Synthesis and in vitro activities of a spacer-containing glycophospholipid ligand of a lipopolysaccharide receptor involved in endotoxin tolerance

Article abstract of DOI:10.1016/S0968-0896(98)00027-3

A glycophospholipid consisting in a derivative of N,N'-acylated and bisphosphorylated 2,3-dideoxy-2,3-diamino-d-glucose, bearing a 6-aminocaproyl side chain as spacer arm at carbon 6 (PPDm2-B), has been synthesized and its effect on murine macrophages evaluated. The synthesis started from 2,3-diamino-d-glucose, which was best obtained from glucosamine essentially by known procedures, since attempts to use another known precursor (3-nitro-glycoside) led to unexpected results. Selective N-acylation was performed with the hydroxysuccinimide ester of (d)-3-benzyloxymyritic acid followed by esterification of the sole primary hydroxyl function by 6-azidocaproylchloride and phosphorylation of the resulting 1,4-diol by treatment with tetrabenzyl pyrophosphate. Hydrogenation on a Pd on carbon catalyst permitted the isolation of 6-(6-aminohexanoyl)-2,3-dideoxy-2,3-di-[(R)-3-hydroxy-tetradecanamido]-α-d-glucopyranose 1,4-diphosphate (PPDm2-B). In mouse macrophages, PPDm2-B enhanced the lipopolysaccharide (LPS)-dependent secretion of tumor necrosis factor alpha (TNF-α), and inhibited the LPS-induced desensitization of these cells. The data suggest that PPDm2-B interacts in a serum-independent way with an LPS receptor different from CD14, and involved in endotoxin tolerance. Binding studies of a fluorescent derivative of PPDm2-B indicated that the expression of this unknown receptor is down-regulated during in vitro culture of the cells. Owing to its spacer arm, PPDm2-B could thus be a promising tool for future studies of this receptor. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00027-3

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 755±765
Synthesis and In Vitro Activities of a Spacer-containing
Glycophospholipid Ligand of a Lipopolysaccharide Receptor
Involved in Endotoxin Tolerance
Daniel Charon,a Michelle Mondange,a Jean-Francois Pons,b
,
Karine Le Blayb and Richard Chabyb,*
a
Â
Equipe ``Endotoxines'', Centre National de la Recherche Scienti®que, Universite de Paris-Sud, Centre d'Etudes Pharmaceutiques,
92290 ChaÃtenay-Malabry, France
b
 Â
Equipe ``Endotoxines'', Centre National de la Recherche Scienti®que, Universite de Paris-Sud, Faculte des Sciences d'Orsay,
91405 Orsay, France
Received 7 August 1997; accepted 8 December 1997
AbstractÐA glycophospholipid consisting in a derivative of N,N0-acylated and bisphosphorylated 2,3-dideoxy-2,3-
diamino-D-glucose, bearing a 6-aminocaproyl side chain as spacer arm at carbon 6 (PPDm2-B), has been synthesized
and its e€ect on murine macrophages evaluated. The synthesis started from 2,3-diamino-D-glucose, which was best
obtained from glucosamine essentially by known procedures, since attempts to use another known precursor (3-nitro-
glycoside) led to unexpected results. Selective N-acylation was performed with the hydroxysuccinimide ester of (D)-3-
benzyloxymyritic acid followed by esteri®cation of the sole primary hydroxyl function by 6-azidocaproylchloride and
phosphorylation of the resulting 1,4-diol by treatment with tetrabenzyl pyrophosphate. Hydrogenation on a Pd on
carbon catalyst permitted the isolation of 6-(6-aminohexanoyl)-2,3-dideoxy-2,3-di-[(R)-3-hydroxy-tetradecanamido]-a-
D-glucopyranose 1,4-diphosphate (PPDm2-B). In mouse macrophages, PPDm2-B enhanced the lipopolysaccharide
(LPS)-dependent secretion of tumor necrosis factor alpha (TNF-a), and inhibited the LPS-induced desensitization of
these cells. The data suggest that PPDm2-B interacts in a serum-independent way with an LPS receptor di€erent from
CD14, and involved in endotoxin tolerance. Binding studies of a ¯uorescent derivative of PPDm2-B indicated that the
expression of this unknown receptor is down-regulated during in vitro culture of the cells. Owing to its spacer arm,
PPDm2-B could thus be a promising tool for future studies of this receptor. # 1998 Elsevier Science Ltd. All rights
reserved.
Introduction
now generally recognized that these pathophysiologic
e€ects are elicited in the host by various mediators, and
particularly by tumor necrosis factor a (TNF-a).2
However, a development of a tolerant state to the
pathophysiologic e€ects of LPS (endotoxin tolerance)
can be induced by pre-exposure to sublethal doses of
LPS.3 A similar e€ect can also be induced in vitro: pre-
exposure of macrophages to LPS desensitizes the cells.45
Little is known about the mechanism of this desensiti-
zation. In a search for an understanding of the cellular
and molecular basis of LPS e€ects (activation or desen-
sitization) it is reasonable to assume that cellular
responses to LPS require preliminary interactions with
LPS receptors. Among a number of cell surface LPS-
binding sites that have been detected6 CD14 is the most
The endotoxin of Gram-negative bacteria, chemically
characterized as a lipopolysaccharide (LPS), produces
an extraordinary array of pathophysiological disorders
(fever, in¯ammation, disseminated intravascular coagu-
lation, adult respiratory distress syndrome, multiple
organ failure, and even death) in infected hosts.1 It is
Key words: Lipopolysaccharide; receptor; TNF-a; tolerance;
¯uorescent ligand.
*Corresponding author: Mailing address: Equipe ``Endotox-
ines'', ERS-0571 du C.N.R.S., Batiment 432, Universite de
Paris-Sud, 91405 Orsay, France. Tel: 33-169154830; Fax: 33-
169853715; E-mail: rchaby@mailhost.pasteur.fr
0968-0896/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00027-3

756
D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
extensively studied since it has clearly been shown to
play an important role in mediating LPS e€ects,7
although its functional activity requires the presence of
a serum constituent, the LPS-binding protein (LBP).8
Baer18 who attained the 2,3-diaminoglucose structure
upon addition of ammonia on a 3-nitro-2-hexenopyrano-
side
3 synthesized from methyl-a-glucopyranoside
(Scheme 1). Indeed we were able to obtain a mixture of
4,6-O-benzylidene-3-deoxy-3-nitro-a-D-gluco and manno-
pyranoside on a mole scale with 40% yield (3 steps) and
a gluco/manno ratio close to 1 (estimated by NMR).
When this mixture was transformed into the 3-nitro
ole®n upon mesylation,19 although the reaction
appeared rather clean when monitored by TLC, the
workup procedure failed several times to give the
expected compound, and appeared unreliable. From the
dark mixture that was shown by TLC to contain many
slow moving products, we were able to identify three
compounds. Isolation of a 1:1 mixture of 2-O-acetyl-4,6-
O-benzylidene-3-deoxy-3-nitro-a-glucopyranoside and
of the desired compound upon acetylation of the slow
moving mixture established that rehydration of the nitro
ole®n occurred during work up. The known elimination
of acetic acid from 2-O-acetyl-3-O-nitro glycoside under
the conditions used to obtain the acetate likely explains
the isolation of the nitro ole®n. Two cristalline com-
pounds (5 and 6) could be separated from the colored
sirupy complex mixture by column chromatography
(Scheme 2). Both showed in their IR spectrum the
absorbtion of a nitro group. Elemental analysis of the
®rst eluted compound 5 (Rf 0.66, EtOAc:cyclohexane,
However, several recent observations9,10 pointed to the
possible existence of other, still undiscovered, LPS
receptors. In this regard we reported previously that a
signaling LPS receptor, di€erent from CD14, is con-
stitutively present on mouse11 and human12 granulocytic
bone marrow cells (BMC). We also showed that a syn-
thetic glycophospholipid, consisting of N,N0-acylated
and bisphosphorylated 2,3-dideoxy-2,3-diamino-D-glu-
cose (compound SDZ 280746, also termed PPDm2) can
block LPS-induced stimulation of pre-B cells13 and
BMC,14 and inhibits the binding of LPS to the con-
stitutive receptor of BMC.9 Therefore, it appeared
important to identify the receptor which interacts with
this glycophospholipid, and to determine whether this
type of LPS receptor is present on other cells of immuno-
logical interest. This requires the use of a labeled or
reactive derivative of this lipidic ligand. To confer to
this compound the ability of easy coupling to various
molecules of interest (enzymes, ¯uorophores, radio-
labelable residues, photoactivatable reagents, drugs), an
analogue of PPDm2 bearing a spacer arm with amino
functionality (6-aminocaproyl chain) was designed. We
report here on the synthesis and biological evaluation of
the spacer-containing glycophospholipid.
