Synthesis, structure-activity relationships, and the effect of polyethylene glycol on inhibitors of phosphatidylinositol-specific phospholipase C from Bacillus cereus

Article abstract of DOI:10.1021/jm960434y

Substrate analog inhibitors of Bacillus cereus phosphatidylinositol- specific phospholipase C (PI-PLC) were synthesized and screened for their suitability to map the active site region of the enzyme by protein crystallography. Analogs of the natural subst

Full text of DOI:10.1021/jm960434y

4366
J . Med. Chem. 1996, 39, 4366-4376
Articles
Syn th esis, Str u ctu r e-Activity Rela tion sh ip s, a n d th e Effect of P olyeth ylen e
Glycol on In h ibitor s of P h osp h a tid ylin ositol-Sp ecific P h osp h olip a se C fr om
Ba cillu s cer eu s
Margret Ryan, Miles P. Smith, Thottumkara K. Vinod, Wai Leung Lau, J ohn F. W. Keana,* and
O. Hayes Griffith*
Department of Chemistry and Institute of Molecular Biology, University of Oregon, Eugene, Oregon 97403-1229
Received J une 12, 1996X
Substrate analog inhibitors of Bacillus cereus phosphatidylinositol-specific phospholipase C
(PI-PLC) were synthesized and screened for their suitability to map the active site region of
the enzyme by protein crystallography. Analogs of the natural substrate phosphatidylinositol
(PI) were designed to examine the importance of the lipid portion and the inositol phosphate
head group for binding to the enzyme. The synthetic compounds contained pentyl, hexyl, or
hexanoyl and octyl lipid chains at the sn-1 and sn-2 positions of the glycerol backbone and
phosphonoinositol, phosphonic acid, methyl phosphonate, phosphatidic acid, or methyl
phosphate at the sn-3 position. The most hydrophobic compound, dioctyl methyl phosphate
14, was also the best inhibitor with an IC50 of 12 µM. In a series of dihexyl lipids, compounds
with phosphonoinositol head groups inhibited more strongly than those that do not contain
inositol but are otherwise identical. Compound 29, a short-chain lipid with a phosphonoinositol
head group, was found to be a competitive inhibitor and the most potent in this series with an
IC50 of 18 µM (Ki )14 µM). Analogs with dihexyl chains were better inhibitors than those
with dihexanoyl chains, presumably because the ether-linked lipids are more hydrophobic than
the ester-linked lipids. No appreciable difference in inhibition was found between a phospho-
noinositol lipid and the corresponding difluorophosphonoinositol lipid. Inositols and inositol
derivatives that do not contain lipid moieties show IC50s about 3 orders of magnitude above
those of the short-chain lipids. In this group, glucosaminyl(R1f6)-D-myo-inositol inhibited more
strongly than myo-inositol, which in turn is a better inhibitor than inositol phosphate. The
addition of polyethylene glycol (PEG-600) resulted in a marked decrease in inhibition by the
short-chain lipids, but had little effect on the water-soluble head group analogs. This is
accounted for in terms of solubilization of the amphipathic inhibitors by PEG. Since PEG is
required in the crystallization, these data indicate that the best strategy for obtaining enzyme
inhibitor complexes is to start by cocrystallizing PI-PLC with the head group analogs. The
next step is to synthetically add the shortest possible hydrophobic moieties to the analogs and
cocrystallize these with the enzyme. This strategy may be applicable to other lipolytic enzymes.
In tr od u ction
of a mammalian PI-PLC.7 Bacterial PI-PLCs do not
require Ca2+ for activity and unlike mammalian PI-
PLCs cleave the glycosylphosphatidylinositol (GPI)
anchor of GPI-anchored proteins8 in addition to phos-
phatidylinositol (PI). The GPI-anchor cleaving activity
is of considerable medical interest as diagnostic tool for
the analysis of proteins presented on the outer face of
cell membranes. These GPI-anchored proteins include
activation antigens of the immune system, scrapie prion
protein, and the carcinoembryonic antigen, a human
tumor marker.9-11
Signaling molecules such as growth factors, neu-
rotransmitters, and mitogens are received on the cell
surface and, via heterotrimeric G proteins or tyrosine
kinases, lead to the activation of phosphatidylinositol-
specific phospholipase C (PI-PLC) on the internal face
of the cell membrane. PI-PLC produces the two second
messengers myo-inositol 1,4,5-trisphosphate and dia-
cylglycerol (DAG). These two molecules initiate signal-
ing cascades that result in profound cellular changes
such as cell proliferation and neuronal activity.1,2 Links
to cancer and Alzheimer’s disease are being inves-
tigated.3-5
Mammalian PI-PLCs are large, Ca2+ dependent en-
zymes that contain one catalytic and several regulatory
domains. The much smaller bacterial PI-PLCs consist
only of the catalytic domain, and the crystal structure
of Bacillus cereus PI-PLC6 shows the same architecture
as the catalytic domain of the recently solved structure
B. cereus PI-PLC is stereospecific for the D-myo-
inositol enantiomer of its PI or GPI substrates. The
corresponding L-enantiomers have been shown to be
neither substrates nor inhibitors, which is fortunate
since it is, therefore, possible to use racemic synthetic
substrates.12-14 The enzyme is insensitive to the ster-
eochemistry around the sn-2 carbon of the lipid portion
of the substrate molecule.8,15 Variations in the lipid
moiety are tolerated, e.g., ether- or ester-linked lipids,
mono- or diacyl lipids, or ceramides, and are usually
X Abstract published in Advance ACS Abstracts, October 1, 1996.
S0022-2623(96)00434-7 CCC: $12.00 © 1996 American Chemical Society

Inhibitors of B. cereus PI-PLC
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22 4367
Sch em e 1a
a
Reagents: (a) NaH/THF, BnBr; (b) HOAc/MeOH/H2O; (c) RBr,
KOH/toluene; (d) H2, 10% Pd/C, MeOH.
monomeric solubility in water and compare the inhibi-
tion by these amphiphiles to that of a series of hydro-
philic compounds resembling the head group of the
natural substrate, PI. Since the crystallization of B.
cereus PI-PLC is carried out in the presence of polyeth-
ylene glycol (PEG), the question of effects caused by this
polymer is also addressed.
Resu lts a n d Discu ssion
Syn th esis. The commercially available racemic ketal
4 was selected as a starting point for the synthesis of
the dialkylglycerols (Scheme 1). The known diol 5 is
readily available from solketal 4 in high yield.27,28 The
diol 5 was readily alkylated with a variety of alkyl
bromides using a modification of previously reported
conditions29 to afford the diethers in >90% yield. Hy-
drogenolysis27 produced the desired di-O-alkylglycerols
6, 7, and 8 in 60-80% overall yield from 4. While the
synthesis of 729 and 830,31 by related procedures has been
reported previously, the full characterization of these
compounds is reported here for the first time.
We next undertook the phosphorylation of the short-
chain dialkylglycerols. The importance of phospholipids
in physiological processes, in particular as mediators in
signal transduction, has stimulated the development of
methods for the efficient chemical synthesis of structur-
ally well defined phospholipids and their analogs as
enzyme inhibitors and drug candidates.32 These routes
include the phosphorylation of alcohols using the phos-
phoramidites originally developed for oligonucleotide
chemistry33,34 and the use of activating reagents such
as 2,4,6-triisopropylbenzenesulfonyl chloride (TPS).14,35
These routes are somewhat limited by the availability
of the required intermediates. To circumvent these
problems the chlorophosphate route described by Eibl
and Woolley36 for the synthesis of long-chain diradyl-
glycerol phospholipids was explored. However, in our
hands this method did not produce satisfactory yields.
The decreased steric bulk of the short-chain dialkyl
glycerols employed in this study apparently greatly
diminishes the reactivity differences between the suc-
cessive electrophiles OPCl3 and OPCl2(OR). Similarly
the use of methyl dichlorophosphate37,38 failed to afford
the desired mixed phosphorus triesters in acceptable
yields.
F igu r e 1. Reaction catalyzed by B. cereus PI-PLC. The
substrate PI is cleaved by the enzyme’s phosphotransferase
activity into two components, diacylglycerol (DAG) and D-myo-
inositol 1,2-cyclic phosphate (Ins(1:2-cyc)P). This cyclic phos-
phate is gradually hydrolyzed to myo-inositol 1-phosphate in
a second reaction by PI-PLC. Mammalian PI-PLCs also cleave
PI, but preferentially in a form phosphorylated at the 4 and 5
positions of myo-inositol.
equally good substrates.16 B. cereus PI-PLC possesses
phosphotransferase as well as phosphodiesterase activi-
ties, as shown by 31P NMR.12 Acting on the natural
substrate PI, the phosphotransferase produces lipid-
soluble DAG and water-soluble D-myo-inositol 1,2-cyclic
phosphate (Ins(1:2-cyc)P) (Figure 1). In a second but
much slower reaction, the phosphodiesterase activity of
B. cereus PI-PLC opens the cyclic phosphate ring to form
D-myo-inositol 1-phosphate (Ins(1)P). The reaction from
PI to inositol 1-phosphate proceeds with overall reten-
tion of configuration at phosphorus.15
With the PI-PLC crystal structures now available,6,7
it is important to design inhibitors for cocrystallization.
Our goal here is to provide insight into the choice of
inhibitors to map the active-site region of B. cereus PI-
PLC, an important step toward arriving at a molecular
mechanism. There are a number of reports of PI-PLC
inhibition in the literature,5,8,17-25 but because of vary-
ing assay conditions, the results are often difficult to
compare. Most of these studies use assays based on the
naturally occurring phospholipid substrate, PI, in a
lipid-like environment, e.g., in detergent micelles or
phospholipid vesicles. This strategy is appropriate for
simulating physiological conditions, where the reference
state is the lipid interface environment. However, to
screen inhibitors for cocrystallization we have chosen a
single-phase aqueous assay system to eliminate any
secondary effects on the enzyme and added inhibitors
which may be introduced by the presence of an interface.
In this assay system the synthetic substrate myo-
inositol 1-(4-nitrophenyl phosphate) (NPIP)26 is used.
With this approach, a panel of inhibitors can be com-
pared with an aqueous environment as the reference
state so that the results are applicable to the crystal-
lization conditions. Here we report the synthesis of a
series of short-chain phospholipids which have high
A modification of a procedure recently reported for
the synthesis of diacylglyceryl phosphates39 allowed the
isolation of the desired mixed triester 10 as a mixture
of diastereomers in 64% yield along with the corre-
sponding symmetric diglyceryl phosphate in 31% yield.
Extending these phosphorylation conditions to 6 and 8
gave the corresponding triesters (9 and 11) in 48 and
68% isolated yields, respectively (see Scheme 2).
Hydrogenolysis of the benzyl protecting group in 9-11
furnished the short-chain methyl phosphates 12-14 as

4368 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22
Ryan et al.
