Xylopyranoside-based agonists of D-myo-inositol 1,4,5-trisphosphate receptors: Synthesis and effect of stereochemistry on biological activity

Article abstract of DOI:10.1016/S0008-6215(01)00067-2

The synthesis of a series of tetrahydrofuranyl α- and β-xylopyranoside trisphosphates, designed by excision of three motifs of adenophostin A is reported. The synthetic route features improved preparations of allyl α-D-xylopyranoside and its 2-O-benzyl ether, and gives access to four diastereoisomeric trisphosphates, which show a range of abilities to mobilise Ca²⁺ from the intracellular stores of hepatocytes. A comparison of the potencies of the four trisphosphates provides useful information relating to the effects of stereochemical variation on the recognition of carbohydrate-based trisphosphates by D-myo-inositol 1,4,5-trisphosphate receptors. 1-O-[(3′S,4′R)-3-hydroxytetrahydrofuran-4-yl] α-D-xylopyranoside 3,4,3′-trisphosphate (8) is the most active member of the series with a potency close to Ins(1,4,5)P₃; a β-linked analogue, 1-O-[(3′R,4′S)-3-hydroxytetrahydrofuran-4-yl] β-D-xylopyranoside 3,4,3′-trisphosphate, is ca. 20-fold weaker than Ins(1,4,5)P₃, and the other compounds are much less active. While no compound attained a potency close to that of adenophostin A, we believe that 8 represents the minimal structure for potent Ca²⁺-releasing activity in this type of carbohydrate-based analogue.

Full text of DOI:10.1016/S0008-6215(01)00067-2

www.elsevier.nl/locate/carres
Carbohydrate Research 332 (2001) 53–66
Xylopyranoside-based agonists of
D-myo-inositol
1,4,5-trisphosphate receptors: synthesis and effect of
stereochemistry on biological activity
Heidi J. Rosenberg,a Andrew M. Riley,a Rachel D. Marwood,a Vanessa Correa,b
Colin W. Taylor,b Barry V.L. Pottera,*
aWolfson Laboratory of Medicinal Chemistry, Department of Pharmacy and Pharmacology, Uni6ersity of Bath,
Cla6erton Down, Bath BA2 7AY, UK
bDepartment of Pharmacology, Tennis Court Road, Uni6ersity of Cambridge, Cambridge CB2 1QJ, UK
Received 23 September 2000; accepted 18 January 2001
Abstract
The synthesis of a series of tetrahydrofuranyl a- and b-xylopyranoside trisphosphates, designed by excision of three
motifs of adenophostin A is reported. The synthetic route features improved preparations of allyl a- -xylopyranoside
D
and its 2-O-benzyl ether, and gives access to four diastereoisomeric trisphosphates, which show a range of abilities
to mobilise Ca2+ from the intracellular stores of hepatocytes. A comparison of the potencies of the four
trisphosphates provides useful information relating to the effects of stereochemical variation on the recognition of
carbohydrate-based trisphosphates by
hydrofuran-4-yl] a- -xylopyranoside 3,4,3%-trisphosphate (8) is the most active member of the series with a potency
close to Ins(1,4,5)P3; a b-linked analogue, 1-O-[(3%R,4%S)-3-hydroxytetrahydrofuran-4-yl] b- -xylopyranoside 3,4,3%-
D-myo-inositol 1,4,5-trisphosphate receptors. 1-O-[(3%S,4%R)-3-hydroxytetra-
D
D
trisphosphate, is ca. 20-fold weaker than Ins(1,4,5)P3, and the other compounds are much less active. While no
compound attained a potency close to that of adenophostin A, we believe that 8 represents the minimal structure for
potent Ca2+-releasing activity in this type of carbohydrate-based analogue. © 2001 Elsevier Science Ltd. All rights
reserved.
Keywords: Adenophostin; myo-Inositol; Selective protection
1. Introduction
-myo-Inositol 1,4,5-trisphosphate [Ins(1,4,5)-
P3, 1] is firmly established as a second messen-
ger that binds to its own intracellular receptor
to cause release of Ca2+ from intracellular
stores.1 In recent years many Ins(1,4,5)P3 ana-
logues have been prepared and from their
biological evaluation has come an understand-
ing of some of the structural motifs responsi-
ble for receptor binding and Ca2+ mobilising
activity.2,3 In 1993 Takahashi and co-workers
isolated two glyconucleotides, adenophostins
A and B (2 and 3) from Penicillium bre6icom-
pactum.4 Analyses of both Ca2+ mobilisation
by adenophostins A and B and their binding
to Ins(1,4,5)P3 receptors suggest they are full
agonists with affinities 10–100 times greater
than Ins(1,4,5)P3.4–11 Adenophostins were the
first examples of agonists of Ins(1,4,5)P3 re-
ceptors that were not based on inositol, al-
though there are several important similari-
D
* Corresponding author. Fax: +44-1225-826114.
E-mail address: b.v.l.potter@bath.ac.uk (B.V.L. Potter).
0008-6215/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(01)00067-2

54
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
ties in structure between Ins(1,4,5)P3 and the
adenophostins. Thus, the 3%%,4%%-(bisphos-
phate)/2%%-hydroxyl grouping of the adeno-
phostins is thought to mimic the crucial 4,5-
(bisphosphate)/6-hydroxyl triad of Ins(1,4,5)-
P3. However, the obvious differences in the
structure of the adenophostins and Ins(1,4,5)-
P3 and the unusual potencies of the former
have stimulated considerable interest in syn-
thesis aimed at identifying the structural basis
for the enhanced activity of adenophostins.
Three total syntheses of adenophostin A have
been reported.11–14
rating a
D
-glucopyranosyl 3,4-bisphosphate
moiety with an a-glycosidic linkage to a sec-
ond sugar containing one or more phosphates
were designed and synthesised. They all
proved to be full agonists, but with a range of
activities.7,8 The most potent was methyl 3-O-
(a -
D
- glucopyranosyl) - b -
D
- ribofuranoside
2,3%,4% trisphosphate (ribophostin, 6)8,18,19 in
which the location of the 2-phosphate group
in space is restricted by a ribofuranoside ring,
as in the adenophostins. The potency of 6 in
Ca2+ release assays was similar to that of
Ins(1,4,5)P3, but still about 20-fold less than
To determine whether the adenine compo-
nent of the adenophostins was essential for
activity, we and others designed and synthe-
that of adenophostin A. The lower affinity of
6 suggests that the adenine component of the
adenophostins is not essential for activity, but
enhances affinity for the Ins(1,4,5)P3 receptor.
This led to further simplification and the syn-
thesis of 1-O-[(3%S,4%R)-3-hydroxytetrahydro-
sised 2-hydroxyethyl a- -glucopyranoside
D
2%,3,4-trisphosphate (4),15,16 and another group
synthesised the xylopyranoside equivalent (5)17
as simplified analogues of the adenophostins.
Both 4 and 5 were shown to be full agonists of
Ins(1,4,5)P3 receptors but with 10–20-fold
lower affinity than Ins(1,4,5)P3. They were,
however, more potent than many inositol-
based Ins(1,4,5)P3 mimics and showed that the
design of chiral Ca2+-releasing polyphos-
phates based on carbohydrates is possible.
furan-4-yl] a- -glucopyranoside 3,4,3%-trispho-
D
sphate (furanophostin, 7)9,20 which lacks both
the O-methyl and the 4-hydroxymethyl moi-
eties but retains the rigidity of the five-mem-
bered ring. Again the activity of 7 was found
to be similar to Ins(1,4,5)P3 and to 6. Recently
we designed and synthesised a- and b-
D
-glu-
The lower affinity of
4
compared to
copyranosylmethanol 3,4,1%-trisphosphate
adenophostin A was thought to be due, at
least in part, to the conformational flexibility
of the ethylphosphate side chain,15 and so
various phosphorylated disaccharides incorpo-
analogues based on 4 but with a shorter side
chain.21 In these molecules the third phos-
phate group is attached to a carbon centre
fixed in the a- and b-positions. However, bio-

