Pentacyclic triterpenes. Part 5: Synthesis and SAR study of corosolic acid derivatives as inhibitors of glycogen phosphorylases

Article abstract of DOI:10.1016/j.bmcl.2007.08.057

The synthesis and biological evaluation of corosolic acid derivatives and related compounds as inhibitors of rabbit muscle glycogen phosphorylase a is described. Within this series of compounds, 8 (IC₅₀ = 7.31 μM), 12d (IC₅₀ = 3.26 μM), and 12e (IC₅₀ = 5.1 μM) exhibited more potent activities than the parent compound 1 (IC₅₀ = 20 μM). SAR of these compounds is also discussed.

Full text of DOI:10.1016/j.bmcl.2007.08.057

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5777–5782
Pentacyclic triterpenes. Part 5: Synthesis and SAR study of
corosolic acid derivatives as inhibitors of glycogen phosphorylases
Xiaoan Wen,a Jun Xia,a Keguang Cheng,a Liying Zhang,a Pu Zhang,a Jun Liu,b
Luyong Zhang,b Peizhou Nia and Hongbin Suna,*
aCenter for Drug Discovery, College of Pharmacy, China Pharmaceutical University, 24 Tongjia Xiang, Nanjing 210009, China
bJiangsu Center for Drug Screening, China Pharmaceutical University, 1 Shennonglu, Nanjing 210038, China
Received 17 May 2007; revised 19 August 2007; accepted 24 August 2007
Available online 28 August 2007
Abstract—The synthesis and biological evaluation of corosolic acid derivatives and related compounds as inhibitors of rabbit muscle
glycogen phosphorylase a is described. Within this series of compounds, 8 (IC50 = 7.31 lM), 12d (IC50 = 3.26 lM), and 12e
(IC50 = 5.1 lM) exhibited more potent activities than the parent compound 1 (IC50 = 20 lM). SAR of these compounds is also
discussed.
Ó 2007 Elsevier Ltd. All rights reserved.
Corosolic acid (1) (Fig. 1) is a naturally occurring penta-
cyclic triterpene acid which has attracted much attention
due to its hypoglycemic activity.1 Recently, the Japanese
H
researchers proved for the first time that 1 exhibited a glu-
cose-lowering effect on postchallenge plasma glucose lev-
els in humans.2 The mechanism of hypoglycemic action of
1 is not clear, nevertheless, according to the reported re-
sults of mechanic studies, it is likely that 1 might exert
its glucose-lowing effect through multiple targets.1a,3
COOH
1
2
R
R
1
2
R
R
OH
OH
OH
OH
corosolic acid
-epi-corosolic acid (2)
2-epi-corosolic acid (3)
(1)
3
OH
OH
We have previously reported that 1 and related pentacy-
clic triterpenes are natural inhibitors of glycogen phos-
phorylase (GP), and they might reduce blood glucose,
at least in part, through inhibiting excessive hepatic gly-
cogenolysis.3c GP is the enzyme responsible for glycogen
breakdown to produce glucose and related metabolites
for energy supply. Recently, new insight into this en-
zyme has stimulated research interest in identifying safe
and effective GP inhibitors as therapeutic agents for
treatment of diabetes.4 Several structural classes of GP
inhibitors have been developed, and at least six binding
sites have been identified in GP.5 Although there is still a
debate about whether GP could be an ideal drug target
for diabetes, in our point of view, the key issue is about
how to find PROPER small molecules to PROPERLY
Figure 1. The structures of corosolic acid, 3-epi-corosolic acid and 2-
epi-corosolic acid.
modulate GP, and to maintain chronic efficiency of
these compounds with less side-effects, rather than about
whether GP is a PROPER drug target.
