Stabilized phosphatidylinositol-5-phosphate analogues as ligands for the nuclear protein ING2: Chemistry, biology, and molecular modeling

Article abstract of DOI:10.1021/ja070195b

The interaction of PtdIns(5)P with the tumor suppressor protein ING2 has been implicated in the regulation of chromatin modification. To enhance the stability of PtdIns(5)P for studies of the biological role in vivo, two phosphatase-resistant moieties wer

Full text of DOI:10.1021/ja070195b

Published on Web 05/01/2007
Stabilized Phosphatidylinositol-5-Phosphate Analogues as
Ligands for the Nuclear Protein ING2: Chemistry, Biology,
and Molecular Modeling
Wei Huang,† Honglu Zhang,† Foteini Davrazou,‡ Tatiana G. Kutateladze,‡
Xiaobing Shi,§ Or Gozani,§ and Glenn D. Prestwich*,†
Contribution from the Department of Medicinal Chemistry, The UniVersity of Utah, 419 Wakara
Way, Suite 205, Salt Lake City, Utah 84108-1257, Department of Pharmacology, UniVersity of
Colorado Health Sciences Center, Aurora, Colorado 80045-0511, and Department of Biological
Sciences, Stanford UniVersity, Palo Alto, California 94305
Received January 10, 2007; E-mail: gprestwich@pharm.utah.edu
Abstract: The interaction of PtdIns(5)P with the tumor suppressor protein ING2 has been implicated in
the regulation of chromatin modification. To enhance the stability of PtdIns(5)P for studies of the biological
role in vivo, two phosphatase-resistant moieties were used to replace the labile 5-phosphate. The total
asymmetric synthesis of the 5-methylenephosphonate (MP) and 5-phosphothionate (PT) analogues of
PtdIns(5)P is described herein, and the resulting metabolically stabilized lipid analogues were evaluated in
three ways. First, liposomes containing either the dioleoyl MP or PT analogues bound to recombinant
ING2 similar to liposomes containing dipalmitoyl PtdIns(5)P, indicating that the replacement of the
hydrolyzable 5-phosphate group does not compromise the binding. Second, the dioleoyl MP and PT
PtdIns(5)P analogues were equivalent to dipalmitoyl PtdIns(5)P in augmenting cell death induced by a
DNA double-strand break in HT1080 cells. Finally, molecular modeling and docking of the MP or PT
analogues to the C-terminus PtdInsP-binding region of ING2 (consisting of a PHD finger and a polybasic
region) revealed a number of complementary surface and electrostatic contacts between the lipids and
ING2.
Introduction
The ING family of proteins is highly conserved from yeast
to humans,7 and different members are native subunits of histone
acetyl transferase (HAT) and histone deacetylase (HDAC)
complexes.8 Mammalian ING proteins are candidate tumor
suppressor proteins, which cooperate with p53 to induce cellular
growth arrest and apoptosis; for example, overexpression of
ING2 stimulates acetylation of p53.9,10 As a result, the ING
family is alleged to link chromatin regulation with p53 function
and tumor suppression. ING2 contains a PHD finger that
recognizes trimethylated at lysine 4 histone H3 tail.11,12 The
ING2 PHD finger is also required for the interaction with
PtdIns(5)P,6 and together with the 18-residue polybasic region
C-terminal to the PHD domain, is necessary and sufficient for
phosphoinositide binding by ING2.13 Thus, a model of ING2
with both of PHD domain and polybasic region is requir-
ed to understand the binding and physiological activities of
PtdIns(5)P.
Phosphoinositide (PtdInsPn) signaling involves a wide variety
proteins that exhibit lipid recognition, kinase, phosphatase, or
phospholipase activities.1,2 One relatively rare phosphoinositide,
PtdIns(5)P, has been implicated as a critical regulator of nuclear
signaling events in cell-cycle progression,3 for DNA damage-
dependent ING2 association with chromatin,4 and to inhibit
ATX1 gene expression programs.5 Still, relatively few details
are understood with respect to the signaling functions of PtdIns-
(5)P due to the paucity of well-characterized nuclear PtdIns-
(5)P binding domains. The interaction discovered between
PtdIns(5)P and ING2 (inhibitor of growth protein 2) was shown
to require both a C-terminus plant homeodomain (PHD) zinc
finger motif and a short stretch of polybasic residues6 and
implicated ING2 as a potential nuclear target for PtdIns(5)P.
† The University of Utah.
‡ University of Colorado Health Sciences Center.
§ Stanford University.
(1) Payrastre, B.; Missy, K.; Giuriato, S.; Bodin, S.; Plantavid, M.; Gratacap,
M. Cell. Signalling 2001, 13, 377-387.
(7) Brunecky, R.; Lee, S.; Rzepecki, P. W.; Overduin, M.; Prestwich, G. D.;
Kutateladze, A. G.; Kutateladze, T. G. Biochemistry 2005, 44, 16064-
16071.
(2) Prestwich, G. D. Chem. Biol. 2004, 11, 619-637.
(3) Clarke, J. H.; Letcher, A. J.; D’Santos, C. S.; Halstead, J. R.; Irvine, R. F.;
Divecha, N. Biochem. J. 2001, 357, 905-910.
(4) Jones, D. R.; Divecha, N. Curr. Opin. Genet. DeV. 2004, 14, 196-202.
(5) Alvarez-Venegas, R.; Sadder, M.; Hlavacka, A.; Baluska, F.; Xia, Y.; Lu,
G.; Firsov, A.; Sarath, G.; Moriyama, H.; Dubrovsky, J.; Avramova, Z.
Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 6049-6054.
(6) Gozani, O.; Karuman, P.; Jones, D. R.; Ivanov, D.; Cha, J.; Lugovskoy,
A. A.; Baird, C. L.; Zhu, H.; Field, S. J.; Lessnick, S. L.; Villasenor, J.;
Mehrotra, B.; Chen, J.; Rao, V. R.; Brugge, J. S.; Ferguson, C. G.; Payrastre,
B.; Myszka, D. G.; Cantley, L. C.; Wagner, G.; Divecha, N.; Prestwich,
G. D.; Yuan, J. Cell 2003, 114, 99-111.
(8) Feng, X.; Hara, Y.; Riabowol, K. Trends Cell Biol. 2002, 12, 532-538.
(9) Garkavtsev, I.; Grigorian, I. A.; Ossovskaya, V. S.; Chernov, M. V.;
Chumakov, P. M.; Gudkov, A. V. Nature 1998, 391, 295-298.
(10) Nagashima, M.; Shiseki, M.; Miura, K.; Hagiwara, K.; Linke, S. P.; Pedeux,
R.; Wang, X. W.; Yokota, J.; Riabowol, K.; Harris, C. C. Proc. Natl. Acad.
Sci. U.S.A. 2001, 98, 9671-9676.
(11) Shi, X.; Gozani, O. J. Cell. Biochem. 2005, 96, 1127-1136.
(12) Pena, P. V.; Davrazou, F.; Shi, X.; Walter, K. L.; Verkhusha, V. V.; Gozani,
O.; Zhao, R.; Kutateladze, T. G. Nature 2006, 442, 100-103.
(13) Kaadige, M. R.; Ayer, K. E. J. Biol. Chem. 2006, 281, 28831-28836.
9
6498
J. AM. CHEM. SOC. 2007, 129, 6498-6506
10.1021/ja070195b CCC: $37.00 © 2007 American Chemical Society

Phosphatidylinositol-5-Phosphate Analogues
A R T I C L E S
contained the following strategic steps: (a) selective protection
of positions of inositol,21 in particular using the TBDPS group
for the 1-position, benzoate group for 5-position, and MOM
group for 2,3,4,6-positions; (b) introduction of the MP or PT
functionality in protected form at the 5-position; (c) introduction
of the diacylglycerol phosphate at the 1-position using phos-
phoramidite chemistry.
Scheme 1 illustrates the preparation of the 5-MP-PtdIns(5)P
analogue. D-Camphor dimethyl acetal22 gave exclusively the
2,3-bornanediyl-myo-inositol,23 which was then regioselectively
silylated at 1-OH with TBDPSCl.24 After reaction with benzoyl
chloride and removal of bornanediyl group, the 5-benzoyl-1-
TBDPS intermediate was isolated from the mixture containing
4-benzoate and 4,5-bisbenzoate products. All the remaining
hydroxyl groups were protected as methoxymethyl (MOM)
ethers to give intermediate 6.21 After removal of benzoate group
with sodium methoxide, key intermediate 7 was obtained from
myo-inositol in six steps.
