Inositolphosphoglycan mediators structurally related to glycosyl phosphatidylinositol anchors: Synthesis, structure and biological activity

Article abstract of DOI:10.1002/1521-3765(20001002)6:19<3608::AID-CHEM3608>3.0.CO;2-Q

The preparation of the pseudopentasaccharide 1a, an inositolphosphoglycan (IPG) that contains the conserved linear structure of glycosyl phosphatidylinositol anchors (GPI anchors), was carried out by using a highly convergent 2+3-block synthesis approach which involves imidate and sulfoxide glycosylation reactions. The preferred solution conformation of this structure was determined by using NMR spectroscopy and molecular dynamics simulations prior to carrying out quantitative structure-activity relationship studies in connection with the insulin signalling process. The ability of 1a to stimulate lipogenesis in rat adipocites as well as to inhibit cAMP dependent protein kinase and to activate pyruvate dehydrogenase phosphatase was investigated. Compound 1a did not show any significant activity, which may be taken as a strong indication that the GPI anchors are not the precursors of the IPG mediators.

Full text of DOI:10.1002/1521-3765(20001002)6:19<3608::AID-CHEM3608>3.0.CO;2-Q

FULL PAPER
Inositolphosphoglycan Mediators Structurally Related to Glycosyl
Phosphatidylinositol Anchors: Synthesis, Structure and Biological Activity
Manuel Martín-Lomas,*[a] Noureddine Khiar,[a] Salud García,[a] Jean-Luc Koessler,[a]
Pedro M. Nieto,[a] and Thomas W. Rademacher[b]
Abstract: The preparation of the pseu-
dopentasaccharide 1a, an inositolphos-
phoglycan (IPG) that contains the con-
served linear structure of glycosyl phos-
phatidylinositol anchors (GPI anchors),
was carried out by using a highly con-
vergent 2‡3-block synthesis approach
which involves imidate and sulfoxide
glycosylation reactions. The preferred
solution conformation of this structure
was determined by using NMR spectros-
copy and molecular dynamics simula-
tions prior to carrying out quantitative
structure ± activity relationship studies
in connection with the insulin signalling
process. The ability of 1a to stimulate
lipogenesis in rat adipocites as well as to
inhibit cAMP dependent protein kinase
and to activate pyruvate dehydrogenase
phosphatase was investigated. Com-
pound 1a did not show any significant
activity, which may be taken as a strong
indication that the GPI anchors are not
the precursors of the IPG mediators.
Keywords: carbohydrates ´ confor-
mation analysis ´ inositols ´ syn-
thesis design ´ transduction
of cellular membranes through a covalent linkage (GPI
anchors).[5] This glycan chain in all protein GPI anchors
presents the conserved linear structure Mana(1 ! 2) Man-
a(1 ! 6) Mana(1 ! 4) GlcNH2a(1 ! 6) myo-Ino (1) so that
the diversity within the anchors is only reflected in the nature
and the location of branching groups.[5] The structural
similarities between GPI anchors and IPG mediators is
supported by immunological evidence, as antibody probes
generated against this GPI glycan chain cross-react with IPG
preparations from rat liver and chicken embryo and com-
pounds with IPG-like activities can be removed by immuno-
precipitation from the culture medium of insulin stimulated
cells.[6±8] These and related data have been taken as an
indication that the GPI anchors could be the precursors of the
IPG mediators by the sequential action of a lipase and a
protease.[2] However, data on the chemical composition of
insulin-sensitive glycolipids[9, 3] and radiomethylation stud-
ies[11, 12] have been interpreted as indicative of the existence of
distinct members of the GPI family, with different structure
fromthe GPI anchors and without anchoring function (free
GPIs), as the real precursors of the intracellular IPG
mediators.[2] In this context, recent reports by Müller and
co-workers have shown that a compound derived from a GPI-
anchored ectoprotein from Saccharomyces cerevisiae, after
protease and phospholipase C cleavage, is able to regulate
cellular glucose and lipid metabolism in an insulin-like fashion
in insulin-resistant cells and tissues.[13±15] The structure as-
signed by Müller to this compound is presented as containing
the basic features of the glycan chain of a GPI anchor in which
the glycosidic linkage between the nonacetylated d-glucos-
Introduction
The initial discovery that insulin causes the cleavage of a
glycolipid to generate diacylglycerol and modulators of the
enzyme cAMP phosphodiesterase[1] has led to the proposal of
a new intracellular signalling systemwhich has been postu-
lated to operate for a number of growth factors, classical
hormones and cytokines.[2] According to this proposal, the
binding of these agonists to their receptors results in the
enzymatic cleavage of some inositol-containing glycolipids to
produce uncharacterised inositol-containing glycans that
modulate the activity of a number of enzymes and mediate
a variety of cellular events.[2] These mediators have been
called inositolphosphoglycans (IPGs) and their precise chem-
ical structures are still unknown. The chemical composition of
these mediators,[3, 4] however, reveals a close structural
similarity with the glycan chain of the glycosyl phosphatidyl-
inositols (GPIs) that serve to attach proteins to the outer face
[a] Prof. Dr. M. Martín-Lomas, Dr. N. Khiar, Dr. S. García,
Dr. J.-L. Koessler, Dr. P. M. Nieto
Grupo de Carbohidratos, Instituto de Investigaciones Químicas
CSIC-UNSE, Americo Vespucio s/n, Isla de la Cartuja
Â
41092 Sevilla (Spain)
Fax : (‡34) 95-446-05-65
E-mail: manuel.martin-lomas@iiq.cartuja.csic.es
[b] Prof. Dr. T. W. Rademacher
Molecular Medicine Unit, Department of Molecular Pathology
University College London Medical School, The Windeyer Building
46 Cleveland Street, London W1P 6DB (UK)
Supporting information for this article is available on the WWW under
http://www.wiley-vch.de/home/chemistry/ or from the author.
3608
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3608 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

3608±3621
amine unit and the myo-inositol moiety posses the b-config-
uration (2).[13] Müller and co-workers have also shown that a
number of synthetic compounds, which consist of the com-
plete structure of this glycan chain and variations thereof,
stimulate lipogenesis by up to 90% of the maximal insulin
response at 20mm concentration.[16] These results, which may
in principle be taken as an indication that the GPI anchors of
membrane proteins are the precursors of the intracellular IPG
mediators, have nevertheless added further uncertainty in this
regard, as the configuration of the glycosidic linkage between
the nonacetylated glucosamine unit and the myo-inositol ring
was shown to be a (1) in all the structures of GPI anchors
investigated so far.[5]
In the context of our previous studies directed towards
establishing the molecular basis of this new intracellular
signalling mechanism,[17, 18] it seemed mandatory to unequiv-
ocally demonstrate at this stage whether a pure sample of the
conserved linear glycan chain of the GPI anchors, prepared by
chemical synthesis and fully characterised, shows any appre-
ciable insulin-like activity. This would greatly contribute to
the disclosure of the nature of the insulin-sensitive glycolipids
either as GPI anchors or as free GPIs. Such an approach
involves the development of an effective synthesis of 1 and the
investigation of its insulin-like activity. Assuming that a
phosphatidylinositol-specific phospholipase C is involved in
the first stage of the GPI-IPG signalling mechanism, the 1,2-
cyclic phosphate 1a was the molecule of choice in the present
investigation. We have, therefore, carried out an effective
synthesis of 1a and have investigated the capacity of this
compound to stimulate lipogenesis in rat adipocites, to inhibit
cAMP-dependent protein kinase (PKA) and to activate
pyruvate dehydrogenase phosphatase (PDH phosphatase).
In addition, we have performed a complete investigation of
the conformational and dynamic properties of 1a in solution
as a prerequisite for further structure ± activity relationship
studies with some key enzymes, such as PKA[19] and glycogen
synthase phosphatase 2CÐconsidered to be an excellent
model for mitochondrial phosphatases such as the PDH
phosphatase[20]Ðwhich possesses a well-established binding
site geometry determined by X-ray structure analysis.[21±23]
Abstract in Spanish: Se ha llevado a cabo la síntesis del
Â
pseudopentasacarido 1aÐun inositolfosfoglicano (IPG) que
contiene la estructura lineal de los glicosil fosfatidilinositoles de
Results and Discussion
Â
anclaje (GPIanchors)Ðsiguiendo un esquema sinte tico [2‡3]
Â
altamente convergente que implica reacciones de glicosilacion
Synthesis: The total synthesis of the diacylglycerol-containing
GPI anchors of Trypanosoma brucei[24, 25] either with the
l-[24, 25a] or d-configuration[25b] of the myo-inositol and rat
brain Thy-1 antigen,[26] and that of the ceramide-containing
GPI anchor fromyeast [27, 28] have been previously reported by
using different synthetic strategies and glycosylation method-
ologies. A number of partial structures have also been
successfully prepared[17, 18, 29±53] and most of this work has
been reviewed.[54] Our synthetic approach to structure 1a is
shown in Scheme 1. It was designed as a convergent and
versatile route to these inositol-containing glycans that may
permit the preparation of other members of the IPG family by
using the appropriate substitution pattern in the correspond-
ing building blocks. The retrosynthetic analysis in Scheme 1
Â
Â
Â
por el metodo del imidato y por el metodo del sulfoxido. Se ha
Â
Â
determinado tambien la conformacion preferida de esta
molecula en solucion utilizando metodos de dinamica mole-
cular y espectroscopía de RMN como un paso necesario para
futuros estudios de relacion estructura-actividad en conexion
con el proceso de senÄalizacion de la insulina. Por ultimo, se ha
estudiado la capacidad del compuesto 1a para estimular la
lipogenesis en adipocitos de rata, inhibir la quinasa depen-
diente de AMP cíclico y activar la fosfatasa de la piruvato
deshidrogenasa. El compuesto 1a no muestra ninguna activi-
dad significativa en estos sistemas lo que indica que los GPIde
anclaje no son los precursores de los segundos menajeros de
tipo IPG.
Â
Â
Â
Â
Â
Â
Â
Â
Â
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3609 $ 17.50+.50/0
3609

M. Martín-Lomas et al.
FULL PAPER
Scheme 1. Retrosynthetic steps from 1a.
shows that the preparation of 1a was envisaged fromdiol 3
after phosphorylation and subsequent deprotection. Diol 3 is
prepared by using a 2‡3-block synthesis approach fromthe
pseudodisaccharide building block 4 and homotrisaccharide
building block 5. The other building blocks involved in the
process include the optically pure myo-inositol derivative 6
(Scheme 2) and the monosaccharide derivatives 7 (Scheme 2),
The preparation of 4 was carried out following the synthetic
strategy previously reported by us[18] as shown in Scheme 2.
The myo-inositol (12) was desymmetrised over the course of a
one-pot reaction as originally reported by Bruzik,[56] and the
optically pure ketal 13 was treated with the bifunctional
protecting agent 1,3-dichloro-1,1,3,3-tetraisopropyldisiloxane
(TIPDSCl2) to give diol 6 (39% overall yield from commer-
cially available 12) which can
be regioselectively glycosylated
at the 6-position.[57] Regio- and
stereoselective glycosylation of
6 is most conveniently per-
formed[18] by using the trichloro-
acetimidate procedure[58] with
2-azido-2-deoxy glycosyl do-
nors that bear protective group
patterns compatible with the
further transformations re-
quired and designed to provide
an acceptable reactivity ± selec-
tivity balance in the subsequent
glycosylation reaction. These
glycosyl donors, such as 7, are
easily accessible from commer-
cially available 2-amino-2-de-
oxy
sugar
hydrochlorides
Scheme 2. a) l-CAMP(OMe)2, H2SO4, DMSO 39%; b) TIPDSCl2, Py, 93%; c) PhSH, BF3 ´ Et2O, CH2Cl2, 74%;
d) NaOMe, MeOH; e) PhCH(OMe)2, TsOH, 92% over the two steps; f) NaH, BnBr, DMF; g) NaCNBH3, HCl,
ET2O; h) Ac2O, NEt3, DMAP, CH2Cl2, 70% over the three steps; i) NBS, H2O, Me2CO, 93%; j) Cl3CCN, K2CO3,
CH2Cl2, 89%; k) 6, TMSOTf, Et2O, 75% (a/b, 3:2); l) NH3, MeOH, quantitative.
through a diazo-transfer reac-
tion fromtriflic azide, previous-
ly reported by us,[59] which af-
fords peracetylated derivatives,
such as 14, in a one-pot manner
8 and 9 (Scheme 4, see below). These last three either have a
phenylthio grouping at the anomeric position (8) or are
obtained froma phenylthioglycoside intermediate ( 7 from 10
and 9 from 11). The synthesis of pseudodisaccharide building
block 4 was based on chemistry developed in our laboratory[18]
while that of homotrisaccharide building block 5 was per-
formed by using the sulfoxide glycosylation reaction.[55]
in three steps. The subsequent transformations of these
peracetates to give glycosyl donors with an appropriate
substitution pattern is most conveniently carried out[18, 60] via
a phenylthioglycoside intermediate, such as 15, owing to the
stability and additional versatility of this function.[60±64] Thus
trichloroacetimidate 7 was readily prepared from 15 by using
the sequence indicated in Scheme 2. Coupling of 6 with 7 in
3610
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3610 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

