Synthesis and antihyperglycemic evaluation of various protoberberine derivatives

Article abstract of DOI:10.1016/j.bmcl.2005.11.045

Various berberine derivatives (2-17) were synthesized and their antihyperglycemic activities were evaluated in a model of β-cell-membrane chromatography and a model of alloxan-induced diabetes mice. The results indicated that compounds 5 and 14 exhibited antihyperglycemic activity. Their structure-activity relationships were discussed.

Full text of DOI:10.1016/j.bmcl.2005.11.045

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1380–1383
Synthesis and antihyperglycemic evaluation
of various protoberberine derivatives
Xiaoli Bian, Langchong He* and Guangde Yang
School of Medicine, XiÕan Jiaotong University, Xi’an 710061, PR China
Received 17 June 2005; revised 29 October 2005; accepted 14 November 2005
Available online 15 December 2005
Abstract—Various berberine derivatives (2–17) were synthesized and their antihyperglycemic activities were evaluated in a model of
b-cell-membrane chromatography and a model of alloxan-induced diabetes mice. The results indicated that compounds 5 and 14
exhibited antihyperglycemic activity. Their structure–activity relationships were discussed.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Type 2 diabetes (T2DM) affects a large proportion of
the adult population and is closely associated with other
conditions, including obesity, dyslipidemia, and cardio-
vascular disease.1–4 It is well known that the loss of
the insulin secretion function and insulin resistance con-
tribute to this disease.5,6 Sulfonylureas, which are widely
used as hypoglycemic agents for T2DM, strongly inhibit
ATP-sensitive K+ channel activity by binding to sulfo-
nylurea receptors in pancreatic b-cells.7,8 Pancreatic
b-cells and related sulfonylurea receptors are the targets
on which sulfonylurea molecules and other hypoglyce-
mic agents act. Pancreatic b-cells and sulfonylurea
receptors were used to investigate the binding
interactions between active ligands and their related
targets.9–11 The reported methods that have been used
to analyze interactions between b-cell sulfonylurea
receptors and drug molecules include: direct measure-
ment of K+ ion currents,12 radioligand binding assays,13
and column-based counting of flow-through radioactiv-
ity.14 In this study, cell-membrane chromatography
(CMC)15 was used to investigate the binding interac-
tions between b-cell-related receptors and drug mole-
cules. CMC is a bio-affinity chromatographic method,
in which bioactive cell membranes are immobilized on
the silica carrier as a stationary phase. In this chromato-
graphic system, the active molecules that have binding
affinity to the cell membrane and its related targets will
be retained. The interaction characteristics are reflected
directly by the elution data.15–18
Berberine 1, a major compound present in Coptis
chinensis France (a Chinese traditional medicine), has
been used for the treatment of diarrhea and other
gastrointestinal diseases in China and other Asian
countries.19,20 It has been reported that berberine 1
has hypoglycemic effects.21 However, it is not applicable
to use berberine 1 for the treatment of diabetes as, when
it is used as a hypoglycemic agent for an extended
period, serious side effects occur due to its antibacterial
activity. In this study, in order to find new structures
that have high a hypoglycemic effect but little antibacte-
rial activity, 16 protoberberine derivatives were synthe-
sized with berberine as a lead compound, and their
hypoglycemic effects were screened in a CMC model
and then in an alloxan-induced diabetic model, and their
structure–activity relationships were analyzed.
Berberine 1 has a dibenzo[a,g]quinolizidine ring system
structure. The quaternary salts that have an aromatic
C ring and a 2,3-methylenedioxy moiety are essential
components for the antibacterial activities of protobe-
rines.22–24 Therefore, compounds 2–4 were derived from
replacing the methylenedioxy function of berberine by
–OH, –OCH3, and –OCOCH3, respectively.
Compounds 5–10, which have different substitutes at
positions C-9, C-10 and C-11, were prepared via the to-
tal synthetic route from the starting material of homo-
piperonylamine and different aromatic aldehydes. The
aromatic quaternary nitrogen was reduced to tertiary
nitrogen (11,16), then alkylated to produce N-alkyl
Keywords: Protoberberine derivatives; Hypoglycemic effect; Cell
membrane chromatography.
*
Corresponding author. Tel.: +86 29 82655451; fax: +86 29
82655451; e-mail: helc@mail.xjtu.edu.cn
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.11.045

