Conformational study of the natural iron chelator myo-inositol 1,2,3-trisphosphate using restrained/flexible analogues and computational analysis

Article abstract of DOI:10.1016/j.tet.2010.09.033

Myo-Inositol 1,2,3-trisphosphate [Ins(1,2,3)P₃], a component in mammalian cells, possesses the correct chemical properties of an intracellular iron transit ligand. Here we have examined the conformation of the Ins(1,2,3)P₃-Fe³⁺ complex. The synthesis and antioxidant properties of 4,6-carbonate-myo-inositol 1,2,3,5-tetrakisphosphate [4,6-carbonate Ins(1,2,3,5)P₄], which is locked in the unstable penta-axial chair conformation and 1,2,3-trisphosphoglycerol, a flexible acyclic analogue of Ins(1,2,3)P₃, are reported. 4,6-Carbonate Ins(1,2,3,5)P₄ caused complete inhibition of iron-catalysed hydroxyl radical (HO^?) formation at 100 μM, thereby resembling Ins(1,2,3)P₃ and supporting a penta-axial chair binding conformation. In contrast, 1,2,3-trisphosphoglycerol was shown to have incomplete antioxidant properties. In support of experimental observations, we have applied high-level density functional calculations to the binding of Ins(1,2,3)P₃ to iron. This study provides evidence that Fe³⁺ binds tightly to the less stable penta-axial conformation of Ins(1,2,3)P₃ using terminal and bridging phosphate oxygens, thought to also contain a tightly bound water molecule or hydroxyl ligand in the complex.

Full text of DOI:10.1016/j.tet.2010.09.033

Tetrahedron 66 (2010) 8949e8957
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Conformational study of the natural iron chelator myo-inositol 1,2,3-
trisphosphate using restrained/flexible analogues and computational analysis
David Mansell a, Nicolás Veiga b, Julia Torres b, Laura L. Etchells a, Richard A. Bryce a, Carlos Kremer b,
,
Sally Freeman a
*
a School of Pharmacy & Pharmaceutical Sciences, University of Manchester, Oxford Road, Manchester, M13 9PT, UK
b Cátedra de Química Inorgánica, Departamento Estrella Campos, Facultad de Química, Universidad de la República, CC 1157, Montevideo, Uruguay
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 June 2010
Received in revised form 19 August 2010
Accepted 13 September 2010
Available online 17 September 2010
myo-Inositol 1,2,3-trisphosphate [Ins(1,2,3)P3], a component in mammalian cells, possesses the correct
chemical properties of an intracellular iron transit ligand. Here we have examined the conformation of
the Ins(1,2,3)P3eFe3þ complex. The synthesis and antioxidant properties of 4,6-carbonate-myo-inositol
1,2,3,5-tetrakisphosphate [4,6-carbonate Ins(1,2,3,5)P4], which is locked in the unstable penta-axial chair
conformation and 1,2,3-trisphosphoglycerol, a flexible acyclic analogue of Ins(1,2,3)P3, are reported. 4,
6-Carbonate Ins(1,2,3,5)P4 caused complete inhibition of iron-catalysed hydroxyl radical (HOꢀ) formation
Keywords:
myo-Inositol 1,2,3-trisphosphate
Iron chelation
Chair conformation
Inositol phosphates
Computational analysis
at 100 mM, thereby resembling Ins(1,2,3)P3 and supporting a penta-axial chair binding conformation. In
contrast, 1,2,3-trisphosphoglycerol was shown to have incomplete antioxidant properties. In support of
experimental observations, we have applied high-level density functional calculations to the binding of
Ins(1,2,3)P3 to iron. This study provides evidence that Fe3þ binds tightly to the less stable penta-axial
conformation of Ins(1,2,3)P3 using terminal and bridging phosphate oxygens, thought to also contain
a tightly bound water molecule or hydroxyl ligand in the complex.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
(1B) (Scheme 1). The crystal structure of the cyclo-
hexylammonium salt of Ins(1,2,3)P3,6 showed the expected sta-
myo-Inositol 1,2,3-trisphosphate [Ins(1,2,3)P3, 1] and other
1,2,3-trisphosphate containing myo-inositol phosphates possess
remarkable iron chelation and antioxidant properties.1e4 We
have recently shown that 1, a cellular component in a variety of
mammalian cells, possesses the correct chemical properties
expected of an intracellular iron transit ligand.5 That is, (i) the
capacity to associate with iron to the full stoichiometric extent
even in the presence of cellular amounts of Mg2þ, (ii) the
capacity to maintain such full association with iron in acidic
and/or Ca2þ-rich compartments, and (iii) the capacity to inhibit
iron redox cycling and associated production of free radicals.
This final property is unique to the myo-inositol phosphates
containing the 1,2,3-trisphosphate motif, and is crucial given the
dangers associated with iron-catalysed free radical formation.
The binding conformation of Ins(1,2,3)P3, and more specifically
the orientation of the phosphate groups around iron, is key to
the binding of iron in a ‘safe’ manner. Ins(1,2,3)P3 can exist in
two chair conformations, orientating the phosphate groups
either equatorialeaxialeequatorial (1A) or axialeequatorialeaxial
ble penta-equatorial chair conformation. Using a pyrene-based
fluorescent probe, we have recently demonstrated that Ins(1,2,3)
P3 undergoes a ring-flip from the penta-equatorial (1A, 1a5e) to
the penta-axial chair conformation (1B, 5a1e) upon association
with Fe3þ (Scheme 1).7,8 This result is consistent with the Ins
(1,2,3)P3eFe3þ structure proposed by Phillippy and Graf, in
which the myo-inositol chair adopts the penta-axial conforma-
tion with hexa-coordination of iron by two terminal oxygens
from each phosphate group.9 This binding geometry was pro-
posed following an observation that the inhibition of Fe3þ-cat-
alysed hydroxyl radical formation by a ligand typically arises
from the occupation of all six coordination sites around iron,
thus preventing its participation in the HabereWeiss redox
cycles.10
* Corresponding author. Tel.: þ44 161 275 2366; fax: þ44 161 275 2396; e-mail
Scheme 1. Conformational ring-flip of Ins(1,2,3)P3 from the penta-equatorial (1A) to
address: sally.freeman@manchester.ac.uk (S. Freeman).
the penta-axial chair (1B) upon association with Fe3þ
.
0040-4020/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2010.09.033

8950
D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
Here, we extend the analysis of the Ins(1,2,3)P3eFe3þ complex
the 4,6-diaxial hydroxyl groups. However, a conformationally re-
stricted myo-inositol 1,2,3,5-tetrakisphosphate analogue, possess-
ing an additional P5 phosphate group, was chosen as a more
feasible synthetic target. The myo-inositol 1,2,3,5-tetraki-
sphosphate analogue was also employed in our pyrene-based
fluorescent probe and its Fe(III) chelation properties indicated that
a 1,2,3,5-tetrakisphosphate analogue was a good model for the
1,2,3-trisphosphate motif.7,8
The primary target was 4,6-methylene-myo-inositol 1,2,3,
5-tetrakisphosphate (2) (Fig. 1), which is locked in the penta-axial
chair conformation by a 4,6-bridging methylene acetal. The syn-
thesis of 2 started from myo-inositol (5), which was locked into the
penta-axial conformation by protection with an orthoformate
group to give 6.12 Selective silylation on the equatorial hydroxyl
group of 6 gave diol 7.11 The methylene acetal was introduced by
slow addition of dibromomethane:50% aqueous NaOH (1:1) and
cetyl-NMe3Br to a solution of 7 in DCM to give 8 (Scheme 2).13e15
Hydrolysis of the orthoformate and silyl groups in 8 was achieved
in the presence of p-toluenesulfonic acid to give 9. Phosphorylation
with the design of conformationally restricted Ins(1,2,3)P3 ana-
logues, which are locked in the penta-axial chair conformation
by either a methylene (2) or carbonate bridge (3) between the 4,
6-hydroxyl groups (Fig. 1). The methylene acetal displays particu-
larly high stability towards acidic cleavage, making it an ideal
candidate for a conformational lock. Alternatively, the cyclic car-
bonate protecting group has previously been used by Angyal to lock
the myo-inositol chair in its unstable penta-axial conformation.11 In
addition, 1,2,3-trisphosphoglycerol (4), a flexible acyclic analogue
of Ins(1,2,3)P3 (1) (Fig. 1), has been synthesised to probe the role of
the rigid cyclohexane framework in the binding of iron. The
straight-chain trisphosphoglycerol (4) structure could mimic the
conformation of the 1,2,3-trisphosphate motif, but lacks the rota-
tional restriction of the cyclohexane ring, allowing considerable
flexibility of the phosphate groups.
of tetraol
9 with dibenzyl N,N-diisopropylphosphoramidite,
followed by oxidation with m-chloroperoxybenzoic acid gave
tetrakisphosphate 10, albeit in a surprisingly low yield (3%). In
addition to the desired product,
a small amount of hex-
aphosphorylated myo-inositol was also formed, which may indicate
that the methylene bridge was labile under the phosphorylation
conditions. Alternatively the low yield may have resulted from in-
complete phosphorylation, however, we were unable to identify
any further products by flash column purification. Furthermore,
benzyl deprotection of 10 by catalytic hydrogenolysis was in-
conclusive, therefore the synthesis of the methylene analogue (2)
could not be pursued further. Consequently, the synthesis of the
4,6-carbonate analogue (3) was proposed as an alternative.
