Imidazopyridine/pyrrole and hydroxybenzimidazole/pyrrole pairs for DNA minor groove recognition

Article abstract of DOI:10.1021/ja0300158

The DNA binding properties of fused heterocycles imidazo[4,5-b]pyridine (Ip) and hydroxybenzimidazole (Hz) paired with pyrrole (Py) in eight-ring hairpin polyamides are reported. The recognition profile of Ip/Py and Hz/Py pairs were compared to the five-m

Full text of DOI:10.1021/ja0300158

Published on Web 04/16/2003
Imidazopyridine/Pyrrole and Hydroxybenzimidazole/Pyrrole
Pairs for DNA Minor Groove Recognition
Dorte Renneberg and Peter B. Dervan*
Contribution from the DiVision of Chemistry and Chemical Engineering, California Institute of
Technology, Pasadena, California 91125
Received January 7, 2003; E-mail: dervan@caltech.edu.
Abstract: The DNA binding properties of fused heterocycles imidazo[4,5-b]pyridine (Ip) and hydroxyben-
zimidazole (Hz) paired with pyrrole (Py) in eight-ring hairpin polyamides are reported. The recognition profile
of Ip/Py and Hz/Py pairs were compared to the five-membered ring pairs Im/Py and Hp/Py on a DNA
restriction fragment at four 6-base pair recognition sites which vary at a single position 5′-TGTNTA-3′,
where N ) G, C, T, A. The Ip/Py pair distinguishes G‚C from C‚G, T‚A, and A‚T, and the Hz/Py pair
distinguishes T‚A from A‚T, G‚C, and C‚G, affording a new set of heterocycle pairs to target the four Watson-
Crick base pairs in the minor groove of DNA.
Introduction
metrical pairs of five-membered rings appear to be the best
solution such that Im/Py is specific for G‚C and Hp/Py for
Many diseases are related to aberrant gene expression and
the ability to reprogram transcription in a cell by chemical
methods could be important in biology and human medicine.1
Minor-groove-binding polyamides which bind predetermined
DNA sequences offers one approach to artificial gene regula-
tion.2 These molecules are based on analogues of the pyrrole
ring Py of the natural products netropsin and distamycin A
which bind in the minor groove of DNA.3 Base-pair specificity
can be altered by changing the functional group(s) presented to
the floor of the DNA minor groove.4 Stabilizing and destabiliz-
ing interactions with the different edges of the four Watson-
Crick base pairs are modulated by specific hydrogen-bonds and
shape complementarity. For example, the imidazole residue Im
presents the DNA with the N-atom and its lone pair electrons
which can accept a H-bond from the exocyclic NH2 of G.
Additionally, the 3-hydroxypyrrole residue Hp presents an OH
group that is sterically accommodated opposite T not A and, in
addition, can donate H-bonds to the O(2) of thymine. For
discrimination of each of the Watson-Crick base pairs, unsym-
T‚A.4
The pairing rules have proven useful for the recognition of
hundreds of DNA sequences by polyamides. However, sequence-
dependent DNA structural variation makes binding affinity and
specificity at many DNA sequences unpredictable, which leads
us to continue our search for new heterocycles of slightly dif-
ferent shape, curvature and twist for minor-groove recognition.5-7
Importantly, we find that polyamides containing Hp residues
can degrade over time in the presence of acid or free radicals,
and a robust replacement for use in biological studies is
desirable. Several five-membered heterocyclic residues other
than Py, Im, and Hp have been investigated, including furan
Fr, thiazole Nt, pyrazole Pz, and 3-hydroxythiophene Ht with
no new specificity uncovered.6d In retrospect, it is remarkable
that a search of new five-membered heterocycles (and new
heterocycle pairs) reveals few new leads for sequence discrimi-
nation. This implies that the solution to DNA recognition by
Im/Py, Py/Im, Hp/Py, and Py/Hp could be a narrow structural
window. We attempt to broaden the repertoire of aromatic
heterocycles for DNA recognition by exploring a family of
benzimidazoles with different atomic substitutions (CH, N, OH)
on the six-membered ring directed toward the floor of the DNA
minor groove. Curvature effects are amplified in contiguous-
ring polyamides where π-conjugation limits conformational
flexibility. The pyrrole ring appears slightly over-curved with
(1) (a) Darnell, J. E. Nature ReV. 2002, 2, 740-748. (b) Pandolfi, P. P.
Oncogene 2001, 20, 3116-3127.
(2) (a) Gottesfeld, J. M.; Neely, L.; Trauger, J. W.; Baird, E. E.; Dervan, P. B.
Nature 1997, 387, 202-205. (b) Dickinson, L. A.; Gulizia, R. J.; Trauger,
J. W.; Baird, E. E.; Mosier, D. E.; Gottesfeld, J. M.; Dervan, P. B. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 12 890-12 895. (c) Janssen, S.; Durussel,
T.; Laemmli, U. K. Mol. Cell. 2000, 6, 999-1011. (d) Ansari, A. Z.; Mapp,
A. K.; Nguyen, D. H.; Dervan, P. B.; Ptashne, M. Chem. Biol. 2001, 8,
583-592. (e) Coull, J. J.; He, G.; Melander, C.; Rucker, V. C.; Dervan, P.
B.; Margolis, D. M. J. Virology 2002, 76, 12 349-12 354.
(5) Ellervik, U.; Wang, C. C. C.; Dervan, P. B. J. Am. Chem. Soc. 2000, 122,
9354-9360.
(3) (a) Dervan, P. B.; Bu¨rli, R. W. Curr. Opin. Chem. Biol. 1999, 3, 688-
693. (b) Wemmer, D. E. Biopolymers 1999, 52, 197-211. (c) Dervan, P.
B. Bioorg. Med. Chem. 2001, 9, 2215-2235.
(4) (a) White, S.; Szewczyk, J. W.; Turner, J. M.; Baird, E. E.; Dervan, P. B.
Nature 1998, 391, 468-471. (b) Kielkopf, C. L.; White, S.; Szewczyk, J.
W.; Turner, J. M.; Baird, E. E.; Dervan, P. B.; Rees, D. C. Science 1998,
282, 111-115. (c) Kielkopf, C. L.; Bremer, R. E.; White, S.; Szewczyk, J.
W.; Turner, J. M.; Baird, E. E.; Dervan, P. B.; Rees, D. C. J. Mol. Biol.
2000, 295, 557-567.
(6) (a) Zhan, Z. Y. J.; Dervan, P. B. Bioorg. Med. Chem. 2000, 8, 2467-
2474. (b) Nguyen, D. H.; Szewczyk, J. W.; Baird, E. E.; Dervan, P. B.
Bioorg. Med. Chem. 2001, 9, 7-17. (c) O’Hare, C. C.; Mack, D.; Tandon,
M.; Sharma, S. K.; Lown, J. W.; Kopka, M. L.; Dickerson, R. E.; Hartley,
J. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 72-77. (d) Marques, M. A.;
Doss, R. M.; Urbach, A. R.; Dervan, P. B. HelV. Chim. Acta 2002, in press.
(7) Kielkopf, C. L.; Baird, E. E.; Dervan, P. B.; Rees, D. C. Nature Struct.
Biol. 1998, 5, 104-109.
9
10.1021/ja0300158 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 5707-5716
5707

A R T I C L E S
Renneberg and Dervan
polyamides. In a formal sense, analogues of benzimidazole-
pyrrole keep the same atomic connectivity along the inside
recognition edge as their parent five-membered ring systems
(Figure 1). It is anticipated that the NH proton N(3)-H of the
fused imidazole unit can form H-bonds with lone pairs of N(3)
of purines and O(2) of pyrimidines. The key unknown issue is
whether the pyridine nitrogen on the inside edge of the imidazo-
[4,5-b]pyridine building block is capable of forming a sequence-
specific H-bond to the exocyclic NH2 of G. Specifically, does
the Ip/Py pair mimic an Im/Py pair by distinguishing G‚C from
C‚G? In the case of the hydroxybenzimidazole Hz, the question
arises whether the hydroxyl group on the inside edge of the
ligand can form a H-bond with the O(2) of thymine. Does the
Hz/Py pair mimic Hp/Py and distinguish T‚A from A‚T? A
schematic drawing of the polyamide-DNA complexes including
the postulated hydrogen bonds is illustrated in Figure 2.
