Synthesis of α-galactosyl ceramide (KRN7000) and analogues thereof via a common precursor and their preliminary biological assessment

Article abstract of DOI:10.1021/jo8019994

(Chemical Equation Presented) A new practical synthesis of α-GalCer and of its analogues is presented, opening the chance to easily modify the sphingosine chain. The common precursor is a disaccharide, obtained by coupling tetra-O-benzyl-D-galactose with allyl 2,3-O-isopropylidene-D-lyxofuranoside. Introduction of alkyl chains via Wittig reaction (for α-GalCer and OCH) or via Williamson reaction (for oxa analogues) followed by standard synthetic steps allows one to efficiently obtain such compounds. The analogues are able to activate iNKT cells when presented by CD1d expressing cells.

Full text of DOI:10.1021/jo8019994

The crystal structure of CD1d alone or complexed with
R-GalCer has been described,2 highlighting the importance of
the length of the acyl and the alkyl chains in modulating TCR
binding affinity and cytokine bias.3 Several syntheses of
R-GalCer have been reported, based on the use of various
galactosyl donors (e.g., phosphates, imidates, halides, thiosugars,
etc.)4 and a sphingosine or a ceramide acceptor, with variable
efficiency.5
Synthesis of r-Galactosyl Ceramide (KRN7000)
and Analogues Thereof via a Common Precursor
and Their Preliminary Biological Assessment
Mario Michieletti,† Antonio Bracci,† Federica Compostella,‡
Gennaro De Libero,§ Lucia Mori,§ Silvia Fallarini,†
Grazia Lombardi,† and Luigi Panza*,†
Here we describe a novel approach to the synthesis of
R-GalCer (11a) and R-GalCer analogues starting from a
disaccharide and building the lipidic part on the reducing end,
thus avoiding the difficulties of glycosylation reactions on
ceramide acceptors and opening an easy access to R-GalCer
and various analogues modified on the sphingosine chain.6
We exploited D-lyxose as precursor, with the suitable stere-
ochemistry, of the phytosphingosine moiety. However, differ-
ently from previous approaches7 in which lyxose was converted
into phytosphingosine before the glycosylation, we reversed the
sequence by previously forming a 5-galactosyl lyxoside and
introducing the lipid chain or mimics thereof on the obtained
disaccharide.
Dipartimento di Scienze Chimiche, Alimentari,
Farmaceutiche e Farmacologiche, UniVersita` del Piemonte
Orientale “Amedeo AVogadro”, Via BoVio 6, 28100 NoVara,
Italy, Dipartimento di Chimica, Biochimica e Biotecnologie
per la Medicina, UniVersita` di Milano, Via Saldini 50, 20133
Milano, Italy, and Experimental Immunology, Department of
Biomedicine, Basel UniVersity Hospital, Hebelstrasse 20,
4031-Basel, Switzerland
panza@pharm.unipmn.it
ReceiVed September 11, 2008
Our synthetic route started from the bromide 1, obtained from
commercial tetra-O-benzyl-D-galactose by reaction with oxalyl
bromide, which was coupled with the D-lyxose derivative 2 in
presence of tri(1-pyrrolidine)phosphine oxide as activating
agent8 (Scheme 1), giving 3 together with traces of its ꢀ-anomer
in 95% yield (R/ꢀ >95:5). Careful chromatography gave pure
3 in 84% yield. Acceptor 2 was easily obtained from D-
mannofuranose following the procedure of Brimacombe et al.9
The obtained disaccharide 3 was then deallylated to give the
key intermediate 4 in 90% yield.
Disaccharide 4 is a versatile compound for the synthesis of
either R-GalCer or its alkyl analogues (e.g., OCH) by Wittig
olefination or for the synthesis of oxa analogues by Williamson
alkylation.
A new practical synthesis of R-GalCer and of its analogues
is presented, opening the chance to easily modify the
sphingosine chain. The common precursor is a disaccharide,
obtained by coupling tetra-O-benzyl-D-galactose with allyl
2,3-O-isopropylidene-D-lyxofuranoside. Introduction of alkyl
chains via Wittig reaction (for R-GalCer and OCH) or via
Williamson reaction (for oxa analogues) followed by standard
synthetic steps allows one to efficiently obtain such com-
pounds. The analogues are able to activate iNKT cells when
presented by CD1d expressing cells.