Fig. 1. Conformationally restricted (2) and (3), and acyclic (4) analogues of Ins(1,2,3)P3 (1).
The extent to which these analogues are able to bind iron and in-
hibit Fe3þ-catalysed hydroxyl radical formation should provide evi-
dence for the binding conformation of the 1,2,3-trisphosphate motif
to Fe3þ. These experimental observations are supported using high-
leveldensity functionalcalculationsoftheIns(1,2,3)P3eFe3þ complex.
2. Results and discussion
Pentasodium 4,6-carbonate-myo-inositol 1,2,3,5-tetrakisphos-
phate (3) was synthesised via diol 7 (Scheme 2). The cyclic car-
bonate bridge was introduced between the 4,6-diaxial hydroxyl
groups by reacting 7 with 1,10-O-carbonyldiimidazole at 65e70 ꢁC
in anhydrous THF to give the fully protected 11.11 Selective depro-
tection of an orthoformate group in a 4,6-carbonate-myo-inositol
based co-polymer has been reported using p-toluenesulfonic acid
2.1. Synthesis of the conformationally restricted penta-axial
analogues (2) and (3)
The initial aim was to prepare conformationally restricted ana-
logues of Ins(1,2,3)P3 (1), which are locked in the energetically less
stable penta-axial conformation by a short bridging group between
Scheme 2. Synthetic approaches towards the conformationally restricted analogues (2) and (3): (a) CH2Br2:50% NaOH (aq) (1:1), cetyl-Me3NBr, CH2Cl2; (b). Chloroform/methanol
(2:1), pTsOH; (c) (BnO)2PNPri2, 0.45 M 1H-tetrazole in MeCN, then mCPBA in anhydrous CH2Cl2; (d) 1,10-O-carbonyldiimidazole, anhydrous THF, 65e70 ꢁC; (e) p-toluenesulfonic acid,
THF/MeOH (2:1); (f) 80% trifluoroacetic acid (aq), 65e70 ꢁC; (g) H2, Pd/C (10%), THF. P¼(BnO)2P(O)ꢂ.

D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
8951
in a THF/methanol solution at room temperature.16 Using these
conditions, however, the deprotection of the orthoformate and
TBDMS groups in 11 was accompanied by partial cleavage of the
carbonate bridge, due to nucleophilic attack by methanol gener-
ating 4-(methoxycarbonyl)-myo-inositol (12). Carbonate 12 was
evident from the 1H NMR spectrum, with a singlet for the methoxy
group at
d 3.82 ppm and large Jax/ax coupling constants for the
inositol protons due to the penta-equatorial conformation. A pro-
cedure for the simultaneous deprotection of the orthoformate and
silyl groups in 11 is described by Angyal using aqueous trifluoro-
acetic acid at 65e70 ꢁC,11 however, under these conditions the
deprotection was accompanied by the formation of some myo-
inositol due to cleavage of the carbonate group.11 Here, depro-
tection of 11 using these conditions gave tetraol 13 contaminated
with a small amount of myo-inositol, which could be removed via
hot filtration with acetonitrile. Pure 4,6-carbonate-myo-inositol
(13) was obtained in an 83% yield. Phosphorylation of tetraol 13
proceeded smoothly to give the tetrakisphosphate 14. Benzyl
deprotection of 14 was achieved by catalytic hydrogenation to
generate the free acid, which was converted to the sodium salt (3)
on a Dowex cation-exchange column.
Scheme 3. Synthesis of 1,2,3-trisphosphoglycerol (4): (a) (BnO)2PNPri2, 0.45 M 1H-
tetrazole in MeCN, then mCPBA in CH2Cl2; (b) Dowex-50-X8 (Hþ form), methanol;
(c) H2, Pd/C, ethanol. P¼(BnO)2P(O)ꢂ
.
2.2. Synthesis of 1,2,3-trisphosphoglycerol (4)
Despite an extensive literature search, the synthesis of 1,2,
3-trisphosphoglycerol (4) has not been reported. A search on the
pharmacogenomics databases, Kyoto Encyclopedia of Genes and
Genomes (KEGG) and Human Metabolome Database (HMDB), did
not detect 4 by mass spectrometry, suggesting that it is not a nat-
ural product. The most direct route to synthesise 1,2,3-tris-
phosphoglycerol (4) would be to simultaneously phosphorylate the
three hydroxyl groups of glycerol (Scheme 3). However, attempted
phosphorylation of glycerol with dibenzyl N,N-diisopropylphos-
phoramidite in the presence of 1H-tetrazole in acetonitrile, fol-
lowed by oxidation with mCPBA did not provide phosphate triester
18. Several unidentified products, which could not be separated by
flash column chromatography, were formed. In contrast, 1,2,3-tri-
sphosphoglycerol 4 was readily prepared utilising commercially
available 1,2-isopropylidene glycerol (19, R¼H) (Scheme 3). Phos-
phorylation of 19 (R¼H) gave 20 [R¼(O)P(OBn)2], followed by
cleavage of the isopropylidene acetal to give diol 21.17 Further
phosphorylation of the diol afforded an efficient route to phosphate
triester 18. Catalytic hydrogenolysis of 18 to remove the benzyl
groups gave 1,2,3-trisphosphoglycerol (4), isolated as its tetraso-
dium salt.
Scheme 4. Synthesis of pentasodium myo-inositol 1,2,3,5-tetrakisphosphate (15): (a)
(BnO)2PNPri2, 0.45 M 1H-tetrazole in MeCN, then mCPBA in CH2Cl2; (b) H2, Pd/C,
ethanol. P¼(BnO)2P(O)ꢂ
.
co-workers4 and Graf and co-workers,10,3 utilising a hypoxanthine/
xanthine oxidase system for HOꢀ generation. The HOꢀ generated
reacts with dimethylsulfoxide (present in the assay). The first in-
termediate formed is the methyl radical, which reacts with O2 to
give the methylperoxy radical (MeeOeOꢀ). This decomposes to
form formaldehyde,20 which is measured using the Hantzsch
reaction with pentan-2,4-dione to give diacetyldihydrolutidine
(absorbance at 410 nm).21 The quantity of formaldehyde produced
reflects the ability of the myo-inositol phosphates [1, 15, 3 and Ins
(1,2,6)P3] and 1,2,3-trisphosphoglycerol (4) to chelate iron and
inhibit HOꢀ formation (Fig. 2).
Pentasodium myo-inositol 1,2,3,5-tetrakisphosphate [Ins
(1,2,3,5)P4, 15] was synthesised by phosphorylation of 4,6-diben-
zyl-myo-inositol (16), via the phosphate ester (17), following
18
ꢀ
a modified route described by Sala and co-workers (Scheme 4).
This provides a standard, alongside Ins(1,2,3)P3, to which the iron
chelation properties of the conformationally restricted (3) and
acyclic (4) analogues can be compared. The iron binding properties
of Ins(1,2,3,5)P4 (15) have previously been reported,1 however the
potassium salt was used19 and the assay conditions were slightly
different to those employed here.1
2.3. Inhibition of iron-catalysed hydroxyl radical formation
Fig. 2. Effects of Ins(1,2,3)P3 (1,
ꢀ
), Ins(1,2,3,5)P4 (15, :), 4,6-carbonate-myo-inositol
The inhibition of iron-catalysed hydroxyl radical formation
by Ins(1,2,3)P3 (1), myo-inositol 1,2,3,5-tetrakisphosphate (15), 4,
6-carbonate-myo-inositol 1,2,3,5-tetrakisphosphate (3) and 1,2,
3-trisphosphoglycerol (4) was studied using an assay for the Hab-
ereWeiss reaction. myo-Inositol 1,2,6-trisphosphate [Ins(1,2,6)P3]
was also included in the study to illustrate the behaviour of inositol
phosphates lacking the 1,2,3-trisphosphate grouping. The assay
conditions were based on those described by Hawkins and
1,2,3,5-tetrakisphosphate (3, -), 1,2,3-trisphosphoglycerol (4, ) and Ins(1,2,6)P3 (>)
on HOꢀ generation. The results shown are typical of three independent experiments.
