Activated ketones as inhibitors of intracellular Ca2+-dependent and Ca2+-independent phospholipase A2

Article abstract of DOI:10.1021/ja953553w

Studies are reported on two different types of activated ketones as inhibitors of two important intracellular phospholipase A₂s (PLA₂): the group IV 85 kDa Ca²⁺-dependent phospholipase A₂ (cPLA₂) and

Full text of DOI:10.1021/ja953553w

J. Am. Chem. Soc. 1996, 118, 5519-5525
5519
Activated Ketones as Inhibitors of Intracellular Ca2+-Dependent
and Ca2+-Independent Phospholipase A2
Kilian Conde-Frieboes,† Laure J. Reynolds,† Yi-Ching Lio,† Michael R. Hale,§
Harry H. Wasserman,§ and Edward A. Dennis*,†
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California,
San Diego, La Jolla, California 92093-0601, and the Department of Chemistry, Yale UniVersity,
New HaVen, Connecticut 06511
ReceiVed October 24, 1995X
Abstract: Studies are reported on two different types of activated ketones as inhibitors of two important intracellular
phospholipase A2s (PLA2): the group IV 85 kDa Ca2+-dependent phospholipase A2 (cPLA2) and the P388D1 Ca2+
-
independent phospholipase A2 (iPLA2). In a mixed micelle assay, we observed that the reaction progress curve of
cPLA2 in the presence of a trifluoromethyl ketone (TFMK) is linear at pH 7.4, while at pH 9.0 it is nonlinear and
slows with time. An investigation of this discrepancy demonstrated that the TFMKs are slow, tight-binding inhibitors
of the cPLA2 at both pH’s, that the rate of dissociation of the enzyme-inhibitor complex is the same at both pH’s,
but that the rate of association of enzyme and inhibitor is slower at pH 7.4 than at pH 9.0. A novel group of
activated ketone inhibitors has been synthesized that contain a fatty acyl tricarbonyl. These compounds also inhibit
the cPLA2 in the mixed micelle assay. The inhibition of cPLA2 by the tricarbonyls is readily reversible upon dilution
and does not involve slow binding. For both types of inhibitor, no preference for fatty acid chain was observed as
the palmityl analogs inhibited as well as the arachidonoyl analogs, despite the fact that the cPLA2 shows a strong
preference for arachidonoyl-containing phospholipid substrates over palmitoyl-containing substrates. With the iPLA2,
the inhibition by TFMKs is reversible and does not involve slow or tight binding. The tricarbonyls also inhibited
the iPLA2, but were less potent than the TFMKs. Unlike the cPLA2, the iPLA2 does exhibit a fatty acid preference
as the palmityl analogs of both compounds inhibit better than the arachidonoyl analogs. The palmityl TFMK displayed
a 10-fold lower IC50 at pH 9.0 than at pH 7.5, whereas the potency of the tricarbonyl was unchanged in this range.
Thus, the TFMKs inhibit both the cPLA2 and the iPLA2, but the mechanism of inhibition of the two enzymes appears
to be quite different. The tricarbonyls also inhibited both enzymes, but in both cases in a reversible manner and as
such appear to be poorer inhibitors than the TFMKs.
Ketones with enhanced electrophilic reactivity (hydrate-
forming ketones), such as trifluoromethyl ketones (TFMKs),
R-diketones, R-keto esters, and R,â-dioxo esters or vicinal
tricarbonyls,1,2 are a well-established class of inhibitors for
proteases, lipases, and other serine esterases. This type of
inhibitor has now been used to study phospholipase A2s (PLA2),
a class of esterases which carry out a reaction similar to that of
lipases. Of particular interest are two intracellular PLA2s,
the Group IV, 85 kDa, Ca2+-dependent PLA2 (cPLA2)3-5
and the murine P388D1 80 kDa, Ca2+-independent PLA2
(iPLA2).6 The mechanisms of action of these two PLA2s are
not well established, although the cPLA2 reaction is believed
to proceed through an acyl-enzyme intermediate.7,8 Sharp
et al.9 have shown that mutagenesis of a single serine residue
at position 228 of cPLA2 causes a loss of both the enzyme’s
PLA2 and lysophospholipase (lysoPLA) activities. Thus, this
enzyme is presumably a serine esterase. cPLA2 is inhibited by
a TFMK analog of arachidonic acid.10 This arachidonoyl
trifluoromethyl ketone (AATFMK) appears to be a slow, tight-
binding inhibitor of cPLA2, consistent with a serine esterase
mechanism.
Much less is known about the mechanism of iPLA2, except
that this enzyme is inactivated by the haloenol lactone (E)-6-
(bromomethylene)tetrahydro-3-(1-naphthalenyl)-2H-pyran-2-
one,11 a compound which was first introduced as an inhibitor
of chymotrypsin.12 While this enzyme is also inhibited by fatty
acyl TFMKs, this inhibition appears to be readily reversible and
not slow or tight-binding.11 We have now compared the
inhibition of the cPLA2 and the iPLA2 by TFMKs. We have
also studied the inhibition of these enzymes by a novel group
of vicinal tricarbonyl compounds (Chart 1), whose synthesis is
reported herein. Since the two PLA2s differ in their substrate
* Author to whom correspondence should be addressed. Phone: 619-
534-3055. Fax: 619-534-7390. email: edennis@ucsd.edu.
† University of California, San Diego.
§ Yale University.
X Abstract published in AdVance ACS Abstracts, June 1, 1996.
(1) Wasserman, H. H.; Ennis, D. S.; Power, P. L.; Ross, M. J.; Gomes,
B. J. Org. Chem. 1993, 58, 4785-4787.
(2) Henke, S. L.; Hadad, C. M.; Morgan, K. M.; Wiberg, K. B.;
Wasserman, H. H. J. Org. Chem. 1993, 58, 2830-2839.
