Facile synthesis of oligopeptide distamycin analogs devoid of hydrogen- bond donors or acceptors at the N-terminus: Sequence-specific duplex DNA binding as a function of peptide chain length

Article abstract of DOI:10.1016/S0040-4039(00)00802-9

The first examples of distamycin analogs, which lack hydrogen bond interactor groups at the N-terminus, have been synthesized. The bispyrrole peptide did not exhibit any detectable binding with double-stranded (ds) DNA. However, all other homologues did bind to ds-DNA strongly, with the binding affinities increasing as a function of the number of repeating pyrrole carboxamide units, implying that a hydrogen bond donor or acceptor atom per se at the N-terminus is not essential for their DNA binding. Studies with poly d(GC) showed that the N-terminal formamide is not a prerequisite for GC binding, contrary to earlier postulations. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00802-9

Tetrahedron Letters 41 (2000) 5571±5575
Facile synthesis of oligopeptide distamycin analogs devoid of
hydrogen bond donors or acceptors at the N-terminus:
sequence-speci®c duplex DNA binding as a function of
peptide chain length
Santanu Bhattacharya* and Mini Thomas
Department of Organic Chemistry, Indian Institute of Science, Bangalore 560 012, India
Received 13 March 2000; revised 2 May 2000; accepted 16 May 2000
Abstract
The ®rst examples of distamycin analogs, which lack hydrogen bond interactor groups at the N-terminus,
have been synthesized. The bispyrrole peptide did not exhibit any detectable binding with double-stranded
(
ds) DNA. However, all other homologues did bind to ds-DNA strongly, with the binding anities
increasing as a function of the number of repeating pyrrole carboxamide units, implying that a hydrogen
bond donor or acceptor atom per se at the N-terminus is not essential for their DNA binding. Studies with
poly d(GC) showed that the N-terminal formamide is not a prerequisite for GC binding, contrary to earlier
postulations. # 2000 Elsevier Science Ltd. All rights reserved.
The design of synthetic organic molecules of low molecular mass that can regulate gene
1
expression in a predictable manner is an area of current research interest. Despite the fact that
there are several minor groove binders such as Hoechst, DAPI, etc. the natural product distamycin
(Dst) remains an ideal candidate for further elaboration of the central theme of designing
synthetic gene-regulators. This is due to the simplicity of its chemical nature, consisting of a hetero-
aromatic polyamide backbone, with a single positive charge emanating from the side chain at the
C-terminus. A number of fundamental questions pertaining to the DNA recognition properties of
Dst derivatives remain unanswered. For instance, is there a minimum size requirement for the
onset of DNA binding? Can Dst analogs bind ds-DNA e€ectively in the absence of an N-terminus
hydrogen bond donor or acceptor? Can one improve the intracellular stability of these com-
pounds when the N-terminal formamide is absent? This is signi®cant on account of the inherent
susceptibility of the terminal amide group of distamycin towards cellular degradation. It is note-
worthy that thioformyl distamycin, an amide isostere of distamycin, was synthesized by Lown et
*
Corresponding author. E-mail: sb@orgchem.iisc.ernet.in
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00802-9

