Introduction of aromatic group on 4′-OH of α-GalCer manipulated NKT cell cytokine production

Article abstract of DOI:10.1016/j.bmc.2010.11.061

The glycosphingolipid α-GalCer has been found to influence mammalian immune system significantly through the natural killer T cells. Unfortunately, the pre-clinical and clinical studies revealed several critical disadvantages that prevented the therapeutic application of α-GalCer in treating cancer and other diseases. Recently, the detailed illustration of the CD1d/α-GalCer/NKT TCR complex crystal structural, together with other latest structural and biological understanding on glycolipid ligands and NKT cells, provided a new platform for developing novel glycolipid ligands with optimized therapeutic effects. Here, we designed a series of novel aromatic group substituted α-GalCer analogues. The biological activity of these analogues was characterized and the results showed the unique substitution group manipulated the immune responses of NKT cells. Computer modeling and simulation study indicated the analogues had unique binding mode when forming CD1d/glycolipid/NKT TCR complex, comparing to original α-GalCer.

Full text of DOI:10.1016/j.bmc.2010.11.061

Bioorganic & Medicinal Chemistry 19 (2011) 2767–2776
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Introduction of aromatic group on 40-OH of
cell cytokine production
a-GalCer manipulated NKT
Wenpeng Zhang a, Chengfeng Xia a,b,c, , Janos Nadas b, Wenlan Chen b, Li Gu a,d, Peng George Wang a,b,
⇑
⇑
a Department of Biochemistry, The Ohio State University, Columbus, OH 43210, USA
b Department of Chemistry, The Ohio State University, Columbus, OH 43210, USA
c State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, Chinese Academy of Sciences, Kunming, Yunnan 650204, China
d State Key Laboratory of Microbial Technology (SKLMT), Shandong University, Jinan, Shandong 250100, China
a r t i c l e i n f o
a b s t r a c t
Article history:
The glycosphingolipid
a
-GalCer has been found to influence mammalian immune system significantly
Received 15 September 2010
Revised 23 November 2010
Accepted 25 November 2010
Available online 2 December 2010
through the natural killer T cells. Unfortunately, the pre-clinical and clinical studies revealed several crit-
ical disadvantages that prevented the therapeutic application of -GalCer in treating cancer and other
diseases. Recently, the detailed illustration of the CD1d/ -GalCer/NKT TCR complex crystal structural,
together with other latest structural and biological understanding on glycolipid ligands and NKT cells,
provided a new platform for developing novel glycolipid ligands with optimized therapeutic effects. Here,
a
a
Keywords:
we designed a series of novel aromatic group substituted a-GalCer analogues. The biological activity of
a-GalCer
these analogues was characterized and the results showed the unique substitution group manipulated
the immune responses of NKT cells. Computer modeling and simulation study indicated the analogues
had unique binding mode when forming CD1d/glycolipid/NKT TCR complex, comparing to original
NKT cell ligand
Computer modeling
CD1d
a-GalCer.
NKT TCR
Ó 2010 Published by Elsevier Ltd.
1. Introduction
such as Class I CD1 (CD1a, -b, -c) and Class II CD1 (CD1d and
CD1e). There is only a single type of CD1 protein in mouse, CD1d,
which is homologous to its human isoform.6–10
As a special sub-population of T cells, natural killer T (NKT) cell
expresses a T cell receptor (TCR) that can recognize glyco-
sphingolipids and other glycolipid antigens when they are being
presented by antigen presenting cells (APC) through a non-conven-
tional, MHC class I-like glycoprotein named as CD1d.1–3 NKT cells
have a broad influence on immune system due to the varied signals
emanating after being stimulated by antigens. Once being acti-
vated, the NKT cells will show a reciprocal signal pattern with a ra-
pid but strong secreting of Th1 and Th2 cytokines and other
chemokines in a short time period. Through these cytokines and
other signal molecules, the NKT cells involve in both innate and
adaptive immunity and regulate a broad range of immune re-
sponses. As the primary contact between NKT cells and APC, the
interaction among NKT TCR, glycolipid antigen and CD1d protein,
initiates and affects the outcome of NKT cell immune responses.
CD1d protein is a monomorphic protein and it can present dif-
ferent glycolipid antigens to be recognized by NKT TCR from differ-
ent individuals, even across species.4,5 It belongs to the CD1 family,
which includes five members in two different classes in human,
a
-Galactosylceramide (a-GalCer) was originally discovered, and
named as agelasphins, by the Pharmaceutical Division of the Kirin
Brewery Company during a screen of components derived from
marine sponge Agelas mauritianus as a reagent to prevent tumor
metastases in mice.11,12 The most striking structural feature of
a
-GalCer is the a linkage between the sugar and ceramide, which
is very uncommon in mammalian systems. Structure–activity rela-
tionship (SAR) studies resulted in an optimized ligand, KRN7000,
with an 18-carbon sphingosine and a 26-carbon acyl chain.13
a
-GalCer can be readily loaded onto CD1d, and the CD1d/a-GalCer
complex on the surface of antigen presenting cells can be recog-
nized by the TCR of all types of NKT cells.14–16
a
-GalCer has been extensively studied in preclinical studies and
clinical trials for treating cancer and various other diseases.13,15,17–
26
However, even the Phase I clinical trials proved that
a-GalCer
was safe and immune-stimulatory, unfortunately, no optimistic re-
sponses in patients have been observed.27–29
Several reasons have been indentified to cause the failure of
a-GalCer in clinical trial, such as the much lower abundance of
NKT cells in human and the ineffectiveness in expanding NKT cells
⇑
in patients.27 Also, both the short response period and long term
Corresponding authors. Tel.: +86 871 5223354; fax: +86 871 522 3035 (C.X.);
tel.: +1 614 292 9884; fax: +1 614 688 3106 (P.G.W.).
anergy of NKT cells triggered by
a-GalCer are not ideal for treating
E-mail addresses: xiachengfeng@mail.kib.ac.cn (C. Xia), wang.892@osu.edu,
pgwang@chemistry.ohio-state.edu (P.G. Wang).
some of the chronic diseases.27,30–35 The antagonizing effects of
0968-0896/$ - see front matter Ó 2010 Published by Elsevier Ltd.
doi:10.1016/j.bmc.2010.11.061

2768
W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
Th1 and Th2 cytokines released after a-GalCer induction is another
2. Results
The new aromatic
main factor impaired its clinical effects. So, a further modified
-GalCer analogue with optimized cytokine releasing profile and
immune response would have a great therapeutic potential.
The crystal structures of CD1d/ -GalCer complex and CD1d/
-GalCer/NKT TCR complex revealed a great amount of detail infor-
a
a
-GalCer analogues (structures shown in
Fig. 1) were synthesized and characterized as described in Section 5.
a
a
2.1. The in vitro stimulation from aromatic glycolipids against
NKT hybridoma cells
mation about the interactions among the proteins and glycolipid,
which would benefit the new analogue design.5,36 From the struc-
ture, a-GalCer interacts with CD1d and NKT TCR mainly through H-
In the in vitro hybridoma assay, all three analogues showed sig-
nificant activity against different NKT hybridoma cell lines
(DN3A4-1.2, DN3A4-1.4 and N38-2H4) (Fig. 2A). Especially, the
compound Ar3-GSL showed an overall stronger stimulation than
bonds and it maintains the conformation by protruding only -OH
groups from galactose ring to the NKT TCR as its lipid tails burying
in the CD1d binding groove.
The a-GalCer maintains most of the key interactions with CD1d
a-GalCer in all three hybridoma cell lines, which is not common
protein through its ceramide moiety. The 3-OH and 4-OH on the
sphingosine chain both form critical H-bonds with CD1d.36,37
Meanwhile, the glycosidic linkage C10-O and the C20-OH on galact-
ose ring are also identified to interact with CD1d protein.36 An ex-
tra H-bond between C30-OH and CD1d as well as multi van der
Waals interactions were also characterized in some of other crystal
structures.37,38
for previous evaluated -GalCer analogues. Meanwhile, the
a
compounds Ar2-GSL and Ar1-GSL were also showed strong stimu-
lation against certain hybridoma cell lines.