Consistent with previous observations, both Ins(1,2,3)P3 (1) and
Ins(1,2,3,5)P4 (15) completely inhibited Fe3þ-catalysed hydroxyl
radical formation at w100 m
M.1 For the conformationally restricted
analogue, 4,6-carbonate Ins(1,2,3,5)P4 (3), a more rapid initial de-
crease in HOꢀ formation was observed, compared to Ins(1,2,3)P3 and

8952
D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
Ins(1,2,3,5)P4, and although HOꢀ formation was not completely
eliminated, it does reach a minimum at a concentration of
of iron and may explain why nature evolved Ins(1,2,3)P3 as an in-
tracellular iron shuttle,5 over the simpler 1,2,3-trisphosphoglycerol.
w100 mM. The iron binding properties observed for the penta-axial
analogue, 4,6-carbonate Ins(1,2,3,5)P4 (3), support the data
obtained using the fluorescent probe 4,6-bispyrenoyl myo-inositol
1,2,3,5-tetrakisphosphate for the complexation of Fe3þ with the
1,2,3-trisphosphate motif.7,8 In contrast, a continual reduction in
HOꢀ formation was observed for Ins(1,2,6)P3, which lacks the 1,2,
3-trisphosphate motif. Residual HOꢀ formation observed for 3 could
be rationalised on the basis of a second binding site being present in
the molecule, in the form of the syn-1,3,5-triaxial trisphosphate
groups (Fig. 3, B).22 Iron bound at site B is likely to possess an
available coordination site and therefore allow its participation in
the HabereWeiss cycles.
2.4. Computational analysis
To obtain detailed structural information, we have completed
quantum mechanical calculations at the UB3LYP/6-31þG
*
level of
theory on the high-spin Fe3þeIns(1,2,3)P3 complex. The proposal
by Graf and co-workers10 that inhibition of Fe3þ-catalysed HOꢀ
formation arises from occupation of all coordination sites around
iron arose from competitive displacement experiments using an
azide ligand. A relationship was observed between the presence of
a free iron coordination site (i.e., occupied by water or an easily
displaceable ligand) and the ability of an iron-chelate to inhibit HOꢀ
formation. However, this result might be compatible with the
presence of a tightly bound stable water molecule in the co-
ordination sphere of Fe3þ. Therefore a molecule of tightly bound
water may complete the coordination of iron in the Ins(1,2,3)
P3eFe3þ complex. Accordingly, we have extended our high-level
quantum mechanical calculations on the Ins(1,2,3)P3eFe3þ com-
plex to include a water molecule.
2.4.1. Comparison between penta-axial (5a1e) and penta-equatorial
(1a5e) conformations of Ins(1,2,3)P3. The penta-axial (5a1e) and
penta-equatorial (1a5e) myo-inositol conformations of Ins(1,2,3)P3
in the absence of Fe3þ were first compared. In Fig. 4, the electro-
static potential was mapped on to an isodensity surface for the
completely ionised form (L6ꢂ) of both conformers. The most neg-
ative zones are shown in red and the least negative in blue. In both
cases, the optimised geometries show that a hydrogen bond occurs
between the phosphate group at position 2 and one OH on C4 (5a1e
conformation) or C5 (1a5e conformation). For 1a5e (Fig. 4A), this
bond is strong enough to produce a little structural distortion,
leading the carbon ring conformation to become closer to a boat
Fig. 3. Two potential Fe3þ binding sites are present in 4,6-carbonate-myo-inositol
1,2,3,5-tetrakisphosphate (3): at the 1,2,3-trisphosphate motif (A) and the syn-1,3,5-
triaxial trisphosphate grouping (B).
In 1,2,3-trisphosphoglycerol (4), the flexibility of the phosphate
groups may facilitate occupation of all coordination sites on iron.
Consequently, 4 might be expected to possess similar iron chelation
and antioxidant properties to Ins(1,2,3)P3 and Ins(1,2,3,5)P4. How-
ever, 4 was unable to completely inhibit HOꢀ formation and instead
closely resembled the behaviour of Ins(1,2,6)P3. Thus, it is likely
that the conformational rigidity offered by the cyclohexane ring in
the myo-inositol 1,2,3-trisphosphates facilitates effective binding
Fig. 4. (A) Penta-equatorial (1a5e) and (B) penta-axial (5a1e) conformations of In(1,2,3)P3 in the absence of Fe3þ. (C) and (D) correspond to the electrostatic potential of (A) and (B)
mapped on to an isodensity surface (isodensity value¼0.0004 e, UB3LYP/6-31þG geometries). The strongest intramolecular hydrogen bonds are shown as dashed lines.
*

D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
8953
rather than a chair. As it is exposed, this conformation has the most
negatively charged zones between the phosphate groups, and a less
negatively charged zone (due to the hydrogen atom at position 2)
pointing towards the position where the metal ion will be bound
(Fig. 4C). In contrast, the 5a1e L6ꢂ conformation has a ‘pre-formed’
geometry for metal complexation, as its most negatively charged
zones are arranged in a ‘hole’, which can bind Fe3þ without sig-
nificant structural distortion (Fig. 4D). It seems that the 5a1e con-
formation of the L6ꢂ species has a geometry that is suitable to
complex the metal ion without a drastic change in the structure.
This was in support of our experimental observations; conse-
quently, we focused our computational study on the penta-axial
conformation of the Ins(1,2,3)P3eFe3þ complex (Fig. 5).
angles, which highlight the distortion of this octahedron, are: 70.7ꢁ
(X¼O2), 74.7ꢁ (X¼O3), 144.0ꢁ (X¼O4), 70.0ꢁ (X¼O5) and 100.0ꢁ
(X¼O6).
2.4.3. With one coordinated water molecule. In order to cover all
possibilities for the coordination of Ins(1,2,3)P3eFe3þ by a water
molecule, four configurations of the complex were explored at the
UB3LYP/LANL2DZ level. In three out of four optimised configura-
tions, a hydroxide group linked to the metal centre is formed, with
Ins(1,2,3)P3 protonated. The most stable geometry from this study
was used for UB3LYP/6-31þG
*
calculations. The optimal UB3LYP/
6-31þG
*
structure (Fig. 5B) differs slightly in metal coordination
environment compared to that obtained at the UB3LYP/LANL2DZ
level: the iron moves away from O1, changing the FeeO1 bond
2.4.2. Penta-axial conformation of Ins(1,2,3)P3-Fe3þ. Initially, con-
formational analysis of the penta-axial conformation of the [Ins
(1,2,3)P3eFe]3ꢂ was performed using the UB3LYP/LANL2DZ level,
*
distance from 2.27 Å (LANL2DZ) to 2.92 Å (6-31þG ). This effect can
be explained by means of the diffuse functions contained in the
Gaussian basis set used, which, in general, supports the existence of
weaker and longer chemical bonds through the space. In addition,
this result also gives an appreciation of the non-negligible volume
occupied by the OH group, which distorts to some extent the metal
coordination environment. Particularly interesting is the observed
proton transfer from the water molecule to the 2-phosphate group
of Ins(1,2,3)P3, which confers extra stability due to the formation of
a hydrogen bond OW/HWO2 (Fig. 5B). This OW/HW distance is
1.80 Å, while the O2-HW/OW angle is 176.9ꢁ and the HW/OW-Fe
angle is 115.3ꢁ.
prior to exploration at the UB3LYP/6-31þG
*
level. Similar structures
were obtained from both models, showing that there is no mac-
roscopic difference in the geometry as the method level increases.
