Synthesis and evaluation of human T cell stimulating activity of an α-sulfatide analogue

Article abstract of DOI:10.1016/j.bmc.2007.05.044

A concise synthesis of α-sulfatide 1, an analogue of natural glycolipid antigens with potential anti-tumor activity, was performed. Two different approaches to the α-glycosidic bond were explored, resulting in a high yield and excellent stereoselectivity.

Full text of DOI:10.1016/j.bmc.2007.05.044

Bioorganic & Medicinal Chemistry 15 (2007) 5529–5536
Synthesis and evaluation of human T cell stimulating activity
of an a-sulfatide analogue
Laura Franchini,a Pamela Matto,a Fiamma Ronchetti,a,* Luigi Panza,b Lucia Barbieri,c
Valeria Costantino,c Alfonso Mangoni,c Marco Cavallari,d Lucia Morid
and Gennaro De Liberod
a
`
Dipartimento di Chimica, Biochimica e Biotecnologie per la Medicina, Universita di Milano, Via Saldini 50, 20133 Milano, Italy
b
`
Dipartimento di Scienze Chimiche, Alimentari, Farmaceutiche e Farmacologiche, Universita del Piemonte Orientale,
Via Bovio 6, 28100 Novara, Italy
c
`
Dipartimento di Chimica delle Sostanze Naturali, Universita di Napoli ‘‘Federico II’’, Via D. Montesano 49, 80131 Napoli, Italy
dExperimental Immunology, Department of Research, Basel University Hospital, Hebelstrasse 20, 4031 Basel, Switzerland
Received 16 January 2007; revised 11 May 2007; accepted 18 May 2007
Available online 23 May 2007
Abstract—A concise synthesis of a-sulfatide 1, an analogue of natural glycolipid antigens with potential anti-tumor activity, was
performed. Two different approaches to the a-glycosidic bond were explored, resulting in a high yield and excellent stereoselectivity.
Compound 1 combines the structural features of sulfated b-GalCer (sulfatide) and a-GalCer, which activate specific T cells. a-Sul-
fatide 1 was stimulatory for CD1d-restricted semi-invariant Natural Killer T (iNKT) cell clones, although less potent than a-Gal-
Cer, while it was not recognized by CD1a-restricted sulfatide-specific T cells.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
mide with an a-anomeric bond (a-GalCer) and when
presented by CD1d it stimulates very efficiently T cells
expressing a semi-invariant T cell receptor (TCR).6
These T cells also express surface markers common to
Natural Killer cells and therefore are indicated as invari-
ant NKT (iNKT) cells.
CD1 proteins constitute a family of highly conserved
antigen presenting molecules (APM) which bind and
present lipid and glycolipid molecules to T lympho-
cytes.1 Diverse immunogenic lipids and glycolipids have
been isolated and characterized from mammalian and
bacterial sources.2
The synthetic a-GalCer analogue KRN7000 is highly
immunostimulatory for iNKT cells and has strong
anti-tumor activities.7 Its potential in the treatment of
several diseases such as tumors, infectious diseases like
malaria and hepatitis B, and autoimmune diseases like
diabetes is currently being investigated.8,9
Many evidences confirm the immunological relevance of
CD1-restricted T cells in bacterial infections, tumor
immunosurveillance, and autoimmune response.3 In
particular T cells specific for self-glycosphingolipids
are increased in the blood of individuals with multiple
sclerosis; moreover, synthesis and presentation of self-
glycosphingolipids is enhanced by bacterial infections.4
Upon activation by a-GalCer, iNKT cells release
proinflammatory (Th1) and immunomodulatory (Th2)
cytokines and participate in immunoregulation, in
anti-tumor immune response, and in autoimmune dis-
eases.10 Many a-GalCer analogues have been designed
and prepared with the aim to determine structure–activ-
ity relationship and to enhance bioactivity; in this
context, sphingosine 3-OH and galactose 2-OH have
been found essential for bioactivity,7,11–13 whereas the
length and structure of the lipid chains might influence
the type of released cytokines.14–16
A highly stimulatory glycolipid is agelasphin, initially
isolated from the marine sponge Agelas mauritiana.5
Agelasphin is composed of a galactose bound to cera-
Keywords: Galactosylceramides; Glycosylation; NKT cell activation;
Immunostimulation.
*
Corresponding author. Tel.: +39 02 5031 6037; fax: +39 02 5031
6036; e-mail: fiamma.ronchetti@unimi.it
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.05.044

5530
L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
Sulfatide, a b-D-galactosylceramide sulfated at position
3 of galactose, is a potent self-glycolipid antigen pro-
duced by mammalian cells.17 Sulfatide is abundant in
myelin and in serum and is synthesized in many different
cell types. Moreover, it is highly immunogenic, as it
stimulates both specific T and B cells. Sulfatide is the
only known promiscuous ligand binding to all human
CD1 proteins and activates specific T cells.17 Like
a-GalCer-reacting iNKT cells, also sulfatide-specific T
cells react in different manner to sulfatides with varia-
tions in alkyl chain length.18,19
In this paper, the synthesis of 1 starting from 3-O-benzo-
yl-D-erythro azidosphingosine22 as well as the biological
tests on this analogue are described.
2. Results and discussion
2.1. Chemistry
The synthetic strategy for the preparation of a-sulfatide
1 involves the assembly of a galactose unit and a pro-
tected azido precursor of sphingosine, followed by azide
reduction, N-acylation with fatty acid, and final selective
sulfation of 3-OH of galactose.
The mode of sulfatide binding to CD1 molecules has
been elucidated by 3D structures of CD1a20 and
CD1d19 in complex with sulfatide: two alkyl chains are
inserted into hydrophobic pockets of CD1 projecting
the 3-O-sulfogalactose moiety out of the binding groove
on the CD1 surface which contacts the T cell receptor.
Structural determinations of CD1d-sulfatide19 and
CD1d-a-GalCer21 complexes allow to compare the
binding mode of the two ligands on CD1d: the galactose
moiety of sulfatide, due to b-anomeric linkage, is more
exposed and projects out of the ligand binding groove
compared to the a-linked galactose of a-GalCer.