Results and Discussion
Chemistry
Scheme 1. Reagents (a) NaIO4; (b) CH3NO2/MeO ; (c) C6H5-
CH(OMe)2, TsOH; (d) CH3SO2Cl, Et3N (2 equiv.).
The target compound 15 is a highly functionalized deri-
vative of 2,3-diaminoglucose, a monosaccharide whose
N,N0-diacetate was found in lipid A of Rhodospiril-
lacae.15 Syntheses of diacetamido hexoses and hexosides
have been performed using 2-acetamidoglycosides as
starting material: for instance Meyer zu Reckendorf16
described the obtention of 2,3-diacetamidoglucose from
Benzyl 2-acetamido-2-deoxy-a-glucopyranoside follow-
ing a sequence of reactions involving a double inversion
at C3, ®rst to a 3-O-acetylalloside, and then to a 3-azido-
glucoside. Keeping in mind ®rstly that N-acetyl group
might require harsh removal conditions, and secondly
that a successful nucleophilic substitution of N-pro-
tected 2-amino-3-O-sulfonated glucoside (restricted by
neighbouring group participation) is dependent on both
the nucleophile and the protecting group,17 we tried the
displacement of the good leaving 3-O-tri¯ate group on
the -b-allyl 4,6-O-benzylidene-glucosaminide with either
an acetate or a bromide ion: both results were dis-
appointing. Consequently, we decided to obtain 2,3-
diaminoglucose following essentially the method of
Scheme 2. Reagents: (a) NH3-toluene; (b) unexpected decom-
position.

D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
4:6) was consistent with a nitro enopyranoside. 1H
757
NMR indicated that it contains the methoxy and 4,6-
benzylidene groups as well as 2 hydrogens whose systems
agreed for H4 and H1, but unlike the 3-nitro-2-eno-
pyranoside 3, the ole®nic H2 could not be identi®ed.
De®nitive assignment of the structure to the isomeric 2-
nitro-hex-2-enopyranoside 5 resulted from X-ray cris-
tallography, which clearly showed the position and
substitution of the double bond. The IR spectrum of the
compound eluted further displayed, beside the nitro
band, the strong absorbtion of
a
keto group
(1680 cm 1). Its elemental analysis corresponded to
C21H25NO10 and was in agreement with a pic at m/e 469
in its chemically (NH3) ionised mass spectrum. The
structure of the second eluted compound, disaccharide
6, was deduced from extensive NMR studies: the gluco
structure of the 4,6-benzylidene methyl pyranoside resi-
due was deduced from the large values of the H2±H3,
H3±H4, and H4±H5, coupling constants. The position of
the keto group on carbon 4 of the second residue was in
agreement with its primary hydrogens being the only
one to be consistently displaced down®eld upon treat-
ment with trichloroacetylisocyanate and with the A2B
system given by the hydrogens linked to C06 and C05.
The C3±C02 linkage between the two residues was
deduced from direct and long range C-H correlations
which indicate that C3 (d=41) and C2 (d=80) are
respectively linked to a carbon and an heteroatom while
=C02- and =C03H- have respectively d=155 and 124.
The mechanism by which these unexpected compounds
were formed is not clearly understood. However, for-
mation of the adduct 6 can be visualized as an acidic
catalysed addition of the 3-nitro-2-enopyranoside 3
(with loss of hydrogen 40) on the 2-nitro-2-enopyrano-
side 5 (Scheme 3). Subsequent easy hydrolysis of the
enolic benzylidene group and 20±30 elimination of
nitrous acid from the ketonic form would yield the iso-
lated compound 6. Keeping in mind that nitrite treat-
ment of a 3-nitro alkene has been shown19 to give a 2,3-
dinitro alkane leading to the 2-nitro-alkene, and that
isomerisation of 3-methyl-2-nitro butene to 3-methyl-1-
nitro butene has been reported,20 it appeared that the
instability of the 3-nitro alkene 3 initially formed could
lead to the 2-nitro alkene 5, and adduct 6 (Scheme 2).
Scheme 3.
was not found easily reproducible, we preferred for
routine syntheses the method of Guthrie22 (Scheme 4)
who
obtained
methyl
3-azido-2-benzamido-2,3-
dideoxy-a-D-glucopyranoside by opening with sodium
azide the methyl 2,3-benzoylepimino-4,6-O-benzyl-
idene-2,3-dideoxy-a-D-glucopyranoside, easily obtained
from N-benzoyl-glucosamine in ®ve high yielding steps.
Contrary to the original report, the major compound
resulting from the opening of the epimine had not the
gluco but rather the altro con®guration (diaxial open-
ing). However despite the rather low yield (25%), the 3-
azido glucoside which was highly crystalline was easily
recovered from the reaction mixture. Furthermore we
were able to obtain the epimine in one step from methyl
2-benzamido-4,6-O-benzylidene-2-deoxy-a-D-glucopyr-
anoside 8 upon treatment of the latter in an hetero-
geneous mixture with tosylchloride and a solid base
(KOH±K2CO3).23 After quantitative benzoylation the
N-benzoylated epimine 9a was isolated (60%). Succes-
sive catalytic hydrogenation and acid hydrolysis of the
3-azido-glucoside led to 2,3-diamino-glucose hydro-
chloride 12 (72%) whose 1H NMR showed that it
formed in D2O a 6:4 a:b equilibrium. It is noteworthy
that the azido function of the 3-azido-2-benzamido-gluco-
side can be reduced to a deoxy function. Indeed, when
we attempted N-debenzoylation by hydrazinolysis, we
were able to isolate two compounds: the minor one 11a
was a 3-amino-2-benzamido-glucoside while the major
crystalline one 11b was methyl 2-benzamido-4,6-O-
benzylidene-2,3-dideoxy-a-D-ribopyranoside. The same
reaction on the isomeric 2-azido-3-benzamido-altroside
In view of the aforementioned results, it appeared more
convenient to proceed directly to the addition of
ammonia without isolating the 3-nitro ole®n. However,
the resulting 2-amino-3-nitro glucoside had to be pur-
i®ed by recrystallisation since it was observed that the
crude product was a mixture of 2-amino gluco and
manno glycosides. Actually, Michael addition of
nucleophiles on 3-nitro alkene has been shown to give
isomeric mixtures under acidic or neutral conditions.21
All together, the nitromethane synthesis permitted the
obtention of 2,3-diamino-a-methyl glucoside but, as it

758
D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
Scheme 4. Reagents: (a) MeONa (1 equiv), (Bz)2O; (b) MeOH/H+ re¯ux; (c) C6H5-CH (OMe)2/TsOH, DMF, 50 ꢀC; (d) TsCl,
K2CO3-KOH, Tol/DMSO; (e) BzCl, Pyr; (f) NaN3, NH4Cl, DMF, 120 ꢀC; (g) H2, Pd/C, MeOH; (h) 80% AcOH, 100 ꢀC; (i) 5 N HCl,
100 ꢀC; (j) (D)-3-OBn-Myr-O-C4H2NO2, (iPr)2NEt 2 equiv, DMF; (k) N3(CH)5-COCl 1 equiv, Pyr-CHCl3; (l) LDA, TBPP, THF
70 ꢀC; (m) H2, Pd(OH)2/C, MeOH.
gave the corresponding 3-benzamido-4,6-O-benzyl-
idene-2,3-dideoxy-a-D-ribopyranoside 10a and the allal
derivative 10b.
