The synthesis of amino-acid functionalized β-carbolines as topoisomerase II inhibitors

Article abstract of DOI:10.1016/S0960-894X(01)00136-6

The synthesis and biological activity of amino acid functionalized β-carboline derivatives, which are structurally related to azatoxin and the tryprostatins, are reported. These compounds were assayed for their growth inhibition properties in H520 and PC3

Full text of DOI:10.1016/S0960-894X(01)00136-6

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1251–1255
The Synthesis of Amino-Acid Functionalized ꢀ-Carbolines as
Topoisomerase II Inhibitors
Amy Morin Deveau, Marc A. Labroli, Christine M. Dieckhaus, Mary T. Barthen,
Kirsten S. Smith and Timothy L. Macdonald*
Department of Chemistry, University of Virginia, McCormick Road, Charlottesville, VA 22901, USA
Received 30 December 1999; accepted 20 February 2001
Abstract—The synthesis and biological activity of amino acid functionalized b-carboline derivatives, which are structurally related
to azatoxin and the tryprostatins, are reported. These compounds were assayed for their growth inhibition properties in H520 and
PC3cell lines and were examined for their abilities to inhibit topoisomerase II-mediated DNA relaxation. # 2001 Elsevier Science
Ltd. All rights reserved.
Topoisomerase II (topo II) is a ubiquitous nuclear
enzyme that modulates topological relationships within
the genome of eukaryotic cells; topo II adjusts and
stabilizes DNA during replication, transcription, and
recombination, and also functions in chromosomal
condensation and segregation reactions.1À7 Unlike other
members of the topoisomerase superfamily, topo II is
required for proper cell function because it possesses the
ability to catenate and decatenate DNA.2,4 The essential
nature of topo II coupled with its elevated expression in
tumor cells make it an exceptional target for chemo-
therapy. In fact, two of the most widely utilized anti-
cancer drugs, etoposide and adriamycin (Fig. 1), inhibit
topo II.2,4,5,8
bioavailabilty and enhance potency. The design concept
for these compounds originates from recognition of the
similarities between azatoxin and the tryprostatins16À18
(Fig. 1). Although both azatoxin and the tryprostatins
possess good in vitro growth inhibition properties and a
characteristic l-tryptophan backbone, the tryprostatins
contain a terminal diketopiperazine moiety in place of
azatoxin’s carbamate region. Functionalizing azatoxin’s
carboline core with amino acids and subsequent cycli-
zation to form a terminal diketopiperazine emerged as a
viable synthetic initiative (Figs. 1 and 3). Amino acids
are attractive substrates not only because they are the
fundamental building blocks of biological systems and
many natural products, but also because they are com-
mercially available and possess structurally diverse side
chains. Therefore, by utilizing both d- and l-AAs con-
taining either nonpolar, acidic, basic, or aromatic side
chains, varied functionalities with absolute stereo-
chemistry are introduced to azatoxin’s tetrahydro-
betacarboline core. The chemodivergent methodology
developed provides access to a wide range of AA-sub-
stituted compounds: the AA-functionalized b-carbolines
8a–w and diketopiperazines 8x–z (Fig. 3). As a result of
this work, SARs detailing stereochemical and sub-
stituent preferences for AA-substituted topo II-directed
agents have been elucidated. The ‘acyl region’ domain
of the topo II-pharmacophore has also been further
defined.
Azatoxin (Fig. 1), a compound designed as a hybrid of
etoposide and ellipticine (Fig. 1), inhibits topo II action
and also halts tubulin polymerization at variable con-
centrations.9À11 The intriguing dual activity of azatoxin
coupled with the disclosure of a model pharmacophore9
for topo II inhibitors (Fig. 2) has facilitated the design
and synthesis of novel derivatives that possess height-
ened efficacy and specificity for topo II.12,13 Despite the
enhanced topo II activity of azatoxin derivatives syn-
thesized in recent years,13À15 undesirable physiochemical
and solubility problems preclude the full realization of
their clinical utility.
Recently, a novel class of amino acid (AA)-based aza-
toxin congeners was engineered in an effort to improve
Synthesis of Target Compounds
Amino acid functionalized b-carboline derivatives 8a–z
were prepared in accord with known procedures as
*Corresponding author. Tel.: +1-804-924-7718; fax: +1-804-982-
2302; e-mail: tlm@virginia.edu
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00136-6

