Pentacyclic triterpenes. Part 1: The first examples of naturally occurring pentacyclic triterpenes as a new class of inhibitors of glycogen phosphorylases

Article abstract of DOI:10.1016/j.bmcl.2005.08.026

The semi-synthesis, in vitro and in vivo biological evaluation of corosolic acid (1) and maslinic acid (2) are described. Compounds 1 and 2 represent a new class of inhibitors of glycogen phosphorylases. Both 1 and 2 inhibit the increase of fasted plasma glucose of diabetic mice induced by adrenaline. It is therefore proposed that naturally occurring pentacyclic triterpenes 1 and 2 might reduce blood glucose, at least in part, through inhibiting hepatic glycogen degradation.

Full text of DOI:10.1016/j.bmcl.2005.08.026

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4944–4948
Pentacyclic triterpenes. Part 1: The first examples of
naturally occurring pentacyclic triterpenes as a new class
of inhibitors of glycogen phosphorylases
Xiaoan Wen,a Hongbin Sun,a,* Jun Liu,b Guanzhong Wu,c Luyong Zhang,b
Xiaoming Wua and Peizhou Nia
aDepartment of Medicinal Chemistry, College of Pharmacy, China Pharmaceutical University,
24 Tongjiaxiang, Nanjing 210009, China
bJiangsu Center for Drug Screening, China Pharmaceutical University, 1 Shennong Road, Nanjing 210038, China
cDepartment of Pharmacology, College of Pharmacy, China Pharmaceutical University, 24 Tongjiaxiang, Nanjing 210009, China
Received 8 June 2005; revised 12 July 2005; accepted 5 August 2005
Available online 16 September 2005
Abstract—The semi-synthesis, in vitro and in vivo biological evaluation of corosolic acid (1) and maslinic acid (2) are described.
Compounds 1 and 2 represent a new class of inhibitors of glycogen phosphorylases. Both 1 and 2 inhibit the increase of fasted plas-
ma glucose of diabetic mice induced by adrenaline. It is therefore proposed that naturally occurring pentacyclic triterpenes 1 and 2
might reduce blood glucose, at least in part, through inhibiting hepatic glycogen degradation.
Ó 2005 Elsevier Ltd. All rights reserved.
Corosolic acid (1) has recently attracted much attention
due to its biological activities, such as anti-diabetes,1
anti-inflammation,2 antiproliferation,3 and protein
kinase C inhibition activities.4 Compound 1 is the pri-
mary active principle of a plant extract (commercially
named Glucosol), which has been marketed in Japan
and the United States for reducing blood glucose levels
and weight-loss.5 Maslinic acid (2), which has a similar
structure with 1, has also drawn much interest due to
its anti-tumor,6 anti-HIV,7 and antioxidant activities.8
Surprisingly, despite the obvious biological importance
of 1 and 2, little was reported regarding a practical syn-
thetic preparation of both compounds.9 On the other
hand, it is very tedious and expensive to obtain pure
compound 110 or 211 by plant extraction. As part of
our projects aimed at developing pentacyclic triterpenes
as preventive and therapeutic agents, we hoped to estab-
lish a practical synthetic preparation of 1 and 2.
way. It is well known that liver GP is the major enzyme
for controlling hepatic glucose output, and inhibition of
this enzyme may afford a useful therapeutic approach
for type 2 diabetes.12a Moreover, GP inhibitors might
also find pharmaceutical applications in treatment of
cardiovascular diseases13 and tumors.14 Several structur-
al classes of GP inhibitors have been reported, whose
binding sites identified in GP include the catalytic site,
the purine inhibitory site (also known as I-site), the allo-
steric site, the glycogen storage site, and a novel alloste-
ric inhibitor site.12b,12c,12d Attention attracted in this
area remains high as evidenced by the frequent appear-
ance of related patents and publications.
H
H
COOH
COOH
HO
HO
HO
HO
H
H
Glycogen phosphorylases (GP), which catalyze the first
step of glycogen breakdown, play an important role in
glucose metabolism, especially in glycogenolytic path-
2
1
Herein, we report an efficient semi-synthesis of 1 and
2, starting from readily available ursolic acid and
oleanolic acid, respectively. Most importantly, we report
here, for the first time, that 1 and 2 represent a new class
of natural inhibitors of glycogen phosphorylases.