In fact, the average bond distance between the metal ion and the
donor atoms of the ligands, D(FeeX), for both structures, are almost
identical at 2.08 Å (LANL2DZ) and 2.09 Å (6-31þG
*
). The optimal
UB3LYP/6-31þG
*
geometry obtained is shown in Fig. 5A. In contrast
to the structure suggested by Phillippy and Graf,9 our structure
predicts the involvement of both terminal oxygens and bridging
phosphoester oxygens from each phosphate. This results in
a strongly distorted octahedral geometry around iron. The six FeeO
bond distances are: 2.20 Å (FeeO1), 1.94 Å (FeeO2), 2.10 Å (FeeO3),
1.98 Å (FeeO4), 2.44 Å (FeeO5) and 1.88 Å (FeeO6). The O1eFeeX
The mean FeeX distance, D(FeeX), for [FeL(H2O)]3ꢂ at the
UB3LYP/LANL2DZ level is the same as that obtained for the anhy-
drous complex (2.08 Å, Table 1). However, for the 6-31þG
*
geom-
etry, the mean FeeX distance is longer when water is added (2.16 Å,
Fig. 5. Penta-axial conformation of Ins(1,2,3)P3 bound to Fe3þ in: (A) the absence of water, (B) the presence of water, and (C) the presence of azide. The strongest intramolecular
hydrogen bonds are shown as dashed lines. In (D), the electrostatic potential is mapped on to an isodensity surface for (C, [FeHLN3]3ꢂ). Isodensity value¼0.0004 e, UB3LYP/6-31þG
*
geometry.

8954
D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
Table 1), due to the previously mentioned distortion that surrounds
the Fe(III). Taking into account only the five shorter FeeX bonds
found in the complex, the mean bond distance is 2.00 Å. This shows
that increasing the FeeO1 bond length allows the other chelating
atoms to approach the metal ion more closely. Indeed, the HOꢂ
ligand is now 0.05 Å closer than in the UB3LYP/LANL2DZ structure.
consistent with the concentration at which Ins(1,2,3)P3 and Ins
(1,2,3,5)P4 achieved complete inhibition of HOꢀ formation. Residual
HOꢀ formation was rationalised on the basis of a second binding site
for Fe3þ (Fig. 3, B) due to the syn-1,3,5-triaxial trisphosphate group
arrangement.22 This provides evidence that Fe3þ binds tightly to
the penta-axial conformation of inositol phosphates possessing the
1,2,3-trisphosphate motif, consistent with previous studies with
a fluorescent probe.7,8 In contrast, 1,2,3-trisphosphoglycerol (4), the
acyclic analogue of Ins(1,2,3)P3, was unable to completely inhibit
iron-catalysed hydroxyl radical formation, making clear the re-
quirement for the rigid cyclohexane frame of myo-inositol on which
the 1,2,3-trisphosphate motif features.
Table 1
Mean distances (D) for UB3LYP/LANL2DZ and UB3LYP/6-31þG
*
optimised structures
of [FeL(H2O)]3ꢂ. X refers to the donor chelating O atoms, either from hydroxide or
phosphate
[FeL(H2O)n]3ꢂ
D(FeeX) (Å)
D(FeeOH) (Å)
UB3LYP/6-31þG
*
n¼0
n¼1
n¼1
2.09a
2.08a
2.16a
2.00b
d
Density functional calculations at the UB3LYP/6-31þG
*
level in-
UB3LYP/LANL2DZ
1.94
1.89
dicate that the penta-axial conformation of Ins(1,2,3)P3 possesses
a ‘pre-formed’ geometry ideal for Fe3þ complexation. Subsequent
calculations on the Ins(1,2,3)P3eFe3þ complex in the penta-axial
conformation predicted a highly distorted octahedral geometry due
to hexa-coordination of iron by a terminal oxygen and phosphoester
oxygen from each phosphate. The inclusion of a water molecule in
the coordination sphere alleviates the distortion around iron and is
consequently energetically favourable. Displacement of this tightly
bound water molecule with azide is energetically unfavourable,
which is in agreement with experimental observations in the liter-
ature.10 Therefore the Ins(1,2,3)P3eFe3þ complex is proposed to
adopt the 5a1e conformation which may contain a tightly bound
hydroxyl ligand originating from a deprotonated water molecule.
UB3LYP/6-31þG
*
a
D(FeeX) averaged over all six FeeX bonds.
D(FeeX) averaged over the five shorter FeeX bonds.
b
Using the calculated absolute values of Gibbs free energies, we
can compute the G for the reaction involving the addition of one
D
water molecule (Eq. 1). The addition of one water molecule to [Ins
(1,2,3)P3eFe]3ꢂ is predicted to be energetically favourable (at least
in the gas phase), with an exothermic
DG of ꢂ17.4 kcal/mol
(LANL2DZ) and ꢂ11.4 kcal/mol (6-31þG
*
).
½FeLꢃ3ꢂþH2O/½FeLðH2OÞꢃ3ꢂ
(1)
2.4.4. Water displacement by an azide ion. After consideration of
the observations by Graf and co-workers,10 we have evaluated the
displacement of the Fe-bound hydroxyl ligand from [FeL(H2O)]3ꢂ
4. Experimental
4.1. General methods
by azide (Eq. 2). The calculated
31þG level show that this reaction is certainly not spontaneous,
with an endothermic
G of þ 33.9 kcal/mol.
DG of reaction at the UB3LYP/6-
*
Reagents and solvents were purchased from Aldrich Chemical
Co. (Dorset, UK), VWR International (Leicestershire, UK) and Fisher
Scientific (Leicestershire, UK). Deuterated solvents and tetrame-
thylsilane (TMS) were obtained from Cambridge Isotope Labora-
tories Inc. (Andover, USA). Thin layer chromatography was
performed on pre-coated silica gel 60 F254 aluminium backed
plates (layer thickness 0.2 mm) (Merck). Spots were visualised by
ultraviolet radiation at 254 nm and 325 nm using a UV GL-58
mineral light lamp or by staining with KMnO4 solution. Melting
points (mp) were determined in open glass capillary tubes on
a GallenKamp MPD.350.BM2.5 apparatus microscope (UK) and are
uncorrected. Infra-red spectra were recorded on neat samples us-
ing a JASCO FT/IR 4100 instrument. NMR spectra were recorded
using a Bruker Avance-300 spectrometer operating at 300 MHz (1H
NMR), 75 MHz (13C NMR) or 121 MHz (31P NMR) at 293 K, unless
otherwise stated. The spectrometer was running TOPSPIN NMR
D
h
i
FeLðH2OÞ3ꢂ þN3ꢂ/½FeHLN3ꢃ3ꢂþOHꢂ
Inspection of the optimised geometries for [FeL(H2O)]3ꢂ and
(2)
[FeHLN3]3ꢂ complexes suggests that substitution of HOꢂ by azide
causes a substantial disturbance of the metal coordination sphere.
In this event, the phosphate at position 2 moves away from the
metal, losing the hydrogen bond with the OHꢂ group (Fig. 5B) and
giving rise to a pentacoordinated metal environment (Fig. 5C).
Accordingly, in view of the electrostatic potential of [FeHL(azide)]3ꢂ
(Fig. 5D), it is clear that the azide group faces the most negatively
charged zone, being largely repelled by adjacent groups. As a result,
the phosphate in position 2 is pulled apart from the metal ion. It
seems this repulsion is significant enough to prevent the expected
O2eHW/Nꢂ3 hydrogen bond from forming.
system software (Version 2.0). Chemical shifts (d) are reported in
parts per million (ppm). 1H NMR and 13C NMR spectra were ref-
erenced to Me4Si (0.00 ppm), unless otherwise stated. 31P NMR
spectra were referenced to 85% phosphoric acid. 1H NMR spectra
were assigned with the aid of COSY experiments and 13C NMR
spectra were assigned with the aid of DEPT experiments. Mass
spectra (MS) [Micromass PLATFORM II (ES) and Thermo Finnigan
MAT95XP (accurate mass)] and elemental analyses [EA 1108 Ele-
mental Analyser (Carlo Erba Instrument)] were recorded in the
School of Chemistry, University of Manchester (UK).
Therefore, from an energetic point of view, two main factors
could explain the positive
DG value determined for displacement
by azide. Firstly, this geometric deformation may be due to the
electrostatic repulsion between the Nꢂ3 ion and the phosphate
groups. In contrast to what happens to the OHꢂ ion, the hydrogen
bond cannot compensate it, and the phosphate groups are sepa-
rated from each other. Secondly, the azide ion is bulkier than the
hydroxyl group, the resulting steric repulsion also disfavouring
substitution by azide.
3. Conclusion
4.2. Synthesis
We report the synthesis and antioxidant properties of con-
formationally restricted pentasodium 4,6-carbonate-myo Ins
(1,2,3,5)P4 (3), which models the unstable axialeequatorialeaxial
conformation of the 1,2,3-trisphosphate motif. Although HOꢀ for-
mation was not completely eliminated in the presence of 3, it does
4.2.1. Attempted synthesis of 4,6-methylene-myo-inositol 1,2,3,
5-tetrakisphosphate (2).
