3 mg of retinoic acid was dissolved in
600 µL of CD2Cl2 and the indicated amounts of reagents
added.
[10] H. Chen, M. R. Juchau, J. Biochem. 1997, 327, 721–726.
[11] T. W. Shih, Y. F. Shealy, D. L. Strother, D. L. Hill, Drug Metab.
Dispos. 1986, 14, 698–702.
Acknowledgments
We are grateful to the Danish Strategic Research Council (NABIIT
Program) for financial support and to Ms. Anne Hector for techni-
cal support.
[12] W. R. Roush, A. P. Spada, Tetrahedron Lett. 1983, 24, 3693–
[1] a) T. L. Andresen, S. S. Jensen, K. Jørgensen, Prog. Lipid Res.
2005, 44, 68–97; b) T. L. Andresen, J. Davidsen, M. Begtrup,
O. G. Mouritsen, K. Jørgensen, J. Med. Chem. 2004, 47, 1694–
1703.
3696.
[13] A. K. Ghosh, Y. Wang, Tetrahedron Lett. 2001, 42, 3399–3401.
[14] O. Mitsunobu, Synthesis 1981, 1–28.
[15] F. Dumoulin, D. Lafont, P. Boullanger, G. Mackenzie, G. H.
Mehl, J. W. Goodby, J. Am. Chem. Soc. 2002, 124, 13737–
13748.
[16] a) Y. Hayakawa, M. Kataoka, J. Am. Chem. Soc. 1998, 120,
12395–12401; b) T. L. Andresen, D. M. Skytte, R. Madsen,
Org. Biomol. Chem. 2004, 2, 2951–2957.
[2] L. Linderoth, G. H. Peters, R. Madsen, T. L. Andresen, Angew.
Chem. Int. Ed. 2009, 48, 1823–1826.
[3] P. J. Pedersen, M. S. Christensen, T. Ruysschaert, L. Linderoth,
T. L. Andresen, F. Melander, O. G. Mouritsen, R. Madsen,
M. H. Clausen, J. Med. Chem. 2009, 52, 3408–3415.
[4] a) S. J. Freemantle, M. J. Spinella, E. Dmitrovsky, Oncogene
2003, 22, 7305–7315; b) J. L. Armstrong, C. P. F. Redfern, G. J.
Veal, Biochem. Pharmacol. 2005, 69, 1299–1306.
[17] P. R. J. Gaffney, C. B. Reese, J. Chem. Soc. Perkin Trans. 1
2001, 192–205.
[18] H. Plück, E. A. Dennis, J. Phys. Chem. 1981, 85, 678–683.
[19] a) A. A. Fawzy, B. S. Wishwanath, R. C. Franson, Agents Ac-
tions 1988, 25, 394–400; b) W. C. Hope, B. J. Patel, C. Fiedler-
Nagy, B. H. Wittreich, Inflammation 1990, 14, 543–559.
Received: October 6, 2009
[5] C. Sangmam, J.-Y. Winum, M. Lucas, J.-L. Montero, C. Chavis,
Synth. Commun. 1998, 28, 2945–2958.
[6] H. Yanaka, H. Kagechika, E. Kawachi, H. Fukasawa, Y. Hash-
imoto, K. Shudo, J. Med. Chem. 1992, 35, 567–572.
[7] K. Holmberg, B. Hansen, Acta Chem. Scand. Ser. B 1979, 33,
410–412.
Published Online: December 14, 2009
724
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 719–724