With 1 as a lead compound for GP inhibition, we carried
out structural modification and SAR study. For example,
in order to determine how the configuration of 2,3-
dihydroxyl groups of 1 affected the enzyme inhibitory
activity, we synthesized 3-epi-corosolic acid (2) and 2-
epi-corosolic acid (3), which are naturally occurring iso-
mers of 1. Moreover, we synthesized a series of derivatives
of 1 in order to determine the structure–activity
relationship.
Keywords: Corosolic acid; Derivatives; Glycogen phosphorylase;
Inhibitor; Synthesis; SAR.
The syntheses of 1, 2, and 3 are summarized in
Scheme 1. Herein, preparation of 1 was based on a
*
Corresponding author. Tel.: +86 25 85327950; fax: +86 25
83271335; e-mail: hbsun2000@yahoo.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.057

5778
X. Wen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5777–5782
H
H
H
COOBn
COOH
COOBn
a, b
h
HO
O
(ref.3c)
4
HO
O
8
ursolic acid
H
c
d
g
H
H
COOBn
COOBn
COOBn
d
HO
HO
HO
O
HO
HO
+ 7 (minor)
5
9
6 (major)
f
e
e
H
H
H
COOH
COOBn
+
COOH
HO
HO
HO
HO
HO
HO
6 (minor)
7 (major)
3
1
2-epi-corosolic acid
e
corosolic acid
H
COOH
HO
HO
2
3-epi-corosolic acid
Scheme 1. Reagents and conditions: (a) BnCl, K2CO3/DMF, 50 °C; (b) PCC, DCM, rt, 80% for two steps; (c) m-CPBA, MeOH/DCM/H2SO4 (cat.),
rt, 70%; (d) NaBH4, MeOH/THF; rt, 68% for 6; 14% for 7; 69% for 9. (e) H2, Pd–C (10%), rt, 97% for 1; 92% for 2; 88% for 3; (f) Al(OPr-i)3, i-PrOH,
reflux, 51% for 7 and 6% for 6; (g) KOH, MeOH/DMF, rt, 62%; (h) t-BuOK, t-BuOH, air, rt, 85%.
ketone-hydroxylation strategy6,7 which was different
from our previously reported method.3c As shown in
Scheme 1, stereoselective hydroxylation of ursonic acid
benzyl ester (4)3c with m-CPBA in MeOH–DCM–
H2SO4 (cat.) gave 2a-hydroxy ketone 5 in 70% yield.
Reduction of 5 with NaBH4 in THF–MeOH gave
2a,3b-diol 6 as the major product (68%), together with
2a,3a-diol 7 as the minor product (14%). On the other
hand, Meerwein–Pondorf reduction8 of 5 gave 7 as
the major product (51%), together with 6 as the minor
product (6%). Hydrogenolysis of 6 and 7 over palla-
dium–carbon in THF furnished corosolic acid (1) and
3-epi-corosolic acid (2) in high yields, respectively.9
Treatment of 5 with KOH in MeOH–DMF gave dike-
tone 8 (62%), which existed in the form of a,b-unsatu-
rated ketone.10 Alternatively, treatment of ketone 4
with a large excess of t-BuOK under air at room temper-
ature directly gave 8 in good yield (85%). Reduction of 8
with NaBH4 gave 2b,3b-diol 9 (69%), which was further
converted to 2-epi-corosolic acid (3) by hydrogenolysis
over palladium–carbon.11
ious alkyl halides or 2,3,4,6-tetra-O-acetyl-a-D-glucosyl
bromide gave the corresponding esters 10a–h in good
H
H
RX
COOH
COOR
10a~h
HO
HO
1
H
O
H
O
O
NaOH
O
OH
O
O
O
10i
10g
AcO
HO
H
H
NaOH
OAc
OAc
OH
OH
O
O
O
O
O
O
OAc
OH
Structural modification of corosolic acid was mainly
focused on C-28 and A-ring. The synthetic routes are
outlined in Schemes 2 and 3. Esterification of 1 with var-
10j
10h
Scheme 2. Synthesis of derivatives 10a–j.

