Synthesis and biological evaluation of bengacarboline derivatives

Article abstract of DOI:10.1016/j.bmcl.2007.11.113

Novel derivatives of the marine alkaloid bengacarboline have been synthesized. The seco derivatives 11 and 12 were evaluated for topoisomerase inhibition, DNA damages, cytotoxicity and cell cycle perturbation. The two synthetic analogs are more potent cyt

Full text of DOI:10.1016/j.bmcl.2007.11.113

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1212–1216
Synthesis and biological evaluation of bengacarboline derivatives
a
a
a,
b
*
`
Annie Pouilhes, Cyrille Kouklovsky, Yves Langlois, Jean-Pierre Baltaze,
c
c
c
´
´
Stephane Vispe, Jean-Philippe Annereau, Jean-Marc Barret,
Anna Kruczynskic and Christian Baillyc
aLaboratoire de Chimie des Proce´de´s et des Substances Naturelles, ICMMO, UMR8182,
Universite´ de Paris-sud 11, 91405 Orsay, France
bLaboratoire de RMN en milieu oriente´, ICMMO, UMR8182, Universite´ de Paris-sud 11, 91405 Orsay, France
cInstitut de Recherche Pierre Fabre, 3 rue des satellites, BP94244, 31432 Toulouse Cedex, France
Received 29 October 2007; revised 27 November 2007; accepted 29 November 2007
Available online 4 December 2007
Abstract—Novel derivatives of the marine alkaloid bengacarboline have been synthesized. The seco derivatives 11 and 12 were eval-
uated for topoisomerase inhibition, DNA damages, cytotoxicity and cell cycle perturbation. The two synthetic analogs are more
potent cytotoxic agents than bengacarboline and they both induce an accumulation of cells in the S phase of DNA synthesis. They
do not function as topoisomerase inhibitors but trigger DNA damages in cells.
ꢀ 2008 Published by Elsevier Ltd.
Bengacarboline 1, a marine alkaloid extracted from the
Fijian ascidian Didemnum sp., has been described as an
inhibitor of topoisomerase II displaying an in vitro cyto-
toxicity on a wide panel of tumor cell lines.1 Our labora-
tory described some years ago the first synthesis of
racemic bengacarboline 1.2 The biological activity of
this alkaloid led us to prepare analogous compounds
in this series.
Under Schotten-Baumann reaction conditions, nucleo-
philic attack of hydroxyl anion on the acyl iminium
intermediate gave rise to the expected 2-acyl indole 7
in 60% overall yield from 6 (Scheme 1).
As for the synthesis of bengacarboline 1, the formation of
the imino intermediate 8 was made difficult by the low
reactivity of acyl indole carbonyl. Thus, the reaction
was performed under microwave irradiation with zinc
chloride as catalyst. The resulting unstable imine 8 was di-
rectly acylated in the presence of trifluoroacetic anhydride
to afford, after cyclization via the reactive acyl iminium
intermediate 9, the anticipated tetrahydro b-carboline
10, a fully protected bengacarboline derivative (Scheme 2).
7-Bromo tryptamine 2 was selected as starting material
for the synthesis of bengacarboline derivatives. Thus,
according to our previous work, indole-3-carboxylic
acid 3 was protected at nitrogen as a sulfonamide and
the resulting compound 4 was reacted, after acyl chlo-
ride formation, with 7-bromo tryptamine 2 (Scheme
1). The expected amide derivative 5 was isolated in
77% overall yield. Classical Bischler-Napieralski
cyclization afforded, in the presence of phosphorus oxy-
chloride, the dihydro b-carboline derivative 6. Trichlo-
roethyl chloroformate was used at this stage as
acylating agent, instead of benzyl chloroformate as in
our previous synthesis,2 in order to allow a selective
deprotection at the end of the synthesis without hydrog-
enolysis of the C–Br bound.3
At this stage, deprotection with indium metal appeared
as the method of choice for the cleavage of the trichlo-
roethyl carbamate in the presence of the aryl bromides.3
This reaction afforded three compounds 11, 12, and 13.4
1H and 13C NMR spectra of compounds 11 and 12, iso-
lated in 27% and 13%, respectively, were characterized
by the presence of C–H signal at 6.00 and 32.98 and
6.05 and 32.89 ppm, respectively, attributed to the ali-
phatic CH after DEPT, HMBC, and HMQC experi-
ments. 1H NMR spectrum of compound 12 was in
addition characterized by the presence of a singlet at
6.49 ppm corresponding to C20-H. The formation of
seco compounds 11 and 12 could result from tetrahydro-
carboline ring opening affording the seco iminium inter-
Keywords: Bengacarboline; Marine alakaloid; Indole; Cytotoxicity;
DNA damages.
