Modeling, chemistry, and biology of the benzolactam analogues of indolactam V (ILV). 2. Identification of the binding site of the benzolactams in the CRD2 activator-binding domain of PKCδ and discovery of an ILV analogue of improved isozyme selectivity

Article abstract of DOI:10.1021/jm960875h

Protein kinase C (PKC) is a complex enzyme system comprised of at least 11 isozymes that serves to mediate numerous extracellular signals which generate lipid second messengers. The discovery of isozyme-selective activators and inhibitors (modulators) of

Full text of DOI:10.1021/jm960875h

1316
J . Med. Chem. 1997, 40, 1316-1326
Articles
Mod elin g, Ch em istr y, a n d Biology of th e Ben zola cta m An a logu es of In d ola cta m
V (ILV). 2. Id en tifica tion of th e Bin d in g Site of th e Ben zola cta m s in th e CRD2
Activa tor -Bin d in g Dom a in of P KCδ a n d Discover y of a n ILV An a logu e of
Im p r oved Isozym e Selectivity
Alan P. Kozikowski,*,† Shaomeng Wang,† Dawei Ma,*,‡ J iangchao Yao,‡ Shakeel Ahmad,§ Robert I. Glazer,§
Krisztina Bogi,| Peter Acs,| Shayan Modarres,| Nancy E. Lewin,| and Peter M. Blumberg|
Institute for Cognitive and Computational Sciences and Department of Pharmacology, Georgetown University Medical Center,
3970 Reservoir Road Northwest, Washington, DC 20007-2197, Shanghai Institute of Organic Chemistry, Chinese Academy of
Sciences, 354 Fengling Lu, Shanghai 200032, China, Molecular Mechanisms of Tumor Promotion Section, Laboratory of
Cellular Carcinogenesis and Tumor Promotion, National Cancer Institute,
Bethesda, Maryland 20892
Received December 27, 1996X
Protein kinase C (PKC) is a complex enzyme system comprised of at least 11 isozymes that
serves to mediate numerous extracellular signals which generate lipid second messengers. The
discovery of isozyme-selective activators and inhibitors (modulators) of PKC is crucial to
ascertaining the role of the individual isozymes in physiological and pathophysiological processes
and to manipulating their function. The discovery of such small molecule modulators of PKC
is at present a largely unmet pharmacological need. Herein we detail our modeling studies
which reveal how the natural product indolactam V (ILV) and its 8-membered ring analogue,
the benzolactam 15, bind to the CRD2 activator domain of PKC. These modeling studies reveal
that not all PKC ligands possess a common pharmacophore, and further suggest an important
role of specific hydrophobic contacts in the PKC-ligand interaction. The modeling studies
find strong experimental support from mutagenesis studies on PKCR that reveal the crucial
role played by the residues proline 11, leucine 20, leucine 24, and glycine 27. Next, we describe
the synthesis of two 8-substituted benzolactams starting from L-phenylalanine and characterize
their isozyme selectivity; one of the two benzolactams exhibits improved isozyme selectivity
relative to the n-octyl-ILV. Lastly, we report inhibition of cellular proliferation of two different
breast carcinoma cell lines by the benzolactam 5 and show that the compound preferentially
down-regulates PKCâ in both cell lines.
In tr od u ction
as well as to examine the effects of such agents on
cellular growth and differentiation.2
Protein kinase C (PKC) is a ubiquitous signal trans-
ducing enzyme system that plays a marked role in
diverse cellular processes including regulation of ion
channels, neurotransmitter release, growth and dif-
ferentiation, apoptosis, and neuronal plasticity.1-8 This
complex enzyme system serves to mediate numerous
signals originating from the induction of lipid hydroly-
sis. The discovery of isozyme-selective activators and
inhibitors (modulators) of protein kinase C (PKC) is
crucial to both dissecting the role of the individual
isozymes in physiological and pathophysiological pro-
cesses and to manipulating these pathways. The dis-
covery of such small molecule modulators of PKC is still
in the early stages. It has been our objective to
synthesize analogues of the tumor-promoting substance
lyngbyatoxin (or more precisely, its simpler congener
indolactam V, ILV) and to assess the ability of these
new materials to modulate the various isozymes of PKC
Specifically, it is our aim to learn more about the
nature of the structure-activity relationships that
govern isozyme-selective binding, activation, inhibition,
translocation, and/or down-regulation of PKC by ana-
logues of indolactam (ILV) and to discover isozyme-
selective modulators of this multifunctional kinase
which function through the phorbol ester binding site.
Biological investigations have revealed that there exist
at least 11 isozymes of protein kinase C (see Figure 1)
bearing closely related amino acid sequences.3,5 The
isozymes exhibit distinct patterns of tissue specific
expression and intracellular localization. Complete
details of the functional significance of these different
isozymes and their substrate specificities remain to be
elucidated; however, recent work reveals that the
“distinct isozymes of PKC can exert different effects on
growth control and malignant transformation in the
same cell type”.1,3
† Institute for Cognitive and Computational Sciences, Georgetown
University Medical Center.
In view of the fact that bryostatin-1, a nonpromoting
PKC activator, is in clinical evaluation as an anticancer
agent, it is likely that isozyme-selective activators of
PKC that are not tumor promoters will find use in
treating neoplastic conditions. The recent structural
‡ Shanghai Institute of Organic Chemistry.
§ Department of Pharmacology, Georgetown University Medical
Center.
| Molecular Mechanisms of Tumor Promotion Section, NCI.
X Abstract published in Advance ACS Abstracts, April 1, 1997.
S0022-2623(96)00875-8 CCC: $14.00 © 1997 American Chemical Society

Benzolactam Analogues of ILV
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9 1317
F igu r e 1. Structural organization of the PKC gene family.
PKC is divided into five variable (V1-V5) and four constant
(C1-C4) domains. The Ca2+- and phospholipid-dependent
isoforms (R, â, γ) differ from the Ca2+-independent and
phospholipid-dependent isoforms by the presence of a Ca2+
binding (C2) domain. The unique PKCµ contains a putative
transmembrane domain (J ohannes, F. J .; Prestle, J .; Eis, S.;
Oberhagemann, P.; Pfizenmaier, K. J . Biol. Chem. 1994, 269,
6140).
gested a commonality of their hydrophobic regions and
certain heteroatoms.5,16-19
On the basis of our previous synthetic efforts aimed
at identifying analogues of ILV that were synthetically
more accessible, we were the first group to report routes
to certain benzolactam analogues of ILV in which the
pyrrole ring was absent.21 Previously, we had found
that the 9-membered benzolactam 4 was not active as
a PKC activator, presumably because of its inability to
adopt the twist conformation. In continuation of this
theme, we now describe a novel synthetic approach to
the 8-membered benzolactams 5 and 6. These com-
pounds, unlike 4, are forced to adopt the twist confor-
mation as a consequence of the smaller ring size
constraining the amide group to adopt solely the cis
stereochemistry. Endo et al. have previously reported
the synthesis of some related compounds.15 Herein we
detail our modeling studies which reveal how ILV and
the 8-membered ring benzolactams bind to the CRD2
activator domain of PKC. Next, we describe the syn-
thesis of two 8-substituted benzolactams starting from
L-phenylalanine and determine the isozyme selectivity
of these molecules. As will be seen, one of the benzolac-
tams exhibits improved isozyme selectivity relative to
n-octyl-ILV. Lastly, we describe the effects of one of
these benzolactams on cellular proliferation.
determination of the complex between phorbol 13-
acetate and the CRD2 activator-binding domain of
protein kinase Cδ,5 coupled with the insights gained
from modeling the interactions of diverse ligands to
PKC, provides unique opportunities for the discovery
of PKC modulators with novel properties.
Ch em istr y
In tr od u ction to th e Teleocid in s, Lyn gbya toxin s,
a n d ILV. Teleocidin was first isolated from the mycelia
of Streptomyces mediocidicus as a mixture of highly
toxic compounds by Sakai et al.9 The structure of one
of these metabolites was assigned by Hirata as shown
by formula 1. Lyngbyatoxin A was first isolated from
the lipid extract of a Hawaiian shallow-water variety
of blue green alga, Lyngbya majuscula Gomont.10 The
structure of this compound is closely related to that of
the teleocidin family. The lyngbyatoxin series can be
obtained together with the teleocidin B group from S.
mediocidicus.11 Therefore, they were also named as
teleocidin A-1 (2a , which is identical to lyngbyatoxin A)
and A-2 (2b) by Sakai.11 Indolactam V (3, ILV), which
contains the basic ring structure of the teleocidins, is
the simplest member of the family and is produced in
large quantities by actinomycetes strain NA34-17.12,13
On the basis of both solution NMR studies and molec-
ular mechanics calculations, it has been reported that
the indolactam portion (indolactam V, 3) of the teleo-
cidins and lyngbyatoxins can exist in two conformational
states, the sofa-like conformation containing a trans
amide bond and the twist-like conformation containing
a cis amide bond. At equilibrium, the ratio of twist/
sofa was 2.8; the twist form of ILV represents the
biologically active conformation as is discussed further
below.14,15
Molecu la r -Modelin g a n d Site-Dir ected Mu ta gen -
esis Stu d ies. Previous molecular-modeling efforts
based upon ligand structures and the structure-activity
relationships of various PKC activators have failed to
fully identify the “PKC pharmacophore” in teleocidin
and ILV.16-19,26-29 This shortcoming can be attributed
in part to the conformational flexibility of the 9-mem-
bered ring system in teleocidin and ILV and to the
unique structural features of this class of compounds.