The analogues were presented to stimulate the hybridoma cells
directly by surface-bound CD1d protein in the plate-bound CD1d
presenting assay. As shown in Figure 2B, for both 1.2 and 2H4
hybridoma cell lines, the compounds Ar1-GSL and Ar3-GSL stimu-
The
groups on galactose ring while CDR1
TCR were found to play critical roles. The major interactions in-
clude the C20- and C40-OH groups forming H-bond with Gly96
Phe29
and C30-OH interacting with Ser30a 5
a-GalCer interacts with NKT TCR mainly through -OH
lated stronger hybridoma responses than
a-GalCer did, which
a
and CDR3 loops of NKT
a
means the modified analogues Ar1-GSL and Ar3-GSL can be loaded
efficiently onto the CD1d protein and bind with NKT TCR with a
high binding affinity to trigger a strong NKT hybridoma response.
a,
a
.
The structure–activity relationship study results are consistent
with the crystal structure studies. Modifications on sphingosine li-
pid -OH groups would change the binding affinity but not abolish-
ing the activity of
on galactose ring disrupted the complex and eliminated the activ-
2.2. The in vivo stimulation profile of aromatic glycolipids
In the splenocytes stimulation assay, all three analogues still
showed significant stimulation activity (Fig. 3A). Among the ana-
logues, the compound Ar3-GSL exhibited an overall comparable
a
-GalCer while modifications on C20- and C30-OH
ity of a-GalCer due to their critical interactions with CD1d and NKT
activity as
cated the Ar3-GSL having certain Th2 cytokine preference. The
other two analogues are not as active as -GalCer, except when
a-GalCer, especially on Th2 (IL-4) cytokine, which indi-
TCR, also the limited surrounding space.39–41 The modifications on
C60-OH of galactose ring did not interrupt the interaction inside the
complex as it points towards the solvent and surrounded by free
space.42
a
the stimulation concentration is very high (1000 ng/mL).
To evaluate their in vivo stimulation properties, the blood cyto-
kine profile was characterized after the mice being treated by the
analogues. Total 5 lg of each analogues and a-GalCer were iv in-
jected into C57BL/6 mice and the blood sample from each mouse
was collected at different time points (0, 2, 4, 6, 12, and 24 h after
injection). The cytokine (IFN-c for Th1; IL-4 for Th2) level in blood
serum was characterized by ELISA (Fig. 3B). All the analogues
showed a similar time-dependent cytokine stimulation profile as
For the C40-OH group, it has been proved that modifications can
be tolerated on this group as it is not as critical as the C20-OH group
although it also maintains direct H-bond with NKT TCR in the com-
plex. The modified analogues showed varied immunogenicity
against NKT cells depending on the substitution groups.41,43
Based on these previous understandings on the interactions
among CD1d protein,
of novel NKT cell ligands with aromatic residues on C40 of
expecting to introduce stacking effects in the complex and
a-GalCer, and NKT TCR, we designed a series
a
-GalCer,
a
-GalCer. Specifically, for the Th1 cytokine, the blood IFN-
concentration reached peak at 12 h after injection. Ar2-GSL and
Ar3-GSL reached a stronger Th1 peak than -GalCer, even they
c
p–p
tune immunogenicity against NKT cells through the extra interac-
tion. Here, we are reporting the biological activity tests demon-
strated the modifications could manipulate the NKT cell immune
response as proposed. The modeling and simulation study revealed
the binding affinity change induced by the modification.
a
did not show superior stimulation in earlier time points. For the
Th2 cytokine, IL-4, Ar3-GSL showed exactly same time profile as
a-GalCer, reaching the peak stimulation 2 h after injection and
decreasing quickly. Impressively, Ar3-GSL stimulated a much
Figure 1. Structures of a
-GalCer and novel C40 aryl-analogues with different length spacer.

W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
2769
Figure 2. The in vitro hybridoma stimulation by
a-GalCer analogues. (A) Three NKT hybridoma cell lines (DN3A4-1.2, DN3A4-1.4, N38-2H4) were stimulated by cell A20/CD1
presented glycolipids. (B) Two NKT hybridoma cell lines (DN3A4-1.2, N38-2H4) were stimulated by plate-bounding CD1d protein presented glycolipids.
stronger Th2 cytokine releasing. For compounds Ar1-GSL and Ar2-
GSL, they showed an extended Th2 stimulation peak that main-
tained a significant Th2 level 4 h after initial injection. These re-
sults could be taken as the evidence of manipulated cytokine
in a spectrum. The crystal structure also shows, although it forms
an H-bond with CDR1
loop, the C40-OH itself, like the C60-OH,
a
appears to be located in a much more open solvent-accessible cav-
ity, comparing to the buried position of C20- and C30-OH.5,41,43
The crystal structure revealed the presence of two aromatic res-
releasing profile by the C40 position modification on
a-GalCer.
idues around the C40-OH position, Phe51
a
of TCR and Trp153 of
CD1d,5 which suggests the possibility of incorporating
p–p
stack-
3. Discussion
ing interactions to increase the binding of glycolipid ligand to both
CD1d and TCR. To validate this hypothesis, the homology models of
several C40 position aromatic modified
a
-GalCer analogues were
-GalCer/NKT TCR ternary
structure. Based on our previous study,45 the computer simulation
results confirmed the possibility of forming stacking between
the aromatic rings from Phe51 -Gal-
and proposed C40 aromatic
The information from crystal structure of CD1d/a-GalCer/NKT
TCR ternary complex and other related complexes, together with
the results from in vitro and in vivo biological studies, has showed
an insight into the recognition interactions between the glycolipid
antigen and involved functional proteins, especially the interac-
tions among the galactose sugar ring, TCR CDR loops, and CD1d
protein. Both the C20- and C30-OH groups forms critical H-bonds
created mainly based on the CD1d/
a
p–p
a
a
Cer analogues while no such interaction was predicted for the aro-
matic ring from Trp153 of CD1d protein,45 which may be beneficial
for avoiding too much conformation disrupt from the extra
interaction.
Previous results illustrated that the presence of a bulky group
on the C40 position will abolish the immunogenicity, and the anti-
gen can be recognized by the NKT TCR only after the bulky group
being removed by glycosidase.5,41,43 The biological activity study
with residues from CDR1a and CDR3a loops of NKT TCR and
CD1d protein. Most of the modifications on these two -OH groups
would abolish the H-bonds, change the binding orientation of
ligand and receptor, and overall eliminate the immunogenicity of
resulted ligands.5,44 On the other hand, the modification on the
C40-OH resulted a series of analogues with activity profiles varied

2770
W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
Figure 3. The in vivo immunogenicity profile of aromatic
a
-GalCer analogues. (A) The cytokines (IFN-
c, IL-4) released by mouse splenocytes after being stimulated by
glycolipid antigens. (B) The time dependent pattern of cytokine (IFN-
c
, IL-4) level in mouse serum after iv treatment of glycolipids.
proved that these aromatic analogues do have the immunogenicity
against the NKT cells. As all of the analogues showed significant
activity in varied assays, it means the introducing of aromatic rings
Ar3-GSL which has a 3-carbon linker between C40-O and the aro-
matic ring, the compound Ar2-GSL with a 2-carbon linker only
showed very limited stimulation. This indicates a correct linker
length for the aromatic ring to locate properly is very critical.
Meanwhile, the significant stimulation from the single-carbon lin-
ker compound Ar1-GSL should be partly credited to the compensa-
tion binding affinity from the extra carbonyl group on the linker
carbon.
onto the C40 position of
a-GalCer did not interrupt the basic bind-
ing and recognition among the glycolipid ligand, CD1d and NKT
TCR.
