Synthesis of linked berberine dimers and their remarkably enhanced DNA-binding affinities

Article abstract of DOI:10.1016/j.bmcl.2004.10.098

This communication describes the facile synthesis of five novel berberine dimers and their strong affinities toward double-stranded DNA. These berberine dimers were synthesized in 37-84% yields from the reaction of berberrubine with dihaloalkanes of varying lengths, and fully characterized by HRMS and ¹H NMR. Compared with the monomeric parent berberine, these dimers showed greatly enhanced binding affinities up to approximately 100-fold, with two double helical oligodeoxynucleotides, d(AAGAATTCTT)₂ and d(TAAGAATTCTTA)₂, which was investigated by means of fluorescence spectrometry.

Full text of DOI:10.1016/j.bmcl.2004.10.098

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2689–2692
Synthesis of linked berberine dimers and their remarkably
enhanced DNA-binding affinities
Wen-Hua Chen,a,* Ji-Yan Pang,a,b Yong Qin,b Qian Peng,a
Zongwei Caic and Zhi-Hong Jiangb,*
aSchool of Chemistry and Chemical Engineering, Sun Yat-Sen University, Guangzhou 510275, PR China
bSchool of Chinese Medicine, Hong Kong Baptist University, Kowloon Tong, Hong Kong
cDepartment of Chemistry, Hong Kong Baptist University, Kowloon Tong, Hong Kong
Received 28 September 2004; accepted 30 October 2004
Available online 12 April 2005
Abstract—This communication describes the facile synthesis of five novel berberine dimers and their strong affinities toward double-
stranded DNA. These berberine dimers were synthesized in 37–84% yields from the reaction of berberrubine with dihaloalkanes of
1
varying lengths, and fully characterized by HRMS and H NMR. Compared with the monomeric parent berberine, these dimers
showed greatly enhanced binding affinities up to approximately 100-fold, with two double helical oligodeoxynucleotides,
d(AAGAATTCTT)2 and d(TAAGAATTCTTA)2, which was investigated by means of fluorescence spectrometry.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Much interest is currently focused on the development
of efficient DNA-binding agents with improved binding
affinities and site specificities toward target DNA, be-
cause such agents have potentially wide applications,
for example, in elucidating the action mechanism of
antitumor and antivirus drugs1 and in developing new
chemotherapeutic agents.2 One sophisticated approach
to obtain more effective DNA-binding candidate com-
pounds is the modification of natural products whose
binding affinities and modes with DNA have been well
established. A typical example of this is the rational
structural modification of the antibiotics netropsin and
distamycin leading to the development of lexitropsins.3
binding modes as a groove binder or an intercalator re-
main controversial. The binding affinities of berberine,
usually in the range of 1.0–2.0 · 104 Mꢀ1 with duplex
DNA, are modest and necessitate the appropriate struc-
tural modification as to serve as a novel DNA-binding
agents with enhanced binding affinities. In this paper,
we describe the remarkably enhanced DNA-binding
affinities of berberine dimers 3a–e (Scheme 1) synthe-
sized in moderate to high yields from the linkage of
berberrubine 25 with alkyl chains of varying lengths,
aiming at developing more effective protoberberine alka-
loid-based DNA-binding agents.
Berberine dimers 3a–e were synthesized as described in
Scheme 1. Pyrolysis of berberine 1 at 190 ꢁC under vac-
uum, according to the reported protocols,5 gave berber-
rubine 2 in 60% yield. Coupling of 2 with dibromo- or
diiodoalkanes of varying lengths from two to six car-
bons in DMF, followed by the exchange of all anions
into chloride, afforded corresponding berberine dimers
3a–e in 37–84% yields. The structures of compounds
3a–e were confirmed by HR-ESI-MS and NMR.6 All
these compounds afforded the two-charged ESI-MS
In these aspects, berberine, a typical naturally-occurring
protoberberine isoquinoline alkaloid, is attractive as a
versatile platform for the development of more efficient
DNA-binding agents, because berberine is known as a
DNA binder and its binding affinities have been exten-
sively explored by several analytical techniques includ-
ing absorption, fluorescence, NMR, and electrospray
ionization mass (ESI-MS) spectrometries,4 though its
1
peaks ([Mꢀ2Cl]2+) in the mass spectra. The H NMR
spectra of these compounds, for example, 3a in Figure
1, showed only one set of berberine subunits, indicating
the formation of a symmetric structure. The ratios of the
integrated areas for the protons of berberine subunits to
those for the alkyl chain protons in the 1H NMR spectra
Keywords: Berberine dimer; DNA; Noncovalent interaction; Fluores-
cence spectrometry.