Inositolphosphoglycan Mediators
3608±3621
diethyl ether at room temperature with trimethylsilyl triflate
(TMSOTf) as a promoter resulted in exclusive 6-O-glycosyl-
ation with good yield (75%), but poor stereoselectivity (a/b
3:2). A better stereochemical outcome (a/b 5.5:1) was
observed when trichloroacetimidate 22 was used as glycosyl
donor, but in this case the regioselectivity decreased and an
appreciable amount (16%) of 5-O-linked pseudodisaccharide
(24) accompanied the desired 6-O-linked regioisomer (23) in
the reaction mixture (Scheme 3). Since the mixture 20a:
20b (Scheme 2) showed a good separation factor and
compound 20b is a convenient building block for the
construction of substances related to the insulin mimetic
structures reported by Müller et al.,[15] the route depicted in
Scheme 2 was finally preferred. The key building block 4 was
prepared by treatment of 20a with liquid ammonia in
The preparation of 5 was carried out as outlined in
Scheme 4 by using the sulfoxide glycosylation reaction.[55]
This method has been successfully used for the synthesis of
glycosides and oligosaccharides both in solution[65] and in the
solid phase.[66] It also appears to be particularly attractive with
regard to its potential application in new chemoselective
glycosylation strategies (such as the ªone-pot sequential
glycosylationº)[67] as previously described for the synthesis
of the cyclamicin O-trisaccharide.[68] These potential advan-
tages and our previous positive experience with this glyco-
sylation methodology[37, 69] encouraged us to explore its scope
and limitations further. The synthesis of 5 was envisaged from
mannose pentaacetate (25) fromwhich both the thioglycoside
glycosyl acceptor 8 and the sulfoxide glycosyl donor 9 were
obtained via the key thioglycoside intermediates 27 and 11,
respectively. To secure the 1,2-trans stereochemistry of the
glycosidic linkage[70] and in order to avoid ortho-ester
formation,[71±73] which is particularly likely to occur when
Â
methanol, as an attempted Zemplen deacetylation of the
latter resulted in simultaneous migration of the silyl protect-
ing group.
Scheme 3. a) NBS, H2O, Me2CO, 90%; b) Cl3CCN, K2CO3, CH2Cl2, 89%; c) 6, TMSOTf, EtO2. 23: 64% (a/b, 3:2) 24: (16%).
Scheme 4. a) PhSH, BF3 ´ Et2O, CH2Cl2, 86%; b) MeONa, MeOH, 94%; c) TBDPSCl, imidazole, THF, 87%; d) NaH, BnBr, DMF, 76%; e) TBAF, THF,
90%; f) HBr, AcOH; g) 2,6-lutidine, CH2Cl2, EtOH; h) NaOMe, MeOH; i) AcOH, Ac2O, Py, 80% over the five steps; j) PhSH, BF3 ´ Et2O, CH2Cl2, 79%;
k) NaOMe, MeOH, quantitative; l) PivCl DMAP, Py, 81%; m) mCPBA, CH2Cl2, 80%; n) 9, Tf2O, DTBMP, Et2O, À608C to rt, 88%; o) MeONa, MeOH,
THF, 608C, 97%; p) 9, Tf2O, DTBMP, Et2O, À608C to rt, 73%; q) mCPBA, CH2Cl2, 90%; r) NBS, H2O, Me2CO, 80%; s) Cl3CCN, DBU, CH2Cl2, 70%.
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3611 $ 17.50+.50/0
3611

M. Martín-Lomas et al.
FULL PAPER
using the sulfoxide glycosylation method,[74] the participating
neighbouring group at C2 in 9 was set up as pivaloate. The
glycosylation reaction of 8 with 9 under diethyl ether with
triflic anhydride as promoter yielded 33 stereoselectively in
88% yield. Removal of the pivaloate group in 33 and
subsequent glycosylation of the resulting disaccharide 34 with
9 in the above conditions yielded 35 with excellent stereo-
selectivity in 73% yield. The thioglycoside group in 35 was
then oxidised to give sulfoxide 5a which can be used directly
as a glycosyl donor in the subsequent glycosylation reaction.
This oxidation, and that previously performed on thioglyco-
side 11 to obtain sulfoxide 9, were carried out with mCPBA in
dichloromethane at low temperature and in both cases
afforded almost exclusively one diastereomer. As a conse-
quence of the higher accessibility of the pro-R lone pair, this
will most likely have the R absolute configuration at the
sulfinyl sulfur, as recently shown by X-ray crystallography.[75]
The complete reaction sequence depicted in Scheme 4 con-
stitutes a further example of the potentiality of the sulfoxide
glycosylation method, it compares well with other glycosyla-
tion procedures and strategies already used to construct the
trimannosidic moiety of the conserved linear structure of the
GPI anchors.[24±28]
indicate the a-configuration of the newly formed glycosidic
linkage. The sequence was then completed by removal of the
TIPDS protection, depivaloylation and perbenzylation to give
39. The l-camphor ketal was then removed and the resulting
[76]
diol (3) reacted with N-pyridiniumphosphodichloridate.
Final hydrogenolysis in the presence of 10% Pd on charcoal in
buffered medium gave pseudopentasaccharide 1a which was
purified by ion exchange chromatography.
Conformation: The three-dimensional solution structure of
the GPI anchor glycan of Trypanosoma brucei VSG was
determined ten years ago by using NMR spectroscopy
together with molecular orbital calculations and restrained
molecular dynamics simulations.[77] Fromthis study it was
concluded that the molecule exists in an extended conforma-
tion that undergoes large torsional oscillations about a(1 ! 6)
linkages, with extreme torsional fluctuations of hydroxyl and
free hydroxymethyl groups on the picosecond timescale.
Advances in NMR instrumentation as well as in computa-
tional power, software and protocols during the last ten years
have now allowed the conformational study of the linear
pseudopentasaccharide 1a to be approached froma more
advantageous position. Furthermore, in the present inves-
tigation, the lack of tetragalactopyranosyl branching in
structure 1a, which is present in the glycan chain of the GPI
anchor of the VSG of Trypanosoma brucei, has simplified the
structural problem and permitted experimental evidence to
be obtained that defines the Mana(1 ! 6)Man linkage
dynamics in 1a as a fast 60:40 equilibriumbetween gg and
gt rotamers. In addition, this reinforces the previous data on
the flexibility of the glycan by simultaneous observation of
positive and negative NOEs as a function of the position of
the different protons along the glycan chain.
The glycosylation of 4 with 5a (Scheme 1) was then
investigated under different experimental conditions since a
stochiometric ratio of 2:1 donor± acceptor, as established for
the sulfoxide glycosylation method is not acceptable when
elaborate donor structures such as 5a are involved. Unfortu-
nately the results were discouraging when no excess of the
donor was used and we turned to the trichloroacetimidate
method for the final glycosylation step (Scheme 5). Thiogly-
coside 35 was transformed in two high-yielding steps into
trichloroacetimidate 5b (Scheme 4). Glycosylation of 4 with
5b (Scheme 5) in methylene chloride at room temperature, in
the presence of TMSOTf as promoter, afforded pseudopen-
tasaccharide 37 with excellent stereoselectivity in 74%
isolated yield. The structure of 37 was confirmed by its
13C NMR spectrum, which showed four signals for anomeric
The solution structure of 1a has now been studied by using
a combination of NMR spectroscopy and molecular mechan-
ics calculations. As a consequence of the expected flexibility
of this structure, the calculation protocol was based on
unrestrained molecular dynamics simulations rather than on
the study of all four single glycosidic linkages by using
adiabatic maps. Molecular mechanics calculations were per-
1
carbons at d ˆ 95.8, 99.2, 99.4 and 99.6 with JCH values
between 171.4 and 175.9 Hz. These data unequivocally
Scheme 5. a) TMSOTf, CH2Cl2, rt, 74%; b) TBAF, THF, 84%; c) CF3CO2H, wet CH2Cl2, 60%; d) NaOMe, MeOH; e) NaH, BnBr, DMF; f) MeOP(O)Cl2,
Py; g) H2, 10% Pd/C, 60% over the two steps.
3612
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3612 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

Inositolphosphoglycan Mediators
3608±3621
formed by means of the Senderovich-Stille AMBER all atoms
reparametrisation for pyranoses[78] optimised for use in
combination with the continuum model for solvent general-
ised Born/solvent accessible surface area (GB/SA).[79] Three
different starting structures, corresponding to the three
possible Mana(1 ! 6)Man rotamers, were considered for the
dynamics simulation. The gt conformer was taken from a
Monte Carlo multiple minimum global search,[80, 81] while the
gg and the tg conformers were manually built from the gt
conformer. A 2.0 ns molecular dynamics simulation was run
for each of these starting structures. The gt conformer was
stable along the whole trajectory, while the gg persisted for
just over one nanosecond, before becoming gt and the tg
conformer flipped over to the gt in a few picoseconds
(Figure 1). This unstable tg rotamer was therefore not
considered in the subsequent investigations. The prediction
of two stable conformers (gt and gg), but not their relative
energies, is in agreement with NMR data.[82]
Figure 2. Instantaneous values of F versus Y during the time course of
molecular dynamics runs for: a) 2-GlcNa(1 ! 6) 1-Ins; b) 3-Mana(1 ! 4)2-
GlcN; c) 4-Mana(1 ! 6)3-Man and d) 5-Mana(1 ! 2)4-Man, for
4-Mana(1 ! 6)3-Man gg conformer, triangles and gt, circles, values.
w
Table 1. F and Y torsional angle mean values and their standard deviation
along the MD trajectories for gg and gt conformers.
Linkage
gg [8]
gt [8]
gg [8]
gt [8]
2-GlcNa(1 ! 5)1-lns F1 À 37 Æ 14 À 38 Æ 16 Y1 À 12 Æ 26 À 14 Æ 29
3-Mana(1 ! 4)2-GlcN F2 À 39 Æ 17 À 37 Æ 20 Y2 À 33 Æ 16 À 34 Æ 15
4-Mana(1 ! 6)3-Man F3 À 147 Æ 13 À 46 Æ 18 Y3 À 143 Æ 40 À 179 Æ 31
5-Mana(1 ! 2)4-Man F4 À 38 Æ 11 À 38 Æ 12 Y4 À 25 Æ 28 À 23 Æ 30
Figure 1. Trajectory plots for 4-Mana(1 ! 6)3-Man linkage, w torsion
angle of the MD simulations of compound 1a for a) tg, b) gt and c) gg
starting structures.
space. These calculations predict a hingelike global movement
for glycan 1a which may be considered to be related to the
structural function of the GPI anchors that act as a bridge
between the anchored protein and the lipid moiety embedded
into the cellular membrane.
The two molecular dynamics runs were in agreement with
an extended structure with the glycosidic linkages fluctuating
within a well-defined area of the conformational space
(Figure 2). All angles resided in a narrow region along the
trajectory, as could be predicted by the exo-anomeric effect,
and the y angles also appeared with a well defined but wider
distribution. As expected, the highest conformational flexi-
bility was predicted for the Mana(1 ! 6)Man connection
(Figure 2c). The torsional angle mean values and their
standard deviations (Table 1) were in agreement with this
extended disposition of the different residues with the
exception of the Mana(1 ! 2)Man structural motif, for which
the values were consistent with a stacked orientation of the
monosaccharide units; this is a consequence of the axial ±
axial arrangement of this glycosidic linkage. The 4000 molec-
ular dynamics structures were grouped by using as criterion
the root mean square distance (rmsd) between the positions of
the heavy atoms. The superposition of representative struc-
tures (Figure 3) shows how the overall contribution of the
individual local fluctuations of the glycosidic linkages causes
the molecule to cover a considerably large conformational
The 1H and 13C NMR spectra of 1a were assigned by using a
combination of standard correlation spectroscopy, HMQC
techniques, and double-pulsed field-gradient spin echo-
(dpfgse) one-dimensional TOCSY and one-dimensional
NOESY. Chemical shift and coupling constant data are given
in the Supporting Information. The observed coupling con-
stants, which indicated a monoconformational behaviour of
the pyranose and cyclitol rings, were in agreement with the
values calculated, with the Altona equation,[83] for the Monte
Carlo minimised minimum-energy conformation. The exper-
imental values for the three-bond 31P-1H couplings were also
in agreement with those calculated by using an equation of the
Karplus type previously reported.[84]
The inter- and intraresidue NOEs for 1a were obtained
through NOESY by using five mixing times between 100 and
500 ms at 283 K and 500 MHz. Three close key contacts not
previously reported[77] have now been detected: 4-Man H1
with 3-Man H6a and H6b, and 4-Man H1 with 5-Man H3.
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3613 $ 17.50+.50/0
3613

M. Martín-Lomas et al.
FULL PAPER
Figure 3. Best fit superimposition of fifty representative structures of the MD runs of a) gg conformer, b) gt conformer and c) both.
Cross-relaxation rates (sNOE) were obtained fromthese values
by extrapolation of [Iij(tm)/Iii(tm)]/tm versus tm to zero mixing
time.[85] The interprotonic distances were evaluated by taking
the nearest H1-H2 distance as a reference to make the
evaluation less sensitive to deviations in the local correlation
times due to segmental motions. These interprotonic distances
were in good agreement with the hrÀ6i average calculated over
the 4000 structures collected during the molecular dynamics
runs (Table 2), thus validating the performed simulations. It
Table 2. Experimental and calculated interprotonic distances [Š] for 1.
Proton pair
Exptl
Calcd
H1 GlcN-2/H6 Ins-1
H1 Man-3/H4 GlnN-2
H1 Man-4/H6b Man-3
H1 Man-4/H6a Man-3
H1 Man-4/H5 Man-5
H1 Man-4/H3 Man-5
H1 Man-5/H2 Man-4
2.36
2.23
2.67
2.20
2.60
3.03
2.30
2.33
2.46
2.83
2.37
2.67
4.59
2.30
Figure 4. Expansion of NOESY spectra of 1a (500 ms) at 298 K showing
the key intraresidue NOE: Dashed lines, negative contours and positive
NOE; filled lines, positive contours and negative NOE.
can therefore be concluded that this unrestrained molecular
dynamics simulation provides a fair description of the solution
structure of glycan 1a.
Although anisotropy can obscure the interpretation of
segmental mobility influence on NOE, in this case, in spite
of the apparent cylindrical shape of the molecules, no
evidence of this behaviour was found as the distances
calculated, by assuming isotropic motion, fit with the pre-
dicted ones. This observation provides experimental evidence
that the faster local movements correspond to the extremes of
the molecule, as could be expected from the predicted
hingelike behaviour of the molecule in the molecular
dynamics simulation.
In summary, the structure 1a may be described as an
extended, relatively flexible structure. This relative flexibility
is the result of the contribution of its four glycosidic linkages,
all of which, with the exception of the w torsion of the 1 ! 6
link, reside in one single global minimum but oscillate within
the potential-energy well. Although in some previous studies
which used more simple but structurally related models, we
characterised an additional local low populated minimum for
the glucosamine-myo inositol a(1 ! 6) glycosidic linkage, this
minimum could not be observed for 1a. This fact, however,
cannot be considered as definitive evidence of a single
The local conformation around the w torsion angle of the
a(1 ! 6) glycosidic linkage could be determined from the
coupling constant values (JH5´H6S ˆ 1.8 Hz and JH5´H6R
ˆ
5.49 Hz) and NOE experimental data (4-Man H1 3-Man
H6R/H6S). The stereochemical assignment of the methylenic
protons was carried out after comparison of the hrÀ6i
molecular dynamics averaged distances with the experimental
distances. A ratio of gg/gt/tg of 60:40:0 was obtained by using
three different sets of equations fromthe experimental
data.[84, 86, 87]
An interesting spectral feature is that a higher temperature
NOESY (298 K) is close to the zero crossing point of the NOE
and weak positive and negative cross peaks were observed
(Figure 4). This behaviour reflects a dispersion of the
correlation times of interprotonic vectors faster and slower
than the NOE-narrowing limit condition (wtc ˆ 1.2). With
regard to the structure, the negative NOEs are related to the
inner protons and the positive ones to the external residues.
3614
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3614 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

