Design, synthesis, and structure-activity relationships of spirolactones bearing 2-ureidobenzothiophene as acetyl-CoA carboxylases inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.08.117

The co-crystal structure of the human acetyl-coenzyme A 2 (ACC2) carboxyl transferase domain and the reported compound CP-640186 (1b) suggested that two carbonyl groups are essential for potent ACC2 inhibition. By focusing on enhancing the interactions between the two carbonyl groups and the amino acid residues Gly²¹⁶² and Glu²²³⁰, we used ligand- and structure-based drug design to discover spirolactones bearing a 2-ureidobenzothiophene moiety

Full text of DOI:10.1016/j.bmcl.2011.08.117

Bioorganic & Medicinal Chemistry Letters 21 (2011) 6314–6318
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis, and structure–activity relationships of spirolactones
bearing 2-ureidobenzothiophene as acetyl-CoA carboxylases inhibitors
Tohru Yamashita a, , Makoto Kamata a, Satoshi Endo a, Mitsuo Yamamoto a, Keiko Kakegawa a,
⇑
Hiroyuki Watanabe a, Katsuhiko Miwa a, Toru Yamano a, Masaaki Funata a, Jyun-ichi Sakamoto a,
Akiyoshi Tani a, Clifford D. Mol b, Hua Zou b, Douglas R. Dougan b, BiChing Sang b, Gyorgy Snell b,
Kohji Fukatsu a
a Pharmaceutical Research Division, Takeda Pharmaceutical Company, Ltd, 2-17-85, Jusohonmachi, Yodogawa-ku, Osaka 532-8686, Japan
b Takeda San Diego, Inc., 10410 Science Center Drive, San Diego, CA 92121, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 June 2011
Revised 22 August 2011
Accepted 29 August 2011
Available online 6 September 2011
The co-crystal structure of the human acetyl-coenzyme A 2 (ACC2) carboxyl transferase domain and the
reported compound CP-640186 (1b) suggested that two carbonyl groups are essential for potent ACC2
inhibition. By focusing on enhancing the interactions between the two carbonyl groups and the amino
acid residues Gly2162 and Glu2230, we used ligand- and structure-based drug design to discover spirolac-
tones bearing a 2-ureidobenzothiophene moiety
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Acetyl-CoA carboxylase
Spirolactone
2-Ureidobenzothiophene
Obesity is being widely recognized as a serious health problem
in recent years, due in part to increased food intake in conjunction
with reduced physical activity. The number of obese adults is 475
million, and more than twice number of adults is overweight. Not
only aged, but also younger individuals are facing this problem,
and 200 million children are overweight. A clear necessity exists
for novel therapeutic agents that can regulate body weight and
obesity.1
Acetyl-CoA carboxylases (ACC) catalyze the biotin-dependent
carboxylation of acetyl-CoA to form malonyl-CoA, a key intermedi-
ate metabolite that regulates the rate of fatty acid metabolism.2
Malonyl-CoA has 2 important functions. It is a substrate for fatty
acid synthase in de novo lipogenesis as well as an allosteric inhib-
itor of carnitine palmitoyltransferase 1, a mitochondrial outer
membrane protein that shuttles long-chain fatty acyl-CoAs into
the mitochondria for oxidation.3 Therefore, ACC inhibition is ex-
pected to lower malonyl-CoA levels, resulting in reduced fatty acid
synthesis, increased fatty acid oxidation, and improvement of
obesity as well as insulin sensitivity. In rodents and humans, there
are 2 ACC isoforms encoded as distinct genes. ACC1, a 265 kDa
cytosolic protein, is primarily expressed in lipogenic tissues (liver
and adipose), and ACC2, a 277 kDa protein located in the mito-
chondrial outer membrane, is highly expressed in oxidative tissues
(muscle, heart, and liver).4 Suppression of ACC1 or ACC2 expression
in rat liver using antisense oligonucleotides increases fatty acid
oxidation in rat hepatocytes, and suppression of both enzymes
using a single antisense oligonucleotide is significantly more effec-
tive in promoting fat oxidation.5 These studies suggest that dual
inhibition of ACC1 and ACC2 is a novel approach to treat obesity
and insulin resistance, particularly type 2 diabetes.6
Harwood and coworkers reported a class of potent and nonselec-
tive ACC1/2 dual inhibitors as exemplified by CP-497485 (1a)7 and
the co-crystal structure of the human ACC2 carboxyl transferase
(CT) domain in complex with CP-640186 (1b) was modeled
(Fig. 1).8 This data clarified that the carbonyl group of 3-piperidine-
carboxamide (= nipecotamide) (right side) interacts with the back-
bone of Gly2162 and that the carbonyl group (left side) is bound to
the backbone of Glu2230. Co-crystal data also indicated that the
Glu2230 side chain is located near the anthracene ring, suggesting
an additional interaction. From this data, the proper fixation of the
carbonyl group direction on the right side led to a stronger hydrogen
bonding interaction with Gly2162. Moreover, conversion of the
anthracene ring, which is able to utilize the side chain of Glu2230
,
might lead to the formation of new hydrogen bonds between a li-
gand and the carboxylate of Glu2230, increasing ACC2 inhibitory
activity.9 Because the binding site on the human ACC1 CT domain
is very similar to that of ACC2, the inhibitory activity of ACC1 should
be well correlated with that of ACC2. Herein, we report the discovery
of a spirolactone derivative 26h bearing a 2-ureidobenzothiophene
⇑
Corresponding author. Tel.: +81 6 6300 6117.
