Synthesis of D- and L-myo-inositol 1,2,4,6-tetrakisphosphate, regioisomers of myo-inositol 1,3,4,5 tetrakisphosphate: Activity against Ins(1,4,5)P 3 binding proteins

Article abstract of DOI:10.1039/b302986g

We report here the synthesis of D- and L-myo-inositol 1,2,4,6-tetrakisphosphate 3a and 3b and the racemic modification 3ab. Racemic myo-inositol 1,2,4,6-tetrakisphosphate 3ab was synthesised from DL-1,2,4,6-tetra-O-allylmyo-inositol 9ab. Benzylation and d

Full text of DOI:10.1039/b302986g

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of D- and L-myo-inositol 1,2,4,6-tetrakisphosphate,
regioisomers of myo-inositol 1,3,4,5 tetrakisphosphate:
activity against Ins(1,4,5)P3 binding proteins
Stephen J. Mills,a Katrien Backers,b Christophe Erneuxb and Barry V. L. Potter*a
a Wolfson Laboratory of Medicinal Chemistry, Department of Pharmacy and Pharmacology,
University of Bath, Claverton Down, Bath, UK BA2 7AY
b Institut de Recherche Interdisciplinaire (IRIBHM), Université Libre de Bruxelles,
Campus Erasme, Bldg. C, 808 route de Lennik, B-1070 Brussels, Belgium
Received 17th March 2003, Accepted 18th August 2003
First published as an Advance Article on the web 18th September 2003
We report here the synthesis of - and -myo-inositol 1,2,4,6-tetrakisphosphate 3a and 3b and the racemic
modification 3ab. Racemic myo-inositol 1,2,4,6-tetrakisphosphate 3ab was synthesised from -1,2,4,6-tetra-O-allyl-
myo-inositol 9ab. Benzylation and de-allylation provided the tetraol 11ab, which was phosphitylated in the presence
of bis(benzyloxy)diisopropylaminophosphine and 1H-tetrazole, then oxidised to give the fully protected 1,2,4,6-
tetrakisphosphate 13ab. Hydrogenolysis of 13ab and purification of product by ion exchange chromatography
gave racemic myo-inositol 1,2,4,6-tetrakisphosphate 3ab, which showed no demonstrable agonism or antagonism
for Ca2ϩ release at 200 µM in permeabilised hepatocytes. The chiral derivatives, -3a and -myo-inositol 1,2,4,6-
tetrakisphosphate 3b were synthesised from 5-O-benzyl-1,4,6-tri-O-p-methoxybenzyl-myo-inositol 19ab, which was
resolved using R-(Ϫ)-O-acetylmandelic acid providing two diastereoisomers 21 and 22 which were separated and
deacylated to give the corresponding enantiomers. Further transformations gave the corresponding chiral 1,2,4,6-
tetraols which were phosphitylated, oxidised, deprotected and purified as for the racemic mixture. The enantiomeric
tetrakisphosphates 3a and 3b were evaluated for inhibition of the metabolic enzymes inositol 1,4,5-trisphosphate
5-phosphatase and 3-kinase in comparison with the enantiomers of another synthetic regioisomer - and -myo-
inositol 1,2,4,5-tetrakisphosphate. Both - and -myo-inositol 1,2,4,6-tetrakisphosphate inhibit 5-phosphatase with
an IC50 value of 3.8 µM and 14 µM, repectively. However, both enantiomers were poorly recognised by the 3-kinase
enzyme, with IC50 values greater than 100 µM. The enantiomers of the 1,2,4,5-tetrakisphosphate showed the same
relative pattern of activity towards the two enzymes but were more potent against 5-phosphatase (0.47 µM and 3 µM
respectively).
myo-inositol 1,3,4,5-tetrakisphosphate [Ins(1,3,4,5)P4 2a] which
acts as a potent bi-modal regulator of cellular sensitivity to
Introduction
-myo-Inositol 1,4,5-trisphosphate [Ins(1,4,5)P3 1a] (Fig. 1) is a
ubiquitous hydrophilic second messenger, which elevates the
levels of cytostolic Ca2ϩ ions.1 After Ca2ϩ release, Ins(1,4,5)P3 is
deactivated and the cell returns to a basal state. This is achieved
in two ways: first, an Ins(1,4,5)P3 5-phosphatase removes
the 5-phosphate group from Ins(1,4,5)P3 to give myo-inositol
Ins(1,4,5)P3 and facilitates the regulation of Ca2ϩ signalling.4
Thus, both products may be regarded as ‘off’ signals. The func-
tion of Ins(1,3,4,5)P4 has not been fully resolved; however, it
could be the natural inhibitor of 5-phosphatase in vivo.4 Studies
have shown that -Ins(1,3,4,5)P4 is the most potent inhibitor of
5-phosphatase4 (IC50 of 0.15 µM). However, the enantiomer of
2a -Ins(1,3,4,5)P4 2b is more than ten-fold weaker (IC50 of 1.8
µM).4 These data support the finding that 5-phosphatase is
inhibited by a number of tetrakisphosphates.5 However, there are
no known inhibitors which are as potent as -Ins(1,3,4,5)P4.
We have previously described6 how myo-inositol 1,4,6-tris-
phosphate, [Ins(1,4,6)P3, 4a] can mimic the Ca2ϩ releasing
activity of Ins(1,4,5)P3, by spatial inversion and rotation of
the former molecule, such that positions 4-, 1- and 6- of
Ins(1,4,6)P3 (Fig. 2) can be superimposed on positions 1-,
4- and 5- of Ins(1,4,5)P3 respectively. The hydroxyl groups at
positions 3- and 2- of Ins(1,4,6)P3, can be superimposed on
positions 2- and 3- of Ins(1,4,5)P3 respectively, but the stereo-
chemistry at these sites is inverted. Ins(1,4,6)P3 can release Ca2ϩ
from intracellular stores, albeit with reduced potency compared
to Ins(1,4,5)P3. Since Ins(1,4,6)P3 can be superimposed on
1,4-bisphosphate [Ins(1,4)P2] which does not release Ca2ϩ
.
Ins(1,4)P2 is an allosteric activator of 6-phosphofructo-1-kinase2
and also activates DNA polymerase-α.3 Second, Ins(1,4,5)P3
can be phosphorylated by an Ins(1,4,5)P3 3-kinase to give
Fig. 1
Fig. 2 Orientation of Ins(1,4,6)P3 to mimic Ins(1,4,5)P3.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
3546

View Article Online
Ins(1,4,5)P3, the phosphates and hydroxyl groups of Ins-
(1,2,4,6)P4 3a can also, in principle, be arranged in a pattern
whichbroadlyoverlapswithIns(1,3,4,5)P4 2a, byspatiallyinvert-
ing and rotating the molecule (Fig. 3). The tetrakisphosphate 3a
has a similar arrangement of phosphates as compound 2a;
however, where the 3-phosphate of Ins(1,3,4,5)P4 2a is equa-
torial, the corresponding position of 3a is axial and the equiv-
alent position-2 is now equatorial. Similarly, if the enantiomer
-Ins(1,2,4,6)P4 3b is inverted and rotated in the same way, the
new 5-position (axial 2-phosphate) of 3b corresponding to the
phosphate of 2a is axial and the 3-position is equatorial and 3b
can be superimposed on -Ins(1,3,4,5)P4 2a. The other three
phosphates are equatorial at pseudo positions 1-, 3- and 4- (Fig.
3). Thus, both -Ins(1,2,4,6)P4 3a and -Ins(1,2,4,6)P4 3b can be
superimposed on -Ins(1,3,4,5)P4 with an axial phosphate
group at position-3 for 3a or at position-5 for 3b. We report here
the synthesis of - and -myo-inositol 1,2,4,6-tetrakisphos-
phate and its racemic mixture, and the biological evaluation of
compounds 3a and 3b and Ins(1,3,4,5)P4 2a, together with
another tetrakisphosphate enantiomeric pair -Ins(1,2,4,5)P4
5a and -Ins(1,2,4,5)P4 5b, with the metabolic enzymes,
Ins(1,4,5)P3 3-kinase and 5-phosphatase.
Scheme 1 Synthesis of racemic Ins(1,2,4,6)P4. Reagents and conditions
(i) NaH, AllBr, DMF, rt, 3 h, (89%); (ii) ethanol–1 mol dmϪ3 aq. HCl
(2 : 1), reflux 3 h, (93%); (iii) AllBr, (Bu4N)2SO4, (1 eq.), CH2Cl2–5% aq.
NaOH, 1 : 1, reflux 24 h, (74%); (iv) NaH, BnBr, DMF, (88%); (v) 10%
Pd/C, p-toluene sulfonic acid, MeOH–H2O (5 : 1), reflux, (49%); (vi)
(BnO)2PNPri2, 1H-tetrazole, CH2Cl2; (vii) MCPBA (50–60%), 0 ЊC, 30
min, (91%); (viii) H2, Pd/C, MeOH–H2O (4 : 1), 20 h, then purification
by ion exchange chromatography on Q-Sepharose Fast Flow, (33%).
Fig.
3
Orientation of - and -Ins(1,2,4,6)P4 to mimic -
Ins(1,3,4,5)P4.
as a syrup, since selective allylation at positions C-1-OH or C-3-
OH leads to the same racemic product 9ab. Penta-O-allyl-myo-
inositol together with starting material were observed by TLC,
but these compounds were not isolated after chromatography.
Benzylation of diol 9ab with benzyl bromide and sodium
hydride in DMF provided the fully blocked compound
Results and discussion
myo-Inositol orthoformate 6 (Scheme 1) was prepared accord-
ing to Vasella and co-workers7 in 73% yield to give a highly
crystalline compound. Allylation of the two axial and one
equatorial hydroxyl groups was accomplished using sodium
hydride and allyl bromide in dry DMF at room temperature.
Purification of the crude product by flash chromatography pro-
vided the meso- compound 2,4,6-tri-O-allyl-myo-inositol ortho-
formate 7 as a syrup in 89% yield. The orthorformate protective
group was then removed from compound 7 by acid hydrolysis
using aq. HCl at reflux temperature to afford 2,4,6-tri-O-allyl-
myo-inositol 8 in excellent yield. When the orthoformate pro-
tecting group was removed from compound 7, the hydroxyl
groups at positions 1-, 3- and 5- were exposed. Under these
conditions it was possible to allylate selectively the more
reactive 1- or 3-hydroxyl position and provide a tetra-O-allyl
derivative that was used to prepare racemic Ins(1,2,4,6)P4 3ab.
Monoallylation of the triol 8 to give the tetra-O-allyl derivative
9ab was envisaged because the starting material was water
soluble, however, the product was soluble in the organic layer.
Thus, monoallylation would release tetra-O-allyl-myo-inositol
9ab into the organic layer and move the equilibrium in favour
of the monoallylated product. Selective allylation of the triol 8
was achieved under phase-transfer conditions,8 using tetra-
butylammonium sulfate and allyl bromide in a 5% aqueous
solution of sodium hydroxide and an equal volume of
dichloromethane at reflux temperature for 24 h. Racemic
1,2,4,6-tetra-O-allyl-myo-inositol 9ab was isolated in 74% yield
-1,2,4,6-tetra-O-allyl-3,5-di-O-benzyl-myo-inositol
10ab.