Through the data, it can be found that the aromatic substitution
at the C40 of
a-GalCer affected the NKT cell function by altering the
binding affinity with CD1d and NKT TCR while the substitution still
kept the interactions strong enough for NKT cell stimulation. Both
the in vitro hybridoma stimulation and plate-bound CD1d present-
ing assay indicated the differences of cytokine releasing strength,
which is resulted by the different binding affinity that induced
by the aromatic modification groups. Especially in the plate-bound
CD1d presenting assay, the binding affinity would be the main fac-
tor causing the stimulation variations since the affects from the
antigen presenting cells were excluded by presenting the ligands
directly through plate-binding CD1d protein. As all the analogues
In the in vivo splenocytes culture, the aromatic analogues
showed stimulation effects on both Th1 (IFN-c) and Th2 (IL-4)
cytokines, but, with weaker relative stimulation strength from
the hybridoma model and plate-bound CD1d presenting model,
comparing to the stimulation strength from
a-GalCer. Among the
analogues, the compound Ar3-GSL is still superior to the others,
with the compound Ar1-GSL following, and then the compound
Ar2-GSL, which is consistent with the plate-bound CD1d present-
ing assay results. This result can be also taken as evidence that
the in vivo stimulation profile of NKT cell is related to the binding
property between glycolipid ligand and NKT TCR. The preference of
Th1 and Th2 cytokines for these analogues had not much differ-
share a same ceramide moiety and the C40-OH on
a-GalCer mainly
interacts with NKT TCR, the loading efficiency for each analogue
onto CD1d should also be similar. This means the hybridoma re-
sponses level of plate-bound CD1d presenting assay was majorly
determined by the binding strength between the presented ligands
and NKT TCR, which was affected by the modification on C40 group
in this case. For example, comparing to the very active compound
ence from the
Th2 preference comparing the relative strength against
In the in vivo mouse stimulation, the analogue compounds gave
a similar blood Th1 cytokine (IFN- ) profile as -GalCer did,
although with strength variations, when some of the analogues
a
-GalCer although the analogues showed a slight
a-GalCer.
c
a

W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
2771
changed significantly on the profile of Th2 cytokine (IL-4) (Fig. 3B).
The compound Ar3-GSL showed the most significant IL-4 stimula-
tion while the time pattern is still same as -GalCer’s. The com-
pounds Ar1-GSL and Ar2-GSL both gave a lower but broader peak
which indicates they extended the stimulation retention on Th2
cytokine. This could be caused by the changed metabolism process
of the analogues due to the modification, or, by the binding affinity
change, which is more reasonable here as the both analogues did
not show time dependent profile change on Th1 cytokine.
Molecular modeling and simulation were also applied to further
characterize the interaction among ligands and proteins. Since the
first crystal structure of CD1d/NKT TCR complex (2PO6) was pub-
lished, five more crystal structures have been solved with mCD1d
or varying the chains of NKT TCR. All the crystal structures showed
a similar binding footprint implying that the interaction between
CD1d and TCR is more of a lock-and-key interaction rather than
an induced-fit interaction.46 A 3D alignment of all the crystal struc-
tures verified this observation. More importantly, different ligands
were shown to induce small conformational changes in TCR upon
TCR binding to ensure that the correct binding footprint would
be formed, and that these minor changes could potentially explain
the weaker potency of the ligands.46
aromatic residues lying posterior to the 40-OH of
The new model possessed the Phe51
above -GalCer but lacked the Trp153 amino acid on CD1d directly
below the 40-OH of
-GalCer. This model underwent an energy
minimization followed by the modification of the -GalCer ligand
to yield Ar1-GSL, Ar2-GSL, and Ar3-GSL and then a subsequent
minimization. The two complexes of mCD1d/Ar2-GSL and Ar3-
GSL/TCR then underwent a 200 picosecond molecular dynamics
simulation in implicit solvent to see whether any conformational
changes could be induced by different 40- substituents.
a
-GalCer (Fig. 4).
a
on the CDR loop directly
a
a
a
a
The molecular dynamics simulations were able to show minute
conformational changes that can potentially explain the varying
cytokine profiles of Ar2-GSL and Ar3-GSL (Fig. 5). The 2 carbon lin-
ker on Ar2-GSL is just flexible and long enough to form the pre-
dicted
hand, the 3 carbon linker of Ar3-GSL creates a length that is too
long to interact with Phe51 , and its presence only displaces
Phe51 slightly from its crystal conformation. All the hydrogen
bonds between the ligands and the two proteins are maintained,
therefore, the added non-covalent interaction should add to
the binding energy in the complex. Non-covalent interactions
p–p stacking interactions with Phe51a. On the other
a
a
p–p
p–p
have been shown to have binding energies of 1–3 kcal/mol with
substituents on the aromatic rings being able to increase binding
energies by 1 kcal/mol.47 However, the experimental results show
that Ar3-GSL is a more potent ligand than Ar2-GSL.
Molecular dynamics simulations were used to determine the
cause in the varying cytokine profiles of the different ligands of
a-GalCer, Ar1-GSL, Ar2-GSL, and Ar3-GSL. The experimental proce-
dures utilized mCD1d to present the four ligands to NKT-TCR.
Therefore, a new model had to be created to analyze the experi-
mental results since no crystal structure exists of mCD1d bound
The early computational studies performed in our laboratory
showed that different ligands were capable of inducing slight
changes in the binding footprint of CD1d and TCR.45 The binding
footprint of Ar2-GSL was found to possess very small differences
from Ar3-GSL even after only the 200 ps simulation (Fig. 5). The
Tyr48b lost its hydrogen bond with Glu83 on CD1d more than
to the human NKT-TCR protein possessing the NKT15
a and b
chains. The difference between mCD1d and hCD1d as well as their
respective NKT-TCR proteins has everything to do with the lack of
Figure 4. The aligned crystal structures of mCD1d and hCD1d presenting
a-GalCer to their respective NKT-TCR proteins, PDB ID 3HE6 and 2PO6, respectively. The major
difference between the two crystal structures lies in the fact that the mCD1d structure lacks both aromatic residues that sit behind the 40-OH position of
a-GalCer. The surface
area representation of hCD1d shows that there may be enough space between the residues for potential interactions with a substituent group on the 40-position.
Figure 5. The binding footprints of Ar2-GSL and Ar3-GSL complexes after 200 ps of molecular dynamics simulations. The two-carbon linker on Ar2-GSL provides enough
flexibility and length to form p–p interactions with Phe51a, whereas in Ar3-GSL, the three-carbon linker is too long. However, the slight displacement of TCR for both ligands
causes the loss of hydrogen bonds involving Glu83 for Ar2-GSL or the maintaining and formation of new hydrogen bonds involving Glu83 for Ar3-GSL. The breaking and
forming of such non-covalent interactions between the two proteins is a possible explanation for the potency of the ligands.

2772
W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
likely due to a tilting upward of TCR upon formation of the
interaction. For Ar3-GSL, this hydrogen bond was maintained as
well as Arg103 swinging over to form an additional hydrogen
bond with Glu83. The gaining of the non-covalent interaction
can not compensate for the loss of a hydrogen bond since each
hydrogen bond will contribute between 2 and 7 kcal/mol energy
to the overall binding energy.48 Since both compounds are still
capable of eliciting a cytokine response, the binding footprint has
to be maintained for both. The difference in their cytokine profiles
though will more than likely be due to slight changes in the TCR
protein upon binding causing the breaking or formation of novel
non-covalent interactions.
This result would further extend the understanding on how
the binding affinity between CD1d presented glycolipid ligands
and NKT TCR could affect the NKT cell response. The recognition
and binding affinity between sugar moiety of glycolipid ligand
and NKT TCR could affect the NKT cell response pattern as it
would affect the stability of CD1d/glycolipid/NKT TCR complex.