Keywords: Triterpenes; Maslinic acid; Corosolic acid; Glycogen
phosphorylase; Inhibitors; Synthesis.
*
Corresponding author. Tel.: +86 25 85327950; fax: +86 25
83271335; e-mail: hbsun2000@yahoo.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.08.026

X. Wen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4944–4948
4945
The syntheses of 1 and 2 are summarized in Scheme 1.
Commercially available ursolic acid 3a was mixed with
benzyl chloride and potassium carbonate in DMF at
85 °C for 2 h to give benzyl ester 4a in 96% yield. Con-
sidering that short chain alkyl esters of the C-28 carbox-
ylic acid, such as C-28 methyl ester or ethyl ester, could
not be hydrolyzed by conventional hydrolysis methods
(e.g., basic or acidic aqueous conditions), we chose ben-
zyl as the protecting group, which can be easily removed
later on by catalytic hydrogenolysis without effecting the
C12–13 double bond. Oxidation of 4a with PCC in
CH2Cl2 at 0 °C to room temperature for 24 h afforded
5a in 89% yield after simple work-up and crystallization
from ethanol. Reaction of 5a with vinyl acetate or iso-
propenyl acetate catalyzed by concentrated sulfuric acid
gave enol acetate 6a in 91% yield. Treatment of 6a with
excess borane–THF solution, followed by adding alka-
line H2O2, afforded 2a,3b-diol 7a (55%) as the major
product after column chromatography purification.
The major side products were 5a (produced by saponifi-
cation of 6a) and 4a (produced by reduction of the re-
formed 5a). Another possible isomer (2b,3a) was only
trace amount due to the high stereospecificity of the
reaction and could be removed by column chromatogra-
phy. Hydrogenolysis of 7a over palladium/carbon in
THF furnished 1 in quantitative yield.15 In a similar
fashion, 2 was synthesized, starting with oleanolic acid
3b.16
R2
R2
R1
R1
H
COOH
H
H
COOBn
i
H
HO
HO
3a R1=CH3, R2=H
3b R1=H, R2=CH3
4a R1=CH3, R2=H
4b R1=H, R2=CH3
R2
R1
H
ii
iii
COOBn
H
5a R1=CH3, R2=H
5b R1=H, R2=CH3
O
R2
R2
R1
H
R1
H
COOBn
COOBn
iv
HO
H
H
HO
AcO
6a R1=CH3, R2=H
6b R1=H, R2=CH3
7a R1=CH3, R2=H
7b R1=H, R2=CH3
R2
R1
H
v
COOH
HO
H
HO
1 R1=CH3, R2=H
2 R1=H, R2=CH3
Scheme 1. Reagents and conditions: (i) K2CO3, BnCl, DMF, 85 °C; (ii) PCC, CH2Cl2; (iii) AcOCH@CH2, concd H2SO4 (cat.), 100 °C; (iv) a—BH3–
THF; b—H2O2, NaOH; (v) H2, Pd/C, THF, rt.

4946
X. Wen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4944–4948
The hypoglycemic mechanism of 1 is not clear,17 there-
fore, more research work is still needed to unanimously
solve the puzzle. In this regard, we tested 1 in enzyme
inhibition assay against GP since we hypothesized that
1 and related triterpenes might lower blood glucose by
inhibiting hepatic glycogen degradation. Considering
that muscle GP and liver GP share considerable se-
quence similarity, we first employed rabbit muscle GPa
for the enzyme assay since this enzyme was commercial-
ly available. As described previously,18 the activity of
rabbit muscle GPa was measured through detecting
the release of phosphate from glucose-1-phosphate
in the direction of glycogen synthesis. The results are
outlined in Figure 1. To our delight, 1 exhibited
moderate inhibitory activity on rabbit muscle GPa
(IC50 = 20 lM). Not surprisingly, 2 also effectively
inhibited this enzyme (IC50 = 28 lM). In our assays, 1
and 2 were at least 4-fold more potent than caffeine
(IC50 = 114 lM), a proverbial GP inhibitor.12a
1.2
1
0.8
0.6
0.4
0.2
0
0
5
10
[Glucose], mM
15
Figure 2. The effect of glucose on the potency of 1 (j), 2 (Ç) or caffeine
(m). IC50 values for rabbit muscle GPa inhibition were determined at
varied glucose concentrations and then normalized by dividing the
values by the IC50 values obtained in the absence of glucose. The
normalized results are plotted as a function of glucose concentration.
