Nonhydrolyzable analogs of phosphatidylinositol as ligands of phospholipases C

Article abstract of DOI:10.1039/b9nj00629j

Phosphatidylinositol-specific phospholipases C (PI-PLCs) are important enzymes participating in transmembrane signal transduction. The structures of the two major species of these enzymes: bacterial Ca²⁺-nondependent enzyme from B. cereus and mammalian Ca²⁺-dependent PLCδ₁ from rat brain in the complexes with the polar head groups of their substrates have been previously solved. Although these structures show few differences as compared to free enzymes, there is a compelling evidence that full catalytic activity of PI-PLC necessitates interaction of the enzyme with the entire substrate, including the hydrophobic fatty acid chains. In this work we develop new tightly binding and cleavage-resistant analogs of phosphatidylinositol, using relatively minor modifications of the structure. Two synthesized analogs, 2-amino-2-deoxy-PI (8) and the conformationally constrained analog (10) had binding affinities (K _i) in 10 μM range. ¹⁵N-¹H HSQC NMR spectra of uniformly ¹⁵N-labeled bacterial Ca²⁺-nondependent and Ca²⁺-dependent phospholipases C, btPLC and saPLC1, respectively, displayed changes upon ligand binding that suggest an occurrence of a conformational change.

Full text of DOI:10.1039/b9nj00629j

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
Nonhydrolyzable analogs of phosphatidylinositol as ligands
of phospholipases Cwz
Cornelia Mihai,ac Xiangjun Yue,a Li Zhao,bd Alex Kravchuk,b Ming-Daw Tsaibe
and Karol S. Bruzik*a
Received (in Montpellier, France) 2nd November 2009, Accepted 20th January 2010
First published as an Advance Article on the web 9th March 2010
DOI: 10.1039/b9nj00629j
Phosphatidylinositol-specific phospholipases C (PI-PLCs) are important enzymes participating in
transmembrane signal transduction. The structures of the two major species of these enzymes:
bacterial Ca2+-nondependent enzyme from B. cereus and mammalian Ca2+-dependent PLCd1
from rat brain in the complexes with the polar head groups of their substrates have been
previously solved. Although these structures show few differences as compared to free enzymes,
there is a compelling evidence that full catalytic activity of PI-PLC necessitates interaction of the
enzyme with the entire substrate, including the hydrophobic fatty acid chains. In this work we
develop new tightly binding and cleavage-resistant analogs of phosphatidylinositol, using
relatively minor modifications of the structure. Two synthesized analogs, 2-amino-2-deoxy-PI (8)
and the conformationally constrained analog (10) had binding affinities (Ki) in 10 mM range.
15N-1H HSQC NMR spectra of uniformly 15N-labeled bacterial Ca2+-nondependent and
Ca2+-dependent phospholipases C, btPLC and saPLC1, respectively, displayed changes upon
ligand binding that suggest an occurrence of a conformational change.
Introduction
Phosphatidylinositol-specific phospholipases C (PI-PLCs) are
key enzymes of signaling cascades that hydrolyze a minor
membrane constituent, phosphatidylinositol 4,5-bisphosphate
(PIP2), and generate two second messengers, inositol 1,4,5-
trisphosphate (IP3) and diacyl glycerol (DAG).1–3 Activation
of PI-PLC occurs in response to numerous extracellular
signals including growth factors, neurotransmitters and
hormones, and involves a broad range of cell surface receptors.
Binding of IP3 to its receptor on the endoplasmic reticulum
results in an increase of intracellular Ca2+ concentration,
whereas DAG activates protein kinase C. Together, these
changes control numerous cellular processes, such as DNA
Scheme 1 General mechanism of bacterial Ca2+-nondependent PI-PLC.
synthesis, cell proliferation, and neuronal activity.1–3
The catalytic mechanism of bacterial Ca2+-nondependent
PLC from Bacillus thuringiensis (btPLC) involves (i) activation
of the 2-hydroxy group of inositol by a composite general base
(GB) dyad, Asp274-His32, and (ii) protonation of the leaving
group by the general acid (GA) His82, networked with (iii) the
Amino acid residue numbering is that of Bacillus thuringiensis enzyme.
negative charge stabilizing residue, Arg69, via an Asp33 inter-
mediary (Scheme 1).4,5 In the case of Ca2+-dependent PLCd1,
the negative charge stabilization function of Arg69 is replaced
by Ca2+ cation, that of His82 GA by His356, whereas the
identity of the GB remains unclear.6 The available X-ray data
for B. cereus PI-PLC and for mammalian PI-PLCd1 indicate
that very small changes (if any) occur to the enzyme confor-
mation upon binding of the polar head group of the
phospholipid substrate.7–9 However, since it is also known
that activity of PI-PLC is reduced by a factor of 105 (kcat/Km)
by the removal of hydrophobic side chains from the
substrate,4,10–12 it is quite possible that the reported X-ray
structures of enzyme-head group complexes8,9 do not repre-
sent the fully active enzyme conformations. For example, we
have found that for btPLC, formation of the catalytic triad
a Department of Medicinal Chemistry and Pharmacognosy, University
of Illinois at Chicago, Chicago, IL 60612. E-mail: kbruzik@uic.edu
b Department of Chemistry, The Ohio State University, Columbus,
Ohio 43210
c Department of Natural Sciences, Northwestern Oklahoma State
University, Alva, OK 73717
d Department of Molecular Biology, The Scripps Research Institute,
La Jolla, CA 92037
e Institute of Biological Chemistry, Academia Sinica, Taipei 115,
Taiwan
w This article is part of a themed issue on Biophosphates.
z Electronic supplementary information (ESI) available: Characteri-
zation of lactone 23 and ligand 10 synthesis of intermediate 17. See
DOI: 10.1039/b9nj00629j
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 925–933 | 925

View Article Online
Arg69-Asp33-His82 occurs only in the presence of the hydro-
phobic leaving group, and that enzyme interactions with
substrate hydrophobic chains strongly affect both the general
acid catalysis and negative charge neutralization functions of
the enzyme.4,11 The goal of this work was to synthesize and
test analogs of phosphatidylinositol that while preserving as
many of the active site interactions as possible, would still
remain nonhydrolyzable to the PI-PLC isozymes.
transient-voltage-gated ion channels, however, the inhibitory
power of these compounds was not determined.19b The goal of
this work was to develop tight-binding and nonhydrolyzable
ligands of phospholipase C using as few modifications of
the native phosphatidylinositol structure as possible. These
ligands would be useful in determination of the structures of
PLC isozymes in complexes with their substrates by NMR and
crystallographic methods, and also as research tools for
investigation of inositol signaling pathways.
