Discovery of novel, potent, and orally active spiro-urea human glucagon receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2005.06.101

A novel class of spiro-ureas has been discovered as potent human glucagon receptor antagonists in both binding and functional assays. Preliminary studies have revealed that compound 15 is an orally active human glucagon receptor antagonist in a transgenic

Full text of DOI:10.1016/j.bmcl.2005.06.101

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4564–4569
Discovery of novel, potent, and orally active spiro-urea
human glucagon receptor antagonists
Dong-Ming Shen,a,* Fengqi Zhang,a Edward J. Brady,b Mari Rios Candelore,c
Qing Dallas-Yang,c Victor D.-H. Ding,c Jasminka Dragovic,a William P. Feeney,d
Guoquiang Jiang,c Peggy E. McCann,b Steve Mock,a Sajjad A. Qureshi,c
Richard Saperstein,b Xiaolan Shen,d Constantin Tamvakopoulos,a Xinchun Tong,a
Laurie M. Tota,c Michael J. Wright,c Xiaodong Yang,c Song Zheng,a Kevin T. Chapman,a
Bei B. Zhang,c James R. Tataa and Emma R. Parmeea
aDepartment of Basic Chemistry, Merck Research Laboratories, PO Box 2000, RY50G-146, Rahway, NJ 07065, USA
bDepartment of Pharmacology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
cDepartment of Metabolic Disorders and Molecular Endocrinology, Merck Research Laboratories, PO Box 2000,
Rahway, NJ 07065, USA
dDepartment of Laboratory Animal Resources, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 24 May 2005; revised 24 June 2005; accepted 29 June 2005
Available online 15 August 2005
Dedicated to the loving memory of a pioneering chemist, inspiring mentor, and generous friend, Prof. Orville L. Chapman.
Abstract—A novel class of spiro-ureas has been discovered as potent human glucagon receptor antagonists in both binding and
functional assays. Preliminary studies have revealed that compound 15 is an orally active human glucagon receptor antagonist in
a transgenic murine pharmacodynamic model at 10 and 30 mpk. Compound 15 is orally bioavailable in several preclinical species
and shows selectivity toward cardiac ion channels and other family B receptors, such as hGIP1 and hGLP.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Type 2 diabetes mellitus (T2DM) is a complex disease1
with increasing prevalence and cost, especially in devel-
oped countries.2 Although numerous therapies are
available for the treatment of diabetes, all of them are
associated with strong side effects. Therefore, many
new drug targets are still being actively pursued.3,4 Glu-
cagon is a 29 AA peptide hormone, which acts through
the glucagon receptor (GCGR) to stimulate gluconeo-
genesis and glycogenolysis, thereby counteracting the
role of insulin in the regulation of glucose homeostasis.
Since hyperglycemia and, paradoxically, increased
glucagon levels are both characteristic of T2DM,
antagonism of the hGCGR is a logical target for the
treatment of T2DM.5–7 A recent report on the blockade
of glucagon-induced hyperglycemia in human subjects
with a small molecule hGCGR antagonist has undoubt-
edly stimulated more interest in this area.8
Recently, trisubstituted ureas, such as 1, were reported
to be human glucagon receptor antagonists.9 We wished
to explore spiro-cyclic variants of these trisubstituted
ureas as potential novel human glucagon receptor
antagonists.10,11
AM1 computer modeling of compound 1 with the three
peripheral substituents truncated for ease of computa-
tion showed conformers A and B to have very similar
energies.12 A quick visual inspection revealed that a
slightly lower energy conformer B was well-suited for
the formation of a spiro-urea moiety, which appeared
to preserve the orientation of the three substituents con-
nected to the original acyclic urea core. AM1 modeling
confirmed this conjecture (Fig. 1). It showed that a
[6.5] spiro-urea overlapped very well with the acyclic
urea having the three carbon atoms marked with arrows
Keywords: Glucagon receptor antagonists.
