Synthesis and evaluation of amino-modified α-galcer analogues

Article abstract of DOI:10.1021/ol100934z

α-GalCer analogues featuring a phytoceramide 3- and 4-amino group have been synthesized. A Mitsunobu reaction involving phthalimide was employed for the introduction of the amino groups at the 3- and 4-positions of suitable phytosphingosine-derived precursors. The influence of these modifications on the interaction with the T-cell receptor of NKT cells was investigated, as well as the capacity of the amino-modified analogues to induce a cytokine response after in vivo administration.

Full text of DOI:10.1021/ol100934z

ORGANIC
LETTERS
2010
Vol. 12, No. 13
2928-2931
Synthesis and Evaluation of
Amino-Modified r-GalCer Analogues
Matthias Trappeniers,† Rene´ Chofor,† Sandrine Aspeslagh,‡ Yali Li,|
Bruno Linclau,¶ Dirk M. Zajonc,| Dirk Elewaut,‡ and
Serge Van Calenbergh*,†
Laboratory for Medicinal Chemistry (FFW), Ghent UniVersity, Harelbekestraat 72,
9000 Ghent, Belgium, Laboratory for Molecular Immunology and Inflammation,
Department of Rheumatology, Ghent UniVersity Hospital, Ghent UniVersity, De
Pintelaan 185, 9000 Ghent, Belgium, DiVision of Cell Biology, La Jolla Institute for
Allergy and Immunology, La Jolla, California, and School of Chemistry, UniVersity of
Southampton, Highfield, Southampton SO17 1BJ, U.K.
serge.Vancalenbergh@ugent.be
Received April 23, 2010
ABSTRACT
r-GalCer analogues featuring a phytoceramide 3- and 4-amino group have been synthesized. A Mitsunobu reaction involving phthalimide was
employed for the introduction of the amino groups at the 3- and 4-positions of suitable phytosphingosine-derived precursors. The influence
of these modifications on the interaction with the T-cell receptor of NKT cells was investigated, as well as the capacity of the amino-modified
analogues to induce a cytokine response after in vivo administration.
In the past decade, glycolipids have gained increasing
interest as immune modulatory agents due to their role in
the defense against microbial infections, tumor immuno-
surveillance, and the maintenance of immune tolerance.1
In that context, the most extensively studied glycolipid is
R-GalCer (1, also known as KRN7000, Figure 1),2 of
which high specificity for CD1d, a cell surface glycopro-
tein constitutively expressed by antigen presenting cells,
is well-established. Upon recognition of the R-GalCer-
CD1d complex by their semi-invariant T-cell receptor
(TCR), natural killer T (NKT) cells are activated, resulting
in the rapid release of T helper 1 (Th1) and T helper 2
(Th2) cytokines. The concomitant release of Th1 and Th2
cytokines, which antagonize each other’s effects, is
believed to limit R-GalCer’s therapeutic outcome. Hence,
the development of analogues capable of inducing a biased
Th1 or Th2 response is highly promising.3
Crystal structures of human4 and mouse5 CD1d com-
plexed with R-GalCer or the ternary structures also
comprising the human6 and mouse7 NKT TCR reveal that
† Laboratory for Medicinal Chemistry, Ghent University.
‡ Department of Rheumatology, Ghent University.
| Division of Cell Biology, La Jolla Institute for Allergy and Immunology.
¶ School of Chemistry, University of Southampton.
(1) Reviews: (a) Wu, D.; Fujio, M.; Wong, C. H. Bioorg. Med. Chem.
2008, 16, 1073. (b) Savage, P. B.; Teyton, L.; Bendelac, A. Chem. Soc.
ReV. 2006, 35, 771.
(3) Berkers, C. R.; Ovaa, H. Trends Pharmacol. Sci. 2005, 26, 252.
(4) Koch, M.; Stronge, V. S.; Shepherd, D.; Gadola, S. D.; Mathew,
B.; Ritter, G.; Fersht, G. S.; Schmidt, R. R.; Jones, E. Y.; Cerundolo, V.
Nat. Immunol. 2005, 6, 819.
(2) (a) Morita, M.; Motoki, K.; Akimoto, K.; Natori, T.; Sakai, T.; Sawa,
E.; Yamaji, K.; Koezuka, Y.; Kobayashi, E.; Fukushima, H. J. Med. Chem.
1995, 38, 2176. (b) Natori, T.; Koezuka, Y.; Higa, T. Tetrahedron Lett.
1993, 34, 5591. (c) Natori, T.; Morita, M.; Akimoto, K.; Koezuka, Y.
Tetrahedron 1994, 50, 2771.
(5) Zajonc, D. M.; Cantu, C.; Mattner, J.; Zhou, D.; Savage, P. B.;
Bendelac, A.; Wilson, I. A.; Teyton, L. Nat. Immunol. 2005, 6, 810.
10.1021/ol100934z  2010 American Chemical Society
Published on Web 06/02/2010