Sch em e 2a
Sch em e 3a
a
Reagents: (a) POCl3, Et3N/Et2O, then BnOH, Et3N/Et2O; (b)
H2, 10% Pd/C, MeOH; (c) Dowex 50W-X8 (H+ form), MeOH; (d)
hexanoyl chloride, i-Pr2NEt/CH2Cl2.
a
Reagents: (a) MeOPCl2, i-Pr2NEt/CH2Cl2, then BnOH, i-Pr2
NEt/CH2Cl2, then t-BuOOH/isooctane; (b) H2, 10% Pd/C, MeOH;
(c) Dowex 50W-X8 (H+ form), MeOH; (d) hexanoyl chloride,
i-Pr2NEt/CH2Cl2.
Sch em e 4a
the free acids. Compound 12 could be readily purified
by acid/base extraction and was isolated as the sodium
salt. Methyl phosphates 13 and 14 could not be purified
by this protocol and required column chromatography
to afford 13 and 14, isolated as the ammonium salts.
Dihexanoyl methyl phosphate 18 was prepared via
an indirect route employing solketal 4 as the starting
alcohol. This route avoids the potential complications
arising from the 2 f 3 acyl migration in diacyl glycer-
ols.36 The phosphorylation of 4, under the conditions
reported above, afforded triester 15 in 48% yield. The
isopropylidene protecting group was removed by the
method of Heeb and Nambiar40 to produce diol 16 in
53% yield after chromatographic isolation. Acylation of
the hydroxyls in 16 afforded the fully protected 17 in
good overall yield. On the basis of 13C NMR analysis
no 3 f 2 phosphoryl migration41 under the conditions
employed for the acylation reaction could be detected.
The benzyl protecting group in 17 was then removed
by hydrogenolysis, followed by ion exchange to afford
the sodium salt of 18 in quantitative yield.
The dihexyl and dihexanoyl phosphatidic acid esters
20 and 24 were prepared as shown in Scheme 3. The
intermediate dibenzyl triesters 19 and 21 were readily
available in high yields (87 and 89%, respectively) using
an excess (1.5 equiv) of phosphorus oxychloride followed
by an excess of benzyl alcohol. By employing the
conditions discussed above, the isopropylidene triester
21 was readily converted to the corresponding dihex-
anoyl analog 23. Cleavage of the phosphate protecting
groups in 19 and 23 afforded the acids 20 and 24 in
nearly quantitative yields.
a
Reagents: (a) NaI, MEK; (b) H2, 20% Pd(OH)2/C, MeOH; (c)
AcOH/water, 1:1.
Cr itica l Micelle Con cen tr a tion (cm c). The short-
chain lipid inhibitors are surfactants which will spon-
taneously aggregate into micelles above their cmc’s. In
order to design the inhibition experiments so that these
molecules remained in their monomeric state, the cmc’s
were determined for representative inhibitor compounds
under the condition of the enzyme assay. The values
are listed in Tables 1 and 2. The lipid chain length has
the largest effect on the cmc. The dependence of the
cmc on the nature of the lipid head group effect is much
smaller, and the values reported here are similar to
published data on short-chain phosphatidylcholines,44
phosphatidic acids,45 and phosphatidylinositols.14,46 Some
variations are expected since the methods of measure-
ments differ, and the cmc’s of ionic amphiphiles are
known to depend on the buffer concentration and pH.45
Most of the representative compounds examined in
Tables 1 and 2 are ether-linked lipids. In compounds
containing the same number of carbon atoms in their
lipid chains, ester-linked lipids are known to have cmc’s
higher than those of ether-linked lipids.29 Thus, select-
ing concentrations below the cmc’s of the ether-linked
lipids will also be appropriate for the ester-linked lipids.
In our enzyme assays, the concentrations of short-chain
The phosphonate analog 26 of 15 was prepared
(Scheme 4) from the dimethyl phosphonate 2525 by
selective mono-demethylation employing NaI in methyl
ethyl ketone (MEK). The remaining phosphonates
employed in the present study (29-32) were prepared
as previously described.25 The glycosylinositol 28 was
prepared from 27 by a modification of the procedure
previously described.42,43

Inhibitors of B. cereus PI-PLC
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22 4369
Ta ble 1. Inhibition of B. cereus PI-PLC by Amphiphilic
Substrate Analogs with myo-Inositol Phosphate, Phosphonate
or Phosphate Head Groups
Ta ble 2. Inhibition of B. cereus PI-PLC by Amphiphilic
Substrate Analogs with Methyl Phosphate or Methyl
Phosphonate Head Groups
a
All compounds shown are racemic and were used as the
sodium salts (13 and 14 were converted from the ammonium to
the sodium salt as described for 12 in Synthetic Procedures). b IC50
values are the mean (( average deviation) of two separate
experiments using 5 nM PI-PLC and 0.25 mM NPIP substrate in
assay buffer (100 mM HEPES, 1 mM EDTA; pH 7) at 25 °C. c The
d
cmc values were measured in assay buffer at 25 °C. This
preparation was slightly less pure than the preparation used to
determine the IC50
.
phonic acid, and methyl phosphonate, respectively. The
inhibition data of Table 1 show that 29 with the myo-
inositol 1-phosphonate head group has the lowest IC50,
which is to be expected since this molecule most closely
resembles the natural substrate, PI. We studied this
model compound over a broad range of concentrations
and determined that it is a competitive inhibitor with
Ki ) 14 µM. Compounds with inositol head groups were
synthesized as mixtures of diastereomers, and previous
studies on the hydrolysis of the synthetic substrate
NPIP showed that only the D-enantiomer is a substrate
for PI-PLC, the L-enantiomer being neither a substrate
nor an inhibitor.13 Therefore, presumably only the
D-enantiomer of 29 binds to the enzyme and the cor-
rected IC50 and Ki values would be 9 and 7 µM,
respectively. Surprisingly, 26 with the methyl group
replacing the myo-inositol moiety is also an effective
inhibitor. Removal of the methyl group as in the free
phosphonic acid 31 increases the IC50 to 60 µM. The
same trend, though to a smaller extent, is seen with
the methyl phosphate 18 and the free phosphatidic acid
24, two short-chain phospholipids with ester linkages.
The corresponding lipid with a myo-inositol head group
was omitted from this series, because it is expected to
a
All compounds shown are racemic and were used as the free
b
acids. IC50 is the concentration of inhibitor which causes a 50%
decrease in B. cereus PI-PLC activity and IC50 values are the mean
(( average deviation) of two separate experiments using 5 nM PI-
PLC and 0.25 mM NPIP substrate in assay buffer (100 mM
HEPES, 1 mM EDTA; pH 7) at 25 °C. The Ki is the mean ((
average deviation) of two determinations in assay buffer but
varying the NPIP concentration from 0.1 mM to 0.5 mM. c The
cmc values given for representative compounds were measured
d
in assay buffer at 25 °C. The cmc values were extrapolated as
described in the Experimental Section, part II.
lipid inhibitors were maintained at least 1 order of
magnitude below the cmc’s, ensuring the inhibitors were
present predominantly as monomers and not as micellar
aggregates.
Str u ctu r e-Activity Rela tion sh ip s. Our strategy
was to synthesize a series of short-chain lipid inhibitors
in which there are subsets with only one structural
difference. For example, the three inhibitors 29, 31, and
26 are all phosphonates, i.e., the oxygen of the scissile
oxygen-phosphorus bond has been replaced by CH2. All
three have in common the six carbon, ether-linked side
chains. The only difference in these three molecules is
the head group: myo-inositol-1-phosphonate, free phos-

4370 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22
Ryan et al.
Ta ble 3. Inhibition of B. cereus PI-PLC by Inositols and
be a substrate rather than an inhibitor. In a third
subset, consisting of the ether-linked phospholipids 20
and 13, the difference between phosphate and methyl
phosphonate head groups is not significant. Taken
together, these data suggest that a component of the
binding is due to an electrostatic interaction in the
region of the phosphorous atom and that the optimal
charge on this series of inhibitors is -1. A fraction of
the free acids, under the conditions examined here,
would be anions with a net charge of -2, which could
account for the reduced inhibition compared to the
corresponding methyl derivatives which are expected to
have a net head group charge of -1 at neutral pH.47
Related Compounds and Inositol Phosphates
Another interesting trend is the effect of the length
of the hydrocarbon chains on inhibition. The three
ether-linked methyl phosphates, 12-14, differ only in
alkyl chain length, having both chains five, six, or eight
carbons long, respectively. Inhibition by these com-
pounds increases with the length of the lipid chains. The
most potent inhibitor in Table 2 is the dioctyl lipid 14,
its IC50 of 12 µM is 20-fold lower than the IC50 for the
dipentyl lipid 12. These data underscore the importance
of hydrophobic interactions in the active site region,
identified in the crystal structure of B. cereus PI-PLC.6
This trend is consistent with the results obtained for
short-chain PI substrates of mammalian PI-PLCδ146 and
the substrates of other phospholipases,32,48 where the
activity is chain length dependent. Lipids with ether
linkages are better inhibitors than the corresponding
short-chain lipids with ester linkages (e.g., compare 20
with 24; 26 with 18), probably because the ether linkage
conveys greater hydrophobicity.
Two modifications were designed to test for the
importance of hydrogen bonding to the oxygen of the
O-P bond cleaved by PI-PLC. One is the substitution
of fluorine atoms for hydrogen atoms in the phospho-
nates (i.e., the pairs 29, 30 and 31, 32). The second is
the substitution of CH2 for the oxygen, i.e., the pairs
20, 31 and 13, 26. As can be seen from Tables 1 and 2,
the differences in inhibition within these pairs are not
significant within experimental error. We conclude that
any loss of hydrogen-bonding potential when the oxygen
atom of the scissile oxygen-phosphorus bond is replaced
by CH2 does not result in a large loss in binding. This
is advantageous, as it indicates that phosphonate
derivatives of PI are good routes to inhibitors of PI-
PLCs. There appears to be little benefit in introducing
CF2 in place of CH2.
a
The inositol phosphates were used as the cyclohexylammo-
b
nium salts. IC50 values are the mean (( average deviation) of
two separate experiments using 5 nM PI-PLC and 0.25 mM NPIP
substrate in assay buffer (100 mM HEPES, 1 mM EDTA; pH 7)
at 25 °C. c The IC50 value was extrapolated from nonlinear
regression analysis of the data with the highest concentration of
scyllo-inositol, 85 mM, resulting in 78% of the uninhibited activity.
A series of water-soluble molecules resembling the
inositol head group were tested with the same assay as
for the short-chain lipids so that the IC50s would be
directly comparable. The results are shown in Table 3.
With IC50s in the mM range, i.e., up to 3 orders of
magnitude higher than many of the short-chain lipids
of Tables 1 and 2, it would appear that these water-
soluble molecules are of little interest because they are
poor inhibitors. However, for reasons discussed below
(i.e., the effect of PEG), these water-soluble molecules
are important in developing a strategy for cocrystalli-
zation with the enzyme.