Bis-N,N0-acylation of 2,3-diamino-glucose was per-
formed in good yield using N-3-(R)-benzyloxymyristoyl-
oxysuccinimide.24 It was best obtained upon acid
catalysed benzylation of the stable 3-hydroxymyristoyl-
oxysuccinimide
with
benzyltrichloroacetimidate.25
Indeed, both basic benzylation of 3-hydroxy-myristate
(BnBr-Ag2O) and saponi®cation of the resulting benzyl-
oxy-ester are tedious. The bis-amidate was selectively
6-O-acylated with 6-azidohexanoylchloride (60%) in
cold pyridine and phosphorylation of both the 1- and
4-position was accomplished by reacting the dilithium
salt with tetrabenzylpyrophosphate according to the
method of Watanabe.26 The fully protected target com-
1
pound 14 was characterized by H and 31P NMR spec-
troscopy which showed that the anomeric con®guration
was a(J H1±H2=3 Hz,
J H1±P=6 Hz). Complete
deprotection of 14 and concomitant reduction of the
azido function to a primary amino group (H2±Pd(OH2))
allowed the isolation of the highly insoluble lithium salt
15. Compound 15 (PPDm2-B), and its ¯uorescein-labeled

D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
759
derivative (PPDm2B-FITC), were used in immunologi-
cal studies.
Immunological studies
Because PPDm2-B has been designed as an analogue of
PPDm2, a compound which exhibited in previous stu-
dies anti-LPS activities in bone marrow cells,14 we
examined in a preliminary experiment whether PPDm2-
B can induce the same e€ect on these cells. We found
that PPDm2-B inhibited the LPS-induced expression of
LPS receptors in bone marrow cells as eciently as
PPDm2 (data not shown). However, since a large
majority of the pathophysiological e€ects of LPS are
due to its interaction with macrophages, we restricted
our studies of the immunological activities of PPDm2-B
to its e€ects in this cell type. All further experiments
were thus performed with mouse peritoneal macro-
phages.
Figure 1. In¯uence of pretreatment with PPDm2-B on TNF-a
responses to LPS. Thioglycolate-elicited mouse peritoneal
macrophages (106 cells/well) were pre-incubated (30 min, 37 ꢀC)
in the absence (*) or in the presence of 5 ng/mL (&), or 10 ng/
mL (~) of PPDm2-B. LPS (50 ng/mL) and IFN-g (50 U/mL)
were then added, and the cells were re-incubated for 18 h at
37 ꢀC in the presence of PPDm2-B. The amounts of TNF-a
secreted were then determined. Results expressed as TNF-a
units ‹SD, are interpolated from data measured in seven serial
dilutions of triplicate cell culture supernatants.
PPDm2-B enhances LPS-induced secretion of TNF-ꢀ in
mouse macrophages. One of the most biologically
important e€ect of LPS is its ability to induce the pro-
duction of TNF-a by monocytes and macrophages.2
Therefore, we ®rst used TNF-a secretion by mouse
macrophages as a test to determine if PPDm2-B can
modify an LPS-induced e€ect. Since we were mainly
interested in interactions with yet uncharacterized (non-
CD14) LPS receptors, the experiment was carried out in
the absence of serum. IFN-g (50 U/mL), reported to
enhance LPS-induced responses of macrophages,27,28
was added in the culture medium used during exposure
of the cells to LPS. The results in Figure 1 show that low
doses of PPDm2-B alone did not induce TNF-a secre-
tion. In conjunction with LPS, PPDm2-B did not block,
but rather enhanced LPS activity. This observation can
be explained either by a synergistic e€ect of the two
agents, or by a blocking e€ect of PPDm2-B on a recep-
tor involved in the down-regulation of LPS-induced
TNF-a production.
blocked by PPDm2-B. However, this antagonistic e€ect
of the synthetic glycophospholipid was much lower
when the desensitization was induced by a higher con-
centration of LPS (5 ng/mL), and almost absent with
10 ng/mL of LPS. This observation likely re¯ects a
competitive interaction of LPS and PPDm2-B with the
same receptor. Therefore, the results of Figure 2 sustain
the second hypothesis mentioned above, and suggest
that PPDm2-B is an antagonist of an LPS receptor
involved in down-regulation of LPS-induced TNF-a
secretion, and further desensitization of the cells.
PPDm2-B inhibits LPS-induced desensitization of macro-
phages. The second hypothesis mentioned above postu-
lates the existence of positive and negative pathways
concomitantly triggered by LPS. The latter is reminis-
cent to the mechanism of desensitization responsible for
endotoxin tolerance. Therefore, to evaluate this
hypothesis, we examined the in¯uence of PPDm2-B on
LPS-induced desensitization, as estimated by a reduced
TNF-a production after a second exposure to LPS. A
partial (50%) desensitization of the cells was observed
(Figure 2) after incubation with 2 ng/mL of LPS alone,
and almost complete desensitization occurred with
higher concentrations of LPS (5 and 10 ng/mL). The
results also showed that the desensitization induced by
the lowest dose of LPS (2 ng/mL) was dose-dependently
Binding of PPDm2B-FITC to macrophages. Since com-
pound PPDm2-B ful®lls the conditions for a biological
probe (blocking of at least one of the biological activ-
ities of LPS on macrophages), we then examined its
potential use as a tool for the analysis of LPS recep-
tors. Fluorescein isothiocyanate (FITC) was thus teth-
ered to the amino group of its spacer arm.
Macrophages were incubated with this ¯uorescent
derivative (PPDm2B-FITC) in the presence of increas-
ing concentrations of LPS, used as an unlabeled com-
petitor. Incubations were carried out at 4 ꢀC to avoid
internalization of the ¯uorescent compound. The
results (Figure 3) show that LPS signi®cantly inhibits
the binding of PPDm2B-FITC. In line with the con-

760
D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
Figure 2. In¯uence of PPD-m2-B on LPS-induced desensitiza-
tion of macrophages. Thioglycolate-elicited mouse peritoneal
macrophages from Swiss mice (106 cells/well) were pre-incu-
bated (30 min, 37 ꢀC) with various concentrations of PPDm2-B.
IFN-g (50 U/mL) and LPS (2 ng/mL, *; 5 ng/mL, &; or 10 ng/
mL, ~) were then added, and the cells were re-incubated for
18 h at 37 ꢀC in the presence of PPDm2-B. The cells were then
washed three times (40 min incubations at 37 ꢀC in 0.5 mL CM
between each washing), and exposed again (18 h, 37 ꢀC) to LPS
(5 mg/mL). Results are expressed as desensitization for TNF-a
responses to the second treatment with LPS.
Figure 3. Binding of PPDm2B-FITC to mouse macrophages.
After adhesion (2 h, 37 ꢀC) to 96-well plastic plates, thioglyco-
late-elicited mouse peritoneal macrophages from Swiss mice
(2Â105 cells/well) were (~) or were not (*) cultured in vitro
for 24 h in CM. The cells were then incubated for 1 h at 37 ꢀC in
the presence of various concentrations of unlabeled LPS (used
as a competitor ligand), equilibrated for 15 min at 4 ꢀC, and
exposed to the FITC-labeled ligand (18 h, 4 ꢀC) in the presence
of the inhibitor. After three washings with cold (4 ꢀC) phos-
phate bu€ered saline, ¯uoresence intensities (arbitrary units)
were determined. Results represent the mean‹SD of four
determinations.
clusion of the preceding paragraph, this result con®rms
that PPDm2-B and LPS compete for the same receptor
of mouse macrophages.