Removal of all four allyl protective groups was achieved using
10% palladium on carbon in the presence of an acid to provide
the racemic 1,2,4,6-tetraol 11ab. Compound 11ab was sub-
jected to phosphitylation using the P() reagent bis-
(benzyloxy)(diisopropylamino)phosphine9 (δP = ϩ147.9 ppm)
in the presence of excess 1H-tetrazole. The reaction of the P()
reagent in the presence of 1H-tetrazole provided a reactive
tetrazolide intermediate (δP = ϩ126.73 ppm).10 The tetraol 11ab
was added to the tetrazolide derivative to provide a tetra-
kisphosphite intermediate 12ab. The 31P NMR spectrum of the
1,2,4,6-tetrakisphosphite 12ab, operating at 36.2 MHz, and a
sweep width of 2500 KHz, showed eight assignable peaks
5
derived from several Jpp coupling systems.10 There was a
5
doublet at δP = 139.0 ppm, for 2-P, Jpp = 2.4 Hz; a doublet of
5
doublets at δP = 140 ppm for 1-P, Jpp = 4.3 and 2.4 Hz;
5
a doublet at 142.5 ppm for 6-P, Jpp = 4.3 Hz, and a singlet at
δP = 141.8 ppm for 4-P. This was definitive proof that phos-
phitylation of the three adjacent positions and the isolated
C-4-OH had occurred. The reaction mixture was then cooled
and oxidised using m-chloroperoxybenzoic acid (MCPBA) to
give the P() fully protected phosphotriester 13ab in excellent
5
yield. After oxidation, there is no Jpp coupling, thus, the
pattern of the 31P NMR spectrum gives four separate phos-
phorus singlets at ca. Ϫ1 to Ϫ3 ppm for the 31P–1H decoupled
NMR spectrum. The decabenzyl derivative 13ab was hydro-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3547

View Article Online
genolysed under pressure, in the presence of 10% palladium on
carbon. The solid components of the reaction mixture were
filtered off over a bed of Celite and the methanolic-aqueous
solution was evaporated to give a syrup. The crude compound
was dissolved in MilliQ and purified by ion-exchange chrom-
atography on Q-Sepharose Fast Flow using a gradient of tri-
ethylammonium hydrogen carbonate (TEAB) buffer. Pure
Ins(1,2,4,6)P4 3ab was obtained as the triethylammonium salt
after ion exchange chromatography, and eluted at ca. 800 mmol
dmϪ3 TEAB buffer and was quantified by phosphate analysis.11
A different approach was used to synthesise the chiral anti-
podes, 3a and 3b (Scheme 2). Racemic 5-O-benzyl-1,4,6-tris-O-
(p-methoxybenzyl)-myo-inositol 19ab was resolved using the
chiral auxillary (R)-(Ϫ)-O-acetylmandelic acid 20 in order to
give suitable intermediates for the preparation of the tetra-
kisphosphates 4a and 4b. Intermediate 19ab was prepared in
three steps from the trans diol 14ab,12 which was selectively
alkylated, at the more reactive C-6-OH.13 Thus, treatment of
diol 14ab with dibutyltin oxide, tetrabutylammonium bromide/
iodide and p-methoxybenzyl chloride in acetonitrile resulted in
three products in an approximate ratio of (8 : 4 : 1) for com-
1
pounds 16ab, 15ab and 17ab. The H NMR spectrum of 16ab
showed a broad triplet at δ = 3.32 ppm for 5-H, which was the
signal for the ring proton furthest upfield. Following D2O
exchange, the broad triplet at δ = 3.32 ppm collapsed to give a
sharp triplet, however, the D2O exchange had no effect on the
signal for 6-H or 4-H. The other products, 2,3-O-isopropyl-
idene-1,4,5-tris-O-p-methoxybenzyl-myo-inositol 15ab and
2,3-O-isopropylidene-1,4,5,6-tetrakis-O-p-methoxybenzyl-myo-
inositol 17ab were isolated in 26% and 6% yield respectively.
Racemic 5-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-ino-
sitol 19ab was prepared by benzylation of compound 16ab to
give 18ab and careful acid hydrolysis of the isopropylidene gave
the cis-diol 19ab.
Numerous methods are available to resolve myo-inositol
phosphate precursors;14 most of these procedures use a chiral
auxiliary such as an acid, that is coupled to one or more of the
hydroxyl groups on the myo-inositol ring to form a pair of
diastereoisomeric ester derivatives. The resulting two dia-
stereoisomers can be separated by flash chromatography or
crystallisation.10,13,14 Ogawa and co-workers15,16 have used (S)-
(ϩ)-O-acetylmandelic acid coupled to -1,4,5,6-tetra-O-
benzyl-myo-inositol15 to then provide precursors to previously
inaccessible hexoses,15 and β-glucosidase and α-mannosidase
inhibitors (ϩ)- and (Ϫ)-norjirimycin.16 They also used (R)-(Ϫ)-
O-acetylmandelic acid to acylate aminocyclopentane-1,2,3,4-
tetraols17 which were then used to synthesise (ϩ) and (Ϫ)-man-
nostatin A. The previous success of this reagent for resolving
racemic myo-inositol derivatives containing a free cis-diol,
prompted us to use it as our reagent of choice. The cis-diol -
5-O-benzyl-1,4,6-tris-O-(p-methoxybenzyl)-myo-inositol 19ab
was resolved by coupling to (R)-(Ϫ)-O-acetylmandelic acid 20,
in the presence of DCCI and DMAP at Ϫ20 ЊC in order to form
two equatorial substituted diastereoisomers 21 and 22. We used
the R-enantiomer of acetylmandelic acid, because from pre-
vious resolutions10,13 we predicted the less polar product (by
TLC) would lead to the -enantiomer 3a and the more polar
product should give the -enantiomer 3b. The critical resolution
step in this synthesis gave diastereoisomers 21 and 22, which
had Rf values of 0.34 and 0.24 respectively in our TLC system
(CHCl3–acetone 15 : 1). Separation and purification of the two
diastereoisomers was then accomplished using flash chromato-
graphy and crystallisation. In the NMR spectrum, the 1-H
proton from both diastereoisomers could not be identified
due to the methylene AB coupling pattern of a benzyl or
p-methoxybenzyl moiety which obscured the doublet of doub-
let signal for 1-H in both compounds. However, it was shifted
downfield due to the carbonyl deshielding effect from the ester.
The slower running diastereoisomer showed an unusual upfield
shift of a methylene AB system from the adjacent p-methoxy-
Scheme
2
Synthesis of - and -Ins(1,2,4,6)P4. Reagents and
conditions (i) Bu2SnO, CH3CN, (Bu)4NBr and (Bu)4NI, p-methoxy-
benzyl chloride, Soxhlet, reflux 40 h, (46% 16ab, 26% 15ab, 6% 17ab);
(ii) BnBr, NaH, DMF, 2 h, (86%); (iii) MeOH–1.0 mol dmϪ3 aq. HCl
(9 : 1), 50 ЊC, 30 min, (79%); (iv) (R)-(Ϫ)-O-acetylmandelic acid, DMAP,
DCCI, Ϫ20 ЊC, overnight, (32% 21, 26% 22); (v) MeOH–NaOH, reflux,
30 min, (96% 19a, 99% 19b); (vi) Bu2SnO, PhCH3, Dean-and-Stark
apparatus, 3 h. Then DMF, CsF, BnBr, rt, overnight, (83% 23a, 95%
23b); (vii) 1 Mol dmϪ3 aq. HCl–ethanol (1 : 2) reflux, 4 h, (92% 11a, 93%
11b); (viii) (BnO)2PNPri2, 1H-tetrazole, CH2Cl2, 10 min, then add
tetraols, 10 min, then add (50–60%) MCPBA, 0 ЊC, 30 min, (79% 13a,
86% 13b); (ix) 10% Pd/C, H2, MeOH–H2O (4 : 1), 20 h, then purification
by ion exchange chromatography on Q-Sepharose Fast Flow, (65% 4a,
36% 4b).
benzyl moiety at δ = 4.05 and 4.36. The myo-inositol ring pro-
tons at position (2-H) for both diastereoisomers were identified,
and their signal was indicative of unacylated products and the
equatorial C-1-OH was acylated under our experimental condi-
tions. The unique singlet at δ = 5.97 for diastereoisomer 21 and
δ = 5.99 for diastereoisomer 22 [for CH3CO2CH(Ph)CO2Ins]
indicated the purity of these compounds. The individual
diastereoisomers 21 and 22 were deacylated using methanolic
sodium hydroxide to give the enantiomers 19a and 19b which
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3548

View Article Online
had equal and opposite optical rotations. The equatorial
hydroxyl group was then selectively benzylated over the axial
hydroxyl moiety via the cis-1,2-O-dibutylstannylene derivative
which was formed by refluxing the individual enantiomers 19a
and 19b separately, in the presence of dibutyltin oxide and tolu-
ene, with continuous removal of water. The cis-1,2-O-dibutyl-
stannylene derivatives were dried then dissolved in dry DMF
followed by the addition of three equivalents of caesium fluor-
ide and the mixtures stirred under nitrogen. Benzyl bromide
was added dropwise and the reactions were stirred overnight,
after which a single product was obtained after work up and
chromatography in each case to give the individual enantiomers
23a and 23b. The three p-methoxybenzyl groups were then
removed in the presence of the benzyl ethers under acidic con-
ditions, to give - and -3,5-di-O-benzyl-myo-inositol 11a
and 11b respectively, which had equal and opposite optical
rotations.
We firmly established the absolute configuration of - and
-3,5-di-O-benzyl-myo-inositol 11a and 11b respectively, by
transforming -3,5-di-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-
myo-inositol 23b to the known compound, -2,3,5-tri-O-benzyl-
myo-inositol13 25b (Scheme 3). The 2-hydroxyl group of 23b
was benzylated to give the fully blocked intermediate -2,3,5-
tri-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol 24b.
The three p-methoxybenzyl groups were removed from 24b by
acidic hydrolysis to give triol 25b and the melting point and
optical rotation agreed well with our literature value.13 The
faster running compound (higher Rf) by TLC led to the syn-
thesis of -3,5-di-O-benzyl-myo-inositol 11a and the slower
running compound provided the -3,5-di-O-benzyl-myo-
inositol 11b. Phosphorylation of the individual chiral tetraols
was carried out in the same way as for the racemic mixture,
using the P() approach and oxidation to P() intermediate
using MCPBA. The benzyl protective groups for both fully pro-
tected tetrakisphosphate enantiomers 13a and 13b were then
removed by hydrogenolysis in the presence of 10% palladium
on carbon. The product was then purified on Q-Sepharose Fast
Flow using a gradient of TEAB as buffer to give the corre-
sponding pure tetrakisphosphates -Ins(1,2,4,6)P4 3a and
-Ins(1,2,4,6)P4 3b respectively. The specific rotations for the
chiral antipodes 3a and 3b were determined in methanol and
found to be Ϫ15.4 and ϩ15 for 3a and 3b respectively. Recently,
syntheses for compounds 3a and 3b have been reported by
Chung et al.,18 using a different route. The magnitude and sign
of the optical rotation for each enantiomer, agree fully with our
data.
experimental conditions: these included the naturally occurring
-Ins(1,3,4,5)P4 2a together with the synthetic regioisomers
-Ins(1,2,4,5)P4 5a and -Ins(1,2,4,5)P4 5b, which differ in the
position of one phosphate.10 This study is the first evaluation of
- and -Ins(1,2,4,6)P4 and - and -Ins(1,2,4,5)P4 as potential
inhibitors of 3-kinase and 5-phosphatase.
Initial studies for synthesising new analogues derived from
Ins(1,4,5)P3 focused on the analogue5 -chiro-Ins(2,3,5)P3 (26 in
Fig. 4) which interacts with the Ins(1,4,5)P3 receptor and
deactivating enzymes, 3-kinase and 5-phosphatase. This mole-
cule is similar to Ins(1,4,5)P3 but the position-3 of Ins(1,4,5)P3
is now axial and the name and numbering of the resulting
inositol phosphate derivative changes. -chiro-Ins(2,3,5)P3 26 is
a full agonist for Ca2ϩ release and inhibits both 5-phosphatase
(Ki value of 7.7 µM), [Km for Ins(1,4,5)P3 is 13.8 µM] and
3-kinase (Ki value of 0.97 µM), [Km value for Ins(1,4,5)P3 is
0.85 µM].5 It was a direct lead to Ins(1,4,6)P3 4a where the axial
hydroxyl group of -chiro-Ins(2,3,5)P3 26, corresponding to
position-2 of Ins(1,4,5)P3, is now equatorial. Ins(1,4,6)P3 4a is a
full agonist for Ca2ϩ release6 and only 2-fold less potent than
Ins(1,4,5)P3 in permeabilised platelets. An enzyme study by
Hirata et al.21 demonstrated that Ins(1,4,6)P3 4a is resistant to
5-phosphatase [Ki value 9.2 µM, (cf Ins(1,4,5)P3, Ki value 15.9
µM)]† and a potent 3-kinase inhibitor: IC50 value of 2 µM
(approx.) in the presence of 30 µM Ca2ϩ and 8 µM at < 0.01 µM
Ca2ϩ. This work shows that the hydroxyl orientation at pos-
ition-2 of Ins(1,4,5)P3 is not significant for 3-kinase recog-
nition. However, inverting the stereochemistry at position-3
of Ins(1,4,5)P3 to give -chiro-Ins(2,3,5)P3, delivers a potent
3-kinase and 5-phosphatase inhibitor and a molecule which
retains Ca2ϩ releasing properties when it binds to the
Ins(1,4,5)P3 receptor.