1252
A. M. Deveau et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1251–1255
depicted in Figure 3. The 1,3-trans-1,2,3,4-tetra-
hydrobetacarboline intermediate 4 was obtained via a
Pictet–Spengler reaction utilizing l-tryptophan methyl
methylsilyl (TBS) ethers. The resulting coupled species
7a–w were refluxed in N,N-dimethylformamide (DMF)
or with catalytic p-toluenesulfonic acid ( p-TsOH) in tol-
uene to promote diketopiperazine formation. Cycliza-
tion was only successful for 7b, d, and w. These
diketopiperazines and the aforementioned AA-sub-
stituted b-carboline derivatives 7a–w were appropriately
deprotected to afford final compounds 8a–z. TBS-ether
cleavage of compounds of 7k–l and q–r with tetra-
butylammonium fluoride (TBAF) was required prior to
hydrogenolysis, which was conveniently utilized as the
final deprotection protocol for all compounds.
ester hydrochloride
2 and benzyl protected syr-
ingaldehyde 3; a chromatographically separable 65:35/
trans/cis mixture was obtained after 20 h of vigorous
reflux.19 The requisite 1,3-trans relationship of 4 was
verified by 13C spectroscopy as previously described by
Cook et al.20 b-Carboline 4 was then saponified with
LiOH in refluxing THF to the acid 5. After a survey of
coupling processes, 5 was subsequently coupled to the
methyl ester-protected AA of choice 6a–w using benzo-
triazolyloxy-tris[pyrrolidino]-phosphonium hexafluoro-
phosphate
(PyBop)
in
CH2Cl2
with
either
diisopropylethylamine (DIEA) or triethylamine (TEA)
as the base. The selective protection of AA side chains
in compounds 6k–l, q–r, and o–p was necessary before
incorporation into the b-carboline core due to the pres-
ence of additional reactive functionalities. Benzyloxy-
carbonyl protection was employed for the amine side
chains of 6o–p, while the phenolic and alcohol bearing
moieties of 6k–l and q–r were masked as tert-butyldi-
Discussion
While AA-substituted b-carbolines 8a–w were readily
realized, diketopiperazines 8x–z proved to be challeng-
ing synthetic targets. We postulate that this difficulty
may be attributed to the inaccessibility of the sterically
hindered secondary amine in the b-carboline ring or to
the bulkiness and/or stereochemistry of the respective
AA side chain. Existence of an unfavorable amide bond
conformation in 7 may also impede diketopiperazine
formation. Despite the synthetic problems encountered,
it is important to note that Fischer cyclization condi-
tions (NH3/MeOH), and a peptide deblocking strategy
using piperidine21 were also effective for the synthesis of
8x–z (Fig. 3). We continue to seek conditions that will
promote the cyclization of the remaining AA derivatives.
Growth inhibition data as an indicator of cytotoxicity
for AA-functionalized b-carbolines 8a–z in H520 and
PC3cell lines are summarized in Figure 4. Of the 26
compounds screened, only those containing the AAs
Trp or Phe possessed significant activity. Interestingly,
there does not appear to be a trend correlating AA
stereochemistry (d or l) to growth inhibition; H520
cells exhibit approximately a 7% cell survival rate when
treated with compounds 8e, f, i, or j. Additionally, we
found the GI50 for 8e in H520 cells to be 22.4 mM. Since
the growth inhibition profiles of 8e and i are similar, the
extrapolated GI50 for 8i would also be expected to be in
this range.
The topo II assay data for 8a–z are presented in Figures
4 and 5. Compounds 8g, h, and i significantly inhibit
topo II’s ability to relax supercoiled DNA. Although 8h
is not cytotoxic in the cell lines examined, the data for
topo II-mediated DNA relaxation suggests that we are
on track with the rational design of our molecules; 8a, c,
Figure 1. Structures of known topo II inhibitors (top), and structures
of the tryprostatins and AA-functionalized synthetic targets (bottom).
Figure 2. Model pharmacophore for topo II inhibitors.