The design of the ligands for the structural work is different
from the design of tight-binding inhibitors, since in the former
case preservation of the substrate’s basic structural features is
important. Ideally, a relatively minor modification should
render the substrate analogs resistant to cleavage, while at
the same time enhancing its binding affinity. For PI-PLC
ligands, the available information indicates that the presence
of the hydrophobic residues is important not only for catalysis,
but also for substrate binding. The affinity of PLC toward the
substrates with short carboxylic acid esters is lower, and is
further drastically reduced when these chains are completely
removed.12,13 Thus, btPLC and PLCd1 bind the deacylated PI
(Gro-PI) and IcP with very low affinity.10,12 We conclude, that
a successful ligand for investigation of the structure of the
PLC-substrate complex should bear parts of diacyl glycerol
residue. An additional parameter that must be taken into
account during ligand design is its critical micellar concentration
(cmc). In order to prevent ligand aggregation and the resulting
complications (e.g. signal broadening in NMR spectra), most
of the ligands synthesized in this work feature short chains,
resulting in the cmc values above the binding constants.13
Two types of modifications have been previously explored
to confer resistance of PI analogs to PLC cleavage: in the first
case a nucleophilic 2-OH group of inositol was inverted (1),14
deoxygenated (2),15,16 fluorinated (3)16 or methylated (4)16,17
(Fig. 1). In general, these compounds were reported to be
resistant to cleavage by PLC, but unfortunately showed only
weak inhibitory properties (in the millimolar range). Alter-
natively, the scissile P–O bond was replaced with a C–P bond
as in the phosphonate 5.18,19a The phosphonate analogs were
potent inhibitors with Ki in the 10 mM range, but we have
deliberately excluded such ligands from our considerations
since our goal was to obtain substrate analogs that would
provide information about enzyme interactions with the
leaving group. Zhang et al. have synthesized a-fluoroalkyl-
phosphonate analogs of PIP2 as a replacement for PLC-
hydrolyzed PIP2 that were able to restore properties of TRPM
Results and discussion
Our earlier studies on the mechanism of bacterial PI-PLC
indicated that interactions of enzyme with the inositol 2-OH
group, the nonbridging phosphoryl oxygen atoms and the
bridging oxygen atom of the leaving group provide the bulk
of catalysis. Hence in our ligand design we strove to retain all
three types of interactions. We have also found that the
replacement of the pro-S oxygen atom in the phosphoryl
group by sulfur reduces the rate of the cleavage reaction by
6 orders of magnitude, while preserving the negative charge of
the phosphoryl residue. Thus, the phosphorothioate analog 6
could prove sufficiently resistant to PI-PLC to permit acquisi-
tion of X-ray diffraction or NMR data. The corresponding
dithio analog 7 had relatively weak inhibitory properties.16 In
contrast to compounds 1–4 altered at the 2-position of
inositol, the analog 8 featuring the axial amino group in this
place should retain some form of interactions between the
protonated 2-amino group and e.g. the general base residue
(His32) of the enzyme (Scheme 1). We have also examined an
analog in which the nonscissile C1–O bond was replaced by a
C–C bond (9). Finally, we hypothesized that the conforma-
tionally constrained PI analog, such as the compound 10,
should have stronger binding affinity than the natural sub-
strate, since due to the proximity of the 2-OH group and the
phosphorus atom in this structure, it should mimic a near-
attack-conformation (NAC)20 directly preceding formation of
the transition state. The structures of the previously synthe-
sized and prospective substrate analogs are shown in Fig. 1.
Synthesis of phosphatidylinositol analogs
The compounds 2 and 4 were prepared previously by
others,16,17 and therefore their syntheses are not described
here. We used these compounds for comparison with the
new ligands 6, 8 and 10. The analog 6 was prepared by the
enzymatic separation of the mixture of Rp- and Sp-isomers of
dihexanoyl phosphorothioate analog of PI, (Rp + Sp)-DHPsI.
This mixture was treated with btPLC, and after hydrolysis of
the Rp-isomer, the remaining Sp-isomer was separated
chromatographically from the products of the reaction.
The main challenge in the synthesis of the amino analog 8
was to retain the axial orientation of the amino group at the
2-position. In order to achieve this goal, we have designed two
synthetic pathways (Scheme 2), both involving scyllo-inosose
12 as a key intermediate obtained in good yield from alcohol
11 by Swern oxidation. Unfortunately, reduction of 12 with
several reagents regenerated the alcohol 11 as a major product,
with very little scyllo-inositol derivative 13 (having an
equatorial 2-OH). Although better yields of an equatorial
Fig. 1 Structures of PI analogs considered in this work.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
926 | New J. Chem., 2010, 34, 925–933

View Article Online
Scheme 2 Synthesis of 2-deoxy-2-amino-DHPI, 8. (i): Ac2O, DMSO;
(ii): NH4OAc, NaBH3CN, MeOH; (iii): K2CO3, Cbz-Cl (R1 = Cbz);
(iv): CAN, acetonitrile–water; (v): Cl–P(OMe)(NiPr2), DIPEA;
(vi): 17, tetrazole; (vii): tert-BuOOH; (viii): Me3N; (ix): H2, Pd/C.
Scheme 3 Synthesis of the analog 10 from the phosphonate 19. (i):
t-BuLi, methyl chloroacetate; (ii): Me3N; (iii): 1-(mesitylene-2-sulfonyl)-
3-nitro-1,2,4-triazole (MSNT), C16H33OH; (iv): TBAF; (v): TsOH,
MeOH.
alcohol could be obtained starting from another substrate with
a benzoyl group at the 1-position,14 the separation of axial and
equatorial products proved difficult. Hence, a strategy relying
on a nuclophilic displacement of an activated equatorial
alcohol 13 with an azide was not be pursued. Since the
reducing agents preferentially attacked the carbonyl group in
2-inosose from the less hindered equatorial direction generat-
ing a product which features an axial group at the 2-position,
we reasoned that this finding may be exploited for a direct
conversion of the carbonyl into the axial amino group.
This assumption was validated by converting 2-inosose 12
into 2-deoxy-2-amino-inositol 14 via reductive amination of
the ketone with ammonium acetate and sodium cyanoboro-
hydride in methanol with very good yield. The amine 14 was
not isolated, but instead directly protected with benzyloxy-
carbonyl group (Cbz) to generate compound 15. The axial
orientation of the 2-amino group was evident from the coupling
pattern of inositol protons assigned by 1H COSY NMR
spectra of 15. Thus, proton H-2 of 15 displayed 1.2 Hz coupling
constant to H-3 and 4.5 Hz to H-1, consistent with
equatoral—axial interaction and indicating equatorial orienta-
tion of H-2. The removal of the PMB group from 15 with
ceric ammonium nitrate (CAN) in acetonitrile–water (4: 1)
produced the amino alcohol 16 in good yield. The subsequent
phosphitylation with the with P-chloro-N,N-diisopropyl-
O-methyl-phosphoramidite, followed by coupling with
1,2-dihexyl-sn-3-glycerol (17) and oxidation with tert-butyl
hydroperoxide afforded the fully protected phosphate 18.
