Synthesis, DNA binding, and cytotoxicity of 1,4-bis(2-amino-ethylamino)anthraquinone-amino acid conjugates

Article abstract of DOI:10.1016/j.bmc.2007.10.012

Two series of 1,4-bis(2-amino-ethylamino)anthraquinone-amino acid conjugates (BACs), ametantrone (AT)-amino acid conjugates (AACs) and mitoxantrone (MX)-amino acid conjugates (MACs), were designed and synthesized. The DNA binding of BACs was evaluated by

Full text of DOI:10.1016/j.bmc.2007.10.012

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 1006–1014
Synthesis, DNA binding, and cytotoxicity of
1,4-bis(2-amino-ethylamino)anthraquinone–amino acid conjugates
Ling-Wei Hsin,a,* Hui-Po Wang,b,ꢀ Pi-Hung Kao,a On Lee,a Wan-Ru Chen,a
Hung-Wei Chen,a Jih-Hwa Guh,a Ya-Ling Chan,a Chin-Ping His,a Ming-Show Yang,b
Tsai-Kun Lic and Chieh-Hua Leec
aInstitute of Pharmaceutical Sciences, College of Medicine, National Taiwan University, No. 1, Section 1, Jen-Ai Road,
Room 1336, Taipei 10018, Taiwan, ROC
bInstitute of Natural Product, Chang Gung University, 259 Wen-Hwa 1st Road, Taoyuan 333, Taiwan, ROC
cDepartment and Graduate Institute of Microbiology, College of Medicine, National Taiwan University,
No. 1, Section 1, Jen-Ai Road, Taipei 10018, Taiwan, ROC
Received 8 July 2007; revised 1 October 2007; accepted 4 October 2007
Available online 10 October 2007
Abstract—Two series of 1,4-bis(2-amino-ethylamino)anthraquinone–amino acid conjugates (BACs), ametantrone (AT)–amino acid
conjugates (AACs) and mitoxantrone (MX)–amino acid conjugates (MACs), were designed and synthesized. The DNA binding of
BACs was evaluated by DNA thermal denaturation experiment. In the series, the methionine-substituted BACs had the weakest
DNA binding, while the lysine-substituted BACs had the highest Tm values. The abilities of BACs to inhibit the growth of
MCF-7, NCI-H460, SF-268, and PC-3 cell lines were determined. L-Met–MAC 16 and L-Lys–MAC 20 were the most potent growth
inhibitors. MAC 16 was more cytotoxic than MX, whereas the Tm of MAC 16 was much lower than that of MX. In contrast to
MAC 16, L-Lys–MAC 20 demonstrated higher Tm than MX. These data suggested that Met–BACs possessed a different pharma-
cological profile, in which the ability to stabilize DNA is not parallel to the ability to kill cancer cells, from that of AT and MX. The
primary mechanism of cytotoxicity for MAC 16 was most likely through TOP2 poisoning. Therefore, MAC 16 may provide a lead
for the development of novel generations of anthraquinone-type antitumor agents.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
the cardiac cells.5 This suggests the opportunities to devel-
op novel anthraquinones with reduced cardiotoxicity.
1,4-Disubstituted anthraquinones, such as ametantrone
(AT) and mitoxantrone (MX), demonstrate potent anti-
tumor activity and have been used widely in clinic since
1980s.1,2 In addition, MX has been the only drug ap-
proved by the FDA for the treatment of worsening relaps-
ing-remitting multiple sclerosis (MS), secondary
progressive MS, and progressive-relapsing MS since
2000.3,4 However, these clinical applications are limited
due to the accumulative and irreversible cardiotoxic-
ity.2–4 Recently, MX was found to induce a progressive in-
crease in mitochondrial mass in the cancer cells but not in
The planar tricyclic structure of anthraquinone is essen-
tial for intercalating into DNA base pairs. The two side
chains of AT and MX could be used to connect with a
variety of substituents, which may form additional inter-
actions with the double-stranded DNA (ds-DNA) or
DNA-topoisomerase II (TOP2) cleavable complex to in-
crease their binding affinities and selectivities. It is well
known that precise recognition of defined DNA se-
quences in biological systems is mediated by enzymes
and proteins having appropriate structure motifs.6 In
the literature, various anthraquinone–peptide conju-
gates7–11 and bis-amino acid-substituted anthraquinone
conjugates12,13 were reported to demonstrate remark-
able DNA binding and exhibit cytostatic or cytotoxic
activities. In addition, mono-amino acid-substituted
anthraquinone conjugates could inhibit the catalytic
activity of TOP1 and TOP2.14,15 Theoretical design of
anthraquinone–oligopeptide conjugates to selectively
Keywords: Anthraquinone; DNA binding; DNA affinity; Conjugate;
Cytotoxicity.
*
Corresponding author. Tel.: +886 2 2395 2310; fax: +886 2 2351
2086; e-mail: lwhsin@ntu.edu.tw
ꢀ
Current address: College of Pharmacy, Taipei Medical University,
250 Wu-Hsing St., Taipei 110, Taiwan, ROC.
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.10.012

L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
1007
target the double-stranded oligonucleotides has also
been published.16,17
H
N
X
O HN
O HN
R
R
With the aim to discover more specific chemotherapeu-
tic agents, a focused library of 1,4-bis(2-amino-ethylami-
no)anthraquinone–amino acid conjugates (BACs) was
designed and synthesized. The a-amino acids possessing
a variety of side chains, which could be used to create
specific interactions by one of the following mecha-
nisms, were chosen for construction of BACs.
X
N
H
1: X = H; R =
N-α-tert-Boc-L-methioninyl
2: X = H; R = L-methioninyl (2TFA)
3: X = H; R = -α-tert-Boc-D-methioninyl
4: X = H; R = D-methioninyl (2TFA)
N
5: X = H; R =
6: X = H; R =
N
O
-α-Fmoc-
O-tert-butyl-L-serinyl
(1) Metal ions, such as magnesium and zinc, are
involved in DNA synthesis and repair processes.
Attaching the amino acids containing potential
metal ion-chelating side chain (i.e., the methylsulfa-
nyl group of methionine) to the core structure may
be useful to interrupt DNA synthesis by stabiliza-
tion of the DNA–TOP2 cleavable complex.
(2) The hydroxyl group of tyrosine residue in TOP2
plays an essential role in TOP2-mediated DNA
strand cleavage.18,19 Furthermore, the side chains
of AT and MX also contain the hydroxyl groups.
Therefore, introducing hydroxyl group-containing
amino acids (i.e., serine and tyrosine) to anthraqui-
none may provide the opportunity to enhance the
cytotoxicity of this series of BACs.
-
tert-butyl-L-serinyl
7: X = H; R = L-serinyl (2TFA)
8: X = H; R = tert-butyl-L-tyrosinyl
9: X = H; R = L-tyrosinyl (2TFA)
10: X = H; R = -α,ε-di-tert-Boc-L-lysinyl
11: X = H; R = L-lysinyl (4TFA)
12: X = H; R =
13: X = H; R =
14: X = H; R = D-lysinyl (4TFA)
O
-
N
N
N
-α-Fmoc-
-ε-tert-Boc-D-lysinyl
N-ε-tert-Boc-D-lysinyl
15: X = OH; R = -α-tert-Boc-L-methioninyl
N
16: X = OH; R = L-methioninyl (2TFA)
17: X = OH; R = -α-tert-Boc-D-methioninyl
18: X = OH; R = D-methioninyl (2TFA)
19: X = OH; R = -α,ε-di-tert-Boc-L-lysinyl
20: X = OH; R = L-lysinyl (4TFA)
N
N
(3) The e-amino group of lysine is protonated under
physiological conditions, which would be expected
to produce strong electrostatic interactions with
the negatively charged phosphate groups on DNA
skeleton.12,13
Chart 2.
