Design, synthesis, and evaluation of phenanthridine derivatives targeting the telomerase RNA/DNA heteroduplex

Article abstract of DOI:10.1016/j.bmcl.2007.01.070

We are targeting molecules to the RNA/DNA heteroduplex that forms during the enzyme telomerase's catalytic cycle. Telomerase is a potential universal anti-cancer target that we have previously shown can be inhibited by molecules that target this heteroduplex. The aim of this work was to make derivatives of our lead, ethidium, that would allow its straightforward incorporation into molecules in both solid and solution phase. The heteroduplex targeting intercalator will act as a scaffold to allow the incorporation of new functionalities that will interact with specific protein surfaces of telomerase, thereby potentially increasing affinity and specificity. In examining multiple new derivatives of ethidium, with literature precedent or novel, we have identified one, a 5-benzylic acid ethidium derivative that is synthesized in three steps as a single isomer, and completely retains the inhibition efficacy of the parent compound. Furthermore, we have demonstrated that it can be effectively incorporated into resin bound amines on the solid phase. As such it represents an ideal monomer for the exploration of telomerse inhibition or for other applications which would benefit from hybrid molecules that can target duplexes.

Full text of DOI:10.1016/j.bmcl.2007.01.070

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2267–2273
Design, synthesis, and evaluation of phenanthridine derivatives
targeting the telomerase RNA/DNA heteroduplex
Subhashree Rangarajan and Simon H. Friedman*
Division of Pharmaceutical Sciences, School of Pharmacy, University of Missouri, Kansas City,
5005 Rockhill Road, Kansas City, MO 64110, USA
Received 25 December 2006; revised 17 January 2007; accepted 18 January 2007
Available online 27 January 2007
Abstract—We are targeting molecules to the RNA/DNA heteroduplex that forms during the enzyme telomerase’s catalytic cycle.
Telomerase is a potential universal anti-cancer target that we have previously shown can be inhibited by molecules that target this
heteroduplex. The aim of this work was to make derivatives of our lead, ethidium, that would allow its straightforward incorpora-
tion into molecules in both solid and solution phase. The heteroduplex targeting intercalator will act as a scaffold to allow the incor-
poration of new functionalities that will interact with specific protein surfaces of telomerase, thereby potentially increasing affinity
and specificity. In examining multiple new derivatives of ethidium, with literature precedent or novel, we have identified one, a
5-benzylic acid ethidium derivative that is synthesized in three steps as a single isomer, and completely retains the inhibition efficacy
of the parent compound. Furthermore, we have demonstrated that it can be effectively incorporated into resin bound amines on the
solid phase. As such it represents an ideal monomer for the exploration of telomerse inhibition or for other applications which
would benefit from hybrid molecules that can target duplexes.
Ó 2007 Elsevier Ltd. All rights reserved.
Telomerase is a ribonucleoprotein that is active in a
large majority of cancer cell types while having minimal
activity in normal cells. In addition to this correlation
there are mechanistic reasons why cancer cells require
telomerase to remain viable after multiple rounds of rep-
lication.1,2 Because of this, telomerase remains a com-
pelling target for the development of anti-cancer
therapeutics and has inspired multiple rational
approaches for identifying inhibitors. We have identified
the RNA/DNA heteroduplex formed during the catalytic
cycle of telomerase as a potential site for inhibition.
Molecules that can bind to this structure could inhibit
telomerase by preventing both extension and transloca-
tion of the DNA substrate.3 We have demonstrated that
known duplex-binding molecules, such as the intercala-
tor ethidium, are able to inhibit the enzyme, and that the
efficacy of inhibition correlates strongly with the affinity
for a model RNA/DNA duplex.3,4 We are using these
intercalators not as therapeutics themselves, but as a
platform upon which to build new functionalities that
can specifically recognize unique telomerase features,
thereby increasing the affinity and specificity of the mol-
ecules for telomerase. The intercalator acts as a scaffold
which provides detectable inhibition, which then allows
incremental improvements introduced by combinatorial
variation to be observed. We anticipate that the scaffold
intercalator may eventually be pared back to limit the
toxicity caused by non-specific nucleic acid binding.