Dimethyl phosphonomethyltriflate, which is required for
installation of methylenephosphonate to the 5-position of 7, was
prepared as described.25,26 The alkylation of 7 with dimethyl
phosphonomethyltriflate was employed using NaH/THF in 86%
yield. Use of n-BuLi17,27 as the base in this reaction resulted in
more byproducts and a lower yield. Then, the TBDPS group of
8 was removed with tetrabutylammonium fluoride trihydrate
(TBAF‚3H2O) at room temperature for 4 h to give alcohol 9.
Separately, the two diacylglyceryl phosphoramidites 10a and
10b were prepared from 1,2-O-isopropylidene-sn-glycerol in five
steps as reported.17,28 Then, in the presence of 1H-tetrazole, 9
was treated with 10a (or 10b), followed by oxidation with
t-BuOOH to give the fully protected intermediate 11a (or 11b).
The cyanoethyl (CE) group was removed in the presence of
triethylamine (TEA) and bis(trimethylsilyl)trifluoroacetamide
(BSTFA),29 and the phosphate methyl ester and MOM ether
were removed using TMSBr and BSTFA.21 After complete
evaporation of organic solvent in vacuo, the crude product was
treated with Dowex ion-exchange resin17 to yield the 5-MP
analogues of PtdIns(5)P (1 and 2, Scheme 1).
Figure 1. 5-Methylenephosphonate and 5-phosphothionate analogues of
PtdIns(5)P.
Although the crystal structure of the ING2 PHD domain has
been determined12 and several homologous structures of ING
proteins are available through the PDB bank, at present no 3D
structure of ING2 that includes the flexible C-terminal polybasic
region is available. Thus, further characterization of the structure
and function of the ING2-C-terminus will benefit from the use
of modified PtdIns(5)P ligands.
In vivo studies of PtdIns(5)P are problematic because of
its rapid metabolism. Specifically, three important meta-
bolic pathways are presently known to utilize PtdIns(5)P
as a substrate. Two pathways involve kinases; PtdIns(4,5)P2
is produced by the action of PtdIns(5)P 4-kinase, and
PtdIns(3,5)P2 can be produced by PtdIns 3-kinase.2,14,15 The third
route involves a ubiquitous cellular phosphatase activity that
metabolizes PtdIns(5)P to PtdIns.
We have recently pursued a program to prepare metabolically
stabilized, e.g., phopholipase- and phosphatase-resistant ana-
logues of the phosphoinositides PtdIns(3)P,16,17 PtdIns(4,5)P2,18
and PtdIns(3,4,5)P3.19 In these examples, as with analogues the
phosphatase-labile lysophosphatidic acid (LPA),20 methyl phos-
phonates, fluoromethyl phosphosphonates, phosphorothionates,
and methylenephosphonates were used to replace the natural
phosphomonoester or phosphodiester moieties. Herein we
describe the total asymmetric synthesis of the 5-methylene-
phosphonate (MP) and 5-phosphothionate (PT) analogues of
PtdIns(5)P (Figure 1) as water-soluble ligands with short
dibutanoyl chains or as lipid-soluble amphiphiles with dioleoyl
chains. We show that PtdIns(5)P analogues interact with the
C-terminal region of ING2 using liposome binding assays.
Further, cellular studies demonstrate that the MP and PT
analogues are as bioactive as PtdIns(5)P in augmenting cell death
induced by DNA damage. Finally, we construct 3D models of
the ING2 polybasic region using homology modeling algorithms
and computationally dock the PtdIns(5)P analogues to these
models.
To obtain the 5-PT analogues of PtdIns(5)P, we first
attempted to use the protected 5-phosphorothioate intermediate
13, prepared by phosphitylation of alcohol 7 with bis(2-
cyanoethyl)diisopropylphosphorodiamidite followed by oxida-
tion with sulfur in CS2/pyridine. The removal of TBDPS group
of obtained product 12 was achieved by using HF-pyridine
complex (Scheme 2) instead of TBAF because the cyanoethyl
group was labile to cleavage by TBAF.17 However, the long
reaction time (3 weeks) proved to be a significant disadvantage
of this method. We thus abandoned this intermediate and solved
the problem by replacing the TBDPS group with the triethylsilyl
(TES) group19 as illustrated in Scheme 3.
Results and Discussion
Chemical Synthesis of Stabilized Phosphatidylinositol-5-
Phosphate Analogues. The syntheses of the metabolically
stabilized PtdIns(5)P analogues were based on modifications
of routes to the corresponding PtdIns(3)P analogues.16,17 Each
(21) Kubiak, R. J.; Bruzik, K. S. J. Org. Chem. 2003, 68, 960-968.
(22) Lindberg, J.; Ohberg, L.; Garegg, P. J.; Konradsson, P. Tetrahedron 2002,
58, 1387-1398.
(14) Whiteford, C. C.; Brearley, C. A.; Ulug, E. T. Biochem. J. 1997, 323 (Part
3), 597-601.
(15) Peng, J.; Prestwich, G. D. Tetrahedron Lett. 1998, 39, 3965-3968.
(16) Xu, Y.; Lee, S. A.; Kutateladze, T. G.; Sbrissa, D.; Shisheva, A.; Prestwich,
G. D. J. Am. Chem. Soc. 2006, 128, 885-897.
(23) Pietrusiewicz, K. M.; Salamonczyk, G. M.; Bruzik, K. S. Tetrahedron 1992,
48, 5523-5542.
(24) Bruzik, K. S.; Tsai, M. J. Am. Chem. Soc. 1992, 114, 6361-6374.
(25) Jeanmaire, T.; Hervaud, Y.; Boutevin, B. Phosphorus, Sulfur Silicon Relat.
Elem. 2002, 177, 1137-1145.
(17) Gajewiak, J.; Xu, Y.; Lee, S. A.; Kutateladze, T. G.; Prestwich, G. D. Org.
Lett. 2006, 8, 2811-2813.
(18) Zhang, H.; Xu, Y.; Zhang, Z.; Liman, E. R.; Prestwich, G. D. J. Am. Chem.
Soc. 2006, 128, 5642-5643.
(26) Xu, Y.; Flavin, M. T.; Xu, Z. Q. J. Org. Chem. 1996, 61, 7697-7701.
(27) Kulagowski, J. J.; Baker, R.; Fletcher, S. R. J. Chem. Soc., Chem. Commun.
1991, 1649-1650.
(19) Zhang, H.; Markadieu, N.; Beawens, R.; Erneux, C.; Prestwich, G. D. J.
Am. Chem. Soc. 2006, 128, 16464-16465.
(28) Chen, J.; Profit, A. A.; Prestwich, G. D. J. Org. Chem. 1996, 61, 6305-
6312.
(20) Prestwich, G. D.; Xu, Y.; Qian, L.; Gajewiak, J.; Jiang, G. Biochem. Soc.
Trans. 2005, 33, 1357-1361.
(29) Heeb, N. V.; Nambiar, K. P. Tetrahedron Lett. 1993, 34, 6193-6196.
9
J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6499

A R T I C L E S
Huang et al.
Scheme 1. Synthesis of 5-Methylenephosphonate Analogues of PtdIns(5)Pa
a Reagents and conditions: (a) NaOMe, MeOH, room temperature, 24 h; (b) dimethyl phosphonomethyltriflate, NaH, THF, 0 °C, then room temperature,
12 h; (c) TBAF‚3H2O, THF, room temperature, 4 h; (d) 1H-tetrazole, CH2Cl2, Ar, room temperature, 24 h, then t-BuOOH, 1 h; (e) TEA, BSTFA, MeCN,
Ar, room temperature, 24 h; (f) TMSBr, BSTFA, Ar, room temperature, 10 h; (g) MeOH, room temperature, 1 h.
Scheme 2. Synthesis of Intermediate 13a
analogues could substitute for PtdIns(5)P in biochemical and
biological assays.
HT1080 Cell Death Assay. ING2 overexpression triggers a
number of molecular events including p53 acetylation that can
culminate in the induction of apoptosis;10 the lipid binding
activity of ING2 is critical for these activities.6 Furthermore,
endogenous generation of nuclear PtdIns(5)P is involved in
cellular responses to genotoxic stress, through regulation of
ING2 and likely a number of additional factors.30 To test
whether the 5-PT and 5-MP analogues of PtdIns(5)P would
retain the same function in this pathway, we measured the
HT1080 cell death response to DNA damage following
treatment with dipalmitoyl PtdIns(5)P and/or with
5-MP-PtdIns(5)P (2) and 5-PT PtdIns(5)P (4) (Figure 3).