The deprotection of 18 by consecutive reactions with
anhydrous trimethylamine and catalytic hydrogenolysis
afforded the analog 8.
the phosphonate 19 with methyl chloroacetate afforded the
derivative 20. The subsequent introduction of the hydro-
phobic ester group was achieved by demethylation of the
phosphonate 20 with trimethylamine followed by the coupling
with hexadecanol using the Narang’s triester approach22 to
give the triester 21. The demethylation of the phosphonate 21
and desilylation of the inositol 4-OH group afforded the
alcohol 22.
The cyclization of 22 was achieved in one step via
acid-catalyzed removal of the isopropylidene groups followed
by lactonization with the 6-OH group. It is noteworthy
that our other numerous approaches to similar cyclizations
including 6-OH alkylation, or formation of a hemiacetal with
the 6-OH group, have proved unsuccessful. Since the analysis
of the final product 10 was difficult due to its poor solubility
and aggregation in aqueous and organic solvents, its stereo-
chemical structure was established indirectly by the cyclization
of the intermediate 20 followed by analysis of the neutral
lactone 23. 1H NMR of 23 showed large vicinal coupling
constants between 31P and the axial protons at the 1- and
a-positions (3JHCCP = 19.0 Hz and 26.7 Hz, respectively),
typical of the trans-orientation, whereas the NOESY spectrum
showed a cross peak between H-2 and H-b in accordance with
their diequatorial orientations (for full characterization of
compound 23 see Supporting Information).
The conformationally constrained analog 10 was synthe-
sized in 6 steps from the phosphonate analog of inositol
1-phosphate 1921 (13 steps overall from myo-inositol) as
shown in Scheme 3. In brief, alkylation of the a-carbon in
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 925–933 | 927

View Article Online
Fig. 2 31P NMR of the cleavage of the phosphonate 10 by btPLC.
Conditions: The solution of 10 (8 mM) in MOPS buffer (50 mM,
pH 7.0, 0.5 mL) containing sodium deoxycholate (32 mM) and EDTA
(2 mM) was treated with btPLC (0.33 mg). The bottom spectrum was
collected immediately after enzyme addition; the top spectrum was
registered after 45 h at 23 1C.
In addition, the correct stereochemical structure of the final
analog 10 was established by its PI-PLC-catalyzed cyclization
to afford the cyclic phosphonate 24, albeit with 106-fold lower
activity than that of PI-PLC with natural substrate (Vmax
=
0.004 mmol mgꢁ1 minꢁ1). The cleavage of the analog 10
by PI-PLC resulted in the formation of the product 24
displaying ca. 18 ppm downfield shift in 31P NMR from that
of the substrate 10, consistent with formation of the
five-membered ring phosphonate. Due to the racemic nature
of the analog 10 synthesized and strict stereochemical specifi-
city of PI-PLC, the cyclization reaction stopped at the 50%
conversion (Fig. 2).
Fig.
3
Lineweaver–Burk plots for btPLC (A–E) and saPLC1
inhibition (F): (A) 2-deoxy-DHPI (2): (J) DOsPI, (m) DOsPI +
2 mM 2; (B) 2-O-methyl-DHPI (4): (J) DOsPI; (K) DOsPI + 5 mM
4; (C) Sp-DHPsI (6): (J) DOsPI, (&) DOsPI + 100 mM 6;
(D) 2-deoxy-2-amino-dhPI (8): (J) DOsPI, (m) DOsPI + 250 mM
8; (E) (10): (J) DOsPI, (K) DOsPI + 100 mM 10; F: (&) DOsPI, (J)
DOsPI + 250 mM 8.
The analog 9 was prepared starting from compound 19,
using demethylation with trimethylamine followed by MSNT-
catalyzed coupling with dipalmitoyl glycerol (see below)
followed by acid-catalyzed deprotection of inositol ring. Not
unexpectedly, the phosphonate analog 9 turned out to be a
good substrate of btPLC with Vmax = 670 mmol mgꢁ1 minꢁ1
.
where Km is the Michaelis constant of the substrate, Kp is the
apparent Michaelis constant of the substrate in the presence of
inhibitor, Ki is the inhibition constant, [I] is the concentration
of inhibitor, Vm is the maximum velocity, and Vp is the
apparent maximum velocity (with inhibitor).
Inhibition of PI-PLC by PI analogs
Two of the ligands synthesized in this work, 2-deoxy-DHPI
and 2-MeO-DHPI, have been previously tested by Martin and
Wagman,16 using p-nitrophenyl inositol phosphate as an assay
substrate. The problem with their measurements was that due
to a high Km of the substrate (between 5 and 15 mM depending
upon the buffer), the assays were performed under non-
saturating substrate concentrations (0.2 mM). Hence, the Ki
values reported for 2 (2.6 mM) and 4 (6.6 mM) are of
uncertain reliability.
Unexpectedly, 2-deoxy-PI, 2, was found to be a poor
substrate with a cleavage rate ca. 106-fold slower than that
of PI. In view of the absence of the nucleophilic 2-OH group,
the instability of compound 2 is interesting since it indicates
that PLC can substitute a water molecule for the missing
hydroxyl group to effect a direct hydrolysis, albeit at a very
low rate. Our experiments also showed that this compound is a
weak uncompetitive inhibitor (Fig. 3A) with similar effects on
both Km and Vmax and the calculated Ki of 5 mM. 2-O-Methyl-
DHPI (4) was found to be one of the most stable of all ligands
compared. It was completely resistant to 1 mg of WT PI-PLC
for more than one week, hence the upper limit of the cleavage
rate is 7.7 ꢂ 10ꢁ5 mmol minꢁ1 mgꢁ1). However, it was a rather
weak non-competitive inhibitor with Ki value of 11 mM
(Fig. 3B). Sp-DHPsI (6) was cleaved by btPLC more than
106 times slower than the natural substrate (Vmax = 9 ꢂ
10ꢁ4 mmol minꢁ1 mgꢁ1). This compound was a partially non-
competitive inhibitor with a significant effect on maximum
velocity of substrate and lower effect on Km (Fig. 3C). Using
the expression for a non-competitive inhibition, the Ki was
In this work, the evaluation of inhibitors was performed by
a continuous UV assay using phosphorothiolate dioctanoyl
analog of PI, DOsPI, as a substrate that has a small Km
(75 mM).13,23 The assays were performed at substrate concen-
trations ranging from 25 mM to 400 mM, while inhibitor
concentrations were kept constant. For each inhibitor, the
type of inhibition was assessed, and the Ki value for each
inhibitor calculated using the following expressions:24
Competitive inhibition: Ki = Km [I]/(Kp ꢁ Km)
Uncompetitive inhibition: Ki = Kp [I]/(Km ꢁ Kp)
Non-competitive inhibition: Ki = Vp [I]/(Vm ꢁ Vp)
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
928 | New J. Chem., 2010, 34, 925–933

View Article Online
calculated at 23 mM. A possible explanation of these inhibitory
behaviors of ligands 2, 4 and 6 is that phenomena such as
allosteric activation by the added inhibitors (increasing Vmax
)
and surface dilution in the micelles (reducing Vmax) are
additional parts of the overall interactions. If the active site-
binding of an inhibitor (competitive inhibition) is not signifi-
cantly stronger than that of the substrate, these phenomena can
lead to apparent modes of inhibition other than competitive.