5
572
2
al. on account of the above considerations. Design of appropriate Dst analogs would also be
useful to test the validity of the hypothesis that the N-terminus formamide group is important for
the GC binding mode.3
However, the seemingly simple synthesis of Dst and its derivatives has been complicated by the
low reactivity of the aromatic carboxylic acids and the extremely unstable nature of the amines.4
To circumvent some of these diculties, a solid-phase-based method was recently developed by
5
Dervan et al. toward the synthesis of higher homologues of distamycin. We felt that a viable
solution-phase method would still be quite desirable, as it would be complimentary and would
o€er considerable general applicability, especially when other bio-active moieties such as chemi-
cal nucleases,1 need to be appended. Herein, we present the convenient synthesis and results of
the DNA binding experiments of the ®rst examples of Dst class of minor groove binders that are
devoid of a hydrogen bond donor or an acceptor group at the N-terminus.
c,6
The key starting material for the synthesis, 5e, was obtained by the nitration and subsequent
esteri®cation of commercially available N-methyl pyrrole 2-carboxylic acid, 5c according to a
published procedure.4b This was then reduced using H -Pd/C in DMF to obtain the free amine 5f,
which was coupled with the acid chloride, 5b, in the presence of Et N in DMF to obtain the nitro-
2
3
dipeptide analog 5h in 87% yield (Scheme 1). Further, elongation of the peptide backbone was
achieved by repetition of the reduction and the amide coupling steps. Thus, the nitro derivatives
6
a and 6b were obtained in 83 and 80% yields, respectively, from their precursors as shown in
Scheme 1. The chain growth was terminated upon coupling with N-methyl pyrrole-2-carbonyl
chloride 5d, once the required length of the polyamide chain was achieved. All the nitro com-
pounds 5e, 5h, 6a and 6b were reduced separately and coupled with the acid chloride, 5d, to obtain
5
g, 8a, 8b and 8c respectively, in ca. 80% yields as illustrated in Schemes 1 and 2. It is important
to note that none of the free amines were isolated for characterization, on account of their
unstable nature. They were separated from Pd/C in the reaction mixture by ®ltration and were
coupled with the appropriate acid chlorides directly. We term the latter step as `end-capping' as it
prevents further extension of the amide chain from the N-terminus. The `end-capped' methyl esters
were hydrolyzed to the corresponding acids, which were subsequently converted to the activated
esters using N-hydroxy succinimide (HOSU) and dicyclohexyl carbodiimide (DCC) (Scheme 2).
DCU was removed from the HOSU esters by ®ltration and were subjected to aminolysis with
N,N-dimethyl-1,3-diaminopropane in CHCl to obtain the ®nal compounds, 1±4 in 80, 88, 90 and
3
ꢀ
Scheme 1. Reagents, conditions and yields: (i) SOCl , THF, re¯ux, 1 h; (ii) SOCl , THF, 0 C 1 h, then rt, 30 min; (iii)
2
2
ꢀ
ꢀ
2
H -Pd/C (5%), 1 atm. rt, 12 h; (iv) 5d, Et
3 3
N, ^5 C, 1 h, then 1 h at rt (85%); (v) 5b, Et N, ^5 C, 1 h, then 1 h at rt
(
87%)

5
573
ꢀ
Scheme 2. Reagents, conditions and yields: (i) H -Pd/C(5%), 1 atm. rt, 18 h; (ii) 1d, Et N, ^5 C, 1 h, then 1 h at rt (80,
2
3
ꢀ
7
8 and 85% for 8a±8c, respectively); (iii) 0.25 (N) NaOH, EtOH/H
ꢀ
2
O, re¯ux, 1 h, then 0 C, 0.5 (N) HCl (96, 94, 92 and
0% for 9a±9d, respectively); (iv) DCC, HOSU, DMF, 0 C, 30 min, then rt, 4 h; (v) N,N-dimethyl-1,3-diaminopropane,
9
rt, 2 h (80, 88, 90 and 82% for 1±4, respectively)
7
8
2% yields, respectively. It is noteworthy that the N,N-dimethyl-1,3-diaminopropionamido
group was introduced at the C-terminus only in the last step of the synthesis since the amines
produced at the N-terminus upon reduction from their corresponding nitro derivatives were
relatively more unstable when the N,N-dimethyl-1,3-diaminopionamido group was already present
at the C-terminus. Column chromatography was employed only for the puri®cation of the `end-
capped' methyl esters and the ®nal compounds.
The DNA binding abilities of these oligopeptides were assessed by thermal helix-to-coil `melting'
temperature (T ) measurements (pH=7.4, 10 mM tris bu€er, 20 mM NaCl, D/P=0.5). Compound
m
1
peptide length for the analogs 2±4. The ÁT values of the complexes of 2±4 and poly d (AT) were
did not alter the T of ds-DNA. However, the ÁT increased progressively as a function of the
m m
m
ꢀ
ca. 6.1, 17.2 and 26.6 C, respectively. With CT (calf thymus) DNA, the ÁT were 1.5, 4.5 and
m
8 C, respectively, for 2±4 at the same [D]/[P] ratio suggesting that these compounds have retained
their AT speci®city despite the removal of the N-terminal amide unit. This was further corroborated
ꢀ
8
by the ethidium bromide displacement assay. The apparent binding constants (Kapp) obtained
4
6
6
^1
for 2±4 with CT DNA were ꢁ5.8Â10 , 1Â10 and 4.5Â10 M , respectively, whereas the corre-
5
7
8
^1
sponding values for poly d (AT) were 4.2Â10 , 3.3Â10 and 1.2Â10 M , respectively. The K
app
5
7
^1
values obtained for Dst under the same conditions were 7.7Â10 and 3.5Â10 M for CT DNA
8
and poly d (AT) respectively, suggesting that the compounds 3 and 4 bind to ds-DNA with
greater anity than their natural counterpart, Dst, apart from preserving the sequence-speci®city.