*
Corresponding author. Tel.: +1 732 594 7303; fax: +1 732 594
8080; e-mail: dongming_shen@merck.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.06.101

D.-M. Shen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4564–4569
4565
OCF3
OCF3
O
O
t-Bu
N
N
N
N
H
H
H
N
HO
H
OH
N
O
O
Conformation A
Conformation B
O
t-Bu
O
1
pounds 8 and 9. The cis isomer of the aminonitrile 4
from a Strecker reaction13 of ketone 3 was added to
an isocyanate, cyclized in situ under basic conditions
to give a 5-imino-imidazolidin-2-one intermediate, and
hydrolyzed to give the hydantoin 5.14 The 4-carbonyl
group in 5 was selectively reduced to a methylene (6)
using LiAlH4–AlCl3.15 Acid 7 was then obtained from
6 by alkylation of the urea NH and deprotection of
the tert-butyl ester. Finally, standard coupling condi-
tions were used to yield amino tetrazole 9 or, following
deprotection, b-alanine 8. Hydantoin analogs 10 and
11 were prepared from intermediate 5 in a similar
manner.16
The preparation of the corresponding [6.6] spiro com-
pounds 14–17 was also performed, as outlined in
Scheme 1, except in this case the required aminonitrile
13a was prepared by a Horner–Emmons reaction17, fol-
lowed by 1,4-addition of ammonia18 (Scheme 2). The
two isomers of 13 were separated on silica gel and their
identities were established by NMR using homonuclear
decoupling and NOE spectroscopy.
Figure 1. Overlay of truncated compound 1 (gray) with [6.5] spiro-urea
replacement (blue) from AM1 calculations (red = O, dark blue = N,
others = C). Hydrogen atoms were omitted in this drawing for clarity.
˚
in 1 placed within 0.6 A of the corresponding atoms in
the spiro-urea replacement. The carbonyl oxygen atoms
˚
in the two compounds were also only 0.5 A apart.
Similar results were obtained with a [6.6] spiro-urea
replacement having an ethylene moiety bridging the urea
NH and the cyclohexane in 1.
The in vitro biological activity of these compounds on
the hGCGR was evaluated in binding and functional as-
says and compared with 1 and its 5-aminotetrazole
amide analog 2 (Table 1). The binding IC50 values were
measured from the inhibition of 125I-glucagon to the
hGCGR expressed in CHO cell membranes. For com-
pounds showing good binding affinity, functional inhibi-
tion of glucagon-induced cAMP accumulation in
These computational results encouraged us to prepare
and test these spiro-ureas as potential novel glucagon
receptor antagonists. The synthesis of the [6.5] spiro-
urea is shown in Scheme 1 with the preparation of com-
OCF3
OCF3
OCF3
CN
NH2
O
a
b, c
d
e, f
N
N
t-Bu
t-Bu
O
N
N
O
O
O
t-Bu
N
H
N
t-Bu
t-Bu
CO2H
3
4
H
6
7
5
OCF3
OCF3
g
g
f
7
N
N
N
N
N
CO2t-Bu
NH
N
NH
N
H2N
O
O
H2N
CO2H
N
HN
HN
hydrochloride
N
t-Bu
t-Bu
N
monohydrate
8
9
O
O
Scheme 1. Reagents and conditions: (a) KCN, NH4Cl, 1:1 MeOH and H2O, rt, 83%; (b) (i) 4-CF3O–C6H4–NCO, benzene, rt, 8 h; (ii) NaH, 30 min,
rt, 81%; (c) 1:2 1.5 M aq HCl and ethanol, reflux, 3 h, 94%; (d) LiAlH4, AlCl3, ether, rt, 85%; (e) tert-butyl 4-(bromomethyl)benzoate, NaH, DMF, rt,
56%; (f) 2:5 (v/v) TFA and DCM, rt, 20 min, 66%; (g) 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC), 1-hydroxybenzo-
triazole hydrate (HOBt), diisopropylethylamine (DIEA), DMF.

4566
D.-M. Shen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4564–4569
CN
NH2
CN
a
b
CN
t-Bu
3
+
t-Bu
NH2
t-Bu
70%
12
13a, 25%
13b, 13%
Scheme 2. Reagents and conditions: (a) diethyl (cyanomethyl)phosphonate, n-BuLi in hexanes, THF, À78 ꢁC 3 h, warm to rt; (b) 29% aq NH3,
MeOH, 100 ꢁC, autoclave, 25 h.