PtdIns(5)P shows a statistically significant, albeit small, en-
hancement in the level of cell death induced by neocarzinostatin
(NCS) treatment alone in HT1080 cells. Both the 5-MP-PtdIns-
(5)P analogue 2 and 5-PT PtdIns(5)P analogue 4 show an
equivalent effect in promoting cell death. These results dem-
onstrate a biological effect of the metabolically stabilized
PtdIns(5)P analogues in this nuclear stress response pathway.
Molecular Modeling. Combining known biochemical, struc-
tural, and functional data on the ING2 PHD domain and
polybasic region activities, we sought to model the C-terminal
tail of ING2 to gain insights into its interactions with
PtdIns(5)P analogues. However, little is known concerning the
3D structure of the last 18 residues, which is likely to be a
random coil. To further characterize the structure of this region,
we built 20 homology models of the ING2 C-terminus (residues
202-281) by employing MODELLER,31-33 a set of algorithms
for homology modeling. Each computed model contained the
conserved PHD domain and different polybasic regions based
on reported crystal structures. These models included multiple
possible stereochemical geometries for the polybasic tail as
anticipated for simulation of a random coil. Next, we employed
a Reagents and conditions: (a) bis(2-cyanoethyl)diisopropylphospho-
rodiamidite, 1H-tetrazole, MeCN, Ar, room temperature, 2 h, then sulfur,
CS2/Py (1:1), room temperature, 3 h; (b) HF-Py, THF, room temperature,
3 weeks.
Scheme 3 shows the preparation of the 5-PT analogues
PtdIns(5)P. Thus, 5-benzoyl-1-TBDPS-2,3,4,6-tetrakis(MOM)-
inositol (6) was treated with TBAF to remove the 1-TBDPS
ether; this was then replaced with TES ether by using TESCl.
Although two additional steps were required, this change of
protecting groups improves overall yield and reduces side
products as found for a 3-PT analogue of PtdIns(3,4,5)P3.19 After
debenzoylation of 14 with diisobutyl aluminum hydride
(DIBAL-H), the phosphorothioate group was introduced in the
5-position as shown in Scheme 2. Deprotection of the TES group
of 16 with NH4F gave intermediate 13 efficiently and in high
yield. Then, the diacylglycerophosphodiester linkage to the
1-position of 13 was created using either 10a (or 10b) and 1H-
tetrazole followed by oxidation. After deprotection of cyanoethyl
group and MOM group, the final products of 5-PT analogues
of PtdIns(5)P (3 and 4) were obtained.
Liposome Binding Assay. To determine the effect of the
PtdIns(5)P head group modifications on the inositol ring
recognition by ING2, PtdIns(5)P analogues were evaluated using
liposome binding assays (Figure 2). The C-terminal region of
ING2 (amino acids 200-280 of ING2) has been shown to
interact with PtdIns(5)P in vitro and in vivo.6 In the absence of
any phosphoinositide in the liposome composition or in the
presence of unphosphorylated PtdIns, the GST-fused C-terminal
region of ING2 (GST-ING2) remains primarily soluble. With
the parent ligand dipalmitoyl PtdIns(5)P embedded in the
liposome, GST-ING2 was strongly associated with the lipo-
some fraction. Both modified lipids, 5-MP-PtdIns(5)P (2) and
5-PT PtdIns(5)P (4), bound GST-ING2 with nearly the same
efficiency as the unmodified lipid, suggesting that these synthetic
(30) Jones, D.; Bultsma, Y.; Kuene, W.-J.; Halstead, J.; Elouarrat, D.; Moham-
med, S.; Heck, A.; D’Santos, C.; Divecha, N. Mol. Cell 2006, 23, 685-
695.
(31) Sali, A.; Blundell, T. L. J. Mol. Biol. 1993, 234, 779-815.
(32) Marti-Renom, M. A.; Stuart, A.; Fiser, A.; Sanchez, R.; Melo, F.; Sali, A.
Annu. ReV. Biophys. Biomol. Struct. 2000, 29, 291-325.
(33) Fiser, A.; Do, R. K.; Sali, A. Protein Sci. 2000, 9, 1753-1773.
9
6500 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Phosphatidylinositol-5-Phosphate Analogues
A R T I C L E S
Scheme 3. Synthesis of 5-Phosphothionate Analogues of PtdIns(5)Pa
a Reagents and conditions: (a) TBAF‚3H2O, DMF, room temperature, 6 h; (b) triethylsilyl chloride, imidazole, CH2Cl2, room temperature, 15 h; (c)
DIBAL-H, CH2Cl2, -78 °C, 2 h; (d) bis(2-cyanoethyl)diisopropylphosphorodiamidite, 1H-tetrazole, MeCN, Ar, room temperature, 2 h, then sulfur, CS2/Py
(1:1), room temperature, 6 h; (e) NH4F, MeOH, room temperature, 20 h; (f) 10a (or 10b), 1H-tetrazole, CH2Cl2, Ar, room temperature, 24 h, then t-BuOOH,
1 h; (g) TEA, BSTFA, MeCN, Ar, room temperature, 24 h; (h) TMSBr, BSTFA, Ar, room temperature, 2 h; (i) MeOH, room temperature, 1 h.
sequence identities by SWISS-MODEL server (Figure 4A).39-41
They are solution structures of ING family proteins in the PDB
bank (http://www.rcsb.org/);42 1wes, 1weu, and 1wen are PHD
domains of ING1-like protein; 2q6g is PHD domain of ING2;
1x4i is C-terminal of ING3 which contains 20 models extending
Figure 2. PtdIns(5)P analogues are efficiently recognized by ING2. The
SDS-PAGE gels show the partitioning of the GST-fused C-terminal tail
of ING2 between the supernatant fraction (S) and the liposome pellet (P).
beyond the PHD domain, and these models provide the diverse
templates for the C-terminal tail. After sequence alignment by
MODELLER, the automodel mode of MODELLER was used,
based on each 1x4i model and the other templates, thus affording
20 primary ING2 structures. Then, after side-chain modeling
and energy minimization for each structure an ensemble of the
20 homology models of the ING2 C-terminus was obtained.
As shown in Figure 4, parts B and C, these models contain a
highly conserved PHD domain which ends with a short helix
(normally from Pro257 to Lys265), followed by a highly
divergent set of orientations for the remainder of the polybasic
region.
To better mimic native membranes, liposomes were prepared from
phospholipids typically found in bilayers including PtdCho, PtdEtn, PtdSer
in a ratio of 65:26:9 or PtdCho, PtdEtn, PtdSer and the corresponding
phosphoinositide in a ratio of 54:22:9:15.
The polybasic region contains three arginine and five lysine
residues that can form a binding pocket for the phosphoinositide.
First, we studied the 3D structure of each model and observed
the key features of each one. The helix C-terminal from the
PHD domain (Pro257 to Lys265) showed good identity for the
majority of models, and the side chain of Lys265 was close to
the region of Tyr255 and Lys253 within the PHD domain
(Figure 5, parts A and B). However, the polybasic tails were
widely divergent in the absence of a phosphoinositide ligand.
For example, in model 1 (Figure 5A), the end residues (Arg281,
Arg279, Arg278) move to 25 Å away from the PHD domain
(Arg278 to Lys253), such that the Arg and Lys residues in the
polybasic tail could not form an effective pocket. In contrast,
in model 2 (Figure 5B), the tail loops back toward the PHD
domain (Tyr255, Lys253) with proper torsion angles (9 Å from
Arg278 to Lys253), forming a more reasonable binding pocket.
Figure 3. PtdIns(5)P and analogues augment cell death induced by
neocarzinostatin (NCS) treatment. Cell death was measured in
HT1080 cells treated with 10 µM of the indicated lipid or no lipid control,
( 45 nM NCS. The error bars represent the mean ( SEM for four
experiments.
AutoDock 3.0.534-36 to dock the ligands to ING2 models
successively and performed LIGPLOT 4.4.237 into study the
preferred binding conformation and association within the
ING2-PtdIns(5)P complex.
The sequence of human ING238 (202-281) was provided as
input data, and five template structures were generated with good
(34) Goodsell, D. S.; Olson, A. J. Proteins 1990, 8, 195-202.
(35) Morris, G. M.; Goodsell, D. S.; Huey, R.; Olson, A. J. J. Comput.-Aided
Mol. Des. 1996, 10, 293-304.