We successfully experienced (2S,3R,4E)-3-O-benzoyl
azidosphingosine 2 as glycosyl acceptor for the prepara-
tion of b-galactosylceramides22,23 and a very efficient
protocol for the preparation of this synthon starting
from D-glyceraldehyde was reported by us.22
So compound 2 was chosen as the lipid acceptor for the
crucial glycosylation leading to a-sulfatide; due to the
experience acquired in Mukaiyama glycosylation,11–13
which has been applied for the preparation of different
2-O-analogues of a-GalCer, we first focused our atten-
tion on this procedure. The Mukaiyama glycosylation
reaction involves a glycosyl fluoride as the glycosyl
donor, and AgClO4 and SnCl2 as a Lewis acid catalytic
system. Therefore, 2,3,4,6-tetra-O-benzyl-galactopyr-
anosyl fluoride (3b) was prepared by reaction of tetrab-
enzylgalactose (3a) (Scheme 1) with DAST (Et2NSF3), a
well-known reagent for mild and direct transformation
of glycosyl hemiacetals into glycosyl fluorides.24 The
glycosyl fluoride was obtained in 80% yield. Following
the protocol for the Mukaiyama glycosylation, the azi-
dosphingosine acceptor 2 and the glycosyl fluoride
donor (3 mol equiv) dissolved in dry THF were added
to a flask containing the Lewis acid system, previously
prepared suspending AgClO4 (3 mol equiv) and SnCl4
(3 mol equiv) in dry Et2O. The mixture was allowed to
react for 12 h at room temperature; after work-up, the
isolation of the reaction product was performed by pre-
parative HPLC. The a-glycoside 4 was obtained with
complete a stereoselectivity, but only in a moderate
35% yield. In addition, unlike that in previous work,
we were not able to recover any unreacted acceptor 2.
With the aim to find more efficient bioactive com-
pounds, the preparation of analogues of the natural gly-
colipid antigens, combining the structural features of
sulfated b-GalCer (sulfatide) and a-GalCer, was
planned. From a chemical point of view sulfatide and
a-GalCer differ in several respects: in the anomeric link-
age, in the presence of a E configured double bond
between C4 and C5, instead of a hydroxyl at C4 of the
sphingoid base (sphingosine in sulfatide versus phyto-
sphingosine in a-GalCer), and the presence of sulfate
at position 3 of galactose.
Considering the structural differences between sulfatide
and a-GalCer, compound 1 (Fig. 1) containing 3-O-sul-
fogalactose a-linked to a sphingosine-derived ceramide
is a hybrid glycolipid analogue, whose immunostimula-
tory activity was studied in the context of CD1a and
CD1d mediated T cell activation.
Biological tests on this analogue should allow to evalu-
ate the influence of these structural modifications on its
immunostimulatory potential.
With the aim to gain better yields, the Lemieux halide-
ion catalytic a-glycosylation as modified by Kobayashi
was explored, which potentially provides a straightfor-
ward and simple access to 1,2-cis glycosides in high
yields and a-selectivity.25,26 According to this procedure,
the glycosyl bromide donor is generated in situ through
the use of Appel agents (Ph3P and CBr4), and can then
perform a halide-ion catalytic a-glycosylation with a
large array of different acceptor alcohols, affording
a-glycosides almost quantitatively.25
O
OH
OH
OH
O
OH
R
O
HN
C25H51
OH
O
HO
NaO3SO
O
C13H27
HN
OH
O
OH
OH
C14H29
OH
α-GalCer
sulfatide
OH
OH
O
O
NaO3SO
C21H43
C13H27
HN
OH
O
OH
1 α-sulfatide
We applied this protocol to azidosphingosine acceptor
2, conducting glycosylation as follows (Scheme 1): tet-
rabenzylgalactose 3a was treated with Ph3P (3 mol
equiv) and CBr4 (3 mol equiv) for 3 h at room tempera-
Figure 1. Structures of sulfatide, a-galactosylceramide, and a-sulfa-
tide 1.

L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
5531
BnO
BnO
BnO
OBn
O
OBn
O
N3
b or c
C13H27
HO
a
BnO
+
X
N3
OBn
BnO
OBz
O
C13H27
OBz
3a: X = OH
3b: X = F
2
4
BnO
OR
OBn
O
OR
O
O
O
d
e
C21H43
C13H27
C21H43
C13H27
BnO
RO
HN
HN
BnO
RO
O
O
OH
OR
6: R = Ac
7: R = H
5
f
OH
OH
O
O
g
C21H43
C13H27
NaO3SO
HN
HO
O
OH
1
Scheme 1. Reagents and conditions: (a) 3a, DAST, THF, rt, 20 min, 80%; (b) 2 and 3b, SnCl2/AgClO4, Et2O, ꢀ15 ꢁC, 12 h, 35%; (c) 3a, Ph3P, CBr4,
˚
then Bu4NBr, 2 (CH2Cl2 solution), N,N-tetramethylurea, 4 A ms, CH2Cl2, rt, 5 days, 92%; (d) (i) MeONa, MeOH, rt, 2 h; (ii) H2S, pyridine, water,
rt, 48 h; (iii) C21H43COCl, 50% aq AcONa, THF, rt, 3 h, 55% (three steps); (e) (i) Na, liquid ammonia, 5 (THF solution), ꢀ50 ꢁC, 2 h; (ii) Ac2O, Py,
rt, 20 h, 78% (two steps); (f) MeONa, MeOH, 3 h, (quant); (g) Bu2SnO, MeOH, reflux, 2 h; Me3NÆSO3, THF, rt, 2 h, 70%.
ture to produce the anomeric bromide, which was
one-pot reacted with 2 (2 mol equiv) in the presence of
tetrabutylammonium bromide (3 mol equiv), N,N-
tetramethylurea, and molecular sieves. The reaction pro-
ceeded smoothly and the bromide donor (as monitored
by TLC) was completely consumed after five days. In
spite of an excess amount of Ph3P in the reaction system,
no azide reduction was observed. This result indicates
that this a-glycosylation procedure can be successfully
applied to 3-O-benzoylazidosphingosine 2 further than
the model acceptor 6-azidohexanol,24 so as to have
access to a-linked glycosphingolipids.