For compound 8, in addition to the replacement of the 2-OH
group with the amino group, which prevents the nucleophilic
attack, this analog was also equipped with ether-linked instead
of ester-linked chains. This modification of DAG moiety should
provide stronger interactions with PLC.19 Compound 8 was
found to be resistant to cleavage by btPLC for more than one
week using 5 mM concentration of 8 and 1 mg of btPLC. The
estimated upper limit of the rate of the cleavage is hence
8.5ꢂ10ꢁ5 mmol minꢁ1 mgꢁ1). Fig. 3D indicates that this
compound is a strong competitive inhibitor with a Ki of 15 mM.
Hence, this compound appears to be one of the most promising
ligands synthesized to date, offering a conservative structure
modification, tight binding and complete resistance to enzymatic
cleavage. Compound 8 was also a weaker inhibitor of saPLC1
with Ki = 0.23 mM (Fig. 3F). The bicyclic compound 10 was
cleaved by btPLC at low rate, and was also a clean competitive
inhibitor with Ki = 21.3 mM (Fig. 3E). Given that btPLC is
strictly stereospecific and that the synthesized compound was
racemic, the Ki for compound 10 is just 10.6 mM. Additional
modifications must be performed to improve stability of this
compound to PI-PLC–catalyzed cleavage.
Fig. 4 Arginine side chain (eNH) region of 1H-15N HSQC spectra of
uniformly 15N-labeled btPLC. (A) WT, (B) R69D, (C) R69D + 8 mM
6, (D) WT + 10 mM 4. All samples were in 10% D2O-90% H2O,
50 mM HEPES, pH 7.5. The protein concentration was ca. 0.4 mM.
The open circles with dashed line represent the original locations of the
arginine cross peaks in the spectrum of WT. Spectra were acquired on
a Bruker DRX-800 spectrometer at 37 1C.
Conformational changes of PI-PLC upon ligand binding
Our goal was to examine binding properties of the synthetic
substrate analogs and detect structural changes in PLC upon
ligand binding. We have used 15N-1H HSQC NMR of
uniformly 15N-labeled enzymes as the means to generate
signature data, and to identify the best ligands for further
structural studies, either by multidimensional NMR or X-ray
crystallography. The partial spectra of the arginine side chains
of btPLC are shown in Fig. 4. Note, that the rates of cleavage
of compounds 2, 6 and 10 by the WT enzyme were somewhat
too high for these ligands to be used in NMR experiments,
therefore an active site mutant, Arg69Asp, was used. Fig. 4A
clearly shows signals from all nine arginine side chains
(eNH region) of btPLC. The assignment of the resonaces of
Arg69, Arg71 and Arg163 was achieved using the approriate
changes in the spectra than those observed for Arg69Asp. In
the arginine region, only Arg69 was shifted downfield. These
results suggest that Arg69Asp mutant undergoes a conforma-
tional change upon ligand binding. It is unclear at the moment
whether the same is true for the WT enzyme, since the
differences observed (with a different ligand) are smaller.
Whether this is related to weak binding of the ligand 4 or
greater conformational stability of WT btPLC is also uncertain.
Another PLC that we are interested in is a Ca2+-dependent
PI-PLC from Streptomyces antibioticus (saPLC1).25 This
enzyme employs a unique mechanism in which the 6-hydroxyl
group, rather than 2-OH, is an attacking nucleophile.26 Ligand
1
mutants. The comparison between the H-15N HSQC spectra
of WT and Arg69Asp (Fig. 4A–B) shows that in the mutant,
the signals of not only Arg69 but also Arg71 disappeared,
while the signal of Arg163 was slightly shifted upfield. Upon
addition of the analog 6 (Fig. 4C) to this mutant, the signal of
Arg71 reappeared in a somewhat shifted position, and the
signal of Arg163 moved downfield by almost 3 ppm (1H). The
small residual signal at the original position of Arg163
indicates that the ligand-bound and unbound protein species
remain in a slow conformational exchange on the NMR time
scale. Similar results were observed for analogs 4 and 8 with
the same mutant (spectra not shown). In contrast, addition of
10 mM 4 to WT (Fig. 4D) produced much less dramatic
8 inhibited saPLC1 in a competitive manner with Ki
=
230 mM, and was completely resistant to this enzyme.
The effect the ligand 8 binding on the conformation
of saPLC1 was qualitatively assessed using 15N-1H HSQC
spectra (Fig. 5). In the absence of Ca2+, no obvious spectro-
scopic changes could be observed in the 15N-1H HSQC
spectrum upon addition of 0.2 mM 8 (data not shown). After
adding Ca2+ to this system, about 30 signals in the amide
backbone region moved to different positions (Fig. 5). These
results indicate that compound 8 binds to the enzyme in a
Ca2+-dependent manner.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 925–933 | 929

View Article Online
reactions were anhydrous, and were purchased from Aldrich.
These solvents were stored in vacuum-tight ampoules over
appropriate drying agents and transferred to the reaction
vessel by molecular distillation under vacuum. Chloroform
and methylene chloride were passed through a column of basic
alumina to remove traces of water and acid, and stored over
calcium hydride or phosphorus pentoxide. Acetonitrile and
toluene were stored over calcium hydride and diisopropylethyl
amine (DIPEA) was stored over sodium hydride. THF
was stored over Na–K alloy or lithium aluminium hydride.
Anhydrous methanol was dried by magnesium turnings, in the
presence of catalytic amounts of iodine, distilled and stored
over 3 A molecular sieves. 1H NMR spectra were obtained
with NMR spectrometers operating at 300 and 500 MHz
frequencies, while 13C NMR spectra were obtained at 75 and
125 MHz. 31P NMR spectra were obtained at 121.5 MHz with
Bruker DPX-300 NMR spectrometer and the chemical shifts
are referenced to external 80% orthophoshoric acid.
