Novel design of a pentacyclic scaffold as structural mimic of saframycin A

Article abstract of DOI:10.1016/j.tet.2003.08.036

The design, synthesis and evaluation of a pentacyclic scaffold, CWO-324 to mimic saframycin A is described. CWO-324 is readily synthesized in five steps from 1,4-diacetyl-piperazine-2,5-dione and 2,5-dimethoxybenzaldehyde. CWO-324 was found to scission DN

Full text of DOI:10.1016/j.tet.2003.08.036

Tetrahedron 59 (2003) 8245–8249
Novel design of a pentacyclic scaffold as structural mimic of
saframycin A
b
a
c
Chi Wi Ong,a Yu An Chang, Jing-Yun Wu and Chien-Chung Cheng
,
*
aDepartment of Chemistry, National Sun Yat Sen University, Kaoshiung 804, Taiwan, ROC
bChung Chou Institute of Technology, Chang-Hwa 501, Taiwan, ROC
cDepartment of Chemistry, Tamkang University, Tamsui 251, Taiwan, ROC
Received 13 June 2003; revised 30 July 2003; accepted 4 August 2003
Abstract—The design, synthesis and evaluation of a pentacyclic scaffold, CWO-324 to mimic saframycin A is described. CWO-324 is
readily synthesized in five step0s from 1,4-diacetyl-piperazine-2,5-dione and 2,5-dimethoxybenzaldehyde. CWO-324 was found to scission
DNA, binds to bases 69–83(5 -GCAGTCAGG CACCGT-30) of Hind III/Rsa I from plasmid pBR322 DNA in a foot-printing study and
possesses anti-tumor activity.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
We have approached the design of saframycin from a
different angle, and our strategy is to simplify the structural
complexity of the natural product basic skeleton but to
retain the biological activity. We identified one particular
bond in saframycin A for disconnection, i.e. the C-3 and
C-11 bond, since this bond connects the two isoquinoine-
5,8-dione units to form the bridging ring B that gives
saframycin its complex skeletal framework. The heart of our
design is to reassemble the two isoquinoline units of
saframycin through the fifth ring by forming the novel
5a,12a-diazapentacene ring system, CWO-324, a rather
novel simplified scaffold. The designed molecule, CWO-
324, is similar to saframycin A in that it also has a
pentacyclic structure with a central ring containing an
amino-nitrile functional group. It is important to note that
the designed molecule CWO-324 has a much greater degree
Saframycins,1 natural antiproliferative agents with a novel
molecular architecture containing two units of 7-methoxy-
6-methyl-1,2,3,4-isoquinoline-5,8-dione connected through
the fifth ring to give the central piperazine ring, and bearing
a methylene group substituted with a pyruvamide group has
been an attractive target for total synthesis2a–g due to both
its structural complexity and excellent biological activity.
The promising clinical efficacy of members of this series for
the treatment of solid tumors has recently led to the
preparation of a large number of diverse analogs with
amendable structural modifications devoted to exploring
variation in the nature of the side chains.3 The chemistry and
biology for this class of tetrahydroisoquinoline antitumor
antibiotics have been reviewed.4
of symmetry as compared to the latent symmetry present in
saframycins. Thus CWO-324 now renders itself to an
aesthetically pleasing disconnection whereby compound 1
becomes a precursor. Our design is perhaps attractive in that
it requires only a simple synthesis.
Keywords: 5a,12a-diazapentacene; saframycin model; DNA binding and
scissioning.