H
N
X
X
O HN
O HN
Boc
X
X
O
OH
OH
a, b
In this paper, we report the synthesis, DNA binding, and
in vitro cytotoxicity of two series of L- or D-amino acid-
conjugated BACs, AT–amino acid conjugates (AACs)
and MX–amino acid conjugates (MACs) (Chart 1).
O
Boc
N
H
21: X = H
23: X = H
22: X = OH
24: X = OH
2. Chemistry
c
2, 4, 11, 16, 18, 20
AACs 1–14 and MACs 15–20 (Chart 2) were synthe-
sized starting from leucoquinizarin (21) and 5,8-dihy-
droxy-leucoquinizarin (22), respectively. As compound
22 was not commercially available, 1,5-diamino-4,8-
dihydroxyanthraquinone was used to prepare 22 accord-
ing to the literature procedure.20 Treatment of 21 and 22
with an excess amount of mono-N-t-Boc-protected eth-
ylenediamine in methanol, followed by air oxidation,
afforded bis-N-t-Boc-protected 23 and 24, respectively,
as shown in Scheme 1.8,21–23 Trifluoroacetic acid
(TFA)-catalyzed deprotection of 23 and 24 provided
key intermediates 1,4-bis(2-amino-ethylamino)anthra-
c
NH2
NH2
1, 3, 10, 15, 17, 19
d
X
X
O HN
O HN
e
c
d
6, 13
5, 12
7, 14
d, e
25: X = H
26: X = OH
c
9
8
Scheme 1. Reagents and conditions: (a) BocNHCH2CH2NH2,
CH3OH, 50 ꢁC; (b) air, CH3OH, 45 ꢁC; (c) TFA, CH2Cl2, rt; (d)
corresponding protected amino acids, DCC, HOBt, CH2Cl2, rt; (e)
piperidine, CH2Cl2, rt.
H
N
H
N
R
X
X
O HN
O HN
OH
OH
X
X
O HN
O HN
NH2
NH2
O
O
N
H
N
H
R
Ametantrone (AT): X = H
Mitoxantrone (MX): X = OH
Ametantrone-Amino acid Conjugate (AAC): X = H
Mitoxantrone-Amino acid Conjugate (MAC): X = OH
Chart 1.

1008
L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
quinone (25) and 1,4-bis(2-amino-ethylamino)-5,8-dihy-
droxy-anthraquinone (26) in quantitative yields.
In previous structure–activity relationship (SAR) stud-
ies, 5,8-dihydroxy-anthraquinones were more potent
than the non-hydroxylated analogs in antitumor activ-
ity.20–22 In this study, the 5,8-dihydroxy-substituted
MACs also had larger DTm values than the correspond-
ing AACs. It is noteworthy that when the solution of L-
Lys–MAC 20 in normal concentration was added to the
ct-DNA solution, deep blue aggregates were formed in-
stantly which prohibited the determination of Tm.
Therefore, the final concentration of 20 in the DNA–
drug solutions for thermal denaturation study was re-
duced to 5 lM to estimate the DNA binding affinity
of 20. Even under this diluted condition, L-Lys–MAC
20 showed a very high DTm value of 27.7 ꢁC. MX dis-
played a similar but less aggregation phenomenon than
MAC 20. At a concentration of 10 lM, MX raised the
melting temperature of ct-DNA by 25.3 ꢁC. The L-
Met- and D-Met-substituted MACs 16 and 18 also had
higher Tm than the corresponding AACs 2 and 4,
although the DTm values for Met-substituted MACs
16 and 18 were much less than that of L-Lys–MAC 20.
Both of the AACs and MACs were prepared via a
general reaction sequence (Scheme 1). Coupling of
compounds 25 and 26 with appropriate N- and side
chain-protected amino acids using N,N0-dicyclohexyl-
carbodiimide (DCC) in the presence of N-hydrox-
ybenzotriazole (HOBt), N-hydroxyphthalimide, or
N-hydroxysuccinimide produced protected BACs in
moderate yields. In consideration of both the shorter
reaction time and easier purification process, HOBt pro-
vided the most satisfactory results. The N-t-Boc and O-
t-Bu protective groups of BACs were removed using 20–
50% TFA in CH2Cl2, while the N-9-fluorenylmethoxy-
carbonyl (Fmoc) protective group was cleaved by 10%
piperidine in CH2Cl2.
3. Pharmacology
Effective intercalating agents, such as anthraquinone
derivatives, are well known for their ability to raise the
melting temperature (Tm) of ds-DNA.24,25 Although the
increases in Tm may not be directly correlated with
DNA binding affinity,26 higher DTm values indicate tigh-
ter interactions between the ligand and ds-DNA.24 There-
fore, the DNA binding of BACs was evaluated by
determination of their ability to alter the thermal denatur-
ation profile of calf thymus DNA (ct-DNA).8,24–28 In gen-
eral, the BACs with ClogP values larger than 3.5 were not
soluble enough in BPE buffer. Therefore, the Tm for those
highly lipophilic BACs could not be determined. The DTm
values of the BACs with sufficient solubility were mea-
sured and the results of melting experiments are shown
in Table 1.
Based on these results, the amino acid residues in BACs
did play crucial role in the modulation of drug binding
to ct-DNA. Smaller Tm shifts were observed for the
non-ionic side chain-containing BACs. The larger DTm
values of Ser– and Tyr–AACs than of Met–AACs sug-
gest that the terminal hydroxyl groups are likely to be
favored for binding just as in the cases of AT and
MX. In both the AAC and MAC series, the D- and L-
BACs showed no significant differences in their DNA
binding properties, which point to a poor chiral discrim-
ination in the BAC–DNA complexes.
The results of in vitro cytotoxicity screening29 of BACs
1–20 against MCF-7 (breast cancer), NCI-H460 (non-
small cell lung carcinoma), SF-268 (glioblastoma), and
PC-3 (prostate cancer) cell lines are reported in Table
2. In the AAC series, L-Met–AAC 2 had the most potent
activity against PC-3 cell line (GI50 = 1.50 lM), while
the corresponding N-protected L-Met–AAC 1 was inac-
tive in this assay. N-Fmoc-O-t-Bu-protected L-Ser–AAC
5 was the most potent growth inhibitor for NCI-H460 in
the AACs, while inactive for the other three cancer cell
lines. Selective N-deprotection of AAC 5 provided O-t-
Bu-protected L-Ser–AAC 6, which showed moderate
cytotoxicity against all cancer cell lines. Removing both
protective groups of AAC 5 produced L-Ser–AAC 7,
which possessed the lowest ClogP value (ꢁ1.74) among
the BACs and was totally inactive in this assay. O-t-Bu-
protected L-Tyr–AAC 8 and unprotected L-Tyr–AAC 9
were 2- to 3-fold more potent than AT in growth inhibi-
tion against PC-3 cell line, while did not show significant
activity against MCF-7, NCI-H460, and SF-268 cell lines.
L-Lys- and D-Lys-substituted AACs 11 and 14 had lar-
ger DTm values than AT, which is a potent DNA inter-
calator. When the distal amino groups (Ne) were
protected (i.e., 13), the DTm was significantly reduced.
