Synthesis, biological evaluation, structural-activity relationship, and docking study for a series of benzoxepin-derived estrogen receptor modulators

Article abstract of DOI:10.1016/j.bmc.2008.09.035

The estrogen receptors ERα and ERβ are recognized as important pharmaceutical targets for a variety of diseases including osteoporosis and breast cancer. A series of novel benzoxepin-derived compounds are described as potent selective modulators of the human estrogen receptor modulators (SERMs). We report the antiproliferative effects of these compounds on human MCF-7 breast tumor cells. These heterocyclic compounds contain the triarylethylene arrangement as exemplified by tamoxifen, conformationally restrained through the incorporation of the benzoxepin ring system. The compounds demonstrate potency at nanomolar concentrations in antiproliferative assays against an MCF-7 human breast cancer cell line with low cytotoxicity together with low nanomolar binding affinity for the estrogen receptor. The compounds also demonstrate potent antiestrogenic properties in the human uterine Ishikawa cell line. The effect of a number of functional group substitutions on the ER binding properties of the benzoxepin molecular scaffold is examined through a detailed docking and 2D-QSAR computational investigation. The best QSAR model developed for ERαβ selectivity yielded R² of 0.84 with an RMSE for the training set of 0.30. The predictive quality of the model was Q² of 0.72 and RMSE of 0.18 for the test set. One particular compound (26b) bearing a 4-fluoro substituent, exhibits 15-fold selectivity for ERβ and both our docking and QSAR studies converge on the correlation between enhanced lipophilicity and enhanced ERβ binding for this benzoxepin ring scaffold.

Full text of DOI:10.1016/j.bmc.2008.09.035

Bioorganic & Medicinal Chemistry 16 (2008) 9554–9573
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis, biological evaluation, structural–activity relationship,
and docking study for a series of benzoxepin-derived estrogen receptor
modulators
a
a
b
Irene Barrett a, Mary J. Meegan a, , Rosario B. Hughes , Miriam Carr , Andrew J. S. Knox ,
*
Natalia Artemenko b, Georgia Golfis b, Daniela M. Zisterer c, David G. Lloyd b
a School of Pharmacy and Pharmaceutical Sciences, Centre for Synthesis and Chemical Biology, Trinity College Dublin, Dublin 2, Ireland
b Molecular Design Group, School of Biochemistry & Immunology, Trinity College Dublin, Dublin 2, Ireland
c School of Biochemistry & Immunology, Trinity College Dublin, Dublin 2, Ireland
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 May 2008
Revised 2 September 2008
Accepted 10 September 2008
Available online 16 September 2008
The estrogen receptors ERa and ERb are recognized as important pharmaceutical targets for a variety of
diseases including osteoporosis and breast cancer. A series of novel benzoxepin-derived compounds are
described as potent selective modulators of the human estrogen receptor modulators (SERMs). We report
the antiproliferative effects of these compounds on human MCF-7 breast tumor cells. These heterocyclic
compounds contain the triarylethylene arrangement as exemplified by tamoxifen, conformationally
restrained through the incorporation of the benzoxepin ring system. The compounds demonstrate
potency at nanomolar concentrations in antiproliferative assays against an MCF-7 human breast cancer
cell line with low cytotoxicity together with low nanomolar binding affinity for the estrogen receptor.
The compounds also demonstrate potent antiestrogenic properties in the human uterine Ishikawa cell
line. The effect of a number of functional group substitutions on the ER binding properties of the benzoxe-
pin molecular scaffold is examined through a detailed docking and 2D-QSAR computational investigation.
Keywords:
Benzoxepin
Estrogen receptor antagonist
Antiproliferative activity
Docking
QSAR
The best QSAR model developed for ERa
b selectivity yielded R2 of 0.84 with an RMSE for the training set
of 0.30. The predictive quality of the model was Q2 of 0.72 and RMSE of 0.18 for the test set. One particular
compound (26b) bearing a 4-fluoro substituent, exhibits 15-fold selectivity for ERb and both our docking
and QSAR studies converge on the correlation between enhanced lipophilicity and enhanced ERb binding
for this benzoxepin ring scaffold.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
ever, the amino acid sequence conservation in the ligand binding
domain is only 59%. The LBD volume for ERb is smaller than for
The estrogen receptor (comprising of two subtypes ER
a
and
ER
a
and differences are found in the amino acids of the LBD includ-
with Ile and Met in
ERb) is a ligand inducible nuclear receptor which play a critical
physiological role as mediator of the actions of the estrogen hor-
mones. The estrogen receptors are widely distributed in the body
tissues and are regarded as attractive therapeutic agents for dis-
ing replacement of Met421 and Leu384 in ER
a
ERb. Design of ligands which can modulate the activity of the ER in
a tissue selective manner continues to attract interest due to the
regulating effects of the ER on many diseased states. Tamoxifen
(1a) is a well-established antagonist for the estrogen receptor
and has been a useful endocrine drug in the treatment of ER posi-
tive breast cancer.
X-ray crystallographic studies have determined the binding
modes of several ligands (e.g., estradiol, 4-hydroxytamoxifen (1b)
and raloxifene (2)) in the ligand binding domain of the ER
eases such as osteoporosis and breast cancer.1–3 ER
a is predomi-
nantly found in the uterus, bone, cardiovascular tissue, and liver
and is the predominant ER expressed in breast cancer. ERb is
expressed in many tissues including prostate, breast, vascular
endothelium, and ovary. The precise function of ERb and its role
in breast pathobiology are not clear. However, recent studies indi-
cate that ERb expression may have a potential protective effect on
a and
ERb.5–9 It is evident that the binding of estrogen agonists, pure
antiestrogens and SERMs (selective estrogen receptor modulators)
to the ER-LBD results in induction of key conformational changes
to Helix 12. Many alternative novel scaffolds for estrogen receptor
modulators have been discovered through ‘scaffold hopping’ proto-
cols which have facilitated the design of several novel agonist and
normal cells against ER
a
induced hyperproliferation.4 The DNA
binding domains of the two ER subtypes is well conserved, how-
* Corresponding author. Tel.: +353 1 8962798; fax: +353 1 8962793.
E-mail address: mmeegan@tcd.ie (M.J. Meegan).
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.09.035

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9555
antagonist-type ligands for the ER.10 Examples of ER modulator
scaffolds based on oxygen containing core heterocyclic systems
have been reported, for example, benzopyrans such as EM-652
(3) which is the active metabolite of EM-800,11 7-oxabicy-
clo[2.2.1]heptene12 (4), chromans,13 bisbenzopyrans,14 dihydro-
benzoxathin (5),15 bicyclo[3.3.1]nonene,16 and oxachrysenol17
while mono18 and bis-benzo[b]oxepines19 have been designed as
agonists of the estrogen receptor. The carbocyclic benzocyclohep-
tene (6) has been investigated as a nonisomerisable analogue of
Z-4-hydroxytamoxifen20 while benxocycloheptnenes have been
reported as tissue selective estrogens.21 The structure of estradiol
together with selected SERMs (2–6) is illustrated in Figure 1. We
have identified a novel estrogen receptor modulator core contain-
ing a central benzoxepin scaffold which demonstrated some poten-
tial as antiproliferative and antagonist ER binding ligands18 and
now report the further investigation of this novel core scaffold
structure as tissue and subtype selective estrogen receptor modu-
lators. We now report the syntheses of 22 benzoxepin derivatives
which are substituted at C-4, C-7, and C-8 positions together with
an evaluation of their antiproliferative activity and the relative
the Rings B and C, and also in the basic side chain component of
the Ring A substituent) are considered to be ring fused analogues
of (Z)-tamoxifen. We were interested in examining the effect on
activity of the inclusion of a fluorine substituent at C-8 which
would increase lipophilicity and also block an expected metabolic
oxidation of these compounds, thus ensuring a longer plasma half
life. Fluorotamoxifen has been investigated as a possible estrogen
receptor imaging agent for positron emission tomography in breast
cancer.22 The inclusion of fluorine has been recently reported to
contribute to the SERM activity in the oxachrysenol series of ER li-
gands.23 Type II compound 32 contains the basic side chain substi-
tuent located on the alternative Ring C site and represents a ring
fused analogue of (E)-tamoxifen. Compounds of Type III, 36a–e,
contain a novel heterocyclic or naphthyl Ring C structure while
in Type IV compounds 41a–e, Ring C is included as a benzylic type
substituent.
To rationalize the observed ER binding selectivity of the ben-
zoxepin products, a detailed 2D-QSAR study was undertaken
involving calculation of MOE 2006.08 2D descriptors24 and also a
novel set of descriptors developed within our group known as Qua-
si-Fragmental descriptors. Finally, a thorough flexible docking
study of the benzoxepin ligands was carried out to facilitate both
visualization and corroboration of the biochemical biochemical
results.
binding affinities for ER
to probe the size, shape, and flexibility of the ER
a
and ERb. These novel ligands can be used
and ERb ligand
a
binding domains and could be potentially developed as pharma-
ceutical agents.
The compounds chosen for synthesis are arranged in four differ-
ent structural classes as we wished to examine the response of the
ER ligand binding domain to specific structural variations. Type I
(compounds 15a–d, 26a–e, and 28) representing the novel ben-
zoxepin scaffold (with variations in the aromatic substituents in
2. Chemistry
The synthesis of the Type I series of benzoxepin analogues con-
taining the p-phenolic substituent in Ring C is illustrated in
N
O
OH
N
O
O
C
D
A
B
OH
HO
S
HO
Estradiol
R
2
Raloxifene
1a
1b
R = H, Tamoxifen
R = OH, 4-Hydroxytamoxifen
O
N
HO
O
OH
O
SO3Ph
HO
HO
3
EM 652
4
HO
O
N
O
S
OH
HO
O
N
5
6
Figure 1. Structure of estradiol and SERMs.

9556
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
Scheme 1. The 2,3,4,5-tetrahydro-1-benzoxepin-2-ones 9a–b were
obtained by acid promoted cyclization of the 4-phenoxy butyric
acids 8a–b (obtained by alkylation of the corresponding phenols
7a–b with ethyl bromobutyrate). Many reagents, for example,
polyphosphoric acid, methane sulfonic acid, trifluoro methane sul-
fonic acid, and the use of hafnium triflate as catalyst25 were ex-
plored to optimize this cyclization reaction: PPA was found to be
the most efficient reagent for the preparation of the unsubstitued
benzoxepin 9a, R = H, while Eatons reagent was found to be supe-
rior for 9b, R = OMe with yields of up to 30% achieved. 9a,b were
converted to the triflates 10a,b by treatment with triflic anhydride
and then subsequently coupled with 4-hydroxyphenylboronic acid
in a Suzuki reaction promoted by Pd(PPh3)4 to afford the 5-aryl-
benzoxepins 11a,b in good yield (67–88%). The vinyl bromides
10c,d was also obtained by reaction of 9a,b with phosphorus tri-
bromide; subsequent arylation of the bromides afforded the prod-
ucts 11a,b, respectively. However, the bromides 10c,d proved to
be unstable and the triflates 10a,b were preferred for subsequent
arylation reactions. The basic pyrrolidine and piperidine ethers
were then readily introduced under standard alkylating conditions
to afford products 12a–c which were converted to the vinylic bro-
mides 13a–c on treatment with pyridine hydrobromide perbro-
mide. This reaction was completed rapidly and required careful
monitoring to prevent the formation of the 4,7-dibromide product
13d. An alternative route to 13a,b involves the initial vinylic bro-
mination of 11a,b followed by alkylation of the bromide products
14a,b to afford the products 13a,b, respectively. A second Suzuki
reaction was then completed on compounds 13a–c to afford the re-
quired products 15a–d in good yield. For synthesis of the diphen-
olic product 18, it was necessary to first demethylate 12b with
pyridinium hydrochloride to produce the phenol 16 as the more
usual demethylating reagent boron tribromide etherate caused
extensive degradation of both 13b and 15b. Subsequent arylation
of the bromide 17 afforded the required diphenolic product 18
(Scheme 2).
The fluorinated benzoxepin series was prepared in a similar
procedure as illustrated in Scheme 3. Reaction of 3-fluorophenol
19 with ethyl 4-bromobutyrate followed by base catalyzed hydro-
lysis of the ester afforded the acid 20 which was cyclized with
polyphosphoric acid to give the 8-fluorobenzoxepin-4-one 21.26
O
R1
O
R1
(d)
O
R1
OH
R1
R1
O
(c)
(b)
(a)
R2
R2
CO2H
O
7a R = H
7b R = OMe
10a R1 = H, R2 = OTf
10b R1 = OMe, R2 = OTf
10c R1 = H, R2 = Br
8a R = H
8b R = OMe
9a R = H
9b R = OMe
HO
11a R1 = R2 = H
11b R1 = OMe, R2 = H
14a R1 = H, R2 = Br
14b R1 = OMe, R2 = Br
10d R1 = OMe, R2 = Br
O
R1
O
R1
R4
O
R1
(d)
(f)
(e)
Br
R4
O
O
O
R2
R2
R2
N
N
N
R3
R3
R3
15a R1 = R4 = H, R2R3=-(CH2)3-
13a R1 = R4 = H, R2R3= -(CH2)3-
12a R1 = H, R2R3= -(CH2)3-
12b R1 = OMe, R2R3= -(CH2)3-
12c R1 = OMe, R2R3= -(CH2)4-
15b R1 = H, R2R3=-(CH2)3-, R4 = OH
13b R1 = OMe, R4 = H, R2R3= -(CH2)3-
13c R1 = OMe, R4 = H, R2R3= -(CH2)4-
13d R1 = OMe, R4 = Br, R2R3= -(CH2)3-
15c R1 = OMe, R2R3= -(CH2)3-, R4 = OH
15d R1 = OMe, R2R3= -(CH2)4-, R4 = OH
Scheme 1. Synthesis of Type I benzoxepin SERMs 15a–d. Reagents and conditions: (a) i—BrCH2CH2CH2CO2Et, K2CO3, DMF; ii—NaOH, EtOH; (b) R = H, F; PPA, 110 °C, 4 h;
R = OMe, Eaton’s reagent, 80 °C, 2 h; (c) (TfO)2O, Na2CO3, 18 h, rt; (d) Pd(PPh3)3, 4-HOC6H5B(OH)2, Na2CO3(aq), THF; (e) K2CO3, 1-(2-chloroethyl)pyrrolidine hydrochloride or
1-(2-chloroethyl)piperidine hydrochloride, acetone, 24 h, 85 °C; (f) PyHBr3, CH2Cl2, 20 °C, 18 h; (g) PyꢀHCl, 190 °C, 4 h.
O
HO
(c)
O
HO
(b)
O
HO
(a)
Br
12b
OH
O
O
N
O
N
N
16
18
17
Scheme 2. Synthesis of Type I benzoxepin SERM 18. Reagents and conditions: (a) PyꢀHCl, 190 °C, 4 h; (b) PyHBr3, CH2Cl2, 20 °C, 18 h; (c) Pd(PPh3)3, 4-HOC6H5B(OH)2,
Na2CO3(aq), THF.

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9557
O
O
F
F
O
F
(b)
F
OH
(c)
(a)
CO2H
TfO
O
19
20
21
22
O
F
O
O
F
F
(d)
(f)
(e)
R
Br
Br
O
HO
HO
N
23
24a
24b
R = H
R = Br
25
(g)
(h)
O
F
O
F
O
F
(d)
R2
R1
Br
O
OH
O
N
O
N
N
26a
26b
26c
26d
26e
R1 = R2 = H
28
R1 = OH, R2 = H
R1 = H, R2 = OH
R1 = F, R2 = H
27
R1 = SO2Me, R2 = H
Scheme 3. Synthesis of Type 1 benzoxepin SERMs 26a–e and 28. Reagents and conditions: (a) i—BrCH2CH2CH2CO2Et, K2CO3, DMF; ii—NaOH, EtOH; BrCH2CH2CH2COOH ; (b)
PPA, 110 °CC, 4 h; (c) (TfO)2O, Na2CO3, 18 h, rt (d) Pd(PPh3)3, 4-HOC6H4B(OH)2, Na2CO3(aq), THF; (e) PyHBr3, CH2Cl2, rt, 18 h; (f) K2CO3, 1-(2-chloroethyl)pyrrolidine
hydrochloride, acetone, 24 h, 85 °C; (g) Pd(PPh3)3, R1R2C6H3B(OH)2, Na2CO3(aq), THF; (h) K2CO3, 1-(2-chloroethyl)piperidine hydrochloride, acetone, 24 h, 85 °C.
The phenol 23 was obtained in yield of >90% by Suzuki arylation of
the triflate 22 catalyzed by Pd(PPh3)4 without the need for isolation
of the intermediate triflate 22. Careful bromination of 23 was re-
quired to obtain the product 24a free from contamination with
the dibrominated compound 24b. The bromide 24a was then alkyl-
ated with 1-(2-chloroethyl)pyrrolidine to give 25 which was sub-
sequently arylated in a second Suzuki reaction with a selection of
boronic acids to afford the products 26a–e. Introduction of the
piperidine type basic substituent required the initial alkylation of
24a to give 27 which was subsequently arylated to afford the phe-
nolic product 28.
The synthesis of the isomeric compound 32, which could be
compared to the (E)-tamoxifen stereochemistry was also accom-
plished (Scheme 4). The phenol 24a was converted to the protected
TBDMS ether 29 by treatment with tert-butyltrimethylsilyl chlo-
ride, in DMF in the presence of imidazole. Subsequently arylation
in a second Suzuki reaction with the phenolic boronic acid pro-
vided the phenol 30. Introduction of the basic substituent onto
Ring C was accompanied by in situ deprotection of 31 to afford
the required product 32 in 60% yield.
at C-4 of the benzoxepin system (Type III compounds) was
achieved as outlined in Scheme 5. The products were obtained by
first bromination of the benzoxepin 11b and then demethylation
of 33 with boron tribromide to afford the phenol 34.18 Subsequent
Suzuki arylation of the bromide 34 with the appropriate boronic
acids afforded the products 35a–e. The basic ethers 36a–e were
then produced by standard alkylation of the phenols 35a–e with
N,N-dimethylaminoethyl chloride.
A flexible methylene linking group was introduced between C-
4 of the benzoxepin system and the aromatic ring C (Type IV
compounds) as illustrated in Scheme 6. Aldol condensation of
the benzoxepin 9a with the appropriate aryl aldehyde in the pres-
ence of HCl afforded the 4-benzylidenebenzoxepin-5-one 3727
which was then hydrogenated to form 38. Direct reaction of 38
with N,N-dimethylaminoethoxyphenyl bromide and n-butyllith-
ium was not successful. However, treatment of 38 with n-butyl-
lithium and 4-methoxybromobenzene followed by acid
promoted dehydration of the intermediate alcohol resulted in iso-
lation of the alkene products 39. The target products were ob-
tained following demethylation of 39 with boron tribromide
and subsequent alkylation of the phenolic products 40 to afford
the desired basic ethers 41a–e.
The introduction of a number of alternative Ring C types, for
example, pyridine, naphthalene, furan, thiophene, and benzofuran

