An efficient, one-pot synthesis of various ceramide 1-phosphates from sphingosine 1-phosphate

Article abstract of DOI:10.1016/j.chemphyslip.2007.10.005

An efficient, one-pot procedure for the synthesis of ceramide 1-phosphates with varying N-acyl substituents, to serve as tool compounds for analytical and biological investigations, was developed. Sphingosine 1-phosphate was silylated in situ to increase its solubility and to protect the 3-hydroxy functionality and then allowed to react with activated acid derivatives in the presence of diisopropylethylamine. Simultaneous cleavage of the silyl protecting groups and separation from reagents and by-products was achieved by medium pressure chromatography on reversed phase material. Thus, ceramide 1-phosphates with various fatty acid chains and with fluorescent and affinity labels attached to the sphingoid backbone were prepared in good yields.

Full text of DOI:10.1016/j.chemphyslip.2007.10.005

Available online at www.sciencedirect.com
Chemistry and Physics of Lipids 151 (2008) 125–128
Short communication
An efficient, one-pot synthesis of various ceramide 1-phosphates
from sphingosine 1-phosphate
Peter Nussbaumer∗, Valentin Hornillos, Michael Ghobrial, Thomas Ullrich
Novartis Institutes for BioMedical Research, Brunner Strasse 59, A-1235 Vienna, Austria
Received 21 August 2007; received in revised form 24 October 2007; accepted 31 October 2007
Available online 26 November 2007
Abstract
An efficient, one-pot procedure for the synthesis of ceramide 1-phosphates with varying N-acyl substituents, to serve as tool compounds for
analytical and biological investigations, was developed. Sphingosine 1-phosphate was silylated in situ to increase its solubility and to protect the
3-hydroxy functionality and then allowed to react with activated acid derivatives in the presence of diisopropylethylamine. Simultaneous cleavage
of the silyl protecting groups and separation from reagents and by-products was achieved by medium pressure chromatography on reversed phase
material. Thus, ceramide 1-phosphates with various fatty acid chains and with fluorescent and affinity labels attached to the sphingoid backbone
were prepared in good yields.
© 2007 Elsevier Ireland Ltd. All rights reserved.
Keywords: Biotin; Ceramide phosphate; Fluorescence; Sphingosine phosphate; Sphingolipids; Sphingolipidomics
1. Introduction
as for measuring C1P concentrations in cells and tissues
(sphingolipidomics).
Ceramide 1-phosphate (C1P) is a sphingosine-based lipid
generated from ceramide, which is the central molecule in the
metabolism of sphingolipids, by ceramide kinase. The impor-
tance of C1P as signaling molecule has only recently begun
to be recognized and investigated (Chalfant and Spiegel, 2005;
Baumruker et al., 2005; Marcu and Chalfant, 2007; Graf et
al., 2007). C1P is postulated to interact directly with cytoso-
lic phospholipase A2 alpha resulting in activation and, hence,
increased release of arachidonic acid and downstream inflam-
matory mediators (Pettus et al., 2004). Ceramide and C1P
are naturally occurring as mixtures of analogues with a con-
stant sphingosine backbone and N-acyl residues varying mainly
in length ranging from C14 to C32 with optional ␣- or ␻-
hydroxy substitution. Therefore, C1P analogues with different,
well defined N-acyl side chains (only few are commercially
available) but also labeled C1P analogues are required as
tool and reference compounds for further biological investi-
gations such as metabolism and interaction studies as well
Recently, we published two related methods to intro-
duce a label into the sphingosine backbone of sphingolipids
(Nussbaumer et al., 2005; Peters et al., 2007). The next goal
was to develop an efficient method for the rapid preparation of
C1P derivatives with various, defined labeled and unlabeled fatty
acid residues for biological investigations. Most of the known
synthesis methods for C1Ps require protection of the 3-hydroxy
group. Besideenzymaticmethodsthataremainlyusedtoprepare
radiolabeled analogues and can deliver only very minute quan-
tities, there is only one published procedure for phosphorylation
of ceramide without the need of protection protocols (Byun et
al., 1994). However, each corresponding ceramide is required
in sufficient quantity as starting material and their low solubil-
ity is often a limiting factor. More recently, phosphorylation of
N-Boc protected sphingosine and acylation of the deprotected
phosphate diester followed by ester cleavage was reported for
the synthesis of C1P (Szulc et al., 2000). We envisaged a fast
and flexible alternative method starting from sphingosine 1-
∗
Corresponding author. Fax: +43 80166 354.