(3) Dennis, E. A. J. Biol. Chem. 1994, 269, 13057-13060.
(4) Clark, J. D.; Lin, L. L.; Kriz, R. W.; Ramesha, C. S.; Sultzman, L.
A.; Lin, A. Y.; Milona, N.; Knopf, J. L. Cell 1991, 65, 1043-1051.
(5) Sharp, J. D.; White, D. L.; Chiou, X. G.; Goodson, T.; Gamboa, G.
C.; McClure, D.; Burgett, S.; Hoskins, J.; Skatrud, P. L.; Sportsman, J. R.;
Becker, G. W.; Kang, L. H.; Roberts, E. F.; Kramer, R. M. J. Biol. Chem.
1991, 266, 14850-14853.
(8) Trimble, L. A.; Street, I. P.; Perrier, H.; Tremblay, N. M.; Weech,
P. K.; Bernstein, M. A. Biochemistry 1993, 32, 12560-12565.
(9) Sharp, J. D.; Pickard, R. T.; Chiou, X. G.; Manetta, J. V.; Kovacevic,
S.; Miller, J. R.; Varshavsky, A. D.; Roberts, E. F.; Strifler, B. A.; Brems,
D. N.; Kramer, R. M. J. Biol. Chem. 1994, 269, 23250-23254.
(10) Street, I. P.; Lin, H. K.; Laliberte, F.; Ghomashchi, F.; Wang, Z.;
Perrier, H.; Tremblay, N. M.; Huang, Z.; Weech, P. K.; Gelb, M. H.
Biochemistry 1993, 32, 5935-5940.
(6) Ackermann, E. J.; Kempner, E. S.; Dennis, E. A. J. Biol. Chem. 1994,
269, 9227-9233.
(11) Ackermann, E. J.; Conde-Frieboes, K.; Dennis, E. A. J. Biol. Chem.
1995, 270, 445-450.
(7) Reynolds, L.; Hughes, L.; Louis, A. I.; Kramer, R. A.; Dennis, E. A.
Biochim. Biophys. Acta 1993, 1167, 272-280.
(12) Daniels, S. B.; Cooney, E.; Sofia, M. J.; Chakravarty, P. K.;
Katzenellenbogen, J. A. J. Biol. Chem. 1983, 258, 15046-15053.
S0002-7863(95)03553-0 CCC: $12.00 © 1996 American Chemical Society

5520 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Conde-Frieboes et al.
39.8, 31.7, 29.5, 28.3, 28.2, 25.9, 22.8, 14.0. HRMS: C44H56O3P
requires 663.3967 amu, observed 663.3936 amu. IR (cm-1): 3442.9,
2920.5, 2856.7, 1732.4, 1670.0, 1469.9, 1440.5, 1368.9, 1305.7, 1255.1,
1128.6, 1088.0.
Chart 1
Two hundred and fifty milligrams (0.38 mmol) of the ylide derived
from arachidonic acid (above) was dissolved in 2.0 mL of anhydrous
tetrahydrofuran (THF). To this solution were added 0.4 mL of
deionized water and 465 mg (0.76 mmol, 2.0 equiv) of OXONE. After
the mixture was stirred at 25 °C for approximately 2 h, the reaction
was quenched by the addition of 3 mL of brine. The crude tricarbonyl
compound was obtained after washing the reaction mixture with 3 ×
8 mL of dichloromethane, drying the combined organic layers over
MgSO4, filtering, and concentrating in Vacuo. Purification by silica
gel chromatography (20:1 to 10:1, ethyl acetate-hexane) provided 160
mg (0.36 mmol, 95%) of the desired product as a pale yellow oil. 1H-
NMR (300 MHz, CDCl3): δ 5.38 (m, 8H), 4.93 (s, OH), 2.80 (m,
6H), 2.56 (t, 2H, J ) 7.5 Hz), 2.06 (m, 4H), 1.71 (m, 6H), 1.48 (s,
9H), 1.29 (m, 6H), 0.87 (t, 3H, J ) 7.3 Hz). 13C-NMR (75 MHz,
CD3OH): δ 205.9, 168.3, 131.1, 130.1, 129.7, 129.4, 129.1, 129.0,
128.8, 128.7, 84.5, 51.0, 37.2, 32.7, 30.5, 28.2, 28.1, 27.6, 26.6, 24.4,
23.7, 14.5. IR (cm-1): 3452.3, 3008.3, 2954.3, 2929.9, 2862.2, 1733.6,
1729.5, 1457.6, 1395.6, 1366.3, 1280.0, 1255.5, 1152.2, 1103.5, 1097.7.
HRMS or C, H analysis was not performed due to the instability of
the compound; AA-TC decomposes at -20 °C in several days if stored
neat or in neutralized CDCl3. However, when stored in methanol only
slight decomposition occurred over several weeks at -20 °C.
2,3-Dioxooctadecanoic Acid tert-Butyl Ester Monohydrate (PA-
TC). PA-TC was prepared from palmitic acid using the procedure
described above. The analytical data for the ylide (51% yield) are as
follows: 1H-NMR (300 MHz, CDCl3): δ 7.42-7.68 (m, 15H), 2.88
(t, 2H, J ) 7.3 Hz), 1.61 (m, 2H), 1.26 (m, 24H), 0.89 (t, 3H, J ) 6.7
Hz). 13C-NMR (75 MHz, CDCl3): δ 198.3, 168.2, 133.2, 133.1, 131.5,
128.6, 128.4, 127.5, 126.3, 70.3, 49.5, 40.4, 31.9, 29.8, 29.5, 26.0, 22.8,
14.2. HRMS: C40H56O3P requires 615.3967 amu, observed 615.3973
amu. The analytical data for PA-TC (90% yield) are as follows: 1H-
NMR (300 MHz, CDCl3): δ 5.01 (s), 2.56 (t, 2H, J ) 8.6 Hz), 1.49
(s, 9H), 1.25 (s, 26H), 0.88 (t, 3H, J ) 6.9 Hz). 13C-NMR (75 MHz,
CDCl3): δ 203.7, 168.3, 92.5, 84.9, 35.7, 31.9, 29.7, 29.6, 29.5, 29.4,
29.3, 29.1, 27.7, 23.4, 22.7, 14.2. HRMS: C20H41O4 requires 369.3004
amu, observed 369.2998 amu.
specificity with respect to the fatty acid hydrolyzed, the
specificity of the enzymes for the fatty acid in the inhibitor was
compared.