H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
55
logical evaluation showed that the a-C-ana-
logue was only comparable to 4, while the
b-C-analogue was even weaker. This finding
suggests that the second ring of 6 and 7 is
required for greater affinity.
ready been prepared and reported,22 the
-xylopyranoside
(12) was restricted by the unavoidable forma-
tion of two butane diacetal protected regioiso-
mers at the first step of the synthetic route. A
overall yield from allyl a-
D
In the present work we have taken the
simplification of 7 a step further by replacing
the glucopyranosyl 3,4-bisphosphate structure
in 7 with the xylopyranosyl equivalent to give
8, a new and potent xylopyranoside-based ag-
onist of Ins(1,4,5)P3 receptors. We suggest
that 8 represents the minimal structure in car-
bohydrate derived polyphosphates sufficient
for high affinity agonist activity of Ins(1,4,5)P3
receptors. In order to explore the effect of
stereochemical variations on the biological ac-
tivity of 8, and in particular the positioning of
the non-vicinal phosphate group for potent
activity, we have also synthesised three of its
diastereoisomers 9, 10 and 11. The series of
trisphosphates 8–11 may be regarded as con-
formationally restricted analogues of the origi-
nal xylopyranoside-based analogue 5 and its
b-linked equivalent.17 A comparison of the
abilities of the four stereoisomers to release
Ca2+ from permeabilised hepatocytes pro-
vides insights into the influence of stereochem-
istry on biological activity in analogues of this
type.
different route was therefore investigated, with
the aim of increasing the overall yield of 12
while reducing the need for time-consuming
purifications by column chromatography.
Thus Fischer glycosidation of xylose by a
modification (see Section 4) of the previously
described method22 gave the allyl glycoside
(12) as colourless crystals in 41% yield over
three crops. The selective acylation of position
2 of 12 was achieved using trimethylacetyl
chloride (pivaloyl chloride).23 After several
systematic changes relating to equivalents
of trimethylacetyl chloride, temperature and
reaction time, the best conditions were estab-
lished (see Section 4). The required 2-O-
trimethylacetyl ester 13 could be isolated by
crystallisation in 53% yield, without the need
for purification by chromatography.
Positions 3 and 4 in 13 were then protected
with an isopropylidene acetal by reaction of
13 with 2-methoxypropene in THF in the
presence of a catalytic amount of p-toluene-
sulphonic acid.24 The product 14 of this reac-
tion was used immediately in the next step
without purification. The trimethylacetyl
group was removed and replaced by benzyl
protecting groups using sodium hydride and
benzyl bromide in DMF and again the
product of these reactions 16 was used with
2. Results and discussion
Although the required regioselectively pro-
tected
D
-xylopyranose 18 (Scheme 1) had al-

56
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
no purification. Finally the isopropylidene
group was removed by stirring 16 with
HCl in MeOH for 30 min to give allyl 2-O-
tion to glycosylate each of two enantiomeric
alcohols. Thus, zinc chloride and silver per-
chlorate promoted glycosylation25 of (+)-
(3R,4S) - 4 - p - methoxybenzyloxy - tetrahydro-
furan-3-ol20 and of (−)-(3S,4R)-4-p-methoxy-
benzyloxy-tetrahydrofuran-3-ol20 with the gly-
benzyl-a- -xylopyranoside (17) in high yield
D
(76% over the four steps from 13) after purifi-
cation by flash chromatography followed by
crystallisation. Compound 17 has been pre-
pared previously but was reported to be an
oil.22 The diol 17 was easily converted in high
yield into 2-O-benzyl-3,4-bis-O-(p-methoxy-
cosyl phosphite gave
a
total of four
diastereoisomeric products 19–22. It is inter-
esting to note that, in each case, a mixture of
a- and b-coupled products was formed using
this xylopyranosyl phosphite glycosyl donor,
whereas only the a-coupled product was de-
tected when the same conditions were applied
to the analogous glucopyranosyl phosphite in
the synthesis of furanophostin (7).20 The a and
b epimers could not be separated at this stage.
However, after removal of the PMB ethers
using 10% trifluoroacetic acid in CH2Cl2, the
resulting a- and b-linked products could, in
each case, be separated then by flash chro-
matography followed by crystallisation to give
the diastereosiomeric triols 23–26. Phosphity-
lation of each triol with 1H-tetrazole-acti-
benzyl)-a,b- -xylopyranose (18), a selectively
D
protected intermediate without the labile trans
isopropylidene group; alkylation of 17 with
sodium hydride and p-methoxybenzyl chloride
in DMF gave the fully protected product,
from which the allyl protection at the
anomeric position was then removed using a
catalytic amount of PdCl2 in MeOH to give
xylopyranose 18.
Activation of 18 with dimethyl N,N-di-
ethylphosphoramidite in the presence of 1H-
tetrazole gave a glycosyl phosphite (Scheme
2), which was used without further purifica-
vated
bis(benzyloxy)diisopropylaminophos-
phine in CH2Cl2 and oxidation of the interme-
diate phosphites with m-chloroperbenzoic acid
(MCPBA) then gave the fully protected
trisphosphates 27–30, respectively. Deprotec-
tion of each by hydrogenation over palladium
on carbon yielded the target trisphosphates
8–11, respectively, which were purified by ion-
exchange chromatography on Q Sepharose
Fast Flow resin and isolated as their triethy-
lammonium salts. Finally, each trisphosphate
was accurately quantified by total phosphate
assay26 before biological evaluation in perme-
abilised rat hepatocytes.
Maximally effective concentrations of
Ins(1,4,5)P3, 8 or 11 released the same fraction
of the intracellular Ca2+ stores (4792, 5195
and 4692%, respectively, Fig. 1 and Table 1).
Furthermore, combined application of 10 mM
Ins(1,4,5)P3 with 10 mM of either 8 or 11
released no more Ca2+ than either of the
agonists applied alone. The concentration of 8
required to cause half-maximal Ca2+ release
(EC50) was only three- to fourfold higher than
that for Ins(1,4,5)P3, while 11 was consider-
ably weaker. The responses to both agonists
were positively co-operative. Interestingly,
there was no significant interaction between 9
Scheme 1. (i) AllOH, HCl, reflux, 16 h; (ii) (CH3)3COCl,
pyridine, −40 °C, 2.5 h; (53%); (iii) 2-methoxypropene,
PTSA, THF, 30 min; (iv) NaOH, MeOH, reflux, 1 h; (v)
NaH, BnBr, DMF, 0 °C, 90 min; (vi) 1 M HCl:MeOH 1:10,
rt, 30 min (76%, over four steps); (vii) (a) NaH, PMBCl,
DMF, rt, 12 h (84%); (b) PdCl2, MeOH, 0 °C to rt, 4 h (90%).
All, allyl; Piv, (CH3)3CCO, (pivaloyl); PMB, p-methoxyben-
zyl; Bn, benzyl.

H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
57
,
Scheme 2. (i) (MeO)2PNEt2, 1H-tetrazole, CH2Cl2; (ii) AgClO4, ZnCl2, dioxane, toluene, 4 A sieves, glycosyl acceptor, (70–76%);
(iii) CF3COOH, CH2Cl2 (18–34%); (iv) (a) (BnO)2PNPr2i , 1H-tetrazole, CH2Cl2; (b) MCPBA, −78 °C to rt (73–99%); (v) H2,
Pd–C, 50 psi, MeOH–H2O, 24 h (65–85%).
(the b epimer of 8) and Ins(1,4,5)P3 receptors;
even at 10 mM 9 failed to cause significant
Ca2+ mobilisation and nor did it affect the
response to a submaximal concentration of
Ins(1,4,5)P3 (Table 2). Finally, the results for
10 are consistent with it being a very low
affinity full agonist (EC50 \10 mM); alone at
10 mM it released less than 50% of the
Ins(1,4,5)P3-sensitive Ca2+ stores (Table 1)
and its effects were approximately additive
ribophostin (6) and furanophostin (7). This
shows that the 5-hydroxymethyl moiety of the
glucose ring in 6 and 7 is not essential for
with
a
submaximal concentration of
Ins(1,4,5)P3 when added in combination with
it (Table 2).
These results demonstrate that both 8 and
11 behave as full agonists in this assay and
that diastereoisomer 8 has a potency in Ca2+
release assays similar to that of Ins(1,4,5)P3,
Fig. 1. Effects of Ins(1,4,5)P3 and xylopyranosides 8 and 11
on 45Ca2+ release from the Ins(1,4,5)P3-sensitive Ca2+ stores
of permeabilised hepatocytes.