However, to a larger extent, this failure is due to
The teleocidins, like the phorbol ester 12-O-tetradeca-
noylphorbol 13-acetate (TPA), bryostatin, and aplysia-
toxin, bind avidly to the regulatory domain of PKC, thus
serving as potent activators of this enzyme. These
compounds serve as structural mimics of the endog-
enous activator of PKC, diacylglycerol. Indeed, computer-
assisted molecular-modeling studies of these tumor
promoters, while somewhat controversial, have sug-

1318 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9
Kozikowski et al.
F igu r e 2. X-ray structure of phorbol 13-acetate in complex
with PKCδ CRD2 together with its specific hydrogen bond
network.
limitations inherent in ligand-based molecular model-
ing. Recently, an X-ray structure of phorbol 13-acetate
in complex with PKCδ CRD2 activator-binding domain
was solved (Figure 2).5 The X-ray structure offered us
an unprecedented opportunity to elucidate how teleo-
cidin and ILV bind to PKC using molecular-modeling
methods. Additionally, and as detailed below, we have
modeled the interaction of the benzolactam analogue of
ILV in complex with PKC. The coordinates of the
phorbol 13-acetate/PKCδ CRD2 complex were kindly
made available to us by Dr. J ames Hurley at the NIH
prior to their deposit to the Brookhaven National
Laboratories databank.
Teleocidin and ILV exist in two major conformations
in solution, namely the trans-sofa (trans refers to the
stereochemistry of the amide bond) and cis-twist.14 At
equilibrium, the ratio of twist/sofa was found to be 2.8.14
To determine unambiguously which conformation rep-
resents the biologically active one, we and others have
designed compounds that exclusively adopt one confor-
mation. The 9-membered ring benzolactam 4 exclu-
sively adopts the sofa conformation, and its analogues
with or without appropriate hydrophobic substituents
show very little or no activity.15,21 On the other hand,
the 8-membered ring benzolactam adopts predomi-
nantly the twist conformation, and its analogues with
appropriate hydrophobic substituents indeed display
high binding affinities for PKC (see the accompanying
Biological Results section). These experiments have
unambiguously established that the cis-twist conforma-
tion of ILV and teleocidin represents the biologically
active conformation when binding to PKC. Figure 3
displays comparisons of the conformations of the 8- and
9-membered ring benzolactams together with the sofa
and the twist conformations of ILV.
F igu r e 3. Comparison of the conformations of the 8- and
9-membered ring benzolactams and the twist and sofa con-
formations of ILV.
desirable to deduce a precise binding model for the
interaction of these ligands with PKC using a modeling
approach based upon the X-ray structure of phorbol 13-
acetate in complex with PKCδ CRD2.
In this X-ray structure,5 the phorbol ester forms four
specific hydrogen bonds with PKC via two hydroxyl
groups and one carbonyl group (Figure 2). Specifically,
the hydroxyl group at C-4 functions as a hydrogen bond
donor toward the carbonyl group of Gly 23. The
carbonyl group at C-3 functions as a hydrogen bond
acceptor toward the amino group of Gly 23. The
hydroxyl group at C-20 functions both as a hydrogen
bond donor and an acceptor. As a donor, it forms a
hydrogen bond with the carbonyl group of Leu 21, while
as an acceptor, it forms a hydrogen bond with the amino
group of Thr 12. The four hydrogen bonds have optimal
or nearly optimal geometries for hydrogen bond forma-
tion between donor and acceptor, with the exception of
the hydrogen bond formed between the hydroxyl at C-4
and the carbonyl group of Gly 23.
As is apparent, ILV, teleocidin, and the 8-membered
ring benzolactam all possess a carbonyl group, an amido
group, and a primary hydroxyl group. Among these
three ligands and in comparison to phorbol 13-acetate,
it would seem reasonable that the primary hydroxyl and
the carbonyl groups should be equivalent, while the
amido NH group in ILV, teleocidin, and the benzolac-
tam, which can only function as a hydrogen bond donor,
would be equivalent to the C-4 hydroxyl group of phorbol
ester. However, using their active cis-twist conforma-
tions, the superposition of ILV, teleocidin, and the
8-membered ring benzolactam on phorbol ester results
in a poor overlap of these crucial hydrogen-bonding
groups, results consistent with those reported by Itai
et al.16 Furthermore, docking of teleocidin, ILV, and the
8-membered ring benzolactam to the phorbol ester
binding site in PKC, using the binding model of phorbol
13-ester revealed by the X-ray structure study,5 con-
firmed that these ligands are unlikely to bind to PKC
Despite the determination of the active conformation
of ILV and teleocidin, it was still not entirely clear what
constitutes the actual pharmacophore in teleocidin, ILV,
and their 8-membered ring analogues, and exactly how
these compounds bind to PKC. Such information would,
of course, be invaluable to our efforts directed toward
the design of isozyme-selective PKC modulators, and to
acquiring a better understanding of the mechanism of
activation of PKC by these compounds. Thus, it became

Benzolactam Analogues of ILV
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9 1319
Sch em e 1. An L-Phenylalanine-Based Route to the Benzolactam Analogue of ILV
using this arrangement of hydrogen bond donor and
acceptor groups.
We therefore decided to fully explore the possible
binding modes of teleocidin, ILV, and the 8-membered
ring benzolactam to PKC. Teleocidin, ILV, and the
benzolactam are known to directly compete with phorbol
ester binding to PKC. Our recent mutagenesis studies
have also clearly shown that a number of residues, such
as Pro 11, Leu 21, and Gln 27, are important to the
binding of both phorbol esters, ILV, and the benzolactam
to PKC.20,30 These results all sugest that ILV, teleoci-
din, the 8-membered ring benzolactam, and the phorbol
esters should occupy the same binding region or that
their binding regions at least partially overlap.
Teleocidin, ILV, the 8-membered ring benzolactam,
phorbol esters, and diacylglycerol (DAG, the natural
ligand for PKC) all contain a primary hydroxyl group
F igu r e 4. The overall features of the binding model for the
which is critical for their activity. All previous molec-
ular-modeling studies and structure-activity relation-
ships have indicated that these primary hydroxyl groups,
shared by many different classes of high-affinity PKC
ligands, are equivalent.16-19,26-28 For these reasons, we
used this particular group as the anchor point in
exploring the binding of the ligands in question. From
the modeling studies, it was found that the hydrophobic
tail attached to the phenyl group in compounds 5 and
6 has very little interaction with the receptor. Presum-
ably, the hydrophobic tail in these ligands contributes
to the binding by modifying ligand hydrophobicity and
by affecting ligand-lipid interactions rather than by
making direct contacts to the receptor. Therefore, for
reasons of simplicity, the interactions between the
8-membered ring benzolactams and PKC were calcu-
lated using a ligand without the hydrophobic tail (15,
see Scheme 1).
8-membered ring benzolactam in complex with PKCδ CRD2.
The protein backbone is shown in a ribbon diagram. The
ligand is displayed as a ball and stick model, and the three
hydrophobic residues in close contact with the ligand are
displayed as a stick model.
lations, a binding model, displayed in Figures 4 and 5,
was identified. This binding model affords a maximum
interaction energy between the 8-membered ring ben-
zolactam and the protein, with both components main-
taining their low-energy conformations. The binding
models calculated for ILV, teleocidin, and their 8-mem-
bered benzolactam analogue are very similar,32 as would
be anticipated. Below we provide details of the binding
model for the 8-membered ring benzolactam. Unless
indicated otherwise, identical interaction features also
apply to ILV and teleocidin. More comprehensive and
quantitative molecular-modeling analyses for these
ligands will be published separately.32
Through extensive manual docking studies, followed
by energy minimization and molecular dynamics simu-
This binding model (Figures 4 and 5) allows us to fully

1320 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9
Kozikowski et al.
obtained from the molecular dynamics simulations are
summarized in Table 1. As can be seen from Table 1,
the hydrogen bond distances between the hydrogen bond
donor and the acceptor for these four hydrogen bonds
are between 2.79 and 3.01 Å, close to the optimum
value. The A-H-D angles are between 147° and 159°,
within 35° of the optimum value (180°). The AA-A-H
angles are between 116° and 153.5°. With the exception
of the hydrogen bond formed between the amido group
of the benzolactam and the carbonyl group of Leu 21,
the other three hydrogen bond angles are within 15° of
the optimum value. Thus, based upon these hydrogen
bond parameters, all four bonds can be characterized
as strong or fairly strong hydrogen bonds.