*
Corresponding author. Tel.: þ886-7-5252000; fax: þ886-7-5253908;
e-mail: cong@mail.nsysu.edu.tw
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.08.036

8246
C. W. Ong et al. / Tetrahedron 59 (2003) 8245–8249
with a large excess of acetaldehyde in the presence of
trifluoroacetic acid in acetic acid provided the desired
pentacyclic product 2 in 85% yield, together with some
mono-cyclized product 3. We were able to obtain a crystal
of 3 suitable for an X-ray crystallographic analysis and this
proved unequivocally its stereochemistry (Fig. 1). The
reaction of 3 with acetaldehyde in the presence of
trifluoroacetic acid in acetic acid also gave 2. The course
of the reaction has been reported to proceed through an
E-iminium intermediate that cyclizes stereoselectively from
the less hindered a-face.2b,6 Selective mono-reduction of 2
with lithium triethoxyaluminum hydride in THF to the
corresponding cyclic aminal followed by addition of sodium
cyanide gave CWO-324 in 75% yield after flash chromato-
graphy on silica gel (Scheme 1).2
It is particularly important that the analog CWO-324 should
retain a certain degree of similar characteristics as safra-
mycin A for host molecular recognition in the non-covalent
binding to the minor groove of DNA. The DNA binding site
of CWO-324 was determined by foot-printing experiments
using a minor-groove cleaving agent ruthenium (III)–picen
complexes, performed on a 135-bp 32P-end-labeled DNA
obtained from a restriction fragment of Hind III/Rsa I from
plasmid pBR322 DNA.7 This restriction fragment has been
used to foot-print the binding site of the antibiotic
saframycin.8 The foot-printing of compound CWO-324
does not show many binding sites except preferential
binding sequence at bases 69–83(50-GCAGTCAGG
CACCGT-30) (Fig. 2). This is in good agreement with the
binding sequence of saframycin A reported by Lown et al.8
at bases 73–85 (TCAGGCACCGTGT) with Hind III/Rsa I
from plasmid pBR322 DNA. This result strongly suggests
that the average orientation/location of the simplified
scaffold of CWO-324 within the base pair pocket of DNA
has not been altered. Clearly, the good mimic of our
simplified pentacyclic CWO-324 to the parent saframycin
scaffold is noteworthy for the future design of saframycin
analogs.
Figure 1. X-ray structure showing the relative stereochemistry of 3.
2. Results and discussions
Here, we report the synthesis, DNA binding and scission
properties, and cytotoxic activity of CWO-324. The
3,6-bis(2,5-dimethoxyphenyl)methylpiperazine-2-5-dione 1
was prepared as a 10:1 mixture of cis and trans isomers
from the condensation of 1,4-diacetyl-piperazine-2,5-dione
with 2,5-dimethoxybenzaldehyde (K2CO3, DMF, 75%
yield), and subsequent catalytic hydrogenation (Pd/C, H2,
98% yield). Catalytic reduction of bis-arylidenepiperazine-
diones has been reported to give bis(aryl)methyl compound
with the predominant product as the cis isomer.5 The cis
isomer of 1 can be obtained by rigorous washing with
cyclohexane as reported by Lieberskind.5 Myers3b has
recently evaluated the versatility of structural modification
at C-1 by variation of a wide range of aliphatic and aromatic
aldehydes in the Pictet–Springler cyclization reaction. We
have previously described the Pictet–Springler cyclization
of mono-arylidenepiperazinediones with aldehyde that
yields the reverse stereochemistry to saframycin. In this
study, acetaldehyde was chosen to reduce any unfavorable
impact on the binding to DNA due to the incorrect
stereochemistry of the methyl group. The reaction of 1
Computer modeling studies with Cerius II program for the
most stable conformation of the design CWO-324 (total
energy, 138.355 J) and saframycin A (total energy
Scheme 1. Synthesis of CWO-324.

C. W. Ong et al. / Tetrahedron 59 (2003) 8245–8249
8247
sixty cell lines tested, CWO-324 showed selective bio-
logical activities against non-small cell lung, colon,
CNS, melanoma, ovarian, renal and prostrate cell lines.
Unfortunately, it displayed a much lower activity relative to
saframycin A (Table 1). However, the absence of the
quinonoid moiety in CWO-324 provides evidence that it is
not a prerequisite for cytotoxic activity. Ecteinascidins,11 a
non-quinonoid natural product has recently been found to be
more cytotoxic than saframycin A.
It has been proposed that the cytotoxicity of saframycin A is
due to DNA alkylation, and it has been demonstrated, for
instance, that saframycin causes PM2-CCC DNA cleavage
in the presence of reducing agents.12 Our investigation on
the simplified model CWO-324 revealed a similar ability to
the scission of f-174 DNA, although at a much higher
concentration (Fig. 4—concentrations from 50 mM to 1 mM
were found to cause increasing single strand cleavage of
f-174 DNA into the open-circular Form II). The mechanism
of DNA cleavage by CWO-324 may be similar to that
reported for quinocarcin13 that mediates DNA cleavage, and
addition of dithiothreitol (DDT) enhanced DNA cleavage.