58
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
Table 1
45Ca2+ release data for Ins(1,4,5)P3, 8, 9, 10 and 11 a
Stereochemistry
EC50 (nM)
h
n
% release at 10 mM
Ins(1,4,5)P3
8
9
10
11
14496
487958
nd
1.6890.26
2.5191.00
nd
5
3
5
5
3
4792
5195
393
1292
4692
a,3%S,4%R
b,3%S,4%R
a,3%R,4%S
b,3%R,4%S
nd
nd
26949276
1.7890.43
a The EC50 values and Hill coefficients (h) were separately determined for n independent experiments by fitting results to logistic
equations. Results are shown as means9S.E.M.
Clearly, the stereochemical differences be-
tween analogues 8, 9, 10 and 11 have a signifi-
cant effect on the recognition of these
molecules by Ins(1,4,5)P3 receptors. The struc-
tural simplicity of 8–11 suggested that a
molecular modelling approach (see Section 4)
might be used to account for the impact of
stereochemical variation on their potencies. It
is important to appreciate, however, that
modelling of adenophostin A10,15 and even of
the simpler analogues 8–11 presents many
difficulties, including those of simulating the
combined influence of charged phosphate
groups, solvent, counter-ions, pseudorota-
tional equilibria in five-membered rings, and
anomeric effects associated with the glycosidic
linkage. It is also likely that the conformation
of these molecules is sensitive to the ionisation
state28,29 of the three phosphate groups under
physiological conditions. With these reserva-
tions in mind, a consideration of predicted
low-energy conformations of 8–11 can be
used to suggest qualitative explanations for
their differing potencies.
potent activity at Ins(1,4,5)P3 receptors, al-
though the very gradually decreasing potency
in the series 6\7\8 suggests that each hy-
droxymethyl group in 6 may have a slight
enhancing effect on activity. The results also
confirm that, at the receptor binding site, the
orientation of 6 and 7 relative to that of
Ins(1,4,5)P3 is such that the 5-hydroxymethyl
group of the glucose moiety mimics the 3-hy-
droxyl of Ins(1,4,5)P3. Had an inverted bind-
ing orientation been involved (in which case
the 5-hydroxymethyl group would be equiva-
lent to the 6-OH of Ins(1,4,5)P3), then a dra-
matic difference in activity between the
glucopyranoside-based analogues (6, 7) and
their xylopyranoside-equivalent (8) would
have been expected, analogous to the large
difference between Ins(1,4,5)P3 and 6-deoxy-
Ins(1,4,5)P3. The fact that 8 (similar to 6 and
7) is still 10–100-fold less potent than the
adenophostins supports previous arguments
that the adenine component of the
adenophostins plays a pivotal role in their
activity. The similar activities of 1, 6 and 7
suggest that 8, with its single hydroxyl group,
is the minimal structure for potent agonism in
a simple carbohydrate-derived agonist of
Ins(1,4,5)P3 receptors. Because this remaining
OH group is equivalent to the crucial 6-OH of
Ins(1,4,5)P3, its removal would undoubtedly
result in a dramatic decrease in potency. In-
deed, while this work was in progress a report
appeared27 on a synthesis of the 2-deoxy
derivative of 5. This molecule, having no re-
maining hydroxyl group, was some 2000-fold
less potent than Ins(1,4,5)P3 in mobilising in-
tracellular Ca2+, showing the expected paral-
lel with deletion of the 6-OH group of
Ins(1,4,5)P3.2
Table 2
45Ca2+ release data for combined stimulation with
Ins(1,4,5)P3, 9 and 10 a
% Ca2+ release n
Ins(1,4,5)P3 (150 nM)
9 (10 mM)
Ins(1,4,5)P3 (150 nM) with 9 (10 mM) 3093
10 (10 mM)
Ins(1,4,5)P3 (150 nM) with 10 (10
mM)
2294
494
3
3
3
3
3
1393
3491
a The percentage of the intracellular Ca2+ stores released
by a submaximal concentration of Ins(1,4,5)P3 alone or in
combination with 9 and 10 are shown. Results are shown as
means9S.E.M for three independent experiments.

H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
59
Fig. 2. X-ray crystal structure of Ins(1,4,5)P3 derived from its complex with the PH domain of phospholipase C-l1 compared with
low energy conformers of diastereoisomers 8–11 as determined by molecular modelling experiments (see Section 4).
The most stable conformers found for 8–11
are shown in Fig. 2, together with the confor-
mation of Ins(1,4,5)P3 identified in an X-ray
crystallographic study of the PH domain of
phospholipase C-l1 in complex with
Ins(1,4,5)P3.30 For diastereoisomer 8, a partic-
ularly well-defined global minimum energy
conformation was identified, similar to that
reported for adenophostin A.† In this con-
former, the tetrahydrofuranyl ring is posi-
tioned below the plane of the xylopyranosyl
ring, presumably in an orientation that avoids
steric clashes with the receptor binding pocket
and does not disrupt binding of the 3%-phos-
phate group. This phosphate group is located
in a similar region of space to the 1-phosphate
group of Ins(1,4,5)P3, although in a slightly
more extended position. Thus, as has been
suggested for adenophostin A,10,15 the non-vic-
inal phosphate in 8 may be held close to some
ideal position for optimal interaction with the
region of the receptor binding site that nor-
mally accommodates the 1-phosphate group
of Ins(1,4,5)P3. As in the case of adenophostin
A,10 a local minimum conformer of 8 (corre-
sponding to a 2%-endo conformation of
adenophostin A) with slightly higher energy
and a different puckering of the five-mem-
bered ring was also found.
In low energy conformations of the other
diastereoisomers 9–11, the 3%-phosphate
group is held in different regions of space
relative to the xylopyranosyl ring. Presumably
the 3%-phosphate groups in the less active iso-
mers 9–11 are, to different degrees, less able
to interact effectively with the Ins(1,4,5)P3 re-
ceptor binding pocket, and can only approach
the ideal position and orientation at greater
energetic costs than can the 3%-phosphate
group in 8. The almost negligible activity of 9
is striking, particularly when it is compared to
that of the relatively potent b-linked analogue
11. A possible explanation for this difference,
in addition to the factors already discussed, is
that in low energy conformations of 9, such as
that shown in Fig. 2, regions of the hydropho-
bic tetrahydrofuranyl ring may interfere with
binding to the receptor.
3. Conclusions
We have described the synthesis of a series
of novel tetrahydrofuranyl xylopyranoside-
based
trisphosphates
related
to
the
adenophostins and to Ins(1,4,5)P3. The most
active member of the series (8) is of compara-
ble potency to Ins(1,4,5)P3 and to the previ-
ously described ribophostin (6)
and
† It should noted, however, that this conformation of 8
furanophostin (7), from which 8 is a logical
development. Further simplification would in-
evitably result in a substantial decrease in
potency, and analogue 8 is therefore likely to
represent the simplest possible structure for
resembles
a 3%-endo conformation of adenophostin A
1
whereas, on the basis of H NMR evidence, the 2%-endo form
of adenophostin A is thought to predominate in solution. It
has been suggested that the 2%-endo conformation might be
essential to the high potency of adenophostin A.10

60
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
potent Ca2+-releasing activity in this type of
carbohydrate-based analogue. The weaker ac-
tivities of the other three diastereoisomers 9,
10 and 11 may be related to the effect of
differing stereochemistry at three stereogenic
centres on the spatial positioning of the non-
Modelling was carried out using the SYBYL
software package (v6.3, Tripos Associates) on
a Silicon Graphics workstation. The TRIPOS
force field was used, with added parameters
for carbohydrates.31 Charges were calculated
using a semi-empirical method (MNDO) for
structures with fully ionised phosphate
groups, and the resulting charged structures
were energy minimised, terminating at a gradi-
vicinal
phosphate
group
and
tetra-
hydrofuranyl ring.
ent of 0.05 kcal mol−1
−1, and using a
,
A
distance-dependent dielectric of 4r. Each
structure was then subjected to 100 cycles of
simulated annealing, with heating at 300 K for
2 ps followed by exponential cooling to 0 K
over 10 ps during each cycle, generating low-
energy conformations for each analogue.
These conformers were then further investi-
gated by systematic searching about the tor-
sions ƒ and „ of the glycosidic linkage to
create relaxed maps, allowing local and global
minimum energy conformers for each ana-
logue 8–11 to be identified.
4. Experimental
General methods.—Chemicals were pur-
chased from Aldrich, Sigma and Fluka. DMF
was distilled from barium oxide under reduced
,
pressure and then stored over 4 A molecular
sieves. Pyridine, CH2Cl2, THF, toluene and
dioxane were purchased in anhydrous form.
Ins(1,4,5)P3 was from American Radiolabeled
Chemicals. TLC was performed on precoated
plates (E. Merck aluminium sheets silica 60
F254, Art. No. 5554). Products were visualised
by dipping plates into phosphomolybdic acid
in MeOH followed by heating. Flash chro-
matography was carried out on silica gel (par-
ticle size 40–63 mm). 1H and 13C NMR spectra
were recorded on JEOL JMN GX-270 or
EX-400 NMR spectrometers. Unless other-
wise stated, chemical shifts were measured in
45Ca2+ release from permeabilised rat hepa-
tocytes.—Permeabilised hepatocytes were
loaded to steady state (5 min at 37 °C) with
45Ca2+ in a cytosol-like medium (CLM: KCl,
140 mM; NaCl, 20 mM; 2 mM MgCl2; 1 mM
EGTA; 300 mM CaCl2; 20 mM pipes; pH 7.0)
containing ATP (1.5 mM), creatine phosphate
(5 mM) creatine phosphokinase (five units per
mL) and FCCP (10 mM). After 5 min, thapsi-
gargin (1.25 mM) was added to the cells (still
at 37 °C) to inhibit further Ca2+ uptake; 30 s
later the cells were added to appropriate con-
centrations of the agonists and after a further
60 s the 45Ca2+ contents of the stores were
determined by rapid filtration. Concentra-
tion–response relationships were fitted to a
four parameter logistic equation using
KALEIDAGRAPH software (Synergy Software,
PA) from which the maximal response, half-
maximally effective agonist concentration
(EC50) and Hill slope (h) were determined. All
results are expressed as means9SEM. CLM,
cytosol-like medium; EGTA, ethylene glycol-
bis(b-aminoethyl ether)N,N,N%,N%-tetraacetic
acid; PIPES, piperazine-N,N%-bis[2-ethanesul-
fonic acid]; ATP, adenosine 5%-triphosphate;
31
ppm relative to internal tetramethylsilane. P
NMR chemical shifts were measured in ppm
and denoted positive downfield from external
85% H3PO4. Mps (uncorrected) were deter-
mined using a Reichert–Jung Thermo Galen
Kofler block. Microanalysis was carried out at
the University of Bath Microanalysis Service.
FAB mass spectra (m-nitrobenzyl alcohol
(mNBA)) were recorded at the University of
Bath Mass Spectrometry Service using a VG
analytical Autospec mass spectrometer. Opti-
cal rotations were measured at ambient tem-
perature using an Optical Activity Ltd. AA-10
polarimeter, and [h]D values are given in 10−1
deg cm2 g−1. Ion-exchange chromatography
was performed on an LKB-Pharmacia
medium
pressure
ion-exchange
chro-
matograph using Q Sepharose Fast Flow resin
and gradients of triethylammonium hydrogen
carbonate (TEAB) as eluent. Compounds con-
taining phosphates were assayed by a modifi-
cation of the Briggs phosphate test.26
FCCP,
carbonyl
cyanide
p-trifuoro-
methoxyphenylhydrazone; EC50, concentra-
tion causing half the maximal effect; h, Hill
coefficient.