the galactose ring is well ordered and extends above the
surface of a lipid-binding groove. Several hydrogen bonds
between 1 and CD1d are assumed to anchor R-GalCer in
a proper orientation for recognition by the TCR.8 Asp-80
was found to interact with both secondary hydroxy groups
(i.e., 3-OH and 4-OH) of the phytosphingosine backbone.
A key interaction of the subsequent NKT TCR binding
involves a hydrogen bond between the TCR Arg-95
residue and the same 3-OH group.
The hypothesis stating that the Th1-Th2 balance
is determined by the overall stability of the CD1d-
glycolipid-TCR complex3 is a matter of debate. Many
analogues, typically obtained by modifying the apolar regions
of 1 (Figure 1), have been synthesized in order to investigate
stability. On the other hand, the interaction with the TCR
would also be affected by such modifications since a
3-ammonium group can no longer act as hydrogen bond
acceptor and the inductive effect of the 4-ammonium
group was expected to severely weaken the hydrogen bond
acceptor capacity of the 3-OH.
Synthetic procedures toward the desired 2,3-diamino-1,4-
butanediol or 2,4-diamino-1,3-butanediol aglycon scaffolds are
scarce. Noteworthy is the recently reported enantioselective
synthesis of a O-1-protected 4-deoxy-4-azido-2-phthalimido-
D-ribo-phytosphingosine, involving a regioselective opening of
a cyclic sulfate.13 The structure of prumycin, an antibiotic
isolated in 1971 that is synthetically accessible,14 contains a
2,4-diamino-1,3-butanediol moiety with a configuration match-
ing that of the desired L-lyxo analogue 5.
However, for the synthesis of the four desired glycosyl
acceptors, it was decided to use a common strategy starting
from D-ribo-phytosphingosine, which is now readily available
from a yeast fermentation process.
The synthesis of the 3-amino-3-deoxy-R-GalCer epimers
is shown in Scheme 1. D-ribo-phytosphingosine was
converted to the two C3-epimeric 1,4-diprotected azido-
phytosphingosine derivatives 6-7.15 Unfortunately, all
attempts to effect invertive phthalimide introduction via
a Mitsunobu reaction resulted in an elimination side-
reaction to give 8, regardless of the configuration at C-3
or the O-1 protecting group (see the Supporting Informa-
tion). The E-configuration of the double bond in 8 was
confirmed by a ROESY experiment.
Given that the electron-withdrawing azido group pre-
sumably promoted the undesired elimination reaction, it
was decided to first convert the azido group to the required
long-chain amido moiety. Hence, Staudinger reduction,
followed by acylation with hexacosanoic acid and EDC,
gave the ceramides 9 and 10, which were now successfully
converted to the phthalimides via a Mitsunobu protocol
to afford the desired 3-phthalimidophytosphingosine ac-
ceptors 11 and 12 after deprotection of the trityl group
with zinc(II) dibromide.
The reduced nucleophilicity of the primary hydroxyl group
of the ceramide acceptors 11 and 12, due to intramolecular
hydrogen bond formation with the NH group of the amide,
forced us to use the Mukaiyama glycosydation16 involving the
reactive galactosyl fluoride 13 as the glycosyl donor.
Reaction of 3-phthalimido-D-ribo-phytosphingosine 11
and 3-phthalimido-D-xylo-phytosphingosine 12 with ga-
lactosyl fluoride 13 afforded the desired 3-amino-R-GalCer
analogues 2 and 3 after final deprotection operations.
Toward the synthesis of the 4-amino-4-deoxy-R-GalCer
analogues 4 and 5, the 4-phtalimido-substituted phytosph-
ingosine precursors with the natural D-ribo- (22) and L-lyxo-
Figure 1
2-5.
. Structures of KRN7000 (1) and amino-modified analogues
this relationship,9-11 but recent studies also point toward the
importance of the mode of glycolipid loading into CD1d and
the rate of glycolipid dissociation from CD1d in the
lysosomes.12
From the X-ray data, it can be concluded that R-GalCer’s
3-OH is a crucial group as it interacts both with CD1d
and the TCR. Hence, we proposed to investigate the 3-
and 4-amino-3-(4-)deoxy analogues 2-5, as it was
anticipated that ionic interactions would be established
with the CD1d Asp-80, leading to increased complex
(6) Borg, N. A.; Wun, K. S.; Kjer-Nielsen, L.; Wilce, M. C. J.; Pellicci,
D. G.; Koh, R.; Besra, G. S.; Bharadwaj, M.; Godfrey, D. I.; McCluskey,
J.; Rossjohn, J. Nature 2007, 448, 44.
(7) Pellicci, D. G.; Patel, O.; Kjer-Nielsen, L.; Pang, S. S.; Sullivan,
L. C.; Kyparissoudis, K.; Brooks, A. G.; Reid, H. H.; Gras, S.; Lucet, I. S.;
Koh, R.; Smyth, M. J.; Mallevaey, T.; Matsuda, J. L.; Gapin, L.; McCluskey,
J.; Godfrey, D. I.; Rossjohn, J. Immunity 2009, 31, 47–59.
(8) Henon, E.; Dauchez, M.; Haudrechy, A.; Banchet, A. Tetrahedron
2008, 64, 9480.
(9) (a) Miyamoto, K.; Miyake, S.; Yamamura, T. Nature 2001, 413,
531. (b) Goff, R. D.; Gao, Y.; Mattner, J.; Zhou, D.; Yin, N.; Cantu, C.;
Teyton, L.; Bendelac, A.; Savage, P. B. J. Am. Chem. Soc. 2004, 126, 13602
.
(10) Yu, K. O. A.; Im, J. S.; Molano, A.; Dutronc, Y.; Illarionov, P. A.;
Forestier, C.; Fujiwara, N.; Arias, I.; Miyake, S.; Yamamura, T.; Chang,
Y. T.; Besra, G. S.; Porcelli, S. A. Proc. Natl. Acad. Sci. U.S.A. 2005, 102,
3383
.
(11) (a) Liang, P. H.; Imamura, M.; Li, X.; Wu, D.; Fujio, M.; Guy,
R. T.; Wu, B. C.; Tsuji, M.; Wong, C. H. J. Am. Chem. Soc. 2008, 130,
12348. (b) 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, 104,
10299
.
(13) Llaveria, J.; Diaz, Y.; Matheu, M. I.; Castillon, S. Org. Lett. 2008,
11, 205.
(12) (a) Im, J. S.; Arora, P.; Bricard, G.; Molano, A.; Venkataswamy,
M. M.; Baine, I.; Jerud, E. S.; Goldberg, M. F.; Baena, A.; Yu, K. O. A.;
Ndonye, R. M.; Howell, A. R.; Yuan, W.; Cresswell, P.; Chang, Y.-t.;
Illarionov, P. A.; Besra, G. S.; Porcelli, S. A. Immunity 2009, 30, 888–898.
(b) Bai, L.; Sagiv, Y.; Freigang, S.; Yu, K. O. A.; Teyton, L.; Porcelli, S.;
Savage, P. B.; Bendelac, A. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10254–
10259.
(14) Hasegawa, A.; Aritake, N.; Kiso, M. Carbohydr. Res. 1976, 51,
C10.
(15) Trappeniers, M.; Goormans, S.; Van Beneden, K.; Decruy, T.;
Linclau, B.; Al-Shamkhani, A.; Elliott, T.; Ottensmeier, C.; Werner, J. M.;
Elewaut, D.; Van Calenbergh, S. ChemMedChem 2008, 3, 1061.
(16) Mukaiyama, T.; Murai, Y.; Shoda, S. Chem. Lett. 1981, 431.
Org. Lett., Vol. 12, No. 13, 2010
2929