5
574
Experiments with 3/poly d (GC) employing circular dichroism showed intense ICD signals and
thus suggested that the N-terminal formamide is not a prerequisite for GC binding, as was pos-
tulated earlier. This complex was highly electrostatic in nature, as the ICD signal disappeared
completely in ca. 200 mM (NaCl).
In summary, a convenient, general and adaptable method has been developed for the synthesis
of four Dst analogs, which form the ®rst examples of Dst class of compounds that bear neither a
hydrogen bond donor nor an acceptor at the N-terminus. Our methodology o€ers several
advantages over the existing procedures. First, the total number of steps in the synthetic scheme is
kept at a minimum, as each intermediate nitro compound is the precursor for the next higher
analog. Secondly one of the least yielding steps in the synthesis of Dst, viz. the introduction of the
terminal formamide group has been completely avoided. Thirdly, this modi®cation, along with
the introduction of the C-terminus side chain only in the last step, permit the synthetic procedure
to have enough ¯exibility such that it would accommodate the introduction of natural amino
acids, etc. in the sequence, if desired, for further modulation of the DNA binding properties. It is
also demonstrated that the presence of a hydrogen bond donor or acceptor at the N-terminus per se,
as in the case of Dst and imidazole/pyrrole oligopeptide analogs, respectively, is not a prerequisite
for the maintenance of DNA binding. The minimum number of pyrrole carboxamide units for
the onset of DNA binding, in the absence of the N-terminus amide is shown to be three. It is also
shown that the N-terminus formamide is not a prerequisite for GC binding. Work is now
underway in our laboratory toward the synthesis of novel minor groove binders with greater
anity and longer sequence recognition properties.
Acknowledgements
This work was supported by the Swarnajayanti Fellowship Grant of the Department of Science
and Technology, Government of India.
References
1
. (a) See, for reviews: Dervan, P. B. Science 1986, 232, 464. (b) Geierstanger, B. H.; Wemmer, D. E. Annu. Rev.
Biophys. Biomol. Struct. 1995, 24, 463. (c) Bailly, C.; Chaires, J. B. Bioconjugate Chem. 1998, 9, 513. (d) Iida, H.;
Jia, G.; Lown, J. W. Curr. Opin. Biotechnology. 1999, 10, 29.
2
. Singh, M. P.; Kumar, S.; Joseph, T.; Pon, R. T.; Lown, J. W. Biochemistry 1992, 31, 6453, and references cited
therein.
3
4
. Luck, G.; Zimmer, C.; Rinert, K. E.; Arcamone, F. Nucleic Acids Res. 1977, 4, 2655.
. (a) Penco, S.; Redaelli, S.; Arcamone, F. Gazz. Chim. Ital. 1967, 97, 1110. (b) Bialer, M.; Yagen, B.; Mecholum, R.
Tetrahedron 1978, 34, 2389. (c) Lown, J. W.; Krowicki, K. J. Org. Chem. 1985, 50, 3774. (d) Grehn, L.;
Ragnarsson, U. J. Org. Chem. 1981, 46, 3492. (e) Rao, K. E.; Bathni, Y.; Lown, J. W. J. Org. Chem. 1990, 55,
7
28. (f) Ding, L.; Grehn, L.; De Clercq, E.; Andrei, G.; Snoeck, R.; Balzarini, J.; Fransson, B.; Ragnarson, U.
Acta Chem. Scand. 1994, 48, 495. (f) Wade, W. S.; Mrksich, M.; Dervan, P. B. J. Am. Chem. Soc. 1992, 114, 8783.
. Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6141.
. (a) Bhattacharya, S.; Mandal, S. S. J. Chem. Soc., Chem. Commun. 1995, 2489. (b) Bhattacharya, S.; Mandal, S. S.
J. Chem. Soc., Chem. Commun. 1996, 1515.
5
6
7
. All the intermediates and ®nal compounds were characterized by IR, NMR and electrospray mass spectrometry.
1
Selected spectral data for compounds 1±4. Compound 1: H NMR (300 MHz, CDCl
3
) ꢀ (ppm) 1.67±1.75 (m,
=4 Hz,
J =3Hz), 6.44 (d, 1H, J=2 Hz), 6.70 (dd, 1H, J =4 Hz, J =3 Hz), 6.75 (t, J=2 Hz, 1H), 7.16 (d, 1H, J=2 Hz),
2
H), 2.27 (s, 6H), 2.42 (t, 2H, J=6 Hz), 3.4±3.46 (m, 2H), 3.84 (s, 3H), 3.91 (s, 3H), 6.11 (dd, 1H, J
1
2
1
2