Fig. 4 Modifications of positions 2- and 3- around a -1,4,5-
trisphosphate environment.
The addition of a phosphate group at the equatorial pos-
ition-3 of Ins(1,4,5)P3 gives Ins(1,3,4,5)P4 2a, the most potent
5-phosphatase inhibitor (albeit a co-substrate, IC50 value of
0.15 µM).4 Kozikowski et al.23 synthesised -chiro-Ins(1,2,3,5)P4
27 which is an analogue of Ins(1,3,4,5)P4 in which a phosphate
has been substituted at position-1 of 26 to give tetrakisphos-
phate 27, resulting in a significant reduction in Ca2ϩ release for
27, which is 700-fold less potent than Ins(1,4,5)P3 at binding to
the Ins(1,4,5)P3 receptor. In this respect, 27 is similar to
Ins(1,3,4,5)P4 2a in its Ca2ϩ releasing properties.4 Ins(1,3,4,5)P4
Scheme 3 Determination of the absolute configuration of compound
23b. Reagents and conditions (i) NaH, BnBr, DMF, rt, 2 h, (82%);
(ii) 1 Mol dmϪ3 aq. HCl–ethanol (1 : 2), reflux 4 h, (88%).
Compounds - and -Ins(1,2,4,6)P4 (3a and 3b), which are
structurally similar to myo-inositol 1,3,4,5-tetrakisphosphate
2a (see Fig. 3), were evaluated in the presence of the
Ins(1,4,5)P3 metabolising enzymes, 5-phosphatase (type I from
human brain)19 and 3-kinase (type A, from rat brain, clone
C5).20 For a complete study, several other structurally related
inositol tetrakisphosphates were evaluated under the same
† IC50 is the concentration of agent inhibiting the enzyme activity by
50% at 1 µM substrate concentration. In these experiments, inhibiting
the phosphorylation of [3H] Ins(1,4,5)P3. Ki is the dissociation constant
for an enzyme–inhibitor complex calculated from IC50 values, using the
Cheng and Prusoff equation.22
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3549

View Article Online
Table 1 IC50 values determined on recombinant enzymes and calculated by non linear regression (curve fit) at 1 µM Ins(1,4,5)P3. Analogues were
evaluated over a range of concentrations 0.1–100 µM. Typical data for 5-phosphatase are given
Recombinant Ins(1,4,5)P3
5-phosphatase from human brain IC50/µM
Recombinant 3-kinase-A
from rat brain IC50/µM
Compound
-Ins(1,3,4,5)P4 2a
-Ins(1,2,4,6)P4 4a
-Ins(1,2,4,6)P4 4b
-Ins(1,2,4,5)P4 5a
-Ins(1,2,4,5)P4 5b
0.19
3.8
14
0.47
3
1.4
>100
>>100
>100
>>100
2a has been evaluated for Ca2ϩ release and the inhibition of 3-
kinase. Some literature (Pollokoff et al.24 and Gawler et al.25)
indicates that it releases Ca2ϩ from intracellular stores, but Bird
and Putney Jr.26 demonstrated that Ins(1,3,4,5)P4 was ineffect-
ive at mobilising Ca2ϩ: the older literature studies probably
used Ins(1,3,4,5)P4 contaminated with small quantities of
Ins(1,4,5)P3. Only a small number of groups have evaluated
racemic Ins(1,3,4,5)P4 against 3-kinase. They found it was a
poor inhibitor of the enzyme with IC50 values of 90 µM24 and
220 µM.27 However, in our experiments, (Table 1 and Fig. 5) -
Ins(1,3,4,5)P4 2a was a potent 3-kinase inhibitor (type-A, from
rat brain, clone C5), with an IC50 value of 1.4 µM. This may
result from the use of recombinant enzyme, different assay
conditions or the use of other Ins(1,4,5)P3 3-kinase isoenzymes.
Moreover, evaluating inositol phosphate analogues in crude
lysates may lead to degradation of some compounds due to
Ins(1,4,5)P3 5-phosphatase activity in the assay. The fact that
Ins(1,3,4,5)P4 is an inhibitor of Ins(1,4,5)P3 3-kinase is not sur-
prising and probably results from product inhibition derived
from our assay conditions.
racemic Ins(1,2,4,6)P4 3ab was synthesised by Chung and
Chang28 and was evaluated for Ca2ϩ release in Chinese Hamster
Ovary cells,29 where it mobilised only 34.7% of the intracellular
Ca2ϩ stores at 100 µM concentration. Its affinity for the
Ins(1,4,5)P3 receptor of bovine adrenal cortical membranes was
some 653-fold lower than Ins(1,4,5)P3.29 Fig. 6 shows the bio-
logically active Ins(1,4,5)P3 regioisomer Ins(1,4,6)P3 super-
imposed upon Ins(1,4,5)P3 and docked into the Ins(1,4,5)P3
binding site of the recently published crystal structure of type-I
Ins(1,4,5)P3 receptor.30 Using such docking studies we can pre-
dict that a phosphate group cannot be accommodated comfort-
ably in the axial direction at position-3 of Ins(1,4,5)P3, which
corresponds to the 2-position of Ins(1,4,6)P3 (Fig. 6). The
resulting Ins(1,2,4,6)P4 should therefore have low affinity for
the Ins(1,4,5)P3 receptor and be ineffective at Ca2ϩ release, as
demonstrated by our results above and those of others.
Fig. 6 Three dimensional structure of the Ins(1,4,5)P3 binding site of
the type 1 receptor based on the X-ray crystal structure of the mouse
receptor core binding domain in complex with Ins(1,4,5)P3.30 Molecular
docking experiments suggest that Ins(1,4,6)P3 (green) may be a
relatively effective mimic of Ins(1,4,5)P3 at the type 1 Ins(1,4,5)P3
binding site because it can bind in an orientation that allows its
phosphate groups to mimic the three phosphate groups of Ins(1,4,5)P3
while its axial 2-hydroxyl group is accepted by an open region close to
Gly-268. The region close to the 2-hydroxyl group of Ins(1,4,6)P3 is
sterically constrained to addition of another phosphate group as
in Ins(1,2,4,6)P4. For clarity, all hydrogen atoms and the six
crystallographic observed water molecules have been omitted.
Fig. 5 The curves illustrate the inhibition of recombinant Ins(1,4,5)P3
5-phosphatase from human brain by Ins(1,3,4,5)P4, - and -
Ins(1,2,4,6)P4 and - and -Ins(1,2,4,5)P4 in the presence of 1 µM
Ins(1,4,5)P3.
Using similar reasoning for the synthesis of Ins(1,4,6)P3
derived from Ins(1,4,5)P3 and -chiro-Ins(2,3,5)P3, addition
of a phosphate at position-2 of Ins(1,4,6)P3 would deliver
Ins(1,2,4,6)P4, a regioisomer of Ins(1,3,4,5)P4. The axial
2-phosphate of Ins(1,2,4,6)P4 can be superimposed on the
3-position of Ins(1,3,4,5)P4 and the axial 2-hydroxyl of
Ins(1,3,4,5)P4 and the 3-hydroxyl in Ins(1,2,4,6)P4 are
equatorial (see Fig. 3): the -enantiomer was also synthesised
and evaluated. Examination of structures 1a, 3a and 3b (Fig. 1)
illustrates that, should -Ins(1,2,4,6)P4 possess Ins(1,4,5)P3-
like Ca2ϩ mobilising activity, we would expect activity to reside
only in the -enantiomer as for the Ins(1,4,5)P3 enantiomers.14
We found that at a concentration of 200 µM there was however,
no demonstrable agonism or antagonism in permeabilised
hepatocytes for Ca2ϩ release for racemic Ins(1,2,4,6)P4 3ab and
in receptor binding it had an IC50 value of 1.06 µM in cere-
bellum {cf IC50 circa 50 nM for [3H]Ins(1,4,5)P3}. Previously,
Racemic Ins(1,2,4,6)P4 3ab, was tested for inhibition of 3-
kinase from rat brain.27 In their study, Choi and co-workers27
reported that racemic Ins(1,2,4,6)P4 showed greater inhibition
of 3-kinase (IC50 value of 42.1 µM) than Ins(1,3,4,5)P4 (IC50
value of 220 µM). Our study is the first in which both enantio-
mers, - and -Ins(1,2,4,6)P4 (3a and 3b), have been evaluated
against both 3-kinase and 5-phosphatase. -Ins(1,2,4,6)P4
(Table 1 and Fig. 5) was found to be a good inhibitor of
5-phosphatase with an IC50 value of 3.8 µM [cf Ins(1,4,6)P3 has
a Ki value of 9.2µM from another study], although it was 20-
fold weaker than Ins(1,3,4,5)P4 and not recognised by 3-kinase.
-Ins(1,2,4,6)P4 (Table 1 and Fig. 5) was a reasonable inhibitor
of 5-phosphatase with an IC50 value of 14 µM and some 74-fold
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3550

View Article Online
weaker than Ins(1,3,4,5)P4, but not recognised by 3-kinase. To
complete this study we also evaluated two other chiral tetra-
kisphosphates, namely, -Ins(1,2,4,5)P4 5a and -Ins(1,2,4,5)P4
5b (Table 1 and Fig. 5).10 Previously, we31 and others32 have
shown that racemic Ins(1,2,4,5)P4 is a full agonist at the
Ins(1,4,5)P3 receptor and a potent 5-phosphatase inhibitor, but
is not recognised by 3-kinase. The biological interaction of
these enantiomers with 3-kinase and 5-phosphatase has, how-
ever, not been examined. We found that -Ins(1,2,4,5)P4 5a is a
potent inhibitor of 5-phosphatase having an IC50 value of 0.47
µM but is not recognised by 3-kinase. -Ins(1,2,4,5)P4 is also a
good inhibitor of 5-phosphatase, with an IC50 value of 3 µM
and does not inhibit 3-kinase.
These results illustrate some recognition elements of
3-kinase. We have found that only Ins(1,3,4,5)P4 inhibits
3-kinase significantly, having an equatorial 3-phosphate and
an axial 2-hydroxyl moiety and a 6-hydroxyl together with a
-1,4,5 arrangement of phosphates. Deviation from these
requirements results in a dramatic loss of recognition by
the 3-kinase, although inverting the stereochemistry at the
3-position of Ins(1,4,5)P3 1a to give -chiro-Ins(2,3,5)P3 26
delivers a potent 3-kinase inhibitor.
For 5-phosphatase inhibition, the -series of inositol phos-
phates possessing a -1,4,5-type of phosphate arrangement
plus an extra phosphate at the 2- or 3-position is more potent
than the -series of inositol tetrakisphosphates. Similarly,
phosphate substitution at position-2 of Ins(1,4,6)P3 to give
Ins(1,2,4,6)P4, delivers a more potent 5-phosphatase inhibitor
than the 1,4,6-trisphosphate analogue. This corresponds with
the general trend that tetrakisphosphates are more potent
5-phosphatase inhibitors than trisphosphate derivatives. How-
ever, trisphosphates having a -1,4,5-configuration and a
6-hydroxyl [derived from -chiro-Ins(2,3,5)P3 and Ins(1,4,6)P3]
are superior inhibitors of 3-kinase than most of the tetra-
kisphosphates.
cedure.10 All final compounds were homogeneous as judged by
standard spectroscopic methods and purified by ion exchange
chromatography and used as their triethylammonium salts.
Ion exchange chromatography was performed on an LKB-
Pharmacia Medium Pressure Ion Exchange Chromatograph
using Q-Sepharose Fast Flow with gradients of triethyl-
ammonium hydrogen carbonate (TEAB) as eluent. Column
fractions containing inositol polyphosphates and phos-
phorothioates were assayed for total phosphate and phos-
phorothioate by
a
modification of the Briggs test.11
(Diisopropylamino)dichlorophosphine was prepared by the
method of Tanaka et al.34 by adding two mole equivalents of
diisopropylamine to a solution of phosphorus trichloride in
dry diethyl ether at Ϫ78 ЊC. The crude product was purified
by distillation under reduced pressure (δP = ϩ169.4 ppm) and
could be stored as a crystalline solid at Ϫ20 ЊC. Two equivalents
of benzyl alcohol in the presence of triethylamine were then
reacted with the purified product in methylene dichloride,
to afford bis(benzyloxy)(diisopropylamino)phosphine9 (δP
=
ϩ147.9 ppm) which was pure by 31P NMR, (Rf = 0.78, hexane–
triethylamine 10 : 1). NMR spectra (proton frequency 270, or
400 MHz) were referenced to SiMe4, (HDO) or [2H6]-dimethyl
sulfoxide ([2H6]DMSO). The 31P NMR shifts were measured in
ppm relative to external 85% phosphoric acid. Melting points
(uncorrected) were determined using a Reichert-Jung Thermo
Galen Kofler block. Microanalysis was carried out by the
Microanalysis Service, at the University of Bath. Mass spectra
were recorded by the University of Bath Mass Spectrometry
Service using positive and negative Fast Atom Bombardment
(FAB) with 3-nitrobenzyl alcohol (NBA) as the matrix. Optical
rotations were measured using an Optical Activity Ltd. AA-10
polarimeter, and [α]D-values are given in 10Ϫ1 deg cm2 gϪ1 and
were measured at ambient temperature.