E-mail address: peter.nussbaumer@novartis.com (P. Nussbaumer).
0009-3084/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.chemphyslip.2007.10.005

126
P. Nussbaumer et al. / Chemistry and Physics of Lipids 151 (2008) 125–128
Scheme 1. Synthesis of ceramide 1-phosphates 2a–f from S1P.
phosphate (S1P) which should provide C1Ps 2 with defined
N-acyl substituents in one step (Scheme 1). S1P has become
commercially available in 100 mg batches at a reasonable
price.
using a Biotage SP4 equipment. HRMS spectra were recorded
on a Bruker Daltonics 9.4T APEX-III FT-MS.
2.2. Synthesis
2. Experimental procedures
General procedure: synthesis of N-[(1S,2R,3E)-2-hydroxy-1-
[(phosphonooxy)methyl]-3-heptadecen-1-yl]-octadecanamide
(C18-ceramide 1-phosphate, 2a).
2.1. Materials
Sphingosine 1-phosphate (10.3 mg, 0.027 mmol) was stirred
with N,O-bis(trimethylsilyl)acetamide (BSA, 400 ␮l) at room
temperature overnight. After evaporation under high vacuum,
the residue was dissolved in dry dichloromethane (DCM,
400 ␮l) and treated at 0 ◦C with diisopropylethylamine (15 ␮l,
0.08 mmol) and stearoyl chloride (400 ␮l from a stock solution
of 82 mg stearoyl chloride in 4 ml of dry DCM, corresponding
to 0.027 mmol). The mixture was stirred for 2 h, concentrated
under vacuum to about 30% of volume and then soaked onto a
reversed phase chromatography samplet (BiotageTM). The sam-
plet was dried, loaded onto an RP-18 reversed phase column,
and the product was isolated by starting the chromatographic
purification with H2O:MeOH:NH4OH = 50:50:1, followed by
an increasing gradient of MeOH to achieve a final solvent mix-
ture of MeOH:NH4OH = 100:1. The fractions were analyzed
by TLC (silica gel, nBuOH:MeOH:H2O:NH4OH = 8:1:1:1)
and MS. Concentration, dissolution in a small amount of
MeOH/dichloromethane and precipitation with acetone gave
13.8 mg (79%) of 2a as a colourless powder.
Stearoyl chloride (1a), palmitoyl chloride (1b), nervonic
acid and biotin were purchased from Sigma–Aldrich, N-
succinimidyl-proprionate (1c) from Wako Pure Chemical
Industries, Ltd., NBD-aminohexanoic acid N-hydroxyl suc-
cinimidyl ester (1f) from Marker Gene Technologies, Inc.,
and d-erythro-sphingosine-1-phosphate (S1P) from Toronto
Research Chemicals, Inc. All reagents were used without further
purification.
Nervonoyl chloride (1d) and biotin chloride (1e) were pre-
pared from the corresponding acids by treatment with oxalyl
chloride for 5 h at room temperature followed by evaporation to
dryness in high vacuum and used immediately.
1
The purity of products 2 was checked by H NMR, and
the structures were further confirmed by 13C NMR, 31P NMR
and HRMS. NMR spectra were recorded on a 400 MHz Bruker
Avance spectrometer and chemical shifts are reported in δ units
(ppm) relative to tetramethylsilane and phosphoric acid, respec-
tively, as internal standard. Chromatography was performed

P. Nussbaumer et al. / Chemistry and Physics of Lipids 151 (2008) 125–128
127
Analogously, the following compounds were pre-
pared, with the only exception that 0.2 equivalents of
4-dimethylaminopyridine (DMAP) was added when N-
hydroxysuccinimide esters 1c,f (1.1 equivalents) were
used.
