Convenient synthesis of indeno[1,2-c]isoquinolines as constrained forms of 3-arylisoquinolines and docking study of a topoisomerase I inhibitor into DNA-topoisomerase I complex

Article abstract of DOI:10.1016/j.bmcl.2007.08.062

11-Hydroxyindeno[1,2-c]isoquinolines 12a-c were prepared as constrained forms of 3-arylisoquinolines through an intramolecular cyclization reaction. Among the synthesized compounds, the 11-ⁱbutoxy analog 15l displayed potent in vitro cytotoxici

Full text of DOI:10.1016/j.bmcl.2007.08.062

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5763–5767
Convenient synthesis of indeno[1,2-c]isoquinolines as constrained
forms of 3-arylisoquinolines and docking study of a topoisomerase I
inhibitor into DNA–topoisomerase I complex
Hue Thi My Van,a Quynh Manh Le,a Kwang Youl Lee,a Eung-Seok Lee,b
Youngjoo Kwon,c Tae Sung Kim,d Thanh Nguyen Le,a Suh-Hee Leea and Won-Jea Choa,*
aCollege of Pharmacy and Research Institute of Drug Development, Chonnam National University, Gwangju 500-757,
Republic of Korea
bCollege of Pharmacy, Yeungnam University, Kyongsan 712-749, Republic of Korea
cCollege of Pharmacy, Ewha Womans University, Seoul 120-750, Republic of Korea
dSchool of Life Sciences and Biotechnology, Korea University, Seoul 136-701, Republic of Korea
Received 26 June 2007; revised 16 August 2007; accepted 24 August 2007
Available online 29 August 2007
Abstract—11-Hydroxyindeno[1,2-c]isoquinolines 12a–c were prepared as constrained forms of 3-arylisoquinolines through an intra-
molecular cyclization reaction. Among the synthesized compounds, the 11-ibutoxy analog 15l displayed potent in vitro cytotoxicity
against four different tumor cell lines as well as topoisomerase 1 inhibitory activity. A FlexX docking study was performed to explain
the topoisomerase 1 activity of 15l.
Ó 2007 Elsevier Ltd. All rights reserved.
Topoisomerase 1 (top 1) inhibitors have emerged as
promising anticancer drugs since topotecan and irino-
tecan were launched.1 Both drugs are camptothecin
(CPT) derivatives, which were developed considering
physicochemical properties of camptothecin (1) such as
water solubility and stability. The critical drawbacks
of camptothecin analogs can be summarized as fol-
lows.2,3 These drugs must be infused for long periods
for cancer treatment as they reverse the CPT-trapped
cleavage complexes within minutes and their inactive
decomposed carboxylates are in equilibrium with active
lactone forms under physiological conditions. More-
over, these analogs can cause resistance because the
drugs are also substrates for the efflux transporters.4
Therefore, the development of novel non-camptothecin
top 1 inhibitors has been actively pursued.5–7
in Figure 1. Generally, constrained structures are con-
sidered to have little conformational entropy com-
pared to flexible forms and can be more efficiently
fitted into the active site of a receptor.8 11-Methylin-
denoisoquinoline analogs of 2 that bear several sub-
stituents on aromatic ring A and on the nitrogen
atom have previously been synthesized and have top
1 activities; their cytotoxicities have also been tested.9
Me
11
O
OH
N
O
R1
O
O
N
N
R2
N
O-
O
O
O
Camptothecin (1)
2
MJ-238 (3)
As a part of our ongoing effort to develop isoquino-
line antitumor agents, we designed indenoisoquinolines
as constrained forms of 3-arylisoquinolines as shown
R'
R'
C
R
R
A
B
NH
NH
O
O
Keywords: Indenoisoquinoline; Topoisomerase 1; Docking study;
Antitumor agents; Synthesis; Cytotoxicity.
3-Arylisoquinoline
Indeno[1,2-c]isoquinoline
*
Figure 1. Structure of camptothecin and constrained form of
3-arylisoquinoline to indeno[1,2-c]isoquinoline.