Appropriate hydrophobicity has been recognized to be
required for PKC ligands to achieve high affinity.15,33
However, prior to this receptor structure-based model-
ing study, specific hydrophobic interactions between the
PKC ligands and PKC were not well understood. Now,
for the first time, our binding model sheds light on the
importance of some very “specific” hydrophobic interac-
tions between the hydrophobic moieties of ILV, teleo-
cidin, and the 8-membered ring benzolactam and hy-
drophobic residues in the protein (Figure 4). In the
binding model for benzolactam (Figure 4), the phenyl
ring resides on the top of Pro 11, parallel to the proline
ring, allowing for effective hydrophobic interactions. The
N-1 methyl group in the ligand interacts well with the
hydrophobic side chain of Leu 20 and part of the Pro
11 ring. The isopropyl group at C-2 in the ligand
interacts nicely with the hydrophobic side chain of Leu
24. In addition to these hydrophobic interactions,
several carbon atoms in the benzolactam ring (C-3, C-5,
C-6), as well as the C-11 atom, appear to engage in
substantial hydrophobic interactions with the receptor.
F igu r e 5. The specific hydrogen bond network formed be-
tween the 8-membered ring benzolactam and PKCδ CRD2.
appreciate the role that each functional group in the
8-membered ring benzolactam plays in its binding to
PKC. The primary hydroxyl group at C-5, acting as a
hydrogen bond donor, forms a hydrogen bond with the
carbonyl group of residue Leu 21, while this same group,
acting as a hydrogen bond acceptor, forms a hydrogen
bond with the amido group of Thr 12 (Figure 5). Thus,
this primary hydroxyl group is functionally equivalent
to the hydroxyl group at C-20 in phorbol 13-acetate
(Figure 2).5 The carbonyl group at position 3 forms a
hydrogen bond with the amido group of Gly 23, which
is equivalent to the carbonyl group at position 3 in
phorbol 13-acetate.5 The amido group at position 4
forms a hydrogen bond with the carbonyl group of Leu
21. It is noteworthy that this hydrogen bond was not
found in the hydrogen bond network formed between
phorbol 13-acetate and PKC (Figure 2), and is thus
unique to the 8-membered ring benzolactam, ILV, and
teleocidin (Figure 5).5 On the other hand, the hydrogen
bond formed between the hydroxyl group at C-4 in
phorbol 13-acetate and the carbonyl group of Gly 23 in
PKC is absent in the hydrogen bond network of the
benzolactam and PKC system (Figure 2).5
To obtain a quantitative measure of the strength of
these four hydrogen bonds formed between benzolactam
and PKC, hydrogen bond geometric parameters were
obtained by analyzing 1000 trajectories recorded during
100 ps molecular dynamics simulations. Hydrogen bond
interactions involve four atoms, a donor atom (D), a
hydrogen atom (H), an acceptor atom (A), and an
acceptor antecedent (AA). The strength of a hydrogen
bond depends upon the nature of the donor and acceptor
groups, and at least three geometric parameters: the
distance between the donor and acceptor atoms; the
angle of the acceptor, the hydrogen atom, and the donor
(A-H-D); and another angle comprised of the acceptor
antecedent, the acceptor, and the hydrogen atom (AA-
A-H). Various studies have shown that the optimum
value for the hydrogen bond distance between the donor
and the acceptor atoms is around 2.8-3.0 Å, the
optimum value for the A-H-D angle is near 180°, and
the optimum value of the AA-A-H angle is 120°. The
hydrogen bond parameters for all four hydrogen bonds
In order to assess these hydrophobic interactions more
quantitatively, we have performed HINT analyses48 on
a total of 11 trajectories recorded at 10 ps intervals
during 100 ps molecular dynamics simulations; the
results are summarized in Table 2. As is apparent from
Table 2, three residues are primarily responsible for the
hydrophobic interactions between the benzolactam and
PKC, namely Pro 11, Leu 20, and Leu 24. Together,
these three residues account for 79% of the total
hydrophobic interactions between the ligand and the
receptor based upon the HINT analyses.48 It is of
interest to note that the backbone carbon atom of Gly
23 and the two carbon atoms of the side chain of Gln
27 also contribute 7% and 8%, respectively, to the total
hydrophobicity.
In regard to the ligand, the two most important
groups involved in making the hydrophobic contacts are
the N-methyl and the isopropyl groups. Together, they
account for nearly half of the total hydrophobic interac-
tions. Indeed, previous structure-activity relationship
studies on ILV have revealed the importance of these
groups, for the conversion of the N-methyl group to
hydrogen results in a greater than 100-fold reduction
in activity,35 while the conversion of the isopropyl group
to methyl diminishes the activity by 1000-fold.35 The
phenyl ring contributes 9% to the total. It is of interest
to note that both C-3 and C-11 make significant
contributions, accounting for 18% and 15%, respectively,
of the total. The latter contributions may be due to the
fact that these two atoms reside deep within the binding

Benzolactam Analogues of ILV
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9 1321
Ta ble 1. Hydrogen Bond Parameters Derived from the Analysis of 1000 Trajectories Recorded during 100 ps Molecular Dynamics
Simulations of the Complex Structure of PKC/Benzolactam
rAD ( deviation
θA-H-D ( deviation
θAA-A-H ( deviation
hydrogen bond donor group
hydrogen bond acceptor group
(Å)
(deg)
(deg)
N-H (Thr 12, PKC)
O-H (primary hydroxyl, ligand)
N-H (ligand)
O-H (primary hydroxyl, ligand)
CdO (Leu 21, PKC)
CdO (Leu 21, PKC)
CdO (ligand)
3.01 ( 0.15
2.79 ( 0.13
2.93 ( 0.13
2.98 ( 0.20
2.8-3.0
158.2 ( 11.2
155.1 ( 11.4
159.6 ( 10.6
147.5 ( 15.5
160-180
115.9 ( 10.1
131.5 ( 8.8
153.5 ( 8.8
121.2 ( 11.4
109-120
N-H (Gly 23, PKC)
optimal value
Ta ble 2. Hydrophobic Interaction Analysis of 11 Trajectories Recorded at the End of Each 10 ps during 100 ps Molecular Dynamics
Simulations of the Complex Structure of PKC/Benzolactam
atom or group in benzolactam
interacting residues in PKCa
HINT hydrophobic score
% of total
N-methyl
isopropyl group
benzo group
C-3
P r o 11, Leu 20
Leu 20, Trp 22, Gly 23, Leu 24, Gln 27
Tyr 8, Ser 10, P r o 11, Leu 24
Leu 20, Trp 22, Gly 23, Leu 24, Gln 27
Tyr 8, Pro 11, Leu 20, Gly 23, Gln 27
119
75
35
71
25
30
19
9
18
4
C-5
C-11
C-6
Tyr 8, Pro 11, Leu 20, Gly 23, Leu 20, Leu 21, Gly 23, Gln 27
Tyr 8, Pro 11, Leu 20, Gly 23, Leu 24, Gln 27
57
21
15
5
all other atoms
0
0
100
sum
393b
residues
interaction with atoms or groups in benzolactam
HINT hydrophobic score
% of total
Tyr 8
Ser 10
C-5, C-6, C-11, benzene ring
ben zen e r in g
13
4
3
1
Pro 11
Thr 12
N-m eth yl, ben zen e r in g, C-5, C-6, C-11
C-11
97
1
25
0
Leu 20
Leu 21
N-m eth yl, isopropyl, C-3, C-5, C-6, C-11
C-11
158
4
40
1
Trp 22
C-3
5
1
Gly 23
Leu 24
Gln 27
all other residues
isop r op yl, C-3, C-5, C-6, C-11
isop r op yl
C-3, C-5, C-6, C-11
28
54
31
0
7
14
8
0
100
sum
395b
a
b
Boldprint indicates the atom, group, or residue that makes significant contributions to the interactions. The difference between
these two values is due to rounding errors.