Since we observed in previous (unpublished) studies that
responses of macrophages to LPS vary after in vitro
culture of the cells, we wished to determine if a variation
in the cell density of LPS receptors is detectable with
PPDm2B-FITC. We found indeed that the speci®c
(LPS-inhibitable) binding of PPDm2B-FITC was higher
in macrophages cultured at 37 ꢀC for 24 h (51‹15%
inhibition of the binding with 40 mg/mL of LPS) than in
uncultured cells (29‹7% inhibition). This can be taken
as an indication that the expression of the receptor
involved in endotoxin tolerance is down regulated dur-
ing in vitro culture of the macrophages.
tolerance. The detection of this interaction in serum-free
medium can be taken as an indication that the receptor
involved is not the well-characterized and serum-depen-
dent LPS receptor CD14. PPDm2-B can be considered
as an ecient tool for the analysis of this unknown
receptor since a ¯uorescent derivative of this glycophos-
pholipid provided a new information on the modulation
of its expression upon in vitro culture of the macro-
phages. The conjugation of PPDm2-B with photo-
activatable reagents, and its use for photolabeling and
identi®cation of LPS receptors will be reported else-
where.
Conclusion
Experimental
Synthesis
The overall results indicate that the spacer-containing
glycophospholipid PPDm2-B interacts with, and blocks,
an LPS receptor of macrophages involved in down-reg-
ulation of LPS-induced secretion of TNF-a which
probably accounts for the phenomenon of endotoxin
General methods. NMR spectra were determined on
Bruker ACX 200 or ARX 400 spectrometers at ambient
temperature. Chemical shifts (d) are expressed in part

D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
761
per million (ppm) relative to the signal of internal stan-
dards: Me4Si, d 0.0 (1H in CDCl3); d 77.0 (13C in
CDCl3); HDO, d 4.75 (1H in D2O); PPh3, d 0.0 (31P in
CDCl3). Assignments of signals was generally based on
2-D H-H and C-H COSY. Melting points (uncorrected)
were measured in capillaries with an Electrothermal
melting point apparatus. Infrared spectra were recorded
on a Perkin±Elmer 841 spectrophotometer in KBr
pellets. Column chromatography was performed on
Kieselgel 60 (Merck, 230±400 mesh). Thin-layer chroma-
tography (TLC) was performed on aluminum backed
precoated silica gel plates (Merck 60 F254), using detec-
tion with UV light and then charring with alcoholic
10% sulfuric acid. Reactions requiring anhydrous con-
ditions were conducted under a dry argon atmosphere.
Solvent removal was accomplished by a rotary eva-
porator operating at 45 ꢀC (bath) under reduced pres-
sure. Unless speci®ed, all commercially available
reagents were used without further puri®cation.
OMe), 3.45 (dd, J=4, 6.5 Hz, 1H, H2), 3.1 (dd, J=2.3,
6.5 Hz, 1H, H3).
Methyl 3-azido-2-benzamido-4,6-O-benzylidene-2,3-di-
deoxy-ꢀ-D-glucopyranoside (11) and methyl 2-azido-3-
benzamido-4,6-O-benzylidene-2,3-dideoxy-ꢀ-D-altropyr-
anoside (10). Sodium azide treatment of 28 g
(76.3 mmol) of the benzoylepiminoglycoside 9a accord-
ing to the original procedure of Guthrie22 a€orded 7.8 g
(25%) of crystalline 3-azido-2-benzamido-glucoside (11)
melting at 220 ꢀC after crystallisation from isopropanol
and recrystallisation from EtOH (Lit.22 233±235 ꢀC
from isopropanol), and 21.8 g (70%) of the 2-azido
derivative (10) as a clear oil after ¯ash chromatography
eluting with EtOAc:cyclohexane, 6:4 (Rf 0.54, same sol-
vent). (11) 1H NMR (200 MHz, CDCl3) d 7.7±7.2 (m,
10H, Ar), 6.35 (d, J=9.4 Hz, 1H, NH), 5.6 (s, 1H,
PhCH), 4.76 (d, J=3.6 Hz, 1H, H1), 4.5 (dt, J=3.6,
10 Hz, 1H, H2), 4.33 (m , 1H, H6), 3.9 (t, J=10 Hz, 1H,
0
H3), 3.85 (m, 2H, H5, H6 ), 3.7 (t, J=10 Hz, 1H, H4),
Methyl 4,6-O-benzylidene-2,3-dideoxy-2,3-epimino-ꢀ-D-
allopyranoside (9). Tosyl chloride (31.5 g, 165 mmol)
was added to a stirred suspension of powdered potas-
sium hydroxide:potassium carbonate (1:2.5, 130 g) in
toluene:dimethylsulfoxide (1 L, 4:1) containing methyl
2-benzamido-4,6-O-benzylidene-2-deoxy-a-D-glucopyr-
anoside (8) (58 g, 150 mmol).29 After 30 min tetra-
butylammonium hydrogen sulfate (0.5 g) was added and
stirring was continued for 4 h. As TLC (EtOAc:cyclo-
hexane, 6:4) showed the absence of starting material the
mixture was diluted with toluene (1 L), the organic layer
was decanted from insoluble salts, washed with water,
dried (Na2SO4) and concentrated to a syrup crystallising
from chloroform (27 g, 70% in 2 crops), mp 145 ꢀC,
identical with the product obtained from a 3-O-mesyl
3.35 (s, 3H, OMe); Anal. calcd for C21H22N4O5: C, 61.5;
H, 5.4; N,13.65; found: C, 61.4; H, 5.2; N, 12.95. (10) 1H
NMR (200 MHz, CDCl3) d 7.8±7.2 (m, 10H, Ar), 7.15
(d, J=9 Hz, 1H, NH), 5.6 (s, 1H, PhCH), 4.94 (t d,
J=3, 8.7 Hz, 1H, H3), 4.75 (s, 1H, H1), 4.33 (dd, J=4.3,
9.9 Hz, 1H, H6), 4.14 (dd, J=4, 9.7 Hz, 1H, H4), 4.02 (
dt, J=4, 10 Hz, 1H, H5), 4.01 (bd, J=3 Hz, 1H, H2),
3.88 (t, J=9.9 Hz, 1H, H6), 3.35 (s, 3H, OMe). The oily
altroside (10) (1.5 g) was hydrogenated in methanolic
hydrogen chloride (50 mL, 0.0125%) at 1 atm and at
room temperature in the presence of 10% palladium on
charcoal (200 mg). Puri®cation by ¯ash chromatography
(CHCl3:MeOH, 9:1) gave 1 g of 2-amino-3-benzamido-
1
altroside as a yellow oil. H NMR (200 MHz, DMSO-
d6) d 7.8±7.2 (m, 10H, Ar), 5.65 (s, 1H, PhCH), 4.52 (s,
1H, H1), 4.5 (dt, 1H, H3), 4.35±4.05 (m, 3H, H4, H5,
derivative by Richardson30 (mp 141±145 ꢀC); H NMR
1
0
H6), 3.75 (t, J=10 Hz, 1H, H6 ), 3.35 (s, 3H, OMe), 3 (d,
(200 MHz, CDCl3) d 7.6±7.45 (m, 2H, Ar), 7.4±7.3 (m,
3H, Ar), 5.6 (s, 1H, PhCH), 4.9 (bs, 1H, H1), 4.2 (dd,
J=5, 10 Hz, 1H, H6), 4.1 (m, 1H, H5), 3.9 (bd,
0
J=10 Hz, 1H, H4), 3.65 ( t, J=10 Hz, 1H, H6 ), 3.45 (s,
3H, OMe), 2.7 ( bd, 2H, H2, H3).
J=2.5 Hz, 1H, H2). Acetylation of the above amino-
altropyranoside gave the 2-acetamido-3-benzamido
1
derivative as an oil. H NMR (200 MHz, CDCl3) d 7.8±
7.1 (m, 11H, Ar, NHCOPh), 6.4 (d, J=7.5 Hz, 1H,
NHAc), 5.45 (s, 1H, PhCH), 4.82 (t d, J=9, 2.5 Hz, 1H,
H3), 4.65 (s, 1H, H1), 4.3 (bd, J=7.5 Hz, 1H, H2), 4.2
(dd, J=10, 5 Hz, 1H, H6), 3.95 (m, 2H, H4, H5), 3.75 (t,
0
J=10 Hz, 1H, H6 ), 3.4 (s, 3H, OMe), 3.35 (s, 3H,
OMe), 1.95 (s, 3H, CH3CO ).
