Rebeccamycin analogues from indolo[2,3-c]carbazole

Article abstract of DOI:10.1016/S0968-0896(00)00251-0

Glycosylated indolocarbazoles related to the antibiotic rebeccamycin represent an important series of antitumor drugs. In the course of structure-activity relationship studies, we report the synthesis of two new derivatives containing an indolo[2,3-c]carb

Full text of DOI:10.1016/S0968-0896(00)00251-0

Bioorganic & Medicinal Chemistry 9 (2001) 357±365
Rebeccamycin Analogues from Indolo[2,3-c]carbazole
Aline Voldoire,a Martine Sancelme,a Michelle Prudhomme,a,* Pierre Colson,b
Claude Houssier,b Christian Bailly,c Stephane Leonced
and Stephanie Lambeld
a
 Á Á
Universite Blaise Pascal, Synthese, Electrosynthese et Etude de Systemes a Interet Biologique, UMR 6504,
63177 Aubiere, France
Á
Á
 Ã
Á
b
Â
Â
Á
Á
Laboratoire de Chimie Macromoleculaire et Chimie Physique, Universite de Liege, Liege 4000, Belgium
cCentre Oscar Lambret and INSERM U-524, IRCL, Place de Verdun, 59045 Lille, France
dInstitut de Recherches SERVIER, 11 Rue des Moulineaux, 92150 Suresnes, France
Received 29 November 1999; accepted 19 September 2000
AbstractÐGlycosylated indolocarbazoles related to the antibiotic rebeccamycin represent an important series of antitumor drugs.
In the course of structure±activity relationship studies, we report the synthesis of two new derivatives containing an indolo[2,3-
c]carbazole chromophore instead of the conventional indolo[2,3-a]carbazole unit found in the natural metabolites. The N-methy-
lated compound 8 containing one glucose residue behaves as a typical DNA intercalating agent, as judged from circular and electric
linear dichroism measurements with puri®ed DNA. In contrast, the bis-glycosylated derivative 7 containing a glucose residue on
each indole nitrogen has lost its capacity to form stable complexes with DNA. DNA relaxation experiments reveal that the two
drugs 7 and 8 have weak e€ects on human DNA topoisomerase I. The modi®ed conformation of the indolocarbazole chromophore
is detrimental to the stabilization of topoisomerase I±DNA complexes. The lack of potent topoisomerase I inhibition leads to
decreased cytotoxicity but, however, we observed that the DNA-intercalating mono-glycosyl derivative 8 is about 5 times more
cytotoxic than the bis-glycosyl analogue 7. The study suggests that the naturally-occurring indolo[2,3-a]carbazole skeleton should
be preserved to maintain the topoisomerase I inhibitory and cytotoxic activities. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
In previous works, we showed that the sugar residue on
the indolocarbazole ring system is a key element for
both DNA binding and topoisomerase I inhibition.9,10
Moreover, the interaction is stereospeci®c. Indolo-
carbazoles linked to the carbohydrate via a b-N-glyco-
sidic linkage are potent DNA intercalators and
topoisomerase I inhibitors, whereas an a-N-glycosidic
linkage results in a complete loss of both DNA
intercalation and topoisomerase I inhibition.11,12
Indolocarbazoles represent an important class of anti-
tumor agents.1 Antibiotics such as staurosporine and K-
252a for which the sugar unit is linked to both indole
nitrogens are potent protein kinase C (PKC) inhibi-
tors.2,3 In contrast, the antibiotics rebeccamycin,4 AT-
2433-A1/B1,5 and the semisynthetic derivatives, ED-
1106 and NB-506,7 for which the carbohydrate residue is
attached to only one indole nitrogen, mainly act as
topoisomerase I inhibitors (Scheme 1). Like the camp-
tothecins, these drugs stabilize topoisomerase I±DNA
covalent complexes.8 Their antitumor activity is attrib-
uted to their capacity to induce topoisomerase I-depen-
dent DNA strand breaks. NB-506 is currently
undergoing clinical trials.
As part of structure±activity relationship studies, we
investigated the synthesis of rebeccamycin analogues 7
and 8 possessing an indolo[2,3-c]carbazole framework
instead of the conventional indolo[2,3-a]carbazole
found in indolocarbazoles from bacterial sources. The
orientation of the indole rings is reversed and, conse-
quently, the sugar residue is now pointing toward the
imide heterocycle. The altered orientation of the carbo-
hydrate moiety is expected to modify the capacity of the
drug to interact with DNA and/or DNA±topoisomerase
I complexes.