(2) Koch, M.; Stronge, V. S.; Shepard, D.; Gadola, S. D.; Mathew, B.; Ritter,
G.; Fersht, A. R.; Besra, G. S.; Schmidt, R. R.; Jones, E. Y.; Cerundolo, V. Nat.
Immunol. 2005, 8, 819–826.
(3) (a) Goff, R. D.; Gao, Y.; Mattner, T.; Zhou, D.; Yin, N.; Cantu, C., III;
Teyton, L.; Bendelac, A.; Savage, P. B. J. Am. Chem. Soc. 2004, 126, 13602–
13603. (b) McCarthy, C.; Shepherd, D.; Fueire, S.; Stronge, V. S.; Koch, M.;
Illarionov, P. A.; Bossi, G.; Salio, M.; Denkberg, G.; Reddington, F.; Tarlton,
A.; Reddy, B. G.; Schmidt, R. R.; Reiter, Y.; Griffiths, G. M.; Van der Merwe,
P. A.; Besra, G. S.; Jones, E. Y.; Batista, F. D.; Cerundolo, V. J. Exp. Med.
2007, 1131–1144.
(4) Boutureira, O.; Morales-Serna, J. A.; D´ıaz, Y.; Matheu, M. I.; Castillo´n,
S. Eur. J. Org. Chem. 2008, 1851–1854, and references cited therein.
(5) Morales-Serna, J. A.; Boutureira, O.; D´ıaz, Y.; Matheu, M. I.; Castillo´n,
S. Carbohydr. Res. 2007, 342, 1595–1612.
R-GalCer (KRN7000) is one of the most powerful activating
glycolipids when presented by CD1d proteins expressed on
antigen-presenting cells (e.g., monocytes, dendritic cells, and
B cells). The stimulation of iNKT cells by engagement of the
T-cell receptor initiates a cascade of events leading to the release
of different Th1- or Th2-type cytokines.1
(6) For some recent examples of analogues, see: (a) Wipf, P.; Pierce, J. G.
Org. Lett. 2006, 8, 3375–3378. (b) Fuhshuku, K.; Hongo, N.; Tashiro, T.; Masuda,
Y.; Nakagawa, R.; Seino, K.; Taniguchi, M.; Mori, K. Bioorg. Med. Chem. 2008,
16, 950–964. (c) Liu, Y.; Deng, S.; Bai, L.; Freigang, S.; Mattner, J.; Teyton,
L.; Bendelac, A.; Savage, P. B. Bioorg. Med. Chem. Lett. 2008, 18, 3052–3055.
(d) Park, J.-J.; Lee, J. H.; Ghosh, S. C.; Bricard, G.; Venkataswamy, M. M.;
Porcelli, S. A.; Chung, S.-K. Bioorg. Med. Chem. Lett. 2008, 18, 3906–3909.
(e) Dere, R. T.; Zhu, X. Org. Lett. 2008, 10, 4641–4644. (f) Leung, L.; Tomassi,
C.; Van Beneden, K.; Decruy, T.; Elewaut, D.; Elliott, T.; Al-Shamkhani, A.;
Ottensmeier, C.; Van Calenbergh, S.; Werner, J.; Williams, T.; Linclau, B Org.
Lett. 2008, 10, 4433–4436.
(7) Fan, G. T.; Pan, Y. S.; Lu, K. C.; Cheng, Y. P.; Lin, W. C.; Lin, S.; Lin,
C. H.; Wong, C. H.; Fang, J. M.; Lin, C. C. Tetrahedron 2005, 1855–1862.
(8) Mukaiyama, T.; Kobashi, Y. Chem. Lett. 2004, 1, 10–11.
(9) Brimacombe, J. S.; Hunedy, F.; Tucker, L. C. N. J. Chem. Soc. 1968,
1381–1384.
† Universita` del Piemonte Orientale.
‡ Universita` di Milano.
§ Basel University Hospital.
(1) (a) Tsuji, M. Cell. Mol. Life Sci. 2006, 1889. (b) Wu, D.; Zajonc, D. M.;
Fujio, M.; Sullivan, B. A.; Kinjo, Y.; Kronenberg, M.; Wilson, I. A.; Wong,
C. H. Proc. Natl. Acad. Sci. U.S.A. 2006, 11, 3972. (c) De Libero, G.; Mori, L.
Nat. ReV. Immunol. 2005, 5, 485. (d) Chang, Y. J.; Huang, J. R.; Tsai, Y. C.;
Hung, J. T.; Wu, D.; Fujio, M.; Wong, C.-H.; Yu, A. L. Proc. Natl. Acad. Sci.