Ins(1,4,5)P3 5-phosphatase and 3-kinase activities
We have synthesised - and -Ins(1,2,4,6)P4 and evaluated
these compounds against, 5-phosphatase and 3-kinase. Broadly,
our results confirm early conclusions that 5-phosphatase has
loose specificity of recognition, where that for 3-kinase is very
stringent.5 In agreement with the 5-phosphatase activity of
other tetrakisphosphates, -Ins(1,2,4,6)P4 is more potent than
the corresponding -Ins(1,4,6)P3 4a and the -series of inositol
trisphosphates derived from a 1,4,5-trisphosphate motif and an
additional 6-hydroxyl group is more potent than the corre-
sponding -derivatives. Substitution of a phosphate group at
position-2 of Ins(1,4,6)P3 to give Ins(1,2,4,6)P4 3a, removes any
interaction of this compound with 3-kinase.
Thus, we have synthesised the enantiomers of Ins(1,2,4,6)P4
and provide the first evaluation of the activity against
Ins(1,4,5)P3 5-phosphatase and 3-kinase. We have also rational-
ised why -Ins(1,2,4,6)P4 is apparently inactive in Ca2ϩ mobilis-
ation at the ins(1,4,5)P3 receptor, relative to the active trisphos-
phate Ins(1,4,6)P3. These results contribute to the developing
structure–activity relationship amongst the soluble inositol
polyphosphates.
The activities of Ins(1,4,5)P3 5-phosphatase and 3-kinase were
determined at 1 µM of Ins(1,4,5)P3 substrate concentration
using published procedures.35,36 We used purified recombinant
enzymes from rat brain for the Ins(1,4,5)P3 3-kinase (clone C5).
Cloning and expression was carried out in Escherichia coli of
the rat brain cDNA encoding a Ca2ϩ/calmodulin-sensitive
myo-inositol 1,4,5-trisphosphate 3-kinase,37 and the source of
Ins(1,4,5)P3 5-phosphatase was derived from human brain
(clone ECH11).38 The Ins(1,4,5)P3 3-kinase activity was deter-
mined in the presence of 0.9 mM EGTA. The incubation period
was 10 min in the presence of [3H]Ins(1,4,5)P3 which was
provided by NEN; 1 µM of cold Ins(1,4,5)P3 was sourced from
EuroBiochem.
Molecular modelling
Molecular docking of Ins(1,4,6)P3 and Ins(1,4,5)P3 to the
ligand binding domain of the Ins(1,4,5)P3 receptor was carried
out similarly to that described by Rosenberg et al.39
2,4,6-Tri-O-allyl-myo-inositol orthoformate 7
A mixture of myo-inositol orthoformate7 6 (3.83 g, 20 mmol)
and 60% sodium hydride (4.0 g, 100 mmol) was stirred in DMF
(50 cm3) at 0 ЊC. Allyl bromide (6.05 cm3, 70 mmol) was added
dropwise over 10 min, and the reaction mixture was stirred at
room temperature for 3 h. TLC (Et2O–hexane 1 : 1), showed a
single product (Rf 0.44). The reaction was cooled with ice–water
and the excess sodium hydride was destroyed with methanol
(10 cm3). The solvents were evaporated in vacuo and the residue
was partitioned between ether (200 cm3) and water (200 cm3).
The organic layer was separated and dried (MgSO4), then evap-
orated to give an oil, which was purified by flash chromato-
graphy (Et2O–pentane 1 : 1) to give the title compound 7 as a
syrup. Yield (5.54 g, 89%); (Found: C, 61.7; H, 7.23. C16H22O6
Experimental
Materials and methods
Chemicals were purchased from Aldrich, Fluka and Lancaster.
Sodium hydride was 60% pure in a dispersion mineral oil. Thin-
layer chromatography (TLC) was performed on precoated
plates (Merck TLC aluminium sheets silica 60 F254): products
were visualised by spraying with phosphomolybdic acid in
methanol, and heated at high temperature. Flash chromato-
graphy refers to the procedure developed by Still et al.33 and was
carried out on Sorbsil C60 silica gel. -Ins(1,3,4,5)P4 was
purchased from Eurobiochem (Belgium) and - and -
Ins(1,2,4,5)P4 were synthesised according to the published pro-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3551

View Article Online
requires C, 61.92; H, 7.15%); δH (270 MHz; CDCl3) 3.91 (1 H,
d, J 1.65, CH), 4.01–4.33 (10 H, m, 3 × OCH2CHCH2 and 4
inositol ring protons), 4.41 (1 H, m, CH), 5.18–5.37 (6 H, m,
3 × OCH2CHCH2), 5.53 (1 H, d, J 1.3, CH), 5.81–6.07 (3 H, m,
3 × OCH2CHCH2); δC (68 MHz; CDCl3) 67.48, 67.58, 68.05,
73.76 (myo-inositol ring carbons), 70.59 (OCH2CHCH2),
103.11 (CH, orthoformate), 117.42, 117.67 (OCH2CHCH2),
134.12, 134.57 (OCH2CHCH2); m/z (FABϩ) 621.3 (46), 446.1
(39), 311.1 (100), 253.1 (20), 153.1 (32), 81.0 (34).
pound 10ab (2.15 g, 88%) as an oil which solidified, but could
not be crystallised; (Found: C, 73.5; H, 7.68. C32H40O6 requires
C, 73.81; H, 7.75%); δH (400 MHz; CDCl3) 3.14 (1 H, dd, J 2.4,
9.8, 1-H or 3-H), 3.25 (1 H, t, J 2.1, 9.8, 1-H or 3-H), 3.32 (1 H,
t, J 9.2, 5-H), 3.77 (1 H, t, J 9.5, 4-H or 6-H), 3.80 (1 H, t, J 9.8,
4-H or 6-H), 3.90 (1 H, t, J 2.1, 2-H), 4.10–4.40 (8 H, m, 4 ×
OCH2CHCH2), 4.66, 4.70 (2 H, AB, J 11.9, OCH2Ph), 4.82
(2 H, s, OCH2Ph), 5.12–5.31 (8 H, m, 4 × OCH2CHCH2), 5.85–
6.03 (4 H, m, 4 × OCH2CHCH2), 7.22–7.38 (10 H, m, 2 ×
OCH2Ph); δC (100 MHz; CDCl3) 71.65, 72.82, 73.24, 74.54,
75.99 (CH2Ph and OCH2CHCH2), 74.12, 80.32, 80.58, 81.33,
83.61 (myo-inositol ring carbons), 116.44, 116.48, 116.55,
116.64 (OCH2CHCH2), 127.54, 128.09, 128.31, 128.34
(CH2Ph), 134.96, 135.50, 135.64, 135.77 (OCH2CHCH2),
138.49, 138.88 (Cq, CH2Ph); m/z (FABϩ) 521.1 (88), 249.0 (12),
181.0 (50), 91.0 (100).
2,4,6-Tri-O-allyl-myo-inositol 8
2,4,6-Tri-O-allyl-myo-inositol orthoformate 7 (5.0 g, 16.1
mmol) was heated at reflux temperature in a mixture of ethanol
and 1.0 mol dmϪ3 aq. HCl (60 cm3 2 : 1) for 3 h, after which
TLC (CH2Cl2–EtOAc 1 : 1) showed a single product (Rf 0.30).
The solvents were evaporated in vacuo to give a syrup, then
co-evaporated with water (2 × 50 cm3) to give the title com-
pound 8 (4.5 g, 93%) as a solid; mp 77–78 ЊC (from EtOAc–
hexane); (Found: C, 60.0; H, 8.15. C15H24O6 requires C, 59.98;
H, 8.05%); δH (400 MHz; CDCl3) 2.55 (2 H, d, J 5.5, D2O ex,
OH), 2.74 (1 H, d, J 2.1, D2O ex, OH), 3.41–3.50 (5 H, m, 1-H,
3-H, 4-H, 5-H and 6-H), 3.90 (1 H, s, 2-H), 4.31–4.40 (6 H, m,
3 × OCH2CHCH2), 5.18–5.33 (6 H, m, 3 × OCH2CHCH2),
5.90–6.02 (3 H, m, 3 × OCH2CHCH2); δC (100 MHz; CDCl3)
73.94, 74.28 (OCH2CHCH2), 72.37, 74.76, 79.17, 81.69 (myo-
inositol ring carbons), 116.93, 117.24 (OCH2CHCH2), 134.96,
135.07 (OCH2CHCH2); m/z (FABϩ) 601.2 (10), 301.1 (88),
243.1 (100), 153.1 (44), 109.0 (88), 81.0 (70).
DL-1,5-Di-O-benzyl-myo-inositol 11ab
A
mixture of -1,2,4,6-tetra-O-allyl-3,5-di-O-benzyl-myo-
inositol 10ab (1.20 g, 2.50 mmol), palladium on activated
charcoal (10% Fluka, 1.0 g) and toluene p-sulfonic acid was
dissolved in methanol–water (120 cm3, 5 : 1) then heated under
reflux for 1.5 h, after which TLC (EtOAc) showed a new
product Rf 0.20. The solution was filtered through Celite and
evaporated to give a solid. The crude product was purified by
flash chromatography (CHCl3–MeOH 10 : 1) to give the title
compound 11ab, (441 mg, 49%) as a solid; mp 139–141 ЊC
(from ethanol); (Found: C, 66.4; H, 6.72. C20H24O6 requires C,
66.64; H, 6.72%); δH (270 MHz; [2H6]DMSO) 3.04 (1 H, t, J 9.0,
5-H), 3.14 (2 H, d, J 9.3, 1-H and 3-H, D2O ex, 2 dd, over-
lapping), 3.58 (1 H, dt, J 9.3, 5.0, 4-H or 6-H), 3.77 (1 H, dt,
J 9.3, 5.5, 4-H or 6-H), 3.98 (1 H, s, 2-H), 4.56 (1 H, d, J 5.5,
D2O ex, OH), 4.58 (1 H, d, J 7.0, D2O ex, OH), 4.64 (2 H, m,
OCH2Ph), 4.73 (1 H, d, J 4.9, D2O ex, OH), 4.79 (2 H, app s,
OCH2Ph), 4.92 (1 H, d, J 5.3, D2O ex, OH), 7.21–7.45 (10 H, m,
2 × OCH2Ph); δC (68 MHz; [2H6]DMSO) 69.12, 71.86, 71.92,
72.42, 80.16, 84.33 (myo-inositol ring carbons), 70.79, 73.80
(CH2Ph), 127.01, 127.19, 127.60, 127.94, 128.08 (CH2Ph),
139.34, 139.95 (Cq, CH2Ph); m/z (FABϩ) 513.1 (76), 359.1 (36),
308.0 (26), 91.0 (100).