PLC Assay. The assay used in this work is a modified
procedure first reported by Hendrickson and co-workers29
and expanded in our recent publication.13 For btPLC assay,
the appropriate amount of a 10 mM stock solution of DOsPI
in DI water to provide the final substrate concentration of
25–400 mM was loaded into a semi-micro UV cuvette. The
buffer solution (40 mL, 50 mM MOPS, 1 mM EGTA, pH 7.0)
was added into the cuvette followed by the solution of 4,40-
dithiopyridine (DTP) in ethanol (10 mL, 100 mM). An aliquot
of the aqueous solution of an inhibitor (1.0–10 mM depending
on a ligand) was added next to obtain a desired concentration
of an inhibitor shown in Fig. 3. The mixture was diluted with
distilled water to 390 mL, and the contents of the cuvette were
stirred and equilibrated for 3 min at room temperature to
achieve optical clarity. Background absorbance was recorded
for 200 s at 324 nm (25 1C) against the MOPS buffer on a
Hitachi U3010 spectrophotometer. The reaction was initiated
by adding 10 mL of a freshly prepared enzyme solution, and
the absorbance at 324 nm was recorded for 200–400 s.
The molar extinction coefficient of 4-pyridyl thiol, e =
19 800 Mꢁ1 cmꢁ1, was used to calculate the initial rates of
reaction. The dependencies of these rates vs. the substrate
concentrations were fitted to kinetic equation, described
above, using either GraFit or KaleidaGraph software to
obtain the kinetic parameters. The assay conditions for
saPLC1 were the same except 3 mM CaCl2 was included in
the assay mixture.
Fig. 5 The backbone amide region of 15N-1H HSQC spectra of
uniformly 15N-labeled saPLC1. Overlaid are spectra for saPLC1
(black), saPLC1 + 1 mM free Ca2+ (green) and saPLC1 + 1 mM
free Ca2+ + 0.2 mM 8 (red). All samples were in 10% D2O–90%
H2O, 50 mM HEPES, 1 mM EDTA, pH 7.0. The protein concentra-
tion was 0.2 mM. Spectra were acquired on a Bruker DRX-800
spectrometer at 37 1C.
As a control, we also examined the changes in the 15N-1H
HSQC spectrum of saPLC1 upon Ca2+ titration. As shown in
Fig. 5, about 50 signals in the amide backbone region have
shifted upon Ca2+ addition. Interestingly, overlaying three
spectra (free enzyme, enzyme + Ca2+, enzyme + Ca2+ + 8)
demonstrates that most of the signals, which were subject to
changes upon inhibitor binding, also experienced shifts upon
calcium binding. This ultimately suggests that Ca2+ and 8 binds
to saPLC1 in a closely related region, possibly the active site.
Conclusions
We have synthesized several analogs of phosphatidylinositol
with the aim of achieving strong binding and complete resistance
to PI-PLC with relatively minor conservative modifications of
the PI structure. Among six compounds prepared, 2-deoxy-2-
amino analog 8, and the conformationally constrained analog 10
were the strongest competitive inhibitors. In addition,
compound 8 was also completely resistant to cleavage by the
enzymes, while compound 10 underwent slow cleavage. The
latter analog will require additional modification to achieve
complete hydrolytic resistance. Using these conservative analogs
of PI and 15N-1H HSQC NMR spectra we have shown that both
btPLC and saPLC1 seem to undergo conformational transition
upon ligand binding. Further studies are needed to solve the
structures of these enzymes in complexes with the ligands.
Synthesis
Alcohol 11 was synthesized according to a published procedure.27
Synthesis of the phosphonate 19 was performed as reported
previously.21 btPLC and its mutants were expressed as
described earlier.4 Expression and purification of saPLC1 is
described in the reference 28.28 Uniformly 15N-labeled WT
saPLC1 was expressed in M9 minimal media using 15NH4Cl
Experimental
as
a single nitrogen source and purified as described
earlier.28 Two-dimensional 15N-1H HSQC spectra of the
uniformly 15N-labeled protein at Ca2+ concentration ranging
from 0 to 5 mM were obtained on a Bruker DRX-800
spectrometer at 37 1C.
General
All reagents were obtained from Aldrich Chemical Company
and Fisher Scientific. All solvents used for moisture sensitive
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
930 | New J. Chem., 2010, 34, 925–933

View Article Online
(Sp)-(1,2-Di-O-hexanoyloxypropane-3-thiophospho)-1-D-myo-
inositol (6). The mixture of (Rp + Sp) DHPsI diastereomers
(0.285 g, 0.52 mmol), synthesized analogously as described
earlier,30 was dispersed by sonication in water (30–40 mL) and
the pH of the solution obtained was adjusted to pH = 7.0 with
50 mM ammonium formate. To this mixture was added
PI-PLC from B. thuringiensis (40 mg) in 50 mM MOPS buffer
(pH 7.0, 50 mL). The mixture was stirred vigorously at room
temperature for 5 h, monitoring the progress of the reaction by
31P NMR. When the Rp-DHPsI was completely cleaved, the
mixture was extracted twice with diethyl ether (2 ꢂ 20 mL) and
the aqueous extract was loaded on a C18-reverse phase
column. Elution with 4 column volumes of water delivered
all the compounds soluble in water, including IcPs, while the
fractions obtained from elution with 4–5 column volumes of
methanol contained Sp-DHPsI. The methanolic fractions were
concentrated, the residue dispersed in water and the solution
passed through a cation-exchange column (ammonium-form).
The freeze-drying of the fraction eluted from the column
afforded pure 6 as an ammonium salt (0.102 g, 90% recovered
from the initial mixture) as a white solid. 1H NMR (CD3OD) d
d 0.90 (t, J = 6.9 Hz, 6H), 1.33–1.40 (m, 8H), 1.59–1,66
(m, 4H), 2.31–2.39 (m, 4H), 3.18–3.25 (m, 1H), 3.45 (dd, 1H),
3.64 (t, J = 9.5 Hz 1H), 3.79 (t, J = 9.5 Hz, 1H), 4.11–4.22
(m, 5H), 4.4 (m, 1H), 5.25 (m, 1H); 31P NMR (D2O) d57.0. 61;
ES-MS (m/z) 545.
Purification by column chromatography (hexane–ethyl
acetate, 2 : 1) afforded pure amine (225 mg, 90%) as a white
solid (Rf 0.24, hexane–ethyl acetate, 2 : 1). To the solution of
the amine in dioxane-water (1.5 : 1, 15 mL) solid potassium
carbonate (51.4 mg, 0.32 mmol) was added in portions and the
mixture was stirred for a few minutes. Benzyl chloroformate
(100 mL, 0.7 mmol) was added stepwise under nitrogen and the
reaction was vigorously stirred for 1 h. The progress of the
reaction was monitored by TLC (hexane–ethyl acetate, 2 : 1).
