Inhibition of the complete set of mammalian secreted phospholipases A 2 by indole analogues: A structure-guided study

Article abstract of DOI:10.1016/j.bmc.2004.01.022

Structure-guided design was employed in a search for potent and selective inhibitors of mammalian secreted phospholipases A₂ (sPLA ₂s). Using the X-ray structures of human groups IIA and X sPLA ₂s (hGIIA and hGX) as templates, homology structural models were made for the other human and mouse sPLA₂s (hGIB, mGIB, mGIIA, mGIIC, hGIID, mGIID, hGIIE, mGIIE, hGIIF, mGIIF, hGV, mGV, and mGX). Me-Indoxam is a previously discovered indole analogue that binds tightly to many sPLA ₂s, and the X-ray structure of the hGX-Me-Indoxam complex was determined at a resolution of 2.0 A. Modeling suggests that the residues near the N₁-substituent of Me-Indoxam vary significantly among the mammalian sPLA₂s, and therefore a library of 83 N₁- variants was prepared by parallel synthesis. Several Me-Indoxam analogues bearing a 4-(2-oxy-ethanoic acid) side chain were potent inhibitors (IC ₅₀ <0.05 μM) of hGIIA, mGIIA, mGIIC, hGIIE, mGIIE, hGV, and mGV, while they displayed intermediate potency (0.05-5 μM) against hGIB, mGIB, hGX, and mGX, and poorly inhibited (>5 μM) hGIID, mGIID, hGIIF, and mGIIF. Me-Indoxam analogues bearing a 5-(4-oxy-butanoic acid) side chain were generally less potent inhibitors. Although no compounds were found to be highly specific for a single human or mouse sPLA₂, combinations of Me-Indoxam analogues were discovered that could be used to distinguish the action of various sPLA₂s in cellular events. For example, Me-Indoxam and compound 5 are approximately 5-fold more potent on hGIIA than on hGV, and compound 21 is 10-fold more potent on hGV versus hGIIA.

Full text of DOI:10.1016/j.bmc.2004.01.022

Bioorganic & Medicinal Chemistry 12 (2004) 1737–1749
Inhibition of the complete set of mammalian secreted
phospholipases A2 by indole analogues:
a structure-guided study§
Brian P. Smart,a Ying H. Pan,b Amanda K. Weeks,a James G. Bollinger,a
Brian J. Bahnsonb,* and Michael H. Gelba,*
aDepartments of Chemistry and Biochemistry, University of Washington, Seattle, WA 98195, USA
bDepartments of Chemistry and Biochemistry, University of Delaware, Newark, DE 19716, USA
Received 20 November 2003; accepted 7 January 2004
Abstract—Structure-guided design was employed in a search for potent and selective inhibitors of mammalian secreted phospho-
lipases A2 (sPLA2s). Using the X-ray structures of human groups IIA and X sPLA2s (hGIIA and hGX) as templates, homology
structural models were made for the other human and mouse sPLA2s (hGIB, mGIB, mGIIA, mGIIC, hGIID, mGIID, hGIIE,
mGIIE, hGIIF, mGIIF, hGV, mGV, and mGX). Me-Indoxam is a previously discovered indole analogue that binds tightly to
many sPLA2s, and the X-ray structure of the hGX-Me-Indoxam complex was determined at a resolution of 2.0 A. Modeling sug-
gests that the residues near the N1-substituent of Me-Indoxam vary significantly among the mammalian sPLA2s, and therefore a
library of 83 N1-variants was prepared by parallel synthesis. Several Me-Indoxam analogues bearing a 4-(2-oxy-ethanoic acid) side
chain were potent inhibitors (IC50 <0.05 mM) of hGIIA, mGIIA, mGIIC, hGIIE, mGIIE, hGV, and mGV, while they displayed
intermediate potency (0.05–5 mM) against hGIB, mGIB, hGX, and mGX, and poorly inhibited (>5 mM) hGIID, mGIID, hGIIF,
and mGIIF. Me-Indoxam analogues bearing a 5-(4-oxy-butanoic acid) side chain were generally less potent inhibitors. Although no
compounds were found to be highly specific for a single human or mouse sPLA2, combinations of Me-Indoxam analogues were
discovered that could be used to distinguish the action of various sPLA2s in cellular events. For example, Me-Indoxam and com-
pound 5 are approximately 5-fold more potent on hGIIA than on hGV, and compound 21 is 10-fold more potent on hGV versus
hGIIA.
# 2004 Elsevier Ltd. All rights reserved.
1. Introduction
in this process is complicated by the fact that mammals
contain a large number of these enzymes. Humans con-
tain the groups IB, IIA, IID, IIE, IIF, III, V, X, and
XIIA sPLA2s (designated hGIB, hGIIA, etc.), whereas
the mouse contains all of these sPLA2s (designated
mGIB, mGIIA, etc.) plus the group IIC enzyme (which
is present in the human genome as a pseudogene).4,5
Deciphering the role played by sPLA2s in arachidonic
acid release as well as in other physiological events
would be greatly aided by the availability of potent and
specific inhibitors of each mammalian homologue.
The phospholipase A2 class of enzymes catalyze the
hydrolysis of the sn-2 ester of glycero-phospholipids to
release fatty acids and lysophospholipids.1 One or more
of these enzymes is thought to initiate pro-inflammatory
cascades by liberating arachidonic acid from membrane
phospholipids for the biosynthesis of the eicosanoids
(prostaglandins, leukotrienes, and others). The current
thinking is that arachidonic acid liberation from the
membranes of mammalian cells is catalyzed by the 87-
kDa cytosolic phospholipase A2 (group IVA) and by
one or more low molecular weight, secreted phospho-
lipases A2 (sPLA2s).2,3 Understanding the role of sPLA2s
Of the numerous known sPLA2 inhibitors, many are
non-specific agents that act by perturbing the physical
structure of the membrane substrate.6 sPLA2s are
water-soluble enzymes that must bind to the membrane
surface to access their water-insoluble substrates,
(interfacial enzymes).7 Methods have been developed
for the proper quantitative analysis of inhibitors of
§Supplementary data associated with this article can be found, in the
online version at, doi:10.1016/j.bmc.2004.01.022
* Corresponding author. Tel.: +1-206-543-7142; fax: +1-206-685-
8665; e-mail: gelb@chem.washington.edu
0968-0896/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.01.022

1738
B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
interfacial enzymes that allow ‘true’, active site-directed,
competitive inhibitors to be investigated.6,8,9 Inhibitors
that have been convincingly established to be potent
and active site-directed competitive inhibitors of sPLA2s
either by proper kinetic analysis or by determination of
the enzyme-inhibitor X-ray structure include phospho-
lipid analogues (i.e., sn-2 amides and phosphonates)10À14
as well as non-phospholipid analogues including the
plant-derived alkaloid aristolochic acid,15 1,3-dioxane-
4,6-dione-5-carboxamides,16,17 fatty acid amides,18 aro-
matic sulfonamides (i.e., SB-203347),19 plant-derived
sterols,20 bis-carboxylates,21,22 complex poly-p-hydroxy
benzoate natural products,23À25 highly substituted pyr-
azoles,26 short peptides,27,28 vitamin E analogues,29 and
indole analogues.30À33 The X-ray structure of some of
these compounds bound to hGIIA, porcine and bovine
pancreatic sPLA2s, and venom sPLA2s have been
determined.15,17,27À30,34À41
vitro, which accounts for their ability to act on the
phosphatidylcholine-rich outer leaflet of the mammalian
cell plasma membrane.42À45 Thus, among the sPLA2
family members, group V and X enzymes seem to be the
strongest candidates for playing a role in arachidonic
acid liberation for eicosanoid production. For example,
a role for mGV in lipopolysaccharide-induced arachi-
donic acid release from a murine macrophage-like
cell line has been reported.49 The group V and X
sPLA2s also show relatively high activity on serum
lipoproteins, and may play a role in the development of
atherosclerosis.50,51
In the present study, we extend our structural studies of
hGX52 by reporting the 2.0 A X-ray crystal structure of
this sPLA2 complexed with Me-Indoxam. This structure
was then used as a guide for preparing analogues of the
lead compound Me-Indoxam with the hope of identify-
ing compounds with improved potency and specificity
for one or more of the various human and mouse
sPLA2s. We report the preparation, by small-scale, par-
allel organic synthesis, of a library of 83 Me-Indoxam
analogues, and provide inhibition values of many of
these compounds toward the full set of human and
mouse sPLA2s.
Of these sPLA2 inhibitors, the indole analogues devel-
oped by workers at Lilly and Shionogi laboratories are
particularly appealing for further study as they appear
to have properties most suitable for use in studies with
cultured mammalian cells and are the most generally
potent among the full set of human and mouse
sPLA2s.42 We have tested the indole analogue Me-
Indoxam (Fig. 1) on all of the human and mouse
sPLA2s except the human and mouse group III and
mouse group XIIA enzymes, and found it to be a potent
inhibitor of mGIIA, hGIIA, mGIIC, mGIIE, hGIIE,
mGV, and hGV (IC50ꢀ0.01–0.02 mM), and a modestly
potent inhibitor of mGIB, hGIB, mGX, and hGX (IC50
ꢀ0.1–1 mM).42 On the other hand Me-Indoxam is a
poor inhibitor of mGIID, hGIID, mGIIF, hGIIF, and
hGXIIA (IC50>10 mM).42
2. Results
2.1. Crystal structure of the complex of hGX with
Me-Indoxam
The X-ray crystal structure of hGX-Me-Indoxam was
solved in a protein crystal form with two independent
subunits in the asymmetric unit. The overall structure is
well refined with good geometry with an Rworking of
0.219 and an Rfree of 0.267. These statistics are similar
to the converged Rfactors for the ligand free form of
hGX that was previously reported from a protein crys-
tal also with symmetry of the C2221 space group and
similar unit cell lengths.52 Furthermore, there are par-
allels between the two structures with one of the two
subunits in each structure exhibiting more disorder than
the other, with a notable increase in the individual
B-factors. Importantly, the 2-phenyl-benzyl group of
the Me-Indoxam inhibitor is well ordered in subunit A
and completely disordered in subunit B. The probable
cause of this phenomenon is a difference in protein–
protein crystal contacts between the two subunits, par-
ticularly near the 2-phenyl of the 2-phenyl-benzyl group
of the inhibitor. This observed disorder in subunit B
indicates the lack of multiple contacts between the pro-
tein and the 2-phenyl group of the N1-substituent of the
inhibitor bound to the same subunit. In our design of
inhibitors, we hoped to take advantage of the N1-sub-
stituent with a goal to achieve inhibitors that are specific
for the various sPLA2 enzymes. For the remainder of
the structure description, the more ordered subunit A
will be used for analyses and figures.
The groups V and X sPLA2s are particularly appealing
for further study because these enzymes display high
enzymatic activity when added exogenously to a variety
of mammalian cells,43À48 whereas the other mammalian
sPLA2s are poorly active in this assay42 (the human and
mouse group III enzymes remain to be analyzed on
mammalian cells). These two sPLA2s display the unique
ability to bind to phosphatidylcholine-rich vesicles in
Figure 1. Structure of Me-Indoxam. Atom labels are as given in the
pdb file (except for carbons which are labeled C1, C2, etc., whereas
they are labeled 1, 2, 3, etc. in the figure). The atom numbering of the
indole ring is different than the conventional IUPAC ring numbering.
The latter is used to name the compounds synthesized in this study.
The Me-Indoxam inhibitor is complexed to the catalytic
calcium through interactions between the amide car-
bonyl oxygen and the carboxyl oxygen atoms. A simu-