9558
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
O
F
O
F
O
F
(a)
(c)
Br
Br
OH
O
O
HO
24a
Si
Si
29
30
O
F
O
F
(d)
(e)
O
O
N
N
HO
O
Si
31
32
Scheme 4. Synthesis of Type II SERM, 32. Reagents and conditions: (a) TBDMSiCl, imidazole, DMF; (b) Pd(PPh3)3, 4-HOC6H5B(OH)2, Na2CO3(aq), THF; (c) K2CO3,
1-(2-chloroethyl)pyrrolidine hydrochloride, acetone, 24 h, 80 °C; (d) (n-Bu)4NF, THF.
fused analogue of 4-hydroxytamoxifen was reported McCague et
al. to have a close correlation of stereochemistry between the triar-
ylbutene and the seven membred ring structure, particularly in the
orientation of the aryl rings as determined by X-ray crystallogra-
phy.20 Compound 6 was shown to have similar antiproliferative ef-
fect as 4-hydroxytamoxifen in MCF-7 breast cancer cells at 10ꢁ8 M
and blocks the growth stimulation caused by 10ꢁ8 M estradiol.
Introduction of the pyrrolidine basic ether on the alternative
Ring C site, (compound 32) resulted in significant reduction of anti-
3. Biochemistry
The compounds were evaluated in a series of estrogen depen-
dent in vitro assays which measured their affinity for the estrogen
receptors and also their ability to act as functional antagonists or
agonists of estrogen.
3.1. Antiproliferative activity in MCF-7 breast cancer cells
proliferative activity (IC50 = 6.503 lM) which is consistent with the
The products were initially assessed for their antiproliferative
action using the ER expressing (ER dependent) human MCF-7 breast
cancer cell line and the results are displayed in Table 1. Examina-
tion of the antiproliferative data of compounds of type I show that
many of the products are active at submicromolar concentrations;
the most active compounds are those containing the OH group in
Ring C and either the F or OH substituent at the C-8 position. Com-
pounds 18 and 26b, with IC50 values of 11 nM and 21 nM, respec-
tively, were the most potent in the series. The replacement of the
OH group on compound 18 (ClogP = 5.53) with lipophilic fluorine
in 26b (ClogP = 6.30) results in a slight decrease in the IC50 value
from 11.9 to 21.4 nM. The nature of the substituent in the Ring B
is therefore critical in determining the antiproliferative activity
and ER binding ability of these compound series. The monohydroxy
compound 15b and difluoro compound 26d were found to have
moderate antiproliferative activity (IC50 446 and 450 nM), although
considerably greater than tamoxifen, see Table 1.28 Introduction of
the phenolic group on Ring C at the meta-position as in compound
26c, reduces the potency of the compound (IC50 304 nM) while the
presence of a methyl sulfone grouping compound 26e, resulted in a
previous reports of the poorer antiproliferative activity of the E-
type tamoxifen and the less favorable stereochemical arrangement
for antagonist binding in the ER-LBD, possibly contributing to ago-
nist effects observed for E-tamoxifen.33
From our previous investigations,18 the nature of ring C was
found to also be crucial in determining the ER affinity of the ben-
zoxepin ligand. Structural type III compounds (36a–e) containing
a heterocyclic (3-pyridyl, 2-furyl, 2-thienyl, and 2-benzofuryl) or
1-naphthyl Ring C substituent displayed poor antiproliferative
activity with IC50 in the range 16–20
lM for MCF-7 cells (e.g., for
compound 36c, IC50 MCF-7 = 16.7 M) and poor ER
l
a
binding va-
lue; IC50 = 828 nM) and were not further investigated. Introduction
of a flexible hinge to link Ring B to the central ethylene core has
been achieved in many SERMs, for example, the carbonyl group
in raloxifene,34 the oxygen ether in arzoxifene,35 the oxygen ether
and methylene linker groups in naphthalene23 and tamoxifen
based SERM examples.36 However, introduction of a lipophilic ben-
zylic Ring C type substituent as in structural type IV (compounds
41a–e) resulted in low antiproliferative activity for the series (e.g.,
compound 41d, IC50 = 10.4
14.2 M).
lM) and poor ERa binding (IC50
dramatic loss of activity for the compound, (IC50 4.56 lM). The
l
methyl sulfone group has been exploited as a Ring C phenol replace-
ment in naphthalene based SERMs to improve pharmacokinetic
properties as it cannot undergo conjugation metabolism to the glu-
curonide or sulfate ester.23
The antiproliferative results obtained in this study compare
favorably with the activity reported for products based on the ben-
zocycloheptene and dibenzothiepin scaffold structures.3,20,29–32
The structurally related benzocycloheptene compound 6, a ring
The most potent examples were also tested in hormone inde-
pendent MDA-MB-231 breast cancer cells. The growth was not
inhibited by the benzoxepin compounds up to 100 nM. At higher
concentrations all compounds showed a cytotoxic effect but the
IC50 value determined in the MDA-MB 231 cell line for the most
potent compounds were at least 10-fold greater than that obtained
for the same compounds in the MCF-7 cell line (Compounds 15b,

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9559
15c, 26b, and 18 gave IC50 values of 5.10
2.60 M, respectively, when evaluated in the MDA-MB-231 cell
line).
lM, 5.24 lM, 6.75 lM,
O
O
O
(c)
(b)
l
(a)
Br
Br
3.2. Cytotoxicity
H3CO
33
The cytotoxicity of the compounds in the MCF-7 cell line was
also determined in the standard LDH assay as previously reported
as we wished to confirm that the antiproliferative effects of the
compounds were due to cytostasis rather than cellular necrosis.
The compounds all demonstrated low cytotoxicity profiles sug-
gesting that their action is cytostatic rather than cytotoxic. Typical
values for cytotoxic induced antiproliferative effects for compound
HO
H3CO
11b
34
O
O
(d)
R
15a at 10
whereas value for tamoxifen is observed to be 24% at testing con-
centration of 10 M. The cytotoxicity values obtained were less
lM is 8%, compound 18 is 11%, and compound 11 is 17%
R
l
O
than that for tamoxifen for all compounds evaluated.
HO
35 a-e
N
3.3. Estrogen receptor binding studies
36a-e
Estrogen receptor binding studies were carried out for the most
potent compounds of the series with ERa and ERb using a fluores-
cence polarization procedure.36 The displacement of fluorescein la-
beled estradiol (fluoromone) in a competitive binding assay from
the human recombinant full length receptor proteins ERa and
ERb expressed from baculovirus-infected insect cells by the syn-
thesized ligands was observed as a decrease in polarization values.
The IC50 values are calculated from resulting sigmoidal inhibition
curves as illustrated in Table 1.
35a
35b
R =
36a
36b
R =
R =
N
N
O
O
R =
S
S
For the Type I benzoxepin scaffold structures, most compounds
35c
35d
35e
R =
R =
R =
36c
36d
36e
R =
R =
R =
showed potent ER
a binding affinity with IC50 < 20 nM. The pres-
ence of a hydroxyl or fluoro substituent at position 8 in Ring B to-
gether with a hydroxyl group at C-4 in Ring C proved to be very
effective as demonstrated in compounds 18 and 26b and with
IC50 values of 8.4 nM and 10.4 nM, respectively. The piperidine ba-
sic group present in compound 28 proved to equally effective
(IC50 = 6.5 nM) as the pyrrolidine in its role as antiestrogenic basic
O
O
binding group for ERa LBD Aspartic acid 351. The values compare
very favorably with the values obtained for tamoxifen and
hydroxytamoxifen, (Table 1). These benzoxepin compounds also
Scheme 5. Synthesis of Type III SERMs 36a–e. Reagents and conditions: (a) PyHBr3,
CH2Cl2, 20 °C, 18 h; (b) BBr3, DCM; (c) Pd(PPh3)4, RB(OH)2, Na2CO3(aq), THF;
(d) ClCH2CH2N(CH3)2, K2CO3, acetone, 24 h, 85 °C.
displayed considerably more potent binding to ERa and ERb than
O
O
O
(b)
(c), (d)
(a)
CH3
CH3
O
O
O
9a
37
38
O
O
O
(e)
(f)
CH3
CH3
CH3
O
H3CO
HO
R1
N
R2
39
40
41a R1 = R2 = CH3
41b R1 = R2 = CH2CH3
-
41c R1R2 = -CH2CH2CH2
-
41d R1R2 = -CH2CH2CH2CH2
41e R1R2 = -CH2CH2OCH2CH2
-
Scheme 6. Synthesis of Type IV SERMs 41a–e. Reagent and conditions: (a) 4-CH3-C6H4CHO, HCl, EtOH; (b) H2/Pd, EtOH; (c) 4-MeO-C6H4Br, nBuLi, THF; (d) H3PO4, EtOH; (e)
BBr3, CH2Cl2; (f) K2CO3, R1R2NCH2CH2Cl, acetone, 24 h, 85 °C.

9560
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
Table 1
Antiproliferative effects and estrogen receptor binding affinities for benzoxepin type SERMs
a
b
b
Compound
Structure type
Yield (%)
MCF-7 IC50
(lM)
ER
a
IC50 (nM)
ERb IC50 (nM)
ER ab
15a
15b
15c
15d
18
26a
26b
26c
26d
26e
28
32
36c
41d
I
I
I
I
I
I
I
I
I
I
I
II
III
IV
—
—
89
37
39
30
12
40
68
50
31
22
84
60
71
31
—
1.584 0.262
0.446 0.111
1.280 0.147
1.540 0.274
0.0119 0.005
1.869 1.140
0.0214 0.0063
0.359 0.163
0.450 0.321
4.560 1.00
0.156 0.050
6.503 0.290
16.70 2.10
10.40 2.70
4.12 0.038
8.77
1.9
6.7
10.3
8.4
46.1
5.9
4.61
3.63
2.9
31.0
0.70
6.0
5.3
3.1
0.69
0.35
0.35
1.9
0.07
0.43
2.07
—
1.35
3.11
—
—
16.2
10.4
14.0
126
>15
6.5
262
l
M
>15
l
M
8.78
17.33
828
14.2
170
20
5.57
828c
14.2
70
c
c
lMc
lM
Tamoxifend(1)
4-Hydroxy tamoxifen (2)
2.42
0.5
—
0.107 0.081
40
a
Experimental values represent the average for two experiments performed in triplicate along with the standard deviation (SD) between the assay values. Values without
SD values were run once. IC50 values are half maximal inhibitory concentrations required to block the growth stimulation of MCF-7 cells.
b
Values are an average of at least nine replicate experiments for ERa with typical standard errors below 15%; and six replicate experiments for ERb with typical standard
errors below 15%.
c
The ER binding values for compounds 36c and 41d were obtained in a competitive radiometric binding assay with [3H]estradiol rat uteri cytosol and are expressed as Ki
values.
d
The value for tamoxifen IC50 4.12 0.38 lM is in good agreement with the reported IC50 value for Tamoxifen using the MTT assay on human MCF-7 cells.
that reported for the binding of related benzocycloheptene com-
pound 6 to estrogen receptors in rat cytosol,20 indicating that the
benzoxepin ring is more favorably bound to the receptor than
the more lipophilic benzocycloheptene structure.
17.33 nM. The Type III and Type IV compounds containing hetero-
cyclic and benzylic Ring C structures were shown to be ineffective
ligands for the ER
a and ERb with IC50 values >15 lM.
It is evident that the presence of the hydroxy group in Ring B is
not essential for the potent binding of these compounds. These
binding studies are also broadly consistent with the antiprolifera-
tive activity of these compounds as observed in the MCF-7 study
in which the most potent compounds were identified as 18 and
3.4. Estrogenic stimulation
The estrogen stimulation and antagonistic properties of the
most active compounds 18 and 26b and was determined in the
estrogen bioassay which is based on the stimulation of alkaline
phosphatase (AlkP) in the Ishikawa human endometrial adenocar-
cinoma cell line.38 Compounds are tested as estrogen antagonists
by their effect on the inhibition of estradiol stimulation in the
Ishikawa cells in a dose-dependent manner (Table 2). Compound
26b was approximately equivalent to tamoxifen in its ability as
an estrogen antagonist with IC50 value of 194 nM. Compound 18
was much more potent than tamoxifen as an estrogen antagonist
with IC50 values of 15.9 nM, respectively, and correlates with the
26b. Since MCF-7 cells predominantly express ER
erative activity may be mediated by the ER receptor subtype,
however the contribution of ERb subtype to antiproliferative activ-
ity is also possible. Pure antiestrogens are known to degrade ER
a, this antiprolif-
a
a
and stabilize ERb and this process can alter the ratio of these sub-
types during treatment. A notable exception in the series was com-
pound 26e, containing a methyl sulfone substituent in Ring B. This
substituent was introduced to prevent unfavorable metabolic
hydroxylation on Ring B.23 However it seems to prevent the usual
potent ERa and ERb binding observed for these compounds. The
antiestrogenic binding of the compound associated with the ER
a
estrogenic stimulatory properties of these compounds could be
determined in the Ishikawa cells by measuring the stimulation of
alkaline phosphatase(AlkP) in these cells in the absence of estra-
diol. Compounds 18 and 26b, the most potent benzoxepin-derived
compound in the series, showed low stimulatory value (6% and
LBD residues Argenine 394 and Glutamic acid 353. For the ben-
zoxepin compounds, the increase in lipophilicity37 on replacement
of the OH (calculated ClogP = 5.53) at C-8 of compound 18 with F
in compound 26b (ClogP = 6.30) resulted in small increase in both
antiproliferative and ER binding activity. For ERb binding, the ben-
zoxepin Type I compounds 15b, 15c, 26c, 15d, and 28 were potent
as ERb ligands with IC50 values of <10 nM. With compounds 18 and
26b being the most potent, with IC50 binding values of 2.9 nM and
0.7 nM, respectively.
9.7%, respectively, at concentration of 1
lM). The reverse com-
pound 32 displayed stronger stimulatory activity of 23% at 1
l
M
concentration in the absence of estradiol. This compound which
is a ring fused analogue of E-tamoxifen, also displayed estrogen
antagonist activity approximately equivalent to tamoxifen with
IC50 value of 216 nM, which is confirmed in the MCF-7 antiprolifer-
ative assay result above. The results from this AlkP assay for estro-
gen antagonism and stimulation provide useful information in
selection of the structural features for optimum antiestrogenic
activity without associated adverse estrogenic effect on tissues
such as the uterus.
Many of the compounds displayed a low selectivity for ER
a, for
example, compounds 15b and 28 with b ratio of 3.1 and 1.35,
a
respectively. The selectivity of some of the compounds for ERb
could be related to the presence of the F substituent on Ring B as
demonstrated for compounds 26b which showed b
a ratio of
15:1. The possible binding orientation of this ligand in the ER-
LBD is explored in the modeling section below. The corresponding
diphenolic compound 18 showed much lower selectivity with b/
ratio of 2.9 which is closer to the b/ ratio value of 1.67 obtained
for 4-hydroxytamoxifen in this experiment.
In the Type II compounds, introduction of the pyrrolidine basic
ether on the alternative Ring C site (compound 32) resulted in
a
3.5. Computational Investigation
a
To rationalize the observed ER
zoxepin compounds (15a–d, 18, 16a–e, 28, and 32) shown in Ta-
ble 1, a QSAR model for ER b selectivity in the benzoxepin
scaffold structure was developed using the general core structure
ab selectivity of the novel ben-
a
unexpected potent ER
a binding result IC50 5.57 nM, and ERb IC50