Inositolphosphoglycan Mediators
3608±3621
minimum situation for this glycosidic linkage in the case of 1a,
since the exclusive NOE for this conformation may become
undetectable as a result of either an unfavourable correlation
time or a smaller conformational population at the lower
temperature used for the NOESY experiments.
In conclusion, the pioneer work by Homans on the
structure of the GPI anchor glycan of Trypanosoma brucei
VSG[77] clearly established that, as a result of torsional
oscillations around the glycosidic linkages, it was not possible
to define a unique conformation for this molecule and
proposed the use of molecular dynamics as a suitable tool
for the prediction of the conformational behaviour of the
glycan. These have since become generally accepted concepts
in structural carbohydrate chemistry. Our present study on the
conformation of 1a has arrived at the same general conclu-
sions, although the major simplicity of this structure and the
high sophistication of the instrumentation presently available
have now permitted the disclosure of some more detailed
structural features. These include the characterisation of the
dynamics of the a(1 ! 6) glycosidic linkage as a 60:40 fast
equilibriumbetween gg and gt rotamers and the experimental
evidence of the flexibility of the core GPI glycan structure on
the basis of the simultaneous observation of positive and
negative NOE values.
Figure 5. Effects of structure 1a on the enzyme activity of PKA. No
statistically significant effect of 1a was found over the concentration ranges
tested (R ˆ 0.58, p ˆ 0.22).
Finally, structure 1a was not found to have any stimulatory
lipogenic activity in the absence of insulin, whereas natural
IPG A-type substances are able to stimulate lipogenesis under
these conditions (Figure 6). However, a small synergistic
The above structural data on compound 1a and those on
other IPG-like molecules previously carried out in our
laboratory[18, 91] are of major importance in structure ± activi-
ty relationship studies involving these molecules and enzymes
with well-defined binding-site geometries such as PKA[19] and
glycogen synthase phosphatase 2C.[20]
Biological activity: The IPG-like activity of compound 1a was
studied. Since naturally occurring IPGs have been shown to
mimic the lipogenic activity of insulin in adipocites and to be
inhibitors of the enzyme PKA (type A IPGs)[3] or activators of
PDH phosphatase (type P IPGs),[4] the capacity of 1a to
produce these biological effects was investigated.
The assay of the activation of PDH phosphatase by 1a was
performed by using the spectrophotometric variant of the
Figure 6. Effects of structure 1a on basal lipogenesis. The parallel dashed
lines indicate the percentage change frombasal ‡/ À 25d. No statistically
[92]
two-stage systemdescribed by Lilley et al.
Compound 1a
~
*
significant values were found for compound 1a alone ( ). Zinc ( ) on its
was examined at concentrations of 10À4 and 10À6 m. This
activity was also investigated in the presence of Zn2‡ (5 Â
10À5 m) and Mn2‡ (10À5 m) as it has been established that the
presence of these metallic cations influences the activity of
natural IPGs.[10] There was no stimulation of PDH phospha-
tase above the base line found in compound 1a (data not
shown). Type P IPG fromrat liver was simultaneously assayed
as a positive control. (Six microliters of this liver IPG, which is
equivalent to 100 mg liver extracted, gave a value of 103%
above the base line.) It can therefore be concluded that
structure 1a does not behave as an IPG P-type mimetic.
Similar negative results were obtained when the ability of
1a to inhibit the enzyme PKA was determined by using a
colorimetric assay kit and a standard PKA preparation.
Structure 1a was found to be inactive in inhibiting this
enzyme, which is inhibited by A type IPGs, and this
unambiguously establishes that structure 1a is not an IPG
A-type mimetic (Figure 5).
own inhibited basal lipogenesis (15%, 10À4 m). This effect was reversed by
&
addition of 1a ( ) which suggests that 1a was able to chelate the zinc and
prevent the zinc induced inhibition. The concentrations in the figure
&
*
indicate the following: ( ) concentration of 1a ( ) concentration of 1a in
~
the presence of a tenfold molar excess of zinc and ( ) concentration of zinc.
stimulation was observed in the presence of insulin (Figure 7).
This synergistic activity of structure 1a could be effected by a
variety of mechanisms that may involve binding to the surface
of the adipocites, but not to stimulation of key metabolic
control enzymes within the cytoplasm, as proposed for
naturally occurring IPGs.
In summary, synthetic structure 1a, which contains the basic
structural features that may be expected for IPGs derived
fromGPI anchors, does not show any significant IPG-like
activity. This may be taken as a strong indication that the
protein GPI anchors are not the precursors of the IPG
mediators.
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3615 $ 17.50+.50/0
3615

M. Martín-Lomas et al.
FULL PAPER
Compound 20b: 1H NMR (CDCl3, 500 MHz): d ˆ 7.33 ± 7.25 (m, 10H,
arom.), 5.05 (t, J ˆ 9.1 Hz, 1H, H-4b), 4.81 (d, J ˆ 8.1 Hz, 1H, H-1b), 4.70
(d, JCH2Ph ˆ 11.4 Hz, 2H, CH2Ph), 4.40 (d, J ˆ 11.9 Hz, 2H,), 4.20 (t, J ˆ
4.6 Hz, 1H, H-2a), 3.90 (m, 2H, H-1a, H-3a), 3.82 (t, J ˆ 9.2 Hz, 1H, H-6a
or H-4a), 3.71 (dd, J ˆ 6.8, 10.1 Hz, 1H, H-4a or H-6a), 3.50 (m, 5H), 3.40
(t, J ˆ 9.5 Hz, 1H, H-3b), 2.77 (s, 1H, OH), 1.56 (s, 3H, Ac), 1.56 ± 0.92 (m,
38H), 0.79 (s, 3H), 0.74 (s, 3H); 13C NMR (CDCl3, 125 MHz): d ˆ 169.6,
137.9, 137.7, 128.3, 128.2, 127.9, 127.7, 127.6, 117.5, 102.0, 82.0, 80.7, 80.3, 76.6,
75.5, 74.7, 74.3, 74.3, 73.5, 72.9, 72.6, 71.7, 70.2, 67.9, 65.8, 51.5, 45.0, 29.4, 27.0,
25.6, 20.7, 20.5, 20.2, 17.5, 17.4, 17.4, 17.3, 17.2, 17.1, 17.0, 13.0, 12.7, 12.3, 12.0,
9.8. [a]2D3 À 16 (c ˆ 1 in CHCl3); elemental analysis calcd for C50H75N3O12Si2:
C 62.15, H 7.82, N 4.35; found: C 61.87, H 7.85, N 4.63.
O-(2-Azido-3,6-di-O-benzyl-2-deoxy-a-d-glucopyranosyl)-(1 ! 6)-3,4-O-
(1,1,3,3-tetraisopropyldisiloxane-1,3-diyl)-1,2-O-(l-1,7,7-trimethyl[2.2.1]bi-
cyclohept-6-ylidene-d-myo-inositol (4): Compound 20a (0.62 g, 6.1 mol)
was conventionally deacylated by treatment with liquid ammonia in
methanol to give compound 4. 1H NMR (CDCl3, 500 MHz): d ˆ 7.42 ± 7.25
~
*
Figure 7. Effects of structure 1a ( ) alone and zinc ( ) alone on insulin
stimulated lipogenesis. At concentrations of 10À6 and 10À5 m of compound
1a, a synergistic effect was seen with insulin. The stimulatory effect was
(m, 10H, arom.), 5.45 (d, J ˆ 3.5 Hz, 1H, H-1b), 4.90 (AB system, JAB
ˆ
11.2 Hz, Dv ˆ 26.4 Hz, 2H, Bn), 4.57 (AB system, JAB ˆ 12.2 Hz, Dv ˆ
31.6 Hz, 2H, Bn), 4.25 (m, 1H, H-5b), 4.20 (t, J ˆ 5.1 Hz, 1H, H-2a), 4.04
(t, J ˆ 5.2 Hz, 1H, H-1a), 3.90 (m, 3H, H-3a, H-4b, H-6b), 3.65 (dd, J ˆ 4.3,
10.5 Hz, 1H, H-6'b), 3.40 (t, J ˆ 9.6 Hz, H-5a), 3.33 (dd, J ˆ 3.5, 10.2 Hz,
1H, H-2b), 2.50 (s, 1H, OH), 1.96 ± 0.97 (m, 38H), 0.84 (s, 3H), 0.82 (s, 3H);
13C NMR (CDCl3, 125 MHz): d ˆ 138.3, 138.0, 128.6, 128.4, 128.0, 127.9,
127.7, 117.8, 96.1, 79.9, 79.4, 77.3, 77.0, 76.8, 76.3, 75.8, 75.0, 73.5, 72.8, 72.6,
72.3, 69.7, 69.5, 63.0, 51.5, 48.0, 45.2, 45.1, 29.6, 27.1, 20.5, 20.2, 17.6, 17.4, 17.3,
17.2, 17.1, 17.0, 13.0, 12.7, 12.3, 12.0, 9.7. [a]2D3 ‡ 45 (c ˆ 1.0 in CHCl3);
elemental analysis calcd (%) for C48H73N3O11Si2: C 62.36, H 7.98, N 4.50;
found: C 62.70, H 7.53, N 4.03.
&
antagonised by the addition of tenfold molar excess of zinc ( ), this
suggests that compound 1a may be able to complex with zinc. This
complex, however, appears to have no activity and further inhibits the
activity of 1a. The concentrations in the figure indicate the following: (
concentration of compound 1a, ( ) concentration of 1a in the presence of a
tenfold molar excess of zinc and ( ) concentration of zinc.
*
)
&
~
Experimental Section
o-(2-Azido-3-O-benzyl-4,6-O-benzylidene-2-deoxy-a-d-glucopyranosyl)-
(1 ! 6)-3,4-O-(1,1,3,3-tetraisopropyldisiloxane-1,3-diyl)-1,2-O-(l-1,7,7-tri-
methyl[2.2.1]bicyclohept-6-ylidene)-d-myo-inositol (23a), o-(2-azido-3-O-
benzyl-4,6-O-benzylidene-2-deoxy-b-d-glucopyranosyl)-(1 ! 6)-3,4-O-
(1,1,3,3-tetraisopropyldisiloxane-1,3-diyl)-1,2-O-(l-1,7,7-trimethyl[2.2.1]bi-
cyclohept-6-ylidene)-D-myo-inositol (23b) and o-(2-azido-3-O-benzyl-4,6-
O-benzylidene-2-deoxy-a-d-glucopyranosyl)-(1 ! 5)-3,4-O-(1,1,3,3-tetra-
isopropyldisiloxane-1,3-diyl)-1,2-O-(l-1,7,7-trimethyl[2.2.1]bicyclohept-6-
ylidene)-d-myo-inositol (24): TMSOTf (29 mL, 0.16 mmol) at À208C was
added to a solution of 22 (1.698 g, 3.22 mmol) and 6 (2.51 g, 4.51 mmol) in
dry diethyl ether (50 mL). The mixture was slowly warmed to 08C (4.30 h),
then neutralized with solid sodiumbicarbonate, stirred for an additional
30 min and filtered over Celite. After concentration under vacuo, the mixture
was purified by flash column chromatography (hexane/EtOAc 95:5 !
75:25), which afforded three pseudodisaccharides 23a (1.642 g, 1.78 mmol,
55%), 23b (0.287 g, 0.31 mmol, 9%) and 24 (0.473 g, 0.51 mmol, 16%).
General methods: All reactions were run under an atmosphere of dry
argon with oven-dried glassware and freshly distilled and dried solvents.
THF and diethyl ethr were distilled fromsodiumand benzophenone.
Dichloromethane and acetonitrile were distilled from calcium hydride.
Thin layer chromatography (TLC) was performed on silica gel GF254
(Merck) with detection by charring with phosphomolibdic acid/EtOH.
For flash chromatography, silica gel (Merck 230 ± 400 mesh) was used.
Columns were eluted with positive air pressure. Chromatographic eluents
are given as volume to volume ratios (v/v). Routine NMR spectra were
recorded with Bruker Avance DPX300 (1H, 300 MHz), Bruker Avance
DRX400 (1H, 400 MHz) and Bruker Avance DRX500 (1H, 500 MHz)
spectrometers. Chemical shifts are reported in ppm and coupling constants
are reported in Hz. Spectra were referenced to the residual proton or
carbon signals of the solvent. High-resolution mass spectra were recorded
on
a Kratos MS-80RFA 241-MC apparatus. Optical rotations were
determined with a Perkin-Elmer 341 polarimeter. Elemental analyses were
performed by using a Leco CHNS-932 apparatus. The organic extracts were
dried over anhydrous sodiumsulfate and concentrated in vacuo.
Phenyl 2,3,4,6-tetra-O-acetyl-1-thio-a-d-mannopyranoside (26): Thiophe-
nol (5.3 mL), followed by boron trifluoride diethyl ether (2.5 mL), was
added to a solution of d-mannose pentaacetate 25 (20 g) in dry dichloro-
methane (60 mL). After 20 h at room temperature, the solution was diluted
with dichloromethane and was washed twice with saturated aqueous
sodiumbicarbonate solution and then with water. The organic layer was
dried over sodiumsulfate and concentrated under vacuum. The crude
mixture obtained was purified by column chromatography (EtOAc/
Hexane 7:3) to give 26 as a colourless oil (18.9 g, 83%). [a]2D3 ‡ 107 (c ˆ
1.07 in CHCl3); 1H NMR (CDCl3), 500 MHz): d ˆ 7.50 ± 7.20 (m, 5H,
arom.), 5.50 (m, 2H, H-1, H-2), 5.35 (m, 2H, H-3, H-4), 4.55 (m, 1H, H-5),
4.30 (dd, J ˆ 5.8, 12.2 Hz, H-6), 4.00 (dd, J ˆ 2.4, 12.2 Hz, 1H, H-6'), 2.15,
2.08, 2.05, 2.02 (4s, 4Ac); 13C NMR (125 MHz, CDCl3): d ˆ 170.3, 169.7,
169.6, 169.5, 132.4, 131.9, 129.0, 127.9, 85.5, 70.7, 69.3, 69.2, 66.1, 62.2, 20.7,
20.5, 20.6; elemental analysis calcd (%) for C20H24O9S: C 54.54, H 5.49;
found: C 54.23, H 5.41.
o-(4-O-Acetyl-2-azido-3,6-di-O-benzyl-2-deoxy-a and b-d-glucopyrano-
syl)-(1 ! 6)-3,4-O-(1,1,3,3-tetraisopropyldisiloxane-1,3-diyl)-1,2-O-(l-1,7,7-
trimethyl[2.2.1]bicyclohept-6-ylidene)-d-myo-inositol (20a and 20b): A
mixture of 6[18, 56] (471 mg, 0.84 mmol) and 7 (352 mg, 0.61 mmol) in diethyl
ether (20 mL), which contained freshly activated 4 Š molecular sieves
(400 mg), was stirred for 1 h at room temperature. The mixture was then
cooled to À408C and treated with TMSOTf in diethyl ether [185 mL of a
0.1m solution (18.3mm)]. After 30 min, solid sodium carbonate was added
and the mixture was filtered over Celite and concentrated. Flash
chromatography of the residue (hexane/EtOAc 9:1) gave 20a (285 mg,
48%) and 20b (152 mg, 26%).
Compound 20a: 1H NMR (CDCl3, 500 MHz) d ˆ 7.50 ± 7.20 (m, 10H,
arom.), 5.50 (d, J ˆ 3.4 Hz, 1H, H-1b), 5.17 (t, J ˆ 9.7 Hz, 1H, H-4b), 4.70
(d, JCH2Ph ˆ 11.0 Hz, 2H, CH2Ph), 4.60 (d, J ˆ 12.1 Hz, 2H), 4.42 (m, 1H,
H-5b), 4.21(t, J ˆ 4.5 Hz, 1H, H-2a), 4.06 (t, J ˆ 5.9 Hz, 1H, H-3a), 4.01 (t,
J ˆ 9.7 Hz, H-3b), 3.95 (dd, J ˆ 3.95, 9.28 Hz, 1H, H-1a), 3.87 (t, J ˆ 9.2 Hz,
1H, H-6a), 3.80 (dd, J ˆ 6.5, 10.1 Hz, 1H, H-4a), 3.53 ± 3.42 (m, 4H), 2.60 (s,
1H, OH), 1.85 (s, 3H, Ac), 1.97 ± 0.88 (m, 38H), 0.84 (s, 3H), 0.83 (s, 3H);
13C NMR (CDCl3, 125 MHz): d ˆ 169.5, 138.1, 137.9, 128.4, 128.2, 127.9,
127.8, 127.5, 117.9, 95.9, 79.7, 78.0, 76.3, 75.8, 74.6, 73.3, 72.9, 72.4, 71.1, 68.7,
68.6, 63.0, 51.5, 48.0, 45.2, 45.1, 36.6, 29.6, 27.1, 24.7, 23.3, 20.9, 20.5, 20.2,
17.5, 17.4, 17.3, 17.3, 17.2, 17.1, 17.0, 13.0, 12.7, 12.2, 12.0, 9.7; [a]2D3 ‡53 (c ˆ 1,
in CHCl3); elemental analysis calcd (%) for C50H75N3O12Si2: C 62.15, H 7.82,
N 4.35; found: C 62.04, H 8.15, N 4.27.
Phenyl 2,3,4-tri-O-benzyl-1-thio-a-d-mannopyranoside (8): A solution of
26 (18.9 g, 42.8 mmol) in dry methanol was treated with a catalytic amount
of MeONa at room temperature. After 1 h the reaction mixture was
‡
neutralised with Amberlite IR-120 (H ), filtered and evaporated to give
the corresponding tetrol (10.9 g, 94%). A solution of this crude product
(7.9 g, 29 mmol) and imidazol (5.96 g) in THF (60 mL) was treated with
TBDPS (9 mL) at 08C. After 20 h saturated aqueous NH4Cl was added and
the mixture was extracted with CH2Cl2, washed with brine, dried over
sodiumsulfate and concentrated to dryness. The residue was purified by
flash chromatography (MeOH/CH2Cl2 5:95) to give 27 (13.4 g, 87%).
3616
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3616 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