N-[(1S,2R,3E)-2-hydroxy-1-[(phosphonooxy)methyl]-3-
heptadecenyl]-hexadecanamide (C16-ceramide 1-phosphate,
2b), 82% yield.
(dd, J = 15.2 + 7.3 Hz, 1H; H-3), 3.97–4.18 (m, 4H), 3.48–3.57
(m, 2H; CH2 NH ), 2.20–2.28 (m, 2H; COCH2 ), 2.00
(qua, J = 7 Hz, 2H; C CH2 ), 1.79 (qui, J = 7.5 Hz, 2H), 1.68
(qui, J = 7.6 Hz, 2H), 1.48 (qui, J = 7.6 Hz, 2H), 1.20–1.38 (m,
22H), 0.88 (t, J = 7 Hz, 3H; CH3); 13C NMR (CD3OD): δ
174.7 (C O), 138.4 (Ar–C), 135.0 ( CH ), 131.0 ( CH ), 99.4
(Ar–C), 72.6 ( CH(OH) ), 66.2, 55.0, 44.2, 36.7, 33.4, 33.0,
30.8. 30.7, 30.5, 30.4, 30.3, 28.8, 27.6, 26.6, 23.7, 14.4; 31P
NMR (CD3OD): δ 2.77; ESI-MS−: 654.4 [M−]; HRMS: calcd
for [M − H] 654.3273, found 654.3279.
N-[(1S,2R,3E)-2-hydroxy-1-[(phosphonooxy)methyl]-3-
heptadecenyl]-propanamide (C3-ceramide 1-phosphate, 2c),
86% yield:
1H NMR (CD3OD): δ 5.72 (dt, J = 15.2 + 6.7 Hz, 1H; H-
4), 5.48 (dd, J = 15.2 + 7.6 Hz, 1H; H-3), 4.10–4.24 (m, 2H),
3.90–3.95 (m, 2H), 2.22 (qua, J = 7.6 Hz, 2H; COCH2 ), 2.04
(qua, J = 7.1 Hz, 2H; C CH2 ), 1.27–1.40 (m, 22H), 1.13 (t,
J = 7.6 Hz, 3H; CH3), 0.94 (t, J = 6.8 Hz, 3H; CH3); 13C NMR
(CD3OD): δ 177.1 (C O), 135.1 ( CH ), 131.5 ( CH ), 73.0
( CH(OH) ), 65.5 ( OCH2 ), 55.9( CH(NH) ), 34.6, 33.8,
33.5, 31.2, 31.1, 31.0, 30.9, 30.8, 30.7, 24.1, 14.8, 10.9; 31P
NMR (CD3OD): δ 2.77; ESI-MS−: 434.2 [M−]; HRMS: calcd
for [M + Na] 458.2642, found 458.2643.
3. Results and discussion
We found only one report of direct acylation of S1P where
excess acetic anhydride and triethylamine were used to gen-
erate C2–C1P (Gijsbers et al., 1999). The same group also
prepared tritium labeled N-acetyl- and N-hexanoyl-sphinganine
1-phosphate from sphinganine 1-phosphate by this method
(De Ceuster et al., 1995). However, in both cases the 3-
hydroxy group in the sphingosine/sphinganine backbone was
also acylated and the ester intermediates had to be selectively
cleaved. Moreover, we expected complications in the reac-
tion and product purification when using longer, less soluble
fatty acid derivatives. In our first experiments, we attempted
to use the N-hydroxysuccinimide ester of octadecanoic acid
to achieve selective acylation, since such activated esters are
known to react highly preferentially with an amino over a
hydroxy group. As solvent, we chose pyridine because S1P
is very sparingly soluble in other organic solvents including
dimethylformamide and dimethylsulfoxide. At room temper-
ature, no conversion of S1P could be detected either in
the presence or absence of triethylamine; only at 60 ◦C
for 48 h some product formation was observed (TLC and
MS detection). When 4-dimethylaminopyridine (DMAP) was
used as catalyst, very slow conversion of S1P was seen
already at room temperature, whereas a reasonable reaction
time (15 h) for nearly complete conversion was observed at
60 ◦C. However, for full consumption of the starting mate-
rial at least 2 equivalents of hydroxysuccinimide ester had
to be applied and under these conditions also some double-
acylated product (N- and 3-O-acylation) was generated. This
by-product and unreacted materials could be separated from
the desired C18-ceramide phosphate (2a, Scheme 1) by tritu-
rating with ethyl acetate followed by chromatography (silica
gel, n-BuOH:MeOH:NH4OH:H2O = 8:1:1:1), but we were not
satisfied with the procedure.