Three features are important for the intercalator mole-
cules that are part of our combinatorial and parallel li-
braries targeting telomerase: (1) the availability of
functional groups on the molecule that are amenable
to modification, (2) the ability of these molecules to
retain telomerase inhibition properties upon modifica-
tion, and (3) the ease and scalability of synthesis of the
molecules. Multiple positions in the structure of ethi-
dium bromide have been explored in the past including
(a) variation of the alkyl chain at position 5 with groups
other than ethyl5,6 (b) modification/replacement of the
exocyclic amino (3- and 8-amino positions),7,8 and (c)
modification/replacement of the phenyl group at posi-
tion 6 with other groups9,10 (Fig. 1). However, the deriv-
atives reported so far are not optimal for our purpose
because of their (1) lack of possible functional groups
for modification,7 (2) generation of mixtures of regioi-
somers upon modification,8,10,11 (3) multi-step synthetic
Keywords: Telomerase; Dulex binding; Solid phase synthesis.
*
Corresponding author. Tel.: +1 816 235 2224; fax: +1 186 235
5190; e-mail: freidmans@umkc.edu
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.070

2268
S. Rangarajan, S. H. Friedman / Bioorg. Med. Chem. Lett. 17 (2007) 2267–2273
sult in incomplete purification.10,8,11 In addition, the
modification of the exocyclic amine can translate into
lower affinity of these derivatives for nucleic acids, aris-
ing from interference with intercalation of the
molecule.8,16
20
19
18
21
16
4
15
17
22
5
6
N
In our search to develop a compound with comparable
efficacy to ethidium bromide toward telomerase that
also possesses functional groups suitable for modifica-
tion, we synthesized, characterized, and tested multiple
modified phenanthridine derivatives. These included
the 5-hexanoic acid derivative, the 5-benzylic acid deriv-
ative, the N-glutaryl derivatives previously reported by
Barton and co-workers,18 as well as the carboxamides
of all of these. All the compounds synthesized were eval-
uated for telomerase inhibition. Only the 5-benzylic acid
derivative retained the inhibition potential of the parent,
while being synthesized as a single regioisomer. In addi-
tion it effectively reacted with resin bound amino groups
using standard solid phase synthesis protocols, thus
demonstrating its utility in the synthesis of libraries of
compounds.
14
13
7
9
NH2
H2N
23
3
12
8
11
24
2
1
10
Figure 1. Structure of ethidium bromide showing the numbering
system.
procedure,12–15 and (4) loss or decrease in nucleic acid
binding affinity following the modifications.16,17
Among the ethidium derivatives reported in the litera-
ture, the majority are phenanthridine derivatives with
modifications of the exocyclic amines. Acylation of
these amino groups results in two regioisomers with
similar physical properties, leading to challenges during
purification. A difference in the reactivity of the two
exocyclic amines has been described in the literature,
and has been attributed to the differences in the elec-
tronic characteristics of the two amines due to the abil-
ity of the lone pair of electrons on the 3- and 8-amines
to delocalize to different extents. The lone pair on the
3-amine can delocalize onto the quaternary nitrogen
making it more electron deficient and less reactive com-
pared to the 8-amine. The separation and isolation of
the regioisomers formed however during modification
of the exocyclic amines is challenging, due to their sim-
ilar chromatographic retention times. This can then re-
The first compound we synthesized was compound 6,
the so called 5-hexanoic acid derivative. This was pre-
pared by using the method of Ross et al.19 namely the
conversion of the precursor alcohol to its triflate, fol-
lowed by reaction with the protected amino phenanthri-
dine precursor (Fig. 2). As a general approach, we tested
both the carboxylic acid product as well as the
carboxamide formed from it against telomerase. The
carboxamide (compound 7) was synthesized to allow
assessment of the compound without the negative char-
ge of the carboxylate present, a situation comparable to
that of the compound after it acylates an amine during
combinatorial synthesis.