*Corresponding author. Tel.: 33-473-40-71-24; fax: 33-473-40-77-17; e-
mail: mprud@chimtp.univ-bpclermont.fr
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00251-0

358
A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
We have shown previously that the introduction of a
methyl group on the imide nitrogen led to stronger
topoisomerase I inhibition.11 For this reason, the newly
designed compounds 7 and 8 were both synthesized with
a methyl group on the imide nitrogen. The original
synthesis, DNA binding and topoisomerase I inhibitory
properties of the mono- and bis-glycosylated indolo[2,3-
c]carbazole derivatives 7 and 8 are reported here. These
compounds were examined for their in vitro anti-
proliferative activities against L1210 leukemia cells, and
their antimicrobial activities against the Gram-positive
bacterium Bacillus cereus.
Scheme 1.
Scheme 2.

A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
359
Results and Discussion
DNA binding
Chemistry
The mono-glycosylated compound 8 has signi®cant
interaction with DNA. The melting temperature (Tm) of
the polynucleotide poly(dA-dT)ꢁ (dA-dT) is markedly
increased in the presence of 8 whereas 7 has absolutely
no e€ect, even at high drug/DNA ratios (Fig. 1). The
Tm value reaches 6.5 ꢀC with 8 whereas the Tm remains
unchanged at 41 ꢀC with 7. The di€erence between the
two compounds was also evident from the absorption
measurements (not shown). The interaction of 8 with
DNA caused a hypochromism in the absorption band
centered at 450 nm whereas the spectrum of 7 was not
a€ected by the addition of DNA. The incorporation of
a second carbohydrate residue seems to abolish the
capacity of the drug to stabilize the double stranded
DNA structure. Complementary biophysical analyses
fully support this view.
3-Hydroxy-2-oxo-1,3-dihydro-10H-3,30 -bisindolyl 1
(Scheme 2) was prepared by reaction of commercial
isatin with indole in the presence of dimethylamine in
ethanol. Intermediate 1 was then reduced to 2 by gen-
erating B2H6 in situ using NaBH4 and BF3/Et2O
according to the procedure described by Berens et al.13
Indolocarbazole 3 was obtained from 2 via a Diels±
Alder cycloaddition in acetic acid at 100 ꢀC.14 N-
Methylation of indolocarbazole 3 with 2N solution of
methylamine in THF in a sealed tube at 70 ꢀC gave 4
with an overall yield of 97.8%. This procedure proved
much more ecient than the conventional method14
based on a Diels±Alder reaction with N-methylmale-
imide. Several methods have been investigated for cou-
pling a carbohydrate unit onto an indolocarbazole
aglycone.11,15À17 The method from 2,3,4,6-tetra-O-ace-
tyl-a-d-glucopyranosyl bromide and silver oxide led
mainly to a-coupling products. But, as mentioned in the
Introduction, only b-coupling products were desired due
to their superior activity pro®les.11,12 The glycosylation
of the indole ring was achieved using the Mitsunobu
reaction with 2,3,4,6-tetra-O-benzyl-d-glucose (a-domi-
nant, Sigma)17 in the presence of both diethylazodi-
carboxylate (DEAD) and triphenylphosphine. Di- and
mono-b-N-glycosylated compounds 5 and 6 were the
major products of the reaction and were isolated in 14.7
and 28.9% yields, respectively. The di- and mono-a-N-
glycosylated compounds were also detected as minor
products of the reaction. Debenzylation of 5 and 6 per-
formed by hydrogenolysis with 10% Pd/C in THF:me-
thanol (1:1 v/v) a€orded analogues 7 and 8 in 46.6 and
43.6% yields, respectively. The b-N-glycosidic bonds
Figure 2shows the circular dichroism (CD) spectra of 7
and 8 either free in solution or bound to calf thymus
DNA. One can immediately observe that the CD spec-
trum of 7 is little modi®ed upon addition of DNA
whereas the spectrum of 8 shows marked changes. In
this case the peak at 458 nm is shifted to 515 nm and the
molar dichroism intensity becomes less negative. The CD
1
were identi®ed from the H NMR coupling constants
0
0
between protons H1 and H2 of the sugar moiety. The
coupling constant of about 9 Hz corresponds to an
axial±axial coupling.
Figure 1. Thermal denaturation curves for (&) poly(dA-dT)ꢁ (dA-dT)
alone or in interaction with (*) 7 or (*) 8 at a drug/DNA ratio of
0.25. The Tm values were determined from the corresponding ®rst-
derivative plots.
Figure 2. Circular dichroism spectra of (solid lines) 7 and 8 free in
solution or (dashed lines) bound to calf thymus DNA. The circular
dichroism amplitude ÁA was measured in 3 mL quartz cuvettes using
20 mM drugÆ500 mM DNA.