Methyl 4,6-O-benzylidene-2,3-benzoylepimino-2,3-dideoxy-
ꢀ-D-allopyranoside (9a). Benzoic anhydride (19 g,
84 mmol) was added to a solution of 9 (17.65 g,
67 mmol) in pyridine (230 mL). After 30 min as TLC
(EtOAc:cyclohexane) showed that benzoylation was
complete pyridine was distilled under vacuo, last traces
being removed by codistillation with toluene, and the
residue was crystallized from EtOH-CHCl3 a€ording
22.1 g (89.8%) of 9a as crystals melting at 190±191 ꢀC.
Lit.31,32,33 190±191 ꢀC, 194±195 ꢀC, 194±198 ꢀC. 1H
NMR (200 MHz, CDCl3) d 8.3 (d, J=7.5 Hz, 2H,
=CH-CO2-), 7.7±7.2 (m, 8H, Ar), 5.6 (s, 1H, PhCH), 5
(d, J=4 Hz, 1H, H1), 4.3 (m, 2H, H5, H6), 4 (dd, J=2.3,
2,3-Diamino-2,3-dideoxy-D-glucopyranose (12). The
azido-glucoside 11 (2.4 g) was hydrogenated in metha-
nolic hydrogen chloride (150 mL, 0.0125%) overnight
with Pd(OH)2 on carbon (20%, 500 mg) at room tem-
perature. The catalyst was removed by ®ltration, the
®ltrate was concentrated to dryness, and the residue
taken up in 5 N aqueous hydrogen chloride (100 mL).
After re¯uxing for 2.5 h as TLC (isopropanol: 20%
ammonia: H2O, 7:2:1) showed that hydrolysis was
0
8.4 Hz, 1H, H4), 3.8 (t, J=11.5 Hz, 1H, H6 ), 3.5 (s, 3H,

762
D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
complete (Rf 0.34) the solution was extracted with ether
(3Â100 mL), the aqueous layer was brought to dryness,
the dark residue, triturated with a small volume of
ethanol and dissolved with addition of water. After half
an hour the cloudy solution was cleared of small
amounts of ¯oculent impurities by ®ltration, the sol-
vents were evaporated in vacuo and the dihydrochloride
of the diaminoglucose (12) was recovered by trituration
with ethanol and centrifugation of the o€-white solid
(1.07 g, 68%), mp 195 ꢀC with decomposition. Lit.16:
196 ꢀC. 1H NMR (200 MHz, D2O) d 5.45 (s, 0.71H,
(4:6) 55% of the 2-deoxy-glycoside 10a as a clear oil. Rf
0.44 (same solvent); 1H NMR (200 MHz, CDCl3) d 7.8±
7.2 (m, 11H, Ar, NH-CO), 5.6 (s, 1H, PhCH), 4.9 (m,
1H, H3), 4.8 (dd, J=2, 3 Hz, 1H, H1), 4.3 (dd, J=4.7,
0
10 Hz, 1H, H6), 4 (m, 1H, H5), 3.8 (m, 2H, H6 , H4), 3.45
(s, 3H, OMe), 2.15 (m, 2H, H3e, H3ax) and 20% of 1,5-
anhydro-3-benzamido-4,6-benzylidene-2,3-dideoxy-D-
ribo-hex-1-enitol (10b); mp 183.5 ꢀC; 1H NMR
(200 MHz, CDCl3) d 7.8±7.2 (m, 10H, Ar), 6.45 (d,
J=6 Hz, 1H, H1), 6.35 (bd, J=6 Hz, 1H, NH-CO), 5.7
(s, 1H, PhCH), 5.2, (t, J=6 Hz, 1H, H2), 4.8 (q,
J=6 Hz, 1H, H3), 4.5 (dd, J=3.4, 9.2 Hz, 1H, H6), 4.1
H1a), 4.95 (d, J=7.3 Hz, 0.28H, H1b), 4±3.3 (m, 6H).
0
(dd, J=6.2, 9.8 Hz, 1H, H4), 3.9 (m, 2H, H5, H6 ); Anal.
Anal. calcd for C6H16Cl2N2O4,1 H2O: C, 26.8; H, 6.7;
N, 10.4. Found: C, 26.9; H, 6.9; N, 10.4. N-acetylation
in 50% aqueous methanol with acetic anhydride in pre-
sence of NaHCO3 gave 2,3-diacetamido-2,3-dideoxy-
glucose as a 4:3 a/b equilibrium. 1H NMR (200 MHz,
D2O) d 5.21 (d, J=3.4 Hz, H1a), 4.84 (d, J=8.7 Hz,
calcd for C20H19NO4: C, 71.2; H, 5.6; N, 4.15. Found:
C, 71.1; H, 5.8; N, 4.1.
Methyl 4,6-O-benzylidene-2,3-dideoxy-2-nitro-ꢀ-D-erythro-
hex-2-enopyranoside (5) and methyl 4,6-O-benzylidene-
2,3-dideoxy-3-(methyl-2,3-dideoxy-4-keto-ꢀ-D-glycero-
hex-2-enopyranosid-2-yl)-2-nitro-ꢀ-D-erythro-hex-2-eno-
pyranoside (6). Methanesulfonyl chloride (40 mL,
517 mol) and triethylamine (140 mL, 1 mol) were added
dropwise, under an atmosphere of argon to a cooled
( 50 ꢀC) solution of methyl 4,6-O- benzylidene-3-nitro-
a-D-gluco and manno-hexopyranoside (2) (155 g,
500 mmol) in dry dichloromethane (500 mL). After
30 min, when addition was terminated, the cooling bath
was removed and the reaction mixture was let to reach
room temperature (1 h). Since TLC showed that the
great majority of organic material was the expected 3-
nitro ole®n, the mixture was diluted with toluene
(500 mL) and percolated through a pad of silica gel
(12Â8 cm) rinsed with EtOAc:cyclohexane (1:1). Con-
centration of the e‚uent gave a dark-brownish syrup
whose TLC indicated that decomposition of the nitro
ole®n gave rise to many tailing spots. Careful chroma-
tography (EtOAc:cyclohexane, 3:7 to 5:5) allowed the
isolation of two crystallising compounds; the ®rst eluted
one, Rf 0.67 (EtOAc:cyclohexane, 4:6) was methyl 4,6-
O-benzylidene-2,3-dideoxy-2-nitro-a-D-erythro-hex-2-
enopyranoside (5) (1.5 g). Isolated as white crystals from
EtOAc- petroleum ether, it had mp 185±190 ꢀC with
H1b), 4.15 (t, J=11 Hz, H3a), 4.0 (dd, J=3, 11 Hz, H2a),
3.7 (dd, J=11, 8.7 Hz, H2b).
Methyl 2-benzamido-4,6-O-benzylidene-2,3-dideoxy-ꢀ-D-
ribopyranoside (11b). The azidoglycoside 11 (200 mg,
0.49 mmol) was heated in anhydrous hydrazine (10 mL)
for 4 h at 100 ꢀC. TLC indicated that starting material
was transformed in a major product having Rf 0.75 and
a minor one with Rf 0.25 (CHCl3:MeOH, 95:5). After
concentration of the mixture in vacuo and column
chromatography (same solvent) 130 mg (72%) of crys-
talline 11b was recovered. Recrystallised from dichloro-
methane-methanol, it had a mp of 275 ꢀC. 1H NMR
(200 MHz, CDCl3) d 7.7±7.25 (m, 10H, Ar), 6.35 (d,
J=9 Hz, 1H, NH), 5.5 (s, 1H, PhCH), 4.65 (d, J=3 Hz,
0
1H, H1), 4.2 (m, 3H, H4, H5, H6 ), 3.35 (s, 3H, OMe),
2.2 (dt, J=11, 3 Hz, 1H, H3e), 1.85 (q, J=11 Hz, 1H,
H3ax). Anal. calcd for C21H23NO5: C, 68.3; H, 6.25; N,
3.8. Found: C, 68.1; H,6.4; N, 3.7. Isolation of the
minor compound yielded 20 mg (10%) of Methyl 3-
amino-2-benzamido-4,6-O-benzylidene-2,3-dideoxy-a-
D-glucopyranoside (11a) as a solid. 1H NMR (200 MHz,
CDCl3) d 7.8±7.25 (m, 10H, Ar), 6.4 (d, J=9 Hz, 1H,
NHCOC6H5), 5.5 (s, 1H, PhCH), 4.75 (d, J=3 Hz, 1H,
1
0
H1), 4.25 (m, 2H, H2, H6), 3.75 (m, 2H, H5, H6 ), 3.35 (s
decomposition. H NMR (200 MHz, CDCl3) d 7.6±7.4
(m, 6H, Ar, C3H=), 5.6 (s, 1H, PhCH), 5.4 (s, 1H, H1),
and t J=9 Hz, 4H, H4, OMe), 3.17 (t, J=9Hz, 1H, H3).
