Design, synthesis, and biological evaluation of new phosphodiesterase type 4 inhibitors

Article abstract of DOI:10.1016/j.bmc.2004.07.040

The mode of interaction with enzyme is hypothesized based on the SAR data. The design, synthesis, and biological evaluation of new phosphodiesterase type 4 inhibitors, which possess new templates instead of a cyclohexane ring, are described. The mode of i

Full text of DOI:10.1016/j.bmc.2004.07.040

Bioorganic & Medicinal Chemistry 12 (2004) 5063–5078
Design, synthesis, and biological evaluation of
new phosphodiesterase type 4 inhibitors
Hiroshi Ochiai,a Yoshihiko Odagaki,a Tazumi Ohtani,a Akiharu Ishida,a
Kensuke Kusumi,a Katuya Kishikawa,b Susumu Yamamoto,a Hiroshi Takeda,a
a
Takaaki Obata,a Kaoru Kobayashi,a Hisao Nakaia, and Masaaki Toda
*
aMinase Research Institute, Ono Pharmaceutical Co., Ltd, 3-1-1 Sakurai, Shimamoto, Mishima, Osaka 618-8585, Japan
bDevelopment Planning, Ono Pharmaceutical Co., Ltd, 2-1-5 Doshomachi, Chuo-ku, Osaka 541-8526, Japan
Received 7 June 2004; revised 16 July 2004; accepted 16 July 2004
Available online 20 August 2004
Abstract—The design, synthesis, and biological evaluation of new phosphodiesterase type 4 inhibitors, which possess new templates
instead of a cyclohexane ring, are described. The mode of interaction with the enzyme is discussed based on the structure–activity
relationship (SAR) data obtained for the synthesized inhibitors. Furthermore, the roles of three pharmacophores, a catechol moiety,
a nitrile moiety, and acidic moieties, are discussed using in silico docking studies. More detailed biological evaluations of selected
compounds are also presented.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
reported so far for PDE4 inhibitors, we focused our
attention on the unique arrangement of three pharmaco-
phores (carboxylic acid, benzylic nitrile, and 3-cyclopent-
yloxy-4-methoxyphenyl moieties) on the cyclohexane
ring of Ariflo 1. It is especially interesting that the cis-
isomer (Ariflo) exhibits more potent PDE4 inhibitory
activity than the corresponding trans-isomer, and that
the methyl ester of Ariflo shows reduced inhibitory
activity relative to Ariflo.7 Thus, the carboxylic acid
moiety may play an important role in increasing the
affinity of the inhibitor for the target enzyme. The nitrile
group could also play a role in promoting this interac-
tion because the decyanated analog of Ariflo 1 exhibits
weaker activity.7
Phosphodiesterase type 4 (PDE4) is an enzyme that is
mainly expressed in airway smooth muscle, immune
cells, and inflammatory cells, where it catalyzes the
hydrolysis of the second messenger adenosine 30,50-
monophosphate (cAMP) to form the inactive nucleotide
50-monophosphate.1–4 An anti-inflammatory effect of
PDE4 inhibitors has been demonstrated in various ani-
mal models,5,6 and inhibition of PDE4 has been pro-
posed as a new therapeutic approach to the treatment
of chronic inflammatory diseases such as asthma,
chronic obstructive pulmonary disease (COPD), and
rheumatoid arthritis. Common side effects of PDE4
inhibitors are nausea and emesis. Arifloꢁ (cilomilast 1)
is the PDE4 inhibitor for which development is most ad-
vanced and it is currently in pre-registration for
COPD.7–9 This drug seems to have an acceptable thera-
peutic index.
To improve the above-described side-effect profiles, we
reported on the design and synthesis for two series of
PDE4 inhibitors, that is, piperidines10 and bicy-
clo[3.3.0]octanes,11 both of which have these three phar-
macophores. The piperidines were designed so as to
decrease their penetration into the central nervous
system (CNS), while the bicyclo[3.3.0]octanes were
expected to improve subtype selectivity by restricting a
spatial arrangement of these three pharmacophores.
Although many efforts have been made to develop
PDE4 inhibitors with acceptable therapeutic indexes,
their common side effects of nausea and emesis have still
not been overcome. Among the innumerable structures
It is of interest to estimate the probable mode of interac-
tion of the new inhibitors with diverse stereochemistries
of these three pharmacophores as shown in Figure 1.
*
Corresponding author. Tel.: +81-75-961-1151; fax: +81-75-962-
9314; e-mail: hi.nakai@ono.co.jp
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.07.040

5064
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
O
MeO
O
O
O
NH
MeO
Rolipram
MeO
MeO
Ring
COY
COOH
NC
1 (Ariflo)
O
X
F
X = CN, H
Y = NHOH, OH
F
O
N
N
H
Zardavenine
O
O
O
H
MeO
MeO
H
MeO
N
n
COY
COY
NC
COY
X
X
H
H
X = CN, H
Y = OH, NHOH
(3ab, 9ab)
n = 1, 2, 3
X = CN, H
Y = OH, NHOH
(10ab, 11ab)
Y = OH, NHOH
(2ab, 4ab, 5ab)
O
O
H
MeO
MeO
MeO
MeO
Me
n
N
COOH
COOH
COOH
NC
H
NC
COY
n = 0, 1
(3c)
Y = OH, NHOH
(6ab, 7ab)
O
H
O
NC
MeO
H
Me
Me
(3e)
N
NC
COY
Y = OH, NHOH
O
O
(8ab)
H
MeO
H
COY
H
NC
H
H
(3d)
Y = OH, NHOH
(9cd)
Figure 1. New templates with stereochemical diversity.
Here we report on the probable mode of interaction of
newly found PDE4 inhibitors with the target enzyme,
based on their SAR and in silico docking studies. Fur-
ther biological evaluation of selected inhibitors is also
presented.
carboxylic acids 4a–7a, respectively. Condensation of
4a–7a with the O-protected hydroxylamine, followed
by acidic deprotection in methanol, produced the corre-
sponding hydroxamic acids 4b–7b, respectively.
Synthesis of 9c–d is described in Scheme 2. Monopro-
tected bicyclo[3.3.0]octane-3,7-dione 1712,13 was con-
verted to a bicyclo[3.3.0]octene intermediate 18 by
nucleophilic addition of the appropriately substituted
phenyl lithium, followed by dehydration through
methanesulfonylation in the presence of triethyl amine.
Catalytic hydrogenation of 18 in the presence of palla-
dium on carbon provided endo-configuration 19, acidic
deprotection of which afforded a ketone 20. Nucleo-
philic addition of trimethylsilyl cyanide to 20 in the
2. Chemistry
Synthesis of 2a–b and 8a–b was reported in the preced-
ing paper,10 while synthesis of 4–7 is outlined in Scheme
1. 1,4-Addition of 1210 to methyl acrylate provided 13.
N-Alkylation of piperidine 12 with an appropriate alkyl
halide in the presence of potassium carbonate afforded
14–16. Alkaline hydrolysis of esters 13–16 gave

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5065
X
HCl
NH a or b
X
N
COOH
N
Y
c
O
O
O
O
O
O
CN
CN
Me
CN
Me
Me
13: X = -(CH2)2- , Y = COOMe (87%)
4a: X = -(CH2)2- (50%)
5a: X = -(CH2)3- (77%)
12
14: X = -(CH2)3- , Y = COOEt (85%)
15: X = -CH(CH3)- , Y = COOEt (97%)
6a: X =-CH(CH3)- (51%)
16: X = -CH(CH2CH3)- , Y = COOEt (quant.)
7a: X = -CH(CH2CH3)- (60%)
X
N
CONHOH
d
O
O
CN
Me
4b: X = -(CH2)2- (35%)
5b: X = -(CH2)3- (71%)
6b: X =-CH(CH3)- (76%)
7b: X = -CH(CH2CH3)- (56%)
Scheme 1. Synthesis of compounds 4–7. Reagents: (a) CH2CHCOOMe, Et3N, THF, 45ꢂC; (b) alkylbromide, K2CO3, DMF, rt to 80ꢂC; (c) 1N
NaOH, MeOH, rt to 80ꢂC; (d) EDC, HOBt, DMF, Et3N, NH2OC(CH3)2OCH3 then 2N HCl, MeOH.
Me
Me
Me
Me
H
O
X
O
H
H
e, f
c
a, b
O
O
O
O
O
H
O
H
Me
H
O
Me
X
17
18
O
O
19: X =
d
Me Me
20: X = O
H
H
COX
h
O
O
H
H
Me
Me
O
O
21: X = CN
22: X = COOH
9c: X = OH
9d: X = NHOH
g
i
Scheme 2. Synthesis of compounds 9c–d. Reagents: (a) nBuLi, 3-cyclopentyloxy-4-methoxy phenylbromide, CeCl3, THF, ꢀ78ꢂC; (b) MsCl, Et3N,
CH2Cl2, 0ꢂC to rt (56% in two steps); (c) H2, 10% Pd/C, dioxane (100%); (d) TsOH, acetone (95%); (e) TMSCN, ZnI2, CH2Cl2; (f) POCl3, py, reflux
(91% in two steps); (g) 40% KOHaq, HOCH2CH2OH, 200ꢂC (64%); (h) H2, 10% Pd/C, MeOH (54%); (i) EDC, HOBt, DMF, Et3N,
H2NOC(CH3)2OCH3 then 2N HCl, MeOH (75%).
presence of zinc iodide followed by dehydration afforded
the conjugated nitrile 21, alkaline hydrolysis of which
produced a carboxylic acid 22. Catalytic hydrogenation
of 22 in the presence of palladium on carbon resulted
in 9c, which was converted to the corresponding
hydroxamic acid 9d by condensation with O-protected
hydroxamine followed by acidic deprotection. The
stereochemistry of 22 and 9c was determined by X-ray
crystallography (Fig. 2) and NMR analysis (Fig. 3),
respectively. As shown in Figure 3, NOESY correlation
of the three protons Ha–c, which are attached to the
benzylic position, the ring juncture, and the carbon
atom adjacent to the carboxylic acid moiety, respec-
tively, was observed.
one of the challenging problems to be solved because
of the exo-configuration of the aromatic moiety.
However, synthesis was successfully carried out by the
catalytic hydrogenation of a key intermediate 24 in
the presence of an iridium(I) catalyst,14–16 as described
below. Compound 23 was obtained by acidic deprotec-
tion of 18. First, hydride reduction of 23 from the less
hindered side gave the hindered alcohol 24. Second, ster-
eoselective catalytic hydrogenation of 24 from the hin-
dered side was achieved without contamination by the
unwanted endo-isomer using an iridium(I) catalyst14,15
to produce the exo-isomer 25 exclusively.16 The hin-
dered hydroxy residue of 24 served as an excellent scaf-
fold for the iridium(I) catalyst.16 Oxidation of the
alcohol 25 with tetrapropylammonium perruthenate
(TPAP) in methylene chloride in the presence of 4-meth-
ylmorpholine N-oxide gave 26. Compound 26 was
Synthesis of 9a–b and 11a–b is outlined in Scheme 3.
Stereoselective synthesis of 9a–b starting from 18 was