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9561
for the benzoxepin antiestrogens 15a–d as illustrated in Scheme
1. A data set of 27 estrogen receptor antagonist compounds was
selected (see Supplementary information, Tables 5 and 6) utiliz-
ing ER active ligands extracted from literature including three li-
number of chiral centers, Petitjean graph shape coefficient, and to-
tal atom information content were deemed as important contribu-
tors. Our best model (Table 3, No. 8) could be described by the
following equation
gands with similar structures to the benzoxepin scaffolds of this
series (benzothiepin-derived SERMs) previously published by
our group.39 A number of different initial descriptor sets, includ-
ing only substructural, 2D-MOE, normalized 2D-MOE descriptors,
and different combinations of all, were chosen to analyze their
n = 20; RMSEtr = 0.303; R2 ¼ 0:836; RMSEval = 0.402; Q2 = 0.715;
tr
RMSEtest = 0.177.
The propensity for a ligand to bind with higher affinity to ERb is
correlated with higher LogP (ER
a binding affinity decreases with
increase in lipophilicity),40 lower van Der Waals surface area, but
most importantly the presence of a seven-membered benzoxepin
heterocycle within the nucleus of the antiestrogenic core.
impact on observed activity. The ratio of selectivity of ER b/
a
binding values (pIC50 values) was used as input for the prediction
of selectivity of the benzoxepin series. This data was represented
in the form of the natural logarithm for convenience. The original
data set was divided several times into a training set of 20 ligands
and an external test set of seven ligands, respectively. Cross-val-
idation procedure was implemented by using leave-one-out
(LOO) cross-validation procedure implying exclusion of each
compound of the training set and the prediction of its activity
by the model developed using the remaining compounds of the
training set.
The variety of PLS (partial least squares) models (1–11) based
on different descriptors sets and utilizing different selections of
training and test sets is represented in Table 3, together with the
predictive ability of each model.
Analysis of selected descriptor sets revealed that substructural
descriptors play a dominant role in prediction of ERab selectivity.
These consisted of mainly chains of 3, 7–15 atoms with different
bond environmental (i.e., single, double, and aromatic) and few
branched fragments of four atoms. Among the list of other 2D
descriptors appeared some physicochemical ones, such as hydro-
phobicity (logP), topological polar surface area, total negative
and positive van der Waals surface area, mass density, molar
refractivity (MR) descriptors. Also some other indices such as the
Figure 2 illustrates correlation between experimental and pre-
dicted values of the activity for all subsets for the best model (Table
3, No. 8). The ER subtype selectivity effect of the 7-memebered
fused-ring system is more apparent when one examines the
docked solutions of 26b in both estrogen receptor isoforms.
3.6. Docking
To further rationalize the observed ERb selectivity (15-fold) of
compound 26b, a flexible ligand–receptor docking investigation
was undertaken. At present there are no crystal structures of hu-
man ERa and ERb with the same antagonist co-crystallized and
thus it is difficult to compare docking of the same ligand in differ-
ent isoforms with accuracy. To overcome this, Autodock4.041 was
employed with movement of several key active site residues al-
lowed. A summary of RMSD differences between active site resi-
dues in ER
a obtained by superposition of data from 10 crystal
structures is presented in Figure 3.
We examined the ability of our procedure in reproducing the
binding position for raloxifene in PDB entry 1ERR, when docked
in another X-ray structure (PDB: 3ERT) with a co-crystallized
antagonist. The procedure was successful in reproducing the bind-
ing position for raloxifene in 3ERT (Glu353, 2.70 Å; Arg394, 2.80 Å;
His524, 3.20 Å; Asp351, 4 Å) when compared with 1ERR (Glu353,
2.40 Å; Arg394, 3.00 Å; His524, 2.70 Å; Asp351, 2.70 Å) exhibiting
the same patterns for hydrogen bonding.
Table 2
Antiestrogenic activity for selected Benzoxepin type SERMs
Compound
Ishikawa IC50 (nM)a
% Stimulation (1 lM)
Next, we examined the possible binding mode of the most ERb
18
26b
32
20.8 5.4
194 36.6
216 141nM
170 0.0
6.0
9.7
23
selective benzoxepin, compound 26b, when docked in ERa (3ERT).
Firstly and importantly, two possible conformations (1 and 2) of
the 7-membered benzoxepin ring system can be observed for com-
pound 26b as illustrated in Figure 4. Both conformations (1 and 2)
were docked in each ER isoform to establish any differences in
binding of the two ring conformations. The most stable solutions
Tamoxifen
16.5%
a
Ishikawa IC50 values are the compound concentrations required to inhibit by
50% the stimulation produced by 1 nM estradiol as determined by the alkaline
phosphatase quantitation. Experimental values represent the average for triplicate
determinations along with the standard deviation (SD) between the assay values.
for 26b complexed in ERa exhibited a propensity towards binding
Table 3
Statistical parameters of PLS models for ERb/ERa binding selectivity (IC50 values) of estrogen receptor active ligands
Descriptor type in the initial set
Model
RMSEtr
R2
RMSEval
Q2
RMSEtest
Substructural
Substructural
1
2
0.305
0.314
0.300
0.312
0.306
0.319
0.320
0.303
0.326
0.355
0.334
0.846
0.821
0.841
0.839
0.832
0.831
0.831
0.836
0.824
0.792
0.789
0.432
0.436
0.402
0.462
0.419
0.439
0.726
0.402
0.476
0.453
0.431
0.705
0.665
0.718
0.656
0.694
0.689
0.158
0.715
0.649
0.668
0.655
0.542
0.231
0.260
0.457
0.240
0.453
0.390
0.177
0.459
0.457
0.164
Substructural + MOE 2D
Substructural + MOE 2D
Substructural + MOE 2D
Substructural + MOE 2D
Substructural + MOE 2D
Substructural + MOE 2D + normalised MOE 2D
Substructural + MOE 2D + normalised 2D
Substructural + MOE 2D + normalised 2D
Substructural + MOE 2D + normalised 2D
3/1
4/2
5/3
6/4
7/5
8/1
9/2
10/3
11/4

9562
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
A
B
9
2.0
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
1.5
1.0
y = 0.8357x + 0.083
y = 0.7775x + 0.077
0.5
2 = 0.7146
R2 = 0.8357
R
0.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
-0.5
-1.0
log (IC50-beta/ IC50-alpha)exp
log (IC50-beta/ IC50-alpha)exp
0.3
C
0.2
0.1
0.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
-0.1
-0.2
-0.3
-0.4
-0.5
y = 0.3542x - 0.0202
R2 = 0.7566
log (IC50-beta/ IC50-alpha)exp
Figure 2. Correlation between experimental and predicted ERb/ER
a
selectivity values (logIC50ERb/logIC50ER
a
) for ensemble of estrogen receptor active ligands: (A) training;
(B) validation; (C) test set.
3.5
3
1SJ0
1UOM
1XP1
1XP6
1XPC
1XQC
1YIM
1YIN
2.5
2
1.5
1
0.5
0
3ERT
Figure 3. Summary of RMSD differences between active site residues in ERa obtained by superposition of data from 10 crystal structures.
as illustrated in Figure 3 irrespective of their ring conformations,
(free energy of binding,
G ꢁ (conformation 1) = ꢁ17.72 kcal/
energy of ꢁ16.73 kcal/mol. Unlike in ER
a, two distinct binding ori-
D
entations of 26b (for conformation 1) are observed in ERb as illus-
trated in Figure 5. Met421 and Ile424 adjust their positions
accordingly to allow both ring conformations of 26b to bind in
mol; (conformation 2) = ꢁ17.50 kcal/mol). Compound 26b clearly
docks in an antiestrogenic manner in ER
tacts as outlined in Table 4.
a making H-bonding con-
the same manner in ERa. In the case of ERb, only Ile328 and
In the case of docking both ring conformations of 26b in ERb
Ile321 move position slightly to accommodate the alternative
(1QKN), conformation 1 is preferred binding mode, with a binding
binding modes.
Virtually all the same binding mode contacts are made in both
orientations of 26b in ERb as detailed in Table 4. However, what is
immediately apparent and different to those binding modes ob-
Table 4
served in ER
with His524 in ERb as illustrated in Figure 6. The substitution of
Met421 (ER ) with Ile328 (ERb) causes steric interaction with
a, is that the ligand is devoid of H-bonding interaction
Hydrogen bonding contacts for 26b in both ring conformations (1 and 2) in both
isoforms ER
a and ERb8
a
Conformation
Isoform
Binding energy (kcal/mol)
Distance (Å) Asp 351 (258)
Glu 353 (260)
Arg 394 (301)
Gly 521 (427)
His 524 (430)
1
1
2
2
the ligand, supplementary to that imposed by Phe332, preventing
the ligand from docking in the same orientation as observed with
ERa. Ile328 aids this process moving out of position to accommo-
date the ligand in its alternative binding mode, and consequently
His430 can make a hydrogen bonding contact with Gly327 instead
of to the ligand. As mentioned earlier the lowest energy binding
mode of 26b is actually the alternative ‘ring flipped’ binding mode
3ERT (
ꢁ17.72
a
)
1QKN (b)
ꢁ16.73
2.7
3ERT (
ꢁ17.50
a
)
1QKN (b)
ꢁ16.24
2.7
2.6
2.7
2.99
2.58
2.7
2.8
2.7
2.9
—
3.67
2.46
2.8
2.71
2.7
2.9
3.2
3.0
—

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9563
in which the para-fluoro of the ligand is directed towards Histidine
524. The strength of the bond between this fluorine and His430
would be less than if it hydrogen bonds to Gly327 so this relocation
of His430 would be preferable.
ability to bind in two alternative ring ‘flipped’ orientations because
of the presence of this mutation from Met421 to Ile328 in ERb.
Both our QSAR and docking analysis converge on the importance
of the position of the fluorine substituent (enhanced lipophilicity)
and the fused benzoxepin ring system in providing binding selec-
tivity towards ERb.
Celik et al.42 have recently carried out a 158 ns unrestrained all-
atom MD simulation of ER
a LBD sampling the conformational
changes upon binding of estradiol and to establish the importance
of His524 in the ligand binding process. Surface exposure of His524
was seen to translate to movement in H12 via disruption of the
conserved hydrogen bonding network between Helix 3 and Helix
11. We suggest from this and our analysis that this rotation of
His430 in ERb may also result in this plasticity in Helix12 and con-
sequently, ease of hydrogen bonding contact between Asp258 and
the amine of the basic side chain of the ligand.
4. Conclusions
We have identified a number of fluoro substituted benzoxepin
compounds which demonstrate antiproliferative activity against
the MCF-7 human breast cancer cell line, and in addition display
up to 15-fold selectivity for ERb together with subnanomolar bind-
ing affinity for the ERb. Initial molecular modeling studies indicate
that these compounds may dock in an unexpected mode in the
binding site of the ER. Metabolic oxidation of these compounds
would be blocked by the presence of the F atom in Ring B thus
ensuring a longer plasma half life. These compounds which are ring
It has been suggested previously by a number of groups43,44 that
enhanced ERb selectivity is achieved through a ligand that can dif-
ferentiate between ERa Met421 and ERb Ile328. Compound 26b
exhibits these preferences through ligand ring conformations and
Figure 4. Left: Two possible conformations (1 and 2) of benzoxepin 26b ring system. Right: Binding mode of 26b in ERa active site.
Figure 5. Left: Two possible conformations of benzoxepin ring system 26b docked in ER
ERb.
a. Right: Alternate ‘ring flipped’orientations of benzoxepin ring system 26b docked in

9564
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
5.2. 5-Bromo-2,3-dihydrobenzo[b]oxepine (10c)
mixture of 3,4-dihydro-2H-benzo[b]oxepin-5-one (9a)
A
(0.33 g, 2.04 mmol) and phosphorus tribromide (1.00 mL,
10.4 mmol) was heated at 90 °C for 24 h. The reaction was cooled
to room temperature, added slowly dropwise to ice-water
(20 mL), and the aqueous layer extracted with ethyl acetate (3ꢂ
50 mL). The combined organic layers were washed with water
(20 mL), dried over sodium sulfate, and the solvent removed under
reduced pressure. The product was purified by chromatography
(5% diethyl ether/hexanes) to give an unstable yellow oil (0.26 g,
56%), which was used immediately in the next reaction without
further purification.
5.3. 4-(2,3-Dihydrobenzo[b]oxepin-5-yl)-phenol (11a)
Method A: Pd(PPh3)4 (40 mg) was added to a solution of 10c
(1.12 g, 5 mmol) in THF (20 mL) and stirred at room temperature
of 30 min. 4-Hydroxyphenylboronic acid (0.76 g, 5.5 mmol) and
2 M Na2CO3 (12.8 mL, 25 mmol) was added and the solution
heated to 80 °C for 24 h. The solution was cooled, water was added
(20 mL), and the aqueous layer extracted with dichloromethane
(4ꢂ 50 mL). The combined organic layers were washed with brine
(20 mL), dried over sodium sulfate, and the solvent removed under
reduced pressure. The crude product was purified by column chro-
matography (5% diethyl ether/hexanes) to give the prooduct as a
Figure 6. Lowest energy docked structures of 26b in ER
heavy atom. Hydrogen bond contact for 26b with His524 in ER
a
and ERb superposed by
clearly shown.
a
fused analogue of (Z)-tamoxifen, expand the structural diversity of
ligands that can act as potent antagonists for the estrogen receptor
and without metabolic complications introduced by the facile
isomerisation of the tamoxifen type triarylethylene antiestrogen
structures. Selectivity for ERb for the most potent compound 26b
has been examined in both docking and QSAR studies and suggest
a correlation between enhanced lipophilicity and enhanced ERb
binding for this benzoxepin ring scaffold.
white solid (0.62 g, 52%).IR mmax (KBr): 3384, 1606 cmꢁ1 1H NMR
;
(400 Mz, CDCl3) d 7.23 (1H, m, ArH), 7.14 (3H, m, ArH), 7.0 (2H,
m, ArH), 6.77 (2H, d, J = 6.0, ArH), 6.27 (1H, t, J = 6.0, CH), 2.51
(2H, t, J = 6.0, OCH2), 2.46 (2H, m, CH2); 13C NMR (101 Mz,
CDC13) d 157.6, 154.8, 140.9, 135.3, 133.0, 131.0, 129.8, 128.4,
126.9, 123.3, 121.9, 114.9, 78.0, 29.7.
Method B: Trifluoromethanesulfonic anhydride (5.0 mL,
30.9 mmol) in dichloromethane (10 mL) was added to a suspension
of 9a (1.66 g, 10.3 mmol) and sodium carbonate (3.28 g,
30.9 mmol) in dichloromethane (30 mL) at 0 °C. The suspension
was stirred for 18 h at room temperature, filtered and the filtrate
washed with water (50 mL), brine (50 mL), dried over sodium sul-
fate, and the solvent removed under reduced pressure. The crude
triflate 10a was dissolved in dry THF (50 mL) and Pd(PPh3)4
(100 mg) was added and the solution stirred under nitrogen for
15 min. 4-Hydroxyphenylboronic acid (1.79 g, 12.9 mmol) and
2 M sodium carbonate (25 mL) was added and the solution re-
fluxed for 6 h. The solution was cooled to room temperature and
acidified with 2 M hydrochloric acid. The aqueous layer was ex-
tracted with dichloromethane and the combined organic layers
were washed with water (30 mL), brine (30 mL), dried over sodium
sulfate, and the solvent removed under reduced pressure. The
product was purified by flash chromatography (5% diethyl ether/
hexanes) to afford the product 11a as a white solid (2.15 g, 88%).
5. Experimental
5.1. Materials and methods: chemistry
All reagents were commercially available and were used with-
out further purification unless otherwise indicated. Anhydrous
THF was obtained by distillation from benzophenone-sodium un-
der nitrogen immediately before use. All reactions were per-
formed under a nitrogen atmosphere unless specifically noted.
IR spectra were recorded as thin films on NaCl plates or as KBr
disks on a Perkin-Elmer Paragon 100 FT-IR spectrometer. 1H
and 13C NMR spectra were obtained on a Bruker Avance DPX
400 instrument at 20 °C, 400.13 MHz for 1H spectra, 100.61 MHz
for 13C spectra, in either CDCl3, CD3COCD3, or CD3OD (internal
standard tetramethylsilane). Low resolution mass spectra were
run on a Hewlett–Packard 5973 MSD GC–MS system in an elec-
tron impact mode, while high resolution accurate mass determi-
nations for all final target compounds were obtained on
a
5.4. 4-(4-Bromo-2,3-dihydro-1-benzo[b]oxepin-5-yl)-phenol
Micromass time of flight mass spectrometer (TOF) equipped with
electrospray ionization (ES) interface operated in the positive ion
mode at the High Resolution Mass Spectrometry Laboratory in the
Department of Chemistry, Trinity College Dublin. Thin-layer chro-
matography was performed using Merck Silica gel 60 TLC alumi-
num sheets with fluorescent indicator visualizing with UV light at
254 nm. Flash chromatography was carried out using standard
silica gel 60 (230–400 mesh) obtained from Merck. All products
isolated were homogenous on TLC. All samples were analyzed
using reversed phase high performance liquid chromatography
(Waters Alliance system). The analysis was performed at
(14a)
Pyridinium tribromide (0.99 g, 2.8 mmol) was added to a solu-
tion of 11a (0.67 g. 2.8 mmol) in dry dichloromethane (30 mL)
0 °C and the solution was warmed to room temperature and stirred
for 1 h. A solution of sodium hydrogen carbonate (10%, 20 mL) was
added and the aqueous layer extracted with dichloromethane (3ꢂ
50 mL), washed with water, dried over sodium sulfate, and the sol-
vent removed under reduced pressure. The crude productwas puri-
fied by flash chromatography (CH2Cl2/hexane 1:1) to give the
product as a white solid (0.59 g, 67%). Mp 116 °C; IR mmax (KBr):
254 nm using a Phenomonex column (4
a mobile phase of acetonitrile/water (0.1% TFA) 70:30 deliverated
l
, 250 ꢂ 4.60 mm) using
3384, 1606 cmꢁ1
m, ArH), 7.14 (2H, d, J = 8.5, ArH), 7.07 (1H, dd, J = 8.0, 1.0, ArH),
;
1H NMR (400 Mz, CDCl3) d 7.20 (1H, t, J = 6.0,
at a flow rate of 1.0 mL/min.