*
Corresponding author. Tel.: +33 1 69 15 76 30; fax: +33 1 69 15 46
79; e-mail: langlois@icmo.u-psud.fr
0960-894X/$ - see front matter ꢀ 2008 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2007.11.113

`
A. Pouilhes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1212–1216
1213
H2N
NH
1
N
N
H
H
NH
Bengacarboline 1
HO
NH
b
NR
O
O
NSO2Ph
N
Br
H
NH2
5
3 : R = H
4 : R = SO2Ph
N
R
a
Br
2
NHTroc
O
N
c
d
N
H
N
Br
Br
NSO2Ph
NSO2Ph
H
6
7
Scheme 1. Reagents and conditions: (a) BuLi (2 equiv), THF, ꢀ60 to ꢀ40 ꢁC, 30 min; PhSO2Cl, ꢀ60 to ꢀ20 ꢁC, 20 h; (b) SOCl2, 20 ꢁC, 3 h; 2,
K2CO3, EtOAc–H2O, 20 ꢁC, 20 h, 77%; (c) POCl3, neat, 80 ꢁC, 3 h; (d) Cl3CCH2OCOCl, Na2CO3, CH2Cl2–H2O, 20 ꢁC, 20 h, 60%.
mediate 14. Indium reduction of the iminium 145,6 and
partial deprotection of the sulfonamide group with
simultaneous deprotection of the Troc carbamate gave
rise to 11 and 12 (path a, Scheme 2). The third minor
product 13, isolated in 8% yield, was probably obtained,
as in the synthesis of bengacarboline 1,2 via a mecha-
nism involving the recyclisation of the common iminium
14 with nucleophilic attack by the Troc protected trypt-
amine chain (path b, Scheme 2). In addition, treatment
of 13 with indium-NH4Cl did not afford the seco deriv-
atives 11 and/or 12. Accordingly, the whole transforma-
tion affording compound 13 from compound 10 seems
to be irreversible. The proposed intermediate 14, gener-
ated from 10, is in agreement with the apparent ex-
change of protecting group. Isolation of seco
compounds 11 and 12 corroborates the proposed mech-
anism. Compound 11 was fully deprotected by alkaline
treatment affording compound 154 (Scheme 2). The sym-
metrical structure of compound 15 was confirmed by its
13C NMR spectrum. As bengacarboline 1 itself is nota-
bly labile, we decided to select the original seco deriva-
tives 11 and 12 for biological evaluation.
Six tumor cell lines originating from different cancer tis-
sues were used to evaluate the anti-proliferative poten-
tial.7 IC50 values measured after 72 h are summarized
in Table 1. Interestingly, in all cases, the new synthetic
products 11 and 12 appeared about two times more
cytotoxic than racemic bengacarboline 1. Surprisingly,
in our hands the latter compound showed a modest
cytotoxicity toward human lung carcinoma A549 epi-
thelial cells but showed no significant cytotoxic action
toward Namalwa cells derived from a human Burkitt
lymphoma and the SKOV3 human ovary carcinoma cell
line. In contrast, compounds 11 and 12 showed a signif-
icant cytotoxic action toward all six selected cell lines,
with IC50 values around 5 lM, including BxPC3, a
p53-mutated human cell line derived from a pancreatic
adenocarcinoma and known to display an intrinsic phe-
notype of resistance. The two compounds show a com-
parable cytotoxicity and their profile is markedly
distinct from that of the reference topoisomerase II
inhibitor etoposide (Table 1). We assumed that the
two compounds affect a common target, similarly ex-
pressed or of similar functional importance in each cell
line.