X. Wen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5777–5782
5779
H
RO
RO
RO
HO
acylation
COOH
or
HO
HO
11a~d
12a~e
1
O
H
HOOC
HOOC
O
O
O
O
O
O
COOBn
HO
HO
11e
6
MsCl or TsCl
RO
HO
NaH
O
12 f R=Ms
12g R=Ts
12h
Scheme 3. Synthesis of derivatives 11a–e and 12a–h.
yields. Hydrolysis of 10g and 10h afforded 10i and 10j in
high yields, respectively. Acylation of 1 or 6 afforded dia-
cylates 11a–e or monoacylates 12a–e. As we expected,
the reactivity was quite different between 2-hydroxyl
and 3-hydroxyl of 1 or 6, and this kind of discrimination
could be used to get 2,3-diacylates or 2-monoacylates. In
this regard, diacylates 11a–e were obtained by using a
large excess of acylating agents, or by heating the reac-
tion mixture in the presence of DMAP as a catalyst.
On the other hand, monoacylates 12a–e were obtained
under relatively mild conditions such as controlled
amount of acylating agents and low reaction tempera-
ture (normally room temperature). Mesylation or tosyla-
tion of 6 in pyridine at room temperature gave mesylate
12f (26%) or tosylate 12g (32%), respectively. Treatment
of both 12f and 12g with NaH afforded epoxide 12h
(from 12f: 32%; from 12g: 73%). The detailed reagents,
conditions, and yields for each reaction are shown in
Tables 1 and 2.
rabbit muscle glycogen phosphorylase a (RMGPa). As
described previously,12 the activity of RMGPa was mea-
sured through detecting the released amount of phos-
phates from glucose-1-phosphates in the direction of
glycogen synthesis. The assay results are outlined in
Tables 3 and 4. Initially, we investigated how the config-
uration of 2-hydroxy and 3-hydroxy groups affected the
GP inhibitory activity. The inhibitory potency decreases
sharply when both 2-hydroxy and 3-hydroxy groups are
in the same side of A-ring (e.g., 1 vs 2, 1 vs 3, 6 vs 7, and
Table 2. Reagents, conditions, and yields for synthesis of 11a–e and
12a–h
Products
Reagents and conditions
Yield (%)
11a
11b
Ac2O (126 equiv), pyridine, rt
Propionic anhydride
(10 equiv),
93
27
pyridine, DMAP, rt
Butyric anhydride (19 equiv),
pyridine, 80 °C
Succinicanhydride (288 equiv),
pyridine, DMAP, 80 °C
Succinicanhydride (20 equiv),
pyridine, DMAP, 80 °C
Ac2O (4.5 equiv), pyridine, rt
(CH3CH2CO)2O (6 equiv),
pyridine, rt
11c
11d
11e
51
56
66
Compounds 1–3, 5–9, 10a–j, 11a–e, and 12a–h were bio-
logically evaluated for their inhibitory activity against
Table 1. Reagents, conditions, and yields for synthesis of 10a–j
Products Reagents and conditions
Yield (%)
12a
12b
6
48
10a
10b
10c
10d
10e
10f
MeI, K2CO3/DMF, rt
BrCH2CH3, K2CO3/DMF, 50 °C
BrCH2CH2CH3, K2CO3/DMF, 60 °C
69
90
92
12c
12d
12e
Butyric anhydride (6 equiv),
pyridine, rt
12
41
27
BrCH2CH2CH2CH3, K2CO3/DMF, 50 °C 89
BrCH2CH2@CH2, K2CO3/DMF, rt
1-(Chloromethyl)-1H-pyrazole,
K2CO3/DMF, rt
65
60
Heptanoic acid anhydride
(5 equiv), pyridine, rt
Benzoic acid (8 equiv),
THF/DCC/DMAP, rt
MsCl (5 equiv), pyridine, rt
TsCl (16 equiv), pyridine, rt
NaH, THF, rt
10g
10h
BrCH2COOEt, K2CO3/DMF, rt
2,3,4,6-Tetra-O-acetyl-a-D-glucosyl
bromide, K2CO3/DMF, rt
NaOH (4N), H2O/THF, rt
NaOH (4N), H2O/THF, rt
90
94
12f
12g
12h
26
32
10i
10j
93
92
32 (from 12f)
73 (from 12g)

5780
X. Wen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5777–5782
Table 3. Effects of configuration of 2,3-dihydoxyl groups of corosolic
acid analogues on inhibitory activity against RMGPa
tor (IC50 = 7.31 lM) than its corresponding 2,3-diol
analogues (e.g., 1, 2, 3, and 6).