The overall in vivo effects on NKT cell response include the re-
sponse strength and also the cytokine preference. Moreover, it
also includes the time profile of NKT cell response as the modifi-
cation on the analogues could not only change the space fitting
and binding strength, but also the time-dependent manipulation
of binding partners’ on-and-off ratio.
p
–
p
5.2. Preparation of the ceramide acceptor
a
5.2.1. (2S,3S,4S)-2-Azido-octadecan-1,3,4-triol (2)
p–p
As shown in Scheme 1, to a mixture solution of CH2Cl2 (5 mL)
and H2O (5 mL) containing NaN3 (2.05 g, 31.5 mmol) cooled at
0 °C was added dropwise of Tf2O (1.1 mL, 6.3 mmol) in 20 min.
After addition the resulting mixture was continued stirring for
3 h. The organic layer was separated and the aqueous was ex-
tracted with CH2Cl2 (2 ꢀ 2 mL). The combined organic layer was
washed with saturated Na2CO3 aqueous.
To a suspension of phytosphingosine 1 (1.0 g, 3.1 mmol), K2CO3
(2.18 g, 15.8 mmol) and Cu2SO4 (20 mg) in a mixture of MeOH
(4 mL) and H2O (3 mL) were added the above organic layer which
contained TfN3. More MeOH was added to make the mixture to be
a homogenous solution. The reaction mixture was stirred over-
night at room temperature. The organic solvent was removed in
vacuo. The aqueous was extracted with ethyl acetate. After dried
over anhydrous Na2SO4, it was purified by silica gel flash chroma-
tography (1:1 EtOAc–hexane) to give 0.96 g of product 2 in 89%
yield. 1H NMR (400 MHz, CDCl3): d 4.02 (dd, J = 11.6, 5.2 Hz, 1H),
3.92 (dd, J = 11.6, 4.4 Hz, 1H), 3.82 (m, 1H), 3.78 (m, 1H), 3.69 (m,
1H), 1.64–1.54 (m, 3H), 1.46–1.28 (m, 23H), 0.91 (t, J = 6.8 Hz,
3H); 13C NMR (100 MHz, CDCl3): d 74.8, 72.6, 63.4, 62.0, 32.0,
31.9, 29.6, 29.5, 29.3, 25.7, 22.6, 14.0.
In summary, we developed and synthesized a novel type of gly-
colipid ligands with the help from previous understandings on
structure and activity properties. From the new group analogues,
we further illustrated how the structure of glycolipid ligands, espe-
cially the sugar moiety, would affect the NKT cell immune re-
sponse. The introducing of aromatic group on the C40 position
with suitable linker does not block the NKT TCR recognizing ligand.
Instead, it increased the interaction affinity of ligand with NKT TCR
5.2.2. (2S,3S,4S)-2-Azido-3,4-di-O-benzyl-octadecan-1,3,4-triol (3)
A solution of azido-sphingosine 2 (0.40 g, 1.2 mmol), trityl
chloride (1.3 g, 4.7 mmol) and catalytic amount of DMAP in dry
pyridine (5 mL) was stirred at 50 °C for 10 h. The reaction mixture
was diluted with CH2Cl2, and washed with cooled 1 N HCl, satu-
rated aqueous NaHCO3 and brine. The organic layer was dried
over anhydrous Na2SO4 and concentrated. The residue was then
dissolved in dry DMF (6 mL) and then treated with 60% NaH
(106 mg, 2.6 mmol) for 10 min. BnCl (0.30 mL, 2.6 mmol) was
added by syringe and the reaction was stirred overnight. After
the solvent was removed in vacuo, the residue was dissolved in
water, extracted with ethyl ether. The organic extraction was con-
centrated and then dissolved in a mixture of CH2Cl2 and MeOH
(2:1, 9 mL). pTsOHꢁH2O (95 mg, 0.5 mmol) was added and the
reaction was stirred for 1 h at room temperature. After concen-
trated, the residue was purified by silica gel flash chromatography
(1:6 EtOAc–hexane) to give 0.43 g of acceptor 3 in 68% yield for
three steps. 1H NMR (500 MHz, CDCl3): d 7.37–7.28 (m, 10H),
4.73 (d, J = 11.3 Hz, 1H), 4.69 (d, J = 11.3 Hz, 1H), 4.65 (d,
by p–p stacking and thus changed the profile of cytokine releasing
both in strength and time-dependence. These results provide a
new understanding of NKT response and give some guide on the
future utility of glycosphingolipids in clinical trials.
4. Conclusion
We designed and synthesized a series of aromatic modified
a-
GalCer analogues. The biological activity study proved the immu-
nogenicity of these novel analogues. Our results also showed the
p–p
stacking effects introduced by the aromatic substitution group
altered the immunogenicity profile of -GalCer, which could help
a
to explore the potential therapeutic application of glycolipids.
J = 11.4 Hz, 1H), 4.59 (d,
J =11.4 Hz, 1H), 3.92 (dd, J = 11.7,
5.1 Hz, 1H), 3.82 (dd, J = 11.7, 5.1 Hz, 1H), 3.73 (m, 1H), 3.70 (m,
1H), 3.66 (m, 1H), 1.71 (m, 1H), 1.59 (m, 1H), 1.45 (m, 1H),
1.37–1.27 (m, 23H), 0.91 (t, J = 6.8 Hz, 3H); 13C NMR (125 MHz,
CDCl3): d 138.0, 137.7, 128.6, 128.5, 128.1, 128.03, 128.0, 127.9,
80.5, 79.0, 73.7, 72.6, 63.1, 62.3, 32.0, 30.3, 29.7, 29.68, 29.6,
29.59, 29.4, 25.5, 22.7, 14.1.
5. Experiment and materials
5.1. General information for cell culture and compounds
synthesis
All solvents were dried with solvent-purification system from
Innovative Technology, Inc. All reagents were obtained from com-
mercial sources and used without further purification. Analytical
TLC was carried out on Silica Gel 60 F254 aluminum-backed plates
(E. Merck). The 230–400 mesh size of the same absorbent was uti-
lized for all chromatographic purifications. 1H and 13C NMR spectra
were recorded at the indicated field strengths. The high-resolution
mass spectra were collected at The Ohio State University Campus
Chemical Instrumentation Center.
The A20/CD1cell and hybridoma cells49,50 are maintained in
RPMI-1640 (Gibco, CA, USA) with 10% FBS (Gibco, CA, USA) and
antibiotics. The cell culture was always maintained in a CO2 incu-
bator at 37 °C with 5% CO2. The synthetic glycolipid compounds
were dissolved in DMSO as stock solution and then diluted with
medium to the indicated working concentrations.
Scheme 1. Preparation of the ceramide acceptor.