Inositolphosphoglycan Mediators
3608±3621
1H NMR (CDCl3, 500 MHz): d ˆ 7.67 ± 7.16 (m, 5H, arom.), 5.45 (brs, H-1),
4.20 ± 3.80 (m, 6H, H-2, H-3, H-4, H-5, H-6, 6') 1.05 (s, 9H); 13C NMR
(125 MHz, CDCl3): d ˆ 135.7, 135.6, 131.5, 129.9, 129.0, 127.8, 127.4, 88.0,
72.7, 72.2, 72.1, 70.1, 64.9, 27.0.
(30 mL) at À788C. After addition of saturated aqueous NaHCO3 (50 mL),
the organic layer was washed with brine, dried and evaporated to dryness.
The residue was purified by column chromatography (EtOAc/hexane,
1:5 ! 1:2) to give 5.76 g of the less polar isomer, 0.63 g of the more polar
isomer and 1.38 g of the mixture of both sulfoxides (total yield 81%). Data
for the major isomer: 1H NMR (CDCl3, 500 MHz): d ˆ 7.65 ± 7.10 (m, 20H,
arom.), 6.00 (dd, J ˆ 4.6, 2.4 Hz, 1H, H-2), 4.86 (d, J ˆ 10.8 Hz, 1H,
CH2Ph), 4.77 (d, J ˆ 11.1 Hz, 1H, CH2Ph), 4.57 (d, J ˆ 11.6 Hz, 1H,
CH2Ph), 4.54 (d, J ˆ 1.7 Hz, 1H, H-2), 4.52 ± 4.42 (m, 4H, CH2Ph), 4.28 (dd,
J ˆ 9.3, 3.2 Hz, 1H, H-3), 4.18 (m, 1H, H-5), 3.84 (t, J ˆ 9.7 Hz, 1H, H-4),
3.73 (dd, J ˆ 10.7, 1.9 Hz, 1H, H-6), 3.65 (dd, J ˆ 10.7, 5.4 Hz, 1H, H-6'), 1.15
(s, 9H, Piv); elemental analysis calcd (%) for C38H42SO7: C, 71.00; H, 6.58;
S, 4.99. Found: C, 71.05; H, 6.38; S, 4.66.
A solution of the above product (27, 6.7 g, 13.1 mmol.) in THF (30 mL) was
added through a canula to a stirred suspension of sodiumhydride (1.9 g)
and TBAI (catalytic amount) in THF (10 mL) at 08C. After 15 min, benzyl
bromide (5.6 mL) was added and stirring was continued at room temper-
ature for 20 h. The mixture was then diluted with CH2Cl2 and washed
successively with saturated aqueous NH4Cl and brine, dried over sodium
sulfate and evaporated to dryness. The residue was purified by column
chromatography (Et2O/hexane 10:90) to give pure 28 (7.70 g, 76%) as a
colourless oil. 1H NMR (CDCl3, 500 MHz): d ˆ 7.70 (m, 4H, arom.), 7.40 ±
7.20 (m, 26H, arom.), 5.59 (s, 1H, H-1), 4.94 ± 4.45 (m, 6H, Bn), 4.20 ± 3.80
(m, 6H), 1.03 (s, 9H); 13C NMR (CDCl3, 75 MHz): d ˆ 138.9, 138.7, 138.6,
138.5, 136.4, 136.0, 135.5, 134.3, 133.7, 131.3, 130.0, 129.9, 129.3, 128.8, 128.7,
128.4, 128.3, 128.1, 128.0, 127.9, 127.4, 86.0, 80.7, 77.6, 77.3, 77.0, 75.8, 75.1,
74.4, 72.6, 72.4, 63.4, 27.1, 19.7, 16.0.
Phenyl 2,3,4-tri-O-benzyl-6-O-(3,4,6-tri-O-benzyl-2-O-pivaloyl-a-d-man-
nopyranosyl)-1-thio-a-d-mannopyranoside (33): A solution of 9 (1.03 g,
1.6 mmol) in diethyl ether (40 mL) at À788C was treated with triflic
anhydride (0.135 mL, 0.8 mmol). After 5 min, a mixture of 8 (0.43 g,
0.8 mmol) and DTBMP (0.98 g, 4.8 mmol) in diethyl ether (10 mL) was
added dropwise. The reaction mixture was stirred while the temperature
was allowed to rise from À60 to À108C, then hydrolysed by addition of
saturated aqueous NaHCO3, extracted with CH2Cl2 (4 Â 100 mL), dried
over sodiumsulfate and concentrated. The crude mixture was purified by
flash column chromatography (Hexanes/EtOAc, 10:1) to give 33 (601 mg,
71%, 88% on 100 mg scale)[a]2D3‡44.68 (c ˆ 1.09 in CHCl3); 1H NMR
(500 MHz, CDCl3): d ˆ 7.40 ± 7.00 (m, 30H, arom.), 5.55 (d, J ˆ 1.2 Hz,
H-1c), 5.41 (t, J ˆ 2.4 Hz, H-2d), 4.88 (d, J ˆ 1.5, 1H, H-1c), 4.73 (d, J ˆ
11.1 Hz, CH2Ph), 4.65 (d, J ˆ 10.8 Hz, CH2Ph), 4.56 (d, J ˆ 12.3 Hz,
CH2Ph), 4.50 (d, J ˆ 12.1 Hz, CH2Ph), 4.60 (m, 4H, CH2Ph), 4.22 (dd,
J ˆ 4.3, 9.7 Hz, H-5d), 3.99 (t, J ˆ 2.6 Hz, 1H, H-2c), 3.93 (m, 3H, H-3d,
H-4c, H-6c), 3.86 (t, J ˆ 9.6 Hz, 1H, H-4c), 3.83 (dd, J ˆ 9.2, 2.9 Hz, H-3c),
3.79 (ddd, J ˆ 3.7, 9.7, 1.5 Hz, H-5c), 3.92 (m, 2H), 3.62 (dd, J ˆ 1.5, 10.8, Hz,
H-6d), 1.19 (s, 9H, Piv); 13C NMR (125 MHz, CDCl3): d ˆ 177.5, 138.6,
138.5, 138.5, 138.4, 138.3, 138.1, 138.0, 134.8, 131.0, 129.2, 128.8, 128.5, 128.4,
128.3, 128.3, 128.2, 128.1, 128.1, 128.0, 128.0, 127.9, 127.9, 127.8, 127.8, 127.7,
127.6, 127.5, 98.1, 85.5, 80.3, 76.3, 75.2, 75.1, 74.8, 74.2, 73.1, 72.2, 72.0, 71.4,
71.1, 68.9, 68.0, 66.8, 27.2; elemental analysis calcd (%) for C65H70O11S: C
73.70, H 6.66; found: C 73.73, H 6.73.
A 1m solution of TBAF in THF (11.5 mL) was added dropwise to a solution
of 28 (7.0 g, 8.96 mmol) in THF (70 mL) at room temperature. After 7 h, the
mixture was treated with a saturated solution of NH4Cl in water and then
extracted with CH2Cl2. The organic layer was dried over sodiumsulfate and
concentrated to dryness, and the residue purified by column chromatog-
raphy (EtOAc/hexane, 1:4) to afford 8 (4.37 g, 90%) as a solid. [a]2D0 ‡73.3
(c ˆ 1.0 in CHCl3); 1H NMR (CDCl3, 500 MHz): d ˆ 7.39 ± 7.27 (m, 20H,
arom.) 5.50 (s, 1H, H-1) 4.95 (d, J ˆ 10.9 Hz, 1H, CH2Ph), 4.50 (m, 1H,
H-5), 4.03 (t, J ˆ 9.4 Hz, 1H, H-4), 3.99 (dd, J ˆ 1.6, 1.1 Hz, 1H, H-2), 3.88
(dd, J ˆ 9.2, 2.9 Hz, 1H, H-3), 1.88 (dd, J ˆ 6.7, 1.0 Hz, 1H, OH); 13C NMR
(125 MHz, CDCl3): d ˆ 138.5, 138.3, 137.9, 134.1, 131.9, 129.2, 129.1, 128.6,
128.5, 128.4, 128.3, 128.2, 128.1, 128.0, 127.9, 127.8, 127.7, 86.1, 80.2, 76.5,
75.4, 74.9, 74.7, 73.5, 72.7, 72.5, 72.3, 62.2; elemental analyis calcd (%) for
C33H34SO5: C 73.04, H 6.31; found: C 72.75, H 6.20.
1-(Phenylsulfinyl)-3,4,6-tri-O-benzyl-2-O-pivaloyl-1-deoxy-a-d-manno-
pyranoside (9): Thiophenol (6.3 mL, 61.7 mmol) and BF3 ´ Et2O (5.21 mL)
were added to a solution of 30[92] (22 g, 41.15 mmol) in CH2Cl2 (60 mL) at
room temperature. The mixture was stirred for 2 h, then diluted with
CH2Cl2 (400 mL) and poured into water 8500 mL). The organic layer was
washed with a saturated aqueous solution of NaHCO3 (2 Â 200 mL) and
water (300 mL), dried over Na2SO4 and concentrated. The residue was
purified by column chromatography (EtOAc/hexane, 1:4) to afford 32
(18 g, 79% a/b, 35:1) as a colourless oil. 1H NMR (CDCl3, 500 MHz): d ˆ
7.60 ± 7.15 (m, 20H, arom.), 5.80 (dd, J ˆ 3.2, 1.1 Hz, 1H, H-2b), 5.65 (dd,
J ˆ 2.8, 1.6 Hz, 1H, H-2a), 5.55 (d, J ˆ 1.6 Hz, 1H, H-1a), 4.88 (d, J ˆ
1.1 Hz, H-1b), 4.48 ± 4.42 (m, 8H, CH2Ph), 4.35 (m, 1H, H-5a), 4.30 ± 3.65
(m, 5H), 3.55 (m, 1H, H-5b), 2.55 (s, 3H, Ac), 2.15 (s, 3H, Ac); 13C NMR a
anomer (CDCl3, 125 MHz): d ˆ 170.2, 138.2, 138.1, 137.5, 133.6, 131.7, 128.9,
128.4, 128.3, 128.2, 128.1, 127.8, 127.7, 127.6, 127.5, 86.1, 78.4, 75.2, 74.4, 73.3,
72.4, 71.8, 70.3, 68.7, 21.0.
Phenyl 2,3,4-tri-O-benzyl-6-O-(3,4,6-tri-O-benzyl-a-d-mannopyranosyl)-
1-thio-a-d-mannopyranoside (34): A solution of 33 (1.26 g, 1.19 mmol) in
dry methanol (15 mL) and THF (5 mL) was treated with a 1m solution of
MeONa in MeOH (2.5 mL). The reaction mixture was heated to 658C
‡
overnight, then neutralised with Amberlite IR-120 (H ), the resin filtered
off and the solution evaporated. The crude mixture was purified by flash
column chromatography (EtOAc/hexane, 1:1.5) to give pure 34 (1.12 g,
97%) as a colourless oil. [a]D20 ‡608 (c ˆ 0.7 in CHCl3); 1H NMR (500 MHz,
CDCl3): d ˆ 7.44 ± 7.13 (m, 30H, arom.), 5.57 (d, J ˆ 1.5 Hz, 1H, H-1c), 5.04
(d, J ˆ 1.5 Hz, H-1d), 4.75 (d, J ˆ 11.0 Hz, 2H, CH2Ph), 4.69 (d, J ˆ 10.8 Hz,
2H, CH2Ph), 4.68 (d, J ˆ 12.2 Hz, 2H, CH2Ph), 4.58 (d, J ˆ 10.8 Hz, 2H,
CH2Ph), 4.60 (m, 4H, CH2Ph), 4.24 (ddd, J ˆ 4.9, 1.5, 9.7 Hz, 1H, H-5c),
4.08 (s, 1H, H-2d), 4.04 (dd, J ˆ 2.9, 1.5 Hz, 1H, H-2c), 3.98 (t, J ˆ 9.7 Hz,
1H, H-4c), 3.96 (dd, J ˆ 11.4, 4.9 Hz, 1H, H-6c), 3.85 (dd, J ˆ 8.9, 3.1 Hz,
H-3d), 3.90 (t, J ˆ 9.4 Hz, H-4d), 3.81 (ddd, J ˆ 4.1, 1.8, 9.4 Hz, H-4d), 3.76
(dd, J ˆ 1.5 Hz, H-6'c), 3.71 (dd, J ˆ 10.7, 4.1 Hz, H-6'd), 3.64 (dd, J ˆ
1.8 Hz, H-6d), 2.39 (s, 1H, OH); 13C NMR (125 MHz, CDCl3): d ˆ 138.6,
138.4, 138.3, 138.1, 137.9, 137.9, 134.7, 131.0, 129.2, 128.5, 128.5, 128.5, 128.4,
128.3, 128.3, 128.2, 128.0, 127.8, 127.7, 127.6, 127.3, 99.6, 85.9, 80.3, 79.7, 76.8,
75.2, 75.1, 74.7, 74.2, 73.4, 72.5, 72.1, 71.6, 71.0, 68.8, 66.5; elemental analysis
calcd (%) for C65H70O11S: C 73.90, H 6.20; found: C 74.19, H 6.25.
A 1m solution of NaOMe in MeOH (4 mL) was added to a solution of 32
(18 g, 32.57 mmol) in MeOH (50 mL). After stirring for 1 h at room
temperature, the reaction mixture was neutralised with Amberlite IR-120
‡
(H ). The resin was then filtered off and the solution concentrated to give
crude 32 as a colourless oil that was dissolved in pyridine (75 mL). A
catalytic amount of DMAP and pivaloyl chloride (12 mL) were added to
this solution with stirring. Stirring was continued for 60 h at room
temperature, then MeOH (50 mL) was added and the solution concen-
trated to dryness. The residue was dissolved in CH2Cl2 (400 mL) and
washed successively with 10% HCl (50 mL), saturated aqueous NaHCO3
and brine. The organic layer was dried over Na2SO4, concentrated and
purified by column chromatography (EtOAc/hexane, 1:15) to give 11
(16.5 g, 81%). Data for the a-anomer: 1H NMR (CDCl3, 500 MHz): d ˆ
7.70 ± 7.30 (m, 20H, arom.), 5.75 (s, 1H, H-2), 5.60 (s, 1H, H-1), 4.52 (AB
system, J ˆ 10.8 Hz, 2H, CH2Ph), 4.75 (AB system, J ˆ 11.1 Hz, CH2Ph),
4.70 (AB system, J ˆ 12.0 Hz, CH2Ph), 4.50 (m, 1H, H-5), 4.10 (m, 3H, H-2,
H-3, H-4), 4.00 (dd, 1H, J ˆ 10.8, 4.5 Hz, H-6), 3.90 (dd, 1H, J ˆ 10.8,
1.26 Hz, H-6'), 1.35 (s, 9H, Piv); 13C NMR (CDCl3, 125 MHz): d ˆ 177.6,
138.5, 138.4, 133.8, 132.3, 129.2, 128.5, 128.4, 128.2, 128.1, 127.9, 127.8, 127.7,
127.6, 127.5, 86.5, 78.9, 75.4, 74.5, 73.3, 72.5, 71.6, 69.9, 69.1, 39.1, 27.3.
Phenyl 2,3,4-tri-O-benzyl-6-O-[3,4,6-tri-O-benzyl-2-O-pivaloyl-a-d-man-
nopyranosyl)-a-d-mannopyranosyl]-1-thio-a-d-mannopyranoside (35): A
mixture of 9 (189 mg, 0.29 mmol) and DTBMP (8230 mg, 1.1 mmol) in
CH2Cl2 (12 mL) was treated with triflic anhydride (24.7 mL) at À788C. The
reaction was stirred for 5 min, then acceptor 34 (319 mg, 0.5 mmol) in
CH2Cl2 (4 mL) was added dropwise at À608C. Stirring was continued for
1 h while the temperature was allowed to rise to 08C. The reaction mixture
was hydrolysed with saturated aqueous NaHCO3 (30 mL), extracted with
CH2Cl2 (3 Â 20 mL), dried over sodium sulfate and concentrated. The
crude mixture was purified by flash column chromatography (hexanes/
EtOAc, 3:1) to give pure 35 (161 mg, 73%). [a]2D3‡37 (c ˆ 1.16 in CHCl3);
1H NMR (500 MHz, CDCl3): d ˆ 7.45 ± 7.00 (m, 45H, arom.), 5.54 (d, J ˆ
1.2 Hz, 1H, H-1c), 5.50 (t, J ˆ 2 Hz, 1H, H-2e), 5.02 (d, J ˆ 1.5 Hz, 1H,
A solution of m-chlorobenzoic acid (m-CPBA) (60%) in CH2Cl2 (30 mL)
was added through a canula to a solution of 33a (9.4 g, 15 mmol) in CH2Cl2
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3617 $ 17.50+.50/0
3617