4.2.1.1. 2-O-(tert-Butyldimethylsilyl)-4,6-methylene-myo-inositol
1,3,5-orthoformate (8). A solution of 2-O-(tert-butyldimethylsilyl)-
myo-inositol 1,3,5-orthoformate12,11 (7) (1.40 g, 4.60 mmol) in DCM
(80 mL) was added dropwise to a mixture of dibromomethane
reach a minimum at a concentration of w100
mM. This was

D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
8955
(100 mL), sodium hydroxide (120 mL, 50% aq) and cetyl-trimethy-
lammonium bromide (0.50 g, 1.38 mmol) at 65 ꢁC, with vigorous
stirring. The reaction mixture was stirred for a further 15 min and
then diluted with DCM (500 mL). The organic layer was washed with
water (2ꢄ50 mL) and brine (1ꢄ50 mL), dried over Na2SO4 and
concentrated under reduced pressure. The crude material was pu-
rified by flash column chromatography (hexane/ethyl acetate,
3:1/1:1) to give 8 as a colourless solid (0.55 g, 38%). Mp
102e102.5 ꢁC. 1H NMR (CDCl3): 0.15 (6H, s, SiMe2), 0.94 (9H, s, tBu),
4.08e4.17 (2H, m, H-1/3), 4.55 (1H, br s, H-2), 4.68 (1H, d, CHAHB, Jgem
5.9 Hz), 4.78 (2H, t, H-4/6, 3J4/6e1/3z3J4/6e5 4.7 Hz), 4.82e4.88 (1H,
4.2.2. Synthesis of 4,6-carbonate-myo-inositol 1,2,3,5-tetrakis-
phosphate 3.
4.2.2.1. 4,6-Carbonate-myo-inositol (13). The deprotection of
2-O-tert-butyldimethyl-4,6-carbonate-myo-inositol
1,3,5-ortho-
formate (11) (0.05 g, 0.15 mmol) in the presence of p-toluene-
sulfonic acid (0.015 g, 0.08 mmol) in chloroform/methanol (2:1,
2 mL) was unsuccessful. The solution was stirred at 45 ꢁC for 48 h
and then cooled on ice. The precipitate was isolated by filtration
and the colourless solid identified as 4-O-methoxycarbonyl-myo-
inositol (12) (0.02 g, 61%). Mp 173.5e175.5 ꢁC. 1H NMR (D2O, ref-
erenced to DMSO at 2.71 ppm): 3.47 (1H,t, H-5, 3J2e1z3J2e3 9.5 Hz),
5
m, H-5), 5.20 (1H, d, CHAHB, Jgem 5.9 Hz), 5.48 (1H, d, H-7, J7e2
0.7 Hz). 13C NMR (CDCl3): ꢂ4.61 (SiMe2), 18.4 (CMe3), 25.9 (CMe3),
61.6 (inositol CH), 61.9 (inositol CH), 66.3 (2ꢄinositol CH), 72.3
(2ꢄinositol CH), 84.7 (CH2), 102.3 (C-7). MS (electrospray) m/e
[MþNa]þ 339. Anal. Calcd for C14H24O6Si: C 53.14; H 7.65%. Found: C
52.90; H 7.64%. IR data: nmax 2947 (m), 2927 (m), 2877 (m), 2857 (m)
(CeH), 1255 (SieMe), 1144 (s), 1131 (s).
3.54 (1H, dd, H-1, J1e6 9.9 Hz, J1e2 2.8 Hz), 3.70 (1H, t, H-6,
3
3
3J6e1z3J6e5 9.9 Hz), 3.74 (1H, dd, H-3, J3e4 10.3 Hz, J3e2 2.6 Hz),
3.82 (3H, s, MeO), 4.08 (1H, t, H-2, 3J2e1z3J2e3 2.8 Hz), 4.85 (1H, t,
H-4, 3J4e3z3J4e5 9.9 Hz). 13C NMR (D2O, referenced to DMSO at
39.4 ppm): 56.1 (CH3), 69.9 (CH, inositol), 71.6 (CH, inositol), 72.7
(CH, inositol), 72.8 (CH, inositol), 72.9 (CH, inositol), 80.0 (CH,
inositol), 157.1 (C]O). MS (electrospray) m/e [MþNa]þ 261. Accu-
rate mass calcd for C8H14O8Na: 261.0586. Found: 261.0581 (error:
ꢂ2.1 ppm). IR data: nmax 3328 (br) (OeH), 1721 (C]O), 1310 (CeH).
4,6-Carbonate-myo-inositol (13) was subsequently prepared
according to the literature procedure.11 Under these conditions
myo-inositol was formed as an impurity and was removed by hot
filtration in acetonitrile. Mp 171e172 ꢁC, lit. mp 172 ꢁC.11 1H NMR
3
3
4.2.1.2. 4,6-Methylene-myo-inositol (9). A solution of 8 (0.09 g,
0.28 mmol) and p-toluenesulfonic acid (0.03 g, 0.14 mmol) in
chloroform/methanol (3 mL, 2:1) was stirred at 40 ꢁC for 48 h.
Upon cooling a precipitate formed containing a mixture of 4,6-
methylene-myo-inositol (9) and a small amount of myo-inositol
(4). The precipitate was filtered, washed with cold diethyl ether
(3ꢄ1 mL) and the product purified by hot filtration in acetonitrile
to give 9 as a colourless solid (0.04 g, 74% yield). Mp 194e196 ꢁC.
1H NMR (D2O, referenced to DMSO at 2.71 ppm): 4.13 (2H, br t, H-
(D2O, referenced to acetonitrile at
d 2.06 ppm): 4.00 (1H, t, H-2,
3
3J2e1/3 4.3 Hz), 4.36e4.44 (2H, m, H-1/3), 4.58 (1H, tt, H-5, J5e4/6
4
3.7 Hz, J5e1/3 1.2 Hz), 4.90 (2H, t, H-4/6, 3
J
4/6e1/3z3J4/6e5 3.8 Hz).
3
1/3, 3J1/3e2z3J1/3e4/6 3.6 Hz), 4.36 (1H, t, H-2, J2e1/3 4.4 Hz), 4.51
(2H, t, H-4/6, 3J4/6e1/3z3J4/6e5 3.8 Hz), 4.68 (1H, d, CHAHB-8, Jgem
4.2.2.2. 4,6-Carbonate-myo-inositol
1,2,3,5-tetrakis(dibenzyl
3
5.9 Hz), 4.85 (1H, t, H-5, J5e4/6 4.1 Hz), 5.17 (1H, d, CHAHB-8, Jgem
phosphate) (14). The phosphorylation of 13 was achieved under the
conditions described for the preparation of compound 10. The
crude residue was purified by flash column chromatography
(hexane/ethyl acetate 3:1/0:1) to give 14 as a colourless oil (0.19 g,
31%). 1H NMR (CDCl3): 4.51e4.65 (2H, m, H-5þH-2), 4.83 (2H, t, H-
4/6, 3J4/6e1/3z3J4/6e5 3.7 Hz), 4.89e5.11 (18H, m, H-1/
3þ8ꢄPeOeCH2ePh), 7.05e7.36 (40H, m, Ph). 13C NMR (CDCl3):
64.1 (d, JPC 4.3 Hz, inositol C-5), 66.5 (dtwq, 3JPCz2JPC 5.3 Hz, ino-
sitol C-2), 69.7 (d, JPC 5.6 Hz, 2ꢄPhCH2OP), 70.0 (d, JPC 5.9 Hz,
2ꢄPhCH2OP), 70.2 (d, JPC 5.6 Hz, 2ꢄPhCH2OP), 70.4 (d, JPC 5.7 Hz,
2ꢄPhCH2OP), 71.1 (t, JPC 4.7 Hz, inositol C-4/6), 72.8 (dd, 3JPC 4.8 Hz,
2JPC 2.9 Hz, inositol C-1/3), 127.9, 128.0, 128.1, 128.2, 128.4, 128.5,
128.5, 128.7, 128.9 (all s, aromatic), 135.1 (Cq, d, JPC 6.7 Hz,
2ꢄPhCH2OP), 135.27 (Cq, d, JPC 6.7 Hz, 2ꢄPhCH2OP), 135.31 (Cq, d,
JPC 7.4 Hz, 2ꢄPhCH2OP), 135.35 (Cq, d, JPC 7.3 Hz, 2ꢄPhCH2OP), 143.8
(C]O). 31P NMR (CDCl3): ꢂ1.45 (2P), ꢂ1.37 (1P), ꢂ0.96 (1P). 31P
NMR (CDCl3, 1H-coupled): ꢂ1.46 (sextet, JPH 8.8 Hz), ꢂ1.39 (sextet,
JPH 8.8 Hz), ꢂ0.97 (sextet, JPH 9.2 Hz). MS (electrospray) m/e
[MþNa]þ 1269.2. IR data: nmax 1789 (C]O), 1092 (PeOeBn).