DL-1,2,4,6-Tetra-O-allyl-myo-inositol 9ab
A mixture of 2,4,6-tri-O-allyl-myo-inositol 8 (4.3 g, 14.3 mmol),
allyl bromide (1.47 cm3 17 mmol), tetrabutylammonium
sulfate8 (5.08 g, 15 mmol) in methylene dichloride (150 cm3)
and 5% aq. sodium hydroxide (150 cm3), was heated at reflux
temperature for 24 h. The organic layer was separated and
washed with water (2 × 100 cm3) then dried (MgSO4) and evap-
orated to give a syrup. Flash chromatography (CH2Cl2–EtOAc
3 : 1) gave the title compound 9ab as a syrup (3.6 g, 74%, Rf
0.62, CH2Cl2–EtOAc 1 : 1), together with other minor products
Rf 0.88, Rf 0.44 and a tiny quantity of starting material Rf 0.24,
which were not isolated; (Found: C, 63.4; H, 8.36. C18H28O6
requires C, 63.49; H, 8.30%); δH (400 MHz; CDCl3) 2.45 (2 H,
br m, D2O ex, OH), 3.23 (1 H, dd, J 2.45, 9.8, 1-H), 3.39 (1 H, t,
J 8.8, 5-H), 3.42 (1 H, dd, J 2.9, 10.7, 3-H), 3.51 (1 H, t, J 9.3,
4-H), 3.64 (1 H, t, J 9.8, 6-H), 3.93 (1 H, t, J 2.9, 2-H), 4.07–4.45
(8 H, m, 4 × OCH2CHCH2), 5.18–5.33 (8 H, m, 4 × OCH2-
CHCH2), 5.90–6.02 (4 H, m, 4 × OCH2CHCH2); δC (100 MHz;
CDCl3) 71.52, 71.94, 73.86, 74.23 (OCH2CHCH2), 71.94,
74.65, 76.72, 80.56, 80.76 (myo-inositol ring carbons), 116.80,
116.90, 117.11 (OCH2CHCH2), 135.20, 135.26 (OCH2-
CHCH2); m/z (FABϩ) 681.9 (68), 341.2 (100), 283.2 (62), 225.2
(6), 111.0 (16), 97 (36).
DL-3,5-Di-O-benzyl-1,2,4,6-tetrakis-O-[di(benzyloxy)-
phosphoryl)]-myo-inositol 13ab
A mixture of bis(benzyloxy)diisopropylaminophosphine (0.69
g, 2 mmol) and 1H-tetrazole (0.28 g, 4 mmol) in dry methylene
dichloride (5 cm3) was stirred at room temperature for 15 min.
-1,5-Di-O-benzyl-myo-inositol 11ab (0.108 g, 0.30 mmol)
was added to the solution which was stirred for a further
10 min. The reaction mixture was cooled with ice–water and
(50–60%) MCPBA (0.80 g, 2.30 mmol) was added slowly and
the mixture was stirred for a further 30 min. The solution was
diluted with ethyl acetate (50 cm3) and washed with 10% aq.
sodium metabisulfite (50 cm3), saturated aq. sodium hydrogen
carbonate, brine and water (50 cm3 of each). The organic layer
was isolated then dried (MgSO4) and evaporated to give an oil.
The crude product was purified by flash chromatography, (Rf
0.20, CHCl3–acetone 5 : 1) then EtOAc–pentane 2 : 1, to give
the title compound 13ab (0.382 g, 91%) as a solid; mp 76–77 ЊC
(from EtOAc–pentane); δH (400 MHz; CDCl3) 3.63 (1 H, d,
J 10.3, 3-H), 3.63 (1 H, t, J 9.8, 5-H), 4.47–5.20 (23 H, m, 1-H,
4-H, 6-H and 10 × OCH2Ph), 5.58 (1 H, d, J 8.85, 2-H), 6.92–
7.46 (50 H, m, 8 × O(O)POCH2Ph, 2 × OCH2Ph); δC (100
MHz; CDCl3) 69.24, 69.29, 69.37, 69.40, 69.55, 69.60, 69.68,
70.00, 70.15, 72.16, 74.56 (CH2Ph), 73.95, 74.40, 75.88,
76.97, 78.50, 79.30 (myo-inositol ring carbons), 127.16, 127.65,
127.80, 127.85, 127.94, 128.05, 128.11, 128.16, 128.27, 128.40,
128.53, 128.64, 129.48, 129.90 (CH2Ph), 133.37, 134.22, 135.39,
135.53, 135.62, 136.56, 137.87 (Cq, CH2Ph); δP (162 MHz; D2O)
Ϫ1.04, Ϫ1.26, Ϫ1.43, Ϫ2.75; m/z (FABϩ) 1401.0 (64), 1311.0
DL-1,2,4,6-Tetra-O-allyl-3,5-di-O-benzyl-myo-inositol 10ab
A mixture of -1,2,4,6-tetra-O-allyl-myo-inositol 9ab (1.60 g,
4.70 mmol) and sodium hydride (1.25 g, 31.25 mmol) was
stirred in dry DMF (40 cm3). Benzyl bromide (1.68 cm3, 14.1
mmol) was added dropwise and the solution was stirred for 2 h
at room temperature. TLC (Et2O–pentane 1 : 3) showed a new
product Rf 0.50. The excess sodium hydride was destroyed with
methanol (10 cm3) and the solvents were evaporated in vacuo to
give an oil. The crude product was partitioned between water
and ether (50 cm3 of each) and washed with 0.10 mol dmϪ3 aq.
hydrochloric acid, then water again (50 cm3 of each). The
organic layer was dried (MgSO4) and the solvent was evapor-
ated to give an oil. The crude product was purified by flash
chromatography (pentane–Et2O 3 : 1) to give the title com-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3552

View Article Online
(15), 1033.0 (10), 459 (16), 361 (40), 301 (10), 181.1 (22), 91
(100); [Found: m/z, 1401.4066 (M ϩ H)ϩ requires m/z
1401.4060].
15ab (Found: C, 68.0; H, 6.90. C33H40O9 requires C, 68.26; H,
6.94%); mp 84–86 ЊC (from Et2O–pentane); δH (270 MHz;
[2H6]DMSO) 1.27, 1.38 (6 H, 2 s, CMe2), 3.32 (1 H, t, J 9.0,
5-H), 3.56–3.60 (2 H, m, 1-H or 3-H and 6-H), 3.73 (3 H, s,
OMe), 3.74 (3 H, s, OMe), 3.75 (3 H, s, OMe), 4.10 (1 H, t,
J 6.0, 1-H or 3-H), 4.37 (1 H, t, J 4.5, 2-H), 4.56–4.72 (6 H, m,
3 × OCH2C6H4OMe), 5.23 (1 H, br s, D2O ex, OH), 6.84–6.91
(6 H, m, OCH2C6H4OMe), 7.20–7.32 (6 H, m, OCH2C6H4-
OMe); δC (68 MHz; [2H6]DMSO) 25.64, 27.53 (CMe2), 55.05
(CH2PhOMe), 71.35, 72.33, 73.15 (CH2PhOMe), 71.66, 73.72,
77.25, 78.34, 81.28, 82.35 (myo-inositol ring carbons), 108.56
(Cq, CMe2), 113.49, 129.14, 129.23, 129.30 (CH2PhOMe),
130.82, 130.89, 131.15 (Cq, CH2PhOMe), 158.67 (Cq,
CH2PhOMe); m/z (FABϪ) 733.5 (100), 626.3 (50), 579.3 (16),
355.1 (20), 299.2 (18), 181.1 (16), 106.0 (16).
17ab (Found: C, 70.0; H, 6.96. C41H48O10 requires C, 70.27;
H, 6.90%); δH (400 MHz; CDCl3) 1.35, 1.51 (6 H, 2 s, CMe2),
3.36 (1 H, t, J 8.85, 5-H), 3.64 (1 H, dd, J 3.7, 8.85, 1-H or 3-H),
3.74–3.79 (1 H, obscured, 1-H or 3-H), 3.77 (3 H, s, OMe), 3.74
(6 H, s, OMe), 3.79 (3 H, s OMe), 3.89 (1 H, t, J 8.85, 4-H or
6-H), 4.06 (1 H, t, J 6.7, 4-H or 6-H), 4.21 (1 H, t, J 4.0, 5.5,
2-H), 4.64–4.80 (8 H, m, 4 × OCH2C6H4OMe), 6.83–6.87 (8 H,
m, OCH2C6H4OMe), 7.20–7.32 (8 H, m, OCH2C6H4OMe);
δC (100 MHz; CDCl3) 25.80, 27.76 (CMe2), 55.21 (CH2-
PhOMe), 72.86, 73.55 (CH2PhOMe), 74.60, 74.94, 79.17,
80.61, 81.87, 82.33 (myo-inositol ring carbons), 109.67 (Cq,
CMe2), 113.62, 113.69, 113.73, 129.52, 129.61 (CH2PhOMe),
130.28, 130.74, 130.83, 130.87 (Cq, CH2PhOMe), 159.12,
159.16, 159.31 (Cq, CH2PhOMe); m/z (FABϩ) 723.3 (22), 699.3
(38), 579.3 (82), 241.1 (90), 121.0 (100).
DL-myo-Inositol 1,2,4,6-tetrakisphosphate 3ab
-3,5-Di-O-benzyl-1,2,4,6-tetrakis-O-[di(benzyloxyphos-
phoryl)]-myo-inositol 13ab (0.165 g, 0.118 mmol) was
hydrogenolysed in a mixture of methanol (40 cm3) and water
(10 cm3), in the presence of palladium on carbon (10%, 0.20 g)
for 20 h. The reaction mixture was filtered over a bed of Celite
to remove the insoluble components, and washed with water
and methanol, then evaporated in vacuo to give a syrup. The
residue was then dissolved in MilliQ water (150 cm3) and
purified by ion exchange chromatography on Q-Sepharose Fast
Flow with a gradient of TEAB buffer (0–1000 mmol dmϪ3) at
pH 8.6. The triethylammonium salt of 3ab eluted at ca. 800
mmol dmϪ3. Yield (39 µmol, 33%); δH (400 MHz; D2O) 3.60
(1 H, t, J 9.2, 5-H), 3.69 (1 H, d, J 9.3, 3-H), 4.10 (1 H, J 9.3,
1-H), 4.23 (1 H, q, J 9.2, 4-H or 6-H), 4.32 (1 H, q, J 9.2, 4-H or
6-H), 4.69 (1 H, d, J 9.2, 2-H); δP (162 MHz; D2O) 2.12 (1 P, d,
J 10.1), 1.75 (1 P, br d, J 13.5), 1.62 (1 P, d, J 10.0), 1.32 (1 P, d,
J 10.0); m/z (FABϪ) 999.1 (5), 499.0 (100), 291.2 (5); [Found:
m/z, 498.9193 (M–H)Ϫ requires m/z 498.9208].
DL-2,3-O-Isopropylidene-1,4,6-tris-O-p-methoxybenzyl-myo-
inositol 16ab, DL-2,3-O-isopropylidene-1,4,5-tris-O-p-methoxy-
benzyl-myo-inositol 15ab and DL-2,3-O-isopropylidene-1,4,5,6-
tetrakis-O-p-methoxybenzyl-myo-inositol 17ab
A mixture of -2,3-O-isopropylidene-1,4-di-O-p-methoxy-
benzyl-myo-inositol12 14ab (9.21 g, 20 mmol), acetonitrile (400
cm3), dibutyltin oxide (5.48 g, 22 mmol), tetrabutylammonium
bromide (6.45 g, 20 mmol) and p-methoxybenzyl chloride
(4.35 cm3, 30 mmol) was heated under reflux in a Soxhlet
apparatus containing 4 Å molecular sieves for 16 h. After this
time a further quantity of p-methoxybenzyl chloride (4.35 cm3,
30 mmol) was added, together with some tetrabutylammonium
iodide (7.38 g, 20 mmol) and heated under reflux for a further
24 h, after which the reaction was complete according to TLC.
The reaction mixture was cooled, the solvent was evaporated
and the orange residue was partitioned between water and ether
(250 cm3 of each). The organic layer was separated and stirred
with a saturated aqueous solution of sodium hydrogen carb-
onate (250 cm3) for 1 h. The solid components were removed by
filtering the solution over Celite then washed with ether and the
organic layer was dried over magnesium sulfate. TLC (CH2Cl2–
EtOAc 1 : 1) showed four products, p-methoxybenzyl halide,
Rf 0.78; -2,3-O-isopropylidene-1,4,5,6-tetrakis-O-p-methoxy-
benzyl-myo-inositol 17ab, Rf 0.44; -2,3-O-isopropylidene-
1,4,6-tris-O-p-methoxybenzyl-myo-inositol 16ab, Rf 0.34 and
-2,3-O-isopropylidene-1,4,5-tris-O-p-methoxybenzyl-myo-
inositol 15ab, Rf 0.22, which were separated by flash chromato-
graphy to give the products as syrups. -2,3-O-Isopropylidene-
1,4,5-tris-O-p-methoxybenzyl-myo-inositol 15ab was recrystal-
lised from Et2O–pentane. 16ab (5.34 g, 46%), 15ab (3.01 g,
26%), 17ab (0.87 g, 6%).