N-[(1S,2R,3E)-2-hydroxy-1-[(phosphonooxy)methyl]-
3-heptadecenyl]-(15Z)-tetracosenamide
1-phosphate, 2d), 74% yield:
(C24:1-ceramide
1H NMR (CD3OD): δ 5.70 (dt, J = 15.2 + 6.7 Hz, 1H; H-4),
5.45 (dd, J = 15.2 + 7.6 Hz, 1H; H-3), 5.30–5.39 (m, 2H; (Z)-
CH CH ), 4.10–4.23 (m, 2H), 3.84–3.91 (m, 2H), 2.13–2.24
(m, 2H; COCH2 ), 1.98–2.07 (m, 6H), 1.54–1.63 (m, 2H),
1.25–1.40 (m, 54H), 0.90 (t, J = 6.8 Hz, 6H; 2 × CH3); 13C
NMR (CD3OD): δ 176.3 (C O), 135.1 ( CH ), 131.6 ( CH ),
131.3 (2 × CH ), 72.8 ( CH(OH) ), 65.5 ( OCH2 ), 55.9
( CH(NH) ), 37.8, 33.9, 33.5, 33.4, 31.3, 31.2, 31.1, 31.0, 30.9,
30.8, 30.7, 28.5, 27.5, 24.1, 14.9; 31P NMR (CD3OD): δ 3.11;
ESI-MS−: 726.4 [M−]; HRMS: calcd for [M + H] 728.5953,
found 728.5951.
N-[(1S,2R,3E)-2-hydroxy-1-[(phosphonooxy)methyl]-
3-heptadecenyl]-5-[(3aS,4S,6aR)-hexahydro-2-oxo-1H-
thieno[3,4-d]imidazol-4-yl]-pentanamide (biotin-C5-ceramide
1-phosphate, 2e), 53% yield:
1H NMR (CD3OD/d6-DMSO): δ 5.70 (dt, J = 15 + 6.7 Hz,
1H; H-4), 5.47 (dd, J = 15.2 + 7.5 Hz, 1H; H-3), 4.46–4.52
(m, 1H), 4.28–4.34 (m, 1H), 4.07–4.27 (m, 2H), 3.91–4.00
(m, 2H), 3.17–3.25 (m, 1H), 2.92 (dd, J = 5 + 12.6 Hz, 1H),
2.70 (d, J = 12.6 Hz, 1H), 2.22 (t, J = 7.4 Hz, 2H; COCH2 ),
2.03 (qua, J = 7 Hz, 2H; C CH2 ), 1.23–1.78 (m, 28H), 0.90
(t, J = 6.8 Hz, 3H; CH3); 13C NMR (CD3OD/d6-DMSO): δ
174.3 (C O), 133.4 ( CH ), 129.7 ( CH ), 71.1 ( CH(OH) ),
64.1 ( OCH2 ), 61.9, 60.2, 55.5, 54.0, 39.7, 35.5, 32.0, 31.7,
29.4, 29.3, 29.0, 28.9, 28.4, 28.0, 25.4, 22.3, 13.1; 31P NMR
(CD3OD): δ 3.11; ESI-MS−: 604.4 [M−]; HRMS: calcd for
[M + Na] 628.3156, found 628.3157.
