Chemically induced dimerization of human nonpancreatic secretory phospholipase A2 by bis-indole derivatives

Article abstract of DOI:10.1021/jm7010707

A series of novel bis-indole compounds, 1,ω-bis(((3-acetamino-5- methoxy-2-methylindole)-2-methylene)phenoxy)alkane, have been designed and synthesized on the basis of the enzyme structure of human nonpancreatic secretory phospholipase A2 (hnps PLA2). The

Full text of DOI:10.1021/jm7010707

3360
J. Med. Chem. 2008, 51, 3360–3366
Articles
Chemically Induced Dimerization of Human Nonpancreatic Secretory Phospholipase A2 by
Bis-indole Derivatives
Lu Zhou,† Chao Fang,† Ping Wei,† Shiyong Liu,†,‡ Ying Liu,*,† and Luhua Lai*,†,‡
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Unstable and Stable Species, College of
Chemistry and Molecular Engineering, Peking UniVersity, Beijing 100871, China, and Center for Theoretical Biology, Peking UniVersity,
Beijing 100871, China
ReceiVed August 30, 2007
A series of novel bis-indole compounds, 1,ω-bis(((3-acetamino-5-methoxy-2-methylindole)-2-methylene)phe-
noxy)alkane, have been designed and synthesized on the basis of the enzyme structure of human nonpancreatic
secretory phospholipase A2 (hnps PLA2). Their inhibition activities against hnps PLA2 were improved
compared to that of the monofunctional protocompound. These bivalent ligands not only inhibited hnps
PLA2 but also drove the dimerization of hnps PLA2. Their dimerization ability correlated with the linker
length and position. Further study on the potent compound 5 (1,5-bis(((3-acetamino-5-methoxy-2-
methylindole)-2-methylene)phenoxy)pentane, IC50 ) 24 nM) revealed that cooperative binding interactions
between the two enzyme molecules also contributed to the stability of the ternary complex. The combination
of bivalent ligands and hnps PLA2 can be used as a novel chemically induced dimerization (CID) system
for designing regulatory inhibitors.
Introduction
useful as a regulatory tool in biological systems, the number of
known CID systems is limited.
Dimerization and oligomerization are general biological
control mechanisms that contribute to the activation of intra-
and extracellular proteins, including cell membrane receptors
and vesicle fusion proteins.1,2 Chemically induced dimerization
(CIDa) systems have been used to manipulate cellular regulatory
pathways from signal transduction to transcription by inducing
dimerization of key proteins. The concept of CID systems was
first exploited by Schreiber and Crabtree in their pioneering
design of FK1012 (27, a dimeric form of tacrolimus (FK506)
(28)).2 This approach has yielded several novel CID systems3–7
and has been applied to important biological systems such as
the RNA-protein complex8 and heterodimeric complex.9 In fact,
inhibitors capable of inducing HIV-1 reverse transcriptase
dimerization exhibited increased binding strength.10 Even though
these examples demonstrate the general feasibility of modulating
protein-protein interactions with small organic molecules, the
application of this principle to drug discovery research has
encountered many obstacles. Berg11 pointed out three main
problems: (1) the identification of lead compounds is difficult;
(2) the protein-protein interface is much larger than the
potential binding area of a low molecular weight compound;
(3) protein-protein interfaces are often flat and may therefore
lack binding sites for small molecules. Because of these
difficulties, it is a challenge to find small molecules that can
modulate protein-protein interactions and to characterize the
cooperative nature of these interactions. Thus, although CID is
Phospholipase A2 (PLA2, EC 3.1.1.4) hydrolyzes the acyl-
ester bond at the sn-2 position of phosphoglycerides and liberates
free fatty acids, predominantly arachidonic acid, and lysophos-
pholipids. Overexpression of PLA2 is believed to play an
important role in the inflammatory process.12,13 As a membrane-
bound enzyme,14 PLA2 has an i-face that has been proposed to
make contact with the substrate interface, leading to the
interfacial catalytic turnover.15 The i-face of PLA2 is a relatively
flat surface of 1600 Å2, which binds tightly to the phospholipid
bilayer (Kd < 10-13 M). There are some polarizable and
hydrophobic residues on the flat surface, through which PLA2
binds to the anionic bilayer. The i-faces have been found to
interact with each other under certain circumstances, and two
crystal structures containing PLA2 dimers have been reported.16–18
The catalytic residues were buried by the dimer interface and
were inaccessible from the bulk solvent. The free-energy
minimization process might exclude the hydrophobic molecular
surface from the aqueous solution, but PLA2s were all detected
as monomers in aqueous solution. It seemed that the interaction
between two PLA2s was weak and not exclusive.
As described above, the application of CID systems to drug
research remains challenging. Fortunately, all of the noted
difficulties were avoided in the present study. The identification
of lead compounds is easy for hnps PLA2. Many potent
inhibitors against hnps PLA2 were reported. The secretory PLA2
is an interfacial enzyme with its active pocket near the substrate
binding i-face. The small molecule could act as an inducer to
pull two protein molecules together and let them interact with
each other. Thus, the last two problems were solved at the same
time. PLA2 simultaneously provides the binding site for small
molecules and the protein-protein interaction surface, providing
good conditions for CID studies. Since the buried i-face plays
an important role in binding to the phospholipid surface, we
* To whom correspondence should be addressed. For Y.L.: phone, 86-
10-62751490; fax, 86-10-62751725; e-mail, liuying@pku.edu.cn. For L.L.:
phone, 86-10-62757486; fax, 86-10-62751725; e-mail, lhlai@pku.edu.cn.