5066
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
converted to a conjugated nitrile 27 by formation of a
O2
O1
cyanohydrin with TMS cyanide in the presence of zinc
iodide, followed by dehydration with phosphorus oxy-
chloride in the presence of pyridine. Alkaline hydrolysis
of 27 gave a conjugated carboxylic acid 11a (Fig. 4),
which was converted to the corresponding hydroxamic
acid 11b according to the same procedures as described
for the conversion of 9c to 9d. Stereochemistry of 11a
was determined by comparison of its 1H NMR spectrum
with that of 22. The two spectra were clearly different
with regard to the d values of the protons at C-4 (H-
4ab) and C-6 (H-6ab). Catalytic hydrogenation of 11a
produced saturated carboxylic acid 9a (Fig. 5), which
was converted to the corresponding hydroxamic acid
analog 9b by the method described above. Stereochem-
istry of 9a was determined by comparison of its 1H
NMR spectrum with that of 9c. The two spectra were
clearly different with regard to the d values of the pro-
tons at C-2 (H-2), C-4 (H-4b), and C-6 (H-6b). Synthesis
of 3a–e was reported in the preceding paper.11
C1
C9
C2
C8
C7
C6
C3
C4
C5
C15
C11
C10
C18
C17
C12
C14
Synthesis of 10a–b from 2811 is described in Scheme 4.
Trapping an enolate anion prepared from 28 with
N,N-bis(trifluoromethylsulfonyl)aniline afforded an enol
triflate 29. Insertion reaction of a carbon monoxide into
29 in the presence of the palladium catalyst and metha-
nol resulted in an a,b-unsaturated ester 30, alkaline
hydrolysis of which produced a carboxylic acid 10a.
Compound 10a was converted to the corresponding
hydroxamic acid 10b according to the same procedures
as described above.
C19
C20
C13
O4
C16
O3
C21
Figure 2. X-ray crystallographic study of 22.
3. Results and discussion
3.1. Inhibition of PDE4
NOESY
Hc
Hb
COOH
Ha
O
All the test compounds listed in Tables 1–4 were synthe-
sized and evaluated for their ability to inhibit PDE4
prepared from U937 cells17 (derived from human
monocytes). The assay data are expressed as IC50 values,
Hb
Me
O
9c
Figure 3. Observed key NOESY correlations of 9c.
H
H
H
X
CN
X
e, f
a
c
O
O
O
O
18
H
H
H
Me
Me
Me
O
O
27
23: X = O
H
OH
25: X =
b
d
H
OH
24: X =
26: X = O
H
COX
i
H
COX
g
O
O
H
H
Me
h
Me
j
O
O
11a: X = OH
11b: X = NHOH
9a: X = OH
9b: X = NHOH
Scheme 3. Synthesis of compounds 11a–b and 9a–b. Reagents: (a) TsOH, acetone (90%); (b) NaBH4, MeOH, THF, 0ꢂC (100%); (c) H2, CH2Cl2,
Ir(cod)py(PCy3)PF3 (89%); (d) TPAP, NMO, CH2Cl2, MS4A (92%); (e) TMSCN, ZnI2, CH2Cl2; (f) POCl3, pyridine, 90ꢂC (55% in two steps); (g)
40% KOHaq, HOCH2CH2OH, 200ꢂC (93%); (h) EDC, HOBt, DMF, Et3N, H2NOC(CH3)2OMe then 2N HCl, MeOH (56%); (i) H2, 10% Pd/C,
MeOH (81%); (j) EDC, HOBt, DMF, Et3N, H2NOC(CH3)2OCH3 then 2N HCl, MeOH (70%).

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5067
3a
H
3a
H
3
3
2
1
2
1
COOH
COOH
4
5
4
5
O
O
H
6a
H
Me
Me
6
11a
6
22
6a
O
O
1H NMR δ (ppm)
1H NMR δ (ppm)
3.39 (1H, m, H-6a)
2.91 (1H, m, H-5)
2.90-2.75 (2H, m, H-3a, 3β)
2.50-2.35 (2H, m, H-3α, 6β)
2.30 (1H, m, H-4β)
3.52 (1H, m, H-6a)
3.10-2.89 (3H, m, H-5, 3a, 3β)
2.39 (1H, m, H-3α)
2.00-1.74 (10H, m, H-6αβ,4αβ, cPen X 3/4)
1.68-1.53 (3H, m, cPen X 1/4 + COOH)
1.98-1.70 (7H, m, cPen X 3/4 + COOH)
1.70-1.50 (2H, m, cPen X 1/4)
1.45-1.27 (2H, m, H-4α, 6α)
Figure 4. 1H NMR comparison of 11a and 22.
3a
H
3a
3
3
2
1
H
2
1
COOH
COOH
4
5
4
O
O
5
H
H
Me
O
Me
6
9c
6
6a
6a
O
9a
1H NMR δ (ppm)
3.08 (1H, m, H-5)
2.92 (1H, m, H-2)
2.66-2.55 (2H, m, H-3a, 6a)
2.30-2.19 (4H, m, H-1β, 3β, 4β, 6β)
1H NMR δ (ppm)
3.10 (1H, m, H-5)
2.73-2.57 (3H, m, H-2, 3a, 6a)
2.34-2.24 (2H, m, H-1β, 3β)
Figure 5. 1H NMR comparison of 9a and 9c.
H
COX
H
H
b
OTf
O
a
O
O
O
H
NC
H
Me
H
NC
Me
NC
Me
O
O
O
29
28
30: X = OMe
10a: X = OH
10b: X = NHOH
c
d
Scheme 4. Synthesis of compound 10a–b. Reagents: (a) LiHMDS, Tf2NPh, THF, ꢀ78 to 0ꢂC; (b) Pd(OAc)2, Et3N, PPh3, CO, MeOH, DMF (94% in
two steps); (c) 2N NaOH, MeOH, THF (85%); (d) EDC, HOBt, Et3N, NH2OC(CH3)2OMe, DMF then 2N HCl, MeOH (56%).
that is, the test compound concentration that achieved
50% inhibition of PDE4 relative to the effect of the
vehicle. In preceding papers,10,11 we reported on the
discovery of two new series of PDE4 inhibitors with
therapeutic potential, starting from the chemical modifi-
cation of Ariflo (Fig. 1). During those studies, the car-
boxylic acid or hydroxamic acid in addition to another
two pharmacophores (a catechol and a benzylic nitrile)
were found to play an important role in the interaction
of these inhibitors with the target enzyme. In the present
report, we discuss the critical role of these pharmaco-
phores (carboxylic/hydroxamic acid, benzylic nitrile,
and catechol) based on SAR data and in silico docking
studies for newly found inhibitors.
gradual decrease in inhibitory activity for each series.
Thus, the chain length between the nitrogen of the pip-
erizine ring and the acidic group was found to be opti-
mized at the monomethylene moiety, as illustrated by
2a–b, which shows the strongest activity in the SAR
study.
Secondly, substitution of the a-position of the carbonyl
group to block rapid hydrolysis of hydroxamic acid was
carried out. Introduction of a methyl group at the a-po-
sition of 2a–b afforded 6a–b, respectively, with more
than 4.5-fold and 5-fold decreases in inhibitory activity,
respectively. Replacement of the methyl moiety of 6a–b
with a more hindered ethyl moiety produced 7a–b,
respectively, with a further decrease in inhibitory activ-
ity for each series. Introduction of another methyl moi-
ety at the a-position of 6a–b gave 8a–b, respectively, also
As shown in Table 1, C1 and C2 homologation of inhib-
itors 2a–b gave 4a–b and 5a–b, respectively, with a

5068
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
Table 1. Activity profiles of piperidine analogs
Table 3. Activity profiles of bicyclo[3.3.0]octane analogs
H
R1
R2
X
N
R3
O
O
MeO
R2
H
CN
Me
O
Compd
X =
PDE4a
IC50 (nM)
Inhibition of
TNF-ab ID50
(mg/kg, po)
Compd R1
R3 PDE4a
Inhibition of
IC50 (nM) TNF-aa ID50
(mg/kg, po)
2a (Y@OH)
66
0.080
>300
18
(74%)c
0.04
COY
3d
3e
9c
9d
H
COOH
COOH
H
COOH
CN >300
CN >300
NTc
NTc
2
2b (Y@NHOH)
4a (Y@OH)
4b (Y@NHOH)
5a (Y@OH)
–(CH2)2COY
–(CH2)3COY
NEd
NEd
NTe
H
H
H
H
80
26
CONHOH
(37%)b
>300
>100
a See corresponding footnotes from Table 1.
b Inhibition % at 3mg/kg, po.
c Not tested.
5b (Y@NHOH)
NTe
Me
6a (Y@OH)
>300
0.41
NTe
0.4
6b (Y@NHOH)
COY
Me
7a (Y@OH)
7b (Y@NHOH)
>1000
37
NTe
NTe
Table 4. Activity profiles of bicyclo[3.3.0]octene analogs
COY
H
R1
MeMe
COY
8a (Y@OH)
8b (Y@NHOH)
>1000
34
NTe
NTe
O
H
a Inhibition of PDE4 prepared from U937 cells (a cell line derived from
human monocytes). IC50 values represent a mean of n = 2.
b ID50 for inhibition of LPS-induced TNF-a production in rats (n = 7)
0.5h after oral dosing of a test compound.
c Inhibition % at 3mg/kg, po.
R2
MeO
Compd R1
R2
PDE4a
IC50 (nM) ID50 (mg/kg, po)
Inhibition of TNF-aa
d Not effective at 1mg/kg, po.
10a
10b
11a
11b
COOH CN
CONHOH CN
42
0.32
100
2.5
0.6
e Not tested.
(41%)b
(84%)b
0.2
COOH
CONHOH
H
H
a See corresponding footnotes from Table 1.
b Inhibition % at 3mg/kg, po.
Table 2. Activity profiles of bicyclo[3.3.0]octane analogs
H
R1
R2
hydroxamic acid analogs relative to their corresponding
carboxylic acid analogs may be due to the greater affin-
ity of the hydroxamic acid group for the metal ion at the
catalytic site of PDE relative to that of the carboxylic
acid group.18
O
MeO
R2
H
R3
Compd R1
R3 PDE4a
Inhibition of
IC50 (nM) TNF-aa ID50
(mg/kg, po)
Among the four possible isomers (3a, 3c, 3d, and 3e) of
bicyclo[3.3.0]octane with three pharmacophores, 3a
showed the most strong PDE4 inhibitory activity. As
shown in Table 2, relatively more active bicy-
clo[3.3.0]octane analogs 3a–b11 and their corresponding
decyanated analogs 9a–b were synthesized and evalu-
ated. The cyanated analog 3a exhibited more potent
activity than the decyanated analog 9a. In fact, an al-
most 16-fold increase in inhibitory activity was obtained
by the conversion of 9a into 3a. This clearly shows that
the newly introduced benzylic nitrile was useful for pro-
moting the interaction of the inhibitor with the enzyme.
However, the hydroxamic acid analog 3b unexpectedly
showed no increase in inhibitory activity relative to that
of the corresponding carboxylic acid analog 3a. Addi-
tionally, 3b did not show an increased activity relative
to 9b. This suggested that there may be two different
modes of interaction between the inhibitor and the en-
zyme. With regard to the series of carboxylic acid ana-
logs 3a and 9a, the benzylic nitrile plays an important
3a
3b
3c
9a
9b
H
COOH CN
CONHOH CN
9.0
9.7
0.4
(52%)b
3.8
H
COOH
H
CN 150
H
H
COOH
CONHOH
H
H
150
5.8
(79%)b
(80%)b
a See corresponding footnotes from Table 1.
b Inhibition % at 3mg/kg, po.
showing a further decrease in inhibitory activity. Thus,
as the acidic groups became more sterically hindered,
the inhibitory activity decreased. These findings strongly
support our hypothesis that the structures around the
acidic group are critical for the interaction between these
inhibitors and the target enzyme. The SAR data indi-
cated that inhibitory activity is impaired if the interac-
tion of the acidic groups with the enzyme is disturbed
by steric repulsion and/or incorrect positioning. The
marked increase in the inhibitory activity of the