This spectrum was identical to that of the product
obtained by catalytic (Pd/C) hydrogenation of methyl
3-azido-2-benzamido-4,6-O-benzylidene-2,3-dideoxy-a-
D-glucopyranoside (11), mp 250 ꢀC. Anal. calcd for
C21H24N2O5, 1H2O: C, 62.7; H, 6.5; N, 7.0. Found: C,
62.7; H, 6.5; N, 7.0.
4.4 (dd and d, 2H, H6, H4), 4.1 (dt, J=4, 9.9 Hz, 1H,
H5), 3.85 (t, J=9.9 Hz, 1H, H6 ), 3.55 (s, 3H, OMe). 13C
0
NMR (CDCl3) d 147.1 (C2), 136.6 (C arom. subst.),
134.13 (C3), 129.5 (C arom. p), 128.4, 126.2 (C arom. o,
m ), 102.7 (Ph-CH), 94.1 (C1), 74.5 (C4), 69.0 (C6), 63.0
(C5), 57.2 (OMe). IR nmax 1532 cm 1. Anal. calcd for
C14H15NO6: C, 57.3; H, 5.1; N, 4.8. Found: C, 57.1; H,
5.3; N, 4.6. The compound eluted next was methyl 4,6-
O-benzylidene-2,3-dideoxy-3-(methyl-2,3-dideoxy-4-keto-
a-D-glycero-hex-2-enopyranosid-2-yl)-2-nitro-a-D-erythro-
hex-2-enopyranoside (6) (4.5 g), mp 177 ꢀC (from EtOAc),
Rf 0.37 (EtOAc:cyclohexane, 5:5). 1H NMR (400 MHz,
Methyl 3-benzamido-4,6-O-benzylidene-2,3-dideoxy-ꢀ-D-
ribopyranoside (10a). Similar treatment with hydrazine
of methyl 2-azido-3-benzamido-4,6-O-benzylidene-2,3-
dideoxy-a-D-altropyranoside (10) gave after chromato-
graphy on a column eluted with EtOAc:cyclohexane

D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
763
0
CDCl3) d 7.4±7.3 (m, 5H, Ar), 6.0 (s,1H, H3 ), 5.5 (s, 1H,
4.2 (dd, J=12, 5.5 Hz, 1H, H6), 3.8 (m, 2H, 2ÂCH-
OBn), 3.25 (t, J=6.6 Hz, 1H, CH2N3), 2.35 (m, 6H,
3ÂCH2-CO-), 1.75±1.15 (m, 46 H, 23Â-CH2 ), 0.8 (t,
J=6.3 Hz, 6 H, 2ÂCH3 ).
0
PhCH), 5.28 (d, J=4 Hz, 1H, H1), 5.19 (s, 1H, H1 ), 4.83
0
(dd, J=4, 11.5 Hz, 1H, H2), 4.4 (t, J=4.5Hz, 1H, H5 ),
0
4.32 (dd, J=4.8, 10.5 Hz, 1H, H6a), 3.95 (m, 3H, H5, H6 a
0
6 b), 3.75 (t, J=10.5 Hz, 1H, H6b), 3.71 (t,
,
H
J=11.4 Hz,1H, H3), 3.56 (t, J=11 Hz, 1H, H4), 3.55 (s,
3H, OMe), 3.45 (s, 3H, OMe). 13C NMR (CDCl3) d 195
6-(6-Azidohexanoyl)-2,3-di-[(R)-3-benzyloxytetradecan-
amido]-4-O-dibenzylphosphoryl-2,3-dideoxy-ꢀ-D-glucopyr-
anosyl-dibenzylphosphate (14). A 2 M solution of lithium
diisopropylamide in heptane (0.75 mL, 1.5 mmol) was
added, under an atmosphere of argon to a cooled
( 70 ꢀC) solution of the dried azido derivative 13a
(500 mg, 5.3 mmol) in tetrahydrofuran freshly distilled
from sodium benzophenone ketyl. After 10 min, 10 mL
of anhydrous tetrahydrofuran containing tetrabenzyl-
pyrophosphate (1 g) was added, after one more hour the
cooling bath was discarded and the reaction mixture
was abandoned overnight at room temperature. The
heterogeneous white mixture was diluted with chloroform
(20 mL) and ®ltered through a pad of silica gel (1Â3 cm).
The ®ltrate was concentrated to give an oil (1g). The
product was puri®ed by chromatography using EtOAc:-
cyclohexane (1:1.5) as eluent. The title compound
(450 mg) was a colorless syrup. 1H NMR (400 MHz,
CDCl3) d 7.4±7.2 (m, 30H, Ar), 6.97 (d, J=7.8Hz, 1H,
C3NHCO), 6.76 (d, J=7.6Hz, 1H, C2NHCO), 5.76 (dd,
J=3, 5.6 Hz, 1H, H1), 4.97 (m, 7 H, P-O-CH-Ph), 4.75
(dd, J=7.8, 11.5 Hz, 1H, P-O-CH-Ph), 4.55 (m, 3H, -C-
O-CH2-Ph, H3), 4.39 (t, J=11 Hz, 2H, C-O-CH2-Ph),
4.31 (m 1H, H4), 4.24 (dd, J=4, 10.4 Hz, 1H, H6), 4.03
0
(C=O, C4 ), 155 (=C-, C2 ), 136.5 (C arom. subst.), 129.3,
0
0
128.4, 125.8 (C arom. p, o, m), 124.3 (=CH, C3 ), 101.8
0
(Ph-CH), 96.9 (C1 ), 96.6 (C1), 82.8 (CNO2, C2), 79.7 (C4),
0
0
73.1 (C5 ), 68.6 (C6), 63.8 (C5), 61.5 (C6 ), 57.2 (OMe), 57
(OMe), 40.7 (C3) IR nmax: NO2, 1540 cm 1, C=O,
1680 cm 1; MS (CI, NH3) m/z 452 (M+H+), 469
(M+NH4+). Anal. calcd for C21H25NO10: C, 55.9; H, 5.5;
N, 3.1. Found: C, 55.9; H, 5.7; N, 3.0.
2,3-di-[(R)-3-Benzyloxytetradecanamido]-2,3-dideoxy-D-
glucose (13). Diisopropylamine (2.2 mL, 20 mmol) was
added to a stirred solution of 2,3-diamino-glucose di-
hydrochloride (1 g, 4 mmol) in DMF (70 mL) containing
N-(R)-3-benzyloxytetradecanoyloxy-succinimide (6.3 g,
14.5 mmol). After 48 h at room temperature the mixture
was evaporated under vacuo (1 mm Hg) to a residue
which gave 1.4 g (43%) of the title compound as a waxy
solid upon column chromatography with CHCl3:MeOH
(95:5) as eluent. Washed with petroleum ether it had a
mp of 161 ꢀC. 1H NMR (200 MHz, CDCl3:MeOH-d4
3:2) d 7.65 (d, J=8 Hz, 1H, NH), 7.3±7.1 (m, 10 H, Ar),
7.1 (d, J=8 Hz 1H, NH), 4.95 (d, J=3.9 Hz, 1H, H1a),
4.45 (b s, 4H, 2ÂCH2-Ph), 4.1 (t, J=9.9 Hz, 1H, H3),
0
(m, 3H, H2, H5, H6 ), 3.8 and 3.74 (2 quintet, 2H, 2ÂCH-
0
3.9±3.5 (m, H2, H5, H6, H6 , OH), 3.6 (m, 2H, 2ÂCH-
OBn), 3.22 (t, J=7 Hz, 2H, CH2N3), 2.4±2.10 (m, 6H,
3Â-CH2-CO-), 1.75±1.0 (m, 23Â-CH2-), 0.87 (t, J=6Hz,
6H, 2ÂCH3-). 31P NMR (CDCl3) d 3 (q, J=8Hz, P-O-
C4), 5.4 (m, P-O-C1).