Therefore, the positively charged side chains of AACs
11 and 14 were important and the multiple ionic interac-
tions with the negatively charged phosphate backbone
of ds-DNA might be involved. In contrast to Lys-substi-
tuted AACs, the Met-substituted AACs2 and 4 showed
very weak DNA binding. Interestingly, the L-Met–AAC
2 and D-Met–AAC 4 possessed similar DNA stabiliza-
tion ability, which means there was no significant chiral-
ity discrimination.
The DTm values for the hydroxyl-containing L-Ser–AAC 7
andL-Tyr–AAC9weremoderateandalmostthesame. Ini-
tially, we thought the additional hydrogen bonding to the
DNA produced by the hydroxyl-containing side chains
was responsible for the Tm change. However, the hydro-
xyl-protected L-Ser(t-Bu)–AAC 6 demonstrated a higher
Tm than that of unprotected L-Ser–AAC 7 (6.5 and
5.4 ꢁC, respectively) and therefore, the side chain hydroxyl
groups might not be important for DNA binding in AAC
series. The magnitude of DTm values in the AAC series was
in the following order: L-Lys, D-Lys >AT ꢀ Ser(t-Bu) >
Ser > Tyr > D-Lys(Boc) > D-Met > L-Met.
L-Lys–AAC 11 did not show significant cytotoxicity
against the four cancer cell lines, although AAC 11
was an extraordinarily strong DNA binder. Interest-
ingly, the Na, Ne-protected L-Lys–AAC 10 showed sub-
stantial growth inhibition activity against MCF-7, SF-
268, and PC-3 cell lines. In PC-3 cell line, AAC 10
was four times more potent than AT (GI50 = 1.83,
8.52 lM, respectively). The lipophilicity and a-amino

L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
Table 1. ClogP and DTm values for BACs
1009
No.
N-a PG
Amino acid
Side chain PG
ClogPa
DTm (ꢁC)b,c
Molecular formula
C38H54N6O8S2
1
Boc
—
Met
Met
D-Met
D-Met
Ser
—
—
4.22
0.57
ND
1.4 0.3
ND
2
C28H38N6O4S2Æ2C2HF3O2
C38H54N6O8S2
3
Boc
—
—
—
4.22
0.57
4
1.6 0.3
ND
C28H38N6O4S2Æ2C2HF3O2
C62H66N6O10
C32H46N6O6
5
Fmoc
—
—
t-Bu
t-Bu
—
>6.00
1.98
6
Ser
Ser
6.5 0.5
5.4 0.1
ND
7
ꢁ1.74
5.42
1.77
C24H30N6O6Æ2C2HF3O2
C44H54N6O6
8
—
—
Tyr
Tyr
t-Bu
—
9
5.2 0.7
ND
>25
C36H38N6O6Æ2C2HF3O2
C50H76N8O12
10
11
12
13
14
15
16
17
18
19
20
AT
MX
Boc
—
Lys
Lys
Boc
—
>6.00
ꢁ0.33
>6.00
3.32
C30H44N8O4Æ4C2HF3O2
C70H80N8O12
C40H60N8O8
Fmoc
—
—
D-Lys
D-Lys
D-Lys
Met
Met
D-Met
D-Met
Lys
Boc
Boc
—
ND
3.1 0.1
>25
ND
ꢁ0.33
3.86
0.22
C30H44N8O4Æ4C2HF3O2
C38H54N6O10S2
Boc
—
—
—
4.7 0.3
ND
C28H38N6O6S2Æ2 C2HF3O2
C38H54N6O10S2
Boc
—
—
—
3.86
0.22
5.0 0.1
ND
27.7 1.0d,e
21.4 0.1
25.3 1.0f
C28H38N6O6S2Æ2C2HF3O2
C50H76N8O14
Boc
—
Boc
—
>6.00
ꢁ0.68
0.59
Lys
—
C30H44N8O6Æ4C2HF3O2
—
—
—
—
—
—
—
0.24
a The predicted ClogP values were calculated using ClogP module in ChemDraw Ultraꢂ version 6.0.1, CambridgeSoft.Com.
b BAC concentrations in the test solutions were 20 lM unless otherwise specified.
c ND, not determined.
d The concentration of compound 20 in the test solutions was 5 lM.
e Biphasic melting curves were recorded.
f The concentration of MX in the test solutions was 10 lM.
Table 2. Growth inhibition of BACs against human cancer cell lines
a
No.
GI50 (lM)
MCF-7b
NCI-H460b
SF-268b
PC-3c
1
ND
ND
ND
>30
1.50
>30
>30
>10
3.67
>10
2.23
3.97
1.83
>10
>10
>30
>30
ND
0.28
ND
>10
>10
0.83
8.52
0.78
2
60%
95%
71%
59%
35%
95%
93%
66%
49%
83%
68%
103%
68%
92%
0%
66%
101%
52%
15%
48%
96%
54%
56%
58%
79%
88%
98%
100%
96%
1%
53%
107%
58%
70%
42%
96%
82%
53%
45%
79%
74%
97%
82%
98%
1%
3
4
5
2.71
6.08
6
8.18
3.85
6.46
2.13
7
8
9
10
11
12
13
14
15
16
17
18
19
20
AT
MX
1.64
0.38
0.35
92%
92%
79%
55%
34%
19%
104%
94%
99%
31%
22%
8%
106%
100%
65%
42%
6%
1.38
2.55
1.29
2.03
2.45
0.97
4.97
3.93
2%
a ND, not determined.
b Initial screening of cytotoxicity for BACs in a concentration of 4 lg/mL was conducted to determine the percentage of growth for MCF-7, NCI-
H460, and SF-268 cancer cell lines. For those BACs that demonstrated growth inhibition more than 50% at the screening concentration (4 lg/mL),
the GI50 values were determined.
c For those BACs demonstrated growth inhibition less than 50% against PC-3 cell line at the range of concentrations tested, the highest concen-
trations of the ligands used in the assay are shown.

1010
L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
groups seemed to be the most critical factors for the
cytotoxicity of AACs for PC-3 cell line. With the excep-
tion of AAC 10, the active AACs (2, 6, 8, and 9) pos-
sessed free a-amino groups and lipophilic terminals
(side chains), while the AACs with either protected a-
amino groups (i.e., 1 and 5) or negative ClogP values
(i.e., 7 and 11) were inactive.
macophore which gives promise to eliminate the severe
problem of cardiotoxicity for MX.
The DNA binding of hydroxyl-containing AACs 7 and
9 was substantially stronger than that of Met–BACs but
weaker than that of AT. Under physiological condi-
tions, the positively charged Lys–BACs were the stron-
gest DNA binders in the series. However, this ability
was not completely reflected in the antitumor activity
of Lys–BACs. In general, more polar amino acid-con-
taining BACs displayed higher Tm. Therefore, the DTm
values of the BACs were in the order of
Lys ꢀ Ser P Tyr > Met. In contrast to DNA binding,
the more lipophilic BACs were more potent in killing
cancer cells in both the AAC and MAC series. More-
over, all the D-BACs were inactive in inhibition of can-
cer cell growth irrespective of their ability to stabilize ct-
DNA. Considering that all inactive Na-unprotected
BACs were either in D-configuration or with negative
ClogP values, these inactive BACs might have poorer
accessibility (e.g., penetration, uptake, etc.) to tumor
cells.