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5069
role in the increase of activity. Alternatively, for the ser-
ies of hydroxamic acid analogs 3b and 9b, the hydroxa-
mic acid group plays the dominant role in the increase
in activity.
seen in Table 4, 10a was more potent than the decya-
nated analog 11a. Hydroxamic acid analog 10b was also
much more potent than expected, presumably because
of the better positioning of the hydroxamic acid group
relative to those of 3b, 9b, and 11b. The smaller increase
in inhibitory activity by conversion of 9a to 9b relative
to conversion of 11a to 11b (25-fold vs 40-fold) was as-
cribed to relatively less favorable positioning of the
hydroxamic acid group.
As shown in Table 3, endo-phenyl isomers 3d–e and their
decyanated analogs 9c–d were synthesized and evalu-
ated. Cyanated analogs 3d and 3e did not show PDE4
inhibitory activity at a concentration of 300nM, regard-
less of the stereochemistry of their carboxylic acid resi-
due, while decyanated analogs 9c and 9d exhibited
IC50 values of 80 and 26nM, respectively. These findings
support our speculation that the benzylic nitrile has an
appropriate interaction site in the enzyme and prevents
an inhibitor from docking into the enzyme if it has unde-
sirable stereochemistry. Regarding decyanated analogs
9c and 9d, which show weaker in vitro activity than
the cyanated analogs 3a–b, respectively, they are consid-
ered to have relatively more flexible interactions with the
enzyme regardless of their stereochemistry.
3.2. In silico docking study
Molecular modeling and graphic manipulations of the
complexes consisting of the enzyme and the inhibitor
were performed using the INSIGHT II (ver. 2000) software
packages, running on a Silicon Graphics OCTANE
work station (Figs. 6A and B, 7A and B and 8A and
B). Proposed main interactions of the PDE4/2a and
PDE4/2b complexes are briefly described in Figure 9.
Model building and conformational analysis were
accomplished with the CHARMm (ver. 23.2) force
field.19 Adopted-Basis Newton Rapson method was ap-
plied for CHARMm minimization step. During the
These analyses are also consistent with the SAR study of
a,b-unsaturated analogs 10a–b and 11a–b. As can be
Figure 6. (A, B) The carboxylic acid residue can coordinate with Mg2+ without distortion of the piperidine ring. A hydrogen bond is formed between
the tertiary amine part of the piperidine ring and the nitrogen atom of the imidazole ring. Replacement of the carboxylic acid residue of 2a with a
hydroxamic acid residue afforded 2b with the increased bidentate coordination with Mg2+. As a result, 2b exhibited a remarkable increase in its
inhibitory activity.
Figure 7. (A, B) Severe distortion of the bicyclo[3.3.0]octane ring of 3a is needed for the carboxylic acid residue to coordinate with Mg2+ because the
carboxylic acid residue has a relatively more hindered endo-stereochemistry. As a result, the corresponding hydroxamic acid analog 3b did not show
increased inhibitory activity.

5070
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
Figure 8. (A, B) The carboxylic acid residue of 10a can coordinate with the Mg2+ without distortion of the bicyclo[3.3.0]octane ring. Thus the
corresponding hydroxamic acid analog 10b can coordinate strongly with Mg2+. Consequently, a remarkable increase (nearly 100-fold) in the
inhibitory activity was obtained.
Figure 9. Scheme of the main interactions observed in the molecular modeling of the PDE4/2a and PDE4/2b complexes.
calculations, the inhibitor atoms and side chain atoms,
which directly interact with the inhibitors were refined,
and other atoms including metal ions and water mole-
cules were fixed at the positions reported in the crystal
structure.20–23
conducted regarding this a-helix as another recognition
site, the presence of which was strongly suggested by the
in vitro SAR study of nitrile or nonnitrile compounds,
as described in Tables 2–4. Docking of the inhibitors
into Apo-PDE4B (PDB ID:1F0J) followed by optimiza-
tion of the amino acid side chains around the recogni-
tion site was performed. The a-helix has a small
hydrophobic pocket consisting of methionine, cysteine,
and valine for the benzylic nitrile to bind. The cyclopen-
tyloxy moiety occupies the region, which can avoid the
methionine residue attached to the a-helix, while the
aromatic moiety occupies the predicted region.
Manipulations regarding crystallographic analysis of
rolipram/PDE4D complex (PDB ID:1Q9M)23 and
zardaverine/PDE4D complex (PDB ID:1OYN)22 were
applied to our in silico docking study because of the
structural analogy among rolipram, zardaverine, and
the new inhibitors (Fig. 1). The aromatic moiety of these
inhibitors was speculated to dock into the same site of
the enzyme as that of rolipram and zardaverine does.
The cyclopentyloxy moiety is slightly rotated from the
corresponding region in the PDE4D/rolipram complex
structure (PDB ID:1OYN).22 This displacement can
avoid the methionine residue attached to the a-helix
while the aromatic moiety occupies the same region as
that in the PDE4D/rolipram complex structure.
Modeling was performed so that the aromatic part of
the inhibitors overlapped the corresponding part of the
reported PDE-bound conformation of rolipram. As a
result, the benzylic nitrile was exposed to the solvent
region.
The acidic residues of the inhibitors were postulated to
coordinate with a metal cation in one of the recognition
sites of the enzyme. However, the most plausible target
metal cation, Zn2+, is already coordinated to the imidaz-
oles of the two histidine residues and the carboxylic acid
of the aspartic acid residue. Thus, Zn2+ is not thought to
be a target metal, which acidic residues can coordinate
with, while another metal cation, Mg2+, which is located
The Apo-PDE4B crystal structure (PDB ID:1F0J) was
reported by Xu et al.,21 and two independent PDEs were
in the crystal structure. In one of these PDE structures,
the substrate binding site was covered by an a-helix. Xu
et al. speculated that the conformation of this helix was
caused by its crystal packing, however our inhibitors are
nicely docked into this structure. A modeling study was

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5071
close to Zn2+ and is loosely bound to aspartic acid res-
idues of the enzyme, is considered to be a possible target
metal instead. The role of this Mg2+ cation is reported in
detail.20 As shown in Figure 6A and B, two oxgen atoms
of the acidic residues of compounds 2a–b are coordi-
nated with Mg2+ ion. As shown in Figure 9, the nitrogen
atoms of piperidines 2a–b were expected to perform the
hydrogen bond to the imidazole of the histidine residue
(His160). Consequently, the conformation of the piper-
idine ring of compound 2a was slightly more distorted
than that of compound 2b. Docking study of 3a–b was
described in Figure 7A and B. Both of the acidic resi-
dues of 3a–b occupy the space, which is not able to coor-
dinate with Mg2+ ion. Docking study of 10a–b was also
conducted as described in Figure 8A and B. Acidic res-
idues of 10a–b are able to occupy closer position to
Mg2+ ion relative to those of 3a–b because of the a,b-
unsaturated flat structures.
inhibitory activity. With regard to carboxylic acid ana-
logs, compounds 9a and 10a exhibited unexpectedly po-
tent in vivo activity in comparison to their relatively
weak in vitro activity.
3.4. Further biological evaluations
Further biological evaluations of 2b and 3a have already
been reported in our previous paper.10,11
Further biological evaluations of 2b, 3a,10,11 10a, and
11b, which were selected based on their oral activity
against LPS-induced TNF-a production assay, were car-
ried out as shown in Table 5. These compounds were
evaluated for their potency to inhibit the slow reacting
substance of anaphylaxis (SRS-A)-mediated broncho-
constriction.25,26 The results were obtained as ID50 val-
ues, that is, the dose that resulted in 50% inhibition
relative to the vehicle. These compounds were also eval-
uated as to the concentration required to inhibit TNF-a
production in human whole blood (HWB)27 to estimate
their clinical potential. The results of the assays were ob-
tained as IC50 values, that is, the test compound concen-
tration that resulted in 50% inhibition relative to the
vehicle. The potency of these compounds for inhibiting
SRS-A-mediated bronchoconstriction in actively sensi-
tized guinea pigs was not always consistent with their
potency for inhibiting LPS-induced TNF-a production
in rats, probably because of differences in the pharmaco-
kinetics due to cross-species comparison. Compound 3a,
the stereochemistry of which is close to that of Ariflo,
showed 50% inhibition of SRS-A-induced bronchocon-
striction at ID50 values of 9.6mg/kg while it showed
an ID50 value for TNF-a production of 0.4mg/kg and
an IC50 value for TNF-a production in HWB of
21lM. Compounds 2b and 11b demonstrated 50% inhi-
bition of SRS-A-induced bronchoconstriction at ID50
values of 0.3 and 0.2mg/kg, respectively, while their
ID50 values for TNF-a production were 0.043 and
0.2mg/kg, respectively. With regard to gastric emptying
in rats,28 the inhibitory activities of 2b and 11b were
weak in comparison to their much more potent IC50
values relative to Ariflo 1 for LPS-induced TNF-a
Thus, modeling and graphic manipulations of the com-
plexes PDE4/2a, PDE4/2b, PDE4/3a, PDE4/3b, PDE4/
10a, and PDE4/10b were performed as shown in Figures
6A and B, 7A and B, and 8A and B, respectively.
3.3. Inhibition of TNF-a production
Test compounds, which were selected based on their
LPDE4 inhibitory activity, were also evaluated for their
potency to inhibit lipopolysaccharide (LPS)-induced
production of tumor necrosis factor-a (TNF-a) in rats.24
The results were expressed as ID50 values, that is, the
dose that resulted in 50% inhibition relative to the vehi-
cle. In contrast to our prediction based on their in vitro
results, a series of hydroxamic acid analogs did not al-
ways exhibit more potency than their corresponding car-
boxylic acid analogs in this assay, presumably because
of their pharmacodynamic problems with oral absorp-
tion, metabolism etc. A series of carboxylic acid analogs
also exhibited results that were distinct from their
LPDE4 inhibitory activity. Hydroxamic acid analogs
2b, 6b, and 11b exhibited relatively potent activity while
other analogs 9b, 9d, and 10b showed unexpectedly
weak potency in comparison to their potent LPDE4
Table 5. Further biological evaluation of 2b, 3a, 10a, and 11b
Compd
Inhibition of
Inhibition of TNF-a
Inhibition of
Inhibition of TNF-a
Ferret emesise
(vomiting/tested)
(mg/kg, po)
bronchoconstrictiona productionb ID50 (mg/kg, po) gastric emptyingc production in HWBd IC50 (lM)
ID50 (mg/kg, po)
ID50 (mg/kg, po)
1
3
10
NTg NTg NTg
2/2
NTg NTg
NTg 0/3
0/3
NTg NTg NTg
0/1 0/2
NTg
1 (Ariflo) 4.5
1.7
0.04
0.4
0.6
0.2
5.7
18
0.0089
21
NTg
0.043
2b
3a
0.3
9.6
7.7
0.2
0.3
(73%)f
(38%)f
(71%)f
10a
11b
a Inhibition of SRS-A-mediated bronchoconstriction and airway microvascular leakage in actively sensitized guinea pigs (n = 3–6); OVA challenge
0.15mg/kg 1h after oral dosing of a test compound.
b See corresponding footnotes from Table 1.
c Inhibition of gastric emptying in rats (n = 5).
d Inhibition of LPS-induced TNF-a production in human whole blood. IC50 values represent a mean of n = 3.
e Vomiting test in fasted ferrets.
f Inhibition % at 10mg/kg, po.
g Not tested.