Experimental Section
Materials. L-R-1-Palmitoyl-2-[14C]arachidonoyl phosphatidylcholine
(57 mCi/mmol) was purchased from DuPont New England Nuclear.
L-R-1-Palmitoyl-2-[14C]palmitoyl phosphatidylcholine (58 mCi/mmol)
was purchased from Amersham. Nonradioactive lipids were purchased
from Avanti Polar Lipids. Ultrol grade 4-(2-hydroxyethyl)-1-piper-
azinylethanesulfonic acid (Hepes, free acid) and Triton X-100 were
from Calbiochem. The silica gel used in Dole assays was Davisil grade
633 (200-425 mesh) from Fisher Scientific. Fatty acid free bovine
serum albumin (BSA) was from Sigma. Purified, recombinant human
cPLA2 was provided by Dr. Ruth Kramer, Lilly Research Laboratories.
iPLA2 was purified from P388D1 cells as described previously6 and
mono-Q eluents (purified 100000-fold) with a specific activity of about
1.3 µmol/(min mg) were utilized. Pentadecyl trifluoromethyl ketone
(palmityl trifluoromethyl ketone; PATFMK) and all-cis-5,8,11,14-
nonadecatetraenyl trifluoromethyl ketone (arachidonoyl trifluoromethyl
ketone; AATFMK) were synthesized as described elsewhere.11
General Methods. Thin-layer chromatography (TLC) was carried
out on Analtech silica gel GHLF 250-µm plates. NMR spectra were
obtained on a GE QE 300-MHz NMR spectrometer or a Y 490-MHz
NMR spectrometer with chemical shifts recorded in parts per million
(ppm) from tetramethylsilane and referenced to residual solvent
1
(CHCl3: 7.24 H-NMR; 77.0 13C-NMR). Infrared (IR) spectra were
Standard cPLA2 Assay. cPLA2 activity was measured in a PC
mixed micelle assay in a standard buffer composed of 80 mM Hepes
(pH 7.4), 150 mM NaCl, 10 mM CaCl2, 1 mg/mL BSA, and 1 mM
dithiothreitol (DTT). The assay also contained 1 mM 1-palmitoyl-2-
arachidonoylphosphatidylcholine (PAPC) (with 100000 cpm [14C]-
PAPC), 2 mM Triton X-100, 30% glycerol, and 3.75 µg/mL cPLA2 in
a volume of 200 µL. To prepare the substrate, an appropriate volume
of cold PAPC in chloroform and [14C]PAPC in toluene/ethanol 1:1
solution was evaporated to dryness under a stream of nitrogen. Triton
X-100 (8 mM) in 3X assay buffer (240 mM Hepes (pH 7.4), 450 mM
NaCl, 30 mM CaCl2, 3 mg/mL BSA) was added to the dried lipid to
give 4-fold concentrated substrate solution (4 mM PAPC). This solution
was probe-sonicated on ice (0.5 s on, 0.5 s off for 6 min). Inhibitors
were dissolved in dimethyl sulfoxide (DMSO) at a concentration twenty
times the desired assay concentration.
recorded on a Nicolet FTIR spectrometer. High-resolution mass spectra
(HRMS) were performed by the University of Illinois, Champaign-
Urbana, Mass Spectroscopy Laboratory (Urbana, IL). Radioactivity
was measured with a Packard Tri-Carb model 1600 TR liquid
scintillation analyzer using a Biosafe II scintillation cocktail (Research
Products International, Corp.).
2,3-Dioxo-all-cis-7,10,13,16-docosatetraenoic Acid tert-Butyl Ester
Monohydrate (AA-TC). Arachidonic acid (421 mg, 1.38 mmol, 70%
from Sigma) was stirred at 25 °C in 7.0 mL of anhydrous dichlo-
romethane under nitrogen. To this solution were added 780 mg (2.1
mmol, 1.5 equiv) of tert-butyl 2-(triphenylphosphoranylidenyl)acetate
and 795 mg (2.5 mmol, 1.8 equiv) of EDCl (1-(3-(dimethylamino)-
propyl)-3-ethylcarbodiimide hydrochloride). After approximately 5
min, a catalytic amount of DMAP (N,N-dimethylaminopyridine) was
added. The reaction was then stirred at 25 °C for 8 h. The coupling
reaction was monitored by TLC (silica; 1:1, ethyl acetate-hexane;
Rf(arachidonic acid) ) 0.6, Rf(product) ) 0.9). Extended reaction times
(>12 h) result in the formation of a lower Rf material, most likely the
result of decomposition of the desired product, as this material does
not react under the above coupling reaction conditions. The reaction
was quenched by the addition of 3 mL of brine and washed with 3 ×
10 mL of dichloromethane. The combined organic layers were then
dried over anhydrous MgSO4, filtered, and concentrated in Vacuo. The
crude oil was purified by silica gel chromatography (20:1, ethyl
acetate-hexane) and provided 555 mg (0.77 mmol, 86%) of the desired
ylide. 1H-NMR (300 MHz, CDCl3): δ 7.4-7.7 (m, 15H), 5.32 (m,
8H), 2.80 (m, 6H), 2.05 (m, 2H), 1.55 (m, 4H), 1.29 (m, 6H), 1.05 (s,
9H), 0.87 (t, 3H, J ) 7.3 Hz). 13C-NMR (75 MHz, CDCl3): δ 197.0,
174.1, 133.1, 132.9, 132.2, 131.2, 131.5, 130.6, 130.5, 128.8, 128.7,
128.6, 128.5, 128.4, 128.1, 128.0, 127.9, 127.2, 127.6, 126.6, 40.0,
The substrate solution (50 µL) was placed in a siliconized glass test
tube7 containing 100 µL of 60% glycerol and 10 µL of the inhibitor/
DMSO solution. The control contained DMSO alone, and no inhibitor.