50 mM inhibitor resulted in 73% of the uninhibited activity. e The
d
IC50 was extrapolated from nonlinear regression analysis of the
data with the highest inhibitor concentration used, 108 mM,
resulting in 54% of the uninhibited activity.
epi-Inositol and epi-inosose have an axial C-6 hydroxyl
or carbonyl group and are slightly better inhibitors. We
have no explanation for the improved inhibition. We
modified the C-6 position by adding glucosamine (28)
to represent the proximal portion of a GPI anchor. The
inhibition by glucosaminyl(R1f6)-D-myo-inositol (GlcN-
(R1f6)Ins, 28) is 5 times greater than by myo-inositol,
arguing the presence of this portion of the GPI anchor
improves binding to the enzyme. These results are in
agreement with the observation that PI-PLC has a
higher affinity for GPI substrates than for PI, demon-
myo-Inositol is a convenient reference compound, as
it is part of the natural substrate, PI. With an IC50 of
10 mM, it is a weak inhibitor, but nevertheless is able
to displace the solvent molecules as cocrystals of PI-
PLC in complex with myo-inositol have been obtained.

Inhibitors of B. cereus PI-PLC
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22 4371
Ta ble 4. Effect of Polyethylene Glycol on Inhibition of PI-PLC
strated by apparent KM values for GPI substrates that
are at least 2 orders of magnitude lower than the KM
for PI.49
activity remaining (%)a
inhibitor
concn (mM)
bufferb,c
PEG-600c,d
scyllo-Inositol, scyllo-inosose, and 2-C-methylene-myo-
inositol oxide are all poor inhibitors (Table 3). These
molecules have alterations at the axial 2-hydroxyl group
of myo-inositol, which is required for the formation of
the cyclic phosphate ester, Ins(1:2-cyc)P (Figure 1). The
crystal structure of PI-PLC in complex with myo-inositol
shows the 2-hydroxyl at hydrogen-bonding distance to
His32, one of the two histidines participating in cataly-
sis. The loss of this hydrogen bond possibly in combina-
tion with steric hindrance by the altered arrangement
around the 2-position would reduce binding to the
enzyme. The three myo-inositol phosphates in Table 3
are also poor inhibitors; with IC50 values of about 50
mM they are 5 times weaker than myo-inositol. Two
of these compounds, Ins(1:2-cyc)P and Ins(1)P, are
products of reactions of B. cereus PI-PLC, and the third
compound, myo-inositol 2-phosphate (Ins(2)P), is an
isomer of a product. Thus, little or no product inhibition
by the inositol phosphates is found, consistent with Ins-
(1:2-cyc)P acting as a very weak substrate for the
phosphodiesterase activity of PI-PLC.12
none
myo-Ins
myo-Ins(2)P
GlcN(R1f6)Ins (28)
14
29
29
100 ((4)
18 ((0)
69 ((5)
71 ((0)
14 ((1)
25 ((5)
ND
100 ((5)
22 ((1)
56 ((3)
69 ((4)
102 ((6)
100e
40.0
16.0
0.8
0.17
0.25
0.5
102e
a
The assays contained 5 nM PI-PLC and 0.25 mM NPIP
substrate and were performed as described in the Experimental
b
Section. Assays in the standard assay buffer (100 mM HEPES,
1 mM EDTA; pH 7) with no PEG present. c Values given are the
mean (( average deviation) of at least two measurements from
separate experiments. The data are normalized, so that the
activity in the absence of inhibitor is 100% which corresponds to
a turnover rate of about 40 µmol min-1 mg-1 in assay buffer and
to about 90 µmol min-1 mg-1 in the 33% PEG solution. Assays
d
were in 90% precipitant solution, which contains 33% PEG-600,
0.16 M trisodium citrate-HCl, pH 7.35, and 46 mM HEPES, pH
7. This corresponds to 90% rather than 100% precipitant solution
so as to avoid phase separation when inhibitor compound and
substrate were added. e Single measurement at the given concen-
tration. ND, not determined.
ing is an entropic effect, and it does not require any
specific interactions between macromolecules. How-
ever, the effect of PEG on proteins is more complex.
Other factors present include the water sorption by
PEG56 and weak attractive interactions of PEG with
hydrophobic amino acid residues.57
The observation that short-chain methyl phosphates
are inhibitors is consistent with results of a recent study
on interfacial binding of B. cereus PI-PLC to phospho-
lipid vesicles of ditetradecylglycerophosphomethanol
(DTPM).50 When B. cereus PI-PLC was exposed to
DTPM vesicles, DTPM was found to act as a competitive
inhibitor.50 In fact, it was this report that encouraged
us to synthesize and examine the short-chain methyl
phosphates reported here.
In contrast to the extensive literature on the effects
of PEG on protein aggregation and crystallization, we
were unable to locate any studies on the effect of PEG
on enzyme inhibitors. For the present work, we selected
PEG with an average molecular weight of 600, a
relatively small polymer with an average n between 12.5
and 13.9,58 because PEG-600 was used to grow crystals
for the structure determination of B. cereus PI-PLC.6,59
For the purpose of measuring inhibition under crys-
tallization conditions, a single concentration of inhibitor
was added to reactions that simulate the conditions at
which PI-PLC crystals were grown (i.e., close to 100%
precipitant solution). The concentration of PEG present
during the assays was 33% by volume. The results are
summarized in Table 4. The control experiment (top
row) shows that PEG has no inhibitory effect on the
enzyme itself. Likewise, PEG has no effect on the
inhibition of the enzymatic reaction by water-soluble
molecules such as myo-inositol, myo-inositol 2 phos-
phate, or GlcN(R1f6)Ins. However, PEG causes a large
decrease in the observed inhibition by the representative
amphipathic inhibitors 14 and 29, even though these
molecules are present at concentrations that inhibit PI-
PLC strongly in the absence of PEG (Tables 1, 2, and
4). This effect cannot be explained by the simple
macromolecular crowding model. It can, however, be
explained by the solubilizing action of PEG. For ex-
ample, 20% PEG-400 is known to solubilize drugs and
cosmetics that have low water solubility60-62 and to
interact with surfactants.63,64 Thus, the PEG polymer
molecules are excluding the protein and at the same
time PEG is solubilizing, or increasing the aqueous
solubility of, the short-chain amphipathic inhibitors. The
net effect is that the inhibitor-binding equilibrium is
shifted and less inhibitor is bound at the active site of
the enzyme. This has important consequences in se-
lecting a strategy for mapping the active site by means
Turning to the water-soluble inhibitors, we found in
an earlier study of inhibition of hydrolysis of PI in the
detergent deoxycholate that myo-inositol was an inhibi-
tor but that epi-inositol was not.24 Recently, Morris et
al. examined the glycan requirements of GPI-PLC from
Trypanosoma brucei and included a comparison with B.
cereus PI-PLC.21 Their assay was based on cleavage of
GPI-anchored variant surface glycoprotein solubilized
in deoxycholate micelles. With this assay, the glu-
cosamine derivative GlcN(R1f6)Ins exhibited no inhibi-
tion of B. cereus PI-PLC.21 We do not view these results
as inconsistent with the data reported here, because
myo-inositol and GlcN(R1f6)Ins are at best weak
inhibitors. The differences serve primarily to point out
the importance of considering the microenvironment in
the assay when comparing inhibition data, e.g., phos-
pholipid vesicles, detergent micelles, or monomers in an
aqueous environment. For the purpose of identifying
compounds for cocrystallizing with the enzyme, an
aqueous reference state, as used here, is the appropriate
choice since our crystallization conditions for PI-PLC
do not include any lipid aggregates.
Effect of P EG on In h ibition . Polyethylene glycol,
HO(CH2CH2O)nH, is widely used as a precipitating
agent in protein crystallography.51,52 PEG lowers the
solubility of proteins primarily by a volume-exclusion
effect, more recently described as macromolecular crowd-
ing.53,54 This same principle is involved in the enhance-
ment of protein-protein and protein-nucleic acid as-
sociation in the concentrated milieu of living cells, and
PEG is used as an inert “crowding agent” to enhance
these interactions in vitro.53-55 Macromolecular crowd-

4372 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22
Ryan et al.
of inhibitor-enzyme complexes. The molecules exhibit-
ing the highest degree of inhibition in a conventional
assay, i.e., the short-chain lipids, turn out to be the
molecules most likely to be solubilized by PEG and fail
in cocrystallization attempts. The water-soluble mol-
ecules representing the polar head group region of the
substrate, although exhibiting only weak inhibition, are
more likely to cocrystallize with the enzyme in an active
site directed manner. This translates into a synthetic
strategy for designing substrate analog inhibitors: in-
stead of beginning with the lipid moiety of the substrate
as the building block for inhibitors, a better strategy is
to begin with the polar head group (i.e., myo-inositol)
and then to synthetically add such residues as needed
to mimic the essential features of the substrate or its
transition state. This is the strategy we are pursuing
because PEG is required for protein crystallization.59
The first structure was of B. cereus PI-PLC complexed
with myo-inositol.6 The next logical step is to add
glucosamine, since this increases the binding (Tables 3
and 4). Recently we have succeeded in cocrystallizing
GlcN(R1f6)Ins with PI-PLC, which confirms this pre-
diction.65 Finally, the phosphate, and hydrophobic resi-
dues of minimum length will be added in order to
complete the mapping of the active site. This strategy
should be useful in studying other lipophilic enzymes
where the presence of PEG is required for crystalliza-
tion.
the weakly inhibiting water-soluble derivatives are
important. In the series examined, myo-inositol with
an IC50 of 10 mM serves as a useful reference as it has
been successfully cocrystallized with B. cereus PI-PLC.
Adding a phosphate at the 1 or 2 position or the cyclic
1,2-phosphate decreases inhibition by a factor of about
5, whereas adding a glucosamine at the 6 position as in
GlcN(R1f6)Ins increases inhibition by the same amount.
This is consistent with the fact that the phosphotrans-
ferase activity is rapid, whereas the cyclic phosphodi-
esterase activity, which requires the binding of the
water-soluble product Ins(1:2-cyc)P, proceeds very slowly.
The increased inhibition observed for GlcN(R1f6)Ins
is consistent with the GPI-cleaving activity of B. cereus
PI-PLC, where the KM of the GPI substrate < PI
substrate.
Exp er im en ta l Section
I. Syn th etic P r oced u r es. Melting points were obtained
on a Thomas-Hoover apparatus and are uncorrected. Infrared
spectra were recorded in KBr pellets (ca. 0.2%) or as thin films
on NaCl plates on a Nicolet Magna-IR 550. 1H and 13C NMR
spectra were taken on a General Electric QE-300 FT spec-
trometer, and 31P NMR spectra were recorded on a Nicolet NT-
360 spectrometer. All chemical shifts are reported in ppm
relative to residual solvent peaks for 1H and 13C and 1% H3-
PO4 (external reference) for 31P. The identity of all compounds
were confirmed by their NMR and IR spectra as well as TLC
Rf values. Analytical TLC utilized silica gel 60 F254 plates.