X. Bian et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1380–1383
1381
derivatives 12–14 and 17, and oxidized to N-oxide com-
pound 15. Their structures are summarized in Table 1.
The general synthetic route that is used to synthesize
the designated compounds is outlined in Scheme 1.
methylenedioxy function was replaced by groups of
hydroxyl (2), methyloxyl (3), and acetoxyl (4),
respectively.
The quaternary salts have an aromatic C ring, which is a
necessary structure of berberine derivatives for antibac-
terial activity.22–24 When the quaternary salts were re-
duced to tertiary amines, their binding affinity to b-cell
membranes disappeared. This suggests that the electro-
static effect of positively charged nitrogen atoms in the
aromatized C ring may play an important role in bind-
ing interactions. However, when tetrahydroberberine
was alkylated to quaternary ammonium salts, the binding
activity changed with different alkyl derivatives. N-Meth-
yl 12 and N-ethyl 13 displayed no binding activity. This
could be explained by the fact that the electrostatic effects
of positively charged nitrogen atoms were shielded by the
methyl or ethyl groups. However, N-benzyl 14 and 17 pos-
sessed high binding affinity. In this case, the interaction
forces may be different due to the hydrophobic properties
of the benzyl group. These results suggested that the elec-
trostatic effects of positively charged nitrogen atoms may
be non-specific binding forces for berberine and that the
quaternary ammonium salt structure (that has an aromat-
ic C ring) is not necessary for the hypoglycemic effects of
berberine derivatives to occur.
Protoberberines 1–17 were tested in vitro against Esche-
richia coli. The minimum inhibitory concentration
(MIC) of each compound was determined using a twofold
serial-dilution technique.25 The MIC value of compound
1 was 5 lg/mL. There was still no antibacterial activity
when the concentration of compounds 2–17 was raised
to 50 lg/mL. The protoberberines 2–17 do not possess
as much antibacterial potential compared with berberine
1. This result corresponds with the findings reported in the
literature.22–24
Compounds 1–17 were screened using a model of pan-
creatic-islet b-cell-membrane chromatography.18,26 The
binding affinity of the compounds on b-cell membrane
and membrane receptors in the CMC system was
reflected by the logarithm of capacity factor (logk0).
The results indicated that compounds 1, 5, 10, 14, 15,
and 17 exhibited retentive behavior similar to that of
gliquidone in the model. The others had no retention
under the same conditions (Table 2). The chromato-
graphic conditions were given in reference.26
The hypoglycemic effect of compounds that have reten-
tive behavior (1, 5, 10, 14, and 15) was tested in alloxan-
induced diabetes mice, as shown in Table 3. Compound
17 was not tested because of its obvious toxicity at
experimental dosages.
Berberine 1 showed two methoxyl groups at positions C-9
and C-10 (ring D); it also possessed high binding affinity
and hypoglycemic activity. Compound 5, without a sub-
stituent on ring D, also possessed high binding affinity
and hypoglycemic activity. However, introduction of a
methoxyl group at the C-9 position of compound 5 led
to a decrease in binding activity. Similar results were
observed when a methoxyl group was introduced at the
C-10 position, the C-11 position and a benzyloxy group
at the C-11 position of compound 5 (compounds 8, 9,
and 10, respectively). This indicated that binding affinity
The results derived from CMC screening indicated that
the methylenedioxy function at C-2 and C-3 may be essen-
tial for the binding activity of protoberberines to b-cell
membranes. The binding activity disappeared when the
methylenedioxy ring of berberine was opened and the
Table 1. Protoberberinum salts
R1
O
R
X
N
X
N
O
R2
R3
R4
R3
R4
R5
Compound
R1
R2
OCH2O
R3
R4
R5
Compound
R3
R4
R
1
2
OMe
OMe
OMe
OMe
H
OMe
OMe
OMe
OMe
H
H
11
12
13
14
15
16
17
OMe
OMe
OMe
OMe
OMe
H
OMe
OMe
OMe
OMe
OMe
H
H
Me
OH
OMe
OCOMe
OH
OMe
OCOMe
H
3
H
Et
CH2C6H5
4
H
5
OCH2O
H
O
6
OCH2O
OCH2O
OCH2O
OCH2O
OCH2O
OMe
H
H
H
H
CH2C6H5
7
OMe
H
H
OMe
H
H
8
H
9
10
H
OH
H
OMe
OCH2C6H5
H