†
College of Chemistry and Molecular Engineering.
Center for Theoretical Biology.
‡
a Abbreviations: hnps PLA2, human nonpancreatic secretory phospho-
lipase A2; CID, chemically induced dimerization ; IC50, inhibition constant;
Ka, association constant.
10.1021/jm7010707 CCC: $40.75
2008 American Chemical Society
Published on Web 06/04/2008

Dimerization of Phospholipase A2
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12 3361
capsule. We docked the known inhibitor 1 (IC50 ) 0.84 µM20)
into the substrate binding pocket of each monomer enzyme. The
hydrophobic group of each inhibitor molecule is close to the
interface and exposed to the other pocket. The minimal length
between the two phenyl rings of the inhibitor is approximately
7 Å, about seven C-C bond lengths. Because of their distance
and conformation, the two phenyl rings can be linked together
to form a homobifunctional inhibitor. An alkyl chain would be
a favorable linker because it can occupy the hydrophobic capsule
inside the PLA2 dimer. Since the exact chemical composition
of an optimal spacer is difficult to predict, we synthesized
bivalent ligands with alkyl spacers ranging from 3 to 9 at various
substitution positions.
Figure 1. Strategy of bivalent inhibitor design. The bivalent inhibitor
can induce PLA2 molecules to form a dimer at the i-face to prevent
their binding to the lipid surface.
Synthesis of Bivalent Ligands. In order to modulate the
spacer length by one-atom increments with homogeneous
chemical compositions, we synthesized a series of compounds
(shown in Schemes 1 and 2). The compounds 3-11 were
synthesized in three steps: (a) Compounds 13-19 were prepared
by nucleophilic substitute reaction of molecule 12 and 1,ω-
dibromoalkane (the mole ratio was 2:1). (b) The hydroxymethyl
of compounds 13-19 was converted to chloromethyl to yield
compounds 20-26. (c) 3-11 were obtained by the reaction of
compounds 20-26 with 5-methoxy-2-methyl-1H-indole-3-ac-
etamide.24 10 and 11 were started from 3-hydroxyphenylmetha-
nol and 4-hydroxyphenylmethanol, respectively, and the linker
was 1,5-dibromopentane. The structures of compounds 3-11
1
were confirmed by H NMR, MS, and elemental analysis. In
order to check for the influence of the linker on the inhibition
activity, a monovalent compound 2 with a linker was synthe-
sized by steps simliar to those for compounds 3-11 but the
dibromoalkane was changed to 1-bromopentane.
Dimer Induction Study. Sedimentation experiments can
study the aggregation state of proteins in native solution, and
no dilution effect exists compared to gel filtration, which may
cause inhibitor deposition. Peter Schuck has developed a method
for the analysis of protein self-association by sedimentation
velocity experiments.27 Sedimentation velocity provides hydro-
dynamic information about molecular size distribution and
conformational changes.25,26 We used sedimentation velocity
experiments to determine the dimerization of hnps PLA2 in the
presence of inhibitors.
Figure 2. Dimer model of hnps PLA2. The protein molecules are
shown as ribbons, and the small molecules are shown as stick models.
The structure is similar to that of the pancreatic IB PLA2 dimer induced
by MJ33 (1-hexadecyl-3-(trifluoroethyl)-sn-glycero-2-phosphometha-
nol).18 The small molecule is half of the dimerizer, and the distance
between the two phenyl rings’ ortho position is 7.4 Å. The figure was
prepared using PyMOL.23
postulated that compounds that can induce dimerization of
PLA2, thus burying the i-face, may increase the binding affinity
of the monomers and regulate enzyme activity (Figure 1).
We have developed a novel CID system using phospholipase
A2 as the target protein and a bivalent inhibitor as the chemical
linker. The design of the bivalent ligand was based on the protein
dimer model, where two molecules of the human nonpancreatic
secretory PLA2 (hnps PLA2) were each bound to an indole
derivative 1 (5-methoxy-2-methyl-1H-1-benzyl-indole-3-aceta-
mide) (PDB code 1DB4)19 reported to be a potent hnps PLA2
inhibitor.20 Bis-indole derivatives 3-11 were synthesized to
contain two compound 1 molecules linked with an alkyl ether
chain of varying length. These compounds could drive the
dimerization of hnps PLA2 as confirmed by analytical ultra-
centrifugation and were found to be potent hnps PLA2 inhibitors
using continuous fluorescence assay.21
All the inhibitors were tested against hnps PLA2 for their
ability to induce dimers (Figure 3). The relative peak areas are
approximately equal to the proportions of monomer and dimer
based on their absorbance. From the result, we can see that 3-6
could induce the dimerization of hnps PLA2 in the order 5 >
3 ∼ 6 > 4. 7 showed only a weak ability to induce dimers,
while the other compounds showed no inducement. Although
10 (meta-substituting) and 11 (para-substituting) have the same
alkyl chain length as 5, they could not induce the dimerization
of hnps PLA2 (data not shown). This means that the dimer-
ization ability of the compounds is influenced by both the
length of the linker and the position of the substituent.
Compound 5 has an optimal alkyl chain length and position
that fits the capsule of the dimer well. Figure 2 shows that
the distance of the two phenyl rings of the monofunctional
inhibitors is about 7.4 Å, which correlates well with the linker
length of compound 5.