Table 1. Binding and functional activity (IC50) of human glucagon receptor antagonists
OCF3
OCF3
Y
O
N
N
N
t-Bu
n
H
t-Bu
N
O
R
H
HN
N
R
A
B
O
O
Compound
Class
R
Y
n
Binding IC50, nM (n)
cAMP IC50, nM (n)
1
2
A
A
B
B
B
B
B
B
B
B
B
B
B
B
—
–CH2CH2CO2H
Tetrazol-5-yl
—
—
—
—
0
21 7 (5)
144 88 (4)
46 14 (5)
459
3.2 1.7 (6)
336 282 (5)
105 56 (5)
856
8
–CH2CH2CO2H
Tetrazol-5-yl
H2
H2
@O
@O
H2
H2
@O
@O
H2
H2
H2
H2
—
9
0
193 145 (3)
NDa
NDa
10
11
14
15
16
17
18
19
20
21
7
–CH2CH2CO2H
Tetrazol-5-yl
0
0
631
–CH2CH2CO2H
Tetrazol-5-yl
1
182 129 (6)
34 32 (10)
1691
282 85 (3)
92 29 (13)
NDa
1
–CH2CH2CO2H
Tetrazol-5-yl
1
1
1965
NDa
–CH2CH2CO2Et
–CH2CH2CONH2
0
23% inh. at 100 lM
13% inh. at 100 lM
303 352 (2)
92 36 (2)
68% inh. at 100 lM
NDa
0
NDa
–(R)-CH2CHOHCO2H
(Tetrazol-5-yl) methyl
—
1
NDa
1
—
1975 673 (2)
NDa
a Not determined.
hGCGR-transfected CHO cells was also measured
(cAMP IC50).19
parable functional assays as that used for hGCGR, 15
was found to be only 1.5- to 4-fold less potent at these
receptors than at the hGCGR (cAMP IC50 = 159 46
(2), 203 66 (3), 307 65 (4), and 398 nM,
respectively).
In our hands, the acyclic ureas 1 and 2 were very potent
binders to the hGCGR (3–21 nM), but exhibited a
somewhat diminished functional antagonism (Table 1).
The most potent of the spiro-cyclic analogs prepared
were the [6.6] spiro-ureas, 14 and 15, while being up to
10-fold less potent than 1 and 2 in the binding assay,
only showed a slightly diminished functional inhibition
of the receptor (ꢀ2-fold). The [6.5] analogs 8 and 9 were
about 4-fold less potent than their acyclic analogs in the
cAMP assay. Containing an extra carbonyl group, the
hydantoins, 10 and 11, and dihydropyrimidine-2,4-
diones, 16 and 17, were only weak glucagon receptor
antagonists (binding IC50 >600 nM). An acidic residue
seems to be required for binding activity since ester 18
and amide 19 were both inactive up to 100 lM. Even
minor modifications to b-alanine or amino tetrazole side
chains of 8 and 9 were not well-tolerated. For example,
hydroxy-acid 20 lost a further 2-fold in binding over 14.
Aminomethyl tetrazole 21, while being potent in the
binding assay, only showed weak functional antagonism
(cAMP IC50 ꢀ2000 nM). The truncated acid 7 also
showed only a very weak binding activity.
These observations prompted us to focus further SAR
studies on the spiro-urea series (Table 2). Initial efforts
centered on the less potent [6.5] spiro-urea series since
its chemistry was more accessible than the [6.6] counter-
parts. A variety of changes to the aryl group were ex-
plored with a tetrazole or b-alanine side chain. The
parent phenyl compound 22 was much less active than
9. On the other hand, cyclohexyl is a reasonable replace-
ment for phenyl itself (22 and 23). Smaller aliphatic
groups, such as tert-butyl and i-propyl, were also not
well-tolerated (24 and 25). Among the three chlorophe-
nyl isomers (26, 27, and 29), the meta isomer 27 was the
most active and was comparable to the 4-trifluorometh-
oxy derivative 9. Addition of a second meta substituent
to give 3,5-dichlorophenyl derivative 31 resulted in a
3-fold boost in activity compared to 27. While the
b-alanine derivatives were usually less active than the
tetrazole analogs, in the 3,5-dichloro series the acid
derivative 32 was a potent binder to the hGCGR
(IC50 = 43 nM), although it shifted 10-fold in the cAMP
assay. Other replacements for the 4-trifluoromethoxy-