360
A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
data corroborate the Tm and UV measurements. To
de®ne more precisely the mode of binding of the two
drugs to DNA, we recorded the electric linear dichroism
spectra of the drug±DNA complexes.
dechlorinated rebeccamycin or synthetic derivatives
including the antitumor drug NB-506.10,18
Further ELD studies were undertaken to compare the
binding of the drugs to the synthetic polynucleotides
poly(dA-dT)ꢁ poly(dA-dT) and poly(dG-dC)ꢁ poly(dG-
dC). Measurements were carried out at a ®xed P/D ratio
of 20 and under the same bu€er and temperature con-
ditions as with calf thymus DNA. Here again, 7 failed
to bind to the polynucleotides (Fig. 4). The spectrum
of the complex between 8 and the AT polymer is rem-
iniscent of that obtained with calf thymus DNA. The
reduced dichroism is negative and the amplitude of the
signal is the same as with the polynucleotide alone at
260 nm (Fig. 4B). This is the expected behavior for a
drug parallel to the DNA base pair plane. But, interest-
ingly, very weak ELD signals were obtained with the
GC polymer with which compound 8 interacts only
poorly. In this case the reduced dichroism is much
weaker than that of the GC polymer alone (Fig. 4B).
From the ELD data, we therefore concluded that 8 can
intercalate into AT sequences but not into GC repeats.
However, no particular sequence selectivity with 8 could be
discerned using footprinting methods. It may be the pecu-
liar conformation of poly(dG-dC) ꢁ poly(dG-dC) rather
than the primary sequence per se which is recognized by 8.
Figure 3 shows the dependence of the reduced dichro-
ism ÁA/A on (A) the wavelength and (B) the electric ®eld
strength. The drug/DNA ratio was ®xed at 20. Very
weak signals were measured with 7. Clearly this bis-gly-
cosyl compound does not form stable complexes with
calf thymus DNA. No particular binding mode could be
determined for this compound, even using synthetic
polynucleotides. In contrast, the ELD spectrum of the
complexes between 8 and calf thymus DNA reveals that
the reduced dichroism is always negative in sign in the
400±500 nm region where the indolocarbazole chromo-
phore absorbs the light (Fig. 3A). The intensity of the
ELD signal is a function of the degree of alignment of
the DNA molecules in the electric ®eld. With 8, there is
a clear parallelism between the electric ®eld dependence
of the reduced dichroism measured at 260 nm for the
DNA bases and at 500 nm for the drug (Fig. 3B). This
indicates that the indolo[2,3-c]carbazole ring is tilted
close to the plane of the DNA bases, consistent with an
intercalative mode of binding. Similar data were pre-
viously reported with indolo[2,3-a]carbazole such as
Figure 3. Dependence of the reduced dichroism ÁA/A on (A) the wavelength and (B) the electric ®eld strength. (E) 7, (J) 8, (G) calf thymus DNA
alone. Conditions: (A) P/D=20, 13 kV/cm, (B) 460 nm for 7, 500 nm for 8, P/D=20 in 1 mM sodium cacodylate bu€er, pH 7.0.
Figure 4. (A) Electric linear dichroism spectra of 8 bound to (G) poly(dA-dT)ꢁ poly(dA-dT) and (B) poly(dG-dC)ꢁ poly(dG-dC). Panel B shows the
®eld strength dependence for (E) drug-free poly(dA-dT)ꢁ poly(dA-dT), (H) drug-free poly(dG-dC)ꢁ poly(dG-dC) and the complexes: (G) 8-poly(dA-
dT)ꢁ poly(dA-dT) and (B) 8-poly(dG-dC)ꢁ poly(dG-dC). In each case the drug was present at 20 mM and the polymer at 400 mM. Conditions: (A)
P/D=20, 13 kV/cm, (B) 500 nm, P/D=20 in 1 mM sodium cacodylate bu€er, pH 7.0.

A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
361
Topoisomerase I inhibition
electrophoresis. The e€ect of 8 may be attributed to a
direct e€ect on the catalytic activity of the enzyme (sta-
bilization of enzyme±DNA complexes) but it may also
result from a nonspeci®c inhibition. Indeed, 8 is an
intercalating agent and as such it can reduce the binding
of the enzyme to the DNA template, especially at high
drug concentrations. Similar e€ects have been recently
observed with an indolo[2,3-a]carbazole derivative.8
We studied the e€ects of 7 and 8 on the relaxation of
plasmid DNA by topoisomerase I. Negatively super-
coiled plasmid pKMp27 was incubated with human
topoisomerase I and increasing concentration of the test
drug, ranging from 1 to 100 mM. DNA samples were
treated with SDS and proteinase K to remove any
covalently bound protein and resolved in a 1% agarose
gel. The gel shown in Figure 5A indicates that 8, but not
7, interferes with the relaxation of DNA mediated by
topoisomerase I. The mono-glycosyl compound shifts
the distribution of the population of the topoisomers.