Results
Compound 5 was further studied for its mechanism of dimer
induction. Carlson and co-workers gave a method to estimate
Keq and Kc values when the dimerizer possessed conformational
equilibria.28 The effects of ligand conformational equilibria (Keq)
and cooperative receptor interactions (Kc) modulate the values
Molecular Modeling Studies. To design bivalent ligands,
we first built a dimer model for hnps PLA2 using the hnps PLA2
monomer (PDB code 1DB4) and the FTDOCK program22
(Figure 2). In the model, the i-faces were buried at the interface
of the dimer and the two substrate binding pockets formed a

3362 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12
Zhou et al.
Scheme 1. Chemical Structures of Bivalent Inhibitors and the Two Lead Compounds
Scheme 2. Synthetic Route of Compounds 3-9a
best inhibitor with an IC50 of 24 nM, approximately 5 times
more potent than compound 2.
Ka1
Ka2
E + I 8 EI; EI + E 8 E2I; Kcoop ) Ka2 ⁄ Ka1
To quantitatively calculate the Kcoop value, another set of
enzyme inhibition experiments were performed under special
conditions where the concentration of compound 5 was very
low: 7.5-40 nM, approximately the same as for the enzyme
(30 nM). The binding model was changed to 1:2. Since in the
sedimentation experiment, the free enzyme and 1:1 binary
complex could not be distinguished, making the Kcoop value hard
to calculate. According to the equations listed in the Experi-
mental Section, the two association constants were calculated
to be Ka1 ) (5.5 ( 0.3) × 107 M-1, Ka2 ) (1.1 ( 0.1) × 108
M-1, and Kcoop ) 2.0.
a (a) Br(CH2)nBr (n ) 3-9), K2CO3, CH3CN; (b) PPh3, CCl4; (c)
5-methoxy-2-methyl-1H-indole-3-acetamide, NaH, DMF.
of Ka1 and Ka2, respectively.28 A concentration-dependent
sedimentation experiment was carried out (Figure 4a) to evaluate
the cooperative interactions at the PLA2 interface. The fraction
of protein in the ternary complex is shown as a function of the
ratio of compound 5 to PLA2. When the ratio of 5/PLA2 was
increased from 0.0 to 0.5, the dimer percentage increased from
zero to a peak value of around 0.75. When the inhibitor
concentration was further increased, the dimer percentage started
to decrease, showing that the binding model of 5 and hnps PLA2
changed from 1:2 to 1:1. Because of the limited solubility of
compound 5, the maximal concentration of 5 was twice that of
PLA2. From the data available, the Kc value was small and the
cooperative interaction was not very strong. A competition
experiment with the monovalent ligand was carried out to assess
the possible impact of the hydrophobic collapse of the bis-indole
compound in water on dimerization (Figure 4b). The fraction
of ternary complex decreased slowly, indicating that Keq/Kc is
relatively small, which implied that the effect of ligand
conformational equilibria played a minor role in the dimerization
process. Because compound 5 contains a linker of five C-C
bonds, it may be too short for the molecule to fold up. We also
checked the conformation of compound 5 by 2D NMR. The
compound showed similar spectra in DMSO and in water/
DMSO mixed solvent, implying an extended conformation (data
not shown).
Discussion
In the current study, we determined the orientation of the
two protein molecules by docking them together using the
FTDOCK program. We have compared this computational
model with the reported crystal structure of porcine pancreatic
IB phospholipase A2 dimer and found that the molecular
orientation, binding interface, and interaction residues were quite
similar. The 19 interface residues in the porcine pancreatic IB
phospholipase A2 dimer (nonpolar, L2, W3, L19, M20, L31,
V65, and L118; polar, N23, N24, D66, N67, Y69, T70, S72,
N117, D119, and T120) were somewhat changed in hnps PLA2
(nonpolar, L2, V3, A18, L19, F23, V30, and F63; polar, H6,
E16, E55, K62, Y111, S113, and K115). The nonpolar residues
surrounded the opening of active site, showing that the pocket
was a hydrophobic capsule. It was possible that the alkyl linker
of ligand interacted with these nonpolar residues while the other
end of the ligand was free. Herein, ligands that could not drive
protein dimerization also showed low IC50 values.
We tested the ability of the bivalent ligands to induce
dimerization. Bivalent inhibitors with linkers of varying lengths
and positions showed different dimer induction abilities. Dimer-
ization ability decreased rapidly when the alkyl chain length
was increased from 5 to 7. The ability was also influenced by
the parity of chain length. This suggested that the conformation
of the dimer and inhibitor was relatively rigid. There is little
room in the interior of the capsule for a redundant linker. The
i-face interacts specifically through electrostatic and hydrophobic
interactions, which might be quite weak between free PLA2
molecules, as no dimer could be detected in solution by the
analytical ultracentrifugation study. These interactions become
significant in the enzyme dimer, with the bivalent molecule
contributing to the stability of the ternary complex. Jain31 and
his co-workers reported that short-range specific binding interac-
tions along the i-face and their effect on the hydrogen bonding
Enzymatic Inhibition Study. The expression and purification
of hnps PLA2 were carried out according to accepted
methods.13,21,29,30 The inhibition studies were carried out using
the reported ANS assay.21 In the IC50 experiment, the concen-
trations of inhibitors were much higher than that of the enzyme
so the binding model was 1:1. Table 1 shows the IC50 values
of lead compounds 1 and bivalent inhibitors. Compound 2 shows
better inhibition ability than 1, which implies that the linker
could increase the binding affinity about 5 times. Also, the
bivalent inhibitors are more potent than 2. Compound 5 is the
-
network provide the basis for the K*S and k*cat allostery. This
would help to explain the stability of the ternary complex. Thus,

Dimerization of Phospholipase A2
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12 3363
Figure 3. (a) Sedimentation coefficient and distribution c(s) of hnps PLA2. The black line is the free enzyme, indicating a monomeric form. The
red line comes from the mixture of hnps PLA2 and 5. The main peak shifted to a larger c(s) corresponding to the dimer. The concentrations of the
enzyme and inhibitor 5 were 50 and 25 µM, respectively. Only the complex of 5 was shown here. For other compounds, refer to the figure in
Supporting Information. (b) Relationship of the length of alkyl chain (n) and the fraction of ternary complex for compounds 3-9. 10 and 11 cannot
induce dimerization of PLA2 (not shown).