B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
1739
Gly-28, Leu-29, Gly-30, Cys-43, Tyr-50, and Ile-94 as
shown in Figure 3. These side chain interactions form
the basis of predicted binding affinity differences of Me-
Indoxam between the various sequences for the sPLA2
enzymes (discussed below).
The surface of the hGX sPLA2 that binds to lipid water
interfaces (i-face) is composed of polar and non-polar
residues and displays a preference to zwitterionic inter-
faces.52 The 2-phenyl group of the 2-phenyl-benzyl
group of the Me-Indoxam inhibitor protrudes out of the
enzyme active site opening. The minimal protein side-
chain interactions with the 2-phenyl group of Me-
Indoxam are consistent with an observation of a dis-
ordered N1-substituent in subunit B of this structure.
Furthermore, the hGX-Me-Indoxam inhibitor complex
suggests that for the enzyme-inhibitor complex bound
to the membrane surface, the 2-phenyl of the 2-phenyl-
benzyl group penetrates into the membrane interface,
possibly the polar head group region or just below this
region.
Figure 2. Stereo ball and stick depiction of the hGX-Me-Indoxam
complex. The active site is shown including residues H46 and D47, the
Ca2+-binding loop and the competitive inhibitor Me-Indoxam. The
Ca2+ is seven coordinate with three carbonyl backbone ligands (F26,
G28, and G30), a bidentate coordination from D47, and coordination
to the amide carbonyl oxygen and carboxyl oxygen atoms of Me-
Indoxam. Hydrogen bonding of the nitrogen and oxygen atoms of
H46 and D47, respectively, are shown by dotted lines. Difference
electron density from a composite simulated annealing omit map con-
toured at 1.0 sigma is displayed around the Me-Indoxam ligand
(coefficients 2Fo–Fc). The figure was rendered using the programs
MOLSCRIPT,62 POVSCRIPT (http://people/brandeis.edu/ꢀfenn/
povscript/), and POVRAY (http://www.povray.org).
2.2. Design of the Me-Indoxam library
We desired to prepare Me-Indoxam analogues with
improved potency and selectivity for the various mam-
malian sPLA2s. The X-ray structures of groups I, II,
and X sPLA2s show that these enzymes are structurally
related,52,53 and based on sequence homology we
anticipate that the structure of group V sPLA2s will also
be similar. We built homology structural models of
several mammalian sPLA2s (hGIB, mGIB, mGIIA,
mGIIC, hGIID, mGIID, hGIIE, mGIIE, hGIIF,
mGIIF, hGV, mGV, and mGX) using the X-ray struc-
tures of hGIIA and hGX as templates. This was done
with the Insight-II software (MSI) simply by replacing
peptide segments of hGIIA and hGX with the homo-
logous segments of the other sPLA2s. Many of the
active site residues are conserved among all the sPLA2s,
and no attempt was made to minimize the energy of the
structural models using molecular dynamics. Addition-
lated annealing omit map of the active site (Fig. 2)
demonstrates a well refined structure and provides
convincing evidence of the proper modeling of each
atom of the Me-Indoxam inhibitor in subunit A of the
crystal structure. The resulting calcium coordination of
hGX sPLA2 shows the typical 7-coordinate geometry of
a ligand bound form of the sPLA2 enzyme37 as shown in
Figure 2. Additional hydrogen bonds exist between the
hydrogen on the amide nitrogen of Me-Indoxam and
the residues His-46 and Asp-47. The only other hydro-
gen bonding interactions come from three water mole-
cules that are interacting with the carbonyl oxygen of
the carboxyl group of Me-Indoxam. The balance of
interactions come from hydrophobic interactions with
side chains from residues Ile-2, Leu-5, Ala-6, Met-21,
Figure 3. hGX-Me-Indoxam contacts. Side chain residues of hGX that lie within 4 A of Me-Indoxam are displayed in ball and stick format for
residues Ile-2, Leu-5, Ala-6, Met-21, Leu-29, Cys-43, Tyr-50, and Ile-94. Coordination of Me-Indoxam to the Ca2+ is shown. The side chain con-
tacts from His-46 and Asp-47 and main chain contacts within residues of the calcium binding loop (residues 26–30) are not shown for clarity and can
be observed in Figure 2. Cys-43 is disulfide linked to another cysteine residue (not shown).