The experiments presented in Figure 5A were per-
formed without ethidium bromide in the gel during the
To distinguish the speci®c and nonspeci®c e€ects, we
repeated the DNA relaxation experiments using gels
pre-stained with ethidium bromide (Fig. 5B). With
ethidium the electrophoretic mobility of relaxed DNA,
but not that of nicked DNA, is markedly changed
because of DNA unwinding e€ects, whereas under the
Figure 5. E€ect of increasing concentrations of 7 and 8 on the relaxation of plasmid DNA by topoisomerase I. Native supercoiled pKMp27 DNA
(0.5 mg) (lane DNA) was incubated for 30 min at 37 ꢀC with 6 units topoisomerase I (lane TopoI) in the absence or presence of drug at the indicated
concentration (mM). The reactions were stopped with sodium dodecylsulfate and treatment with proteinase K. DNA samples were separated by
electrophoresis on an agarose gel (A) without or (B) with ethidium bromide. Nck, nicked; Rel, relaxed; Sc, supercoiled. The histograms in panel C
compare the topoisomerase I-mediated cleavage eciency of the di€erent drugs. Data were compiled from quantitative analysis of three gels as the
one shown in (B) and must be considered as a set of averaged values.

362
A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
standard conditions used in Figure 5A, nicked DNA and
relaxed (closed circular) DNA comigrate at the same
level in the agarose gel matrix. The gel shown in Figure
5B and the quanti®cation in Figure 5C reveal that in
fact the two drugs have minor e€ects on topoisomerase
I. In the presence of 8, the amount of nicked DNA
produced by topoisomerase I is about the same as that
formed without drug. With 7, the percentage of nicked
DNA is slightly higher indicating that this compound
weakly stabilizes DNA±topoisomerase I covalent com-
plexes. However, the e€ect is poor compared to what can
be achieved with the reference inhibitor camptothecin.
Therefore we must conclude that the inhibition of
topoisomerase I observed with 8 under standard (ethi-
dium-free) conditions (Fig. 5A) is nonspeci®c and arises
solely from DNA binding rather than from interaction
between the drug and covalent enzyme±DNA species. It
is worth mentioning here that we also studied the e€ect
of the drugs on the cleavage activity of the enzyme using
radiolabeled DNA substrates but the pro®les of DNA
cleavage were not a€ected by the two compounds, even
using concentrations as high as 100 mM. The indolo[2,3-
c]carbazole derivatives 7 and 8 cannot be considered as
good topoisomerase I poisons, in contrast to the corre-
sponding indolo[2,3-a]carbazole.
Biological activities
Table 1. In vitro antiproliferative activities against murine leukemia
L1210 cells: percentage of L1210 cells in the G2+M phase (for a drug
concentration of 1 mM) and antimicrobial activities against Bacillus cereus
The antiproliferative activities were tested in vitro
against L1210 leukemia cells. The IC50 values reported
in Table 1 reveal that the two new compounds are much
less cytotoxic than the parent drugs rebeccamycin and
dechlorinated rebeccamycin. However, the mono-glyco-
syl molecule is about 5 times more toxic to leukemia
cells than the corresponding bis-glycosyl analogue. The
di€erences between the two compounds may be attrib-
uted to the higher propensity of 8 to interact with DNA
as compared to 7. The e€ect on the cell cycle of rebec-
camycin and dechlorinated rebeccamycin, that exhibited
the strongest antiproliferative activities, were studied
and it was observed that, at a concentration of 1 mM,
about 70% of L1210 cells were accumulated in the G2+M
phase. The antimicrobial assay shows that the two com-
pounds are equally active at inhibiting the growth of the
Gram-positive bacterium B. cereus (Table 1).
Compd
L1210
IC50 mM
% of L1210
cells in the
B. cereus
MIC mM
G2+M phasea
Rebeccamycin
Dechlorinated rebeccamycin
7
8
0.14
0.11
48.9
10.4
69 (1 mM)
71 (1 mM)
n.e.
10.9
>97
19
n.e.
20
a24% of untreated control cells were in the G2+M phase of the cell
cycle; n.e.: not evaluated.
Discussion
All rebeccamycin derivatives synthesized thus far con-
tained a carbohydrate moiety on one of the two indole
nitrogens. Substitution of both indoles was not possible
due to steric hindrance. In the present case, the change
of the indolocarbazole structure allows for the intro-
duction of two sugar units on each indole nitrogen. The
reversal of the indole rings markedly modi®es the shape
of the molecules. The indolo[2,3-a]carbazole structure is
extended whereas the corresponding indolo[2,3-c]carba-
zole chromophore is essentially curved (Fig. 6). The
geometrical changes may modify signi®cantly stacking
interactions with DNA bases or amino acid residues of
topoisomerase I (e.g., with the catalytic tyrosine 723
residue).