Figure 4. (a) Fraction of protein in the ternary complex is shown as a function of the ratio of dimerizer (5) to protein (PLA2). (b) Competitive
displacement of 5 by 1.
Table 1. Enzyme (hnps PLA2) Inhibition Activities of Bivalent Ligands
with Different Chain Lengths and Positions
monomer probably because of the hydrophobic nature of the
binding pocket (Lu Zhou et al., unpublished results). Therefore,
an enzymatic inhibition study was performed to calculate the
two association constants of 5. The enzymatic activities were
measured at different 5 concentrations to fit Ka1 (5.5 × 107 M-1),
Ka2 (1.1 × 108 M-1), and Kcoop(Ka2/Ka1). Kcoop was 2.0,
indicating that the protein-protein contacts also contribute to
the stability of the ternary complex, as noncooperative binding
should give a Kcoop of 1.
compd
n
linker position
IC50, nM
1
2
3
4
5
6
7
8
540 ( 110
120 ( 20
48 ( 10
60 ( 10
24 ( 8
39 ( 10
52 ( 4
58 ( 7
49 ( 5
76 ( 20
86 ( 4
5
3
4
5
6
7
8
9
5
5
o
o
o
o
o
o
o
o
m
p
The ability to induce protein dimerization and the difference
between the two association constants make compound 5
potentially useful in clinical applications. Because it was
reported that a high concentration of hnps PLA2 was detected
in rheumatoid arthritis patients, a compound like 5 could be
used to regulate the concentration of effective hnps PLA2. When
the enzyme concentration is high, the compound can form a
ternary complex with excess enzyme. When the enzyme
concentration is low, as the first order binding constant Ka1 is
not that strong, some of the enzyme molecules may form binary
complexes with the compound while the rest remain free to
perform their normal physiological functions. Of course, there
is still a long way to go before putting this regulatory method
into clinical usage.
9
10
11
the interaction at the i-face is complicated but important to dimer
formation and enzymatic activity.
To establish a binding model, the two association constants
of 5 with the enzyme molecule were calculated. Kopytek, et al.
recommended a method to simulate the Kcoop value of the
bivalent inhibitor bisMTX.4 In their work, the Ka1 value for the
binding of bisMTX with E. coli DHFR was taken approximately
as the Ka value for the binding of MTX and DHFR. However,
the method is not suitable here. For the case of PLA2 inhibitors,
the Ka1 value changed when the linker was added to the

3364 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12
Zhou et al.
3.67 (s, 4H), 2.24 (s, 6H), 2.01-2.04 (m, 4H), 1.78 (m, 2H). MS
717 (M + 1). Anal. (C43H48N4O6).
Here, we have synthesized a series of bivalent compounds,
1,ω-bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)alkane, to drive the CID of hnps PLA2. The ability
of these dimeric compounds to drive the assembly of monomeric
hnps PLA2 into the ternary complex was shown in vitro by
analytical ultracentrifugation analysis. The inhibition strengths
of these compounds were tested using a continuous fluorescence
assay. The dimerization ability of the inhibitors was found to
be related to the linker length and position. Quantitative analysis
of the association constants for the strongest binding compound,
5, revealed that the ternary complex forms mainly by the binding
of the bivalent 5 to two enzyme molecules, with some
contribution from interactions between the two enzyme mol-
ecules. The combination of bis-indole derivatives and hnps
PLA2 can be used as a novel CID system for regulating
biological networks.
3, 4, 6-11 were synthesized by a similar procedure.
1,3-Bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)propane (3). The product was recrystallized by EtOAc,
and the yield was 66%, mp 123-125 °C. 1H NMR (CDCl3) δ
7.19-7.23 (m, 2H), 6.96-7.01 (m, 6H), 6.69-6.74 (m, 6H),
6.25-6.28 (d, 2H), 5.64 (s, 2H), 5.36 (s, 2H), 5.26 (s, 4H), 4.28 (t,
4H, J ) 3.8), 3.83 (s, 6H), 3.68 (s, 4H), 2.22 (s, 6H), 1.84 (m,
2H). Anal. (C41H44N4O6).
1,4-Bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)butane (4). The product was recrystallized by EtOAc,
and the yield was 24%, mp 138-141 °C. 1H NMR (CDCl3): δ
7.19-7.23 (m, 6H), 6.93-6.96 (m, 4H), 6.69-6.74 (m, 2H), 6.24
(d, 2H), 5.64 (s, 2H), 5.32 (s, 2H), 5.22 (s, 4H), 4.65 (t, 4H), 4.20(s,
6H), 4.10 (s, 4H), 2.42 (s, 6H), 2.09 (m, 4H). Anal. (C42H46N4O6).