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9565
6.95 (1H, t, J = 6.5, ArH), 6.85–6.77 (3H, m, ArH), 4.59 (2H, t, J = 6.0,
OCH2), 3.00 (2H, t, J = 6.0, CH2); 13C NMR (101 Mz, CDC13) d 156.4,
154.8, 131.4, 128.9, 123.4, 122.0, 114.9, 77.4, 41.1; HRMS (ESI) cal-
culated for C16H14BrO2 (M+) 317.1848; found: 317.1842.
stirred for 30 min. Ethyl 4-bromobutyrate (7.9 mL, 55 mmol) was
added dropwise via syringe and catalytic KI was added. The reac-
tion mixture was refluxed for 24 h. The reaction mixture was
cooled to ambient and the solid was removed by filtration. The so-
lid was washed with acetone (50 mL) and the combined filtrate
and washings were concentrated under reduced pressure. The res-
idue was taken up in diethyl ether (150 mL), washed with 5% so-
dium hydroxide (50 mL) and the solvent removed under reduced
pressure. The residue was dissolved in ethanol (20 mL) and treated
with 10% sodium hydroxide in water (100 mL) and heated at
110 °C for 3 h or until the solution went clear. The solution was
cooled and acidified with concd hydrochloric acid and the product
which precipitated was filtered and dried (8.05 g, 76%). mp 89–
5.5. 1-{2-[4-(2,3-Dihydro-benzo[b]oxepin-5-yl)-phenoxy]-ethyl}-
pyrrolidine (12a)
Potassium carbonate (0.94 g, 6.82 mmol) was added to a solu-
tion of 14a (0.46 g, 1.36 mmol) in acetone (50 mL). The suspension
was stirred for 15 min and 1-(2-chloroethyl)pyrrolidine hydrochlo-
ride (0.34 g, 2.04 mmol) was added and the mixture refluxed for
6 h. The solution was cooled, filtered, and the solvent removed un-
der reduced pressure. The residue was dissolved in dichlorometh-
ane (50 mL), washed with water (30 mL), brine (30 mL), dried over
sodium sulfate, and the solvent removed under reduced pressure.
The crude product was purified by flash chromatography (2.5%
methanol/ CH2Cl2) to give the product as a yellow oil (0.45 g,
91 °C19; IR vmax (film): 3220, 1725, 1559 cmꢁ1 1H NMR (400 Mz,
;
CDCl3) d 7.16 (1H, m, ArH), 6.51–6.45 (3H, m, ArH), 3.99 (2H, t,
J = 6.02 Hz, OCH2CH2CH2), 3.77 (3H, s, OCH3), 2.58 (2H, t, J = ?,
OCH2CH2CH2), 2.11 (2H, m, OCH2CH2CH2); 13C NMR (101 Mz,
CDC13) d 179.6, 160.8, 159.9, 129.8, 106.6, 106.4, 100.8, 66.4,
55.2, 30.5, 24.3; HRMS (ESI) calculated for C11H14O4Na (M++Na)
233.0790; found: 233.0787.
95%). IR tmax (film): 2962, 1606 cmꢁ1 1H NMR (400 Mz, CDCl3) d
;
7.21–7.14 (3H, m, ArH), 7.06 (1H, d, J = 8.0, ArH), 6.95–6.89 (3H,
m, ArH), 6.78 (1H, d, J = 8.0, ArH), 4.57 (2H, t, J = 6.0, OCH2), 4.13
(2H, J = 6.0, OCH2), 2.99 (2H, t, J = 6.0, CH2), 2.92 (2H, t, J = 6.0,
CH2), 2.65–2.62 (4H, m, CH2NCH2), 1.83–1.79 (4H, m, CH2CH2);
13C NMR (101 Mz, CDC13) d 157.7, 156.4, 139.3, 134.9, 133.5,
131.0, 129.4, 128.8, 123.4, 121.9, 121.8, 114.5, 113.9, 66.1, 54.8,
54.6, 41.1, 23.4; HRMS (ESI) calculated for C22H24BrNO2 (M+)
414.3461; found:, 414.3395.
5.9. 8-Methoxy-3,4-dihydro-2H-benzo[b]oxepin-5-one (9b)
A mixture of 8b (3.50 g, 16.7 mmol) and Eaton’s reagent (27 mL)
was stirred for 2 h at 100 °C under nitrogen. The mixture was
poured into water (100 mL) and extracted into ethyl acetate (4ꢂ
200 mL). The organic layer was washed with water (100 mL), brine
(100 mL), dried over sodium sulfate, and the solvent removed un-
der reduced pressure. The crude oil was purified by flash chroma-
tography (5% diethyl ether in hexanes) to give the product as a
yellow oil (0.94 g, 29%).19,26,45 A similar procedure using polyphos-
phoric acid gave a 14% of the benzoxepinone product. IR tmax
5.6. 1-{2-[4-(4-Phenyl-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenoxy]-ethyl}-pyrrolidine (15a)
This compound was prepared from 1-{2-[4-(4-bromo-2,3-dihy-
dro-benzo[b]oxepin-5-yl)-phenoxy]-ethyl}-pyrrolidine18 (13a)
using a procedure similar to that described for 11a. The crude
product was purified by flash chromatography (hexane/dichloro-
methane/methanol 48:48:2) to give the product as a brown oil
(film): 2934, 1607 cmꢁ1 1H NMR (400 Mz, CDCl3) d (2H, d,), 7.76
;
(2H, d, J = 6.0 HZ, ArH), 6.66 (1H, d, J = 9.0 HZ, ArH), 6.56 (1H, s,
ArH), 4.24 (2H, t, J = 6.0 HZ, OCH2), 3.83 (3H, s, CH3), 2.87 (2H, t,
J = 7.0 Hz, CH2), 2.0 (2H, m, CH2); 13C NMR (101 Mz, CDC13) d
157.7, 156.4, 139.3, 134.9, 133.5, 131.0, 129.4, 128.8, 123.4,
121.9, 121.8, 114.5, 113.9, 66.1, 54.8, 54.6, 41.1, 23.4; HRMS (ESI)
calculated for C11H13O3 193.0865 (M+); found: 193.0858.
(76 mg, 89%). IR tmax (film): 2962, 1606 cmꢁ1
;
HPLC
(tr = 2.422 min; 100%); 1H NMR (400 Mz, CDCl3) d 6.64 (2H, d,
J = 8.5, ArH), 4.61 (2H, J = 6.0, OCH2), 4.11 (2H, t, J = 5.5, CH2), 2.99
(2H, t, J = 5.5, CH2), 2.79 (4H, m, CH2NCH2), 2.70 (2H, t, J = 6.0,
CH2), 1.89–1.86 (4H, m, CH2CH2); 13C NMR (101 Mz, CDC13) d
157.0, 156.0, 142.5, 137.8, 137.2, 137.1, 134.0, 132.5, 130.9,
129.5, 128.4, 127.9, 126.3, 123.5, 122.0, 113.6, 80.5, 65.9, 54.8,
54.6, 35.6, 23.3; HRMS (ESI) calculated for C28H30NO2 (M+):
412.2277; found: 412.2277.
5.10. 5-Bromo-8-methoxy-2,3-dihydro-benzo[b]oxepine (10d)
A mixture of 9b (0.48 g, 2.5 mmol) and phosphorus tribromide
(1.20 mL, 12.5 mmol) was heated at 90 °C under nitrogen for
22 h. The solution was cooled and was added carefully dropwise
to ice-water (25 mL). The aqueous layer was extracted with ethyl
acetate (3ꢂ 50 mL), washed with brine (20 mL), dried over sodium
sulfate, and the solvent removed under reduced pressure. The yel-
low oil, which decolorises rapidly was purified by flash chromatog-
raphy (5% diethyl ether in hexanes) to give the product 10d (0.31 g,
50%), which was used immediately in the next reaction.
5.7. 4-{5-[4-(2-Pyrrolidin-1-yl-ethoxy)-phenyl]-2,3-dihydro-
benzo[b]oxepin-4-yl}-phenol (15b)
This compound was prepared from 13a18 using a procedure
similar to that described for 11a. The crude product was purified
by flash chromatography (hexane/CH2Cl2/CH3OH; 48:48:2) to give
the product as a beige solid (0.17 g, 37%). IR tmax (film): 2962,
1606 cmꢁ1 1H NMR (400 Mz, CDCl3) d 7.21–7.14 (3H, m, ArH),
;
5.11. 4-(8-Methoxy-2,3-dihydro-benzo[b]oxepin-5-yl)-phenol
(11b)
7.06 (1H, d, J = 8.0, ArH), 6.95–6.89 (3H, m, ArH), 6.78 (1H, d,
J = 8.0, ArH), 4.57 (2H, t, J = 6.0, OCH2), 4.13 (2H, J = 6.0, OCH2),
2.99 (2H, t, J = 6.0, CH2), 2.92 (2H, t, J = 6.0, CH2), 2.65–2.62 (4H,
m, CH2NCH2), 1.83–1.79 (4H, m, CH2CH2); 13C NMR (101 Mz,
CDC13) d 157.7, 156.4, 139.3, 134.9, 133.5, 131.0, 129.4, 128.8,
123.4, 121.9, 121.8, 114.5, 113.9, 66.1, 54.8, 54.6, 41.1, 23.4; HRMS
(ESI) calculated for C28H29NO3 (M+) 414.3461; found: 427.5479.
Method A: This compound was prepared from 10d using a pro-
cedure similar to that described for 11a. The residue was purified
by flash chromatography (5% diethyl ether in hexanes) to give
the product as a colorless gel (0.62 g, 52%) which was used in the
following reaction without further purification. IR tmax (film):
3423, 2939, 1609 cmꢁ1 1H NMR (400 Mz, CDCl3) d 7.11 (2H, d,
;
5.8. 4-(3-Methoxy-phenoxy)-butyric acid (8b)19
J = 8.5, ArH), 6.90 (1H, dd, J = 6.5, 8.5 Hz, ArH), 6.80 (1H, m, ArH),
6.7 (2H, m, ArH), 6.77 (2H, d, J = 8.5 Hz, ArH), 6.72 (1H, m, ArH),
6.20 (1H, t, J = 6.3 Hz, = CH), 5.47 (1H, br s, OH), 4.50 (2H, t,
J = 6.0 Hz, OCH2CH2), 2.42 (2H, m, OCH2CH2); 13C NMR (101 Mz,
A mixture of 3-methoxyphenol (7b) (5.5 mL, 50 mmol) and
potassium carbonate (6.9 g, 55 mmol) in acetone (100 mL) was

9566
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
CDC13) d 162.2, 159.7, 157.5, 153.8, 139.0, 133.9, 130.8, 128.6,
127.8, 125.4, 114.1, 109.2, 108.1, 107.9, 107.6, 107.6, 77.8, 28.4;
19F NMR (376 MHz, CDCl3) d ꢁ113.65 ppm.
(400 Mz, CDCl3) d 7.06 (2H, d, J = 8.56 Hz, ArH), 6.86–6.83 (3H, q,
J = 6.52 Hz, ArH), 6.62 (1H, d, J = 3Hz, ArH), 6.54–6.51 (1H, q,
J = 3.84 Hz, ArH), 6.08 (1H, t, J = 6.28 Hz, @CH), 4.43 (2H, t,
J = 6.02 Hz, CH2), 4.09 (2H, t, J = 6.02 Hz, OCH2), 3.76 (3H, s, O–
CH3), 2.87 (2H, t, J = 6.02 Hz, CH2); 2.79 (2H, t, J = 7.04 HZ, CH2),
2.53 (s, 4H, (CH2)2), 1.78 (s, 4H, (CH2)2).
Method B: This compound was prepared from 9b using a proce-
dure similar to that described for 11a. The product was purified by
column chromatography (silica, 5% diethyl ether/hexane) to afford
the product (0.35 g, 44%).
5.16. 1-{2-[4-(8-Methoxy-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenoxy]-ethyl}-piperidine (12c)
5.12. 4-(4-Bromo-8-methoxy-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenol (14b)
This compound was prepared from 11b using a procedure simi-
lar to that described for 12a. The residue was purified by flash chro-
matography (55:35:10 dichloromethane/hexane/methanol) to give
This compound was prepared from 11b using a procedure sim-
ilar to that described for 14a. The residue was purified using flash
chromatography (10% diethyl ether in hexanes) to give the product
as a yellow oil (0.63 g, 63%) which was used in the following reac-
a yellow oil (292 mg, 65%). IR t
max (film): 2934, 1607 cmꢁ1; 1H NMR
(400 Mz, CDCl3) d 7.17 (2H, d, J = 8.5 Hz, ArH), 6.89–6.84 (3H, m,
ArH), 6.65 (1H, d, J = 2.5 Hz, ArH), 6.56 (1H, dd, J = 8.5, 2.5 Hz,
ArH), 6.11 (2H, t, J = 6.0 Hz, CH), 4.46 (2H, t, J = 6.0 Hz, OCH2), 4.11
(2H, t, J = 6.0 Hz, CH2), 3.79 (3H, s, CH3), 2.78 (2H, t, J = 6.0 Hz,
CH2), 2.51–2.47 (6H, m, CH2NCH2), 1.64–1.59 (4H, m, CH2CH2),
1.45 (2H, m, CH2); 13C NMR (101 Mz, CDC13) d 159.6, 159.0,
157.9, 140.4, 135.6, 131.9, 129.6, 125.6, 124.9, 114.1, 109.2, 106.7,
77.2, 65.8, 57.9, 55.3, 54.9, 30.2, 25.9, 24.1; HRMS (ESI) calculated
for C24H30NO3 380.2226 (M+); found: 380.2233.
tion without further purification. IR tmax (film): 3310, 1234 cmꢁ1
;
1H NMR (400 Mz, CDCl3) d 7.09 (2H, d, J = 8.52 Hz, ArH), 6.82 (2H,
d, J = 8.04 Hz, ArH), 6.67 (2H, d, J = 8.56 Hz, ArH), 6.62 (2H, d,
J = 2.52 Hz, ArH), 6.48–6.51 (1H, q (dd), J = 2.52 Hz, 7.04 Hz, ArH),
4.56 (2H, t, J = 5.76 Hz, CH2), 3.77 (3H, s, OCH3), 3.02 (2H, t,
J = 5.76 Hz, CH2); 13C NMR (101 Mz, CDC13) d 159.8, 157.5, 154.8,
138.9, 135.0, 132.4, 131.2, 125.4, 120.3, 114.9, 109.7, 106.8, 77.3,
55.3, 41.3; 19F NMR (376 MHz, CDCl3) d ꢁ113.65.
5.17. 1-{2-[4-(4-Bromo-8-methoxy-2,3-dihydro-benzo[b]xepin-
5-yl)-phenoxy]-ethyl}-piperidine (13c)
5.13. 1-{2-[4-(4-Bromo-8-methoxy-2,3-dihydrobenzo[b]oxepin-
5-yl)-phenoxy]-ethyl}-pyrrolidine (13b)
Method A: This compound was prepared from 12c using a pro-
cedure similar to that described for 14a. The residue was purified
using flash chromatography (10% diethyl ether in hexanes) to give
the product as a yellow oil (0.81 g, 89%).
This compound was prepared from 12b using a procedure sim-
ilar to that described for 12a. Flash chromatography (1% methanol
in dichloromethane) afforded the product as a yellow oil, (0.77 g).
IR tmax (film): 2934, 1607 cmꢁ1 1H NMR (400 Mz, CDCl3) d 6.91
;
Method B: This compound was prepared from 14b using a pro-
cedure similar to that described for 12a. The residue was purified
by flash chromatography (silica, 55:35:10 dichloromethane/hex-
ane/methanol) to give a yellow oil (187 mg, 52%). IR tmax (film):
(2H, d, J = 6.0, ArH), 6.89 (2H, d, J = 6.0 Hz, ArH), 6.61 (1H, s, ArH),
6.50 (1H, d, J = 6.0 Hz, ArH), 6.47 (1H, d, J = 6.0 Hz, ArH), 4.55 (2H,
t, J = 6.0 Hz, OCH2), 4.17 (2H, t, J = 6.0 Hz, CH2), 3.77 (3H, s, CH3),
3.04 (2H, t, J = 6.0 Hz, CH2), 3.01 (2H, t, J = 6.0, 6.0 Hz, CH2), 2.97
(4H, m, CH2NCH2), 1.84 (4H, m, CH2CH2); 13C NMR (101 Mz,
CDC13) d 157.8, 157.8, 135.2, 132.4, 130.9, 125.2, 120.4, 113.9,
109.5, 106.7, 76.5, 66.5, 56.2, 55.3, 56.4, 41.5, 23.4; HRMS (ESI) cal-
culated for C23H26BrNO3 444.3726 (M+); found: 444.3723.
; d 7.13 (2H, d,
2933, 1606 cmꢁ1 1H NMR (400 Mz, CDCl3)
J = 8.5 Hz, ArH), 6.89 (3H, m, ArH), 6.67–6.60 (2H, m, ArH), 6.48
(1H, dd, J = 8.0, 2.5 Hz, ArH), 4.54 (2H, t, J = 6.0, OCH2), 4.14 (2H,
t, J = 6.0 Hz, OCH2), 3.76 (3H, s, CH3), 3.04 (2H, t, J = 6.0 Hz, CH2),
2.81 (2H, t, J = 6.0 Hz, CH2), 2.54 (4H, m, CH2NCH2), 1.66–1.60
(4H, m, CH2CH2), 1.46 (2H, m, CH2); 13C NMR (101 Mz, CDC13) d
159.8, 157.9, 156.8, 155.7, 138.9, 135.1, 132.3, 130.9, 129.1,
125.2, 120.3, 113.9, 109.5, 106.7, 65.6, 57.8, 56.2, 55.2, 54.9, 41.4,
25.7, 24.0; HRMS (ESI) calculated for C24H28BrNO3 (M+)
458.3997; found: 458.4001.
5.14. 4-{8-Methoxy-5-[4-(2-pyrrolidin-1-yl-ethoxy)-phenyl]-
2,3-dihydro-benzo[b]oxepin-4-yl}-phenol (15c)
This compound was prepared from 13b using a procedure sim-
ilar to that described for 11a. The residue was purified by flash
chromatography (10% methanol in dichloromethane) to give the
product as a brown solid (0.31 g, 39%). HPLC (tr = 2.354 min;
5.18. 4-{8-Methoxy-5-[4-(2-piperidin-1-yl-ethoxy)-phenyl]-2,3-
dihydro-benzo[b]oxepin-4-yl}-phenol (15d)
93.2%); IR tmax (film): 2934, 1607 cmꢁ1 1H NMR (400 Mz, CDCl3)
;
d 6.95 (2H, d, J = 6.0 Hz, ArH), 6.85 (2H, d, ArH), 7.01 (4H, m,
ArH), 6.57 (1H, d, J = 6.0 Hz, ArH), 6.49 (2H, d, J = 6.0 Hz, ArH),
4.59 (2H, t, J = 6.0 Hz, OCH2), 4.08 (2H, t, J = 6.0 Hz, CH2), 3.80
(3H, s, CH3), 3.26 (2H, br t, CH2), 3.14 (4H, m, CH2–CH2), 2.67 (2H,
br t, CH2), 1.97 (4H, m, CH2CH2); 13C NMR (101 Mz, CDC13) d
157.8, 157.3, 139.2, 133.5, 132.4, 130.9, 125.2, 120.4, 113.9,
109.5, 106.7, 66.1, 55.3, 54.8, 54.6, 41.1, 23.4; HRMS (ESI) calcu-
lated for C29H32NO4 458.2331 (M+); found: 458.2338.
This compound was prepared from 13c using a procedure similar
to that described for 11a. The crude product was purified by flash
chromatography (hexane/diethyl ether 1:1) to give the product as
a yellow solid (0.25 g, 30%). HPLC (tr = 2.667 min; 88.1%); IR tmax
(KBr) 2970, 1609 cmꢁ1; 1H NMR (400 Mz, DMSO-d) d 6.83 (10H, m,
ArH), 6.53 (2H, d, J = 6.0, ArH), 4.60 (2H, t, J = 6.0, CH2), 4.33 (2H, t,
J = 6.0 Hz, CH2), 3.55 (2H, t, J = 6.0 Hz, CH2), 3.30 (4H, m, CH2–CH2),
2.72 (2H, t, J = 6.0 Hz, CH2), 1.67 (4H, m, CH2–CH2); 13C NMR
(101 Mz, CDC13) d 156.8, 155.8, 155.4, 138.1, 134.9, 132.9, 131.1,
129.9, 114.9, 113.9, 66.1, 55.3, 55.4, 45.5, 25.8, 23.2; HRMS (ESI) cal-
culated for C30H34NO4 (M+) 472.2480; found: 472.2490.
5.15. 1-{2-[4-(8-Methoxy-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenoxy]-ethyl}-pyrrolidine (12b)
This compound was prepared from 11b using a procedure sim-
ilar to that described for 12a. The crude product was purified by
flash chromatography (2.5% methanol/CH2Cl2) to give the product
as an oil which was used in the subsequent reaction without fur-
5.19. 5-[4-(2-Pyrrolidin-1-yl-ethoxy)-phenyl]-2,3-dihydro-
benzo[b]oxepin-8-ol (16)
A mixture of pyridinium hydrochloride (1.41 g, 12.2 mmol) and
12b (0.89 g, 2.44 mmol) was heated at 180–190 °C for 3 h. The hot
ther purification (0.45 g, 95%). IR tmax (KBr) cmꢁ1 1H NMR
;