The pharmacological profile of the two racemic benga-
carboline derivatives 11 and 12 was determined using a
panel of tumor cell lines and selected biochemical assays
to investigate their effects on topoisomerases, in cellulo
DNA cleavage activities, and cell cycle effects.
The effects of the compounds on the functional activity
of DNA topoisomerases were investigated at the molec-
ular and cellular level. Both 11 and 12 were found to af-
fect the relaxation of a supercoiled plasmid by either

`
A. Pouilhes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1212–1216
1214
NHTroc
N
a
N
7
Br
H
NH
NSO2Ph
Br
8
+
N
NHTroc
COCF3
b
N
H
Br
NH
NSO2Ph
Br
9
NHCOCF3
path b
NHTroc
path a
NHTroc
NCOCF3
c
1
N
+
N
H
Br
HN
H
Br
HN
NSO2Ph
Br
NSO2Ph
Br
10
14
path b
path a
reduction with Indium
NHCOCF3
NHR2
NH2
9"
8
4
7
3
9
NTroc
3a
8"
5
3"
2
2"
3"a
1
4"
N
H
Br
N
HN
Indium
Br
6
2a
HN
2'
5"
6"
Br
H
3'
2"a
4' 3'a
NSO2Ph
Br
7"
5'
6'
NR1
2'a
13
7'
11 : R1= SO2Ph; R2 = COCF3
12 : R1= H; R2 = COCF3
15 : R1= H; R2 = H
d
Scheme 2. Reagents and conditions: (a) 2, ZnCl2, 10%, (w/w), 1,4-Me2C6H4, microwave, 130 ꢁC, 1 h, 40%; (b) TFAA, CH2Cl2, reflux, 15 h, 53%; (c)
In, NH4Cl, EtOH–H2O, reflux, 20 h; (d) 11, K2CO3, MeOH–H2O, 60 ꢁC, 6 h, 31%.
topoisomerase I or topoisomerase II in vitro.8 A typical
electrophoresis gel obtained with topoisomerase I is pre-
sented in Figure 1. The relaxation of the plasmid by
purified topoisomerase I is not perturbed by bengacarb-
oline 1, whereas the formation of topoisomers is signif-
icantly reduced in the presence of 11 and 12. IC50 values
of 9 and 8 lM were determined with 11 and 12, respec-
tively, using topoisomerase I. Similar results were ob-
tained with topoisomerase II, with IC50 values of 19
and 13 lM for 11 and 12, respectively (and >100 lM
for 1). The two compounds were also found to be
equally potent in a decatenation assay using kinetoplast

`
A. Pouilhes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1212–1216
1215
Table 1. Cytotoxicity (IC50, lM)a
Compound
1
11
12
Cell line
A549
7.1
10
5.0
3.6
4.5
4.9
3.5
6.5
4.6
5.1
5.0
4.2
5.4
6.5
BxPC3
LoVo
9.9
8.6
>10
>10
MCF7
Namalwa
SKOV3
a SD never exceeded 15%.
1
12
11
Ct T1
Figure 2. DNA damage evaluated by the comet assay. Cells were
treated with etoposide, compound 11 or compound 12 (10 lM each for
1 h). The control sample contained 1% DMSO, as for the drug
samples. Each photomicrograph (10·) is a representative view of a full
slide.
Rel
Sc-
Figure 1. Inhibition of topoisomerase I-mediated DNA relaxation.
Plasmid DNA (200 ng, lane Ct) was incubated for 20 min at 37 ꢁC with
the minimum amount of purified topoisomerase I required to relax
DNA, in the presence or absence of drug at 1, 3.2, 10, or 100 lM.