Ta ble 3. Site-Directed Mutagenesis Analysis: Ki Values for the Inhibition of [3H]PDBu Binding to PKC Mutants by Benzolactam 5
and PDBu
Ki (nM) (5)
ratio (mutant/wild type)
Ki (nM) (PDBu)
ratio (mutant/wild type)
wild-type
10.7
730
134
ndb
nd
0.8a
Pro 11 f Gly
Leu 20 f Gly
Leu 24 f Gly
Gln 27 f Gly
Gln 27 f Val
68
13
-
100a
125
15
-
-
27
12a
no bindinga
no bindinga
21.6a
-
933
88
a
b
Data taken from ref 30. No data since the reference ligand, PDBu, does not bind to the mutant.
pocket, and each of them is surrounded by a number of
receptor carbon atoms.
As expected, the Pro 11, Leu 20, and Leu 24 f Gly
mutations all resulted in a significant reduction in the
binding affinity of both 5 and PDBu. The mutation of
Pro 11 f Gly reduces the binding affinity of 5 to the
receptor by 68-fold and the binding affinity of PDBu by
125-fold. The loss in the binding affinity of these two
ligands may be accounted for by two factors: (1) the loss
of the hydrophobic contacts between the ligands and the
receptor and (2) possible conformational changes in the
binding site resulting from the mutation. The later
factor is particularly important in the case of Pro, as it
is a fairly rigid amino acid which is known to be
important for protein folding. The mutation of Leu 20
to Gly also greatly reduces the binding affinities of 5
and PDBu, although not as much as observed for Pro
11 f Gly. In this case, since the side chain of Leu 20
resides outside of the binding groove formed by the
backbone of residues 8-12 and 21-27, this mutation
should not affect the conformation of the binding groove.
Therefore, the reduction in the binding affinity of 5 and
PDBu to this mutant should reflect solely the loss of
As can be seen from the HINT analyses, Pro 11, Leu
20, and Leu 24 are primarily responsible for the
hydrophobic interactions between the ligand and the
receptor. To further investigate the effect of these
hydrophobic interactions on ligand binding affinity, site-
directed mutagenesis studies were performed using the
PKCδ CRD2 activator binding domain. In these experi-
ments, each of these three residues was individually
mutated into Gly. Since Gln 27 has been shown to play
a fundamental role in maintaining the binding site
conformation of PKC5 and to be somewhat important
for the hydrophobic interactions (8%), this residue was
also mutated into either glycine or valine. Compound
5 was then assayed for its ability to bind to these
mutants in competition with [3H]PDBu. The results of
these site-directed mutagenesis experiments are sum-
marized in Table 3. For comparison, the binding
affinities of PDBu to these mutants are also provided
in the table.

1322 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9
Kozikowski et al.
Ta ble 4. Ki Values ( SEM for the Inhibition of [3H]PBDu Binding by the Compounds Tested
compd
R
â
γ
δ
ꢀ
5
6
14.7 ( 1.3
46.6 ( 8.0
334.0 ( 14.0
11.0
17.4 ( 2.2
40.7 ( 8.9
140 ( 25
122 ( 22
185 ( 30
142 ( 3
58.2 ( 12.6
187 ( 22
15
ILVa
6.1
0.4
19.4
1.2
8.2
0.8
21.9
1.0
n∼octyl-ILVa
0.5
a
Data taken from ref 4.
the hydrophobic interactions between ligand and recep-
tor. The effect of the Leu 24 f Gly mutation is more
dramatic. No PDBu binding is observed with this
mutant, and as a consequence, the binding of 5 cannot
be measured since PDBu is used as the reference ligand.
The complete loss of PDBu binding to this mutant
cannot be explained simply by the loss of the hydropho-
bic interactions between the ligand and receptor, since
we have shown that both Pro 11 and Leu 20 make larger
contributions to the total hydrophobic interaction pro-
file, while at the same time the mutation of these
residues to Gly does not lead to a total loss in binding.
Thus, a significant conformational change is likely to
have taken place in the Leu 24 f Gly mutant. Although
the backbone of Leu 24 does not participate in hydrogen
bond interactions, its adjacent residue, Gly 23, forms
two hydrogen bonds with phorbol ester as observed in
the X-ray structure,5 and one hydrogen bond with the
benzolactam, as shown in our modeling study. The
mutation of Leu 24 to Gly essentially places two Gly
residues adjacent to one another, which could change
the conformation of Gly 23 significantly, since Gly has
been shown to be able to adopt diverse conformations
in protein folding. Gln 27 has been shown to play an
important role in maintaining the binding site confor-
mation through formation of three hydrogen bonds with
Tyr 8 and Gly 23.5 Indeed, the mutation of this residue
to Gly completely abolishes its ability to bind PDBu.30
However, the mutation of Gln 27 to Val reduces the
binding affinity of PDBu by 27-fold and the binding
affinity of 5 by 88-fold. These results suggest that Val
can more or less preserve the binding site conformation,
perhaps because its size is somewhat similar to that of
Gln.
a large contribution in the binding of these ligands to
PKC through specific hydrophobic interactions with the
protein.
Syn th etic Ch em istr y. An enantiospecific route to
these 8-membered benzolactam analogues bearing a 10-
carbon chain that uses L-phenylalanine as the starting
material is provided in Scheme 1. The L-phenylalanine
was subjected to reduction with sodium borohydride/
iodine,36 and the intermediate amino alcohol 7 was
reacted first with ethyl chloroformate and then with
acetic anhydride to provide urethane 8. Nitration of 8
with nitric acid afforded an inseparable mixture of the
ortho- and para-substituted products, in which the
undesired para isomer predominated. Fortunately, we
found that further transformation of these nitro com-
pounds to the corresponding anilines 9 and 10 by
hydrogenation over palladium on carbon allowed for
ready separation of the regioisomers. Next, reaction of
the aniline 9 with the valine-derived triflate 11 by
refluxing in dichloroethane,37 followed by the base-
assisted removal of all three protecting groups, yielded
13. The aliphatic amino group was reprotected as its
BOC derivative, the acid activated as its NHS ester
derivative 14 by reaction with N-hydroxysuccinimide/
DCC/HOBt,38 the BOC group removed, and lactam
formation initiated by exposure to aqueous NaHCO3.
Reductive N-methylation of the aniline nitrogen then
provided the parent benzolactam 15. Compound 15 was
functionalized further by iodination39 to give 16 which
was subjected to a palladium-catalyzed coupling reac-
tion with 1-decyne40 to provide acetylene 5. Hydrogena-
tion of 5 over Pd/C provided the 8-decyl derivative 6.
Biologica l Resu lts
In summary, our calculated binding models for the
8-membered ring benzolactam, ILV, and teleocidin are
fully consistent with all known structure-activity data
for teleocidin, ILV, and its analogues,32 as well as with
our site-directed mutagenesis results. The mutagenesis
data help to quantitate the contributions of the indi-
vidual residues present in the binding site in their
binding to benzolactam. The present studies clearly
show for the first time how teleocidin, ILV, and the
benzolactam bind to their receptor and thus provide us
the structural basis required for the design of novel PKC
ligands of improved potency and enhanced isozyme
selectivity. Through our studies, it is apparent that the
definition of the PKC ligand pharmacophore should be
expanded not only to include those hydrophilic groups
that form specific hydrogen bonds to the protein but also
to encompass those hydrophobic groups that make fairly
specific hydrophobic contacts with the hydrophobic
residues present in PKC. Thus, for first time it is
possible to appreciate that these hydrophobic groups in
ILV, teleocidin, and the 8-membered ring benzolactam
not only contribute to the overall hydrophobicity of the
ligands but also, and perhaps more importantly, make
Isozym e Stu d ies. The two new benzolactams 5 and
6 have been tested for isozyme selectivity by investigat-
ing their ability to displace [3H]PDBu binding to recom-
binant PKC isozymes expressed in the baculovirus
system.4 These data are presented in Table 4 along
with comparison data for ILV and n-octyl-ILV. As is
apparent, the acetylenic derivative 5 shows a rather nice
level of selectivity, with the compound showing a higher
affinity for the R and â isozymes, in comparison to γ, δ,
and ꢀ. The greatest selectivity of 5 is the approximate
10-fold difference in affinity between PKCR and -ꢀ. In
contrast, n-octyl-ILV shows a pairwise selectivity of only
about 2 for R vs ꢀ and at best a 3-fold selectivity for â
vs γ. As is also apparent, the electronic effect of the
acetylenic group of 5 plays some role in its isozyme
selectivity, for its saturated alkyl counterpart 6 shows
poorer selectivity with only a 4-fold difference in the Ki
for R vs ꢀ.