Scheme 1. Synthesis of the 3-Amino-3-deoxy-R-GalCer Analogues 2 and 3
configurations (23) were first prepared in five steps from the
advanced intermediates 16 and 17 (Scheme 2), again
accessible from D-ribo-phytosphingosine.15
without notable formation of the ꢀ-glycoside (Scheme 3).
Interestingly, Staudinger reduction of the glycoside 25 led
to the intramolecular formation of the amidine 27,
Scheme 2
.
Synthesis of the D-Ribo and L-Lyxo Glycosidation
Acceptors 22 and 23
Scheme 3
.
Synthesis of the 4-Amino-4-deoxy-R-GalCer
Analogues 4 and 5
The more nucleophilic 4-phthalimidophytosphingosine
acceptors could now easily be glycosylated with the less
reactive trichloroacetimidate donor 24.17 TMSOTf-
promoted galactosidation of the 4-deoxy-4-phthalimido-
D-ribo-phytosphingosine 22 afforded the R-glycoside
(17) (a) Plettenburg, O.; Bodmer-Narkevitch, V.; Wong, C. H. J. Org.
Chem. 2002, 67, 4559. (b) Koeman, F. A. W.; Kamerling, J. P.; Vliegenthart,
J. F. G. Tetrahedron 1993, 49, 5291. (c) Figueroa-Pere´z, S.; Schmidt, R. R.
Carbohydr. Res. 2000, 328, 95.
2930
Org. Lett., Vol. 12, No. 13, 2010