16ab (Found: C, 68.1; H, 6.99. C33H40O9 requires C, 68.26; H,
6.94%); δH (400 MHz; [2H6]DMSO) 1.28, 1.38 (6 H, 2 s, CMe2),
3.32 (1 H, br t, D2O ex, t, J 9.2, 5-H), 3.50 (2 H, m, 1-H or 3-H
and 6-H), 3.66 (1 H, dd, J 3.7, 7.6, 4-H), 3.74 (9 H, s, OMe),
4.05 (1 H, t, J 6.7, 1-H or 3-H), 4.36 (1 H, t, J 4.0, 2-H), 4.52–
4.69 (6 H, m, 3 × OCH2C6H4OMe), 5.15 (1 H, br s, D2O ex,
OH), 6.86–6.89 (6 H, m, OCH2C6H4OMe), 7.23–7.32 (6 H,
m, OCH2C6H4OMe); δC (100 MHz; [2H6]DMSO) 25.67, 27.56
(CMe2), 55.03 (CH2PhOMe), 72.20, 73.06, 73.20 (CH2-
PhOMe), 71.16, 73.66, 76.46, 78.23, 80.79, 81.84 (myo-inositol
ring carbons), 108.50 (Cq, CMe2), 113.38, 113.51, 127.90,
129.04, 129.20 (CH2PhOMe), 130.57, 130.99, 131.09 (Cq,
CH2PhOMe), 158.55, 158.63 (Cq, CH2PhOMe); m/z (FABϪ)
733.3 (100), 579.2 (60), 467.1 (50), 258.1 (44), 92 (30).
DL-5-O-Benzyl-2,3-O-isopropylidene-1,4,6-tris-O-p-methoxy-
benzyl-myo-inositol 18ab
A mixture of -2,3-O-isopropylidene-1,4,6-tris-O-p-methoxy-
benzyl-myo-inositol 16ab (5.34 g, 9.18 mmol) and sodium
hydride (0.96 g, 24 mmol) in dry DMF (60 cm3) was stirred at
room temperature. Benzyl bromide (2.38 cm3, 20 mmol) was
added to the solution, which was then stirred for a further 2 h.
The excess sodium hydride was destroyed with methanol
(10 cm3) and the solvents were evaporated in vacuo. The crude
product was partitioned between ether and water (100 cm3 of
each) and the organic layer was separated and dried (MgSO4).
The fully protected product 18ab was purified by flash chrom-
atography (Rf 0.22, Et2O–pentane 1 : 1). Yield (5.29 g, 86%);
(Found: C, 71.7; H, 7.03. C40H46O9 requires C, 71.62; H,
6.91%); δH (270 MHz; CDCl3) 1.35, 1.52 (6 H, 2 s, CMe2), 3.37
(1 H, t, J 9.2, 5-H), 3.64 (1 H, dd, J 3.7, 9.0, 1-H or 3-H), 3.75–
3.79 (1 H, obscured, 1-H or 3-H), 3.77 (3 H, s, OMe), 3.78 (3 H,
s, OMe), 3.79 (3 H, s, OMe), 3.90 (1 H, t, J 8.8, H-4 or H-6),
4.06 (1 H, t, J 6.6, H-4 or H-6), 4.21 (1 H, dd, J 4.0, 5.3, H-2),
4.63–4.80 (8 H, m, OCH2Ph and 3 × OCH2C6H4OMe), 6.83–
7.34 (17 H, m, CH2Ph and 3 × OCH2C6H4OMe); δC (68 MHz;
CDCl3) 25.70, 27.67 (CMe2), 55.09, 55.19 (CH2PhOMe), 72.80,
73.46, 74.87, 75.13 (CH2Ph and CH2PhOMe), 74.53, 76.66,
79.09, 80.50, 82.11 (myo-inositol ring carbons), 109.63 (CMe2),
113.67, 126.82, 127.44, 127.83, 128.20 (CH2Ph and CH2-
PhOMe), 130.23, 130.63, 130.74 (Cq, CH2PhOMe), 138.63 (Cq,
CH2Ph), 159.04, 159.11, 159.24 (Cq, CH2PhOMe); m/z (FABϩ)
639.3 (20), 669.3 (22), 549.2 (76), 241.1 (22), 211.1 (52), 121.0
(100).
DL-5-O-Benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol 19ab
-5-O-Benzyl-2,3-O-isopropylidene-1,4,6-tris-O-p-methoxy-
benzyl-myo-inositol 18ab (9.46 g, 14.08 mmol) was dissolved
in methanol–1.0 mol dmϪ3 aq. HCl (100 cm3, 9 : 1) and kept at
50 ЊC for 30 min. TLC (CH2Cl2–EtOAc 2 : 1) showed a new
product Rf 0.40, which precipitated from the reaction mixture
as a white solid. The methanolic solution was cooled and the
solid was filtered off. The remaining solution was kept at 50 ЊC
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3553

View Article Online
for a further 30 min, after which, most of the starting material
had disappeared. The reaction mixture was cooled and the
insoluble solid was filtered off. The acidic solution was neutral-
ised with triethylamine (10 cm3) and the solvents were evapor-
ated. The remaining solid was partitioned between methylene
dichloride and water (50 cm3 of each) and the organic layer was
dried (MgSO4) and evaporated. The crude product was purified
by flash chromatography (CH2Cl2–EtOAc 2 : 1) to give the pure
title compound 19ab (7.0 g, 79%); mp 148–150 ЊC (from
EtOAc–hexane); (Found: C, 70.2; H, 6.77. C37H42O9 requires C,
70.46; H, 6.71%); δH (400 MHz; [2H6]DMSO) 3.32 (1 H, t, J 9.5,
5-H), 3.36–3.59 (2 H, br, D2O ex, OH), 3.37 (2 H, app dt, J 3.35,
9.5, D2O ex, 2 dd, J 2.45, 9.8, 1-H, J 2.75, 9.8, 3-H), 3.59 (1 H, t,
J 9.5, 4-H or 6-H), 3.66 (3 H, s, OMe), 3.67 (3 H, s, OMe), 3.69
(3 H, OMe), 3.67–3.71 (1 H, obscured, 4-H or 6-H), 3.99 (1 H,
t, J 2.4, 2-H), 4.44–4.78 (8 H, m, OCH2Ph and 3 × OCH2-
C6H4OMe), 6.79–7.50 (17 H, m, CH2Ph and 3 × OCH2-
C6H4OMe); δC (100 MHz; [2H6]DMSO) 55.27 (CH2PhOMe),
72.40, 75.25, 75.62 (CH2PhOMe and CH2Ph), 69.05, 71.58,
79.74, 80.94, 81.49, 83.92 (myo-inositol ring carbons), 113.77,
113.92, 113.99, 127.59, 127.70, 128.42, 129.53, 129.64, 129.94
(CH2Ph and CH2PhOMe), 130.65, 130.85 (Cq, CH2PhOMe),
138.66 (Cq, CH2Ph), 159.18, 159.34, 159.40 (Cq, CH2PhOMe);
m/z (FABϩ) 653.3 (54), 629.3 (14), 509.2 (88), 419.3 (44), 329.1
(68), 242.2 (64), 167.1 (56), 121.0 (100).
4.05 and 4.36 (2 H, AB, J 10.3, OCH2C6H4OMe), 4.37 (1 H,
2-H, br, obscured), 4.57–4.82 (9 H, m, OCH2Ph and 3 ×
OCH2C6H4OMe and 1-H), 5.99 [1 H, s, O2CCH(OAc)Ph],
6.63–7.51 (22 H, m, CH2Ph, OCH2C6H4OMe and O2CCH-
(OAc)Ph); δC (100 MHz; CDCl3) 20.66 [OC(O)Me], 55.17,
55.23 (CH2PhOMe), 72.49, 74.82, 75.58, 75.80 (CH2PhOMe
and CH2Ph), 67.43, 74.94, 75.55, 78.20, 79.48, 80.67, 82.92
(O2CCH(OAc)Ph, and myo-inositol ring carbons), 113.42,
113.72, 113.90, 127.43, 127.59, 128.03, 128.25, 128.93, 129.09,
129.55, 129.73 (CH2PhOMe, CH2Ph and O2CCH(OAc)Ph),
130.30, 130.78, 132.99, 138.59, 158.87, 159.12, 159.38 (Cq,
CH2PhOMe, CH2Ph and O2CCH(OAc)Ph), 168.56, 170.66
(O2CCH(OAc)Ph, OC(O)Me); m/z (FABϪ) 959.5 (100), 805.4
(95), 629.3 (25), 419.3 (15), 335.1 (35), 106.0 (35).
D-5-O-Benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol 19a
A
mixture of -1-O-[R-(Ϫ)-O-acetylmandelyl]-5-O-benzyl-
3,4,6-tris-O-p-methoxybenzyl-myo-inositol 22 (2.59 g, 4.10
mmol) and sodium hydroxide (0.8 g, 20 mmol) in methanol (100
cm3) was heated under reflux for 30 min. The mixture was
cooled and neutralised with carbon dioxide. The remaining
solution was diluted with water (100 cm3) and evaporated to
dryness in vacuo. The title compound was extracted with
methylene dichloride (4 × 100 cm3) and the solvent was evapor-
ated to give the product 19a (1.86 g, 96%), (Rf 0.40, CH2Cl2–
EtOAc 2 : 1); mp 123–125 ЊC (from Et2O–pentane); [α]D ϩ24
(c 1 in CHCl3); (Found: C, 70.4; H, 6.60. C37H42O9 requires C,
70.46; H, 6.71%); Mass spectrum and NMR data were identical
with those of the racemic mixture 19ab.
L-(21) and D-1-O-[R-(؊)-O-Acetylmandelyl]-5-O-benzyl-3,4,6-
tris-O-p-methoxybenzyl-myo-inositol 22
A
mixture of -5-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-
myo-inositol 19ab (6.60 g, 10.45 mmol), (R)-(Ϫ)-O-acetyl-
mandelic acid 20 (2.06 g, 10.6 mmol) and 4-(dimethyl-
amino)pyridine (0.03 g, 0.25 mmol) was stirred in methylene
dichloride (100 cm3) at Ϫ20 ЊC. A solution of dicyclo-
hexylcarbodiimide (DCCI) (2.15 g, 10.6 mmol) in methylene
dichloride (20 cm3) was added dropwise over a period of 2 h
with stirring, then stirred overnight at room temperature. TLC
(CHCl3–acetone 15 : 1) showed two main products, Rf 0.34 and
0.24, together with minor products at Rf 0.44 and 0.14. The
minor products were not investigated further. The solution was
filtered over a bed of Celite and washed with methylene dichlor-
ide (2 × 100 cm3). The solvent was evaporated off to give a solid
and the individual diastereoisomers were isolated by flash
chromatography (CHCl3–acetone 15 : 1) to give 21 Rf 0.34
(2.74 g, 32%); mp 105–107 ЊC (from EtOAc–hexane); [α]D Ϫ15
(c 1 in CHCl3) and 22 Rf 0.24 (2.19 g, 26%); mp 127–129 ЊC
from EtOAc–hexane; [α]D Ϫ37 (c 1 in CHCl3).
21 (Found: C, 69.5; H, 6.18. C47H50O12 requires C, 69.96; H,
6.25%); δH (400 MHz; CDCl3) 2.13 (1 H, s, D2O ex, OH), 2.22 [3
H, s, O2CCH(OAc)Ph], 3.41 (1 H, dd, J 2.4, 9.8, 3-H), 3.45 (1 H,
t, J 9.5, 5-H), 3.77 (3 H, s, OMe), 3.78 (3 H, s, OMe), 3.79 (3 H,
s, OMe), 3.88 (1 H, t, J 9.5, 4-H), 4.04 (1 H, t, J 9.5, 6-H),
4.11 (1 H, t, J 2.4, 2-H), 4.54–4.86 (9 H, m, 1-H, OCH2Ph and
3 × OCH2C6H4OMe), 5.97 [1 H, s, O2CCH(OAc)Ph], 6.76–7.47
(22 H, m, CH2Ph, 3 × OCH2C6H4OMe and O2CCH(OAc)Ph);
δC (100 MHz; CDCl3) 20.72 [OC(O)Me], 55.23 (CH2PhOMe),
74.72, 75.33, 75.57, 75.78 (CH2PhOMe, CH2Ph), 67.34, 72.51,
75.20, 78.11, 79.28, 80.65, 82.95 (O2CCH(OAc)Ph and myo-
inositol ring carbons), 113.68, 113.73, 113.88, 127.30, 127.49,
127.61, 128.31, 128.84, 129.28, 129.44, 129.61, 129.70, 129.79,
129.86 (CH2PhOMe, CH2Ph and O2CCH(OAc)Ph), 130.43,
130.76, 133.39, 138.68, 159.14, 159.36 (Cq, CH2PhOMe, CH2Ph
and O2CCH(OAc)Ph), 168.25, 170.74 (OC(O)Me and O2CCH-
(OAc)Ph); m/z (FABϪ) 959.5 (100), 805.4 (95), 629.3 (30), 419.3
(30), 331.1 (45), 258.1 (33), 181.1 (40), 149.1 (35).