The unprotected L-Met–MAC 16 and L-Lys–MAC 20
were much more potent than their corresponding AACs
2 and 11. The L-Met-substituted MAC 16 was the most
potent compound among the BACs in this study and
demonstrated much better growth-inhibition activity
than AT and MX (3–30- and 2–3-fold, respectively) in
all tumor cell lines.
In contrast to Lys–AACs 10 and 11, Na, Ne-protected
L-Lys-MAC 19 did not demonstrate significant cytotox-
icity against cancer cell lines, whereas unprotected L-
Lys–MAC 20 showed much stronger growth inhibition
activity than AT and as potent as MX against NCI-
H460, SF-268, and PC-3 cell lines. Previously, it has
been reported that Lys–MAC 20 showed potent DNA
binding and was effective in HL60 tumor cell killing.12,13
Both AT and MX have been well established as TOP2-
targeting anticancer drugs and induce TOP2-mediated
DNA breakage. We therefore evaluated the abilities of
MAC 16 in generation of DNA damage and cytotoxicity
by the comet assay30 and MTT assay.31 The potential
contribution of TOP2 in mediating the effect of MAC
16 was further investigated with the mutant HL-60/
MX-2 cell line that was known to have reduced levels
of both TOP2 isozymes and consequently leading to
cross-resistance to all known TOP2 poisons, but not
other anticancer drugs.30,32 In Figure 1, the extents of
DNA breakage of chromosome DNA in HL-60/MX-2
cells treated with etoposide (VP-16), AT, MX, and
MAC 16 were greatly reduced compared to that in
parental HL-60 cells. In marked contrast, the TOP1-tar-
geting drug, camptothecin (CPT), induced similar extent
of DNA damage in both cell lines. Similar conclusion
was obtained using PFGE analysis to examine the abil-
ity of MAC in TOP2-mediated excision of the loop-sized
DNA fragments31 (data not shown). These results
The cytotoxicity of D-amino acid-containing BACs was
also studied, as the D-enantiomers not only possessed
comparable DNA binding properties to the L-enantio-
mers, but may also be less susceptible to enzymatic
hydrolysis. However, all D-amino acid-conjugated
AACs and MACs were inactive in the growth inhibition
against various human cancer cell lines. In contrast to
the DNA thermal denaturation study, in which the chi-
rality did not demonstrate any significant influence on
the Tm, the chirality played a critical role in the cytotox-
icity of BACs. A clear discrimination in the results
between the L- and D-enantiomers was observed, in
which the L-isomers were always much more effective
than the corresponding D-isomers.
4. Discussion
In this study, three strategies were used to design BACs
with higher antitumor activity. The sulfur-containing
Met–MAC 16 had higher cytotoxicity than MX, while
16 was a much weaker DNA binder than MX. These re-
sults suggested that the conjugation of methionine to the
anthraquinone moiety altered the pharmacological pro-
files of AT and MX. Classical antitumor anthraquinone
derivatives, such as AT and MX, can bind to DNA
tightly by intercalation, whereas it is not clear whether
or how this binding results in their antiproliferative ef-
fects. In addition, the notion that the ability to increase
the melting temperature of DNA did not correlate with
the antitumor activity was also reported.10,13 Therefore,
the potent antiproliferative effects of Met–MAC 16
against tumor cells may be more a function of their abil-
ity to interfere with other cellular targets (e.g., TOP2)
than their ability to intercalate with DNA. The quite dif-
ferent pharmacological profile of the Met–MAC 16, in
which different mechanisms of cytotoxicity might be in-
volved, pointed to a novel type of antineoplastic phar-
100
80
60
40
20
0
HL-60
MX-2
Figure 1. Drug-induced strand breaks in chromosomal DNA in HL-60
and TOP2-deficient HL-60/MX-2 cell lines.33 DNA damage (%): the
percentage of cells showing the comet image, which corresponded to
damaged chromosomal DNA.

L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
1011
Table 3. Growth inhibition against HL-60 and TOP2-deficient HL-60/
MX-2 cell linesa,b
mass spectrometer. Elemental analyses were performed
with a Heraeus varioIII-NCSH instrument and were
within 0.4% for the elements indicated. Thin-layer
chromatography (TLC) was performed on Merck (art.
5715) silica gel plates and visualized under UV light
(254 nm), upon treatment with iodine vapor, or upon
heating after treatment with 5% phosphomolybdic acid
in ethanol. Flash column chromatography was per-
formed with Merck (art. 9385) 40–63 lm silica gel 60.
Anhydrous tetrahydrofuran was distilled from sodium-
benzophenone prior to use. No attempt was made to
optimize yields.
Compoundc
GI50 (nM)
HL-60/MX-2
GI50 ratio
HL-60
CPT
1.7
1.5
5980
430
26
0.9
43
15
79
38
VP-16
AT
MX
140
29
0.33
0.32
MAC 16
12
a Determined by the MTT cytotoxic assay.31
b HL-60: leukemia cell line.
c CPT, camptothecin; VP-16, etoposide.
6.2. General synthetic methods
strongly suggest that TOP2s are primarily responsible
for DNA breakage induced by MAC 16 treatment.
Consistently, HL-60/MX-2 cells were also shown to be
cross-resistant to VP-16, AT, MX, and MAC 16 (GI50
ratio = 15- to 79-fold), but not to CPT (GI50
ratio = 0.9-fold) as shown in Table 3. Taken together,
we conclude that the cytotoxic action of these BACs
was most likely through poisoning of the TOP2s.
6.2.1. Synthetic method A. (S,S)-1,4-Bis[2-[2,6-bis(tert-
butoxycarbonylamino)hexanoylamino]ethylamino]-9,10-
anthracenedione (10). N-a,e-Di-t-Boc-L-lysine dicyclo-
hexylammonium salt (1.35 g, 2.55 mmol), dicyclohexyl-
carbodiimide (DCC, 551 mg, 2.67 mmol), and HOBt
(375 mg, 2.78 mmol) were dissolved in CH2Cl2 (10 mL)
and stirred for 1 h at room temperature. The white pre-
cipitate (N,N0-dicyclohexylurea) was removed by filtra-
tion, and then the filtrate was added to a stirred
solution of 25 (376 mg, 1.16 mmol) and triethylamine
(1.30 mL) in CH2Cl2 (5 mL). After being stirred for
17 h at room temperature, the solvent was removed un-
der reduced pressure. The residue was purified by flash
5. Conclusion
L-Met–MAC 16 was the most potent inhibitor of tumor
cell growth in the BACs, whereas 16 only showed limited
DNA binding. The cytotoxicity of MAC 16 was higher
than that of MX, while the Tm of MAC 16 was much lower
than that of MX. These results suggested a quite different
pharmacological profile, which may provide opportuni-
ties to develop novel anthraquinone derivatives with in-
creased potency and lower toxicity. In contrast to MAC
16, L-Lys–MAC 20 demonstrated higher Tm than MX
and a similar cytotoxicity as MX. These data supported
that the ability to stabilize DNA is not parallel to the abil-
ity to kill cancer cells. Non-selective DNA intercalation of
classical anthraquinone antitumor agents has resulted in
various adverse reactions in clinic. Based on the SAR
established in this study, the DNA binding and the cyto-
toxicity could be separated and thus, MAC 16 may pro-
vide a lead for the development of new generations of
anthraquinone-type antitumor agents. Additional data
from further pharmacological studies suggest that the
cytotoxicity of MAC 16 was most likely through poison-
ing of the TOP2s.
