Synthesis and DNA binding properties of saturated distamycin analogues

Article abstract of DOI:10.1016/S0960-894X(02)00467-5

A series of saturated heterocyclic analogues of distamycin were prepared and examined. A fluorescent intercalator displacement (FID) assay conducted on p[dA]-p[dT] DNA to obtain C₅₀ values and a hairpin deoxyoligonucleotide containing an A/T-rich binding site was used to evaluate DNA binding affinity. It is observed that saturated heterocycles greatly reduce the DNA binding relative to distamycin.

Full text of DOI:10.1016/S0960-894X(02)00467-5

Bioorganic & Medicinal Chemistry Letters 12 (2002) 2647–2650
Synthesis and DNA Binding Properties of Saturated Distamycin
Analogues
Craig R. Woods,a Nicolas Faucher,a Bernd Eschgfaller,b Kenneth W. Bairb,*
and Dale L. Bogera,*
aDepartment of Chemistry and the Skaggs Institute for Chemical Biology, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, CA 92037, USA
bNovartis Institute for Biomedical Research, Novartis Pharmaceuticals Corporation—Oncology, 556 Morris Avenue, Summit,
NJ 07901-1398, USA
Received 8 April 2002; accepted 28 May 2002
Abstract—Aseries of saturated heterocyclic analogues of distamycin were prepared and examined. Afluorescent intercalator dis-
placement (FID) assay conducted on p[dA]–p[dT] DNA to obtain C50 values and a hairpin deoxyoligonucleotide containing an A/T-
rich binding site was used to evaluate DNAbinding affinity. It is observed that saturated heterocycles greatly reduce the DNAbinding
relative to distamycin. # 2002 Elsevier Science Ltd. All rights reserved.
Polyamides composed of N-methylpyrrole (Py),
N-methylimidazole (Im), and a growing set of structural
analogues bind in the DNAminor groove with pre-
dictable sequence selectivity and high affinities.1 Several
factors contribute to the success of polyamide binding
including hydrogen bonding, curvature, Van der Waals
contacts, and charged end groups. Each component has
been explored through the preparation of analogous
systems designed to probe the magnitude, requirements,
and relative importance of each feature. Yet, the bind-
ing effectiveness for polyamides lacking the p-system of
the integral heteroaromatic rings has been relatively
unexplored. Herein we report the synthesis and evalua-
tion of a series of saturated cyclic polyamides 1–5 (Fig. 1),
based on the known, effective DNAminor groove
binding agent distamycin.2 Saturation removes the
p-system, alters the conformation, and increases the
thickness of the compound, yet maintains the relative
structure of distamycin and was used to assess the
importance of the p-system. Afluorescent intercalator
displacement (FID) assay conducted on p[dA]–p[dT]
ethylcarbodiimide hydrochloride (EDCI) mediated cou-
pling reactions as described previously3 where workup,
isolation, and purification could be addressed princi-
pally by liquid–liquid acid–base extractions. Each series
of analogues (1–5) are discussed separately below.
N-CBz pyrrolidine based system (1a and b). N-Boc
deprotection of 64 (HCl–EtOAc) provided the amine 7
as the hydrochloride salt and methyl ester hydrolysis of
64(LiOH, THF–MeOH–H2O) provided carboxylic acid
8 (Scheme 1).
After deprotection, 7 and 8 were coupled to produce
dimer 95 (EDCI, HOBt, DMF, 25 ꢀC, 13–20 h, 66%)
DNAto obtain
C50 values and a hairpin deox-
yoligonucleotide containing an A/T-rich binding site
was used to evaluate DNAtheir binding affinity.
Asolution-phase synthesis of the polyamides was con-
ducted using a series of 1-(3-dimethylaminopropyl)-3-
*Corresponding author. Tel.: +1-858-784-7522; fax: +1-858-784-
7550; e-mail: boger@scripps.edu
Figure 1. Distamycin and saturated heterocyclic polyamides 1–5.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00467-5