For the present study, six eight-ring polyamides, ImPyIpPy-
(R)H2Nγ-PyPyPyPy-CONHMe (1), ImPyHzPy-(R)H2Nγ-PyPy-
PyPy-CONHMe (2), and ImPyBiPy-(R)H2Nγ-PyPyPyPy-CON-
HMe (3) as well as the corresponding all five-membered ring
hairpins ImPyImPy-(R)H2Nγ-PyPyPyPy-CONHMe (4), ImPy-
HpPy-(R)H2Nγ-PyPyPyPy-CONHMe (5), and ImPyPyPy-(R)H-
2Nγ-PyPyPyPy-CONHMe (6) were synthesized (Figure 3). The
plasmid pDR1 was designed containing four 6-base pair binding
sites 5′-TGTNTA-3′ (N ) G, C, T, A) which differ by a single
internal binding position allowing us to determine their match
and single base pair mismatch DNA binding properties for the
four Watson-Crick base pairs in the minor groove of DNA
(Figure 4).
Figure 1. Structures of the (a) imidazo[4,5-b]pyridine,(b) hydroxybenz-
imidazole, and (c) benzimidazole building blocks in comparison with their
respective five-membered ring systems. H-bonding surfaces along the
recognition sites are highlighted.
respect to the DNA helix, likely leading to less optimal contact.7
We would anticipate that polyamides with benzimidazole
substitutions would have a smaller degree of curvature compared
with the five-membered rings8-11 which might improve the
polyamide-DNA minor groove fit. Recently, we found that the
incorporation of a benzimidazole (Bi) into an eight-ring hairpin
polyamide leads to minor groove binding ligands with no loss
in affinity when paired opposite a five-membered ring Im and
that the Im/Bi pair specifies for G‚C similar to the Im/Py pair.11
We report here the synthesis of two benzimidazole analogues,
imidazo[4,5-b]pyridine (Ip), hydroxybenzimidazole (Hz) amino
acids and their incorporation into eight-ring hairpin polyamides
by solid-phase methods.12 Their DNA binding affinities and
relative sequence specificities were analyzed by quantitative
DNase I footprint titration on a DNA restriction fragment and
compared to the well-characterized Im, Hp and Py hairpin
Results and Discussion
Monomer Syntheses. For the synthesis of the N-Boc-
protected imidazo[4,5-b]pyridine-pyrrole building block (Boc-
PyIp-OH, 15) we chose the commercially available 2,6-dichloro-
Figure 2. Postulated hydrogen-bonding models of the 1:1 polyamide-DNA complexes formed between the imidazo[4,5-b]pyridine-containing polyamide
ImPyIpPy-(R)H2Nγ-PyPyPyPyCONHMe (1), the hydroxybenzimidazole-containing polyamide ImPyHzPy-(R)H2Nγ-PyPyPyPyCONHMe (3), the benzimidazole-
containing polyamide ImPyBiPy-(R)H2Nγ-PyPyPyPyCONHMe (5) as well as their parent five-membered ring-containing polyamides ImPyImPy-(R)H2Nγ-
PyPyPyPyCONHMe (2), ImPyHpPy-(R)H2Nγ-PyPyPyPyCONHMe (4), and ImPyPyPy-(R)H2Nγ-PyPyPyPyCONHMe (6) and their respective match sequences
5′-TGTGTA-3′ (for 1 and 2), 5′-TGTTTA-3′ (for 3 and 4), and 5′-TGTATA-3′ (for 5 and 6), respectively. Circles with two dots represent the lone pairs of
N(3) of purines and O(2) of pyrimidines. Circles containing an “H” represent the N(2) hydrogen of G. Putative hydrogen bonds are illustrated by dotted
lines. The incorporated imidazo[4,5-b]pyridine-/imidazole-pyrrole, hydroxybenzimidazole-/hydroxypyrrole-pyrrole, and benzimidazole-/pyrrole-pyrrole
moieties are shown in bold. Below each hydrogen-bonding model are the ball-and-stick representations of the polyamides.
9
5708 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

DNA Minor Groove Recognition
A R T I C L E S
Figure 3. Structures of the six-membered-ring fused heterocycles containing eight-ring polyamides 1, 3, and 5, and the corresponding parent polyamides
2, 4, and 6. Ball-and-stick models are shown with shaded and nonshaded circles indicating imidazole and pyrrole carboxamides, respectively, whereas
hydroxypyrrole rings are annotated with an “H” in the center of an unshaded sphere. Two touching squares represent the six-membered ring systems attached
to a pyrrole: the shaded square represents the imidazo[4,5-b] unit, the nonshaded square with an “H” in the center the hydroxybenzimidazole unit, and the
nonshaded square represents the benzimidazole unit, the adjacent pyrrole is depicted as a nonshaded circle inside the second square as described above.
(R)-Diaminobutyric acid (DAB (a) is depicted as a curved line and a plus sign, the methyl amide tail is annotated as a triangle.
Figure 4. Illustration and complete sequences of the EcoRI/PVuII restriction fragment derived from plasmid pDR1. The four designed 6-base pair binding
sites that were analyzed in quantitative footprint titrations are shown in bold and surrounded by a box.
3-nitropyridine (7) as starting material (Scheme 1). Nitropyridine
7 was converted to 6-bromo-3-nitropyridin-2-ylamine (8) via
halogen exchange13 in both ortho-positions with HBr in acetic
acid followed by treatment with NH3/MeOH to regioselectively
substitute one of the two bromo substituents.14 The introduction
of the methyl ester functionality was then accomplished via a
palladium-catalyzed carbonylation.15,16 In the presence of carbon
monoxide, MeOH, Et3N, and catalytic amounts of Pd(OAc)2
and PPh3, the conversion of 8 proceeded smoothly and methyl
ester 9 could be isolated in 61% yield. Systematic variation of
the reaction conditions, e.g., CO pressure, temperature and
catalytic system (PdCl2(PPh3)3; Pd(OAc)2, Dppp) did not show
any effect on reaction time or yield. Reduction of the nitro group
in 9 with Pd/C in the presence of hydrogen gave the corre-
sponding ortho-diamine 10 in 63% yield.
The key step for the synthesis of the imidazo[4,5-b]pyridine-
pyrrole unit is a cyclocondensation. The first step, a HBTU
mediated coupling of diamine 10 with carboxylic acid 11
resulted in a mixture of the two isomeric amino amides 12a
and 12b which were then transformed to the desired imidazo-
[4,5-b]pyridine 13 by heating at 80-90 °C in glacial acetic acid.
Note that similar cyclocondensation steps of amino amides
usually require higher temperatures of 120-140 °C.17 However,
compounds 12a and 12b were unstable under these conditions,
and no cyclization product could be isolated. Due to purification
(8) (a) Minehan, T. G.; Gottwald, K.; Dervan, P. B. HelV. Chim. Acta 2000,
83, 2197-2213. (b) Ji, Y.-H.; Bur, D.; Haesler, W.; Schmitt, V. R.; Dorn,
A.; Bailly, C.; Waring, M. J.; Hochstrasser, R.; Leupin, W. Bioorg. Med.
Chem. 2001, 9, 2905-2919. (c) Singh, M. P.; Joseph, T.; Kumar, S.;
Bathini, Y.; Lown, J. W. Chem. Res. Toxicol. 1992, 5, 579-607.
(9) Singh, A. K.; Lown, J. W. Anti-Cancer Drug Design 2000, 15, 265-275.
(10) Bathini, Y.; Rao, K. E.; Shea, R. G.; Lown, J. W. Chem. Res. Toxicol.
1990, 3, 268-280.
(12) (a) Baird, E. E.; Dervan, P. B. J. Am. Chem. Soc. 1996, 118, 6141-6146.
(b) Belitsky, J. M.; Nguyen, D. H.; Wurtz, N. R.; Dervan, P. B. Bioorg.
Med. Chem. 2002, 10, 2767-2774.
(13) Mutterer, F.; Weis, C. D. HelV. Chim. Acta 1976, 59, 229-235.
(14) EP0995 742 AI
(11) Briehn, C. A.; Weyermann, P.; Dervan, P. B. Eur. J. Chem., in press.