U.S.A. 2007, 25, 10299.
9192 J. Org. Chem. 2008, 73, 9192–9195
10.1021/jo8019994 CCC: $40.75 2008 American Chemical Society
Published on Web 10/29/2008

SCHEME 1. Key Intermediate Synthesis
SCHEME 2. r-GalCer and OCH Synthesis
To obtain R-GalCer and OCH, Wittig olefination was
performed using the ylids derived from phosphonium salt
obtained from n-tridecyl- or n-butylbromide. Opening of the
furanosidic ring due to the Wittig reaction let the 2R-hydroxyl
group ready for the introduction of the azido group via SN2
after reduction of the resulting double bond via catalytic
hydrogenation with Pt/C. Attempts to introduce the azido group
before the double bond reduction in order to reduce both
functional groups in a single step gave sluggish results during
the substitution with NaN3, possibly due to an intramolecular
cycloaddition between the azido group and the double bond.
Activation of the hydroxyl group as chloromethanesulfonyl
ester10 allowed the subsequent substitution with NaN3 to give
the desired 2S-azide with inversion of configuration.
The azido group was then reduced by mild catalytic hydro-
genation with Lindlar catalyst to the amines 9a,b, which were
condensed with hexacosanoic acid in presence of EDC, DIPEA,
and HOBt11 to give 10a,b.
Hydrolysis of the isopropylidene and catalytic hydrogenolysis
of the benzyl groups gave pure R-GalCer 11a and OCH 11b
(Scheme 2), with an overall yield (from compound 4) of 40%
for R-GalCer and 34% for OCH.
In a similar way, we have been able to perform the synthesis
of a series of oxa analogues simply by alkylation of the alcohol
14, derived from 4, then following a similar strategy to that
used for the alkyl derivatives.
Compound 14 is the starting point to obtain a new family of
analogues of R-GalCer, with the presence of oxygen atom(s)
on the alkyl chain.
Such isosteric substitution is expected not to change signifi-
cantly the structure of the final compounds, but it could modify
the electronic properties, allowing a possible modulation of their
biological properties. Using different alkyl halides and an alkyl-
O-mesylate (see Scheme 3), we built a small library of the
above-mentioned compounds.
introduced using the same protocol shown for compound 7.
Removal of pivaloyl ester with tetrabutylammonium hydroxide
in dioxane let the primary OH free for the subsequent alkylations
(Scheme 3).
Alkylations were performed following the classical William-
son’s synthesis of ethers, where alcohol 14 was deprotonated
and alkylated by an alkyl halide or mesylate (Scheme 3).
The last steps to each new compound followed the same
strategy previously shown for R-GalCer: reduction of the azide
by catalytic hydrogenation (using the Lindlar catalyst) to amine
and subsequent condensation with hexacosanoic acid in presence
of EDC, DIPEA, and HOBt.11
Final deprotection of the isopropylidene group using hydro-
chloric acid in dioxane and benzyl groups by catalytic hydro-
genation with Pd(OH)2 on charcoal afforded the target analogues
16a-d.
The biological activities of the newly synthesized compounds
were determined by a classical T-cell antigen presentation assay,
using R-GalCer as positive control. Such preliminary screening
allows one to evaluate how modifications on the sphingosine
alkyl chain affect their activity with respect to R-GalCer. Figure
1 shows that when R-GalCer-specific iNKT hybridoma cells
are stimulated with CD1d-transfected THP-1 cells, previously
exposed to increasing concentrations (10-7-10-5 g/mL) of
either R-GalCer or compounds under evaluation for 2 h, a
significant increase of IL-2 is detected in the medium 48 h later.
The analysis of these curves demonstrates that all compounds
were active in stimulating IL-2 release from iNKT cells (Figure
1).
To explore different possibilities of substitution, we began
synthesizing four analogues: one with a second atom of oxygen
in the chain (16a), one with a saturated cycle (16b), one with
an aromatic cycle (16c), and one with a simple alkylic chain
(16d). To obtain such analogues, compound 4 was reduced with
NaBH4 to give the corresponding galactosyl lyxitol 12, and the
primary alcohol was then selectively protected as pivaloyl ester
to give 13. On compound 13, the azido group in position 2 was
(10) Matto, P.; Modica, E.; Franchini, L.; Facciotti, F.; Mori, L.; De Libero,
G.; Lombardi, G.; Fallarini, S.; Panza, L.; Compostella, F.; Ronchetti, F. J. Org.