Column chromatography was performed using Mallinckrodt
silica gel, Davisil, grade 62 (60-200 mesh). Tetrahydrofuran
(THF) was distilled under dry nitrogen from sodium benzophe-
none. Methylene chloride was distilled from calcium hydride.
All reactions were carried out under a dry nitrogen atmosphere
with anhydrous, freshly distilled solvents, unless otherwise
stated.
(()-3-O-Ben zyl-1,2-d i-O-p en tyl-sn -glycer ol. A suspen-
sion of (()-1-(benzyloxy)propane-2,3-diol (5)29 (10.3 g, 56.5
mmol), 1-bromopentane (34.1 g, 226 mmol), and powdered
KOH (12.7 g, 226 mmol) in toluene (150 mL) was heated at
reflux with a Dean-Stark drying tube attached for 24 h. The
reaction mixture was cooled to 25 °C and washed with water
(2 × 100 mL). The solvents were removed in vacuo, and the
resulting residue was purified by column chromatography (9:1
hexanes/EtOAc) to afford (()-3-O-benzyl-1,2-di-O-pentyl-sn-
glycerol (17.4 g, 95%) as a colorless oil: 1H NMR (CDCl3) δ
7.33 (m, 5 H), 4.56 (s, 2 H), 3.60-3.54 (m, 5 H), 3.43 (t, 4 H),
1.58 (m, 4 H), 1.31 (m, 8 H), 0.89 (t, 6 H); 13C NMR (CDCl3) δ
138.5, 128.2, 127.5, 78.0, 73.3, 71.6, 70.8, 70.6, 70.4, 29.8, 29.4,
28.3, 22.5, 14.0.
(()-3-O-Ben zyl-1,2-d i-O-h exyl-sn -glycer ol. Prepared as
described above, using 1-bromohexane, to afford (()-3-O-
benzyl-1,2-di-O-hexyl-sn-glycerol29 as a colorless oil (91%): 1H
NMR (CDCl3) δ 7.34 (m, 5 H), 4.57 (s, 2 H), 3.65-3.51 (m, 7
H), 3.44 (t, 2 H), 1.56 (m, 4 H), 1.30 (m, 12 H), 0.90 (t, 6 H);
13C NMR (CDCl3) δ 138.5, 128.2, 127.5, 78.0, 73.3, 71.6, 70.8,
70.6, 70.4, 31.7, 30.1, 29.6, 25.8, 22.6, 14.0.
(()-3-O-Ben zyl-1,2-d i-O-octyl-sn -glycer ol. Prepared as
described above, using 1-bromooctane, to afford (()-3-O-benzyl-
1,2-di-O-octyl-sn-glycerol31 as a colorless oil (96%): 1H NMR
(CDCl3) δ 7.34 (m, 5 H), 4.56 (s, 2 H), 3.68-3.35 (m, 9 H), 1.55
(m, 4 H), 1.27 (m, 20 H), 0.88 (t, 6 H); 13C NMR (CDCl3) δ
138.5, 128.3, 127.5, 78.1, 73.3, 71.6, 70.8, 70.6, 70.3, 31.7, 30.2,
29.7, 29.4, 29.1, 26.0, 22.3, 13.7.
Con clu sion s
The synthesis of a series of amphiphilic PI-PLC
substrate analogs with methyl phosphate, phosphate,
methyl phosphonate, and phosphonate head groups was
accomplished. The decreased steric bulk of the diradyl-
glycerols precluded the use of standard methodologies
for the synthesis of the intermediate phosphate tri-
esters. A protocol that allowed for the efficient synthesis
of short-chain dialkylglyceryl phosphates was developed.
The inhibition of B. cereus PI-PLC by these short-chain
lipids was measured below their cmc’s in a single-phase
aqueous assay using a water-soluble synthetic sub-
strate. Inhibition was most strongly dependent on the
hydrocarbon chain length of the inhibitors: among
compounds with identical methyl phosphate head groups
the potency of the inhibitor followed the order dioctyl
> dihexyl > dipentyl. This trend is consistent with the
structure of B. cereus PI-PLC and parallels observations
for short-chain lipid substrates for mammalian PI-
PLC.46 Also supporting this trend are results showing
ether-linked inhibitors are more potent than the corre-
sponding ester-linked inhibitors. Together, these data
stress the importance of hydrophobic interactions for
inhibitor or substrate binding. The hydrophilic head
group contributes to the inhibition, and the order of
potency is myo-inositol phosphonate > methyl phos-
ph(on)ate > phosph(on)ate. Consequently, the strongest
inhibitor among the dihexyl compounds is the phospho-
nate most closely resembling the substrate, PI. In
contrast, as a group, the water-soluble inositols and
other molecules resembling the head group of PI exhib-
ited a reduction in inhibition by two to three orders of
magnitude. Surprisingly, the presence of PEG es-
sentially abolished the inhibition by the short-chain
lipids, but had little effect on the water-soluble inositol
derivatives. Thus, when the presence of PEG is re-
quired, as in the crystallization of the B. cereus PI-PLC,
(()-1,2-Di-O-p en tyl-sn -glycer ol (6). A suspension of (()-
3-O-benzyl-1,2-di-O-pentyl-sn-glycerol (13.8 g, 42.8 mmol) and
10% Pd/C (0.2 g) in MeOH (50 mL) was shaken under an H2
atmosphere (50 psi) on a Parr hydrogenation apparatus for 2
h. The resulting suspension was filtered through a plug of
Celite, and the solvents were removed in vacuo to afford 6 as
a colorless oil (9.88 g, 99%): 1H NMR (CDCl3) δ 3.74-3.39 (m,
9 H), 2.36 (br s, 1 H), 1.58 (m, 4 H), 1.30 (m, 8 H), 0.89 (t, 6

Inhibitors of B. cereus PI-PLC
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22 4373
H); 13C NMR (CDCl3) δ 78.6, 71.6, 70.7, 70.2, 62.6, 29.6, 29.2,
28.1, 22.3, 13.8. Anal. (C13H28O3) C, H.
(()-1,2-Di-O-h exyl-sn -glycer ol (7). Prepared as described
above for 6, to afford 7,29 as a colorless oil (94%): 1H NMR
(CDCl3) δ 3.69-3.36 (m, 9 H), 2.54 (br s, 1 H), 1.49 (m, 4 H),
1.29 (m, 12 H), 0.82 (t, 6 H); 13C NMR (CDCl3) δ 78.5, 71.7,
70.8, 70.3, 62.8, 31.7, 30.0, 29.5, 25.7, 22.5, 13.8.
washed with CH2Cl2 (3 × 25 mL). The pooled organic extracts
were concentrated in vacuo to afford the free acid of 12. This
residue was suspended in MeOH (5 mL) and stirred with ion
exchange resin (BIO-REX 70, sodium form, 1.5 g). The
suspension was filtered and stirred with a fresh sample of ion
exchange resin. The suspension was filtered, and the solvents
were removed in vacuo to afford the sodium salt of 12 (430
mg, 89%) as a clear colorless oil: 1H NMR (CDCl3) δ 3.85 (m,
2 H), 3.74-3.21 (m, 10 H), 1.54 (m, 4 H), 1.28 (m, 8 H), 0.87
(t, 6 H); 13C NMR (CDCl3) δ 77.9 (d, J CP 4.7 Hz), 71.6, 70.5,
70.1, 64.9, 52.6 (d, J CP 5.7 Hz), 29.6, 29.4, 28.3, 28.2, 22.5, 14.0;
31P NMR (CDCl3) δ -1.76 (m). Anal. (C14H30O6NaP) C, H.
(()-Meth yl 1,2-Di-O-h exylglycer -3-yl P h osp h a te, Am -
m on iu m Sa lt of 13. The free acid of 13 was prepared as
described above for the free acid of 12. This colorless oil was
purified by column chromatography (CHCl3/MeOH/30% NH4-
OH, gradient) to afford the ammonium salt of 13 (301 mg,
52%), as a slightly yellow oil:66 1H NMR (CDCl3) δ 7.47 (br s,
NH4), 3.81 (m, 2 H), 3.72-3.34 (m, 10 H), 1.51 (m, 4 H), 1.24
(m, 12 H), 0.85 (t, 6 H); 13C NMR (CDCl3) δ 77.6 (d, J CP 5.7
Hz), 71.3, 70.4, 70.1, 53.6 (m), 31.7, 29.9, 29.6, 25.7, 25.6, 22.6,
14.0; 31P NMR (CDCl3) δ -0.21 (m). Anal. (C16H38O6NP‚H2O)
C, H.
(()-Meth yl 1,2-Di-O-octylglycer -3-yl P h osp h a te, Am -
m on iu m Sa lt of 14. The free acid of 14 was prepared as
described above for the free acid of 12. This colorless oil was
purified by column chromatography (89:10:1 CHCl3/MeOH/
30% NH4OH) to afford the ammonium salt of 14 (526 mg, 28%)
as a colorless oil: 1H NMR (D2O) δ 3.80 (m, 2 H), 3.72-3.39
(m, 10 H), 1.55 (m, 4 H), 1.27 (m, 20 H), 0.85 (m, 6 H); 13C
NMR (D2O) δ 78.3 (m), 71.7, 71.1, 70.6, 65.0, 53.0 (m), 32.2,
30.1, 29.9, 29.6, 26.4, 26.3, 22.8, 14.0; 31P NMR (D2O) δ 0.23
(m). Anal. (C20H46O6NP‚0.5H2O) C, H.
(()-Ben zyl Met h yl (2,2-Dim et h yl-1,3-d ioxola n -4-yl)-
m eth yl P h osp h a te (15). Prepared as described above for 9.
The residue was purified by column chromatography (1:1
hexanes/EtOAc) to afford 15 (3.05 g, 48%) as a mixture of
diastereomers: 1H NMR (CDCl3) δ 7.34 (m, 5 H), 5.05 (d, 2 H,
J HP 8.4 Hz), 4.22 (m, 1 H), 3.96 (m, 3 H), 3.68 (d, 3 H, J HP 11
Hz), 3.74 (m, 1 H), 1.37 (s, 3 H), 1.30 (s, 3 H); 13C NMR (CDCl3)
δ 135.7, 128.5, 127.8, 109.7, 73.9 (d, J CP 7.6 Hz), 69.4, 69.3,
67.4 (d, J CP 5.8 Hz), 65.9, 54.2 (d, J CP 5.6 Hz), 26.6, 25.9, 25.1;
31P NMR (CDCl3) δ 0.48 (m).
(()-Ben zyl Meth yl Glycer -3-yl P h osp h a te (16). A sus-
pension of 15 (1.49 g, 4.71 mmol) and ion exchange resin
(Dowex 50W-X8, acid form, 1.25 g) was stirred at room
temperature for 3.5 h under N2. The suspension was filtered,
and the solids obtained were washed with CH2Cl2 (2 × 5 mL).