1740
B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
ally, no attempt to model the structure of hGIII and
hGXIIA was made since these enzymes show amino
acid sequence homology to the groups I, II, V and X
sPLA2s only in short stretches that include the catalytic
residues. With notable exceptions (see below), the
residues that contact the indole core, the glyoxamide
side chain, and the 4-(2-oxy-ethanoic acid) side chain of
Me-Indoxam are well conserved among the groups I, II,
V, and X sPLA2s. On the other hand, residues that are
likely to contact the 2-phenyl-benzyl side chain attached
to the N1-position of Me-Indoxam differ significantly
among the various group I, II, V, and X sPLA2s. Thus,
it seemed reasonable to prepare a library of Me-
Indoxam analogues in which the N1-substituent is var-
ied. Since the analogue of Me-Indoxam in which the 4-
(2-oxy-ethanoic acid) side chain is replaced with hydro-
gen and the longer 4-oxy-butanoic acid side chain is
present at the 5-position of the indole core is also
reported to be a potent inhibitor of hGIIA,54 we also
prepared a library of these compounds in which the N1-
substituent was varied.
Next, the acyl chloride intermediate was converted to
the amide by bubbling ammonia gas into the solution.
Common intermediates (C) and (G) were used to pre-
pare the library of Me-Indoxam analogues (D) and (H),
respectively. A single solution of (C) or (G) in THF was
treated with NaH leading to deprotonation of the N1-
hydrogen. This mixture was distributed into several
small vials and allowed to react with various alkyl bro-
mides to form the N1-substituted indole analogues.
Treatment with neat trifluoroacetic acid led to depro-
tection of the tert-butyl ester of the side chain at the 4-
or 5-position of the indole ring to give the desired
library members (D) and (H). Each compound was
purified by HPLC and confirmed to have the correct
structure by electrospray ionization mass spectrometry.
The structure of the 48 indole analogues with a 4-(2-
oxy-ethanoic acid) chain (compounds 1–48) are shown
in Table 1, and the structures of the 35 indole analogues
with a 5-(4-oxy-butanoic acid) chain (compounds 49–
83) are shown in Table 3.
We desired an accurate and rapid method to quantify
the amount of Me-Indoxam analogues that were pre-
pared by multiple, small-scale, parallel organic synthesis.
Thus, we used low specific activity, carbon-14-labeled
oxalyl chloride to form the glyoxamide side chain,
which permitted the concentration of Me-Indoxam
analogues in stock solutions to be quantified by scintil-
lation counting. Labeled oxalyl chloride was synthesized
from commercially available [14C]oxalic acid by reaction
with phosphorous pentachloride.
2.3. Chemistry
Figure 4 shows the synthetic scheme used to prepare a
library of indole analogues containing the common
functionalities, a 2-methyl group, a 3-glyoxamide side
chain and either a 4-(2-oxy-ethanoic acid) or a 5-(4-oxy-
butanoic acid) side chain, and with varying substituents
at the N1-position. The N1-substituent is introduced late
in the synthesis, which facilitated the preparation of a
large number of analogues in which this substituent was
varied. Commercially available 4- and 5-hydroxy-2-
methyl indoles (A) and (E), respectively (Fig. 4), were
treated with potassium carbonate in refluxing acetone,
resulting in selective deprotonation of the hydroxyl
group. Alkylation of the alkoxide with tert-butyl
2-bromoacetate or tert-butyl 4-bromobutanoate provided
indoles (B) and (F), respectively. The 3-glyoxamide side
chain was added to the indole in two steps. First, indole
(B) or (F) was added to a dilute solution of oxalyl chlo-
ride in CH2Cl2. High dilution conditions were necessary
to avoid the troublesome dimerization of the indole.
2.4. sPLA2 inhibition studies
We developed a high throughput, fluorometric assay of
sPLA2s that can be carried out in multi-well plates.43
The full set of 48 Me-Indoxam analogues with a 4-(2-
oxy-ethanoic acid) side chain were tested on the set of
human sPLA2s (hGIB, hGIIA, hGIID, hGIIE, hGIIF,
hGV, hGX, and hGXIIA) (Table 1) and on the set of
mouse sPLA2s (mGIB, mGIIA, mGIIC, mGIID,
mGIIE, mGIIF, mGV, and mGX) (Table 2). Indole
analogues were tested at a single concentration and re-
tested at a second concentration if inhibition approa-
ched 100% or if inhibition was <30%. Studies with
mGXIIA were not carried out because the Me-Indoxam
analogues were uniformly poor inhibitors of hGXIIA,
and mGXIIA is 96% identical in amino acid sequence
to hGXIIA (the active site residues of these two
enzymes are identical). Studies with hGIII were limited
to testing only analogue 21 because of limited amount
of enzyme. It is anticipated that other indole analogues
would not be potent against hGIII given that indole 21
is a weak inhibitor, and that the active site structure of
hGIII is predicted to be different from that of groups I,
II, V, and X sPLA2s (based on the X-ray structure of
the related group III sPLA2 from honey bee venom36).
Me-Indoxam analogues with a 5-(4-oxy-butanoic acid)
side chain were tested (Table 3) and were found to be
much less potent sPLA2 inhibitors than the 4-(2-oxy-
ethanoic acid)-containing analogues. Thus the former
were tested only on hGIIA and hGX sPLA2s.
Figure 4. Synthesis of Me-Indoxam analogues. The reagents used are:
(i) K2CO3, Br(CH2)nCO2t-Bu (n=1 or 3), acetone, reflux; (ii) oxalyl
chloride, CH2Cl2 followed by NH3; (iii) NaH, RBr, THF followed by
neat CF3CO2H.

B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
1741
Table 1. Inhibition of human sPLA2s by indoles bearing a 4-(2-oxy-ethanoic acid) side chain
R Group
Compd
% Inhibition against human sPLA2
IB
0.5
mM
IIA
0.33
mM
IIA
0.025
mM
IID
5
IIE
0.5
IIE
0.025
mM
IIF
5
III
5
V
V
X
XIIA
5
0.33
mM
0.10
mM
0.33
mM
mM
mM
mM
mM
mM
2,6-Cl2-C6H3CH2
Phthalimido(CH2)4
C6F5CH2
1
71
25
63
78
14
37
59
58
36
10
9
100
2
45
5
17
9
98
86
68
17
0
46
37
14
65
22
44
38
60
52
23
22
30
16
44
45
40
41
29
12
52
37
43
17
31
41
22
44
31
37
22
8
95
50
32
93
10
82
93
93
69
47
71
63
4
86
34
3
30
15
9
6
7
2
3
33
100
61
86
96
97
29
82
86
64
7
6
20
63
0
72
96
9
4-Br-C6H4CH2
C6H5COCH2
4
53
27
29
30
33
4
80
0
85
8
29
0
31
14
2
5
36
96
3-(CH2Br)-C6H4CH2
3-Cl-C6H4CH2
6
9
81
86
80
42
84
47
13
0
59
81
84
44
20
37
33
0
19
33
15
13
0
7
31
49
34
0
100
100
100
100
95
16
24
13
27
11
24
0
2-NapthaleneCH2
4-CN-C6H4CH2
2-CN-C6H4CH2
3-CN-C6H4CH2
(CH3)2CHCH2
8
9
10
23
29
31
0
11
12
0
1
28
37
59
50
70
67
24
18
24
55
30
9
20
16
70
26
41
52
5
81
49
26
2
2-(OCH3)-5-(NO2)-C6H3CH2
3-Br-C6H4CH2
2-Br-C6H4CH2
13
14
95
100
100
80
57
9
27
22
37
34
32
0
100
100
100
97
88
96
96
69
46
2
93
78
90
88
82
2
84
49
75
73
46
20
5
37
26
40
24
30
0
13
18
9
15
16
2-F-C6H4CH2
4-F-C6H4CH2
17
18
9
3-(NO2)-C6H4CH2
4-(7-CH3O)-coumarinylCH2
1-(8-CH2Br)-naphthaleneCH2
(S)-(+)-CH3CH2CH(CH3)CH2
2-Cl-C6H4CH2
97
1
19
20
44
37
0
6
91
2
24
22
91
0
2
18
0
23
86
88
6
18
33
36
0
8
21
22
0
94
8
57
57
2
19
12
3
31
9
52
38
23
89
34
34
48
29
60
0
100
53
99
0
PhthalimidoCH2
2-(CF3)-C6H4CH2
2,4-F2-C6H3CH2
2,5-F2-C6H3CH2
2,6-F2-C6H3CH2
3,4-F2-C6H3CH2
2,3-F2-C6H3CH2
3,5-F2-C6H3CH2
3-(CF3)-C6H4CH2
4-(CF3)-C6H4CH2
2-((2-CH2Br)-C6H4)-C6H4CH2
3-(OCF3)-C6H4CH2
3-I-C6H4CH2
23
24
29
78
62
79
32
73
55
55
60
83
49
59
36
56
15
46
0
93
84
76
91
84
96
84
97
83
97
97
98
93
84
7
30
7
100
98
93
82
52
82
54
85
16
81
62
23
89
89
77
63
0
79
93
80
92
94
98
55
92
79
95
94
92
64
76
2
48
70
44
68
60
68
4
19
27
13
16
12
47
24
18
3
14
13
10
11
7
25
26
1
98
98
27
28
12
22
15
20
30
15
25
53
51
45
9
96
99
29
30
16
6
90
95
31
32
76
49
66
75
75
43
20
0
10
19
33
17
19
13
15
14
14
10
5
23
43
55
59
7
100
87
41
51
49
58
26
19
23
38
3
33
34
17
32
20
0
94
96
35
36
2-(CF3)-4-F-C6H4CH2
2,3,6-F3-C6H2CH2
2,5-(CF3)2-C6H3CH2
4-(CH3)-C6H4CH2
CH3CH2
95
94
37
38
8
4
1
45
30
27
45
33
77
78
62
100
29
54
0
39
40
88
39
51
69
57
79
89
100
82
9
100
17
85
0
85
1
62
1
5
13
10
18
4
0
CH3(CH2)2
CH3(CH2)3
41
42
15
72
63
66
75
80
68
66
84
0
5
37
52
26
23
60
64
73
2
0
0
26
26
43
50
60
42
31
21
35
49
53
65
4
0
CH3(CH2)4
CH3(CH2)5
43
44
71
84
0
15
15
35
42
22
5
9
3
8
CH3(CH2)6
CH3(CH2)7
C6H5CH2
45
46
22
59
24
90
92
21
31
84
8
15
0
47
48
97
2-(C6H5)-C6H4CH2
Me-Indoxam
84
100
58
9
100
91
34
91
75
57
25
Analyses were performed in at least duplicate for hGIIA, hGV, and hGX sPLA2 enzymes.