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5709

A R T I C L E S
Renneberg and Dervan
Scheme 2. Synthesis of Boc-protected O-methyl
Scheme 1. Synthesis of Boc-protected Imidazo[4,5-b]pyridine
Amino Acid 15
Hydroxybenzimidazole Amino Acid 23
(a) HBr, AcOH; (b) NH3, MeOH; (c) Pd(OAc)2, PPh3, MeOH, Et3N,
DMF, CO; (d) Pd/C, MeOH, H2; (e) HBTU, DIEA, DMF; (f) AcOH; (g)
SnCl2•2H2O, DMF; (h) Boc2O, DIEA; (i) NaOH, Dioxane/H2O.
(a) HNO3; (b) H2SO4, MeOH; (c) Pd/C, Et3N, MeOH, H2; (d) HBTU,
DIEA, DMF; (e) AcOH; (f) Pd/C, DMF, H2; (g) Boc2O, DIEA; (h) NaOH,
Dioxane/H2O.
problems with 13 we decided to use the crude product directly
in the next step. Thus, after evaporation of the glacial acetic
acid, reduction of the nitro group in 13 with tin(II)chloride in
DMF and in situ Boc protection yielded in the aromatic amino
acid ester 14. Final saponification of methyl ester 14 with
sodium hydroxide in dioxane/H2O afforded the desired N-Boc-
protected imidazo[4,5-b]pyridine-pyrrole amino acid (Boc-PyIp-
OH, 15) in 81% yield.
For the synthesis of the N-Boc O-methyl-protected hydroxy-
benzimidazole-pyrrole building block (Boc-PyOz-OH, 23) we
also envisioned a cyclocondensation of the corresponding ortho-
diamine and the carboxylic acid for the construction of the
hydroxybenzimidazole unit. As starting material we chose
4-acetylamino-5-chloro-2-methoxy-benzoic acid methyl ester
(16), which is commercially available (Scheme 2). A regiose-
lective nitration18 17 followed by deacetylation of the amine
under acidic conditions afforded methyl ester 18 in 66% yield
over two steps, starting from 16. Hydrogenation of 18 on 10%
Pd/C in the presence of triethylamine in MeOH18 at r.t. afforded
cleanly ortho-diamine 19. The HBTU mediated coupling of
diamine 19 and carboxylic acid 11 resulted in the two isomeric
amino amides 20a and 20b, which cyclized by heating at 90
°C in glacial acetic acid 21. Reduction of the nitro group in 21
with Pd/C in the presence of hydrogen, in situ Boc-protection
of the free amine to obtain 22 and final hydrolysis of the methyl
ester with sodium hydroxide in dioxane/H2O gave the desired
N-Boc-O-methyl-protected hydroxybenzimidazole-pyrrole (Boc-
PyOz-OH, 23). The synthesis of the N-Boc-protected benzimi-
dazole-pyrrole (Boc-PyBi-OH, 24) will be reported elsewhere.11
Polyamide Syntheses. The parent polyamides 2, 6, and the
O-protected derivative of 4 (namely 32) were synthesized
following manual solid-phase methods12 on Kaiser’s oxime resin
(0.48 mmol/g) in 17 steps. The building blocks employed for
the stepwise oligomer elongation procedure were Boc-N-methyl
pyrrole (Boc-Py-OBt, 25), Boc-N-methyl imidazole (Boc-Im-
OH, 26), Boc-N-methyl methoxypyrrole (Boc-Op-OH, 27)
monomers, N-methyl imidazole (Im-OH, 28) cap and R-Fmoc-
γ-Boc-(R)-2,4-diaminobutyric acid (R-Fmoc-γ-Boc-(R)-DABA,
29). For the synthesis of the imidazo[4,5-b]pyridine, hydroxy-
benzimidazole and benzimidazole containing polyamides 1, 5,
and 31, respectively, the same protocol could be applied
successfully. Compounds 15, 23, and 24 were coupled in NMP/
DIEA for 3 h at r.t. after activation with HBTU. Deprotection
was achieved with 20% TFA/CH2Cl2. Following the general
protocol summarized in Scheme 3, the synthesis of hairpin
polyamides containing the fused heterocycles was carried out
in 15 steps.
Cleavage from the resin and removal of the Fmoc group of
the DABA turn was achieved by treatment with methylamine/
CH2Cl2 (12 h, 37 °C). Following filtration of the resin, the crude
mixtures were purified by reversed-phase HPLC to yield
polyamides ImPyIpPy-(R)H2Nγ-PyPyPyPy-CONHMe (1), Im-
PyImPy-(R)H2Nγ-PyPyPyPy-CONHMe (2), ImPyOzPy-(R)H2Nγ-
PyPyPyPy-CONHMe (31), ImPyOpPy-(R)H2Nγ-PyPyPyPy-
CONHMe (32), ImPyBiPy-(R)H2Nγ-PyPyPyPy-CONHMe (5),
and ImPyPyPy-(R)H2Nγ-PyPyPyPy-CONHMe (6). Subse-
quently, both O-methyl protected polyamides 31 and 32 were
(15) Schoenberg, A.; Bartoletti, I.; Heck, R. F. J. Org. Chem. 1974, 39, 3318-
3326.
(16) Ma¨gerlein, W.; Indolese, A. F.; Beller, M. Angew. Chem., Int. Ed. Engl.
2001, 40, 2856-2859.
(17) Wright, J. B. Chem. ReV. 1951, 48, 396-541.
(18) Tanaka, A.; Ito, K.; Nishino, S.; Motoyama, Y.; Takasugi, H. Chem. Pharm
Bull. 1994, 42, 560.
9
5710 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

DNA Minor Groove Recognition
A R T I C L E S
Scheme 3. Solid-phase Synthesis of Hairpin Polyamides 1, 3, and 5.
Reaction conditions: (a) 25, DIEA, NMP; (b) 20% TFA/CH2Cl2; (c) 25, DIEA, NMP; (d) 20% TFA/CH2Cl2; (e) 25, Diea, NMP; (f) 20% TFA/CH2Cl2;
(g) 25, DIEA, NMP; (h) 20% TFA/CH2Cl2; (i) 29, HBTU, DIEA, NMP; (j) 20% TFA/CH2Cl2; (k) 15 (for 1 where X ) N), 24 (for 5 where X ) CH), and
23 (for 31 where X ) COCH3), DIEA, NMP;(l) 20% TFA/CH2Cl2; (m) 17, HBTU, DIEA, NMP; (n) 20% TFA/CH2Cl2; (o) 28, HBTU, DIEA, NMP; (p)
2.0 M CH3NH2/THF, CH2Cl2; (q) EtSH, NaH, DMF. Inset: Amino Acid Building Blocks for the Polyamide Synthesis. Building Blocks 26 and 27 Are
Required for the Synthesis of Polyamides 2 and 4.
O-demethylated under identical conditions using ethanethiol/
NaH protocol5 (80 °C, 30 min) and purified by reversed-phase
HPLC to provide ImPyHzPy-(R)H2Nγ-PyPyPyPy-CONHMe (3)
and ImPyHpPy-(R)H2Nγ-PyPyPyPy-CONHMe (4).
binds the 5′-TGTGTA-3′ site with an equilibrium association
constant of Ka ) 1.2 × 1010 M-1, a 5-fold higher affinity than
its parent compound 2 (Table 1. Although the A/T base pair
mismatch recognition is similar, the introduction of the Ip/Py
pair is accompanied with a slight decrease in binding specificity
for the C‚G base pair site. This encouraging result supports a
model wherein the pyridine nitrogen of the imidazo[4,5-b]-
pyridine forms a specific H-bond to the exocyclic NH2 of G.
Imidazo[4,5-b]pyridine/Pyrrole Pair. To examine whether
an imidazo[4,5-b]pyridine unit Ip can replace a five-membered
imidazole ring Im within an eight-ring hairpin polyamide
regarding DNA binding affinity and specificity, polyamide 1
and its parent all five-membered ring hairpin 2 were evaluated
by DNase I footprint titrations under identical conditions.