The combined filtrates were concentrated in vacuo, and the
residue was purified by column chromatography (9:1 EtOAc/
MeOH) to afford 16 (693 mg, 53%) as a colorless oil: 1H NMR
(CDCl3) δ 7.36 (m, 5 H), 5.09 (d, 2 H, J HP 13 Hz), 4.08 (m, 2
H), 3.87 (m, 1 H), 3.73 (d, 3 H, J HP 11 Hz), 3.64 (m, 2 H), 3.00
(br s, 2 H).
(()-Ben zyl Meth yl 1,2-Di-O-h exa n oylglycer -3-yl P h os-
p h a te (17). Hexanoyl chloride (1.17 g, 6.86 mmol) was added
dropwise to a rapidly stirred solution of 16 (600 mg, 2.17 mmol)
and N,N-diisopropylethylamine (1.12 g, 8.68 mmol) in CH2-
Cl2 at 0 °C. The reaction mixture was allowed to warm to 25
°C over 14 h. The solution was flushed through a short column
of silica gel (1:1 hexanes/EtOAc), and the solvents were
removed in vacuo. The residue obtained was purified by
column chromatography (3:1 hexanes/EtOAc) to afford 17 (833
mg, 81%) as a colorless oil: 1H NMR (CDCl3) δ 7.37 (m, 5 H),
5.19 (t, 1 H), 5.07 (d, 2 H, J HP 8.4 Hz), 4.27 (dd, 1 H), 4.13 (m,
3 H), 3.72 (d, 3 H, J HP 11 Hz), 2.29 (t, 4 H), 1.60 (m, 4 H), 1.29
(m, 8 H), 0.88 (t, 6 H); 13C NMR (CDCl3) δ 173.0, 172.6, 135.7,
128.5, 127.8, 69.4 (m), 65.4 (d, J CP 4.7 Hz), 61.5, 54.3 (m), 34.0,
33.9, 31.1, 31.0, 24.2, 22.1, 13.7; 31P NMR (CDCl3) δ 0.32 (m).
(()-Meth yl 1,2-Di-O-h exa n oylglycer -3-yl P h osp h a te,
Sod iu m Sa lt of 18. The free acid of 18 was prepared as
described above for the free acid of 12. This residue was
stirred with ion exchange resin (Dowex 50W-X8, sodium form,
1.0 g) for 18 h and filtered. The solvents were removed in
vacuo to afford a yellow oil. This oil was suspended in hexanes
(50 mL) and filtered through a pad of Celite. The solvents
(()-1,2-Di-O-octyl-sn -glycer ol (8). Prepared as described
above for 6, to afford 8,30 as a colorless oil (94%): 1H NMR
(CDCl3) δ 3.61-3.31 (m, 9 H), 2.66 (br s, 1 H), 1.47 (m, 4 H),
1.20 (m, 20 H), 0.80 (t, 6 H); 13C NMR (CDCl3) δ 78.5, 71.6,
70.8, 70.3, 62.7, 31.7, 30.0, 29.6, 29.3, 29.1, 26.0, 22.5, 13.9.
(()-Ben zyl Meth yl 1,2-Di-O-p en tylglycer -3-yl P h os-
p h a te (9). Methyl dichlorophosphite (731 mg, 5.5 mmol) was
added via syringe (ca. 3 min) to a rapidly stirred solution of
N,N-diisopropylethylamine (1.94 g, 15 mmol) in CH2Cl2 (2.5
mL) at -78 °C. A solution of 6 (1.16 g, 5.0 mmol) in CH2Cl2
(25 mL) was then added dropwise over 45 min. After an
additional 1 h at this temperature, a solution of benzyl alcohol
(649 mg, 6.0 mmol) was added dropwise over 20 min. The
resulting mixture was stirred at -78 °C for 1 h and then
allowed to warm to 25 °C for 1 h. The solvents were removed
in vacuo, and the residue obtained was suspended in EtOAc
(10 mL). The resulting suspension was filtered through a plug
of Celite, and the solids obtained were washed with EtOAc (2
× 5 mL). The combined filtrates were concentrated, and the
residue was suspended in CH2Cl2 (10 mL) and cooled under
N2 to 0 °C. A 3 M solution of t-BuOOH in isooctane (1.5 mL)
was then added dropwise over 5 min. The reaction mixture
was then removed from the cooling bath and allowed to warm
to room temperature over 2 h. The reaction was filtered
through a pad of silica gel, and the filter was washed with
hexanes/EtOAc (100 mL, 4:1). The combined filtrates were
concentrated in vacuo to a yellow oil. This residue was purified
by column chromatography (3:2 hexanes/EtOAc) to give 9 as
a colorless oil (972 mg, 48%): 1H NMR (CDCl3) δ 7.33 (m, 5
H), 5.05 (d, 2 H, J HP 7.8 Hz), 4.20-3.99 (m, 2 H), 3.70 (d, 3 H,
J HP 11 Hz), 3.68-3.37 (m, 7 H), 1.52 (m, 4 H), 1.28 (m, 8 H),
0.85 (t, 6 H); 13C NMR (CDCl3) δ 135.8, 128.2, 126.7, 77.2 (m,
obscured by solvent peak), 71.6, 70.6, 70.3, 69.2, 69.1, 67.1 (d,
J CP 11 Hz), 64.8, 54.1 (d, J CP 12 Hz), 29.6, 29.2, 29.1, 28.2,
28.1, 22.4, 13.9; 31P NMR (CDCl3) δ 0.18 (m).
(()-Ben zyl Met h yl 1,2-Di-O-h exylglycer -3-yl P h os-
p h a te (10). Prepared as described above for 9. The crude
residue obtained was purified by RPTLC (3:2 hexanes/EtOAc)
to give 10 (541 mg, 64%) as a colorless oil: 1H NMR (CDCl3)
δ 7.33 (m, 5 H), 5.05 (d, 2 H, J HP 7.8 Hz), 4.20-3.99 (m, 2 H),
3.70 (d, 3 H, J HP 11 Hz), 3.68-3.37 (m, 7 H), 1.52 (m, 4 H),
1.28 (m, 12 H), 0.85 (t, 6 H); 13C NMR (CDCl3) δ 135.9, 128.5,
127.8, 77.2 (d, J CP 7.0 Hz), 71.7, 70.6, 69.7, 69.2, 69.1, 67.1 (d,
J CP 5.0 Hz), 54.1 (m), 31.6, 29.9, 29.3, 25.7, 25.6, 22.5, 13.9;
31P NMR (CDCl3) δ 0.58 (br m), along with methyl bis(1,2-di-
O-hexylglyceryl-3-yl) phosphate (177 mg, 31.2%) as a colorless
oil: 1H NMR (CDCl3) δ 4.19-4.01 (m, 4 H), 3.76 (d, 3 H, J HP
11 Hz), 3.83-3.38 (m, 14 H), 1.57 (m, 8 H), 1.29 (m, 24 H),
0.83 (m, 12 H).
(()-Ben zyl Meth yl 1,2-Di-O-octylglycer -3-yl P h osp h a te
(11). Prepared as described above for 9. The residue was
purified by column chromatography (3:2 hexanes/EtOAc) to
give 11 as a colorless oil (555 mg, 68%): 1H NMR (CDCl3) δ
7.35 (m, 5 H), 5.06 (d, 2 H, J HP 8.0 Hz), 4.20-3.95 (m, 2 H),
3.71 (d, 3 H, J HP 11 Hz), 3.60-3.30 (m, 7 H), 1.52 (m, 4 H),
1.24 (m, 20 H), 0.86 (t, 6 H); 13C NMR (CDCl3) δ 135.9, 128.5,
127.8, 77.2 (m), 71.7, 70.6, 69.6, 69.2, 67.0 (m), 54.1 (m), 31.8,
30.0, 29.6, 29.4, 29.2, 26.0, 22.6, 14.0; 31P NMR (CDCl3) δ 0.23
(br m).
(()-Meth yl 1,2-Di-O-p en tylglycer -3-yl P h osp h a te, So-
d iu m Sa lt of 12. A suspension of 9 (560 mg, 1.38 mmol) and
20% Pd(OH)2/C (150 mg) in MeOH (25 mL) was shaken on a
Parr hydrogenation apparatus for 1.5 h under H2 (50 psi). The
resulting suspension was filtered through a plug of Celite, and
the solvents were removed in vacuo to afford the crude free
acid of 12 as a yellow oil. This crude residue was suspended
in 2.5% aqueous NaHCO3 (50 mL) and washed with CH2Cl2
(3 × 25 mL). The organic layers were discarded, and the
aqueous solution was acidified to ca. pH ) 1 with HCl and

4374 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22
Ryan et al.
were removed in vacuo, and the residue was suspended in
water (10 mL) and filtered through a pad of Celite. The
resulting soapy solution was lyophilized to afford the sodium
salt of 18 (468 mg, 100%) as a colorless leather:66 1H NMR
(CDCl3) δ 5.32 (m, 1 H) 4.45 (dd, 1 H), 4.30 (m, 2 H), 4.02 (m,
2 H), 3.60 (d, 3 H, J HP 10 Hz), 2.36 (m, 4 H), 1.63 (m, 4 H),
1.33 (m, 8 H), 0.91 (t, 6 H); 13C NMR (CDCl3) δ 173.5, 173.4,
70.8 (d, J CP 7.7 Hz), 63.4, 62.7, 52.6 (d, J CP 5.8 Hz), 34.2, 34.0,
31.2, 24.4, 22.2, 13.8; 31P NMR (CDCl3) δ 1.38 (m). Anal.
(C16H30O8NaP) C, H.
concentrated in vacuo, and the resulting residue was sus-
pended in 1% NaHCO3 (100 mL). The resulting milky suspen-
sion was washed with CH2Cl2 (3 × 25 mL). The organic layers
were discarded, and the aqueous suspension was acidified
(HCl) to ca. pH ) 1 and washed with CH2Cl2 (3 × 25 mL).
The pooled organic phases were concentrated in vacuo, and
the resulting residue was suspended in water (10 mL), stirred
with ion exchange resin (Dowex 50W-X8, sodium form, 2.0 g)
for 2 h, and filtered. The filtrates were lyophilized to afford
26 (134 mg, 19%) as a white waxy residue: 1H NMR (D2O) δ
3.74-3.37 (m, 10 H), 1.87-1.64 (m, 8 H), 1.36 (m, 12 H), 0.86
(m, 6 H); 13C NMR (D2O) δ 179.5, 79.3, 73.3, 70.8 (d, J CP 79
Hz), 51.3 (m), 31.6, 29.6 (d, J CP 17 Hz), 26.7, 25.6, 25.4, 22.6,
13.8; 31P NMR (D2O) δ 28.9 (m). Anal. (C17H35O5NaP) C, H.