1382
X. Bian et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1380–1383
R3
CHO
R4
R5
R3
R4
CH2CH2N CH
CH2CH2NH2
O
O
O
O
b
a
R5
R3
O
O
H
CH2CH2N CH2
R4
R5
X
N
O
O
c
d
CHO
CHO
R3
R4
R5
O
O
X
R
N
O
N
O
e
R3
R4
R3
R4
R5
R5
Scheme 1. Synthesis of compounds 2–17. Reagents and conditions: (a) 110 ꢁC, 1 h, 85%; (b) NaBH4, CH3OH, reflux, 2 h, 85%; (c) HOAc, CuSO4,
NaCl, 50 ꢁC 0.5 h, then 85 ꢁC, 6 h, 60%; (d) Zn, HCl, reflux, 3 h, 65%; (e) 42 ꢁC, 2 h, 86%.
Table 2. Retention features of compounds 1–17 in the model of b-cell membrane chromatography
Compound:
1
2–4
5
6–9
10
11–13
14
15
16
17
GQ
logk0
1.32
—
1.44
—
1.09
—
1.73
1.32
—
1.56
1.59
Table 3. Effect of compounds on hyperglycemia in mice with alloxan-
induced diabetes ðꢀx ꢀ sÞ
14, and 15 was 0.8714. This indicated that there was a
good correlation between the b-CMC model and the
mouse model with alloxan-induced diabetes.
Group
Dose (mg/kg)
n
Blood glucose (mg/dL)
Initial Final Reduction
As a bio-affinity screen model, the CMC model could be
used to investigate ligands that have potential binding
affinity with targets on the cell membrane.15,16,18 The
retentive behavior of the active compounds in the pan-
creatic-islet CMC model (shown in Table 1) indicated
that these compounds have a binding affinity for targets
in pancreatic-islet b-cell membranes that are similar to
those of gliquidone. They may produce hypoglycemic
effects through similar mechanisms as sulfonylurea.
The hypoglycemic effects can be confirmed in diabetes
animal experiments. Although it cannot be confirmed
that the binding site is a sulfonylurea receptor, com-
pounds that have no interaction with b-cell membranes
were excluded rapidly from the CMC model. Therefore,
the number of compounds that are available for animal
experiments is greatly reduced. In the CMC model, the
separation and purification of receptor proteins are
not necessary, and the fast and efficient characteristics
of chromatography make it a time- and resource-effi-
cient method for screening new drugs.
Normal
Control
10 122 23 110 20 ꢁ14.4 2.3
13 580 16 558
9
7
14 572 511 11
ꢁ22 8*
1
200
200
200
200
200
200
ꢁ61
3
5
13 591 14 536
12 562 11 574
4
9
5
4
3
ꢁ67 4*
10
14
15
GQ
ꢁ12
5
13 558
14 572
13 563
3
6
5
412
546
408
ꢁ146 3**
ꢁ23
3
5
ꢁ162
* p < 0.01 versus normal.
** p < 0.05 versus control.
may relate to substituents on ring D. This feature could
not be concluded here, as further study results are
required.
Of the compounds that had retentive behavior, com-
pounds 1, 5, and 14 showed significant hypoglycemic
activity. Hypoglycemic activity of compound 17 was
not tested due to its obvious toxicity at experimental
dosages. Compound 14 presented the highest hypoglyce-
mic activity. The relative coefficient between the logk0
values and the reduction values of compounds 1, 5, 10,
From the binding activity–structure point of view, some
features can be pointed out. On the basis of the results