The mixture was vortexed and then bath sonicated for about 10 s. The
reaction was initiated by the addition of 40 µL of enzyme solution
(18.75 ng/µL in a 1.25X assay buffer) and incubated at 40 °C for 40
min. The reaction was quenched by the addition of 2.5 mL of Dole
reagent (2-propanol, heptane, 0.5 M H2SO4; 400/100/20, v/v/v).13 The
amount of hydrolysis was determined using a modified Dole procedure14
as described earlier.15 Blank reactions lacking enzyme were routinely
run and the resulting background hydrolysis was subtracted from the
(13) Dole, V. P. J. Clin. InVest. 1956, 35, 150-154.
(14) Ibrahim, S. A. Biochim. Biophys. Acta 1967, 137, 413-419.
(15) Ulevitch, R. J.; Sano, M.; Watanabe, Y.; Lister, M. D.; Deems, R.
A.; Dennis, E. A. J. Biol. Chem. 1988, 263, 3079-3085.

Inhibition of Phospholipase A2
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5521
reported activities. Inhibitor concentrations expressed as mole fractions
(øs)16 were calculated utilizing the total concentration of lipid present
in the assay (i.e. PC plus Triton X-100 plus inhibitor).
cPLA2 Assay after Preincubation. The activity of cPLA2 in the
presence of AATFMK was also determined in the cPLA2 assay after a
4-h preincubation. Solutions were prepared as described for the
standard cPLA2 assay above, except that the buffer also contained 80
mM glycine. The enzyme solution, 40 µL was incubated in the
presence of inhibitor in 5 µL of DMSO and 100 µL of 60% glycerol
for 4 h prior to the initiation of the reaction by the addition of 50 µL
of substrate solution. The inhibitor concentrations reported represent
that in the final assay.
Progress Curves. Time courses for the cPLA2 were performed
similar to the standard assay but utilized a mixed buffer, prepared by
adding 80 mM glycine to the standard assay buffer, so that time courses
at both pH 7.4 and 9.0 were performed in similar buffers. Higher
amounts of radioactivity were also utilized (300000 cpm/200 µL) so
that the rates at the earlier time points could be more accurately
measured. Other assay components were the same as described above,
except that reactions were initiated in a volume of 2.0 mL. At each
time point a 200-µL aliquot was removed, quenched by being added
to a test tube containing 2.5 mL of Dole reagent, and then extracted as
described above.
iPLA2 Assay. Assays for the iPLA2 contained 100 µM 1,2-
dipalmitoylphosphatidylcholine (DPPC) (with 200000 cpm [14C]-
DPPC), 400 µM Triton X-100, 100 mM Hepes (pH 7.5), 5 mM EDTA,
and 0.1 mM ATP in a final volume of 500 µL. The substrate was
prepared as described previously.6 The inhibitors, in 5 µL of DMSO,
were added to 445 µL of the mixed micellar substrate, vortexed for 10
s, bath sonicated for 10 s, and vortexed again for 10 s. Assays were
initiated by the addition of 50 µL of iPLA2 (an amount sufficient to
generate 2000 cpm in 30 min). Assay solutions were incubated at 40
°C for 30 min, then quenched and extracted using the Dole procedure
described above for the cPLA2 assay. Assays at pH 9.0 were run in a
mixed buffer composed of the assay buffer with 100 mM glycine added.
For time courses, the reactions were initiated in a volume of 5 mL and
at each time point 500 µL of the assay mix (containing 200000 cpm
[14C]-DPPC) were removed, quenched with Dole reagent, and extracted
as before.
Inhibition Recovery Curves. For experiments monitoring the
recovery of cPLA2 from inhibition, 1 mg/mL of enzyme was incubated
for 1 h at 40 °C with 1 mM AATFMK or PA-TC in a volume of 50
µL. The buffer was a combination of enzyme storage buffer and assay
buffer and was composed of 160 mM glycine, 160 mM Hepes, 6.25
mM Tris-HCl, 338 mM NaCl, 20 mM CaCl2, 0.5 mM EDTA, 2.25
mM DTT, 2 mg/mL BSA, 8.75% glycerol, and 6.5% DMSO at pH 7.4
or 9.0. After preincubation, enzymatic activity was followed as
described above for the progress curve experiments. To initiate the
reaction, 3 µL of the preincubation solution was added to 3 mL of the
standard cPLA2 assay mixture (at the same pH as the preincubation)
resulting in a 1000-fold dilution of enzyme and inhibitor into the assay.
From this solution, 200 µL were removed at appropriate times and the
amount of product quantified as described above.
Figure 1. Reaction progress curves of cPLA2 at pH 7.4 in the presence
of PA-TC and AATFMK. The cPLA2 was assayed in the cPLA2 assay
at pH 7.4 in the absence of inhibitor (2) or in the presence of either 50
µM AATFMK (b) or 50 µM PA-TC (0). Each point represents the
average of duplicates.
following equations. For Figures 2 and 5 the IC50s were determined
with:
Vi
Vo
IC50
)
(1)
IC50 + [I]
Where Vi is the inhibited velocity, Vo is the non-inhibited velocity, [I]
is the inhibitor concentration, and IC50 is the concentration of the
inhibitor that produces 50% inhibition.
For the progress and recovery experiments (Figures 3 and 4), the
data were fit to eq 2.17
Vo - Vs
kobs
obst
P ) Vst +
(1 - e-k
)
(2)
Where P is the product, Vs is the equilibrium velocity, Vo is the initial
velocity, kobs is the apparent rate constant for the approach to
equilibrium, and t is time.