(36) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart W. E.;
Belew, R. K.; Olson, A. J. J. Comput. Chem. 1998, 19, 1639-1662.
(37) Wallace, A. C.; Laskowski, R. A.; Thornton, J. M. Protein Eng. 1995, 8,
127-134.
(39) Schwede, T.; Kopp, J.; Guex, N.; Peitsch, M. C. Nucleic Acids Res. 2003,
31, 3381-3385.
(40) Kopp, J.; Schwede, T. Nucleic Acids Res. 2004, 32, D230-D234.
(41) Guex, N.; Peitsch, M. C. Electrophoresis 1997, 18, 2714-2723.
(42) Westbrook, J.; Feng, Z.; Chen, L.; Yang, H.; Berman, H. M. Nucleic Acids
Res. 2003, 31, 489-491.
(38) Wagner, M. J.; Gogela-Spehar, M.; Skirrow, R. C.; Johnston, R. N.;
Riabowol, K.; Helbing, C. C. J. Biol. Chem. 2001, 276, 47013-47020.
9
J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6501

A R T I C L E S
Huang et al.
Figure 4. (A) Sequence alignment of ING2 (202-281) with related family members from PDB. The alignment was performed under the module of
ALIGNMENT in MODELLER. (B) Structural alignment of the ING2 model (one of 20 models, red), 2g6q (white), 1x4i (blue), 1weu (purple), 1wes (cyan),
and 1wen (yellow). PHD domains are the conserved fractions in the center of the structures, end with a short helix, and the following tails on the left side
are polybasic regions. (C) Twenty ING2 models built by MODELLER. The multiorientation coils on the left side are polybasic C-termini.
Figure 5. Structural analysis of two representative models from 20 models generated by homology modeling: (A) model 1; (B) model 2.
Indeed, Arg279, Arg278, and Lys276 form a proper site with
Lys265 and the PHD domain contributing to the pocket. In this
case, Lys265 is located in the site center and serves as a “bridge”
to connect the PHD domain and polybasic tail. Among the 20
models, six contained a pocket similar to that of model 2. The
other 14 models had extended tails (as in model 1) but with
widely divergent orientations.
Furthermore, we docked diC4-PtdIns(5)P to each of the 20
models using AutoDock3.0.5. The models containing a reason-
able binding pocket that included the polybasic tail, helix, and
PHD domains provided better binding scores than other models.
Supporting Information Figure S1A-S1D presents the binding
modes of diC4-PtdIns(5)P with models 2, 7, 9, and 10 which
give the highest scores and most reasonable conformations.
These models all gave similar binding modes with
diC4-PtdIns(5)P. The inositol ring was located in the area of
Arg279, Arg278, Asp277, and Lys265, and the phosphate and
hydroxyl groups formed good hydrogen bond networks with
these residues. The diacyl groups of lipid approached Tyr255
and Lys253 of the PHD domain.
Finally, after overall examination on binding energy and
structure rationality of these four models, we selected model 2
as the ING2 C-terminal model to compare the binding features
of diC18:1-PtdIns(5)P with the corresponding metabolically
stabilized analogues. The binding conformations of diC18:1
-
PtdIns(5)P, 5-MP-diC18:1-PtdIns(5)P (2), and 5-PT-diC18:1
-
PtdIns(5)P (4) are shown in Figure 6. These three compounds
have similar conformations. The inositol ring is located in the
pocket, and one of acyl groups is oriented toward the PHD
domain (Tyr255, Lys253) and the other orients to another part
of polybasic tail (Arg260, Lys273). The 5-phosphate of the
inositol head group is located proximal to Arg278, while the
1-phosphodiester linkage is proximal to Arg279 (Figure 6).
These results indicated a good steric simulation of chemical
geometry. We employed LIGPLOT4.4.2 to test the hydrogen
bonds and hydrophobic contacts between ligands and ING2.
Supporting Information Figures S2-S4 show that the hydrogen
bonds for the MP and PT analogues are almost identical with
those for diC18:1-PtdIns(5)P. However, there are more hydro-
phobic contacts observed for the longer acyl chains, thus
9
6502 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Phosphatidylinositol-5-Phosphate Analogues
A R T I C L E S
Figure 6. Docking conformations of diC18:1-PtdIns(5)P (red), diC18:1-MP-PtdIns(5)P (2, green), and diC18:1-PT-PtdIns(5)P (4, cyan) in the surface of the
ING2 C-terminal model.
shifts are reported in ppm with TMS as internal standard () 0.00);
31P, 85% H3PO4 () 0.00); 19F, CFCl3 () 0.00). Low- and high-
resolution mass spectra were obtained on HP5971A MSD and Finnigan
MAT95 double-focusing mass spectrometer (MS) instruments, respec-
tively.
enhancing interactions with the PHD domain. This can rational-
ize our observations (unpublished results) that the dioleoyl
analogues showed higher affinity than the dibutanoyl analogues
in each of the binding assay formats tested.
Finally, in agreement with the modeling studies, Lys 251,
Lys 253, W254, and G261 residues of the PHD finger and the
polybasic region of ING2 were perturbed in NMR spectra of
the protein during gradual addition of inositol 1,5-bisphosphate
(Ins(1,5)P2), the head group of PtdIns(5)P (Supporting Informa-
tion Figures S5 and S6). Supporting Information Figure S5
shows normalized backbone amide chemical shift changes in
0.2 mM ING2 induced by 24 mM Ins(1,5)P2, while Supporting
Information Figure S6 plots chemical shift changes versus Ins-
(1,5)P2 concentration from 2 to 24 mM. Thus, both of the
predicted basic regions of ING2 are involved in the coordination
of the inositol head group.
In conclusion, the phosphatase-resistant PtdIns(5)P analogues
should provide useful tools for further in vivo research on
several physiologically important pathways, including those
governed by ING2. The molecular modeling presented herein
provides the first model of the full ING2 C-terminus and reveals
the possible binding mode of PtdIns(5)P (and its analogues)
with ING2. These interactions support the biological results,
provide new 3D structural information, and offer testable
structure-activity hypotheses for further studies on the physiol-
ogy of PtdIns(5)P binding proteins.
1D-1-O-(tert-Butyldiphenylsilyl)-2,3,4,6-O-tetrakis(methoxymeth-
ylene)-myo-inositol (7). To a solution of 6 (440 mg, 0.63 mmol) in
methanol (5 mL) was added sodium methoxide (4 mL, 25% solution
in anhydrous methanol). The mixture was stirred at room temperature
for 24 h, then 10 mL of ethyl acetate was added, and the mixture was
stirred for 20 min. The organic layer was washed with water and dried
over anhydrous sodium sulfate. After filtration and concentration, the
subsequent chromatography on silica gel (acetone/hexanes, 1:3) afforded
1
7 (330 mg, 88%) as a pale yellow oil. H NMR (400 MHz, CDCl3)
7.71 (m, 4H), 7.41 (m, 6H), 4.78 (m, 2H), 4.74 (d, J ) 6.0 Hz, 1H),
4.67 (d, J ) 6.0 Hz, 1H), 4.62 (d, J ) 6.0 Hz, 1H), 4.52 (d, J ) 6.0
Hz, 1H), 4.42 (d, J ) 6.0 Hz, 1H), 4.39 (d, J ) 6.0 Hz, 1H), 3.77 (m,
3H), 3.57 (m, 1H), 3.42 (s, 3H), 3.35 (s, 3H), 3.31 (s, 3H), 3.26 (m,
2H), 3.23 (s, 3H), 1.07 (m, 9H). MALDI-HRMS [M + Na]+ calcd for
C30H46O10SiNa, 617.2752; found, 617.2749.