5072
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
production in HWB. Based on the above-mentioned
biological data, 2b and 11b are likely to have improved
therapeutic potential and an improved side-effect profile.
In a ferret emesis model, which is known as an estab-
lished evaluation method for the side effects of PDE4
inhibitors, 3a and 11b did not cause emesis up to oral
doses of 10 and 3mg/kg, respectively, although 2b
caused emesis at an oral dose of 1mg/kg. Regarding
inhibition of the SRS-A-induced bronchoconstriction
assay, 10a showed the highest ID50 value (7.7mg/kg,
po), which was more than 10-fold less potent than that
of LPS-induced TNF-a production in rats, while it
showed the least potent inhibition against gastric empty-
ing in rats among the compounds 2b, 3a, 10a, and 11b
selected for further evaluation.
(APCI) mass spectra were determined by a Hitachi M-
1200H spectrometer. Matrix assisted laser desorption
ionization-time of flight (MALDI-TOF) mass spectra
were obtained on a PerSeptive Voyager Elite spectrom-
eter. IR spectra were measured using a Perkin–Elmer
FTIR 1760X or JASCO FTIR-430 spectrometer. Ele-
mental analyses were performed with a Perkin–Elmer
PE2400 Series II CHNS/O Analyzer and are only indi-
cated as the elements within 0.4% of the theoretical
values unless otherwise noted. Column chromatography
was carried out using silica gel [Merck silica gel 60
(0.063–0.200mm), Wako Gel C200, Fuji Silysia
FL60D, or Fuji Silysia BW-235]. TLC was also per-
formed on silica gel (Merck TLC plate, silica gel 60
F254).
5.2. Synthesis of compounds 4a–7a and 4b–7b
4. Conclusion
5.2.1. Methyl 3-{4-cyano-4-[3-(cyclopentyloxy)-4-meth-
oxyphenyl]piperidin-1-yl}propanate (13). To a stirred
solution of 12 (0.45g, 1.34mmol) in THF (5.0mL) were
added Et3N (0.38mL, 2.68mmol) and methylacrylate
(0.36mL, 4.0mmol). After being stirred at 45ꢂC for
24h, the reaction mixture was poured into H2O and
extracted with EtOAc. The organic layer was washed
with H2O and then brine, dried over MgSO4, and con-
centrated in vacuo. The residue was purified by column
chromatography on silica gel (n-hexane/EtOAc, 1/1) to
give 13 (0.45g, 1.17mmol, 87%) as colorless oil: TLC
Rf = 0.42, (n-hexane/EtOAc, 2/3); MS (APCI, Pos. 20
In summary, the possible mode of interaction of the new
inhibitors of PDE4 with the target enzyme was discussed
based on both the SAR of the synthesized compounds
and their in silico docking studies. The cyanated analogs
seem to interact with the target enzyme via three phar-
macophores, which are a catechol moiety, a benzylic ni-
trile moiety, and acidic moieties. There seem to be two
different modes of interaction between these inhibitors
and PDE4. With the carboxylic acid analogs, the benz-
ylic nitrile plays a dominant role in the increase in
PDE4 inhibitory activity, while the hydroxamic acid
group plays a dominant role in the hydroxamic acid
analogs. Appropriate spatial positioning and reduced
steric hindrance around the acidic group were found
to be critical for achieving potent inhibitory activity,
as illustrated by the SAR data in Tables 1–4. As illus-
trated in Tables 2 and 3, decyanated carboxylic acid
analogs 9a and 9c, which could not have an interaction
with the enzyme via a benzylic nitrile moiety, appeared
to have a relatively loose interaction with the enzyme
regardless of their stereochemistry. The hydroxamic acid
group, a well-known metal chelator, achieved a striking
increase in PDE4 inhibitory activity in both the analogs.
1
V) m/z = 387 (M+H)+; H NMR (300MHz, CDCl3) d
7.02–6.97 (m, 2H), 6.87–6.83 (m, 1H), 4.83–4.76 (m,
1H), 3.85 (s, 3H), 3.70 (s, 3H), 3.05–2.96 (m, 2H), 2.81
(t, J = 7.2Hz, 2H), 2.59–2.49 (m, 2H), 2.55 (t,
J = 7.2Hz, 2H), 2.12–2.02 (m, 4H), 2.02–1.76 (m, 6H),
1.70–1.50 (m, 2H).
5.2.2. Ethyl 4-{4-cyano-4-[3-(cyclopentyloxy)-4-methoxy-
phenyl]piperidin-1-yl}butanate (14) (Method A). To a
stirred solution of 12 (0.45g, 1.34mmol) in DMF
(10mL) were added K2CO3 (554mg, 4.01mmol) and
ethyl 4-bromobutyrate (0.21mL, 1.47mmol). After
being stirred at 50ꢂC for 24h, the reaction mixture
was poured into H2O and extracted with EtOAc. The
organic layer was washed with H2O and then brine,
dried over MgSO4, and concentrated in vacuo. The res-
idue was purified by column chromatography on silica
gel (n-hexane/EtOAc, 2/1–1/1) to give 14 (0.470mg,
1.13mmol, 85%) as a colorless oil: TLC Rf = 0.48 (n-
hexane/EtOAc, 2/3); MS (APCI, Pos. 20 V) m/z = 415
Further biological evaluation of selected compounds
based on their inhibition of TNF-a production assay
identified another two orally active inhibitors, 2b and
11b, with therapeutic potential.
5. Experimental
1
(M+H)+; H NMR (300MHz, CDCl3) d 7.02–6.98 (m,
5.1. General procedure
2H), 6.87–6.84 (m, 1H), 4.83–4.76 (m, 1H), 4.15 (q,
J = 7.2Hz, 2H), 3.85 (s, 3H), 3.05–2.97 (m, 2H), 2.52–
2.43 (m, 4H), 2.36 (t, J = 7.2Hz, 2H), 2.12–2.03 (m,
4H), 2.01–1.76 (m, 8H), 1.68–1.55 (m, 2H), 1.27 (t,
J = 7.2Hz, 3H).
Analytical samples were homogeneous as confirmed by
thin-layer chromatography (TLC), and yielded spectro-
scopic data consistent with the assigned structures. All
1H NMR spectra were obtained with a Varian Gem-
ini-200 or MERCURY-300 spectrometer. The chemical
shift values are reported in ppm (d) and coupling con-
stants (J) in Hertz (Hz). Fast atom bombardment
(FAB) and electron ionization (EI) mass spectra were
obtained with a JEOL JMS-DX303HF or JMS-700
spectrometer. Atmospheric pressure chemical ionization
5.2.3. Ethyl 2-{4-cyano-4-[3-(cyclopentyloxy)-4-methoxy-
phenyl]piperidin-1-yl}propanate (15). The title compound
was prepared from the corresponding alkylbromide
according to Method A to give a colorless oil: Yield
1
97%; TLC Rf = 0.46 (n-hexane/EtOAc, 2/1); H NMR
(300MHz, CDCl3) d 7.05–6.95 (m, 2H), 6.85 (d,

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5073
J = 8.7Hz, 1H), 4.79 (m, 1H), 4.30–4.15 (m, 2H), 3.84 (s,
3H), 3.37 (q, J = 7.4Hz, 1H), 3.10–2.95 (m, 2H), 2.90–
2.65 (m, 2H), 2.20–2.00 (m, 4H), 2.00–1.80 (m, 6H),
1.75–1.50 (m, 2H), 1.35 (d, J = 7.4Hz, 3H), 1.32 (t,
J = 7.2Hz, 3H).
3.73 (s, 3H), 3.30 (q, J = 7.1Hz, 1H), 4.00–2.70 (br,
1H), 3.00–2.90 (m, 2H), 2.78–2.68 (m, 1H), 2.66–2.56
(m, 1H), 2.13–2.04 (m, 2H), 2.02–1.80 (m, 4H), 1.76–
1.63 (m, 4H), 1.63–1.50 (m, 2H), 1.19 (d, J = 7.1Hz,
3H); Anal. found C21H28N2O4ÆHCl (C, H, N).
5.2.4. Ethyl 2-{4-cyano-4-[3-(cyclopentyloxy)-4-methoxy-
phenyl]piperidin-1-yl}butanate (16). The title compound
was prepared from the corresponding alkylbromide
according to Method A to give a colorless oil: Yield
quant; TLC Rf = 0.55 (n-hexane/EtOAc, 2/1); MS
(APCI, Pos. 20 V) m/z = 415 (M+H)+; 1H NMR
(300MHz, CDCl3) d 7.02–6.98 (m, 2H), 6.88–6.83 (m,
1H), 4.83–4.76 (m, 1H), 4.24 (q, J = 7.4Hz, 2H), 3.84
(s, 3H), 3.16–3.10 (m, 1H), 3.04–2.95 (m, 2H), 2.95–
2.84 (m, 1H), 2.79–2.69 (m, 1H), 2.13–2.01 (m, 4H),
2.00–1.54 (m, 10H), 1.26 (t, J = 7.1Hz, 3H), 0.95 (t,
J = 7.4Hz, 3H).
5.2.8. 2-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}butanoic acid (7a). The title compound
was prepared according to Method B to give a white
powder: Yield 60%; TLC Rf = 0.52 (CHCl3/MeOH, 9/
1); IR (KBr) 3457, 2957, 2231, 1619, 1523, 1467, 1446,
1419, 1374, 1303, 1262, 1152, 1107, 1065, 1022, 957;
1
MS (APCI, Neg. 20 V) m/z = 385 (MꢀH)ꢀ; H NMR
(300MHz, DMSO-d6) d 7.02–6.93 (m, 3H), 4.87–4.80
(m, 1H), 3.72 (s, 3H), 3.31 (br, 1H), 3.05 (t, J = 7.4Hz,
1H), 2.97–2.86 (m, 2H), 2.81–2.70 (m, 1H), 2.64–2.54
(m, 1H), 2.13–2.04 (m, 2H), 1.99–1.80 (m, 4H), 1.75–
1.63 (m, 6H), 1.63–1.51 (m, 2H), 0.87 (t, J = 7.4Hz,
3H); Anal. found C22H30N2O4Æ2/3HClÆ1/3H2O (C, H,
N).
5.2.5. 3-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}propanoic acid (4a) (Method B). To a
stirred solution of 13 (451mg, 1.17mmol) in MeOH
(10mL) was added 1N NaOH (2.33mL, 2.33mmol).
After being stirred at room temperature for 3.5h, the
reaction mixture was acidified with 1N HCl (2.33mL,
2.33mmol), and extracted with EtOAc/THF. The organ-
ic layer was washed with H2O and brine, dried over
MgSO4, and concentrated in vacuo. The residue was
triturated with EtOAc/Et2O/n-hexane to give 4a
(217mg, 0.58mmol, 50%) as a white powder: TLC
Rf = 0.43 (CHCl3/MeOH, 9/1); IR (KBr) 3431, 2957,
2872, 2657, 2570, 2231, 1731, 1608, 1519, 1455, 1416,
1373, 1336, 1258, 1171, 1147, 1133, 1073, 1026, 993;
5.2.9. 3-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}-N-hydroxypropanamide hydrochloride
(4b) (Method C). To a stirred solution of 4a (378mg,
1.01mmol) in DMF (5.0mL) were added Et3N
(0.21mL, 1.5mmol), EDCÆHCl (292mg, 1.52mmol),
HOBt (206mg, 1.52mmol), and (1-methoxy-1-methy-
ethyl)oxyamine (533mg, 5.07mmol). After being stirred
at room temperature for 24h, the mixture was poured
into H2O, and extracted with EtOAc. The organic layer
was washed with brine, dried over MgSO4, and concen-
trated in vacuo. The residue was purified by column
chromatography on silica gel (EtOAc/MeOH, 1/0–20/
1) to give an amide (190mg, 0.410mmol) as a pale yel-
low oil: TLC Rf = 0.50 (EtOAc/MeOH, 15/1).
1
MS (APCI, Neg. 20 V) m/z = 371 (MꢀH)ꢀ; H NMR
(300MHz, DMSO-d6) d 7.03–6.95 (m, 3H), 4.87–4.80
(m, 1H), 3.74 (s, 3H), 3.31 (br, 1H), 3.24–3.19 (m,
2H), 2.96–2.85 (m, 2H), 2.66–2.53 (m, 4H), 2.27–2.20
(m, 2H), 2.16–2.05 (m, 2H), 1.96–1.80 (m, 2H), 1.75–
1.63 (m, 4H), 1.63–1.48 (m, 2H); Anal. found
C21H28N2O4Æ3/2H2O (C, H, N).
To a stirred solution of an amide (190mg, 0.410mmol)
in MeOH (4.0mL) was added 2N HCl (0.25mL,
0.50mmol). After being stirred at room temperature
for 1h, the reaction mixture was concentrated in vacuo.
The residue was triturated with Et2O/MeOH to give 4b
(149mg, 0.350mmol, 35% in two steps) as a white pow-
der: TLC Rf = 0.52 (CHCl3/MeOH, 9/1); IR (KBr) 3466,
3141, 2954, 2665, 2611, 2596, 2243, 1670, 1607, 1594,
1518, 1469, 1442, 1387, 1369, 1337, 1320, 1290, 1247,
1217, 1163, 1139, 1077, 1025, 992; MS (FAB, Pos.) m/
z = 388 (M+H)+; 1H NMR (300MHz, pyridine-
d5+CDCl3) d 7.23 (d, J = 2.2Hz, 1H), 7.11 (dd,
J = 8.5, 2.2Hz, 1H), 6.88 (d, J = 8.5Hz, 1H), 6.53 (br,
3H), 4.86–4.79 (m, 1H), 3.72 (s, 3H), 3.37–3.28 (m,
2H), 3.28 (t, J = 7.2Hz, 2H), 2.93–2.83 (m, 2H), 2.85
(t, J = 7.2Hz, 2H), 2.59–2.48 (m, 2H), 2.15–2.10 (m,
2H), 1.92–1.84 (m, 4H), 1.80–1.65 (m, 2H), 1.54–1.41
(m, 2H); Anal. found C21H29N3O4ÆHClÆ3/4H2O (C, H,
N).
5.2.6. 4-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}butanoic acid (5a). The title compound
was prepared according to Method B to give a white
powder: Yield 77%; TLC Rf = 0.55 (CHCl3/MeOH, 8/
1); IR (KBr) 3449, 2958, 2842, 2236, 1944, 1696, 1593,
1519, 1474, 1445, 1421, 1363, 1336, 1255, 1176, 1134,
1025, 963; MS (APCI, Neg. 20 V) m/z = 385 (MꢀH)ꢀ;
1H NMR (300MHz, DMSO-d6) d 7.03–6.94 (m, 3H),
4.87–4.81 (m, 1H), 3.74 (s, 3H), 3.31 (br, 1H), 3.01–
2.96 (m, 2H), 2.44–2.38 (m, 2H), 2.33–2.21 (m, 4H),
2.13–2.08 (m, 2H), 2.00–1.94 (m, 2H), 1.92–1.83 (m,
2H), 1.76–1.63 (m, 6H), 1.62–1.54 (m, 2H); Anal. found
C22H30N2O4Æ1/4H2O (C, H, N).
5.2.7. 2-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}propanoic acid (6a). The title compound
was prepared according to Method B to give a white
powder: Yield 51%; TLC Rf = 0.67 (CHCl3/MeOH/
AcOH, 15/2/1); IR (KBr) 3439, 2957, 2231, 1615, 1521,
1446, 1397, 1366, 1261, 1151, 1022, 993; MS (APCI,
5.2.10. 4-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}-N-hydroxybutanamide
hydrochloride
(5b). The title compound was prepared according to
Method C to give a white powder: Yield 71% in two
steps; TLC Rf = 0.63 (CHCl3/MeOH, 9/1); IR (KBr)
3437, 3143, 2975, 2817, 2599, 2246, 1656, 1567, 1517,
1469, 1444, 1421, 1357, 1333, 1305, 1255, 1243, 1157,
1
Neg. 20 V) m/z = 371 (MꢀH)ꢀ; H NMR (300MHz,
DMSO-d6) d 7.03–6.94 (m, 3H), 4.88–4.80 (m, 1H),