As shown in Table 4, the inhibitory potency decreases as
the size of C-28 substituents increases (e.g., inhibitory
potency: 10a > 10b > 10c > 10 ꢀ 10e > 10g > 10h), indi-
cating that small C-28 substituents of corosolic acid
derivatives may be preferred for GP inhibition. This
trend seems different from that of maslinic acid deriva-
tives which somehow prefer big C-28 side chains for
good potency.6 Corosolic acid 28-O-b-D-glucopyranosyl
ester (10j) was about 3-fold more potent than its 2,3,4,6-
tetra-O-acetyl ester 10h. In the series of A-ring modified
triterpenes, the potency of 2,3-diacyloxyl triterpenes was
not quite different from that of the corresponding 2-
monoacyloxyl triterpenes (e.g., 11a vs 12a, 11b vs 12b,
and 11c vs 12c). An increase in polarity at 2,3-diacyl
substituents resulted in a significant loss of potency
(e.g., 11c vs 11dand 11e). The SAR of 2-monoacyl side
chains was less predictive (12a–e). Within this series of
H
3
COOR
1
2
R
R
a
Compound
R1
R2
R3
IC50 (lM)
1
a-OH
a-OH
b-OH
a-OH
a-OH
a-OH
@Ob
b-OH
a-OH
b-OH
@O
b-OH
a-OH
@O
H
20
2
H
212.5
115.5
31.8
41.2
323.5
7.31
107.9
114
3
H
5
Bn
Bn
Bn
Bn
Bn
6
7
8
9
caffeine
b-OH
b-OH
a Values are means of three experiments.
b Existed in the form of 1,2-enol as shown in Scheme 1.
compounds,
12d
(IC50 = 3.26 lM)
and
12e
(IC50 = 5.1 lM) exhibited more potent inhibitory activ-
ity than the parent compound 1, indicating that modifi-
cation on A-ring of 1 might improve the potency of the
triterpenes against GP.
6 vs 9). Surprisingly, 2,3-diketone 8 (existing in the form
of a,b-unsaturated ketone) is a more potent GP inhibi-
Table 4. Effects of variation of substituents in corosolic acid derivatives on RMGPa
H
COOR3
R1O
R2O
R1
R2
R3
IC50 (lM)
a
Compound
10a
10b
10c
10d
10e
10f
H
H
H
H
H
H
H
H
H
H
H
H
Me
Et
33.7
46.7
82.5
92.1
91.2
62.1
Pr
Bu
Vinyl
(1H-Pyrazol-
1-yl)methyl
CH2COOEt
10g
10h
10i
H
H
102.0
373.1
NIb
H
H
b-Glc(Ac)4
CH2COOH
H
H
10j
H
Ac
H
Ac
b-Glc
H
105.6
35.0
73.7
78.7
NIb
11a
11b
11c
11d
11e
12a
12b
12c
12d
12e
12f
12g
12h
COCH2CH3
CO(CH2)2CH3
CO(CH2)2COOH
CO(CH2)2COOH
Ac
COCH2CH3
CO(CH2)2CH3
H
H
CO(CH2)2COOH
CO(CH2)2COOH
H
Bn
H
722.5
37.7
96.7
77.9
3.26
5.1
H
H
H
H
H
H
H
COCH2CH3
CO(CH2)2CH3
CO(CH2)4CH3
Bz
H
H
H
H
Ms
Ts
Bn
Bn
Bn
43.4
443.8
69.5
2b,3b-epoxy
a Values are means of three experiments.
b NI, no inhibition.