1D-1-O-(tert-Butyldiphenylsilyl)-5-(dimethyl methylenephospho-
nate)-2,3,4,6-O-tetrakis(methoxymethylene)-myo-inositol (8). To a
solution of 7 (115 mg, 0.19 mmol) in THF (10 mL) was added NaH
(100 mg) at 0 °C, and the mixture was stirred for 20 min. Then,
dimethylphosphonomethyltriflate was added, and the mixture was
warmed to room temperature and stirred for 12 h. Ethyl acetate was
added, and the organic layer was washed with water, dried over
anhydrous sodium sulfate, concentrated, and chromatographed on silica
gel (acetone/hexanes, 1:1) giving product 8 (120 mg, 86%) as a colorless
oil. 1H NMR (400 MHz, CDCl3) 7.70 (m, 4H), 7.41 (m, 6H), 5.02 (d,
J ) 6.4 Hz, 1H), 4.81 (m, 3H), 4.55 (d, J ) 6.4 Hz, 1H), 4.46 (d, J )
6.4 Hz, 1H), 4.39 (d, J ) 6.4 Hz, 1H), 4.29 (m, 1H), 4.18-4.12 (m,
2H), 4.01 (t, J ) 9.6 Hz, 1H), 3.85-3.74 (m, 8H), 3.49 (s, 3H), 3.38
(s, 3H), 3.24 (s, 3H), 3.20 (s, 1H), 3.12 (s, 3H), 3.08 (m, 2H), 1.09 (s,
9H); 13C NMR (101 MHz, CDCl3) 136.3, 136.1, 134.3, 133.0, 130.3,
130.0, 128.1, 127.9, 99.2, 98.6, 97.4, 95.2, 85.6, 85.5, 79.0, 77.8, 75.6,
Experimental Methods
General Chemical Reagents and Methods. Chemicals were
purchased from Aldrich and Acros Chemical Corp. and used without
prior purification. Solvents were reagent-grade and distilled before
use: CH2Cl2 was distilled from CaH2, and THF was distilled from
sodium wire. TLC used precoated silica gel aluminum sheets (EM
Science silica gel 60F254). Flash chromatography (FC) employed
Whatman 230∼400 mesh ASTM silica gel. NMR spectra were
recorded on a Varian INOVA 400 at 400 MHz (1H), 101 MHz
(13C), 162 MHz (31P), and 376 MHz (19F) at 25 °C. Chemical
75.8, 74.1, 67.4, 65.8, 57.0, 56.6, 55.8, 55.6, 53.1, 53.0, 27.4, 19.3; 31
P
NMR (162 MHz, CDCl3) 24.7. MALDI-HRMS [M + Na]+ calcd for
C33H53O13SiPNa, 739.2885; found, 739.2888.
9
J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6503

A R T I C L E S
Huang et al.
1D-5-(dimethyl methylenephosphonate)-2,3,4,6-O-tetrakis(meth-
oxymethylene)-myo-inositol (9). To a solution of 8 (125 mg, 0.17
mmol) in THF (5 mL) was added TBAF‚3H2O (150 mg). Then, the
mixture was stirred at room temperature for 4 h. The solvent was
evaporated off, and the residue was chromatographed on silica gel
(acetone/hexanes, 2:1) giving product 9 (80 mg, 96%) as a colorless
(101 MHz, D2O) 173.3, 172.9, 85.4, 85.3, 81.2, 77.9, 75.0, 74.6, 71.9,
71.6, 70.2, 68.2, 67.7, 66.3, 62.4, 61.5, 35.7, 35.6, 18.1, 12.9, 12.8; 31
P
NMR (162 MHz, D2O) 19.3, 0.73 (m). MALDI-HRMS [M]+ calcd for
C18H34O16P2, 568.1322; found, 568.1299.
1D-O-(1,2-Di-O-oleoyl-sn-(2S)-glycerol-3-phospho)-5-(methylene-
phosphonate)-myo-inositol (2) was obtained from 11b as described
for 1, yield 70%. 1H NMR (400 MHz, CD3OD-CDCl3) 5.31 (m, 5H),
4.19-4.07 (m, 5H), 4.00 (m, 2H), 3.93 (t, J ) 9.2 Hz, 1H), 3.74 (m,
2H), 3.39 (m, 1H), 3.11 (t, J ) 9.2 Hz, 1H), 2.23 (m, 4H), 2.01 (m,
8H), 1.56 (m, 4H), 1.13 (m, 40H), 0.87 (t, J ) 6.8 Hz, 6H); 13C NMR
(101 MHz, CD3OD-CDCl3) 173.2, 172.7, 129.7, 129.6, 85.9, 78.1,
77.7, 75.5, 74.9, 72.5, 72.3, 71.5, 71.3, 69.5, 68.6, 66.4, 62.5, 61.9,
33.8, 33.7, 31.8, 29.68, 29.60, 29.45, 29.30, 29.23, 29.19, 29.12, 29.09,
29.02, 27.0, 24.8, 22.6, 13.5, 13.4; 31P NMR (162 MHz, CD3OD-
CDCl3) 21.2, 0.68 (m). MALDI-HRMS [M - H]- calcd for C46H85O16P2,
955.5318; found, 955.5345.
1D-1-O-(tert-Butyldiphenylsilyl)-5-(bis(cyanoethyl)phosphothio-
nate)-2,3,4,6-O-tetrakis(methoxymethylene)-myo-inositol (12). To a
solution of 7 (100 mg, 0.17 mmol) in acetonitrile (5 mL) was added
1H-tetrazole (3% in MeCN, 0.09 mL, 0.39 mmol) and bis(2-cyanoethyl)-
diisopropylphosphorodiamidite (50 mg, 0.185 mmol) at room temper-
ature under Ar, and the mixture was stirred for 2 h. Then, sulfur (180
mg) and CS2/Py (1:1, 1.8 mL) were added, and the mixture was stirred
at room temperature for 3 h. After filtration, the filtrate was concentrated
and chromatographed on silica gel (acetone/hexanes, 1:3) giving product
12 (100 mg, 75%) as a colorless oil. 1H NMR (400 MHz, CDCl3) 7.70
(m, 4H), 7.41 (m, 6H), 4.94 (d, J ) 6.0 Hz, 1H), 4.88 (d, J ) 6.0 Hz,
1H), 4.74 (m, 2H), 4.57 (d, J ) 6.4 Hz, 1H), 4.48 (d, J ) 6.0 Hz, 1H),
4.39-4.29 (m, 6H), 4.18 (d, J ) 7.2 Hz, 1H), 4.07 (m, 1H), 3.82 (m,
2H), 3.49 (s, 3H), 3.37 (s, 3H), 3.25 (s, 3H), 3.22 (m, 1H), 3.15 (m,
1H), 3.10 (s, 3H), 2.76 (m, 4H), 1.09 (s, 9H); 13C NMR (101 MHz,
CDCl3) 136.2, 136.0, 134.1, 132.7, 130.4, 130.2, 128.2, 128.0, 117.0,
99.3, 98.6, 97.5, 95.4, 80.7, 80.6, 77.3, 76.6, 75.4, 75.2, 74.1, 62.84,
62.83, 62.78, 62.77, 57.3, 57.0, 55.9, 55.7, 27.5, 19.5, 19.4, 19.3; 31P
NMR (162 MHz, CDCl3) 69.1. MALDI-HRMS [M + Na]+ calcd for
C36H53N2O12SSiPNa, 819.2718; found, 819.2708.
1
oil. H NMR (400 MHz, CDCl3) 4.83-4.69 (m, 8H), 4.16 (d, J )
10.0 Hz, 2H), 4.05 (m, 1H), 3.91 (t, J ) 9.2 Hz, 1H), 3.79 (d, J )
10.4 Hz, 6H), 3.70 (t, J ) 9.2 Hz, 1H), 3.38-3.27 (m, 14H), 3.18 (t,
J ) 9.2 Hz, 1H); 13C NMR (101 MHz, CDCl3) 98.8, 98.6, 98.2, 96.4,
85.7, 85.5, 83.3, 77.8, 77.6, 76.3, 71.0, 67.7, 66.0, 56.6, 56.3, 55.9,
53.1 (m); 31P NMR (162 MHz, CDCl3) 24.4. MALDI-HRMS [M +
Na]+ calcd for C17H35O13PNa, 501.1708; found, 501.1712.