5074
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
1137, 1071, 1058, 1036, 1022, 1009, 995; MS (FAB, Pos.)
m/z = 402 (M+H)+; 1H NMR (300MHz, pyridine-
d5+CDCl3) d 7.29 (d, J = 2.6Hz, 1H), 7.17 (dd,
J = 8.7, 2.6Hz, 1H), 6.87 (d, J = 8.7Hz, 1H), 6.04 (br,
3H), 4.89–4.83 (m, 1H), 3.72 (s, 3H), 3.32–3.28 (m,
2H), 2.90–2.63 (m, 6H), 2.46 (t, J = 6.9Hz, 2H), 2.26–
2.14 (m, 4H), 1.92–1.85 (m, 4H), 1.80–1.66 (m, 2H),
1.56–1.45 (m, 2H); Anal. found C22H31N3O4ÆHClÆ1/
4H2O (C, H, N).
To a stirred solution of an alcohol (2.24g, 5.38mmol) in
CH2Cl2 (30mL) were added Et3N (1.80mL, 12.9mmol)
and MsCl (0.50mL, 6.45mmol) at 0ꢂC. After being stir-
red at room temperature for 24h, the reaction mixture
was poured into saturated aqueous NaHCO3 and ex-
tracted with CH2Cl2. The organic layer was washed with
H2O and brine, dried over MgSO4, and concentrated
in vacuo. The residue was purified by column chromato-
graphy on silica gel (n-hexane/EtOAc, 6/1) to give 18
(1.99g, 4.99mmol, 56% in two steps) as a pale yellow
oil: TLC Rf = 0.49 (n-hexane/EtOAc, 5/1); MS (APCI,
Pos. 20 V) m/z = 399 (M+H)+; 1H NMR (200MHz,
CD3OD) d 6.99 (d, J = 2.1Hz, 1H), 6.93 (dd, J = 8.4,
2.1Hz, 1H), 6.80 (d, J = 8.4Hz, 1H), 5.93 (m, 1H),
4.83–4.75 (m, 1H), 3.84 (s, 3H), 3.53 (s, 2H), 3.33 (m,
1H), 2.97–2.80 (2H, H), 2.51–2.37 (m, 3H), 1.98–1.54
(m, 8H), 0.97 (s, 6H).
5.2.11. 2-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}-N-hydroxypropanamide hydrochloride
(6b). The title compound was prepared according to
Method C to give a white powder: Yield 76% in two
steps; TLC Rf = 0.44 (CHCl3/MeOH, 10/1); IR (KBr)
3439, 3118, 2960, 2720, 2239, 1671, 1640, 1518, 1443,
1417, 1393, 1362, 1266, 1157, 1126, 1038, 991; MS
(APCI, Pos. 20 V) m/z = 388 (M+H)+; 1H NMR
(300MHz, pyridine-d5+CDCl3) d 7.20 (d, J = 2.3Hz,
1H), 7.09 (dd, J = 8.3, 2.3Hz, 1H), 6.94 (d, J = 8.3Hz,
1H), 6.45 (br, 3H), 4.82–4.75 (m, 1H), 3.74 (s, 3H),
3.53 (q, J = 6.9Hz, 1H), 3.23–3.02 (m, 3H), 2.93–2.82
(m, 1H), 2.26–2.10 (m, 4H), 1.95–1.65 (m, 6H), 1.53–
1.41 (m, 2H), 1.47 (d, J = 6.9Hz, 3H); HRMS (FAB)
calcd for C21H30N3O4 388.2236, found 388.2238.
5.3.2. (10S,30S,50R)-5,5-Dimethyl-30-(3-cyclopentyloxy-4-
methoxyphenyl)spiro[1,3-dioxine-2,70-bicyclo[3.3.0]octane]
(19). A solution of 18 (683mg, 1.71mmol) in dioxane
(16mL) was hydrogenated under atmospheric pressure
of H2 gas in the presence of 10% Pd/C (68mg) for 3h.
The catalyst was removed by filtration through a pad
of Celite, and washed with MeOH. The filtrates were
concentrated in vacuo. The residue was purified by col-
umn chromatography on silica gel (n-hexane/EtOAc, 7/
1) to give 19 (689mg, 1.71mmol, 100%) as a white pow-
der: TLC Rf = 0.65 (n-hexane/EtOAc, 7/1); MS (FAB,
5.2.12. 2-{4-Cyano-4-[3-(cyclopentyloxy)-4-methoxyphen-
yl]piperidin-1-yl}-N-hydroxybutanamide
hydrochloride
(7b). The title compound was prepared according to
Method C to give a white powder: Yield 56% in two
steps; TLC Rf = 0.54 (CHCl3/MeOH, 9/1); IR (KBr)
3438, 3107, 2957, 2876, 2750, 2713, 2239, 1672, 1605,
1590, 1519, 1460, 1447, 1420, 1362, 1318, 1258, 1197,
1172, 1136, 1106, 1032, 993; MS (FAB, Pos.)
m/z = 402 (M+H)+; 1H NMR (300MHz, pyridine-
d5+CDCl3) d 7.20 (d, J = 2.4Hz, 1H), 7.10 (dd,
J = 8.5, 2.4Hz, 1H), 6.92 (d, J = 8.5Hz, 1H), 6.35 (br,
3H), 4.83–4.76 (m, 1H), 3.72 (s, 3H), 3.34–3.15 (m,
4H), 3.06–2.96 (m, 1H), 2.27–2.00 (m, 5H), 1.95–1.62
(m, 7H), 1.54–1.39 (m, 2H), 0.99 (t, J = 7.2Hz, 3H);
Anal. found C22H31N3O4ÆHCl (C, H, N).
1
Pos.) m/z = 401 (M+H)+; H NMR (200MHz, CDCl3)
d 6.82–6.72 (m, 3H), 4.76 (m, 1H), 3.81 (s, 3H), 3.51
(s, 4H), 2.93 (m, 1H), 2.68–2.48 (m, 2H), 2.31–2.14 (m,
4H), 2.00–1.73 (m, 8H), 1.66–1.39 (m, 2H), 0.98 (s, 6H).
5.3.3. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]hexahydro-2(1H)-pentalenone (20). To a stirred solu-
tion of 19 (2.49g, 6.22mmol) in acetone (60mL) was
added TsOHÆH2O (118mg, 0.62mmol). After being stir-
red at room temperature for 24h, the reaction mixture
was poured into saturated aqueous NaHCO3 and ex-
tracted with EtOAc. The organic layer was washed with
H2O and brine, dried over Na2SO4, and concentrated in
vacuo. The residue was purified by column chromato-
graphy on silica gel (n-hexane/EtOAc, 5/1) to give 20
(1.85g, 5.89mmol, 95%) as a pale yellow oil: TLC
Rf = 0.48 (n-hexane/EtOAc, 3/1); MS (APCI, Pos. 20
5.3. Synthesis of compounds 9c–d
5.3.1. An enantiomeric mixture (18) of (10S,50R)-5,5-
dimethyl-30-(3-cyclopentyloxy-4-methoxyphenyl)spiro-
[1,3-dioxine-2,70-bicyclo[3.3.0]octan-20-en] and (10R,50S)-
5,5-dimethyl-30-(3-cyclopentyloxy-4-methoxyphenyl)spiro-
[1,3-dioxine-2,70-bicyclo[3.3.0]octan-20-en]. The follow-
ing reaction was carried out under argon atmosphere.
To a stirred solution of 3-cyclopentyloxy-4-methoxy
phenylbromide (5.66g, 20.9mmol) in THF (80mL) were
added n-BuLi (1.53M solution in n-hexane, 14.5mL,
22.3mmol) and CeCl3 (5.48g, 22.3mmol) at ꢀ78ꢂC.
To this suspension was added a solution of 17 (3.12g,
13.9mmol) in THF (60mL) at ꢀ78ꢂC. After being stir-
red at ꢀ78ꢂC for 3h, the reaction mixture was quenched
with brine, and extracted with EtOAc. The organic layer
was washed with brine, dried over anhydrous MgSO4,
and concentrated in vacuo. The residue was purified
by column chromatography on silica gel (n-hexane/
EtOAc, 5/1) to give an alcohol (3.48g, 8.35mmol) as a
pale yellow oil: TLC Rf = 0.39 (n-hexane/EtOAc, 2/1).
1
V) m/z = 315 (M+H)+; H NMR (300MHz, CDCl3) d
6.82–6.72 (m, 3H), 4.76 (m, 1H), 3.82 (s, 3H), 3.10 (m,
1H), 2.91–2.76 (m, 2H), 2.63–2.52 (m, 2H), 2.49–2.38
(m, 2H), 2.19–2.11 (m, 2H), 1.95–1.76 (m, 6H), 1.68–
1.53 (m, 2H), 1.44 (m, 2H).
5.3.4. An enantiomeric mixture (21) of (3aR,5R,6aR)-5-
[3-(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,6,6a-hex-
ahydro-2-pentalenecarbonitrile and (3aS,5S,6aS)-5-[3-
(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,6,6a-hexa-
hydro-2-pentalenecarbonitrile (Method D). To a stirred
solution of 20 (1.41g, 4.47mmol) in CH2Cl2 (60mL)
were added ZnI2 (143mg, 0.447mmol) and TMSCN
(577mg, 5.81mmol). After being stirred at room temper-
ature for 2h, the mixture was concentrated in vacuo.
The residue was used for the next step without further
purification.