We also examined the effect of varying glucose concen-
trations on the IC50 values for 1, 2, and caffeine, because
some GP inhibitors were reported to be more potent in
the presence of high glucose concentrations.19,20 As
shown in Figure 2, at high glucose concentrations, the
relative IC50 values for 1, 2, and caffeine were signifi-
cantly decreased, indicating the increases in potency
for the test compounds.
100
90
80
70
60
50
40
30
20
10
0
Encouraged by the results obtained with muscle GPa,
we further tested 1 and 2 on purified liver GPa obtained
by extraction from rat liver as described by Fosgerau
et al.21 The assay results showed that both
1
(IC50 = 101 lM) and 2 (IC50 = 99 lM) were inhibitors
of rat liver GPa (Fig. 3), whereas caffeine
(IC50 = 648 lM) was a 6-fold less potent inhibitor of
rat liver GPa. Although 1 and 2 were less potent as rat
1
10
100
1000
10000
Concentration,
M
Figure 3. Inhibition of rat liver GPa by 1, 2, and caffeine. GPa activity
was measured at varied concentrations of 1 (m), 2 (j) or caffeine (Ç).
The data are plotted as percent inhibition versus concentration of
compound. Results are means SEM for three independent
experiments.
110
100
90
80
70
60
50
40
30
20
10
0
liver GPa inhibitors than as rabbit muscle GPa inhibi-
tors in our assays, we could not conclude that 1 and 2
were more selective for muscle GPa than for liver
GPa, since the rabbit muscle GPa used for the assay
was pure commercial product, while the rat liver GPa
used for assays was self-prepared and thus had a differ-
ent quality from the muscle enzyme. However, in either
muscle GPa or liver GPa assays, both 1 and 2 were at
least 4-fold more potent than caffeine.
Based on their potency in the enzyme assays, 1 and 2
were evaluated for their hypoglycemic activity in adren-
aline-induced diabetic mice since it is well known that
adrenaline is thought to induce high blood glucose by
indirect stimulating glycogenolysis and therefore
increasing hepatic glucose production. Test compounds
were administered orally at 100 mg/kg/day for 7 days.
Not surprisingly, the preliminary animal study results
-10
1
10
100
1000
Concentration, µM
Figure 1. Inhibition of rabbit muscle GPa by 1, 2, and caffeine. GPa
activity was measured at varied concentrations of 1 (j), 2 (Ç) or
caffeine (m). The data are plotted as percent inhibition versus
concentration of compound. Results are means SEM for three
independent experiments.

X. Wen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4944–4948
4947
Table 1. Effect of 1 and 2 on fasted plasma glucose of hyperglycemic
mice induced by adrenaline (n = 10)
4. Ahn, K. S.; Hahm, M. S.; Park, E. J.; Lee, H. K.; Kim, I.
H. Planta Med. 1998, 64, 468.
5. Judy, W. V.; Hari, S. P.; Stogsdill, W. W.; Judy, J. S.;
Naguib, Y. M. A.; Psswater, R. J. Ethnopharmacol. 2003,
87, 115.
Compound Dose
(mg/kg)
Fasted plasma glucose (OD)
2 h 4 h
0 h
6. Taniguchi, S.; Imayoshi, Y.; Kobayashi, E.; Takamatsu,
Y.; Ito, H.; Hatano, T.; Sakagami, H.; Tokuda, H.;
Nishino, H.; Sugita, D.; Shimura, S.; Yoshida, T. Phyto-
chemistry 2002, 59, 315.
Vehicle
0.138 0.028 0.149 0.037 0.124 0.019
0.111 0.019 0.076 0.026 0.083 0.017
0.117 0.032 0.081 0.027 0.079 0.023
1
2
100
100
7. Xu, H. X.; Zeng, F. Q.; Wan, M.; Sim, K. Y. J. Nat. Prod.
1996, 59, 643.