L-5-O-Benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol 19b
A
mixture of -1-O-[R-(Ϫ)-O-acetylmandelyl]-5-O-benzyl-
3,4,6-tris-O-p-methoxybenzyl-myo-inositol 22 (2.00 g, 4.10
mmol) and sodium hydroxide (0.8 g, 20 mmol) in methanol (100
cm3) was heated under reflux for 30 min. Work up as for the
-enantiomer gave the title compound 19b (1.48 g, 99%), Rf
0.40 (CH2Cl2–EtOAc 2 : 1); mp 123–125 ЊC (from Et2O–
pentane); [α]D Ϫ24 (c 1 in CHCl3); (Found: C, 70.4; H, 6.65.
C37H42O9 requires C, 70.46; H, 6.71%); The mass spectrum and
NMR data were identical with those of the racemic mixture
19ab.
D-3,5-Di-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol
23a
A mixture of -5-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-
inositol 19a (1.64 g, 2.6 mmol) and dibutyltin oxide (0.875 g, 3.5
mmol) was heated under reflux in toluene (250 cm3) using a
Dean-and-Stark apparatus for 3 h. The reaction mixture was
cooled and the solvent was evaporated to give a syrup which
was dried under vacuum for a further 2 h. Caesium fluoride
(1.52 g, 10 mmol) and dry DMF (50 cm3) were added to the
dried residue under an atmosphere of nitrogen. Benzyl bromide
(0.71 cm3, 6 mmol) was added to the mixture and the reaction
was stirred overnight at room temperature. TLC (CH2Cl2–
EtOAc 10 : 1), showed a product with Rf 0.40. The solvent was
evaporated under reduced pressure and the residue was taken
up in methylene dichloride (100 cm3), washed with water (100
cm3) and stirred with a saturated aqueous solution of sodium
hydrogen carbonate (100 cm3, 10% w/v) for 30 min. The organic
layer was separated, washed with water, dried over MgSO4 and
the solvent was evaporated to give the crude product. The title
compound 23a was obtained by flash chromatography (CH2Cl2–
EtOAc 10 : 1), to give the pure enantiomer as a solid. Yield
(1.55 g, 83%); mp 100–102 ЊC (from EtOAc–hexane); [α]D
Ϫ3 (c 1 in CHCl3); (Found: C, 73.3; H, 6.70. C44H48O9 requires
C, 73.31; H, 6.71%); δH (400 MHz; CDCl3) 2.50 (1 H, br s, D2O
ex, OH), 3.34 (1 H, dd, J 2.75, 9.5, 1-H or 3-H), 3.35 (1 H, dd,
J 2.75, 9.5, 1-H or 3-H), 3.41 (1 H, t, J 9.5, 5-H), 3.76 (3 H, s,
22 (Found: C, 69.6; H, 6.18. C47H50O12 requires C, 69.96; H,
6.25%); δH (400 MHz; CDCl3) 2.19 [3 H, s, O2CCH(OAc)Ph],
2.72 (1 H, s, D2O ex, OH), 3.41 (1 H, t, J 9.15, 5-H), 3.47 (1 H,
dd, J 2.4, 9.5, 3-H), 3.75 (3 H, s, OMe), 3.76 (3 H, s, OMe), 3.79
(3 H, s, OMe), 3.92 (1 H, t, J 9.2, 4-H), 4.04 (1 H, t, J 9.8, 6-H),
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3554

View Article Online
OMe), 3.77 (3 H, s, OMe), 3.79 (3 H, s, OMe), 3.95 (1 H, t,
J 9.2, 4-H or 6-H), 3.98 (1 H, t, J 9.5, 4-H or 6-H), 4.18 (1 H, t,
J 2.7, 2-H), 4.63–4.87 (10 H, m, 2 × OCH2Ph and 3 ×
OCH2C6H4OMe), 6.79–7.36 (22 H, m, 2 × CH2Ph and 3 ×
OCH2C6H4OMe); δC (100 MHz; CDCl3) 55.23 (CH2PhOMe),
67.56, 72.37, 72.69, 75.62, 75.82 (CH2Ph and CH2PhOMe),
67.56, 79.52, 79.83, 80.94, 83.19 (myo-inositol ring carbons),
113.73, 113.82, 127.47, 127.67, 127.80, 128.34, 128.44, 129.48,
129.68 (CH2Ph and CH2PhOMe), 130.06, 130.91, 130.94,
138.02, 138.82, 159.12, 159.31 (Cq, CH2Ph and CH2PhOMe);
m/z (FABϪ) 873.1 (100), 719.1 (44), 599.4 (22), 329.0 (20), 106.0
(16).
The solution was cooled to 0 ЊC then (50–60%) MCPBA (1.60
g, 4.6 mmol) was added to the solution and the reaction mixture
was stirred for a further 30 min. Work up and purification as for
the racemate 13ab gave the title compound 13a as a syrup (661
mg, 79%); [α]D Ϫ2.20 (c 7.68 in CHCl3); The mass spectrum and
NMR data were identical with those of racemate 13ab.
L-3,5-Di-O-benzyl-1,2,4,6-tetrakis-O-[di(benzyloxy)-
phosphoryl]-myo-inositol 13b
A mixture of bis(benzyloxy)diisopropylaminophosphine (0.69
g, 2 mmol) and 1H-tetrazole (0.28 g, 4 mmol) in dry methylene
dichloride (10 cm3) was stirred at room temperature for 10 min.
-3,5-Di-O-benzyl-myo-inositol 11b (0.108 g, 0.30 mmol) was
added to the reaction mixture, which was stirred for a further 10
min. The solution was cooled to 0 ЊC then (50–60%) MCPBA
(0.80 g, 2.3 mmol) was added to the solution and the reaction
mixture was stirred for a further 30 min. Work up and purifi-
cation as for the racemate 13ab gave the title compound 13b as a
syrup (0.360 g, 86%); [α]D ϩ2.37 (c 5.05 in CHCl3); The mass
spectrum and NMR data were identical with those of racemate
13ab.
L-3,5-Di-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol
23b
A mixture of -5-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-
inositol 19b (0.90 g, 1.43 mmol) and dibutyltin oxide (0.425 g,
1.7 mmol) was heated under reflux in toluene (100 cm3) using a
Dean-and-Stark apparatus for 3 h. The reaction mixture was
cooled and the solvent was evaporated to give a syrup which
was dried under vacuum for a further 2 h. Caesium fluoride
(0.65 g, 4.29 mmol) and dry DMF (30 cm3) were added to the
dried residue under an atmosphere of nitrogen. Benzyl bromide
(0.35 cm3, 3 mmol) was added to the mixture and the reaction
was stirred overnight at room temperature. TLC (CH2Cl2–
EtOAc 10 : 1) showed a product with Rf 0.40. Work up and
purification as for the -enantiomer provided the title
compound 23b (0.98 g, 95%); mp 100–102 ЊC (from EtOAc–
hexane); [α]D ϩ2 (c 1 in CHCl3); (Found: C, 73.0; H, 6.74.
C44H48O9 requires C, 73.31; H, 6.71%); The mass spectrum and
NMR data were identical to those of the -derivative.
D-myo-Inositol 1,2,4,6-tetrakisphosphate 3a
-3,5-Di-O-benzyl-1,2,4,6-tetrakis-O-[di(benzyloxy)phos-
phoryl]-myo-inositol 13a (0.160 g, 0.114 mmol) was hydro-
genolysed in a mixture of methanol (40 cm3) and water (10
cm3), in the presence of palladium on carbon (10%, 0.20 g) for
20 h. The reaction mixture was filtered over a bed of Celite to
remove the insoluble components and washed with water and
methanol then evaporated in vacuo to give a syrup. The residue
was then dissolved in MilliQ water (150 cm3) and purified by
ion exchange chromatography on Q-Sepharose Fast Flow with
a gradient of TEAB buffer (0–1000 mmol dmϪ3) at pH 8.6. The
D-3,5-Di-O-benzyl-myo-inositol 11a
-3,5-Di-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol
23a (1.31 g, 1.82 mmol) was added to 1 mol dmϪ3 aq. hydro-
chloric acid–ethanol (60 cm3, 1 : 2). The mixture was heated
under reflux for 4 h, cooled and the solvents were evaporated
in vacuo to give a solid, which was co-evaporated with methanol
(50 cm3). TLC (EtOAc) showed a product Rf 0.20 and depro-
tected p-methoxybenzyl derivative Rf 1.00. The remaining solid
was purified by flash chromatography (Rf 0.40 CHCl3–MeOH
10 : 1), to give the title compound 11a as a solid (654 mg, 92%);
mp 162–163 ЊC (from ethanol); [α]D ϩ5 (c 1 in MeOH); (Found:
C, 66.6; H, 6.70. C20H24O6 requires C, 66.64; H, 6.72%); The
mass spectrum and NMR data were identical with those of
racemate 11ab.
triethylammonium salt of 4a eluted at ca. 800 mmol dmϪ3
.
Yield (74 µmol, 65%); [α]D Ϫ15.4 (c 3.12 in MeOH), lit.18 [α]D
Ϫ15.2 (c 2.10, H2O, pH 9.5); δH (270 MHz; D2O) 3.58 (1 H, t,
J 9.2, 5-H), 3.67 (1 H, d, J 9.3, 3-H), 4.08 (1 H, t, J 9.3, 1-H),
4.19 (1 H, q, J 9.0, 4-H or 6-H), 4.29 (1 H, q, J 9.2, 4-H or 6-H),
4.67 (1 H, d, J 9.2, 2-H); δP (109 MHz; D2O) 1.78 (1 P, d, J 6.7),
1.32 (2 P, d, J 6.7), 1.07 (1 P, d, J 6.7); m/z (FABϪ) 999.1 (5),
499.0 (100), 291.2 (5); [Found: m/z, 498.9189 (M Ϫ H)Ϫ requires
m/z 498.9208].
L-myo-Inositol 1,2,4,6-tetrakisphosphate 4b
-3,5-Di-O-benzyl-1,2,4,6-tetrakis-O-[di(benzyloxy)phos-
phoryl]-myo-inositol 13b (0.150 g, 0.107 mmol) was hydro-
genolysed in a mixture of methanol (40 cm3) and water (10 cm3)
over palladium on carbon (10%, 0.20 g) for 20 h. The reaction
mixture was filtered over a bed of Celite to remove the insoluble
components and washed with water and methanol then evapor-
ated in vacuo to give a syrup. The residue was then dissolved in
MilliQ water (150 cm3) and purified by ion exchange chromato-
graphy on Q-Sepharose Fast Flow with a gradient of TEAB
buffer (0–1000 mmol dmϪ3) at pH 8.6. The triethylammonium
salt of 4b eluted at ca. 800 mmol dmϪ3. Yield (38 µmol, 36%);
[α]D ϩ15 (c 1.6 in MeOH), lit.18 [α]D ϩ 14.7 (c 1.77, H2O, pH
9.5); δH (270 MHz; D2O) 3.57 (1 H, t, J 9.2, 5-H), 3.67 (1 H, d,
J 10.1, 3-H), 4.08 (1 H, t, J 9.2, 1-H), 4.21 (1 H, q, J 9.15, 4-H or
6-H), 4.30 (1 H, q, J 9.3, 4-H or 6-H), 4.67 (1 H, d, J 9.2, 2-H);
δP (109 MHz; D2O) 2.05 (1 P, d, J 10.0), 1.72 (1 P, d, J 10.1), 1.56
(1 P, d, J 10.1), 1.29 (1 P, d, J 6.8); m/z (FABϪ) 999.1 (5), 499.0
(100), 291.2 (5); [Found: m/z, 498.9212 (M Ϫ H)Ϫ requires m/z
498.9208].