Quantitative DNase I footprint titration experiments (10 mM
Tris-HCl, 10 mM KCl, 10 mM MgCl2, 5 mM CaCl2, pH 7.0,
22 °C, equilibration time: 12 h)19 were performed on a 32P-3′-
labeled EcoRI/PVuII restriction fragment of plasmid pDR1
0(Figure 4). This DNA fragment contains four 6-base pair
binding sites which differ at one position, 5′-TGTCTA-3′, 5′-
TGTTTA-3′, 5′-TGTATA-3′, and 5′-TGTGTA-3′. The latter
one was regarded as the possible match site having a G‚C base
pair under the Ip/Py or Im/Py pair, respectively. Indeed, both
polyamides 1 and 2 exhibit the same binding site preference
(Figure 5). Parent 2 binds its match site 5′-TGTGTA-3′ with
an affinity of Ka ) 2.3 × 109 M-1 and shows a clear preference
versus the C‚G site by nearly a factor of 10. The affinities to
A‚T and T‚A sites are lower, revealing modest specificity (4-
8-fold). The imidazo[4,5-b]pyridine-containing polyamide 1
Hydroxybenzimidazole/Pyrrole Pair. To examine whether
the hydroxybenzimidazole unit Hz can replace a five-membered
hydroxypyrrole ring Hp, polyamide 3 and its parent all five-
membered ring hairpin 4 were compared. The all five-membered
ring polyamide 4 binds preferentially to the site 5′-TGTTTA-
3′ as expected, revealing a preference for the T‚A base pair by
Hp/Py. The binding affinity determined for the T‚A and A‚T
sites were Ka ) 4.1 × 108 M-1, and Ka e 2 × 107 M-1
,
respectively, a specificity of at least 20-fold. No binding was
detected in the concentration range up to 200 nM for the two
G‚C and C‚G sites. Remarkably, analysis of the hydroxyben-
zimidazole-containing polyamide 3 showed that this hairpin
binds with similar affinity and specificity as 4. Polyamide 3
binds the site 5′-TGTTTA-3′ with an affinity of Ka ) 5.7 ×
108 M-1, an 18-fold selectivity of T‚A over A‚T. The sequence
5′-TGTGTA-3′, representing a G‚C mismatch site, was bound
with at least 80-fold lower affinity. No binding was observed
for the C‚G mismatch site. Thus, sequence specificity can be
(19) Trauger, J. W.; Dervan, P. B. Methods Enzymol. 2001, 340, 450-466.
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5711

A R T I C L E S
Renneberg and Dervan
Figure 5. Quantitative DNase I footprint titration experiments on the 3′-
32P-labeled 284-bp EcoRI/PVuII restriction fragment from plasmid pDR1.
Left: polyamide 1: lane 1, intact DNA; lane 2, A-specific reaction; lane
3, G-specific reaction; lane 4, DNase I standard; lane 5-15, 5, 10, 20, 50,
100, 200, 500 pM and 1, 2, 5, 10 nM polyamide, respectively. Right:
polyamide 2: lane 1, intact DNA; lane 2, A-specific reaction; lane 3,
G-specific reaction; lane 4, DNase I standard; lane 5-15, 10, 20, 50, 100,
200, 500 pM and 1, 2, 5, 10, 20 nM polyamide, respectively. The analyzed
6-base pair binding site locations are designated in brackets along the right
side of each autoradiogram with their respective unique base pairs indicated.
Schematic binding models of 1 and 2 with their putative binding sites are
shown on the top side of the autoradiograms. Flanking sequences are
designated in lower case, regular type, while the binding site is given in
capital bold type. The boxed X•Y base pair indicates the position that was
examined in the experiments.
Figure 6. Quantitative DNase I footprint titration experiments on the 3′-
32P-labeled 284-bp EcoRI/PVuII restriction fragment from plasmid pDR1.
Left: polyamide 3: lane 1, intact DNA; lane 2, A-specific reaction; lane
3, G-specific reaction; lane 4, DNase I standard; lane 5-14, 0.2, 0.5, 1, 2,
5, 10, 20, 50, 100, 200 nM polyamide, respectively. Right: polyamide 4:
lane 1, intact DNA; lane 2, A-specific reaction; lane 3, G-specific reaction;
lane 4, DNase I standard; lane 5-15, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 50,
100 nM polyamide, respectively. The analyzed 6-base pair binding site
locations are designated in brackets along the right side of each autorad-
iogram with their respective unique base pairs indicated. Schematic binding
models of 3 and 4 with their putative binding sites are shown on the top
side of the autoradiograms. Flanking sequences are designated in lower
case, regular type, while the binding site is given in capital bold type. The
boxed X•Y base pair indicates the position that was examined in the
experiments.
retained by replacing an Hp with an Hz unit. Presumably the
T‚A selectivity arises in this DNA sequence context in part from
shape selection in the minor groove by the phenolic OH group
against the O(2) of thymine versus the C(2)-H of adenine.
Importantly, the hydroxybenzimidazole-containing polyamide
3 is stable at acidic pH.
Benzimidazole/Pyrrole Pair. To complete this study a third
polyamide containing the benzimidazole building block was
included. This would allow a comparison of the specificity of
Bi/Py pairs with Py/Py pairs and serve as a control as well,
confirming that the different specificity of Hz/Py vs Bi/Py is
due to the single atomic substitution of CH to COH on the
benzene ring (Table 2. For the all five-membered ring polyamide
6 the DNA binding affinity is highest at match sites 5′-
TGTTTA-3′ and 5′-TGTATA-3′ within a factor of 2. The
binding affinity to the G‚C and C‚G sites is lower by a factor
of 23. The same binding site preference was also found for the
benzimidazole containing polyamide 5 revealing that Bi/Py
mimics Py/Py and binds both A‚T and T‚A.
are major contributors. Fused heterocycle subunits, such as
imidazo[4,5-b]pyridine-pyrrole or hydroxybenzimidazole-pyr-
role, are conformationally restricted compared with bis-pyr-
rolecarboxamides leading to entropically stabilized complexes
with DNA. Because Ip and Hz contain one amide linkage less
than their respective distamycin analogues, the resulting higher
hydrophobicity of the curved ligand with a different hydration
behavior could contribute to a deeper position of the polyamide
in the minor groove allowing stronger hydrogen bonding and
van der Waals interactions with the floor and walls of the DNA
groove.
Inspection of the curvature of Ip-, Hz-, and Bi-containing
polyamides as obtained from gas-phase calculations of ImPyXPy
fragments, where X ) Ip, Hz, and Bi compared to their parent
fragments ImPyXPy, where X ) Im, Hp, and Py indicate that
the strands bearing a six-membered fused heterocycle have a
smaller degree of curvature than the corresponding five-
membered ring systems (see Supporting Information Figures
S1, S2, S3). Due to the fact that hairpin polyamides containing
Im, Hp, and Py are slightly over-curved with respect to the
DNA helix, one could assume that the Ip-, Hz-, or Bi-containing
Several factors are considered to play a role in stabilizing
ligand interactions in the minor groove of DNA. Hydrogen
bonding, shape complementarity and electrostatic interactions
9
5712 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

DNA Minor Groove Recognition
A R T I C L E S
our biological gene regulation studies. The DNA binding
properties of polyamides containing multiple Hz/Py pairs, as
well structural studies, will be reported in due course.
Experimental Section
General. 1H NMR and 13C NMR spectra were recorded on a Varian
Mercury 300 and 500 instrument. Chemical shift values were recorded
as parts per million relative to solvent and coupling constants in hertz.
All NMR spectra were measured at room temperature (unless otherwise
stated). UV spectra were measured on a Beckman Coulter DU 7400
diode array spectrophotometer. Mass spectra were recorded on the
following mass spectrometer: matrix-assisted, laser desorption/ioniza-
tion time-of-flight (MALDI-TOF) on Voyager DE-PRO from Applied
Biosystems, fast atom bombardment (FAB) on a JEOL JMS-600H
double focusing high-resolution magnetic sector, and electrospray
injection (ESI) LCQ ion trap on a LCQ classic, Thermofinnigan and
were carried out at the Protein and Peptide Microanalytical Facility at
the California Institute of Technology. Precoated plates silica gel 60F254
(Merck) were used for TLC and silica gel 60 (40 µm) for flash
chromatography. Visualization was realized by UV and/or by using a
solution of Ce(SO4)2, phosphomolybdic acid, H2SO4, and H2O. HPLC
analysis was performed on a Beckman Gold system using a Varian
C18, Microsorb-MV 100-5, 250 × 4.6 mm reversed-phase column in
0.1% (w/v) TFA with acetonitrile as eluent and a flow rate of 1.0 mL/
min, gradient elution 1.25% acetonitrile/min. Preparatory HPLC was
carried out on a Beckman HPLC using a Waters DeltaPak 100 × 25
mm, 100 µm C18 column, 0.1% (w/v) TFA, 0.25% acetonitrile/min.