Corresponding author. Tel.: +82 62 530 2933; fax: +82 62 530
2911; e-mail: wjcho@jnu.ac.kr
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.062

5764
H. T. M. Van et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5763–5767
Although 11-methylindenoisoquinolines 2 have weak
top 1 inhibitory activity, their potent cytotoxicities
against tumor cell lines led us to explore the struc-
ture–activity relationships of indenoisoquinolines.
Next, our research focused on introducing an oxygen
functionality at C 11 because the carbonyl group
was known to be essential for H-bonding with Arg
364 of top 1.10 Moreover, modification of the car-
bonyl group to another group, such as hydroxy,
reduction of carbonyl group or its replacement with
an alkoxy group would provide detailed information
of the structure–activity relationship of indeno[1,2-
c]isoquinolines in the binding pocket.
In 2005, the X-ray crystal structure of the indenoiso-
quinoline analog (MJ238)-DNA–top 1 ternary com-
plex was revealed.12 Interestingly, in this paper, the
camptothecin analog (topotecan), the indenoisoquino-
line derivative (MJ238) (3), and the indolocarbazole
analog bound to the same binding sites, despite the
structural difference of these compounds.12 Disclosure
of the detailed binding pocket in the cleavage site of
the top 1–DNA complex enabled researchers to do
computational investigations such as docking studies
and virtual screening using databases to find novel
ligands.
The previously reported lithiated toluamide-benzoni-
trile cycloaddition method was used to synthesize the
3-arylisoquinolines 8a, b. N-Methyl-o-toluamides 6a,
b were treated with n-BuLi to give the anions, which
were then reacted with benzonitrile 7 to afford the
3-arylisoquinolines 8a, b in 39% and 42% yield,
The C ring of indeno[1,2-c]isoquinoline 4 could be con-
structed through intramolecular enamide aldehyde cycli-
zation of compound 5. 3-Arylisoquinoline 5 could be
synthesized via toluamide-benzonitrile cycloaddition
reaction from 6 and 7 as depicted in Schemes 1 and 2.11
H
O
HO
1
R
1
R
Me
NHMe
R
PMBO
N
2
+
R
N
2
R
NC
O
O
O
4
6
7
5
Scheme 1. Retrosynthesis of indenoisoquinoline.
PMBO
PMBO
MeI/NaH
or BnCl/NaH
or PMBCl/K2CO3
R1
Me
R1
PMBO
R1
n-BuLi
+
NHMe
THF, -70 o
C
NH
NC
N
R2
O
7
6
O
O
a: R1=H, R2=Me (57%)
b: R1=Me, R2=Me (85%)
c: R1=Me, R2=PMB (92%)
d: R1=Me, R2=Bn (68%)
a: R1=H (39%)
b: R1=Me (42%)
a: R1=H
b: R1=Me
8
9
HO
HO
OHC
R1
1
1
1
H2,Pd/C
CH3COOH
DDQ
H2O
R
R
R
Acetone
10% HCl
PDC
CH2Cl2
N
R2
EtOH,
80 psi, r.t.
N
N
O
N
2
2
CH2Cl2
R
R2
R
10
11
O
16
12
O
O
a: R1=H, R2=Me (59%)
b: R1=Me, R2=Me (93%)
c: R1=Me, R2=PMB (78%)
a: R1=H, R2=Me (83%)
b: R1=Me, R2=Me (77%)
c: R1=Me, R2=PMB (72%)
a: R1=H, R2=Me (59%)
b: R1=Me, R2=Me (73%)
c: R1=Me, R2=PMB (85%)
a: R1=H, R2=Me (59%)
b: R1=Me, R2=Me (77%)
c: R1=Me, R2=PMB (95%)
R1OH
10%HCl
PDC
CH2Cl2
R1OH
H
+
R1O
O
H
O
1
1
1
R
R
R
Me
2
a: R1=H, R2=Me (98%)
b: R1=Me, R2=Me (89%)
c: R1=Me, R2=PMB (99%)
N
R
N
+
N
N
2
R2
R
2
R
O
17
O
14
O
O
15
13
Scheme 2. The synthesis of indeno[1,2-c]isoquinolines.