The results indicate that the mono-glycosyl analogue 8
retains the capacity to intercalate into DNA. Therefore,
the curvature introduced into the chromophore is
apparently not an obstacle for intercalation. In contrast,
the indolocarbazole structural change is clearly detri-
mental to the inhibition of topoisomerase I. The carbo-
hydrate pointing toward the imide nitrogen may not be
recognized by the enzyme or the overall con®guration of
the chromophore may not permit optimal stacking
interactions with certain amino acid residues of topo-
isomerase I (e.g., with the tyrosine 723 residue which is
believed to stack with camptothecin).19
Figure 6. Conformation of the (A) indolo[2,3-a]carbazole and (B)
indolo[2,3-c]carbazole chromophores. The energy minimized struc-
tures were built using the softwares HyperChemTM 5.01 and Alchemy
20001
.

A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
363
The presence of a sugar residue on each indole ring
abolishes the capacity of the drug to intercalate into
DNA. The fact that the bis-glycosyl derivative 7 is con-
siderably less cytotoxic than the mono-glycosyl analo-
gue 8 suggests indirectly that DNA-binding contributes
to the cytotoxicity. This bis-glycosyl compound is remi-
niscent of the molecule ED-954 (also called RBM-1)
identi®ed as an 11-O-glucuronide biliary metabolite of
the antitumor drug NB-506.20 At this stage, substitution
of the indolocarbazole structure with two sugar residues
does not seem to represent a pro®table route for the
design of tumor active compounds.
146 mL). The reaction mixture was stirred at À55 to 0 ꢀC
until TLC indicated complete consumption of the indole
(2h). The mixture was poured into 0.2 M aq HCl, then
extracted with EtOAc. The organic phase was washed
with NaHCO3 and brine, and dried over MgSO4. The
solvent was removed and the residue puri®ed by ¯ash
chromatography (eluent cyclohexane:EtOAc, 7:3, then
EtOAc:methanol, 95:5) to give 5 (63 mg, 0.045 mmol,
14.7% yield) and 6 (77 mg, 0.090 mmol, 29% yield) as
red powders. The 1H NMR spectra of 5 and 6 showed a
mixture of conformers.
7-Methyl-8,9-di(ꢀ-D-glucopyranosyl)-6,7,8,9-tetrahydro-
5H-indolo[2,3-c]pyrrolo[3,4-a]carbazole-6,8-dione
In recent years, we have elaborated over 100 rebecca-
mycin derivatives with various side chains on the di€er-
ent rings of the indolocarbazole unit. We have
elucidated important structure±activity relationships
and also a few compounds endowed with marked cyto-
toxic activities were obtained. Here, for the ®rst time,
we modi®ed the chromophore structure. This work
opens a new area to de®ne novel SARs in the
rebeccamycin series.
7.
Hydrogenolysis of compound 5 (53 mg, 0.038 mmol) in
THF:methanol (1:1 v/v, 10 mL) using 10% Pd/C
(15 mg) with hydrogen at atmospheric pressure for 24 h.
After ®ltration over Celite, the solvent was removed and
the residue puri®ed by ¯ash chromatography (eluent
CH2Cl2:methanol, 85:15) to give 7 (12mg, 0.023 mmol,
47% yield) as a red solid. Mp 225±227 ꢀC. IR (KBr) nCO
1738 cmÀ1, nOH 3200±3600 cmÀ1. HRMS (FAB+) calcd
1
for C33H33N3O12 (M+) 633.2059, found 633.2069. H
NMR (400 MHz, DMSO-d6) d 3.24 (3H, s, NCH3),
3.42±3.55 (4H, m), 3.66±3.78 (4H, m), 3.90 (2H, m), 4.00
(2H, m), 4.75 (2H, br s, 2 OH), 4.85 (2H, br s, 2 OH),
Experimental
Chemistry
0
5.19 (4H, br s, 4 OH), 7.51 (2H, d, J=9.1 Hz, 2H 1 ),
7.53 (2H, t, J=7.7 Hz), 7.68 (2H, t, J=7.7 Hz), 8.12
(2H, d, J=8.4 Hz), 8.97 (2H, d, J=8.2Hz); 13C NMR
IR spectra were recorded on a Perkin±Elmer 881 spec-
trometer (n in cmÀ1). NMR spectra were performed on
a Bruker AC 400 (1H: 400 MHz, 13C: 100 MHz) (che-
mical shifts d in ppm, the following abbreviations are
used: singlet (s), doublet (d), triplet (t), multiplet (m),
tertiary carbons (C tert), quaternary carbons (C quat)).