Results
cPLA2. Using the cPLA2 assay, we investigated the action
of several fatty acid analogs as inhibitors of cPLA2. Figure 1
shows the reaction progress curve for cPLA2 at pH 7.4 along
with the progress curves observed in the presence of 50 µM
PA-TC or 50 µM AATFMK. All three reactions are linear with
time for over 40 min. This result is in striking contrast to the
nonlinear reaction progress curve observed previously with
AATFMK10 which was indicative of a slow, tight-binding
inhibition.
To determine the reversibility of iPLA2 inhibition by PATFMK, the
iPLA2 was first concentrated 10-fold using a Centricon-10 concentrator
from Amicon. The concentrated iPLA2 (20 µL) in 10 mM Tris-HCl
(pH 7.5), 1 mM EDTA, 1 mM Triton X-100, 1 mM DTT, 1 mM ATP,
and 10% glycerol was preincubated with either 30 µM PATFMK in
DMSO or DMSO alone (2 µL) for 1 h at 40 °C. A 2-µL aliquot was
then removed and diluted 1500-fold into 3 mL of assay mixture
containing 100 µM DPPC (with 400000 cpm per 100 µL), 400 µM
Triton X-100, 5 mM EDTA, 1 mM DTT, 0.8 mM ATP, 100 mM Hepes,
and 100 mM glycine at pH 9.0. At selected time points, a 100-µL
aliquot was removed and the amount of product was quantified as
described above.
The concentration-dependent inhibition of cPLA2 by the fatty
acid tricarbonyls and the fatty acid TFMKs is shown in Figure
2. IC50s were determined by fitting the data to eq 1 using
nonlinear regression and are as follows: AATC ) 67 ( 3 µM
(0.022 mole fraction), PATC ) 57 ( 8 µM (0.019 mole
fraction), AATFMK ) 50 ( 6 µM (0.016 mole fraction), and
PATFMK ) 45 ( 7 µM (0.015 mole fraction). The potency
of the tricarbonyl inhibitors was approximately the same as that
of the trifluoromethyl ketone inhibitors. Note that for both types
of compounds, the palmitic acid analog inhibited as well as the
arachidonic acid analog. Thus, this enzyme did not show a
preference for the fatty alkyl chain attached to the functional
group. Since the reaction progress curves in Figure 1 were
linear, the data shown in Figure 2 were obtained without first
preincubating the enzyme with the inhibitor; preincubation of
the enzyme with the inhibitors for 15 min prior to assaying did
not affect the amount of inhibition observed. The data shown
Data Analysis. Linear least-squares lines were drawn through all
linear data, and the correlation coefficients were all greater than 0.99.
For nonlinear plots the nonlinear regression algorithms of the Sigmaplot
Graphing program (Jandel Scientific) were used to fit the data to the
(16) Lichtenberg, D.; Robson, R. J.; Dennis, E. A. Biochim. Biophys.
Acta ReV. Biomembr. 1983, 737, 285-304.

5522 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Conde-Frieboes et al.
Figure 2. Concentration-dependent inhibition of cPLA2 by tricarbonyls
and trifluoromethyl ketones. cPLA2 was assayed in the cPLA2 assay
at pH 7.4 in the presence of increasing concentrations of inhibitor: (A)
PA-TC (0) or AA-TC (9); (B) PATFMK (O) or AATFMK (b). The
enzyme activity is plotted as the percentage of the control activity,
which was assayed in the absence of inhibitor; 100% activity represents
0.15 µmol/(min mg). Each point represents the average of triplicates.
Figure 3. Reaction progress curves of cPLA2 at pH 9.0 in the presence
of PA-TC, PATFMK, and AATFMK. cPLA2 was assayed in the cPLA2
assay at pH 9.0 in the absence of inhibitor (2) or in the presence of
either 50 µM AATFMK (b), PATFMK (O) or PA-TC (0). Each point
represents the average of duplicates.
in this figure were obtained in the standard buffer but in the
absence of DTT; when the experiments were repeated in the
presence of DTT the results were the same. Also, since fatty
acids bind to BSA we were concerned that the inhibitors under
study might also bind to BSA. However, experiments run in
the absence of BSA showed the same IC50s.
To determine if the pH of the assay affects the nature of the
inhibition, we analyzed the reaction progress curves at pH 9.0.
As shown in Figure 3, the control and PA-TC displayed linear
progress curves at pH 9.0. The presence of 50 µM PA-TC
caused a 58% inhibition of the control activity which was similar
to the 53% inhibition observed at pH 7.4 (Figure 1). Thus, the
potency of this compound is not affected by the pH of the assay.
The reaction progress curves with both AATFMK and
PATFMK are not linear at pH 9.0 since the reaction rates
decrease with time; this is similar to the results reported by
Street et al.10 The parameters were obtained from the nonlinear
regression of the AATFMK data to eq 2 and are Vo ) 0.32 (
0.01 nmol/min, Vs ) 0.41 ( 0.03 nmol/min, and kobs ) 0.16 (
0.01 min-1 while those for PATFMK are Vo ) 0.41 ( 0.03
nmol/min, Vs ) 0.036 ( 0.003 nmol/min, and kobs ) 0.22 (
0.02 min-1. Thus, pH does affect the nature of the inhibition
by these compounds.
Since the reaction progress curve of cPLA2 in the presence
of AATFMK was nonlinear at pH 9.0 but linear at pH 7.4, the
mechanism of AATFMK inhibition is ambiguous. Thus, we
investigated further the behavior of these inhibitors at pH 7.4
and 9.0. The reversibility of cPLA2 inhibition by AATFMK
and PA-TC at pH 7.4 and 9.0 is shown in Figure 4. The enzyme
was preincubated with a high concentration of inhibitor (1 mM)
for 1 h and then diluted 1000-fold into an assay mixture. The
hydrolysis of substrate with time following dilution is shown.