18 MΩ water was obtained from a Millipore MilliQ water purification
system, and all buffers were 0.2 µm filtered. DNA oligonucleotides
were synthesized by the Biopolymer Synthesis Center at the California
Institute of Technology and used without further purification. Plasmids
were sequenced by the Sequence/Structure Analysis Facility (SAF) at
the California Institute of Technology. DNA manipulations were
performed according to standard protocols. Autoradiography was
performed with a Molecular Dynamics Typhoon Phosphorimager.
ReactionswerecarriedoutunderArinanhydroussolvents.N,N′-Dicyclohexyl-
carbodiimide (DCC), N-hydroxybenzotriazole (HOBt), and 2-(1H-
benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate
(HBTU) were purchased from Peptides International. Oxime resin was
purchased from Novabiochem (0.48 mmol/g). (R)-2-Fmoc-4-Boc-
diaminobutyric acid (R-Fmoc-γ-Boc-(R)-DABA, 29) was from Bachem,
dichloromethane (DCM) was reagent grade from EM, and trifluoroacetic
acid (TFA) was from Halocarbon. Bis(triphenylphosphine)palladium-
(II)dichloride was from Fluka, all other reagents were from Aldrich
(highest quality available). All enzymes (unless otherwise stated) were
purchased from Roche Diagnostics and used with their supplied buffers.
pUC19 was from New England Biolabs. [R-32P]-Deoxyadenosine
triphosphate and [R-32P]-thymine triphosphate was purchased from New
England Nucleotides. RNase-free water (used for all DNA manipula-
tions) was from US Biochemicals. Ethanol (200 proof) was from
Equistar, 2-propanol from Mallinckrodt. Premixed tris-borate-EDTA
(Gel-Mate, used for gel running buffer) was from Gibco. Bromophenol
blue and xylene cyanol FF were from Acros. 3-Methoxy-1-methyl-4-
nitro-1H-pyrrole-2-carboxylic acid (11) was synthesized as reported
earlier.11
Figure 7. Quantitative DNase I footprint titration experiments on the 3′-
32P-labeled 284-bp EcoRI/PVuII restriction fragment from plasmid pDR1.
Left: polyamide 5: lane 1, intact DNA; lane 2, A-specific reaction; lane
3, G-specific reaction; lane 4, DNase I standard; lane 5-15, 5, 10, 20, 50,
100, 200, 500 pM and 1, 2, 5, 10 nM polyamide, respectively. Right:
polyamide 6: lane 1, intact DNA; lane 2, DNase I; lane 3, A-specific
reaction; lane 4, G-specific reaction; lane 5-15, 5, 10, 20, 50, 100, 200,
500 pM and 1, 2, 5, 10 nM polyamide, respectively. The analyzed 6-base
pair binding site locations are designated in brackets along the right side of
each autoradiogram with their respective unique base pairs indicated.
Schematic binding models of 5 and 6 with their putative binding sites are
shown on the top side of the autoradiograms. Flanking sequences are
designated in lower case, regular type, while the binding site is given in
capital bold type. The boxed X•Y base pair indicates the position that was
examined in the experiments.
strand of the hairpin polyamides 1, 3, and 5 have a better shape
complementarity with the minor groove of DNA. Indeed, the
increased DNA binding affinity of the polyamides comprising
Ip, Hz, or Bi units might be a result of the decreased curvature
and therefore deeper position into the minor groove. Neverthe-
less, in this first exploratory study only one strand of the hairpin
was substituted, whereas the opposite one remained five-
membered rings. On the basis of the data presented here, it will
be interesting to examine whether antiparallel unsymmetrical
pairs of the benzimidazole analogues, Ip/Bi and Hz/Bi, mimic
Im/Py and Hp/Py, and code for G‚C and T‚A base pairs,
respectively.
Monomer Syntheses. 2-Amino-3-nitro-6-bromo-pyridine (8) was
synthesized in 2 steps following described procedures.13,14 Data of 8:
Rf 0.30 (hexane/EtOAc 1:1).1H NMR (300 MHz, DMSO-d6): δ 6.87
(d, J ) 8.2 Hz, 1H), 8.23 (d, J ) 8.2 Hz, 1H), 8.25 (brs, 2H).
6-Amino-5-nitro-pyridine-2-carboxylic Acid Methyl Ester (9). To
a solution of bromide 8 (8.0 g, 36.6 mmol) in DMF (130 mL) were
added methanol (130 mL), Pd(OAc)2 (248 mg, 1.1 mmol), triph-
enylphosphine (316 mg, 1.2 mmol) and Et3N (20 mL). The mixture
was stirred for 10 min and then transferred into a parr apparatus. After
evacuating 3 times with CO, the reaction mixture was heated under a
CO pressure of 15 atm to 60 °C for 15 h. The red/brown solution was
Conclusion
Six-membered ring heterocycles Ip, Hz, and Bi, paired
opposite the five-membered ring Py, bind the minor groove of
DNA and distinguish the edges of the four Watson-Crick base
pairs. The Ip/Py pair mimics an Im/Py and exhibits the same
preference for G‚C. The Bi/Py pair mimics Py/Py and exhibits
the same preferences for A/T vs G/C base pairs. Most
importantly, the Hz/Py pair distinguishes a T‚A base pair from
an A‚T base pair and is a strong candidate to replace Hp/Py in
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5713

A R T I C L E S
Renneberg and Dervan
Table 1. Equilibrium Association Constants Ka [M-1] for Polyamides 1, 2, 3, 4, 5, and 6
a The reported association constants Ka are the average values obtained from three DNase I footprint titration experiments, with the standard deviation for
each data set indicated in parentheses. b The assays were carried out at 22 °C at pH 7.0 in the presence of 10 mM Tris-HCl, 10 mM KCl, 10 mM MgCl2,
and 5 mM CaCl2 with an equilibration time of 12 h. c Specificities are given in brackets under the Ka values and calculated as Ka(match)/Ka(mismatch).
Table 2. Specificity of Bi, Hz, and Ip Pairingsa
(2.5 mL) were added and the mixture was stirred under heating to 40
°C for 2 days. The reaction mixture was poured into ice and the
precipitate formed was collected by filtration. Drying under hv provided
pair
T‚A
A‚T
G‚C
C‚G
Bi/Py
Hz/Py
Ip/Py
+
+
-
+
-
-
-
-
+
-
-
-
2.8 g (82%) as a mixture of both isomers 12a and 12b in a ratio of 3:2
1
(as determined by H NMR) as a red/brown solid which were used
without separation. Data of 12a and 12b: Rf 0.55, 0.65 (hexane/EtOAc
1
a Favored (+), disfavored (-).
1:5). H NMR (300 MHz, DMSO-d6): δ 3.76, 3.79, 3.91, 3.92 (4s,
6H), 5.99 (brs, 0.6H), 6.35 (brs, 0.4H), 7.10 (d, J ) 8.2 Hz, 0.6H),
7.30 (d, J ) 8.2 Hz, 0.4H), 7.72-7.77 (m, 4H), 8.21 (s, 2H), 9.59
(brs, 0.4H), 10.33 (brs, 0.6H). 13C NMR (75 MHz, DMSO-d6): δ 38.27,
38.45, 52.32, 52.70, 109.73, 109.90, 114.54, 119.42, 121.69, 121.95,
126.18, 126.62, 129.04, 129.13, 132.86, 133.59, 134.46, 136.23, 142.80,
144.78, 159.54, 162.87, 165.43, 165.83. MS (ESI): m/z (rel intensity)
320 (65). HRMS (FAB): m/z calcd for C13H14N5O5, 320.0987; found,
320.0994.
then cooled to r.t., filtered through a pad of Celite, rinsed with EtOAc
and concentrated. Purification by flash chromatography (hexane/EtOAc
1:1) afforded 4.4 g (61%) of 9 as a bright yellow solid. Data of 9: Rf
0.57 (hexane/EtOAc 1:1). 1H NMR (300 MHz, DMSO-d6): δ 3.85 (s,
3H), 7.27 (d, J ) 8.8 Hz, 1H), 8.09 (brs, 2H), 8.52 (d, J ) 8.8 Hz,
1H). 13C NMR (75 MHz, DMSO-d6): δ 53.52, 113.10, 129.43, 137.48,
152.41, 153.79, 164.53. MS (ESI): m/z (rel intensity) 198 (100).