Since the structure of quinocarcin differs from saframycins
in that it does not possess a quinoid moiety, the mechanism
by which it mediates DNA cleavage has not been apparent.
Figure 2. Densitometry of autoradiograph of a 10% denaturing poly-
acrylamide gel showing foot-printing of CWO-324 in a 135-bp restriction
fragment (45–101) induced by ruthenium-(III)–picen complexes and
H2O2. The upper and lower region represents different cleavage intensity in
the absence (upper) and presence (lower) of CWO-324.
118.106 J) showed reasonable compatibility in shape
(Fig. 3). The two quinone rings of saframycin has been
reported to be at 758 to each other,9 a close resemblance to
that obtained by our simulation. Overlay of CWO-324 and
saframycin A showed an optimal overlap for ring D and E,
but ring A, B and C are in a slightly different orientation.
The presence of a shallow-cleft within CWO-324 suggests
that it is indeed mimicking the shape of saframycin A and
this further support the similarity in their binding sites.
3. Conclusion
In summary, a strategy of saframycin A mimicry based on
the 5a,12a-diazapentacene scaffold has been successfully
developed. Considering the exceptional similarity in bind-
ing sites to DNA of CWO-324, this new scaffold can be
The saframycin analog CWO-324 was tested in the NCI
human cancer cell panel,10 and the in vitro inhibitory
activities, LC50 (mM), are represented in Table 1. Of the
Figure 3. Computer simulation of the most stable conformation of saframycin A and CWO-324 whereby ring A is placed perpendicular to the plane of the
paper.
Table 1. Inhibition of in vitro human tumor cell growth by CWO-324 (mM)
NCI-H522
COLO-205
U251
SK-MEL-2
OVCAR-8
UO-31
PC-3
BT549
SR
GI50
LC50
TGI
1.74
5.75
3.09
1.85
6.46
3.47
1.72
5.89
3.24
1.73
5.89
3.24
1.71
7.76
5.44
1.90
6.76
3.46
2.91
29
8.71
2.40
6.31
3.09
1.60
.50
5.89
GI50, LC50, TGI values are from the National Cancer Institute of USA. Cell type: NCI-H522, non-small cell lung cancer line; COLO 205, colon cancer cell line;
U251, CNS cancer cell line; SK-MEL-2, melanoma cell line; OVCAR-8, ovarian cancer cell line; UO-31, renal cancer cell line; PC-3, prostate cancer; BT594,
breast cancer cell line; SR, leukemia cell line.

8248
C. W. Ong et al. / Tetrahedron 59 (2003) 8245–8249
Figure 4. Agarose gel electrophoresis of f174 (RF1) DNA(100 mM per base pair) treated with CWO-324 and incubated in 50 mM Tris HCl buffer, pH 8.5 and
1 mM DTT at 378C for 14 h. Lane 1. DNA plasmid as received. Lanes 2–7. 10, 50, 100, 250, 500 and 1000 mM of CWO-324, respectively.
considered a new prototype for the development of novel
potential anti-tumor agents. The ease in the synthesis of this
new scaffold provides a practical and scalable synthesis for
related analogs in the future. Further works are in progress
to improve the cytotoxicity and the preparation of
enantiomerically pure compound.
4.1.3.
1,4,8,11-Tetramethoxy-5,12-dimethyl-
5,6a,7,12,13a,14-octadydro-6,13-dioxo-5a,12a-diaza-
pentacene (2) and 3-(3,5-dimethoxybenzyl)-7,10-
dimethoxy-6-methyl-2,3,11,11a-tetrahydro-5H-pyra-
zino(1,2-b)-isoquinoline-1,4-diones (3). To a solution of
compound 1 (6.02 g, 14.50 mmol) in acetic acid/trifluoro-
acetic acid mixture (50 mL in a 1:4 mixture) was added
acetaldehyde (8.41 mL, 145.00 mmol) and the mixture
stirred at room temperature for 1 h followed by heating to
reflux for 5 h. The reaction was cooled, diluted with cold
water and extracted with chloroform. The organic layer was
washed with sodium bicarbonate and brine, dried (MgSO4)
and the solvent evaporated in vacuo. Purification of the
crude product by column chromatography on silica gel
(ethyl acetate/hexane 1:1) gave 2 (5.82 g, 85%) as a white
solid; mp 231–2328C. [Found: C, 66.86; H, 6.52; N, 6.03.