II. P h ysica l P r op er ties. Deter m in a tion of th e cm c.
The cmc of synthetic amphiphilic compounds was determined
by recording the solute concentration dependent change in
surface tension. The surface tension was measured by capil-
lary rise67 adapted to small volumes as described by Rebecchi
and co-workers.46 A glass cup (i.d. 21 mm) filled with at least
0.5 mL of the solution to be measured was placed onto a
vertically adjustable table. The cup was raised to immerse
the tip of a 10 µL glass capillary (Gold Seal glass capillaries,
Clay Adams). The capillary had been clamped to a clear block
of acrylic held by an adjustable arm and positioned perpen-
dicular to the surface of the liquid. Latex tubing fitted to the
distal end of the capillary served to apply suction as well as
pressure to the liquid column risen in the capillary, allowing
the inner wall of the capillary to be wetted before measuring
falling and rising menisci. A cathetometer (Eberbach Corp.)
was used to measure the distance between the surface of the
solution and the meniscus of the liquid column in the capillary.
In a series of surface tension measurements, the highest
concentration of solute was measured first, and subsequent
dilutions were made directly in the glass cup. The height of
the capillary rise resulting from at least one rising and one
falling meniscus were recorded for each dilution after allowing
the meniscus to stabilize for 10-20 min. A fresh capillary was
used to measure each dilution.
(()-Diben zyl 1,2-Di-O-h exylglycer -3-yl P h osp h a te (19).
To a rapidly stirred suspension of POCl3 (3.45 g, 22.5 mmol)
and triethylamine (2.28 g, 22.5 mmol) in anhydrous ether (15
mL) at -15 °C was added a solution of 7 (3.26 g, 12.5 mmol)
in anhydrous ether (5 mL) over 15 min. The reaction was
stirred at this temperature for 1 h and filtered under N2. The
solids obtained were washed with toluene (25 mL), and the
pooled filtrates were concentrated in vacuo at <40 °C. The
resulting oil was suspended in anhydrous ether (40 mL), and
triethylamine (8.34 g, 82.4 mmol) was added. The resulting
mixture was cooled to 0 °C, and benzyl alcohol (8.93 g, 82.4
mmol) was added. The reaction was stirred at this tempera-
ture for 1.5 h, allowed to warm to 25 °C for 2 h and filtered,
and the filtrates were concentrated to a red oil. This crude
residue was purified by column chromatography (7:3 hexanes/
EtOAc) to give 19 (5.63 g, 87%) as a colorless oil: 1H NMR
(CDCl3) δ 7.33 (m, 1 H), 5.05 (m, 4 H), 4.16-3.98 (m, 2 H),
3.58-3.48 (m, 3 H), 3.41-3.37 (m, 4 H), 1.52 (m, 4 H), 1.27
(m, 12 H), 0.89 (t, 6 H); 13C NMR (CDCl3) δ 135.9, 128.5, 128.4,
127.8, 71.7, 70.7, 69.7, 69.2, 69.1, 67.1 (d, J CP 6.0 Hz), 31.6,
29.9, 29.6, 25.7, 22.6, 14.0; 31P NMR (CDCl3) δ -3.26 (m).
(()-1,2-Di-O-h exylglycer -3-yl P h osp h a te, F r ee Acid of
20. The free acid of 20 was prepared as described above for
12. The residue obtained (329 mg, 98%) was analytically pure
without further purification: 1H NMR (CDCl3) δ 9.35 (br s, 2
H) 4.10 (m, 2 H), 3.70-3.55 (m, 3 H), 3.49-3.40 (m, 4 H), 1.56
(m, 4 H), 1.28 (m, 12 H), 0.88 (t, 6 H); 13C NMR (CDCl3) δ
72.1, 71.4, 69.8, 66.2, 31.6, 29.6, 29.4, 25.6, 25.5, 22.5, 14.0;
31P NMR (CDCl3) δ -0.97 (m). Anal. (C15H28O6P‚H2O) C, H.
(()-Dib en zyl (2,2-Dim et h yl-1,3-d ioxola n -4-yl)m et h yl
P h osp h a te (21). Prepared as described above for 19. The
residue was purified by column chromatography (3:2 hexanes/
EtOAc) to afford 21 (8.72 g, 89%): 1H NMR (CDCl3) δ 7.35
(m, 10 H), 5.05 (d, 4 H, J HP 8.4 Hz), 4.21 (m, 1 H), 3.99-3.89
(m, 2 H), 3.80-3.70 (m, 2 H), 1.38 (s, 3 H), 1.33 (s, 3 H).
(()-Diben zyl Glycer -3-yl P h osp h a te (22). Prepared as
described above for 16. The crude residue obtained was
purified by column chromatography (EtOAc) to afford 22 (1.83
g, 70.2%) as a colorless oil: 1H NMR (CDCl3) δ 7.36 (m, 10 H),
5.06 (d, 4 H, J HP 8.7 Hz), 4.03 (dd, 2 H), 3.82 (m, 1 H), 3.64-
3.53 (m, 2 H), 2.75-2.60 (br s, 2 H).
(()-Diben zyl 1,2-Di-O-h exa n oylglycer -3-yl P h osp h a te
(23). Prepared as described for 17. The crude residue
obtained was purified by column chromatography (9:1 hexanes/
EtOAc) to afford 23 (1.83 g, 83%) as a colorless oil: 1H NMR
(CDCl3) δ 7.34 (m, 10 H), 5.16 (m, 1 H), 5.07 (d, 4 H, J HP 8.1
Hz), 4.27 (m, 1 H), 4.14 (m, 3 H), 2.30 (t, 4 H), 1.61 (m, 4 H),
1.31 (m, 8 H), 0.95 (t, 6 H); 13C NMR (CDCl3) δ 173.1, 172.8,
138.7, 128.7, 128.6, 127.9, 69.5 (m), 65.1, 61.6, 34.0, 33.9, 31.2,
31.0, 24.5, 22.2, 13.8; 31P NMR (CDCl3) δ -2.22 (m).
(()-1,2-Di-O-h exa n oylglycer -3-yl P h osp h a te, F r ee Acid
of 24. The free acid of 24 was prepared as described above
for 12. The residue obtained (428 mg, 100%) was analytically
pure without further purification: 1H NMR (CDCl3) δ 5.24 (m,
1 H), 4.37-4.10 (m, 5 H), 2.17 (m, 4 H), 1.59 (m, 4 H), 1.30
(m, 8 H), 0.86 (m, 6 H); 13C NMR (CDCl3) δ 174.0, 173.7, 69.7
(d, J CP 13 Hz), 65.0, 64.9, 62.1, 34.1, 33.9, 31.2, 31.1, 22.4, 22.2,
13.8; 31P NMR (CDCl3) δ -2.37 (m).
Assuming a 0° contact angle between solution and capillary
wall, the surface tension γ (mN m-1
) was calculated using the
equation γ ) (1/2)hrpg, where h (cm) is the distance between
the surface of the solution in the cup and the meniscus of the
solution in the capillary with the internal radius r (cm), p (g
cm-3) is the density of the solution, and g (cm s-2) is the
acceleration due to gravity. A graph of γ versus ln(solute
concentration) gave the estimated cmc value as that concen-
tration of solute above which the slope of the line fitted to the
measurements is changed significantly.
Using this procedure, the cmc values determined for Triton
X-100 (Boehringer, Mannheim) and SDS (BDH Chemicals
Ltd.), 1.1 and 9.0 mM, were close to those reported in the
literature,68 0.2-0.9 and 7-10 mM, respectively, and were
higher than those we found using the du Nouy ring detach-
ment method (Surface Tensiomat Model 21, Fisher), 0.3 and
5.0 mM, respectively.
For compounds 20, 24, and 31, there was no clear transition
in the slope of the curve of γ as a function of the concentration
due to insufficient data at high concentrations. The estimated
cmc’s listed in Table 1 were obtained by placing γ ) 25.5-
28.5 mN m-1 as the surface tension for solute concentrations
at the cmc that we have observed for other compounds listed
in Tables 1 and 2 (see also results by Bian and Roberts44 and
Garigapati et al.,45 obtained for similar compounds).
III. Bioch em ica l Assa ys. En zym e. The PI-PLC from B.
cereus was isolated from recombinant E. coli over-expressing
the cloned enzyme as described69 but modifying the procedure
as follows: the recombinant enzyme was released from the
periplasmic space of the bacterial cells by treatment with
lysozyme in combination with mild osmotic shock using the
protocol given by Witholt et al.70 The purified enzyme was
stored at -20 °C in 10 mM Tris-HCl, pH 8.5, 1 mM EDTA,
0.01% NaN3, and 50% glycerol. (Essentially the same enzyme
preparation is now available from Molecular Probes Inc.,
Eugene, OR.) Enzyme dilutions were prepared in 0.1% PEG-
8000 (Sigma), 6 mM HEPES, 0.06 mM EDTA, pH 7. HEPES,
Meth yl [3,4-Bis(h exyloxy)bu tyl]p h osp h on a te, Sod iu m
Sa lt of 26. A suspension of 2525 (700 mg, 1.9 mmol) and NaI
(2.0 g) in 5 mL of MEK was heated at reflux under N2 for 9 h.
The resulting mixture was poured into water (50 mL), and the
resulting milky suspension was washed with CH2Cl2 (3 × 25
mL). The organic layers were discarded, and the aqueous
suspension was acidified (HCl) to ca. pH ) 1 and washed with
CH2Cl2 (3 × 25 mL). The pooled organic phases were

Inhibitors of B. cereus PI-PLC
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22 4375
free acid, was from Calbiochem and Tris base from Life
Technologies.
Refer en ces
(1) Berridge, M. J . Inositol Trisphosphate and Calcium Signalling.
Nature 1993, 361, 315-325.
(2) Majerus, P. W. Inositol Phosphate Biochemistry. Annu. Rev.
Biochem. 1992, 61, 225-250.
(3) Noh, D.-J .; Shin, S. H.; Rhee, S. G. Phosphoinositide-Specific
Phospholipase C and Mitogenic Signaling. Biochim. Biophys.
Acta 1995, 1242, 99-114.
Su bst r a t e. The chromogenic substrate, racemic myo-
inositol 1-(4-nitrophenyl phosphate) (NPIP) was prepared as
reported and used in all kinetic analyses.26 Substrate con-
centrations given refer to the D-enantiomer only. The L-
enantiomer was shown to be neither a substrate nor an
inhibitor for PI-PLC.13 The kinetic parameters for the hy-
(4) Shimohama, S.; Perry, G.; Richey, P.; Praprotnik, D.; Takenawa,
T.; Fukami, K.; Whitehouse, P. J .; Kimura, J . Characterization
of the Association of Phospholipase C-δ with Alzheimer Neu-
rofibrillary Tangles. Brain Res. 1995, 669, 217-224.