The inhibition of cPLA2 by PA-TC was readily reversible at
both pH’s and full enzymatic activity was recovered immediately
after dilution. In contrast, the TFMK exhibited slow-binding
at both pH’s and little activity was observed immediately after
Figure 4. Reversibility of cPLA2 inhibition by PA-TC and AAT-
FMK: panel A, pH 7.4; panel B, pH 9.0. cPLA2 was incubated for 1
h in the absence of inhibitor (2) or in the presence of PA-TC (0) or
AATFMK (b). After preincubation, the enzyme was diluted 1000-
fold into a substrate solution and the amount of product formed was
monitored with time. Each point represents the average of duplicates.
dilution. The initial velocity, Vo, in the recovery experiments
was <10-10 nmol/min (because of the low value the error in
determining this value was very high) at pH 7.4 and 0.004 (
0.002 nmol/min at pH 9. Enzyme activity recovered with time
following dilution and full enzymatic activity was observed
within 1 h. The final equilibrium velocities were close to the
controls, 0.075 ( 0.003 nmol/min at pH 7.4 and 0.045 ( 0.0008
nmol/min at pH 9.0. Note that this rate of recovery is con-
siderably faster than that observed by Street et al.,10 who saw

Inhibition of Phospholipase A2
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5523
only 14% recovery of activity in 4 h after dilution. This faster
recovery rate could be due to the higher substrate and Triton
concentrations used in this current study as well as to the lack
of 1,2-ditetradecyl-sn-glycero-3-phosphomethanol (DTPM) which
was present in the previous study and which may cause an
increase in the fraction of cPLA2 bound to the micelles.10
In Figure 4, enzymatic activity is recovered very rapidly after
the 1000-fold dilution and appears to reach equilibrium. The
rate constant of this recovery (kobs) is given by eq 3:17
kon
kobs ) koff
+
[I] ) koff + k′on
(3)
(
)
1 + [A]/KA
where KA is the Michaelis constant for the substrate whose
concentration is A, kon is the association rate constant, k′on is
the apparent association rate constant, and koff is the dissociation
rate constant. Since the enzyme-inhibitor complex is diluted
1000-fold into the assay mix, the concentration of inhibitor I
present during recovery is 1 µM which is well below the
estimated IC50 of 50 µM. Therefore the term {kon/(1 + [A]/
KA)}[I] should be small compared to koff. Thus, kobs is approxi-
mately equal to koff. The kobs obtained from the recovery of
cPLA2 from AATFMK inhibition is 0.025 ( 0.007 min-1 at
pH 7.4 and 0.021 ( 0.003 min-1 at pH 9.0. Thus, the off rate
for AATFMK is the same at both pH 7.4 and 9.0.
At pH 7.4, the progress and the recovery experiments for
AATFMK do not seem to agree. The progress experiment
indicates that AATFMK is not a slow-binding inhibitor while
the recovery experiment indicates that it is. One way in which
these differences can be reconciled is if AATFMK reaches
equilibrium much faster or much slower than the rate of substrate
hydrolysis, i.e. equilibrium is achieved either before the first
assay data point or long after the last data point is taken so that
during the assay no appreciable change in rate is observed. To
distinguish between these two possibilities, the concentration-
dependent inhibition of cPLA2 by AATFMK was repeated as
described above with a 4-h preincubation of enzyme and
inhibitor prior to assay. At pH 9.0, after preincubation the
apparent IC50 was 5.6 ( 0.7 µM; at pH 7.4, preincubation
lowered the apparent IC50 from 50 ( 6 to 15 ( 1 µM (data not
shown). The decrease in IC50 after the long preincubation
demonstrates that the amount of inhibition is slowly increasing
and indicates that the rate of reaction of AATFMK with cPLA2
is very slow at physiological pH. Since the koff has been shown
to be the same, this change in kobs is due to the association rate
being slower at pH 7.4, slow enough so that no decrease of the
enzymatic rate is perceived during the reaction progress curve
in Figure 1, thus implying that k′on at pH 7.4 is less than at pH
9.0.
Figure 5. Concentration-dependent inhibition of iPLA2 by PA-TC and
PATFMK. P388D1 iPLA2 was assayed at pH 7.5 (A) or 9.0 (B) in the
iPLA2 assay in the presence of increasing concentrations of PATFMK
(O) or PA-TC (0). The enzyme activity is plotted as the percentage
of the control activity, which was assayed in the absence of inhibitor;
100% activity represents 20 pmol/min in both A and B. Each point
represents the average of duplicates.
surface concentration terms, this IC50 is 2-3-fold lower than
that observed by the TFMKs with the cPLA2. Thus, at pH 7.5
without preincubation the inhibitor appears to work equally well
on both enzymes. The tricarbonyl was a poorer inhibitor for
the iPLA2 than the TFMK, with an IC50 of 16 ( 3 µM (0.031
mole fraction). The AA-TC (not shown) was an even weaker
inhibitor with an IC50 of 52 µM (0.094 mole fraction), consistent
with the fatty acid preference observed with the TFMKs. At
pH 9.0 the IC50 of the tricarbonyl is unchanged, 20 ( 2 µM.
The TFMK, however, was significantly more potent at this pH
with an IC50 of 0.36 ( 0.02 µM (0.0006 mole fraction), over
10-fold lower than was observed at pH 7.5.
The reaction progress curve for iPLA2 in the presence of the
fatty acyl tricarbonyls was linear with time at both pH 7.5 and
9.0, as shown in Figure 6. The reaction progress was also
monitored at pH 9.0 in the presence of PATFMK. This reaction
was also linear with time. This is in contrast to the time-
dependent inhibition of cPLA2 at this pH. Also, after 1 h of
preincubation the inhibition of iPLA2 by PATFMK was fully
reversible at pH 9.0 immediately following dilution (Figure 7).