When no more starting material was observed by TLC, the
mixture was concentrated to dryness, the residue redissolved in
methylene chloride and the solution washed with brine, dried
and concentrated. Purification by column chromatography on
silica gel (hexane–ethyl acetate, 2 : 1) gave pure 15 (207 mg,
90% from the amine) as a white solid. Rf 0.55 (hexane–ethyl
acetate, 2 : 1); [a]2D0 = ꢁ51 (c 1.6, chloroform); 1H NMR
(CDCl3:CD3OD, 1 : 1) d 3.55–3.71 (m, 5H), 3.87 (s, 3H),
4.56 (d, J = 10.7 Hz, 1H), 4.63 (d, J = 11 Hz, 1H),
4.83–5.00 (m, 10H), 5.26 (s, 2H), 6.91 (d, 2H), 7.28–7.45
(m, 27H); 13C NMR (CDCl3) d 55.43, 71.82, 72.15, 75.85,
76.17, 77.91, 78.22, 81.61, 82.75, 113.96, 127.77, 127.89,
127.99, 128.17, 128.24, 128.28, 128.49, 128.53, 128.69,
129.36, 130.12, 136.58,138.02, 138.68, 138.74, 157.20, 159.46.
1-D-2-Deoxy-2-N-benzyloxycarbonylamino-3,4,5,6-tetra-O-
benzyl-myo-inositol (16). Compound 15 (200 mg, 0.25 mmol)
was dissolved in acetonitrile–water (4 : 1, v/v, 25 mL). To this
solution, CAN (690 mg, 1.26 mmol) was added in portions at
room temperature, and the mixture stirred for 30 min. When
TLC showed complete consumption of the starting material,
the mixture was concentrated, the residue redissolved in
chloroform and the organic extract was washed with brine
several times. Organic extract was separated, dried and
concentrated. Purification by silica gel chromatography
(hexane–ethyl acetate, 2 : 1 - 1 : 1) afforded pure compound
16 (144 mg, 85%) as a white solid. Rf 0.25 (hexane–ethyl
acetate, 2 : 1). 1H NMR (CDCl3) d 3.55–3.71 (m, 5H), 4.55
(m, 1H), 4.64–4.93 (m, 8H), 5.17 (s, 2H), 5.27 (d, 1H),
7.24–7.42 (m, 25H).
1-D-1-O-(4-Methoxybenzyl)-3,4,5,6-tetra-O-benzyl-scyllo-
inosose (12). A solution of alcohol 11 (0.5 g, 0.75 mmol) in
anhydrous DMSO (10 mL) was added dropwise with acetic
anhydride (1 mL, 10.6 mmol) under nitrogen and stirred at
room temperature for about 14 h. The reaction mixture was
added stepwise to the stirred aqueous solution of sodium
bicarbonate (25 g in 125 mL water) during 1 h and then stirred
for 2 more hours. The precipitate formed was filtered off and
washed with water. The recrystallization from hexane–ethyl
acetate (1 : 1) afforded pure 2-inosose 12 (0.435 g, 87%).
Rf 0.48 (hexane–ethyl acetate, 2 : 1). 1H NMR (CDCl3) d
3.61–3.65 (m, 2H), 3.82 (s, 3H), 3.87 (t, J = 9 Hz, 1H),
4.12–4.17 (m, 2H), 4.48 (d, J = 12 Hz, 1H), 4.52 (d, J = 12
Hz, 1H), 4.75–4.94 (m, 8H), 6.86 (d, 2H), 7.28–7.39 (m, 22 H);
13C NMR (CDCl3) d 53.16, 72.98, 73.29, 75.83, 75.87, 75.98,
82.16, 83.54, 83.85, 113.81, 127.47, 127.64, 127.68, 127.88,
127.93, 128.31, 128.43, 129.56, 129.86, 137.55, 138.32, 138.37,
138.42, 159.54, 202.34.
1,2-Di-O-hexyl-sn-glycero-3-phospho)-D-2-deoxy-2-benzyloxy-
carbonylamino-3,4,5,6-tetra-O-benzyl-myo-inositol (18). Into
the solution of alcohol 16 (144 mg, 0.21 mmol) in chloroform
(2 mL) was added via syringe P-chloro-N,N-diisopropyl-O-
methyl-phosphoramidite (0.25 mmol) at room temperature,
under nitrogen. The reaction was stirred for 24 h, solvents
were removed under vacuum, and the residue dried in vacuo
for a few more hours. Tetrazole (1 mmol) and 1,2-di-O-hexyl-
sn-glycerol 17 (86 mg, 0.33 mmol) (see Supporting Informa-
tion for description of synthesis of 17z) were added and a
minimum volume of anhydrous acetonitrile was transferred
into the reaction vessel via distillation under vacuum. The
suspension obtained was stirred for 24 h at room temperature
1-D-1-O-(4-Methoxybenzyl)-2-deoxy-2-benzyloxycarbonyl-
amino-3,4,5,6-tetra-O-benzyl-myo-inositol (15). The inosose 12
(250 mg, 0.38 mmol) was dispersed in anhydrous THF (5 mL)
and stirred until the ketone dissolved completely. Ammonium
acetate (293 mg, 3.8 mmol) was added and the mixture stirred
for a few minutes until solution was clear. A solution of
sodium cyanoborohydride (27.6 mg, 0.42 mmol) was added
dropwise at room temperature and the mixture stirred for 2 h.
The progress of the reaction was monitored by TLC
(hexane–ethyl acetate, 2 : 1). The reaction mixture was
concentrated, the residue redissolved in ethyl acetate, and
washed with sodium bicarbonate and brine. The organic phase
was separated, dried over sodium sulfate and concentrated.
under nitrogen, monitoring the progress of the reaction by 31
P
NMR. The solution of crude phosphite triester in dry
chloroform, was cooled to ꢁ10 1C and tert-butyl hydro-
peroxide in decane (0.5 mmol) was added stepwise via a
syringe. The reaction mixture was warmed up to room
temperature and aqueous sodium bicarbonate solution was
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 925–933 | 931

View Article Online
added. The organic extract was separated, washed with brine
and dried over sodium sulfate, filtered and concentrated. The
crude product was purified on silica gel (hexane–ethyl acetate,
4 : 1 - 1 : 1) to give pure 18 as a mixture of two diastereomers
(100 mg, 72%) in a form of a colorless oil. Rf 0.20 (hexane–
as an eluent gave the phosphonate diester 21 (550 mg, 36%) as
1
a colorless oil. H NMR (300 MHz, CDCl3): d 0.10 (s, 6H,
SiMe2), 0.89 (m, 3H, CH3), 0.90 (s, 9H, SiBut), 1.26 (s, 28H,
CH2), 1.34 (s, 3H, CH3), 1.37 (s, 3H, CH3), 1.40 (s, 3H, CH3),
1.46 (s, 3H, CH3), 2.33 (m, 1H, PCH), 2.50–3.20 (m, 3H, CH,
CH2CO2Me), 3.26 (dd, 1H, J = 4.9 Hz, CH), 3.66 (s, 3H,
OCH3), 3.60–3.80 (m, 5H, P(OMe), P(OCH2R), 3.91 (dd, 1H,
J = 6.4 Hz, CH), 4.03 (m, 2H, CH), 4.64 (m, 1H, CH).