With amide 5 in hand, removal of benzyl protecting
groups was easily achieved with sodium in liquid ammo-
nia at ꢀ50 ꢁC; the reaction was complete within 2 h and
the crude was acetylated to facilitate chromatographic
purification; the fully acetylated derivative 6 was ob-
tained in satisfactory 78% yield.
Removal of acetyl groups on compound 6 afforded
a-GalCer 7 and finally the sulfate was introduced at
position 3 of galactose according to Flitsch et al.27 to
give the target compound 1 in 70% yield. Selective func-
tionalization at position 3 of the 1,2-cis glycoside 7 was
successfully accomplished through reaction of dibutylst-
annilene acetal with SO3ÆMe3N, as indicated by the
downfield shift of the H-30 signal18,22,23 in the 1H
NMR spectrum of a-sulfatide 1. In this way, tedious
manipulations of protecting groups on galactose
donor28 were avoided.
The stereoselectivity of the glycosylation reaction (a/b
ratio: 95:5) was established by integration of the H-1
doublets for the a-anomer (d 4.87, J = 3.5 Hz) and for
1
the b-anomer (d 4.37, J = 7.7 Hz) in the H NMR spec-
trum of the crude mixture. The a-glycoside 4 was recov-
ered pure in 92% yield (based on 3a) after work-up and
purification by silica gel column chromatography; in
addition, practically all the excess acceptor 2 was recov-
ered from the reaction mixture.
2.2. Biological evaluation
The biological activity of a-sulfatide 1 was tested by using
CD1-restricted T cell antigen presentation assays in com-
parison with synthetic sulfatide or a-GalCer. We used dif-
ferent types of human T cells. The clone K34B9.1, which
was obtained by stimulation with naturally occurring
b-sulfatide,3 is restricted by CD1a, and recognizes also
synthetic analogues of b-sulfatide like a C-analogue29
and a 4-O-sulfated analogue.23 The iNKT cell clones
BGA01 and BGA89 which recognize a-GalCer presented
by CD1d, were recently obtained in our laboratories.
Removal of benzyl ethers, benzoyl ester and azide reduc-
tion could have been performed simultaneously by treat-
ment with sodium and liquid ammonia without affecting
the olefinic moiety of sphingosine; but when glycoside 4
was subjected to these conditions a complex mixture of
products was formed, from which it was not possible to
isolate the deprotected amine.
Therefore, the procedure was conducted stepwise. To
avoid acyl shift on amino group during azide reduction
a-Sulfatide 1 is unable to stimulate CD1a-restricted and
sulfatide-specific T cells when presented by CD1a
(Fig. 2), probably due to the alpha configuration of
the sulfated galactose moiety. Both interferon gamma
(IFNc) and interleukin-4 (IL-4) cytokines were not re-
leased after stimulation with this compound, whereas
´
benzoyl ester was first removed by Zemplen transesteri-
fication. After that, azide reduction and N-acylation
with docosanoyl chloride were carried out as previously
described18 to give a-GalCer 5 in 55% yield over three
steps.

5532
L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
Figure 2. C1R cells expressing human CD1a were preincubated with cis-tetracosenoyl (C24:1) sulfatide18 (d) or with a-sulfatide 1 (s) at the doses
indicated, before addition of the sulfatide-specific CD1a-restricted T cell clone K34B9.1. Released IFN-c (a) and IL-4 (b) were measured by ELISA.
Data are expressed as mean pg/ml SD of duplicates.
synthetic18 but naturally occurring b-sulfatide was very
active. Hence the b configuration is an important requi-
site of sulfatide to become recognized by CD1a-re-
stricted T cells.
Cer (Fig. 3). Using CD1d-restricted iNKT cell clones, a
dose–response shift was observed. a-Sulfatide 1 induced
the same half-maximal stimulation at doses about 1000
times larger than a-GalCer. This was observed with two
iNKT cell clones, and was detected by measuring both
IFNc and IL-4 cytokines. These findings indicate that
a-sulfatide 1 has the capacity to bind to CD1d and
a-Sulfatide 1 is instead stimulatory for CD1d-restricted
iNKT cell clones, although it is less potent than a-Gal-
Figure 3. C1R cells expressing human CD1d were preincubated with a-GalCer (d) or with a-sulfatide 1 (s) at the doses indicated, before addition of
the CD1d-restricted iNKT cell clone BGA01 (a and b) or BGA89 (c and d). Released IFN-c (a and c) and IL-4 (b and d) were measured by ELISA.
Data are expressed as mean pg/ml SD of duplicates.

L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
5533
˚
stimulate iNKT cells. However, the presence of the sul-
fate group and the absence of phytosphingosine induce a
striking reduction of the stimulatory capacity, although
for the synthetic 3-O-sulfo-a-GalCer analogue28 the
presence of the sulfate itself does not seem to alter
NKT cell activation. According to the recent crystal
structure of the complex formed by CD1d and a-Gal-
Cer21 the sulfate group most likely hinders the appropri-
ate contact with the TCR. Moreover, an important role
of the phytosphingosine in stabilizing CD1d binding or
in ligand solubilization cannot be excluded.
and methanol were dried on 4 A molecular sieves. All
reagents were purchased from Aldrich. Docosanoyl chlo-
ride was prepared from docosanoic acid following a
literature procedure.30 The elemental analyses were
consistent with the theoretical ones.
4.1.1. 2,3,4,6-Tetra-O-benzyl-D-galactopyranosyl fluoride
(3b). Tetrabenzylgalactose 3a (0.82 g, 1.52 mmol) was
dissolved in dry THF (15 mL) and cooled at ꢀ30 ꢁC.
DAST (0.24 mL, 1.8 mmol) was rapidly added and the
cooling bath removed immediately. After stirring for
30 min at room temperature, the mixture was cooled
again at ꢀ30 ꢁC and quenched with MeOH (2 mL).