H. T. M. Van et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5763–5767
5765
respectively.13,14 Careful treatment of alkyl halides
such as MeI, BnCl, and PMBCl with 8a, b in the
presence of NaH or K2CO3 produced the correspond-
ing N-alkylated compounds 9a–d in 57–92% yield.
Deprotection of the benzyl group on the hydroxy-
methyl on 9a–d was achieved by treatment of DDQ
in methylene chloride. Interestingly, under these reac-
tion conditions, the PMB group attached to the amide
nitrogen on 9c was retained. PDC oxidation of 10a–c
provided the corresponding aldehydes 11a–c, which
were then treated with 10% HCl in acetone to give
the desired cyclization adducts 12a–c in 59–93% yield.
Interestingly, when the alcohols 12a–c were reacted
with various alcohols in the presence of 10% HCl,
the corresponding alkoxy compounds 15a–m were ob-
tained in good yield. This result could be explained by
the successive reactions: dehydration of 13 in the
acidic condition and consecutive nucleophilic attack
of alcohols at C-11 position to provide C-11 alkoxy
compounds 15a–m. The catalytic hydrogenation of
12a–c with 5% Pd/C under 80 psi hydrogen gas in
EtOH afforded 16a–c in 59–95% yield. The hydroxyl
group at C-11 was oxidized by PDC in methylene
chloride to provide the corresponding 11-keto indeno-
isoquinolines 17a–c in excellent yield.
the synthesized compounds for 30 min at 37 °C. The
reaction mixtures were analyzed on 1% agarose gel fol-
lowed by ethidium bromide staining.16
In Table 1, the IC50 cytotoxicity values obtained with
cell lines and the relative potencies of the compounds
are expressed semi-quantitatively as follows: +, weak
activity; +++, lower activity than 0.1 lM camptothecin;
++++, similar or greater activity than 0.1 lM
camptothecin.
Among the various cytotoxic functions of indenoiso-
quinolines, the carbonyl group at C11 position is known
to contribute to hydrogen bonding with Arg 364 of top
1. As shown in Table 1, the cytotoxic activities of
indenoisoquinoline analogs are highly influenced by
the alkoxy analogs, rather than by ketone or hydroxyl
derivatives. 11-Hydroxy analogs 12b–c did not exhibit
significant cytotoxicities against the four tumor cell
lines. These results were not surprising due to the fact
that the hydroxyl group does not work well as a hydro-
gen-bonding donor with Arg 364 of top 1. Compounds
16a–c also did not display potent cytotoxicity. Com-
pound 16c showed low potency (8.9 lmol) against the
HCT 15 cell line. Unexpectedly, 11-keto analogs 17a–c
exhibited weak cytotoxicity (14–30 lmol) or even worse
activity than 16a–c. Furthermore, these compounds did
not have any top 1–DNA inhibitory activity. These re-
sults could not be explained by their low aqueous solu-
bility or poor membrane permeability. However,
dramatic enhancement of cytotoxicity and top 1 inhibi-
tory activity was observed when the hydroxyl groups
were transformed to alkoxy analogs, especially com-
pounds 15g–m. These compounds contain p-methoxyb-
enzyl group at C6 nitrogen and homologous alkoxy
The in vitro cytotoxicity experiments of the synthesized
compounds were performed against four human tumor
cell lines including A 549 (lung), SKOV-3 (ovarian),
SK-MEL-2 (melanoma), and HCT 15 (colon) using sul-
forhodamine B (SRB) assays.15 The top 1 inhibitory
activity assays were carried out using the supercoiled
DNA unwinding method. Five hundred nanograms of
supercoiled pBR 322 DNA was incubated with 1 U
top 1 in the absence or presence of camptothecin or
Table 1. Synthetic yield, IC50 cytotoxicity (lM), and top 1 activity of compounds
No.