1742
B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
Table 2. Inhibition of mouse sPLA2s by indoles bearing a 4-(2-oxy-ethanoic acid) side chain
R Group
Compd
% Inhibition against mouse sPLA2
IB
0.5
mM
IIA
0.025
mM
IIC
0.025
mM
IID
5
IIE
0.025
mM
IIF
5
V
X
0.333
mM
1.0
mM
mM
mM
2,6-Cl2-C6H3CH2
Phthalimido(CH2)4
C6F5CH2
1
96
66
69
89
17
52
84
86
54
31
27
59
57
78
81
81
76
31
20
73
60
81
9
85
17
8
70
42
55
78
15
64
83
82
42
52
11
38
14
86
93
99
82
33
12
27
59
95
18
78
86
88
83
61
92
67
67
48
89
73
83
68
73
16
83
6
27
17
14
85
15
16
59
92
25
11
12
64
12
56
33
50
39
21
18
12
58
44
19
3
56
21
29
61
19
62
31
27
6
2
4
95
55
31
97
7
70
35
38
61
7
2
3
8
4-Br-C6H4CH2
C6H5COCH2
4
88
0
30
6
5
3-(CH2Br)-C6H4CH2
3-Cl-C6H4CH2
6
47
75
83
29
34
17
10
4
5
87
92
87
75
66
63
73
5
51
68
64
49
42
33
34
4
7
5
2-NapthaleneCH2
4-CN-C6H4CH2
2-CN-C6H4CH2
3-CN-C6H4CH2
(CH3)2CHCH2
8
30
0
9
10
45
12
0
0
11
12
4
15
8
2-(OCH3)-5-(NO2)-C6H3CH2
3-Br-C6H4CH2
2-Br-C6H4CH2
13
14
4
75
74
78
78
36
0
26
45
51
32
20
0
7
96
86
94
88
74
8
73
73
77
47
35
7
15
16
12
8
2-F-C6H4CH2
4-F-C6H4CH2
17
18
3
3-(NO2)-C6H4CH2
4-(7-CH3O)-coumarinylCH2
1-(8-CH2Br)-naphthaleneCH2
(S)-(+)-CH3CH2CH(CH3)CH2
2-Cl-C6H4CH2
1
19
20
13
3
8
0
29
87
91
4
21
55
69
13
69
54
58
80
41
70
47
63
33
71
51
64
23
62
10
59
2
21
22
16
75
11
46
66
49
72
45
73
33
54
56
43
66
67
47
57
0
16
62
10
60
56
34
45
26
57
13
47
35
31
37
35
37
27
7
5
8
PhthalimidoCH2
2-(CF3)-C6H4CH2
2,4-F2-C6H3CH2
2,5-F2-C6H3CH2
2,6-F2-C6H3CH2
3,4-F2-C6H3CH2
2,3-F2-C6H3CH2
3,5-F2-C6H3CH2
3-(CF3)-C6H4CH2
4-(CF3)-C6H4CH2
2-((2-CH2Br)-C6H4)-C6H4CH2
3-(OCF3)-C6H4CH2
3-I-C6H4CH2
23
24
2
62
78
74
77
73
72
63
63
74
84
68
71
71
68
37
67
18
58
79
79
80
87
86
72
4
84
95
77
89
76
93
56
92
73
90
87
91
66
71
14
85
12
44
70
58
65
74
82
85
25
26
40
32
30
25
29
30
49
45
11
62
56
17
16
13
46
48
74
71
40
71
89
95
34
8
0
27
28
0
0
29
30
8
4
31
32
10
5
33
34
40
16
32
13
5
35
36
2-(CF3)-4-F-C6H4CH2
2,3,6-F3-C6H2CH2
2,5-(CF3)2-C6H3CH2
4-(CH3)-C6H4CH2
CH3CH2
37
38
24
9
39
40
74
0
43
4
11
10
4
CH3(CH2)2
CH3(CH2)3
41
42
5
31
55
56
69
80
86
85
8
8
35
37
59
71
82
72
10
0
40
30
39
40
54
69
CH3(CH2)4
CH3(CH2)5
43
44
11
13
23
36
11
0
CH3(CH2)6
CH3(CH2)7
C6H5CH2
45
46
17
6
47
48
69
2-(C6H5)-C6H4CH2
Me-Indoxam
90
81
92
39
28
34
94
67

B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
1743
Table 3. Inhibition of sPLA2s by indoles bearing a 5-(4-oxy-butanoic
acid) side chain
inhibition among the various enzyme species is readily
apparent from inspection of these arrays and is sum-
marized as follows. For the human sPLA2s, several
members of the 48-compound, Me-Indoxam analogue
library are potent inhibitors (IC50 <0.05 mM) of hGIIA,
hGIIE and hGV. These compounds show intermediate
potency against hGIB (several compounds with an
IC50=0.05–0.5 mM) and hGX (several compounds with
an IC50=0.5–5 mM), and are poor inhibitors of hGIID,
hGIIF, and hGXIIA (IC50>5 mM). The same set of
Me-Indoxam analogues are potent inhibitors of mGIIA,
mGIIC, mGIIE, and mGV (several compounds with an
IC50<0.05 mM), they show intermediate potency
toward mGIB (several compounds with an IC50=0.05–
0.5 mM) and mGX (several compounds with an
IC50=0.5–5 mM), and are poor inhibitors of mGIID
and mGIIF (IC50>5 mM). It can be seen from Figure
5C that for each human/mouse orthologue pair, the
inhibition potencies for the Me-Indoxam analogues
acting on the human sPLA2 matches closely to that
measured for its mouse orthologue (hGIB versus mGIB,
hGIIA versus mGIIA, hGIID versus mGIID, hGIIE
versus mGIIE, hGIIF versus mGIIF, hGV versus mGV,
and hGX versus mGX). This is expected given that for
each orthologue pair, the active site residues are highly
conserved (based on the hGIIA and hGX X-ray struc-
tures and the active site models for the remaining
groups I, II, V, and X enzymes).
R Group
Compd
% Inhibition
hGIIA
1 mM
hGX
1 mM
2,6-Cl2-C6H3CH2
C6F5CH2
4-Br-C6H4CH2
C6H5COCH2
3-Cl-C6H4CH2
2-NapthaleneCH2
4-CN-C6H4CH2
2-CN-C6H4CH2
3-CN-C6H4CH2
(CH3)2CHCH2
2-(OCH3)-5-(NO2)-C6H3CH2
3-Br-C6H4CH2
2-Br-C6H4CH2
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
85
26
83
37
87
93
33
0
10
29
63
83
75
49
5
11
4
20
43
36
4
20
13
10
0
20
11
43
26
10
18
0
4-F-C6H4CH2
3-(NO2)-C6H4CH2
4-(7-CH3O)-coumarinylCH2
2-Cl-C6H4CH2
0
77
0
PhthalimidoCH2
2-(CF3)-C6H4CH2
2,4-F2-C6H3CH2
2,5-F2-C6H3CH2
2,6-F2-C6H3CH2
3,4-F2-C6H3CH2
2,3-F2-C6H3CH2
3,5-F2-C6H3CH2
3-(CF3)-C6H4CH2
4-(CF3)-C6H4CH2
2-((2-CH2Br)-C6H4)-C6H4CH2
3-(OCF3)-C6H4CH2
2-(CF3)-4-F-C6H4CH2
4-(CH3)-C6H4CH2
CH3(CH2)3
18
6
8
90
27
13
16
55
79
46
23
0
77
3
89
29
7
78
30
26
44
13
68
59
30
32
32
15
27
28
25
19
36
3.2. Interactions between enzyme residues and the indole
core of Me-Indoxam and its analogues
The trends in inhibition potency among the different
sPLA2 species noted above tend to hold as the N1-sub-
stituent is varied. In addition, the contacts between the
glyoxamide and carboxymethyl side chains of Me-
Indoxam and the enzyme are conserved among all of
the sPLA2s. This suggests that it is the residues that
interact with the remaining portion of Me-Indoxam, the
indole core, that are responsible for the differences in
inhibition potency among the sPLA2 enzyme species.
For the hGX-Me-Indoxam complex, these residues are
Ile-2, Leu-5, Gly-28, and Leu-29 (Fig. 3). The residues
at these positions for all of the human and mouse
groups I, II, V, and X sPLA2s are listed in Table 4
(supplemental information). For all of the enzymes,
glycine is conserved at position-28. This leaves residues
2, 5, and 29 for further consideration. When residue 5 is
a Phe, the edge of this aromatic group would form a
favorable interaction with the aromatic face of the
indole ring of Me-Indoxam. Such an interaction
between two aromatic rings that are perpendicular to
each other is thought to be highly favorable.55 This
could explain why Me-Indoxam and many of its analo-
gues are generally potent inhibitors of hGIB, mGIB,
hGIIA, mGIIA, mGIIC, hGIIE, and mGIIE, all of
which have a Phe-5 (Table 4, supplemental informa-
tion). Within this group of enzymes, less potent inhibi-
tion is seen with hGIB and mGIB. This may be due to
the fact that the group IB enzymes have a Val at posi-
tion-2, which is not quite long enough to contact the
indole core of the inhibitor. In contrast, the group IIA,
IIC, and IIE sPLA2s contain longer aliphatic side chains
CH3(CH2)4
CH3(CH2)5
CH3(CH2)6
8
31
Analyses were performed in at least duplicate for hGIIA and hGX
sPLA2 enzymes.
3. Discussion
3.1. Inhibition patterns among sPLA2 enzyme species
The large amount of inhibition data obtained with the
4-(2-oxy-ethanoic acid) analogues and the human
sPLA2s (potency of 48 compounds on eight group I, II,
V, and X enzymes) is conveniently summarized by the
gray scale array shown in Figure 5A. Figure 5B shows
the array for the same compounds as inhibitors of the
eight mouse sPLA2s. Figure 5C shows a comparison of
the results obtained with human and mouse sPLA2s
(mGIIC is removed since there is no hGIIC and
mGXIIA is removed since inhibition data is not avail-
able, see above). The general pattern for potency of