Thus, the mechanism of inhibition of iPLA2 by the trifluorom-
ethyl ketones appears to be different from that of the cPLA2.
iPLA2. In addition to inhibiting the cPLA2, we have
previously demonstrated that the fatty acid TFMKs also inhibit
the mouse macrophage iPLA2.11 Unlike the cPLA2, the iPLA2
showed a preference for the fatty acid moiety and was about
4-fold more potent on PATFMK than on AATFMK.11 These
previous studies with the iPLA2 were performed at pH 7.5 where
the enzyme displays a linear progress curve in the presence of
the TFMKs. In light of the results described above for cPLA2,
the behavior of iPLA2 was investigated with the TFMKs at pH
9.0. Also, the ability of the tricarbonyls to inhibit the iPLA2
was tested. The concentration-dependent inhibition of iPLA2
by PATFMK and PA-TC is shown in Figure 5 at pH 7.5 (A)
and 9.0 (B). As before,11 PATFMK inhibited the iPLA2 at pH
7.5 with an IC50 of 3.3 ( 0.3 µM (0.006 mole fraction). In
Discussion
The use of trifluoromethyl ketones as inhibitors of proteases
and other hydrolases was first described by Abeles and
co-workers.18,19 These activated ketones form hydrates in
aqueous solution. They can also form an enzyme-bound
hemiketal upon reaction with the hydroxyl group of a serine
residue. Both the hemiketal and the hydrate forms of the
TFMKs have been observed with the proteases; the hemiketal
form inhibits the serine proteases19,20 and the tetrahedral hydrate
(18) Gelb, M. H.; Svaren, J. P.; Abeles, R. H. Biochemistry 1985, 24,
1813-1817.
(19) Imperiali, B.; Abeles, R. H. Biochemistry 1986, 25, 3760-3767.
(17) Williams, J. W.; Morrison, J. F. Methods Enzymol. 1979, 63, 437-
467.

5524 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Conde-Frieboes et al.
Scheme 1
utilized as controls for comparison with the fatty acyl tricar-
bonyls; however, the unexpected linear reaction progress curve
with the TFMKs at pH 7.4 led us to investigate this inhibition
further. We have found that AATFMK is a slow, tight-binding
inhibitor of cPLA2 at both pH 7.4 and 9.0, but at pH 7.4 the
inhibition is considerably slower than is observed at pH 9.0.
Our results indicate that the rate of dissociation of the
enzyme-inhibitor complex is the same at pH 7.4 as at pH 9.0.
This result is similar to that observed for the dissociation of
the chymotrypsin/TFMK complex21 where koff showed no pH
dependence between pH 4.0 and 9.5 indicating the lack of an
ionizable group in the enzyme-inhibitor complex in this pH
range. Unlike koff, the apparent association rate constant (k′on)
of AATFMK with cPLA2 is smaller at pH 7.4 than at pH 9.0.
As shown in Scheme 1, the binding of the inhibitor to the
enzyme involves several steps:22 the dehydration of the hydrate
form of the inhibitor (which is predominant in aqueous solution)
to the free ketone, the initial binding of the inhibitor to the
enzyme, and a second enzymatic step, presumably the formation
of a covalent enzyme-inhibitor hemiketal. k′on is a function
of all these steps. The different association rates observed at
different pH’s could reflect an effect of pH on any of these
steps, particularly on the rate of dehydration of the hydrate to
the ketone form of the inhibitor or on the ionization of certain
groups in the enzyme active site. The effect of pH may also
be on substrate binding to the active site since k′on contains the
term 1 + [A]/KA.
Figure 6. Reaction progress curves of iPLA2 in the presence of
tricarbonyls and trifluoromethyl ketones. iPLA2 activity was measured
at pH 7.5 (A) in the absence of inhibitor (2) or in the presence of 32
µM AA-TC (9) or 32 µM PA-TC (0) or at pH 9.0 (B) in the absence
of inhibitor (2) or in the presence of 24 µM PA-TC (0) or 0.4 µM
PATFMK (O). Each point represents the average of duplicates.
The fatty acyl tricarbonyls appeared to behave similarly to
the TFMKs. In particular, the reaction progress curves and the
concentration-dependent inhibition by the tricarbonyls at pH 7.4
(in the absence of preincubation) were practically superimpos-
able with those of the TFMKs. However, the linear reaction
progress curve at pH 9.0 and the immediate reversibility of PA-
TC inhibition following preincubation clearly differentiate these
inhibitors from the TFMKs and indicate that the tricarbonyls
are not slow, tight-binding inhibitors.
According to Morrison and Walsh,23 the majority of slow,
tight-binding inhibitors involve the rapid formation of an
inhibitor-enzyme complex followed by a slower isomerization
or conformational change. In the case of the TFMKs, the
inhibitor could bind rapidly followed by the slower formation
of the enzyme-inhibitor hemiketal as shown by Street et al.10
With the tricarbonyls, the inhibitor may bind quickly but be
unable to form the hemiketal. As evidenced by 13C-NMR data,
the tricarbonyls are isolated as stable hydrates of the central
ketone, similar to that observed previously with the peptidyl
tricarbonyls.1,2,24 The most reactive carbonyl for this inhibitor
may not be positioned properly in the active site for attack by
the serine residue. Such a mechanism would explain why the
TFMKs and tricarbonyls appear to behave so similarly at pH
7.4 in the absence of preincubation. Alternatively, it is possible
that the tricarbonyls bind differently in the active site or that
they do not act at the catalytic site but rather bind to another
site on the enzyme or interfere with the binding of the cPLA2
to the surface.
Figure 7. Reversible inhibition of iPLA2 by PATFMK. iPLA2 was
preincubated for 1 h in the absence (2) or presence (O) of 30 µM
PATFMK and then diluted 1500-fold into an assay mixture at pH 9.0.