Chem. 2007, 72, 7757–7760.
(11) Murata, K.; Toba, T.; Nakanishi, K.; Takahashi, B.; Yamamura, T.;
Miyake, S.; Annoura, H. J. Org. Chem. 2005, 70, 2398–2401.
J. Org. Chem. Vol. 73, No. 22, 2008 9193

SCHEME 3. Synthesis of Oxa Analogues of r-GalCer
FIGURE 1. IL-2 release by iNKT hybridoma cells upon stimulation
with the new glycolipid analogues.
galactopyranosyl bromide,12 dissolved in 65 mL of anhydrous
dichloromethane, was added, giving a colorless solution; 19.45
mmol (5.00 g, 4.46 mL) of tri(1-pyrrolidine)phosphine oxide was
added dropwise, and the reaction was stirred under Ar atmosphere
for 24 h at room temperature. Solution was diluted with AcOEt
and filtered on a plug of Celite. The crude obtained after evaporation
of the solvent was purified by flash chromatography to give 4.12 g
of pure allyl 2,3,4,6-tetra-O-benzyl-R-D-galactopyranosyl-(1f5)-
2,3-O-isopropylidene-R-D-lyxofuranoside 3 in 84% yield.
Rf 0.68 (ETP/AcOEt ) 8/2); [R]rt ) +7.0 (c ) 1.0, CHCl3);
D
1H NMR (300 MHz, CDCl3) δ 7.40-7.24 (m, 20 H), 5.85 (ddd, 1
H, J ) 5.2, 6.1, 10.4 Hz), 5.27-5.14 (m, 2 H), 5.02 (s, 1 H),
4.97-4.56 (m, 7 H), 4.47 and 4.42 (2d, 2 H, J ) 11.9 Hz), 4.26
(m, 1 H), 4.14-3.88 (m, 7 H), 3.86 (dd, 1 H, J ) 2.8, 6.1 Hz),
3.55 (m, 2 H), 1.41 (s, 3 H), 1.29 (s, 3 H); 13C NMR (75 MHz,
CDCl3) δ 139.0, 138.8, 138.1, 134.0, 128.5-127.6, 117.6, 112.6,
105.4, 97.8, 85.2, 79.9, 79.0, 78.4, 76.5, 75.1, 74.9, 73.5, 73.2, 73.1,
69.2, 68.8, 26.2, 25.1. Anal. Calcd for C45H52O10: C, 71.79; H, 6.96.
Found: C, 71.90; H, 7.01.
These results suggest that the performed structural modifica-
tions on the sphingoid chain led to derivatives that retain strong
iNKT cell-stimulatory activity similar to that of R-GalCer. It is
worthy to note that the presence of the oxygen seems not to
have a major impact on the biological activity since compound
16a, which contains two oxygen atoms, is as active as R-GalCer.
Moreover, the activity is only slightly decreased when the chain
ends with a ring, either aliphatic (as in 16b) or aromatic (as in
16c).
In conclusion, the synthetic route shown here allows an easy
access to different carba (such as OCH) or oxa analogues of
R-GalCer. The flexibility of our approach permits, in principle,
introduction of any kind of side chain on the ceramide moiety
and provides a powerful tool for analogue development and for
elucidation of biological functions. The preliminary biological
testing showed that the novel oxa analogues are able to activate
efficiently iNKT cells, opening new possibilities for studying
the role of these immunoregulatory cells.
Under Ar atmosphere, 1.39 mmol (1.05 g) of 3 was dissolved in
14 mL of anhydrous DMSO, and then 2.08 mmol (0.23 g) of tBuOK
was added, giving a brown solution. The reaction was stirred at 80
°C for 1 h, and then it was allowed to cool to ambient temperature.
t
A drop of water was added to destroy the excess of BuOK, and
the reaction was diluted with AcOEt. Organic phase was washed
three times with water and once with brine. After anhydrification
with Na2SO4, solvent was evaporated on vacuum. The crude was
then dissolved in 28 mL of THF and 2.78 mmol (0.71 g) of I2,
5.56 mmol (0.44 g, 0.45 mL) of pyridine, and 5 mL of water was
added. After 1 h stirring at room temperature, the volume of solvent
was reduced, and solution was diluted with AcOEt. Organic phase
was washed three times with a 5% solution of Na2S2O3, with 1 N
HCl, with a saturated solution of NaHCO3, and with water and
brine.