1744
B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
at position-2 (Leu, Ile, and Phe) which are predicted to
make a direct contact with the indole core of Me-
Indoxam, based on the contact between Ile-2 and inhi-
bitor observed in the hGX-Me-Indoxam X-ray struc-
ture. Thus, the combination of Phe-5 and Leu, Ile, or
Phe at position-2 seems to provide the most favorable
interaction with the indole core of Me-Indoxam. The
enzymes hGIID, mGIID, hGIIF, mGIIF, hGX, and
mGX all have Leu at position-5 instead of Phe, and this
may be the reason that Me-Indoxam and its analogues
are generally poor inhibitors of these enzymes. The
group V sPLA2s represent an exception to these trends.
They have Leu-2, which favors Me-Indoxam binding,
but they contain Leu-5 instead of Phe-5, which should
disfavor binding and yet many of the Me-Indoxam
analogues are potent hGV and mGV inhibitors.
Figure 5. Array of potencies of inhibition of human and mouse sPLA2s by Me-Indoxam and its analogues. Panel A is for the human enzymes
indicated along the horizontal top edge of the array, panel B is for the mouse enzymes, and panel C is human and mouse enzymes combined. The
vertical edge of the array gives the compound numbers for the Me-Indoxam analogues with the 4-(2-oxy-ethanoic acid) side chain (see Table 1).
Inhibition potency was classified into 4 gray scale categories as follows: black is IC50<0.05 mM, dark gray is IC50=0.05–0.5 mM, light gray is
IC50=0.5–5 mM, and white is IC50>5 mM.

B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
1745
However, the group V sPLA2s are unique in containing
a Trp at position-29, whereas all of the other sPLA2s
contain an aliphatic residue at this position (Ile, Leu,
Val). The homology models of the group V sPLA2
active sites suggest that the edge of the indole ring of
Trp-29 could form a favorable interaction with the face
of the indole ring of Me-Indoxam.
Me-Indoxam pointing into the hydrophobic cavity of
the enzyme and interacting favorably with Ile-2 and
Ala-6. However, the non-benzyl substituents either
sterically clash with Ala-6 (8 and 19) or lose their
favorable interaction with Ile-2 and Ala-6 (2, 5, and 12)
of hGX as seen from modeling of these substituents
with the hGX-Me-Indoxam X-ray structure and models
of the human and mouse groups I, II, V, and X sPLA2s.
Indole analogue 23 does not contact Ala-6, but instead
its phthalimido group sterically clashes with Leu-29 and
Gly-30 of hGX because the methylene group between
the N1-position and the phthalimido group is too short
to allow the phthalimido group to protrude out of the
active site. Analogue 21 is similar to analogue 12
(branched alkyl group at the N1-position) except that
the side chain is lengthened by one carbon. This re-
establishes the favorable interaction with Ala-6 of hGX,
as the geometry of the chain rotates to mimic the benzyl
substituent of Me-Indoxam. Indole analogues 40–46
display increasing inhibition potency against human
and mouse group sPLA2s. This is due to the filling of
the hydrophobic cavity by the increasing size of the
saturated alkyl chains. The favorable interactions with
Ile-2 and Ala-6 of hGX (and the corresponding residues
of the other sPLA2s modeled) are realized as the chain
length is increased to eight carbons. Presumably, after
eight carbons this trend would no longer hold, as the
remaining methylene groups would protrude out of the
cavity and not interact with any protein residues.
3.3. Interaction between enzyme residues and the N1-sub-
stituent of Me-Indoxam and its analogues
The X-ray structure of Me-Indoxam bound to hGX
shows that the 2-phenyl group on the N1-benzyl chain
does not form significant interactions with the enzyme
(discussed above). This is predicted to be the case for all
group I, II, V, and X sPLA2s, based on modeling. This
presumably explains why the potency of Me-Indoxam
and 47, which lacks the 2-phenyl group, are similar
for all sPLA2s (maximum difference in potency of
ꢀ3-fold).
Inspection of the hGX-Me-Indoxam X-ray structure
suggests that a 2-, 3- or 4-substituent on the N1-benzyl
group would not influence sPLA2 inhibition potency
since these substituents protrude away from the enzyme
surface. Indeed compounds 4, 6, 7, 9–11, 14–18, 22, 24,
31–35, 39, and 48, which contain a single substitution at
either the 2-, 3- or 4-position of the N1-benzyl chain,
display similar inhibition potency as compound 47,
which contains an unsubstituted N1-benzyl chain. This
trend is true for all sPLA2s as predicted from the active
site modeling. The same applies for disubstituted
N1-benzyl chains that include 2-, 3-, and 4-substituents
(compounds 25, 28, 29, and 36). It is clear from the
X-ray structure of the hGX-Me-Indoxam complex as
well as the models for the other sPLA2 active sites that
compounds 13 and 38 are expected to be generally
poor inhibitors, and this was found to be the case.
These compounds have an N1-benzyl group with a 2,5-
disubstitution pattern. This leads to a steric clash with
the enzyme. If the 2-substituent points away from the
enzyme surface, the 5-substituent would clash with Ala-
6 of hGX (as well as corresponding residues of the other
sPLA2s). Flipping the aromatic ring would relieve the
steric clash with the 5-substituent and Ala-6, but would
position the 2-substituent in steric clash with the indole
ring of Me-Indoxam. Poor inhibition was not observed
for compound 26, which contains the smaller fluorine
atoms at the 2- and 5-position. Compounds 1 and 27,
which contain a 2,6-dichloro and 2,6-difluoro pattern of
substitution on the benzyl group display similar potency
as compound 47, which contains the unsubstituted ben-
zyl chain. Presumably the fluorine and chlorine sub-
stituents cause less steric clash because they are smaller
than those present on the poorly active compounds with
2,5-disubstituted benzyl chains (13 and 38 which con-
tain larger substituents OMe, NO2, and CF3).
3.4. Discovery of compounds useful for cellular studies of
sPLA2 function
With the series of Me-Indoxam analogues prepared in
the current study, we have not been able to generate
compounds that are highly specific for a single human
or mouse sPLA2. However, some compounds that will
be useful in cell culture studies have been found. For
example, the original lead compound Me-Indoxam and
compound 5 are ꢀ5-fold more potent on hGIIA than
on hGV, and compound 21 was discovered to have a
reverse preference (10-fold more potent on hGV com-
pared to hGIIA). Compound 21 was re-tested on all of
the human sPLA2s to determine the precise IC50 values
(based on enzyme activity measured in the presence of
five inhibitor concentrations): hGIB, IC50=0.77 mM;
hGIIA, IC50=1.0 mM; hGIID, no inhibition at
5 mM; hGIIE, IC50=0.04 mM; hGIIF, no inhibition at
5 mM; hGV, IC50=0.1 mM; hGX, IC50=5 mM; hGXIIA,
<20% inhibition at 5 mM. Thus, Me-Indoxam or com-
pound 5, when used in combination with compound 21,
should be useful for distinguishing cellular events
caused by hGIIA versus hGV.
4. Conclusion
In an attempt to find potent and selective inhibitors of
mammalian secretory phospholipases A2, a library of 83
indole analogues was synthesized and tested against the
full set of mouse and human sPLA2s. Homology models
of the human and mouse sPLA2s were made based on
the X-ray structures of the hGIIA and hGX and
Indole analogues containing non-benzyl substituents at
the N1-position (2, 5, 8, 12, 19, 20, 21, 23, 40, 41, 42, 43,
44, 45, and 46) generally display poor inhibition potency
against human and mouse sPLA2s. The hGX-Me-
Indoxam X-ray structure shows the benzyl group of