The most limiting factor we experienced in these initial
experiments was the very poor solubility of S1P which did
not allow much variation of reaction conditions. Therefore, we
planned to apply in situ intermediate protection of the phos-
phate group, still avoiding a multi-step sequence, and opted
for a silyl protecting group which would be cleaved during
work-up. For amino acids this strategy had been known for a
long time (Birkofer et al., 1959) and during our investigations
the application to phosphinate amino acids was reported (Li et
al., 2007). The in situ silylation was achieved by reacting S1P
with excess of neat N,O-bis(trimethylsilyl)acetamide (BSA).
N-[(1S,2R,3E)-2-hydroxy-1-[(phosphonooxy)methyl]-3-
heptadecenyl]-6-[(7-nitro-2,1,3-benzoxadiazol-4-yl)amino]-
hexanamide (NBD-C6-ceramide 1-phosphate, 2f), 76%
yield:
1H NMR (CD3OD): δ 8.53 (d, J = 8.9 Hz, 1H; Ar–H), 6.37 (d,
J = 8.9 Hz, 1H; Ar–H), 5.70 (dt, J = 15.3 + 6.7 Hz, 1H; H-4), 5.45

128
P. Nussbaumer et al. / Chemistry and Physics of Lipids 151 (2008) 125–128
Within minutes, S1P dissolved completely and after evapora-
tion under high vacuum the silylated S1P also dissolved well
in dry dichloromethane. Now the conversion with octadecanoyl
chloride (1a) in the presence of diisopropylethylamine (DIEA)
proceeded smoothly at room temperature without formation of
the double-acylated product and with nearly complete consump-
tion of S1P (the silylated intermediates could not be analysed
due to their instability). The only remaining issue was the isola-
tion and purification of the ceramide phosphate 2a since aqueous
treatment to remove the silyl protecting groups resulted in a very
poorly soluble crude product and simple precipitation with ace-
tone from dichloromethane/methanol solution did not provide
pure material. The optimised procedure we developed was to
directly load the crude reaction mixture together with the sol-
vent on a RP-18 reversed phase chromatography samplet and
to perform the chromatographic purification on reversed phase
material without any work-up. The presence of water in the start-
ing eluent was sufficient to cleave the silyl protecting groups and
we never isolated a silylated S1P or C1P derivative by applying
this procedure. Thus, the final one-pot procedure became very
simple, i.e. silylation of S1P, concentration, acylation and finally
chromatographic purification.
In summary, the intermediate in situ silylation of S1P
increases its solubility in organic aprotic solvents dramatically
and simultaneously renders the amino group readily accessi-
ble for acylation with structurally diverse acid chlorides and
hydroxysuccinimide esters. The cleavage of the silyl protecting
groups during the chromatographic purification avoids handling
of poorly soluble crude products. All reactions studied pro-
ceeded well with moderate to good yields (Scheme 1) and we
never observed a double-acylated product, as the 3-hydroxy
group of the sphingosine backbone was also intermediately
silylated under the described conditions (confirmed by NMR
experiments of the silylated S1P). Occasionally, the isolated
products 2 contained some residual DIEA which could be easily
removed under high vacuum. Thus, this new method provides
quick and easy access to ceramide 1-phosphate derivatives with
various acyl residues.
References
Baumruker, T., Bornancin, F., Billich, A., 2005. The role of sphingosine and
ceramide kinases in inflammatory responses. Immunol. Lett. 96, 175–185.
Birkofer, L., Konkol, W., Ritter, A., 1959. Peptide synthesis with silyl amino
acids. Angew. Chem. 71, 701.