HN
NH
O
H2N
NH2
O
O
a
O
O
O
N
N
O
O
HN
NH
2
c
N
1
O
O
5
O
O
b
HO
O
O
O
S
O
O
CF3
3
4
d
H2N
NH2
H2N
NH2
e
N
N
OH
O
NH2
O
6
7
Figure 2. Reagents and conditions: (a) 2.15 equiv of ethylchloroformate, dry pyridine, 5 h, rt; (b) lutidine, DCM, 10 min, on ice, 20 min, rt; (c)
nitrobenzene, 19 h, rt; (d) 48% HBr solution, reflux, 3 h; (e) HOBt, HBTU, DIEA in NMP, reaction with rink amide resin, rt 22 h.

S. Rangarajan, S. H. Friedman / Bioorg. Med. Chem. Lett. 17 (2007) 2267–2273
2269
Both of the resulting compounds were tested for their
ability to inhibit telomerase and both showed a reduc-
tion in inhibition potential, when compared to the ethi-
dium parent. Compound 6 had an IC50 of 23.3 lM and
its carboxamide 7 had an IC50 of 30.3 lM representing a
reduction in inhibition of approximately 10-fold
and 14-fold for compounds 6 and 7, respectively
(Table 1,Figs. 5 and 6).
ative (compound 10) were obtained by reaction of ethi-
dium with glutaric anhydride as reported by Barton and
co-workers (Fig. 3). HPLC analysis of the reaction mix-
ture showed that 9 and 10 were formed in the ratio
ꢀ90:10, respectively, consistent with previous reports.18
We attempted to separate the two regioisomers using re-
verse-phase HPLC, a process made challenging because
the retention times of the two regio-isomers differed by
only 0.2 min on HPLC. We were able to isolate 9 as a
pure compound, whereas the similar retention times of
the regioisomers and low yield of 10 (the 3-isomer),
made it difficult to completely purify. We were able to
enrich this compound by fivefold, giving rise to a
ꢀ50:50 mixture of the 3- and 8-glutaryl derivatives, as
In addition, we examined the N-glutaryl ethidium deriv-
atives previously described by Barton and co-workers.18
The advantage of these compounds is that they are syn-
thesized in a single step from unprotected ethidium. The
8-glutaryl derivative (compound 9) and 3-gluaryl deriv-
Table 1. List of Compounds tested with their IC50 values
Compound
Name
Telomerase IC50 (lM)
Standard error
6
7
3,8-Bisamino 5-hexanoyl 6-phenyl phenanthridine
3,8-Bisamino 5-(5-carboxyaminopentyl) 6-phenyl phenanthridine
Ethidium bromidea
23.3
30.3
2.7
>80b
>80b
68.2
5.6
5.1
4.7
0.8
—
8
9
8-Glutaryl ethidium
3-Glutaryl ethidiumc
8-Glutarylamido ethidium
3-Glutarylamido ethidiumc
3,8-Bisglutaryl ethidium
10
11
12
13
14
16
17
—
13.4
1.4
10.3
—
63.1
>80b
27.3
2.9
3,8-Bisglutarylamido ethidium
3,8-Bisamino 5-(4-(carboxyl)benzyl) 6-phenyl phenanthridine
3,8-Bisamino 5-(4-(carboxylamido)benzyl) 6-phenyl phenanthridine
4.4
0.5
a Ethidium bromide is reported as a control. The inhibition constant of ethidium toward telomerase has been reported previously.
b Indicates the highest concentration the compounds were tested.
c Indicates that the compounds were present as a 50:50 ratio of the 3-derivative and 8-derivative.
HN
O
NH2
HN
O
NH2
b
N
N
a
HOOC
H2NOC
H2N
NH2
9
11
N
COOH
a
CONH2
O
8
O
b
H2N
NH
H2N
NH
N
N
12
10
c
COOH
CONH2
O
O
HN
O
NH
HN
O
NH
d
N
N
HOOC
H2NOC
13
14
Figure 3. Reagents and conditions: (a) 1.05 equiv of glutaric anhydride, DMF, 75 °C, 24 h; (b) activated using 1:1:2 equiv of HBTU:HOBt:DIEA,
NMP; mixture added to rink amide resin, 24 h product cleaved from resin; (c) 3 equiv of glutaric anhydride, DMF, 75 °C, 24 h; (d) activated using
6:6:12 equiv of HBTU:HOBt:DIEA, NMP; mixture added to 20:40 molar excess of ammonium chloride: DIEA, 24 h.