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5075
1
To a stirred solution of cyanohydrin in pyridine (20mL)
was added P(O)Cl3 (1.03mL, 11.2mmol). After being
stirred at 120ꢂC for 1.5h, the reaction mixture was
poured into ice water, and extracted with EtOAc. The
organic layer was washed with H2O and brine, dried
over MgSO4, and concentrated in vacuo. The residue
was purified by column chromatography on silica gel
(n-hexane/EtOAc, 4/1) to give 21 (1.31g, 4.05mmol,
91% in two steps) as a white powder: TLC Rf = 0.62
(n-hexane/EtOAc, 2/1); MS (APCI, Pos. 20 V) m/
(M+Na)+, 359 (M)+; H NMR (300MHz, DMSO-d6)
d 10.35 (s, 1H), 8.64 (br, 1H), 6.82 (d, J = 8.4Hz, 1H),
6.78 (d, J = 1.8Hz, 1H), 6.72 (dd, J = 8.4, 1.8Hz, 1H),
4.76 (m, 1H), 3.68 (s, 3H), 3.02 (m, 1H), 2.60–2.40 (m,
3H), 2.17–2.07 (m, 2H), 1.98–1.76 (m, 4H), 1.76–1.61
(m, 4H), 1.61–1.44 (m, 4H), 1.40–1.28 (m, 2H).
5.4. Synthesis of compounds 9a–b and 11a–b
1
z = 324 (M+H)+; H NMR (300MHz, CDCl3) d 6.81–
5.4.1. An enantiomeric mixture (23) of (3aR,6aS)-5-[3-
(cyclopentyloxy)-4-methoxyphenyl]-3,3a,4,6a-tetrahydro-
2(1H)-pentalenone and (3aS,6aR)-5-[3-(cyclopentyloxy)-
4-methoxyphenyl]-3,3a,4,6a-tetrahydro-2(1H)-pentalenone.
To a stirred solution of 18 (4.61g, 11.6mmol) in acetone
(116mL) was added TsOHÆH2O (220mg). After being
stirred at room temperature for 3.5h, the reaction mix-
ture was poured into saturated aqueous NaHCO3, and
extracted with EtOAc. The organic layer was washed
with H2O and brine, dried over Na2SO4, and concen-
trated in vacuo. The residue was purified by column
chromatography on silica gel (n-hexane/EtOAc, 3/1) to
give 23 (3.24g, 10.4mmol, 90%) as a pale yellow oil:
TLC Rf = 0.48 (n-hexane/EtOAc, 3/1); MS (APCI, Pos.
6.77 (m, 1H), 6.75–6.70 (m, 2H), 6.59 (m, 1H), 4.77
(m, 1H), 3.82 (s, 3H), 3.39 (m, 1H), 2.99–2.77 (m, 3H),
2.48–2.24 (m, 3H), 1.96–1.76 (m, 6H), 1.68–1.54 (m,
2H), 1.43–1.31 (m, 2H).
5.3.5. An enantiomeric mixture (22) of (3aR,5R,6aR)-5-
[3-(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,6,6a-hex-
ahydro-2-pentalenecarboxylic acid and (3aS,5S,6aS)-5-
[3-(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,6,6a-hexa-
hydro-2-pentalenecarboxylic acid (Method E). To a stir-
red solution of 21 (941mg, 2.91mmol) in ethylene glycol
(6.5mL) was added 40% aqueous KOH (5.4mL,
58.2mmol). After being stirred at 200ꢂC for 4h, the
reaction mixture was washed with Et2O, neutralized by
2N HCl, and extracted with EtOAc. The organic layer
was washed with H2O and brine, dried over MgSO4,
and concentrated in vacuo. The residue was triturated
with EtOAc to give 22 (635mg, 1.85mmol, 64%) as a
pale powder: TLC Rf = 0.46 (CHCl3/MeOH, 10/1); IR
(KBr) 3436, 2944, 1981, 1626, 1517, 1424, 1292, 1241,
1143, 1034, 850; MS (APCI, Neg. 20 V) m/z = 341
20 V) m/z = 313 (M+H)+; H NMR (300MHz, CDCl3)
1
d 6.99 (d, J = 2.1Hz, 1H), 6.93 (dd, J = 8.4, 2.1Hz,
1H), 6.82 (d, J = 8.4Hz, 1H), 5.89 (m, 1H), 4.86–4.76
(m, 1H), 3.85 (s, 3H), 3.60 (m, 1H), 3.16–3.01 (m, 2H),
2.63–2.47 (m, 3H), 2.33 (m, 1H), 2.14–2.04 (m, 1H),
1.97–1.81 (m, 6H).
5.4.2. An enantiomeric mixture (24) of (2S,3aR,6aS)-5-[3-
(cyclopentyloxy)-4-methoxyphenyl]-1,2,3,3a,4,6a-hexa-
hydro-2-pentalenol and (2R,3aS,6aR)-5-[3-(cyclopentyl-
oxy)-4-methoxyphenyl]-1,2,3,3a,4,6a-hexahydro-2-pen-
talenol. To a stirred solution of 23 (3.24g, 10.4mmol)
and THF (275mL) in MeOH (275mL) was added
NaBH4 (395mg, 10.4mmol). After being stirred at 0ꢂC
for 1h, the reaction mixture was quenched with acetone
and H2O. The mixture was extracted with EtOAc and
washed with H2O and brine. The organic layer was dried
over anhydrous MgSO4, and concentrated in vacuo to
give 24 (3.27g, 10.4mmol, 100%) as a pale yellow oil:
TLC Rf = 0.67 (n-hexane/EtOAc, 1/1); MS (APCI, Pos.
1
(MꢀH)ꢀ; H NMR (200MHz, CDCl3) d 6.86 (m, 1H),
6.82–6.70 (m, 1H), 4.77 (m, 1H), 3.81 (s, 3H), 3.39 (m,
1H), 2.91 (m, 1H), 2.90–2.75 (m, 2H), 2.50–2.35 (m,
2H), 2.30 (m, 1H), 1.98–1.70 (m, 7H), 1.70–1.50 (m,
2H), 1.45–1.27 (m, 2H).
5.3.6. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]octahydropentalene-2-carboxylic acid (9c) (Method
F). A solution of 22 (670mg, 1.96mmol) in MeOH
(16mL) was hydrogenated under atmospheric pressure
of H2 gas in the presence of 10% Pd/C (70mg) for 4h.
The catalyst was removed by filtration through a pad
of Celite, and washed with MeOH. The filtrates were
concentrated in vacuo. The residue was purified by col-
umn chromatography on silica gel (CHCl3/MeOH, 100/
1) to give 9c (689mg, 1.11mmol, 54%) as a white pow-
der: TLC Rf = 0.78 (CHCl3/MeOH, 10/1); IR (KBr)
3439, 2960, 1698, 1516, 1467, 1421, 1303, 1246, 1161,
1139, 1036, 1001, 855; MS (APCI, Neg. 20 V)
m/z = 343 (MꢀH)ꢀ; 1H NMR (300MHz, CDCl3) d
6.81–6.73 (m, 3H), 4.81–4.74 (m, 1H), 3.82 (s, 3H),
3.08 (m, 1H), 2.92 (m, 1H), 2.66–2.55 (m, 2H), 2.30–
2.19 (m, 4H), 1.97–1.78 (m, 6H), 1.78–1.37 (m, 7H);
Anal. found C21H28O4Æ1/4H2O (C, H).
20 V) m/z = 315 (M+H)+; H NMR (300MHz, CDCl3)
1
d 7.00 (d, J = 2.0Hz, 1H), 6.94 (dd, J = 8.4, 2.0Hz,
1H), 6.81 (d, J = 8.4Hz, 1H), 6.02 (m, 1H), 4.79 (m,
1H), 4.23 (m, 1H), 3.84 (s, 3H), 3.40–3.30 (m, 2H),
3.04 (m, 1H), 2.85 (m, 1H), 2.63 (m, 1H), 2.20 (m,
1H), 2.07 (m, 1H), 2.00–1.76 (m, 6H), 1.69–1.53 (m, 4H).
5.4.3. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]octahydro-2-pentalenol (25). A solution of 24 (39mg,
0.098mmol) in CH2Cl2 (690mL) was hydrogenated un-
der atmospheric pressure of H2 gas in the presence of
Ir(cod)py(PCy3)PF6 (16mg, 0.020mmol) for 1h. The
reaction mixture was concentrated in vacuo. The residue
was purified by column chromatography on silica gel
(n-hexane/EtOAc, 3/2) to give 25 (28mg, 0.087mmol,
89%) as a pale yellow oil: TLC Rf = 0.64, (n-hexane/
5.3.7. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]-N-hydroxyoctahydropentalene-2-carboxamide
(9d).
The title compound was prepared according to Method
C to give a pale pink powder: Yield 75%; TLC Rf = 0.48
(CHCl3/MeOH, 10/1); IR (KBr) 3432, 2946, 1629, 1515,
1464, 1265, 1137, 1030; MS (FAB, Pos.) m/z = 382
1
EtOAc, 1/1); H NMR (300MHz, CDCl3) d 6.84–6.75
(m 3H), 4.78 (m, 1H), 3.95 (m, 1H), 3.76 (s, 3H), 3.15