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9567
mixture was poured into a 5% aqueous solution of sodium carbon-
ate (40 mL) and extracted into ethyl acetate (5ꢂ 100 mL) and
methanol (20 mL). The organic extracts were combined, washed
with water, and the solvent removed under reduced pressure.
The residue was purified by flash chromatography (2.5% metha-
nol/dichlorormethane) to give a brown oil (111 mg, 13%). 1H
NMR (400 Mz, CDCl3) d 7.10 (2H, d, J = 8.5 Hz, ArH), 6.74–6.67
(4H, m, ArH), 6.44 (2H, dd, J = 8.5, 2.5 Hz, ArH), 6.02 (1H, t,
J = 6.0 Hz, ArH), 4.40 (2H, t, J = 6.0 Hz, CH2), 4.12 (2H, t, J = 5.5 Hz,
CH2), 3.03 (2H, t, J = 5.5 Hz, CH2), 2.82 (4H, m, CH2NCH2), 2.46
(2H, t, J = 6.0 Hz, CH2), 1.89 (4H, m, CH2CH2); 13C NMR (101 Mz,
CDC13) d 159.0, 157.4, 157.3, 140.5, 135.9, 131.9, 129.5, 125.1,
123.8, 113.9, 111.0, 108.9, 65.6, 54.8, 54.4, 30.2, 23.2; HRMS (ESI)
calculated for C22H26NO3 (M+) 352.1913; found: 352.1914.
(101 Mz, CDC13) d 179.2, 164.8, 162..4, 160.4, 159.9, 130.2, 110.2,
107.5, 102.2, 66.7, 30.5, 24.2; 19F NMR (376 MHz, CDCl3) d
ꢁ112.1; HRMS (ESI) calculated for C10H11O3NaF (M+Na+)
221.2042; found: 221.0484.
5.23. 8-Fluoro-3,4-dihydro-2H-benzo[b]oxepin-5-one (21)
A mixture of 20 (4.75 g, 27.8 mmol) and polyphosphoric acid
(51 g) was heated at 80 °C for 3 h and then poured into ice-water.
The product was extracted with ethyl acetate (4ꢂ 100 mL) and the
combined organic layers were washed with 10% sodium hydroxide
(100 mL), brine (100 mL), dried over sodium sulfate, and the sol-
vent removed under reduced pressure. The crude product was
purified by column chromatography (silica, 5% diethyl ether/hex-
ane) to give the product as a yellow oil (1.56 g, 31%).26 IR tmax
(film): 1685; 1H NMR (400 Mz, CDCl3) d 7.85 (1H, m, ArH), 6.82
(2H, m, ArH), 4.26 (2H, t, avg. J = 6.8 Hz, OCH2CH2CH2), 2.90 (2H,
t, avg. J = 7.0 Hz, OCH2CH2CH2), 2.22 (2H, m, OCH2CH2CH2); 13C
NMR (101 Mz, CDC13) d 198.9, 166.9, 164.3, 163.8, 131.5, 125.6,
110.3, 107.6, 73.2, 40.4, 26.1; 19F NMR (376 MHz, CDCl3) d
ꢁ105.3 ppm; HRMS (ESI) calculated for C10H9O2NaF (M+Na+)
203.18; found: 203.0479.
5.20. 4-Bromo-5-[4-(2-pyrrolidin-1-yl-ethoxy)-phenyl]-2,3-
dihydro-benzo[b]oxepin-8-ol (17)
This compound was prepared from 16 using a procedure similar
to that described for 14a. The residue was purified using flash chro-
matography (10% diethyl ether in hexanes) to give the product as a
yellow oil (84 mg, 64%). 1H NMR (400 Mz, CDCl3) d 7.01 (2H, d,
J = 8.0 Hz, ArH), 6.86 (2H, d, J = 8.0 Hz, ArH), 6.52–6.42 (2H, m,
ArH), 6.30 (1H, d, J = 6.5 Hz, ArH), 4.41 (2H, t, J = 6.0 Hz, CH2),
4.11 (2H, br t, CH2), 3.05 (2H, t, CH2), 2.93 (2H, br t, CH2), 2.92–
2.79 (4H, m, CH2NCH2), 1.84 (4H, m, CH2CH2); 13C NMR (101 Mz,
CDC13) d 157.7, 156.4, 139.3, 134.9, 133.5, 131.0, 129.4, 128.8,
123.4, 121.9, 121.8, 114.5, 113.9, 66.1, 54.8, 54.6, 41.1, 23.4; HRMS
(ESI) calculated for C22H24BrNO3 (M+) 458.2331; found: 430.3450.
5.24. 4-(8-Fluoro-2,3-dihydro-benzo[b]oxepin-5-yl)-phenol (23)
This compound was prepared from 21 and 22 using a procedure
similar to that described for 11a, Method B: The product was puri-
fied by flash chromatography (5% diethyl ether/hexanes) to give
the product was a yellow oil (1.50 g, 67%, 96%) which was used
in the following reaction without further purification. IR tmax
(film): 3310 cmꢁ1
; d 7.11 (2H, d,
1H NMR (400 Mz, CDCl3)
5.21. 4-(4-Hydroxyphenyl)-5-[4-(2-pyrrolidin-1-yl-ethoxy)-
phenyl]-2,3-dihydro-benzo[b]oxepin-8-ol (18)
J = 8.5 Hz, ArH), 6.90 (1H, dd, J = 6.5, 8.5 Hz, ArH), 6.80 (1H, m,
ArH), 6.70 (2H, m, ArH), 6.77 (2H, d, J = 8.5 Hz, ArH), 6.72 (1H, m,
ArH), 6.20 (1H, t, J = 6.3 Hz, @CH), 5.47 (1H, br s, OH), 4.50 (2H, t,
J = 6.0 Hz, OCH2CH2), 2.42 (2H, m, OCH2CH2); 13C NMR (101 Mz,
CDC13) d 162.2, 159.7, 157.5, 153.8, 139.0, 133.9, 130.8, 128.6,
127.8, 125.4, 114.1, 109.2, 108.1, 107.9, 107.6, 107.6, 77.8, 28.4;
19F NMR (376 MHz, CDCl3) d: ꢁ113.65.
This compound was prepared from 17 using a procedure similar
to that described for 11a. The residue was purified by flash chro-
matography (5% diethyl ether in hexanes) to give the product as
a beige solid (11 mg, 12%), Mp 148 °C. HPLC (tr = 2.404 min;
87%); 1H (400 MHz, DMSO-d, Me4Si): 6.95 (2H, d, J = 6.5, ArH),
6.89 (2H, d, J = 6.5 Hz, ArH), 6.76 (2H, d, J = 8.5 Hz, ArH), 6.67–
6.54 (4H, m, ArH), 6.47 (1H, dd, J = 8.5, 2.5 Hz, CH), 4.53 (2H, t,
J = 6.0, CH2), 4.22 (2H, t, J = 5.0 Hz, CH2), 3.45 (2H, t, J = 5 Hz,
CH2), 3.36–3.27 (6H, m, CH2NCH2), 2.68 (2H, t, J = 6.0 Hz, CH2),
2.05 (4H, m, CH2CH2); 13C NMR (101 Mz, CDC13) d 157.7, 156.4,
139.3, 134.9, 133.5, 131.0, 129.4, 128.8, 123.4, 121.9, 121.8,
114.5, 113.9, 66.1, 54.8, 54.6, 34.9, 21.9; HRMS (ESI) calculated
for C28H30NO4 (M+) 444.2175; found: 444.2194.
5.25. 4-(4-Bromo-8-fluoro-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenol (24a)
This compound was prepared from 23 using a procedure similar
to that described for 14a The crude product purified by flash chro-
matography (5% diethyl ether/hexanes) to give the pure product as
a yellow oil (1.79 g, 91%). IR tmax (film) 3402 cmꢁ1 1H NMR
;
(400 Mz, CDCl3) d 7.09 (2H, d, J = 9.0 Hz, ArH), 6.84–6.73 (4H, m,
ArH), 6.68–6.63 (1H, m, ArH), 4.59 (2H, t, J = 6.0 Hz, OCH2), 3.00
(2H, t, J = 6.0 Hz, CH2); 13C NMR (101 Mz, CDC13) d 163.5, 161.0,
157.4, 151.6, 138.6, 134.7, 132.5, 130.9, 129.3, 121.5, 115.0,
110.8, 110.5, 109.4, 109.2, 77.4, 41.0; 19F NMR (376 MHz, CDCl3)
5.22. 4-(3-Fluoro-phenoxy)-butyric acid (20)
To a suspension of potassium carbonate (7.59 g, 55 mmol) in
acetone (100 mL) was added 3-fluorophenol (4.5 mL, 50 mmol)
and the mixture stirred under nitrogen for 15 min. Ethyl-4-bromo-
butyrate (7.2 mL, 50 mmol) was added and the mixture refluxed
for 6 h, cooled to ambient then filtered. The filtrate was concen-
trated under reduced pressure and the residue taken up in diethyl
ether (100 mL). The organic layer was washed with 10% sodium
hydroxide solution (50 mL), dried over sodium sulfate, and the sol-
vent removed under reduced pressure. The oily residue refluxed in
10% sodium hydroxide (100 mL) and ethanol (20 mL) until the
solution became clear. The solution was cooled, acidified with con-
cd hydrochloric acid, and beige crystals were collected by filtration
d
ꢁ112.47; HRMS (ESI) calculated for C16H12FO2BrNa (M+)
359.1650; found: 359.0339.
5.26. 1-{2-[4-(4-Bromo-8-fluoro-2,3-dihydro-benzo[b]oxepin-
5-yl)-phenoxy]-ethyl}-pyrrolidine (25)
This compound was prepared from 24a using a procedure sim-
ilar to that described for 12a. The crude product was purified by
flash chromatography (2.5% methanol/dichloromethane) to give
the product as a yellow oil (0.48 g, 81%). 1H NMR (400 Mz, CDCl3)
d 7.13 (2H, d, J = 8.5 Hz, ArH), 6.79 (2H, d, J = 8.5 Hz, ArH), 6.68
(1H, m, ArH), 4.58 (2H, t, avg. J = 5.8 Hz, OCH2CH2N–), 4.13 (2H, t,
avg. J = 5.8 Hz, OCH2CH2), 3.01 (2H, t, J = 6.0 Hz, OCH2CH2N–),
2.90 (2H, J = 6.0 Hz, OCH2CH2), 2.63 (4H, m, CH2NCH2), 1.80, (4H,
m, CH2–CH2), 13C NMR (101 Mz, CDC13) d 163.4, 160.9, 158.1,
(4.75 g, 79%), mp 42–45 °C.26 IR tmax (KBr): 3436, 1713 cmꢁ1 1H
;
NMR (400 Mz, CDCl3) d 7.87 (1H, br s, OH), 7.22 (1H, m, ArH),
6.61 (3H, m, ArH), 4.00 (2H, t, J = 6.0 Hz, OCH2CH2CH2), 2.58 (2H,
t, avg. J = 7.3, OCH2CH2CH2), 2.10 (2H, m, OCH2CH2CH2); 13C NMR

9568
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
157.6, 138.6, 134.6, 132.6, 130.9, 121.4, 114.0, 110.6, 110.4, 109.3,
109.0, 77.3, 66.8, 54.9, 54.6, 41.0, 23.5; 19F NMR (376 MHz, CDCl3) d
ꢁ112.66; HRMS (ESI) calculated for C22H24NO3FBr (M+) 432.0974;
found: 432.0958.
matography (1:10:30 methanol/dichloromethane/hexanes) to give
the product as a beige solid (82 mg, 40%). HPLC (tr = 3.037min;
93.9%); IR tmax (KBr) 2970, 1609 cmꢁ1 1H NMR (400 Mz, CDCl3)
;
d 6.96 (5H, m, ArH), 6.86–6.67 (4H, m, ArH), 6.56 (1H, m, ArH),
4.60 (2H, d, J = 6.2 Hz, CH2), 4.20 (2H, t, J = 6 Hz avg., OCH2CH2),
(2H, t, J = 6 Hz avg., OCH2CH2), 3.22 (2H, t, 4ꢂ CH2NCH2, 2ꢂ
CH2NCH2), 2.68 (4H, m, CH2–CH2); 13C NMR (101 Mz, CDC13) d
165, 157.5, 157.1, 142.3, 137.5, 136.5, 133.5, 132.9, 132.3, 132.0,
129.3, 127.9, 127.5, 126.3, 113.7, 110.7, 110.5, 109.4, 109.2 (dd,
C–F), 80.7, 66.8, 55.0, 54.7, 35.5, 23.4; 19F NMR (376 MHz, CDCl3)
d 113.45; HRMS (ESI) calculated for C28H29NO2F (M+) 430.2181;
found: 430.2173.
5.27. 4-{8-Fluoro-5-[4-(2-pyrrolidin-1-yl-ethoxy)-phenyl]-2,3-
dihydro-benzo[b]oxepin-4-yl}-phenol (26b)
This compound was prepared from 25 using a procedure similar
to that described for 11a. The product was purified by flash chro-
matography (2.5% methanol/dichloromethane) to give the product
as a foamy yellow solid (152 mg, 68%), mp 112–120 °C. HPLC
(tr = 2.499 min; 97.6%); IR tmax (KBr) 3431 (OH) cmꢁ1 1H NMR
;
(400 Mz, CDCl3) d 6.96 (2H, d, J = 8.5 Hz, ArH), 6.86–6.67 (7H, m,
ArH), 6.56 (2H, d, J = 9.0 Hz, ArH), 4.60 (2H, t, avg. J = 6.2 Hz,
OCH2CH2N–), 4.20 (2H, t, J = avg. 5.0 Hz, OCH2CH2), 3.22 (6H, m,
4ꢂ CH2NCH2, 2ꢂ CH2NCH2), 2.68 (4H, m, CH2–CH2); 13C NMR
(101 Mz, CDC13) d 157.1, 156.9, 156.0, 155.4, 138.2, 135.1, 133.6,
132.5, 131.8, 130.6, 115.2, 113.7, 110.8, 110.5, 109.5, 109.3, 80.9,
63.9, 54.4, 54.1, 35.6, 23.1; 19F NMR (376 MHz, CDCl3) d ꢁ113.7;
HRMS (ESI) calculated for C28H29NO3F (M+) 446.2131; found:
446.2114.
5.31. 1-(2-{4-[8-Fluoro-4-(4-methanesulfonyl-phenyl)-2,3-
dihydro-benzo[b]oxepin-5-yl]-phenoxy}-ethyl)-pyrrolidine (26e)
This compound was prepared from 25 using a procedure sim-
ilar to that described for 11a. The product was purified by flash
chromatography (5% diethyl ether/hexanes) to give the product
as a white solid (66 mg, 22%). 1H NMR (400 Mz, CDCl3) d 7.74
(2H, d, J = 8.5 Hz, ArH), 6.88–6.78 (5H, m, ArH), 6.67 (2H, d,
J = 8.9 Hz, ArH), 4.60 (2H, t, J = 6.0, OCH2), 4.06 (2H, t, J = 6.0 Hz,
OCH2CH2), 3.04 (3H, s, CH3), 2.91 (2H, t, J = 6.0 Hz, OCH2CH2),
2.73 (2H, t, J = 6.0 Hz, OCH2CH2), 2.71–2.65 (4H, m, CH2NCH2),
1.82 (4H, m, CH2–CH2); 13C NMR (101 Mz, CDC13) d 163.9,
161.3, 157.8, 157.1, 149.5, 139.2, 135.9, 132.5, 132.2, 130.3,
127.1, 114.0, 111.0, 110.8, 109.6, 109.4 (dd, C–F), 80.5, 66.3,
5.28. 3-{8-Fluoro-5-[4-(2-pyrrolidin-1-yl-ethoxy)-phenyl]-2,3-
dihydro-benzo[b]oxepin-4-yl}-phenol (26c)
This compound was prepared from 25 using a procedure similar
to that described for 11a. The product was purified by flash chro-
matography (2.5% methanol/dichloromethane) to give the product
as brown crystals, mp 105 °C. HPLC (tr = 2.489 min; 93.7%); IR tmax
54.8, 54.6, 44.4, 35.3, 23.4; 19F NMR (376 MHz, CDCl3)
d
ꢁ112.27; HRMS (ESI) calculated for C29H30NO4FS (M+) 507.6292;
found: 507.6100.
(KBr) 3435, 2922, 1607 cmꢁ1 1H NMR (400 Mz, CDCl3) d 7.02 (1H,
;
5.32. 1-{2-[4-(4-Bromo-8-fluoro-2,3-dihydro-benzo[b]oxepin-
5-yl)-phenoxy]-ethyl}-piperidine (27)
t, J = 7.78 Hz, ArH), 6.83 (4H, m,ArH), 6.72 (1H, m, ArH), 6.66 (1H, d,
J = 7.52 Hz, ArH), 6.60 (1H, d, J = 8.52 Hz, ArH), 6.55–6.53 (3H, m,
ArH), 4.76 (1H, br s, OH), 4.59 (2H, t, J = 5.78 Hz, OCH2), 4.06 (2H,
t, J = 4.78 Hz, OCH2), 3.04 (2H, t, J = 5.04 Hz, OCH2), 2.90 (4H, s,
CH2–N–CH2), 2.67 (2H, t, J = 5.78 Hz, CH2); 1.90 (4H, s, CH2–CH2);
13C NMR (101 Mz, CDC13) d 163.4, 160.9, 157.2, 157.0, 156.3,
143.8, 137.9, 136.2, 134.2, 132.9, 132.1, 131.9, 129.21, 120.8,
116.8, 113.9, 113.7, 110.5, 109.3, 80.8, 64.9, 54.5, 54.4, 35.6,
23.2,; 19F NMR (376 MHz, CDCl3) d ꢁ112.9; HRMS (ESI) calculated
for C28H28NO3F (M+) 446.2131; found: 446.2130.
This compound was prepared from 24a using a procedure sim-
ilar to that described for 12a. The product was purified by flash
chromatography (5% diethyl ether/hexanes) to give the product
as an orange gel. IR tmax (film) cmꢁ1 1H NMR (400 Mz, CDCl3) d
;
7.66 (2H, d, J = 8.56 Hz, ArH), 6.91 (2H, d, J = 8.52 Hz, ArH), 6.77–
6.66 (2H, m, ArH), 6.64–6.62 (1H, m, ArH), 4.59 (2H, t, J = 5.78 Hz,
OCH2), 4.12 (2H, t, J = 6.02 Hz, OCH2), 3.01 (2H, t, J = 5.76 Hz,
OCH2), 2.79 (2H, t, J = 6.02 Hz, CH2); 2.53 (4H, s, CH2–N–CH2),
1.62 (6H, s, CH2–CH2–CH2); 13C NMR (101 Mz, CDC13) d 163.1,
160.6, 157.7, 138.3, 134.2, 132.1, 132.0, 131.8, 129.6, 128.9,
120.9, 113.7, 110.3, 110.0, 108.9, 108.7, 76.9, 65.4, 57.5, 57.2,
54.6, 40.6, 25.6, 23.8; HRMS (ESI) calculated for C23H25BrFNO2
(M+) 446.1131; found: 446.1129.
5.29. 1-(2-{4-[8-Fluoro-4-(4-fluorophenyl)-2,3-dihydro-benzo[b]-
oxepin-5-yl]-phenoxy}-ethyl)-pyrrolidine (26d)
This compound was prepared from 25 using a procedure similar
to that described for 11a. The product was purified by flash chro-
matography (2.5% methanol/dichloromethane) to give the product
5.33. 4-{8-Fluoro-5-[4-(2-piperidin-1-yl-ethoxy)-phenyl]-2,3-
dihydro-benzo[b]oxepin-4-yl}-phenol (28)
as a yellow solid (67 mg, 31%), mp 130–135 °C. IR
tmax (KBr) 3431
(OH) cmꢁ1 1H NMR (400 Mz, CDCl3) d 7.12 (2H, m, ArH) 6.85 (6H,
;
m, ArH), 6.73 (1H, m, ArH), 6.65 (2H, d, J = 8.0 Hz, ArH), 4.60 (2H, t,
J = 6.0, OCH2CH2N), 4.20 (2H, t, J = 6.0 Hz, CH2), 3.19 (2H, t,
J = 6.0 Hz, CH2), 2.94 (4H, m, CH2NCH2) 2.68 (2H, t, J = 6.0 Hz,
CH2), 1.93 (4H, m, CH2CH2); 13C NMR (101 Mz, CDC13) d 157.5,
138.2, 136.8, 136.5, 133.4, 132.4, 132.0, 131.9, 131.0, 130.9,
115.0, 114.8, 113.8, 110.8, 110.6, 109.5, 109.3, 80.7, 66.7, 55.0,
54.7, 35.6, 23.4; 19F NMR (376 MHz, CDCl3) d ꢁ113.3, 116.2; HRMS
(ESI) calculated for C28H28NO2F2 (M+) 448.2088; found: 448.2072.
This compound was prepared from 27 using a procedure similar
to that described for 11a. The residue was purified by flash chro-
matography (5% diethyl ether in hexanes) to give the product as
a beige solid (0.84 g, 84%), mp 188 °C. HPLC (tr = 2.614 min;
96.2%); IR tmax (KBr) 2970, 1609 cmꢁ1 1H NMR (400 Mz, CDCl3)
;
d
1H NMR (400 MHz, acetone-d) d 6.83 (10H, m, ArH), 6.53 (2H,
d, J = 6.0 Hz, ArH), 4.60 (2H, t, J = 6.0 Hz, CH2), 4.33 (2H, t,
J = 6.0 Hz, CH2), 3.55 (2H, t, J = 6.0 Hz, CH2), 3.30 (4H, m, CH2–
CH2), 2.72 (2H, t, J = 6.0 Hz, CH2), 1.91 (4H, m, CH2 CH2); 13C NMR
(101 Mz, CDC13) d 156.8, 155.8, 155.4, 138.1, 134.9, 132.9, 131.1,
129.9, 114.9, 113.9, 109.7, 109.7, 108.5, 108.2 (dd, C–F), 80.1,
66.1, 55.3, 52.9, 34.9, 22.2, 20.7; 19F (376 MHz, CDCl3) d ꢁ115.8;
HRMS (ESI) calculated for C28H29NO2F (M+) 430.2181; found:
430.2173.
5.30. 1-{2-[4-(8-Fluoro-4-phenyl-2,3-dihydro-benzo[b]oxepin-
5-yl)-phenoxy]-ethyl}-pyrrolidine (26a)
This compound was prepared from 25 using a procedure similar
to that described for 11a. The product was purified by flash chro-