C26H30N2O6 requires C, 66.94; H, 6.48; N, 6.00%]; nmax
(CHCl3) 1650 cm21; dH (300 MHz, CDCl3) 6.68 (4H, d,
J¼5.8 Hz, ArH), 6.00 (2H, q, J¼6.6 Hz, CH£2), 4.38 (2H,
dd, J¼4.5, 12.6 Hz, CH), 3.78–3.81 (12H, close s,
OCH3£4), 3.11 (2H, dd, J¼4.5, 12.6 Hz, CH2), 2.65 (2H,
dd, J¼12.6, 17.4 Hz, CH2), 1.44 (6H, d, J¼6.6 Hz, CH3£2);
dC (75 MHz, CDCl3) 163.7, 150.8 149.7 126.5, 121.9,
108.0, 107.9 55.6, 55.5, 50.2, 44.6, 29.6, 18.7; m/z (EI,
70 eV) 466 (100, Mþ), 451 (90), 423 (70%). 3a (minor,
320 mg, 5%); mp 201–2028C. [Found: C, 65.18; H, 6.42; N,
6.32. C24H28N2O6 requires C, 65.44, H, 6.41; N, 6.36%];
nmax (CHCl3) 1678, 1652 cm21; dH (300 MHz, CDCl3)
6.66–6.80 (4H, m, ArH), 6.62 (1H, d, J¼1.2 Hz, ArH), 6.10
(1H, br s, NH), 5.94 (1H, q, J¼6.6 Hz, CH), 4.38 (1H, dd,
J¼4.8, 5.6 Hz, CH), 4.23 (1H, dd, J¼4.8, 12.4 Hz, CH),
3.82 (3H, s, OCH3), 380 (3H, s, OCH3), 3.74 (3H, s, OCH3),
3.37 (1H, dd, J¼4.8, 13.8 Hz, CH), 3.28 (3H, s, OCH3), 3.09
(1H, dd, J¼4.8, 17.6 Hz, CH), 2.99 (1H, dd, J¼5.6, 13.8 Hz,
CH), 1.48 (1H, dd, J¼12.4, 17.6 Hz, CH), 1.41 (3H, d,
J¼6.6 Hz, CH3); m/z (EI, 70 eV) 440 (100, Mþ), 425 (60),
397 (20%). Crystallographic data. Mr: 439.49 g/mol;
crystal system: monoclinic; space group: P21/n; a:
4. Experimental
4.1. General procedure
4.1.1. 3,6-Bis[(2,5-dimethoxyphenyl)methylene]-2,5-
piperazinedione. A mixture of 1,4-diacetylpiperazinedione
(5.7 g, 50 mmol), 2,5-dimethoxybenzaldehyde (20.00 g,
0.12 mol) and potassium carbonate (14.65 g, 0.12 mol) in
dry dimethylformamide (150 mL) was stirred under a
nitrogen atmosphere in an oil bath at 808C overnight. The
reaction mixture was cooled and most of the solvent
evaporated in vacuo. Water was added to the oily residue to
give a yellow solid. The crude product was further washed
with ether to give the title compound (15.36 g, 75%) as
yellow solid; mp 2858C (decomposed); nmax (CHCl3) 1680,
1623 cm21; dH (300 MHz, CDCl3) 9.29 (2H, br, NH), 6.94–
7.06 (6H, m, ArH), 6.77 2 H, s, C¼CH£2), 6.28, 3.80 (6H, s,
OCH3£2), 3.74 (6H, s, CH3£2); m/z (EI, 70 eV) 410 (55,
Mþ), 379 (70%), 348 (65%); HRMS (ESI, MeOH): MHþ
found 411.1554. C22H23N2O6 requires 411.1551.