1D-O-(1,2-Di-O-butanoyl-sn-(2S)-glycerol-3-O-(cyanoethyl)phos-
pho)-5-(dimethyl methylenephosphonate)-2,3,4,6-O-tetrakis(methoxy-
methylene)-myo-inositol (11a). To a solution of 9 (40 mg, 0.084 mmol)
in dichloromethane (2 mL) was added 1H-tetrazole (20 mg) and
N,N-diisopropyl-O-(cyanoethyl)-O-(dibutanoyl-sn-(2S)-glycerol)-phos-
phonamidite (10a, 100 mg) at room temperature under Ar. After stirring
the mixture at room temperature for 24 h, t-BuOOH (0.1 mL) was
added for oxidation, and the mixture was stirred for 1 h. The solution
was diluted by CH2Cl2 (20 mL) and washed with 10% sodium bisulfite,
then dried over anhydrous sodium sulfate. After filtration and concen-
tration, the residue was chromatographed on silica gel (acetone/hexanes,
1
2:1 then 3:1) giving product 11a (45 mg, 65%) as a colorless oil. H
NMR (400 MHz, CDCl3) 5.17 (m, 1H), 4.74-4.58 (m, 8H), 4.26-
4.02 (m, 10H), 3.89-3.80 (m, 2H), 3.70 (d, J ) 10.8 Hz, 6H), 3.39-
3.27 (m, 13H), 3.07 (t, J ) 9.2 Hz, 1H), 2.67 (m, 2H), 2.20 (m, 4H),
1.52 (m, 4 H), 0.83 (m, 6H); 13C NMR (101 MHz, CDCl3) 173.2, 172.8,
116.7, 98.8, 98.7, 98.6, 97.7, 96.2, 85.5, 85.4, 77.5, 76.7, 75.6, 75.5,
74.7, 69.4, 69.3, 67.6, 66.29, 66.28, 66.25, 66.24, 66.0, 62.56, 62.55,
62.52, 62.51, 61.6, 56.8, 56.7, 56.6, 56.0, 55.9, 53.1, 53.0, 36.1, 35.9,
19.8, 19.7, 18.4, 13.7, 13.6; 31P NMR (162 MHz, CDCl3) 23.9, -1.09,
-1.15. MALDI-HRMS [M + Na]+ calcd for C31H57O20NP2Na,
848.2841; found, 848.2849.
1D-O-(1,2-Di-O-oleoyl-sn-(2S)-glycerol-3-O-(cyanoethyl)phospho)-
5-(dimethyl methylenephosphonate)-2,3,4,6-O-tetrakis(methoxymeth-
ylene)-myo-inositol (11b) was obtained from 9 and 10b as described
1D-5-(Bis(cyanoethyl)phosphothionate)-2,3,4,6-O-tetrakis-
(methoxymethylene)-myo-inositol (13). Method A. A solution of 12
(100 mg, 0.125 mmol) in THF (5 mL) was added to hydrogen fluoride-
pyridine complex (70%, 0.6 mL) at room temperature in a Teflon
container. After stirring for 3 weeks, the solution was diluted with ethyl
acetate and washed with 10% Na2CO3. The organic layer was collected
and dried over anhydrous sodium sulfate. After filtration, the filtrate
was concentrated and chromatographed on silica gel (acetone/hexanes,
1:2) giving product 13 (50 mg, 71%) as a colorless oil.
1
for 11a, yield 45%. H NMR (400 MHz, CDCl3) 5.32 (m, 4H), 5.25
(m, 1H), 4.84-4.67 (m, 8H), 4.35-4.08 (m, 10H), 3.98-3.90 (m, 2H),
3.70 (dd, J ) 2.8, 10.8 Hz, 6H), 3.49 (d, J ) 10.0 Hz, 1H), 3.44 (s,
3H), 3.43 (s, 3H), 3.39 (s, 3H), 3.37 (s, 3H), 3.17 (dt, J ) 2.8, 9.2 Hz,
1H), 2.76 (m, 2H), 2.31 (m, 4H), 1.98 (m, 8H), 1.58 (m, 4H), 1.28 (m,
40H), 0.85 (m, 6H); 13C NMR (101 MHz, CDCl3) 173.4, 173.0, 130.2,
129.9, 116.6, 98.9, 98.8, 98.7, 97.7, 96.2, 96.1, 85.5, 85.4, 77.2, 76.7,
75.6, 75.5, 74.7, 69.4, 69.3, 67.8, 66.3, 66.2, 66.1, 62.5, 62.4, 61.7,
56.9, 56.8, 56.7 56.1, 55.9, 55.8, 53.1, 53.0, 34.2, 34.1, 31.7, 29.97,
29.94, 29.73, 29.53, 29.45, 29.42, 29.35, 29.32, 29.29, 27.4, 25.0, 22.9,
19.8, 19.7, 14.3; 31P NMR (162 MHz, CDCl3) 24.0, -0.97, -0.99.
MALDI-HRMS [M + Na]+ calcd for C59H109O20NP2Na, 1236.6910;
found, 1236.6934.
Method B. To a solution of 16 (20 mg, 0.030 mmol) in methanol
(0.5 mL) was added ammonium fluoride (20 mg), and the mixture was
stirred at room temperature for 20 h. After removal of the solvent, the
residue was chromatographed on silica gel (acetone/hexanes, 1:2) to
1
afford product 13 (15 mg, 90%). H NMR (400 MHz, CDCl3) 4.82-
4.68 (m, 8H), 4.22-4.12 (m, 5H), 4.05 (s, 1H), 3.94 (d, J ) 7.2 Hz,
1H), 3.69 (d, J ) 7.2 Hz, 1H), 3.49-3.34 (m, 14H), 2.72 (m, 4H); 13
C
1D-O-(1,2-Di-O-butanoyl-sn-(2S)-glycerol-3-phospho)-5-(methyl-
enephosphonate)-myo-inositol (1). To a solution of 11a (18 mg, 0.021
mmol) in MeCN (1 mL) was added triethylamine (0.5 mL) and bis-
(trimethylsilyl)trifluoroacetamide (BSTFA, 0.5 mL) at room temperature
under Ar. After stirring the mixture at room temperature for 24 h, the
solvent was removed completely in vacuum. The residue was dissolved
in CH2Cl2 (1 mL), added to trimethylsilyl bromide (TMSBr, 0.5 mL)
and BSTFA (0.5 mL), then stirred at room temperature under Ar for
10 h. After concentration, methanol (1 mL) was added, and the mixture
was stirred for 1 h. The solution was evaporated and treated with cation-
exchange resin (Dowex 50W-X8, 400 mesh) to give 1 (9 mg, 75%).
1H NMR (400 MHz, CD3OD) 5.24 (m, 1H), 4.36 (m, 1H), 4.26 (m,
1H), 4.14 (m, 3H), 3.99 (m, 1H), 3.80 (m, 3H), 3.66 (m, 1H), 3.39 (t,
J ) 9.2 Hz, 1H), 3.15 (t, J ) 9.2 Hz, 1H), 2.28 (dd, J ) 7.6, 16.0 Hz,
4H), 1.58 (p, J ) 6.8 Hz, 4H), 0.89 (q, J ) 6.8 Hz, 6H); 13C NMR
NMR (101 MHz, CDCl3) 116.9, 98.8, 98.7, 98.1, 96.5, 82.2, 81.3, 76.8,
76.7, 76.3, 70.7, 62.85, 62.84, 62.77, 62.76, 57.0, 56.4, 56.0, 55.9, 19.6,
19.5; 31P NMR (162 MHz, CDCl3) 69.3. MALDI-HRMS [M + Na]+
calcd for C20H35N2O12SPNa, 581.1541; found, 581.1526.
1D-1-O-(Triethylsilyl)-5-O-benzoxyl-2,3,4,6-O-tetrakis(methoxy-
methylene)-myo-inositol (14). To a solution of 6 (400 mg, 0.57 mmol)
in DMF (5 mL) was added TBAF‚3H2O (360 mg). The mixture was
stirred at room temperature for 6 h, then diluted with 30 mL of ethyl
acetate. The solution was washed with 1 N HCl, water, then dried over
anhydrous sodium sulfate. After filtration and concentration, the
subsequent chromatography on silica gel (acetone/hexanes, 1:5) afforded
1
pure product (250 mg, 95%) as a colorless oil. H NMR (400 MHz,
CDCl3) 8.11 (d, J ) 6.8 Hz, 2H), 7.55 (t, J ) 7.2 Hz, 1H), 7.43 (m,
2H), 5.27 (t, J ) 9.6 Hz, 1H), 4.87 (d, J ) 6.8 Hz, 1H), 4.82-4.68
9
6504 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Phosphatidylinositol-5-Phosphate Analogues
A R T I C L E S
(m, 5H), 4.58 (t, J ) 6.8 Hz, 2H), 4.13 (m, 2H), 3.87 (t, J ) 9.6 Hz,
1H), 3.69 (dd, J ) 2.4, 9.6 Hz, 1H), 3.61 (d, J ) 9.6 Hz, 1H), 3.46 (s,
3H), 3.39 (s, 3H), 3.25 (s, 3H), 3.02 (s, 3H); 13C NMR (101 MHz,
CDCl3) 166.0, 133.3, 130.3, 129.9, 128.6, 98.4, 98.3, 98.1, 96.3, 81.2,
77.9, 76.9, 76.6, 74.0, 71.3, 56.2, 56.1, 56.0, 55.9. MALDI-HRMS [M
+ Na]+ calcd for C21H32O11Na, 483.1837; found, 483.1829. The
obtained alcohol (150 mg, 0.33 mmol) was dissolved in CH2Cl2 (8
mL), then imidazol (100 mg, 1.47 mmol) and triethylsilyl chloride (0.15
mL, 0.89 mmol) were added. After the mixture was stirred at room
temperature for 15 h, it was washed with water and dried over
anhydrous sodium sulfate. After filtration and concentration, the residue
was chromatographed on silica gel (acetone/hexanes, 1:4) giving product
was obtained as a colorless oil. 1H NMR (400 MHz, CDCl3) 5.22 (m,
1H), 4.75-4.67 (m, 8H), 4.30-4.08 (m, 13H), 3.99 (t, J ) 9.6 Hz,
1H), 3.93 (t, J ) 9.6 Hz, 1H), 3.48 (d, J ) 9.6 Hz, 1H), 3.39 (s, 3H),
3.37 (s, 3H), 3.34 (s, 3H), 3.33 (s, 3H), 2.72 (t, J ) 6.0 Hz, 6H), 2.25
(m, 4H), 1.59 (m, 4H), 0.89 (m, 6H); 13C NMR (101 MHz, CDCl3)
173.2, 172.9, 117.0, 116.7, 98.8, 98.7, 97.8, 96.6, 80.6, 76.9, 76.1, 75.8,
74.7, 69.4, 66.3, 66.1, 63.1, 63.0, 62.7, 62.4, 61.6, 57.1, 57.0, 56.2,
55.9, 36.1, 36.0, 19.8, 19.6, 19.5, 18.4, 13.8, 13.7; 31P NMR (162 MHz,
CDCl3) 68.99, 68.97, -1.25, -1.35. MALDI-HRMS [M + Na]+ calcd
for C34H57O19N3P2SNa, 928.2674; found, 928.2717.