W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
2773
5.3. Preparation of 4-hydroxyl glycolipid
5.3.1. 2,3-Di-O-benzyl-4,6-O-benzyliden-a-D-galactopyranosyl-
(1,1)-(2S,3S,4R)-2-azido-3,4-di-O-benzyl-octadecan-1,3,4-triol
(5)
As shown in Scheme 2, Powered 4 Å molecule sieves (3.0 g)
were added to a stirred solution of benzylidene protected trichlo-
roacetimide donor 4 (1.25 g, 2.10 mmol) and azido lipid 3 (1.00 g,
1.91 mmol) in fresh dried CH2Cl2 (20 mL). After 30 min, the mix-
J = 11.7 Hz, 1H), 4.71 (d, J = 12.4 Hz, 1H), 4.70 (d, J = 11.5 Hz, 1H),
4.67 (d, J = 11.4 Hz, 1H), 4.61 (d, J = 11.6 Hz, 1H), 4.59 (d,
J = 11.9 Hz, 1H), 4.55 (d, J = 12.2 Hz, 1H), 4.53 (d, J = 11.6 Hz, 1H),
4.10 (m, 1H), 4.08 (dd, J = 10.3, 1.9 Hz, 1H), 3.97 (t, J = 5.9 Hz,
1H), 3.95 (m, 2H), 3.80–3.75 (m, 3H), 3.72 (m, 1H), 3.68–3.64 (m,
2H), 2.67 (br, 1H), 1.70 (m, 1H), 1.58 (m, 1H), 1.44 (m, 1H), 1.36–
1.25 (m, 23H), 0.93 (t, J = 6.7 Hz, 3H); 13C NMR (125 MHz, CDCl3)
d 138.6, 138.5, 138.2, 138.16, 138.0, 128.5, 128.4, 128.37, 128.35,
128.3, 127.9, 127.8, 127.7, 127.68, 127.6, 98.4, 79.4, 79.1, 77.5,
75.7, 73.8, 73.6, 73.2, 72.6, 72.0, 69.6, 68.9, 68.1, 68.1, 62.0, 32.0,
30.0, 29.8, 29.74, 29.72, 29.70, 29.68, 29.65, 29.4, 25.4, 22.7, 14.2;
ture was cooled to ꢂ40 °C, TMSOTf (57
lL, 0.27 mmol) was added
by syringe and the resulting mixture was stirred for 2 h. The reac-
tion was quenched by addition of Et3N (0.2 mL), and the mixture
was filtered through Celite pad. The filtered was concentrated
and purified by column chromatography with hexane/ethyl acetate
(8:1) to give 1.35 g of product in 74% yield. 1H NMR (500 MHz,
CDCl3) d 7.56–7.24 (m, 25H), 5.48 (s, 1H), 5.00 (d, J = 3.4 Hz, 1H),
4.89 (d, J = 11.8 Hz, 1H), 4.84 (d, J = 12.4 Hz, 1H), 4.77 (d,
J = 12.4 Hz, 1H), 4.70 (d, J = 12.1 Hz, 1H), 4.69 (d, J = 11.3 Hz, 1H),
4.64 (d, J = 11.6 Hz, 1H), 4.61 (d, J = 11.8 Hz, 1H), 4.53 (d,
J = 11.6 Hz, 1H), 4.19 (d, J = 3.2 Hz, 1H), 4.13 (dd, J = 10.3, 3.4 Hz,
1H), 4.10 (d, J = 11.5 Hz, 1H), 4.06 (dd, J = 10.4, 3.3 Hz, 1H), 4.03
(d, J = 7.9 Hz, 1H), 3.91 (d, J = 12.0 Hz, 1H), 3.78–3.72 (m, 3H),
3.66 (m, 1H), 3.60 (br, 1H), 1.70 (m, 1H), 1.57 (m, 1H), 1.44 (m,
1H), 1.34–1.25 (m, 23H), 0.92 (t, J = 6.9 Hz, 3H); 13C NMR
(125 MHz, CDCl3) d 138.8, 138.4, 138.1, 137.9, 128.9, 128.4,
128.38, 128.3, 128.2, 128.1, 127.9, 127.8, 127.78, 127.73, 127.70,
127.6, 127.5, 127.48, 126.4, 101.1, 99.2, 79.5, 79.0, 75.9, 75.5,
74.7, 73.8, 73.5, 72.1, 72.0, 69.4, 68.5, 63.0, 61.7, 31.9, 30.1, 29.8,
29.7, 29.68, 29.66, 29.6, 29.4, 25.5, 22.7, 14.1; HRMS calcd for
HRMS calcd for
978.5589.
C
59H77N3O8Na ([M+Na]+) 978.5603, found
5.4. Preparation of Ar1-GSL
5.4.1. 4-O-Benzoyl-2,3,6-tri-O-benzyl-a-D-galactopyranosyl-
(1,1)-(2S,3S,4R)-2-azido-3,4-di-O-benzyl-octadecan-1,3,4-triol
(7)
As shown in Scheme 3, to a solution of alcohol 6 (80 mg,
0.084 mmol), Et3N (19
lL, 0.14 mmol) and catalytic amount of
DMAP in dry CH2Cl2 (2 mL) was added BzCl (14
lL, 0.12 mmol).
The resulting mixture was stirred for 3 h at room temperature
and then diluted with CH2Cl2 (10 mL), washed with brine. The or-
ganic layer was dried over anhydrous Na2SO4, filtered and concen-
trated in vacuo. The residue was purified by column
chromatography with hexane/ethyl acetate (12:1) to give 82 mg
of product in 92% yield. 1H NMR (500 MHz, CDCl3) d 8.06 (d,
J = 7.5 Hz, 2H), 7.61 (t, J = 7.4 Hz, 1H), 7.48 (t, J = 7.7 Hz, 2H),
7.38–7.22 (m, 25H), 5.88 (d, J = 2.1 Hz, 1H), 5.01 (d, J = 3.3 Hz,
1H), 4.89 (d, J = 11.4 Hz, 1H), 4.84 (d, J = 10.0 Hz, 1H), 4.73–4.60
(m, 6H), 4.53 (d, J = 11.4 Hz, 1H), 4.51 (d, J = 11.6 Hz, 1H), 4.41 (d,
J = 11.9 Hz, 1H), 4.24 (t, J = 6.2 Hz, 1H), 4.16 (dd, J = 10.1, 3.1 Hz,
1H), 4.11 (d, J = 9.3 Hz, 1H), 3.97 (dd, J = 10.0, 3.4 Hz, 1H), 3.83–
3.78 (m, 3H), 3.65 (m, 1H), 3.54 (d, J = 6.3 Hz, 2H), 1.70 (m, 1H),
1.58 (m, 1H), 1.44 (m, 1H), 1.37–1.26 (m, 22H), 0.93 (t, J = 6.8 Hz,
3H); 13C NMR (125 MHz, CDCl3) d 165.8, 138.6, 138.5, 138.2,
137.8, 133.0, 130.1, 129.9, 128.38, 128.37, 128.34, 128.25, 128.21,
127.92, 127.91, 127.87, 127.76, 127.70, 127.64, 127.61, 127.5,
127.4, 98.9, 79.4, 79.0, 76.2, 75.2, 73.8, 73.6, 73.5, 72.0, 71.8,
68.8, 68.7, 68.5, 61.9, 32.0, 30.0, 29.8, 29.74, 29.72, 29.70, 29.68,
C
59H75N3O8Na ([M+Na]+) 976.5446, found 976.5467.
5.3.2. 2,3,6-Tri-O-benzyl- -galactopyranosyl-(1,1)-(2S,3S,4R)-
a-D
2-azido-3,4-di-O-benzyl-octadecan-1,3,4-triol (6)
Powered 4 Å molecular sieves (0.5 g) were added to a stirred
solution of compound 5 (395 mg, 0.41 mmol) in freshly dried
CH2Cl2 (5 mL). After 30 min, the mixture was cooled to ꢂ78 °C, Et3-
SiH (0.20 mL, 1.24 mmol) and TfOH (72 lL, 0.82 mmol) were added
by syringes respectively. The reaction mixture was stirred for 2 h at
ꢂ78 °C and then was quenched by addition of solid NaHCO3 (0.5 g)
and MeOH (1 mL). The mixture was allowed to warm to room tem-
perature and filtered through a Celite pad. The solids were washed
three times with CH2Cl2. The combined filtrates were washed with
saturated aqueous NaHCO3 and brine. After dried over anhydrous
Na2SO4, filtered and concentrated, the residue was purified by col-
umn chromatography with hexane/ethyl acetate (5:1) to give
368 mg of product as colorless oil in 94% yield. 1H NMR
(500 MHz, CDCl3) d 7.42–7.25 (m, 25H), 4.94 (d, J = 2.8 Hz, 1H),
4.82 (d, J = 11.7 Hz, 1H), 4.81 (d, J = 12.0 Hz, 1H), 4.75 (d,
29.66, 29.4, 25.4, 22.7, 14.2; HRMS calcd for
C66H81N3O9Na
([M+Na]+) 1082.5865, found 1082.5829.