(5) Hill, S. R.; Bonjouklian, R.; Powis, G.; Abraham, R. T.; Ashendel,
C. L.; Zalkow, L. H. A Multisample Assay for Inhibitors of
Phosphatidylinositol Phospholipase C: Identification of Natu-
rally Occurring Peptide Inhibitors with Antiproliferative Activ-
ity. Anti-Cancer Drug Des. 1994, 9, 353-361.
(6) Heinz, D. W.; Ryan, M.; Griffith, O. H. Crystal Structure of the
Phosphatidylinositol-Specific Phospholipase C from Bacillus
cereus in Complex with myo-Inositol. EMBO J . 1995, 14, 3855-
3863.
(7) Essen, L.-O.; Perisic, O.; Cheung, R.; Katan, M.; Williams, R. L.
Crystal Structure of a Mammalian Phosphoinositide-specific
Phospholipase Cδ. Nature 1996, 380, 595-602.
(8) Bruzik, K. S.; Tsai, M.-D. Toward the Mechanism of Phospho-
inositide-Specific Phospholipases C. Bioorg. Med. Chem. 1994,
2, 49-72.
(9) Rosenberry, T. L.; Toutant, J .-P.; Haas, R.; Roberts, W. L.
Identification and Analysis of Glycoinositol Phospholipid An-
chors in Membrane Proteins. Methods Cell Biol. 1989, 32, 231-
255.
(10) Low, M. G. Degradation of Glycosyl-Phosphatidylinositol An-
chors by Specific Phospholipases. In Molecular and Cell Biology
of Membrane Proteins. Glycolipid Anchors of Cell-Surface Pro-
teins; Turner A. J ., Ed.; Ellis Horwood Ltd.: Chichester, West
Sussex, U.K., 1990; pp 35-63.
(11) Prusiner, S. B. The Prion Diseases. Sci. Am. 1995, 272, 48-57.
(12) Volwerk, J . J .; Shashidhar, M. S.; Kuppe, A.; Griffith, O. H.
Phosphatidylinositol-Specific Phospholipase C from Bacillus
cereus Combines Intrinsic Phosphotransferase and Cyclic Phos-
phodiesterase Activities: A 31P NMR Study. Biochemistry 1990,
29, 8056-8062.
(13) Leigh, A. J .; Volwerk, J . J .; Griffith, O. H.; Keana, J . F. W.
Substrate Stereospecificity of Phosphatidylinositol-Specific Phos-
pholipase C from Bacillus cereus Examined Using the Resolved
Enantiomers of Synthetic myo-Inositol-(4-Nitrophenyl Phos-
phate). Biochemistry 1992, 31, 8979-8983.
drolysis of D-NPIP by B. cereus PI-PLC are KM ) 5 mM and
13
Vmax ) 650 µmol min-1 mg-1
.
In h ibitor s. The short-chain phospholipids of Tables 1 and
2 and GlcN(R1f6)Ins (28) (Table 3) were synthesized as
described above. All inositols, inososes, and inositol phos-
phates were commercial products obtained from Sigma Chemi-
cal Co. All compounds evaluated for their inhibitory properties
were prepared in 1.25× assay buffer (125 mM HEPES, 1.25
mM EDTA; pH 7). These stock solutions were sonicated and
stored at -20 °C. Dilutions of the stock solutions were
prepared directly before use in the same buffer with 0.1% PEG-
8000 added.
En zym e Assa ys. The rate of substrate hydrolysis by PI-
PLC was measured in 0.5 mL reactions thermostated at 25
°C. All reactions contained final concentrations of 5 nM PI-
PLC and 0.25 mM NPIP substrate. Duplicate control reactions
were included in each set of inhibition assays. In a disposable
methacrylate semi-microcuvette 360 µL of 1.25× assay buffer
(125 mM HEPES, 1.25 mM EDTA; pH 7), 40 µL of 0.1% PEG-
8000 in 1.25× assay buffer and H2O to 492.5 µL were
combined. A few minutes before the reaction was started, 2.5
µL of 50 mM NPIP was added and mixed. The reaction was
initiated by adding 5 µL of diluted enzyme and mixing
immediately by rapid inversion. Reactions containing inhibi-
tor were combined in the same fashion except inhibitor was
added a few minutes prior to the addition of NPIP and enzyme.
Of inhibitor stock solutions in 1.25× assay buffer up to 360
µL could be added or up to 40 µL of diluted inhibitor in 0.1%
PEG-8000, 1.25× assay buffer. In each case, the 0.5 mL
reactions contained final concentrations of 1× assay buffer and
0.008% PEG-8000.
When PI-PLC activity was measured under conditions
routinely used to grow PI-PLC crystals, the reactions contained
450 µL of 100% precipitant solution (35.6% PEG-600 (Sigma),
0.18 M trisodium citrate-HCl, pH 7.35, and 40 mM HEPES,
pH 7), 40 µL of 1.25× assay buffer with or without inhibitor,
H2O to 492.5 µL, 2.5 µL of 50 mM NPIP, and 5 µL of diluted
enzyme. The reactions were combined as described above with
the enzyme added last.
(14) Lewis, K. A.; Garigapati, V. R.; Zhou, C.; Roberts., M. F.
Substrate Requirements of Bacterial Phosphatidylinositol-
Specific Phospholipase C. Biochemistry 1993, 32, 8836-8841.
(15) Bruzik, K. S.; Morocho, A. M.; J hon, D.-Y.; Rhee, S. G.; Tsai,
M.-D. Phospholipids Chiral at Phosphorus. Stereochemical
Mechanism for the Formation of Inositol 1-Phosphate Catalyzed
by Phosphatidylinositol-Specific Phospholipase C. Biochemistry
1992, 31, 5183-5193.
The progress of NPIP hydrolysis was followed by monitoring
the accumulation of the p-nitrophenolate anion hydrolysis
product in a spectrophotometer at 399 nm for 4 min, com-
mencing within 15 s after addition of enzyme. The molar
extinction coefficient of the p-nitrophenolate anion was deter-
mined for the different reaction conditions used. To obtain
the initial velocity of the reaction, v0, the progress curve was
fitted by nonlinear regression to a first-order rate equation
using GraFit Vs. 3.0 software.71 To determine the IC50 of an
inhibitor, v0 values were obtained for uninhibited reactions and
in the presence of at least five different concentrations of
inhibitor. The plot of v0 against inhibitor concentration was
fitted by nonlinear regression to yield the concentration of
inhibitor where v0 ) 50% of the uninhibited rate. When it
was not possible to achieve inhibitor concentrations high
enough to result in a 50% reduction of the initial rate, IC50
values were extrapolated from available data points if the
trend of the curve was clearly defined.
The Ki was determined by measuring initial rates in the
presence of varying concentrations of substrate without inhibi-
tor added and in the presence of two different concentrations
of inhibitor. Rate is monitored versus the concentration of
substrate and inhibitor and the data fitted to a multidimen-
sional equation for competitive inhibition using GraFit.71 The
Ki obtained in this fashion compared well with results obtained
from linearized data in Lineweaver-Burke plots.
(16) Guther, M. L. S.; Cardoso de Almeida, M. L.; Rosenberry, T. L.;
Ferguson, M. A. J . The Detection of Phospholipase-Resistant and
-Sensitive Glycosyl-Phosphatidylinositol Membrane Anchors by
Western Blotting. Anal. Biochem. 1994, 219, 249-255.
(17) Brufani, M.; Cellai, L.; Cesta, M. C.; Filocamo, L.; Iannelli, M.
A.; Lappa, S. Synthesis of a Deoxythiophosphate Analog of
Phosphatidyl-myo-Inositol as Potential Inhibitor of Phosphati-
dylinositol Specific Phospholipase C. Pharm. Pharmacol. Lett.
1995, 5, 35-37.
(18) Campbell, A. S.; Thatcher, G. R. J . Synthesis of an Analogue of
D,L-myo-Inositol-1,2-Cyclic Phosphate: Inhibition of Phosphati-
dylinositol-Specific Phospholipase C. Tetrahedron Lett. 1992, 32,
2207-2210.
(19) Campbell, A. S.; Thatcher, G. R. J . Tetravanadate is an Inhibitor
of Phosphatidyl Inositol-Specific Phospholipase C. Bioorg. Med.
Chem. Lett. 1992, 2, 655-658.
(20) Garigapati, V. R.; Roberts, M. F. Synthesis of Phosphatidyl-2-
O-Alkylinositols as Potential Inhibitors for PI Specific PLC.
Tetrahedron Lett. 1993, 34, 5579-5582.
(21) Morris, J . C.; Ping-Sheng, L.; Shen, T.-Y.; Mensa-Wilmot, K.
Glycan Requirements of Glycosylphosphatidylinositol Phospho-
lipase C from Trypanosoma brucei. Glucosaminylinositol Deriva-
tives Inhibit Phosphatidylinositol Phospholipase C. J . Biol.
Chem. 1995, 270, 2517-2524.
(22) Ogawara, H.; Higashi, K.; Manita, S.; Hidaka, M.; Kato, H.;
Takenawa, T. An Inhibitor of Inositol-Phospholipid-Specific
Phospholipase C. Biochim. Biophys. Acta 1993, 1175, 289-292.
(23) Perrella, F. W.; Chen, S.-F.; Behrens, D. L.; Kaltenback, R. F.,
III; Seitz, S. P. Phospholipase C Inhibitors: a New Class of
Cytotoxic Agents. J . Med. Chem. 1994, 37, 2232-2237.
(24) Shashidhar, M. S.; Volwerk, J . J .; Keana, J . F. W.; Griffith, O.
H. Inhibition of Phosphatidylinositol-Specific Phospholipase C
by Phosphonate Substrate Analogues. Biochim. Biophys. Acta
1990, 1042, 410-412.
Ack n ow led gm en t. This work was supported by
NIH Grants GM 25698 and GM 27137.

4376 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 22
Ryan et al.
(25) Vinod, T. K.; Griffith, O. H.; Keana, J . F. W. Isosteric Phospho-
nate and Isopolar Fluorophosphonate Substrate Analogue In-
hibitors for Phosphatidylinositol-Specific Phospholipase C from
Bacillus cereus. Tetrahedron Lett. 1994, 35, 7193-7196.
(26) Shashidhar, M. S.; Volwerk, J . J .; Griffith, O. H.; Keana, J . F.
W. A Chromogenic Substrate for Phosphatidylinositol-Specific
Phospholipase C: 4-Nitophenyl myo-Inositol-1-Phosphate. Chem.
Phys. Lipids 1991, 60, 101-110.
(49) Stieger, S.; Brodbeck, U. Glycosyl-Phosphatidylinositol Anchored
Acetylcholinesterase as Substrate for Phosphatidylinositol-
Specific Phospholipase C from Bacillus cereus. Biochimie 1991,
73, 1179-1186.
(50) Volwerk, J . J .; Filthuth, E.; Griffith, O. H.; J ain, M. K. Phos-
phatidylinositol-Specific Phospholipase C from Bacillus cereus
at the Lipid-Water Interface: Interfacial Binding, Catalysis, and
Activation. Biochemistry 1994, 33, 3464-3474.