M. Martín-Lomas et al.
FULL PAPER
H-1e), 4.96 (d, J ˆ 1.4 Hz, 1H, H-1d), 4.70 ± 4.30 (m, 18H, CH2Ph), 4.20 (m,
1H, H-5c), 4.02 (s, 1H, H-2d), 3.97 (m, 2H, H-2c, H-3d), 3.95 ± 3.55 (m,
13H), 1.11 (s, 9H, Piv); 13C NMR (125 MHz, CDCl3): d ˆ 177.3, 138.7, 138.6
(Â2), 138.4 (Â2), 138.3, 138.1, 137.9, 99.5, 99.1, 84.4, 128.6, 128.5, 128.3,
128.3, 128.2, 128.1, 128.0, 128.0, 127.9, 127.8, 127.7, 127.6, 127.6, 127.5, 127.4,
127.3, 127.2, 79.6, 79.1, 78.3, 76.8, 75.1, 75.1, 75.0, 74.8, 74.7, 74.5, 74.2, 73.2,
73.1, 72.3, 72.0, 71.9, 71.8, 71.7, 71.4, 69.0, 68.2, 67.3, 27.2; elemental analysis
calcd (%) for C92H98O16S: C 74.07, H 6.62; found: C 74.17, H 6.42.
purified by flash column chromatography (EtOAc/hexanes, 1:5) to give 37
(210 mg, 74%) as a colourless oil. H NMR (500 MHz, CDCl3): d ˆ 7.30 ±
1
7.05 (m, 55H, arom), 5.49 (t, J ˆ 2.9 Hz, 1H, H-2e), 5.47 (d, J ˆ 3.5 Hz, 1H,
H-1a), 5.31 (d, J ˆ 1.6 Hz, 1H, H-1c), 4.99 (d, J ˆ 1.4 Hz, 1H, H-1d), 4.90 ±
4.75 (m, 3H, CH2Ph), 4.60 ± 4.17 (m, 21H), 4.05 (m, 2H, H-1a, H-2d), 3.96
(dd, J ˆ 8.7 Hz, 2.9 Hz, 1H, H-3e), 3.90 (t, 1H, J ˆ 10.0, H-3b), 3.77 (m, 2H,
H-4b, H-6b), 3.72 (t, J ˆ 2.2, 1H, H-2c), 3.66 (m, 1H, H-6'b), 3.27 (dd, J ˆ
10.3, 3.5 Hz, 1H, H-2b), 2.61 (s, 1H, OH), 1.98 ± 0.90 (m, 38H), 0.8 (s, 6H);
13C NMR (125 MHz, CDCl3): d ˆ 177.2, 138.7, 138.6, 138.6, 138.5, 138.4,
138.3, 128.2, 137.9, 135.2, 129.7, 128.7, 128.4, 128.3, 128.3, 128.2, 128.1, 128.1,
128.0, 127.9, 127.6, 127.6, 127.5, 127.4, 127.4, 127.3, 127.3, 127.2, 127.0, 117.8,
99.9, 99.4, 99.2, 95.8, 80.2, 80.0, 79.4, 78.3, 76.3, 75.9, 75.8, 75.0, 74.8, 74.1,
73.9, 73.2, 73.1, 72.9, 72.3, 72.2, 72.1, 71.9, 71.9, 71.5, 71.4, 69.7, 69.1, 69.0,
68.2, 66.7, 63.2, 51.6, 48.0, 45.3, 45.1, 38.9, 29.7, 29.5, 29.4, 27.2, 27.1, 20.5, 20.2,
17.6, 17.5, 17.4, 17.3, 17.2, 17.1, 17.0, 13.0, 12.7, 12.2, 12.0, 9.7; elemental
analysis calcd (%) for C134H165N3O27Si2: C 69.80, H 7.21, N 1.82; found: C
69.66, H 7.22, N 1.99.
S and R (phenylsulfinyl) 2,3,4-tri-O-benzyl-6-O-[3,4,6-tri-O-benzyl-2-O-
(3,4,6-tri-O-benzyl-2-O-pivaloyl-a-d-mannopyranosyl)-a-d-mannopyrno-
syl]-1-deoxy-a-d-mannopyranoside (5a): A solution of m-CPBA (60%) in
CH2Cl2 (0.30 mL) was added by syringe to a solution of 35 (36.9 mg,
0.0247 mmol) in CH2Cl2 (1.5 mL) at À788C. After 1.5 h, the reaction was
stopped by addition of saturated aqueous NaHCO3, the organic layer was
washed with brine, dried over sodiumsulfate and evaporated under vacuo.
The residue was purified by column chromatography (EtOAc/hexanes, 1:3)
to give 30.8 mg (83%) of the less polar (major) isomer. 1H NMR (CDCl3,
500 MHz): d ˆ 7.50 (d, J ˆ 7.6 Hz, 2H, arom.), 7.40 (t, J ˆ 7.6 Hz, 2H,
arom.), 7.30 ± 7.11 (m, 46H, arom.), 5.49 (t, J ˆ 2.2 Hz, 1H, H-2e), 5.03 (s,
1H, H-1e), 4.90 (s, 1H, H-1d), 4.86 (d, J ˆ 10.96 Hz, 1H, CH2Ph), 4.85 (d,
J ˆ 10.7 Hz, 1H, CH2Ph), 4.80 (d, J ˆ 11.0 Hz, 1H, CH2Ph), 4.63 ± 4.34 (m,
16H), 4.16 (dd, J ˆ 3.1, 9.1 Hz, H-3c), 4.05 (m, 2H), 3.97 (dd, J ˆ 3.0, 9.1 Hz,
1H, H-3e), 3.85 (m, 7H), 3.72 (dd, 1H), 3.65 (m, 3H), 3.55 (m, 2H), 1.17 (s,
9H, Piv); 13C NMR (125 MHz, CDCl3): d ˆ 177.3, 142.2, 138.6, 138.5, 138.3,
138.2, 138.1, 138.0, 137.5, 131.2, 129.4, 128.5, 128.4, 128.4, 128.3, 128.2, 128.1,
128.1, 128.0, 127.9, 127.8, 127.7, 127.6, 127.5, 127.4, 127.3, 124.0, 99.5, 99.3,
95.9, 79.6, 79.1, 78.3, 77.1, 75.1, 75.0, 74.5, 74.4, 74.1, 73.7, 73.2, 73.1, 72.6,
72.1, 71.9, 71.8, 71.7, 71.4, 69.0, 68.2, 67.3, 60.4, 27.2.
O-(3,4,6-Tri-O-benzyl-2-O-pivaloyl-a-d-mannopyranosyl)-(1! 2)-O-(3,4,6-
tri-O-benzyl-a-d-mannopyranosyl)-(1 ! 6)-O-(2,3,4-tri-O-benzyl-a-d-
mannopyranosyl)-(1 ! 4)-O-(2-azido-3,6-tri-O-benzyl-2-deoxy-a-d-gluco-
pyranosyl)-(1 ! 6)-1,2-O-(L-1,7,7-trimethyl[2.2.1]bicyclohept-6-ylidene)-
d-myo-inositol (38): A solution of 37 (232 mg, 0.1 mmol) in THF (12 mL)
was treated with TBAF (1m solution in THF, 173 mL, 0.173 mmol) at
À158C. After 20 min, the reaction was hydrolysed by addition of saturated
aqueous NH4Cl (50 mL), extracted with EtOAc (5 Â 20 mL), dried over
sodium sulfate and evaporated. The mixture was purified by column
chromatography (EtOAc/hexanes, 1:1) to give pure 38 (8180 mg, 87%) as a
colourless oil. 1H NMR (500 MHz, CDCl3): d ˆ 7.30 ± 7.10 (m, 55H, arom),
5.50 (t, J ˆ 2.9 Hz, 1H, H-2e), 5.24 (d, J ˆ 3.7 Hz, 1H, H-1b), 5.18 (d, J ˆ
1.5 Hz, 1H, H-1c), 5.09 (d, J ˆ 1.5 Hz, 1H, H-1e), 4.91 (d, J ˆ 1.3 Hz, 1H,
H-1d), 4.82 (d, J ˆ 10.9 Hz, 2H, CH2Ph), 4.76 (d, J ˆ 11.0 Hz, 1H, CH2Ph),
4.74 (d, J ˆ 11.7 Hz, 1H, CH2Ph), 4.63 ± 4.27 (m, 19H), 4.26 (m, 1H, H-2a),
4.08 (t, J ˆ 2.83 Hz, 1H, H-2d), 3.97 (dd, J ˆ 9.5, 2.9 Hz, H-3e), 3.88 (t, J ˆ
9.5 Hz, 1H, H-4d), 3.8 (dd, J ˆ 9.5, 2.8 Hz, H-3d), 3.90 (m, 2H), 3.77 (m,
1H, H-3b), 3.77 ± 3.58 (m, 17H), 3.75 (m, 1H, H-5d), 3.43 (t, J ˆ 9.3 Hz, 1H,
H-5a), 3.27 (dd, J ˆ 9.9, 3.7 Hz, 1H, H-2b), 3.22 (td, J ˆ 9.2, 3.6 Hz, 1H,
H-4a), 3.10 (s, 1H, OH), 2.10 (s, 1H, OH), 1.90 (m, 2H), 1.15 (s, 9H, Piv),
0.65 (s, 3H), 1.70 ± 0.70 (m, 29H); 13C NMR (125 MHz, CDCl3): d ˆ 177.2,
138.7, 138.5, 138.3, 138.2, 138.1, 137.9, 137.9, 137.8, 128.2, 128.1, 127.9, 127.7,
127.6, 126.8, 118.4, 100.2, 99.4, 99.0, 97.1, 82.4, 80.1, 80.0, 79.6, 78.4, 78.2,
76.0, 75.6, 75.5, 75.1, 75.0, 74.9, 74.4, 74.3, 74.2, 73.4, 73.3, 73.2, 73.1, 73.0,
72.8, 72.7, 72.5, 72.0, 71.8, 71.6, 71.5, 71.3, 70.4, 70.2, 69.1, 69.0, 68.8, 67.9, 67.4,
63.0, 51.6, 47.9, 45.1, 45.0, 38.9, 29.7, 29.6, 29.4, 29.1, 27.0, 20.5, 20.4, 14.1, 9.8;
2,3,4-Tri-O-benzyl-6-O-[3,4,6-tri-O-benzyl-2-O-(3,4,6-tri-O-benzyl-2-O-
pivaloyl-a-d-mannopyranosyl)-a-d-mannopyranosyl]-d-mannopyranoside
(36): A solution of 35 (381 mg, 0.255 mmol) in 99% wet acetone (20 mL)
was treated portionwise with NBS (60 mg, 0.322 mmol) at À158C. The
reaction mixture was stirred for 30 min with exclusion of light, then
neutralised with saturated aqueous NaHCO3, extracted with CH2Cl2 (3 Â
20 mL), dried over sodium sulfate and concentrated under vacuo. The
crude mixture was purified by flash column chromatography (EtOAc/
hexanes, 10:90) which gave 36 (348 mg, 98%) as a colourless oil in a 4:1 a/b
mixture. 1H NMR (500 MHz, CDCl3): d ˆ 5.00 (s, H-2eb), 5.47 (d, J ˆ
1.4 Hz, H-2ea), 5.05 (s, H-1eb), 5.02 (s, H-1db), 4.96 (d, J ˆ 2.2 Hz,
H-1db), 4.90 (s, H-1ca), 4.90 ± 4.20 (m, 18H, CH2Ph), 4.09 (t, J ˆ 2.6 Hz,
H-2da), 4.20 (s, H-2db), 3.98 (dd, J ˆ 10.4 Hz H-3ea), 3.89 (m, 1H, H-3da),
3.80 ± 3.50 (m, 14H), 1.18(s, 9H, Piv); 13C NMR (125 MHz, CDCl3): d ˆ
177.3, 138.5, 138.5, 138.4, 138.3, 138.2, 128.5, 128.4, 128.3, 128.2, 128.1, 128.0,
127.9, 127.8, 127.8, 127.7, 127.6, 127.6, 127.5, 127.4, 98.6, 92.1, 79.9, 78.2, 75.3,
75.2, 75.0, 74.9, 74.4, 73.7, 73.3, 73.2, 72.5, 72.0, 71.8, 71.7, 71.5, 70.8, 69.7,
69.3, 68.1, 68.0, 27.7, 27.2; HRFABMS calcd for C86H94O17NA: 1421.638872;
found: 1421.6650540.
‡
FABMS calcd for C122H128N3O26Na [M‡Na] : 2086; found 2086.
O-(2,3,4,6-Tetra-O-benzyl-a-d-mannopyranosyl-(1 ! 2)-O-(3,4,6-tri-O-
benzyl-a-d-mannopyranosyl)-(1 ! 6)-O-(2,3,4-tri-O-benzyl-a-d-manno-
pyranosyl)-(1 ! 4)-O-(2-azido-3,6-tri-O-benzyl-2-deoxy-a-d-glucopyrano-
syl)-(1 ! 6)-3,4,5-tri-O-benzyl-d-myo-inositol (3):
A solution of 38
2,3,4-Tri-O-benzyl-6-O-[3,4,6-tri-O-benzyl-2-O-(3,4,6-tri-O-benzyl-2-O-
pivaloyl-a-d-mannopyranosyl)-a-d-mannopyranoside-trichloroacetimi-
date (5b): NaH (60%, 15 mg), previously washed with hexanes, was added
(150 mg, 0.007 mmol) in MeOH/THF (1:1, 2 mL) was treated with a 1m
solution of MeONa in MeOH (150 mL). The mixture was stirred at 608C for
‡
24 h, cooled to room temperature, neutralised with Amberlite IR-120 (H )
to
a solution of 36 (334 mg, 0.239 mmol) and trichloroacetonitrile
and evaporated. The residue was dissolved in DMF (2.5 mL) and NaH
(17.5 mg, 0.43 mmol) was added to this solution. The mixture was stirred for
30 min at room temperature and then benzyl bromide (70 mL, 0.584 mmol)
was added dropwise. After 1 h, the mixture was cooled to 08C, MeOH
(5 mL) was added followed by CH2Cl2 (100 mL). The organic layer was
washed with saturated aqueous NH4Cl, dried and concentrated. The
residue was purified by column chromatography (EtOAc/hexane, 1:4.5) to
(0.359 mL) in CH2Cl2 (12 mL). After 2 h, the solvent was evaporated and
the mixture purified by column chromatography by using an Et3N
pretreated silica gel to give 5b (314 mg, 85%) as a colourless oil. Rf ˆ
0.86; 1H NMR (400 MHz, C6D6): d ˆ 8.65 (s, 1H, NH), 7.50 ± 7.00 (m, 45H,
arom.), 6.72 (d, J ˆ 1.8 Hz, 1H, H-1c), 5.95 (dd, J ˆ 3.0 Hz, 1H, H-1e), 5.10
(d, J ˆ 11.1 Hz, 1H, CH2Ph), 5.08 (d, J ˆ 10.8 Hz, 2H, CH2Ph), 4.77 ± 4.60
(m, 10H), 4.50 (m, 6H), 4.43 (t, 1H), 4.35 (m, 2H), 4.2 (m, 7H), 4.1 (dd,
1H), 4.03 (t, 1H), 4.0 (m, 3H), 3.86 (d, 1H), 3.76 (d, 1H), 1.2 (s, 9H, Piv);
elemental analysis calcd (%) for C88H94NO17Cl3: C 68.45, H 6.14, N 0.9;
found: C 68.55, H 6.38, N 1.02.
1
give 39. H NMR (500 MHz, CDCl3): d ˆ 7.03 ± 7.34 (m, 75H, arom), 5.57
(d, J ˆ 3.5 Hz, 1H, H-1b), 5.25 (s, 1H, H-1c), 5.11 (s, 1H, H-1e), 4.88 (s, 1H,
H-1d), 4.51 ± 4.88 (m, 13H, CH2PH), 4.32 ± 4.47 (m, 15H, CH2Ph), 4.46 (dd,
1H), 4.20 (dd, 2H, CH2Ph), 4.06 ± 4.10 (m, 2H), 3.98 ± 4.04 (m, 3H), 3.91 (t,
1H), 3.59 ± 3.86 (m, 16H), 3.48 ± 3.54 (m, 3H), 3.38 ± 3.42 (m, 3H), 3.26 (dd,
J ˆ 9.7, 3.5 Hz, 1H, H-2b), 185 ± 1.90 (m, 2H), 1.67 ± 1.72 (m, 2H), 1.43 (d,
1H), 1.16 ± 1.35 (m, 2H), 1.03 (s, 3H), 0.82 (s, 3h), 0.80 (s, 3H); 13C NMR:
d ˆ 138.9, 138.8, 138.6, 138.5, 138.4, 138.2, 138.1, 137.8, 128.6, 128.5, 128.4,
128.3, 128.2, 128.1, 128.0, 127.9, 127.8, 127.7, 127.6, 127.4, 127.3, 127.2, 127.0,
118.2, 100.3, 99.4, 95.3, 80.9, 80.7, 80.0, 79.9, 79.7, 77.9, 77.2, 77.0, 76.2, 75.2,
75.0, 74.9, 74.8, 74.7, 74.2, 73.9, 73.8, 73.3, 73.2, 73.1, 72.6, 72.3, 72.2, 72.1,
72.0, 71.9, 69.9, 69.2, 68.8, 66.7, 63.0, 51.6, 48.0, 45.2, 44.9, 29.9, 29.7, 27.0,
20.6, 20.5, 20.4, 9.7.
O-(3,4,6-tri-O-benzyl-2-O-pivaloyl-a-d-mannopyranosyl)-(1 ! 2)-O-
(3,4,6-tri-O-benzyl-a-d-mannopyranosyl)-(1 ! 6)-O-(2,3,4-tri-O-benzyl-a-
d-mannopyranosyl)-(1 ! 4)-O-(2-azido-3,6-tri-O-benzyl-2-deoxy-a-d-glu-
copyranosyl)-(1 ! 6)-3,4-O-(1,1,3,3-tetraisopropyldisiloxanedi-1,3-yl)-1,2-
O-(l-1,7,7-trimethyl[2.2.1]bicyclohept-6-ylidene)-d-myo-inositol
(37):
TMSOTf (1.6 mL, 0.008 mmol) was added to a mixture of 4 (114 mg,
1 equiv), 5b (210 mg) and 4 Š molecular sieves in CH2Cl2 (200 mg) at room
temperature. After 1 h, the reaction mixture was neutralised with solid
NaHCO3, filtered over Celite and evaporated. The crude mixture was
3618
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3618 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