Also provided in Table 4 is the Ki at PKCR for the
parent benzolactam 15. This compound lacks the
hydrophobic tail which is important for membrane
association and which accordingly is coupled to an

Benzolactam Analogues of ILV
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9 1323
antiproliferative activity correlated with down-regula-
tion of PKCâ in both cell lines, although there was not
complete isozyme selectivity. Further studies of the
effects of other substituent modifications on the isozyme
selectivity of the benzolactams will be reported in due
course.
Exp er im en ta l Section
1. Molecu la r -Mod elin g Meth od s. All of the molecular-
modeling studies were carried out using the QUANTA molec-
ular package42 running on a Silicon Graphics Indigo 2/R10000
with IRIS 6.2. All of the energy calculations and molecular
dynamics simulations were performed using a stand-alone
version of the CHARMM program43 (version 24), with the
version 22 MSI parameter set, running on a Deck Alpha
machine with two CPU processors. An adopted-basis New-
ton-Raphson algorithm, implemented in the CHARMM pro-
gram, was used in the energy minimizations. Energy was
minimized for 10 000 steps for the ligand-receptor complex,
or until convergence, defined as an energy gradient e0.001
kcal mol-1 Å-1. A constant dielectric was used and set to 1.
The nonbonded cutoff distance was set to 14.0 Å. A shifted
smoothing function was used for the van der Waals interaction
and a switch function for the electrostatic energy. The cutoff
distance parameters used in the smoothing functions are as
follows: CTOFNB ) 12.0 Å; CTONNB ) 8.0 Å.
The X-ray structure of PKCδ CRD2 in complex with phorbol
13-acetate5 was used as the initial structure in our docking,
energy minimization, and molecular dynamics simulation
studies. Docking studies were done manually using the
QUANTA program. Each complex was then fully minimized.
In the molecular dynamics simulations, the complex, which
includes the protein and the ligand, was solvated using 538
TIP3P44 water molecules. The system was heated to 300 K in
a period of 5 ps and equilibrated for 20 ps at 300 K. Finally,
100 ps or longer constant temperature simulations were
performed at 300 K with a step size of 0.001 ps. Trajectories
were recorded every 0.1 ps during the simulations and
analyzed. A shake algorithm was used to constrain bonds to
hydrogen.45 The three-dimensional structure of the benzolac-
tam was built starting from the X-ray structure of teleocidin46
using the QUANTA program. The structure of the benzolac-
tam was first minimized using the CHARMM program. The
CHARMM minimized structure was then fully reminimized
using the Gaussian 92 program47 with the 3-21g* basis set.
The Mulliken atomic charges for the benzolactam were cal-
culated using the Gaussian 92 program47 with the 6-31g* basis
set and used to build the residue topology file (RTF) for the
ligand.
F igu r e 6. Antiproliferative activity of benzolactam 5 and ILV
against two breast carcinoma cell lines.
enhancement in compound potency. While the parent
compound still binds to PKCR, it is 7-fold less potent
than 6 and about 30-fold less potent than ILV. This
result reveals that a molecule possessing the appropri-
ate twist-like conformation of ILV is capable of retaining
some PKC activity, in spite of the absence of the
hydrophobic tail.
Cell P r olifer a tion Assa y a n d P KC Dow n -Regu la -
tion . The acetylenic benzolactam 5 and ILV were
tested for antiproliferative activity against breast car-
cinoma cell lines MCF-7 and MDA-MB-231 (Figure 6).7
Exposure of MCF-7 and MDA-MB-231 cells to 5 for 4
days resulted in IC50 values of 20 and 30 µM, respec-
tively, whereas ILV was inactive.
PKC isozyme levels were determined in MCF-7 and
MDA-MB-231 cells exposed to 50 µM 5 for 24 h (Figure
7). In MCF-7 cells, PKCâ and PKCꢀ were virtually
eliminated, PKCδ was reduced to a lesser extent, and
PKCú was unchanged. MDA-MB-231 cells exhibited a
similar reduction in PKCâ, whereas PKCδ and -ú were
slightly reduced, while PKCR and PKCꢀ remained
unchanged. These results indicate that 5, while not
completely selective, preferentially down-regulates PKCâ
in both cell lines. The varying degree of selectivity of 5
for other PKC isozymes was to some degree cell type
specific. This result was not completely unexpected,
since different tumors would be expected to exhibit
different PKC isozyme patterns as well as different
pathways governing their stability and turnover.
The hydrophobic analysis was performed using the HINT
program48 running under the Chem-X program49 on a Silicon
Graphics Indigo 2/R10000 with IRIS 6.2. The log P value for
the ligand was calculated using the “calculate” option, and the
log P value for the protein was calculated using the “dictio-
nary” option in the HINT program. The intermolecular
hydrophobic interaction map and table were generated for the
trajectories recorded at the end of each 10 ps simulation period
during the 100 ps CHARMM molecular dynamics simulations.
The HINT values were calculated using the following equation:
Su m m a r y
The present work delineates a simple route to the
8-substituted benzolactam analogues of ILV starting
from L-phenylalanine. The molecular-modeling studies
and site-directed mutagenesis analyses reveal precisely
how these benzolactams bind to the CRD2 activator-
binding domain of PKC and provide a structural basis
for the design of high-affinity PKC ligands of improved
isozyme selectivity. It is noteworthy that the acetylene
bearing benzolactam 5 has a 10-fold higher affinity for
PKCR and -â than for PKCδ and -ꢀ, and such isozyme
selectivity has not been reported previously for the ILV-
based PKC activators. Of further interest is the anti-
proliferative action of 5 on two breast carcinoma cell
lines in comparison to the lack of activity of ILV. The
bij ) siaisjajRij summed for all i, j
where bij is a microinteraction constant representing the
attraction/interaction between atoms i and j, si and sj are the
solvent accessible surface areas for i and j, respectively, ai and
aj are the hydrophobic atom constants for i and j, respectively,
and Rij is the functional distance behavior for the interaction
of i and j and was set as equal to exp(-r). The region definition
was set as molecular extents and the ligand was selected as
the molecule. All other parameters were set as the default
values, as recommended in the HINT program.
2. Site-Dir ected Mu ta gen esis An a lysis. Site-directed
mutagenesis analysis was performed using the same proce-
dures as those described previously.30

1324 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9
Kozikowski et al.
F igu r e 7. Western blots showing PKC down-regulation by the benzolactam 5 in MCF-7 and MDA-231 cells.
3. Syn th esis Meth od s. (S)-O-Acetyl-2-[(eth oxyca r bo-
n yl)a m in o]-3-p h en yl-1-p r op a n ol (8). To a suspension of 7
(25.0 g, 0.165 mol) and Na2CO3 (20.1 g, 0.19 mol) in 80 mL of
water was added EtOCOCl (26.0 mL, 0.27 mol) dropwise at
room temperature. The solution was stirred at room temper-
ature for 4 h and extracted with CH2Cl2 (4 × 250 mL). The
organic layer was dried over Na2SO4 and evaporated to give
an oil which was dissolved in Et3N (100 mL, 0.72 mol) and
Ac2O (60 mL, 0.58 mol). The mixture was stirred overnight
at room temperature and partitioned between 400 mL of
EtOAc and 150 mL of water. The organic layer was washed
with 50 mL of brine and dried over Na2SO4. Evaporation and
chromatography (1/2 ethyl acetate/petroleum ether as eluent)
CN was added dropwise at 0 °C, and the mixture was stirred
at room temperature overnight. Filtration from the precipi-
tate, evaporation, and chromatography on silica gel (1/1 CH2-
Cl2/EtOAc as eluent) provided 1.79 g of 14. Without further
purification, this product was directly dissolved in 10 mL of
dried CH2Cl2, and the solution was cooled to 0 °C before 10
mL of CF3COOH was added. The mixture was stirred at 0 °C
for 2 h, and then the volatiles were removed in vacuo below
30 °C. The residue was dissolved in 100 mL of EtOAc, and 5
mL of saturated aqueous NaHCO3 solution was added. The
mixture was heated at 80 °C for 24 h with vigorous stirring.