6.0 Hz). 13C NMR (D2O, referenced to DMSO at 39.4 ppm): 62.6
(inositol CH), 65.5 (inositol CH), 72.0 (2ꢄinositol CH), 72.3
(2ꢄinositol CH), 85.1 (CH2). MS (electrospray þve) m/e [MþNa]þ
215.1; (electrospray ꢂve) m/e [Mþe]ꢂ 190.9. Anal. Calcd for
C7H12O6: C 43.75; H 6.29%. Found: C 44.14; H 5.94%. IR data: nmax
3417 (m), 3300 (m) (OeH), 2902 (w) (CeH).
4.2.1.3. 4,6-Methylene-myo-inositol
1,2,3,5-tetrakis(dibenzyl
phosphate) (10). Dibenzyl N,N-diisopropylphosphoramidite
(2.38 mL, 7.08 mmol) was added to a mixture of 4,6-methylene-
myo-inositol (9) (0.17 g, 0.88 mmol) and 1H-tetrazole (w0.45 M
solution in MeCN) (31.5 mL, 14.2 mmol). The reaction mixture was
stirred under argon for 4 h at room temperature, then cooled to
ꢂ40 ꢁC.
A solution of 3-chloroperoxybenzoic acid (3.0 g,
w13.5 mmol) in anhydrous DCM (80 mL) was added dropwise.
The reaction mixture was stirred at 0 ꢁC for 3 h, followed by di-
lution with DCM (150 mL) and then washed with 10% Na2SO3
(2ꢄ80 mL), satd NaHCO3 (2ꢄ80 mL) and brine (1ꢄ80 mL). The
organic layer was dried over MgSO4, concentrated under reduced
pressure and purified by flash column chromatography (hexane/
ethyl acetate, 2:1/0:1) to give 10 as a colourless oil (0.03 g, 3%).
1H NMR (CDCl3): 4.25 (2H, br s, H-4/6), 4.57 (2H, br s, methyl-
eneeCH2), 4.88e5.11 (20H, m, 8ꢄPeOeCH2PhþH-1/3þH-2þH-5),
7.02e7.28 (40H, m, Ph). 13C NMR (100 MHz, CDCl3): 69.0e69.2 (m,
inositol CH), 69.4 (d, JPC 5.5 Hz, 2ꢄPhCH2OP), 69.6 (d, JPC 5.9 Hz,
2ꢄPhCH2OP), 69.7 (d, JPC 5.7 Hz, 2ꢄPhCH2OP), 69.9 (d, JPC 5.5 Hz,
2ꢄPhCH2OP), 70.8e71.2 (m, inositol CH), 74.6 (br s, C-1/3), 82.8 (s,
C-4/6), 104.0 (CH2), 127.7, 128.0, 128.19, 128.23, 128.27, 128.32,
128.37, 128.43, 128.6 (all s, aromatic CH), 135.6 (d, JPC 7.4 Hz, 2ꢄCq
PhCH2OP), 135.6 (d, JPC 6.6 Hz, 2ꢄCq PhCH2OP), 135.7 (d, JPC 7.8 Hz,
4ꢄCq PhCH2OP). 31P NMR (CDCl3): ꢂ1.48 (2P), ꢂ1.19 (1P), ꢂ0.50
(1P). 31P NMR (CDCl3, 1H-coupled): ꢂ1.50 (sextet, JPH 7.5 Hz, 2P),
ꢂ1.20 (dpentet, JPH 11.6 Hz, JPH 7.3 Hz, 1P), ꢂ0.53 (br q, JPH 7.3 Hz,
1P). MS (electrospray) m/e [MþNa]þ 1256. IR data: nmax 1277 (s,
P]O), 997 (PeOeBn).
4.2.2.3. Pentasodium 4,6-carbonate-myo-inositol 1,2,3,5-tetrakis-
phosphate (3). A mixture of 14 (0.048 g, 0.038 mmol) and PdeC
(10%, 0.013 g) was stirred in tetrahydrofuran (10 mL) under a hy-
drogen atmosphere for 24 h at room temperature. The catalyst was
removed by filtration through a pad of Celite and the solvent re-
moved under reduced pressure. The residue was dissolved in water
(2 mL) and the solution applied to a cation-exchange column
(Dowex-50-X8, mesh 20e50, 5 mL, Naþ form) and eluted with
water (25 mL). The eluent was concentrated to approximately 3 mL
and lyophilised to give 3 as a colourless solid (0.021 g, 79%). Mp
>300 ꢁC. 1H NMR (500 MHz, D2O, referenced to DMSO at
2.71 ppm): 4.47e4.54 (1H, m, H-2), 4.82e4.91 (3H, m, H-1/3þH-5),
5.16 (2H, t, H-4/6, 3J4/6e1/3z3J4/6e5 2.8 Hz). 13C NMR (D2O, refer-
enced to DMSO at 39.39 ppm): 63.6 (d, JPC 4.3 Hz, C-5), 66.0 (dtwq,
2JPCz3JPC 5.4 Hz, C-2), 70.6 (t, 3JPC 4.8 Hz, C-4/6), 75.4 (d, JPC 4.2 Hz,
C-1/3), 149.5 (C]O). 31P NMR (D2O): 1.17 (1P), 1.43 (1P), 1.51 (2P).

8956
D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
31P NMR (D2O, 1H-coupled): 1.15 (d, JPH 8.7 Hz), 1.47 (br s). Anal.
Calcd for C7H9O19P4Na5.3H2O: C 12.18; H 2.19; P 17.96%. Found: C
12.35; H 2.29; P 17.60%. IR data: nmax 3336 (br) (residual water),
1743 (C]O), 1650 (br).
3.19 (2H, s, 2ꢄOH), 3.56 (1H, dd, Jgem 11.6 Hz, 3J3A-2 5.2 Hz, CHAHB-3),
3.61 (1H, dd, Jgem 11.6 Hz, 3J3B-2 4.3 Hz, CHAHB-3), 3.82 (1H, pentet,
3J2e1z3J2e3 5.0 Hz, H-2), 4.03 (2H, dd, JPH 9.0 Hz, 3J1e2 5.3 Hz, CH2-
1), 5.02 (2H, dd, Jgem 11.4 Hz, JPH 7.5 Hz, 2ꢄPeOeCHAHBPh), 5.05
(2H, dd, Jgem 11.4 Hz, JPH 7.8 Hz, 2ꢄPeOeCHAHBPh), 7.30e7.40 (10H,
m, Ph). 13C NMR (CDCl3): 62.7 (CH2-3), 68.6 (d, JPC 5.9 Hz, CH2-1),
69.7 (d, JPC 5.5 Hz, 2ꢄPhCH2OP), 70.6 (d, JPC 6.1 Hz, CH-2), 128.0
(Ph), 128.6 (Ph) 128.7 (Ph), 135.5 (d, JPC 6.7 Hz, 2ꢄquat. PhCH2OP).
31P NMR (CDCl3): 0.56 (1P). 31P NMR (CDCl3, 1H-coupled): 0.56 (1P,
septet, JPH 8.3 Hz).
4.2.3. Synthesis
of
myo-inositol
1,2,3,5-tetrakisphosphate
(15). 4.2.3.1. 4,6-Dibenzyl-myo-inositol (16). p-Toluenesulfonic acid
monohydrate (0.16 g, 0.83 mmol) was added to 2-O-(tert-butyldi-
methylsilyl)-4,6-dibenzyl-myo-inositol
1,3,5-orthoformate11,12,23
(0.40 g, 0.83 mmol) in methanol (10 mL). The mixture was stirred
at room temperature for 48 h. The reaction mixture was concen-
trated and the residue purified by flash column chromatography
(DCM/methanol 9:1) to give 16 as a colourless solid (0.24 g, 81%).