1,6-Bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)hexane (6). The product was recrystallized by EtOAc,
and the yield was 29%, mp 196-198 °C. 1H NMR (CDCl3): δ
7.06-7.17 (m, 4H), 6.86-6.91 (m, 4H), 6.72-6.77 (m, 4H),
6.23-6.27 (d, 2H), 5.66 (s, 2H), 5.41 (s, 2H), 5.27 (s, 4H), 4.10 (t,
4H, J ) 6.4), 3.84 (s, 6H), 3.68 (s, 4H), 2.27 (s, 6H), 1.94 (m,
4H), 1.65 (m, 4H). Anal. (C44H50N4O6).
1,7-Bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)heptane (7). The product was recrystallized by EtOAc,
mp 146-149 °C. 1H NMR (CDCl3): δ 7.27 (s, 2H), 7.12(m, 8H),
6.84 (s, 2H),6.65 (m, 4H), 6.18-6.22 (d, 2H), 5.23 (s, 4H), 4.09
(t, 4H, J ) 7.0), 3.73 (s, 6H), 3.43 (s, 4H), 2.24 (s, 6H), 1.82 (m,
4H), 1.51 (m, 6H). Anal. (C45H52N4O6).
1,8-Bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)octane (8). The product was recrystallized by EtOAc,
mp 108-110 °C. 1H NMR (CDCl3): δ 7.27 (s, 2H), 7.12 (m, 8H),
6.84 (s, 2H), 6.64-6.70 (m, 4H), 6.18-6.21 (d, 2H), 5.23 (s, 4H),
4.06 (t, 4H, J ) 5.6), 3.73 (s, 6H), 3.43 (s, 4H), 2.24 (s, 6H), 1.81
(m, 4H), 1.45 (m, 8H). Anal. (C46H54N4O6).
1,9-Bis(((3-acetamino-5-methoxy-2-methylindole)-2-methyle-
ne)phenoxy)nonane (9). The product was recrystallized by EtOAc,
mp 112-115 °C. 1H NMR (CDCl3): δ 7.27 (s, 2H), 7.12 (m, 6H),
6.97-7.01 (m, 2H), 6.85 (s, 2H), 6.65-6.71 (m, 4H), 6.18-6.21
(d, 2H), 5.24 (s, 4H), 4.05 (t, 4H, J ) 4.6), 3.73 (s, 6H), 3.43 (s,
4H), 2.25 (s, 6H), 1.79 (m, 4H), 1.39 (m, 10H). Anal. (C47H56N4O6).
1,5-Bis(((3-acetamino-5-methoxy-2-methylindole)-3-methyle-
ne)phenoxy)pentane (10). The product was recrystallized by EtOAc,
and the yield was 88%, mp 110-112 °C. 1H NMR (CDCl3): δ
7.12-7.17 (m, 4H), 6.96-6.97 (m, 2H), 6.73-6.82 (m, 4H),
6.49-6.53 (m, 4H), 5.66 (s, 2H), 5.54 (s, 2H), 5.24 (s, 4H), 3.86
(m, 10H), 3.68 (s, 4H), 2.31 (s, 6H), 1.73-1.78 (m, 4H), 1.54-1.56
(m, 2H). Anal. (C43H48N4O6).
1,5-Bis(((3-acetamino-5-methoxy-2-methylindole)-4-methyle-
ne)phenoxy)pentane (11). The product was recrystallized by EtOAc,
and the yield was 32%, mp 160-165 °C. 1H NMR (CDCl3): δ
7.12-7.17 (d, 2H) 6.97 (d, 2H, J ) 2.2), 6.90 (m, 4H), 6.78-6.83
(m, 6H), 5.41 (s, 2H), 5.62 (s, 2H), 5.21 (s, 4H), 3.90 (t, 4H), 3.84
(s, 6H), 3.67 (s, 4H), 2.30 (s, 6H), 1.80 (m, 4H), 1.65 (m, 2H).
Anal. (C43H48N4O6).
Experimental Section
Chemistry. Melting points were recorded on an X4 apparatus
and are uncorrected. Yields refer to isolated products. 1H NMR
spectra were measured on a Varian Mercury 300 M spectrometer
using TMS as internal standard. Mass spectra were recorded on a
VG-ZAB-HS spectrometer. Elemental analyses were preformed on
an Elementar Vario EL instrument. All reactions were monitored
by thin layer chromatography, carried out on silica gel 60 F-254
aluminum sheets using UV light (254 and 366 nm). The reagents
and solvents were commercially available and purified according
to conventional methods.
General Procedure for the Preparation of 1,5-Bis((2-hydroxym-
ethyl)phenoxy)pentane (Compund 15). A solution of 2-hydroxy-
benzyl alcohol (12, 1.86 g, 15mmol) and anhydrous potassium
carbonate (2.07 g, 15mmol) in 30 mL of CH3CN was added into 1,
5-dibromopentane (1.15 g, 5 mmol). Then the solution was heated
slowly to reflux and maintained at reflux until 1a was consumed (TLC
monitor, about 6 h). The reaction mixture was cooled and concentrated
at reduced pressure, and the residue was separated by column
chromatography (silica gel, the eluting solvent ranged from hexane to
50% EtOAc/hexane). An amount of 1.44 g of 15 was obtained, with
1
a yield of 91%. H NMR (CDCl3) δ7.22-7.25 (m, 4H), 6.90-6.95
(m, 4H), 4.68 (s, 4H), 4.05 (t, 3H), 2.8 (br s, 1H), 1.85-1.92 (m, 4H),
1.69-1.72 (m, 2H).