(51) Ducruix, A., Giege´, R., Eds. Crystallization of Nucleic Acids and
Proteins. A Practical Approach; Oxford University Press: Oxford,
1992.
(52) McPherson, A. Preparation and Analysis of Protein Crystals;
Robert E. Krieger Publishing Co.: Malabar, FL, 1989.
(53) Garner, M. M.; Burg, M. B. Macromolecular Crowding and
Confinement in Cells Exposed to Hypertonicity. Am. J . Physiol.
1994, 266, C877-C892.
(54) Zimmerman, S. B.; Minton, A. P. Macromolecular Crowding:
Biochemical, Biophysical, and Physiological Consequences. Annu.
Rev. Biophys. Biomol. Struct. 1993, 22, 27-65.
(27) Kates, M.; Chan, T. H.; Stanacev, N. Z. Aliphatic Diether Analogs
of Glyceride-Derived Lipids. I. Synthesis of D-R,â-Dialkyl Gly-
ceryl Ethers. Biochemistry 1963, 2, 394-397.
(28) Howe, R. J .; Malkin T. A Simple Synthesis of Choline Alkyl
Phosphates. J . Chem. Soc. 1951, 2663-2667.
(29) Burns, R. A., J r.; Friedman, J . M.; Roberts, M. F. Characteriza-
tion of Short-Chain Alkyl Ether Lecithin Analogs: 13C NMR and
Phospholipase Studies. Biochemistry 1980, 20, 5945-5950.
(30) Paltauf, F.; Spencer, F. An Improved Synthesis of 1,2-Dialkyl
Glycerol Ethers and the Synthesis of 14C-Labeled Trialkyl
Glycerol Ethers. Chem. Phys. Lipids 1968, 2, 168-172.
(31) Ganog, B. R.; Loomis, C. R.; Hannun, Y. A.; Bell, R. M. Specificity
and Mechanism of Protein C Activation by sn-1,2-Diacylglycer-
ols. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 1184-1188.
(32) Wheeler, T. N.; Blanchard, S. G.; Andrews, R. C.; Fang, F.; Gray-
Nunez, Y.; Harris, C. O.; Lambert, M. H.; Mehrotra, M. M.;
Parks, D. J .; Ray, J . A.; Smalley, T. L. Substrate Specificity in
Short-Chain Phospholipid Analogs at the Active Site of Human
Synovial Phospholipase A2. J . Med. Chem. 1994, 37, 4118.
(33) Beaucage, S. L.; Iyer, R. P. The Synthesis of Specific Ribonucle-
otides and Unrelated Phosphorylated Biomolecules by the Phos-
phoramidite Method. Tetrahedron 1993, 49, 10441-10488.
(34) Reddy, K. K.; Saady, M.; Falk, J . R.; Whited, G. Intracellular
Mediators: Synthesis of L-R-Phosphatidyl-D-myo-Inositol 3,4,5-
Triphosphate Glyceryl Ether Analogs. J . Org. Chem. 1995, 60,
3385-3390.
(55) Reddy, M. K.; Weitzel, S. E.; Daube, S. S.; J arvis, T. C.; von
Hippel, P. H. Using Macromolecular Crowding Agents to Identify
Weak Interactions within DNA Replication Complexes. Methods
Enzymol. 1995, 262, 466-476.
(56) Cohen, S.; Marcus, Y.; Migron, Y.; Dikstein, S.; Shafran, A. Water
Sorption, Binding and Solubility of Polyols. J . Chem. Soc.,
Faraday Trans. 1993, 89, 3271-3275.
(57) Sasahara, K.; Uedaira, H. Solubility of Amino Acids in Aqueous
Poly(ethylene glycol) Solutions. Colloid Polym. Sci. 1993, 271,
1035-1041.
(58) The Merck Index: An Encyclopedia of Chemicals, Drugs, and
Biologicals, 10th ed.; Windholz, M., Ed.; Merck & Co.: Rahway,
NJ , 1983; p 1092.
(59) Bullock, T. L.; Ryan, M.; Kim, S. L.; Remington, J .; Griffith, O.
H. Crystallization of Phosphatidylinositol-Specific Phospholipase
C from Bacillus cereus. Biophys. J . 1994, 64, 784-791.
(60) Serajuddin, A. T. M.; Sheen, P. C.; Augustine, M. A. Water
Migration from Soft Gelatin Capsule Shell to Fill Material and
its Effect on Drug Solubility. J . Pharm. Sci. 1986, 75, 62-64.
(61) Gould, P. L.; Goodman, M.; Hanson, P. A. Investigation of the
Solubility Relationships of Polar, Semi-Polar and Non-Polar
Drugs in Mixed Co-Solvent Systems. Int. J . Pharm. 1984, 19,
149-159.
(62) Shihab, F.; Sheffield, W.; Sprowls, J .; Nematollahi, J . Solubility
of Alkyl Benzoates I: Effect of some Alkyl p-Hydroxybenzoates
(Parabens) on the Solubility of Benzyl p-Hydroxybenzoate. J .
Pharm. Sci. 1970, 59, 1574-1577.
(35) Stepanov, A. E.; Shevets, V. I. Formation of Phosphoester Bonds
in Phosphoglyceride Synthesis. Chem. Phys. Lipids 1986, 41,
1-51.
(36) Eibl, H.; Woolley, P. Synthesis of Enantiomerically Pure Glyceryl
Esters and Ethers II. Methods Employing the Precursor 3,4-
Isopropylidene-D-Mannitol. Chem. Phys. Lipids 1988, 47, 47-
55.
(37) Lacey, C. J .; Loew, L. M. A Simple Synthesis of Choline Alkyl
Phosphates. Tetrahedron Lett. 1980, 21, 2017-2020.
(38) Lacey, C. J .; Loew, L. M. Phospholipid Synthesis Based on New
Sequential Phosphate and Carboxylate Ester Bond Formation
Steps. J . Org. Chem. 1983, 48, 5214-5221.
(39) Martin, S. F.; J osey, J . A.; Wong, Y.-L.; Dean, D. W. General
Method for the Synthesis of Phospholipid Derivatives of 1,2-O-
diacyl-sn-glycerols. J . Org. Chem. 1994, 59, 4805-4820.
(40) Heeb, N. V.; Nambiar, K. P. Synthesis of (R)-Lysothioposphatidic
Acid and (R)-Thiophosphatidic Acid. Tetrahedron Lett. 1993, 34,
6193-6196.
(41) Lemmen, P.; Buchweitz, K. M.; Stumpf, R. A Synthesis of
Phospholipids Using a Phosphite Triester Approach with Uni-
formly Cleavable â-Halogenated Protecting Groups. Chem. Phys.
Lipids 1990, 53, 65-75.
(42) Plourde, R.; d’Alarcao, M. Synthesis of a Potentially Insulin-
Mimetic Phosphodisaccharide. Tetrahedron Lett. 1990, 31, 2693-
2696.
(63) Pandit, N. K.; Kanjia, J .; Patel, K.; Pontikes, D. G. Phase
Behavior of Aqueous Solutions Containing Nonionic Surfactant-
Polyethylene Glycol Mixtures. Int. J . Pharm. 1995, 122, 27-
33.
(64) Molyneux, P. Water-Soluble Synthetic Polymers: Properties and
Behavior, Vol. II; CRC Press, Inc.: Boca Raton, FL, 1984; pp
77-157.
(65) Heinz, D. W.; Ryan, M.; Smith, M. P.; Weaver, L. H.; Keana, J .
F. W.; Griffith, O. H. Crystal Structure of Phosphatidylinositol-
Specific Phospholipase C from Bacillus cereus in Complex with
Glucosaminyl(R1f6)-D-myo-Inositol, an Essential Fragment of
GPI-Anchors. Biochemistry 1996, 35, 9496-9504.
(66) J ain, M. K.; Rogers, J .; Marecek, J . F.; Ramirez, F.; Eibl, F. Effect
of Structure of Phospholipid on Kinetics of Intravesicle Scooting
of Phospholipase A2. Biochim. Biophys. Acta 1986, 860, 462-
474.
(67) Adamson, A. W. Physical Chemistry of Surfaces, 5th ed.; J ohn
Wiley & Sons, Inc.: New York, 1990.
(68) Neugebauer, J . A Guide to the Properties and Uses of Detergents
in Biology and Biochemistry; Calbiochem Corp.: San Diego, CA,
1988.
(69) Koke, J . A.; Yang, M.; Henner, D. J .; Volwerk, J . J .; Griffith, O.
H. High-Level Expression in Escherichia coli and Rapid Puri-
fication of Phosphatidylinositol-Specific Phospholipase C from
Bacillus cereus and Bacillus thuringiensis. Protein Expression
Purif. 1991, 2, 51-58.
(70) Witholt, B.; Boekhout, M.; Brock, M.; Kingma, J .; van Heer-
ikhuizen, H.; de Leij, L. An Efficient and Reproducible Procedure
for the Formation of Spheroplasts from Variously Grown Es-
cherichia coli. Anal. Biochem. 1976, 74, 160-170.
(71) Leatherbarrow, R. J . GraFit Version 3.0; Erithacus Software
Ltd.: Staines, U.K., 1992.
(43) Plourde, R.; d’Alarcao, M.; Saltiel, A. R. Synthesis and Charac-
terization of an Insulin-Mimetic Disaccharide. J . Org. Chem.
1992, 57, 2606-2610.
(44) Bian, J .; Roberts, M. F. Comparison of Surface Properties and
Thermodynamic Behavior of Lyso- and Diacylphosphatidylcho-
lines. J . Colloid Interface Sci. 1992, 153, 420-428.
(45) Garigapati, V. R.; Bian, J .; Roberts, M. F. Synthesis and
Characterization of Short-Chain Diacylphosphatidic Acids. J .
Colloid Interface Sci. 1995, 169, 486-492.
(46) Rebecchi, M. J .; Eberhardt, R.; Delaney, T.; Ali, S.; Bittman, R.
Hydrolysis of Short Acyl Chain Inositol Lipids by Phospholipase
C-δ1. J . Biol. Chem. 1993, 268, 1735-1741.
(47) Nieschalk, J .; O’Hagan, D. Monofluorophosphonates as Phos-
phate Mimics in Bioorganic Chemistry: a Comparative Study
of CH2- CHF- and CF2-Phosphonate Analogues of sn-Glycerol-
3-Phosphate as Substrates for sn-Glycerol-3-Phosphate Dehy-
drogenase. J . Chem. Soc., Chem. Commun. 1995, 719-720.
(48) El-Sayed, J . Y.; DeBose, C. D.; Coury, L. A.; Roberts, M. F.
Sensitivity of Phospholipase C (Bacillus cereus) Activity to
Phosphatidylcholine Structural Modifications. Biochim. Biophys.
Acta. 1985, 837, 325-335.
J M960434Y