Mp 138e139 ꢁC, lit. mp 138.5e139 ꢁC.24 1H NMR (CDCl3): 2.60 (3H,
br s, 3ꢄOH), 2.89 (1H, br s, OH), 3.52 (2H, dd, H-1/3, 3J1/3e4/6 9.2 Hz,
3J1/3e2 2.3 Hz), 3.53 (1H, t, H-5, 3J5e4/6 9.5 Hz), 3.67 (2H, t, H-4/6, 3J4/
4.2.4.3. 1,2,3-Tris(dibenzyl phospho)glycerol (18). The phos-
phorylation of 21 was performed as described for the preparation of
compound 10. The crude residue was purified by flash column
chromatography (hexane/ethyl acetate 2:1/1:2) to give 18 as
a colourless oil (0.44 g, 60%). 1H NMR (500 MHz, CDCl3): 4.02 (2H,
dddwdt, Jgem 11.2 Hz, JPHw3JHH 5.6 Hz, CHAHB-1, CHAHB-3), 4.09 (2H,
dddwdt, Jgem 11.0 Hz, JP3Hw3JHH 5.4 Hz, CHAHB-1, CHAHB-3), 4.55 (1H,
dpentet, JPH 9.1 Hz, J2e1/3 4.5 Hz, H-2), 4.94e5.03 (12H, m,
6ꢄPeOeCH2Ph), 7.22e7.34 (30H, m, Ph). 13C NMR (CDCl3): 65.5 (t,
JPC 4.9 Hz, 2ꢄCH2-1/3), 69.6 (br d, JPC 5.6 Hz, 3ꢄPhCH2OP), 74.6 (td,
3JPC 8.7 Hz, 2JPC 5.3 Hz, CH-2), 128.0 (Ph), 128.1 (Ph), 128.6 (br s, Ph),
128.7 (br s, Ph), 135.7 (br d, JPC 6.7 Hz, 3ꢄquat. PhCH2OP). 31P NMR
(CDCl3): ꢂ1.01 (2P, P-1/3), ꢂ1.68 (1P, P-2). 31P NMR (CDCl3, 1H-
coupled), ꢂ1.01 (2P, septet, JPH 7.6 Hz, P-1/3), ꢂ1.69 (1P, sextet, JPH
7.9 Hz, P-2). MS (electrospray) m/e [MþNa]þ 895. Accurate mass
calcd for C45H47O12P3Na: 895.2173. Found 895.2176 (error:
0.4 ppm). IR data: nmax 1266 (s, P]O), 992 (PeOeBn).
3
6e1/3z3J4/6e5 9.3 Hz), 4.08 (1H, t, H-2, J2e1/3 2.6 Hz), 4.83 (2H, d,
2ꢄOeCHAHBPh, Jgem 11.4 Hz), 4.89 (2H, d, 2ꢄOeCHAHBPh, Jgem
11.4 Hz), 7.25e7.44 (10H, m).
4.2.3.2. 4,6-Dibenzyl-myo-inositol 1,2,3,5-tetrakis(dibenzyl phos-
phate) (17). The phosphorylation of 16 was performed as described
for the preparation of compound 10. The crude residue was purified
by flash column chromatography (hexane/ethyl acetate, 3:1/0:1)
to give 17 (0.42 g, 54%) as a colourless oil. 1H NMR (400 MHz,
CDCl3): 3.93 (2H, t, H-4/6, 3J4/6e1/3z3J4/6e5 9.5 Hz), 4.38 (2H, tt, H-1/
3, 3J1/3e4/6z3JPH 9.6 Hz, 3J1/3e2z4JPH 1.9 Hz), 4.45 (1H, q, H-5, 3J5e4/
6z3JPH 9.3 Hz), 4.63 (2H, dd, PeOeCH2Ph, Jgem 11.8 Hz, 3JPH 8.9 Hz),
4.67e4.88 (8H, m, 4ꢄCH2Ph), 4.90e5.05 (6H, m, 3ꢄCH2Ph), 5.12
3
3
(4H, d, 2ꢄCH2Ph, JPH 7.6 Hz), 5.42 (1H, dt, H-2, JPH 8.5 Hz, J2e1/3
2.1 Hz), 6.95e7.00 (4H, m, Ph), 7.05e7.10 (4H, m, Ph), 7.12e7.42
(42H, m, Ph).
4.2.4.4. Tetrasodium 1,2,3-trisphosphoglycerol (4). Hydrogenation
of compound 18 was performed as described for the preparation of 3
to give tetrasodium 1,2,3-trisphosphoglycerol (4) as a hydroscopic
gum (0.11 g, quantitative yield). 1H NMR (D2O, referenced to acetone
3
4.2.3.3. Pentasodium
myo-inositol
1,2,3,5-tetrakisphosphate
at 2.22 ppm): 4.39 (1H, dpentet, JPH 9.6 Hz, J2e1/3 4.8 Hz, H-2),
(15). Hydrogenation of compound 17 was performed as described
for the preparation of 3 to give pentasodium myo-inositol 1,2,3,5-
tetrakisphosphate (15) as a colourless solid in a quantitative yield.
Mp >300 ꢁC, lit. mp >300 ꢁC.18 1H NMR (400 MHz, D2O, water
suppression, referenced to acetone at 2.22 ppm): 3.94 (2H, t, H-4/6,
3J4/6e5z3J4/6e1/3 9.3 Hz), 4.00 (1H, q, H-5, 3J5e4/6zJPH 8.4 Hz), 4.11
(2H, dddwbr t, H-1/3, 3J1/3e4/6zJPH 8.8 Hz, coupling to H-2 not
4.07e3.94 (4H, m, CH2-1/3). 13C NMR (D2O, referenced to acetone at
2
3
30.9): 64.7 (t, 2JCPz3JCP 4.7 Hz, CH2-1/3), 73.9 (dt, JCP 8.1 Hz, JCP
5.1 Hz, CH-2). 31P NMR (D2O): 1.38 (s, P-1/3), 0.53 (s, P-2). 31P NMR
(D2O, 1H-coupled): 1.38 (br s, P-1/3), 0.53 (d, JPH 8.1 Hz, P-2). Anal.
Calcd for C3H7P3O12Na4$4H2O: C 7.32; H 3.07; P 18.89%. Found: C
7.36%, H 3.53%, P 18.79%. IR data: nmax 3357 (br, residual water), 1684
(br, residual water), 1184 (m, P]O).
3
observed), 4.93 (1H, td, H-2, JPH 9.8 Hz, J2e1/3 2.3 Hz). 13C NMR
3
(D2O, referenced to DMSO at 39.39 ppm): 71.4 (dd, C-4/6, JPC
6.4 Hz, 3JPC 3.4 Hz), 74.8 (dd, C-1/3, 2JPC 5.2 Hz, 3JPC 2.8 Hz), 75.5 (d,
C-5, JPC 5.8 Hz), 79.3 (d, C-2, 2JPC 6.3 Hz, 3JPC not observed). 31P NMR
(D2O): 0.34 (3P), 0.98 (1P). 31P NMR (D2O, 1H-coupled): 0.35 (br s),
0.98 (d, JPH 4.8 Hz). Anal. Calcd for C6H11O18P4Na5$4H2O: C 10.57; H.
2.81; P 18.17. Found: C 10.90; H. 2.79; P 17.78%.
4.3. Hydroxyl radical assay
Ultra-violet (UV) absorbance spectra were recorded on a Ther-
moSpectronic Unicam-300 spectrophotometer (UK) running Vision
32 software (version 1.25), using disposable 1.5 mL semi-micro
cuvettes (path length 1 cm). The assay conditions employed were
based on literature described by Hawkins and co-workers4 and Graf
and co-workers.10,3 The standard assay mixture contained: 20 mM
Trizma (Sigma) buffered to pH 7.5 with HCl, 50 mM dimethylsulf-
4.2.4. Synthesis of 1,2,3-trisphosphoglycerol (4). 4.2.4.1. 1-(Dibenzyl
phospho)-2,3-isopropylidene glycerol (20). The phosphorylation of
19 was performed as described for the preparation of compound 10.