5.4.2. Ar1-GSL
Triphenylphosphine (12 mg, 0.045 mmol) was added to a solu-
tion of azide 7 (42 mg, 0.039 mmol) in benzene (3 mL) and water
(0.1 mL). The reaction mixture was stirred at 50 °C for 10 h. The
Scheme 2. Preparation of 4-hydroxyl glycolipid.

2774
W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
Scheme 3. Preparation of Ar1-GSL.
solvent was evaporated under reduced pressure and azeotroped
with benzene (2 ꢀ 10 mL). Then it was dissolved in dry THF
(2 mL) and treated with cerotic acid (18 mg, 0.045 mmol) and EDCI
(9 mg, 0.047 mmol). After stirring for 10 h at room temperature,
the solvent was evaporated and the residue was partitioned be-
tween CH2Cl2 (10 mL) and water (10 mL). The organic layer was
separated and dried over Na2SO4. After concentrated, the residue
was purified by silica gel flash chromatography (1:3 EtOAc–hex-
ane) to provide 34 mg as foam.
HRMS calcd for
1082.6257.
C
67H83N3O8Na ([M+Na]+) 1082.6229, found
5.5.2. Ar2-GSL
Triphenylphosphine (8 mg, 0.030 mmol) was added to a solu-
tion of azide 8 (20 mg, 0.018 mmol) in benzene (3 mL) and water
(0.1 mL). The reaction mixture was stirred at 50 °C for 10 h. The
solvent was evaporated under reduced pressure and azeotroped
with benzene (2 ꢀ 10 mL). Then it was dissolved in dry THF
(2 mL) and treated with cerotic acid (8 mg, 0.020 mmol) and EDCI
(4 mg, 0.020 mmol). After stirring for 10 h at room temperature,
the solvent was evaporated and the residue was partitioned be-
tween CH2Cl2 (10 mL) and water (10 mL). The organic layer was
separated and dried over Na2SO4. After concentrated, the residue
was purified by silica gel flash chromatography (1:4 EtOAc–hex-
ane) to provide 14 mg of product.
A suspension of the above product (34 mg, 0.024 mmol) and
5 mg of Pd(OH)2 in EtOH/CH2Cl2 (3:1, 3 mL) was shaken for 4 h un-
der H2 atmosphere (40 psi). The catalytic Pd/C was filtered off
through a Celite pad and the filtrate was concentrated to give
18 mg of product in 44% yield as powder. HRMS calcd for
C
57H103NO10Na ([M+Na]+) 984.7474, found 984.7502.
5.5. Preparation of Ar2-GSL
A suspension of the above product (14 mg, 0.01 mmol) and
5 mg of Pd(OH)2 in EtOH/CH2Cl2 (3:1, 2 mL) was shaken for 4 h un-
der H2 atmosphere (40 psi). The catalytic Pd/C was filtered off
through a Celite pad and the filtrate was concentrated to give
6 mg of product as powder in 35% yield for three steps. HRMS calcd
for C58H107NO9Na ([M+Na]+) 984.7838, found 984.7806.
5.5.1. 4-O-Phenylethyl-2,3,6-tri-O-benzyl-a-D-galactopyranosyl-(1,
1)-(2S,3S,4R)-2-azido-3,4-di-O-benzyl-octadecan-1,3,4-triol (8)
As shown in Scheme 4, a solution of alcohol 6 (114 mg,
0.12 mmol) and 2-phenylethyl iodide (0.17 mL, 1.2 mmol) in dry
DMF (2 mL) was treated with Ag2O (220 mg, 0.95 mmol) at 90 °C
for 10 h. After cooled to room temperature, the mixture was di-
luted with ethyl acetate (10 mL), washed twice with water and
then brine. The organic layer was dried over anhydrous Na2SO4, fil-
tered and concentrated in vacuo. The residue was purified by col-
umn chromatography with hexane/ethyl acetate (15:1) to give
24 mg of product in 19% yield and recover 83 mg of compound 6.
5.6. Preparation of Ar3-GSL
5.6.1. 4-O-Phenylpropyl-2,3,6-tri-O-benzyl-a-D-galactopyranosyl-
(1,1)-(2S,3S,4R)-2-azido-3,4-di-O-benzyl-octadecan-1,3,4-triol (9)
As shown in Scheme 5, a solution of alcohol 6 (126 mg,
0.13 mmol) in dry DMF (2 mL) was treated with 60% NaH (8 mg,
Scheme 4. Preparation of Ar2-GSL.

W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
2775
Scheme 5. Preparation of Ar3-GSL.
0.30 mmol) for 30 min. 3-Iodopropylbenzene (30
lL, 0.18 mmol)
CD1 cells were mixed with 50,000 hybridoma cells and co-cultured
was added via syringe. The resulting mixture was stirred overnight.
After diluted with ethyl acetate (10 mL), the mixture was washed
twice with water and then brine. The organic layer was dried over
anhydrous Na2SO4, filtered and concentrated in vacuo. The residue
was purified by column chromatography with hexane/ethyl ace-
tate (15:1) to give 111 mg of product in 79% yield. 1H NMR
(500 MHz, CDCl3) d 7.39–7.17 (m, 30H), 4.94 (d, J = 3.3 Hz, 1H),
4.48 (d, J = 12.1 Hz, 1H), 4.81 (d, J = 12.1 Hz, 1H), 4.74–4.69 (m,
3H), 4.66 (d, J = 11.4 Hz, 1H), 4.61 (d, J = 11.5 Hz, 1H), 4.54 (d,
J = 8.8 Hz, 1H), 4.51 (d, J = 8.5 Hz, 1H), 4.46 (d, J = 11.8 Hz, 1H),
4.06–4.02 (m, 2H), 4.01–3.92 (m, 3H), 3.83 (m, 1H), 3.78–3.75
(m, 3H), 3.70–3.64 (m, 2H), 3.56–3.50 (m, 2H), 2.67 (t, J = 7.7 Hz,
2H), 1.90 (m, 2H), 1.70 (m, 1H), 1.58 (m, 1H), 1.44 (m, 1H), 1.36–
1.26 (m, 23H), 0.93 (t, J = 6.9 Hz, 3H); 13C NMR (125 MHz, CDCl3)
in total volume of 200 lL. The supernatant was collected for IL-2
ELISA after another 24–48 h as described.51
5.8. Plate-bound CD1d presenting assay
The assay was carried out according to previous published pro-
cedure.52 First, each well of the 96-well plate was coated by 0.5
lg
mice CD1d protein (BD Pharmingen, CA, USA) in 100
at 37 °C for 2 h. Glycolipid antigens in gradient concentrations
(1000, 100, 10, and 1 ng/mL in total volume of 100 L) were then
lL pH 7.4 PBS
l
incubated in the coated wells for 24 h. After washing out the com-
pounds, 1.0 ꢀ 105 NKT hybridoma cells was cultured in each trea-
ted well for 24 h. The supernatant was collected and the IL-2
concentration was measured by ELISA as described.
d
142.2, 138.83, 138.80, 138.5, 138.2, 138.0, 128.5, 128.41,
128.36, 128.30, 128.28, 128.25, 127.9, 127.8, 127.71, 127.68,
127.59, 127.50, 127.45, 127.41, 125.68, 98.8, 79.4, 79.0, 78.7,
76.3, 75.8, 73.7, 73.5, 73.3, 72.9, 72.7, 72.1, 69.7, 68.8, 68.5, 62.0,
32.4, 32.0, 31.9, 30.0, 29.8, 29.74, 29.69, 29.67, 29.65, 29.4, 25.5,
22.7, 14.1; HRMS calcd for C68H87N3O8Na ([M+Na]+) 1096.6385,
found 1096.6369.