31P NMR (121 MHz, CDCl3) d 33.76 ppm.
1
ethyl acetate, 2 : 1); H NMR (CDCl3) d 0.89 (m, 12H), 1.31
(m, 24H), 1.56 (m, 8H), 3.25–3.78 (m, 24H), 4.02–4.06
(m, 4H), 4.51 (4.55 (m, 2H), 4.74–4.92 (m, 16H), 5.17 (s, 2H),
7.21–7.37 (m, 50H); 31P NMR (CDCl3) d 1.25, 1.45 (1 : 1).
Formation of the alcohol 22. A solution of the phosphonate
21 (530 mg, 0.7 mmol) was treated with trimethylamine (5 mL)
analogously as described above. Evaporation of trimethyl-
amine gave the demethylated product quantitatively, which
was further dissolved in the solution of tetra-n-butylammonium
fluoride in THF (1 M, 5 mL) and stored at room temperature
for 2 h. Evaporation of the solvent followed by chromato-
graphy on silica gel using chloroform–methanol-ammonium
hydroxide (40 : 10 : 1, v/v) as an eluent gave the phosphonate
monoester 22 (300 mg, 65%) as a colorless oil. This compound
was used for further steps without additional purification as
described below.
1,2-Di-O-hexyl-sn-3-glycero-3-phospho-D-2-deoxy-2-amino-myo-
inositol (8). The triester 18 (100 mg, 0.098 mmol) was
demethylated with anhydrous trimethylamine during 48 h at
50 1C to afford pure product, which was then subjected to
hydrogenolysis with palladium on charcoal in dry methanol.
After 24 h the reaction was complete, the mixture was filtered
through Celite and the solution concentrated to dryness. The
residue was redissolved in water, and the solution was passed
through a reverse phase column. The product was eluted off
with methanol and concentrated. The residue was dissolved in
water and converted into an ammonium salt using a cation
exchange column. Freeze-drying gave the pure compound 8
(38 mg, 70%) as a white powder. 1H NMR (CDCl3:CD3OD:D2O,
3 : 3 : 1) d 0.8 (m, 6H), 1.20–1.30 (m, 12H), 1.48 (m, 4H), 3.23
(m, 2H), 3.39–3.45 (m, 5H), 3.50–3.68 (m, 6H), 3.82 (t, J =
4.5, 1H), 3.90 (m, 2H), 4.09 (m, 1H); 31P NMR (D2O) d 0.70;
ES-MS (m/z) 501.
Formation of the d-lactone 10. A mixture of the demethylated
phosphonate 22 (65 mg, 0.1 mmol) and p-toluenesulfonic acid
monohydrate (5 mg) in methanol (2 mL) was heated at 60 1C
for 2 days. Evaporation of all solvents followed by chromato-
graphy on silica gel eluting with chloroform–methanol–water-
ammonium hydroxide (60 : 40 : 5 : 1, v/v) gave the bicyclic
phosphonate 10 (27 mg, 50%) as an amorphous colorless
solid. 1H NMR (500 MHz, CD3OD): d 4.25 (dd, J = 9.3,
11.2 Hz, H-6, 1H), 4.19 (tr, J 3.0 Hz, H-2, 1H), 3.89 (m, OCH2,
2H), 3.65 (tr, J = 10.6 Hz, H-4, 1H), 3.49 (tr, J = 9.2 Hz, H-5,
1H), 3.47 (dd, 3.0 Hz, 9.8 Hz, H-3, 1H), 3.22 (m, 1H),
2.76–2.87 (m, 2H), 2.52 (m, 1H), 1.97 (m, 1H), 1.62 (m, 2H),
1.31 (m, 26H), 0.89 (tr, 3H). 31P NMR (121 MHz, D2O): d 25.2
ppm. HR ESMS: Calcd. for C25H46O9P: 521.2879; Found:
521.2869.
Alkylation of phosphonate 19. Into a solution of phospho-
nate 19 (3.36 g, 7 mmol) in anhydrous THF (30 mL) was
added dropwise a solution of 1.7 M t-BuLi in pentane
(4.94 mL, 8.4 mmol, 1.2 equiv.). The mixture was stirred at
ꢁ78 1C for 1.5 h methyl chloroacetate (0.920 mL, 10.5 mmol,
1.5 equiv.) and the temperature was allowed to warm up to
20 1C. The reaction mixture was diluted with chloroform
(300 mL), washed with brine, and dried over Na2SO4.
Evaporation of all solvents followed by chromatography on
silica gel using hexane-acetone (1 : 1, v/v) as an eluting solvent
gave the product 20 (3.55 g, 92%) as a colorless oil. 1H NMR
(300 MHz, CDCl3): d 0.09 (s, 6H, SiMe2), 0.90 (s, 9H, SiBut),
1.31 (s, 3H, CH3), 1.32 (s, 3H, CH3), 1.34 (s, 3H, CH3), 1.47
(s, 3H, CH3), 2.31 (m, 1H, PCH), 2.50–3.10 (m, 3H, CH,
CH2CO2Me), 3.21 (dd, 1H, J = 4.9 Hz, CH), 3.63 (s, 3H,
OCH3), 3.60–3.80 (m, 6H, P(OMe)2), 3.90 (dd, 1H, J = 6.4
Hz, CH), 4.55 (dd, 1H, J = 4.3 Hz, CH). 31P NMR (121 MHz,
CDCl3): d 35.40 ppm.