Compound
R1
R2
A549
HCT15
OV-3
MEL-2
Top 1a
1
12b
12c
Me
Me
Me
Et
nPr
nBu
iBu
nPt
Me
Et
nPr
iPr
nBu
iBu
nPt
H
Me
300.47
100.41
130.47
90.03
40.60
110.24
28.15
16.19
10.25
6.71
20.88
20.25
20.86
20.61
10.89
10.19
27.39
22.01
1.11
60.11
70.99
60.06
70.94
70.68
30.48
38.37
13.33
1.36
110.97
180.72
70.26
40.85
80.34
20.91
10.52
26.50
15.16
3.60
À
2
PMB
Me
Me
À
3
15a
15b
À
4
À
5
15c
15d
Me
Me
À
6
À
7
15e
15f
Me
Me
À
8
À
9
15g
15h
PMB
PMB
PMB
PMB
PMB
PMB
PMB
Me
+
10
11
12
13
14
15
16
17
18
19
20
21
22
23
1.93
1.42
5.89
5.76
+
15i
15j
3.45
6.22
6.26
2.43
+++
+++
À
0.91
4.40
1.21
2.01
15k
15l
3.24
1.87
3.37
2.07
9.92
5.59
1.63
5.70
+++++
À
15m
16a
11.21
30.09
130.18
50.98
23.51
20.05
155.30
0.067
0.97
11.25
130.31
80.26
31.9
20.50
20.14
8.9
30.56
260.79
23.2
À
16b
16c
Me
Me
H
Me
À
PMB
Me
Me
À
17a
17b
30.55
14.34
171.36
0.080
1.67
80.95
16.29
102.22
0.024
1.17
90.20
17.14
157.54
0.075
4.78
À
Me
Me
À
17c
CPT
Doxo rubicin
PMB
À
+++
a Activity is expressed semi-quantitatively as follows: À, very weak activity; +, weak activity; +++, similar activity to camptothecin; +++++, stronger
activity than camptothecin.

5766
H. T. M. Van et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5763–5767
was developed as a new technique for structure-based
drug design. Fragments of the ligand are automatically
placed into the active site using a new algorithmic ap-
proach based on a pattern recognition technique called
pose clustering. Placement of the ligand is scored based
on protein–ligand interactions. Finally, the binding en-
ergy is estimated, and placements are ranked.
Figure 2. Top 1 inhibitory activities of the compounds. Lane P,
pBR322; lane T, pBR322 + topoisomerase 1; lane C, pBR322 + topo-
isomerase 1 + camptothecin (0.01 mg/ml); lanes 1–10 (prepared com-
pound number, 0.1 mg/ml): 1 (15 g), 2 (15 d), 3 (15h), 4 (15e), 5 (15f), 6
(15k), 7 (15m), 8 (15i), 9 (15j), and 10 (15l).
The structure of the inhibitor 15l was drawn into the
Sybyl package with standard bond lengths and angles
and minimized using the conjugate gradient method
until the gradient was 0.001 kcal/mol with the Tripos
force field. The Gasteiger–Huckel charge, with a dis-
tance-dependent dielectric function, was applied for
the minimization process. We chose the 1SC7 (PDB
code) structure in Protein Data Bank and the struc-
ture was refined as follows. The phosphoester bond
of G12 in 1SC7 was rebuilt and the SH of G11 on
the scissile strand was changed to OH. After the ac-
n
groups from methoxy to pentoxy at the C11 position.
The isobutoxy compound 15l exhibited the most potent
top 1 activity as well as strong cytotoxicity (1.63–
9.92 lmol) against all four tumor cell lines. Interestingly,
compounds 15g–m, which contain p-methoxybenzyl
group at C4 nitrogen, showed more potent cytotoxicities
than the N-methyl substituted compounds 15a–f. Top 1
inhibitory activity of the compounds is depicted in
Figure 2. The semi-quantitative assay was carried out
to show the relative top 1 potency of the compounds.
Compounds 15i and 15j had the same potency as the ref-
erence camptothecin. However, compound 15l exhibited
much more potent inhibition activity than camptothe-
cin. In many cases, the top 1 activity does not correlate
well with cytotoxicities. However, compound 15l
showed potent cytotoxicity and potent top 1 activity.
˚
tive site was defined with a 6.5 A radius, DNA nucle-
otides such as G12, G11, T10, and T9 on the scissile
strand and C112, A113, and A114 on the non-scissile
strand were selected as heteroatoms for the RDF file.