2648
C. R. Woods et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2647–2650
before efficient conditions were found (H2, 10% Pd/C,
10 equiv TFA, MeOH, 25 ꢀC, 4 h, entry 6). Thus,
although N-CBz deprotection was slow and typically
problematic, acid catalysis with inclusion of TFApro-
vided 2 in good conversions.
Furan based system (3a and b). Transformation of the
monomer 124 and its enantiomer into their respective
functionalized trimers 3a9 and b12 was performed using the
strategy detailed above for 1a and b (Schemes 1 and 2).
However, the tetrahydrofuran-based compounds
proved highly water soluble. Thus, removal of excess
starting materials, reagents, and side products by
simple acid–base extraction was not possible and
chromatographic purifications were required.
Scheme 1.
(Scheme 2). N-Boc deprotection of 9 (HCl–EtOAc, 25ꢀC,
1 h) and coupling with 8 (EDCI, HOAt, DMF, 25ꢀC,
20h, 73%) provided trimer 10.6 The methyl ester of 10
was then hydrolyzed (2.5 equiv NaOH, THF–H2O, 1.5 h,
91%) to give the free carboxylic acid which was coupled
with 3-aminopropionitrile to afford 117 (EDCI, HOAt,
DMF, 25ꢀC, 20h, 75%). End group functionalization of
11 was accomplished by simultaneous N-Boc deprotec-
tion and imidate formation (HCl–EtOH), amidine for-
mation (NH3–EtOH, 25 ꢀC, 2–24 h), and N-formylation
(N-formyl imidazole, THF–MeOH, ꢁ40ꢀC, 2 h) without
purification of the intermediates yielding 1a8 in 56% over
the three steps. The enantiomer 1b8 was produced in the
same manner starting from the enantiomer of 6.4
Cyclopentane based system (4a). Transformation of the
monomer 184 into 4a was performed uneventfully using
the strategy detailed above (Schemes 1 and 2).
N-Me pyrrolidine based system (5a). N-Boc deprotec-
tion (HCl–dioxane) of 244 provided the amine 25 as the
hydrochloride salt and methyl ester hydrolysis of 244
(LiOH, THF–MeOH) provided the lithium carboxylate
2613 (Scheme 3).
Pyrrolidine based system (2a and b). Several hydro-
genolysis conditions for N-CBz deprotection of 1a and b
were examined to produce 2a and b (Table 1, entries 1–5)
Following the strategy detailed above, the lithium salt
26 was used to provide trimer 28 (Scheme 4). However,
coupling of the carboxylic acid derived from 28 with
3-aminopropionitrile afforded 29 in only moderate yield
Table 1. Hydrogenolysis conditions for 1a using 10% Pd/C
Table 2. C50 values
Compd
C50 (mM)
Distamycin
0.46
4.70
5.10
6.37
8.55
5.40
5.50
7.91
6.70
1a
1b
2a
2b
3a
3b
4a
5a
Entry
Conditions
Result
1
2
3
4
5
6
H2, MeOH, 25 ꢀC, 24–48 h
H2, MeOH, 2 N HCl, 25 ꢀC, 24 h10
4% HCOOH, MeOH, 25 ꢀC, 14 h11
H2, acetone–MeOH, 25 ꢀC, 24 h12
H2, 2 equiv HOAc, MeOH, 25 ꢀC, 24 h
H2, 10 equiv TFA, MeOH, 25 ꢀC, 4 h
No reaction
Decomposition
Decomposition
No reaction
No reaction
2a, 64%
p[dA]–p[dT] 8.8 mM bp, EtBr 4.4 mM.
Scheme 2.