OBn), 3.2 (t, J=9.9 Hz, 1H, H4), 2.25 (m, 4H, 2ÂCH2-
CON), 1.4 (bm, 4H, 2Â-CH2-COBn-), 1.15 (bs, 36H,
2Â-(CH2)9-Me), 0.8 (bt, 6 H, 2ÂCH3-). Upon acetyl-
ation (Ac2O-pyridine) H1a was deshielded to d 6.05,
trace of H1b was detected at d 5.4 (J=7.7 Hz) and three
acetoxy groups gave singlets at d 1.8, 1.85, and 2.0.
6-(6-Aminohexanoyl)-2,3-dideoxy-2,3-di-[(R)-3-hydroxy-
tetradecanamido]-ꢀ-D-glucopyranose 1,4-diphosphate
(15). A methanolic solution (100 mL) of the bis dibenzyl-
phosphate 14 (425 mg ) was stirred, in presence of pal-
ladium hydroxide (20%) on carbon (Pearlman's
catalyst) under a slight overpressure of hydrogen. After
2 h at room temperature, as white precipitate started to
deposit, the mixture was diluted with tetrahydrofuran
(100 mL) and brought to a pH of roughly 5 with a 0.004 M
lithium hydroxide solution and hydrogenation was con-
tinued overnight. The catalyst was then removed by ®ltra-
tion and extensively washed with chloroform: isopropanol
(50:50). The ®ltrate and washing were brought to pH 6
(pH paper) and concentrated to give a waxy solid (200 mg)
melting with decomposition at 219 ꢀC. Anal. calcd for
C40H77Li2N3O15P2, 2 H2O: C, 50.5; H, 8.5; N, 4.4; P, 6.5.
Found: C, 50.1; H, 8.6; N, 4.0; P, 6.5.
6-(6-Azidohexanoyl)-2,3-di-[(R)-3-benzyloxytetradecan-
amido]-2,3-dideoxy-D-glucose (13a). An anhydrous
chloroformic solution (1 mL) of 6-azidohexanoylchloride
(116 mg, 0.66 mmol) was added with a seringe to a
cooled ( 50 ꢀC) solution, protected from moisture, of
the diamido derivative 13 (450 mg, 0.55 mmol) in a
mixture of chloroform and pyridine (2:1, 9 mL). The
temperature was allowed to reach slowly 20 ꢀC (2 h) and
the mixture was abandoned overnight. It was then dilu-
ted with chloroform, washed successively with solutions
of sodium hydrogen sulfate, sodium hydrogeno-
carbonate and water. The dried (Na2SO4) organic phase
was concentrated to a glass (500 mg), which was puri®ed
by chromatography (CCl4:acetone, 25:5) to give pure
13a (400 mg, 76%) as a colorless varnish melting at
140 ꢀC. 1H NMR (200 MHz, CDCl3) d 7.3 (m, 10H, Ar),
7.22, 6.75 (2d, 2H, 2ÂNHCO), 5.05 (d, J=3 Hz, 1H,
N-(R)-3-Hydroxytetradecanoyloxysuccinimide. Equimo-
lar amounts of (R)-3-hydroxy myristic acid and hydroxy-
succinimide, in ethyl acetate were treated with
H1a), 4.6, 4.52, 4.42 (d, s, d 1H, 2H, 1H, 2ÂCH2-Ph),

764
D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
dicyclocabodiimide at 0 ꢀC. After 16 h, the dicyclohexyl
urea formed was discarded by ®ltration, the ®ltrate
concentrated and the product crystallised from metha-
nol at 4 ꢀC. (80%) mp 109.5 ꢀC. 1H NMR (200 MHz,
CDCl3-) d 4.05 (m, 1H, CH-O), 2.9±2.5 (m, 6H, CH2-
CO-), 1.7±1 (m, 20H, 10Â-CH2-), 0.8 (t, J=6 Hz, 3H,
CH3). Anal. calcd for C18H31NO5: C, 63.3; H, 9.1; N,
4.1. Found: C, 63.2; H, 9.0; N, 4.0.
and 2-mercaptoethanol (5Â10 5 M). Heparin, actino-
mycin D, crystal violet, and neutral red were purchased
from Sigma Chemical Co. (St. Louis, MO, USA). A
solution of trypsin in EDTA was purchased from Bio-
chrom KG (Berlin, Germany). The lipopolysaccharide
from Salmonella typhimurium LT2 was prepared in our
laboratory, by the phenol-water extraction procedure,
as described previously.34 Thioglycolate broth and
MEM were from Diagnostic Pasteur (Ville d'Avray,
France), and fetal calf serum (FCS) was from Boehrin-
ger (Mannheim, Germany). Mouse recombinant inter-
feron gamma (iFN-g was purchased from Gibco
(Gaithersburg, MD, USA).
N-(R)-3-Benzyloxytetradecanoyloxysuccinimide.
The
hydroxysuccinimide ester of 3-hydroxymyristic acid
(10 g) was suspended in 100 mL of anhydrous toluene
and treated with benzyltrichloroacetimidate (15 g) in the
presence of a catalytic amount of tri¯uoromethane sul-
fonic acid (0.7 mL). After 3 h at room temperature the
mixture containing the product (Rf 0.42 in toluene:
EtOAc, 9:1) was directly puri®ed by chromatography
(same solvent) to give the title compound (6.2 g, 50%)
identical (mp 42 ꢀC) to the product obtained by Macher
upon esteri®cation of the 3-benzyl ether of 3-hydroxy-
myristic acid.24
Cell culture. The murine ®broblast cell line L929-a was
from Dr M. Parant (Paris, France). Suspensions of L929
cells were prepared by treatment (1 min) of con¯uent
cultures (75 cm2) with trypsin (2 mL, 0.05% in 0.02%
EDTA), followed by addition of heat-inactivated horse
serum (1 mL) to stop the reaction. Swiss mice (4±8
weeks old, from R. Janvier, Le Genest Saint-Isle,
France), were used as sources of peritoneal macro-
phages. Peritoneal exudates were harvested ®ve days
after ip injection of 1.7 mL of thioglycolate broth, by
peritoneal washes with MEM containing 10 U/mL of
sodium heparinate. Suspensions of peritoneal exudate
cells (106 neutral red-positive cells/well in RPMI buf-
fered with 25 mM HEPES), were incubated (2 h, 37 ꢀC)
in plates with 24 wells. Nonadherent cells were removed
by four washings with MEM (500 mL/well). The adher-
ent cell population contained 92 to 96% neutral red-
positive macrophages.
6-Azidohexanoylchloride. 6-Bromohexanoic acid (9.8 g,
50 mmol) was heated in 2-methoxyethanol, at 120 ꢀC
with sodium azide (6.5 g, 100 mmol) for 12 h. After
dilution with dichloromethane the solution was washed
with water, aqueous sodium hydrogen sulfate, and
brine. Concentration of the dried organic layer gave an
oil, which was distilled (160 ꢀC, 14 mm Hg) to give 6 g
(64%) of 6-azidohexanoic acid. The azido acid (1.5 g,
9.55 mmol) was treated, at 40 ꢀC, with thionyl chloride
(3.5 mL) for 1 h with a drop of dimethylformamide.