The signals were assigned from exhange with D2O and
inverse gate decoupling. Mass spectra (FAB+) were
determined at CESAMO (Talence, France) on a high
resolution Fisons Autospec-Q spectrometer. Chromato-
graphic puri®cations were performed by ¯ash silica gel
Geduran SI 60 (Merck) 0.040±0.063 mm or Kieselgel 60
(Merck) 0.063±0.200 mm column chromatography. For
purity tests, TLC was performed on ¯uorescent silica gel
plates (60 F254 from Merck). Rebeccamycin was from
our laboratory stock sample.
0
(100 MHz, CH3OD) d 16.5 (NCH3), 63.2(C 6 ), 72.1,
0
0
0
0
0
72.3, 79.7, 81.3, 89.0 (C1 , C2 , C3 , C4 , C5 ), 114.4, 125.3,
126.2, 135.4, 143.4 (C quat arom), 116.7, 122.3, 125.2,
128.9 (C tert arom), 170.3 (CˆO).
7-Methyl-8-(ꢀ-D-glucopyranosyl)-6,7,8,9-tetrahydro-5H-
indolo[2,3-c]pyrrolo[3,4-a]carbazole-6,8-dione 8. Hydro-
genolysis of 6 (67 mg, 0.078 mmol) in the same condi-
tions as described above gave, after 4 days, compound 8
(17 mg, 0.034 mmol, 44% yield) as a red solid. Mp
>300 ꢀC. IR (KBr) nCO 1681, 1697 cmÀ1, nNH,OH 3200±
3600 cmÀ1. HRMS (FAB+) calcd for C27H23N3O7
1
(M+) 501.1531, found 501.1521. H NMR (400 MHz,
DMSO-d6) d 3.22 (3H, s, NCH3), 3.50 (2H, m), 3.71
(2H, m), 3.90 (1H, m), 4.06 (1H, m), 4.74 (1H, t,
0
J=4.9 Hz, OH6 ), 4.92(1H, d, J=5.3 Hz, OH), 5.16
(1H, s, OH), 5.21 (1H, s, OH), 7.46 (1H, t, J=7.3 Hz),
0
7.53 (1H, t, J=7.3 Hz), 7.57 (1H, d, J=9.5 Hz, H1 ),
7.64 (1H, t, J=7.7 Hz), 7.66 (1H, t, J=7.4 Hz), 7.88
(1H, d, J=8.2Hz), 8.08 (1H, d, J=8.3 Hz), 8.90 (1H, d,
J=8.2Hz), 8.95 (1H, d, J=8.1 Hz), 12.01 (1H, s, NH);
13C NMR (100 MHz, DMSO-d6) d 24.0 (NCH3), 61.2
7-Methyl-6,7,8,9-tetrahydro-5H-indolo[2,3-c]pyrrolo[3,4-
a]carbazole-6,8-dione 4. Indolocarbazole 3 (300 mg,
0.92mmol) and a 2 M solution of dimethylamine in
THF was heated at 70 ꢀC in a sealed tube for 4 days.
After cooling, the mixture was poured into water. The
precipitate was ®ltered o€, washed with water and with
Et2O to give 4 (305 mg, 0.90 mmol, 97.8%) as an orange
solid. Mp, IR and NMR spectra were identical with the
descriptions in the literature.11
0
0
0
0
0
0
(C6 ), 70.1, 70.2, 77.7, 80.1, 87.1 (C1 ,C2 ,C3 ,C4 ,C5 ),
110.7, 111.9, 120.3, 122.6, 123.0, 123.1, 129.4, 132.4,
140.2, 143.2 (C quat arom), 112.7, 114.9, 119.9, 120.7,
123.5, 123.8, 126.9, 127.6 (C tert arom), 168.2, 168.4
(CˆO).
7-Methyl-8,9-di(2,3,4,6-tetra-O-benzyl-ꢀ-D-glucopyrano-
syl)-6,7,8,9-tetrahydro-5H-indolo[2,3-c]pyrrolo[3,4-a]car-
bazole -6,8-dione 5 and 7-methyl-8-(2,3,4,6-tetra-O-
benzyl-ꢀ-D-glucopyranosyl)-6,7,8,9-tetrahydro-5H-in-
dolo[2,3-c]pyrrolo[3,4-a]carbazole-6,8-dione 6. To a mix-
Biochemicals. Human topoisomerase I was purchased
from TopoGen Inc. (Columbus, OH). Calf thymus
DNA and the double-stranded polymers poly(dA-
dT)ꢁ poly(dA-dT) and poly(dG-dC)ꢁ poly(dG-dC) were
from Pharmacia (Uppsala, Sweden). Calf thymus
DNA was deproteinized with sodium dodecylsulfate
ture of
4
(100 mg, 0.31 mmol), PPh3 (244 mg,
0.93 mmol), tetra-O-benzyl-d-glucopyranose and THF
(20 mL) at À55 ꢀC was added dropwise DEAD (162mg,

364
A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
(protein content <0.2%) and all nucleic acids were
dialyzed against 1 mM sodium cacodylate bu€er, pH
7.0. All solutions were prepared using doubly deionized,
Millipore ®ltered water.
bu€er (pH 7.4). Cytotoxicity was measured by the
microculture tetrazolium assay as described.23 Cells
were exposed to graded concentrations of the com-
pounds for 48 h and results expressed as IC50 (con-
centration which reduced by 50% the optical density of
treated cells with respect to untreated controls).