2270
S. Rangarajan, S. H. Friedman / Bioorg. Med. Chem. Lett. 17 (2007) 2267–2273
O
O
O
O
H2N
NH2
HN
O
NH
O
HN
NH
a
b
O
O
N
N
N
O
O
15
2
1
c
H2N
NH2
d
N
NH2
O
H2N
NH2
N
OH
17
O
16
Figure 4. Reagents and conditions: (a) 2.15 equiv of ethylchloroformate, dry pyridine, 5 h, rt; (b) 1.5 equiv of methyl 4-(bromomethyl)benzoate,
acetonitrile, 75 °C, 16 h; (c) 48% hydrobromic acid solution, reflux, 20 h; (d) activated using 1:1:2 equiv of HBTU:HOBt:DIEA, NMP; mixture
added to rink amide resin, 24 h; product cleaved from resin using 95:2.5:2.5 TFA/TIS/water mixture for 1.5 h.
compared to the 1:9 mixture before purification. Com-
pounds 9 and 10 were then converted to their corre-
sponding carboxamides. Again in this case, compound
11, the 3-glutarylamido derivative, was present in small
amounts compared to 5, the 8-glutarylamido derivative.
Following purification, 12 again was enriched by five-
fold in comparison to the original amount formed,
and thus present as a ꢀ50:50 mixture of 12 and 11,
whereas compound 11 was obtained as a pure sample.
We further synthesized compounds 13 (3,8-bisglutaryl
ethidium) and its corresponding bis-amide 14, and puri-
fied these compounds by reverse phase HPLC. All the
glutaryl derivatives were then assayed for telomerase
inhibition using our previously described fluorescence
based assay.20
Ethidium bromide, which was used as a control, had the
expected inhibition constant of 2.7 lM as was reported
in our previous publications.3,4,20 Compounds 9, 10,
and 13 are carboxylic acid derivatives of ethidium bro-
mide. Compounds 9 and 10 had IC50 values >80 lM
and compound 13 had an IC50 value of 63.1 lM. The
Figure 5. IC50 plot of all the carboxylic acid derivatives.
Figure 6. IC50 plot of all the carboxamide derivatives.

S. Rangarajan, S. H. Friedman / Bioorg. Med. Chem. Lett. 17 (2007) 2267–2273
2271
loss in affinity of these compounds compared to ethi-
dium bromide can be potentially attributed to the nega-
tive charge on the carboxylic acid as well as possibly
unfavorable steric interactions of the glutaryl side chains
with the nucleic acid. This result is further corroborated
by publications from other research groups, wherein a
loss of affinity following modifications of exocyclic
amines is observed.16,21,8 Interestingly, 13, the bisderiva-
tive while still significantly worse than the parent ethi-
dium was found to be a better inhibitor compared to
any of the mono-acid derivatives (3- or 8-derivative)
with an IC50 of 63 lM (Table 1, Fig. 5). A similar obser-
vation has been made by Yielding et al.16 specifically
that compounds with two identical substituents at the
3- and 8-position had better affinity to DNA than those
with substitution at only one position.
sized in low amounts and difficult to purify regioiso-
meric mixtures (compound 12), we examined other
ethidium derivatives for synthesis. The so called 5-ben-
zylic acid ethidium, compound 16, was the most
promising of these and its synthesis was carried out
in three steps (Fig. 4). The exocyclic amines of com-
pound 1 were protected with ethylchloroformate to
generate compound 2 in 73% yield. Compound 2
was then treated with methyl 4-(bromomethyl) benzo-
ate, which reacted with the tertiary amine on the phe-
nanthridine ring to generate a quaternary ammonium
salt. The reaction mixture was purified by flash
chromatography and compound 15, the quaternary
ammonium salt with bromide counter ion, was ob-
tained in 57% yield. Compound 15 was then refluxed
with 48% hydrobromic acid solution to hydrolyze
the protecting groups on the exocyclic amines and
on the carboxybenzyl side chain. A side reaction of
the hydrolysis reaction was dealkylation of the carb-
oxybenzyl group to regenerate 1. While shorter reac-
tion duration will result in incomplete hydrolysis of
the ester functionality of 15, prolonged hydrolysis will
regenerate 1 completely via dealkylation. A 20-h reac-
tion period was optimal and ensured complete hydro-
lysis of 15 to generate 16. The small amount of 1
formed was removed by basifying the aqueous solu-
tion with N,N-diisopropylethylamine and then extract-
ing the deprotonated and uncharged compound 1 into
ether. The zwitterionic compound 16 was retained in
the aqueous layer in 87% yield. An equal amount of
diisopropylethylammonium bromide salt was also
present along with 16 as a result of the base extrac-
tion, but the salt however did not interfere with the
derivatization reaction as discussed later.