phenyl group, which showed similar in vitro activities,
Compound 15 was also tested for activity at the GCGRs
of the rhesus monkey, dog, rat, and mouse. Using com-

D.-M. Shen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4564–4569
4567
Table 2. Binding and functional activity (IC50) of the human glucagon receptor antagonists
Y
Y
N
N
O
O
N
N
N
t-Bu
t-Bu
NH
N
H
N
OH
HN
O
N
O
C
D
O
Compound
Class
Y
Binding IC50, nM (n)
cAMP IC50, nM (n)
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
C
C
C
C
C
C
D
C
C
C
D
C
C
C
C
C
C
C
D
C
Ph
894
700
NDa
Cyclohexyl
t-Bu
i-Pr
NDa
1663
2556
NDa
NDa
2-ClPh
3-ClPh
3-ClPh
4-ClPh
2,4-Cl2Ph
3,5-Cl2Ph
3,5-Cl2Ph
4-FPh
739
94 42 (2)
351
NDa
353 153 (2)
601
116 55 (2)
286 209 (2)
36 26 (2)
43 23 (3)
178 6 (2)
135 111 (2)
427
446 52 (2)
1416 188 (2)
116 51 (2)
423 41 (2)
611 405 (2)
113 72 (2)
NDa
4-CF3Ph
4-CH3Ph
4-BrPh
156 39 (2)
30 5 (2)
407
655 122 (2)
164 89 (2)
NDa
( )-1-(4-BrPh)Et
4-CH3SO2Ph
3-CF3OPh
33 15 (2)
60 22 (2)
65 7 (2)
234 89 (2)
515 69 (2)
227 25 (2)
3-CF3OPh
3,4-F2Ph
a Not determined.
Table 3. Pharmacokinetic profiles of human glucagon receptor antagonists in the mousea
Compound
Clp (mL/min/kg)
Vdss (L/kg)
t1/2 (h)
AUCNpo (lM · h/dose)
Cmax (lM)
%F
9
15
21
32
8.1
24
2.6
4.3
7.3
3.6
5.5
3.5
4.7
2.3
0.99
0.25
NDb
0.25
0.20
0.07
0.01
0.11
27
20
40
34
NDb
27
a Compounds were dosed at 1.0 mpk IV and 2.0 mpk PO formulated with a 5:10:85 mixture of DMSO, polysorbate 80, and water.
b Not determined due to plasma concentration below LOQ of 0.0083 lM at most time points in 3/3 animals.
include 4-trifluoromethylphenyl (34), ( )-1-(4-bromo-
phenyl)ethyl (37), 3-trifluoromethoxyphenyl (39), and
3,4-difluorophenyl (41). More polar substituents, such
as methylsulfonyl (38), reduced potency significantly.
Replacing the 4-tert-butylcyclohex-1-yl in 9 with 1-ben-
zylpiperidin-4-yl moiety afforded 42, which exhibited a
similar binding affinity to 9 (IC50 = 97 7 nM), but
was less potent in the functional assay having an IC50
of 892 nM.
full characterization of these mice and their response
to known antagonists have been reported elsewhere.20
Oral administration of the antagonists (nine mice per
group) was followed 60 min later by an IP injection of
glucagon (15 lg/kg). Blood glucose levels were moni-
tored 12, 24, 36, and 48 min later. Compounds 9, 14,
15, and 32 fully suppressed glucose excursion following
the glucagon challenge at 30 mpk. Compounds 8, 21,
and 34 were not effective.
Pharmacokinetic profiles of selected compounds were
obtained in the mouse (Table 3). Tetrazole analogues 9
and 15 and b-alanine analogue 32 exhibited reasonable
2.3–5.5 h half-lives and >20% oral bioavailability. The
oral exposure of aminomethyl tetrazole 21 was found
to be much lower.
Based on its in vitro potency and PD activity, com-
pound 15 was selected for further evaluation including
off-target profile, titration in the pharmacodynamic
assay, evaluation in an in vivo receptor occupancy
assay,20b and measurement of pharmacokinetic proper-
ties in other species. Among the related family B
GPCRs, compound 15 showed modest functional selec-
Compounds 8, 9, 14, 15, 21, 32, and 34 were evaluated
in vivo in a pharmacodynamic (PD) assay for their abil-
ity to block glucagon-induced glucose excursion at
30 mpk using transgenic mice expressing only a func-
tional human glucagon receptor (hGCGR mice). The
tivity
against
hGIP
and
hGLP1
(cAMP
IC50 = 0.43 0.60 lM and 4.6 1.9 lM, respectively).