A. M. Deveau et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1251–1255
1253
Figure 3. Synthesis of AA-functionalized b-carboline derivatives. Reagents: (A) TEA, MeOH; (B) p-TsOH (cat), toluene, reflux; (C) LiOH, THF,
reflux; (D) ByBop, TEA, CH2Cl2; (E) DMF, reflux; (F) TBAF, THF; (G) H2, Pd/C, THF/MeOH.
e, f, j, k, m, o, s, and u also interact with topo II but to a
lesser extent. Most significantly, compounds 8a, c, g, h,
k, m, o, and u exhibit measurable topo II activity, but
do not inhibit cancer cell growth. It is conceivable that
growth inhibition may not be observed despite the
documented inhibition of topo II in the DNA relaxation
assay; these polar derivatives may be unable to cross the
plasma membrane. Additionally, it is possible that these
compounds may be degraded by peptides or acyl
transferases under cellular assay conditions.
upon examination of the topo II data for 8a, c, k, m, o,
s, and u (Fig. 4). While compounds 8e–j do not follow
this pattern, the hydrophobic and aromatic nature of
their side chains may alone satisfy the binding require-
ments to generate topo II-active compounds.
The diketopiperazine analogues 8x–z do not inhibit
H520/PC3cell growth or topo II-mediated DNA
relaxation (Fig. 4). While these results are negative, they
help us to further define the acyl region of the proposed
topo II model pharmacophore (Fig. 2). There appears
to be a preference for the five-membered carbamate
moiety of azatoxin9,15 over the six-membered diketo-
piperazine ring present in compounds 8x–z. It is possible
that the ketone functionalities of the diketopiperazine
may force a conformation that ultimately restricts rota-
tion of the pendant group (Fig. 2). The spatial orientation
and mobility of the pendant group is known to be
important in drug recognition and DNA binding.14
The compounds possessing the greatest topo II activity
incorporate either nonpolar or aromatic AAs into their
structural framework. When comparing 8h to 8f, the
addition or removal of one carbon in the AA side chain
causes a marked difference in the respective compound’s
growth and topo II-inhibition abilities (Fig. 4). This
observation is consistent with a hydrophobic and/or
aromatic binding region on the DNA/topo II complex
that enhances activity against topo II when occupied by
a compound. Previous work in our laboratory has
found that azatoxin derivatives substituted with aniline
functionalities in the variable substituent domain are
more potent topo II toxins than azatoxin itself.13 Thus,
the terminal AA region of 8a–w may indeed mimic the
variable substituent domain of our model pharmaco-
phore (Fig. 2). Data from compounds 8a, c, e–j, s, and u
further support this hypothesis. The remaining topo II-
active compounds (8k, m, and o) also follow this trend
because a segment of their side chain is either nonpolar
or aromatic.
Conclusion
We have realized the synthesis of AA-functionalized
b-carbolines 8a–w and diketopiperazines 8x–z. These
compounds were analyzed for in vitro growth inhibition
of H520 and PC3cell lines, and for their abilities to
inhibit topo II-mediated DNA relaxation. Of the com-
pounds synthesized, 8e, f, i, and j inhibit both cell
growth and topo II action. 8h does not inhibit cell
growth but was found to be active in our topo II-medi-
ated DNA relaxation assay. Data from these com-
pounds suggest that a hydrophobic and/or aromatic
binding region may exist and promote further refine-
ment of our model pharmacophore for topo II/drug
In contrast to the growth inhibition data, a stereo-
chemical preference within the topo II data for d- over
l-amino acids appears to exist. This trend is evident

1254
A. M. Deveau et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1251–1255
Figure 4. Summary of in vitro cell growth and topo II22 assays for 8a–z. Compounds that inhibit both cell growth and topo II-mediated DNA
relaxation are highlighted in red. Compounds that only inhibit topo II-mediated DNA relaxation are highlighted in blue. aCellTiter 96TM Aqueous
Non-Radioactive Cell Proliferation Assay (Promega) was used to determine growth inhibition. All data points run in quadruplicate. Controls were
DMSO (0.20%) and m-AMSA23 (20, 40 mM). Standard error is Æ5 mM. b Densitometer qualitation of topo II-mediated DNA relaxation assay. Gel
was run in triplicate. Scale: (+), drug interacts with topo II; (À), drug does not interact with topo II. A representative gel is shown in Figure
c
5.22 Amino acid noted was used in synthesis as described in Figure 3.
References
1. Wang, J. C. Q. Rev. Biophys. 1998, 31, 107.
2. Osheroff, N.; Froelich-Ammon, S. J. J. Biol. Chem. 1995,
270, 21429.
3. Capranico, G.; Binaschi, M.; Borgnetto, M. E.; Zunino, F.;
22
Palumbo, M. Trends Pharmacol. Sci. 1997, 18, 323.
4. Hsieh, T. In DNA Topology and Its Biologic Effects; Coz-
zarelli, N. R, Wang, J. C, Eds.; Cold Spring Harbor: New
York, 1990; Chapter 7.
5. Liu, L. F. Annu. Rev. Biochem. 1989, 58, 351.
6. Osheroff, N. Pharmacol. Ther. 1989, 41, 223.
7. Robert, J.; Larsen, A. K. Biochimie 1998, 80, 247.
8. Liu, L. F. In DNA Topology and Its Biologic Effects; Coz-
zarelli, N. R., Wang, J. C., Eds.; Cold Spring Harbor: New
York, Chapter 14, p 90.
Figure 5.
Representative agarose gel from the topo II-mediated
DNA relaxation assay. Data for all compounds are not shown. Lane
A: DNA only; lane B: DNA+topo II; lane C: DNA, topo II, and m-
AMSA; lane D: DNA, topo II, and DMSO; lane E: 8y; lane F: 8a;
lane G: 8c; lane H: 8e; lane I: 8h; lane J: 8g; lane K: 8i; lane L: 8k; lane
M: 8m; lane N: 8o.
interactions. Because the AA-functionalized agents syn-
thesized are structurally similar to azatoxin and the
podophyllotoxins, it is important to note that our com-
pounds could also inhibit either topo II-mediated
cleavable complex formation or tubulin polymerization.
9. Leteurtre, F.; Madalengoitia, J.; Orr, A.; Guzi, T.; Lehnert,
E.; Macdonald, T. L.; Pommier, Y. Cancer Res. 1992, 52,
4478.
10. Solary, E.; Leteurtre, F.; Paull, K. D.; Scudiero, D.;
Hamel, E.; Pommier, Y. Biochem. Pharmacol. 1993, 45, 2449.