column
chromatography
(silica
gel,
MeOH/
CH2Cl2 = 1:10) to afford 10 (723 mg, 64%) as a deep
blue solid. Mp 196–198 ꢁC; 1H NMR (200 MHz,
DMSO-d6) d 1.05–1.42 (m, 44H), 1.45–1.70 (m, 4H),
2.70–2.92 (m, 4H), 3.12–3.41 (m, 4H), 3.45–3.65 (m,
4H), 3.73–3.96 (m, 2H), 6.64–6.86 (m, 4H), 7.55–7.65
(m, 2H), 7.72–7.84 (m, 2H), 8.08 (br s, 2H), 8.17–8.28
(m, 2H), 10.81 (br s, 2H); 13C NMR (100 MHz,
DMSO-d6) d 22.8, 28.1, 28.2, 29.2, 31.7, 38.9, 39.5,
41.2, 54.4, 77.2, 77.9, 108.7, 124.4, 125.7, 132.3, 133.8,
145.9, 155.3, 155.5, 172.7, 180.8; FABHRMS Calcd
for C50H76N8O12 [M]+ 980.5583. Found: 980.5547.
6.2.2. Synthetic method B. (S,S)-1,4-Bis[2-(2,6-diamino-
hexanoyl-carbamoyl)ethylamino]-9,10-anthracenedione tri-
fluoroacetate (11). Compound 10 (255 mg, 0.260 mmol)
was treated with a mixture of CH2Cl2 (5 mL) and triflu-
oroacetic acid (5 mL) and stirred at room temperature
for 30 min. The resulted solution was evaporated to dry-
ness under high vacuum to give 11 (225 mg, 84%) as a
1
blue oil. H NMR (200 MHz, DMSO-d6) d 1.25–1.42
6. Experimental
(m, 4H), 1.42–1.61 (m, 4H), 1.61–1.82 (m, 4H), 2.65–
2.86 (m, 4H), 3.44–3.68 (m, 8H), 3.68–3.85 (m, 2H),
7.59 (s, 2H), 7.80–7.84 (m, 8H), 8.22–8.24 (m, 8H),
8.82 (br s, 2H), 10.81 (br s, 2H); 13C NMR (100 MHz,
DMSO-d6) d 21.3, 26.6, 30.5, 38.4, 38.8, 41.2, 52.2,
109.0, 124.4, 125.8, 132.6, 133.8, 145.8, 169.0, 181.1;
FABHRMS Calcd for C30H45N8O4 [M+H]+ 581.3564.
Found: 581.3574.
6.1. General procedures
Melting points were determined on a MEL-TEMP II
apparatus by Laboratory Devices and are uncorrected.
NMR spectra were recorded on Bruker DPX-200 and
AMX-400 FT–NMR spectrometers. Chemical shifts
are expressed in parts per million (ppm) on the d scale
relative to a tetramethylsilane (TMS) internal standard.
Fast atom bombardment mass (FABMS) spectra were
recorded on a JEOL SX-102A GC/LC/Mass spectrome-
ter. High-resolution FABMS measurements were ob-
tained using a Finnigan/Thermo Quest MAT 95XL
6.2.3. Synthetic method C. (S,S)-l,4-Bis[2-(2-amino-3-
tert-butoxy-propionylamino)ethylamino]-9,10-anthracenedi-
one (6). Compound 5 (120 mg, 0.114 mmol) was treated
with a solution of 10% piperidine in CH2Cl2 (5 mL) for
40 min at room temperature, and then the solvent was

1012
L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
removed under reduced pressure. The crude residue was
chromatographed (silica gel, MeOH/CH2Cl2 = 1:5) to
afford 6 (50.4 mg, 73%) as a blue solid. Mp 159 ꢁC; H
NMR (200 MHz, CDCl3) d 1.17 (s, 18H), 1.66–1.85
(m, 4H), 3.42–3.62 (m, 10H), 3.62–3.76 (m, 4H), 7.30
(s, 2H), 7.62–7.67 (m, 2H), 8.01 (s, 2H), 8.19–8.24 (m,
2H), 10.72 (s, 2H); 13C NMR (100 MHz, CDCl3) d
27.5, 39.0, 42.0, 55.6, 63.9, 73.4, 110.4, 123.6, 126.1,
132.2, 134.4, 146.1, 173.9, 182.8; FABHRMS Calcd
for C32H47N6O6 [M+H]+ 611.3557. Found: 611.3555.
A as a blue solid in a yield of 54%. Mp 156–158 ꢁC;
1H NMR (200 MHz, DMSO-d6) d 1.05 (s, 18H), 3.31–
3.32 (m, 4H), 3.38–3.42 (m, 4H), 3.48–3.52 (m, 4H),
4.05–4.10 (m, 4H), 4.20–4.24 (m, 2H), 4.24–4.30 (m,
2H), 7.28–7.56 (m, 16H), 7.68–7.72 (m, 6H), 7.84–7.88
(m, 2H), 8.17–8.19 (m, 2H), 10.75–10.79 (m, 2H); FAB-
HRMS Calcd for C62H66N6O10 [M]+ 1054.4840. Found:
1054.4838.
1
6.8. (S,S)-l,4-Bis[2-(2-amino-3-hydroxy-propionylami-
no)ethylamino]-9,10-anthracenedione trifluoroacetate (7)
6.3. (S,S)-l,4-Bis[2-(2-tert-butoxycarbonylamino-4-methyl-
sulfanyl-butyrylamino)ethylamino]-9,10-anthracenedione (1)
Compound 7 was synthesized using compound 6
according to the synthetic method B as a blue oil in a
yield of 100%. 1H NMR (200 MHz, DMSO-d6) d
1.58–1.63 (m, 4H), 2.83–3.05 (m, 6H), 3.49–3.62 (m,
4H), 3.75–3.79 (m, 2H), 7.57 (s, 2H), 7.81 (dd, J = 5.9,
3.3 Hz, 2H), 8.12 (br s, 6H), 8.24 (dd, J = 5.9, 3.3 Hz,
2H), 8.71 (t, J = 6.0 Hz, 2H), 10.80 (t, J= 6.0 Hz, 2H);
FABHRMS Calcd for C24H31N6O6 [M+H]+ 499.2305.
Found: 499.2302.
Compound 1 was synthesized using N-a-t-Boc-L-methi-
onine and 25 according to the synthetic method A as a
blue solid in a yield of 64%. Mp 236–237 ꢁC; 1H
NMR (200 MHz, DMSO-d6) d 1.36 (s, 18H), 1.61–1.87
(m, 4H), 1.96 (s, 6H), 2.30–2.45 (m, 4H), 3.25–3.30 (m,
4H), 3.40–3.60 (m, 4H), 3.81–4.00 (m, 2H), 6.96 (d,
J = 8.0 Hz, 2H), 7.59 (s, 2H), 7.77–7.79 (m, 2H), 8.01–
8.12 (m, 2H), 8.22–8.26 (m, 2H), 10.79–10.82 (m, 2H);
FABHRMS Calcd for C38H54N6O8S2 [M]+ 786.3445.
Found: 786.3439. Anal. Calcd for C38H54N6O8S2: C,
57.99;H,6.92;N,10.68. Found:C, 57.94;H,6.96;N,10.44.