The actual goal was to develop a method allowing broad vari-
ation of the N-acyl residue. Therefore, we reacted the in situ
generated silylated S1P with additional fatty acid chlorides 1b
and 1d (Scheme 1). The corresponding products 2b and 2d were
again obtained in high purity and good yield (82% and 74%,
respectively). Then we applied the hydroxysuccinimide ester of
propanoic acid (1c) as a model reagent, in combination with
DMAP as catalyst to guarantee fast conversion. In this reac-
tion we also obtained the desired ceramide phosphate derivative
(2c) in high yield (86%) demonstrating that less activated acid
derivatives can be used as acylation reagents without the need
for more drastic reaction conditions. This result is in contrast
with the observation in an initial experiment that acylation of
(non-silylated) S1P in pyridine with the hydroxysuccinimide
ester of octadecanoic acid was only achieved at 60 ◦C. Next,
we investigated the synthesis of C1P derivatives featuring two
different labels in the fatty acid residue (biotin and the fluo-
rescent dye 7-nitro-4-benzofurazanamine (NBD); Scheme 1).
The use of biotin chloride (1e) gave product 2e in moderate
yield (53%), which was mainly caused by the very poor sol-
ubility of the product and not due to a lower conversion. The
introduction of the NBD-label was accomplished by using the
activated ester 1f, again in good yield (76%), confirming that
under the optimised conditions with the in situ silylation, the
conversion of hydroxysuccinimide esters occurs readily at room
temperature. Compound 2f is a useful tool compound to comple-
ment metabolism studies with the corresponding, commercially
available NBD-C6-ceramide.
Byun, H.-S., Erukulla, R.K., Bittman, R., 1994. Synthesis of sphingomyelin
and ceramide 1-phosphate from ceramide without protection of the allylic
hydroxyl group. J. Org. Chem. 59, 6495–6498.
Chalfant, C.E., Spiegel, S., 2005. Sphingosine 1-phosphate and ceramide 1-
phosphate: expanding roles in cell signaling. J. Cell Sci. 118, 4605–4612.
De Ceuster, P., Mannaerts, G.P., Van Veldhoven, P.P., 1995. Identification and
subcellular localization of sphinganine-phosphatases in rat liver. Biochem.
J. 311, 139–146.
Gijsbers, S., Mannaerts, G.P., Himpens, B., Van Veldoven, P.P., 1999. N-Acetyl-
sphingenine-1-phosphate is a potent calcium mobilizing agent. FEBS Lett.
453, 269–272.
Graf, C., Rovina, P., Tauzin, L., Schanzer, A., Bornancin, F., 2007. Enhanced
ceramide-induced apoptosis in ceramide kinase overexpressing cells.
Biochem. Biophys. Res. Commun. 354, 309–314.
Li, S., Whitehead, J.K., Hammer, R.P., 2007. Application of in situ silylation
for improved, convenient preparation of fluorenylmethoxycarbonyl (Fmoc)-
protected phosphinate amino acids. J. Org. Chem. 72, 3116–3118.
Marcu, A.C., Chalfant, C.E., 2007. Ceramide-1-phosphate, a new bioactive
sphingolipid in regulating cell signaling. Future Lipidol. 2, 75–84.
Nussbaumer, P., Ettmayer, P., Peters, C., Rosenbeiger, D., Ho¨genauer, K., 2005.
One-step labelling of sphingolipids via a scrambling cross-metathesis reac-
tion. Chem. Commun., 5086–5087.
Peters, C., Billich, A., Ghobrial, M., Ho¨genauer, K., Ullrich, T., Nussbaumer,
P., 2007. Synthesis of BODIPY-labeled sphingosine derivatives by cross-
metathesis reaction. J. Org. Chem. 72, 1842–1845.
Pettus, B.J., Bielawska, A., Subramanian, P., Wijesinghe, D.S., Maceyka, M.,
Leslie, C.C., Evans, J.H., Freiberg, J., Roddy, P., Hannun, Y.A., Chalfant,
C.E., 2004. Ceramide 1-phosphate is a direct activator of cytosolic phospho-
lipase A2. J. Biol. Chem. 279, 11320–11326.
Szulc, Z.M., Hannun, Y.A., Bielawska, A., 2000. A facile regioselective syn-
thesis of sphingosine 1-phosphate and ceramide 1-phosphate. Tetrahedron
Lett. 41, 7821–7824.