After evaporation of solvent under reduced pressure,
water (200 mL) was added to the residue, that was
extracted with CH2Cl2 (3· 300 mL). The combined
organic layers were dried over Na2SO4 and concentrated
under reduced pressure. The residue was purified by col-
umn chromatography on SiO2 (CH2Cl2) giving 3b
(0.58 g, 70% yield) as mixture of anomers, a/b about
3. Conclusions
In summary, this study has allowed to prepare a new
interesting hybrid glycolipid analogue through a concise
synthesis resulting in satisfactory yield and excellent
stereoselectivity.
1
In the crucial a-galactosylation step performed through
the Lemieux halide-ion catalytic procedure as modified
by Kobayashi, the desired a-galactosylceramide deriva-
tive 4 was obtained in 92% yield.
1:1, colorless oil; H NMR (CDCl3), a anomer: d 3.54
(d, 2H, J = 6.5 Hz, H2-6), 3.94 (dd, 1H, J3,2 = 10.0 Hz,
J3,4 = 2.5 Hz, H-3), 4.02 (br d, 1H, J4,3 = 2.5 Hz, H-4),
4.03 (ddd, 1H, J2,F = 25.1 Hz, J2,3 = 10.0 Hz,
J1,2 = 2.5 Hz, H-2), 4.10 (br t, 1H, J = 6.5 Hz, H-5),
4.41 (d, 1H, J = 11.9 Hz, CHPh), 4.47 (d, 1H,
J = 11.9 Hz, CHPh), 4.57 (d, 1H, J = 11.7 Hz, CHPh),
4.72 (d, 1H, J = 11.8 Hz, CHPh), 4.75 (d, 1H,
J = 11.4 Hz, CHPh), 4.81 (d, 1H, J = 11.4 Hz, CHPh),
4.84 (d, 1H, J = 11.8 Hz, CHPh), 4.93 (d, 1H,
J = 11.7 Hz, CHPh), 5.59 (1H, dd, J1,F = 53.8 Hz,
J1,2 = 2.5 Hz, H-1), 7.39–7.24 (m, 20H, Ph); b anomer:
d 3.53 (1H, dd, J3,2 = 9.7 Hz, J3,4 = 2.3 Hz, H-3), 3.66–
3.58 (3H, m, H-5, H2-6), 3.90 (1H, br d, J4,3 = 2.3 Hz,
H-4), 3.93 (1H, ddd, J2,F = 13.3, J2,3 = 9.7 Hz,
J2,1 = 7.1 Hz, H-2), 4.40 (d, 1H, J = 11.7 Hz, CHPh),
4.46 (d, 1H, J = 11.7 Hz, CHPh), 4.58 (d, 1H,
J = 11.5 Hz, CHPh), 4.69 (d, 1H, J = 12.1 Hz, CHPh),
4.73 (d, 1H, J = 12.1 Hz, CHPh), 4.75 (d, 1H,
J = 11.0 Hz, CHPh), 4.83 (d, 1H, J = 11.0 Hz, CHPh),
4.92 (d, 1H, J = 11.5 Hz, CHPh), 5.16 (dd, 1H,
J1,F = 53.2 Hz, J1,2 = 7.1 Hz, H-1), 7.37–7.23 (m, 20H,
Ph). 13C NMR (CDCl3), a anomer: d 68.2, 71.7, 73.1,
73.4, 73.6, 74.3, 74.8, 75.7 (CH, d, JC–F = 24 Hz, C-2),
78.4, 106.0 (CH, d, JC–F = 232 Hz, C-1), 127.5–128.5,
138.0–137.5; b anomer: d 68.3, 73.1, 73.2, 73.5, 73.6,
74.6, 75.0, 79.1 (CH, d, JC–F = 25 Hz, C-2), 81.0, 110.1
(CH, d, JC–F = 227 Hz, C-1), 127.5–128.5, 138.0–137.5.
ESI-MS (positive ion mode): m/z 565 [M+Na]+.
The immunostimulatory activity of a-sulfatide 1 was
studied in the context of CD1a and CD1d mediated T
cell activation, indicating that T cells display a very spe-
cific recognition of the CD1-antigen complexes, when
either CD1a or CD1d are the restriction elements. These
findings show the fine discriminatory capacity of the
TCR interacting with glycolipid antigens.
4. Experimental
4.1. General methods
Optical rotations were measured with a 241 Perkin-El-
mer polarimeter at 20 ꢁC. All NMR spectra were
recorded at 303 K with a Bruker FT NMR Avance
DRX500 spectrometer in CDCl3 or CDCl3/CD3OD
solutions with TMS as internal standard; chemical shifts
are reported as d (ppm) relative to CHCl3 fixed at
7.24 ppm for CDCl3 and 7.60 ppm for CDCl3/CD3OD
1
solutions for H NMR spectra and relative to CDCl3
fixed at 77.00 ppm for 13C NMR spectra.
MS spectra were recorded in the negative or positive ion-
modes on a Thermo Quest Finnigan LCQTMDECA ion
trap mass spectrometer; the mass spectrometer was
equipped with a Finnigan ESI interface; data were pro-
cessed by Finnigan Xcalibur software system. All reac-
tions were monitored by TLC on Silica Gel 60 F-254
plates (Merck), spots being developed with 5% sulfuric
acid in methanol/water (1:1), or with phosphomolybdate
based reagent. Flash column chromatography was per-
formed on Silica Gel 60 (230–400 mesh, Merck). Organic
solutions were dried over sodium sulfate. All evapora-
tions were carried out under reduced pressure at 40 ꢁC.
Dry solvents and liquid reagents were distilled prior to
use: THF was distilled from sodium; dichloromethane
and pyridine were distilled from calcium hydride; DMF
The 1H and 13C NMR data of galactopyranosyl fluoride
3b were in agreement with those reported in the litera-
ture by Nicolaou et al.31 and Graziani et al.32 who ob-
tained the a- or b-fluoride through different synthetic
approaches.