*
Corresponding authors. Tel.: +852 3411 2906; fax: +852 3411 2461
(Z.-H.J.); e-mail addresses: cescwh@zsu.edu.cn; zhjiang@hkbu.edu.
hk
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.10.098

2690
W.-H. Chen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2689–2692
O
O
O
Cl-
Cl-
190oC
N
N
O
OMe
OMe
OH
OMe
1
2
O
O
O
O
X
X
1)
n
DMF, 60oC, 12 h
Cl-
Cl- N+
N
3a: n = 0, X = Br (46%)
3b: n = 1, X = Br (37%)
3c: n = 2, X = Br (84%)
3d: n = 3, X = I (47%)
3e: n = 4, X = I (74%)
2) Anion Exchange
O
O
n
MeO
OMe
Scheme 1. Synthetic route for berberine dimers 3a–e.
a
O
O
4
5
O
O
b
6
8
1
N
N+
13
12
1'
O
O
a
11
OCH3
b
H3CO
4
1
'
1
13
11
6
5
12
8
8
ppm
0
10
9
7
6
5
4
3
2
1
Figure 1. 1H NMR (300 MHz) spectrum of 3a in DMSO-d6.
were in full accord with the linked berberine dimeric
structures.
200
200
150
100
50
The binding affinities of 3a–e toward two self-comple-
mentary double-stranded oligodeoxynucleotides, d(AA-
GAATTCTT)2 and d(TAAGAATTCTTA)2, were
evaluated by means of fluorescence spectrometry. Figure
2 shows the fluorescence titration spectra of 3b with
d(TAAGAATTCTTA)2. It is observed that the weak
fluorescence of 3b was largely enhanced upon the addi-
tion of DNA, suggesting the interaction of 3b with
DNA added. The 1:1 binding stoichiometry between
3b and d(TAAGAATTCTTA)2 was determined by mo-
lar ratio methods (inset in Fig. 2). Analyses of the rela-
tionship between the fluorescence intensities and the
DNA concentrations by nonlinear curve fitting meth-
ods7 afforded the association constants (KaÕs) of 3a–e
with d(AAGAATTCTT)2 and d(TAAGAATTCTTA)2
(Fig. 3 and Table 1). For comparison, the binding con-
stants of 1 with the two oligodeoxynucleotide duplexes
were measured in a similar way. The binding affinities
of 3a–e relative to 1 with d(AAGAATTCTT)2 and
d(TAAGAATTCTTA)2 are illustrated in Figure 4.
150
100
50
0
0
1
2
3
4
[d(TAAGAATTCTTA)2] / [3b])
[d(TAAGAATTCTTA)2]
0
400
450
500
550
600
650
Wavelength (nm)
Figure 2. Fluorescence spectra of 3b (2.47 · 10ꢀ6 M) with d(TAA-
GAATTCTTA)2 of increasing concentrations (0–8.89 · 10ꢀ6 M) in
50 mM Tris–HCl (pH 6.35) at room temperature, ex 355 nm. The inset
indicates the relationship between the fluorescence intensities (at em
516.7 nm) and the molar ratio of d(TAAGAATTCTTA)2 to 3b.