13, 14, 16-19 were synthesized following a similar procedure.
General Procedure for the Preparation of 1,5-Bis((2-chlorom-
ethyl)phenoxy)pentane (Compound 22). A solution of 15 (1.26 g,
4mmol) and triphenylphosphine (2.62 g, 10mmol) in 20 mL of
anhydrous CCl4 was heated to reflux and maintained at reflux for
about 5 h. After the reaction mixture was cooled to room
temperature, the mixture was concentrated at reduced pressure and
the residue was separated by column chromatography (silica gel,
the eluting solvent ranged from hexane to 33% EtOAc/hexane).
Pure 22 weighed 0.75 g (yield 53%). 1H NMR(CDCl3) δ 7.19-7.33
(m, 4H), 6.86-6.97 (m, 4H), 4.67 (s, 4H), 4.00-4.06 (t, 3H),
1.90-1.96 (m, 4H), 1.75-1.81 (m, 2H).
5-Methoxy-2-methyl-1H-1-(o-pentoxy)benzylindole-3-aceta-
20, 21, 23-26 were synthesized by a similar procedure.
mide (2). The product was recrystallized by ether, and the yield
1
General Procedure for the Preparation of 1,5-Bis(((3-acetamino-
5-methoxy-2-methylindole)-2-methylene)phenoxy)pentane (Com-
pound 5). Under the protection of nitrogen, a solution of 5-methoxy-
2-methyl-1H-indole-3-acetamide (0.77 g, 3.5mmol) and 50%
sodium hydride (0.25 g, 5 mmol) in 10 mL of anhydrous DMF
was stirred at room temperature for about 0.5 h. Then 22 (0.57 g,
1.6mmol) was slowly added, and the solution was stirred for 6 h at
room temperature. The solution was poured into 70 mL of water
and then stirred for about 2 h. The solid was filtered with suction
and washed with cold diethyl ether. The product was recrystallized
was 75%, mp 136-138 °C. H NMR (CDCl3): δ 7.16 (m, 3H),
6.85 (m, 2H), 6.66-6.72 (m, 2H), 6.23 (d, 1H), 5.25 (s, 2H), 4.07
(m, 2H), 3.75 (s, 3H), 3.44 (s, 3H), 2.25 (s, 3H), 1.82 (m, 2H),
1.36-1.50 (m, 6H), 0.92 (t, 3H). Anal. (C24H30N2O3).
Analytical Ultracentrifugation Analysis. Sedimentation velocity
experiments were conducted on a Beckman Optima XLA analytical
ultracentrifuge equipped with absorbance optics. Data were analyzed
with the software Sedfit, version 8.9g. Different inhibitors were
tested to examine their ability to induce enzyme dimerization. The
concentrations of enzyme and inhibitors were 50 and 25 µM,
respectively. The buffer contained 100 mM NaCl, 50 mM Tris-
HCl (pH 8.0), and 2% DMSO (v/v).
1
by EtOAc and weighed 0.60 g (yield 52%), mp 168-171 °C. H
NMR (CDCl3): δ 7.19-7.26 (t, 2H) 7.16 (d, 2H, J ) 1.2),
6.91-7.07 (m, 4H), 6.69-6.74 (m, 4H), 6.23-6.26 (d, 2H),
5.5-5.6 (d, 4H), 5.26 (s, 4H), 4.13 (t, 4H, J ) 4.2), 3.81 (s, 6H),
5 was tested in detail to obtain the data of concentration-
dependent and competitive displacement (Figure 4). In the con-

Dimerization of Phospholipase A2
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12 3365
centration-dependent experiment, the concentration of enzyme was
50 µM and that of 5 was changed from 6.25 to 100 µM. In the
competitive displacement experiment, the concentrations of enzyme
and 5 were 50 and 25 µM and the concentration of competitor was
changed from 12.5 to 100 µM.
Because the velocity of the enzyme hydrolysis reaction and substrate
could be described as V ) k[E][S] and [S] is almost constant during
the experiment, the reaction can be regarded as pseudo-first-order.
Thus, V/V0 ) [E]/E]t, where V is the velocity of the reaction with 5
and V0 is the velocity of the reaction without 5. According to the
equation
Enzymatic Inhibition Assay. The assay was carried out in a 96-
well plate utilizing a multiwell fluorometer (SpectraMax GeminiXS,
Molecular Devices). The hnps PLA2 was expressed as reported
previously.13,21,29,30 In brief, a synthesized gene was ligated to a
pET21a vector and transformed into Escherichia coli cells
(BL21(DE3) strain). And there was a refolding process in the
purification.21 All solutions were prepared with high purity (18 MX
conductivity) water. The buffer used in all the experiments was 50
mM Tris, 100 mM NaCl, pH 8.0, 10 µM ANS, 10 mM CaCl2. The
substrate was 2 mM dimyristoylphosphatidylcholine (DMPC)-
deoxycholic acid (DCA) buffer solution, which was sonicated
for 1 min to get the micelle suspension. All the samples were
dissolved in DMSO. A 200 µL reaction solution contained 15
nM PLA2, 20 µL of substrate, 0.1 mg/mL BSA, and 10 µL of
inhibitor solution. The reaction was monitored by excitation at
377 nm and emission at 470 nm. Fluorescent signals were
monitored by a kinetics mode program. Reported IC50 values,
determined by plotting concentration-velocity curves, are the
mean of at least three separate experiments. The initial velocities
of the change in fluorescence intensity with various inhibitor
concentrations were used to calculate the IC50 values.