1746
B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
inhibitors were designed with structure-guided tech-
niques. The X-ray structure was determined of Me-
Indoxam, an indole analogue that is a competitive tight
binding inhibitor of several sPLA2s, and modeling of
this structure with the sPLA2 homology models suggests
a good deal of variation in the residues surrounding the
N1-position of the indole. Based on this structure and a
series of previously published indoles bound to the
hGIIA sPLA2, a library was generated by varying the
N1-substituent of the indole and the position and length
of the calcium coordinating carboxylate arm. All com-
pounds generated were tested in a fluorometric assay
employing vesicles with a pyrenedecanoic acid at the
sn-2 position of the phospholipid. The structure–activity
relationship suggests that a 4-(2-oxy-ethanoic acid)
moiety and an N1-position carrying a substituted benzyl
group are the most highly potent against the sPLA2
enzymes. However, 2,5- and 3,6-disubstituted benzyl
groups at the N1-position cause steric clash and sig-
nificantly diminish the inhibition potency. Although no
single compound was highly selective for a single sPLA2
enzyme, a combination of compounds with a 5 to
10-fold inverse selectivity preference for the hGV versus
hGIIA was found and should prove useful in cell culture
studies.
ture of hGX52 (pdb code 1LE6) as the search model.
Standard refinement techniques in the program CNS-
1.158 such as rigid body, positional and simulated
annealing refinement were employed during cycles of
refinement. Manual model improvement was done using
the graphic programs CHAIN59 and O.60 The Me-
Indoxam ligand was built into one of the two subunits of
the protein when the Rfree and Rworking were at 0.339 and
0.279, respectively at resolution 2.2 A. Following a
round of refinement, the electron density difference
maps (2Fo-Fc) showed clear electron density for the
active site ligand. Next, one round of positional and
simulated annealing refinement was performed using X-
ray diffraction data that was extended to a resolution of
2.0 A. Electron density difference maps directed the
building of a second Me-Indoxam molecule into the
second subunit’s active site and was refined. Finally, a
total of 164 solvent molecules were built into difference
electron density (Fo–Fc), and iterative rounds of model
adjustment and refinement converged. The final model
has a value of Rfree and Rworking of 0.269 and 0.221,
respectively with good geometry.
5.3. Synthesis of Me-Indoxam analogues
5.3.1. General methods. Unless otherwise noted, all
reagents were obtained from commercial suppliers and
used without further purification and all non-aqueous
reactions were carried out under an argon atmosphere
with oven-dried glassware. THF was distilled under
argon from sodium/benzophenone and CH2Cl2 was
distilled over CaH2 immediately prior to use. Reaction
progress was monitored by thin-layer chromatography
on Merck 60F254 silica plates. Unless otherwise speci-
fied, extracts were dried over MgSO4 and solvents were
removed with a rotary evaporator at aspirator pressure.
All compounds were purified by column chromato-
graphy using Merck 60 230–400 mesh silica gel. NMR
spectra were recorded on Bruker AC-200 (200 MHz) or
AC-300 (300 MHz) spectrometers. Unless otherwise
specified, all spectra were obtained in CDCl3 and che-
mical shifts are reported in parts per million relative to
TMS. Low-resolution mass spectral data was obtained
for all compounds using a Bruker Daltonics Esquire
electrospray–ion trap mass spectrometer. High-resolu-
tion mass spectral data was obtained on a Bruker Dal-
tonics APEX III 47e Fourier transform (Ion Cyclotron
Resonance) Mass Spectrometer for two key compounds
of the library, compounds 5 and 21. (Compound 5: HR-
ESI-MS m/z 369.1426 calculated for C18H22N2O5Na
[M+Na]+, found 369.1427. Compound 21: HR-ESI-
MS m/z 417.1063 calculated for C21H18N2O6Na
[M+Na]+, found 417.1068.)
5. Experimental
5.1. Crystallization and X-ray data collection
The purified recombinant hGX sPLA2 was prepared as
described elsewhere.52 The hGX-Me-Indoxam complex
was co-crystallized at ambient temperature in 2 mL
hanging drops of a mixture of 1 mL of a pre-equilibrated
solution containing 15 mg/mL protein, 0.56 mM Me-
Indoxam, 10 mM CaCl2, 10% (v/v) MPD, 0.25% (v/v)
DMSO, and 1 mL of a crystallization reservoir solution
of 10 mM CaCl2, 12% (w/v) PEG 3500, 10% (v/v)
MPD, 2% (v/v) ethylene glycol and 0.2 M Hepes buffer,
pH 7.4. Crystals of rectangular shape appeared within a
week and were allowed to grow for two weeks before
data collection.
X-ray diffraction data were collected on a Rigaku RU-
H3R rotating anode generator with a RAXIS IV image
plate area detector. Crystals were pre-treated with a
cryo-protecting and stabilizing solution made of the
crystallization condition reported above with the addi-
tion of 10% (v/v) glycerol and flash frozen in the
À180 ꢁC nitrogen cryo-stream. X-ray diffraction data
were collected from a single crystal that diffracted
to a resolution of 1.97 A. The programs DENZO and
SCALEPAK were used for data processing.56 The data
collection statistics are shown in Table 5 (supplemental
information). The protein crystal belongs to space
group C2221 with 2 copies of the hGX-Me-Indoxam
complex in the asymmetric unit.
({3-[Amino(oxo)acetyl]-1-(biphenyl-2-ylmethyl)-2-
methyl-1H-indol-4-yl}oxy) acetic acid (Me-Indoxam).
Prepared according to the literature procedure33 except
that the starting material used was the commercially
available 4-hydroxy-2-methylindole (A) from TCI
1
America. H NMR (300 MHz, DMSO-d6, d): 2.35 (s,
5.2. X-ray structure solution and refinement
3H), 4.66 (s, 2H), 6.36 (s, 2H), 6.40 (d, 1H), 6.51 (d,
1H), 6.88 (d, 1H), 7.02 (t, 1H), 7.19–7.58 (m, 8H), 7.73
(s, 1H), 12.90 (br s, 1H).
The structure was solved by molecular replacement
using the program AMORE57 using a ligand free struc-

B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
1747
tert-Butyl [(2-methyl-1H-indol-4-yl)oxy]acetate (B). The
commercially available 4-hydroxy-2-methylindole (A)
(3.00 g, 20.38 mmol) and t-butylbromoacetate (3.98 g,
20.38 mmol) were dissolved in 100 mL acetone (freshly
distilled from CaSO4) in a 250 mL round bottom flask.
To this solution at room temperature was added anhy-
drous potassium carbonate (5.66 g, 40.80 mmol) and the
stirred mixture brought to reflux for 12 h. The solid was
removed by filtration and washed with two 20 mL por-
tions of acetone. The solvent was removed by rotary
evaporation to leave an off-white solid. This solid was
dissolved in 50 mL EtOAc and washed with 50 mL of a
brine solution in a separatory funnel. The aqueous
phase was extracted twice with 20 mL portions of
EtOAc and the organic layers combined, dried over
MgSO4 and evaporated to leave an off-white solid. The
product was purified by flash column chromatography
on silica gel using 20% EtOAc/hexanes as eluant to give
4.131 g of a white powder (77% yield). This compound
decomposes quickly in solution when not under argon.
The product of decomposition is a bright green color.
Therefore, it is imperative to work quickly, including
at À20 ꢁC in jars containing CaSO4 as desiccant. Fifty mL
of each of these solutions was added to individual 4 mL
vials awaiting a reaction with compound (C). A 4 mL
vial was charged with compound (C) (20.0 mg, 0.06
mmol) and flushed with argon. The solid was dissolved
in 2.25 mL freshly distilled THF. To this solution was
added NaH (3.5 mg, 0.15 mmol) without mineral oil
and the reaction left to stir at room temperature for 10
min, at which point it became a bright yellow color. The
entire contents of the vial were removed by gas-tight
syringe and 25 mL aliquots were added to each of the
vials containing the alkyl bromide/THF solutions.
These vials were then heated at 60 ꢁC for 2 h followed
by removal of solvent by entrainment with air. The vials
containing the crude products were stored free of sol-
vent at À20 ꢁC until purified by HPLC. To each vial was
added neat trifluoroacetic acid (100 mL, 1.30 mmol), vials
were mixed on a vortex mixer for 15 s and left for 10
min at room temperature. The excess trifluoroacetic acid
was removed by entrainment with air. The compounds
were then purified using high performance liquid chro-
matography and eluting with a methanol/water/0.06%
TFA system on a Vydac C-18 reverse phase semi-pre-
parative 218TP1010 column using UV–Vis detection at
262 nm. The purified compounds were identified by
electrospray mass spectrometry with characteristic frag-
mentation patterns. A speed vac (Savant Instruments)
was used to remove the HPLC solvents from the frac-
tions containing the desired products to leave solids.
These solids were dissolved in 50 mL DMSO and trans-
ferred to Eppendorf tubes. An aliquot of this DMSO
solution was used to quantify the compound using a
liquid scintillation counter. Dilutions were made of each
of the stock DMSO/inhibitor solutions in preparation
for inhibition assays and stored at À20 ꢁC.
1
the purification step. H NMR (200 MHz, CDCl3, d):
1.51 (s, 9H), 2.42 (s, 3H), 4.65 (s, 2H), 6.40–6.47 (m,
2H), 6.85 (d, 1H), 7.07 (t, 1H), 7.91 (br s, 1H).
14C-Labeled tert-butyl ({3-[amino(oxo)acetyl]-2-methyl-
1H-indol-4-yl}oxy)acetate (C). To 14C-labeled oxalic acid
solid (50 mCi, 10 mCi/mmol) from American Radio-
labeled Chemicals Inc. in an amber vial under argon
atmosphere was added freshly distilled, unlabeled oxalyl
chloride (185 mL, 2.12 mmol) and approximately 3 mg
of phosphorus pentachloride solid. The exact weight of
the phosphorus pentachloride could not be obtained, as
it is sensitive to moisture, which causes the weight to
change rapidly. The mixture was refluxed for 10 min
until the evolution of bubbles had ceased upon standing
at room temperature. The entire contents of the vial were
transferred to a 1 L round bottom flask containing 400
mL freshly distilled CH2Cl2. Compound (B) (0.490 g,
1.86 mmol) was added as a solution in 100 mL freshly
distilled CH2Cl2 over 1 h under an argon atmosphere via
addition funnel. The bright yellow solution was left to stir
for 8 h at room temperature before ammonia gas was
bubbled into the solution for 15 min, at which point the
mixture became pale yellow and a fluffy white solid
precipitated. The precipitate was removed by filtration and
the solvent removed by rotary evaporation to leave a
yellow solid. The crude solid was dissolved in 150 mL
EtOAc and washed with water followed by brine. The
organic layer was dried over MgSO4 and purified by
column chromatography on silica gel using EtOAc as the
eluant. The product was obtained in 87% yield as a bright
yellow powder with a specific activity of 7160 cpm/
tert-Butyl 4-[(2-methyl-1H-indol-5-yl)oxy]butanoate (F).
A portion of 5-hydroxy-2-methylindole (E) (200.0 mg,
1.36 mmol) from Biosynth AG and tert-butyl 4-bromo-
butanoate (455 mg, 2.04 mmol) (prepared according to
the literature procedure61) were dissolved in 20 mL ace-
tone (freshly distilled from CaSO4) in a 50 mL round
bottom flask under argon. To this solution at room
temperature was added anhydrous potassium carbonate
(1.00 g, 6.8 mmol) and the stirred mixture brought to
reflux for 3 days. The solid was removed by filtration
and washed twice with 10 mL portions of acetone. The
solvent was removed by rotary evaporation to leave a
yellow oil. This oil was dissolved in 50 mL EtOAc and
washed with 50 mL of brine in a separatory funnel. The
aqueous phase was extracted with three 20 mL portions
of EtOAc and the organic layers combined, dried over
MgSO4 and evaporated to leave a yellow oil. The pro-
duct was purified by flash column chromatography on
silica gel using 20% EtOAc/hexanes as eluant to give
1
mmol. H NMR (200 MHz, CD3OD, d): 1.46 (s, 9H),
2.58 (s, 3H), 4.61 (s, 2H), 6.50 (dd, 1H), 6.98–7.12 (m, 2H).
1
359 mg (77% yield) of yellow oil. H NMR (300 MHz,
CDCl3, d): 1.49 (s, 9H), 2.04 (m, 2H), 2.42 (s, 3H), 2.46
(s, 2H), 4.03 (t, 2H), 6.12 (br s, 1H), 6.75 (dd, 1H), 6.99
(d, 1H), 7.17 (dd, 1H), 7.80 (br s, 1H).
Library synthesis of indole analogues (D). A collection
of various alkyl bromides from Aldrich Chemicals was
purchased in small quantities (10–12 mg each) and each
was delivered in a separate 4 mL amber vial. These
compounds were dissolved in 1.00 mL of freshly distilled
THF and capped. When not in use the vials were stored
14C-Labeled tert-butyl 4-({3-[amino(oxo)acetyl]-2-methyl-
1H-indol-5-yl}oxy)butanoate (G). To 14C-labeled oxalic
acid solid (50 mCi, 3 mCi/mmol) from American Radio-