After anhydrification with Na2SO4, solvent was evaporated under
vacuum, and the crude was purified by flash chromatography, giving
0.891 g of pure 4 in 90% yield.
Rf 0.09 (ETP/AcOEt ) 8/2); [R]rt ) +30.9 (c ) 2.5, CHCl3);
D
1H NMR (300 MHz, CDCl3) δ 7.40-7.23 (m, 20 H), 5.36 (s, 1
H), 4.98-4.67 (m, 4 H), 4.88 (d, 1 H, J ) 3.0 Hz), 4.55 (d, 1 H,
J ) 11.5 Hz), 4.54 (d, 1 H, J ) 5.8 Hz), 4.45 and 4.40 (2d, 2 H,
J ) 11.9 Hz), 4.44-4.42 (m, 1 H), 4.06-3.98 (m, 4 H), 3.90-3.77
(m, 1 H), 3.53 (d, 2 H, J ) 6.3 Hz), 3.29 (br s, 1 H), 1.41 (s, 3 H),
Experimental Section
Key Intermediate (4) Synthesis. A quantity of 6.48 mmol (1.50
g) of allyl 2,3-O-isopropylidene-R-D-lyxofuranoside was coevapo-
rated twice with anhydrous toluene and left under high vacuum
for 2 h in the presence of activated molecular sieves (0.4 nm). Under
Ar atmosphere, 8.43 mmol (5.09 g) of 2,3,4,6-tetra-O-benzyl-D-
(12) Grayson, E. J.; Ward, S. J.; Hall, A. L.; Rendle, P. M.; Gamblin, D. P.;
Batsanov, A. S.; Davis, B. G. J. Org. Chem. 2005, 70, 9740–9754.
9194 J. Org. Chem. Vol. 73, No. 22, 2008

1.29 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 139.0, 138.8, 138.7,
128.5-128.4, 112.6, 101.1, 98.4, 98.0, 96.8, 85.5, 79.1, 78.8, 73.5,
73.3, 73.1, 69.3, 68.8, 66.5, 66.1, 60.6, 26.2, 25.2. Anal. Calcd for
C42H48O10: C, 70.77; H, 6.79. Found: C, 70.68; H, 6.65.
General Procedure for Wittig Reaction. Under Ar atmosphere,
7.00 mmol of the appropriate phosphonium bromide was dissolved
in 30 mL of anhydrous THF in the presence of activated molecular
sieves (0.4 nm) and cooled to 0 °C; 7.00 mmol (4.38 mL) of nBuLi
(solution 1.6 M in hexane) was added. After 15 min, 2.80 mmol
(2.00 g) of compound 4, dissolved in 10 mL of anhydrous THF,
was added dropwise, and the solution was allowed to warm to
ambient temperature.
The solution was stirred at room temperature for 2 h and
quenched by addition of solid NH4Cl, diluted with AcOEt, washed
four times with a saturated solution of NH4Cl, and with water and
brine. After anhydrification with Na2SO4, solvent was removed
under vacuum, and the crude was purified by flash chromatography
to give the product.
was added, and after 15 min, 0.160 mmol of alkyl bromide (or
alkyl methanesulfonate) was added dropwise. Solution was allowed
to warm to rt and stirred overnight. The reaction was quenched
with 1 mL of an aqueous saturated solution of NH4Cl. Solution
was diluted with AcOEt and washed three times with water and
brine. After anhydrification with Na2SO4, the solvent was removed
under vacuum, and the product was purified by flash chromatog-
raphy (62-75% yield).
Acknowledgment. We thank University of “Piemonte Ori-
entale”, Vercelli (Italy), Interdisciplinary Research Center of
Autoimmune Diseases (IRCAD) Novara (Italy), Compagnia di
San Paolo, Torino (Italy), and Novartis Vaccines, Siena (Italy)
for financial support, and Regione Piemonte (L.R. 4/2006 and
Ricerca Scientifica Applicata) for a research fellowship to A.B.
Supporting Information Available: General experimental
The 13C NMR analysis showed a 95:5 ratio of the E/Z isomers.
Since the double bond would be reduced in the next step, the single
isomers were not separated.
General Procedure for Alkylation of Compound 14. Under
Ar atmosphere, 0.135 mmol (100 mg) of 14 was dissolved in 1.5
mL of anhydrous DMF and cooled to 0 °C; 0.20 mmol of NaH
methods, additional experimental procedures, analytical data,
1
and copies of H and 13C NMR of all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JO8019994
J. Org. Chem. Vol. 73, No. 22, 2008 9195