berberine dimers 3a–e exhibited much higher affinities
than their monomeric parent compound 1. It can be
seen that 3a–e bound to d(AAGAATTCTT)2 at least
Some interesting observations can be extracted from Ta-
ble 1 and Figure 4. The first observation is that linked

W.-H. Chen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2689–2692
2691
comparable with those of some natural antibiotics. This
was further supported by the ESI-MS results on the
comparison of the binding affinities of dimers 3a–e
and berberine 1. This great increase in the binding affini-
ties may be ascribed to the cooperative interactions of
the two berberine subunits in 3a–e. Secondly, com-
pounds 3a–e showed a prominent structure–activity cor-
relation and their binding abilities with both duplex
DNA could be modulated by the lengths of the alkyl
chains. Five dimers 3a–e showed similar relative binding
affinities toward both oligodeoxynucleotide duplexes in
the order of 3b > a > c > d > e. Dimer 3b linked by a
propyl chain exhibited the highest affinity in both cases,
indicating that propyl chain may be the most suitable
linker to bridge the two berberine units. This result
may guide future rational design efforts. Thirdly, com-
pounds 3a–e showed higher affinities with 12-mer
d(TAAGAATTCTTA)2 than with 10-mer d(AA-
GAATTCTT)2. For example, the binding constants of
3a–b with d(TAAGAATTCTTA)2 were over 10-fold
greater than with d(AAGAATTCTT)2. This result sug-
gested that dimeric berberines would necessarily occupy
a greater number of base pairs,8 with potentially more
stringent sequence recognition relative to the mono-
meric berberine.
200
160
120
80
Berberine
3b
200
160
120
80
40
40
0
0
2
4
6
8
10
[d(TAAGAATTCTTA)2] (uM)
0
0
20
40
60
80
100
[d(TAAGAATTCTTA)2] (uM)
Figure 3. Fluorescence titrations of
1
(2.08 · 10ꢀ6 M) and 3b
(2.47 · 10ꢀ6 M) with d(TAAGAATTCTTA)2 in 50 mM Tris–HCl
(pH 6.35) at room temperature, ex 355 nm, em 516.7 nm. Inset shows
data for 3b with an expanded x-axis.
Table 1. Association constants (KaÕs, Mꢀ1
d(AAGAATTCTT)2 and d(TAAGAATTCTTA)2
)
a
of 3a–e with
Compound
d(AAGAATTCTT)2
Ka
RAb Ka
(1.24 0.12) · 104 (2.93 0.31) · 104
(1.18 0.10) · 105
d(TA AGAATTCTTA)2
In summary, five novel spacer-linked berberine dimers
have been successfully synthesized in moderate to high
yields. These berberine dimers show much higher
DNA-binding affinities than their monomeric parent
berberine. Great efforts are being currently made to sys-
tematically explore the DNA-binding mechanisms and
sequence selectivities of these linked berberine dimers.
RAb
1
1
1
3a
3b
3c
3d
3e
9.5 (1.62 0.35) · 106 54.9
(2.46 0.07) · 105 19.8 (2.76 0.37) · 106 94.2
(8.78 0.26) · 104
(5.77 0.13) · 104
(3.42 0.07) · 104
7.1 (3.49 0.60) · 105 11.9
4.7 (3.21 0.20) · 105 11.0
8.5
2.8 (2.48 0.34) · 105
a In 50 mM Tris–HCl (pH 6.35) at room temperature.
b RA denotes relative affinity.
Acknowledgements
The authors are grateful to Professor Irving H. Gold-
berg, Department of Biological Chemistry and Molecu-
lar Pharmacology, Harvard Medical School, for his
critical reading of the manuscript. This work was finan-
cially supported by the Faculty Research Grants from
Hong Kong Baptist University and Research Grants
Council, University Grants Committee of Hong Kong.
100
d(AAGAATTCTT)2
d(TAAGAATTCTTA)2
80
60
40
20
0
References and notes
1. Wolkenberg, S. E.; Boger, D. L. Chem. Rev. 2002, 102,
2477–2495.
2. Barton, J. K. Science 1986, 233, 727–734; Chrisey, L. A.;
Bonjar, G. H. S.; Hecht, S. M. J. Am. Chem. Soc. 1988, 110,
644–646; Zein, N.; Sinha, A. M.; McGarhren, W. J.;
Ellestad, G. A. Science 1988, 240, 1198–1201.