C. R. Woods et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2647–2650
2649
(20%, Scheme 2). Consequently, an alternative con-
vergent synthesis was used to provide 29 (Scheme 4).
Coupling 26 with 3-aminopropionitrile (HATU,
i-Pr2NEt, DMF, 25 ꢀC, 80%) yielded nitrile monomer
30, which was subsequently N-Boc deprotected (HCl–
EtOAc) and coupled with the carboxylic acid derived
from 27 (LiOH, THF–MeOH–H2O) using standard
coupling conditions (HATU, i-Pr2NEt, DMF, 25 ꢀC,
60%) to provide 29 in a more satisfactory yield. Treat-
ment of 29 with freshly prepared 8 N HCl–EtOH led to
N-Boc deprotection and imidate formation. Subsequent
treatment of the imidate with NH3–EtOH provided
the corresponding amidine which was formylated using
N-formyl imidazole freshly prepared by treating
carbonyldiimidazole with formic acid to provide 5a.14
flat fluorescence reduction curves too shallow for accu-
rate Scatchard analysis.16 Qualitatively, their behavior
can be compared using the ratio of fluorescence present
before (F100%) and after the addition of one equivalent
of agent (F1),17 using the F1/F100% value (Table 3). With
distamycin, a F1/F100% of 0.48 is observed, whereas
much higher values are observed with the saturated
systems (0.77–0.92). The trends for hairpin and p[dA]–
p[dT] DNAbinding (F /F100% versus C50) are similar
1
with the N-CBz derivatives (1a and b) exhibiting the
better binding properties than the remaining analogues.
However, additional insights are difficult to extrapolate
due to the low binding ability.
Discussion
Initial assessment of DNA binding. The compounds 1–5
were initially screened for DNAbinding using p[dA]–
p[dT] in a fluorescence intercalator displacement (FID)
assay.15 The relative binding affinities were established
by monitoring the loss of fluorescence derived from
titration displacement of prebound ethidium bromide
from p[dA]–d[dT]. A comparison the compound con-
centration required for 50% displacement of the ethi-
dium bromide (C50) revealed that none of the saturated
analogues compared well to distamycin (Table 2). The
best of the analogues, N-CBz pyrrolidine 1a, required a
10-fold greater concentration relative to distamycin
(4.70 vs 0.46 mM), while the remaining analogues
required even higher concentrations (5.10–8.55 mM).
Interestingly, one stereochemical configuration, (R)C-
ring(S)N-ring (the a analogues), consistently performed
better than their respective enantiomers.
One major impact of saturating the distamycin pyrrole
rings with 2 or 5 is the introduction of three additional
basic nitrogens. Those of 2 are present as secondary
amines while those of 5 are tertiary amines. Although
not observed, one might have suspected that pH 7 pro-
tonation of these amines might provide positively
charged centers that would benefit from stabilizing
electrostatic interactions with DNAenhancing their
affinity. In addition, the comparable behavior of 2 and 5
with either 3 or 4, suggest that the presence of the basic
amines do not enhance or inhibit binding relative to the
impact of introduction of the saturated versus unsaturated
Table 3. F1/F100% values of distamycin agents 1–5 with hairpin DNA
Binding to a hairpin deoxyoligonucleotide containing an
A/T-rich binding site. In an attempt to assess binding
constant comparisons of 1–5, a FID titration assay was
performed using a DNAhairpin with an A/T-rich
binding site of 5 bp. As above, the analogues exhibited
low binding affinity toward the hairpin DNA, yielding
a
Compd
F1 / F100%
Curve description
Distamycin
0.48
0.88
0.77
0.90
0.88
0.90
0.90
0.92
0.83
Good
Flat
Flat
Flat
Flat
Flat
Flat
Flat
Flat
1a
1b
2a
2b
3a
3b
4a
5a
aF1/F100% is the ratio of fluorescence before (F100%) and after the
addition of 1 equivalent of agent (F1).
Scheme 3.
Scheme 4.