H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
61
Allyl
acetyl-h-
3,4-O-isopropylidene-2-O-trimethyl-
-xylopyranoside (14).—A solution
Preparation of allyl h- -xylopyranoside
D
(12)22.—Allyl alcohol (1000 mL) and acetyl
D
of 13 (5.0 g, 18 mmol) in THF (50 mL) was
stirred with p-toluenesulphonic acid (20 mg)
and 2-methoxypropene (3.49 mL, 36 mmol) at
rt for 30 min under N2. TLC (EtOAc) indi-
cated consumption of the starting material
and showed the presence of one product. The
mixture was diluted with Et2O (50 mL) and
washed with satd aq NaHCO3 (30 mL). The
organic layer was dried (MgSO4) and filtered,
a few drops of Et3N were added before the
organic layer was concentrated. The residue
was used directly in the next step without
purification. A small sample was purified for
analysis by flash chromatography (1:7 eluent
Et2O–hexane); [h]2D0 +143° (c 0.4, CHCl3); 1H
NMR (CDCl3; 400 MHz) lH 5.80–5.90 (m, 1
chloride (6.0 mL) were stirred together for 30
min, after which time
D
-xylose (100.00 g, 0.67
mol) was added. The mixture was heated at
reflux for 16 h and then cooled to rt. Solid
NaHCO3 (7.00 g) was slowly added, and stir-
ring continued for 30 min. The mixture was
then filtered and the filtrate was concentrated
by evaporation under reduced pressure, leav-
ing an off-white solid. The solid was dissolved
in a minimum of EtOH (700 mL) and the
solution was kept at −20 °C for 24 h. Diiso-
propyl ether (ca 150 mL) was added immedi-
ately before the collection of the title
compound over three crops as white fluffy
crystals (52.25 g, 41%); mp 100–103 °C; Lit
mp 101–103 °C;22 [h]2D0 +140° (c 3.19,
CHCl3); Lit [h]2D0 +149° (c 3.2, CHCl3).22
3
H, CH2CHꢀCH2), 5.29 (dd, 1 H, J 17.3 Hz,
3
Allyl
2-O-trimethylacetyl-h-
D
-xylopyran-
CH2CHꢀCHcisCHtrans), 5.19 (dd, 1 H, J 10.2
oside (13).—A solution of 12 (10.0 g, 52.6
mmol) in dry pyridine (200 mL) was stirred at
−40 °C under N2. Trimethylacetyl chloride
(6.8 mL, 55.2 mmol) was added dropwise over
1 h. The mixture was stirred at −40 °C for a
further 90 min. The reaction was quenched by
adding water (100 mL) and then allowed to
reach rt and extracted with Et2O (100 mL).
The organic layer was washed with 1 M HCl
(100 mL) and satd aq NaHCO3 solution (100
mL), dried (MgSO4), filtered and concentrated
to give an oil which was dissolved in hexane
and refrigerated at −20 °C. The title com-
pound was collected as crystals over three
crops (7.6 g, 53%); mp 76–77 °C; [h]2D0 +154°
Hz, CH2CHꢀCHcisCHtrans), 5.16 (d, 1 H, J1,2
3.8 Hz, H-1), 4.82 (dd, 1 H, H-2), 4.16–4.21
(m, 1 H, CHHCHꢀCH2), 3.90–3.96 (m, 3 H,
3
H-3, H-5a, CHHCHꢀCH2), 3.76 (dd, 1 H, J
10.2 Hz, H-5b), 3.48–3.54 (m, 1 H, H-4), 1.46
(s, 6 H, 2×isopropylidene CH3) and 1.23 (s, 9
H, C(O)C(CH3)3); 13C NMR (CDCl3; 100.4
MHz) lC 177.96 (C(O)C(CH3)3), 133.50
(CH2CHꢀCH2),
117.83
(CH2CHꢀCH2),
110.86 (isopropylidene C(CH3)2), 95.10 (C-1),
75.75 (C-2), 74.12 (C-3), 72.88 (C-4), 68.96
(CH2CHꢀCH2),
61.27
(C-5),
38.77
(C(O)C(CH3)3) and 26.54, 26.81, 27.10 (2×
isopropylidene CH3, C(O)C(CH3)3); Anal.
Calcd for C16H26O6: C, 61.13; H, 8.34. Found:
C, 61.00; H, 8.29.
1
(c 1.7, CHCl3); H NMR (CDCl3; 270 MHz)
Allyl 3,4-O-isopropylidene-h- -xylopyran-
D
lH 5.77–5.89 (m, 1 H, CH2CHꢀCH2), 5.29
(dd,
1
H, 2J 1.6, 3J 17.0 Hz,
2 3
oside (15).—A solution of crude 14 (5.72 g,
18.0 mmol), NaOH (1.46 g, 36 mmol) in
MeOH (100 mL) was heated under reflux for
1 h. The mixture was cooled to rt and neu-
tralised by careful addition of solid CO2. The
solvents were evaporated off and the residue
was partitioned between Et2O (100 mL) and
water (50 mL). The aqueous layer was back-
extracted with Et2O (100 mL) and the com-
bined organic fractions were dried (MgSO4).
A few drops of Et3N were added before the
organic layer was concentrated. The residue
was used directly in the next step without
purification. For analytical purposes, a small
CH2CHꢀCHcisHtrans), 5.18 (dd, 1 H, J 1.5, J
11.7 Hz, CH2CHꢀCHcisHtrans), 4.96 (d, 1 H,
J1,2 3.7 Hz, H-1), 4.62 (dd, 1 H, J2,1 3.7, J2,3
9.9 Hz, H-2), 4.13–4.20 (m,
2
H,
CH2CHꢀCH2), 3.44–3.76 (m, 4 H, H-3, H-4,
H-52) and 1.22 (s, 9 H, C(O)C(CH3)3); 13C
NMR (CDCl3; 67 MHz) lC 178.60
(C(O)C(CH3)3), 133.57 (CH2CHꢀCH2), 117.51
(CH2CHꢀCH2), 95.28 (C-1), 73.07, 72.14,
70.44 (C-2, C-3, C-4), 68.39 (CH2CHꢀCH2),
61.42 (C-5), 38.89 (C(O)C(CH3)3) and 27.08
(C(O)C(CH3)3); Anal. Calcd for C13H22O6: C,
56.92; H, 8.08. Found: C, 56.9; H, 8.08.