Calculation of Kcoop Value. For the testing of inhibition activities,
the concentration of the inhibitor was much higher than that of the
enzyme (50-1000 nM vs 15 nM). Therefore, the interaction of
the inhibitor and the enzyme should follow the 1:1 model. However,
if the concentrations of the bivalent ligand and the enzyme were
comparable (7.5-40 nM vs 30 nM), they may follow the 1:2
binding model; that is, both binary and ternary complexes can form
under assay conditions. The equations below describe the possible
interactions:
1
B)-Ka2A +
Ka1
we can calculate that Ka1 ) (5.5 ( 0.3) × 107 M-1, Ka2 ) (1.1 (
0.1) × 108 M-1, and Kcoop ) 2.0.
Acknowledgment. This project was supported in part by
the Ministry of Science and Technology of China, the National
Natural Science Foundation of China (Grants 30490245,
90403001, 20473001, 20773002), and China Postdoctoral Sci-
ence Foundation (Grant 20070410433).
Supporting Information Available: Elemental analysis and
analytical ultracentrifugation analysis data for Figure 3 and gel
electrophoresis of hnps-PLA2 purification. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
(1) Ullrich, A.; Schlessinger, J. Signal transduction by receptors with
tyrosine kinase-activity. Cell 1990, 61, 203–212.
(2) Spencer, D. M.; Wandless, T. J.; Schreiber, S. L.; Crabtree, G. R.
Controlling signal-transduction with synthetic ligands. Science 1993,
262, 1019–1024.
(3) Ho, S. N.; Biggar, S. R.; Spencer, D. M.; Schreiber, S. L.; Crabtree,
G. R. Dimeric ligands define a role for transcriptional activation
domains in reinitiation. Nature 1996, 382, 822–826.
(4) Belshaw, P. J.; Ho, S. N.; Crabtree, G. R.; Schreiber, S. L. Controlling
protein association and subcellular localization with a synthetic ligand
that induces heterodimerization of proteins. Proc. Natl Acad. Sci.
U.S.A. 1996, 93, 4604–4607.
Ka1
Ka2
E + I 8 EI; EI + E 8 E2I
(5) Rivera, V. M.; Clackson, T.; Natesan, S.; Pollock, R.; Amara, J. F.;
Keenan, T.; Magari, S. R.; Phillips, T.; Courage, N. L.; Cerasoli, F.;
Holt, D. A.; Gilman, M. A humanized system for pharmacologic
control of gene expression. Nat. Med. 1996, 2, 1028–1032.
(6) Farrar, M. A.; Alberol, I.; Perlmutter, R. M. Activation of the Raf-1
kinase cascade by coumermycin-induced dimerization. Nature 1996,
383, 178–181.
(7) Kopytek, S. J.; Standaert, R. F.; Dyer, J. C. D.; Hu, J. C. Chemically
induced dimerization of dihydrofolate reductase by a homobifunctional
dimer of methotrexate. Chem. Biol. 2000, 7, 313–321.
[I]t ) [I] + [EI] + [E2I]
[E]t ) [E] + [EI] + 2[E2I]
[EI] ) Ka1[E][I]
[E2I] ) Ka2[EI][E]
(8) Harvey, I.; Garneau, P.; Pelletier, J. Forced engagement of a RNA/
protein complex by a chemical inducer of dimerization to modulate
gene expression. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 1882–1887.
(9) Liu, J. Y.; Zhang, Z. S.; Tan, X. J.; Hol, W. G. J.; Verlinde, C. L. M. J.;
Fan, E. K. Protein heterodimerization through ligand-bridged multi-
valent pre-organization: enhancing ligand binding toward both protein
targets. J. Am. Chem. Soc. 2005, 127, 2044–2045.
(10) Tachedjian, G.; Orlova, M.; Sarafianos, S. G.; Arnold, E.; Goff, S. P.
Nonnucleoside reverse transcriptase inhibitors are chemical enhancers
of dimerization of the HIV type 1 reverse transcriptase. Proc. Natl.
Acad. Sci. U.S.A. 2001, 98, 7188–7193.
(11) Berg, T. Modulation of protein-protein interactions with small organic
molecules. Angew. Chem., Int. Ed. 2003, 42, 2462–2481.
(12) Nevalainen, T. Serum phospholipases A2 in inflammatory diseases.
J. Clin. Chem. 1993, 39, 2453–2457.
(13) Pruzanski, W.; Vadas, P.; Browning, J. Secretory non-pancreatic group
II phospholipase A2: role in physiologic and inflammatory processes.
J. Lipid Mediators 1993, 8, 161–167.
where [I]t is the total concentration of bivalent ligands, [I] is the
concentration of free bivalent ligands in the solution, [E]t is the
total concentration of enzyme, [E] is the concentration of free
enzyme in the solution, [EI] is the concentration of the binary
complex of bivalent ligands and the enzyme, and [E2I] is the
concentration of the tertiary complex of bivalent ligands and two
enzyme molecules.
Here, the interaction between the substrate and PLA2 was
neglected because the [S]/Km was less than 1 (Km ) 330 µM and
[S] ) 100 µM). After some mathematic conversion, we get the
equation below, which could be used to calculate the two association
constants Ka1 and Ka2. And Kcoop ) Ka2/Ka1.