After the mixture was cooled to room temperature, 40 mL of
water was added. The organic layer was separated, and the
aqueous layer was extracted with EtOAc (4 × 20 mL). Drying
over Na2SO4 and evaporation produced a yellow oil, which was
dissolved in 10 mL of CH3CN. This solution was cooled to 0
°C, and 4.0 mL (40 mmol) of formalin, 1.0 g (16 mmol) of
NaBH3CN, and 0.27 mL of AcOH were added. The resulting
mixture was stirred at 0 °C for 2 h before being quenched with
phosphate buffer (pH ) 2). The solvent was evaporated, and
the residue was dissolved in EtOAc, washed with water,
saturated NaHCO3, and brine, and dried over Na2SO4. Evapo-
ration and chromatography on silica gel (1/10 MeOH/CH2Cl2
as eluent) gave 15 (480 mg, 44% overall from 12): [R]20D -271°
afforded 40.2 g (91%) of 8: [R]20 -17.9° (c 0.9, CHCl3); IR
D
(KBr) 2980, 1700, 1500, 1200 cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.31-7.15 (m, 4H), 4.72 (br s, 1H), 4.09 (q, J ) 7.3 Hz, 2H),
4.07-4.02 (m, 3H), 2.80 (m, 2H), 2.04 (s, 3H), 1.18 (t, J ) 7.3
Hz, 3H); MS m/z 265 (M+), 165, 91. Anal. (C14H19NO4) C, H,
N.
(S)-O-Acet yl-3-(2-a m in op h en yl)-2-[(et h oxyca r b on yl)-
a m in o]-1-p r op a n ol (9). To a solution of 8 (25.0 g, 0.094 mol)
in Ac2O (25 mL) was added nitric acid (10 mL, 0.24 mol)
dropwise at 0 °C. The mixture was stirred at room temper-
ature for 24 h, poured into 200 mL of ice-water, and extracted
with EtOAc (3 × 250 mL). The organic layers were washed
with brine, dried over Na2SO4, and evaporated. The residue
was dissolved in 200 mL of EtOAc, and 500 mg of Pd/C was
added. The mixture was stirred under 1 atm of H2 at room
temperature for 24 h. Filtration from the catalyst, evapora-
tion, and chromatography (2/3 ethyl acetate/petroleum ether
as eluent) provided 9 (6.02 g, 23%) and its isomer 10 (13.2 g,
50%). Compound 9: [R]20D +4.6° (c 0.5, CHCl3); IR (KBr) 3360,
1720, 1700, 1500, 1200, 730 cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.06 (t, J ) 7.5 Hz, 1H), 6.93 (d, J ) 7.6 Hz, 1H), 6.67 (m,
2H), 5.26 (br s, 1H), 4.14 (q, J ) 7.3 Hz, 2H), 4.10-4.05 (m,
3H), 4.01 (br s, 2H), 2.94 and 2.61 (AB q, 2 H, J ) 13.6 Hz),
2.10 (s, 3H), 1.25 (t, J ) 7.3 Hz, 3H); MS m/z 280 (M+), 191,
132, 106. Anal. (C14H20N2O4) C, H, N.
(2S ,2′S )-N -[3′-Ace t oxy-2′-[(e t h oxyca r b on yl)a m in o]-
p h en yl]va lin e Ben zyl Ester (12). A mixture of (R)-benzyl
R-[[(trifluoromethyl)sulfonyl]oxy]isovalerate (11, 5.03 g, 18.1
mmol), 9 (5.02 g, 18.1 mmol), and 2,6-lutidine (2.4 mL, 20
mmol) in 40 mL of 1,2-dichloroethane was stirred at 70 °C for
10 h. Evaporation and chromatography on silica gel (1/5 ethyl
acetate/petroleum as eluent) gave compound 12 (6.05 g, 70%)
as a colorless oil: [R]20D -20.6° (c 0.07, CHCl3); IR (KBr) 3300,
1725, 1710, 1520 cm-1; 1H NMR (300 MHz, CDCl3) δ 7.29 (m,
5H), 7.05 (t, J ) 7.5 Hz, 1H), 6.95 (d, J ) 7.6 Hz, 1H), 6.67 (t,
J ) 7.5 Hz, 1H), 6.68 (d, J ) 7.6 Hz, 1H), 5.14 (m, 3H), 4.12
(m, 5H), 3.91 (d, J ) 6.2 Hz, 1H), 2.98 (m, 1H), 2.67 (m, 1H),
2.23 (m, 1H), 2.02 (s, 3H), 1.16 (t, J ) 7.2 Hz, 3H), 1.03 (d, J
) 7.0 Hz, 3H), 0.84 (d, J ) 7.0 Hz, 3H); MS m/z 471 (M + H+),
335, 275, 186, 91; HRMS calcd for C26H34N2O6 470.241, found
470.241.
1
(c 0.08, CHCl3); H NMR (300 MHz, CDCl3) δ 6.80-7.20 (m,
4H), 6.72 (br s, 1H), 4.10 (m, 1H), 3.73 (m, 1H), 3.61 (m, 1H),
3.46 (d, J ) 8.1 Hz, 1H), 3.11 (dd, J ) 15.8, 8.1 Hz, 1H), 2.85
(dd, J ) 15.8, 3.2 Hz, 1H), 2.84 (s, 3H), 2.48 (m, 1H), 1.15 (d,
J ) 6.4 Hz, 3H), 0.94 (d, J ) 6.6 Hz, 3H); MS m/z 262 (M+),
245, 117, 91.
(2S,5S)-8-Iod oben zola cta m V (16). A mixture of 15 (850
mg, 3.2 mmol), Et3N (15 mL), and Ac2O (5 mL) was stirred at
room temperature for 24 h. The mixture was poured into 100
mL of water, extracted with EtOAc (4 × 100 mL), washed with
brine, and dried over Na2SO4. After evaporation, the residue
was dissolved in 10 mL of 1,4-dioxane and 2 mL of pyridine.
To this solution was added 1.70 g (6.7 mmol) of I2, and the
deep brown solution was stirred at room temperature for 3 d.
The mixture was partitioned between EtOAc and water, and
the organic layer was washed with 10 mL of 10% aqueous
NaHSO3 and dried over Na2SO4. Chromatography on silica
gel provided 598 mg of 16 together with 338 mg of unreacted
starting material (83% of 16 based on conversion): [R]20D -279°
(c 0.21, CHCl3); IR (KBr) 3200, 1740, 1680, 1260, 1240, 1065
1
cm-1; H NMR (300 MHz, CDCl3) δ 7.18-7.00 (m, 3H), 6.05
(s, 1H), 4.54 (m, 1H), 4.13 and 3.98 (AB q, 2 H, J ) 11.1 Hz,
both parts d with J ) 8.2 Hz), 3.45 (d, J ) 8.2 Hz, 1H), 3.01-
2.87 (m, 2H), 2.76 (s, 3H), 2.46 (m, 1H), 2.10 (s, 3H), 1.08 (d,
J ) 7.2 Hz, 3H), 0.92 (d, J ) 7.2 Hz, 3H); MS m/z 430 (M+),
304, 261, 233, 158, 132; HRMS calcd for C17H23N2O3I 430.076,
found 430.075.
(2S,5S)-8-(1-Decyn yl)ben zola cta m V (5). To a mixture
of 16 (116 mg, 0.27 mmol), Et2NH (2 mL), and PdCl2(PPh3)2
(26 mg, 0.018 mmol) were added 1-decyne (0.13 mL, 0.74
mmol) and CuI (4 mg, 0.01 mmol). The resulting solution was
stirred at room temperature for 24 h. The solvent was
evaporated, the residue was dissolved in 1 mL of 20% aqueous
NaOH and 4 mL of MeOH, and the solution was stirred at
room temperature for 0.5 h. The mixture was partitioned
between 10 mL of water and 100 mL of EtOAc. The organic
layer was separated, washed with brine, and dried over
Na2SO4. Evaporation and chromatography on silica gel (2/1
ethyl acetate/petroleum ether as eluent) afforded 87 mg (81%)
of 5: [R]20D -303.3° (c 0.5, ethanol); 1H NMR (300 MHz, CDCl3)
δ 7.18 (d, J ) 7.8 Hz, 1H), 7.09 (s, 1H), 6.85 (d, J ) 7.8 Hz,
(2S,5S)-Ben zola cta m V (15). A mixture of 12 (6.0 g, 12.8
mmol) and KOH (5.0 g, 89 mmol) in 40 mL of MeOH/H2O (1/
1) was stirred at 30 °C for 3 d. After neutralization to pH ∼7
with concentrated HCl, di-tert-butyl dicarbonate (2.7 g, 12.5
mmol) and NaHCO3 (1.5 g, 18 mmol) were added. The mixture
was stirred at room temperature for 24 h, washed with
petroleum ether, and adjusted to pH ∼2. Extraction with
EtOAc (4 × 150 mL), drying over Na2SO4, and evaporation
gave 4.1 g of crude 13. This product (1.61 g, 4.37 mmol) and
N-hydroxysuccinimide (0.52 g, 4.41 mmol) were dissolved in
20 mL of CH3CN, DCC (1.21 g, 5.81 mmol) in 20 mL of CH3-

Benzolactam Analogues of ILV
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9 1325
(11) Sakai, S.; Hitotsuyanagi, Y.; Aimi, N.; Fujiki, H.; Suganuma,
M.; Sugimura, T.; Endo, Y.; Shudo, K. Absolute configuration of
lyngbyatoxin A (teleocidin A-1) and teleocidin A-2. Tetrahedron
Lett. 1986, 27, 5219-5220.