5
575
+
1
7.55 (bs, 1H), 7.70 (s, 1H). LRMS (EI; m/z) calcd for C17
25 5 2
H N O (M ): 331; found: 331. Compound 2: H NMR
(
300 MHz, CDCl ) ꢀ (ppm) 1.71±1.77 (m, 3H), 2.31 (s, 6H), 2.44 (t, 2H, J= 6 Hz), 3.43±3.49 (m, 2H), 3.92 (s, 3H),
3
3
Hz), 6.73 (d, J=2 Hz), 6.77 (t, 1H, J=2 Hz), 7.09 (d, 1H, J=2 Hz), 7.18 (d, 1H, J=2 Hz), 7.52 (s, 1H), 7.62 (s,
1 2 1 2
.94 (s, 3H), 3.99 (s, 3H), 6.14 (dd, 1H, J =4 Hz, J =3 Hz), 6.41 (d, 1H, J=2 Hz), 6.67 (dd, 1H, J =4 Hz, J =3
+
1
1
H), 7.74 (bs, 1H). ESI-MS; calcd for C H N O (MH ): 454.5; found: 454.5. Compound 3: H NMR (300
23 32 7 3
MHz, CDCl
3
) ꢀ (ppm) 1.68±1.74 (m, 2H), 2.26 (s, 6H), 2.42 (t, 2H, J=6 Hz), 3.40±3.43 (m, 2H), 3.86 (s, 3H), 3.87
=4 Hz, J =3 Hz), 6.54 (d, 1H, J=2 Hz), 6.60 (d, 1H, J=2 Hz),
.73 (d, 1H, J=2 Hz), 6.75±6.76 (m, 2H), 7.12 (t, 2H, J=2 Hz), 7.2 (d, 1H, J=2 Hz), 7.68 (s, 1H), 7.70 (bs, 1H),
(
6
8
s, 3H), 3.90 (s, 3H), 4.0 (s, 3H), 6.11 (dd, 1H, J
1
2
+
1
.05 (bs, 2H). ESI-MS calcd for C H N O (MH ): 576.3; found: 576.3. Compound 4: H NMR (500 MHz,
29 38 9 4
3
CDCl ) ꢀ (ppm) 1.66±1.69 (m, 2H), 2.20 (s, 6H), 2.36 (t, 2H, J=6 Hz), 3.37±3.39 (m, 2H), 3.77 (s, 6H), 3.81 (s,
3
1
1
H), 3.83 (s, 3H), 3.93 (s, 3H), 6.06 (bs, 1H), 6.57 (bs, 1H), 6.61 (bs, 1H), 6.65 (bs, 1H), 6.73 (bs, 1H), 6.74 (bs,
H), 6.78 (bs, 1H), 7.04 (bs, 1H), 7.13 (bs, 1H), 7.15 (bs, 1H), 7.18 (bs, 1H), 7.61 (bs, 1H), 8.09 (s, 1H), 8.18 (s,
+
44 11 5
H), 8.49 (bs, 2H). ESI-MS calcd for C35H N O (MH ): 698.4; found: 698.4.
8
. Lee, M.; Rhodes, A. L.; Wyatt, M. D.; Forrow, S.; Hartley, J. A. Biochemistry 1993, 32, 4237.