3. Lown, J. W. Anti-Cancer Drug Des. 1988, 3, 25–40.
4. (a) Krey, A. K.; Halm, F. E. Science 1969, 166, 755–757; (b)
Saran, A.; Srivastava, S.; Coutinho, E.; Maiti, M. Indian J.
Biochem. Biophys. 1995, 32, 74–77; (c) Li, W.-Y.; Lu, H.;
Xu, C.-X.; Zhang, J.-B.; Lu, Z.-H. Spectrosc. Lett. 1998,
31, 1287–1298; (d) Li, W.-Y.; Lu, Z.-H. Microchem. J.
1998, 60, 84–88; (e) Gong, G.-Q.; Zong, Z.-X.; Song, Y.-M.
Spectrochim. Acta Part A 1999, 55, 1903–1907; (f) Mazzini,
S.; Bellucci, M. C.; Mondelli, R. Bioorg. Med. Chem. 2003,
11, 505–514; (g) Kumar, G. S.; Das, S.; Bhadra, K.; Maiti,
K. Bioorg. Med. Chem. 2003, 11, 4861–4870; (h) Kluza, J.;
3b
1
3a
3c
3e
3d
Figure 4. Relative binding affinities of 3a–e with d(AAGAATTCTT)2
and d(TAAGAATTCTTA)2 in 50 mM Tris–HCl (pH 6.35) at room
temperature.
2–20 times and to d(TAAGAATTCTTA)2 at least 10–90
times more strongly than 1. More significantly, the bind-
ing constant of 3b with d(TAAGAATTCTTA)2 did not
only increase by a factor of 94 relative to 1, but was

2692
W.-H. Chen et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2689–2692
Baldeyron, B.; Colson, P.; Rasoanaivo, P.; Mambu, L.;
349.1309. Found: 349.1318. Compound 3d: 1H NMR
(DMSO-d6, 300 MHz) d 1.65–1.80 (m, 2H, CH2CH2CH2O),
1.99 (m, 4H, CH2CH2CH2O), 3.15–3.23 (m, 4H, 5-H), 4.04
(s, 6H, OCH3), 4.32 (t, 4H, J = 6.3 Hz, OCH2CH2), 4.93
(t, 4H, J = 6.4 Hz, 6-H), 6.16 (s, 4H, OCH2O), 7.07 (s, 2H,
4-H), 7.75 (s, 2H, 1-H), 7.97 (d, 2H, J = 8.7 Hz, 12-H), 8.18
(d, 2H, J = 8.7 Hz, 11-H), 8.90 (s, 2H, 13-H), 9.75 (s, 2H,
8-H); HRMS for C43H40N2O8 ([Mꢀ2Cl]2+) calcd: 356.1387.
Found: 356.1385. Compound 3e: 1H NMR (DMSO-d6,
300 MHz) d 1.60 (s, 4H, CH2CH2CH2O), 1.90–2.00 (m, 4H,
CH2CH2CH2O), 3.15–3.23 (m, 4H, 5-H), 4.03 (s, 6H,
OCH3), 4.29 (t, 4H,J = 5.8 Hz, OCH2CH2), 4.92 (t, 4H,
J = 6.9 Hz, 6-H), 6.16 (s, 4H, OCH2O), 7.07 (s, 2H, 4-H),
7.75 (s, 2H, 1-H), 7.97 (d, 2H, J = 9.3 Hz, 12-H), 8.18 (d,
2H, J = 9.3 Hz, 11-H), 8.90 (s, 2H, 13-H), 9.73 (s, 2H, 8-H);
HRMS for C44H42N2O8 ([Mꢀ2Cl]2+) calcd: 363.1471.
Found: 363.1504.
Frappier, F.; Bailly, C. Eur. J. Pharm. Sci. 2003, 20, 381–
391; (i) Park, H.; Kim, E. H.; Sung, Y.-H.; Kang, M. R.;
Chung, I. K.; Cheong, C.; Lee, W. Bull. Korean Chem. Soc.