Formation of lactone 23. Lactone 23 was obtained from
1
compound 20 analogously as described for compound 21. H
NMR (400 MHz, Py-d5) d 5.05 (d, J = 9.4, 11.0 Hz, H-6, 1H),
4.82 (brs, H-2, 1H), 4.65 (tr, J = 8.9 Hz, H-4, 1H), 4.23
(tr, J = 9.0 Hz, H-5, 1H), 4.14 (dd, J = 2.7, 9.5 Hz, H-3, 1H),
3.68 (d, J = 9.1 Hz, 3H), 3.65 (d, J = 8.5 Hz, 3H), 3.50 (dddd,
J = 2.3, 8.8, 11.0, 17.0 Hz, Hb, 1H), 3.26 (ddd, J = 10.7 Hz,
16.6 Hz, 26.7 Hz, Haa, 1H), 3.07 (dtr, J = 2.5 Hz, 16.1 Hz,
Hae, 1H), 2.61 (dddd, J = 2.2, 9.0, 11.2, 19.0 Hz, H-1, 1H). 13
C
Exchange of the ester grop in the phosphonate 20. The
solution of the phosphonate 20 (1.1 g, 2 mmol) in trimethyl-
amine (5 mL) was heated in a sealed ampoule at 80 1C for 4
days. Evaporation of Me3N gave the demethylated product
quantitatively as indicated by 31P NMR, which was used
directly in the next step. The mixture of the demethylated
product from above, 1-(mesilytene-2-sulfonyl)-3-nitro-1,2,4-
triazole (1.18 g, 4 mmol, 2 equiv.) and 1-hexadecanol
(720 mg, 3 mmol, 1.5 equiv.) in anhydrous pyridine (5 mL)
was stirred at room temperature for 24 h. The reaction mixture
was diluted with chloroform (300 mL), washed with brine, and
dried over Na2SO4. Evaporation of all solvents followed by
chromatography on silica gel using hexanes-acetone (2 : 1, v/v)
NMR (Py-d5) d 171.6, 171.5, 80.65, 80.53, 76.2, 75.8, 74.4,
71.3, 53.2, 53.1, 53.0., 39.4, 30.2, 29.7, 29.6, 28.2. 31P NMR
(Py-d5) d 36.0.
Acknowledgements
This work was supported by the grant from NIGMS
GM 57568.
References
1 S. G. Rhee, Annu. Rev. Biochem., 2001, 70, 281.
2 M. J. Rebecchi and S. N. Pentyala, Physiol. Rev., 2000, 80, 1291.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
932 | New J. Chem., 2010, 34, 925–933

View Article Online
3 P. G. Suh, J. I. Park, L. Manzoli, L. Cocco, J. C. Peak, M. Katan,
K. Fukami, T. Kataoka, S. Yun and S. H. Ryu, BMB Rep., 2008,
41, 415.
18 M. S. Shashidhar, J. J. Volwerk, J. F. W. Keana and O. H.
Griffith, Biochim. Biophys. Acta, Lipids Lipid Metab., 1990, 1042,
410.
4 R. J. Kubiak, X. Yue, C. Mihai, R. J. Hondal, M.-D. Tsai and
K. S. Bruzik, Biochemistry, 2001, 40, 5422.
5 O. H. Griffith and M. Ryan, Biochim. Biophys. Acta, Mol. Cell
Biol. Lipids, 1999, 1441, 237.
6 D. W. Heinz, L.-O. Essen and R. L. Williams, J. Mol. Biol., 1998,
275, 635.
7 C. S. Gassler, M. Ryan, T. Liu, O. H. Griffith and D. W. Heinz,
¨
Biochemistry, 1997, 36, 12802.
8 L. O. Essen, L. Perisic, M. Katan, Y. Wu, M. F. Roberts and
R. L. Williams, Biochemistry, 1997, 36, 1704.
9 D. W. Heinz, M. Ryan, T. L. Bullock and O. H. Griffith, EMBO
J., 1995, 14, 3855.
10 H. Wehbi, J. Feng and M. F. Roberts, Biochim. Biophys. Acta,
Biomembr., 2003, 1613, 15.
11 C. Mihai, A. V. Kravchuk, M.-D. Tsai and K. S. Bruzik, J. Am.
Chem. Soc., 2003, 125, 3236.
12 Y. Wu, O. Perisic, R. L. Williams, M. Katan and M. F. Roberts,
Biochemistry, 1997, 36, 11223.
19 (a) M. Ryan, M. P. Smith, T. K. Vinod, W. L. Lau, J. F. W. Keana
and O. H. Griffith, J. Med. Chem., 1996, 39, 4366; (b) H. Zhang,
Y. Xu, Z. Zhang, E. R. Liman and G. D. Prestwich, J. Am. Chem.
Soc., 2006, 128, 5642.
20 (a) T. C. Bruice and S. J. Benkovic, Biochemistry, 2000, 39, 6267;
(b) T. C. Bruice and F. C. Lightstone, Acc. Chem. Res., 1999, 32,
127.
21 J. J. Kulagowski, Tetrahedron Lett., 1989, 30, 3869.
22 N. Katagiri, K. Itakura and S. A. Narang, J. Am. Chem. Soc.,
1975, 97, 7332.
23 C. Mihai, J. Mataka, S. R. Riddle, M.-D. Tsai and K. S. Bruzik,
Bioorg. Med. Chem. Lett., 1997, 7, 1235.
24 M. Dixon and E. C. Webb, Enzymes., Academic Press, London,
1979, pp. 47–138; 332–381.
25 Y. Iwasaki, Y. Tsubouchi, A. Ichihashi, H. Nakano,
T. Kobayashi, H. Ikezawa and T. Yamane, Biochim. Biophys.
Acta, Lipids Lipid Metab., 1998, 1391, 52.
26 C. Bai, L. Zhao, M. Rebecchi, M.-D. Tsai and K. S. Bruzik, J. Am.
Chem. Soc., 2009, 131, 8362.
13 Y. Liu, C. Mihai, R. J. Kubiak, M. Rebecchi and K. S. Bruzik,
Chem. Bio. Chem., 2007, 8, 1430–9.
27 J. Gigg, R. Gigg, S. Payne and R. Conant, J. Chem. Soc., Perkin
Trans. 1, 1987, 2411.
14 H.-X. Zhai, P.-S. Lei, J. C. Morris, K. Mensa-Wilmot and
T. Y. Shen, Tetrahedron Lett., 1995, 36, 7403.
28 L. Zhao, Y. Liu, K. S. Bruzik and M.-D. Tsai, J. Am. Chem. Soc.,
2003, 125, 22.
29 H. S. Hendrickson, E. K. Hendrickson, J. L. Johnson, T. H. Khan
and H. J. Chial, Biochemistry, 1992, 31, 12169–12172.
30 K. S. Bruzik, A. M. Morocho, D.-Y. Jhon, S. G. Rhee and
M.-D. Tsai, Biochemistry, 1992, 31, 5183.
15 S. P. Seitz, R. F. Kaltenbach, R. H. Vreekamp, J. C. Calabreses
and F. W. Perrella, Bioorg. Med. Chem. Lett., 1992, 2, 171.
16 S. F. Martin and A. S. Wagman, J. Org. Chem., 1996, 61, 8016.
17 V. R. Garigapati and M. F. Roberts, Tetrahedron Lett., 1993, 34,
5579; K. A. Lewis, V. R. Garigapati, C. Zhou and M. F. Roberts,
Biochemistry, 1993, 32, 8836.
ꢀc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
New J. Chem., 2010, 34, 925–933 | 933