1H), 6.55 (s, 1H), 3.82 (m, 1H), 3.58 (m, 1H), 3.48 (m, 2H),
3.12 (dd, J ) 16.3, 8.1 Hz, 1H), 2.80 (s, 3H), 2.73 (d, J ) 16.2
Hz, 1H), 2.33 (m, 1H), 2.30 (t, J ) 7.8 Hz, 2H), 1.68-1.15 (m,
12H), 1.05 (d, J ) 7.2 Hz, 3H), 0.88 (t, J ) 7.4 Hz, 3H), 0.85
(d, J ) 7.2 Hz, 3H); MS m/z 398 (M+), 312, 91; HRMS calcd
for C25H38N2O2 398.293, found 398.294. Anal. (C25H38N2O2)
C, H, N.
(12) Sugimura, T.; Fujiki, H.; Mori, M.; Nakayasu, M.; Terada, M.;
Umezawa, K.; Moore, R. E. Teleocidin: new naturally occurring
tumor promoter. Carcinogenesis 1982, 7, 69-73.
(13) Moore, R. E. Toxins, anticancer agents, and tumor promoters
from marine prokaryotes. Pure Appl. Chem. 1982, 54, 1919-
1934; Irie, K.; Koshimizu, K. Chemistry of indole alkaloid tumor
promoter teleocidins. Comments Agric. Food Chem. 1993, 3,
1-25. Irie, K.; Koshimizu, K. Structure-activity studies of indole
alkaloid tumor promoters. Memoirs of the College of Agriculture,
Kyoto University, No. 132, 1988.
(14) Endo, Y.; Hasegawa, M.; Itai, A.; Shudo, K.; Tori, M.; Asakawa,
Y.; Sakai, S. Tumor promoters in two conformational states in
solution. Stereochemistry of (()-Indolactam-V. Tetrahedron
Lett. 1985, 26, 1069-1072.
(2S,5S)-8-Decylben zola cta m V (6). A mixture of 5 (20
mg, 0.05 mmol), Pd/C (10 mg), and EtOAc (5 mL) was
hydrogenated under 10 atm of H2 at room temperature for 2
h. Filtration from the catalyst, evaporation, and chromatog-
raphy gave 18 mg (90%) of 6: [R]20 -247.5° (c 1.0, ethanol);
D
1H NMR (300 MHz, CDCl3) δ 6.93 (m, 2H), 6.87 (s, 1H), 6.45
(s, 1H), 4.14 (m, 1H), 3.78 and 3.51 (AB q, 2H, J ) 15.8 Hz,
both parts d, J ) 12.4 and 8.6 Hz, respectively), 3.39 (d, J )
8.3 Hz, 1H), 2.96 (dd, J ) 16.2, 10.6 Hz, 1H), 2.83 (d, J ) 16.2
Hz, 1H), 2.70 (s, 3H), 2.36 (t, J ) 7.6 Hz, 2H), 2.30 (m, 1H),
1.60-1.06 (m, 16H), 1.01 (d, J ) 7.6 Hz, 3H), 0.82 (d, J ) 7.6
Hz, 3H), 0.79 (t, J ) 7.4 Hz, 3H); MS m/z 402 (M+), 359, 331,
316; HRMS calcd for C25H42N2O2 402.324, found 402.325.
Anal. (C25H42N2O2) C, H, N.
(15) Endo, Y.; Ohno, M.; Hirano, M.; Itai, A.; Shudo, K. Synthesis,
conformation, and biological activity of teleocidin mimics, ben-
zolactams. Problem in structure-activity studies of teleocidins.
J . Am. Chem. Soc. 1996, 118, 1841-1855.
(16) Itai, A.; Kato, Y.; Tomioka, N.; Iitaka, Y.; Endo, Y.; Hasegawa,
M.; Shudo, K.; Fujiki, H.; Sakai, S. A receptor model for tumor
promoters: rational superposition of teleocidins and phorbol
esters. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 3688-3692.
Ack n ow led gm en t. We are indebted to the National
Natural Science Foundation of China, the Chinese
Academy of Sciences, and Alexis Biochemicals for sup-
port of this work.
(17) Rando, R.; Kishi, Y. Structural basis of protein kinase
C
activation by diacylglycerols and tumor promoters. Biochemistry
1992, 31, 2211-2218.
(18) Thomson, C.; Wilkie, J . The conformations and electrostatic
potential maps of phorbol esters, teleocidins and ingenols.
Carcinogenesis 1989, 10, 531-540.
Refer en ces
(19) Wender, P. A.; Cribbs, C. M.; Koehler, K. F.; Sharkey, N. A.;
Herald, C. L.; Kamano, Y.; Pettit, G. R.; Blumberg, P. M. Proc.
Nat. Acad. Sci. U.S.A. 1988, 86, 7197-7201.
(1) Nishizuka, Y. The molecular heterogeneity of protein kinase C
and its implications for cellular recognition. Nature 1988, 334,
(20) Blumberg, P. M.; Lewin, N. E.; Wang, S. Unpublished work.
(21) Kozikowski, A. P.; Ma, D.; Pang, Y.-P.; Shum, P.; Likic, V.;
Mishra, P. K.; Macura, S.; Basu, A.; Lazo, J . S.; Ball, R. G.
Synthesis, molecular modeling, 2-D-NMR, and biological evalu-
ation of ILV mimics as potential modulators of protein kinase
C. J . Am. Chem. Soc. 1993, 115, 3957-3965.
661-65. Dekker, L. V.; Parker, P. J . Protein kinase C:
a
question of specificity. Trends Biochem. Sci. 1994, 19, 73-77.
Protein Kinase C, Current Concepts and Future Perspectives;
Lester, D. S., Epand, R. M., Eds.; Ellis Horwood Ltd.: New York,
1992. Basu, A. The potential of protein kinase C for anticancer
treatment. Pharmacol. Ther. 1993, 59, 257-280. Glazer, R. I.
Protein kinase C in neoplastic transformation and differentia-
tion. In Protein Kinase C; Kuo, J . F., Ed.; Oxford Press: Oxford,
England, 1994; pp 171-198.
(22) Irie, K.; Hagiwara, N.; Tokuda, H.; Koshimizu, K. Structure-
activity studies of the indole alkaloid tumor promoter teleocidins.
Carcinogenesis 1987, 8, 547-552 and 963-965.
(23) Irie, I.; Okuno, S.; Kajiyama, S.; Koshimizu, K.; Nishino, H.;
Iwashima, A. Quantitative structure-activity studies on indole
alkaloid tumor promoter indolactam congeners. Carcinogenesis
1991, 12, 1883-1886.
(24) Kozikowski, A. P.; Ma, D.; Du, L.; Lewin, N. E.; Blumberg, P.
M. Effect of alteration of the heterocyclic nucleus of ILV on its
isozyme selectivity for PKC. Palladium-catalyzed route to the
benzofuran analogues of ILV. J . Am. Chem. Soc. 1995, 117,
6666-6672. Kozikowski, A. P.; Ma, D.; Du, L.; Lewin, N. E.;
Blumberg, P. M. Synthesis of the benzofuran analogue of ILV,
a new protein kinase C (PKC) activator. Bioorg. Med. Chem.
Lett. 1994, 4, 637-640.
(25) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J .;
Vistica, D.; Warren, J . T.; Bokesch, H.; Kenney, S.; Boyd, M. R.
New colorimetric cytotoxicity assay for anticancer-drug screen-
ing. J . Natl. Cancer Inst. 1990, 82, 1107-1112. Ahmad, S;
Glazer, R. I. Expression of the antisense cDNA to protein kinase
CR attenuates resistance in doxorubicin-resistant MCF-7 breast
carcinoma cells. Mol. Pharmacol. 1993, 43, 858-862.
(26) J effrey, A. M.; Liskamp, R. M. Computer-assisted molecular
modelling of tumour promoters: rationale for the activity of
phorbol esters, teleocidin B, and aplysiatoxin. Proc. Natl. Acad.
Sci. U.S.A. 1986, 83, 241-245.
(27) Wender, P. A.; Koehler, K. F.; Sharkey, N. A.; Dell’Aquila, M.
L.; Blumberg, P. M. Analysis of the phorbol ester pharmacophore
on protein kinase C as a guide to the rational design of new
classes of analogs. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 4214-
4218.