(Table 1) showed that both compounds significantly
inhibited the increase of fasted plasma glucose of diabet-
ic mice induced by adrenaline. To the best of our knowl-
edge, this is the first report that 2 is capable of reducing
blood glucose of diabetic mice, while 1 is a well-known
hypoglycemic agent. Further in vivo studies on 1, 2,
and related products as hypoglycemic agents are
ongoing.
8. Montilla, M. P.; Agil, A.; Navarro, M. C.; Jimenez, M. I.;
Garcia-Granados, A.; Parra, A.; Cabo, M. M. Planta
Med. 2003, 69, 472.
9. Caglioti reported a partial synthesis of methyl maslinate
based on a hydroboration-oxidation reaction, however,
the overall yield was very poor, and moreover, it is very
difficult (if possible) to convert methyl maslinate to
maslinic acid by conventional acid or base catalyzed
hydrolysis Caglioti, L.; Cainelli, G. Tetrahedron 1962, 18,
1061.
In summary, we have identified corosolic acid (1) and
maslinic acid (2) as a new class of inhibitors of glyco-
gen phosphorylases. This discovery affords novel lead
compounds for developing potent GP inhibitors. As
an effort to identify the possible binding sites in GP
for 1 and 2, molecular modeling study is still in pro-
cess. Based on the in vitro and in vivo studies, it is
therefore proposed that naturally occurring pentacyclic
triterpenes 1 and 2 might reduce blood glucose, at least
in part, through inhibiting hepatic glycogen degrada-
tion. The advantages of these natural GP inhibitors
as anti-diabetic agents are obvious: (a) they are mild
GP inhibitors, and thus side effects caused by strong
and nonselective GP inhibition by synthetic GP inhib-
itors are avoidable; (b) they are nontoxic, for example,
2 is an abundant constituent of olive fruit;11 (c) except
for lowering blood glucose, they also exhibit other
therapeutic benefits such as anti-inflammation2 and
antioxidant activities,8 etc. Extensive research on lead
optimization and biological evaluation of pentacyclic
triterpenoids as a new class of GP inhibitors is in pro-
gress in our laboratory. Further reports will describe
structure–activity relationships in a wider series of pen-
tacyclic triterpenoids.
10. Wang, P.; Lei, L. U.S. Patent Application 20030165581,
2003.
11. Noriyasu, N.; Shinohara, G. PCT Patent, WO 0212159,
2002.
12. (a) Oikonomaos, N. G. Curr. Protein Pept. Sci. 2002, 3,
561; (b) 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.; Willough-
by, C. A.; Yao, J.; Chapman, K. Bioorg. Med. Chem. Lett.
2003, 13, 4125; (c) 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; (d)
Treadway, J. L.; Mendys, P.; Hoover, D. J. Expert Opin.
Investig. Drugs 2001, 10, 439, and references therein.
13. Tracey, W. R.; Treadway, J. L.; Magee, W. P.; Sutt, J. C.;
McPherson, R. K.; Levy, C. B.; Wilder, D. E.; Yu, L. J.;
Chen, Y.; Shanker, R. M.; Mutchler, A. K.; Smith, David
M.; Flynn, A. H.; Knight, D. R. Am. J. Physiol. Heart
Circ. Physiol. 2004, 286, H1177.
14. Schnier, J. B.; Nishi, K.; Monks, A.; Gorin, F. A.;
Bradbury, E. M. Biochem. Biophys. Res. Commun. 2003,
309, 126.
15. Analytical data for 1: mp 253–255 °C (mp 255–258 °C,
1
Ref. 6); IR (KBr, cmÀ1) 3414, 2945, 1695, 1456, 1049; H
NMR (pyridine-d5, 300 MHz): d 0.94 (3H, s, H-29), 0.96
(3H, s, H-30), 0.97 (3H, s, H-25), 1.02 (3H, s, H-24), 1.05
(3H, s, H-26), 1.19 (3H, s, H-27), 1.25 (3H, s, H-23), 2.60
(1H, d, J = 11.3 Hz, H-18), 3.36 (1H, d, J = 9.4 Hz, H-3a),
4.06 (1H, ddd, J = 4.3, 9.4, 11.1 Hz, H-2b), 5.44 (1H, t,
J = 3.3 Hz, H-12); 13C NMR (pyridine-d5, 300 MHz): d
17.0 (C-25), 17.5 (2C, C-26, C-30), 17.7 (C-24), 18.9 (C-6),
21.4 (C-29), 23.8 (C-11), 23.9 (C-27), 24.9 (C-16), 28.7 (C-
15), 29.4 (C-23), 31.1 (C-21), 33.5 (C-7), 37.5 (C-22), 38.5
(C-10), 39.4 (C-19), 39.5 (C-20), 39.8 (C-4), 40.1 (C-8),
42.6 (C-14), 48.0 (C-1), 48.1 (C-17), 53.6 (C-18), 60.0 (C-
5), 68.6 (C-2), 83.8 (C-3), 125.5 (C-12), 139.3 (C-13), 179.9
(C-28); MS: 495 [M+Na]+. The spectrum data of 1 were
identical with the reported data: see Ref. 6.