E-mail address: Yamashita_Tohru@takeda.co.jp (T. Yamashita).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.08.117

T. Yamashita et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6314–6318
6315
Me
N
N
Me
N
O
O
O
O
CP-497485 1a
(
)
O
N
N
N
CP-640186 (1b)
Figure 1. Chemical structure of CP-497485 (1a) and CP-640186 (1b) and co-crystal structure of the human ACC2 CT domain in complex with CP-640186 (1b).
ring formed by focusing interactions with Gly2162 and Glu2230 to in-
crease inhibitory activity.
10 through deprotection of the Cbz group and subsequent conden-
sation of antheracene-9-carboxylic acid. The strictly planar com-
pound was synthesized as illustrated in Scheme 2. Introduction of
tert-butyl piperazine-1-carboxylate into 12, followed by catalytic
hydrogenation of the cyano group over 10% Pd–C under heat affor-
ded isoindolone derivative 14. Alkylation of 14 with iodoethane
gave intermediate 15, which was deprotected under acidic condi-
tions and subsequently acylated with anthracene-9-carboxylic acid
to produce 2. Synthesis of spirolactams and lactones is outlined in
Scheme 3. Cynanomethylation or cyanoethylation of ester 16, fol-
lowed by reduction of the cyano group afforded spirolactams,
which was alkylated using iodoethane to give 18a–b. Finally, the
desired spirolactams 3a–b were obtained by deprotection of the
Boc group followed by reductive amination with 23, prepared from
anthracene-9-carboxylic acid and a 4-piperidone derivative. Intro-
duction of an allyl group into nipecotate 16, followed by oxidative
cleavage of the olefin using ozone afforded ketone derivative 19.
Subsequent reduction using NaBH4 or alkylation using methyl Grig-
nard reagent directly afforded spirolactones 20a–b. Deprotection of
20a–b under acidic conditions followed by reductive amination
with 23 afforded spirolactones 21a–b.
Designs based on the nipecotamide and anthracene regions are
shown in Figure 2. For the nipecotamide region, spatial orientation
of the carbonyl group was fixed as coplanar or perpendicular by
constructing a condensed bicyclic (2) or a spiro ring (3). For con-
version of the anthracene region, we designed 2-acylaminobicyclic
systems and 2-ureidobicyclic systems to introduce additional
interactions with the side chain of Glu2230. The docking studies of
compounds 4–7 with ACC2 enzyme revealed that naphthalene 4
or benzofuran 5 was not suitable for the interaction with the side
chain of Glu2230. On the other hand, the plane of benzothiophene,
including the acetamide at 2-position (6), was coplanar with that
of anthracene due to the spatial anchorage by the S–O interac-
tion.10 This suggested that the methyl group of acetamide 6 could
be close to the side chain of Glu2230; therefore, the conversion of
the 2-acetamide group (6) to a 2-ureido group (7) could induce
the formation of an additional interaction.
Conversion of the nipecotamide moiety was accomplished
according to the methods described in Schemes 1–3. As shown in
Scheme 1, incorporation of benzyl piperazine-1-carboxylate into
the fluorobenzene derivative 8 followed by reduction of the nitro
group using zinc powder and subsequently deamination via the
Sandmeyer reaction of the resulting aniline unexpectedly afforded
monoethylamide derivative 9, which was then alkylated using
iodoethane to produce 10. We speculated that unexpected dealky-
lation was occurred by radical reaction during cupper-mediate
reduction of the diazo compound. Compound 11 was prepared from
Benzothiophene derivatives were synthesized as shown in
Scheme 4. Amine 24, which was easily prepared from 20b, was
acylated with various benzothiophene-3-carboxylic acids to afford
25a–c. The amine 25c was converted to 26a–g with acylating re-
agents or sulfonylating reagent.