Docking simulations were carried out using FlexX
Single Receptor mode with a Mol2 file molecule as
a Ligand Source. After running FlexX, 30 docked
conformers were displayed in a molecular spread sheet
to rank the scores. We selected the best total score
conformer (À19.188) and speculated regarding the de-
tailed binding patterns in the cavity. The resulting
docking model revealed a very different binding mode
compared to the former 11-methylindenoisoquinoline
model.9 In our model, the benzene ring of p-meth-
oxybenzyl group intercalated between the À1 and +1
bases, parallel to the plane of the base pairs, and
the indenoisoquinoline skeleton, which was positioned
between the À1 and +1 bases in the 11-methylindeno-
isoquinoline model, was placed in the cavity between
the DNA and the top 1 residues, Ala 351, Asn 352,
and Lys 425, perpendicular to the DNA base pairs
as depicted in Figure 3. The oxygen of the p-meth-
oxybenzyl group was H-bonded to Arg 364, which is
considered an essential amino acid that interacts with
the ligand in the DNA–top 1 active site.
Given the X-ray crystallographic structure of top 1–
DNA complex with indenoisoquinoline (MJ238), dock-
ing studies of indenoisoquinolines into the active site
have been considered more convincing than that of mol-
ecules to non-clarified binding sites. To understand the
binding mode of action of the most potent top1 inhibi-
tor 15l, we performed a docking study using FlexX in
the Sybyl 7.2.5 version by Tripos Associates, operating
under Red Hat Linux 4.0 with an IBM computer (Intel
Pentium 4, 2.8 GHz CPU, 1GB memory).
FlexX docking into the DNA–top 1 active site cavity
consisted of three steps: (1) defining the active site; (2)
constructing the ligand structure and, if needed, building
a ligand database for multi-ligand docking process; (3)
defining the receptor description file (RDF). FlexX
Figure 3. Wall-eyed viewing docked model of compound 15l.

H. T. M. Van et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5763–5767
5767
5. Howard-Jones, A. R.; Walsh, C. T. Biochemistry 2005, 44,
15652.
In our model, the p-methoxybenzyl group worked as a
DNA intercalator and as a blocker of the religation step
of the phosphoester. From this docking study, we ob-
served that the indenoisoquinoline ring could be posi-
tioned in the active site, not as a DNA intercalator,
and the other aromatic ring could replace it.
6. Nagarajan, M.; Morrell, A.; Ioanoviciu, A.; Antony, S.;
Kohlhagen, G.; Agama, K.; Hollingshead, M.; Pommier,
Y.; Cushman, M. J. Med. Chem. 2006, 49, 6283.
7. Takai, N.; Ueda, T.; Nishida, M.; Nasu, K.; Fukuda, J.;
Miyakawa, I. Oncol. Rep. 2005, 14, 141.
8. Merabet, N.; Dumond, J.; Collinet, B.; Van Baelinghem,
L.; Boggetto, N.; Ongeri, S.; Ressad, F.; Reboud-Ravaux,
M.; Sicsic, S. J. Med. Chem. 2004, 47, 6392.
9. Cho, W. J.; Le, Q. M.; Van, H. T. M.; Lee, K. Y.; Kang,
B. Y.; Lee, E. S.; Lee, S. K.; Kwon, Y. Bioorg. Med. Chem.
Lett. 2007, 17, 3531.
10. Xiao, X. S.; Antony, S.; Pommier, Y.; Cushman, M.
J. Med. Chem. 2005, 48, 3231.
11. Le, T. N.; Gang, S. G.; Cho, W. J. J. Org. Chem. 2004, 69,
2768.
12. Staker, B. L.; Feese, M. D.; Cushman, M.; Pommier, Y.;
Zembower, D.; Stewart, L.; Burgin, A. B. J. Med. Chem.
2005, 48, 2336.
13. Le, T. N.; Gang, S. G.; Cho, W. J. Tetrahedron Lett. 2004,
45, 2763.