5,6-Diamino-pyridine-2-carboxylic Acid Methyl Ester (10). Meth-
yl ester 9 (3.5 g, 18.0 mmol) was dissolved in methanol (300 mL) and
EtOAc (300 mL) and the solution was degassed with Ar. After the
addition of Pd/C (10 wt %, 0.7 g) the reaction mixture was stirred for
4 h under a hydrogen atmosphere. Filtration through a pad of Celite
with EtOAc, evaporation of the solvent and purification by flash
chromatography (EtOAc/MeOH 15:1) yielded 2.4 g (79%) of 10 as a
deep red powder. Data of 10: Rf 0.42 (EtOAc/MeOH 15:1). 1H NMR
(300 MHz, DMSO-d6): δ 3.68 (s, 3H), 5.51 (brs, 2H), 5.78 (brs, 2H),
6.67 (d, J1 ) 7.7 Hz, 1H), 7.21 (d, J1 ) 7.7 Hz, 1H). 13C NMR (75
MHz, DMSO-d6): δ 51.88, 116.32, 117.98, 132.40, 135.09, 147.93,
166.30. MS (ESI): m/z (rel intensity) 168 (100).
2-(4-tert-Butoxycarbonylamino-1-methyl-1H-pyrrole-2-yl)-3H-
imidazo[4,5-b]pyridine-5-carboxylic Acid Methyl Ester (14). The
aforementioned mixture of amino amides 12a and 12b (0.80 g, 2.4
mmol) was suspended in glacial acetic acid (20 mL) and heated to 80
°C for 7 h. The redish/grey suspension was then concentrated and dried
under hv to obtain crude 13 as a brown solid in a ca. 1:1 mixture with
1
unreacted amino amide 12a (as determined with H NMR). Data of
1
13: Rf 0.62 (hexane/EtOAc 1:5). H NMR (300 MHz, DMSO-d6): δ
3.86 (s, 3H), 3.92 (s, 3H), 7.40 (d, J ) 8.2, 1H), 7.98 (d, J ) 2.2, 1H),
8.26 (d, J ) 1.6, 1H), 8.41 (d, J ) 8.2, 1H), 10.26 (s, 1H). MS (ESI):
m/z (rel intensity) 302 (13).
To a solution of unpurified 13 (0.75 g, 2.5 mmol) in DMF (13 mL)
was added SnCl2‚2H2O (3.4 g, 15.0 mmol) and the mixture was stirred
under heating to 50 °C for 36 h. After completion of the reaction
detected by TLC, Boc2O (830 mg, 3.7 mmol) and DIEA (0.4 mL) were
added to the reaction mixture. After stirring for 20 h another portion
of Boc2O (280 mg, 1.25 mmol) was added and stirring was continued
for an additional 18 h. The mixture was concentrated, the residue
redissolved in EtOAc, and washed with brine. The organic phase was
5-Amino-6-[(1-methyl-4-nitro-1H-pyrrole-2-carbonyl)-amino]-py-
ridine-2-carboxylic acid methyl ester and 6-amino-5-[(1-methyl-4-
nitro-1H-pyrrole-2-carbonyl)-amino]-pyridine-2-carboxylic acid meth-
yl ester (12a/12b). Carboxylic acid 11 (2.0 g, 11.8 mmol) and diamine
10 (1.8 g, 10.8 mmol) were dissolved in DMF (26 mL). 2-(1H-
benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate
(HBTU) (4.4 g, 11.2 mmol) and N,N-diisopropylethylamine (DIEA)
9
5714 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003

DNA Minor Groove Recognition
A R T I C L E S
dried over MgSO4 and the solvent evaporated. Flash chromatography
(hexane/EtOAc 1:4) afforded 130 mg (14%, 3 steps, based on 12a/
12b) of 14 as a yellow powder. Data of 14: Rf 0.45 (hexane/EtOAc
1:4). Compound 14 appears as two tautomers (1H/3H) on the NMR
time scale at r.t. in a ratio of about 1:1;21 heating the NMR probe to 70
°C results in only one set of signals. UV (MeOH): λmax 364. 1H NMR
(300 MHz, DMSO-d6, 70 °C): δ 1.43 (s, 9H), 3.85 (s, 3H), 4.04 (s,
3H), 6.94 (brs, 2H, D2O exchange), 7.08 (brs, 1H), 7.88-7.97 (m, 3H),
9.24 (s, 2H, D2O exchange). 13C NMR (75 MHz, DMSO-d6): δ 28.85,
31.95, 37.42, 52.81, 79.07, 86.25, 104.53, 118.52, 119.90, 125.39,
146.79, 151.23, 153.37. MS (ESI): m/z (rel intensity) 372 (100). HRMS
(FAB): m/z calcd for C18H22N5O4, 372.1671; found, 372.1660.
DMSO-d6): δ 51.77, 60.53, 109.36, 111.27, 121.35, 127.78, 141.95,
147.61. MS (ESI): m/z (rel intensity) 197 (55).
4-Amino-2-methoxy-3-[(1-methyl-4-nitro-1H-pyrrole-2-carbonyl)-
amino]-benzoic Acid Methyl Ester (20a/20b). Carboxylic acid 11 (3.0
g, 17.6 mmol) and the aforementioned crude diamine 19 were dissolved
in DMF (50 mL). HBTU (6.5 g, 16.5 mmol) and DIEA (3.4 mL) were
added, and the reaction mixture was stirred for 1.5 days. The mixture
was poured into ice and the precipitate collected by filtration. After
washing with cold water and drying under hv provided 4.4 g (68%) as
a mixture of the two isomeric amino amides 20a/20b in a ratio of about
1:1 (as determined by 1H NMR) as a beige solid. Data of 20a and 20b:
Rf 0.58, 0.74 (hexane/EtOAc 1:4). 1H NMR (300 MHz, DMSO-d6): δ
3.61, 3.69, 3.75, 3.90 (4s, 9H), 5.04 (brs, 1H), 5.89 (brs, 1H), 6.47 (d,
J ) 8.1 Hz, 0.5H), 6.94 (d, J ) 8.2 Hz, 0.5H), 7.09 (d, J ) 8.2 Hz,
0.5H), 7.49 (d, J ) 8.1 Hz, 0.5H), 7.63-7.69 (m, 1H), 8.17-8.20 (m,
1H), 9.30 (brs, 0.5H), 9.65 (brs, 0.5H). 13C NMR (75 MHz, DMSO-
d6): δ 38.18, 38.30, 51.98, 52.67, 61.23, 62.01, 108.98, 109.54, 110.07,
110.38, 115.48, 117.35, 121.92, 122.19, 127.19, 128.57, 131.83, 134.44,
147.52, 152.26, 159.57, 160.08, 165.67, 166.49. MS (ESI): m/z (rel
intensity) 349 (100). HRMS (FAB): m/z calcd for C15H16N4O6,
348.1069; found, 348.1085.
2-(4-tert-Butoxycarbonylamino-1-methyl-1H-pyrrole-2-yl)-3H-
imidazo[4,5-b]pyridine-5-carboxylic acid (15). Compound 14 (60 mg,
0.6 mmol) was dissolved in 0.2 M NaOH in H2O/dioxane 1:1 (7 mL)
and stirred for 5 h at r.t. while the solution turned orange. The solution
was cooled to 0 °C and 1N HCl was added dropwise until pH 1-2,
while an orange precipitate developed. The product was separated by
centrifugation, the supernatant removed, and the formed precipitate
washed with cold H2O (2 × 2 mL). Lyophilization afforded 46 mg
(81%) of 15 as an orange solid. Data of 15: UV (MeOH): λmax 359.