4.1.2. 3,6-Bis[(2,5-dimethoxyphenyl)methyl]-2,5-piperi-
zenedione (1). This was prepared according to the method
of Liebeskind.5 A mixture of 3,6-bis[(2,5-dimethoxy-
phenyl)methylene]-2,5-piperazinedione (9.5 g, 23 mmol)
and 10% palladium on charcoal (1 g) in acetic acid
(150 mL) was shake under a hydrogen atmosphere (Parr
hydrogenator) at 808C for 6 h. The clear and colorless
reaction mixture was filtered through Celite and concen-
trated. Water was added and the white solid was collected
by filtration and dried in vacuo (9.2 g, 98%). The white solid
obtained was triturated with hot cyclohexane several times
to give 1 (8.4 g, 89%) as a white solid; mp 179–1818C; nmax
(CHCl3) 1670 cm21; dH (300 MHz, CDCl3) 7.90 (2H, br,
NH), 6.70–6.85 (6H, m, ArH), 4.23 (2H, br dd, J¼3.9,
9.0 Hz), 3.80 (6H, s, OCH3), 3.75 (6H, s, OCH3), 3.85 (2H,
dd, J¼3.9, 13.5 Hz, CH2), 2.51 (2H, dd, J¼9.0, 13.5 Hz,
CH2); m/z (EI, 70 eV) 414 (65, Mþ), 383 (75%); HRMS
(EI): Mþ, found 414.1793. C22H26N2O6 requires 414.1791.
˚
˚
˚
12.576(2) A; b: 10.899(2) A; c: 15.596(2) A; a¼90.008,
3
˚
b¼90.66(1)8, g¼90.008, V: 2137.6(5) A ; Z: 4; D:
1.278 g/cm3; crystal size: 0.40£0.40£0.60 mm3; R¼0.046,
Rw¼0.039, GOF¼8.80 for 3003 reflections with I¼
˚
.3.00s(1); radiation type: Cu Ka wavelength: 1.54178 A;
diffractometer: Rigaku AFC6S; CCDC-number: 213200.
4.1.4. 1,4,8,11-Tetramethoxy-5,12-dimethyl-
5,6a,7,12,13a,14-octadydro-13-dioxo-5a,12a-diaza-pen-
tac ene-6-carbonitrile (CWO-324). Compound 2 (1 mmol)

C. W. Ong et al. / Tetrahedron 59 (2003) 8245–8249
8249
1015–1018. (b) Kubo, A.; Saito, N. Studies in Natural
Products Chemistry, Atta-ur-Rahman, Ed.; Elsevier: Amster-
dam, 1992; Vol. 10, pp 77–145.
in THF was added dropwise to a solution of lithium
triethoxyaluminum hydride (0.6 mmol) in THF and the
mixture was allowed to react for 3 h at 2108C under
nitrogen. To this reaction mixture was then added a
concentrated solution of sodium cyanide (1.5 mmol) and
the mixture was stirred for 5 h at room temperature.
Purification of the crude product by flash column chromato-
graphy on silica gel (ether/hexane as eluent, 3:1) gave the
title compound CWO-324 in 75% yield as a light yellowish
brown solid; mp 164–1658C. [Found: C, 68.09; H, 6.30; N,
8.93. C27H31N3O5 requires C, 67.91; H, 6.54, N 8.80%];
2. Saframycin, B: (a) Fukuyama, T.; Sachleben, R. A. J. Am.
Chem. Soc. 1982, 104, 4957–4958. (b) Kubo, A.; Saito, N.;
Yamato, H.; Musubuchi, K.; Nakamura, M. J. Org. Chem.
1988, 53, 4295–4310. Saframycin, A: (c) Fukuyama, T.;
Yang, L.; Ajeck, K.; Sachleben, R. A. J. Am. Chem. Soc. 1990,
112, 3712–3713. (d) Myers, A. G.; Kung, W.; Zhong, B.;