In conclusion, we prepared various indeno[1,2-c]iso-
quinoline analogs as constrained 3-arylisoquinoline
structures. An intramolecular cycloaddition reaction
was employed to efficiently generate 11-hydroxyindeno-
isoquinolines. In order to investigate the structure–activ-
ity relationships, the 11-hydroxy group of the
compounds was modified to another group such as a ke-
tone, dihydro or alkoxy group. The cytotoxic activity of
these analogs was then measured in various cancer cells.
The alkoxy derivatives displayed higher cytotoxicity and
top 1 inhibitory activity than the 11-hydroxy and 11-
keto compounds. Although the reason for these higher
cytotoxicities and top 1 activity is presently not clear,
the top 1 activity could be explained by a docking study
using FlexX in the Sybyl program. To this end, we are
currently investigating the structure–activity relation-
ships of diverse substituted indenoisoquinolines, and
the results will be reported in due course.
14. All synthesized compounds were fully characterized by
spectroscopy. Selected data for some compounds: com-
pound 12a; mp: 217–219 °C. IR (cmÀ1): 3338, 1621 1H
NMR (DMSO-d6): d 8.22 (d, J = 8.1, 1H), 7.72–7.37 (m,
7H), 5.80 (d, J = 7.4, 1H), 5.50 (d, J = 8.3, 1H), 3.94 (s,
3H). EIMS m/z (%) 263 (M+, 100). Compound 12b; mp:
235–237 °C. 1H NMR (CDCl3): d 8.13 (d, J = 8.2, 1H),
7.94–7.28 (m, 6H), 5.80 (d, J = 8.5, 1H), 5.52 (d, J = 8.4,
1H), 3.95 (s, 3H), 2.46 (s, 3H). EIMS m/z (%) 277 (M+,
Acknowledgment
1
86). Compound 12c; mp: 226–228 °C. H NMR (DMSO-
d6): d 8.18 (d, J = 8.2, 1H), 7.89 (s, 1H), 7.64–6.85 (m, 9H),
5.82 (d, J = 8.6, 1H), 5.69 (s, 2H), 5.55 (d, J = 8.5, 1H),
3.69 (s, 3H), 3.35 (s, 3H). EIMS m/z (%) 383 (M+, 64).
Compound 15l; mp: 175–176 °C. 1H NMR (CDCl3): d
8.37 (d, J = 8.3, 1H), 7.89–6.82 (m, 10H), 5.80–5.70 (m,
2H), 5.73 (s, 1H), 3.74 (s, 3H), 2.92–2.87 (m, 1H), 2.79–
2.74 (m, 1H), 2.53 (s, 3H), 1.81–1.76 (m, 1H), 0.86–0.83
(m, 6H). EIMS m/z (%) 439 (M+, 100). Compound 17a;
mp: 218- 221 °C. 1H NMR (CDCl3): d 8.58 (d, J = 8.0,
1H), 8.28–7.35 (m, 7H), 4.00 (s, 3H). EIMS m/z (%) 261
(M+, 75).
This work was supported by Korea Research Founda-
tion Grant (KRF-2004-013-E00031).
References and notes
1. Hardman, W. E.; Moyer, M. P.; Cameron, I. L. Antican-
cer Res. 1999, 19, 2269.
2. Adams, D. J.; da Silva, M. W.; Flowers, J. L.; Kohlhagen,
G.; Pommier, Y.; Colvin, O. M.; Manikumar, G.; Wani,
M. C. Cancer Chemother. Pharmacol. 2006, 57, 135.
3. Gibson, R. J.; Bowen, J. M.; Keefe, D. M. Int. J. Cancer
2005, 116, 464.
15. Rubinstein, L. V.; Shoemaker, R. H.; Paull, K. D.; Simon,
R. M.; Tosini, S.; Skehan, P.; Scudiero, D. A.; Monks, A.;
Boyd, M. R. J. Nat. Cancer Inst. 1990, 82, 1113.
16. Cho, W. J.; Min, S. Y.; Le, T. N.; Kim, T. S. Bioorg. Med.
Chem. Lett. 2003, 13, 4451.
4. Pommier, Y.; Pourquier, P.; Fan, Y.; Strumberg, D.
Biochim. Biophys. Acta 1998, 1400, 83.