62
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
CH3); 13C NMR (CDCl3; 100.4 MHz) lC
138.05 (C-1 of benzyl ether ring), 133.69
(CH2CHꢀCH2), 128.3, 127.9, 127.6 (ArCH),
117.80 (CH2CHꢀCH2), 110.54 (isopropylidene
C(CH3)2), 96.24 (C-1), 77.9, 77.91, 74.1 (C-4,
sample was purified by flash chromatography
(1:1 eluent Et2O–hexane); [h]2D0 +133° (c 0.4,
1
CHCl3); H NMR (CDCl3; 400 MHz) lH
5.87–5.97 (m, 1 H, CH2CHꢀCH2), 5.32 (dd, 1
3
H, J 17.3 Hz, CH2CHꢀCHcisCHtrans), 5.24
3
(dd, 1 H, J 10.2 Hz, CH2CHꢀCHcisCHtrans),
C-3,
C-2),
71.8
(OCH2Ar),
68.4
4.95 (d, 1 H, J1,2 3.8 Hz, H-1), 4.23–4.28 (m,
1 H, CH2CHꢀCH2), 4.02–4.07 (m, 1 H,
CH2CHꢀCH2), 3.94 (t, 1 H, J3,4 9.7 Hz, H-3),
3.83 (dd, 1 H, H-2), 3.66–3.74 (m, 2 H, H-5),
3.38–3.44 (m, 1 H, H-4), 2.43 (d, 1 H, J 10.8
Hz, 2-OH) and 1.45, 1.46 (2s, 6 H, 2×iso-
propylidene CH3); 13C NMR (CDCl3; 100.4
MHz) lC 133.35 (CH2CHꢀCH2), 118.07
(CH2CHꢀCH2), 61.1 (C-5) and 26.8, 26.4 (2×
isopropylidene CH3); Anal. Calcd for
C18H24O5: C, 67.48; H, 7.55. Found C,
67.40%; H, 7.53.
Preparation of allyl 2-O-benzyl-h- -xylopy-
D
ranoside (17)22.—A solution of 16 (5.86 g, 18
mmol) in MeOH (25 mL) was stirred with 1
M HCl (2.5 mL) for 30 min. TLC (4:1 Et2O–
hexane) indicated consumption of starting ma-
terial. Solid NaHCO3 was added until the
mixture was neutral. The solvents were evapo-
rated off and the residue was partitioned be-
tween CH2Cl2 (50 mL) and water (25 mL).
The aqueous layer was back-extracted with
CH2Cl2 (100 mL) and the combined organic
layers were dried (MgSO4), filtered and con-
centrated. The syrup obtained was subjected
to flash chromatography (4:1 eluent Et2O–
hexane then Et2O) to give the title compound
as a colourless syrup (3.87 g, 76% over the
previous four steps) which slowly crystallised;
mp 46–47 °C (from isopropyl ether); Lit.22 oil.
[h]2D0 +115° (c 2.3, CHCl3); Lit [h]2D0 +113° (c
2.3, CHCl3).22
(CH2CHꢀCH2),
110.70
(isopropylidene
C(CH3)2), 97.50 (C-1), 79.3 (C-2), 73.6 and
71.9 (C-4, C-3, C-2), 68.84 (CH2CHꢀCH2),
61.57 (C-5) and 26.40, 26.73 (2×isopropyli-
dene CH3); Anal. Calcd for C11H18O5: C,
57.38; H, 7.94. Found C, 57.10; H, 7.94.
Allyl 2-O-benzyl-3,4-O-isopropylidene-h- -
D
xylopyranoside (16).—A solution of crude 15
(4.18 g, 18 mmol) in dry DMF (25 mL) was
stirred at 0 °C with NaH (760 mg of an 60%
w/w dispersion in mineral oil, 19 mmol) and
BnBr (2.17 mL, 19 mmol) was added slowly
under N2. The mixture was stirred at rt for 90
min, after which time TLC (1:2 Et2O–hexane)
showed complete consumption of the starting
material. Water (25 mL) was added carefully
and stirring continued for 60 min. The sol-
vents were evaporated and the residue was
dissolved in Et2O (50 mL). The extract was
washed with water (25 mL), dried (MgSO4),
filtered and concentrated. The product of this
reaction was used directly in the next step
without purification. A small sample was
purified for analysis by flash chromatography
(1:9 eluent Et2O–hexane); [h]2D0 +30.6° (c 0.5,
Preparation of 2-O-benzyl-3,4-bis-O-(p-
methoxybenzyl)-
lution of allyl
(p-methoxybenzyl)-a-
D
-xylopyranose (18)22.—A so-
2-O-benzyl-3,4-bis-O-
D
-xylopyranoside (6.87
g, 13.2 mmol), obtained from 17 as previously
described,22 in MeOH (60 mL) was cooled to
0 °C and PdCl2 (0.47 g, 2.64 mmol) was
added. The flask was fitted with a drying tube,
the cooling bath removed and the reaction
mixture was stirred vigorously for 4 h, after
which time TLC (3:2 ether–hexane) indicated
almost complete conversion of starting mate-
rial into a product (Rf 0.24). The reaction
mixture was quenched with Et3N and filtered
through a Celite pad. The filtrate was concen-
trated and the dark brown residue subjected
to flash chromatography (3:2 eluent ether–
hexane) to yield the title compound (mixture
of a and b anomers) as a white solid (5.72 g,
90%).
1
CHCl3); H NMR (CDCl3; 400 MHz) lH
7.24–7.39 (m, 5 H, ArCH), 5.86–5.96 (m, 1
3
H, CH2CHꢀCH2), 5.34 (dd, 1 H, J 17.3 Hz,
3
CH2CHꢀCHcisCHtrans), 5.21 (dd, 1 H, J 10.2
Hz, CH2CHꢀCHcisCHtrans), 4.83–4.86 (m, 2 H,
H-1, OCH2Ar), 4.64 (AB, 1 H, JAB 12.3 Hz,
OCH2Ar), 4.15–4.20 (m, 1 H, CH2CHꢀCH2),
3.91–4.00 (m, 2 H, H-3, CH2CHꢀCH2), 3.86
2
3
(dd, 1 H, J 9.7, J 4.7 Hz, H-5a), 3.73 (dd, 1
H, J 10.2 Hz, H-5b), 3.65 (dd, 1 H, J2,1 3.5,
J2,3 10.2 Hz, H-2), 3.35–3.41 (m, 1 H, H-4)
and 1.46, 1.47 (2s, 6 H, 2×isopropylidene