Reaction was stopped with an SDS–proteinase K treatment. The DNA
was then analysed by agarose gel electrophoresis. The gel was stained
with ethidium bromide and photographed under UV light. Rel, relaxed
DNA; Sc, supercoiled DNA.
strand breaks. The appearance of ‘comets’ characteriz-
ing DNA damage was observed after 1 h of treatment
with each compound at 10 lM (Fig. 2). The amount
of strand breaks induced by 11 or 12 was characterized
by Tail Moment (TM) values of 3.6 2.0 and 7.3 1.7,
respectively. As comparison, TM values with 1% DMSO
(control) or 10 lM etoposide were 1.2 0.7 and
15.5 3.2.
DNA but the effect is modest with IC50 values of about
20 lM for both compounds (data not shown). In our
hands, racemic bengacarboline 1 showed no effect on
DNA decatenation, even at 100 lM. To investigate fur-
ther the effect on topoisomerase I, we compared the
cytotoxicity of 12 toward CEM and CEM/C2 leukemia
cells, respectively, sensitive and resistant to the reference
topoisomerase I poison camptothecin (CPT). CEM/C2
cells carry a mutation on the top1 gene that confers a
18,000-fold resistance to CPT. With 12, IC50 of 5.5
and 5.6 lM were measured with CEM and CEM/C2
cells, respectively, indicating therefore that topoisomer-
ase I is not a key cellular target for this compound. Sim-
ilarly, we tested the cytotoxicity of 12 toward HL-60 and
HL-60/MX2 leukemia cells, respectively, sensitive and
resistant to the reference topoisomerase II poison etopo-
side. The HL-60/MX-2 cell line displays an atypical mul-
tidrug resistance profile with a decreased expression and
activity of topoisomerase II. Again, no major difference
was observed (IC50 of 8.8 and 10 lM with HL-60 and
HL-60/MX2 cells). Therefore, we concluded that topoi-
somerases are not the determinant targets for this cyto-
toxic bengacarboline derivative.
Drug-induced cell cycle perturbations were analyzed by
flow cytometry using A549 cells.10 The data with com-
pound 12 are presented as an example in Figure 3. Com-
pounds 11 and 12 induce relatively similar cell cycle
changes in a dose-dependent manner but the effects are
more pronounced with 12 compared to bengacarboline
1. The increase of the S phase population is much more
100
90
80
70
60
50
40
30
20
10
0
4
DNA-binding studies indicated that neither 11 nor 12
binds to DNA. Fluorescence measurements showed that
these two compounds could not displace intercalated
ethidium bromide or the minor groove binding dye Hoe-
chst 33258 from their binding sites. However, the com-
pounds induce DNA strand breaks in cells. This was
evidenced by the Comet assay (Fig. 2).9 The effect of
12 is much weaker than that of etoposide but neverthe-
less, both 11 and 12 trigger a noticeable amount of DNA
2
1
1/2
G1
1/4
1/8
S
Concentration (x IC50)
0
G2/M
Figure 3. Drug-induced cell cycle effects induced by compound 12.
Percentage of A549 cells in each phase of the cell cycle after 48 h of
incubation in the presence of increasing concentrations of the test
compounds, ranging from 0.25- to 4-fold the IC50 of proliferation at
72 h.

`
A. Pouilhes et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1212–1216
1216
1H), 4.6 (dd, 2H), 4.2 (bd, 1H), 3.6 (bt, 1H), 3.4 (m, 2H),
3.1 (bd, 2H), 2.3 (m, 1H), 2.1 (m, 1H). 13C NMR
(62.5 MHz, CDCl3): 158,9 (COCF3), 154.8 (NCO2R),
138.0–114.5 (C and C–H aromatics), 109.7 (C), 95.2
(CCl3), 74.6 (OCH2), 61.9 (C1), 42.6 (CH2), 41.6 (CH2),
23.8 (CH2), 21.4 (CH2). MS (ESI +) m/z: 1034 (M + Na+),
1036, 1038, 1040. IR, m cmꢀ1: 1720, 1692.
pronounced with the new derivative, which effect is rem-
iniscent to those previously described with antimetabo-
lites such as gemcitabine or arabinosyl-cytosine (Ara-
C). It is plausible that 12 affects DNA synthesis so as
to block or slow down cell cycle during the S phase.