Inositolphosphoglycan Mediators
3608±3621
Trifluoroacetic acid (0.4 mL) was added to a solution of 39 (149 mg,
0.06 mmol) in wet chloroform (3 mL) and the mixture kept for 18 h at room
temperature. Saturated aqueous NaHCO3 was then added at 08C, the
aqueous layer extracted with CH2Cl2 (3 Â 10 mL) and the combined
organic extracts were dried and concentrated. The residue was purified by
column chromatography (EtOAc/hexane, 1:25) to afford pure 3 (85 mg,
60%) as a colourless oil. 1H NMR (500 MHz, C6D6): d ˆ 7.50 ± 6.96 (m,
75H, arom), 5.66 (d, J ˆ 3.5 Hz, 1H, H-1b), 5.59 (d, J ˆ 1.6 Hz, 1H, H-1c),
5.49 (d, J ˆ 1.3 Hz, 1H, H-1e), 5.33 (d, J ˆ 1.5 Hz, 1H, H-1d), 5.11 ± 5.04 (m,
3H, CH2Ph), 4.98 ± 4.94 (m, 2H, CH2Ph), 4.91 ± 4.87 (m, 2H, CH2Ph),
4.79 ± 4.65 (m, 5H, CH2Ph), 4.61 ± 4.18 (m, 29H), 4.13 ± 4.00 (m, 7H), 3.90
(t, J ˆ 2.4 Hz, 1H, H-2c), 3.89 ± 3.65 (m, 8H), 3.55 ± 3.51 (m, 1H, H-1a), 3.37
(t, J ˆ 9.5 Hz, 1H, H-5a), 3.12 (dd, J ˆ 9.6, 2.6 Hz, 1H, H-3a), 3.08 (dd, J ˆ
10.2, 3.5 Hz, 1H, H-2b), 2.97 (d, J ˆ 7.3 Hz, 1H, OH); 13C NMR (125 MHz,
CDCl3): d ˆ 138.8, 138.7, 138.6, 138.5, 138.4, 138.2, 138.1, 137.7, 137.6, 128.6,
128.5, 128.4, 128.3, 128.2, 128.1, 128.0, 127.9, 127.8, 127.7, 127.6, 127.5, 127.4,
127.3, 127.2, 127.0, 100.2, 99.4, 98.0, 81.6, 81.0, 80.6, 79.9, 79.8, 79.7, 76.9, 76.1,
75.8, 75.2, 74.9, 74.6, 74.4, 74.1, 73.3, 73.2, 72.8, 72.3, 72.1, 72.0, 71.9, 70.8,
temperature of 300 K with a thermal-bath coupling constant of 25 ps, and
SHAKE was used for hydrogen bonds; the structures were saved each
picosecond. The conditions were optimised by running previous 500 ±
1000 ps dynamics and by varying the integration step values (1, 1.5 and
2 fs) and thermal bath coupling constants (5, 10, 25 and 50 ps), and by using
a constant dielectric electrostatic treatment (e ˆ 80). Trajectories were
analysed with software written in-house.[90] The molecular dynamic
structures, excluding the first 20 ps, were grouped into two sets. The first
one included the first 1020 ps structures of the gg starting conformation
simulation, for which this rotamer is stable, and the second one
corresponded to the structures fromthe whole gt simulation. Both sets
were clustered in fifty groups according their heavy atomrmsd and then the
representative structure of each cluster was extracted.
Biological assays: Coenzyme A, lithium salt, sodium pyruvate, thiamine
pyrophosphate and dithiothreitol were obtained fromSigma (Poole,
Dorset, UK). ATP disodiumsalt, NAD and EGTA were obtained from
Boehringer Mannheim(Germany); pyruvate dehydrogenase and pyruvate
dehydrogenase phosphatase assays were carried out with a Jasco v-560 UV/
Vis spectrophotometer, the circulating temperature in the cell housing was
308C. d-[U-14C] glucose and d-[3-3H] glucose were purchased fromICN
Biomedicals (Ohio). Collagenase d was fromBoehringer and Insulin was
purchased from Sigma. Flaco Primaria microtest tissue culture plates were
purchased fromBecton Dickinson. All reagents were of analytical grade or
better. Male Wistar rats (120 ± 140 g) were purchased fromHarlan Olac
(Bicester, UK).
‡
69.6, 69.2, 68.9, 66.6, 64.3; FABMS calcd for C135H141N3O25Na[M‡Na] :
2228; found: 2228.
O-a-d-Mannopyranosyl-(1 ! 2)-O-a-d-mannopyranosyl-(1 ! 6)-O-a-d-
mannopyranosyl-(1 ! 4)-O-2-ammonio-2-deoxy-a-d-glucopyranosyl-(1 !
6)-D-myo-inositol-1,2-cyclic phosphate (1a): A solution of dichlorome-
thylphosphate (10 mL, 19.9 mmol) in pyridine (180 mL) was stirred at room
temperature until formation of a white precipitate occured(30 min). A
solution of 3 in pyridine (120 mL) was added dropwise. The mixture was
stirred for 30 min and then saturated aqueous NaHCO3 (0.7 mL) was
added. The whole mixture was then evaporated and the residue dissolved in
water, acidified with aqueous HCl (2m) until the solution reached pH 1,
extracted with EtOAc and the extract was dried over sodiumsulfate and
concentrated under vacuo. The residue was dissolved in THF/EtOH/H2O,
1:11:2 (1.8 mL) that contained NH4OAc (7 mg) and 10% Pd/C. The
reaction mixture was stirred for 18 h under atmospheric pressure of H2,
filtered over Celite and concentrated. The crude mixture was passed
through Sephadex C-18, loaded on Sephadex G-25 and eluted with 10%
H2O/EtOH. Liophylisation gave 1a as a white powder. 1H NMR (500 MHz,
D2O): d ˆ 5.45 (d, J ˆ 3.6 Hz, H-1b), 5.18 (s, H-1c), 5.10 (s, H-1d), 4.97 (s,
H-1e), 4.65 (t, J ˆ 4.2 Hz, H-1a), 4.52 (ddd, J ˆ 4.5, 8.0, 20.4 Hz, H-2a),
4.08 ± 3.61 (m, 25H), 3.5 (t, J ˆ 9.8 Hz, 1H, H-4e), 3.36 (t, J ˆ 9.9 Hz, H-5a),
3.31 (dd, J ˆ 3.4, 10.8 Hz, H-2b).
NMR measurements: NMR experiments were recorded on a DRX500
Bruker spectrometer with standard or 200 mL microsample NMR tubes.
DQF-COSY, HSQC, HMQC and TOCSY measurements used for the full
assignment of 1a were recorded by using the standard z-pulsed-field
gradient-enhanced or selected-pulse sequences when possible. NOESY
measurements were recorded with mixing times of 100, 200, 300 and
400 ms. Selective one-dimensional NOESY and TOCSY experiments were
recorded with the dpfgse technique.[88] Due to severe overlapping, 1H-1H
and 31P-1H coupling constants were extracted fromthe dqf-COSY by
deconvolution of the two-dimensional antiphase peaks. Calculation of the
w rotamer populations were done with a program written in-house.[89]
Activation of PDH phosphatase: The biological activity of 1a was
determined by using the activation of PDH phosphatase.[92] The assay of
the activation of the phosphatase was performed by the spectrophotometric
variant of the two-stage systemdescribed. The pyruvate dehydrogenase
complex (PDC) and the PDH phosphatase were prepared from bovine
heart and stored at À808C until use. The assay for PDH phosphatase, in the
presence or the absence of 1a and liver IPG, was based upon the initial rate
of the activation of the inactivated phosphorylated PDH complex. After
inactivation with ATP, the activity of the PDC was reduced to less than 1%
of the original value.
A two-stage assay was used to quantify the
phosphatase activity. A sample of inactivated PDC (50 mL) was pre-
incubated at 308, for 3 min C in a solution that contained 1 mgmLÀ1 fat-
free bovine serumalbumin, 10m m MgCl2, 0.1mm CaCl2, 1mm dithiothrei-
tol in 20m potassiumphosphate buffer, pH 7.0 (total volume 250 mL). After
this time, 5 mL of metal, 10 mL of analogue or IPG and 10 mL of the PDH
phosphatase were added and the incubation was continued for a further
3 min. At the end of this time, NaF (300mm, 135 mL)was added. The
activated PDH was determined at the second stage spectrophotometrically
by measuring the rate of production of the reduced form of NADH. Two
hundred microliters of the stopped reaction were added to 1 mL of reaction
mixture that contained 50mm potassiumphosphate buffer at pH 8.0,
2.5mm coenzyme A, 0.32mm dithiothreitol and 2mm sodiumpyruvate. The
production of NADH was followed by using an absorbance at 340 nmfor
5 min.
Inbition of cAMP-dependent protein kinase (PKA): The ability of 1a to
inhibit PKA activity was determined by using a colorimetric assay kit and a
standard PKA preparation (Pierce, Rockford, IL, USA). A PKA inhibitory
peptide was used as a positive control.
Molecular modelling and molecular dynamics: Glycosidic torsion angles
were defined as F H-1-C-1-O-1-C-X, Y C-1-O-1-C-X-H-X and w H-5-C-5-
C-6-O-6. All simulations: single minimisation, Monte Carlo (MC) multiple
minimum conformational search or molecular dynamics (MD) were
performed by using AMBER* with force-field parameters for pyranose
oligosaccharides. Solvent effects were included by using the GB/SA
continuummodel for water.
Isolation of adipocytes: Adipocytes were isolated fromepididymal fat pads
of male rats by collagenase digestion as described by Rodbell with minor
modifications.[93] Briefly, fat pads fromtwo rats were finely minced with
scissors and incubated in 10 mL Krebs Ringer Hepes (KRH) buffer that
contained Hepes (9.2mm), NaH2PO4 ´ 2H2O (2.2mm), NaHCO3 (10mm),
NaCl (132mm), KCl (4.7mm), MgSO4 ´ 7H2O (1.2mm), CaCl2 ´ 6H2O
(2.5mm), 2% BSA and glucose (5mm), pH 7.4, and collagenase (20 mg)
for 20 ± 30 min at 378C in a shaking water bath with continuous gassing with
95% O2 and 5% CO2.
All minimisations were run by using up to 9999 Polak Ribiere conjugate
gradient iterations until the convergence criterion (rmsd derivative lower
than 0.001 kJŠÀ1 molÀ1) was achieved. A multiple minimum conforma-
tional search, which varied all the torsional angles involved in the glycosidic
linkages simultaneously, was performed with 1000 Monte Carlo steps, with
a previously accepted structure used as the starting structure in the next
step. After every MC step, the resulting geometry was minimised with
3000 iterations of steepest descent. The structure was checked against
duplication and accepted within a window of 50 kJmolÀ1 from the minima.
The minimum energy structure was a Mana(1 ! 6)3-Man gt rotamer; gg
and tg starting structures used in MD were manually built from the first one
and minimised. Molecular dynamics simulations for the above mentioned
structures were run for 2.0 ns, with an integration step of 1.5 fs, at a constant
Multiwell plate lipogenesis: Lipogenesis was determined as the incorpo-
ration of [3-3H] glucose into toluene-extractable lipids. Briefly, 100 mL of
adipocyte suspension (3.5 Â 105 mLÀ1 KRH, with or without 10À8m insulin)
was incubated in a 96-multiwell plate for 30 minutes at 378C in a CO2
incubator with 2 mL of various concentrations of 1a, with or without a
tenfold excess of Zn2‡. Lipogenesis was initiated by the addition of 100 mL
KRH containing 0.2 mCi d-[3-3H] glucose and the incubation continued for
2 hours. Adipocytes were harvested onto glassfibre filter mats by using a
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3619 $ 17.50+.50/0
3619