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9569
5.34. 4-(4-Bromo-8-fluoro-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenoxy]-tert-butyldimethyl- silane (29)
2.71 (2H, t, J = 6.04 Hz, CH2), 4.61 (2H, t, J = 6.02 Hz, CH2), 6.66–
7.71 (12H, m, ArH); 13C NMR (101 Mz, CDC13) d 34.7, 79.8, 114.7,
121.6, 127.1, 128.0, 130.7, 130.9, 131.51, 132.92, 136.39, 139.67,
146.45, 149.65; MS (EI) m/z 315 (M+).
To a mixture of 24a (8.05 mmol) and imidazole (10.5 mmol)
were dissolved in DMF (10 mL). tert-Butyldimethylsilyl chloride
(9.66 mmol) was added in six portions over 4 h. Stirring was con-
tinued at room temperature for a further 10 h. The reaction mix-
ture was diluted with ethyl acetate (150 mL) and quenched with
10% HCl (20 mL). The organic layer was separated, washed with
water (40 mL) and brine (40 mL) and dried over Na2SO4. The sol-
vent was removed under reduced pressure and the residue was
chromatographed on silica gel (hexane/diethyl ether = 20:1) to
yield the product 8 (1.54 g, 69%) as a brown solid, (69%) which
was used in the following reaction without further purification.
1H NMR (400 Mz, CDCl3) d 7.06 (2H, d, J = 8.52 Hz, ArH), 6.84 (2H,
d, J = 10.52 Hz, ArH), 6.80–6.73 (4H, m, ArH), 6.71–6.62 (1H, m,
ArH), 4.58 (2H, t, J = 6.02 Hz, OCH2), 3.01 (2H, t, J = 6.02 Hz, CH2);
0.95 (9H, s, C-(CH3)3), 0.23 (6H, s, Si-(CH3)2), 13C NMR (101 Mz,
CDC13) d 163.1, 160.6, 157.1, 138.4, 134.7, 132.3, 130.6, 129.0,
120.9, 119.2, 110.3, 108.9, 76.9, 40.7, 25.3, 20.3, 17.8, ꢁ4.8; 19F
NMR (376 MHz, CDCl3) d ꢁ112.7;
5.38. 4-(4-Furan-2-yl-2,3-dihydro-benzo[b]oxepin-5-yl)-phenol
(35b)
This compound was prepared from 34 using a procedure similar
to that described for 11a. The product was purified by flash chro-
matography (C6H12/CH2Cl2/EtOAc; 55.45:5) to afford an oil (81%).
This compound was used in subsequent reactions without further
purification. IRmmax (film) 3429, 2953–2929, 2853, 1601 cmꢁ1 1H
;
NMR (400 Mz, CDCl3) d 2.86 (2H, t, J = 6.04 Hz, CH2), 4.70 (2H, t,
J = 6.02 Hz, CH2), 5.74, (1H, m, ArH), 6.25 (1H, m, ArH), 6.70–7.24
(9H, m, ArH); 13C NMR (101 Mz, CDC13) d 35.7, 80.5, 109.2,
110.5, 113.0, 114.9, 120.0, 121.7, 123.6, 127.9, 128.0, 129.9,
131.1, 133.9, 136.6, 140.3, 141.4; MS (EI) m/z 304 (M+).
5.39. 4-(4-Thiophen-2-yl-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenol (35c)
5.35. [4-{5-(4-tert-Butyldimethylsilanyloxy)-phenyl]-8-fluoro-
2,3-dihydro-benzo[b]oxepin-5-yl}-phenol (30)
This compound was prepared from 34 using a procedure similar
to that described for 11a. The product was purified by flash chroma-
tography (C6H12/CH2Cl2/EtOAc; 30:60:10) to afford an oil (64%). This
compound was used in subsequent reactions without further purifi-
This compound was prepared from 29 using a procedure similar
to that described for 11a. The crude product was purified by column
chromatography (5% diethyl ether/hexanes) to give 4-(2,3-dihydro-
benzo[b]oxepin-5-yl)-phenol as a yellow solid (52%) which was
used in the following reaction without further purification. 1H
NMR (400 Mz, CDCl3) d 6.99 (2H, d, J = 8.52 Hz, ArH), 6.85–6.78
(4H, m, ArH), 6.76–6.71 (1H, m, ArH), 6.62–6.58 (4H, m, ArH),
4.61 (2H, t, J = 5.78 Hz, OCH2), 2.68 (2H, t, J = 6.02 Hz, CH2); 0.95
(9H, s, C-(CH3)3), 0.15 (6H, s, Si-(CH3)2), 13C NMR (101 Mz, CDC13)
d 162.8, 160.4, 156.4, 153.8, 136.8, 134.1, 134.0, 132.8, 131.9,
131.5, 131.4, 130.4, 130.3, 119.0, 114.4, 110.4, 109.1, 108.8, 80.5,
35.0, 25.2, 17.8, ꢁ4.9; 19F NMR (376 MHz, CDCl3) d ꢁ113.7;
cation. IR m
max (film) 3397, 2926, 2882–2854, 2227, 1617 cmꢁ1; 1H
NMR (400 Mz, CDCl3) d 2.73 (2H, t, J = 6.02 Hz, CH2), 4.65 (2H, t,
J = 6.26 Hz, CH2), 6.66–7.24 (11H, m, ArH); 13C NMR (101 Mz,
CDC13) d 34.8, 80.2, 114.4, 121.3, 121.6, 123.8, 125.4, 127.9,
128.2, 128.6, 129.7, 130.4, 131.3, 132.0, 136.8, 142.1; MS (EI) m/z
320 (M+).
5.40. 4-(4-Benzofuran-2-yl-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenol (35d)
This compound was prepared from 34 using a procedure similar
to that described for 11a. Purification by flash chromatography
(C6H12/CH2Cl2/EtOAc; 55:45:5) afforded the product as an oil
(46%) which was used without further purification. IR mmax (film)
5.36. [4-{8-Fluoro-4-[4-(2-pyrrolidin-1-yl-ethoxy)-phenyl]-2,3-
dihydro-benzo[b]oxepin-5-yl}-phenol (32)
3408, 2926, 2853, 1603 cmꢁ1 1H NMR (400 Mz, CDCl3) d 2.93
;
This compound was prepared from 30 using a procedure similar
to that described for 12a. The crude product was purified by flash
chromatography (2.5% methanol/dichloromethane) to give the
deprotected product 32 as yellow crystals (81%). HPLC
(2H, t, J = 6.02 Hz, CH2), 4.72 (2H, t, J = 6.26 Hz, CH2), 6.16 (1H, s,
CH), 6.79–7.50 (12H, m, ArH); 13C NMR (101 Mz, CDC13) d 41.0,
81.1, 110.7, 114.9, 115.6, 120.7, 121.6, 123.14, 124.0, 125.6,
128.5, 129.2, 131.0, 132.0, 155.14, 155.7; MS (EI) m/z 354 (M+).
(tr = 2.481 min; 91.9%); IR tmax (KBr) 3445, 2927, 1608 cmꢁ1 1H
;
NMR (400 Mz, CDCl3) d 7.01 (1H, d, J = 8.56 Hz, ArH), 6.94 (1H, d,
J = 8.52 Hz, ArH), 6.85–6.80 (3H, m, ArH), 6.76–6.71 (2H, m, ArH),
6.61–6.47 (4H, m, ArH), 4.61 (2H, t, J = 5.26 Hz, OCH2), 4.08-4.03
(2H, m, OCH2), 3.01–2.98 (2H, m, OCH2), 2.79 (4H, s, CH2–N–CH2),
2.67 (2H, t, J = 5.02 Hz, CH2); 1.87 (4H, s, CH2–CH2); 13C NMR
(101 Mz, CDC13) d 162.6, 160.1, 156.4, 153.1, 135.6, 134.6, 133.9,
132.1, 131.9, 131.6, 131.5, 130.3, 130.1, 114.9, 113.4, 110.3,
108.8, 80.6, 65.0, 54.4, 53.9, 35.1, 22.8; 19F NMR (376 MHz, CDCl3)
5.41. 4-(4-Naphthalen-2-yl-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenol (35e)
This compound was prepared from 34 using a procedure similar
to that described for 11a. The product was purified by flash chro-
matography (C6H12/CH2Cl2/EtOAc; 60:40:2) to afford an oil (50%)
which was used without further purification. IR mmax (film) 3397,
d
ꢁ113.9; HRMS (ESI) calculated for C28H28NO3F (M+) 446.2131;
3080–2924, 2852, 1601 cmꢁ1 1H NMR (400 Mz, CDCl3) d 2.64
;
found: 446.2137.
(2H, m, CH2), 4.64 (2H, m, CH2), 6.40–6.42 (2H, d, J = 8.56 Hz,
ArH), 6.80–7.71 (13H, m, ArH); 13C NMR (101 Mz, CDC13) d 35.7,
79.7, 113.9, 114.6, 121.6, 123.1, 124.9, 125.5, 126.1, 127.9128.1,
131.0, 132.2133.3; MS (EI) m/z 364 (M+).
5.37. 4-(4-Pyridin-3-yl-2,3-dihydro-benzo[b]oxepin-5-yl)-phenol
(35a)
This compound was prepared from 4-(4-bromo-2,3-dihydro-
benzo[b]oxepin-5-yl)-phenol (34)18 using a procedure similar to
that described for 11a. The product was purified by flash chroma-
tography (C6H12/CH2Cl2/EtOAc; 40:50:10) to afford a yellow solid
(83%) and used without further purification. IR mmax (film) 3430,
5.42. Dimethyl-{2-[4-(4-pyridin-3-yl-2,3-dihydro-benzo[b]oxepin-
5-yl)-phenoxy]-ethyl}-amine (36a)
This compound was prepared from 35a using a procedure sim-
ilar to that described for 12a. The product was purified by flash
3068–2970, 2926, 1658, 1590 cmꢁ1 1H NMR (400 Mz, CDCl3) d
;