Compound 15 exhibited only mediocre binding to the
hERG K+ channel (IC50 = 5.4 1.5 lM) and did not
bind to the human Na+ and rabbit DLZ sensitive Ca2+

4568
D.-M. Shen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4564–4569
netic profile in several preclinical species. Spiro-urea 15
Vehicle
showed good oral activity in a pharmacodynamic assay
using transgenic mice by blocking the effect of glucagon-
induced hyperglycemia at doses which correlated with
partial blockade of the receptor in an in vivo receptor
occupancy assay. Further studies in this and related ser-
ies will be reported in due course.
5000
4500
4000
3500
3000
2500
2000
Glucagon
15 @ 3 mpk
15 @ 10 mpk
*
a
a
Acknowledgments
The authors thank Dr. Judith E. Fenyk-Melody and
Mr. Christian N. Nunes for some of the in vivo studies
described here, and Mr. Frederick Wong for synthetic
assistance.
* p<0.05 vs vehicle
a p<0.05 vs glucagon
Figure 2. Titration of compound 15 in pharmacodynamic assay in the
hGCGR mouse.
References and notes
Table 4. Concentration of 15 in the plasma of the hGCGR mouse
during pharmacodynamic assay (n = 3 for each group)
1. DeFronzo, R. A. Med. Clin. N. Am. 2004, 88, 787.
2. Ettaro, L.; Songer, T. J.; Zhang, P.; Engelgau, M. M.
Pharmacoeconomics 2004, 22, 149.
3. Nourparvar, A.; Bulotta, A.; Di-Mario, U.; Perfetti, R.
Trends Pharmacol. Sci. 2004, 25, 86.
Minutes post dose
3 mpk dose (lM)
10 mpk dose (lM)
45
120
0.180 0.074
0.123 0.066
0.272 0.139
0.394 0.211
4. Morral, N. Trends Endocrinol. Metab. 2003, 14, 169.
5. Sloop, K. W.; Michael, M. D. Drugs Future 2004, 29, 835.
6. Ling, A. L.; Wasserman, J. I. Expert Opin. Ther. Pat.
2003, 13, 15.
7. Ling, A. L. Drugs Future 2002, 27, 987.
8. Petersen, K. F.; Sullivan, J. T. Diabetologia 2001, 44,
2018.
channels at concentrations up to 10 lM. The results
from dose titration of 15 in the PD assay using an iden-
tical protocol are shown in Figure 2. It was fully active
in suppressing hyperglycemia following a glucagon chal-
lenge at 10 mpk and partially active at 3 mpk. The plas-
ma concentrations of 15 during the PD experiment are
given in Table 4. This compound also showed good
duration of action, as it was fully active in this assay
when dosed orally at 30 mpk 5 h prior to the glucagon
challenge (data not shown). Drug levels in plasma
at the time of glucagon administration were
0.503 0.198 lM.
9. Ling, A.; Plewe, M. B.; Truesdale, L. K.; Lau, J.; Madsen,
P.; Sams, C.; Behrens, C.; Vagner, J.; Christensen, I. T.;
Lundt, B. F.; Sidelmann, U. G.; Thøgersen, H. U.S.
Patent 6,503,949, 2003. Chem. Abstr. 2001, 134, 4764.
10. This compliments efforts following up on in-house HTS
leads in pursuit of orally active hGCGR antagonists
reported recently. See Duffy, J. L.; Kirk, B. A.; Konteatis,
Z.; Campbell, E. L.; Liang, R.; Brady, E. J.; Candelore,
M. R.; Ding, V. D.-H.; Jiang, G.; Liu, F.; Qureshi, S. A.;
Saperstein, R.; Szalkowski, D.; Tong, S.; Tota, L. M.; Xie,
D.; Yang, X.; Zafian, P.; Zheng, S.; Chapman, K. T.;
Zhang, B. B.; Tata, J. R. Bioorg. Med. Chem. Lett. 2005,
15, 1401.
11. The following patent claiming non-spiro cyclic urea
analogues of the same compounds published after the
studies were completed: Kurukulasuriya, R.; Link, J. T.;
Patel, J. R.; Sorensen, B. K. U.S. Patent Application 2004/
0209928, 2004; Chem. Abstr. 2004, 141, 1366227.