X. Wen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5777–5782
5781
5. (a) Lu, Z. J.; Bohn, J.; Bergeron, R.; Deng, Q. L.;
Ellsworth, K. P.; Geissler, W. M.; Harris, G.; McCann, P.
E.; Mckeever, B.; Myers, R. W.; Saperstein, R.; Wil-
loughby, C. A.; Yao, J.; Chapman, K. Bioorg. Med. Chem.
Lett. 2003, 13, 4125; (b) Wright, S. W.; Rath, V. L.;
Genereux, P. E.; Hageman, D. L.; Levy, C. B.; McClure,
L. D.; McCoid, S. C.; McPherson, R. K.; Schelhorn, T.
M.; Wilder, D. E.; Zavadoski, W. J.; Gibbs, E. M.;
Treadway, J. L. Bioorg. Med. Chem. Lett. 2005, 15, 459;
(c) Treadway, J. L.; Mendys, P.; Hoover, D. J. Expert
Opin. Invest. Drugs 2001, 10, 439, and references therein;
(d) Chrysina, E. D.; Kosmopolou, M. N.; Tiraidis, C.;
Kardarakis, R.; Bischler, N.; Leonidas, D. D.; Hadady,
Z.; Somsak, L.; Docsa, T.; Gergely, P.; Oikonomakos, N.
G. Protein Sci. 2005, 14, 873; (e) Hampson, L. J.; Arden,
C.; Agius, L.; Ganotidis, M.; Kosmopoulou, M. N.;
Tiraidis, C.; Elemes, Y.; Sakarellos, C.; Leonidas, D. D.;
Oikonomakos, N. G. Bioorg. Med. Chem. 2006, 14, 7835;
(f) Juhasz, L.; Docsa, T.; Brunyaszki, A.; Gergely, P.;
Antus, S. Bioorg. Med. Chem 2007, 15, 4048; (g) Birch, A.
M.; Kenny, P. W.; Oikonomakos, N. G.; Otterbein, L.;
Schofield, P.; Whittamore, P. R. O.; Whalley, D. P.
Bioorg. Med. Chem. Lett. 2007, 17, 394.
Control of glycogen metabolism plays a key role in
maintaining blood glucose homeostasis in both fed
and fasted states.13 Modulation of key elements in gly-
cogen metabolism such as GP and related signaling mol-
ecules represents a promising therapeutic approach to
diabetes.14 Corosolic acid and other pentacyclic triter-
penes may exert their glucose-lowering effect through
multiple targets including GP. Further studies are
needed to address the detailed molecular mechanisms
and to biologically evaluate pentacyclic triterpenes as
promising anti-diabetic agents with preventive effects
against diabetic complications.
In summary, two steric isomers of corosolic acid and a
series of corosolic acid derivatives have been synthesized
and tested for inhibitory activity against RMGPa.
Among the tested compounds, 8, 12d, and 12e are much
more potent than the lead compound 1. SAR studies
show that A-ring modification of corosolic acid may
improve the potency for GP inhibition. Extensive struc-
tural modifications, biological evaluation, and mechanis-
tic studies on corosolic acid and other pentacyclic
triterpenes are ongoing in this laboratory in order to find
potent and low-toxic anti-diabetic agents with preventive
and therapeutic effects against diabetic complications.
6. Wen, X. A.; Zhang, P.; Liu, J.; Zhang, L. Y.; Wu, X. M.;
Ni, P. Z.; Sun, H. B. Bioorg. Med. Chem. Lett. 2006, 16,
722.