Acknowledgments
We hope to thank Mr. Shanzhi Wang and Ms. Rui Niu
for their technical supports in enzyme assay studies.
References and notes
1. Miura, T.; Itoh, Y.; Kaneko, T.; Ueda, N.; Ishida, T.;
Fukushima, M.; Matsuyama, F.; Seino, Y. Biol. Pharm.
Bull. 2004, 27, 1103.
16. Analytical data for 2: mp 269–271 °C (mp 266–269 °C,
Ref. 22); IR (KBr, cmÀ1) 3414, 2943, 1695, 1460, 1051; 1H
NMR (pyridine-d5, 300 MHz): d 0.93, 0.98, 0.99, 1.01,
1.06, 1.25, 1.26 (each, 3H, s), 3.28 (1H, dd, J = 3.9,
13.6 Hz, H-18), 3.37 (1H, d, J = 9.3 Hz, H-3a), 4.07 (1H,
ddd, J = 4.2, 9.3, 11.0 Hz, H-2b), 5.46 (1H, br s, H-18);
13C NMR (pyridine-d5, 300 MHz): d 16.9 (C-24), 17.5 (C-
25), 17.7 (C-26), 18.9 (C-6), 23.7 (C-16), 23.8 (C-30), 23.9
2. Banno, N.; Akihisa, T.; Tokuda, H.; Yasukawa, K.;
Higashihara, H.; Ukiya, M.; Watanabe, K.; Kimura, Y.;
Hasegawa, J.; Nishino, H. Biosci. Biotechnol. Biochem.
2004, 68, 85.
3. Miyako, Y.; Masahiro, F.; Tsuneatsu, N.; Hikaru, O.;
Kazuhisa, M.; Jiro, T.; Yoshiharu, K.; Ryota, T.; Junei,
K.; Kunihide, M.; Toshihiro, F. Biol. Pharm. Bull. 2005,
28, 173.

4948
X. Wen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4944–4948
(C-30), 26.2 (C-27), 28.3 (C-15), 29.3 (C-23), 31.0 (C-20), 18. Engers, H. D.; Shechosky, S.; Madsen, N. B. Can. J.
33.2 (C-7), 33.3 (2C, C-22, C-29), 34.3 (C-21), 38.5 (C-10),
39.8 (C-4), 42.0 (C-19), 42.2 (C-14), 46.7 (C-17), 47.8 (C-
1), 48.2 (2C, C-8, C-9), 55.9 (C-5), 68.6 (C-2), 83.8 (C-3),
122.5 (C-12), 144.9 (C-13), 180.2 (C-28); MS: 495
[M+Na]+. The spectrum data of 2 were identical with
the reported data: see Ref. 6.
Biochem. 1970, 48, 746.
19. Kasvinsky, P. J.; Shechosky, S.; Fletterick, R. J. J. Biol.
Chem. 1978, 253, 9102.
20. Ercan-Fang, N.; Nuttall, F. Q. J. Pharmcol. Exp. Ther.
1997, 280, 1312.
21. Fosgerau, K.; Westergaard, N.; Quistorff, B.; Grunnet, N.;
Kristiansen, M.; Lundgren, K. Arch. Biochem. Biophys.
2000, 830, 274.
22. Goswami, A.; Khastgir, H. N. Indian J. Chem., Sect. B
1980, 19, 315.
17. Miura reported that 1 reduced blood glucose of KK-Ay
mice, at least in part, through increasing GLUT4 trans-
location in muscle and thus stimulating glucose uptake:
see Ref. 1.