6.9. (S,S)-l,4-Bis[2-[2-amino-3-(4-tert-butoxyphenyl)-pro-
pionylamino]ethylamino]-9,10-anthracenedione (8)
Compound 8 was synthesized using N-a-Fmoc-O-t-bu-
tyl-L-tyrosine and 25 according to the synthetic method
A followed by the general procedure C, as a blue solid in
a yield of 48%. Mp 150–152 ꢁC; H NMR (200 MHz,
CDCl3) d 1.32 (s, 18H), 2.67 (dd, J = 13.8, 9.5 Hz,
2H), 3.24 (dd, J = 13.7, 3.9 Hz, 2H), 3.49 (s, 4H),
3.55–3.65 (m, 10H), 6.92 (d, J = 8.4 Hz, 4H), 7.11 (d,
J = 8.4 Hz, 4H), 7.35 (s, 2H), 7.65–7.70 (m, 4H), 8.26–
8.31 (m, 2H), 10.78 (br s, 2H); FABHRMS Calcd for
C44H55N6O6 [M+H]+ 763.4183. Found: 763.4172.
6.4. (S,S)-l,4-Bis[2-(2-amino-4-methylsulfanyl-butyrylami-
no)ethylamino]-9,10-anthracenedione trifluoroacetate (2)
1
Compound 2 was synthesized using compound 1
according to the synthetic method B as a blue solid in
1
a yield of 98%. Mp 178–180 ꢁC; H NMR (200 MHz,
DMSO-d6) d 1.95 (s, 6H), 1.99–2.03 (m, 4H), 2.41–
2.45 (m, 4H), 3.32–3.36 (m, 4H), 3.54–3.72 (m, 4H),
3.74–3.88 (m, 2H), 7.57 (s, 2H), 7.79–7.83 (m, 2H),
8.20–8.23 (m, 2H), 8.25–8.27 (m, 6H), 8.73–8.82 (m,
2H), 10.82 (br s, 2H); ESIMS m/z 587 (MH+). Anal.
Calcd for C28H38N6O4S22CF3CO2H1.5H2O: C, 45.66;
H, 5.15; N, 9.98. Found: C, 45.55; H, 5.09; N, 9.84.
6.10. (S,S)-l,4-Bis[2-[2-amino-3-(4-hydroxyphenyl)-propi-
onylamino]ethylamino]-9,10-anthracenedione trifluoro-
acetate (9)
6.5. (R,R)-l,4-Bis[2-(2-tert-butoxycarbonylamino-4-meth-
ylsulfanyl-butyrylamino)ethylamino]-9,10-anthracenedi-
one (3)
Compound 9 was synthesized using compound 8 accord-
ing to the synthetic method B as a blue solid in a yield of
99%. Mp 169–171 ꢁC; 1H NMR (200 MHz, DMSO-d6) d
2.80–3.03 (m, 4H), 3.25–3.55 (m, 8H), 3.85–3.96 (m, 4H),
6.68 (d, J = 8.4 Hz, 4H), 7.02 (d, J = 8.4 Hz, 4H), 7.55 (s,
2H), 7.77–7.83 (m, 2H), 8.13–8.28 (m, 8H), 8.70–8.77 (m,
2H), 10.77 (br s, 2H); FABHRMS Calcd for C36H39N6O6
[M+H]+ 651.2931. Found: 651.2917.
Compound 3 was synthesized using N-a-t-Boc-D-methio-
nine and 25 according to the synthetic method A as a blue
solid in a yield of 63%. Mp 233–236 ꢁC; FABHRMS
Calcd for C38H54N6O8S2 [M]+ 786.3445. Found: 786.3429.
6.6. (R,R)-l,4-Bis[2-(2-amino-4-methylsulfanyl-butyryla-
mino)ethylamino]-9,10-anthracenedione trifluoroacetate (4)
6.11. (R,R)-l,4-Bis[2-[6-tert-butoxycarbonylamino-2-(9H-
fluoren-9-ylmethoxycarbonylamino)-hexanoylamino]eth-
ylamino]-9,10-anthracenedione (12)
Compound 4 was synthesized using compound 3 accord-
ing to the synthetic method B as a blue solid in a yield of
100%. Mp 68–71 ꢁC (free base); FABHRMS Calcd for
C28H39N6O4S2 [M+H]+ 587.2474. Found: 587.2484.
Compound 12 was synthesized using N-a-Fmoc-N-e-t-
Boc-D-lysine and 25 according to the synthetic method
A as a blue solid in a yield of 41%. Mp 155–156 ꢁC;
1H NMR (200 MHz, DMSO-d6) d 1.33 (s, 18H), 1.50–
1.70 (m, 4H), 2.79–2.90 (m, 4H), 3.08–3.19 (m, 8H),
3.28–3.30 (m, 4H), 3.51–3.53 (m, 4H), 3.81–3.98 (m,
2H), 4.04–4.20 (m, 6H), 6.72–6.75 (m, 2H), 7.18–7.56
(m, 16H), 7.63–7.78 (m, 4H), 7.84–7.88 (m, 4H), 8.10–
8.22 (m, 2H), 10.65–10.80 (m, 2H); FABHRMS Calcd
for C70H80N8O12 [M]+ 1224.5896. Found: 1224.5895.
6.7. (S,S)-l,4-Bis[2-[3-tert-butoxy-2-(9H-fluoren-9-yl-
methoxycarbonylamino)-propionylamino]ethylamino]-9,
10-anthracenedione (5)
Compound 5 was synthesized using N-a-Fmoc-O-t-bu-
tyl-L-Serine and 25 according to the synthetic method

L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
1013
6.12. (R,R)-l,4-Bis[2-(2-amino-6-tert-butoxycarbonylami-
no-hexanoylamino)-ethylamino]-9,10-anthracenedione (13)
A as a blue solid in a yield of 69%. Mp 214–215 ꢁC;
FABHRMS Calcd for C38H54N6O10S2 [M]+ 818.3343.
Found: 818.3340.
Compound 13 was synthesized using compound 12
according to the synthetic method C as a blue solid in
a yield of 78%. Mp 150 ꢁC; 1H NMR (200 MHz,
DMSO-d6) d 1.10–1.40 (m, 26H), 1.45–1.60 (m, 4H),
2.70–2.90 (m, 4H), 3.01–3.16 (m, 4H), 3.32–3.33 (m,
4H), 3.47–3.63 (m, 6H), 6.64–6.78 (m, 2H), 7.62 (s,
2H), 7.76–7.80 (m, 2H), 8.16–8.25 (m, 4H), 10.71–
10.85 (m, 2H); FABHRMS Calcd for C40H61N8O8
[M+H]+ 781.4612. Found: 781.4601.
6.17. (R,R)-l,4-Bis[2-(2-amino-4-methylsulfanyl-buty-
rylamino)ethylamino]-5,8-dihydroxy-9,10-anthracenedi-
one trifluoroacetate (18)
Compound 18 was synthesized using compound 17
according to the synthetic method B as a blue solid in
a yield of 100%. Mp 95 ꢁC (free base); FABHRMS
Calcd for C28H39N6O6S2 [M+H]+ 619.2372. Found:
619.2371.