Absorption spectra and melting temperature studies.
Melting curves were measured using an Uvikon 943
spectrophotometer coupled to a Neslab RTE111 cryo-
stat. For each series of measurements, 12samples were
placed in a thermostatically controlled cell-holder, and
the quartz cuvettes (10 mm pathlength) were heated by
circulating water. Measurements were performed in
BPE bu€er, pH 7.1 (6 mM Na2HPO4, 2mM NaH 2PO4,
1 mM EDTA). The temperature inside the cuvette was
measured with a platinum probe; it was increased over
the range 20±100 ꢀC with a heating rate of 1 ꢀC/min.
The ``melting'' temperature Tm was taken as the mid-
point of the hyperchromic transition.
Cell cycle analysis. For the cell cycle analysis, L1210
cells (2.5Â105 cells/mL) were incubated for 21 h with
various concentrations of the compounds, then ®xed by
70% ethanol (v/v), washed and incubated in PBS con-
taining 100 mg/mL RNꢀase and 25 mg/mL propidium
iodide for 30 min at 20 C. For each sample, 104 cells
were analyzed on an XL/MCL ¯ow cytometer (Beck-
man Coulter). The ¯uorescence of propidium iodide was
collected through a 615 nm long-pass ®lter. Data are
displayed as linear histograms and results are expressed
as the percentage of cells found in the G2+M phase of
the cell cycle.
Circular dichroism (CD). CD spectra were recorded on a
Jobin-Yvon CD 6 dichrograph interfaced to a micro-
computer. Solutions of drugs, nucleic acids and their
complexes in 1 mM sodium cacodylate bu€er (pH 7.0)
were scanned in 2cm quartz cuvettes. Measurements
were made by progressive dilution of drug±DNA com-
plex at a high P/D (phosphate/drug) ratio with a pure
ligand solution to yield the desired drug/DNA ratio.
Three scans were accumulated and automatically aver-
aged.
Antibiogram tests and MIC determination. MICs of 7
and 8 were determined on B. cereus ATCC 14579 in
Mueller±Hilton broth, pH 7.4 (Difco), after 24 h incu-
bation at 27 ꢀC. The compounds diluted in DMSO were
added to 12tubes; the concentration range was from
100 mg/mL to 0.05 mg/mL.
Acknowledgements
Electric linear dichroism (ELD). ELD measurements
were performed according to the procedures outlined
previously.21 All experiments were conducted at 20 ꢀC
with a 10 mm pathlength Kerr cell having 1.5 mm elec-
trode separation, in 1 mM sodium cacodylate bu€er, pH
7.0. The DNA samples were oriented under an electric
®eld strength of 13 kV/cm, and the drug under test was
present at 10 mM together with the DNA or poly-
nucleotide at 100 mM unless otherwise stated. This elec-
tro-optical method has proved most useful as a means
of determining the orientation of drugs bound to DNA,
and has the additional advantage that it senses only the
orientation of the polymer-bound ligand: free ligand is
isotropic and does not contribute to the signal.22
This work was done under the support of research
grants (to C.B.) from the Ligue Nationale FrancË aise
Contre le Cancer (Comite du Nord) and the Association
pour la Recherche sur le Cancer (to C. H. and P. C.),
from the Actions de Recherches Concertees contract 95/
00-193 and the FNRS, Televie 7/4526/96. Support by
the ``convention INSERM-CFB'' is acknowledged.
References
1. Prudhomme, M. Curr. Pharm. Des. 1997, 3, 265.
2. Tamaoki, T.; Nomoto, H.; Takahashi, I.; Kato, Y.; Mor-
imoto, M.; Tomita, F. Biochem. Biophys. Res. Commun. 1986,
135, 397.
3. Kase, H.; Iwahashi, K.; Matsuda, Y. J. Antibiot. 1986, 39,
1059.
4. Bush, J. A.; Long, B. H.; Catino, J. J.; Bradner, W. T.;
Tomita, K. J. Antibiot. 1987, 40, 668.