The carboxylic acid derivatives were converted to their
corresponding carboxamides (compounds 11, 12, and
14) to (a) mask the negative charge from the carboxylic
acid and (b) to simulate peptide derivatization and con-
jugation with other moieties that contain amino groups.
Among the carboxamide compounds, only the 3-glu-
tarylamido ethidium (12) inhibits telomerase at a level
approaching that of the parent (IC50 value of 5.6 lM)
(Table 1, Fig. 6). However since the compound was only
50% pure, with the remaining being the 8-glutarylamido
derivative, we found it difficult to obtain a more accu-
rate inhibition constant. The pure 8-glutarylamido ethi-
dium had an IC50 value of 68.2 lM (Table 1, Fig. 6) and
therefore contributed little to the inhibition of the mix-
ture. Interestingly, compound 14, the 3,8-bisglutarylam-
ido derivative, is a worse inhibitor than 13, its bis-acid
precursor (IC50 > 80 lM), even though the negative
charges were eliminated.
As with the other compounds, 16 was also converted to
its corresponding carboxamide, compound 17.
Although 16 has the required synthetic handle for deriv-
atization, unlike ethidium, it has a bulky benzyl group at
Because the derivatives based on literature precedent
either lost telomerase inhibition efficacy (compounds
9–11, 13, and 14) or retained efficacy but were synthe-
H2N
NH2
O
N
NH
H2N
HN
H2N
NH2
O
N
NH2
O
O
a, b, c
H2N
O
HN
H2N
16
N
NH2
18
Figure 7. Reagents and conditions: (a) Rink amide D-series lantern, 20% piperidine in NMP, Fmoc-lys(Mtt)-OH, HBTU:HOBt:DIEA 1:1:2, NMP;
(b) 20% piperidine in NMP, Fmoc-lys(Mtt)-OH, HBTU:HOBt:DIEA 1:1:2, NMP; 95:5 DCM:TFA, 5 min; (c) HATU, DIEA, NMP, 20 h; 20%
piperidine in NMP, 3 min; compound cleaved from the lantern using 95:2.5:2.5 TFA/TIS/water mixture for 1.5 h.

2272
S. Rangarajan, S. H. Friedman / Bioorg. Med. Chem. Lett. 17 (2007) 2267–2273
Figure 8. MS spectrum of compound 18.
the N5 position that might obstruct the intercalation of
the molecule. However, a crystal structure of ethidium
bound to a dinucleotide suggested that the benzyl group
will be pointing out of the minor groove of the helix,
acylation. We are therefore exploring its utility to act
as a foundation upon which to build unique functional
groups that can interact with the unique telomerase pro-
tein surfaces that surround the telomerase RNA/DNA
duplex. We believe that it may also have utility for other
laboratories interested in attaching functionalities to
ethidium to impart other properties.
much like the ethyl group in ethidium bromide.22
A
modeled complex depicting this (based on the crystal
structure of ethidium bound to a GC dinucleotide in
Ref. 22) is included in Supplemental data.
Compound 16 had a telomerase inhibition constant of
27.3 lM, (Table 1, Fig. 5), a value which represents a
10-fold decrease in efficacy in comparison with the par-
ent ethidium bromide. However, its corresponding
amide derivative, 17, had an IC30 value of 2.9 lM, essen-
tially identical to that of ethidium bromide (Fig. 6).