H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
63
1-O-[(3%S,4%R)-3-hydroxytetrahydrofuran-4-yl]
2-O-benzyl-i- -xylopyranoside (24).—A solu-
1 - O - [(3%S,4%R) - 3 - (p - Methoxybenzyloxy)-
tetrahydrofuran-4-yl] 2-O-benzyl 3,4-di-O-(p-
methoxybenzyl)-h and i- -xylopyranosides (19
D
D
tion of the mixture of 19 and 20 (1.60 g, 2.32
mmol) and TFA (2 mL) in CH2Cl2 (20 mL)
was stirred for 20 min, TLC (10:1 CH2Cl2–
acetone) showed consumption of the starting
material. Satd aq NaHCO3 was added to neu-
tralise the reaction. The mixture was extracted
with EtOAc (4×100 mL) and the combined
organic extracts were dried (MgSO4), filtered
and concentrated. The oil obtained was sub-
jected to flash chromatography (20:1 eluent
EtOAc–MeOH) to give the title compound 23
which was crystallised from EtOAc–hexane
(261 mg, 34%); mp 119–120 °C (EtOAc–hex-
and 20).—To a stirred mixture of 18 (2.0 g,
4.6 mmol) and 1H-tetrazole (440 mg, 6.25
mmol) in CH2Cl2 (10 mL) was added bis(-
methoxy)(diethylamino)phosphine (0.9 mL,
5.4 mmol). After 20 min, TLC (1:4 EtOAc–
toluene) indicated complete conversion into
product (Rf 0.63). The reaction mixture was
partitioned between Et2O (100 mL) and water
(75 mL). The resulting ethereal layer was
washed with brine (75 mL), dried (MgSO4),
filtered and concentrated to give crude 2-O-
benzyl-3,4-bis-O-(p-methoxybenzyl)- -xylo-
D
pyranosyl dimethyl phosphite14 as a colourless
oil (2.38 g), which was used without further
purification. A solution of the dimethyl phos-
phite (2.38 g), (+)-(3R,4S)-4-p-methoxyben-
zyloxy-tetrahydrofuran-3-ol20 (747 mg, 3.3
1
ane); [h]2D0 +87.2° (c 0.5, CHCl3); H NMR
(CDCl3; 400 MHz) lH 7.26–7.37 (m, 5 H, Ar),
4.77 (d, 1 H, J1,2 2.9 Hz, H-1), 4.75, 4.71 (AB,
2 H, JAB 11.7 Hz, ArCH2O), 4.25–4.26 (m, 1
H, H-3%), 4.07–4.11 (m, 1 H, H-4%), 3.96 (dd, 1
H, J5%b,4% 5.9, J5%b,5%a 9.4 Hz, H-5%b), 3.88–3.91
(m, 2 H, H-3, H-2%b), 3.80 (dd, 1 H, J5%a,4% 4.4
Hz, H-5%a), 3.68–3.72 (m, 1 H, H-2%a), 3.51–
3.65 (m, 3 H, H-4, H-5) and 3.35 (dd, 1 H, J2,3
,
mmol) and 4 A molecular sieves (3 g) in (21:7
mL) dioxane–toluene was stirred for 2 h un-
der N2. Zinc chloride (679 mg, 4.99 mmol)
and silver perchlorate (2.04 g, 9.88 mmol)
were added to the reaction mixture which was
stirred in the dark for a further 2 h, TLC (4:1
toluene–EtOAc) showed consumption of the
starting material. Sodium bicarbonate (3 g),
EtOAc (100 mL) and water (75 mL) were
added to the mixture and it was stirred vigor-
ously for a further 30 min. The mixture was
filtered through a Celite pad and the organic
layer was washed with brine (100 mL), dried
(MgSO4), filtered and concentrated. The oil
obtained was subjected to flash chromatogra-
phy (7:3 eluent EtOAc–hexane) to give a ꢀ
13
8.8 Hz, H-2); C NMR (CDCl3; 100.4 MHz)
lC 137.34 (C-1 of benzyl ether ring), 128.92,
128.73, 128.68 (3×Ar), 98.82 (C-1), 79.40,
79.00 (2×CH), 74.59 (CH2), 73.78 (CH),
73.13 (CH2), 71.11 (CH), 71.11 (ArCH2O),
70.28 (CH) and 62.41 (C-5); Anal. Calcd for
C16H22O7: C, 58.88; H, 6.79. Found C, 58.80;
H, 6.76.
Further elution gave compound 24 as an
oil (150 mg, 20%); [h]2D0 −8.33° (c 0.6,
1
CHCl3); H NMR (CDCl3; 400 MHz) lH
1
7.26–7.34 (m, 5 H, Ar), 4.85, 4.65 (AB, 2 H,
JAB 11.4 Hz, ArCH2O), 4.45 (d, 1 H, J1,2 7.0
Hz, H-1), 4.19–4.25 (m, 2 H, H-4%, H-3%),
3.94–3.98 (m, 3 H, H-5b, H-2%b, H-5%b), 3.77–
3.80 (m, 2 H, H-2%a, H-5%a), 3.62–3.63 (m, 1
H, H-4), 3.45–3.52 (m, 1 H, H-3), 3.23–
3.30 (m, 4 H, H-2, H-5a, 2×OH) and 3.18
2:1 (judged by H NMR) anomeric mixture of
the title compounds (1.60 g, 70%).
1 - O - [(3%R,4%S) - 3 - (p - Methoxybenzyloxy)-
tetrahydrofuran-4-yl] 2-O-benzyl 3,4-bis-O-(p-
methoxybenzyl)-h and i- -xylopyranosides (21
D
and 22).—The title compounds (1.75 g, 76%,
1
ꢀ1:1 anomeric mixture judged by H NMR)
13
(broad s, 1 H, OH); C NMR (CDCl3; 100.4
were obtained from 2-O-benzyl-3,4-bis-O-(p-
MHz) lC 137.96 (C-1 of benzyl ether ring),
128.84, 128.36, 128.22 (Ar), 104.02 (C-1),
81.43, 80.63 and 75.75 (3×CH), 74.96, 73.29
(CH2), 71.04 (CH), 70.19 (ArCH2O), 69.51
(CH) and 65.60 (C-5); MS: (FAB) m/z Calcd
for C16H22O7, [M+H]+327.1443. Found
327.1455.
methoxybenzyl)-
D
-xylopyranosyl
dimethyl
phosphite (2.38 g, 4.16 mmol) and (−)-
(3S,4R) - 4 - p - methoxybenzyloxy - tetrahydro-
furan-3-ol20 (747 mg, 3.3 mmol) as described
for the synthesis of 19 and 20.
1-O-[(3%S,4%R)-3-Hydroxytetrahydrofuran-
4-yl] 2-O-benzyl-h- -xylopyranoside (23) and
D

64
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
upon triol 23 (83 mg, 0.25 mmol) was added
1-O-[(3%R,4%S)-3-Hydroxytetrahydrofuran-
4-yl] 2-O-benzyl-h- -xylopyranoside (25) and
1-O-[(3%R,4%S)-3-hydroxytetrahydrofuran-4-yl]
2-O-benzyl-i- -xylopyranoside (26).—Com-
D
and stirring was continued for 30 min. The
mixture was cooled to −78 °C and MCPBA
(863 mg, 60%, 3.0 mmol) was added. The
mixture was stirred at rt for 10 min, then was
diluted with CH2Cl2 (50 mL). The solution
was washed successively with 10% (w/v) aq
Na2SO3, satd aq NaHCO3 (25 mL) and satd
aq NaCl (25 mL), dried (MgSO4), filtered and
concentrated. The residue was purified by
flash chromatography (4:1 eluent EtOAc–hex-
ane) to give the title compound as an oil (279
D
pounds 25 (145 mg, 18%) and 26 (155 mg,
19%) were obtained from the mixture of 21
and 22 (1.75 g, 2.50 mmol) as described for
the synthesis of 23 and 24.
25: mp 144–145 °C; [h]2D0 +132° (c 0.2,
1
CDCl3); H NMR (CDCl3; 400 MHz) lH
7.25–7.35 (m, 5 H, Ar), 4.74 (d, 1 H, J1,2 3.5
Hz, H-1), 4.66, 4.57 (AB, 2 H, JAB 11.4 Hz,
ArCH2O), 4.22–4.25 (m, 1 H, H-3%), 4.12–
4.16 (m, 1 H, H-4%), 3.89–3.99 (m, 2 H, H-3,
H-5%b), 3.85 (dd, 1 H, J2%b,3% 5.6, J2%b,2%a 9.4 Hz,
H-2%b), 3.72–3.76 (m, 2 H, H-2%a, H-5%a),
3.46–3.63 (m, 3 H, H-4, H-5) and 3.33 (dd, 1
H, J2,3 9.4 Hz, H-2); 13C NMR (CDCl3; 67
MHz) lC 137.28 (C-1 of benzyl ether ring),
128.63, 128.43 (ArCH), 95.08 (C-1), 79.44,
76.53 (2×CH), 74.79 (CH2), 73.49 (CH),
72.57 (CH2), 71.40, 69.63 (2×CH), 68.89
(ArCH2O) and 61.85 (C-5); Anal. Calcd for
C16H22O7: C, 58.88; H, 6.79. Found: C, 58.70;
H, 6.79.
1
mg, 99%); [h]2D0 +0.3° (c 1.9, CHCl3); H
NMR (CDCl3; 400 MHz) lH 7.04–7.39 (m, 35
H, Ar), 4.66–5.06 (m, 16 H, H-1, H-3, H-3%,
6.5×ArCH2O), 4.31–4.47 (m, 2 H, H-4,
0.5×ArCH2O), 4.08–4.12 (m, 1 H, H-4%),
3.84–3.96 (m, 4 H, H-5a, H-5b, H-2%b, H-5%b),
3.79 (dd, 1 H, J 5%a,4%5.3, J5%a,5%b 9.4 Hz, H-5%a)
and 3.43–3.49 (m, 2 H, H-2, H-2%a); 13C NMR
(CDCl3; 67 MHz) lC 137.589 (C-1 of benzyl
ether ring), 136.02–135.50 (C-1 of benzyl ester
rings), 128.50, 128.50, 128.48, 128.41, 128.61,
128.22, 128.13, 128.06, 127.94, 127.88, 127.71,
127.60, 127.49 (ArCH), 95.68 (C-1), 77.09,
75.35, 75.26, 74.03 and 73.93 (5×CH), 71.97,
69.99 (2×CH2), 68.91–69.71 (ArCH2O) and
59.69 (C-5); 31P NMR (CDCl3; 162 MHz) lp
−0.70, −0.62 and 0.11; MS: (FAB) m/z
Calcd for C58H61O16P3 [M+H]+ 1107.3250.
Found 1107.3233.
Further elution gave compound 26 which
was crystallised from EtOAc–hexane; mp
1
166–167 °C; [h]2D0 −28.3° (c 0.4, MeOH); H
NMR (CD3OD; 400 MHz) lH 7.27–7.43 (m, 5
H, Ar), 4.95, 4.81 (AB, 2 H, JAB 11.1 Hz,
ArCH2O), 4.55 (d, 1 H, J1,2 7.6 Hz, H-1),
4.23–4.31 (m, 2 H, H-2%b and H-5%b), 3.94–
3.98 (m, 1 H, H-3% or H-4%), 3.86–3.9 (m, 2 H,
H-3, H-2%a or H-5%a), 3.76–3.80 (m, 1 H, H-3%
or H-4%), 3.63 (dd, 1 H, J 4.1, Jgem 9.1 Hz,
H-2%a or H-5%a), 3.46–3.57 (m, 2 H, H-4,
H-5b) and 3.21–3.3 (m, 2 H, H-2, H-5a); 13C
NMR (CDCl3; 67 MHz) lC 140.21 (C-1 of
benzyl ether ring), 129.70, 129.54 and 128.96
(Ar), 105.10 (C-1), 83.02, 80.18, 78.02 (3×
CH), 75.98, 73.66 (2×CH2), 72.23 (ArCH2O),
72.13, 71.60 (2×CH) and 67.17 (C-5); Anal.
Calcd for C16H22O7: C, 58.88; H, 6.79. Found:
C, 58.50; H, 6.80.
1-O-[(3%S,4%R)-3-Hydroxytetrahydrofuran-
4-yl] 2-O-benzyl-i- -xylopyranoside 3,4,3%-
D
tris(dibenzylphosphate) (28).—Compound 28
(89 mg, 87%) was obtained from 24 (30 mg,
0.09 mmol) as described above for the synthe-
sis of 27.
[h]2D0 −10.4° (c 1.4, CHCl3); 1H NMR
(CDCl3; 400 MHz) lH 7.02–7.34 (m, 35 H,
Ar), 4.75–5.14 (m, 14 H, H-3%, ArCH2O),
4.49–4.58 (m, 2 H, H-3, ArCH2O), 4.39 (d, 1
H, J1,2 7.0 Hz, H-1), 4.31–4.35 (m, 1 H, H-4%),
4.23–4.30 (m, 1 H, H-4), 4.19 (m, 1 H, H-5b),
3.85–3.93 (m, 3 H, H-5%b, H-2%a, H-2%b), 3.74
(dd, 1 H, J5%a,4% 6.5 Hz, H-5%a) and 3.23–3.28
1-O-[(3%S,4%R)-3-Hydroxytetrahydrofuran-
4-yl] 2-O-benzyl-h- -xylopyranoside 3,4,3%-
D
13
(m, 2 H, H-2, H-5a); C NMR (CDCl3; 100.4
tris(dibenzylphosphate) (27).—A mixture of
bis(benzyloxy)(diisopropylamino)phosphine
(518 mg, 1.50 mmol), 1H-tetrazole (158 mg,
2.25 mmol) and dry CH2Cl2 (5 mL) was vigor-
ously stirred at rt for 30 min under N2, where-
MHz) lC 137.89 (C-1 of benzyl ether ring),
136.12–135.64 (C-1 of benzyl ester ring),
128.83, 128.80, 128.75, 128.61, 128.59, 128.46,
128.23, 128.21, 128.12, 128.03, 127.97, 127.91,