presumably via the aza-Wittig mechanism.18 To circum-
vent this side reaction, it was decided to perform a
protecting group switch to give the corresponding diben-
zylamino group (28), which allowed uncomplicated
Staudinger reduction to give the 2-NH2 group. Subsequent
acylation with hexacosanoic acid and final deprotection
afforded the desired 4-amino-4-deoxy-R-galactosyl-D-ribo-
phytoceramide 4.
Similarly, glycosidation of the 4-deoxy-4-phthalimido-L-
lyxo-phytosphingosine 23 afforded the R-glycoside 26 in 81%
yield. In this case, Staudinger reduction led to the desired
free amine and a relatively small amount of the amidine,
which could be efficiently removed after acylation of the
former. Debenzylation, followed by final deprotection of the
phthalimide with methylamine, gave the desired 4-amino-
4-deoxy-R-galactosyl-L-lyxo-phytoceramide 5.
complex of mCD1d plus 1, 2, and 4. Compared to R-GalCer
(KD ) 12.8 nM), the TCR showed a 13-fold reduced binding
affinity to CD1d loaded with the 3-amino analogue 2 (KD )
163 nM) and a 125-fold drop in affinity for the 4-CD1d
complex (KD ) 1.7 µM). While the association of the TCR
is 2× slower for analogue 4 (ka ) 2.21 × 104 1/Ms) and
6.5× slower for 2 (ka ) 6.1 × 104 1/Ms), the TCR dissociates
25× faster from the amino analog 4 (kd ) 0.037 × 1/s) and
7.5× faster for 2 (0.01 × 1/s). Therefore, the 3- and 4-amino
modification significantly affect the overall stability of the
TCR interaction.
In summary, we showed that, despite its reasonable affinity
for the TCR when complexed with CD1d, analogue 2 shows
a very small cytokine response, while an opposite trend is
observed for analogue 4. These findings conflict with current
hypotheses, which, however have been inferred mainly from
studies involving analogues with altered aliphatic chains.
While such alterations are hidden from the TCR contact
surface, this is different for the amino modifications in this
study. Indeed, as highlighted by the crystal structure of the
ternary mCD1d-R-GalCer-TCR,7 the 3-OH group is not only
interacting with CD1d but also exposed to the TCR to form
a hydrogen bond with Arg-95 (of the CDR3R unit). Possibly,
when protonated, the 3-amino group of 2 interacts unfavor-
ably with Arg-95, thereby leading to suboptimal activation
of the TCR.
To compare the cytokine profile induced by the amino-
modified analogues with R-GalCer, serum cytokine levels
were measured after IP injection of glycolipids 1-5 into
C57Bl/6 mice (Figure 2). Following injection of the 3-amino
In conclusion, the present amine-modified analogues
indicate that, although TCR affinity probably accounts to
some degree for the quality of responses mediated by iNKT
cells, subtle changes to R-GalCer may afford analogues with
diminished capacity to provoke activation of the iNKT cells,
despite good affinity for the TCR.
Figure 2. INF-γ and IL-4 secretion after intraperitoneal injection
Acknowledgment. M.T. and S.A. are aspirants of the Fund
for Scientific Research-Flanders (F.W.O.-Vlaanderen). Fi-
nancial support by F.W.O., Stichting tegen Kanker, and
Cancer Research Technologies is gratefully acknowledged.
D.M.Z. is the recipient of an investigator award from the
Cancer Research Institute and supported by NIH Grant No.
AI074952.
of R-GalCer (1) and 2-5 in mice (*P < 0.01 vs R-GalCer).
epimers 2 and 3, low levels of IL-4 were detected, and also
IFN-γ levels were drastically lower compared to R-GalCer.
On the other hand, a trend toward a more moderate reduction
of IFN-γ was observed after stimulation with the 4-amino
epimers 4 and 5.
Surface plasmon resonance was used to measure the
binding kinetics of a soluble mouse VR14i TCR for the
Supporting Information Available: Experimental pro-
cedures for the preparation of 2-5 for the in vivo stimulation
with R-GalCer-analogues and the SPR studies. This material
is available free of charge via the Internet at http://pubs.acs.org.
(18) Molina, P.; Vilaplana, M. J. Synthesis 1994, 1197.
OL100934Z
Org. Lett., Vol. 12, No. 13, 2010
2931