[I]t + [E] - [E]
[E]t - [E]
2[I]t + [E] - [E]t
[E]t - [E]
t[E] ) -Ka2
[E]2 +
1
Ka1
Here,
and
(14) Gelb, M. H.; Min, J. H.; Jain, M. K. Do membrane-bound enzymes
access their substrates from the membrane or aqueous phase: interfacial
versus non-interfacial enzymes. Biochim. Biophys. Acta 2000, 1488,
20–27.
(15) Berg, O. G.; Yu, B. Z.; Rogers, J.; Jain, M. K. Interfacial catalysis by
phospholipase-A2. Determination of the interfacial kinetic rate con-
stants. Biochemistry 1991, 30, 7283–7297.
2[I]t + [E] - [E]t
A)
[E]2
[E]t - [E]
[ ] [ ] [ ]
I t + E - E t[ ]
(16) Brunie, S.; Bolin, J.; Gewirth, D.; Sigler, P. B. The refined crystal
structure of dimeric phospholipase A2 at 2.5 Å. Access to a shielded
catalytic center. J. Biol. Chem. 1985, 260, 9742–9749.
B )
E
[ ] [ ]
E t - E
While [E]t was a constant (30 nM) and [I]t changed from 40 to 7.5
nM, [E] was calculated according to the experimental result.
(17) Suzuki, A.; Matsueda, E.; Yamane, T.; Ashida, T.; Kihara, H.; Ohno,
M. Crystal-structure analysis of phospholipase a(2) from trimeresurus-

3366 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 12
Zhou et al.
flavoviridis (Habu Snake) venom at 1.5 angstrom resolution. J. Bio-
chem. 1995, 117, 730–740.
(18) Pan, Y. H.; Epstein, T. M.; Jain, M. K.; Bahnson, B. J. Five coplanar
anion binding sites on one face of phospholipase A(2). Relationship
to interface binding. Biochemistry 2001, 40, 609–617.
(19) Schevitz, R. W.; Bach, N. J.; Carlson, D. G.; Chirgadze, N. Y.;
Clawson, D. K.; Dillard, R. D.; Draheim, S. E.; Hartley, L. W.; Jones,
N. D.; Mihelich, E. D.; Olkowski, J. L.; Snyder, D. W.; Sommers,
C.; Wery, J. P. Structure-based design of the first potent and selective
inhibitor of human nonpancreatic secretory phospholipase-a(2). Nat.
Struct. Biol. 1995, 2, 458–465.
(20) Dillard, R. D.; Bach, N. J.; Draheim, S. E.; Berry, D. R.; Carlson,
D. G.; Chirgadze, N. Y.; Clawson, D. K.; Hartley, L. W.; Johnson,
L. M.; Jones, N. D.; McKinney, E. R.; Mihelich, E. D.; Olkowski,
J. L.; Schevitz, R. W.; Smith, A. C.; Snyder, D. W.; Sommers, C. D.;
Wery, J. P. Indole inhibitors of human nonpancreatic secretory
phospholipase A2. 2. Indole-3-acetamides with additional functionality.
J. Med. Chem. 1996, 39, 5137–5158.
(24) Liu, Y.; Han, X. F.; Huang, C. K.; Hao, X.; Lai, L. H. Indole-5-
phenylcarbamate derivatives as human non-pancreatic secretory phos-
pholipase A2 inhibitor. Bioorg. Med. Chem. Lett. 2005, 15, 4540–4542.
(25) Lebowitz, J.; Lewis, M. S.; Schuck, P. Modern analytical ultracen-
trifugation in protein science: a tutorial review. Protein Sci. 2002, 11,
2067–2079.
(26) Laue, T. M.; Stafford, W. F. Modern applications of analytical ultracen-
trifugation. Annu. ReV. Biophys. Biomol. Struct. 1999, 28, 75–100.
(27) Schuck, P. On the analysis of protein self-association by sedimentation
velocity analytical ultracentrifugation. Anal. Biochem. 2003, 320, 104–
124.
(28) Carlson, J. C. T.; Kanter, A.; Thuduppathy, G. R.; Cody, V.; Pineda,
P. E.; McIvor, R. S.; Wagner, C. R. Designing protein dimerizers:
the importance of ligand conformational equilibria. J. Am. Chem. Soc.
2003, 125, 1501–1507.
(29) Othman, R.; Baker, S.; Li, Y.; Worrall, A. F.; Wilton, D. C. Human
non-pancreatic (group II) secreted phospholipase A(2) expressed from
a synthetic gene in Escherichia coli: characterisation of N-terminal
mutants. Biochim. Biophys. Acta 1996, 2, 92–102.
(30) Han, S. K.; Lee, B. I.; Cho, W. H. Bacterial expression and
characterization of human pancreatic phospholipase A(2). Biochim.
Biophys. Acta 1997, 1346, 185–192.
(31) Jain, M. K.; Berg, O. G. Coupling of the i-face and the active site of
phospholipase A(2) for interfacial activation. Curr. Opin. Chem. Biol
2006, 10, 473–479.
(21) Huang, C. K.; Zhou, L.; Liu, Y.; Lai, L. H. A continuous fluorescence
assay for phospholipase A2 with nontagged lipid. Anal. Biochem. 2006,
351, 11–17.
(22) Jackson, R. M.; Gabb, H. A.; Sternberg, M. J. E. Rapid refinement of
protein interfaces incorporating solvation: application to the docking
problem. J. Mol. Biol. 1998, 276, 265–285.
(23) Warren, S. C.; DeLano, L. The PyMOL Molecular Graphics System;
DeLano Scientific LLC: Palo Alto, CA; http://pymol.sourceforge.net/.
JM7010707