The crude residue was purified by flash column chromatography
(hexane/ethyl acetateþ1% triethylamine 3:1/1:1) to give 20 as
a colourless oil (0.26 g, 86%). 1H NMR (500 MHz, CDCl3): 1.33 (3H, s,
oxide, 0.3 mM hypoxanthine, 5
mM FeCl3 (added from a freshly
prepared 50 M stock solution) and 18 m-units of xanthine oxidase
m
(Sigma; lyophilised powder from bovine milk). For the stock hy-
poxanthine solution, addition of 0.1 M NaOH solution prior to di-
lution with buffer was required to aid solubility. For each
measurement, triplicate 1.0 mL samples were prepared containing
the assay mixture and varying concentrations of chelator (added
from 1 mM stock solution). The fractions were incubated at 37 ꢁC
for 30 min, and then quenched by immersion in a boiling-water
3
CH3), 1.38 (3H, s, CH3), 3.73 (1H, dd, CHAHB-3, Jgem 8.5 Hz, J3e2
5.5 Hz), 3.93 (1H, dddwdt, CHAHB-1, Jgem 11.2 Hz, JPHw3J1e2 6.4 Hz),
3.96e4.00 (1H, m, 1ꢄCHAHB-1), 4.01 (1H, dd, CHAHB-3, Jgem 6.2 Hz,
3J3e2 4.7 Hz), 4.21 (1H, pentet, H-2, 3J2e1z3J2e3 5.8 Hz), 5.04 (2H,
dd, 2ꢄPeOeCHAHBPh, Jgem 11.6 Hz, JPH 8.4 Hz), 5.07 (2H, dd,
2ꢄPeOeCHAHBPh, Jgem 11.6, JPH 8.3 Hz), 7.31e7.39 (10H, m, 2ꢄPh).
bath for 2 min, followed by addition of 250 mL formaldehyde de-
4.2.4.2. 1-(Dibenzyl phospho)glycerol (21). Compound 21 was
prepared according to the literature procedure.17 The crude residue
was purified by flash column chromatography (ethyl acetate) to
give 21 as a colourless oil (0.34 g, 96%). 1H NMR (500 MHz, CDCl3):
tection reagent (prepared from 15 g of ammonium acetate, 0.3 mL
of acetic acid and 0.2 mL of acetylacetone in a final volume of
100 mL of water). The fractions were incubated at 37 ꢁC for 40 min
to develop the colour attributed to the formation of

D. Mansell et al. / Tetrahedron 66 (2010) 8949e8957
8957
6. Spiers, I. D.; Freeman, S.; Schwalbe, C. H. J. Chem. Soc., Chem. Commun. 1995,
2219e2220.
7. Mansell, D.; Rattray, N.; Etchells, L. L.; Schwalbe, C. H.; Blake, A. J.; Bichenkova,
E. V.; Bryce, R. A.; Barker, C. J.; Diaz, A.; Kremer, C.; Freeman, S. Chem. Commun.
2008, 5161e5163.
8. Mansell, D.; Rattray, N.; Etchells, L.L.; Schwalbe, C.H.; Blake, A.J.; Torres, J.;
Kremer, C.; Bichenkova, E.V.; Barker, C.J.; Freeman, S. Org. Biomol. Chem. 2010, 8,
2850e2858.
9. Phillippy, B. Q.; Graf, E. Free Radic. Biol. Med. 1997, 22, 939e946.
10. Graf, E.; Mahoney, J. R.; Bryant, R. G.; Eaton, J. W. J. Biol. Chem. 1984, 259,
3620e3624.
diacetyldihydrolutidine, and the absorbance measured at 410 nm.
The blanks were incubated with heat-inactivated enzyme.
4.4. Computational methods
The current density functional theoretical study focuses on fully
deprotonated Ins(1,2,3)P3 and its high-spin Fe3þ complexes,
using the hybrid UB3LYP functional.25,26 The basis sets LANL2DZ27
and 6-31þG
*
were applied throughout this work. All geometry op-
11. Angyal, S. J. Carbohydr. Res. 2000, 325, 313e320.
timisations were performed via the modified GDIIS algorithm in gas
phase as implemented in Gaussian 03.28 The effectof the solvent was
studied through a discrete model with one water molecule. The final
structures obtained were all minima in the potential energy surface,
the natureof the stationary points being verified through vibrational
analysis. ZPE and thermochemical corrections derived from this
analysis were used to compute free energies at 298 K.
12. Praveen, T.; Shashidhar, M. S. Carbohydr. Res. 2001, 330, 409e411.
13. Norman, D. G.; Reese, C. B.; Serafinowska, H. T. Synthesis 1985, 751e754.
14. Kim, K. S.; Szarek, W. A. Synthesis 1978, 48e50.
15. Nicolaou, K. C.; Mitchell, H. J.; Fylaktakidou, K. C.; Rodriguez, R. M.; Suzuki, H.
Chem.dEur. J. 2000, 6, 3116e3148.
16. Kim, T.-H.; Holmes, A. B. J. Korean Chem. Soc. 2006, 50, 129e136.
17. Ryan, M.; Smith, M. P.; Vinod, T. K.; Lau, W. L.; Keana, J. F. W.; Griffith, O. H.
J. Med. Chem. 1996, 39, 4366e4376.
ꢀ
ꢀ
ꢀ
18. Sala, M.; Kolar, J.; Strlic, M.; Kocevar, M. Carbohydr. Res. 2006, 341, 897e902.
19. Chung, S.-K.; Chang, Y.-T. J. Chem. Soc., Chem. Commun. 1995, 11e12.
20. Klein, S. M.; Cohen, G.; Cederbaum, A. I. Biochemistry 1981, 20, 6006e6012.
21. Nash, T. Biochem. J. 1953, 55, 416e421.
Acknowledgements
22. Volkmann, C. J.; Chateauneuf, G. M.; Pradhan, J.; Bauman, A. T.; Brown, R. E.;
Murthy, P. P. N. Tetrahedron Lett. 2002, 43, 4853e4856.
23. Lee, H. W.; Kishi, Y. J. Org. Chem. 1985, 50, 4402e4404.
24. Laumen, K.; Ghisalba, O. Biosci. Biotechnol. Biochem. 1994, 58, 2046e2049.
25. Becke, A. D. J. Chem. Phys. 1993, 98, 5648e5652.
26. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter 1988, 37, 785e789.
EPSRC and the School of Pharmacy and Pharmaceutical Sciences
are thanked for a studentship and PhDþ award (D.M.). N.V. is in-
debted to PEDECIBA-Química and ANII for a scholarship. We thank
Professor Christopher J. Barker (Karolinska Institutet) and Dr. Rick
Dunn (The University of Manchester).
27. Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270e283.
28. Frisch, M.J.; Trucks, G.W.; Schlegel, H.B.; Scuseria, G.E.; Robb, M.A.; Cheeseman,
J.R.; Montgomery, J.A.; Vreven, T.; Kudin, K.N.; Burant, J.C.; Millam, J.M.; Iyengar,
S.S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Pe-
tersson, G.A.; Nakatsuji, H.; Hada, M.; Ehara, K.; Toyota, K.; Fukuda, R.; Hase-
gawa, J.; Ishida, T.; Nakajima, Y.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.;
Knox, J.E.; Hratchian, H.P.; Cross, J.B.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R.E.; Yazyev, O.; Austin, A.J.; Cammi, R.; Pomelli, C.; Ochterski, J.;
Ayala, P.Y.; Morokuma, K.; Voth, G.A.; Salvador, P.; Dannenberg, J.J.; Zakrzekski,
V.G.; Dapprich, S.; Daniels, A.D.; Strain, M.C.; Farkas, O.; Malick, D.K.; Rabuck, A.
D.; Raghavachari, K.; Foresman, J.B.; Ortiz, J.V.; Cui, Q.; Baboul, A.G.; Clifford, S.;
Cioslowski, J.; Stefanov, B.B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.I.;
Martin, R.L.; Fox, D.J.; Keith, T.; Al-Laham, M.A.; Peng, C.Y.; Nanayakkara, A.;
Challacombe, M.; Gill, P.M.W.; Johnson, B.G.; Chen, W.; Wong, M.W.; Gonzalez,
C.; Pople, J.A. Gaussian 03, revision B.01, Gaussian Inc.: Pittsburgh, PA, 2003.
References and notes
1. Spiers, I. D.; Barker, C. J.; Chung, S.-K.; Chang, Y.-T.; Freeman, S.; Gardiner, J. M.;
Hirst, P. H.; Lambert, P. A.; Michell, R. H.; Poyner, D. R.; Schwalbe, C. H.; Smith,
A. W.; Solomons, K. R. H. Carbohydr. Res. 1996, 282, 81e99.
2. Spiers, I. D.; Freeman, S.; Poyner, D. R.; Schwalbe, C. H. Tetrahedron Lett. 1995,
36, 2125e2128.
3. Graf, E.; Empson, K. L.; Eaton, J. W. J. Biol. Chem. 1987, 262, 11647e11650.
4. Hawkins, P. T.; Poyner, D. R.; Jackson, T. R.; Letcher, A. J.; Lander, D. A.; Irvine, R.
F. Biochem. J. 1993, 294, 929e934.
5. Veiga, N.; Torres, J.; Mansell, D.; Freeman, S.; Dominguez, S.; Barker, C. J.; Diaz,
A.; Kremer, C. J. Biol. Inorg. Chem. 2009, 14, 51e59.