4.1.2. (2S,3R,4E)-2-Azido-3-benzoyloxy-1-(2,3,4,6-tetra-
O-benzyl-a-D-galactopyranosyloxy)-4-octadecene (4). (a)
A solution of the azidosphingosine 2 (0.06 g, 0.15 mmol)
and the fluoride 3b (0.24 g, 0.45 mmol) in dry THF
(4 mL) was added at ꢀ15 ꢁC to a flask containing solid
AgClO4 (0.09 g, 0.45 mmol) and solid SnCl2 (0.08 g,
0.45 mmol). After stirring for 12 h at room temperature,

5534
L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
the mixture was filtered through Celite and evaporated.
The reaction mixture was diluted with CH2Cl2 (30 mL),
washed with a saturated NaHCO3 solution (20 mL),
dried with Na2SO4, and taken to dryness. The residue
was chromatographed by HPLC (n-hexane/EtOAc 9:1)
giving 4 (0.05 g, 35% yield) as a colorless oil.
and subjected to azeotropic distillation with toluene.
The crude amine was dissolved in THF (20 mL) and
treated with aqueous sodium acetate (50%, 18 mL) and
freshly prepared docosanoyl chloride (0.16 g,
0.44 mmol) under vigorous stirring at room temperature
for 4 h. The aqueous phase was extracted with THF (2·
40 mL). The combined organic layers were dried and
concentrated. The residue was purified by flash chroma-
tography (n-hexane/EtOAc 8:2 containing 0.1% Et3N)
affording compound 5 (0.25 g, 55% yield from 4) as a
foam.
(b) 2,3,4,6-Tetra-O-benzyl-D-galactose 3a (0.30 g,
0.56 mmol) in dry CH2Cl2 (12 mL) was treated with
Ph3P (0.44 g, 1.67 mmol) and CBr4 (0.55 g, 1.67 mmol)
for 3 h at room temperature under argon. Then N,N-
tetramethylurea (0.6 mL), Bu4NBr (0.54 g, 1.67 mmol),
20
D
1
˚
molecular sieves (MS, 4 A, 0.6 g), and 3-O-benzoyl azi-
½aꢁ þ 24:8 (c 1, CHCl3). H NMR (CDCl3): d 0.90 (t,
dosphingosine 2 (0.48 g, 1.11 mmol) in dry CH2Cl2
(6 mL) were added and stirred at room temperature
for five days; at this time the bromide donor was com-
pletely consumed as evidenced by TLC analysis. Trieth-
ylamine (3 mL) was added and the mixture was diluted
with CH2Cl2 (250 mL), filtered on Celite pad, washed
with saturated aq NaHCO3 (250 mL) and NaCl solution
(250 mL), dried, and concentrated. The residue was
purified by flash chromatography (petroleum ether/
EtOAc 9:1) affording first compound 4 (0.49 g, 92%
yield based on the amount of 3a) as a colorless oil, then
unreacted 2 (0.25 g) as a clear oil.
6H, J = 7.5 Hz, 2 CH3), 1.20–1.40 (m, 58H, 29 CH2),
1.57–1.65 (m, 2H, COCH2CH2), 2.02 (m, 2H,
CH@CHCH2), 2.15 (m, 2H, COCH2), 3.48–3.57 (m,
2H, H2-60), 3.72 (dd, 1H, J1a,1b = 10.5 Hz,
J1a,2 = 4.0 Hz, H-1a), 3.83–3.92 (m, 3H, H-1b, H-30,
H-50), 3.98–4.04 (m, 2H, H-40, H-2), 4.06 (dd, 1H,
J2 ;3 ¼ 10:0 Hz, J2 ;1 ¼ 3:5 Hz, H-20), 4.16 (m, 1H, H-
3), 4.40 (d, 1H, J = 12.0 Hz, CHPh), 4.50 (d, 1H,
J = 12.0 Hz, CHPh), 4.58 (d, 1H, J = 11.5 Hz, CHPh),
4.73 (d, 1H, J = 11.5 Hz, CHPh), 4.75–4.82 (m, 2H,
0
0
0
0
CHPh), 4.77 (d, 1H, J1 ;2 ¼ 3:5 Hz, H-10), 4.88 (d, 1H,
0
0
J = 12.0 Hz, CHPh), 4.94 (d, 1H, J = 11.5 Hz, CHPh),
J4,3 = 5.5 Hz,
5.45
(ddt,
1H,
J4,5 = 15.5 Hz,
20
D
½aꢁ þ 11 (CHCl3, c = 0.2); 1H NMR (CDCl3): d 0.88 (t,
J4,6 = 1.7 Hz, H-4), 5.75 (dt, 1H, J5,6 = 15.5 Hz,
J5,6 = 7.5 Hz), 6.43 (d, 1H, JNH,2 = 8.0 Hz, NH), 7.20–
7.60 (m, 20H, Ph). 13C NMR (CDCl3): d 14.1 (2C),
22.7 (2C), 25.8, 29.3–30.0 (24C), 31.9 (3C), 32.4, 36.7,
52.8, 68.7, 69.0, 69.8, 72.7, 73.6, 74.0, 74.1, 74.4, 74.8,
75.8, 79.2, 99.1, 125.4–128.5 (Ph), 129.2, 132.9, 137.6,
138.0; 138.4, 138.5, 173.3. ESI-MS (negative-ion mode):
m/z 1142.5 [MꢀH]ꢀ.