7. Takayama, H.; Kitajima, M.; Ishizuka, T.; Seo, S. U.S.
Patent Application: US2005020681 A1, 2005.
8. (a) Huneck, S.; Snatzke, G. Chem. Ber. 1965, 98, 120;
(b) Bore, L.; Honda, T.; Gribble, G. W. J. Org. Chem.
2000, 65, 6278.
Acknowledgments
9. Analytical data for compound 2: IR (KBr, cmÀ1) 3454,
2928, 1688, 1630, 1456, 1383, 1038, 665; 1H NMR
(pyridine-d5, 300 MHz): d 0.88, 0.94, 1.03, 1.11, 1.25
(each 3H, s), 0.92 (3H, d, J = 6.2 Hz), 0.96 (3H, d,
J = 6.3 Hz), 2.61 (1H, d, J = 11.1 Hz, H-18), 3.74 (1H,
d, J = 2.6 Hz, H-3b), 4.28 (1H, m, H-2b), 5.44 (1H, t,
J = 3.3 Hz, H-12); 13C NMR (pyridine-d5, 300 MHz): d
16.8, 17.47, 17.50, 18.5, 21.4, 22.3, 23.7, 23.9, 24.9,
28.7, 29.5, 31.1, 33.5, 37.4, 38.6, 38.8, 39.4, 39.5, 40.2,
42.6, 43.0, 47.9, 48.1, 48.7, 53.6, 66.1, 79.4, 125.6,
139.3, 179.9; MS: 495 [M+Na]+; HRMS: Calcd for
C30H48O4–H: 471.3474; Found: 471.3485.
This program was financially supported by National
Natural Science Foundation (Grant 30672523), Chinese
University Ph.D Program Foundation (Grant 2005031
6008), Jiangsu Outstanding Researcher Grant, and pro-
gram for New Century Excellent Talents in University
(NCET-05-0495).
References and notes
10. Analytical data for compound 8: IR (KBr, cmÀ1) 3435,
2970, 2926, 2870, 1724, 1664, 1454, 1404, 1232, 1144,
1. (a) Miura, T.; Itoh, Y.; Kaneko, T.; Ueda, N.; Ishida, T.;
Fukushima, M.; Matsuyama, F.; Seino, Y. Biol. Pharm.
Bull. 2004, 27, 1103; (b) Liu, F.; Kim, J.; Li, Y.; Liu, X.; Li,
J.; Chen, X. J. Nutr. 2001, 131, 2242; (c) Judy, W. V.; Hari,
S. P.; Stogsdill, W. W.; Judy, J. S.; Naguib, Y. M. A.;
Psswater, R. J. Ethnopharmacol. 2003, 87, 115.
2. Fukushima, M.; Matsuyama, F.; Ueda, N.; Egawa,
K.; Takemoto, J.; Kajimoto, Y.; Yonaha, N.; Miura,
T.; Kaneko, T.; Nishi, Y.; Mitsui, R.; Fujita, Y.;
Yamada, Y.; Seino, Y. Diabetes Res. Clin. Pract.
2006, 73, 174.
3. (a) Na, M. K.; Yang, S. M.; He, L.; Oh, H.; Kim,
B. S.; Oh, W. K.; Kim, B. Y.; Ahn, J. S. Planta
Med. 2006, 72, 261; (b) Murakami, C.; Myoga, K.;
Kasai, R.; Ohtani, K.; Kurokawa, T.; Ishibashi, S.;
Dayrit, F.; Padolina, W. G.; Yamasaki, K. Chem.
Pharm. Bull. 1993, 41, 2129; (c) Wen, X. A.; Sun, H.
B.; Liu, J.; Wu, G. Z.; Zhang, L. Y.; Wu, X. M.;
Ni, P. Z. Bioorg. Chem. Med. Lett. 2005, 15, 4944.