Since 17 also represents the modified version of 16, we
can expect derivatives of 16 to have a basal inhibition
constant of approximately 2.9 lM. To determine if 16
could be used in solid phase synthesis, we reacted it with
a resin-bound model dipeptide (Fig. 7). This dipeptide
consisted of two lysine (lys) monomers condensed using
standard fmoc peptide chemistry. Compound 16 was
then attached to the orthogonal amines on the side-
chains of both lys groups to generate a bisintercalator
molecule with a lys–lys dipeptide linker between them.
The identity of this molecule was confirmed by mass
spectrometry and is shown in Figure 8.
Acknowledgments
This work was supported by NIH Grant
1RO1CA090410-01A2.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2007.01.070.
References and notes
1. Lichtsteiner, S. P.; Lebkowski, J. S.; Vasserot, A. P. Ann.
N. Y. Acad. Sci. 1999, 886, 1.
2. Bachand, F.; Autexier, C. Mol. Cell Biol. 2001, 21, 1888.
3. Francis, R.; West, C.; Friedman, S. H. Bioorg. Chem.
2001, 29, 107.
4. West, C.; Francis, R.; Friedman, S. H. Bioorg. Med.
Chem. Lett. 2001, 11, 2727.
5. Watkins, T. I.; Woolfe, G. Nature 1952, 169, 506.
6. Watkins, T. I. J. Chem. Soc. 1952, 3059.
7. Firth, J. W.; Watkins, L. C.; Graves, D. E.; Yielding, L.
W. J. Heterocycl. Chem. 1983, 20, 759.
Compound 16 represents an ideal base molecule for
modification in the search for specific telomerase inhib-
itors. It is synthesized in high yield in relatively few
steps. Upon conversion to the amide, it has essentially
an identical IC50 to that of the parent compound. In
addition, it efficiently modifies resin bound amines via

S. Rangarajan, S. H. Friedman / Bioorg. Med. Chem. Lett. 17 (2007) 2267–2273
2273
8. Luedtke, N. W.; Liu, Q.; Tor, Y. Bioorg. Med. Chem.
2003, 11, 5235.
15. Letsinger, L. R.; Schott, M. E. J. Am. Chem. Soc. 1981,
103, 9294.
9. Wakelin, L. P.; Waring, M. J. Mol. Pharmacol. 1974, 10,
544.
16. Yielding, L. W.; Yielding, K. L.; Donoghue, J. E.
Biopolymers 1984, 23, 83.
10. Jacquemin-Sablon, H.; Le Bret, M.; Jacquemin-Sablon,
A.; Paoletti, C. Biochemistry 1979, 18, 128.
11. Loccufier, J.; Schacht, E. Tetrahedron 1989, 45, 3385.
12. Carrasco, C.; Helissey, P.; Haroun, M.; Baldeyrou, B.;
Lansiaux, A.; Colson, P.; Houssier, C.; Giorgi-Renault, S.;
Bailly, C. Chembiochem 2003, 4, 50.
17. Luedtke, N. W.; Hwang, S. J.; Nava, E.; Gut, D.; Kol, M.;
Tor, Y. Nucleic Acids Res. 2003, 31, 5732.
18. Kelley, O. S.; Holmlin, E. R.; Stemp, D. A.; Barton, J. K.
J. Am. Chem. Soc. 1997, 119, 9861.
19. Ross, S. A.; Pitie, M.; Meunier, B. J. Chem. Soc., Perkin
Trans. I 2000, 571.
13. Peytou, V.; Condom, R.; Patino, N.; Guedj, R.; Aubertin,
A. M.; Gelus, N.; Bailly, C.; Terreux, R.; Cabrol-Bass, D.
J. Med. Chem. 1999, 42, 4042.
20. Francis, R.; Friedman, S. H. Anal. Biochem. 2003, 323, 65.
21. Graves, D. E.; Watkins, C. L.; Yielding, L. W. Biochem-
istry 1981, 20, 1887.
14. Ratmeyer, L. S.; Vinayak, R.; Zon, G.; Wilson, W. D.
J. Med. Chem. 1992, 35, 966.
22. Tsai, C. C.; Jain, S. C.; Sobell, H. M. Proc. Natl. Acad.
Sci. U.S.A. 1975, 72, 628.