5076
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
(m, 1H), 2.58–2.45 (m, 2H), 2.21–2.11 (m, 2H), 1.88–
1.56 (m, 12H), 1.27–1.16 (m, 2H).
2.34–2.24 (m, 2H), 1.96–1.78 (m, 8H), 1.75–1.45 (m,
7H); Anal. found C21H28O4 (C, H).
5.4.9. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]-N-hydroxyoctahydropentalene-2-carboxamide
5.4.4. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]hexahydro-2(1H)-pentalenone (26). To a stirred solu-
tion of 25 (1.3g, 4.1mmol) in CH2Cl2 (20mL) were
added NMO (1.44g, 12.3mmol), TPAP (2.6g,
8.2mmol), and MS4A (1g). After being stirred at room
temperature for 24h, the reaction mixture was filtrated
with column chromatography on silica gel (CH2Cl2)
and purified by column chromatography on silica gel
(n-hexane/EtOAc, 4/1) to give 26 (1.18g, 3.76mmol,
92%) as a white powder: TLC Rf = 0.57, (n-hexane/
(9b).
The title compound was prepared according to Method
C to give a pale pink powder: Yield 70%; TLC Rf = 0.49
(CHCl3/MeOH, 10/1); IR (KBr) 3210, 2940, 1630, 1515,
1442, 1263, 1160, 1138, 1034, 998; MS (FAB, Pos.)
1
m/z = 360 (M+H)+; H NMR (300MHz, DMSO-d6) d
10.36 (s, 1H), 8.67 (s, 1H), 6.84–6.80 (m, 2H), 6.74
(dd, J = 8.4, 1.8Hz, 1H), 4.77 (m, 1H), 3.68 (s, 3H),
3.01 (m, 1H), 2.60–2.40 (m, 1H), 2.29–2.16 (m, 2H),
2.00–1.89 (m, 2H), 1.89–1.77 (m, 2H), 1.77–1.50 (m,
10H), 1.44–1.31 (m, 2H); Anal. found C21H29NO4Æ
1/2H2O (C, H, N).
1
EtOAc, 2/1); H NMR (300MHz, CDCl3) d 6.82–6.73
(m, 3H), 4.77 (m, 1H), 3.82 (s, 3H), 3.18 (m, 1H),
3.03–2.92 (m, 2H), 2.65–2.54 (m, 2H), 2.16–2.08 (m,
2H), 2.06–1.76 (m, 10H), 1.67–1.55 (m, 2H).
5.5. Synthesis of compounds 10a–b
5.4.5. An enantiomeric mixture (27) of (3aR,5S,6aR)-5-
[3-(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,6,6a-hex-
ahydro-2-pentalenecarbonitrile and (3aS,5R,6aS)-5-[3-
(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,6,6a-hexahy-
dro-2-pentalenecarbonitrile. The title compound was pre-
pared according to Method D to give a white powder:
Yield 55% in two steps.
5.5.1. An enantiomeric mixture (29) of (3aS,5S,6aS)-5-[3-
(cyclopentyloxy)-4-methoxyphenyl]-5-cyano-1,3a,4,5,6,
6a-hexahydropentalen-2-yl trifluoromethanesulfonate and
(3aR,5R,6aR)-5-[3-(cyclopentyloxy)-4-methoxyphenyl]-5-
cyano-1,3a,4,5,6,6a-hexahydropentalen-2-yl trifluorome-
thanesulfonate. The following reaction was carried out
under argon atmosphere. To a stirred solution of 28
(587mg, 1.73mmol) in THF (10mL) was added
LiHMDS (1.0M solution in THF, 1.73mL, 1.73mmol)
at ꢀ78ꢂC. After being stirred at ꢀ78ꢂC for 10min,
Tf2NPh (618mg, 1.73mmol) was added. After being
stirred at 0ꢂC for 30min, the reaction mixture was di-
luted with brine, and extracted with EtOAc. The organic
layer was washed with brine, dried over anhydrous
MgSO4, and concentrated in vacuo to give 29 (1.54g)
as a yellow oil: TLC Rf = 0.68 (n-hexane/EtOAc, 2/1);
1H NMR (200MHz, CDCl3) d 7.00–6.75 (m, 3H), 5.67
(m, 1H), 4.79 (m, 1H), 3.84 (s, 3H), 3.50–3.30 (m, 2H),
3.00–2.80 (m, 2H), 2.70–2.40 (m, 2H), 2.40–2.20 (m,
2H), 2.00–1.70 (m, 6H), 1.70–1.50 (m, 2H).
5.4.6. (3aR,5S,6aR)-5-[3-(Cyclopentyloxy)-4-methoxy-
phenyl]-1,3a,4,5,6,6a-hexahydropentalene-2-carboxylic
acid (11a). The title compound was prepared according
to Method E to give a white powder: Yield 93%; TLC
Rf = 0.56 (CHCl3/MeOH, 10/1); IR (KBr) 3433, 2933,
1686, 1630, 1518, 1421, 1270, 1139, 1032, 801; MS
(APCI, Neg. 20 V) m/z = 341 (MꢀH)ꢀ; 1H NMR
(300MHz, CDCl3) d 6.82–6.71 (m, 4H), 4.76 (m, 1H),
3.82 (s, 3H), 3.52 (m, 1H), 3.10–2.89 (m, 3H), 2.39 (m,
1H), 2.00–1.74 (m, 10H), 1.68–1.53 (m, 3H); Anal.
found C21H26O4 (C, H).
5.4.7. (3aR,5S,6aR)-5-[3-(Cyclopentyloxy)-4-methoxy-
phenyl]-N-hydroxy-1,3a,4,5,6,6a-hexahydropentalene-2-
carboxamide (11b). The title compound was prepared
according to Method C to give a pale brown powder:
Yield 56%; TLC Rf = 0.64 (CHCl3/MeOH, 10/1); IR
(KBr) 3424, 2934, 1655, 1514, 1261, 1136, 1030, 795;
MS (FAB, Pos.) m/z = 358 (M+H)+; 1H NMR
(300MHz, DMSO-d6) d 10.52 (br, 1H), 8.79 (br, 1H),
6.81 (d, J = 8.3Hz, 1H), 6.77 (d, J = 1.7Hz, 1H), 6.71
(dd, J = 8.3, 1.7Hz, 1H), 6.19 (m, 1H), 4.76 (m, 1H),
3.68 (s, 3H), 3.40–3.26 (m, 1H), 2.96–2.70 (m, 3H),
2.24 (m, 1H), 1.88–1.61 (m, 10H) 1.61–1.48 (m, 2H);
Anal. found C21H27NO4Æ1/4H2O (C, H, N).
5.5.2. An enantiomeric mixture (30) of methyl
(3aS,5S,6aS)-5-[3-(cyclopentyloxy)-4-methoxyphenyl]-5-
cyano-1,3a,4,5,6,6a-hexahydropentalen-2-carboxylate and
methyl (3aR,5R,6aR)-5-[3-(cyclopentyloxy)-4-methoxy-
phenyl]-5-cyano-1,3a,4,5,6,6a-hexahydropentalen-2-car-
boxylate. To a stirred solution of 29 (1.54g) in DMF/
MeOH (2/1, 6mL) were added PPh3 (27.3mg,
1.04mmol), Pd(OAc)2 (11.7mg, 0.0519mmol), and
Et3N (0.48mL, 3.5mmol). After being stirred at room
temperature for 19h under CO atmosphere, the reaction
mixture was diluted with H2O and extracted with
EtOAc/n-hexane. The organic layer was washed with
H2O, dried over MgSO4, and concentrated in vacuo.
The residue was purified by column chromatography
on silica gel (n-hexane/EtOAc, 9/1–3/1) to give 30
(621mg, 1.63mmol, 94% in two steps) as a yellow oil:
TLC Rf = 0.50 (n-hexane/EtOAc, 2/1); MS (APCI, Pos.
5.4.8. (3aR,6aS)-5-[3-(Cyclopentyloxy)-4-methoxyphen-
yl]octahydropentalene-2-carboxylic acid (9a). The title
compound was prepared according to Method F to give
a white powder: Yield 81%; TLC Rf = 0.59 (CHCl3/
MeOH, 10/1); IR (KBr) 2936, 1699, 1588, 1517, 1443,
1357, 1321, 1269, 1226, 1200, 1161, 1137, 1029, 850;
1
40 V) m/z = 382 (M+H)+; H NMR (300MHz, CDCl3)
d 6.95–6.85 (m, 2H), 6.83 (d, J = 8.4Hz, 1H), 6.68 (m,
1H), 4.79 (m, 1H), 3.84 (s, 3H), 3.76 (s, 3H), 3.51 (m,
1H), 3.00–2.80 (m, 2H), 2.65–2.45 (m, 3H), 2.40–2.10
(m, 2H), 2.00–1.75 (m, 6H), 1.70–1.50 (m, 2H).
1
MS (APCI, Neg. 20 V) m/z = 343 (MꢀH)ꢀ; H NMR
(300MHz, CDCl3) d 6.82–6.74 (m, 3H), 4.80–4.73 (m,
1H), 3.82 (s, 3H), 3.10 (m, 1H), 2.73–2.57 (m, 3H),