2004, 25, 539–544; (j) Chen, W.-H.; Chan, C.-L.; Cai, Z.;
Luo, G.-A.; Jiang, Z.-H. Bioorg. Med. Chem. Lett. 2004,
14, 4955–4959; (k) Chen, W.-H.; Qin, Y.; Cai, Z.; Chan, C.-
L.; Luo, G.-A.; Jiang, Z.-H. Bioorg. Med. Chem. 2005, 13,
1859–1866.
5. Iwasa, K.; Kamigauchi, M.; Ueki, M.; Taniguchi, M. Eur.
J. Med. Chem. 1996, 31, 469–478.
6. Structural data for compound 3a: 1H NMR (DMSO-d6,
300 MHz) d 3.17 (s, 4H, 5-H), 4.02 (s, 6H, OCH3), 4.79 (s,
4H, OCH2), 4.85 (s, 4H, 6-H), 6.16 (s, 4H, OCH2O), 7.08 (s,
2H, 4-H), 7.75 (s, 2H, 1-H), 7.98 (d, 2H, J = 9.6 Hz, 12-H),
8.19 (d, 2H, J = 9.6 Hz, 11-H), 8.91 (s, 2H, 13-H), 9.83 (s,
2H, 8-H); HRMS for C40H34N2O8 ([Mꢀ2Cl]2+) calcd:
335.1152. Found: 335.1129. Compound 3b: 1H NMR
(CD3OD–CDCl3, 300 MHz,) d 2.65 (m, 2H, CH2CH2O),
3.23 (t, 4H, J = 6.3 Hz, 5-H), 4.09 (s, 6H, OCH3), 4.76 (t,
4H, OCH2CH2), 4.92 (t, 4H, 6-H), 6.09 (s, 4H, OCH2O),
6.91 (s, 2H, 4-H), 7.59 (s, 2H, 1-H), 7.98 (d, 2H, J = 9.0 Hz,
12-H), 8.07 (d, 2H, J = 9.0 Hz, 11-H), 8.65 (s, 2H, 13-H),
7. Association constants are derived from nonlinear
curve fitting, using the equation I = I0 + ((I1 ꢀ I0)/
2[B]0) {([DNA]0 + [B]0 + 1/Ka) ꢀ (([DNA]0 + [B]0 + 1/Ka)2 ꢀ
4[DNA]0[B]0)1/2}, wherein [DNA]0 and [B]0 are the initial
concentrations of DNA and 3a–e, respectively. I, I0, and I
1
represent the fluorescence intensities of the sample, 3a–e
alone and the intensity when 3a–e are totally bound,
respectively. For reference, see: (a) Schneider, H.-J.; Yat-
simirski, A. K. In Principles and Methods in Supramolecular
Chemistry; J Wiley: New York, 2000, pp 137–143; (b) Xi,
Z.; Jones, G. B.; Qabaja, G.; Wright, J.; Johnson, F.;
Goldberg, I. H. Org. Lett. 1999, 1, 1375–1377.
9.87 (s, 2H, 8-H); HRMS for C41H36N2O8 ([Mꢀ2Cl]2+
)
calcd: 342.1231. Found: 342.1213. Compound 3c: 1H NMR
(DMSO-d6, 300 MHz) d 2.14 (s, 4H, CH2CH2O), 3.15–3.24
(m, 4H, 5-H), 4.04 (s, 6H, OCH3), 4.40 (t, 4H, J = 5.6 Hz,
OCH2CH2), 4.93 (s, 4H, 6-H), 6.17 (s, 4H, OCH2O), 7.08 (s,
2H, 4-H), 7.78 (s, 2H, 1-H), 7.98 (d, 2H, J = 9.0 Hz, 12-H),
8.19 (d, 2H, J = 9.0 Hz, 11-H), 8.92 (s, 2H, 13-H), 9.78 (s,
2H, 8-H); HRMS for C42H38N2O8 ([Mꢀ2Cl]2+) calcd:
8. Leng, F.; Priebe, W.; Chaires, J. B. Biochemistry 1998, 37,
1743–1752.