1D-1-O-(1,2-Di-O-oleoyl-sn-(2S)-glycerol-3-O-(cyanoethyl)phos-
pho)-5-(bis(cyanoethyl)phosphothionate)-2,3,4,6-O-tetrakis(methoxy-
methylene)-myo-inositol (17b) was obtained from 13 and 10b as
1
14 (180 mg, 96%) as a pale yellow oil. H NMR (400 MHz, CDCl3)
1
8.09 (m, 2H), 7.48 (m, 1H), 7.35 (m, 2H), 5.16 (t, J ) 9.6 Hz, 1H),
4.83 (d, J ) 7.2 Hz, 1H), 4.74 (d, J ) 7.2 Hz, 1H), 4.71 (d, J ) 7.2
Hz, 1H), 4.70 (d, J ) 7.2 Hz, 1H), 4.66 (s, 2H), 4.50 (d, J ) 7.2 Hz,
1H), 4.44 (d, J ) 7.2 Hz, 1H), 4.06 (t, J ) 9.6 Hz, 1H), 3.94 (m, 2H),
3.61 (m, 2H), 3.36 (s, 3H), 3.35 (s, 3H), 2.95 (s, 3H), 2.88 (s, 3H),
0.90 (m, 9H), 0.54 (m, 6H); 13C NMR (101 MHz, CDCl3) 166.2, 133.1,
133.0, 130.7, 129.9, 128.5, 98.3, 98.1, 97.2, 95.9, 82.9, 80.2, 78.2, 76.9,
76.2, 73.9, 56.2, 56.1, 55.8, 55.7, 7.02, 5.04. MALDI-HRMS [M +
Na]+ calcd for C27H46O11SiNa, 597.2702; found, 597.2704.
described for 17a, yield 43%. H NMR (400 MHz, CDCl3) 5.33 (m,
4H), 5.25 (m, 1H), 4.74 (m, 8H), 4.41-4.09 (m, 13H), 4.04 (t, J ) 9.6
Hz, 1H), 3.96 (t, J ) 9.6 Hz, 1H), 3.55 (d, J ) 9.6 Hz, 1H), 3.43 (m,
12H), 2.77 (m, 6H), 2.32 (m, 4H), 2.00 (m, 8H), 1.59 (m, 4H), 1.28
(m, 40H), 0.87 (m, 6H); 13C NMR (101 MHz, CDCl3) 173.4, 173.0,
130.2, 129.9, 117.0, 116.7, 98.8, 98.7, 97.8, 96.6, 80.6, 76.1, 75.9, 75.8,
74.7, 69.5, 66.3, 66.1, 63.0, 62.6, 62.4, 61.7, 57.2, 57.0, 56.3, 56.0,
34.2, 34.1, 32.1, 29.97, 29.94, 29.73, 29.53, 29.41, 29.35, 29.30, 27.42,
27.39, 25.0, 22.9, 19.9, 19.8, 19.6, 19.5, 14.3; 31P NMR (162 MHz,
CDCl3) 69.01, 68.99, -1.15, -1.28. MALDI-HRMS [M + Na]+ calcd
for C62H109O19N3P2SNa, 1316.6743; found, 1316.6707.
1D-1-O-(Triethylsilyl)-2,3,4,6-O-tetrakis(methoxymethylene)-myo-
inositol (15). A solution of 14 (180 mg, 0.31 mmol) in CH2Cl2 (5 mL)
was cooled to -78 °C, then added to diisobutyl aluminum hydride
(1.0 M in hexane, 1.5 mL, 1.5 mmol). After the mixture was stirred at
-78 °C for 2 h, methanol (3 mL) was added, and the mixture was
stirred for 20 min. The residue was warmed to room temperature and
poured to wet sodium sulfate. After filtration and concentration, the
subsequent chromatography on silica gel (acetone/hexanes, 1:10)
1D-O-(1,2-Di-O-butanoyl-sn-(2S)-glycerol-3-phospho)-5-(phos-
phothionate)-myo-inositol (3). To a solution of 17a (20 mg, 0.022
mmol) in MeCN (1 mL) was added triethylamine (0.5 mL) and bis-
(trimethylsilyl)trifluoroacetamide (BSTFA, 0.5 mL) at room temperature
under Ar. After stirring at room temperature for 24 h, the solvent was
removed completely in vacuum. The residue was dissolved in CH2Cl2
(1 mL), added to trimethylsilyl bromide (TMSBr, 0.5 mL) and BSTFA
(0.5 mL), then stirred at room temperature under Ar for 2 h. After
concentration, methanol (1 mL) was added, and the mixture was stirred
for 1 h. The solution was evaporated and treated with cation-exchange
1
afforded product 15 (120 mg, 82%) as a colorless oil. H NMR (400
MHz, CDCl3) 4.84-4.70 (m, 8H), 3.94 (s, 1H), 3.81 (t, J ) 9.6 Hz,
1H), 3.66 (t, J ) 9.6 Hz, 1H), 3.52-3.32 (m, 15H), 0.94 (t, J ) 8.0
Hz, 9H), 0.59 (q, J ) 8.0 Hz, 6H); 13C NMR (101 MHz, CDCl3) 98.7,
98.4, 97.2, 96.0, 83.0, 80.7, 76.2, 75.4, 73.9, 72.8, 56.2, 56.1, 55.7,
55.6, 7.00, 5.05. MALDI-HRMS [M + Na]+ calcd for C20H42O10SiNa,
493.2445; found, 493.2316.
1
resin (Dowex 50W-X8, 400 mesh) to give 3 (8 mg, 64%). H NMR
(400 MHz, D2O) 5.16 (m, 1H), 4.28 (dd, J ) 3.2, 12.8 Hz, 1H), 4.10
(m, 2H), 3.94 (m, 2H), 3.83 (t, J ) 9.6 Hz, 1H), 3.73 (t, J ) 9.6 Hz,
1H), 3.65 (t, J ) 9.6 Hz, 1H), 3.45 (m, 2H), 2.25 (m, 4H), 1.45 (m, 4
H), 0.75 (m, 6H); 13C NMR (101 MHz, D2O) 176.6, 176.2, 84.1, 80.1,
79.5, 76.0, 75.4, 74.1, 72.3, 71.6, 70.7, 69.5, 65.6, 63.9, 63.6, 62.8,
61.5, 35.9, 35.7, 18.1, 18.0, 12.9; 31P NMR (162 MHz, D2O) 64.7,
-0.25. MALDI-HRMS [M + 3Na]3+ calcd for C17H33O16P2SNa3,
656.0658; found, 656.0604.