At the indicated time points, an aliquot was removed and the amount
of product formed was measured. Each point represents the average
of duplicates.
form inhibits the aspartyl proteases and the zinc metallopro-
teases.18 At basic pH, an anion of the hydrate or hemiketal
can be formed. The TFMKs have recently been applied to the
study of phospholipases by Street et al.,10 who demonstrated
that AATFMK is a reversible, slow, tight-binding inhibitor of
the Group IV cytosolic, 85 kDa, Ca2+-dependent PLA2. NMR
studies8 suggested that the inhibitor was bound as the anion of
the hemiketal; however, the possibility that the bound species
was the monoanion of the hydrate could not be excluded.
Our studies with activated ketone inhibitors of cPLA2 began
with the evaluation of a novel set of fatty acyl tricarbonyl
compounds. These compounds were designed based on the
peptidyl tricarbonyls which have previously been described as
inhibitors of the proteases.1 The fatty acyl TFMKs were initially
(21) Brady, K.; Liang, T. C.; Abeles, R. H. Biochemistry 1989, 28, 9066-
9070.
(22) Brady, K.; Abeles, R. H. Biochemistry 1990, 29, 7608-7617.
(23) Morrison, J. F.; Walsh, C. T. AdV. Enzymol. 1988, 61, 201-301.
(24) Wiberg, K. B.; Morgan, K. M.; Maltz, H. J. Am. Chem. Soc. 1994,
116, 11067-11077.
(20) Parisi, M. F.; Abeles, R. H. Biochemistry 1992, 31, 9429-9435.

Inhibition of Phospholipase A2
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5525
One striking fact observed with the TFMKs and tricarbonyls
is that for both types of inhibitors the palmityl analog inhibited
as well as the arachidonoyl analog. cPLA2 has been shown by
several groups to prefer arachidonoyl-containing phospholipids
as substrates.25-28 However, the cPLA2 also exhibits a lyso-
phospholipase activity7,29 and readily hydrolyzes palmitoyl
lysophosphatidylcholine in a Triton/phosphatidylcholine/lyso-
phosphatidylcholine mixed micelle assay at a rate comparable
to the PLA2 activity (K. Conde-Frieboes and E. A. Dennis,
unpublished data). Since the enzyme will hydrolyze both fatty
acid esters, it is reasonable that a fatty acid chain preference
for these inhibitors was not observed.
In preliminary experiments with the iPLA2 from the murine
macrophage-like cell line, P388D1, we found that the inhibition
of the enzyme by TFMKs did not appear to involve slow
binding. In light of the results with the cPLA2, we have now
reexamined this inhibition under more stringent conditions. We
show here that even at pH 9.0 the reaction progress curve is
linear in the presence of PATFMK and this inhibition is readily
reversible after a 1-h preincubation. Also, we found the TFMK
to be 10-fold more potent at pH 9.0 than at pH 7.5. Thus, the
mechanism of inhibition of the fluoromethyl ketones with the
iPLA2 differs from that observed with the cPLA2. While a
serine hemiketal of AATFMK appears to be the form of
inhibition with the cPLA2,8 it is possible that the iPLA2 is
inhibited by the tetrahedral hydrate, or perhaps it forms a
hemiketal with a different amino acid residue.
In comparison with the TFMKs, the fatty acyl tricarbonyls
were weaker inhibitors for the iPLA2. The palmityl tricarbonyl
inhibited better than the arachidonoyl tricarbonyl, consistent with
the TFMKs and with the previously noted substrate preferences
for this enzyme.6 As with the TFMKs, the reaction progress
curves with the tricarbonyls were linear; however, the potency
of the tricarbonyls was unaffected by pH. Thus, the tricarbonyls
were poorer inhibitors than the TFMKs for both the cPLA2 and
the iPLA2.
An important conclusion of this study is that compounds that
were thought to be specific inhibitors of the cPLA2 are also
capable of inhibiting the iPLA2. Thus, particular caution should
be taken when using cPLA2 or iPLA2 inhibitors in whole cell
studies, as these inhibitors may affect both enzymes. Also,
remarkably, while the cPLA2 is specific for arachidonoyl-
containing phospholipid substrates, arachidonoyl-containing
inhibitors in these studies were no better than palmityl-
containing inhibitors.
Acknowledgment. This study was supported by grants to
E.A.D. from NIH (GM 20,501, GM 51,606, and HD 26,171)
and from Lilly Research Laboratories and by grants to H.H.W.
from NIH. K.C.-F. is the recipient of a Deutsche Forschungs-
gemeinschaft (DFG) postdoctoral fellowship. Y.-C.L. is the
recipient of a Lucille P. Markey Charitable Trust Fellowship.
We would like to thank Dr. Richard Ulevitch and Lois Kline
(Research Institute of Scripps Clinic) for generously growing
the P388D1 cells employed in the preparation of the iPLA2 and
Dr. Ruth Kramer for generously providing the cPLA2. We thank
Ray Deems for many helpful discussions during the course of
this work and Dr. Kenneth Wiberg for a helpful discussion on
hydration effects.
(25) Leslie, C. C.; Voelker, D. R.; Channon, J. Y.; Wall, M. M.; Zelarney,
P. T. Biochim. Biophys. Acta 1988, 963, 476-492.
(26) Clark, J. D.; Milona, N.; Knopf, J. L. Proc. Natl. Acad. Sci. U.S.A.
1990, 87, 7708-7712.
(27) Kramer, R. M.; Roberts, E. F.; Manetta, J.; Putnam, J. E. J. Biol.
Chem. 1991, 266, 5268-5272.
(28) Takayama, K.; Kudo, I.; Kim, D. K.; Nagata, K.; Nozawa, Y.; Inoue,
K. FEBS Lett. 1991, 282, 326-330.
(29) Leslie, C. C. J. Biol. Chem. 1991, 266, 11366-11371.
JA953553W