4. (a) Greenberg, C. C.; Jurczak, M. J.; Danos, A. M.;
Brady, M. J. Am. J. Physiol. Endocrinol. Metab. 2006, 291,
E1; (b) Baker, D. J.; Greenhaff, P. L.; Timmons, J. A.
Expert Opin. Ther. Patents 2006, 16, 459; (c) Oikonoma-
kos, N. G. Curr. Protein Pept. Sci. 2002, 3, 561;
1
1053, 746, 696, 534; H NMR (CDCl3, 300 MHz): d 0.69,
1.06, 1.12, 1.20, 1.21 (each 3H, s), 0.86 (3H, d, J =
6.4 Hz), 0.95 (3H, d, J = 6.1 Hz), 2.31 (1H, d, J = 11.1 Hz,
H-18), 5.01 and 5.09 (each 1H, d, J = 12.4 Hz, CH2Ar),
5.29 (1H, t, J = 3.5 Hz, H-12), 5.93 (1H, s, HO-2), 6.35
(1H, s, H-1), 7.34 (5H, m, H-Ar); 13C NMR (CDCl3,
300 MHz): d 17.0, 17.6, 18.7, 19.7, 21.1, 21.8, 23.3, 23.5,
24.3, 27.2, 28.0, 30.7, 32.9, 36.6, 38.3, 38.9, 39.1, 40.3,
42.5, 43.1, 43.9, 48.2, 53.1, 54.0, 66.0, 125.1, 128.0, 128.2,
128.4, 136.4, 138.7, 143.8, 177.1, 201.1; MS: 581
[M+Na]+; HRMS: Calcd for C37H50O4–H–H: 557.3631;
Found: 557.3651.
11. Analytical data for compound 3: IR (KBr, cmÀ1) 3439,
2926, 1691, 1454, 1381, 1051, 787, 766, 663; 1H NMR
(pyridine-d5, 300 MHz): d 0.97 (3H, d, J = 5.7 Hz), 1.03
(3H, d, J = 6.4 Hz), 1.11, 1.25, 1.27, 1.36, 1.51 (each
3H, s), 2.66 (1H, d, J = 11.3 Hz, H-18), 3.46 (1H, d,
J = 3.9 Hz, H-3a), 4.41 (1H, m, H-2a), 5.50 (1H, m, H-
12); 13C NMR (pyridine-d5, 300 MHz): d 16.8, 17.5,
18.2, 18.7, 21.4, 23.9, 24.0, 25.0, 28.7, 30.3, 31.2, 33.7,
37.3, 37.5, 38.8, 39.5, 39.6, 40.2, 42.8, 45.2, 48.1, 48.5,
53.7, 56.0, 71.5, 78.4, 125.9, 139.3, 179.9; MS: 495
[M+Na]+; HRMS: Calcd for C30H48O4–H: 471.3474;
Found: 471.3481.
´
(d) Somsak, L.; Nagy, V.; Hadady, Z.; Docsa, T.; Gergely,
P. Curr. Pharm. Des. 2003, 9, 1177.

5782
X. Wen et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5777–5782
12. Martin, W. H.; Hoover, D. J.; Armento, S. J.; Stock, I. A.;
Mcpherson, R. K.; Danley, D. E.; Stevenson, R. W.;
Barretti, E. J.; Treadway, J. L. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 1776.
13. Bollen, M.; Keppens, S.; Stalmans, W. Biochem. J. 1998,
336, 19, Part 1.
14. (a) Aiston, S.; Coghlan, M. P.; Agius, L. Eur. J.
Biochem. 2003, 270, 2773; (b) Aiston, S.; Hampson, L.
J.; Arden, C.; Iynedjian, P. B.; Agius, L. Diabetologia
2006, 49, 174; (c) Aiston, S.; Hampson, L.; Gomez-
Foix, A. M.; Guinovart, J. J.; Agius, L. J. Biol. Chem.
2001, 276, 23858.