2650
C. R. Woods et al. / Bioorg. Med. Chem. Lett. 12 (2002) 2647–2650
1
heterocycles. Interestingly, the most effective analogue
in the series is 1 bearing the hydrophobic N-CBz pro-
tecting group. However, even 1 would not be considered
an effective DNAbinding agent. Finally, the all carbon
cyclopentyl-based analogue 4 emerged in the DNAbind-
ing assays as the poorest analogue in the series.
6. For 10a, H NMR (acetone-d6, 400 MHz) d 7.93 (m, 2H),
7.35–7.29 (m, 15H), 6.39 (m, 1H), 5.11–4.95 (m, 6H), 4.49–
4.23 (m, 6H), 3.72 and 3.60 (s, 3H), 3.72 and 3.37 (m, 6H),
2.18 (m, 6H), 1.38 (s, 9H); IR (NaCl, film) ꢀmax 3311, 2933,
1704, 1537, 1420, 1356, 1169 cmꢁ1; MALDI–HRMS (DHB)
m/z 893.3688 (M+Na+, C45H54N6O12 requires 893.3692);
25
25
½ꢁꢂD +24 (c 1, THF). For 10b, ½ꢁꢂD ꢁ17 (c 1, THF).
1
7. For 11a, H NMR (acetone-d6, 400 MHz) d 7.76 (m, 1H),
Although several features imparted by the structural
changes could be responsible for the poor binding of 1–
5, the magnitude of the reductions is notable. From
molecular modeling studies, a minor groove com-
plementary shape, hydrogen bonding, and stabilizing
electrostatic interactions analogous to those observed
with distamycin are possible with 1–5 and the hydro-
phobic character covered by the series brackets that of
distamycin. Nonetheless, no saturated analogue in the ser-
ies came close to the DNAbinding affinity observed with
distamycin illustrating the importance of the unsaturated
heterocycles.18
7.58 (m, 2H), 7.35–7.30 (m, 15H), 6.32 (brs, 1H), 5.14–4.92 (m,
6H), 4.4–4.3 (m, 6H), 3.73 and 3.42 (m, 8H), 2.54 and 2.56 (m,
2H), 2.20 (m, 6H), 1.38 (s, 9H); IR (NaCl, film) ꢀmax 3309,
2964, 2340, 1604, 1539, 1420, 1357, 1169 cmꢁ1; MALDI–
HRMS (DHB) m/z 931.3929 (M+Na+, C47H56N8O11
25
25
requires 931.3961); ½ꢁꢂD +29 (c 0.1, THF). For 11b, ½ꢁꢂD ꢁ29
(c 0.1, THF).
8. For 1a, 1H NMR (methanol-d4, 400 MHz) d 8.31 (brs, 2H),
8.02 (brs, 1H), 7.33 (m, 15H), 5.12–4.88 (m, 6H), 4.5–4.3 (m,
6H), 3.81 (m, 3H), 3.43 (m, 5H), 2.67 (m, 2H), 2.19 (m, 6H):
IR (NaCl, film) ꢀmax 3243, 1662, 1542, 1542, 1356, 1126 cmꢁ1
;
MALDI–HRMS
(DHB)
m/z
854.3834
(M+H+,
25
C43H51N9O10 requires 854.3831); ½ꢁꢂD +26 (c 0.1, MeOH).
25
For 1b, ½ꢁꢂD ꢁ27 (c 0.1, MeOH).
1
9. For 3a, H NMR (methanol-d4, 500 MHz) d 8.02 (s, 1H),
4.47 (m, 6H), 4.11 (m, 3H), 3.76 (m, 3H), 3.47 (m, 2H), 2.69
Acknowledgements
(m, 2H), 2.31 (m, 6H); IR (NaCl, film) ꢀmax 3128, 2851, 1668,
1630, 1085 cmꢁ1
; MALDI–HRMS (DHB) m/z 455.2244
We gratefully acknowledge the financial support of the
National Institutes of Health (CA41986, CA78045),
Novartis, and the Department of Army (DAMD17–02–
1-0135).
25
(M+H+, C19H30N6O7 requires 455.2249); ½ꢁꢂD ꢁ25 (c 0.1,
25
MeOH). For 3b, ½ꢁꢂD +22 (c 0.1, MeOH).
10. Oleksyszyn, J.; Boduszek, B.; Chih-Min, K.; Powers, J. C.
J. Med. Chem. 1994, 37, 226.
11. ElAmin, B.; Anantharamaiah, G. M.; Royer, G. P.;
Means, G. E. J. Org. Chem. 1979, 44, 3442.
12. Griffin, K. G.; Hawker, S.; Bhatti, M. A. Chem. Ind. 1996,
68, 325.
References and Notes
13. The lithium salt 26 was used instead of the carboxylic acid
to avoid potential coupling problems with the zwitterionic
form of the amino acid.
14. For 5a, H NMR (methanol-d4, 400 MHz) d 8.01 (s, 1H),
4.37 and 3.99 (m, 6H), 3.56 and 3.43 (m, 3H), 3.18 (m, 3H),
1. (a) Dervan, P. B. Bioorg. Med. Chem. 2001, 9, 2215. (b)
Neidle, S. Nat. Prod. Rep. 2001, 18, 291.
2. Johnson, D. S.; Boger, D. L. In DNA Binding Agents;
Murakami, Y., Ed.; Comprehensive Supramolecular Chem-
istry; Elsevier Science: New York, 1996; Vol. 4 p 81.
3. Boger, D. L.; Fink, B. E.; Hedrick, M. P. J. Am. Chem.
Soc. 2000, 122, 6382.
1
2.81 and 2.68 (m, 2H), 2.40 (m, 11H), 2.14 (m, 6H); IR (NaCl,
film) ꢀmax 2939, 2852, 1643, 1659, 1580, 1256, 1107 cmꢁ1
;
MALDI–HRMS
(DHB)
m/z
516.3022
(M+Na+,
4. Experimental details and characterization for intermediates
and final products can be provided upon request.
25
C22H39N9O4 requires 516.3017); ½ꢁꢂD +88 (c 0.1, MeOH).
5. For 9a, 1H NMR (acetone-d6, 400 MHz) d 7.66 (m, 1H),
7.40–7.30 (m, 10H), 6.34 (brs, 1H), 5.13–4.95 (m, 4H), 4.47–
4.24 (m, 4H), 3.79 and 3.39 (s, 3H), 3.75 and 3.38 (m, 4H),
2.20 (m, 4H), 1.38 (s, 9H); IR (NaCl, film) ꢀmax 3315, 2975,
1700, 1644, 1531, 1417, 1356, 1169 cmꢁ1; MALDI–HRMS
(DHB) m/z 647.2686 (M+Na+, C32H40N4O9 requires
15. Boger, D. L.; Fink, B. E.; Brunette, S. R.; Tse, W. C.;
Hedrick, M. P. J. Am. Chem. Soc. 2001, 123, 5878.
16. Scatchard, G. N.Y. Acad. Sci. 1949, 51, 660.
17. Addition of one equivalent of agent from the titration
curves was chosen arbitrarily for the comparisons.
18. 1a, 1b, 3a, and 3b were inactive in an L1210 cytotoxic
assay (IC50 >10 mM).
25
25
647.2687); ½ꢁꢂD +23 (c 1, THF). For 9b, ½ꢁꢂD ꢁ22 (c 1, THF).