In summary, the two bengacarboline derivatives 11 and
12 both present marked cytotoxic properties toward a
panel of cancer cells. These molecules cannot be consid-
ered as topoisomerase poisons but they induce DNA-
strand breaks in cells. These compounds trigger cell cy-
cle perturbations, mainly an accumulation of cells in the
S phase of DNA synthesis. They represent novel chem-
otypes for the design of antitumor agents derived from
the marine natural product bengacarboline 1. It is worth
to mention that compounds 11 and 12 were tested
in vivo in the P388 murine leukemia model. No antitu-
mor activity (and no toxicity) was observed when the
compounds were injected ip at 10 or 20 mg/kg/injection,
daily for 4 days. Nevertheless, this work has opened the
door to the design of more potent analoges.
Preparation of compound 15: K2CO3 (16 mg, 11.4 mmol)
was added to a solution of compound 11 (24 mg,
0.029 mmol) in MeOH-H2O (3:1, 1.6 mL). The resulting
solution was stirred at 60 ꢁC for 6 h. After evaporation of
MeOH, the residue was extracted with dichloromethane
and the organic layer washed with Na2CO3–H2O (10%).
Evaporation afforded a crude compound (18 mg). After
purification on silica gel (CH2Cl2–MeOH 90/10, NH4OH
1
vap.) compound 15 (5.5 mg, 31%) was isolated. H NMR
400 MHz, CDCl3: 11.3 (bs, 2 H), 8.1 (bs, 1H), 7.4–7.1 (m,
9H), 7.0 (t, J = 12, 1H, C60H), 6.7 (s, 1H, C20H), 6.2 (s,
1H, C1H), 3.1–2.9 (m, 4H, C9H and C900H), 2.8–2.50 (m,
4H, C8H and C800H), 1.5 (bs, 4H). 13C NMR (100 MHz,
CDCl3): 136.9 (C2 and C200), 136.7 (C2a and C200a), 136.6
(C20a), 127.3 (C3a and C300a), 126.7 (C30a), 123.3 (C20),
122.5 (C40), 121.7 (C5 and C500), 119.8 (C60), 119.5 (C50),
119.4 (C4 and C400), 115.4 (C30), 114.5 (C6 and C600), 113.6
(C7 and C700), 111,1 (C70), 109.2 (C3 and C300), 41.5 (C9
and C900), 34.5 (C1), 26.2 (C8 and C800). MS (ESI+) m/z:
604 (M + 1), 606, 608.
Acknowledgments
5. For recent examples of the reductive properties of In-
NH4Cl, see: (a) Banik, B. K.; Banik, I.; Becker, F. F. Org.
Synth. 2005, 81, 188; (b) Vidya Sagar Reddy, G.; Venkat
Rao, G.; Iyengar, D. S. Synth. Commun. 2000, 30, 859,
and references there in.
The authors thank L. Lacastaigneratte, A. Stennevin,
M.-L. Marionneau, J. Filiol, and S. Gras (IRPF) for ex-
pert technical assistance the biological tests.
6. For the indium mediated reduction of imines, see Banik,
B. K.; Hackfeld, L.; Becker, F. F. Synth. Commun. 2001,
31, 1581.
References and notes
7. The human cell lines A549 (lung), BxPC3 (pancreas),
LoVo (colon), MCF7 (breast), Namalwa (lymphoma),
SKOV3 (ovary), and CEM and CEM/C2 (leukemia) were
purchased from the ATCC. Compounds were assayed at
different concentrations ranging from 10 lM to 10 nM.