1D-1-O-(Triethylsilyl)-5-(bis(cyanoethyl) phosphothionate)-2,3,4,6-
O-tetrakis(methoxymethylene)-myo-inositol (16). Bis(2-cyanethyl)
diisopropyl-phosphorodiamidite (100 mg, 0.37 mmol) was added under
an argon atmosphere to a solution of 15 (40 mg, 0.085 mmol) and
1H-tetrazole (3 wt % in MeCN, 1.2 mL, 0.52 mmol) in MeCN (2 mL).
After the mixture was stirred at room temperature for 6 h, sulfur (72
mg) and CS2/Py (1:1, 0.72 mL) were added, and the mixture was stirred
for 3 h. The residue was filtered, and the filtrate was concentrated.
After subsequent chromatography on silica gel (acetone/hexanes, 1:4),
1D-O-(1,2-Di-O-oleoyl-sn-(2S)-glycerol-3-phospho)-5-(phospho-
thionate)-myo-inositol (4) was obtained from 17b as described for 3,
1
yield 67%. H NMR (400 MHz, CD3OD) 5.32 (m, 5H), 4.20-3.96
1
(m, 3H), 3.82-3.71 (m, 3H), 3.60-3.51 (m, 2H), 3.42 (m, 2H), 2.30
product 16 (45 mg, 79%) was obtained as a colorless oil. H NMR
(m, 4H), 2.00 (m, 8H), 1.59 (m, 4H), 1.28 (m, 40H), 0.88 (s, 6H); 13
C
(400 MHz, CDCl3) 4.85-4.70 (m, 8H), 4.34 (m, 5H), 3.95 (m, 3H),
3.55 (dd, J ) 2.4, 9.6 Hz, 1H), 3.48 (dd, J ) 2,4, 9.6 Hz, 1H), 3.45 (s,
3H), 3,44 (s, 3H), 3,43 (s, 3H), 3.41 (s, 3H), 2.77 (t, J ) 6.0 Hz, 4H),
0.97 (t, J ) 8.0 Hz, 9H), 0.62 (q, J ) 8.0 Hz, 6H); 13C NMR (101
MHz, CDCl3) 117.0, 98.7, 98.6, 97.2, 96.7, 81.0, 80.9, 76.9, 76.4, 76.2,
73.5, 62.8, 62.7, 57.1, 57.0, 56.0, 55.7, 19.5, 19.4, 7.04, 5.05; 31P NMR
(162 MHz, CDCl3) 69.1. MALDI-HRMS [M + Na]+ calcd for
C26H49N2O12PSSiNa, 695.2405; found, 695.2392.
NMR (101 MHz, CD3OD) 173.0, 172.7, 129.7, 129.6, 81.0, 79.5, 78.7,
76.9, 75.7, 75.2, 73.2, 72.1, 70.9, 68.2, 66.8, 66.2, 63.1, 61.7, 33.7,
31.8, 29.67, 29.59, 29.44, 29.28, 29.17, 29.08, 29.01, 28.97, 26.99,
26.95, 24.8, 22.5, 13.4; 31P NMR (162 MHz, CD3OD-CDCl3) 63.5,
0.78 (m). MALDI-HRMS [M - H + Na]+ calcd for C45H83O15P2SNa,
979.4753; found, 979.4791.
Liposome Binding Assay. The liposome binding assays were
performed as described.43 Briefly, solutions of 1-palmitoyl-2-oleoyl-
sn-glycero-3-phosphocholine (DPPC), 1-palmitoyl-2-oleoyl-sn-glycero-
3-phosphoethanolamine (DPPE), 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphoserine (DPPS) (Avanti) containing either PtdIns, PtdIns(5)P,
5-MP PtdIns(5)P (2), 5-PT-PtdIns(5)P (4), or no phosphoinositide were
dissolved in CHCl3/MeOH/H2O (65:25:4) and dried down under
vacuum. The lipids were resuspended in 800 µL of 20 mM Tris, 100
mM KCl, pH 7.3, and sonicated using a bath sonicator for 15 min at
1D-1-O-(1,2-Di-O-butanoyl-sn-(2S)-glycerol-3-O-(cyanoethyl)-
phospho)-5-(bis(cyanoethyl)phosphothionate)-2,3,4,6-O-tetrakis-
(methoxymethylene)-myo-inositol (17a). To a solution of 13 (25 mg,
0.045 mmol) in dichloromethane (2 mL) was added 1H-tetrazole (20
mg) and N,N-diisopropyl-O-(cyanoethyl)-O-(dibutanoyl-sn-(2S)-glycerol)-
phosphonamidite (10a, 75 mg) at room temperature under Ar. After
stirring at room temperature for 24 h, t-BuOOH (0.1 mL) was added
for oxidation, and the mixture was stirred for 1 h. The solution was
diluted by CH2Cl2 (20 mL) and washed with 10% sodium bisulfite,
then dried over anhydrous sodium sulfate. After subsequent chroma-
tography on silica gel (acetone/hexanes, 1:1), product 17a (20 mg, 50%)
(43) Lee, S. A.; Eyeson, R.; Cheever, M. L.; Geng, J.; Verkhusha, V. V.; Burd,
C.; Overduin, M.; Kutateladze, T. G. Proc. Natl. Acad. Sci. U.S.A. 2005,
102, 13052-13057.
9
J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6505

A R T I C L E S
Huang et al.
25 °C. Liposomes were collected by centrifugation at 25 000g for 30
min and finally resuspended in 95 µL of buffer by vortexing and
additional sonication for 3 min. Liposomes were incubated with 5 µL
of 2 µg/µL GST-ING2 or GST for 1 h at room temperature and then
collected again by centrifugation. The liposome pellets were separated
from the supernatant and resuspended in 100 µL of buffer. The pellet
and supernatant fractions were analyzed by SDS-PAGE using coo-
massie brilliant blue staining.
HT1080 Cell Death Assay. HT1080 cells were incubated with
vehicle or 45 nM neocarzinostatin (Sigma), at 16 h with 10 µM of
either buffer, or dipalmitoyl PtdIns, dipalmitoyl PtdIns(5)P, dioleoyl
5-MP PtdIns(5)P (2), or dioleolyl 5-PT-PtdIns(5)P (4) were added, and
cells were incubated in the combined mixture for 12 additional hours.
After incubation, dead cells were collected and added to the cells
collected after trypsin treatment of the plates and total cells were spun
down. Cell death was measured by trypan blue exclusion following
the manufacturer’s protocol (Invitrogen).
Molecular Modeling. The sequence of human ING2 was submitted
to the SWISS-MODEL server (http://swissmodel.expasy.org/) in “first
approach mode”. The server provided five homologous templates in
the PDB bank: 1wes, 1weu, 1wen, 2g6q, and 1x4i). Then a sequence
alignment was made by the alignment module of MODELLER 8v2.
On the basis of 20 models of 1x4i and 4 other templates, 20 homology
models were built by MODELLER 8v2 in automodel mode. After
reconstruction of the model side chains by isosterically replacing the
template side chains and energy minimization by the GROMOS96 force
field, the model of ING2 C-terminal was obtain as a pdb file.
AutoDock3.0 software package, with the graphical user interface
AutoDockTools (ADT), was used to dock small molecules
(PtdIns(5)P and analogues) to the ING2 C-terminal domain. The zinc
atoms in the active site were each manually assigned a charge of +2,
and all other atom values were generated automatically by ADT. With
the use of AutoGrid, a grid of 72 Å × 72 Å × 80 Å with 0.375 Å
spacing was calculated around the docking area for the relative ligand
atom types of the PtdIns(5)P and analogues. Separate docking computa-
tions were performed for each ligand using the Lamarckian genetic
algorithm local search (GALS) method. The docking conformations
of each ligand were clustered on the basis of root-mean-square deviation
(rmsd) and ranked on the basis of free energy of binding. The top-
ranked conformations were visually studied for good chemical geom-
etry. On the basis of the complex of ING2-PtdIns(5)P, the binding
interactions were determined by LIGPLOT 4.4.2 giving the hydrogen
bonds and hydrophobic contacts information.
Acknowledgment. We thank the NIH (NS 29632 to G.D.P.,
GM 079641 to O.G., and GM 071424 and CA 113472 to
T.G.K.). O.G. is a recipient of a Burroughs Wellcome CABS
award and a Kimmel Scholar award. We also thank P. Neilsen
(Echelon Biosciences, Inc.) for generous gifts of PtdIns(5)P and
PtdIns.
Supporting Information Available: Computed docking of
ligands with ING2 models; proton, carbon, and phosphorus
NMR data for characterization of new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA070195B
9
6506 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007