3H, J = 6.8 Hz, CH3), 1.30–1.40 (m, 22H, 11 CH2), 2.02
(q, 2H, J = 7.0 Hz, CH@CHCH2), 3.55–3.45 (m, 3H, H-
1a, H2-60), 3.74 (dd, 1H, J1a,1b = 10.9, J1a,2 = 4.4 Hz, H-
1a), 3.98–3.93 (m, 3H, H-50, H-40, H-30), 4.00 (ddd, 1H,
0
0
J = 7.7, 4.4, 4.4 Hz, H-2), 4.05 (dd, 1H, J2 ;3 ¼ 9:5 Hz,
J2 ;1 ¼ 3:6 Hz, H-20), 4.39 (d, 1H, J = 11.9 Hz, CHPh),
4.46 (d, 1H, J = 11.9 Hz, CHPh), 4.56 (d, 1H,
J = 11.4 Hz, CHPh), 4.94 (d, 1H, J = 11.4 Hz, CHPh),
4.69 (d, 1H, J = 12.0 Hz, CHPh), 4.80 (d, 1H,
J = 12.0 Hz, CHPh), 4.74 (d, 1H, J = 11.8 Hz, CHPh),
4.85 (d, 1H, J = 11.8 Hz, CHPh), 4.86 (d, 1H,
0
0
4.1.4. (2S,3R,4E)-3-Acetoxy-1-(2,3,4,6-tetra-O-acetyl-a-
D-galactopyranosyloxy)-2-(docosanoylamino)-4-octade-
cene (6). Sodium (0.06 g) was added to cold (ꢀ50 ꢁC),
stirred liquid ammonia (12 mL) and after a few minutes
deep blue solution was obtained, then compound 5
(0.16 g, 0.14 mmol) in anhydrous THF (2.4 mL) was
added via cannula at such a rate that the solution was
not discolored. The solution was stirred at ꢀ50 ꢁC for
2 h while the blue color persisted; after that methanol
(12 mL) was added and ammonia was allowed to evap-
orate. The mixture was concentrated and the residue
was dissolved in pyridine–Ac2O (6 mL/3 mL) and stirred
for 20 h at room temperature. The reaction was
quenched by addition of methanol, water was added
(40 mL), and the aqueous layer was extracted with
EtOAc (4· 60 mL). Combined organic layers were dried
and concentrated. Flash chromatography of the crude
(n-hexane/EtOAc 7:3) afforded compound 6 (0.10 g,
78%) as an amorphous solid.
J1 ;2 ¼ 3:8 Hz, H-10), 5.53 (ddt, 1H, J4,5 = 15.2 Hz,
J4,3 = 7.9 Hz, J4,6 = 1.4 Hz, H-4), 5.60 (dd, 1H,
J3,4 = 7.9 Hz, J3,2 = 4.5 Hz, H-3), 5.88 (dt, 1H,
J5,4 = 15.2 Hz, J5,6 = 6.8 Hz, H-5), 7.40–7.20 (m, 20 H,
Ph), 7.45 (br t, 2H, J = 7.8 Hz, Ph), 7.57 (br t, 2H,
J = 7.5 Hz, Ph), 8.06 (br d, 2H, J = 8.0 Hz, Ph). 13C
NMR (CDCl3): d 14.1, 22.7, 28.7, 29.2–29.7 (7C), 31.9,
32.3, 64.0, 67.9, 69.0, 69.9, 73.2, 73.3, 73.4, 74.7, 74.9,
75.1, 76.5, 78.6, 98.9, 122.9, 127.4–128.3 (Ph), 128.4
(Ph) 129.8 (Ph), 130.0 (Ph), 133.2, 137.9–138.8 (Ph),
138.6, 165.2. ESI-MS (positive-ion mode): m/z 974.5
[M+Na]+.
0
0
4.1.3. (2S,3R,4E)-1-(2,3,4,6-Tetra-O-benzyl-a-D-galacto-
pyranosyloxy)-2-(docosanoylamino)-3-hydroxy-4-octade-
cene (5). To a stirred solution of compound 4 (0.38 g,
0.40 mmol) in dry CH2Cl2 (10 mL) sodium methoxide
in dry methanol (0.05 M solution, 4.8 mL) was added
and the solution was stirred at room temperature for
2 h. The solution was neutralized with an ion exchange
resin (Dowex 50 · 8, H+ form), filtered, and concen-
trated. The crude was dissolved in pyridine/water (1:1,
12 mL). Hydrogen sulfide was bubbled into the solution
for 20 min and the solution was stirred at room temper-
ature for 48 h. The reaction mixture was concentrated
20
D
1
½aꢁ þ 46:7 (c 1, CHCl3). H NMR (CDCl3): d 0.90 (t,
6H, J = 7.5 Hz, 2 CH3), 1.18–1.45 (m, 58H, 29 CH2),
1.57–1.68 (m, 2H, COCH2CH2), 2.02, 2.05, 2.06, 2.14,
2.16 (5s, 15H, 5 COCH3), 2.04 (m, 2H, CH@CHCH2),
2.15 (m, 2H, COCH2), 3.57 (dd, 1H, J1a,1b = 10.5 Hz,
J1a,2 = 4.0 Hz, H-1a), 3.74 (dd, 1H, J1b,2 = 3.5 Hz, H-
1b), 4.03–4.20 (m, 3H, H-50, H2-60), 4.38 (m, 1H, H-
1b), 5.06 (d, 1H, J4 ;3 ¼ 3:5 Hz, H-40), 5.16 (dd, 1H,
0
0

L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
5535
J3 ;2 ¼ 11:0 Hz, J3 ;4 ¼ 3:5 Hz, H-30), 5.30 (t, 1H,
J4,5 = 15.5 Hz, J4,3 = 7.5 Hz, J4,6 = 1.0 Hz, H-4), 5.71
(dt, J5,4 = 15.5 Hz, J5,6 = 6.5 Hz, H-5). 13C NMR
(CDCl3/CD3OD, 1:1): d 13.6 (2C), 22.5 (2C), 25.9,
29.2–29.5 (24C), 31.8 (3C), 32.3, 36.2, 53.6, 61.5, 66.9,
67.1, 68.1, 70.7, 71.2, 78.1, 99.6, 129.5, 134.0, 174.8.
ESI-MS (negative-ion mode): m/z 862.8 [MꢀNa]ꢀ.