Movassaghi, M.; Kwon, S. J. Am. Chem. Soc. 1999, 121,
10828–10829. (e) Martinez, E. J.; Corey, E. J. Org. Lett. 1999,
1, 75–77. (f) Martinez, E. J.; Corey, E. J. Org. Lett. 2000, 2,
993–996. Saframycin C and D: (g) Saito, N.; Wada, Y.; Kubo,
A. Tetrahedron 1990, 46, 7711–7728.
n
max (CHCl3) 2198, 1624 cm21; dH (300 MHz, CDCl3) 6.72
(2H, J¼6.6 Hz, ArH), 6.62 (2H, d, J¼6.6 Hz, ArH), 5.90
(1H, q, J¼5.8 Hz, CH), 5.50 (1H, d, J¼3.6 Hz, CH), 4.35
(1H, q, J¼5.2 Hz, CH), 4.23–4.16 (1H, m, CH), 3.80 (6H, s,
OCH3), 3.71 (6H, s, OCH3), 3.70 (1H, dd, J¼5.8, 15 Hz,
CH), 3.43–3.36 (2H, m, CH2), 2.82 (2H, m, CH2), 1.42 (3H,
d, J¼5.8 Hz, CH3), 1.38 (3H, d, J¼5.2 Hz, CH3); dC
(75 MHz, CDCl3) 18.3, 20.0, 21.6, 22.9, 44.1, 51.4, 53.7,
55.4, 55.5, 55.6, 55.8, 107.0, 107.5, 108.1, 108.9, 110.8,
116.3, 122.8, 123.6, 128.3, 128.9, 149.2, 149.5 (close
doublet), 151.2, 164.5; m/z (EI, 70 eV) 450 (100, Mþ2CN),
435 (80).
3. (a) Myers, A. G.; Plowright, A. T. J. Am. Chem. Soc. 2001,
123, 5114–5115. (b) Myers, A. G.; Lanman, B. A. J. Am.
Chem. Soc. 2002, 124, 12969–12970.
4. Scott, J. D.; Williams, R. M. Chem. Rev. 2002, 102,
1669–1730.
5. Shawe, S. T.; Lieberskind, L. S. Tetrahedron 1991, 47,
5643–5666.
6. Ong, C. W.; Lee, H. C. Aust. J. Chem. 1990, 43, 773–775.
7. Cheng, C. C.; Jian, Y. H.; Lo, C. J.; Cheng, J. W. J. Chin.
Chem. Soc. 1998, 45, 619–624.
4.2. Scissioning and footprinting experiments
8. Lown, J. W.; Joshua, A. V.; Lee, J. S. Biochemistry 1982, 21,
419–428.
l-Phage fX-174 supercoiled plasmid DNA was purchased
from Life Technologies (Gibco BRL). No purification was
needed prior to use for scisioning assay.
9. Hill, G. C.; Remers, W. A. J. Med. Chem. 1991, 34,
1990–1998.
10. Monks, A.; Scudiero, D.; Skehaan, P.; Shoemaker, R.; Paull,
K.; Vistica, D.; Hose, C.; Langley, J.; Cronise, P.; Vaigro-
Wolff, A.; Gray,-Goodrich, M.; Campbell, H.; Mayo, J.; Boyd,
M. J. Natl. Cancer Inst. 1991, 83, 757–766.
A 135-bp DNA was isolated from the plasmid pBR 322
using restrictive endonuclease Hind III/Rsa I (New England
Biolabs) and purified for footprinting experiment by a
sequential procedure described.7
11. (a) Rinechart, K. L.; Holt, T. G.; Freggeau, N. L.; Stroh, J. G.;
Kieffer, P. A.; Sun, F.; Li, L. H.; Martin, D. G. J. Org. Chem.
1990, 55, 4512–4515. (b) Rinehart, K. L. Med. Res. Rev. 2000,
20, 1–27.
Acknowledgements
12. Rao, K. E.; Lown, J. W. Biochemistry 1992, 31,
12076–12082. Ishigo, K.; Takahashi, K.; Yazawa, K.;
Sakiyama, S.; Arai, T. Biol. Chem. 1981, 256, 2162–2167.
13. Remers, W. A.; Flanagan, M. E.; Glinka, T.; Gallegos, R.;
Coffman, H.; Pei, D. J. Am. Chem Soc. 1992, 114, 733–740.
We thank Dr Michael Chang at NSYSU for his assistance
with the X-ray crystallographic studies.
References
1. (a) Arai, T.; Takahashi, K.; Kubo, A. J. Antibiot. 1977, 30,