H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
65
138.11 (C-1 of benzyl ether ring), 136.13–
135.71 (6×C-1 of benzyl ester ring), 128.88,
128.82, 128.80, 128.75, 128.61, 128.56, 128.44,
128.41, 128.34, 128.32, 128.25, 128.22, 128.20,
128.05, 128.00, 127.89, 127.59 (ArCH), 102.86
(C-1), 79.20, 79.10 (2×CH2), 76.32, 76.26,
75.58, 75.54 and 74.29 (5×CH), 73.79–69.43
(7×ArCH2O) and 62.89 (C-5); 31P NMR
(CDCl3; 162 MHz) lp −0.25, −0.83 and
−1.07; MS: (FAB) m/z Calcd for
C58H61O16P3 [M+H]+ 1107.3250. Found
1107.3232.
127.81 (Ar), 103.09 (C-1), 79.55, 79.26, (2×
CH), 76.40 (CH2), 74.58, 73.86, 71.36 (3×
CH), 70.10 (CH2), 68.87–70.04 (7×ArCH2O)
31
and 63.42 (C-5); P NMR (CDCl3; 162 MHz)
lp −0.76, −0.60 and −0.43; MS: (FAB)
m/z Calcd for C58H61O16P3 [M+H]+
1107.3238. Found 1107.3238.
1-O-[(3%R,4%S)-3-Hydroxytetrahydrofuran-
4-yl] 2-O-benzyl-h- -xylopyranoside 3,4,3%-
D
tris(dibenzylphosphate) (29).—Compound 29
(210 mg, 73%) was obtained from 25 (85 mg,
0.23 mmol) as described above for the synthe-
sis of 27.
1-O-[(3%S,4%R)-3-Hydroxytetrahydrofuran-
4-yl] h- -xylopyranoside 3,4,3%-trisphosphate
D
[h]2D0 +9.4° (c 0.3, CHCl3); 1H NMR
(CDCl3; 400 MHz) lH 7.13–7.40 (m, 35 H,
Ar), 4.87–5.16 (m, 13 H, H-1, 6×ArCH2O),
4.77–4.82 (m, 1 H, H-3%), 4.66–4.69 (m, 2 H,
H-3, ArCH2O), 4.4–4.49 (m, 2 H, H-4,
ArCH2O), 4.10–4.18 (m, 1 H, H-2%a), 3.92–
4.03 (m, 2 H, H-5a, H-5b), 3.85–3.91 (m, 3 H,
H-2%b, H-4%, H-5%b), 3.74 (dd, 1 H, J5%a,4% 6.3,
J5%a,5%b 9.1 Hz, H-5%a) and 3.51 (dd, 1 H, J2,1
3.5, J2,3 9.7 Hz, H-2); 13C NMR (CDCl3; 100.4
MHz) lC 137.80 (C-1 of benzyl ether ring),
136.13–134.49 (6×C-1 of benzyl ester rings),
132.95, 130.17, 128.72, 128.70, 128.63, 128.55,
128.43, 128.36, 128.29, 128.22, 128.16, 128.01,
127.83 (Ar), 96.90 (C-1), 78.53, 77.99, 76.32,
76.26, 75.92 (5×CH), 74.14, 73.87 (CH2)
70.79–70.79 (OCH2Ar) and 60.33 (C-5); 31P
NMR (CDCl3; 109 MHz)lp −0.55, −0.80
and −0.95; MS: (FAB) m/z Calcd for
C58H61O16P3 (M+1)+ 1107.3250. Found
1107.3252.
(8).—10% palladium on activated charcoal
(200 mg) was added to a solution of com-
pound 27 (94 mg, 0.084 mmol) in MeOH (20
mL) and water (5 mL) and the mixture was
shaken under an atmosphere of H2 at 50 psi at
rt for 24 h. The suspension was filtered and
washed well with de-ionised water. The com-
bined filtrate was concentrated to a glassy
solid. The residue was dissolved in de-ionised
water (300 mL) and purified by ion-exchange
chromatography on Q Sepharose Fast flow
resin, eluting with a gradient of TEAB buffer
(0–1 M), pH 8. The triethylammonium salt of
8 eluted between 68 and 82%. Fractions con-
taining the title compound, as judged by phos-
phate assay, were combined and evaporated
to give a residue from which MeOH (3×100
mL) was evaporated to give the title trisphos-
phate as its triethylammonium salt (0.071
1
mmol, 85%); [h]2D0 +30.2° (c 1.3, MeOH); H
NMR (D2O; 400 MHz) lH 4.99 (d, 1 H, J1,2
3.5 Hz, H-1), 4.58–4.61 (m, 1 H, H-3%), 4.25–
4.29 (m, 1 H, H-4%), 4.14–4.22 (m, 1 H, H-3),
3.91–3.97 (m, 1 H, H-4), 3.82–3.90 (m, 2 H,
H-2%a, H-5%a), 3.67–3.75 (m, 3 H, H-2%b, H-
5%b, H-5a), 3.58 (dd, 1 H, J2,3 8.8 Hz, H-2) and
3.45–3.52 (m, 1 H, H-5b); 31P NMR (D2O;
162 MHz) lP 1.05 and 0.89 (2P); MS: (FAB)
m/z Calcd for C9H18O16P3 [M−H]− 474.9807.
Found 474.9801.
1-O-[(3%R,4%S)-3-Hydroxytetrahydrofuran-
4-yl] 2-O-benzyl-i- -xylopyranoside 3,4,3%-
D
tris(dibenzylphosphate) (30).—Compound 30
(107 mg, 79%) was obtained from 26 (30 mg,
0.09 mmol) as described above for the synthe-
sis of 27.
[h]2D0 −10.8° (c 1.2, CHCl3); 1H NMR
(CDCl3; 400 MHz) lH 7.04–7.37 (m, 35 H,
Ar), 4.71–5.02 (m, 14 H, H-3%, 6.5×
ArCH2O), 4.52–4.61 (m, 3 H, H-1, H-3, 0.5×
ArCH2O), 4.29–4.37 (m, 1 H, H-4), 4.23 (dd,
1 H, J 5.3, J 10.2, H-2%a), 4.10 (dd, 1 H, J 5.3,
J 12.0, H-5b), 3.89–3.97 (m, 2 H, H-4%, H-5%a),
3.78–3.82 (m, 2 H, H-2%b, H-5%b), 3.36–3.40
(m, 1 H, H-2) and 3.14 (dd, 1 H, J 9.4, J 12.0,
H-5a); 13C NMR (CDCl3; 100.4 MHz) lC
1-O-[(3%S,4%R)-3-Hydroxytetrahydrofuran-
4-yl] i- -xylopyranoside 3,4,3%-trisphosphate
D
(9).—The title compound (0.049 mmol, 77%)
as its triethylammonium salt was obtained
from compound 28 (72 mg, 0.064 mmol) as
described above for the synthesis of 8.
1
[h]2D0 −17.7° (c 1.2, MeOH); H NMR
(CD3OD; 400 MHz) lH 4.54–4.60 (m, 1 H,