Concentration to a syrup and distillation (130 ꢀC) under
vacuum (14 mm Hg) gave 1.3 g (72%) of the 6-azido-
hexanoylchloride which was stored under argon. 1H
NMR (200 MHz, CDCl3) d 3.15 (t, J=6 Hz, 2H,
CH2N3), 2.8 (t, J=7 Hz, 2H, CH2-CO-), 1.8±1.2 (m, 6H,
3Â-CH2-).
Estimation of TNF-ꢀ secretion and macrophage desensi-
tization. The macrophage culture supernatants were
recovered, centrifuged (10 min, 900 g), and rapidly
assayed (without storage) for TNF-a activity, by deter-
mination of the reciprocal of the dilution that gave 50%
killing of actinomycin-D sensitized L-929 cells. The esti-
mation of the target cell viability was performed by
staining viable cells with crystal violet as described pre-
viously.35 Macrophage desensitization was de®ned as
the percentage of inhibition of LPS-induced production
of TNF-a by macrophages desensitized by a preliminary
treatment with LPS was computed by the formula:
Coupling of ¯uorescein isothiocyanate (FITC). FITC
(50 mL, 1 mg/mL in DMSO) was added to a suspension
of 15 (PPDm2B) in a triethylamine bu€er (200 mL,
55 mM, pH 8.6). After magnetic stirring at room tem-
perature for 150 min, in a plastic tube protected from
light, the mixture was exhaustively dialyzed at 4 ꢀC, ®rst
against PBS, and further against distilled water. TLC
(5:3, isobutyric acid: 1 M ammonium hydroxide)
showed the disappearance of compound 15 (Rf 0.63,
visualized by charring), and the appearance of a single
¯uorescent conjugate (PPDm2B-FITC, Rf 0.71).
% desensitization ˆ ꢀ1 ꢀCxy=Cy0††  100
Where terms of the C xy type represent the measured
mean of the amount of TNF-a produced in response to
y mg/mL of LPS, by macrophages which have been pre-
treated with x ng/mL of the same agent. A computer
program written in our laboratory (and successfully
tested on arti®cial sets of variables) was used to com-
pute the standard deviation (SD) of the values obtained
by the above formula.
Immunology
Media and reagents. Culture medium (CM) was RPMI-
1640 (Gibco, Grand Island, NY, USA) containing 2 mM
L-glutamine, 100 IU/mL penicillin, 100 mg/mL streptomycin,

D. Charon et al./Bioorg. Med. Chem. 6 (1998) 755±765
765
Binding studies. Adherent cells (2Â105 cells per well in
96-well plates) were preincubated for 1 h at 37 ꢀC in
CM, in the presence or absence of unlabeled LPS used
as inhibitor. The cells were then equilibrated for 15 min
at 4 ꢀC, and exposed to the FITC-labeled ligand (18 h,
4 ꢀC) in the presence of the inhibitor. After three wash-
ings with cold (4 ꢀC) phosphate bu€ered saline, the
plates were excited at 485 nm and scanned at 530 nm in
a microplate ¯uorometer (Model 7600, Cambridge
Technology, Watertown, MA, USA), and ¯uorescence
intensities (arbitrary units) were determined.
11. Girard, R.; Pedron, T.; Chaby, R. J. Immunol. 1993, 150,
4504.
12. Pedron, T.; Girard, R.; Turco, S. J.; Chaby, R. J. Biol.
Chem. 1994, 269, 2426.
13. Pedron, T.; Girard, R.; Eustache, J.; Bulusu, A. R. C. M.;
Macher, I.; Radzyner-Vyplel, H.; Stutz, P. L.; Chaby, R.
Internatl. Immunol. 1992, 4, 533.
14. Chaby, R.; Pedron, T.; Stutz, P. L.; Girard, R. J. Immunol.
1993, 151, 4476.
15. Roppel, J.; Mayer, H.; Weckesser, J. Carbohydr. Res. 1975,
40, 31.
16. Meyer zu Reckendorf, W. Chem. Ber. 1969, 102, 4207.
17. Meyer zu Reckendorf, W.; Lenzen, H. J. Liebigs Ann.
Chem. 1982, 265.
Acknowledgements
18. Baer, H. H.; Rajabalee, F. Carbohydr. Res. 1970, 12, 241.
This work was supported by the Centre National de la
Recherche Scienti®que, and by two grants from the
Direction des Recherches Etudes et Techniques (con-
tracts 94/2522A and 96/2537A). We thank Professor M.
Mauguen (Chatenay-Malabry, France) for X-ray ana-
lysis of compound 5.
19. Tachimori, Y.; Sakakibara, T.; Sudoh, R. Carbohydr. Res.
1980, 82, 51.
20. Leseticky, L.; Fidler, V.; Prochazka, M. Collection Cze-
choslov. Chem. Commun. 1973, 38, 459.
21. Sakakibara, T.; Sudoh, R. Carbohydr. Res. 1976, 50, 197.
22. Guthrie, R. D.; Murphy, D. J. Chem. Soc. 1965, 3828.
23. Szeja, W. Carbohydr. Res. 1986, 158, 245.
References
24. Macher, I. Carbohydr. Res. 1987, 162, 79.
25. Wessel, H. P.; Iversen, T.; Bundle, D. R. J. Chem. Soc.
1985, 2247.
1. Parillo, J. E.; Parker, M. M.; Natanson, C.; Su€redini, A.
F.; Danner, R. L.; Cunnion, R. E.; Ognibene, F. P. Ann. Intern.
Med. 1990, 113, 227.
26. Watanabe, Y.; Nakahira, H.; Bunya, M.; Osaki, S. Tetra-
hedron Lett. 1987, 28, 4179.
2. Manthey, C. L.; Vogel, S. N. Rev. Med. Microbiol. 1992, 3,
72.
27. Gi€ord, G. E.; Lohmann-Matthes, M. L. J. Natl. Cancer
Inst. 1987, 78, 121.
3. Johnston, C. A.; Greisman, S. E. Mechanisms of endotoxin
tolerance.; Johnston, C. A.; Greisman, S. E., Eds.; Elsevier
Science: Amsterdam, 1985; Vol. 2, p 359.
28. Haas, J. G.; Meyer, N.; Riethmuller, G.; Ziegler-Heit-
brock, H. W. Eur. J. Immunol. 1990, 20, 1181.
29. Gibbs, C. F.; Hough, L.; Richardson, A. C. Carbohydr.
Res. 1965, 1, 290.
4. Virca, G. D.; Kim, S. Y.; Glaser, K. B.; Ulevitch, R. J. J.
Biol. Chem. 1989, 264, 21951.
30. Ali, Y.; Richardson, A. C.; Gibbs, C. F.; Hough, L. Car-
bohydr. Res. 1968, 7, 255.
5. Fahmi, H.; Chaby, R. Cell. Immunol. 1993, 150, 219.
6. Chaby, R.; Girard, R. Eur. Cytokine Netw. 1993, 4, 399.
31. Buss, D. H.; Hough, L.; Richardson, A. C. J. Chem. Soc.
1963, 5295.
7. Lee, J. D.; Kato, K.; Tobias, P. S.; Kirkland, T. N.; Ule-
vitch, R. J. J. Exp. Med. 1992, 175, 1697.
32. Guthrie, R. D.; Murphy, D. J. Chem. Soc. 1963, 5288.
8. Gegner, J. A.; Ulevitch, R. J.; Tobias, P. S. J. Biol. Chem.
1995, 270, 5320.
33. Christensen, J. E.; Goodman, L. J. Amer. Chem. Soc. 1960,
82, 4738.
9. Girard, R.; Pedron, T.; Chaby, R. Immunology 1997, 91,
391.
34. Westphal, O.; Jann, K. Meth. Carbohydr. Chem. 1965, 5,
83.
10. Bellezzo, J. M.; Britton, R. S.; Bacon, B. R.; Fox, E. S. Am.
J. Physiol. 1996, 270, G959.
35. Lasfargues, A.; Chaby, R. Cell. Immunol. 1988, 115, 165.