1748
B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
labeled Chemicals Inc. in 1 mL dry CH2Cl2 in an amber
vial was added freshly distilled, unlabeled oxalyl chlo-
ride (100 mL, 1.15 mmol) and approximately 3 mg of
phosphorus pentachloride solid (see compound (C)).
The mixture was refluxed for 30 min under an argon
atmosphere and then the entire contents of the vial were
transferred to a 1-liter round bottom flask containing
400 mL freshly distilled CH2Cl2. Compound (F) (219
mg, 0.756 mmol) was dissolved in 100 mL dry CH2Cl2
and added over 1 h under an argon atmosphere. The
solution was left to stir for 12 h at room temperature
before ammonia gas was bubbled into the solution for
15 min, at which point a white solid precipitated. The
solid was removed by filtration and the solvent removed
by rotary evaporation to leave an off white solid. The
crude solid was dissolved in 100 mL EtOAc and washed
with water followed by brine. The organic layer was
dried over MgSO4 and purified by column chromato-
graphy on silica gel using 75% EtOAc/hexanes as
eluant. The product was obtained as an off-white pow-
der with a specific activity of 11,723 cpm/mmol (85%
In place of Solution B was added an additional 100 mL
portion of Solution A as a minus enzyme control to the
first of the eight wells in the assay. Quickly after the
addition of Solution B, the assay was initiated by the
addition of 100 mL of Solution C (4.2 mM 1-hexa-
decanoyl-2-(1-pyrenedecanoyl)-sn-glycero-3-phospho-
glycerol (Molecular Probes) vesicles in assay buffer)
with a repeating pipettor to all eight wells. The fluores-
cence (excitation=342 nm, emission=395 nm) was read
with a microtiter plate spectrophotometer (Fluor-
ocount, Packard Instruments). Control reactions with-
out enzyme or inhibitor were run with each assay of
eight wells and the percent inhibition calculated from
the initial slopes of fluorescence versus time. The
amount of enzyme used per well are as follows: hGIB,
0.5 ng; hGIIA, 0.5 ng; hGIID, 200 ng; hGIIE, 500 ng;
hGIIF, 30 ng; hGIII, 5 ng; hGIII, 5 ng; hgV, 1 ng; hGX,
1 ng; hGXIIA, 1,500 ng; mGIB, 1.6 ng; mGIIA, 8 ng;
mGIIC, 12 ng; mGIID, 2,000 ng; mGIIE, 5 ng; mGIIF,
100 ng; mGV, 6 ng; mGX, 3.5 ng. All of the recombi-
nant mouse and human sPLA2s were prepared as
described.42
1
yield). H NMR (300 MHz, CD3OD, d): 1.43 (s, 9H),
2.03 (m, 2H), 2.42 (t, 2H), 2.64 (s, 3H), 4.00 (t, 2H), 6.81
(dd, 1H), 7.22 (dd, 1H), 7.61 (d, 1H).
References and notes
Library synthesis of indole analogues (H). The alkyl
bromides used in the generation of the first library were
also used to generate this library of compounds. The
alkyl bromides were stored as solutions in dry THF in a
dessicated jar at À20 ꢁC. Twenty-five mL aliquots of
each of these solutions were added to individual 4 mL
vials awaiting a reaction with compound (G). A 4 mL
vial was charged with compound (G) (20.0 mg, 0.06
mmol) and flushed with argon. The solid was dissolved
in 4 mL freshly distilled THF. To this solution was
added NaH (3.5 mg, 0.15 mmol) without mineral oil
and the reaction left to stir at room temperature for 10
min, at which point the solution turned yellow-green.
The entire contents of the vial were removed by syringe
and 50 mL aliquots were added to each of the vials con-
taining alkyl bromide/THF solutions. These vials were
then placed in an oven at 60 ꢁC for 2 h. The resultant
compounds were treated in the same manner as for the
generation of the first library, purified by HPLC, and
characterized by electrospray ionization mass spectro-
metry. The compounds were quantified using liquid
scintillation counting and DMSO/inhibitor stock solutions
were made as for the first set of library compounds.
1. Waite, M. The Phospholipases; Plenum: New York, 1987.
2. Murakami, M.; Kudo, I. J. Biochem. 2002, 131, 285.
3. Dennis, E. A. Am. J. Respir. Crit. Care Med. 2000, 161,
S32.
4. Valentin, E.; Lambeau, G. Biochim. Biophys. Acta 2000,
1488, 59.
5. Six, D. A.; Dennis, E. A. Biochim. Biophys. Acta 2000,
1488, 1.
6. Jain, M. K.; Yuan, W.; Gelb, M. H. Biochemistry 1989,
28, 4135.
7. Gelb, M. H.; Jain, M. K.; Hanel, A. M.; Berg, O. Ann.
Rev. Biochem. 1995, 64, 653.
8. Gelb, M. H.; Jain, M. K.; Berg, O. Faseb J. 1994, 8, 916.
9. Berg, O. G.; Jain, M. K. Interfacial Enzyme Kinetics;
Wiley: West Sussex, England, 2002.
10. Beaton, H. G.; Bennion, C.; Connolly, S.; Cook, A. R.;
Gensmantel, N. P.; Hallam, C.; Hardy, K.; Hitchin, B.;
Jackson, C. G.; Robinson, D. H. J. Med. Chem. 1994, 37,
557.
11. Lin, H.-K.; Gelb, M. H. J. Am. Chem. Soc. 1993, 115,
3932.
12. De Haas, G. H.; Dijkman, R.; Lugtigheid, R. B.; Dekker,
N.; Van Den Berg, L.; Egmond, M. R.; Verheij, H. M.
Biochim. Biophys. Acta 1993, 1167, 281.
13. Jain, M. K.; Tao, W.; Rogers, J.; Arenson, C.; Eibl, H. et
al. Biochemistry 1991, 30, 10256.
14. Davidson, F. F.; Hajdu, J.; Dennis, E. A. Biochem. Bio-
phys. Res. Commun. 1986, 137, 587.
Microtiter plate assay of sPLA2s for inhibition analy-
sis. To each well of a 96-well microtiter plate was added
100 mL of solution A in assay buffer (27 mM bovine
serum albumin, 50 mM KCl, 1 mM CaCl2, 50 mM Tris–
HCl, pH 8.0) followed by the desired concentration of
sPLA2 inhibitor (ꢂ3 mL in DMSO from serial diluted
stock solutions) or 3 mL of DMSO only for control
reactions. Solution B was prepared immediately prior to
each set of assays, which consisted of only eight wells, to
avoid loss of enzymatic activity due to sticking to the
walls of the container. Solution B was delivered in 100
mL portions to all eight wells except the first well and
had the same composition as Solution A plus 5–2000 ng
of sPLA2 (depending on specific enzymatic activity42).
15. Chandra, V.; Jasti, J.; Kaur, P.; Srinivasan, A.; Betzel, C.;
Singh, T. P. Biochemistry 2002, 41, 10914.
16. Breitenstein, W.; Marki, F.; Roggo, S.; Wiesenberg, I.;
Pfeilschifter, J. et al. Eur. J. Med. Chem. 1994, 29.
17. Bryant, M. D.; Flick, K. E. S. K.R.; Wilton, D. C.;
Stoddard, B. L.; Gelb, M. H. Bioorg. Med. Chem. Lett.
1999, 9, 1097.
18. Jain, M. K.; Ghomashchi, F.; Yu, B.-Z.; Bayburt, T.;
Murphy, D. et al. J. Med. Chem. 1992, 3584.
19. Marshall, L. A.; Hall, R. H.; Winkler, J. D.; Badger, A.;
Bolognese, B. et al. J. Pharmacol. Exp. Ther. 1995, 274,
1254.