0
0
0
0
0
0
J = 8.0 Hz, H-3), 5.35 (dd, 1H, 1H, J2 ;3 ¼ 11:0 Hz,
J2 ;1 ¼ 3:5 Hz, H-20), 5.41 (dd, 1H, J4,5 = 15.5 Hz,
0
0
J4,5 = 5.5 Hz, H-4), 5.48 (d, 1H, J1 ;2 ¼ 3:5 Hz, H-10),
5.67 (d, 1H, JNH,2 = 9.0 Hz, NH), 5.82 (dt, 1H,
J4,5 = 15.5 Hz, J5,6 = 7.5 Hz, H-5). 13C NMR (CDCl3):
d 14.1 (2C), 20.5, 20.6 (2C), 20.7, 21.1, 22.7 (2C), 25.8,
29.0–29.8 (24C), 31.9 (3C), 32.3, 36.9, 50.5, 61.8, 66.6,
67.4, 67.5, 67.8, 67.9, 72.9, 97.0, 124.8, 138.0, 169.6,
170.1, 170.2, 170.4, 170.5, 172.7. ESI-MS (positive-ion
mode): m/z 1016.6 [M+Na]+.
0
0
4.2. Generation of T cell clones and activation assays
The K34B9.1 T cell clone specific for sulfatide was gen-
erated as previously described.17 The iNKT cell clones
BGA01 and BGA82 were obtained by a-GalCer stimu-
lation of peripheral blood mononuclear cells of healthy
donors are type I iNKT. To investigate their capability
to react to a-sulfatide 1, T cells (105/well) were incubated
for 48 h together with C1R-CD1d-transfected cells
(2.5 · 104/well) and increasing doses of glycolipids. T
cell activation was detected by measuring the amounts
of IFNc and IL-4 released in the supernatant, using
ELISA as previously described.17
4.1.5. (2S,3R,4E)-1-(a-D-Galactopyranosyloxy)-2-(doc-
osanoylamino)-3-hydroxy-4-octadecene (7). To a stirred
solution of compound 6 (0.10 g, 0.10 mmol) in dry
CH2Cl2 (4 mL) sodium methoxide in dry methanol
(0.05 M solution, 0.8 mL) was added and the solution
was stirred at room temperature for 3 h. The solution
was neutralized with an ion exchange resin (Dowex
50 · 8, H+ form), filtered, and concentrated affording 7
(0.08 g, quantitative) as a foam.
20
D
½aꢁ þ 30:6 (c 0.5, CHCl3/CH3OH, 1:1). 1H NMR
Acknowledgments
(CDCl3/CD3OD, 1:1): d 0.88 (t, 6H, J = 7.5 Hz, 2
CH3), 1.20–1.40 (m, 58H, 29 CH2), 1.55–1.64 (m, 2H,
COCH2CH2), 2.02 (m, 2H, CH@CHCH2), 2.18 (t, 2H,
J = 7.5 Hz, COCH2), 3.68–3.83 (m, 7H, H-1a, H-1b,
This work was supported by MIUR-Italy, COFIN 2004:
‘‘Structure and synthesis of glycolipids’’ (to F.R. and
A.M.), the Swiss National Foundation Grant 3100A0-
109918 (to G.D.L.), the MOLSTROKE project, Contract
No. LSH-CT-2004-005206, funded under the EU
FP6-2003 LIFESCIHEALTH programme (to G.D.L.),
and the Swiss Society of Multiple Sclerosis, the University
of Piemonte Orientale and IRCAD-Novara (to L.P.).
H-20, H-30, H-50, H2-60), 3.92 (br d, 1H, J4 ;3 ¼ 2:8 Hz,
0
0
H-40), 3.95 (m, 1H, H-2), 4.07 (br t, 1H, J = 7.0 Hz,
H-3), 4.87 (d, 1H, J1 ;2 ¼ 3:5 Hz, H-10), 5.44 (ddt, 1H,
J4,5 = 15.5 Hz, J4,3 = 7.5 Hz, J4,6 = 1.0 Hz, H-4), 5.71
(dt, 1H, J5,4 = 15.5 Hz, J5,6 = 6.5 Hz, H-5). 13C NMR
(CDCl3/CD3OD, 1:1): d 13.7 (2C), 22.5 (2C), 25.9,
29.2–29.5 (24C), 31.8 (3C), 32.3, 36.3, 53.7, 61.6, 67.3,
69.0, 69.6, 70.1, 70.7, 72.0, 99.9, 129.2, 134.0, 174.7.
ESI-MS (negative-ion mode): m/z 782.6 [MꢀH]ꢀ.
0
0
References and notes
1. Porcelli, S. A. Adv. Immunol. 1995, 59, 1.
2. De Libero, G.; Mori, L. Nat. Rev. Immunol. 2005, 5, 485.
3. Shamshiev, A.; Donda, A.; Carena, I.; Mori, L.; Kappos,
L.; De Libero, G. Eur. J. Immunol. 1999, 29, 1667.
4. De Libero, G.; Moran, A. P.; Gober, H.-J.; Rossy, E.;
Shamshiev, A.; Chelnokova, O.; Mazorra, Z.; Vendetti, S.;
Sacchi, A.; Prendergast, M. M.; Sansano, S.; Tonevitsky,
A.; Landmann, R.; Mori, L. Immunity 2005, 22, 763.
5. Natori, T.; Morita, M.; Akimoto, K.; Koezuka, Y.
Tetrahedron 1994, 50, 2771.
6. Kawano, T.; Cui, J.; Koezuka, Y.; Toura, I.; Kaneko, Y.;
Motoki, K.; Ueno, H.; Nakagawa, R.; Sato, H.; Kondo,
E.; Koseki, H.; Taniguchi, M. Science 1997, 278, 1626.
7. 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.
8. Van Kaer, L. Nat. Rev. Immunol. 2005, 5, 31.
9. Berkers, C. R.; Ovaa, H. Trends Pharmacol. Sci. 2005, 26,
252.
10. Taniguchi, M.; Harada, M.; Kojo, S.; Nakayama, T.;
Wakao, H. Annu. Rev. Immunol. 2003, 21, 483.
11. Costantino, V.; Fattorusso, E.; Imperatore, C.; Mangoni,
A. Tetrahedron 2002, 58, 369.
12. Barbieri, L.; Costantino, V.; Fattorusso, E.; Mangoni, A.;
Aru, E.; Parapini, S.; Taramelli, D. Eur. J. Org. Chem.
2004, 468.