Cells were seeded in 96-well plates at a density to ensure
logarithmic cell growth phase throughout the 72 h drug
treatment in RPMI medium supplemented with 10% FCS,
glutamine (4 mM), antibiotics penicillin/streptomycin
(100 IU/100 lg/ml), and fungizone (1.25 lg/ml). Cells were
maintained in an incubator at 37 ꢁC under 5% CO2. After
72-h incubation, cell viability was evaluated by dosing
ATP released by viable cells using the ATPLite kit
(Perkin-Elmer, USA). IC50 values correspond to the dose
required to reduce 50% of ATP release by the cell
population.
1. Foderato, T. A.; Barrows, L. R.; Lassota, P.; Ireland, C.
M. J. Org. Chem. 1997, 62, 6064.
`
2. Pouilhes, A.; Langlois, Y.; Chiaroni, A. Synlett 2003,
1488.
3. Mineno, T.; Choi, S.-R.; Avery, M. A. Synlett 2002, 883.
4. Preparation of compounds 11, 12, and 13: To a solution of
compound 10 (347 mg, 0.34 mmol) in EtOH–H2O (64:36,
23 mL) were added at room temperature indium powder
(118 mg, 1.03 mmol, 3 equiv) and ammonium chloride
(55 mg, 1.03 mmol, 3 equiv). Reaction medium was
refluxed for 20 h. After cooling, the reaction medium
was filtered through Celite, extracted with dichlorometh-
ane. Organic layer was successively washed with Na2CO3–
H2O 10% and H2O. The crude mixture of compounds was
purified by preparative TLC (eluant: CH2Cl2–MeOH 90/
10). Compounds 13, 11, and 12 were eluted in that order.
1
8. Topoisomerase inhibition assays were performed as pre-
viously described: Bailly, C. Methods Enzymol. 2001, 340,
610.
9. Comet experiments were performed as previously
described: Barret, J.-M.; Hill, B. T.; Olive, P. Br. J.
Cancer 2000, 83, 1740.
Compound 11: H NMR 250 MHz, CDCl3, d, ppm: 8.7
(bs, 1H), 8.0 (d, J = 9, 1H), 7.8 (d, J = 9, 3H), 7.6–7.0 (m,
13H), 6,0 (s, 1H), 3.6–3.3 (m, 2H), 3.1–2.7 (m, 4H), 2.7–2.5
(m, 2H). 13C NMR (75 MHz, CDCl3, d, ppm): 137.5–
106.9 (C and CH aromatics), 40.3 (CH2), 39.9 (CH2), 33.0
(C1), 23.5 (CH2) and 22.9 (CH2). MS (ESI +) m/z:
840(M + 1), 842, 844. IR, m cmꢀ1: 1713. Compound 12: 1H
NMR 250 MHz: 9.0 (bs, 1H), 8.6 (bs, 1H), 7.8 (d, J = 9,
1H) 7.6–6.9 (m, 11H) 6.5 (s, 1H), 6.0 (s, 1H), 3.9 (bt, 1H),
3.3–3.1 (m, 2H), 3.0–2.6 (m, 8H). 13C NMR (100 MHz,
CDCl3): 136.7–107.1 (C and C–H aromatics), 40.4 (CH2),
39.8 (CH2), 32.9 (C1H), 23.9 (CH2), 23 (CH2). MS (ESI +)
m/z: 700 (M + 1), 702, 704. IR, m cmꢀ1: 1693. Compound
10. Cell cycle measurements. A549 cells were seeded at J0 at
1.3 · 104 cells/cm2. Cells were treated as reported previ-
´
ously (van Hille, B.; Etievant, C.; Barret, J.-M.; Kruczyn-
ski, A.; Hill, B.T. Anticancer Drugs 2000, 11, 829) prior to
the flow cytometry analysis. Fluorescence of propidium
iodide was monitored by the Coulter Epics XL flow
cytometer on the FL-1 channel. Distribution of the cell
populations in G1, S, and G2/M phases was calculated
using a deconvolution algorithm (McCycle software).
1
13: H NMR 400 MHz: 9.5 (bs, 1H), 8.1 (bs, 1H), 7.9 (d,
J = 10, 1H), 7.9 (d, J = 10, 1H), 7.7 to 6.9 (m), 4.7 (bs,