A. M. Deveau et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1251–1255
1255
11. Leteurtre, F.; Sackett, D. L.; Madalengoitia, J.; Kohlha-
gan, G.; Macdonald, T. L.; Hamel, E.; Paull, K. D.; Pommier,
Y. Biochem. Pharmacol. 1995, 19, 1283.
12. Cho, S. J.; Tropsha, A.; Suffness, M.; Cheng, Y. C.; Lee,
K. H. J. Med. Chem. 1996, 39, 1383.
13. Tepe, J. J.; Madalengoitia, J.; Slunt, K. M.; Werbovetz,
K. W.; Spoors, P. G.; Macdonald, T. L. J. Med. Chem. 1996,
39, 2188.
14. Madalengoitia, J.; Tepe, J. J.; Werbovetz, K. W.; Lehnert,
E.; Macdonald, T. L. Bioorg. Med. Chem. 1997, 5, 1807.
15. Miller, T. A.; Vachaspati, P. R.; Labroli, M. A.; Thomp-
son, C. D.; Bulman, A. L.; Macdonald, T. L. Bioorg. Med.
Chem. Lett. 1998, 8, 1065.
16. Cui, C.; Kakeya, H.; Okada, G.; Onose, R.; Osada, H. J.
Antibiot. 1996, 49, 527.
17. Cook, J. M.; Gan, T. Tetrahedron Lett. 1997, 38, 1301.
18. Danishefsky, S. J.; Depew, K. M.; Rosen, N.; Lorenzino-
Sepp, L. J. Am. Chem. Soc. 1996, 118, 12463.
19. Cook, J. M.; Cox, E. D. Chem. Rev. 1995, 95, 1797.
20. Cox, E. D.; Hamaker, L.; Li, J.; Yu, P.; Czerwinski, K.;
Deng, L.; Bennett, D. W.; Cook, J. M. J. Org. Chem. 1997, 62, 44.
21. Wang, H.; Ganesan, A. Tetrahedron Lett. 1997, 38, 4327.
22. The DNA relaxation assay tests the ability of topo II to
relax supercoiled DNA in the presence of drug. Each incuba-
tion contained 1 mL 10Â incubation assay buffer (0.5 M Tris–
Cl, pH 8.0, 1.2 M KCl, 0.1 M MgCl2, 5 mM ATP, 5 mM DTT,
300 mg/mL BSA), 0.5 mL of 2.0 units topo II (Note: 1 unit of
topo II can decatenate 0.2 microgram KDNA in 30 min at
37 ꢀC. Topo II was obtained from TopoGEN, Inc., Columbus,
OH, USA), 0.5 mL of 0.125 mg/mL plasmid pUC18 DNA from
Escherichia coli (obtained from Boehringer Mannheim GmbH,
Germany), 1 mL of 1 mM drug in 10% DMSO and H2O to
10 mL. The samples were prepared with enzyme added last and
incubated at 37 ꢀC for 30–45 min. The reaction was stopped
with 2.5 mL of decatenation buffer (5% sarkosyl, 25% gly-
cerol, and 0.0025% bromophenol blue). The drugs were
extracted from the incubation with 10 mL of 24:1 chloro-
form/isoamyl alcohol. The aqueous DNA containing sam-
ples were loaded on a 1% agarose gel and run for 90 min at
90 V. The gel was stained with ethidium bromide and
destained with H2O. The DNA bands were detected on a
UV light box and photographed with Polaroid 525 film.
Controls were no enzyme, enzyme, m-AMSA (100 mM), and
DMSO (1%).
23. m-AMSA, the topo II-inhibitor utilized as a control in our
DNA relaxation assay, is also a known DNA intercalator