B. P. Smart et al. / Bioorg. Med. Chem. 12 (2004) 1737–1749
1749
20. Jain, M. K.; Yu, B.-Z.; Rogers, J. M.; Smith, A. E.;
Boger, E. T. et al. Phytochemistry 1995, 39, 537.
21. LeMahieu, R. A.; Carson, M.; Han, R.-J.; Madison, V. S.;
Hope, W. C. et al. J. Med. Chem. 1993, 36, 3029.
22. Burke, J. R.; Gregor, K. R.; Tramposch, K. M. J. Biol.
Chem. 1995, 270, 274.
23. Yoshida, T.; Nakamoto, S.; Sakazaki, R.; Matsumoto,
K.; Terui, Y. et al. J. Antibiot. Tokyo 1991, 44, 1467.
24. Tanaka, K.; Matsutani, S.; Kanda, A.; Kato, T.; Yoshida,
T. J. Antibiot. 1994, 47, 631.
41. Ying, H. P.; Epstein, T. M.; Jain, M. K.; Bahnson, B. J.
Biochemistry 2001, 40, 609.
42. Singer, A. G.; Ghomashchi, F.; Le Calvez, C.; Bollinger,
J. G.; Bezzine, S. et al. J. Biol. Chem. 2002, 277, 48535.
43. Bezzine, S.; Koduri, R. S.; Valentin, E.; Murakami, M.;
Kudo, I. et al. J. Biol. Chem. 2000, 275, 3179.
44. Bezzine, S.; Bollinger, J. G.; Veatch, S. L.; Keller, S. L.;
Gelb, M. H. J. Biol. Chem. 2002, 277, 48523.
45. Han, S. K.; Kim, K. P.; Koduri, R. S.; Bittova, L.;
Munoz, N. M. et al. J. Biol. Chem. 1999, 274, 11881.
46. Kim, K. P.; Rafter, J. D.; Bittova, L.; Han, S. K.; Snitko,
Y. et al. J. Biol. Chem. 2001, 276, 11126.
25. Genisson, Y.; Tyler, P. C.; Young, R. N. J. Am. Chem.
Soc. 1994, 116, 759.
26. Hite, G. A.; Mihelich, E. D.; Tulio, S.; Edmund, W. S.
European Patent; 0 846 687 A1, 1998.
47. Hanasaki, K.; Ono, T.; Saiga, A.; Morioka, Y.; Ikeda, M.
et al. J. Biol. Chem. 1999, 274, 34203.
27. Chandra, V.; Jasti, J.; Kaur, P.; Dey, S.; Srinivasan, A. et
al. Acta Crystallogr. D Biol. Crystallogr. 2002, 58, 1813.
28. Chandra, V.; Jasti, J.; Kaur, P.; Dey, S.; Perbandt, M. et
al. J. Biol. Chem. 2002, 277, 41079.
29. Chandra, V.; Jasti, J.; Kaur, P.; Betzel, C.; Srinivasan, A.
et al. J. Mol. Biol. 2002, 320, 215.
30. Schevitz, R. W.; Bach, N. J.; Carlson, D. G.; Chirgadze,
N. Y.; Clawson, D. K. et al. Nat. Struct. Biol. 1995, 2, 458.
31. Mihelich, E. D.; Schevitz, R. W. Biochim. Biophys. Acta
1999, 1441, 223.
32. Dillard, R. D.; Bach, N. J.; Draheim, S. E.; Berry, D. R.;
Carlson, D. G. et al. J. Med. Chem. 1996, 39, 5119.
33. Draheim, S. E.; Bach, N. J.; Dillard, R. D.; Berry, D. R.;
Carlson, D. G. et al. J. Med. Chem. 1996, 39, 5159.
34. Scott, D. L.; White, S. P.; Otwinowski, Z.; Yuan, W.;
Gelb, M. H. et al. Science 1990, 250, 1541.
48. Morioka, Y.; Saiga, A.; Yokota, Y.; Suzuki, N.; Ikeda,
M. et al. Arch. Biochem. Biophys. 2000, 381, 31.
49. Balboa, M. A.; Balsinde, J.; Winstead, M. V.; Tischfield,
J. A.; Dennis, E. A. J. Biol. Chem. 1996, 271, 32381.
50. Hanasaki, K.; Yamada, K.; Yamamoto, S.; Ishimoto, Y.;
Saiga, A.; Ono, T.; Ikeda, M.; Notoya, M.; Kamitani, S.;
Arita, H. J. Biol. Chem. 2002, 277, 29116.
51. Gesquiere, L.; Cho, W.; Subbaiah, P. V. Biochemistry
2002, 41, 4911.
52. Pan, Y. H.; Yu, B.-Z.; Singer, A. G.; Ghomashchi, F.;
Lambeau, G. et al. J. Biol. Chem. 2002, 277, 29086.
53. Scott, D. L.; Sigler, P. B. Adv. Protein Chem. 1994, 45,
53.
54. Dillard, R. D.; Bach, N. J.; Draheim, S. E.; Berry, D. R.;
Carlson, D. G.; Chirgadze, N. Y.; Clawson, D. K. et al. J.
Med. Chem. 1996, 39, 5137.
35. White, S. P.; Scott, D. L.; Otwinowski, Z.; Gelb, M. H.;
Sigler, P. B. Science 1990, 250, 1560.
36. Scott, D. L.; Otwinowski, Z.; Gelb, M. H.; Sigler, P. B.
Science 1990, 250, 1563.
55. Burley, S. K.; Petsko, G. A. Science 1985, 229, 23.
56. Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276,
307.
57. Navazza, J. Acta Cryst. 1994, A50, 157.
37. Scott, D. L.; White, S. P.; Browning, J. L.; Rosa, J. J.;
Gelb, M. H. et al. Science 1991, 254, 1007.
38. Thunnissen, M. M. G. M.; Ab, E.; Kalk, K. H.; Drenth,
J.; Dijkstra, B. W. et al. Nature 1990, 347, 689.
39. Cha, S. S.; Lee, D.; Adams, J.; Kurdyla, J. T.; Jones, C. S.
et al. J. Med. Chem. 1996, 39, 3878.
58. Brunger, A. T.; Adams, P. D.; Clore, G. M.; Delano,
W. L.; Gros, P. et al. Acta Cryst. 1998, D54, 905.
59. Sack, J. S.; Quiocho, F. A. Methods Enzymol. 1997, 277,
158.
60. Jones, T. A.; Zou, J.-Y.; Cowan, S. W. Acta Cryst. 1991,
A47, 110.
40. Kitadokoro, K.; Hagishita, S.; Sato, T.; Ohtani, M.; Miki,
K. J. Biochem. 1998, 123, 619.
61. Morin, C.; Vidal, M. Tetrahedron 1992, 48, 9277.
62. Kraulis, P. J. Appl. Cryst. 1991, 24, 946.