1H NMR (300 MHz, DMSO-d6): δ 1.46 (s, 9H), 4.05 (s, 3H), 7.01
(brs, 1H, D2O exchange), 7.15 (s, 1H), 7.98 (m, 3H), 9.30 (s, 1H, D2O
exchange). 13C NMR (75 MHz, DMSO-d6): δ 28.96, 37.47, 79.21,
106.38, 118.65, 119.93, 120.44, 122.32, 124.99, 141.58, 150.32, 153.38,
166.53. MS (ESI): m/z (rel intensity) 358 (100). HRMS (FAB): m/z
calcd for C17H20N5O4, 358.1515; found, 358.1506.
4-Methoxy-2-(1-methyl-4-nitro-1H-pyrrol-2-yl)-3H-benzimidazole-
5-carboxylic Acid Methyl Ester (21). The aforementioned mixture
of amides 20a/20b (500 mg, 1.43 mmol) was suspended in glacial acetic
acid (8 mL) and heated to 90 °C for 7 h. Most of the solvent was
evaporated, and the yellow residue was suspended in Et2O. After being
vigorously stirred for 20 min, the precipitate was filtered and washed
with Et2O. Drying under hv gave 404 mg (85%) of 21 as a beige solid.
Data of 21: Rf 0.80 (hexane/EtOAc 1:5). 1H NMR (300 MHz, DMSO-
d6): δ 3.76 (s, 3H), 4.14 (s, 3H), 4.34 (s, 3H), 7.14-7.17 (m, 1H),
7.48-7.55 (m, 3H), 8.25 (s, 1H). MS (ESI): m/z (rel intensity) 331
(98).
4-Acetylamino-5-chloro-2-methoxy-3-nitro-benzoic Acid Methyl
Ester (17). The title compound was synthesized according to a literature
procedure applying some modifications.18 Fuming nitric acid (33 mL)
was added dropwise at 20 °C to compound 16 (13.0 g, 0.05 mol) over
15 min. After stirring for an additional 10 min at that temperature, the
mixture was poured into ice water and EtOAc (100 mL) was added.
After separating the organic phase, the aqueous phase was extracted
with EtOAc. The combined organic phases were washed with brine,
dried over MgSO4 and concentrated. Drying under hv yielded 14.0 g
(91%) of 17 as a beige solid. 1H NMR (300 MHz, DMSO-d6): δ 2.01
(s, 3H), 3.84 (s, 3H), 3.87 (s, 3H), 8.12 (s, 1H), 10.30 (s, 1H).
2-(4-tert-Butoxycarbonylamino-1-methyl-1H-pyrrol-2-yl)-4-meth-
oxy-3H-benzimidazole-5-carboxylic Acid Methyl Ester (22). Com-
pound 21 (400 mg, 1.21 mmol) was dissolved in DMF (10 mL) and
Pd/C (10 wt %, 150 mg) was added. The mixture was transferred into
a parr apparatus and stirred for 5 h under a hydrogen atmosphere of 10
atm. After the TLC revealed no starting material, the mixture was
transferred into a round-bottom flask and Boc2O (400 mg, 1.81 mmol)
and DIEA (0.5 mL) were directly added. The reaction mixture was
stirred overnight, filtered through a pad of Celite, rinsed with EtOAc,
and concentrated in vacuo. Purification by flash chromatography
(hexane/EtOAc 1:2) afforded 248 mg (51%) of 22 as a beige solid.
Data of 22: Rf 0.34 (hexane/EtOAc 1:2). 1H NMR (300 MHz, DMSO-
d6): δ 1.43 (s, 9H), 3.75 (s, 3H), 3.99 (s, 3H), 4.31 (s, 3H), 6.78 (s,
2H, CH, D2O exchange), 6.95 (s, 1H), 7.05 (d, J ) 8.2 Hz, 1H), 7.45
(d, J ) 8.6 Hz, 1H), 9.18 (brs, 1H, D2O exchange). 13C NMR (75
MHz, DMSO-d6): δ 28.62, 38.20, 51.98, 62.01, 108.98, 110.08, 115.48,
127.19, 128.57, 131.84, 152.25, 160.09, 165.67. UV (MeOH): λmax
327. MS (ESI): m/z (rel intensity) 401 (100).
4-Amino-5-chloro-2-methoxy-3-nitro-benzoic Acid Methyl Ester
(18). Compound 17 (6.2 g, 20.5 mmol) was dissolved in methanol (70
mL), H2SO4 concentrated (4 mL) was added and the solution heated
to reflux for 9 h. After cooling to r.t., the mixture was poured into ice
(600 mL) and the precipitate formed was collected by filtration. Drying
under hv gave 5.0 g (94%) of 18 as a yellow solid. Data of 18: Rf 0.85
(hexane/EtOAc 1:4). 1H NMR (300 MHz, DMSO-d6): δ 3.77 (s, 6H),
6.85 (brs, 2H), 7.82 (s, 1H). 13C NMR (75 MHz, DMSO-d6): δ 52.80,
64.68, 110.40, 114.21, 133.80, 142.29, 153.60, 163.48. MS (ESI): m/z
(rel intensity) 261 (100). HRMS (FAB): m/z calcd for C9H9N2O5Cl,
260.0200; found, 260.0197.
3,4-Diamino-2-methoxy-benzoic Acid Methyl Ester (19). Amine
18 (4.8 g, 18.4 mmol) was dissolved in methanol (250 mL) and the
resulting solution was intensively degassed for 10 min with Ar. Pd/C
(10 wt %, 4.5 g) was added, followed by Et3N (30 mL) and the resulting
mixture was transferred to a parr apparatus. After being stirred for 6 h
under a hydrogen atmosphere at 5 atm, the reaction mixture was filtered
through a pad of Celite and rinsed with methanol. Evaporation of the
solvent yielded crude diamine 19 as a brown gum (with residual Et3N).
Data for 19: Rf 0.43 (hexane/EtOAc 1:4). 1H NMR (300 MHz, DMSO-
d6): δ 3.62 (s, 3H), 3.67 (s, 3H), 4.38 (brs, 2H), 5.38 (brs, 2H), 6.30
(d, J ) 8.2 Hz, 1H), 6.96 (d, J ) 8.2 Hz, 1H). 13C NMR (75 MHz,
2-(4-tert-Butoxycarbonylamino-1-methyl-1H-pyrrol-2-yl)-4-meth-
oxy-3H-benzoimidazole-5-carboxylic Acid 23. Methyl ester 22 (51
mg, 0.13 mmol) was dissolved in 1 M NaOH (1:1 dioxane/H2O, 5 mL)
and stirred overnight at r.t and an additional 3 h at 50 °C. The reaction
mixture was cooled to 0 °C and 1N HCl was added dropwise to adjust
the pH to 3-4, while a white precipitate was formed. Separation of
the product was accomplished by centrifugation and decanting of the
supernatant. Lyophilization gave 36 mg (73%) of 23 as a beige solid.
Data of 23: 1H NMR (500 MHz, DMSO-d6): δ 1.40 (s, 9H), 3.97 (s,
3H), 4.28 (s, 3H), 6.76 (s, 1H), 6.92 (s, 1H), 7.02 (d, J ) 8.2 Hz, 1H),
7.44 (d, J ) 8.2 Hz, 1H), 9.13 (s, 1H), 12.20 (brs, 1H). 13C NMR (125
MHz, DMSO-d6): δ 27.80, 36.34, 60.78, 78.29, 102.88, 104.47, 114.67,
116.12, 119.79, 123.54, 124.96, 135.00, 139.02, 145.90, 150.62, 152.78,
167.70. UV (H2O): λmax 319. MS (ESI): m/z (rel intensity) 387 (100).
HRMS (FAB): calcd for C19H23N4O5, 387.1668; found, 387.1658.
(20) Besides the designed four binding sites there are several overlapping 1-base
pair and 2-base pair mismatch sites present on the plasmid to which
polyamides 3 and 4 show comparable binding behavior.
(21) Troschu¨tz, R.; Lu¨ckel, A. Arch. Pharm. 1992, 325, 617-619.
(22) (a) Iverson, B. L.; Dervan, P. B. Nucl. Acids Res. 1987, 15, 7823-7830.
(b) Maxam, A. M.; Gilbert, W. S. Methods Enzymol. 1980, 65, 499-560.