9570
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
chromatography (CH2Cl2/EtOAc/MeOH; 70.25:5) to afford an oil
(49%). IRmmax (film) 3047–2919, 2850, 1662, 1570 cmꢁ1 1H NMR
5.47. 4-(4-Methylbenzyl)-3,4-dihydro-2H-benzo[b]oxepin-5-
one (38)
;
(400 Mz, CDCl3) d 2.36 (6H, s, N(CH3)2), 2.72–2.75 (4H, m, NCH2,
CH2), 4.02 (2H, t, J = 5.76 Hz, CH2O), 4.63 (2H, t, J = 6.02 Hz, CH2),
6.68–7.61 (12H, m, ArH; 13C NMR (101 Mz, CDC13) d 34.8, 45.2,
57.6, 65.2, 79.7, 113.6, 121.6, 123.2, 128.4, 129.3, 130.6, 132.0,
136.3, 146.7, 139.5, 149.8; HRMS (ESI) calculated for C25H26N2O2
386.1998 (M+); found: 386.1994.
4-(4-Methylbenzylidene)-3,4-dihydro-2H-benzo[b]oxepin-5-
one (37)27 (18 mmol) was dissolved in ethanol (40 mL). Ten per-
cent of Pd (17mmol) on activated charcoal was added and the reac-
tion mixture was stirred at room temperature under an
atmosphere of H2. Stirring was maintained until TLC analysis veri-
fied that hydrogenation of the starting material was complete. The
catalyst was then removed via filtration, washed with ethanol, and
the solvent was evaporated under reduced pressure. The product
was isolated using column chromatography (eluant C6H12/EtOAc;
90:10) as yellow oil in (83%) yield. IR mmax (film) 3080–2996,
5.43. {2-[4-(4-Furan-2-yl-2,3-dihydro-benzo[b]oxepin-5-yl)-
phenoxy]-ethyl}-dimethyl-amine (36b)
This compound was prepared from 35b using a procedure sim-
ilar to that described for 12a. Flash chromatography (CH2Cl2/
EtOAc/MeOH; 60:40:2) afforded the product as an oil (38%). IR mmax
2977–2827 (CH), 1687 (C@O), 1600 (C@C) cmꢁ1
;
1H NMR
(400 Mz, CDCl3) d 1.79 (1H, m, CH), 2.35 (3H, s, CH3), 2.37–2.49
(1H, m, CH), 2.79 (1H, dd, J = 13.89, 8.06 Hz, CH), 3.29 (1H, dd,
J = 13.98, 6.31 Hz, CH), 3.40 (1H, m, CH), 3.92–3.99 (1H, dt,
J = 12.10, 4.99 Hz, CH), 4.49 (1H, ddd, J = 12.42, 6.95, 2.59 Hz, CH),
6.94–7.12 (6H, m, ArH), 7.42 (1H, dt, J = 7.88, 2.00 Hz, ArH), 7.74
(1H, dd, J = 7.79, 1.67 Hz, ArH). 13C NMR (101 Mz, CDC13) d 20.9
(CH3), 34.4, 34.8, 51.3, 72.3, 119.9, 122.2, 128.9, 128.9, 129.0,
132.7, 135.2, 136.1, 161.9, 201.5; HRMS (ESI) calculated for
C19H20O (M+) 264.1574; found: 375.1514.
(film) 3050–2916, 2848–2775, 1606 cmꢁ1
;
1H NMR (400 Mz,
CDCl3) d 2.40 (6H, s, N(CH3)2), 2.76 (4H, m, NCH2, H-3), 4.11 (2H,
m, CH2O), 4.68 (2H, m, CH2), 5.71-5.75 (1H, m, ArH), 6.23–6.26
(1H, m, ArH), 6.77–7.25 (9H, m, ArH); 13C NMR (101 Mz, CDC13)
d 32.3, 45.7, 58.1, 65.6, 80.0, 109.7, 111.5, 113.4114.3, 120.4,
121.1, 123.5, 128.3, 128.5, 129.2, 131.6, 132.3, 140.5; HRMS (ESI)
calculated for C24H25NO3 (M+) 375.1845; found: 375.1834.
5.44. Dimethyl-{2-[4-(4-thiophen-2-yl-2,3-dihydro-
benzo[b]oxepin-5-yl)-phenoxy]-ethyl}-amine (36c)
5.48. 5-(4-Methoxy-phenyl)-4-(4-methylbenzyl)-2,3-dihydro-
benzo[b]oxepine (39)
This compound was prepared from 35c using a procedure sim-
ilar to that described for 12a. The product was purified by flash
chromatography (CH2Cl2/EtOAc/MeOH; 50:40:10) to afford an oil
n-BuLi (11 mmol, 2.5 M in hexane) was added dropwise to a
stirred solution of 4-bromoanisole (4.1 mmol) in dry THF (25 mL)
at ꢁ78 °C under an atmosphere of nitrogen. This solution was stir-
red for 1 h and then a solution of 38 (2.1 mmol) in dry THF (20 mL)
was added slowly at ꢁ78 °C. This mixture was stirred for 1 h at
ꢁ78 °C and then stirred at ambient temperature for 12 h. The reac-
tion mixture was extracted with diethyl ether (3ꢂ 20 mL), washed
with water, saturated NaC1 (20 mL), dried over MgSO4, and con-
centrated. The intermediate alcohol product was isolated using
flash chromatography (CH2Cl2/EtOAc; 98:2) as an oil (74%) then
dissolved in EtOH (30 mL). Eighty five percent of polyphosphoric
acid (120 mmol) was added and the mixture refluxed for 2–3 h.
The resulting solution was neutralized with NaOH (20% w/v,
20 mL), washed with water, extracted with ethyl acetate (3ꢂ
25 mL), dried over Na2SO4, and concentrated. Flash chromatogra-
phy (C6H12/CH2Cl2; 60:40) afforded the product an oil (79%). IR
(71%). IR mmax 2927–2869, 2821–2772, 1606 cmꢁ1
.
1H NMR
(400 Mz, CDCl3) d 2.41 (6H, s, (CH3)2), 2.73 (2H, t, J = 6.04 Hz,
NCH2), 2.81 (2H, t, J = 5.78 Hz, CH2), 4.09 (2H, t, J = 5.78 Hz,
CH2O), 4.65 (2H, t, J = 6.02 Hz, CH2), 6.73-7.28 (11H, m, ArH); 13C
NMR (101 Mz, CDC13) d 34.8, 45.1, 57.6, 65.1, 80.2, 113.5, 121.6,
122.5, 123.0, 124.24, 127.8, 128.6, 130.4, 131.4, 132.0; HRMS
(ESI) calculated for C24H25NO2S, 391.1609 (M+); found: 391.1606.
5.45. {2-[4-(4-Benzofuran-2-yl-2,3-dihydro-benzo[b]oxepin-5-
yl)-phenoxy]-ethyl}-dimethylamine (36d)
This compound was prepared from 35d using a procedure sim-
ilar to that described for 12a. Flash chromatography (CH2Cl2/
EtOAc/MeOH; 60:40:2) afforded the product as an oil (68%). IR mmax
mmax (film) 2984–2881, 2872–2809, 1605 cmꢁ1 1H NMR (400 Mz,
.
(film) 2924, 2853–2772, 1606 cmꢁ1 1H NMR (400 Mz, CDCl3) d
;
CDCl3) d 2.30 (2H, t, J = 6.28 Hz, CH2), 2.35 (3H, s, CH3), 3.64 (2H,
s, CH2), 3.83 (3H, s, OCH3), 4.34 (1H, t, J = 6.26 Hz, CH), 6.83–6.87
(2H, dd, J = 9.04, 1.70 Hz, ArH), 6.98-7.28 (11H, m, ArH); 13C NMR
(101 Mz, CDC13) d 20.5, 31.9, 40.0, 54.7, 79.9, 113.0, 121.4, 122.7,
127.3, 128.3, 128.6, 130.5, 130.4, 133.3, 136.0, 136.1, 136.3,
136.4, 155.3; MS (EI) m/z 305 (M+). The product was used without
further purification.
2.45 (6H, s, N(CH3)2), 2.80 (2H, m, NCH2), 2.94 (2H, t, J = 5.76 Hz,
CH2), 4.40 (2H, t, J = 5.52 Hz, CH2O), 4.61 (2H, t, J = 5.78 Hz, CH2),
6.82–7.89 (12H, m, ArH); 13C NMR (101 Mz, CDC13) d 41.0 (C-3),
45.6, 58.0, 65.9, 78.3, 111.7, 114.0, 116.2, 121.3, 121.9, 123.3,
124.0, 125.9, 128.2, 131.0, 132.0, 155.22 156.3; HRMS (ESI) calcu-
lated for C28H27NO3, 425.1990 (M+); found: 425.1991.
5.46. Dimethyl-{2-[4-(4-naphthalen-2-yl-2,3-dihydro-benzo[b]-
5.49. 4-[4-(4-Methylbenzyl)-2,3-dihydro-benzo[b]oxepin-5-yl]-
oxepin-5-yl)-phenoxy]-ethyl}-amine (36e)
phenol (40)
This compound was prepared from 35e using a procedure sim-
ilar to that described for 12a. The product was purified by flash
chromatography (CH2Cl2/EtOAc/MeOH; 70:20:5) to afford an oil
To a solution of 39 (0.4 mmol) in dry DCM (15 mL) was added
dropwise BBr3 (8.2 mmol, 1 M solution in DCM) which was diluted
with dry DCM (3 mL). The mixture was stirred for 1 h at ꢁ78 °C and
then quenched with 10% (w/w) NaHCO3 (20 mL). The dark precip-
itate was dissolved in EtOAc/MeOH 10:1 (v/v) (20 mL) with vigor-
ous stirring. The aqueous phase was separated and extracted twice
with EtOAc/MeOH. The combined organic fractions were dried
with MgSO4 and concentrated in vacuo. Purification by flash chro-
matography (CH2Cl2/EtOAc; 99:1) afforded the product as oil
(98%). IR mmax (film) 3678–3202, 3078–2924, 2914–2839,
(48%). IR mmax (film) 3056–2928, 2864–2771, 1606 cmꢁ1 1H NMR
;
(400 Mz, CDCl3) d 2.90 (6H, s, N(CH3)2), 2.64 (2H, t, J = 5.76 Hz,
NCH2), 2.89–2.96 (2H, m, CH2), 3.91 (2H, t, J = 5.78 Hz, CH2O),
4.61–4.68 (2H, m, CH2), 6.49–8.03 (15H, m, ArH); 13C NMR
(101 Mz, CDC13) d 35.7, 45.3, 57.7, 65.2, 79.6, 113.0, 113.3, 121.6,
123.0, 124.0–125.5, 126.4, 127.9, 128.0, 131.9, 133.2, 133.3; HRMS
(ESI) calculated for C30H29NO2 435.2208 (M+); found: 435.2198.

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9571
1608 cmꢁ1; 1H NMR (400 Mz, CDCl3) d 2.28 (2H, t, J = 6.28 Hz, CH2),
2.45 (3H, s, CH3), 3.63 (2H, s, CH2), 4.33 (2H, t, J = 6.28 Hz, CH2),
6.79-7.12 (11H, m, ArH); 13C NMR (101 Mz, CDC13) d 21.0, 32.4,
40.48, 80.45, 114.9, 121.9, 123.2, 127.8, 128.7, 129.1, 130.9,
154.4, 155.7; HRMS (ESI) calculated for C24H22O2 (M+) 342.1619;
found: 342.1620.
136.0, 136.1, 157.2; HRMS (ESI): calculated for C31H35NO2 (M+)
453.2660; found: 453.2668.
5.54. 4-(2-{4-[4-(4-Methylbenzyl)-2,3-dihydro-benzo[b]oxepin-
5-yl]-phenoxy}-ethyl)-morpholine (41e)
This compound was prepared from 40 using a procedure similar
to that described for 12a. The product was purified as an oil by pre-
parative thin layer chromatography (CH2Cl2/tOAc/eOH; 50:44:6),
5.50. Dimethyl-(2-{4-[4-(4-methylbenzyl)-2,3-dihydro-benzo[b]-
oxepin-5-yl]-phenoxy}-ethyl)-amine (41a)
(19%). IR mmax (film) 2955–2917, 2849, 1604 cmꢁ1 1H NMR
.
(400 Mz, CDCl3) d 2.03 (2H, m, CH2), 2.34 (3H, s, CH3), 2.64 (2H,
m, NCH2), 2.86 (4H, m, CH2–CH2), 3.67 (2H, s, CH2), 3.76 (4H, m,
CH2–CH2), 4.18 (2H, m, CH2O), 4.34 (2H, m, CH2), 6.63–6.66 (2H,
dd, J = 7.52, 1.50 Hz, ArH), 6.73–7.11 (10H, m, ArH); 13C NMR
(101 Mz, CDC13) d 22.6, 31.9, 39.5, 54.1, 57.6, 67.0, 69.0, 113.3,
121.6, 128.8, 130.9, 132.4, 136.4, 137.5; HRMS (ESI): calculated
for C30H33NO3 (M+) 455.2464; found: 455.2460.
This compound was prepared from 40 using a procedure similar
to that described for 12a. Preparative thin layer chromatography
(CH2Cl2/EtOAc/ MeOH; 50:50:10) afforded an oil (20%). IR mmax
(film) 3002–2889, 2873–2791, 1605 cmꢁ1
;
1H NMR (400 Mz,
CDCl3) d 2.28 (2H, t, J = 8.48 Hz, CH2), 2.32 (6H, s, N(CH3)2), 2.34
(3H, s, CH3), 2.75 (2H, m, NCH2), 3.60 (2H, s, CH2), 4.05 (2H, m,
(CH2O), 4.30 (2H, t, J = 8.34 Hz, CH2), 6.78–6.83 (2H, d, J = 9.04 Hz,
ArH), 6.86–7.15 (10H, m, ArH); 13C NMR (101 Mz, CDC13) d 20.3,
29.6, 40.3, 45.6, 58.1, 65.8, 80.4, 114.0, 120.4, 120.9, 127.7,
127.9–128.3, 130.2, 130.3, 130.7, 134.5, 138.1, 158.3, 159.6; HRMS
(ESI) calculated for C28H31NO2 (M+) 413.2355; found: 413.2355.
5.55. Biochemical evaluation of activity
5.55.1. Antiproliferation studies
All assays were performed in triplicate for the determination of
mean values reported. Compounds were assayed as the free bases
isolated from reaction. The human breast tumor cell line MCF-7
was cultured in Eagle’s minimum essential medium in a 95%O2/
5% CO2 atmosphere with 10% fetal calf serum. The medium was
supplemented with 1% non-essential amino acids. MDA-MB-231
cells were maintained in Dulbecco’s modified Eagle’s medium
(DMEM), supplemented with 10% (v/v) fetal bovine serum (FBS),
5.51. Diethyl-(2-{4-[4-(4-methylbenzyl)-2,3-dihydro-benzo[b]-
oxepin-5-yl]-phenoxy}-ethyl)-amine (41b)
This compound was prepared from 40 using a procedure similar
to that described for 12a. The product was isolated by preparative
thin layer chromatography (CH2Cl2/EtOAc/MeOH; 50:50:10) as an
oil (27%). IR mmax (film) 3063–2985, 2881–2807, 1608 cmꢁ1 1H
.
2 mM L-glutamine and 100 lg/mL penicillin/streptomycin (com-
NMR (400 Mz, CDCl3) d 2.19 (5H, s, N(CH3–CH2)), 2.29 (2H, t,
J = 6.28 Hz, CH2), 2.34 (8H, s, N(CH3CH2), CH3), 2.73 (2H, m,
NCH2), 3.62 (2H, s, H–CH2), 4.11 (2H, m, CH2O), 4.32 (2H, t,
6.82 Hz, CH2), 6.86–6.89 (2H, dd, J = 9.04, 2.48 Hz, ArH), 6.86–
7.15 (10H, m, ArH); 13C NMR (101 Mz, CDC13) d 22.6, 29.6, 32.3,
40.4, 47.5, 53.4, 70.7, 80.7, 113.3, 114.0, 121.9, 123.2, 127.8,
128.7, 129.2, 130.7, 130.9, 135.6, 136.4, 136.8, 155.7, 157.6; HRMS
(ESI) calculated for C30H35NO2 (M+) 441.2669; found: 441.2668.
plete medium). Cells were trypsinised and seeded at a density of
1.5 ꢂ 104 into a 96-well plate and incubated at 37 °C, 95%O2/5%
CO2 atmosphere for 24 h. After this time they were treated with
2
l
L volumes of test compound which had been pre-prepared as
stock solutions in ethanol to furnish the concentration range of
study, 1 nM–100 M, and re-incubated for a further 72 h. Control
l
wells contained the equivalent volume of the vehicle ethanol (1%
v/v). The culture medium was then removed and the cells washed
with 100 lL phosphate-buffered saline (PBS) and 50 lL MTT added,
to reach a final concentration of 1 mg/mL MTT added. Cells were
incubated for 2 h in darkness at 37 °C. At this point solubilization
5.52. 1-(2-{4-[4-(4-Methylbenzyl)-2,3-dihydro-benzo[b]oxepin-
5-yl]-phenoxy}-ethyl)-pyrrolidine (41c)
was begun through the addition of 200 lL DMSO and the cells
This compound was prepared from 40 using a procedure similar
to that described for 12a. The product was obtained as an oil by
preparative thin layer chromatography (CH2Cl2/EtOAc/MeOH;
maintained at room temperature in darkness for 20 min to ensure
thorough color diffusion before reading the absorbance. The absor-
bance value of control cells (no added compound) was set to 100%
cell viability and from this graphs of absorbance versus cell density
per well were prepared to assess cell viability and from these,
graphs of percentage cell viability versus concentration of subject
compound added were drawn.
55:44:6) (21%). IR mmax (film) 2955–2917, 2849 cmꢁ1 1H NMR
.
(400 Mz, CDCl3) d 1.73 (4H, m, CH2CH2), 2.08 (2H, m, CH2), 2.19
(3H, s, CH3), 2.82 (4H, m, CH2CH2), 2.89 (2H, m, NCH2), 3.67 (2H,
s, CH2), 4.12 (2H, m, CH2O), 4.19 (2H, m, CH2), 6.54–6.55 (2H, d,
J = 4.52 Hz, ArH), 6.81–7.56 (10H, m, ArH). HRMS (ESI) calculated
for C30H33NO2 439.2511 (M+); found: 439.2511.
5.56. Cytotoxicity studies
5.53. 1-(2-{4-[4-(4-Methylbenzyl)-2,3-dihydro-benzo[b]oxepin-
5-yl]-phenoxy}-ethyl)-piperidine (41d)
Human MCF-7 breast cancer cells were plated at a density of
1.5 ꢂ 104 per well in a 96-well plate, then incubated at 37 °C,
95%O2/5% CO2 atmosphere for 24 h. Cells were treated with the
This compound was prepared from 40 using a procedure similar
to that described for 12a. The product was obtained as an oil by
preparative thin layer chromatography (CH2Cl2/EtOAc/MeOH;
50:60:8), (31%). IR mmax (film) 3093–2993, 2976–2828,
compound of choice at varying concentrations (1 nM–100
then incubated for a further 72 h. Following incubation 50
quots of medium were removed to a fresh 96-well plate. Cytotox-
icity was determined using and LDH assay kit obtained from
Promega, following the manufacturer’s instructions for use. A
l
L ali-
M),
l
1607 cmꢁ1 1H NMR (400 Mz, CDCl3) d 1.62 (6H, m, CH2, CH2–
;
CH2), 2.29 (2H, t, J = 6.26 Hz, CH2), 2.34 (3H, s, CH3), 2.54 (4H, m,
CH2, CH2), 2.80 (2H, t, J = 6.02 Hz, NCH2), 3.63 (2H, s, CH2), 4.13
(2H, t, J = 6.04 Hz, CH2O), 4.32 (2H, t, J = 6.02 Hz, CH2), 6.84–6.87
(2H, dd, J = 10.56, 2.00 Hz, ArH), 6.97–7.25 (10H, m, ArH); 13C
NMR (101 Mz, CDC13) d 22.5, 25.4, 31.9, 40.0, 54.5, 57.5, 70.7,
80.7, 113.6, 121.4, 122.8, 127.3, 128.3, 128.7, 130.4, 130.8, 135.1,
50
in darkness at room temperature for equilibration. Stop solution
(50 L) was added to all wells before reading the absorbance at
490 nm. A control of 100% lysis was determined for a set of un-
treated cells which were lysed through the addition of 20 L lysis
solution to the media 45 min prior to harvesting. Data were pre-
lL per well LDH substrate mixture was added and the plate left
l
l