X. Bian et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1380–1383
1383
9. Gribble, F. M.; Ashcroft, F. M. Metab. Clin. Exp. 2000,
49, 3.
10. Andrew, J. P.; Claire, T.; Craig Mckay, M. Bioorg. Med.
Chem. Lett. 2004, 14, 813.
11. Ashcroft, F. M.; Gribble, F. M. Diabetes 1999, 42, 903.
12. Gribble, F. M.; Tucker, S. J. Diabetes 1998, 47, 1412.
13. Mikhailov, M. V.; Mikhailov, E. A. FEBS Lett. 2000, 482,
59.
14. Mikhailov, M. V.; Ashcroft, S. J. H. J. Biol. Chem. 2000,
275, 3360.
15. Langchong, H.; Xindu, G. New Progr. Biomed. Chroma-
togr. 1996, 3, 8.
16. Langchong, H.; Guangde, Y.; Xindu, G. Chin. Sci. Bull.
1999, 44, 826.
17. Langchong, H.; Sicen, W.; Geng, X. Chromatographia
2001, 54, 77.
derived from CMC screening with protoberberinium, it
seems that the substations that exist on ring A are
important for binding activity. The binding affinity dis-
appeared when the methylenedioxy function at C-2 and
C-3 was replaced by other groups. The binding interac-
tions were affected by the station and environment of
N-7. The aromatic quaternary nitrogen may not be the
structure of berberine derivatives that is required for
hypoglycemic effects to occur. For quaternary salts,
which were produced by alkylation of tetrahydrogen
berberine, the hydrophobic force of alkyl groups may
play an important role in their binding interaction. Of
all the synthesized protoberberiniums, compound 14
was the most effective in lowering blood glucose.
18. Guangde, Y.; Langchong, H.; Xiaoli, B. Chin. Sci. Bull.
2005, 50, 1709.
19. Liu, D.; Yang, L. X.; Cui, S.-L. China J. Chin. Mater.
Med. 1997, 22, 79.
20. Iwasa, K.; Hye-Sook, K.; Yusuke, W. Eur. J. Med. Chem.
1998, 33, 65.
Acknowledgment
This work was supported by the Natural Science
Foundation of Shaanxi Province (No. 2003C205).
21. Jie, Z.; Liqin, W. Chinese J. Intergrative Med. Cardio/
Cerebrovasc. Dis. 2003, 1, 160.
22. Iwasa, K.; Nanba, H.; Lee, D.-U. Planta Med. 1998, 64,
748.
References and notes
1. DeFronzo, R. A.; Bonadonna, R. C.; Ferrannini, E.
Diabetes Care 1992, 15, 318.
23. Iwasa, K.; Kam imauchi, M.; Ueki, M. Eur. J. Med.
Chem. 1996, 31, 469.
2. Kahn, S. E.; Porte, D. J. Diabetes Mellitus 1996, 45, 487.
3. Ruderman, N. B.; Schneider, S. H. Diabetes Care 1992, 15,
1784.
4. Satyanarayana, M.; Priti, T.; Brajendra, K. Bioorg. Med.
Chem. 2004, 12, 883.
5. Bennett, P. H. Diabet. Metab. Rev. 1997, 13, 583.
6. Ward, W. K.; Beard, J. C.; Halter, J. B.; Pfeifer, M. A.;
Porte, D. J. R., Jr. Diabetes Care 1984, 7, 491.
7. Bailey, C. J. Proc. Nutr. Soc. 1991, 50, 619.
8. Steven, E.; Kahn, M. B. Am. J. Med. 2000, 108, 2s.
24. Iwasa, K.; Kam imauchi, M. Planta Med. 1997, 63, 196.
25. Shuyun, X.; Rulian B.; Xiu C. In Method of Pharmaco-
logical Experiment. The PeopleÕs Medical Publishing
House, Beijing, 2002; pp 1037.
26. Note: the chromatographic conditions in the CMC model:
rabbit pancreatic-islet b-cell membrane column
(50 mm · 2.0 mm i.d.) was used. The mobile phase was
25 mM ammonium sulfate buffer (pH 7.4), with a flow rate
of 0.5 mL/min. The detection wavelength was 262 nm, at a
column temperature of 37 0.5 ꢁC.