9
J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003 5715

A R T I C L E S
Renneberg and Dervan
Polyamide Syntheses. Polyamides 2, 4, 6 were synthesized in a
stepwise fashion from 0.48 mmol/g oxime resin by manual solid-phase
methods using Boc-protected amino acid building blocks as reported
earlier.12 The incorporation of the imidazo[4,5-b]pyridine, hydroxy-
benzimidazole and benzimidazole Boc-protected building blocks, to
synthesize polyamides 1, 3, and 5, was realized under identical
conditions. Activation of 15, 23, 24 was achieved with HBTU (1 eq.
in NMP, DIEA) for 10 min and coupling was realized for 3 h at r.t.
Deprotection of the immobilized polyamides-containing building blocks
was successful using 20% TFA in CH2Cl2 for 25 min.
dissolved in DMF (0.25 mL), added to the ethanethiolate solution and
the mixture was heated to 80 °C for 30 min in a sealed tube under
periodic agitation. After cooling to 0 °C, glacial acetic acid (3 mL)
was added and all volatiles removed in vacuo. The residue was dissolved
in CH3CN (0.5 mL) and 0.1% TFA (2 mL) and purified by preparative
reversed-phase HPLC to yield the pure polyamides.
ImPyHzPy-(R)H2Nγ-PyPyPyPy-CONHMe (3). The title compound
was recovered upon lyophilization of the appropriate fractions as a
yellow powder (0.4 mg). UV (H2O): λmax 318, 254. MALDI-TOF-
MS: calcd. for C53H58N19O9, 1105.1; found, 1105.1.
Extinction coefficients were calculated based on ꢀ ) 8690 cm L
mol-1 per ring at 310 nm for the polyamides 2, 4, and 6.19 The extinction
coefficient for polyamides 1, 3, and 5 was determined as 28 500 cm L
mol-1 (at 330 nm), 33 914 cm L mol-1 (at 318 nm), and 40 081 cm L
mol-1 (at 326 nm), respectively.
Procedure for MeNH2 Cleavage from the Resin.12b After comple-
tion of the synthesis of the polyamides the resin was filtered off the
reaction and washed with DMF, DCM, MeOH, Et2O, and dried in
vacuo. A sample of the resin (150 mg) was swollen in DCM (3 mL)
and treated with MeNH2 (2M in THF, 3 mL) under periodic agitation
at 37 °C for 16 h. The resin was then removed by filtration, washed
with DCM and DMF, and concentrated in vacuo. After diluting with
20% MeCN/0.1% TFA the crude mixture was purified by preparative
reversed-phase HPLC to yield the pure polyamides.
ImPyHpPy-(R)H2Nγ-PyPyPyPy-CONHMe (4). The title compound
was recovered upon lyophilization of the appropriate fractions as a white
powder (0.1 mg). UV (H2O): λmax 316, 244. MALDI-TOF-MS: calcd.
for C52H60N19O10, 1111.1; found, 1111.7, 1133.7 (M+Na).
Construction of plasmid DNA. The plasmid pDR1 was constructed
by hybridization of the two oligonucleotides 5′-GATCAGGCTGTC-
TATGGACGGCTGCTGTGTATGGACG-GCTGCTGTTTATGGACG-
GCTGCTGTATATGGACGG-CTGGGCGA-3′ and 5′-AGCTTCGC-
CCAGCCGTCCATATACAGCAGCCGTCCATAAACAGCA-
GCCGTCCATACACAGCAGCCGTCCATAGACAGCCT-3′ followed
by ligation into the BamHI/HinDIII restriction fragment of pUC19 using
T4 DNA ligase. The resultant plasmid was then used to transform E.
Coli XL-1 Blue Supercompetent cells. Ampicillin-resistant white
colonies were selected from 25 mL Luria-Bertani agar plates containing
50 mg/mL ampicillin, treated with XGAL and IPTG solutions and
grown overnight at 37 °C. Cells were harvested after overnight growth
at 37 °C. Large-scale plasmid purification was performed with Qiagen
purification kits. The presence of the desired insert was determined by
dideoxy sequencing. DNA concentrations were determined at 260 nm
using the relationship 1 OD unit ) 50 µg/mL duplex DNA.
Preparation of 3′-End-Labeled DNA Restriction Fragments. The
plasmid pDR1 was linearized with EcoRI and PVuII and treated with
the Klenow fragment of DNA polymerase II, [R-32P]-dATP, and [R-32P]-
TTP for 3′-end labeling. The labeled 3′-fragment was purified on a
7% nondenaturing polyacrylamide gel (5% cross-linkage) and the
desired 284 base pair band isolated after visualization by autoradiog-
raphy. The DNA was precipitated with 2-propanol, the pellet washed,
lyophilized, and resuspended in RNase-free H2O. Chemical sequencing
reactions were performed according to published protocols.21
ImPyIpPy-(R)H2Nγ-PyPyPyPy-CONHMe (1). Starting with 100 mg
of resin the title compound was synthesized. It was recovered upon
lyophilization of the appropriate fractions as an orange powder (0.4
mg). UV (H2O): λmax 375, 330, 248. MALDI-TOF-MS: calcd. for
C52H57N20O8, 1090.1; found, 1090.1, 1110.8 (M+Na).
ImPyImPy-(R)H2Nγ-PyPyPyPy-CONHMe (2). Starting with 150
mg of resin the title compound was synthesized. It was recovered upon
lyophilization of the appropriate fractions as a white powder (0.5 mg).
UV (H2O): λmax 314, 246. MALDI-TOF-MS: calcd. for C51H59N20O9,
1096.1; found, 1096.1, 1117.1 (M+Na).
ImPyBiPy-(R)H2Nγ-PyPyPyPy-CONHMe (5). Starting with 150 mg
of resin the title compound was synthesized. It was recovered upon
lyophilization of the appropriate fractions as a beige powder (0.6 mg).
UV (H2O): λmax 326, 246. MALDI-TOF-MS: calcd. for C53H58N19O8,
1089.1; found, 1089.1.
ImPyPyPy-(R)H2Nγ-PyPyPyPy-CONHMe (6). Starting with 150
mg of resin the title compound was synthesized. It was recovered upon
lyophilization of the appropriate fractions as a white powder (0.7 mg).
UV (H2O): λmax 316, 246. MALDI-TOF-MS: calcd. for C52H59N19O9,
1094.1; found, 1094.2.
DNase I Footprinting. All reactions were carried out in a volume
of 400 µL according to published procedures. Quantitation by storage
phosphor autoradiography and determination of equilibrium association
constants were as previously described.19
ImPyOzPy-(R)H2Nγ-PyPyPyPy-CONHMe (31). Starting with 200
mg of resin the title compound was synthesized. It was recovered upon
lyophilization of the appropriate fractions as a white powder (1.6 mg).
UV (H2O): λmax 320, 246. MALDI-TOF-MS: calcd. for C54H60N19O9,
1119.1; found, 1119.2, 1142.17 (M+Na).
Acknowledgment. We are grateful to the National Institute
of Health (GM27681) for research support and the Swiss
National Science Foundation for a postdoctoral fellowship to
D.R. We thank M. A. Marques for guidance with the ab initio
calculations.
ImPyOpPy-(R)H2Nγ-PyPyPyPy-CONHMe (32). Starting with 200
mg of resin the title compound was synthesized. It was recovered upon
lyophilization of the appropriate fractions as a white powder (0.8 mg).
UV (H2O): λmax 318, 245. MALDI-TOF-MS: calcd. for C53H62N19O10,
1125.1, found, 1125.6.
Supporting Information Available: Ab initio calculations for
four-ring polyamide subunits PyImPyIm and PyIpPyIm, PyH-
pPyIm and PyHzPyIm, and PyPyPyIm and PyBiPyIm (Figures
S1, S2, S3). This material is available free of charge via the
Internet at http://pubs.acs.org.
Deprotection of the O-Methyl-Protected Polyamides. Following
a previously reported procedure,7 to a slurry of NaH (80 mg, 60% in
mineral oil) in 0.5 mL DMF was added ethanethiol (0.32 mL), and the
mixture was heated to 80 °C for 5 min. The polyamide (0.4 µmol) was
JA0300158
9
5716 J. AM. CHEM. SOC. VOL. 125, NO. 19, 2003