5.9. In vitro splenocytes stimulation assay
Fresh extracted spleens from 6 week C57BL/6 mice were grin-
ded down to single cell suspension. For each well, 1.0 ꢀ 106 cells
were cultured with glycolipid antigen at different concentrations
as indicated in a 200
lL total volume. The mixtures were cultured
for 72 h at 37 °C before the supernatants were collected. The IFN-
c
5.6.2. Ar3-GSL
and IL-4 concentration in the supernatants was measured by ELISA.
Triphenylphosphine (10 mg, 0.040 mmol) was added to a solu-
tion of azide 9 (35 mg, 0.033 mmol) in benzene (3 mL) and water
(0.1 mL). The reaction mixture was stirred at 50 °C for 10 h. The
solvent was evaporated under reduced pressure and azeotroped
with benzene (2 ꢀ 10 mL). Then it was dissolved in dry THF
(2 mL) and treated with cerotic acid (14 mg, 0.035 mmol) and EDCI
(7 mg, 0.037 mmol). After stirring for 10 h at room temperature,
the solvent was evaporated and the residue was partitioned be-
tween CH2Cl2 (10 mL) and water (10 mL). The organic layer was
separated and dried over Na2SO4. After concentrated, the residue
was purified by silica gel flash chromatography (1:3 EtOAc–hex-
ane) to provide 26 mg of product.
5.10. In vivo cytokine stimulation
All the animal housing and handling procedures were executed
according to the animal usage protocol approved by The Ohio State
University. Male C57BL/6 mice (4–6 weeks, The Jackson Labora-
tory, Maine, USA) were used here. The compounds were diluted
to 10 lg/mL in PBS from the stock solution and 5 lg of each ana-
logue was intravenously injected through the mouse tail vein,
three mice a group. At each time point (Starting point, 2, 4, 6,
and 24 h after injection), blood samples were collected from each
mouse. The serum was separated and the IFN-c and IL-4 concen-
A suspension of the above product (26 mg, 0.018 mmol) and
6 mg of Pd(OH)2 in EtOH/CH2Cl2 (3:1, 3 mL) was shaken for 4 h un-
der H2 atmosphere (40 psi). The catalytic Pd/C was filtered off
through a Celite pad and the filtrate was concentrated to give
14 mg of product as powder in 43% yield for three steps. HRMS
calcd for C59H109NO9Na ([M+Na]+) 998.7995, found 998.8012.
trations in the serum were measured by ELISA.
5.11. Computer modeling and simulation
The crystal structures of both human and mouse CD1D (hCD1d
and mCD1d) proteins presenting a-GalCer to their respective NKT-
TCR proteins were retrieved from the Protein Databank,53 PDB ID#
2PO6 and 3HE6, respectively. The crystal structures were aligned
and a new complex was built using the iNKT-TCR protein from
5.7. NKT hybridoma stimulation assay
As described before,51 100,000 A20/CD1 cells was pulsed over-
night by the glycolipid compounds at different concentrations
(1000 ng mLꢂ1, 100 ng mLꢂ1, 10 ng mLꢂ1 and 1 ng mLꢂ1) in a total
2PO6 and the mCD1d/a-GalCer from 3HE6. This new complex
underwent an energy minimization using Schrödinger’s Macro-
model software (Schrödinger, USA). The environment in which
the minimization was performed involved the OPLS_2000554
volume of 100 lL. After washed with medium, the pulsed A20/

2776
W. Zhang et al. / Bioorg. Med. Chem. 19 (2011) 2767–2776
25. Nieuwenhuis, E. E. S.; Matsumoto, T.; Exley, M.; Schleipman, R. A.; Glickman, J.;
Bailey, D. T.; Corazza, N.; Colgan, S. P.; Onderdonk, A. B.; Blumberg, R. S. Nat.
Med. 2002, 8, 588.
26. Hazlett, L. D.; Li, Q.; Liu, J.; McClellan, S.; Du, W.; Barrett, R. P. J. Immunol. 2007,
179, 1138.
27. Giaccone, G.; Punt Cornelis, J. A.; Ando, Y.; Ruijter, R.; Nishi, N.; Peters, M.; von
Blomberg, B. M. E.; Scheper Rik, J.; van der Vliet, H. J. J.; van den Eertwegh, A. J.
M.; Roelvink, M.; Beijnen, J.; Zwierzina, H.; Pinedo Herbert, M. Clin. Cancer Res.
2002, 8, 3702.
28. Crul, M.; Mathot, R. A. A.; Giaccone, G.; Punt, C. J. A.; Rosing, H.; Hillebrand, M. J.
X.; Ando, Y.; Nishi, N.; Tanaka, H.; Schellens, J. H. M.; Beijnen, J. H. Cancer
Chemother. Pharmacol. 2002, 49, 287.
29. Veldt, B. J.; van der Vliet, H. J. J.; von Blomberg, B. M. E.; van Vlierberghe, H.;
Gerken, G.; Nishi, N.; Hayashi, K.; Scheper, R. J.; de Knegt, R. J.; van den
Eertwegh, A. J. M.; Janssen, H. L. A.; van Nieuwkerk, C. M. J. J. Hepatol. 2007, 47,
356.
30. Parekh, V. V.; Wilson, M. T.; Olivares-Villagomez, D.; Singh, A. K.; Wu, L.; Wang,
C. R.; Joyce, S.; Van Kaer, L. J. Clin. Invest. 2005, 115, 2572.
31. Uldrich, A. P.; Crowe, N. Y.; Kyparissoudis, K.; Pellicci, D. G.; Zhan, Y.; Lew, A.
M.; Bouillet, P.; Strasser, A.; Smyth, M. J.; Godfrey, D. I. J. Immunol. 2005, 175,
3092.
forcefield in a solvent of water with a constant dielectric of 1.0, and
the actual minimization parameters used the Low-memory Broy-
den-Fletcher-Goldfarb-Shanno (LBFGS) method55 with maximum
iterations of 500 converging on the gradient with a threshold of
0.05. The
a-GalCer ligand in mCD1d/a-GalCer/TCR complex was
then modified to form the ligands Ar1-GSL, Ar2-GSL, and Ar3-
GSL, thereby yielding a total of four complexes. An additional en-
ergy minimization using the same parameters were performed in
each of the systems. The mCD1d/Ar2-GSL and Ar3-GSL/NKT com-
plexes then underwent 200 ps of molecular dynamics simulations
in the same save environment with the OPLS_2005 forcefield.
Acknowledgments
This work was supported by National Institutes of Health
R21AI083513 (P.G.W.) and State Key Laboratory of Phytochemistry
& Plant Resources in West China (C.X). The authors thank Dr.
Mitchell Kronenberg (LIAI, La Jolla, CA, USA) for providing us the
A20/CD1cells and hybridoma cell lines.
32. Fujii, S.-i.; Shimizu, K.; Kronenberg, M.; Steinman, R. M. Nat. Immunol. 2002, 3,
867.
33. Nieda, M.; Okai, M.; Tazbirkova, A.; Lin, H.; Yamaura, A.; Ide, K.; Abraham, R.;
Juji, T.; Macfarlane David, J.; Nicol Andrew, J. Blood 2004, 103, 383.
34. Ishikawa, A.; Motohashi, S.; Ishikawa, E.; Fuchida, H.; Higashino, K.; Otsuji, M.;
Iizasa, T.; Nakayama, T.; Taniguchi, M.; Fujisawa, T. Clin. Cancer Res. 2005, 11,
1910.
References and notes
35. Chang, D. H.; Osman, K.; Connolly, J.; Kukreja, A.; Krasovsky, J.; Pack, M.;
Hutchinson, A.; Geller, M.; Liu, N.; Annable, R.; Shay, J.; Kirchhoff, K.; Nishi, N.;
Ando, Y.; Hayashi, K.; Hassoun, H.; Steinman, R. M.; Dhodapkar, M. V. Exp. Med.
2005, 201, 1503.