6.13. (R,R)-l,4-Bis[2-(2,6-diamino-hexanoylamino)ethyl-
amino]-9,10-anthracenedione trifluoroacetate (14)
6.18. (S,S)-1,4-Bis[2-[2,6-bis(tert-butoxycarbonylamino)-
hexanoylamino]ethylamino]-5,8-dihydroxy-9,10-anthra-
cenedione (19)
Compound 14 was synthesized using compound 13
according to the synthetic method B as a blue oil in a
1
yield of 100%. H NMR (200 Hz, DMSO-d6) d 1.25–
Compound 19 was synthesized using N-a,e-di-t-Boc-L-
lysine dicyclohexylammonium salt and 26 according to
the synthetic method A as a blue solid in a yield of
1.41 (m, 4H), 1.41–1.61 (m, 4H), 1.61–1.80 (m, 4H),
2.67–2.87 (m, 4H), 3.44–3.50 (m,4H), 3.50–3.67 (m,
4H), 3.74 (s, 2H), 7.58 (s, 2H), 7.74 (br s, 6H), 7.82
(dd, J = 5.9, 3.3 Hz, 2H), 8.15 (br s, 6H), 8.24 (dd,
J = 5.8, 3.2 Hz, 2H), 8.77 (t, J = 5.0 Hz, 2H), 10.80 (t,
J = 4.0 Hz, 2H); FABHRMS Calcd for C30H45N8O4
[M+H]+ 581.3564. Found: 581.3552.
1
21%. Mp 174–176 ꢁC; H NMR (200 MHz, DMSO-d6)
d 1.05–1.40 (m, 44H), 1.42–1.60 (m, 4H), 2.81 (d,
J = 2.3 Hz, 4H), 3.18–3.41 (m, 4H), 3.45–3.63 (m, 4H),
3.82 (d, J = 2.3 Hz, 2H), 6.62–6.80 (m, 4H), 7.17 (s,
2H), 7.63 (s, 2H), 8.28 (br s, 2H), 10.55 (br s, 2H),
13.51 (br s, 2H); FABHRMS Calcd for C50H76N8O14
[M]+ 1012.5481. Found: 1012.5523.
6.14. (S,S)-l,4-Bis[2-(2-tert-butoxycarbonylamino-4-methyl-
sulfanyl-butyrylamino)ethylamino]-5,8-dihydroxy-9,10-
anthracenedione (15)
6.19. (S,S)-1,4-Bis[2-(2,6-diamino-hexanoylamino)ethyl-
amino]-5, 8-dihydroxy-9,10-anthracenedione trifluoroace-
tate (20)
Compound 15 was synthesized using N-a-t-Boc-L-methio-
nine and 26 according to the synthetic method A aa a blue
solid in a yield of 54%. Mp 219–220 ꢁC; 1H NMR
(200 MHz, DMSO-d6) d 1.36 (s, 18H), 1.76–1.80 (m, 4H),
1.95 (s, 6H), 2.38–2.41 (m, 4H), 2.49–2.51 (m, 4H), 3.58–
3.62 (m, 4H), 3.85–3.97 (m, 2H), 6.96 (d, J = 7.4 Hz, 2H),
7.15 (s, 2H), 7.65 (s, 2H), 8.08–8.16 (m, 2H), 10.52–10.68
(m, 2H), 13.52 (s, 2H); FABHRMS Calcd for
C38H54N6O10S2 [M]+ 818.3343. Found: 818.3365.
Compound 20 was synthesized using compound 19
according to the synthetic method B as a blue oil in a
1
yield of 99%. H NMR (200 MHz, DMSO-d6) d 1.16–
1.40 (m, 4H), 1.40–1.61 (m, 4H), 1.61–1.82 (m, 4H),
2.58–2.82 (m, 4H), 3.24–3.54 (m, 4H), 3.54–3.82 (m,
6H), 7.18 (s, 2H), 7.64 (s, 2H), 7.85 (br s, 6H), 8.24
(br s, 6H), 8.86 (br s, 2H), 10.56 (br s, 2H), 13.50 (br
s, 2H); 13C NMR (100 MHz, DMSO-d6) d 21.2, 26.4,
30.4, 38.4, 38.7, 41.5, 52.1, 107.8, 114.8, 124.8, 125.4,
147.0, 154.7, 169.1, 183.8; FABHRMS Calcd for
C30H45N8O6 [M+H]+ 613.3462. Found: 613.3456.
6.15. (S,S)-l,4-Bis[2-(2-amino-4-methylsulfanyl-butyryla-
mino)ethylamino]-5,8-dihydroxy-9,10-anthracenedione
trifluoroacetate (16)
Compound 16 was synthesized using compound 15
according to the synthetic method B as a blue solid in
a yield of 98%. Mp 266–268 ꢁC; H NMR (400 MHz,
DMSO-d6) d 1.90–1.97 (m, 10H), 2.42 (t, J = 7.9 Hz,
4H), 3.58–3.67 (m, 8H), 3.76–3.81(m, 2H), 7.17 (s,
2H), 7.62 (s, 2H), 8.18 (s, 6H), 8.74–8.76 (m, 2H),
10.53–10.55 (m, 2H), 13.46 (s, 2H); 13C NMR
(100 MHz, DMSO-d6) d 25.5, 28.6, 31.5, 41.4, 52.1,
62.0, 107.5, 114.8, 124.4, 125.2, 146.9, 154.6, 169.8,
183.5; FABHRMS Calcd for C28H39N6O6S2 [M+H]+
619.2372. Found: 619.2380.
6.20. Thermal denaturation of DNA8,23,24
1
Deoxyribonucleic acid sodium salt from calf thymus (ct-
DNA, Type I, fibrous,) was purchased from Sigma
Chemical Co., Ltd, and used without further purifica-
tion. All the working solutions of ct-DNA and ligands
were prepared in BPE buffer, containing 7.5 mM Na2H-
PO4, 2.5 mM NaH2PO4, and 1 mM EDTA, adjusted to
pH 7.00 0.01 using diluted hydrochloric acid. Ligands
without appropriate solubility in BPE buffer were not
tested. Final DNA–ligand solutions containing 100 lM
of ct-DNA and 20 lM of ligands unless otherwise spec-
ified were prepared by adding concentrated ligand solu-
tions to working solution of ct-DNA. Melting
temperatures were measured by using a JASCO V-550
UV–vis spectrophotometer coupled to a JASCO temper-
ature controller (EHC-441 and UHC-443). Samples
were placed in a JASCO thermostatically controlled
6.16. (R,R)-l,4-Bis[2-(2-tert-butoxycarbonylamino-4-
methylsulfanyl-butyrylamino)ethylamino]-5,8-dihydroxy-
9,10-anthracenedione (17)
Compound 17 was synthesized using N-a-t-Boc-D-
methionine and 26 according to the synthetic method

1014
L.-W. Hsin et al. / Bioorg. Med. Chem. 16 (2008) 1006–1014
8. Morier-Teissier, E.; Boitte, N.; Helbecque, N.; Bernier, J.-
L.; Pommery, N.; Duvalet, J.-L.; Fournier, C.; Hecquet,
B.; Catteau, J.-P.; Henichart, J.-P. J. Med. Chem. 1993, 36,
2084.
9. Meikle, I.; Cummings, J.; Macpherson, J. S.; Smyth, J. F.
Anti-Cancer Drug Des. 1995, 10, 515.
10. Gatto, B.; Zagotto, G.; Sissi, C.; Cera, C.; Uriarte, E.;
Palu, G.; Capranico, G.; Palumbo, M. J. Med. Chem.
1996, 39, 3114.