H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
5077
5.5.3. An enantiomeric mixture (10a) of (3aS,5S,6aS)-5-
cyano-5-[3-(cyclopentyloxy)-4-methoxyphenyl]-1,3a,4,5,
6,6a-hexahydropentalene-2-carboxylic acid and (3aR,5R,
6aR)-5-cyano-5-[3-(cyclopentyloxy)-4-methoxyphenyl]-
1,3a,4,5,6,6a-hexahydropentalene-2-carboxylic acid. To a
stirred solution of 30 (488mg, 1.28mmol) in THF/
MeOH (1/2, 7.5mL) was added 2N NaOH (2.0mL,
4.0mmol). After being stirred at room temperature for
1.5h, the reaction mixture was concentrated and washed
Et2O. The water layer was acidified with 2N HCl, and
extracted with EtOAc. The organic layer was dried over
Na2SO4 and concentrated in vacuo. The residue was trit-
urated with Et2O to give 10a (400mg, 1.09mmol, 85%)
as a white powder: TLC Rf = 0.43 (CHCl3/MeOH, 9/
1); IR (KBr) 2961, 2871, 2594, 2232, 1687, 1634, 1519,
1517, 1467, 1444, 1416, 1343, 1299, 1254, 1145, 1132,
1026, 982; MS (APCI, Neg. 40 V) m/z = 366 (MꢀH)ꢀ;
1H NMR (300MHz, CDCl3) d 6.95–6.85 (m, 2H),
6.90–6.80 (m, 2H), 4.79 (m, 1H), 3.84 (s, 3H), 3.55 (m,
1H), 3.00–2.80 (m, 2H), 2.70–2.50 (m, 3H), 2.32 (dd
J = 13.5, 5.7Hz, 1H), 2.22 (dd, J = 12.9, 7.2Hz, 1H),
2.05–1.70 (m, 7H), 1.70–1.55 (m, 2H); HRMS (EI) calcd
for C22H25NO4 367.1784, found 367.1771.
5.7. Inhibition of LPS-induced plasma TNF-a production
in rats
Male Crj:CD(SD)IGS rats aged 6weeks (n = 7) were
fasted overnight, and the test compounds (0.01–0.1mg/
10mL/kg) were administered orally at 1h before intrave-
nous injection of 1lg/kg of LPS (Escherichia coli Sero-
type 055 B5). The plasma TNF-a level was measured
with a commercially available ELISA kit (RD Systems)
at 90 min after LPS challenge. The percent inhibition
(the dosage required to inhibit plasma TNF-a produc-
tion by 50%) was determined by the following formula:
% Inhibition ¼ 100 ꢀ ðC ꢀ SÞ=ðL ꢀ SÞ ꢁ 100
C: Plasma TNF-a concentration in LPS-treated animals
pretreated with a test compound, L: plasma TNF-a con-
centration in LPS-treated animals pretreated with saline,
S: plasma TNF-a concentration in saline-treated ani-
mals pretreated with saline.
5.8. SRS-A mediated bronchoconstriction in guinea pigs
Male Hartley guinea pigs aged 7weeks (n = 5) were ac-
tively sensitized by intraperitoneal administration of
1mg of ovalbumin (OVA) containing 5 · 109 killed Bor-
detella pertussis organisms on day 0. On day 14, the bron-
choconstrictor response was measured using a modified
version of the method of Konzett and Ro¨ssler. Broncho-
constriction was induced by an intravenous injection of
OVA (0.15–0.5mg/kg). Sensitized animals were treated
with both a cyclooxygenase inhibitor (indomethacin at
5mg/kg i.v., 3min before OVA) and an antihistamine
(pyrilamine at 1mg/kg i.v., 1min before OVA) to ensure
that endogenous SRS-A was solely responsible for bron-
choconstriction. Test compounds were administered or-
ally at 1h before antigen challenge. Bronchoconstrictor
response was measured for 15min and the result was rep-
resented as the area under the curve (AUC 0–15min).
5.5.4. An enantiomeric mixture (10b) of (3aS,5S,6aS)-5-
cyano-5-[3-(cyclopentyloxy)-4-methoxyphenyl]-N-hydroxy-
1,3a,4,5,6,6a-hexahydropentalene-2-carboxamide
and
(3aR,5R,6aR)-5-cyano-5-[3-(cyclopentyloxy)-4-methoxy-
phenyl]-N-hydroxy-1,3a,4,5,6,6a-hexahydropentalene-2-
carboxamide. The title compound was prepared accord-
ing to Method C to give a white powder: Yield 56%;
TLC Rf = 0.44 (CHCl3/MeOH, 10/1); IR (KBr) 3208,
2957, 2229, 1644, 1605, 1517, 1444, 1416, 1258, 1143,
1020, 979; MS (APCI, Neg. 40 V) m/z = 381 (MꢀH)ꢀ;
1H NMR (300MHz, CDCl3) d 8.80–8.20 (br, 1H),
6.95–6.85 (m, 2H), 6.83 (d, J = 9.3Hz, 1H), 6.42 (m,
1H), 4.79 (m, 1H), 3.84 (s, 3H), 3.59 (m, 1H), 3.10–
2.90 (m, 2H), 2.65–2.20 (m, 5H), 2.00–1.80 (m, 6H),
1.70–1.55 (m, 3H); HRMS (FAB) calcd for
C22H27N2O4 383.1971, found 38.1996.
5.9. Gastric emptying in rats
5.6. Assay of human PDE4 activity
Male Sprague-Dawley rats were fasted overnight and
were orally administered test compounds or 0.5 w/v%
methylcellulose (10mL/kg). In addition, 0.05mg/mL of
phenol red solution was orally administered in a volume
of 1.5mL at 20min after dosing with the test com-
pounds. Forty minutes after administration of the test
compounds, both the cardia and pylorus of the stomach
were ligated under anesthesia with sodium pentobarbital
(75mg/kg, i.p.), and then the stomach was isolated with-
out leakage of phenol red. The stomach was cut open
and the phenol red solution was drained into a beaker
containing 100mL of 0.1N NaOH. Part of the solution
was filtrated (pore size: 0.45lm) and the absorbance at
546nm was measured to determine the amount of dye
remaining in the stomach. Then the gastric emptying
rate was calculated by the following formula:
The method of Reeves et al.29 was modified to isolate
phosphodiesterase type 4 isozyme (PDE4). The enzyme
was prepared from U937 cells derived from human
monocytes, and was stored at ꢀ20ꢂC after preparation.
Measurement of PDE4 activity was performed using this
stored enzyme after it was diluted with distilled water
containing bovine serum albumin. The substrate solu-
3
tion was prepared by adding H-cAMP (300,000dpm
(5000Bq)/assay) and 100lmol/L cAMP solution to
100mmol/L Tris–HCl (pH8.0) containing 5mmol/L eth-
ylene glycol-bis (b-aminoethyl ether) and O,O0-bis(2-
aminoethyl)ethyleneglycol-N,N,N0,N0-tetraacetic acid.
The substrate solution was mixed with the enzyme solu-
tion containing a test compound dissolved in dimethyl-
sulfoxide (DMSO), and incubation was done for
30min at 30ꢂC. Assays were performed in duplicate at
three to four different concentrations of each test com-
pound, and the IC50 values were determined.
Gastric emptying rate ¼ 100 ꢁ ð0:75
ꢀ concentration of dye in the stomachÞ=0:75

5078
H. Ochiai et al. / Bioorg. Med. Chem. 12 (2004) 5063–5078
8. Hele, D. J.; Belvisi, M. G. Exp. Opin. Investig. Drugs 2003,
12, 5.
9. Giembycz, M. A. Exp. Opin. Investig. Drugs 2001, 10,
1361.
10. Ochiai, H.; Ohtani, T.; Ishida, A.; Kusumi, K.; Kato, M.;
Kohno, H.; Kishikawa, K.; Obata, T.; Nakai, H.; Toda,
M. Bioorg. Med. Chem. Lett. 2004, 14, 207.
11. Ochiai, H.; Ohtani, T.; Ishida, A.; Kishikawa, K.; Obata,
T.; Nakai, H.; Toda, M. Bioorg. Med. Chem. Lett. 2004,
14, 1323.
A value of 0.75lg/mL was equal to a concentration of
0.05mg/mL, which was achieved by adding 1.5mL of
phenol red to 100mL of 0.1N NaOH. The 50% inhibi-
tion rate for gastric emptying by the test compounds
was calculated by defining gastric emptying after vehicle
administration as 100%.
5.10. Inhibitory activity on LPS-induced TNF-a
production in human whole blood
12. Vaulont, I.; Gais, H. J.; Reuter, N.; Schmitz, E.; Ossenk-
amp, R. K. L. Eur. J. Org. Chem. 1998, 805.
13. Shibasaki, M.; Ueda, J.; Ikegami, S. Tetrahedron Lett.
1979, 20, 433.
14. Stork, G.; Kahne, D. E. J. Am. Chem. Soc. 1983, 105,
1072.
15. Evans, D. A.; Morrissey, M. M. J. Am. Chem. Soc. 1984,
106, 3866.
16. Lautens, M.; Rovis, T. Tetrahedron 1999, 55, 8967.
17. Torphy, T. J.; Zhou, H.; Cieslinski, L. B. J. Pharmacol.
Exp. Ther. 1992, 263, 1195.
18. Kleinman, E. F.; Campbell, E.; Giordano, L. A.; Cohan,
V. L.; Jenkinson, T. H.; Cheng, J. B.; Shirley, J. T.;
Pettipher, E. R.; Salter, E. D.; Hibbs, T. A.; Dicapua, F.
M.; Bordner, J. J. Med. Chem. 1998, 41, 266.
19. Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D.
J.; Swaminathan, S.; Karplus, M. J. Comput. Chem. 1983,
4, 187.
Under the supervision of a physician, blood was col-
lected into a heparinized tube (final concentration:
10U/mL heparin sodium) from a forearm vein in three
healthy male donors. A solution of the test compound
(10lL) dissolved in DMSO was added to 180lL of
whole blood, and the mixture was pre-incubated for
30min at 37ꢂC. Then 10lL of 2lg/mL of LPS was
added and incubated for 6h at 37ꢂC, after which the
plasma TNF-a concentration was measured with a
human TNF-a ELISA kit (DIACLONE). Assays were
performed in duplicate at three to four different concen-
trations of each test compound, and the IC50 values
were determined.
5.11. Ferret emetic study
20. Liu, S.; Laliberte, F.; Bobechko, B.; Bartlett, A.; Lario, P.;
Gorseth, E.; Hamme, J. V.; Gresser, M. J.; Huang, Z.
Biochemistry 2001, 40, 10179.
21. Xu, R. X.; Hassell, A. M.; Vanderwall, D.; Lambert, M.
H.; Holmes, W. D.; Luther, M. A.; Rocque, W. J.;
Milburn, M. V.; Zhao, Y.; Ke, H.; Nolte, R. T. Science
2000, 288, 1822.
Male ferrets (weighting about 1.2kg) were fasted over-
night and test compounds were administered orally.
Their behavior was observed throughout a 1h period
after gavage. Results were expressed as the number of
animals that vomited relative to the animals tested.
22. Lee, M. E.; Markowitz, J.; Lee, J. O.; Lee, H. FEBS Lett.
2002, 530, 53.
23. Huai, Q.; Wang, H.; Sun, Y.; Kim, H. Y.; Liu, Y.; Ke, H.
Structure 2003, 11, 865.
References and notes
24. Tracey, K. J.; Cerami, A. Annu. Rev. Med. 1994, 45, 491.
25. Nakagawa, N.; Obata, T.; Kobayashi, T.; Okada, Y.;
Nambu, F.; Terawaki, T.; Aishita, H. Jpn. J. Pharmacol.
1992, 60, 217.
26. Nakagawa, N.; Obata, T.; Kobayashi, T.; Okada, Y.;
Nambu, F.; Terawaki, T.; Furuya, T.; Muryobayashi, K.;
Sawada, M.; Aishita, H. Eur. J. Pharmacol. 1993, 235,
211.
27. Brideau, C.; Staden, C. V.; Sthyler, A.; Rodger, I. W.;
Chan, C. C. Br. J. Pharmacol. 1999, 126, 979.
28. Andres, J. I.; Alonso, J. M.; Diaz, A.; Fernandez, J.;
Iturrino, L.; Martinez, P.; Matesanz, E.; Freyne, E. J.;
Deroose, F.; Boeckx, G.; Petit, D.; Diels, G.; Megens, A.;
Somers, M.; Wauwe, J. V.; Stoppie, P.; Cools, M.; Clerck,
F. D.; Peeters, D.; Chaffoy, D. D. Bioorg. Med. Chem.
Lett. 2002, 12, 653.
1. Sondering, S. H.; Beavo, J. A. Curr. Opin. Cell Biol. 2000,
12, 174.
2. Houslay, M. D. Prog. Nucleic Acid Res. Mol. Biol. 2001,
69, 249.
3. Torphy, T. J. Am. J. Respir. Crit. Care Med. 1998, 157,
351.
4. Essayan, D. M. J. Allergy Clin. Immunol. 2001, 108, 671.
5. Teixeira, M. M.; Gristwood, R. W.; Cooper, N.; Helle-
well, P. G. Trends Pharmacol. Sci. 1997, 18, 164.
6. Doherty, A. M. Curr. Opin. Chem. Biol. 1999, 8, 65.
7. Christensen, S. B.; Guider, A.; Forster, C. F.; Gleason, J.
G.; Bender, P. E.; Karponski, J. M.; Dewolf, W. E.;
Barnette, M. S.; Underwood, D. C.; Griswold, D. E.;
Cieslinski, L. B.; Burman, M.; Bochnowicz, S.; Osborn, R.
R.; Manning, C. D.; Grous, M.; Hillegas, L. M.; Bartus, J.
O.; Ryan, M. D.; Eggleston, D. S.; Haltiwanger, R. C.;
Torphy, T. J. J. Med. Chem. 1998, 41, 821.
29. Reeves, M. L.; Leigh, B. K.; England, P. J. Biochem. J.
1987, 241, 535.