L-3,5-Di-O-benzyl-myo-inositol 11b
-3,5-Di-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol
23b (0.675 g, 0.936 mmol) was added to 1 mol dmϪ3 aq. hydro-
chloric acid–ethanol (90 cm3, 1 : 2). The mixture was heated
under reflux for 4 h, cooled, and the solvents were evaporated in
vacuo to give a solid, which was co-evaporated with methanol
(50 cm3). TLC (ethyl acetate) showed a product Rf 0.20 and a
deprotected p-methoxybenzyl derivative Rf 1.00. The remaining
solid was purified by flash chromatography (CHCl3–MeOH
10 : 1) Rf 0.40, to give the title compound 11b as a solid (0.315 g,
93%); mp 162–163 ЊC (from ethanol); [α]D Ϫ5 (c 1 in MeOH);
(Found: C, 66.6; H, 6.68. C20H24O6 requires C, 66.64; H,
6.72%); The mass spectrum and NMR data were identical with
those of racemate 11ab.
D-3,5-Di-O-benzyl-1,2,4,6-tetrakis-O-[di(benzyloxy)-
phosphoryl]-myo-inositol 13a
A mixture of bis(benzyloxy)diisopropylaminophosphine (1.38
g, 4 mmol) and 1H-tetrazole (0.56 g, 8 mmol) in dry methylene
dichloride (10 cm3) was stirred at room temperature for 10 min.
-3,5-Di-O-benzyl-myo-inositol 11a (0.216 g, 0.60 mmol) was
added to the reaction mixture and stirred for a further 10 min.
L-2,3,5-Tri-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-inositol
24b
A mixture of -5-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-
inositol 23b (0.50 g, 0.79 mmol) and sodium hydride (0.33 g,
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3555

View Article Online
8.25 mmol) was stirred at room temperature in DMF (25 cm3).
Benzyl bromide (0.24 cm3, 2.0 mmol) was added dropwise and
stirred for a further 2 h. TLC (CHCl3–pentane–EtOAc 3 : 3 : 1)
showed a new product Rf 0.50. The excess sodium hydride was
destroyed with methanol (10 cm3) and the solvents were
evaporated off in vacuo. The residue was partitioned between
methylene dichloride and water (50 cm3 of each) and the
organic layer was separated, dried (MgSO4) and the solvent was
evaporated off to give the crude product. Flash chromato-
graphy using the TLC solvent, provided the pure title com-
pound 24b (0.523 g, 82%); mp 84–86 ЊC (from Et2O–pentane);
[α]D Ϫ5 (c 1 in CHCl3); (Found: C, 75.6; H, 6.71. C51H54O9
requires C, 75.53; H, 6.71%); δH (400 MHz; CDCl3) 3.31 (1 H,
dd, J 2.4, 9.8, 1-H or 3-H), 3.32 (1 H, dd, J 2.1, 9.8, 1-H or 3-H),
3.44 (1 H, t, J 9.15, 5-H), 3.76 (3 H, s, OMe), 3.77 (3 H, s,
OMe), 3.81 (3 H, s, OMe), 4.00 (1 H, t, J 2.1, 2-H), 4.05 (1 H, t,
J 9.5, 4-H or 6-H), 4.06 (1 H, t, J 9.5, 4-H or 6-H), 4.52–4.88
(12 H, m, 3 × OCH2Ph and 3 × OCH2C6H4OMe), 6.77–7.42 (27
H, m, 3 × CH2Ph and 3 × OCH2C6H4OMe); δC (100 MHz;
CDCl3) 55.23 (CH2PhOMe), 72.44, 72.75, 74.08, 75.49, 75.82
(CH2Ph, CH2PhOMe), 74.47, 80.72, 80.97, 81.42, 83.76 (myo-
inositol ring carbons), 113.70, 113.75, 127.30, 127.41, 127.50,
127.56, 127.65, 127.76, 128.13, 128.33, 129.15, 129.73 (CH2Ph
and CH2PhOMe), 130.54, 131.03, 131.09, 138.48, 139.01,
159.09, 159.14 (Cq, CH2PhOMe and CH2Ph); m/z (FABϪ) 833.0
(30), 689.0 (90), 391.1 (14), 241.0 (24), 211.0 (78), 121.0 (100).
4 M. C. Hermosura, H. Takeuchi, A. Fleig, A. M. Riley, B. V. L.
Potter, M. Hirata and R. Penner, Nature (London), 2000, 408,
735–740.
5 S. T. Safrany, S. J. Mills, C. Liu, D. Lampe, N. J. Noble,
S. R. Nahorski and B. V. L. Potter, Biochemistry, 1994, 33,
10763–10769.
6 C. T. Murphy, A. J. Bullock, C. J. Lindley, S. J. Mills, A. M. Riley,
B. V. L. Potter and J. Westwick, Mol. Pharmacol., 1996, 50,
1223–1230.
7 G. Baudin, B. I. Glänzer, K. S. Swaminathan and A Vasella, Helv.
Chim. Acta, 1988, 71, 1367–1378.
8 C. E. Dreef, J.-P. Jansze, C. J. J. Elie, G. A. van der Marel and
J. H. van Boom, Carbohydr. Res., 1992, 234, 37–50.
9 K.-L. Yu and B. Fraser-Reid, Tetrahedron Lett., 1988, 29,
979–982.
10 S. J. Mills and B. V. L. Potter, J. Chem. Soc., Perkin Trans. 1, 1997,
1279–1286.
11 D. Lampe, C. Liu and B. V. L. Potter, J. Med. Chem., 1994, 37,
907–912.
12 J. Gigg, R. Gigg, S. Payne and R. Conant, J. Chem. Soc., Perkin
Trans. 1, 1987, 423–429.
13 S. J. Mills and B. V. L. Potter, J. Org. Chem., 1996, 61, 8980–8987.
14 B. V. L. Potter and D. Lampe, Angew. Chem., Int. Ed. Engl., 1995,
34, 1933–1972.
15 N. Chida, E. Yamada and S. Ogawa, J. Carbohydr. Chem., 1988, 7,
555–570.
16 N. Chida, Y. Furuno and S. Ogawa, J. Chem. Soc., Chem. Commun.,
1989, 1230–1231.
17 S. Ogawa, H. Kimura, C. Uchida and T. Ohashi, J. Chem. Soc.,
Perkin Trans. 1, 1995, 1695–1705.
18 S. K. Chung, Y. U. Kwon, J. H. Shin, Y. T. Chang, C. Lee, B. G. Shin,
K. C. Kim and M. J. Kim, J. Org. Chem., 2002, 67, 5626–5637.
19 F. De Smedt, B. Verjans, P. Mailleux and C. Erneux, FEBS Lett.,
1994, 347, 69–72.
L-2,3,5-Tri-O-benzyl-myo-inositol 25b
-2,3,5-Tri-O-benzyl-1,4,6-tris-O-p-methoxybenzyl-myo-ino-
sitol 24b (0.390 g, 0.48 mmol) was added to 1 mol dmϪ3 aq.
hydrochloric acid–ethanol (30 cm3, 1 : 2). The mixture was
heated under reflux for 4 h, cooled, and the solvents were evap-
orated in vacuo then the remaining solid was co-evaporated
with methanol (50 cm3). TLC (ether) showed a new product Rf
0.30 and a deprotected p-methoxybenzyl derivative Rf 0.80. The
remaining solid was partitioned between methylene dichloride
and water (50 cm3 of each) and the organic layer was separated
and dried (MgSO4) then evaporated to give the crude product.
The title compound 25b was obtained by flash chromatography
(EtOAc–CH2Cl2 1 : 1) and isolated as a solid (0.190 g, 88%); mp
175–176.5 ЊC (from EtOAc–hexane); lit.,13 176–177 ЊC; [α]D Ϫ29
(c 1 in CHCl3), lit.,13 [α]D Ϫ34 (c 1 in CH2Cl2); [same sample as
lit.,13 [α]D Ϫ28 (c 1 in CHCl3) for the -enantiomer]; δH (270
MHz; CDCl3) 2.30–2.70 (3 H, br s, D2O ex, OH), 3.24 (1 H, t,
J 9.2, 5-H), 3.30 (1 H, dd, J 2.4, 9.7, 1-H or 3-H), 3.40 (1 H, dd,
J 2.75, 9.7, 1-H or 3-H), 3.82 (1 H, t, J 9.5, 4-H or 6-H), 4.06
(1 H, t, J 2.6, 2-H), 4.12 (1 H, t, J 9.5, 4-H or 6-H), 4.59–4.99
(6 H, m, 3 × OCH2Ph), 7.25–7.40 (15 H, m, 3 × OCH2Ph).
20 K. Takazawa, M. Lemos, A. Delvaux, C. Lejeune, J. E. Dumont and
C. Erneux, Biochem. J., 1990, 268, 213–217.
21 M. Hirata, Y. Watanabe, M. Yoshida, T. Koga and S. Ozaki, J. Biol.
Chem., 1993, 268, 19260–19266.
22 Y.-C. Cheng and W. H. Prusoff, Biochem. Pharmacol., 1973, 22,
3099–3108.
23 A. P. Kozikowski, V. I. Ognyanov, A. H. Fauq, R. A. Wilcox and
S. R. Nahorski, J. Chem. Soc., Chem. Commun., 1994, 599–600.
24 M. A. Pollokoff, G. H. Bencen, J. P. Vacca, S. J. deSolms,
S. D. Young and J. R. Huff, J. Biol. Chem., 1988, 263, 11922–11927.
25 D. J. Gawler, B. V. L. Potter and S. R. Nahorski, Biochem. J., 1990,
272, 519–524.
26 G. St. J. Bird and J. W. Putney Jr., J. Biol. Chem., 1996, 271,
6766–6770.
27 G. Choi, Y.-T. Chang, S.-K. Chung and K. Y. Choi, Korean J. Med.
Chem., 1997, 7, 106–114.
28 S. K. Chung and Y. T. Chang, J. Chem. Soc., Chem. Commun., 1995,
11–12.
29 N. T. Burford, S. R. Nahorski, S.-K. Chung, Y.-T. Chang and
R. A. Wilcox, Cell Calcium, 1997, 21, 301–310.
30 I. Bosanac, J. R. Alattia, T. K. Mai, J. Chan, S. Talarico, F. K. Tong,
K. I. Tong, F. Yoshikawa, T. Furuichi, M. Iwai, T. Michikawa,
K. Mikoshiba and M. Ikura, Nature (London), 2002, 420, 696–700.
31 S. J. Mills, S. T. Safrany, R. A. Wilcox, S. R. Nahorski and B. V. L.
Potter, Bioorg. Med. Chem. Lett., 1993, 3, 1505–1510.
32 M. Hirata, N. Narumoto, Y. Watanabe, T. Kanematsu, T. Koga and
S. Ozaki, Mol. Pharmacol., 1994, 45, 271–276.
33 W. C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43,
2923–2925.
34 T. Tanaka, S. Tamatsukuri and M. Ikehara, Tetrahedron Lett., 1986,
27, 199–202.
Acknowledgements
We thank the Wellcome Trust for Programme Grant support
(060554 to B. V. L. P.). This work was also supported by a grant
from the Action de Rechere Concertée of the Communauté
Française de Belgique (to C. E.). We also thank Professor Colin
Taylor (University of Cambridge) for preliminary Ca2ϩ release
studies and Dr Andrew M. Riley for discussion.
35 C. Erneux, M. Lemos, B. Verjans, P. Vanderhaeghen, A. Delvaux
and J. E. Dumont, Eur. J. Biochem., 1989, 181, 317–322.
36 K. Takazawa, H. Passareiro, J. E. Dumont and C. Erneux, Biochem.
Biophys. Res. Commun., 1988, 153, 632–641.
37 K. Takazawa, J. Vandekerckhove, J. Dumont and C. Erneux,
Biochem. J., 1990, 272, 107–112.
38 F. De Smedt, A. Boom, X. Pesesse, S. Schiffmann and C. Erneux,
J. Biol. Chem., 1996, 271, 10419–10424.
39 H. J. Rosenberg, A. M. Riley, A. J. Laude, C. W. Taylor and B. V. L.
Potter, J. Med. Chem., 2003, in press.
References
1 M. J. Berridge, Nature (London), 1993, 315–325.
2 G. W. Mayr, Biochem. J., 1989, 259, 463–470.
3 V. Sylvia, G. Curtin, J. Norman, J. Stec and D. Busbee, Cell, 1988,
54, 651–658.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 4 6 – 3 5 5 6
3556