11. Gatto, B.; Zagotto, G.; Sissi, C.; Palumbo, M. Int. J. Biol.
Macromol. 1997, 21, 319.
cell-holder (EHC-363, 10-mm pathlength). The reduced
volume quartz cuvettes (1 mL, 10-mm pathlength) were
heated by circulating water. The absorbance at 260 nm
was measured over the range 25–95 ꢁC with a heating
rate of 1 ꢁC/min. Data were collected and processed
using the JASCO V-550 spectra manager. The melting
temperature (Tm) was taken to be the midpoint of the
hyperchromic transition determined from first deriva-
tives plots. All DTm values are reported as the
means SD from at least three determinations.
12. Zagotto, G.; Mitaritonna, G.; Sissi, C.; Palumbo, M.
Nucleosides Nucleotides 1998, 17, 2135.
13. Zagotto, G.; Sissi, C.; Gatto, B.; Palumbo, M. ARKIVOC
2004, 5, 204.
6.21. Sulforhodamine B (SRB) assays29
Cells were seeded in 96-well plates in medium with 5%
FBS. After 24 h, cells were fixed with 10% trichloroacetic
acid (TCA) to represent cell population at the time of
compound addition (T0). After additional incubation
of vehicle or compound for 48 h, cells were fixed with
10% TCA and SRB at 0.4% (w/v) in 1% acetic acid was
added to stain cells. Unbound SRB was washed out by
1% acetic acid and SRB bound cells were solubilized with
10 mM Trizma base. The absorbance was read at a wave-
length of 515 nm. Using the following absorbance mea-
surements, such as time zero (T0), control growth (C),
and cell growth in the presence of compound (TX), the
percentage growth was calculated at each of the com-
pound concentration levels. Percentage growth inhibi-
tion was calculated as: [(TX ꢁ T0)/(C ꢁ T0)] · 100 for
concentrations for which TX P q T0. Growth inhibition
of 50% (GI50) was determined at the drug concentration
which resulted in 50% reduction of total protein increase
in control cells during the compound incubation.
14. Meikle, I.; Cummings, J.; Macpherson, J. S.; Hadfield, J.
A.; Smyth, J. F. Biochem. Pharmacol. 1995, 49, 1747.
15. Cummings, J.; Macpherson, J. S.; Meikle, I.; Smyth, J. F.
Biochem. Pharmacol. 1996, 52, 979.
16. Gresh, N.; Kahn, P. H. J. Biomol. Struct. Dyn. 1990, 7,
1141.
17. Gresh, N.; Kahn, P. H. J. Biomol. Struct. Dyn. 1991, 8,
827.
18. Silverman, R. B. DNA and DNA-Interactive Agents. In
The Organic Chemistry of Drug Design and Drug Action,
Academic Press, 1992; pp 229–232.
19. Palumbo, M.; Gatto, B.; Moro, S.; Sissi, C.; Zagotto, G.
Biochim. Biophys. Acta 2002, 1587, 145.
20. Johnson, M. G.; Kiyokawa, H.; Tani, S.; Koyama, J.;
Morris-Natschke, S. L.; Mauger, A.; Bowers-Daines, M.
M.; Lange, B. C.; Lee, K.-H. Bioorg. Med. Chem. 1997, 5,
1469.
21. Routier, S.; Cotelle, N.; Catteau, J.-P.; Bernier, J.-L.;
Waring, M. J.; Riou, J.-F.; Bailly, C. Bioorg. Med. Chem.
1996, 4, 1185.
22. Murdock, K. C.; Durr, F. E.; Citarella, R. V. J. Med.
Chem. 1979, 22, 1024.
Acknowledgments
23. Zee-Cheng, R. K.-Y.; Cheng, C. C. J. Med. Chem. 1978,
21, 291.
24. Cory, M.; McKee, D. D.; Kagan, J.; Henry, D. W.; Miller,
J. A. J. Am. Chem. Soc. 1985, 107, 2528.
25. Foye, W. O.; Karnik, P. S.; Sengupta, S. K. Anti-Cancer
Drug Des. 1986, 1, 65.
26. Sissi, C.; Bolgan, L.; Moro, S.; Zagotto, G.; Bailly, C.;
Menta, E.; Capranico, G.; Palumbo, M. Mol. Pharmacol.
1998, 54, 1036.
27. Gelus, N.; Bailly, C.; Hamy, F.; Klimkait, T.; Wilson, W.
D.; Boykin, D. W. Bioorg. Med. Chem. 1999, 7, 1089.
28. Kamal, A.; Ramu, R.; Khanna, G. B. R.; Saxena, A. K.;
Shanmugavel, M.; Pandita, R. M. Bioorg. Med. Chem.
Lett. 2004, 14, 4907.
We thank Professor Shwu-Bin Lin (School of Medical
Technology, College of Medicine, National Taiwan Uni-
versity) for the assistance in determination of Tm values
and Dr. Jen-Shin Song (National Health Research Insti-
tutes) for the assistance in determination of growth inhi-
bition of tumor cells, and acknowledge the National
Science Council of the Republic of China (Grant Nos.
NSC 93-2323-B-002-016 and NSC 94-2323-B-002-004)
for partial financial support of this work.
References and notes
29. Kim, H. M.; Han, S. B.; Kim, M. S.; Kang, J. S.; Oh, G.
T.; Hong, D. H. J. Pharmacol. Toxicol. Methods 1996, 36,
163.
30. Zhou, N.; Xiao, H.; Li, T.-K.; Nur-E-Kamal, A.; Liu, L.
F. J. Biol. Chem. 2003, 278, 29532.
31. Li, A.; Chen, A. Y.; Yu, C.; Yong, M.; Wang, H.; Liu, L.
F. Genes Dev. 1999, 13, 1553.
32. Harker, W. G.; Slade, D. L.; Parr, R. L.; Feldhoff, P.
W.; Sullivan, D. M.; Holguin, H. M. Cancer Res. 1995,
55, 1707.
1. Willette, R. E. Analgesic Agents. In Wilson and Gisvold’s
Textbook of Organic Medicinal and Pharmaceutical Chem-
istry; Delgado, J. N., Remers, W. A., Eds., 10th Ed.;
Lippincott-Raven: Philadelphia, 1998; pp 687–708.
2. Lown, J. W. Ed., Anthracycline and Anthracenedione-
Based Anticancer Agents, 1988, p 402.
3. Scott, L. J.; Figgitt, D. P. CNS Drugs 2004, 18, 379.
4. Galetta, S. L.; Markowitz, C. CNS Drugs 2005, 19, 239.
5. Kluza, J.; Marchetti, P.; Gallego, M.-A.; Lancel, S.;
Fournier, C.; Loyens, A.; Beauvillain, J.-C.; Bailly, C.
Oncogene 2004, 23, 7018.
33. 2.5 · 105 cells of HL-60 and its TOP2-deficient variant
HL-60/MX-2 were treated with CPT (10 lM), VP-16
(25 lM), AT (25 lM), MX (25 lM), or MAC 16
(25 lM) for 1 h at 37 ꢁC. Cells were then pelleted and
comet assay was performed to analyze chromosome
DNA integrity. At least 50 images of nuclei were
captured for each drug treatment by a CCD camera for
quantitative analysis.
6. Alberts, B.; Bray, D.; Lewis, J.; Raff, M.; Roberts, K.;
Watson, J. D. In Molecular Biology of the Cell, 3rd Ed.;
Garland Publishing Inc.: New York, 1994; pp 404–416.
7. Morier-Teissier, E.; Bernier, J. L.; Lohez, M.; Catteau, J.
´
P.; Henichart, J. P. Anti-Cancer Drug Des. 1990, 5, 291.