36. Koch, M.; Stronge, V. S.; Shepherd, 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, 6, 819.
37. Zajonc, D. M.; Cantu, C.; Mattner, J.; Zhou, D.; Savage, P. B.; Bendelac, A.;
Wilson, I. A.; Teyton, L. Nat. Immunol. 2005, 6, 810.
38. 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, 103, 3972.
39. Iijima, H.; Kimura, K.; Sakai, T.; Uchimura, A.; Shimizu, T.; Ueno, H.; Natori, T.;
Koezuka, Y. Bioorg. Med. Chem. 1998, 6, 1905.
40. Zeng, Z. H.; Castano, A. R.; Segelke, B. W.; Stura, E. A.; Peterson, P. A.; Wilson, I.
A. Science 1997, 277, 339.
41. Xia, C.; Zhang, W.; Zhang, Y.; Chen, W.; Nadas, J.; Severin, R.; Woodward, R.;
Wang, B.; Wang, X.; Kronenberg, M.; Wang, P. G. ChemMedChem 2009, 4, 1810.
42. Zhou, X.-T.; Forestier, C.; Goff, R. D.; Li, C.; Teyton, L.; Bendelac, A.; Savage, P. B.
Org. Lett. 2002, 4, 1267.
43. 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.
44. 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.
45. Nadas, J.; Li, C.; Wang, P. G. J. Chem. Inf. Model. 2009, 49, 410.
46. Li, Y.; Girardi, E.; Wang, J.; Yu, E. D.; Painter, G. F.; Kronenberg, M.; Zajonc, D. M.
J. Exp. Med. 2010, 207, 2383.
47. Sinnokrot, M. O.; Sherrill, C. D. J. Am. Chem. Soc. 2004, 126, 7690.
48. Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2005, 1, 415.
49. Sidobre, S.; Naidenko, O. V.; Sim, B.-C.; Gascoigne, N. R. J.; Garcia, K. C.;
Kronenberg, M. J. Immunol. 2002, 169, 1340.
50. Kinjo, Y.; Wu, D.; Kim, G.; Xing, G. W.; Poles, M. A.; Ho, D. D.; Tsuji, M.;
Kawahara, K.; Wong, C. H.; Kronenberg, M. Nature 2005, 434, 520.
51. Zhang, W.; Zheng, X.; Xia, C.; Perali, R. S.; Yao, Q.; Liu, Y.; Zheng, P.; Wang, P. G.
ChemBioChem 2008, 9, 1423.
52. Kinjo, Y.; Tupin, E.; Wu, D.; Fujio, M.; Garcia-Navarro, R.; Benhnia, M. R.-E.-I.;
Zajonc, D. M.; Ben-Menachem, G.; Ainge, G. D.; Painter, G. F.; Khurana, A.;
Hoebe, K.; Behar, S. M.; Beutler, B.; Wilson, I. A.; Tsuji, M.; Sellati, T. J.; Wong,
C.-H.; Kronenberg, M. Nat. Immunol. 2006, 7, 978.
1. Kronenberg, M.; Gapin, L. Nat. Rev. Immunol. 2002, 2, 557.
2. Kronenberg, M. Annu. Rev. Immunol. 2005, 23, 877.
3. Bendelac, A.; Savage, P. B.; Teyton, L. Annu. Rev. Immunol. 2007, 25, 297.
4. Brigl, M.; Brenner, M. B. Annu. Rev. Immunol. 2004, 22, 817.
5. Borg, N. A.; Wun, K. S.; Kjer-Nielsen, L.; Wilce, M. C.; Pellicci, D. G.; Koh, R.;
Besra, G. S.; Bharadwaj, M.; Godfrey, D. I.; McCluskey, J.; Rossjohn, J. Nature
2007, 448, 44.
6. Gadola, S. D.; Zaccai, N. R.; Harlos, K.; Shepherd, D.; Castro-Palomino, J. C.;
Ritter, G.; Schmidt, R. R.; Jones, E. Y.; Cerundolo, V. Nat. Immunol. 2002, 3, 721.
7. Martin, L. H.; Calabi, F.; Lefebvre, F. A.; Bilsland, C. A. G.; Milstein, C. Proc. Natl.
Acad. Sci. U.S.A. 1987, 84, 9189.
8. Martin, L. H.; Calabi, F.; Milstein, C. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 9154.
9. Calabi, F.; Milstein, C. Nature 1986, 323, 540.
10. Calabi, F.; Jarvis, J. M.; Martin, L.; Milstein, C. Eur. J. Immunol. 1989, 19, 285.
11. Natori, T.; Koezuka, Y.; Higa, T. Tetrahedron Lett. 1993, 34, 5591.
12. Natori, T.; Morita, M.; Akimoto, K.; Koezuka, Y. Tetrahedron 1994, 50, 2771.
13. 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.
14. Shimosaka, A. Int. J. Hematol. 2002, 76, 277.
15. Naumov, Y. N.; Bahjat, K. S.; Gausling, R.; Abraham, R.; Exley, M. A.; Koezuka,
Y.; Balk, S. B.; Strominger, J. L.; Clare-Salzer, M.; Wilson, S. B. Proc. Natl. Acad.
Sci. U.S.A. 2001, 98, 13838.
16. Taniguchi, M.; Seino, K.-i.; Nakayama, T. Nat. Immunol. 2003, 4, 1164.
17. Nakagawa, R.; Serizawa, I.; Motoki, K.; Sato, M.; Ueno, H.; Iijima, R.; Nakamura,
H.; Shimosaka, A.; Koezuka, Y. Oncol. Res. 2000, 12, 51.
18. Fuji, N.; Ueda, Y.; Fujiwara, H.; Toh, T.; Yoshimura, T.; Yamagishi, H. Clin. Cancer
Res. 2000, 6, 3380.
19. Kobayashi, E.; Motoki, K.; Uchida, T.; Fukushima, H.; Koezuka, Y. Oncol. Res.
1995, 7, 529.
20. Forestier, C.; Takaki, T.; Molano, A.; Im, J. S.; Baine, I.; Jerud, E. S.; Illarionov, P.;
Ndonye, R.; Howell, A. R.; Santamaria, P.; Besra, G. S.; DiLorenzo, T. P.; Porcelli,
S. A. J. Immunol. 2007, 178, 1415.
21. Chen, Y.-G.; Choisy-Rossi, C.-M.; Holl, T. M.; Chapman, H. D.; Besra, G. S.;
Porcelli, S. A.; Shaffer, D. J.; Roopenian, D.; Wilson, S. B.; Serreze, D. V. J.
Immunol. 2005, 174, 1196.
22. Hong, S.; Wilson, M. T.; Serizawa, I.; Wu, I.; Singh, N.; Naidenko, O. V.; Miura,
T.; Haba, T.; Scherer, D. C.; Wei, J.; Kronenberg, M.; Koezuka, Y.; Van Kaer, L.
Nat. Med. 2001, 7, 1052.
23. Sharif, S.; Arreaza, G. A.; Zucker, P.; Mi, Q.-S.; Sondhi, J.; Naidenko, O. V.;
Kronenberg, M.; Koezuka, Y.; DeLovitch, T. L.; Gombert, J.-M.; Leite-de-Moraes,
M.; Gouarin, C.; Zhu, B.; Hameg, A.; Nakayama, T.; Taniguchi, M.; LePault, F.;
Lehuen, A.; Bach, J.-F.; Herbelin, A. Nat. Med. 2001, 7, 1057.
24. Kakimi, K.; Guidotti, L. G.; Koezuka, Y.; Chisari, F. V. J. Exp. Med. 2000, 192, 921.
53. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.;
Shindyalov, I. N.; Bourne, P. E. Nucleic Acids Res. 2000, 28, 235.
54. Jorgensen, W. L.; Tiradorives, J. J. Am. Chem. Soc. 1988, 110, 1657.
55. Head, J. D.; Zerner, M. C. Chem. Phys. Lett. 1985, 122, 264.