M. Martín-Lomas et al.
FULL PAPER
Â
cell harvester and rinsed with 5 mm glucose in 0.154m NaCl. A toluene-
based scintillation cocktail (3 ml) was added to each filter disc for counting
the radioactivity incorporated into lipids.
[18] H. Dietrich, J. F. Espinosa, J. L. Chiara, J. Jimenez-Barbero,Y. Leon, I.
Varela-Nieto, J. M. Mato, F. H. Cano, C. Foces-Foces, M. Martín-
Lomas, Chem. Eur. J. 1999, 5, 320 ± 336 and references cited therein.
[19] S. S. Taylor, J. A. Buechler, W. Yonemoto, Ann. Rev. Biochem. 1990,
59, 971 ± 1005.
[20] S. Tamura, K. R. Lynch, J. Larner, J. Fox, A. Yasui, K. Kikuchi, Y.
Suzuki, S. Tsuiki, Proc. Nat. Acad. Sci. USA 1989, 86, 1796 ± 1800.
[21] P. H. Hünenberger, V. Helms, N. Narayana, S. S. Taylor, J. A.
McCammon, Biochemistry 1999, 38, 2358 ± 2366.
[22] N. Narayana, T. C. Diller, K. Koide, M. E. Bunnage, K. C. Nicolau,
L. J. Brunton, N.-H. Xuong, L. F. T. Eyck, S. S. Taylor, Biochemistry
1999, 38, 2377 ± 2386.
[23] A. K. Das, N. R. Helps, P. T. W. Cohen, D. Barford, EMBO J. 1996, 15,
6798 ± 6809.
[24] T. Ogawa, C. Murakata, Tetrahedron Lett. 1991, 32, 671 ± 674.
[25] a) T. Ogawa, C. Murakata, Carbohyd. Res. 1992, 235, 95 ± 114;
b) D. A. Baeschlin, A. R. Chaperon. L. G. Green, M. G. Hahn, S. J.
Ince, S. V. Ley, Chem. Eur. J. 2000, 6, 172 ± 186.
[26] B. Fraser-Reid, A. S. Campbell, J. Am. Chem. Soc. 1995, 117, 10387 ±
10388.
Statistical analysis: Data are presented as the mean standard error of the
percentage changes frombasal and statistical significance was tested by
Studentꢁs two-tailed t-test for unpaired observations. For the multiwell
plate lipogenesis, statistical significance of changes, the mean and the
standard deviation of the natural logarithms of the net counts per minute
were calculated. The standard deviations for insulin-treated and untreated
cells were separately pooled by averaging their squares, calculating the
square root of the average and plotting the difference of the mean of each
treatment from the basal on a log scale. The error lines above and below the
zero are given by 0 ‡ 2d, in which d ˆ s (1/n ‡ 1/m), s is the pooled standard
deviation, n is the number of replicates in each treatment mean and m is the
number of basal replicates subtracted from the mean. They represent the
range outside which a change is significantly different fromthe basal. The
percent change fromthe basal was obtained as ( ex À 1), in which x is the
difference fromthe basal and multiplied by 100.
[27] T. G. Mayer, B. Kratzer, R. R. Schmidt, Angew. Chem. 1994, 106,
2289 ± 2293; Angew. Chem. Int. Engl. 1994, 33, 2177 ± 2181.
[28] T. G. Mayer, R. R. Schmidt, Eur. J. Org. Chem. 1999, 1153 ± 1165.
[29] R. R. Motoo, P. Konradsson, B. Fraser-Reid, J. Am. Chem. Soc. 1989,
111, 8540 ± 8542.
Acknowledgements
Thanks are given to DGES (Grant PB96 0820) and The Rademacher
Group for financial support. J.-L.K. is grateful to the European Union
(TMR Programme: CARENET-2) for a postdoctoral fellowship. The
contribution of Dr. María Flores-Mosquera and Dr. Emmanuel Poirot to
the experimental part of this work is also gratefully acknowledged.
[30] U. D. Udodong, R. Madsen, C. Roberts, B. Fraser-Reid, J. Am. Chem.
Soc. 1993, 115, 7886 ± 7887.
[31] A. Asarpan, B. Fraser-Reid, J. Org. Chem. 1996, 61, 2401 ± 2406.
[32] R. Verduyn, C. J. J. Elie, C. E. Dreef, G. van der Marel, J. H.
van Boom, Rec. Trav. Chim. Pays Bas 1990, 109, 591 ± 593.
[33] R. Plourde, M. dꢁAlarcao, A. R. Saltiel, J. Org. Chem. 1992, 57, 2606 ±
2610.
[1] A. R. Saltiel, J. A. Fox, P. Sherline, P. Cuatrecasas, Science 1986, 233,
967 ± 972.
[34] A. Zapata, M. Martín-Lomas, Carbohyd. Res. 1992, 234, 93 ± 106.
Â
Â
[35] A. Zapata, Y. Leon, J. M. Mato, I. Varela-Nieto, S. Penades, M.
Martín-Lomas, Carbohyd. Res. 1994, 264, 21 ± 31.
[36] C. Jaramillo, J. L. Chiara, M. Martín-Lomas, J. Org. Chem. 1994, 59,
3135 ± 3141.
[2] For recent reviews see: a) D. R. Jones, I. Varela-Nieto, Int. J. Biochem.
Cell Biol. 1998, 30, 313 ± 326; b) P. Stralfors, BioEssays, 1997, 19, 327 ±
335; c) M. C. Field, Glycobiology, 1997, 7, 161 ± 168; d) I. Varela-Nieto,
Â
Y. Leon, H. N. Caro, Comp. Biochem. Physiol. B 1996, 115, 223 ± 241.
[37] N. Khiar, M. Martín-Lomas, J. Org. Chem. 1995, 60, 7017 ± 7021.
[38] S. V. Ley, L. L. Young, Synlett 1992, 997 ± 998.
[39] G. J. Boons, P. Grice, R. Leslie, S. V. Ley, L. L. Young, Tetrahedron
Lett. 1993, 34, 8523 ± 8526.
[40] B. Kratzer, T. G. Mayer, R. R. Schmidt, Tetrahedron Lett. 1993, 34,
6881 ± 6884.
[41] B. Kratzer, T. G. Mayer, R. R. Schmidt, Eur. J. Org. Chem. 1998, 291 ±
298.
[42] A. I. Frantova, A. E. Stepanov, A. S. Bushnev, E. N. Zvonkova, V. I.
Shvets, Tetrahedron Lett. 1992, 33, 3539 ± 3542.
[43] T. Ziegler, R. Dettman, M. Duszenko, V. Kolb, Carbohyd. Res. 1996,
295, 7 ± 23.
[3] J. M. Mato, K. Kelly, A. Abler, L. Jarret, B. E. Corkey, J. A. Cashel, D.
Zopf, Biochem. Biophys. Res. Commun. 1987, 146, 764 ± 770.
[4] J. Larner, L. C. Huang, C. F. W. Schwartz, A. S. Oswald, T. Y. Shen, M.
Kinter, G. Tang, K. Zeller, Biochem. Biophys. Res. Commun. 1988,
155, 1416 ± 1426.
[5] For reviews on GPI anchors see: a) P. T. Englund, Annu. Rev.
Biochem. 1993, 62, 121 ± 138; b) M. J. McConville, M. A. J. Ferguson,
Biochem. J. 1993, 294, 305 ± 324; c) V. L. Stevens, Biochem. J. 1995,
310, 361 ± 370.
Â
[6] G. Romero, G. Gomez, L. C. Huang, K. Lilley, L. Luttrell, Proc. Nat.
Acad. Sci. USA, 1990, 87, 1476 ± 1480.
[7] J. Represa, M. A. Avila, C. Miner, F. Giraldez, G. Romero, R.
Clemente, J. M. Mato, I. Varela-Nieto, Proc. Natl. Acad. Sci. USA,
1991, 88, 8016 ± 8019.
[8] J. Represa, M. A. Avila, G. Romero, J. M. Mato, F. Giraldez I. Varela-
Nieto, Dev. Biol. 1993, 159, 257 ± 265.
[44] M. M. Silva, J. Cleophax, A. Benicio, M. V. Almeida, J. M. Delaumey,
A. S. Machado, S. D. Gero, Synlett 1996, 764 ± 766.
[45] P. J. Garegg, P. Konradsson, S. Oscarson, K. Ruda, Tetrahedron 1997,
53, 17727 ± 17734.
[46] S. Cottaz, J. S. Brimacombe, M. A. J. Ferguson, J. Chem. Soc. Perkin
Trans. 1 1993, 2945 ± 2951.
[9] J. M. Mato, K. L. Kelly, A. Abler, L. Jarrett, J. Biol. Chem. 1987, 262,
2131 ± 2137.
[47] S. Cottaz, J. S. Brimacombe, M. A. J. Ferguson, Carbohydr. Res. 1995,
270, 85 ± 91.
[48] S. Cottaz, J. S. Brimacombe, M. A. J. Ferguson, J. Chem. Soc. Perkin
Trans. 1 1995, 1673 ± 1678.
[10] S. Kunjara, D. Y. Wang, A. L. Greenbaum, P. McLean, A. Kurtz, T. W.
Rademacher, Mol. Gen. Metab. 1999, 68, 488 ± 502.
[11] M. A. Deeg, N. R. Murray, T. L. Rosenberry, J. Biol. Chem. 1992, 267,
18581 ± 18588.
[49] F. Tagliaferri, S.-N. Wang, W. K. Berlin, R. A. Outten, T. Y. Shen,
Tetrahedron Lett. 1990, 31, 1105 ± 1108.
[12] T. L. Rosenberry, D. Sevlever, M. E. Medof, Brazilian J. Med. Res.
1994, 27, 151 ± 159.
[50] W. K. Berlin, S. N. Wang, T. Y. Shen, Tetrahedron Lett. 1990, 31, 1109 ±
1112.
[13] G. Müller, S. Wied, C. Piossek, A. Bauer, J. Bauer, W. Frick, Mol. Med.
1998, 4, 299 ± 323.
[51] W. K. Berlin, S. N. Wang, T. Y. Shen, Tetrahedron 1991, 47, 1 ± 20.
[52] J. Ye, W. T. Doerrier, M. A. Lehrman, J. R. Falk, Bioorg. Med. Chem.
Lett. 1996, 6, 1715 ± 1718.
[53] K. K. Reddy, J. R. Falk, J. Capdevilla, Tetrahedron Lett. 1993, 34,
7869 ± 7872.
[54] R. Gigg, J. Gigg in Glycopeptides and Related Compounds (Eds: G. D.
Large, C. D. Warren), Marcel Dekker, New York, 1997, pp. 327 ± 392.
[55] D. Khane, S. Walker, Y. Cheng, D. Van Engen, J. Am. Chem. Soc.
1989, 111, 6881 ± 6882.
[14] W. Frick, A. Bauer, J. Bauer, S. Wied, G. Müller, Biochem. J. 1998, 336,
163 ± 181.
[15] A. Kessler, G. Müller, S. Wied, A. Crecelius, J. Eckel, Biochem. J.
1998, 330, 277 ± 286.
[16] W. Frick, A. Bauer, J. Bauer, S. Wied, G. Müller, Biochemistry 1998,
37, 1421 ± 1436.
[17] N. Khiar, M. Martín-Lomas in Carbohydrate Mimics. Concepts and
Methods (Ed.: Y. Chapleur), Wiley-VCH, Weinheim, 1998, pp. 443 ±
462 and references cited therein.
3620
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3620 $ 17.50+.50/0
Chem. Eur. J. 2000, 6, No. 19