13. Barbieri, L.; Costantino, V.; Fattorusso, E.; Mangoni, A.;
Basilico, N.; Mondani, N. Eur. J. Org. Chem. 2005, 15,
3279.
4.1.6. (2S,3R,4E)-1-[3-O-(Sodium oxysulfonyl)-a-D-galac-
topyranosyloxy]-2-(docosanoylamino)-3-hydroxy-4-octade-
cene (1). Compound 7 (0.08 g, 0.06 mmol) and Bu2SnO
(0.02 g, 0.1 mmol) were stirred in MeOH (4 mL) at
reflux under argon for 2 h. The solvent was evaporated
off under reduced pressure, and the dibutylstannylene
complex was treated with Me3NÆSO3 (0.02 g, 0.14 mmol)
in THF (4 mL) for 2 h. The solvent was removed under
reduced pressure, then the residue was dissolved in
CHCl3/MeOH, 1:1 (2 mL), loaded onto a cation
exchange resin column (Dowex 50 · 8 Na+ form, 0.5 ·
6 cm), eluted with CHCl3/MeOH (1:1), concentrated
under reduced pressure and subjected to flash chroma-
tography (CHCl3/MeOH, 9:1) to give the target
compound 1 (70%, 0.06 g) as a foam.
20
D
½aꢁ þ 42:0 (c 0.5, CH3OH). 1H NMR (CDCl3/CD3OD,
1:1): d 0.88 (t, 6H, J = 7.5 Hz, 2 CH3), 1.20–1.40 (m,
58H, 29 CH2), 1.55–1.64 (m, 2H, COCH2CH2), 2.03
(m, 2H, CH@CHCH2), 2.20 (t, 2H, J = 7.5 Hz,
COCH2), 3.69–3.82 (m, 4H, H-1a, H-1b, H2-60), 3.85
(br t, 1H, J = 5.5 Hz, H-50), 3.96 (m, 1H, H-2), 4.05
(dd, 1H, J2 ;3 ¼ 10:5 Hz, J2 ;1 ¼ 4:0 Hz H-20), 4.12 (t,
0
0
0
0
0
0
1H, J = 7.5 Hz, H-3), 4.34 (br d, 1H, J4 ;3 ¼ 2:5 Hz,
H-40), 4.50 (dd, 1H, J3 ;2 ¼ 10:5 Hz, J3 ;4 ¼ 2:5 Hz, H-
0
0
0
0
30), 4.87 (d, 1H, J1 ;2 ¼ 3:5 Hz, H-10), 5.44 (ddt, 1H,
0
0

5536
L. Franchini et al. / Bioorg. Med. Chem. 15 (2007) 5529–5536
14. Wu, D.; Xing, G. W.; Poles, M. A.; Horowitz, A.; Kinjo,
Y.; Sullivan, B.; Bodmer-Narkevitch, V.; Plettenburg, O.;
Kronenberg, M.; Tsuji, M.; Ho, D. D.; Wong, C. H. Proc.
Natl. Acad. Sci. U.S.A. 2005, 102, 1351.
15. 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.
16. 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.; Bersa, G. S.; Porcelli, S. A.
Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 3383.
17. Shamshiev, A.; Gober, H.-J.; Donda, A.; Mazorra, Z.;
Mori, L.; De Libero, G. J. Exp. Med. 2002, 195,
1013.
18. Compostella, F.; Franchini, L.; De Libero, G.;
Palmisano, G.; Ronchetti, F.; Panza, L. Tetrahedron
2002, 58, 8703.
19. Zajonc, D. M.; Maricic, I.; Wu, D.; Halder, R.; Roy, K.;
Wong, C.-H.; Kumar, V.; Wilson, I. A. J. Exp. Med. 2005,
202, 1517.
20. Zajonc, D. M.; Elsliger, M. A.; Teyton, L.; Wilson, I. A.
Nat. Immunol. 2003, 4, 808.
21. Koch, M.; Stronge, D. S.; Gadola, S. D.; Mathew, G.;
Ritter, A. R.; Fersht, G. S.; Schmidt, R. R.; Jones, E. Y.;
Cerundolo, V. Nat. Immunol. 2005, 6, 819.
22. Franchini, L.; Compostella, F.; Donda, A.; Mori, L.;
Colombo, D.; De Libero, G.; Matto, P.; Ronchetti, F.;
Panza, L. Eur. J. Org. Chem. 2004, 4755.
23. Compostella, F.; Ronchi, S.; Panza, L.; Mariotti, S.; Mori,
L.; De Libero, G.; Ronchetti, F. Chem.: Eur. J. 2006, 12,
5587.
24. Rosembroox, W.; Riley, D. A.; Lartey, P. A. Tetrahedron
Lett. 1985, 26, 3.
25. Shingu, Y.; Nishida, Y.; Dohi, H.; Kobayashi, K. Org.
Biomol. Chem. 2003, 1, 2518.
26. Nishida, Y.; Shingu, Y.; Dohi, H.; Kobayashi, K. Org.
Lett. 2003, 5, 2377.
27. Guilbert, B.; Davis, N. J.; Pearce, M.; Aplin, R. T.;
Flitsch, S. L. Tetrahedron: Asymmetry 1994, 5, 2163.
28. Xing, G.-W.; Wu, D.; Poles, M. A.; Horowitz, A.; Tsuji, M.;
Ho, D. D.; Wong, C.-H. Bioorg. Med. Chem. 2005, 13, 2907.
29. Modica, E.; Compostella, F.; Colombo, D.; Franchini, L.;
Cavallari, M.; Mori, L.; De Libero, G.; Panza, L.;
Ronchetti, F. Org. Lett. 2006, 8, 3255.
30. Wild, R.; Schmidt, R. R. Tetrahedron: Asymmetry 1994, 5,
2195.
31. Nicolaou, K. C.; Caulfield, T. J.; Katoaka, H. Carbohydr.
Res. 1990, 202, 177.
32. Graziani, A.; Passacantilli, P.; Piancatelli, G.; Tani, S.
Tetrahedron: Asymmetry 2000, 11, 3921.