We initially examined the effects on the carbonyl group of nipe-
cotamide; these results are summarized in Table 1. Replacement of
O
N
N
N
N
Me
or
O
N
N
O
N
O
3
2
Me
O
O
N
O
S
S
N
O
O
NH
Me
O
NH
Me
O
NH
NH2
O
NH
Me
NH
O
O
R
4
5
6
7
Figure 2. Modification of nipecotamide and candidates as alternates for the anthracene region.

6316
T. Yamashita et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6314–6318
Me
Me
Me
NO2
Me
Me
N
H
N
d, e
N
Me
c
N
a, b
Me
N
N
N
O
N
O
F
N
O
N
O
Cbz
Cbz
O
9
8
10
11
Scheme 1. Reagents and conditions: (a) Benzyl piperazine-1-carboxylate, K2CO3, DMSO; (b) (i) Zn, AcOH–H2O; (ii) NaNO2, H2SO4, EtOH; (iii) Cu powder, 78% (four steps); (c)
EtI, NaH, THF, 49%; (d) H2/Pd–C, THF; (e) anthracene-9-carboxylic acid, WSC, HOBt, DMF, 72% (two steps).
N
CN
CN
a
b
c
d, e
Me
N
NH
N
Me
OMe
O
N
N
N
OMe
N
F
O
O
N
N
N
O
Boc
Boc
O
O
Boc
12
13
14
15
2
Scheme 2. Reagents and conditions: (a) tert-Butyl piperazine-1-carboxylate, K2CO3, DMSO, 46%; (b) H2, Pd–C, THF, heat, 34%; (c) EtI, NaH, DMF, 71%; (d) 4 M HCl–EtOAc; (e)
anthracene-9-carboxylic acid, WSC, HOBt, Et3N, DMF, 63% (two steps).
b, c
a
CN
d
N
n
N
n
N
OEt
Boc
d
N
n
Boc
Boc
N
N
N
O
O
O
16
O
OEt
Me
Me
O
17a
18a b
3a b
-
(n = 1)
17b (n = 2)
-
OH
OH
R
Me
h
g
e, f
N
Me
R
O
N
N
Boc
Me
Boc
HN
O
N
O
O
N
O
O
20a
O
EtO
19
HCl
O
(R = H)
20b (R = Me)
O
21a b
-
22
23
Scheme 3. Reagents and conditions: (a) LHMDS, bromoacetonitrile or 3-bromopropionitrile, THF, ꢀ78 °C, 24% (17a), 31% (17b); (b) H2 (0.5 MPa), Ra–Ni, aq NH3, MeOH; (c)
NaH, EtI, THF, 60 °C, 70% (18a), 6.7% (18b); (d) (i) 4NHCl–AcOEt; (ii) 23, NaBH(OAc)3, Et3N, THF, 3.1% (3a), 39% (3b), 19% (21b), 40% (21b); (e) (i) LHMDS, 3-Bromo-2-
methylpropene, THF, ꢀ78 °C; (f) O3, EtOAc, ꢀ78 °C, 99%; (g) NaBH4, THF–MeOH, 36%; or MeMgBr, THF, 51% (diastereomixture); (h) anthracene-9-carboxylic acid, WSC, HOBt,
Et3N, DMF.
Me
Me
Me
Me
c
N
a, b
N
Me
Me
N
S
Boc
O
N
O
O
HN
2HCl
O
O
O
R
25a
(R = H)
25b (R = Me)
25c
O
20b
24
d
e
a
R =
: cHexNHCONH
f: nPrSO2NH
: Et2NCONH
: MeCONH
b: H2NCONH
26a g
-
(R = NH2)
g
c
: MeNHCONH
d: EtNHCONH
Scheme 4. Reagents and conditions: (a) (i) 4NHCl, AcOEt; (ii) 4-Boc-pipreridone, NaBH(OAc)3, Et3N, CH2Cl2; (b) (i) 4NHCl, AcOEt, 99%; (c) various benzothiophene-3-
carboxylic acids, WSC, HOBt, Et3N, DMF, 81% (25a), 91% (25b), 74% (25c); d) AcCl, Et3N, THF–CH2Cl2 (for 26a); Cl3CCONCO, THF then NH3 in MeOH (for 26b); RNCO, pyridine
(for 26c–e); nPrSO2Cl, Et3N, THF–CH2Cl2 (for 26f); Et2NCOCl, Et3N, THF–CH2Cl2 (for 26g).