66
H.J. Rosenberg et al. / Carbohydrate Research 332 (2001) 53–66
References
H-3%), 4.37–4.42 (m, 2 H, H-1, H-4%), 3.92–
3.97 (m, 3 H, H-5a, H-5b, H-3), 3.83–3.88
(m, 2 H, H-2%b, H-5%b), 3.69–3.73 (m, 2 H,
H-2a%, H-5%a), 3.34–3.38 (m, 1 H, H-2) and
[1] Berridge, M. J. Nature 1993, 361, 315–325.
[2] Potter, B. V. L.; Lampe, D. Angew. Chem. Int. Ed. Engl.
1995, 34, 1933–1972.
[3] Wilcox, R. A.; Primrose, W. U.; Nahorski, S. R.; Challiss,
R. A. J. Trends Pharmacol. Sci. 1998, 19, 467–475.
[4] Takahashi, S.; Kinoshita, T.; Takahashi, M. J. Antibiot.
1994, 47, 95–100.
[5] Takahashi, M.; Tanzawa, K.; Takahashi, S. J. Biol. Chem.
1994, 269, 369–372.
[6] Hirota, J.; Michikawa, T.; Miyawaki, A.; Takahashi, M.;
Tanzawa, K.; Okura, I.; Furuichi, T.; Mikoshiba, K. FEBS
Lett. 1995, 368, 248–252.
[7] Murphy, C. T.; Bullock, A. J.; Lindley, C. J.; Mills, S. J.;
Riley, A. M.; Potter, B. V. L.; Westwick, J. Mol. Pharma-
col. 1996, 50, 1223–1230.
[8] Marchant, J. S.; Beecroft, M. D.; Riley, A. M.; Jenkins,
D. J.; Marwood, R. D.; Taylor, C. W.; Potter, B. V. L.
Biochemistry 1997, 36, 12780–12790.
[9] Shuto, S.; Tatani, K.; Ueno, Y.; Matsuda, A. J. Org. Chem.
1998, 63, 8815–8824.
[10] Hotoda, H.; Murayama, K.; Iwata, Y.; Takahashi, M.;
Kawase, Y.; Tanzawa, K.; Keneko, M. Biochemistry 1999,
38, 9234–9241.
[11] Marwood, R. D.; Correa, V.; Taylor, C. W.; Potter, B. V.
L. Tetrahedron: Asymmetry 2000, 11, 397–403.
[12] Hotoda, H.; Takahashi, M.; Tanzawa, K.; Takahashi, S.;
Kaneko, M. Tetrahedron Lett. 1995, 36, 5037–5040.
[13] van Straten, N. C. R.; van der Marel, G. A.; van Boom,
J. H. Tetrahedron 1997, 53, 6509–6522.
[14] Marwood, R. D.; Riley, A. M.; Jenkins, D. J.; Potter, B.
V. L. J. Chem. Soc., Perkin Trans. 1 2000, 1935–1947.
[15] Wilcox, R. A.; Erneux, C.; Primrose, W. U.; Gigg, R.;
Nahorski, S. R. Mol. Pharmacol. 1995, 47, 1204–1211.
[16] Jenkins, D. J.; Potter, B. V. L. J. Chem. Soc., Chem.
Commun. 1995, 1169–1170.
31
3.23–3.31 (m, 1 H, H-4); P NMR (CD3OD;
162 MHz) lp 0.81, 1.05, 1.20; MS: (FAB)
m/z Calcd for C9H18O16P3 [M−H]−
474.9807. Found 474.9801.
1-O-[(3%R,4%S)-3-Hydroxytetrahydrofuran-
4-yl] h- -xylopyranoside 3,4,3%-trisphosphate
D
(10).—The title compound as its triethylam-
monium salt (0.064 mmol, 65%) was ob-
tained from compound 29 (110 mg, 0.099
mmol) as described above for the synthesis
1
of 8. [h]2D0 +40.8° (c 1.5, MeOH); H NMR
(D2O; 270 MHz) lH 4.79 (d, 1 H, J1,2 3.7 Hz,
H-1), 4.53–4.57 (m, 1 H, H-3%), 4.10–4.23
(m, 2 H, H-3, H-4%), 3.63–3.95 (m, 7 H, H-5a,
H-5b, H-5%a, H-5%b, H-2%a, H-2%b) and
3.50 (dd, 1 H, J2,3 9.2 Hz, H-2); 31P NMR
(D2O; 109 MHz) lp 0.34 and 0.80 (2P); MS:
(FAB) m/z Calcd for C9H18O16P3 [M−H]−
474.9807. Found 474.9802.
1-O-[(3%R,4%S)-3-Hydroxytetrahydrofuran-
4-yl] i- -xylopyranoside 3,4,3%-trisphosphate
D
(11).—The title compound (0.071 mmol,
85%) as its triethylammonium salt was ob-
tained from compound 30 (94 mg, 0.084
mmol) as described above for the synthesis
[17] Moitessier, N.; Chre´tien, F.; Chapleur, Y.; Humeau, C.
Tetrahedron Lett. 1995, 36, 8023–8026.
[18] Desai, T.; Gigg, J.; Gigg, R. Carbohydr. Res. 1996, 280,
209–221.
1
of 8. [h]2D0 −21.1° (c 0.52, MeOH); H NMR
(CD3OD; 400 MHz) lH 4.78–4.81 (m, 1 H,
H-3%), 4.54 (d, 1 H, J1,2 7.6 Hz, H-1), 4.40–
4.45 (m, 1 H, H-4%), 4.19–4.3 (m, 1 H, H-
2%b), 4.06–4.08 (m, 1 H, H-3), 3.92–3.98 (m,
1 H, H-5%b), 3.86 (dd, 1 H, J2%a,3% 3.5, J2%a,2%b
9.6, H-2%a), 3.73 (dd, 1 H, J5%a,4% 6.2, J5%a,5%b
8.8, H-5%a), 3.34–3.44 (m, 2 H, H-2, H-4)
and 3.01–3.08 (m, 3 H, H-5a, H-5b, OH);
31P NMR (CD3OD; 162 MHz) lp 0.71, 1.32
and 1.47; MS: (FAB) m/z Calcd for
C9H18O16P3 [M−H]− 474.9807. Found
474.9815.
[19] Jenkins, D. J.; Marwood, R. D.; Potter, B. V. L. J. Chem.
Soc., Chem. Commun. 1997, 449–450.
[20] Marwood, R. D.; Riley, A. M.; Correa, V.; Taylor, C. W.;
Potter, B. V. L. Bioorg. Med. Chem. Lett. 1999, 9, 453–458.
[21] Rosenberg, H. J.; Riley, A. M.; Correa, V.; Taylor, C. W.;
Potter, B. V. L. Carbohydr. Res. 2000, 329, 7–16.
[22] Jenkins, D. J.; Potter, B. V. L. J. Chem. Soc., Perkin
Trans.1 1998, 41–49.
[23] Jiang, L.; Chan, T.-H. J. Org. Chem. 1998, 63, 6035–6038.
[24] Cai, J.; Davison, B. E.; Ganellin, R. C.; Thaisrivongs, S.
Tetrahedron Lett. 1995, 36, 6535–6536.
[25] Ogawa, T.; Takahashi, Y.; Matsui, M. Carbohydr. Res.
1982, 102, 207–215.
[26] Sawyer, D. A.; Potter, B. V. L. J. Chem. Soc., Perkin
Trans.1 1992, 923–932.
[27] Roussel, F.; Hilly, M.; Chre´tien, F.; Mauger, Y.; Chapleur,
Y. J. Carbohydr. Chem. 1999, 697–707.
[28] Felemez, M.; Marwood, R. D.; Potter, B. V. L.; Spiess, B.
Biochem. Biophys. Res. Commun. 1999, 266, 334–340.
[29] Felemez, M.; Bernard, P.; Schlewer, G.; Spiess, B. J. Am.
Chem. Soc. 2000, 122, 3156–3165.
[30] Ferguson, K. M.; Lemmon, M. A.; Schlessinger, J.; Sigler,
P. B. Cell 1995, 83, 1037–1046.
[31] Pe´rez, S.; Meyer, C.; Imberty, A. In Modelling of Biomolec-
ular Structures and Mechanisms; Pullman, A.; Jortner, J.;
Pullman, B., Eds.; Kluwer: Dordrecht, 1995; pp. 425–454.
Acknowledgements
We thank the Wellcome Trust for Prize
Studentships (H.J.R. and R.D.M.) and Pro-
gramme Grant Support (B.V.L.P. and
C.W.T.).