5. Matson, J. A.; Claridge, C.; Bush, J. A.; Titus, J.; Bradner,
W. T.; Doyle, T. W.; Horan, A. C.; Patel, M. J. Antibiot. 1989,
42, 1547.
6. Yoshinari, T.; Yamada, A.; Uemura, D.; Nomura, K.;
Arakawa, H.; Kojiri, K.; Yoshida, E.; Suda, H.; Okura, A.
Cancer Res. 1993, 53, 490.
7. Arakawa, H.; Iguchi, T.; Morita, M.; Yoshinari, T.; Kojiri,
K.; Suda, H.; Okura, A.; Nishimura, S. Cancer Res. 1995, 55,
1316.
8. Bailly, C.; Carrasco, C.; Hamy, F.; Vezin, H.; Prudhomme,
M.; Saleem, A.; Rubin, E. Biochemistry 1999, 38, 8605.
9. Rodrigues-Pereira, E.; Belin, L.; Sancelme, M.; Prud-
homme, M.; Ollier, M.; Rapp, M.; Severe, D.; Riou, J. F.;
Á
Fabbro, D.; Meyer, T. J. Med. Chem. 1996, 39, 4471.
10. Bailly, C.; Riou, J.-F.; Colson, P.; Houssier, C.; Rodri-
gues-Pereira, E.; Prudhomme, M. Biochemistry 1997, 36, 3917.
DNA relaxation experiments. Supercoiled pKMp27
DNA (0.5 mg) was incubated with 6 units human topo-
isomerase I at 37 ꢀC for 1 h in relaxation bu€er (50 mM
Tris, pH 7.8, 50 mM KCl, 10 mM MgCl2, 1 mM dithio-
threitol, 1 mM EDTA) in the presence of varying con-
centrations of the drug under study. Reactions were
terminated by adding SDS to 0.25% and proteinase K
to 250 mg/mL. DNA samples were then added to the
electrophoresis dye mixture (3 mL) and electrophoresed
in a 1% agarose gel with or without ethidium bromide
(1 mg/mL), at room temperature for 4 h. Gels were
washed and photographed under UV light.
Growthinhibition assay. L1210 cells (Murine Leukemia)
provided by the NCI, Frederik, USA were cultivated in
RPMI 1640 medium (Gibco) supplemented with 10%
fetal calf serum, 2mM l-glutamine, 100 units/mL peni-
cillin, 100 mg/mL streptomycin, and 10 mM HEPES

A. Voldoire et al. / Bioorg. Med. Chem. 9 (2001) 357±365
365
11. Anizon, F.; Belin, L.; Moreau, P.; Sancelme, M.; Voldoire,
A.; Prudhomme, M.; Ollier, M.; Severe, D.; Riou, J. F.; Bailly,
C.; Fabbro, D.; Meyer, T. J. Med. Chem. 1997, 40, 3456.
12. Bailly, C.; Qu, X.; Graves, D. E.; Prudhomme, M.;
Chaires, J. B. Chem. Biol. 1999, 6, 277.
13. Berens, U.; Brown, J. M.; Long, J.; Selke, R. Tetrahedron:
Asymmetry 1996, 7, 285.
14. Desarbre, E.; Bergman, J. J. Chem. Soc., Perkin Trans. 1
1998, 2009.
15. Kaneko, T.; Wong, H.; Okamato, K. T.; Clardy, J. Tet-
rahedron Lett. 1985, 26, 4015.
16. Bailly, C.; Qu, X.; Anizon, F.; Prudhomme, M.; Riou, J.
F.; Chaires, J. B. Mol. Pharmacol. 1999, 55, 377.
17. Ohkubo, M.; Nishimura, T.; Jona, H.; Honma, T.; Ito, S.
Tetrahedron 1997, 53, 5937.
18. Bailly, C.; Dassonneville, L.; Colson, P.; Houssier, C.;
Fukasawa, K.; Nishimura, S.; Yoshinari, T. Cancer Res. 1999,
59, 2853.
19. Fan, Y.; Weinstein, J. N.; Kohn, K. W.; Shi, L. M.;
Pommier, Y. J. Med. Chem. 1998, 41, 2216.
20. Takenaga, N.; Ishii, Y.; Nakajima, S.; Hasegawa, T.;
Iwasa, R.; Ishizaki, H.; Kamei, T. Drug Metab. Dispos. 1999,
27, 111.
21. Bailly, C.; Henichart, J. P.; Colson, P.; Houssier, C. J.
Mol. Recognit. 1992, 5, 155.
22. Colson, P.; Bailly, C.; Houssier, C. Biophys. Chem. 1996,
58, 125.
23. Leonce, S.; Perez, V.; Casabianca-Pignede, M. R.;
Á
Á
Anstett, M.; Bisagni, E.; Atassi, G. Invest. New Drugs 1996,
14, 169.