12. As implemented on a desktop computer using Chem3D
Proꢂ Version 5.0 from CambridgeSoft Corp., MA, USA.
Conformer B was 1.5 kcal/mol lower in energy than
conformer A. This difference is well within the error of
computation, especially considering the fact that no
solvation effects were taken into account.
In vivo receptor occupancy of 15 in the hGCGR mice
was measured at 1 h post oral dose to be 52.2 2.9,
39.4 4.8, and 25.7 4.0 % at 10, 3, and 1 mpk, respec-
tively.20b Therefore, receptor occupancy of 40–50% cor-
relates with observed inhibition of glucagon-induced
hyperglycemia in the PD model. Finally, the pharmaco-
kinetic profile of 15 was evaluated in additional preclin-
ical species (Table 5). In the rat and dog, the half-life
was somewhat shorter than in the mouse and oral bio-
availability ranged from 13% to 38 %.
Thus, a novel class of spiro-ureas been discovered as po-
tent human glucagon receptor antagonists. SAR studies
have identified compound 15 as a potent human gluca-
gon receptor antagonist with an acceptable pharmacoki-
13. Edward, J. T.; Jitrangsri, C. Can. J. Chem. 1975, 53,
3339.
Table 5. Pharmacokinetic profiles of human glucagon receptor antagonist 15 in additional animal species
Species
Clp (mL/min/kg)
Vdss (L/kg)
t1/2 (h)
AUCNpo (lM · h/dose)
Cmax (lM)
%F
Rata
hGCGR mousea
Dogb
15
16
6.8
1.3
4.2
1.8
4.6
2.4
0.26
0.47
1.6
0.15
0.13
0.66
13
24
38
0.48
a Dosed at 1.0 mpk IV and 2.0 mpk PO.
b Dosed at 0.5 mpk IV and 1.0 mpk PO.

D.-M. Shen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4564–4569
4569
14. (a) For reviews of hydantoin chemistry, see Avendan˜o, C.;
confirmed its identity. We thank Ms. Nancy Tsou for the
X-ray structure determination.
17. Kumamoto, T.; Koga, K. Chem. Pharm. Bull. 1997, 45,
753.
´
Menendez, J. C., 4th ed. In Kirk-Othmer Encycl. Chem.
Technol., 1995; Vol. 13, p 512; (b) Avendan˜o Lopez, C.;
´
´
Gonzalez Trigo, G. Adv. Heterocycl. Chem. 1985, 38, 177.
15. Knabe, J.; Buch, H. P.; Biwersi, J. Arch. Pharm. 1993, 326,
18. Suzuki, T.; Imannishi, N.; Itahana, H.; Watanuki, S.;
Ohta, M.; Mase, T. Synth. Commun. 1998, 28, 701.
19. Qureshi, S. A.; Candelore, M. R.; Xie, D.; Yang, X.; Tota,
L. M.; Ding, V. D.-H.; Li, Z.; Bansal, A.; Miller, C.;
Cohen, S.; Jiang, G.; Brady, E.; Saperstein, R.; Duffy, J.
L.; Tata, J. R.; Chapman, K. T.; Moller, D. E.; Zhang, B.
B. Diabetes 2004, 53, 3267.
20. (a) Shiao, L.-L.; Cascieri, M. A.; Trumbauer, M.; Chen,
H.; Sullivan, K. A. Transgenic Res. 1999, 8, 295; (b)
Dallas-Yang, Q.; Shen, X.; Strowski, M.; Brady, E.;
Saperstein, R.; Gibson, R. E.; Szalkowski, D.; Candelore,
M. R.; Fenyk-Melody, J. E.; Parmee, E. R.; Zhang, B. B.;
Jiang, G. Eur. J. Pharmacol. 2004, 501, 225.
¨
331.
16. All intermediates and final products described here
1
1
showed expected H NMR and LC-MS. In addition, H,
COSY, and NOESY spectra of the alkylation product of 5
with 4-bromobenzyl bromide showed it had the desired
stereochemistry on the cyclohexane ring. Only one isomer
of 5 was observed when cyclization was conducted at
room temperature. At higher temperatures, a minor
amount of another isomer of 5 was isolated in some cases.
For example, detailed NMR studies of both isomers of
2-chlorophenyl analogues of 6 confirmed our structure
assignments. Finally, a single crystal X-ray structure of 9