9572
I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
sented following calculation, as percentage cell lysis versus con-
centration of subject compound.
The plates were warmed to room temperature (time zero), and
the yellow color from the production of p-nitrophenol was al-
lowed to develop. The plates were monitored at 405 nm until
maximum stimulation of the cells showed an absorbance of ap-
proximately 1.2.
5.57. Estrogen receptor binding assay
ERa and ERb fluorescence polarization based competitor assay
kits were obtained from Panvera at Invitrogen Life Technologies.
The recombinant ER (insect expressed, full length, untagged hu-
man ER obtained from recombinant baculovirus-infected insect
cells) and the fluorescent estrogen ligand were removed from the
ꢁ80 °C freezer and thawed on ice for 1 h prior to use. The fluores-
5.59. Computational details
5.59.1. QSAR model
5.59.1.1. Data set. A data set of 27 compounds was selected (see
Supplementary information, Fig. 6) utilizing known ER active
ligands extracted from literature including three ligands with simi-
lar structures to the benzoxepin scaffolds of this series (benzothie-
pin-derived SERMs) previously published by our group.39 The
criteria for selection were 2-fold: (1) accuracy of the chemical
structure and (2) of the determined IC50 of the compounds. The
cent estrogen ligand (2 nM) was added to the ER (30 nM for ER
a
and 20 nM for ERb) and screening buffer (100 mM potassium phos-
phate (pH 7.4), 100 g/mL BGG, 0.02%NaN3 was added to make up
l
to a final volume that was dependant on the number of tubes used
(number of tubes (e.g., 50) ꢂ volume of complex in each tube
(50
l
L) = total volume (e.g., 2500
l
L). Test compound (1
l
L, con-
L screening
buffer in each borosilicate tube (6mm diameter). To this 50 L of
the fluorescent estrogen/ER complex was added to make up a fi-
nal volume of 100 L and final concentration range for the test
compound of 0.01 nM to 1 M. A vehicle control contained 1%
(v/v) of ethanol and a negative control contained 50 L screening
buffer and 50 L fluorescent estrogen/ER complex. The negative
ratio of selectivity ER b/a was used as input for the prediction of
centration range 1 nM) to 100 lM) was added to 49 l
selectivity of the benzoxepin series. This data was represented in
the form of the natural logarithm for convenience. The original
data set was divided several times into a training set of 20 ligands
and an external test set of seven ligands, respectively. Cross-valida-
tion procedure was implemented by using leave-one-out (LOO)
cross-validation procedure implying exclusion of each compound
of the training set and the prediction of its activity by the model
developed using the remaining compounds of the training set.
l
l
l
l
l
control was used to determine the polarisation value when no
competitor was present (theoretical maximum polarization). The
tubes were incubated in the dark at room temperature for 2 h
and were mixed by shaking on a plate shaker. The polarization
values were read on a Beacon single-tube fluorescent polarization
instrument with 485 nm excitation and 530 nm emission interfer-
5.59.1.2. Descriptor set. A standard set of 2D descriptors (e.g.,
logP, topological indices, BCUT and GCUT descriptors, TPSA), the
ratio of those descriptor values to the number of heavy atoms in
the molecule, including substructural ones (numbers of occur-
rences of fragments), were generated. Two software packages,
one developed ‘in-house’ (Quasi-fragmental descriptors imple-
mented through Pipeline Pilot) and the QSAR module in MOE
2006.0824 package, respectively, permitted calculation of the set.
The first, Quasi-Fragmental descriptors, as described previously
so only a brief discussion of changes to the implementation will be
detailed.46,47 Four classification methods are available to the user,
the first classification level corresponds to the general type of
atom; the second level corresponds to a type of chemical element.
Atoms on the third and fourth classification levels characterize
type of hybridization, bond environment for each atom, atomic for-
mal charge, and the number of hydrogen neighbours. The fourth le-
vel of classification was employed with a maximum fragment size
of 15 heavy (non-hydrogen) atoms.
ence filters. For ERa and ERb, graphs of anisotropy (mA) versus
competitor concentration were obtained for determination of
IC50 values.
5.58. Estrogenic activity: alkaline phosphatase assay
Following the procedure of Littlefield et al.38, human Ishikawa
cells were maintained in Eagle’s minimum essential medium
(MEM containing 10% vol/vol fetal bovine serum (FBS) and sup-
plemented with 100 U/mL penicillin and 10 lg/mL streptomycin,
2 mM glutamine, and 1 mM sodium pyruvate. Twenty four hours
before the start of the experiment, near confluent cells were
changed to an estrogen-free medium (EFBM), A 1:1 mixture of
phenol-free Ham’s F-12 and Dulbecco’s modified Eagle’s medium,
together with the supplements listed above, and 5% calf serum,
stripped of endogenous estrogens with dextran coated charcoal.
On the day of the experiment, cells were harvested with 0.25%
trypsin and plated in 96-well flat bottomed microtitre plates in
EFBM at a density of 2.5 ꢂ 104 cells/well. Test compounds were
dissolved in ethanol at 10ꢁ3 M, diluted with EFBM (final concen-
tration of ethanol 0.1%) and filter sterilised. After addition of the
5.59.1.3. Model building. Our modeling was based on two appli-
cations of statistical methods—multiple linear regression (MLR)
and partial least squares (PLS) method. The MLR with implemented
genetic algorithm (GA) was used to converge on those descriptors
not strongly correlated with one another. A number of different
initial descriptor sets, including only substructural, 2D-MOE, nor-
malized 2D-MOE descriptors, and different combinations of all,
were chosen to analyze their impact on studied activity. At this
stage, to introduce some nonlinearity, and to increase the predict-
ability of the linear models, some nonlinear modifications, such as
the square, square root, and common logarithm (log Di, calculated
only for those fragments that were found in all structures con-
tained in the database), were applied to the initial descriptors.
The top subsets consisting of 4–5 descriptors and based on MLR
modeling for each partition were used for further modeling. The
main part of our modeling was performed using the PLS method
with LOO cross-validation. For each model, the LOO cross-validated
predictions were examined and outliers were identified. The PLS
procedure was subsequently repeated without cross-validation
inputting the correct components contributing to the final model.
test compounds, (plated in 50
lL, added estradiol in 50
lL, and
blank medium to give a final volume 150
lL) the cells were incu-
bated at 37 C in a humidified atmosphere containing 95%O2/5%
CO2 for 72 h. All experimental values were obtained in triplicate.
The microtitre plates were then inverted and the growth medium
removed. The plates were then rinsed by gentle immersion and
swirling in 2 L of PBS (0.15 M NaCl, 10 mM sodium phosphate,
pH 7.4). The plates were removed from the container, the residual
saline in the plate was not removed, and the wash was repeated.
The buffered saline was then shaken out, and the plate blotted on
paper towel. The covers were replaced and the plates were placed
at ꢁ80 °C for at least 15 min then thawed at room temperature
for 5–10 min. The plates were then placed on ice and 50 lL ice
cold solution containing 50 mM p-nitrophenyl phosphate,
0.24 mM MgCl2 and 1 M diethanolamine (pH 9.8) was added.

I. Barrett et al. / Bioorg. Med. Chem. 16 (2008) 9554–9573
9573
7. Pike, A. C.; Brzozowski, A. M.; Hubbard, R. E.; Bonn, T.; Thorsell, A. G.; Engstrom,
O.; Ljunggren, J.; Gustafsson, J. A.; Carlquist, M. EMBO J. 1999, 18, 4608.
8. Pike, A. C.; Brzozowski, A. M.; Walton, J.; Hubbard, R. E.; Thorsell, A. G.; Li, Y. L.;
Gustafsson, J. A.; Carlquist, M. Structure 2001, 9, 145.
9. Wu, Y. L.; Yang, X.; Ren, Z.; McDonnell, D. P.; Norris, J. D.; Willson, T. M.;
Greene, G. L. Mol. Cell 2005, 18, 413.
5.59.2. Docking
5.59.2.1. Ligand preparation. Compound 26b was drawn using
ACD/Chemsketch v1048 and SMILES strings generated. A single
conformer was generated ensuring a final MMFF optimization step
for refinement of the compound, using Omega v2.1.49
10. Lloyd, D. G.; Buenemann, C. L.; Todorov, N. P.; Manallack, D. T.; Dean, P. M. J.
Med. Chem. 2004, 47, 493.
5.59.2.2. Receptor preparation. PDB entries 3ERT and 1QKN
were downloaded from the Protein Data Bank (PDB). All waters
were retained in both isoforms. Addition and optimization of
hydrogen positions for these waters was carried out using MOE
2006.0824 ensuring all other atom positions remained fixed. The
resulting structure was imported into ADT (Autodock Tools)50 with
Geisteger charges and atoms types subsequently added. Impor-
tantly, the correct tautomeric position for His524 was chosen,
11. Labrie, F.; Labrie, C.; Belanger, A.; Simard, J.; Gauthier, S.; Luu-The, V.; Merand,
Y.; Giguere, V.; Candas, B.; Luo, S.; Martel, C.; Singh, S. M.; Fournier, M.; Coquet,
A.; Richard, V.; Charbonneau, R.; Charpenet, G.; Tremblay, A.; Tremblay, G.;
Cusan, L.; Veilleux, R. J. Steroid Biochem. Mol. Biol. 1999, 69, 51.
12. Zhou, H. B.; Comninos, J. S.; Stossi, F.; Katzenellenbogen, B. S.;
Katzenellenbogen, J. A. J. Med. Chem. 2005, 48, 7261.
13. Kanbe, Y.; Kim, M. H.; Nishimoto, M.; Ohtake, Y.; Kato, N.; Tsunenari, T.;
Taniguchi, K.; Ohizumi, I.; Kaiho, S.; Morikawa, K.; Jo, J. C.; Lim, H. S.; Kim, H. Y.
Bioorg. Med. Chem. 2006, 14, 4803.
14. Jain, N.; Kanojia, R. M.; Xu, J.; Jian-Zhong, G.; Pacia, E.; Lai, M. T.; Du, F.; Musto,
A.; Allan, G.; Hahn, D.; Lundeen, S.; Sui, Z. J. Med. Chem. 2006, 49, 3056.
15. Kim, S.; Wu, J. Y.; Birzin, E. T.; Frisch, K.; Chan, W.; Pai, L. Y.; Yang, Y. T.; Mosley,
R. T.; Fitzgerald, P. M.; Sharma, N.; Dahllund, J.; Thorsell, A. G.; DiNinno, F.;
Rohrer, S. P.; Schaeffer, J. M.; Hammond, M. L. J. Med. Chem. 2004, 47, 2171.
16. Sibley, R.; Hatoum-Mokdad, H.; Schoenleber, R.; Musza, L.; Stirtan, W.; Marrero,
D.; Carley, W.; Xiao, H.; Dumas, J. Bioorg. Med. Chem. Lett. 2003, 13, 1919.
17. Wallace, O. B.; Lauwers, K. S.; Dodge, J. A.; May, S. A.; Calvin, J. R.; Hinklin, R.;
Bryant, H. U.; Shetler, P. K.; Adrian, M. D.; Geiser, A. G.; Sato, M.; Burris, T. P. J.
Med. Chem. 2006, 49, 843.
18. Lloyd, D. G.; Hughes, R. B.; Zisterer, D. M.; Williams, D. C.; Fattorusso, C.;
Catalanotti, B.; Campiani, G.; Meegan, M. J. J. Med. Chem. 2004, 47, 5612.
19. Sarkhel, S.; Sharon, A.; Trivedi, V.; Maulik, P. R.; Singh, M. M.; Venugopalan, P.;
Ray, S. Bioorg. Med. Chem. 2003, 11, 5025.
20. McCague, R.; Leclercq, G.; Jordan, V. C. J. Med. Chem. 1988, 31, 1285.
21. Platzek, J.; BeckmannW; Geisler, J.; Kirstein, H.; Niedballa, U.; Ottow, E.; Radau,
S.; Schulz, C.; Wessa, T. (Schering Aktiengesellschaft, Germany). PCT Int. Appl.
WO 2003033461, 2003, 48 pp..
22. Shani, J.; Gazit, A.; Livshitz, T.; Biran, S. J. Med. Chem. 1985, 28, 1504.
23. Hummel, C. W.; Geiser, A. G.; Bryant, H. U.; Cohen, I. R.; Dally, R. D.; Fong, K. C.;
Frank, S. A.; Hinklin, R.; Jones, S. A.; Lewis, G.; McCann, D. J.; Rudmann, D. G.;
Shepherd, T. A.; Tian, H.; Wallace, O. B.; Wang, M.; Wang, Y.; Dodge, J. A. J. Med.
Chem. 2005, 48, 6772.
Ne. Flexible residues were selected for the fully flexible docking
routine as outlined in the next section.
5.59.2.3. Flexible Residue selection. Ten crystallographic struc-
tures displaying a resolution <2.28 Å with a bound antagonist were
downloaded from the Protein Data Bank (PDB_ID: 1SJ0, 1UOM,
1XP1, 1XP6, 1XP9, 1XPC, 1XQC, 1YIM, 1YIN, and 3ERT) and a set
of 36 estrogen actives with activities ranging from nanomolar to
low micromolar potency were docked in all using a rigid docking
protocol previously optimized.51 The receptor which docked the
least number of actives was chosen as the template for overlay of
the remaining X-ray structures. Superposition of the 10 crystal
structures was executed and a svl script (MOE.2006.08) was used
to analyze the rmsd differences between all residues of the active
site as summarized in Figure 2. From the analysis, the following
residues were chosen to be flexible in the docking process—ERa:
Met421, Ile424, Phe425, His524, and Leu525. Equivalent residues
were also selected by a similar process in ERb: Ile328, Ile331,
Phe332, His430, and Leu431.
24. Molecular Operating Environment (MOE), developed and distributed by
Chemical Computing Group. http://www.chemcomp.com.
25. Hachiya, I.; Moriwaki, M.; Kobayashi, S. Tetrahedron Lett. 1995, 36, 409.
26. Freedman, J.; Stewart, K. T. J. Heterocyc. Chem. 1989, 26, 1547.
27. Orlov, V. D.; Mikhed’ kina, E. I.; Lavrushin, V. F. Zhurnal Obshchei Khimii 1984,
54(1), 168.
28. Teo, C. C.; Kon, O. L.; Sim, K. Y.; Ng, S. C. J. Med. Chem. 1992, 35, 1330.
29. Acton, D.; Hill, G.; Tait, B. S. J. Med. Chem. 1983, 26, 1131.
30. Catherino, W. H.; Wolf, D. M.; Jordan, V. C. Mol. Endocrinol. 1995, 9, 1053.
31. Mccague, R. Tetrahedron: Asymmetry 1990, 1, 97.
32. McCague, R.; Jarman, M.; Leung, O. T.; Foster, A. B.; Leclercq, G.; Stoessel, S. J.
Steroid Biochem. 1988, 31, 545.
5.59.2.4. Docking. Autodock 4 through ADT52 was employed as the
docking tool as it allows both ligand and receptor flexibility to be
modelled. A grid parameter file was set up with a grid box of
175,000 grid points per map centered on the calculated center of
the co-crystallised ligands of each isoform. Docking was carried out
using the Lamarckian genetic algorithm (GALS) and changing only
ga_run 50, ga_num_evals 25,00,000, ga_num_generations 5000.
33. McCague, R.; Leclercq, G. J. Med. Chem. 1987, 30, 1761.
34. Grese, T. A.; Cho, S.; Finley, D. R.; Godfrey, A. G.; Jones, C. D.; Lugar, C. W., 3rd;
Martin, M. J.; Matsumoto, K.; Pennington, L. D.; Winter, M. A.; Adrian, M. D.;
Cole, H. W.; Magee, D. E.; Phillips, D. L.; Rowley, E. R.; Short, L. L.; Glasebrook, A.
L.; Bryant, H. U. J. Med. Chem. 1997, 40, 146.
Acknowledgments
35. Qin, Z.; Kastrati, I.; Chandrasena, R. E.; Liu, H.; Yao, P.; Petukhov, P. A.; Bolton, J.
L.; Thatcher, G. R. J. Med. Chem. 2007, 50, 2682.
36. Lloyd, D. G.; Smith, H. M.; O’ Sullivan, T.; Zisterer, D. M. M. J. M. Med. Chem.
2005, 4, 335.
37. C-QSAR; Biobyte Corp., W. t. S., Suite 204, Claremont, CA.
38. Littlefield, B. A.; Gurpide, E.; Markiewicz, L.; McKinley, B.; Hochberg, R. B.
Endocrinology 1990, 127, 2757.
39. Meegan, M. J.; Barrett, I.; Zimmermann, J.; Knox, A. J.; Zisterer, D. M.; Lloyd, D.
G. J. Enzyme Inhib. Med. Chem. 2007, 22, 655.
40. Mukherjee, S.; Saha, A.; Roy, K. Bioorg. Med. Chem. Lett. 2005, 15, 957.
41. Osterberg, F.; Morris, G. M.; Sanner, M. F.; Olson, A. J.; Goodsell, D. S. Proteins
2002, 46, 34.
42. Celik, L.; Lund, J. D.; Schiott, B. Biochemistry 2007, 46, 1743.
43. Malamas, M. S.; Manas, E. S.; McDevitt, R. E.; Gunawan, I.; Xu, Z. B.; Collini, M.
D.; Miller, C. P.; Dinh, T.; Henderson, R. A.; Keith, J. C., Jr.; Harris, H. A. J. Med.
Chem. 2004, 47, 5021.
44. Yang, C.; Edsall, R., Jr.; Harris, H. A.; Zhang, X.; Manas, E. S.; Mewshaw, R. E.
Bioorg. Med. Chem. 2004, 12, 2553.
45. (a) Nitikin, K. V.; Andryukhova, N. P. Can. J. Chem. 2004, 82, 571; (b) Ghosh, A.;
Bhattacharya, S.;Raychaudhuri, S. R.;Chatterjee, A. Indian. J. Chem B. 1992, 31, 299.
46. Artemenko, N. V.;Baskin, I. I.;Palyulin, V. A.;Zefirov, N. S. DokladyChem. 2001,381,317.
47. Artemenko, N. V.; Palyulin, V. A.; Zefirov, N. S. Doklady Chem. 2002, 386, 114.
48. ACD/Chemsketch v10, Advanced Chemistry Labs. http://www.acdlabs.com/
download/chemsk.html.
49. OMEGA 2.1, distributed by Openeye Scientific Software.
50. Autodock Tools (ADT), developed and distributed by The Scripps Research
Institute. http://autodock.scripps.edu/resources/adt.
We are very grateful to Professor Richard Hochberg at Yale Uni-
versity Medical School, for kindly facilitating the alkaline phospha-
tase experiments with the generous gift of the Ishikawa cells. This
work was supported through funding from the Trinity College IITAC
research initiative (HEA PRTLI), EnterpriseIreland (EI), Science Foun-
dationIreland(SFI), and theHealthResearchBoard(HRD), withaddi-
tional support for computational facilities from the Wellcome Trust.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2008.09.035.
References and notes
1. Jordan, V. C. J. Med. Chem. 2003, 46, 1081.
2. Jordan, V. C. J. Med. Chem. 2003, 46, 883.
3. Park, W. C.; Jordan, V. C. Trends Mol. Med. 2002, 8, 82.
4. Bardin, A.; Boulle, N.; Lazennec, G.; Vignon, F.; Pujol, P. Endocr. Relat. Cancer
2004, 11, 537.
5. Brzozowski, A. M.; Pike, A. C.; Dauter, Z.; Hubbard, R. E.; Bonn, T.; Engstrom, O.;
Ohman, L.; Greene, G. L.; Gustafsson, J. A.; Carlquist, M. Nature 1997, 389, 753.
6. Shiau, A. K.; Barstad, D.; Loria, P. M.; Cheng, L.; Kushner, P. J.; Agard, D. A.;
Greene, G. L. Cell 1998, 95, 927.
[51]. Knox,A.J.;Meegan,M.J.;Carta,G.;Lloyd,D.G.J.Chem.Inf.Model.2005,45,1908.
[52]. Sanner, M. F. J. Mol. Graph. Model. 1999, 17, 57.