(2) Kozikowski, A. P.; Sato, K.; Basu, A.; Lazo, J . S. Synthesis and
biological studies of simplified analogues of lyngbyatoxin A: use
of an isoxazoline-based indole synthesis. Quest for protein
kinase C modulators. J . Am. Chem. Soc. 1989, 111, 6228-6234.
Kozikowski, A. P.; Shum, P. W.; Basu, A.; Lazo, J . S. Synthesis
of structural analogues of lyngbyatoxin A and their evaluation
as activators of protein kinase C. J . Med. Chem. 1991, 34, 2420-
2430.
(3) Stabel, S.; Parker, P. J . Protein kinase C. Pharmacol. Ther. 1991,
51, 71-95. Cacace, A. M.; Guadagno, S. N.; Krauss, R. S.;
Fabbro, D.; Weinstein, I. B. The epsilon isozyme of protein kinase
C is an oncogene when overexpressed in rat fibroblasts. Onco-
gene 1993, 8, 2095-2104. Powis, G.; Kozikowski, A. P. Growth
factor and oncogene signaling pathways as targets for rational
anticancer drug development. Clin. Biochem. 1991, 24, 385-
397.
(4) Kazanietz, M. G.; Areces, L. B.; Bahador, A.; Mischak, H.;
Goodnight, J .; Mushinski, F.; Blumberg, P. M. Characterization
of ligand and substrate specificity for the calcium-dependent and
calcium-independent PKC isozymes. Mol. Pharmacol. 1993, 44,
298-307.
(5) Zhang, G.; Kazanietz, M. G.; Blumberg, P. M.; Hurley, J . H.
Crystal structure of the Cys2 activator-binding domain of protein
kinase Cδ in complex with phorbol ester. Cell. 1995, 81, 917-
924.
(6) Merrill, A. H. J r.; Sereni, A. M.; Stevens, V. L.; Hannun, Y. A.;
Bell, R. M.; Kinkade, J . M. J r. Inhibition of phorbol ester-
dependent differentiation of human promyelocytic leukemic (HL-
60) cells by sphinganine and other long-chain bases. J . Biol.
Chem. 1986, 261, 12610-12615.
(7) Yu, G.; Ahmad, S.; Aquine, A.; Fairchild, C. R.; Trepel, J . B.;
Ohno, S.; Suzuki, K.; Tsuruo, T.; Cowan, K. H.; Glazer, R. I.
Transfection with protein kinase C R confers increased multi-
drug resistance to MCF-7 cells expressing P-glycoprotein. Can-
cer Commun. 1991, 3, 181-189.
(28) Nakamura, H.; Kishi, Y.; Pajares, M. A.; Rando, R. R. Structural
basis of protein kinase C activation by tumour promoters. Proc.
Natl. Acad. Sci. U.S.A. 1989, 86, 9672-9676.
(29) Leli, U.; Hauser, G.; Froimowitz, M. Requirements for the
activation of protein kinase C: comparison of the molecular
geometries of phorbol and diacylglycerols. Mol. Pharmacol.
1990, 37, 286-295.
(30) Kazanietz, M. G.; Wang, W.; Milne, G. W. A.; Lewin, N. E.; Liu,
H. L.; Blumberg, P. M. Residues in the second cysteine-rich
region of PKC relevant to phorbol ester binding as revealed by
site-directed mutagenesis. J . Biol. Chem. 1995, 270, 21852-
21859.
(31) Blumberg, P. M.; Lewin, N. E.; et al. Manuscript in preparation.
(32) Wang, S.; et al. Manuscript in preparation.
(33) Wang, S.; Milne, G. W. A.; Nicklaus, M. C.; Marquez, V. E.; Lee,
J .; Blumberg, P. M. Protein kinase C. Modeling of the binding
site and prediction of binding constants. J . Med. Chem. 1994,
37, 1326-1338.
(8) Basu, A.; Kozikowski, A. P.; Sato, K.; Lazo, J . S. Cellular
sensitization to cis-diamminedichloroplatinum(II) by novel ana-
logues of the protein kinase C activator lyngbyatoxin A. Cancer
Res. 1991, 51, 2511-2514.
(9) Takashima, M.; Sakai, H. A new toxic substance, teleocidin
produced by Streptomyces. Part I. Production, isolation, and
chemical studies. Bull. Agric. Chem. Soc. J pn. 1960, 24, 647-
651.
(10) Cardellina, J . H., II; Marner, F.-J .; Moore, R. E. Seaweed
dermatitis: structure of lyngbyatoxin A. Science 1979, 204,
1193-1195.

1326 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 9
Kozikowski et al.
tions. J . Comput. Chem. 1983, 4, 187-217.
(34) Blumberg, P. M.; et. al. Unpublished results.
(35) Irie, I.; Okuno, S.; Kajiyama, S.; Koshimizu, K.; Nishino, H.;
Iwashima, A. Quantitative structure-activity studies on indole
alkaloid tumor promoter indolactam congeners. Carcinogenesis
1991, 12, 1883-1886.
(44) J orgensen, W. L.; Chandrasekhar, J .; Madura, J . D.; Impey, R.
W.; Klein, M. L. Comparison of simple potential functions for
simulating liquid water. J . Chem. Phys. 1983, 79, 926-935.
(45) van Gunsteren, W. F.; Berendsen, H. J . C. Algorithms for
macromolecular dynamics and constraint dynamics. Mol. Phys.
1977, 34, 1311-1327.
(46) Harada, H.; Sakabe, N.; Hirata, Y.; Tomiie, Y.; Nitta, I. The X-ray
structure determination of dihydroteleocidin B monobromoac-
etate. Bull. Chem. Soc. J pn. 1966, 39, 1773-1775. The X-ray
coordinates can be found in the Cambridge Structural Database
under the code name TLOBDH10.
(47) Gaussian 92/DFT, Revision G.4: Frisch, M. J .; Truck, G. W.;
Schlegel, H. B.; Robb, M. A.; Head-Gordon, M.; Replogle, E. S.;
Gomperts, C.; Martin, R. L.; Fox, D. J .; Defrees, D. J .; Baker,
J .; Stewart, J . J . P.; Pople, J . A., Gaussian, Inc., Pittsburgh, PA,
1993.
(48) The HINT program is supplied by eduSoft, LC, P.O. Box 1811,
Ashland, VA 23005. Kellogg, G. E.; J oshi, G. S.; Abraham, D.
J . New tools for modeling and understanding hydrophobicity and
hydrophobic interactions. Med. Chem. Res. 1992, 1, 444-453.
Meng, E. C.; Kuntz, I. D.; Abraham, D. J .; Kellogg, G. E.
Evaluating docked complexes with the HINT exponential func-
tion and empirical atomic hydrophobicities. J . Compt.-Aided
Mol. Des. 1994, 8, 299-306.
(36) McKennon, M. J .; Meyers, A. I.; Drauz, K.; Schwarm, M. A
convenient reduction of amino acids and their derivatives. J .
Org. Chem. 1993, 58, 3568-3571.
(37) Kogan, T. P.; Somers, T. C.; Venuti, M. C. A regio- and
stereocontrolled total synthesis of (-)-indolactam-V. Tetrahe-
dron 1990, 46, 6623-6632.
(38) Endo, Y.; Shudo, K.; Itai, A.; Hasegawa, M.; Sakai, S. Synthesis
and stereochemistry of indolactam-V, an active fragment of
teleocidins. Structural requirements for tumor-promoting activ-
ity. Tetrahedron 1986, 42, 5905-5924.
(39) Brewster, R. Q. p-Iodoaniline. Organic Syntheses; Wiley: New
York, 1943; Collect. Vol. 2, pp 347-348.
(40) Kikukawa, K.; Abe, A.; Wada, F.; Matsuda, T. Palladium-
catalyzed reaction of 4′-iodobenzocrown ethers with acetylenes.
Convenient synthesis of alkyl-substituted benzocrown ethers and
bis(benzocrown ethers) via alkynyl-substituted benzocrown ethers.
Bull. Chem. Soc. J pn. 1983, 57, 961-962.
(41) Endo, Y.; Imada, T.; Yamaguchi, K.; Shudo, K. Conformational
states of indolactams: Structures of 13-N-Desmethylindolactam-
V and 13-O-indolactam-V. Heterocycles 1994, 39, 571-579.
(42) QUANTA, a molecular modeling system, is supplied by Molec-
ular Simulations Inc., 200 Fifth Ave., Waltham, MA 01803-
5279.
(43) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J .;
Swaminathan, S.; Karplus, M. CHARMM: A program for
macromolecular energy, minimization, and dynamics calcula
(49) Chem-X is a product of Chemical Design Ltd., Roundway House,
Cromwell Park, Chipping Norton, Oxon OX7 5SR, U.K.
J M960875H