the piperidine ring of 1a with a benzene ring (11) resulted in equi-
potent activity from the superposition of the structures. However,
the conversion to isoindolone derivative 2, which was cyclized
from 11, significantly decreased potency. This potency decrease
was determined from the superposition of the compounds as
shown in Figure 3 (left), which suggested that the lactam ring moi-
ety of 2 fits poorly with the enzyme, clashing with an amino acid
residue (Ser2161). On the other hand, five-membered spirolactam
3a and lactone 21a showed equipotent activity to 1a. These results
imply that spiro-fixation is preferable (Fig. 3, right), indicating that
the spirolactam moiety markedly overlaps with the nipecotamide
moiety of 1b. Moreover, six-membered ring 3b showed slightly de-
creased activity, which suggested that the direction of the carbonyl
was unfavorable compared with that of the five-membered ring.
These results showed that the fixed carbonyl group of the spiro
ring is effective in increasing the inhibitory activity. Incorporation
of an additional methyl group into 21a led to 21b, which exhibited
the most potent inhibitory activity against human ACC2, with an
IC50 value of 43 nM. Methyl groups on the lactone ring appeared
to be in preferred position and packed in the hydrophobic pocket
of Val2128, Leu1965, and Leu1969 at the morphorine region of 1b.
As expected, the inhibitory activity of ACC1 was correlated with
that of ACC2.
Next, our efforts were focused on the conversion of the anthra-
cene moiety (Table 2). Nonsubstituted benzothiophene 25a, methyl
benzothiophene 25b, and amino benzothiophene 25c were weaker
than the anthracene derivative 21b. The acetamide derivative 26a
showed more potent activity than 25c. This indicates that the
acylated amino group (acetamide) could interact with the side chain
of Glu2230 more tightly than a nonsubstituted amine (vs 25c).
Moreover, as expected, the ureas 26b–d showed potent inhibitory
activities compared with 25c. Introduction of a bulky group to the

T. Yamashita et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6314–6318
6317
Table 1
Table 2
SAR of nipecotamide derivatives
Conversion of the anthracene moiety; benzothiophene derivatives
A
Me
Me
N
X
S
N
O
N
O
O
R
O
Compound
A
X
IC50 (nM)
ACC1
Compound
R
IC50 (nM)
ACC2
ACC1
ACC2
O
25a
25b
225c
26a
26b
26c
26d
26e
26f
H
Me
NH2
MeCONH
H2NCONH
MeNHCONH
EtNHCONH
cHexNHCONH
nPrSO2NH
>100,000
4300
>10,000
2000
84
95
63
220
>10,000
>10,000
880
360
750
100
15
15
13
45
1000
1200
N
N
N
1b (CP-640186)
1a (CP-497485)
CH2
CH2
550
720
34
R
O
O
O
Me
N
N
Me
110
Me
Me
26g
Et2NCONH
11
2
N
N
1000
220
21b
—
330
43
Me
Me
*
N
N
>10,000
8000
S
Me
O
N
O
O
O
NH
O
NH
N
N
Me
3a
3b
CH2
1900
3600
140
280
N
N
O
O
Me
Compound
Steriochemistry
IC50 (nM)
ACC1
32
>10,000
63
ACC2
26h
26i
S
R
5.4
1600
13
CH2
26d
Racemate
Me
hydrogen bonds of the 2 NH groups play an important role in the po-
tency of ACC1/2 inhibitory activity. In addition, steric hindrance of
the diethyl group in 26g is a negative factordecreasing the inhibitory
activity. The stereospecificity of 26d on the inhibitory activity was
investigated, by utilizing each optical isomer, to find (S)-form
(26h)12 was more potent than (R)-form (26i). The X-ray co-crystal
structure model of 3-ethylurea 26h and the human ACC2 CT domain
was determined, and tight hydrogen bonds between the 2 NH
groups of the urea and the carboxylate of Glu2230 were observed
(Fig. 4). The importance of the stereochemistry was also supported
by co-crystal structure models, which were only determined for
(S)-form. In addition, the co-crystal structure model provides a clear
rationale for a strong interaction between the carbonyl group of the
N
N
21a
21b
CH2
CH2
1100
330
110
43
Me
O
O
O
O
Me
Me
Steriochemistry: racemate or none except 1b.