Inositolphosphoglycan Mediators
3608±3621
[56] K. S. Bruzic, M. D. Tsai, J. Am. Chem. Soc. 1992, 114, 6361 ± 6374.
[57] Y. Watanabe, T. Yamamoto, S. Ozaki, J. Org. Chem. 1996, 61, 14 ± 15.
[58] R. R. Schmidt, W. Kinzi, Adv. Carbohydr. Chem. Biochem. 1994, 50,
21 ± 123.
[59] A. Vasella, C. Witzig, J. L. Chiara, M. Martín-Lomas, Helv. Chim. Acta
1991, 74, 2073 ± 2077.
[77] S. W. Homans, C. J. Edge, M. A. J. Ferguson, R. A. Dwek, T. W.
Rademacher, Biochemistry 1989, 28, 2881 ± 2887.
[78] H. Senderowitz, W. C. Still, J. Org. Chem. 1997, 62, 1427 ± 1438.
[79] W. C. Still, A. Tempczyk, R. C. Hawley, T. Hendrickson, J. Am. Chem.
Soc. 1990, 112, 6127 ± 6129.
[80] G. Chang, W. C. Guida, W. C. Still, J. Am. Chem. Soc. 1989, 111, 4379 ±
4386.
[60] M. Martín-Lomas, M. Flores-Mosquera, J. L. Chiara, Eur. J. Org.
Chem. 2000, 1547 ± 1562.
[81] M. Saunders, K. N. Houk, Y.-D. Wu, W. C. Still, M. Lipton, G. Chang,
W. C. Guida, J. Am. Chem. Soc. 1990, 112, 1419 ± 1427.
[82] S. Perez, A. Imberty, S. E. Engelsen, J. Gruza, K. Mazcan, J. Jimenez-
Barbero, A. Poveda, J. F. Espinosa, B. P. Van Eyk, G. Johnson, A. D.
French, M. L. C. E. KouwiJzer, P. D. Grootenuis, A. Bernardi, L.
Raimondi, H. Senderowitz, V. Durier, G. Vergoten, K. Rasmusen,
Carbohydr. Res. 1998, 314, 141 ± 155.
[61] R. D. Groneberg, T. Nuyazaki, N. A. Stylianides, T. J. Schulze, W.
Stahl, E. P. Schreiner, T. Suzuki, Y. Iwabuchi, A. L. Smith, K. C.
Nicolau, J. Am. Chem. Soc. 1993, 115, 7593 ± 7611.
[62] K. C. Nicolau, R. E. Dolle, D. P. PapakatJis, J. L. Randall, J. Am.
Chem. Soc. 1984, 106, 4189 ± 4192.
[63] V. Poszgay, H. J. Jennings, J. Org. Chem. 1988, 53, 4042 ± 4052.
[64] D. S. Brocon, S. V. Ley, S. Vile, M. Thompson, Tetrahedron 1991, 47,
1329 ± 1342.
[83] C. A. G. Haasnot, F. A. A. M. De Leeuw, C. Altona, Tetrahedron
1980, 36, 2783 ± 2792.
[65] L. Yan, D. Kahne, J. Am. Chem. Soc. 1996, 118, 9239 ± 9248 and
references cited therein.
[84] P. P. Lanhorst, C. A. G. Haasnot, C. E. Erkelens, C. Altona, J. Biomol.
Struct. Dyn. 1984, 1, 1387 ± 1405.
[66] R. Liang, L. Yan, J. Loebach, M. Ge, Y. Uozomi, K. Sekanina, N.
Horan, J. Gildersleeve, C. Thompson, A. Smith, K. Biswas, W. C. Still,
D.Kahne, Science 1996, 274, 1520 ± 1522.
[67] Z. Zhang, I. R. Ollmann, X.-S. Ye, R. Wischnat, T. Baasov, C.-H.
Wong, J. Am. Chem. Soc. 1999, 121, 734 ± 753 and references cited
therein.
[85] D. Neuhaus, M. Williamson, The Nuclear Overhauser Effect in
Structural and Conformational Analysis, VCH, New York, 1989.
[86] Y. Nishida, H. Hori, H. Ohrui, H. Meguro, J. Carbohydr. Chem. 1988,
7, 239 ± 250.
[87] P. C. Manor, W. Saenger, D. B. Davis, K. Jankowski, A. Rabazenko,
Biochim. Biophys. Acta 1974, 340, 472 ± 483.
[68] S. Raghavan, D. Kahne, J. Am. Chem. Soc. 1993, 115, 1580 ± 1581.
[69] I. Alonso, N. Khiar, M. Martín-Lomas, Tetrahedron Lett. 1996, 37,
1477 ± 1480.
[70] K. Toshima, K. Tatsuta, Chem. Rev. 1993, 93, 1503 ± 1531.
[71] H. Kunz, A. Harreus, Liebigs Ann. Chem. 1982, 41 ± 53.
[72] S. Sato, S. Nonumura, T. Nakano, Y. Ito, T. Ogawa, Tetrahedron Lett.
1988, 29, 4097 ± 4100.
[73] P. H. Seeberger, M. Eckhardt, C. E. Gutteridge, S. J. Danishefsky, J.
Am. Chem. Soc. 1997, 119, 10064 ± 10072.
[74] C. Thompson, M. Ge, D. Kahne, J. Am. Chem. Soc. 1999, 121, 1237 ±
1244.
[88] K. Stott, J. Keeler, Q. N. Van, A. J. Shaka, J. Magn. Reson. 1997, 125,
302 ± 324.
[89] J. I. Padron, E. Q. Morales, J. T. Vazquez, J. Org. Chem. 1998, 63,
8247 ± 8258.
[90] M. Martín-Pastor, unpublished results.
Â
[91] M. Martin-Lomas, P. M. Nieto, N. Khiar, S. García, M. Flores-
Mosquera, E. Poirot, J. Angulo, J. L. MunÄoz, Tetrahedron: Asymmetry
2000, 11, 37 ± 51.
[92] K. Lilley, C. L. Zhang, C. Villar-Palasí, J. Larner, L. Huang, Arch.
Biochem. Biophys. 1992, 296, 170 ± 174.
[93] S. Kunjara, D. Y. Wang, A. L. Gleenbaum, P. McLean, A. Kurtz, T. W.
Rademacher, Mol. Gen. Metab. 1999, 68, 488 ± 502.
[75] D. Crich, J. Mataka, S. Sun, K.-C. Lam, A. L. Rheingold, D. J. Wink,
Chem. Commun. 1998, 2763 ± 2764.
[76] J. Smrt, J. Catlin, Tetrahedron Lett. 1970, 58, 5081 ± 5082.
Received: February 16, 2000 [F2303]
Chem. Eur. J. 2000, 6, No. 19
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0947-6539/00/0619-3621 $ 17.50+.50/0
3621