30-position of urea resulted in
(cyclohexyl urea 26e), and the incorporation of a sulfonamide group
decreased the activity (26f vs 26d). The 3,3-diethylurea 26g showed
lower activity than the 3-monoethylurea 26d, indicating that the
a slightly decreased activity
Figure 3. Superposition of the structures of 2 (left, blue) and 3a11 (right, blue) with 1b (yellow): piperidine anthranyl amide (left part) was fixed using 1b.

6318
T. Yamashita et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6314–6318
ACC1/2 dual inhibitor 26h, which exhibits excellent inhibitory activ-
ity against ACC2 (IC50 = 5.4 nM) and ACC1 (IC50 = 32 nM).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.08.117.
References and notes
1. Turkoglu, C.; Duman, B. S.; Gunay, D.; Cagatay, P.; Ozcan, R.; Buyukdevrim, A. S.
Obes. Surg. 2003, 13, 699.
2. McGarry, J. D.; Foster, D. W. Annu. Rev. Biochem. 1980, 49, 395.
3. McGarry, J. D.; Mannaerts, G. P.; Foster, D. W. J. Clin. Invest. 1977, 60, 265.
4. (a) Abu-Elheiga, L.; Jayakumar, A.; Baldini, A.; Chirala, S. S.; Wakil, S. J. Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 4011; (b) Abu-Elheiga, L.; Almarza-Ortega, D. B.;
Baldini, A.; Wakil, S. J. J. Biol. Chem. 1997, 272, 10669; (c) Abu-Elheiga, L.;
Brinkley, W. R.; Zhong, L.; Chirala, S. S.; Woldegiorgis, G.; Wakil, S. J. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, 1444; (d) Munday, M. R. Biochem. Soc. Trans. 2002, 30,
1059.
5. Savage, D. B.; Choi, C. S.; Samuel, V. T.; Liu, Z. X.; Zhang, D.; Wang, A.; Zhang, X.
M.; Cline, G. W.; Yu, X. X.; Geisler, J. G.; Bhanot, S.; Monia, B. P.; Shulman, G. I. J.
Clin. Invest. 2006, 116, 817.
6. Harwood, H. J., Jr. Curr. Opin. Investig. Drugs 2004, 5, 283.
7. Harwood, H. J., Jr.; Petras, S. F.; Shelly, L. D.; Zaccaro, L. M.; Perry, D. A.;
Makowski, M. R.; Hargrove, D. M.; Martin, K. A.; Tracey, W. R.; Chapman, J. G.;
Magee, W. P.; Dalvie, D. K.; Soliman, V. F.; Martin, W. H.; Mularski, C. J.;
Eisenbeis, S. A. J. Biol. Chem. 2003, 278, 37099.
8. Madauss, K. P.; Burkhart, W. A.; Consler, T. G.; Cowan, D. J.; Gottschalk, W. K.;
Miller, A. B.; Short, S. A.; Tran, T. B.; Williams, S. P. Acta Crystallogr., Sect. D 2009,
65, 449.
9. There are only six amino acid differences between human ACC1 and ACC2 in CT
domain and none of which interacted with our ACC1/2 dual inhibitors.
10. Nagao, Y.; Hirata, T.; Goto, S.; Sano, S.; Kakehi, A.; Iizuka, K.; Shir, M. J. Am.
Chem. Soc. 1998, 120, 3104.
Figure 4. Crystal structure of human ACC2 CT domain in complex with compound
26h.
spirolactone and Gly2162. Following conversion of the anthracene
region, ACC1 inhibitory activity was correlated with that of ACC2.
In conclusion, lead generation and an initial SAR study of ACC1/2
dual inhibitors have been performed based on the crystal structure
model of the human ACC2 CT domain with 1b. We focused on
enhancing the interactions between the two carbonyl groups and
the amino acid residues Gly2162 and Glu2230. A co-crystal structure
model with 26h showed that proper fixation of the direction with
respect to the carbonyl group of nipecotamide strengthened the
interaction, leading to the discovery of a novel spirolactone that
interacts with Gly2162. In addition, the benzothiophene moiety,
bearing a ureido group at the 2-position, was found to be a promising
structure aimed at utilizing the side chain of Glu2230. Spatial man-
ners were overlapping with the anthracene ring due to the spatial
anchorage contributed by an S–O interaction. The combination of
the spirolactone moiety on the right region and the 2-ureidobenzo-
thiophenyl group on the left region led to the discovery of a novel
11. (S)-form was used for the docking study.
12. After Cbz-protection of piperidine 24, optical resolution using the chiral HPLC
method was performed. Deprotection of Cbz group followed by acylation with
4-bromobenzoic acid afforded 4-bromobenzyl amide as single crystals whose
stereochemistry was determined using X-ray single crystal structural analysis.